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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
a2fdf4ab-a559-4320-b741-cf62ae293f42 | sequential-stochastic-optimization-in | 2108.09585 | null | https://arxiv.org/abs/2108.09585v1 | https://arxiv.org/pdf/2108.09585v1.pdf | Sequential Stochastic Optimization in Separable Learning Environments | We consider a class of sequential decision-making problems under uncertainty that can encompass various types of supervised learning concepts. These problems have a completely observed state process and a partially observed modulation process, where the state process is affected by the modulation process only through an observation process, the observation process only observes the modulation process, and the modulation process is exogenous to control. We model this broad class of problems as a partially observed Markov decision process (POMDP). The belief function for the modulation process is control invariant, thus separating the estimation of the modulation process from the control of the state process. We call this specially structured POMDP the separable POMDP, or SEP-POMDP, and show it (i) can serve as a model for a broad class of application areas, e.g., inventory control, finance, healthcare systems, (ii) inherits value function and optimal policy structure from a set of completely observed MDPs, (iii) can serve as a bridge between classical models of sequential decision making under uncertainty having fully specified model artifacts and such models that are not fully specified and require the use of predictive methods from statistics and machine learning, and (iv) allows for specialized approximate solution procedures. | ['Chelsea C. White III', 'R. Reid Bishop'] | 2021-08-21 | null | null | null | null | ['decision-making-under-uncertainty', 'decision-making-under-uncertainty'] | ['medical', 'reasoning'] | [ 4.42437440e-01 6.08003259e-01 -5.59049964e-01 -8.20314810e-02
-4.30539787e-01 -5.79896927e-01 8.77201796e-01 5.11398196e-01
-5.56276679e-01 1.12898135e+00 1.42339040e-02 -4.11626756e-01
-5.19636452e-01 -7.23818660e-01 -6.11300170e-01 -1.08563185e+00
-4.13383245e-01 1.26083565e+00 1.40835270e-01 1.27518937e-01
1.09298140e-01 4.07568395e-01 -1.12493491e+00 -1.36118442e-01
4.68562692e-01 1.28759515e+00 3.35752010e-01 6.72136486e-01
-7.74305984e-02 1.06013465e+00 -4.05781358e-01 1.29974589e-01
3.47461522e-01 -4.08050209e-01 -7.74482787e-01 4.69976932e-01
-9.53317225e-01 -1.95028171e-01 -1.84932202e-01 1.14464021e+00
7.80349970e-02 3.09391111e-01 8.83567095e-01 -1.35342860e+00
-1.94889847e-02 4.70425576e-01 -1.90521151e-01 5.52320220e-02
3.04028034e-01 1.64645240e-01 7.92809606e-01 2.20567763e-01
5.73615432e-01 1.57549846e+00 -1.11302294e-01 5.30977607e-01
-1.43848062e+00 -1.93228930e-01 4.47282404e-01 1.32683590e-01
-9.37399864e-01 -1.59997821e-01 3.31236094e-01 -3.32176805e-01
4.65723038e-01 1.27565295e-01 7.30178654e-01 8.88245881e-01
9.74017739e-01 1.10005927e+00 1.62266016e+00 -4.29494470e-01
1.20758700e+00 -1.72977492e-01 4.84376214e-02 3.11367214e-01
3.32107216e-01 5.37835121e-01 3.12531516e-02 -4.18764383e-01
6.93805337e-01 3.96901459e-01 -1.34923324e-01 -6.51225626e-01
-1.06317234e+00 9.28743005e-01 -4.84918982e-01 -2.41888389e-01
-8.05592716e-01 1.28830194e-01 1.64446324e-01 7.25510359e-01
-4.24686959e-03 3.25168312e-01 -5.14945030e-01 -7.40587190e-02
-5.23497045e-01 5.58924854e-01 1.27784526e+00 8.93023431e-01
5.00720024e-01 -1.15894660e-01 -4.00740832e-01 -1.23969831e-01
4.19483483e-01 7.98393965e-01 2.31651276e-01 -1.35818648e+00
5.19222200e-01 -2.68569384e-02 1.03022802e+00 -5.15774451e-02
-5.25293887e-01 1.48045141e-02 -6.35170281e-01 4.51552719e-01
5.52991152e-01 -3.65803301e-01 -1.03553081e+00 1.78508651e+00
3.47133696e-01 -2.74060667e-01 1.40681669e-01 5.29096782e-01
-3.26799303e-01 8.94590020e-01 -1.64093673e-01 -9.40571070e-01
1.20266187e+00 -4.23333883e-01 -1.12263584e+00 -1.76683009e-01
1.29432485e-01 -2.71060884e-01 2.29649827e-01 8.84548903e-01
-1.36751902e+00 -4.29596826e-02 -1.17306733e+00 6.66066408e-01
4.13928591e-02 -6.19618654e-01 4.67997491e-01 3.28946710e-01
-8.48600864e-01 8.61194253e-01 -1.50574255e+00 -1.40433267e-01
9.53494981e-02 5.83261788e-01 -6.09664693e-02 -3.15073013e-01
-1.13532519e+00 1.15822196e+00 6.90886080e-01 -5.16603328e-02
-1.57575595e+00 3.00130337e-01 -7.89682925e-01 1.32656112e-01
1.04399645e+00 -6.14364803e-01 1.69567907e+00 -5.26567519e-01
-1.67359591e+00 1.60488263e-01 -5.56655712e-02 -5.61738074e-01
7.77078390e-01 5.64561933e-02 -3.88696283e-01 2.54091740e-01
1.25230715e-01 1.49225757e-01 1.27410138e+00 -1.12920284e+00
-8.62947226e-01 -5.63308001e-01 8.83810669e-02 4.25935417e-01
6.94478631e-01 -1.35048002e-01 2.02999383e-01 -2.11678684e-01
4.12444502e-01 -1.30248868e+00 -7.46937335e-01 -4.04332757e-01
-3.96927088e-01 7.54306167e-02 7.23010302e-01 -3.15548033e-01
8.64471436e-01 -1.64410782e+00 2.94178903e-01 1.90282285e-01
-1.66172802e-01 -3.93015802e-01 2.36281306e-01 8.44576240e-01
7.15182722e-03 -5.93416952e-02 -4.71565396e-01 -3.46352041e-01
1.99524552e-01 7.50280559e-01 -4.33751285e-01 8.65903854e-01
1.45207942e-01 5.73185205e-01 -8.61025929e-01 -2.27929860e-01
2.83913851e-01 -3.84318292e-01 -1.65872484e-01 2.23585457e-01
-7.39046931e-01 9.00307059e-01 -9.69176292e-01 4.76437718e-01
3.21249753e-01 -6.27186000e-02 4.95126367e-01 7.60962367e-01
-1.93723544e-01 1.56816676e-01 -1.77111435e+00 1.16766381e+00
-1.76626042e-01 -1.90885320e-01 3.98045719e-01 -1.07765877e+00
5.24155498e-01 6.58615470e-01 6.38597548e-01 -2.25330174e-01
3.25740457e-01 -1.47983536e-01 8.56401920e-02 -3.57838333e-01
2.62543827e-01 -7.24382699e-01 -4.38029796e-01 8.33975852e-01
7.98567757e-02 -3.86428148e-01 2.94489805e-02 -2.14001819e-01
1.28709674e+00 -3.35815400e-02 8.28280330e-01 -2.76067168e-01
7.14322403e-02 -8.86781514e-02 8.45575511e-01 1.16062152e+00
-4.83323336e-01 1.41804367e-02 8.97228241e-01 -1.19845524e-01
-6.97898448e-01 -1.04995036e+00 -1.55723557e-01 6.81197703e-01
3.87217134e-01 2.62940794e-01 -3.77847314e-01 -3.83004010e-01
9.35896933e-02 9.35108244e-01 -5.89992046e-01 -1.02873281e-01
-2.57981718e-01 -7.31678426e-01 -2.93112546e-01 2.67731667e-01
3.29871714e-01 -9.27760899e-01 -6.16109312e-01 5.20802677e-01
6.61771968e-02 -8.60120177e-01 -4.96371314e-02 8.33453774e-01
-1.03483248e+00 -1.09788370e+00 -3.86935651e-01 -8.96930993e-02
3.32972527e-01 -2.59551376e-01 7.52942145e-01 -7.14530051e-01
3.95799041e-01 7.62018859e-01 -3.83955650e-02 -9.84010637e-01
-6.69870496e-01 -5.77786326e-01 4.65530843e-01 1.87347636e-01
-5.46242371e-02 -3.09398472e-01 -1.72330990e-01 1.94904462e-01
-1.04224420e+00 -2.71848798e-01 3.25445682e-01 7.10852027e-01
8.61492157e-01 6.66417301e-01 1.70074180e-01 -5.63798010e-01
5.72849691e-01 -5.74183166e-01 -1.06629419e+00 2.13465706e-01
-4.30846959e-01 4.99194026e-01 3.54451686e-01 -3.27368706e-01
-1.20784223e+00 8.09293240e-02 3.65733176e-01 -1.99162349e-01
-4.50512320e-01 2.98602581e-01 -6.32139862e-01 4.47684228e-01
-2.35169157e-02 1.89614564e-01 2.69850105e-01 -4.74342585e-01
-1.28907524e-03 5.30730426e-01 2.54653871e-01 -7.78172374e-01
3.57794374e-01 8.70834291e-01 5.91347277e-01 -6.53463125e-01
-6.17465317e-01 -2.80538738e-01 -5.41049182e-01 -3.07673961e-02
1.01688004e+00 -8.09474289e-01 -9.76669371e-01 2.29044318e-01
-9.26854789e-01 -4.41731721e-01 -7.86653996e-01 8.48392367e-01
-1.35521448e+00 6.73345849e-02 -7.44245827e-01 -1.62009883e+00
4.77522939e-01 -1.18813694e+00 9.95784581e-01 9.39268172e-02
-1.51275128e-01 -1.22943795e+00 1.86497584e-01 2.83363517e-02
-1.48953125e-01 4.22636151e-01 1.05003250e+00 -8.76069844e-01
-7.58687198e-01 -4.67325181e-01 6.85634673e-01 2.37357616e-01
1.60068609e-02 -5.13063312e-01 -6.62092805e-01 -4.75921065e-01
8.76242638e-01 -8.45364407e-02 5.75510800e-01 8.81977797e-01
8.77282262e-01 -3.86646777e-01 -4.55853611e-01 -1.45197868e-01
1.30584049e+00 9.62324202e-01 1.81055352e-01 2.82685071e-01
1.94355994e-02 7.63977528e-01 8.50616336e-01 7.25189984e-01
2.94412196e-01 5.79196393e-01 8.37009251e-01 5.90853035e-01
9.15809810e-01 -3.27243388e-01 5.21937251e-01 3.88206601e-01
-1.49849907e-01 -3.68245006e-01 -7.21343279e-01 2.42570817e-01
-2.30125403e+00 -9.84765351e-01 2.24908054e-01 2.66834044e+00
3.68441403e-01 3.86947125e-01 1.54489964e-01 1.80239230e-01
8.35558891e-01 1.09621659e-01 -7.93090045e-01 -5.72121084e-01
1.90446839e-01 1.05974235e-01 8.90232682e-01 6.63428903e-01
-1.13872087e+00 3.87187809e-01 6.61357927e+00 5.89229941e-01
-4.64815915e-01 2.30496094e-01 4.19878542e-01 -5.15625365e-02
-3.16675194e-02 2.05001950e-01 -7.61554182e-01 5.97125292e-01
1.13103318e+00 -1.39724016e-01 4.37055111e-01 7.22499132e-01
5.39575398e-01 -4.91702288e-01 -1.52607799e+00 5.57934940e-01
-9.10585165e-01 -7.57497787e-01 -2.55783111e-01 6.52726173e-01
7.84108877e-01 -1.85849279e-01 -1.49027631e-01 3.68060499e-01
1.17206919e+00 -8.83605003e-01 9.81540501e-01 6.43970609e-01
3.77205878e-01 -8.68937552e-01 8.07013512e-01 9.38441992e-01
-7.99366057e-01 -6.66271210e-01 -2.14693114e-01 -6.63654029e-01
6.70206130e-01 5.11442780e-01 -3.35497200e-01 5.52573681e-01
2.11058080e-01 1.33618861e-01 5.47366083e-01 8.23799014e-01
-5.03175795e-01 5.66482604e-01 -3.77698421e-01 1.45815918e-03
4.57894772e-01 -4.63235527e-01 7.89910614e-01 5.67515433e-01
1.35608748e-01 3.17789465e-01 7.30393648e-01 5.11862755e-01
4.91665512e-01 -5.08051872e-01 -4.73352015e-01 -2.48255059e-01
2.77404279e-01 4.94595081e-01 -9.95171845e-01 -3.31792891e-01
-6.11137562e-02 7.54373789e-01 -2.32607991e-01 4.79041249e-01
-5.67214906e-01 1.39528513e-01 8.60040247e-01 -8.79235417e-02
4.02093709e-01 -3.03145140e-01 -1.05402954e-01 -1.12148249e+00
-2.81280547e-01 -5.90038896e-01 5.45690775e-01 -3.53303432e-01
-1.06738174e+00 -1.03925705e-01 4.05983448e-01 -1.16034138e+00
-7.75317550e-01 -4.94614393e-01 -4.16756600e-01 8.31950724e-01
-1.19862700e+00 -3.90036047e-01 6.38993919e-01 7.38625348e-01
2.96461016e-01 5.33122895e-03 8.09889019e-01 -7.11997449e-01
-3.86235386e-01 -6.49463236e-01 6.94430888e-01 -4.23211634e-01
-8.03735666e-03 -1.56950021e+00 1.12756062e-02 5.66128433e-01
-3.03567201e-01 1.33680299e-01 1.04955924e+00 -8.23330879e-01
-1.50795722e+00 -1.05126512e+00 4.65132356e-01 -5.23059607e-01
6.84183359e-01 -2.09067330e-01 -6.06045127e-01 9.27596211e-01
-9.84415263e-02 -2.16149315e-01 3.85204613e-01 -1.51499778e-01
5.28191209e-01 1.41074494e-01 -1.45596957e+00 3.90839219e-01
5.78364074e-01 -1.29060626e-01 -9.74337518e-01 3.91801178e-01
6.68169081e-01 -4.59567726e-01 -7.24252224e-01 2.08327964e-01
3.40612382e-01 -5.39173007e-01 4.39669788e-01 -1.06140602e+00
1.23118095e-01 -1.55928373e-01 -3.04428875e-01 -1.36496687e+00
-4.42214310e-01 -1.00987542e+00 -6.28042340e-01 6.86949968e-01
4.73017842e-02 -8.65568817e-01 4.60652202e-01 8.25940907e-01
1.89250410e-01 -7.38754034e-01 -1.32561374e+00 -1.11068940e+00
5.52918538e-02 -6.32992327e-01 6.55813336e-01 6.36465073e-01
8.53803977e-02 6.85166717e-02 -3.39664042e-01 2.35332534e-01
8.86564434e-01 1.94484279e-01 1.50858283e-01 -1.17616880e+00
-8.53715897e-01 5.02486117e-02 -1.84964716e-01 -1.08325851e+00
9.58828107e-02 -2.70997822e-01 3.97815853e-01 -1.51865900e+00
1.44393876e-01 -3.73730242e-01 -4.07045305e-01 2.24034160e-01
2.60866106e-01 -8.72683287e-01 4.31173623e-01 4.32130247e-01
-7.87713468e-01 6.63589060e-01 1.38594437e+00 -8.25674459e-03
-3.41387600e-01 8.96271229e-01 -5.76287150e-01 8.77657115e-01
6.35760307e-01 -7.03950822e-01 -5.70176244e-01 1.88334301e-01
1.20918057e-03 1.09631431e+00 1.55175822e-02 -6.99829459e-01
-3.31396833e-02 -8.95564854e-01 4.25975397e-02 -3.09401184e-01
5.39026320e-01 -1.04269052e+00 3.31710994e-01 9.28569555e-01
-3.62355292e-01 -1.04018927e-01 -2.11460128e-01 1.15925753e+00
-5.93594350e-02 -5.84887981e-01 7.14317799e-01 -5.12562215e-01
-4.82067794e-01 4.31453824e-01 -1.00144386e+00 -6.86515048e-02
1.15861630e+00 5.92449047e-02 7.86583200e-02 -8.39785874e-01
-1.45409751e+00 5.87004185e-01 4.04321313e-01 3.22893821e-02
1.56855077e-01 -1.04797065e+00 -1.69277608e-01 -2.05367342e-01
-4.43384439e-01 3.01406860e-01 4.28123139e-02 6.59680784e-01
2.18199313e-01 7.21356452e-01 -1.27353117e-01 -4.80266988e-01
-6.56865180e-01 9.00364876e-01 4.39298041e-02 -7.93238342e-01
-3.69903296e-01 2.73491412e-01 1.72000542e-01 -1.91274196e-01
6.22124709e-02 -4.84870702e-01 -6.99228272e-02 1.51145056e-01
7.13560283e-01 5.57065129e-01 -2.04369351e-01 -3.16613019e-01
-4.85298969e-02 -1.94687709e-01 1.50025666e-01 -7.27542877e-01
1.23618793e+00 -4.85391915e-01 2.83268429e-02 1.08378768e+00
4.96385843e-01 -5.50972700e-01 -1.74456787e+00 -3.64677221e-01
3.38459074e-01 -5.39198809e-04 -3.81539017e-02 -6.02336645e-01
-3.80202234e-01 6.86438739e-01 3.04186404e-01 6.19575560e-01
8.68902743e-01 5.77012673e-02 2.10014954e-01 4.14481848e-01
1.02314258e+00 -1.19330740e+00 -2.65261859e-01 6.38219237e-01
6.17436886e-01 -1.01114118e+00 -1.62683919e-01 -1.78050213e-02
-8.80532086e-01 8.04858029e-01 -2.00653732e-01 -1.53931573e-01
7.99575388e-01 6.00663066e-01 -5.52574754e-01 9.01281238e-02
-1.14833927e+00 -4.08853590e-01 -2.72793293e-01 6.57199681e-01
-4.07956600e-01 4.66447920e-01 -1.69976696e-01 7.70645678e-01
7.22532347e-02 3.22885096e-01 6.49218976e-01 1.61563373e+00
-7.46533990e-01 -1.01290822e+00 -9.00796294e-01 5.47693074e-01
-5.14697909e-01 3.98145705e-01 1.87037662e-02 8.11482012e-01
-1.72391452e-03 1.05397403e+00 1.71448439e-01 1.98366120e-01
2.14026973e-01 3.07390243e-01 5.73242724e-01 -8.28043997e-01
2.98477501e-01 1.24293111e-01 1.16980292e-01 -7.10626721e-01
-3.00328344e-01 -1.09654605e+00 -1.12186837e+00 -1.02847658e-01
-1.52218759e-01 3.64563525e-01 6.74859703e-01 1.60410011e+00
-1.79787233e-01 4.14607555e-01 6.54203594e-01 -9.53946233e-01
-1.37890434e+00 -8.25052321e-01 -1.33801448e+00 -1.56259984e-01
5.35373569e-01 -8.34727585e-01 -4.07654941e-01 -2.71773010e-01] | [4.362250804901123, 2.277177333831787] |
acc1381f-74f0-4182-bd10-346201ca89b6 | a-multi-hypothesis-classification-approach-to | 2002.12896 | null | https://arxiv.org/abs/2002.12896v2 | https://arxiv.org/pdf/2002.12896v2.pdf | A Multi-Hypothesis Approach to Color Constancy | Contemporary approaches frame the color constancy problem as learning camera specific illuminant mappings. While high accuracy can be achieved on camera specific data, these models depend on camera spectral sensitivity and typically exhibit poor generalisation to new devices. Additionally, regression methods produce point estimates that do not explicitly account for potential ambiguities among plausible illuminant solutions, due to the ill-posed nature of the problem. We propose a Bayesian framework that naturally handles color constancy ambiguity via a multi-hypothesis strategy. Firstly, we select a set of candidate scene illuminants in a data-driven fashion and apply them to a target image to generate of set of corrected images. Secondly, we estimate, for each corrected image, the likelihood of the light source being achromatic using a camera-agnostic CNN. Finally, our method explicitly learns a final illumination estimate from the generated posterior probability distribution. Our likelihood estimator learns to answer a camera-agnostic question and thus enables effective multi-camera training by disentangling illuminant estimation from the supervised learning task. We extensively evaluate our proposed approach and additionally set a benchmark for novel sensor generalisation without re-training. Our method provides state-of-the-art accuracy on multiple public datasets (up to 11% median angular error improvement) while maintaining real-time execution. | ['Daniel Hernandez-Juarez', 'Steven McDonagh', 'Benjamin Busam', 'Ales Leonardis', 'Sarah Parisot', 'Gregory Slabaugh'] | 2020-02-28 | a-multi-hypothesis-approach-to-color | http://openaccess.thecvf.com/content_CVPR_2020/html/Hernandez-Juarez_A_Multi-Hypothesis_Approach_to_Color_Constancy_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Hernandez-Juarez_A_Multi-Hypothesis_Approach_to_Color_Constancy_CVPR_2020_paper.pdf | cvpr-2020-6 | ['color-constancy'] | ['computer-vision'] | [ 6.40303791e-01 -4.65938389e-01 -4.85329665e-02 -4.83772308e-01
-1.27128887e+00 -9.53765333e-01 6.37592673e-01 -1.59197718e-01
-4.63131607e-01 6.98559403e-01 -1.76885314e-02 -1.36814892e-01
-7.18512684e-02 -6.17223978e-01 -1.11348164e+00 -9.67344522e-01
5.89178264e-01 2.50822335e-01 1.57934893e-02 3.25260252e-01
3.47412258e-01 3.49312335e-01 -1.65168309e+00 -1.40680507e-01
9.45387006e-01 1.13901460e+00 4.41785544e-01 1.11537254e+00
3.31862181e-01 3.30935091e-01 -4.24669921e-01 -3.25564682e-01
4.97817069e-01 -2.95485377e-01 -1.06974222e-01 2.47818753e-01
9.31350172e-01 -2.71369189e-01 8.21893737e-02 1.20266998e+00
4.16888297e-01 -9.97298062e-02 7.42601633e-01 -1.11364949e+00
-3.80399257e-01 -1.36554033e-01 -6.21103585e-01 -1.35320351e-01
3.68345022e-01 4.41894710e-01 9.37873185e-01 -6.81164622e-01
6.98794797e-02 7.19643593e-01 5.82202971e-01 2.85767943e-01
-1.34523749e+00 -6.55158401e-01 -1.85363218e-02 -3.33164409e-02
-1.26300168e+00 -4.69409257e-01 7.58865118e-01 -3.44579697e-01
7.12004960e-01 2.26261437e-01 4.61251199e-01 1.14639854e+00
1.47237808e-01 3.47101271e-01 1.55642831e+00 -3.98693740e-01
3.73615891e-01 2.10283220e-01 -4.38421667e-01 5.25254071e-01
2.90089279e-01 2.59394377e-01 -6.44719601e-01 9.86235589e-02
8.30447435e-01 -4.29254472e-01 -3.32600325e-01 -4.23167139e-01
-1.19365096e+00 5.72201073e-01 5.54670513e-01 -5.10511339e-01
-2.46882588e-01 4.03705508e-01 -3.48455347e-02 -1.60519004e-01
2.52895951e-01 5.48727453e-01 -7.18065798e-01 -1.12226848e-02
-6.24289095e-01 -2.63222326e-02 5.68711996e-01 9.84769046e-01
1.05229509e+00 -1.76687241e-01 7.14028180e-02 6.28315985e-01
4.25182581e-01 9.53637660e-01 -8.94391090e-02 -9.04182851e-01
3.77354324e-01 1.79113951e-02 3.61033350e-01 -6.37966394e-01
-4.12272304e-01 -5.03436565e-01 -6.48029208e-01 3.36468160e-01
5.98203599e-01 -1.61370635e-01 -9.05300558e-01 1.69242799e+00
1.50489390e-01 3.04077208e-01 4.82761562e-02 1.13925028e+00
3.48596513e-01 5.28155625e-01 1.00554831e-01 -1.80483609e-01
1.36571622e+00 -7.93732285e-01 -2.56016999e-01 -6.44141078e-01
-7.70148858e-02 -9.95991647e-01 1.01576042e+00 9.39606726e-01
-7.22236514e-01 -5.24078727e-01 -1.12108922e+00 -3.81115377e-02
-9.37901959e-02 4.91903484e-01 8.02819848e-01 1.07176387e+00
-8.98872554e-01 1.36893585e-01 -5.31087816e-01 -3.94023567e-01
2.82677084e-01 2.25117549e-01 -1.08785242e-01 -2.92134583e-01
-6.05981886e-01 9.86028075e-01 3.31725597e-01 6.85524717e-02
-9.42356646e-01 -7.17529178e-01 -7.34252214e-01 -2.15762407e-01
5.89354873e-01 -7.70068049e-01 1.13865602e+00 -1.12131739e+00
-1.88397217e+00 5.77239037e-01 -2.09213912e-01 -2.23745257e-01
4.28735197e-01 -2.82440007e-01 -3.80683511e-01 4.14406769e-02
-8.00630078e-03 6.64787233e-01 1.16319215e+00 -1.59138286e+00
-6.88779175e-01 -6.98076263e-02 2.36497194e-01 2.75808215e-01
5.81865199e-02 -2.96399117e-01 -6.82024002e-01 -1.70615792e-01
1.67406071e-02 -1.02000201e+00 -9.34900269e-02 1.78636625e-01
-5.29887617e-01 2.94023424e-01 3.53719920e-01 -2.46431574e-01
5.70504248e-01 -2.06091595e+00 1.95803437e-02 9.99136865e-02
-3.32863927e-02 -1.22137517e-01 -3.06587905e-01 -1.40877375e-02
-1.31358057e-02 -2.90271282e-01 -2.54796118e-01 -2.83903033e-01
-7.93293342e-02 4.71970066e-02 -1.51981741e-01 7.71996319e-01
3.77371907e-01 4.49772209e-01 -9.89159346e-01 -1.35725960e-01
9.54372287e-01 6.03614986e-01 -6.78955674e-01 5.01543701e-01
-5.26170611e-01 5.12950718e-01 1.37990061e-02 6.27532959e-01
9.62027788e-01 -7.97166377e-02 1.29633978e-01 -7.51195908e-01
-3.95904243e-01 4.02818844e-02 -1.32567489e+00 1.86346936e+00
-9.77510631e-01 7.05177963e-01 -1.73616081e-01 -4.93046671e-01
9.03652310e-01 -2.13710681e-01 2.72172898e-01 -5.16427338e-01
1.69712394e-01 2.29656428e-01 -3.14625323e-01 -5.26156604e-01
4.88036454e-01 -3.04854482e-01 3.36773731e-02 9.77107361e-02
1.43178165e-01 -6.70164466e-01 -3.09804082e-01 -2.40967572e-01
7.06715286e-01 6.47470176e-01 3.81228864e-01 -8.69313404e-02
3.26063335e-01 -2.58831024e-01 3.41591984e-01 8.27867866e-01
-2.33313031e-02 9.82526302e-01 1.67477980e-01 -2.57488370e-01
-1.06834841e+00 -1.45532477e+00 -2.51443028e-01 7.78401256e-01
3.06406975e-01 -1.57009006e-01 -5.43080986e-01 -3.38223726e-01
-2.12727502e-01 1.01649821e+00 -3.53366077e-01 4.50597964e-02
-1.20947741e-01 -1.14709771e+00 1.13859706e-01 3.00852180e-01
6.67335272e-01 -3.18421304e-01 -9.76189494e-01 -6.81191608e-02
-1.58896178e-01 -1.45976985e+00 -2.44742185e-01 4.19399261e-01
-3.29940975e-01 -1.31486523e+00 -2.17776045e-01 -1.27748564e-01
5.72145879e-01 5.31165957e-01 1.15461111e+00 -4.88825768e-01
-4.70149130e-01 7.29017556e-01 -5.29249273e-02 -5.86250007e-01
-2.08903551e-01 -1.79651916e-01 5.19192852e-02 1.36088774e-01
3.58861297e-01 -2.09752843e-01 -9.41729188e-01 3.23816121e-01
-7.55205154e-01 2.15769306e-01 7.73440897e-01 6.37327671e-01
5.01529157e-01 7.61502609e-02 2.03167349e-01 -6.16740942e-01
9.48851258e-02 -3.77857149e-01 -1.27665925e+00 2.48582885e-01
-6.79741323e-01 1.96975261e-01 6.14700258e-01 -3.75277400e-01
-1.47825539e+00 6.87034130e-01 2.29472056e-01 -2.61247963e-01
-4.03410167e-01 7.39591494e-02 -3.30811322e-01 -2.17716202e-01
9.38509762e-01 9.40637589e-02 -3.96471828e-01 -8.36758092e-02
6.71241760e-01 4.35650110e-01 1.01542103e+00 -7.73856640e-01
9.33981001e-01 5.82304358e-01 4.55099404e-01 -8.68122220e-01
-1.10559952e+00 -5.37253857e-01 -6.09558940e-01 -4.90196317e-01
1.00882769e+00 -1.25078535e+00 -8.73738289e-01 4.80353087e-01
-1.17131257e+00 -2.12388232e-01 1.41674086e-01 5.14852107e-01
-5.17342150e-01 2.25604504e-01 1.97694346e-01 -1.01764619e+00
1.14991322e-01 -1.25183368e+00 1.46943200e+00 4.58569169e-01
2.13955894e-01 -9.46627319e-01 6.17981330e-02 4.18949574e-01
3.61942858e-01 4.07045901e-01 6.53542697e-01 2.31207266e-01
-1.03101850e+00 -2.18656525e-01 -7.23694086e-01 3.70801896e-01
3.17470610e-01 1.12542622e-01 -1.72086537e+00 -1.84825912e-01
-1.35298565e-01 -3.31048906e-01 7.67706215e-01 6.26653671e-01
1.26038134e+00 -2.44041849e-02 1.49438716e-02 8.93610239e-01
2.08055472e+00 -3.34893048e-01 5.50639033e-01 4.56616223e-01
8.82869720e-01 4.63416398e-01 4.55286413e-01 6.63912594e-01
4.62370902e-01 7.27696955e-01 1.00621271e+00 -1.73063092e-02
-1.37969762e-01 -7.79499188e-02 2.06474707e-01 1.63539145e-02
-2.43235659e-02 -4.88866627e-01 -6.20537162e-01 3.29659730e-01
-1.51467717e+00 -6.61668181e-01 -3.89871091e-01 2.60887218e+00
7.45154619e-01 6.67828098e-02 5.87610789e-02 -1.51790634e-01
5.14634728e-01 1.24514483e-01 -7.63000488e-01 -1.71064783e-03
-3.65936875e-01 1.40629858e-01 1.22343469e+00 6.39348745e-01
-1.04106474e+00 7.01793790e-01 6.14864349e+00 3.89563620e-01
-1.15857053e+00 -3.15396748e-02 5.49729228e-01 -1.98700652e-01
-3.82007599e-01 2.22377688e-01 -6.75300479e-01 1.97903022e-01
9.26292956e-01 1.95085987e-01 8.90631914e-01 5.33198178e-01
1.27368480e-01 -6.85636342e-01 -1.22945285e+00 1.41918099e+00
4.83757049e-01 -1.12250817e+00 -3.85900289e-01 -6.21313192e-02
7.77766526e-01 2.75037050e-01 1.96239933e-01 -2.34214664e-01
2.45533884e-01 -9.24575210e-01 9.34156656e-01 7.70326316e-01
9.89926755e-01 -6.84127867e-01 3.63212675e-01 7.72502720e-02
-9.59946215e-01 -1.90453365e-01 -4.65474129e-01 -5.70990555e-02
8.12936481e-03 7.47944295e-01 -9.84346211e-01 6.40025795e-01
6.02496147e-01 5.08630633e-01 -7.45136976e-01 1.28084922e+00
-5.60478687e-01 3.83669496e-01 -5.33931255e-01 -1.28184482e-02
-4.94014435e-02 -2.96424031e-01 3.98582160e-01 1.16871119e+00
4.31463212e-01 -1.25642121e-01 -2.18020648e-01 9.18401301e-01
3.69305313e-02 -3.84951621e-01 -4.16755885e-01 4.23521131e-01
4.49687332e-01 1.51708364e+00 -6.10739768e-01 1.75747424e-01
-6.46129489e-01 1.13064420e+00 1.10003121e-01 4.77566540e-01
-9.61465955e-01 1.47120161e-02 6.09510064e-01 -1.56658202e-01
1.61919311e-01 -2.41557211e-01 -3.96657437e-01 -1.20528114e+00
3.25119356e-03 -6.38276100e-01 1.28169179e-01 -1.33468270e+00
-1.38484323e+00 2.22864375e-01 1.29572436e-01 -1.26512504e+00
-1.21785596e-01 -1.15776086e+00 -4.90766436e-01 9.20306325e-01
-1.91532445e+00 -1.34112811e+00 -7.51352906e-01 4.00160968e-01
5.94165862e-01 1.31016031e-01 8.04182053e-01 6.27706125e-02
-4.38887626e-01 1.99352011e-01 1.31056920e-01 -3.73722583e-01
1.10527694e+00 -1.52962196e+00 7.48463199e-02 1.14027560e+00
7.19477907e-02 3.39046896e-01 1.12246275e+00 -2.21979529e-01
-1.80414736e+00 -1.14291775e+00 1.76897898e-01 -7.95915067e-01
6.57094836e-01 -6.64043069e-01 -3.83619934e-01 3.40538591e-01
2.44254470e-01 2.58969218e-01 5.45455992e-01 8.22457075e-02
-7.57372618e-01 -4.16807681e-01 -1.05834913e+00 4.93207902e-01
7.69996583e-01 -6.37828588e-01 -6.61822259e-02 5.01193106e-01
3.38043243e-01 -5.40823758e-01 -4.41600859e-01 1.64879709e-01
7.32823133e-01 -1.31303775e+00 9.88068938e-01 1.04184151e-01
3.73444051e-01 -5.40038228e-01 -5.65241456e-01 -1.35403657e+00
-7.85584450e-02 -5.58688700e-01 2.32707933e-01 1.20842469e+00
3.65702808e-01 -5.31089902e-01 4.61844981e-01 6.34880781e-01
-3.05017438e-02 -2.00157449e-01 -7.84911692e-01 -7.21123099e-01
-1.94278166e-01 -8.74464035e-01 5.31861067e-01 4.88804579e-01
-6.37235582e-01 2.69817084e-01 -6.55643046e-01 8.57219756e-01
1.16448331e+00 -9.39455058e-04 1.08130586e+00 -1.06825566e+00
-5.88804722e-01 -2.81296492e-01 -2.77397074e-02 -9.74888265e-01
7.93606266e-02 -3.81390452e-01 5.72705925e-01 -1.25047934e+00
3.46328974e-01 -3.72504920e-01 -1.12904735e-01 8.14333111e-02
-1.91198230e-01 4.73869413e-01 -1.96346827e-03 -5.54528199e-02
-6.65583014e-01 2.88455993e-01 7.20474303e-01 -2.93276813e-02
-2.16401573e-02 -6.51095212e-02 -7.76145637e-01 6.53054059e-01
6.80784941e-01 -2.38613322e-01 -4.66332257e-01 -7.35000432e-01
8.37292194e-01 -2.70833313e-01 8.65548670e-01 -1.16127074e+00
1.91608682e-01 -4.20748383e-01 6.56461716e-01 -2.69094735e-01
3.74358267e-01 -1.04341197e+00 1.31048918e-01 -4.90417480e-02
-3.21475789e-02 -3.33049089e-01 2.77663410e-01 7.47374177e-01
3.48004699e-01 -1.77301407e-01 1.00761378e+00 5.84057570e-02
-6.49671555e-01 8.86411145e-02 -8.07744786e-02 -2.15819567e-01
6.58572376e-01 -1.33135855e-01 -4.49791014e-01 -2.84103483e-01
1.23139307e-01 -1.98759735e-01 6.45161152e-01 3.33517313e-01
4.02017653e-01 -1.00362742e+00 -5.22835076e-01 1.37973443e-01
6.15994632e-01 1.71348840e-01 8.22402537e-02 5.58750868e-01
-4.31117684e-01 2.58590460e-01 1.33341933e-02 -9.80307758e-01
-8.80220711e-01 4.22062308e-01 5.63445389e-01 4.87632722e-01
-1.27620563e-01 8.57295215e-01 2.86054492e-01 -2.57053465e-01
-1.09931596e-01 -4.94567871e-01 3.04798335e-01 -2.89032340e-01
2.33365238e-01 1.61960557e-01 7.25600421e-02 -4.83272642e-01
-3.65558952e-01 8.66143823e-01 4.15831208e-01 -1.12434685e-01
1.26771104e+00 -4.92851406e-01 8.04084986e-02 5.27339756e-01
1.16630471e+00 1.93694979e-01 -2.00341988e+00 -1.12609237e-01
-4.36024666e-01 -8.07568073e-01 4.45438087e-01 -1.20415008e+00
-8.01949441e-01 6.02140725e-01 8.04422438e-01 -1.03170857e-01
1.39287579e+00 -2.94572301e-02 1.64634153e-01 3.32604319e-01
2.08643034e-01 -1.09292972e+00 2.04712421e-01 1.56362891e-01
6.00819170e-01 -1.91226196e+00 3.18492860e-01 -3.70284736e-01
-5.35806715e-01 1.32036483e+00 5.91847956e-01 4.26193476e-02
2.94517606e-01 3.29255879e-01 1.96432367e-01 -7.56316185e-02
-3.34859043e-01 -1.35324866e-01 4.33201045e-01 8.61443162e-01
3.33509117e-01 5.23386821e-02 5.08361280e-01 1.24746636e-01
-2.54076481e-01 -2.51536489e-01 6.17506266e-01 3.31525713e-01
-2.74928272e-01 -7.81674623e-01 -6.05709136e-01 2.36712813e-01
6.34767711e-02 -1.48391172e-01 -6.12458549e-02 5.90692520e-01
2.78729975e-01 1.16284811e+00 7.03773946e-02 -1.94516361e-01
2.15801343e-01 -2.65254557e-01 9.09242511e-01 -4.57112938e-01
2.03632358e-02 3.21565092e-01 -5.59194200e-02 -5.95991671e-01
-5.87879539e-01 -9.40927088e-01 -5.79816282e-01 1.19165733e-01
-5.21932006e-01 -4.77959901e-01 1.18521929e+00 9.84205365e-01
6.39858516e-03 3.75053048e-01 8.29830468e-01 -1.07889783e+00
-2.18976438e-01 -6.70185685e-01 -2.82659739e-01 1.05874747e-01
6.73393309e-01 -6.35547996e-01 -5.69626391e-01 1.89984903e-01] | [10.20667552947998, -2.6251718997955322] |
daafb498-073f-481c-8e2e-4ded16f4f3c5 | music-representing-corpus-virtual-an-open | 2305.14948 | null | https://arxiv.org/abs/2305.14948v1 | https://arxiv.org/pdf/2305.14948v1.pdf | Music Representing Corpus Virtual: An Open Sourced Library for Explorative Music Generation, Sound Design, and Instrument Creation with Artificial Intelligence and Machine Learning | Music Representing Corpus Virtual (MRCV) is an open source software suite designed to explore the capabilities of Artificial Intelligence (AI) and Machine Learning (ML) in Music Generation, Sound Design, and Virtual Instrument Creation (MGSDIC). The software is accessible to users of varying levels of experience, with an emphasis on providing an explorative approach to MGSDIC. The main aim of MRCV is to facilitate creativity, allowing users to customize input datasets for training the neural networks, and offering a range of options for each neural network (thoroughly documented in the Github Wiki). The software suite is designed to be accessible to musicians, audio professionals, sound designers, and composers, regardless of their prior experience in AI or ML. The documentation is prepared in such a way as to abstract technical details, thereby making it easy to understand. The software is open source, meaning users can contribute to its development, and the community can collectively benefit from the insights and experience of other users. | ['Christopher Johann Clarke'] | 2023-05-24 | null | null | null | null | ['music-generation', 'music-generation'] | ['audio', 'music'] | [ 4.42875847e-02 1.64345831e-01 -2.49909181e-02 4.40995306e-01
-4.78700310e-01 -9.89215851e-01 2.00793356e-01 -2.16434360e-01
-4.10043821e-02 2.90322632e-01 5.16817808e-01 -3.65778089e-01
-3.29080611e-01 -5.44059932e-01 -6.03948571e-02 -3.82941753e-01
-1.86384432e-02 4.54046667e-01 -3.71556908e-01 -2.82855153e-01
5.17394185e-01 3.28957915e-01 -1.76267254e+00 2.29550719e-01
6.24845862e-01 5.84252656e-01 4.91040289e-01 6.99900925e-01
-2.33588949e-01 7.99880743e-01 -7.30265915e-01 -2.87930101e-01
2.67678171e-01 -5.66676557e-01 -8.91215146e-01 -3.27039957e-01
1.59473926e-01 1.95478737e-01 2.29607537e-01 8.02181244e-01
7.89636493e-01 2.13489577e-01 2.30420545e-01 -1.03751242e+00
-7.30410397e-01 1.03891337e+00 1.84266329e-01 -8.74373093e-02
4.47718978e-01 3.92208695e-01 1.24111140e+00 -7.47432530e-01
7.67054737e-01 8.73557448e-01 7.03216612e-01 5.68058789e-01
-1.12244987e+00 -7.87621915e-01 -1.99352071e-01 6.37507737e-02
-1.21229732e+00 -6.59263134e-01 9.11773741e-01 -6.98154211e-01
8.17603290e-01 4.47631657e-01 1.23902643e+00 9.16167140e-01
-5.10639310e-01 7.99115360e-01 5.66230536e-01 -8.63708317e-01
3.99172157e-01 2.22495392e-01 -2.49134704e-01 4.02418405e-01
-1.04088724e-01 4.46037278e-02 -7.93297350e-01 -3.59546214e-01
1.03788614e+00 -4.88984287e-01 -4.71186012e-01 -1.42075390e-01
-1.56795454e+00 6.04823589e-01 1.21274218e-01 6.46693349e-01
-5.43635070e-01 8.60778317e-02 4.98170108e-01 3.51389468e-01
-9.48368460e-02 1.40902472e+00 -6.57344460e-01 -7.99050450e-01
-6.85065031e-01 5.47753930e-01 9.70968068e-01 6.74601376e-01
1.81567341e-01 5.20095468e-01 2.15136066e-01 1.24346924e+00
3.68456423e-01 1.13999553e-01 6.95978820e-01 -1.56345356e+00
-1.12395383e-01 5.45426786e-01 -1.23268832e-03 -8.94204557e-01
-2.40314931e-01 -5.61916471e-01 -4.97219384e-01 6.39545262e-01
3.87815952e-01 -1.20519280e-01 -4.05369699e-01 1.59559143e+00
-9.20702741e-02 -1.12609468e-01 1.40688885e-02 7.56778657e-01
1.18923962e+00 4.93682265e-01 -7.90244862e-02 -3.09409909e-02
1.14346886e+00 -7.25708306e-01 -5.46803057e-01 -2.31174707e-01
4.10961598e-01 -1.00156248e+00 1.53892303e+00 9.52015996e-01
-1.47569704e+00 -5.36982834e-01 -1.03049684e+00 5.67172132e-02
-2.81009614e-01 4.79163378e-02 9.04674292e-01 5.51863968e-01
-8.84755731e-01 9.81558502e-01 -5.33013105e-01 -3.42052072e-01
3.16816866e-01 2.33279660e-01 -6.89520314e-03 3.41008574e-01
-1.01592422e+00 8.14804971e-01 5.15569150e-01 -1.91508383e-01
-1.24909379e-01 -1.01399159e+00 -5.83349764e-01 -1.79419354e-01
1.92256361e-01 -8.60802174e-01 1.56363559e+00 -1.30456018e+00
-1.53967619e+00 6.83301568e-01 2.48933658e-01 1.66919082e-02
1.82306826e-01 7.76289180e-02 -3.60618770e-01 1.51374973e-02
1.66834399e-01 5.03436267e-01 3.71659458e-01 -1.08629274e+00
-4.85948682e-01 -2.02086002e-01 -9.79831815e-02 3.85348290e-01
-2.04447731e-01 1.42720491e-01 -4.72503066e-01 -1.06194222e+00
-4.48105671e-02 -9.38465357e-01 -1.72498375e-02 3.74574848e-02
-2.57385969e-01 2.73712501e-02 5.79828024e-01 -8.25332344e-01
1.38652766e+00 -2.21052670e+00 2.49923244e-01 1.64431125e-01
-7.77274370e-02 2.65893817e-01 -2.50630006e-02 5.98034203e-01
7.27861971e-02 2.79386044e-01 -1.12561202e-02 4.82643582e-02
1.74587846e-01 -4.85458896e-02 -1.37191089e-02 -2.88625002e-01
-2.56500483e-01 6.37689650e-01 -9.47235942e-01 -1.88364327e-01
3.08513671e-01 4.34322059e-01 -4.69887942e-01 -8.22224170e-02
-4.00603175e-01 6.50702059e-01 -3.20423007e-01 6.62953019e-01
1.02498710e-01 -1.49292350e-01 3.02078217e-01 -3.19869071e-02
-4.83931780e-01 4.78167444e-01 -1.55928397e+00 1.86768460e+00
-7.56420612e-01 9.47184324e-01 1.66464746e-01 -2.01398626e-01
9.11115468e-01 6.41621888e-01 1.73691660e-01 -2.37181306e-01
3.44205014e-02 3.38971525e-01 2.09262967e-01 -4.54235315e-01
3.74671042e-01 -8.21143314e-02 2.41997346e-01 1.06914926e+00
1.09702528e-01 -1.34115472e-01 3.81551206e-01 3.70166600e-02
8.24822068e-01 3.92540902e-01 2.76183605e-01 4.20537125e-03
4.04238105e-02 6.06794581e-02 2.52216011e-01 4.68319833e-01
4.17993218e-01 3.06567788e-01 1.80556536e-01 -2.70777136e-01
-1.22300327e+00 -9.62161183e-01 -2.23957285e-01 1.26764643e+00
-5.48020661e-01 -6.57108843e-01 -4.96251285e-01 3.99478942e-01
-2.17390060e-01 9.93151188e-01 -3.17442358e-01 1.29732177e-01
-2.92980641e-01 3.73278409e-02 5.83050191e-01 4.92278814e-01
2.74812520e-01 -1.87230837e+00 -7.86189675e-01 2.48739883e-01
-3.31970185e-01 -4.48516399e-01 -2.96637118e-01 -3.59279499e-03
-7.68093944e-01 -8.43488574e-01 -4.79150951e-01 -9.74519074e-01
9.99594405e-02 -1.50887787e-01 1.11918104e+00 2.19436154e-01
-4.91803110e-01 5.72101057e-01 -4.36433762e-01 -6.29829347e-01
-7.28689134e-01 1.02360070e-01 -1.16080418e-01 -5.78196049e-01
-3.42884809e-02 -1.02118552e+00 -2.12949768e-01 -3.85954715e-02
-8.06444526e-01 4.02030617e-01 3.21300268e-01 5.07748783e-01
3.82435590e-01 -8.13372582e-02 7.76484489e-01 -6.79475963e-01
9.31642711e-01 -1.93389609e-01 -2.73473173e-01 2.27533490e-03
-8.36768210e-01 -2.91607440e-01 3.83042514e-01 -4.46866095e-01
-6.94765151e-01 -1.14989892e-01 -1.66801363e-01 -2.82705694e-01
-3.43702495e-01 7.88380265e-01 -2.54694998e-01 -1.19932346e-01
7.86825061e-01 3.10772397e-02 1.12967417e-01 -8.39949787e-01
5.03087580e-01 8.27699482e-01 7.51091003e-01 -3.95002186e-01
5.39195418e-01 -2.02188224e-01 -3.75519544e-01 -8.15872371e-01
-2.00564966e-01 5.04767429e-03 -6.25282645e-01 -4.84621316e-01
5.71107805e-01 -5.42385340e-01 -7.99249828e-01 3.47098887e-01
-7.04043686e-01 -4.65427488e-01 -5.59711814e-01 4.68421668e-01
-5.17419636e-01 -6.61500171e-02 -4.47131872e-01 -6.07953966e-01
-5.53955853e-01 -9.90769327e-01 3.41636807e-01 4.70238626e-01
-1.06010783e+00 -1.15719056e+00 1.05294697e-01 3.61293793e-01
3.05114478e-01 3.96880090e-01 1.26973093e+00 -6.71827674e-01
-4.22494054e-01 -2.71983266e-01 4.07087147e-01 2.51860142e-01
5.32030389e-02 3.82123232e-01 -1.14348638e+00 2.77909189e-01
-5.90415597e-01 -3.00554544e-01 2.72034734e-01 4.18880731e-01
9.48226333e-01 -5.42634130e-01 3.74113694e-02 4.15807456e-01
1.08219683e+00 5.04719019e-01 5.78953147e-01 7.86211073e-01
5.41374683e-01 4.86646742e-01 1.31232381e-01 5.83565891e-01
1.23384021e-01 6.69820130e-01 1.93610623e-01 -4.88103181e-02
-1.45305917e-01 -1.89599022e-01 1.46596208e-01 7.76669085e-01
-4.94121850e-01 2.54210413e-01 -8.48873615e-01 4.31557745e-01
-1.78884590e+00 -1.20971644e+00 -2.41132289e-01 2.36285043e+00
8.44421148e-01 -1.37202919e-01 3.47071618e-01 5.39635897e-01
3.79136354e-01 3.11032198e-02 -2.61612266e-01 -8.13683748e-01
-4.58053499e-03 6.24652922e-01 -3.08005899e-01 3.45931768e-01
-6.34689152e-01 9.13715899e-01 6.64419985e+00 6.10962570e-01
-1.08608019e+00 -1.41188219e-01 4.86400793e-04 -5.22685528e-01
-5.13540089e-01 8.43763128e-02 -6.94053695e-02 1.93986624e-01
6.48087740e-01 -6.25425875e-01 1.14141417e+00 7.78624058e-01
4.45625991e-01 1.70735151e-01 -7.48785198e-01 8.87133121e-01
-1.37127742e-01 -1.96463227e+00 -2.01527938e-01 -4.01275493e-02
5.59926033e-01 -2.64076386e-02 -2.47641397e-03 1.09779857e-01
1.44242927e-01 -9.67283666e-01 9.24783289e-01 4.59280878e-01
7.30304539e-01 -9.38367188e-01 1.93650231e-01 2.31985092e-01
-1.04901648e+00 -3.00029725e-01 9.85857472e-02 -4.71485287e-01
1.60558134e-01 9.65056345e-02 -8.04710984e-01 3.06969136e-01
7.45038211e-01 4.18127537e-01 -4.90107507e-01 1.19840491e+00
-1.13126487e-01 7.05395401e-01 -1.16023906e-01 -1.26648486e-01
-1.87570006e-01 -1.51951849e-01 7.11130798e-01 1.07748914e+00
1.39613509e-01 4.77645099e-02 4.46423963e-02 1.02695715e+00
2.81084746e-01 3.90828311e-01 -4.83486623e-01 -5.75598717e-01
8.38736117e-01 1.30776680e+00 -6.17484748e-01 1.89019412e-01
-1.41078278e-01 7.05944300e-01 -1.49350643e-01 2.53529876e-01
-7.66883940e-02 -7.96570301e-01 6.37920320e-01 2.28048608e-01
1.64833128e-01 -1.67610601e-01 -6.18838787e-01 -4.97382998e-01
-1.60115048e-01 -1.34499240e+00 1.65673643e-01 -1.30974185e+00
-8.14835846e-01 6.91209078e-01 -2.32242063e-01 -1.12001753e+00
-4.79767472e-01 -3.68454576e-01 -8.23998690e-01 9.84884799e-01
-3.38303149e-01 -1.21823597e+00 -2.55948633e-01 1.27011776e-01
2.35869497e-01 -5.73381186e-01 1.26599514e+00 2.92549163e-01
-4.63486016e-01 1.87405646e-01 1.68150961e-02 4.91636544e-02
5.48347831e-01 -9.96841908e-01 3.57643038e-01 3.55375111e-01
5.05299866e-01 7.57532537e-01 6.74942195e-01 -5.27352512e-01
-1.36768496e+00 -6.79520845e-01 8.22147131e-01 -4.83889997e-01
6.09905839e-01 7.70183653e-02 -5.67959726e-01 7.13867247e-01
1.07374303e-01 -7.93458402e-01 1.25171351e+00 4.79534924e-01
-1.29063591e-01 1.73535779e-01 -7.01180100e-01 8.49913061e-01
8.74624074e-01 -3.95180792e-01 -5.54784298e-01 1.46387935e-01
4.08345461e-01 -4.02188629e-01 -9.88592207e-01 1.35716349e-01
9.82729971e-01 -7.65282273e-01 8.77774954e-01 -4.45274204e-01
4.89222616e-01 -5.08113027e-01 7.03150555e-02 -1.15190518e+00
-6.00672305e-01 -9.20794785e-01 1.17816404e-01 1.35269296e+00
8.20197165e-01 -6.19191170e-01 4.18074667e-01 4.74628329e-01
-4.49663639e-01 -5.02630413e-01 -5.65461278e-01 -2.84152627e-01
-1.22720055e-01 -1.08043635e+00 7.57725358e-01 1.11639559e+00
5.73296666e-01 3.22746426e-01 -2.17236087e-01 -1.79546207e-01
1.53570309e-01 1.94362581e-01 9.54426348e-01 -1.65294302e+00
-8.90454829e-01 -1.00532460e+00 -2.48779744e-01 -3.86979669e-01
-4.34275389e-01 -1.16355038e+00 -3.45818192e-01 -1.79545593e+00
-1.94804132e-01 -3.89719725e-01 -6.03398159e-02 7.46183634e-01
9.74115729e-02 5.00486016e-01 5.54780185e-01 4.82058257e-01
-1.66431647e-02 7.45141283e-02 1.37751198e+00 3.07498008e-01
-7.85614610e-01 2.30052441e-01 -1.41883075e+00 9.13409948e-01
9.32672560e-01 -2.10311964e-01 -5.32474577e-01 -2.55040884e-01
5.56958020e-01 -2.99131900e-01 5.50742269e-01 -1.19327664e+00
1.30124055e-02 -8.62437785e-02 5.94080925e-01 -2.80078471e-01
2.54172534e-01 -1.92922592e-01 8.60626817e-01 2.48932257e-01
-5.17750978e-01 -1.47807643e-01 5.72590947e-01 -1.77298620e-01
3.88634689e-02 -4.65544522e-01 5.57296157e-01 -4.61888283e-01
-5.38623571e-01 -2.17666626e-01 -4.43217546e-01 3.38600390e-02
6.45361900e-01 -4.25697625e-01 -1.88635231e-03 -7.79634714e-01
-1.03295445e+00 -6.58024549e-02 7.73955047e-01 4.46535975e-01
2.77854890e-01 -1.41605842e+00 -3.92806321e-01 2.35057309e-01
2.67076075e-01 -1.43834576e-01 2.12296322e-01 5.17262936e-01
-5.89600503e-01 1.88096285e-01 -2.48821348e-01 4.40589041e-02
-1.15976882e+00 1.74789265e-01 2.93126971e-01 3.15503448e-01
-9.09634829e-01 1.04973471e+00 -4.04841036e-01 -6.61517262e-01
4.31718707e-01 4.08901572e-02 -2.50970453e-01 5.54782115e-02
7.02222764e-01 4.32981759e-01 -1.13953874e-01 -2.69693494e-01
3.03456131e-02 3.77292544e-01 3.27813864e-01 -5.96278787e-01
1.42821538e+00 2.61291355e-01 3.79894301e-02 9.49801922e-01
5.96848667e-01 1.25750840e-01 -7.86578596e-01 3.95034589e-02
-3.08393449e-01 -2.44834229e-01 1.85129046e-01 -1.30872250e+00
-8.47784758e-01 5.90958774e-01 4.95476663e-01 1.13879748e-01
1.12578654e+00 8.79602805e-02 5.33626318e-01 3.02833676e-01
1.78307265e-01 -1.20606005e+00 -3.07854265e-02 2.79959947e-01
1.25769556e+00 -4.85155970e-01 3.95710133e-02 1.63010061e-01
-1.03545868e+00 1.37876618e+00 2.82646358e-01 2.07684010e-01
6.88876629e-01 2.61986613e-01 5.17999291e-01 -7.06033483e-02
-8.15186501e-01 -9.81009603e-02 2.96837121e-01 8.51771533e-01
9.25335228e-01 1.52553506e-02 -1.78192630e-01 8.90813172e-01
-9.65996623e-01 3.77612442e-01 2.95072526e-01 9.56073105e-01
-2.38876387e-01 -1.46040976e+00 -4.36973751e-01 4.33958918e-01
-3.63570720e-01 -3.41298759e-01 -6.94563627e-01 6.24948323e-01
3.49081606e-01 8.20391774e-01 2.92813759e-02 -4.47845995e-01
1.73621565e-01 1.84872240e-01 5.19157946e-01 -6.31903410e-01
-9.64710712e-01 3.88128042e-01 5.54033279e-01 -3.89899611e-02
5.83582819e-02 -7.53066123e-01 -1.29610133e+00 -3.30402762e-01
-2.30308384e-01 1.32930741e-01 9.19874310e-01 5.74068427e-01
5.48803449e-01 5.37422717e-01 1.91667408e-01 -1.11249053e+00
1.17197465e-02 -1.01945102e+00 -8.21691930e-01 1.99470147e-02
-1.96465835e-01 -3.88931423e-01 -1.90768600e-01 2.60565996e-01] | [15.987963676452637, 5.464305400848389] |
70e7a5f1-0fda-4f28-9e10-ae3d3c764ff2 | target-free-text-guided-image-manipulation | 2211.14544 | null | https://arxiv.org/abs/2211.14544v2 | https://arxiv.org/pdf/2211.14544v2.pdf | Target-Free Text-guided Image Manipulation | We tackle the problem of target-free text-guided image manipulation, which requires one to modify the input reference image based on the given text instruction, while no ground truth target image is observed during training. To address this challenging task, we propose a Cyclic-Manipulation GAN (cManiGAN) in this paper, which is able to realize where and how to edit the image regions of interest. Specifically, the image editor in cManiGAN learns to identify and complete the input image, while cross-modal interpreter and reasoner are deployed to verify the semantic correctness of the output image based on the input instruction. While the former utilizes factual/counterfactual description learning for authenticating the image semantics, the latter predicts the "undo" instruction and provides pixel-level supervision for the training of cManiGAN. With such operational cycle-consistency, our cManiGAN can be trained in the above weakly supervised setting. We conduct extensive experiments on the datasets of CLEVR and COCO, and the effectiveness and generalizability of our proposed method can be successfully verified. Project page: https://sites.google.com/view/wancyuanfan/projects/cmanigan. | ['Yu-Chiang Frank Wang', 'Chiao-An Yang', 'Cheng-Fu Yang', 'Wan-Cyuan Fan'] | 2022-11-26 | null | null | null | null | ['image-manipulation'] | ['computer-vision'] | [ 6.38215780e-01 4.64970529e-01 -3.48902941e-01 -2.35855252e-01
-7.04653144e-01 -6.43669188e-01 7.49738634e-01 -5.05372941e-01
-3.11354995e-01 5.73998272e-01 -1.34051189e-01 -6.18636310e-01
3.20062727e-01 -8.30173135e-01 -1.37230361e+00 -7.18103945e-01
6.54567361e-01 2.88761169e-01 -8.06560665e-02 7.95076694e-03
2.13862970e-01 3.75199616e-02 -1.34513533e+00 3.17904592e-01
8.07055235e-01 6.93396688e-01 4.51818317e-01 8.05091143e-01
2.55942762e-01 1.12417924e+00 -5.30382335e-01 -5.23251891e-01
2.90797025e-01 -7.79448688e-01 -7.53510892e-01 4.21113461e-01
5.32823086e-01 -5.18622220e-01 -5.90975404e-01 1.45532501e+00
1.92240760e-01 -2.41282862e-02 2.93551743e-01 -1.62197995e+00
-8.79417539e-01 9.26936090e-01 -6.17675960e-01 -5.63737750e-02
2.11854711e-01 8.55489075e-01 8.47109079e-01 -5.09081125e-01
8.54305744e-01 1.02394795e+00 -1.99261293e-01 9.05296147e-01
-9.44612682e-01 -8.55616271e-01 2.77064145e-01 1.07592903e-01
-1.22807193e+00 -5.25145352e-01 7.36053467e-01 -3.50262880e-01
3.47891062e-01 2.15137586e-01 2.85448909e-01 1.24761963e+00
2.26865277e-01 8.39151442e-01 1.19136703e+00 -6.53703153e-01
2.79509798e-02 1.60441756e-01 -4.29361314e-01 1.09921277e+00
1.77099079e-01 3.32095146e-01 -4.57004189e-01 3.37714851e-01
7.32955515e-01 1.50686115e-01 -6.22577310e-01 -2.97999412e-01
-1.61125433e+00 4.21734899e-01 4.59707946e-01 1.89211607e-01
-1.40854865e-01 5.36831439e-01 3.46700519e-01 3.53465617e-01
4.24186327e-02 2.97009557e-01 -2.90888876e-01 3.74622904e-02
-8.16298664e-01 4.33585905e-02 4.44139808e-01 1.39828110e+00
8.00949514e-01 1.85084611e-01 -3.69758010e-01 -4.07016799e-02
1.77640200e-01 8.57345164e-01 3.42243940e-01 -1.10730636e+00
7.15885341e-01 5.46065331e-01 -7.06167594e-02 -6.34523153e-01
2.84267634e-01 -1.87988028e-01 -8.77626300e-01 2.63486683e-01
4.49059576e-01 -9.88395289e-02 -1.03837228e+00 1.78152478e+00
3.90579700e-01 2.32838407e-01 2.42133364e-01 1.00568461e+00
4.35189307e-01 7.33284473e-01 -1.60638228e-01 -1.57456622e-01
1.25872374e+00 -1.19058299e+00 -6.84927821e-01 -3.91757071e-01
7.69755244e-01 -4.07506287e-01 1.37262237e+00 2.25568250e-01
-1.08490407e+00 -4.42292094e-01 -1.13456130e+00 -1.33650303e-01
1.25965634e-02 4.57306117e-01 3.87836874e-01 3.88988763e-01
-8.52570474e-01 4.70823310e-02 -7.73499787e-01 3.97118852e-02
6.32581413e-01 1.96575701e-01 -2.90244848e-01 -5.11944175e-01
-9.62641597e-01 5.06125212e-01 7.22429872e-01 2.90300369e-01
-1.60087872e+00 -5.83647072e-01 -8.92847836e-01 6.00508787e-02
8.64179254e-01 -5.96003592e-01 1.24341357e+00 -1.47511113e+00
-1.56398976e+00 1.18277752e+00 -2.21480299e-02 -3.44274193e-01
9.00873542e-01 2.32623756e-01 -2.68009603e-01 2.60067105e-01
2.37808421e-01 7.90788889e-01 1.07448304e+00 -1.40982926e+00
-7.58129954e-01 -3.07317376e-01 3.86455387e-01 1.15204006e-01
8.84936154e-02 -2.40315318e-01 -8.91702950e-01 -4.75566477e-01
-2.55546153e-01 -1.08814085e+00 5.57562560e-02 2.97801532e-02
-7.51524746e-01 2.07205117e-01 7.51585782e-01 -7.45247424e-01
8.73087645e-01 -2.38207746e+00 1.06568187e-01 1.99155673e-01
2.07156911e-01 3.91546637e-02 -2.04768836e-01 9.91969034e-02
1.09188445e-01 1.45625383e-01 -4.64722306e-01 -1.80194080e-02
-7.12479814e-04 1.93029568e-01 -5.85857749e-01 5.85349560e-01
3.11583607e-03 1.31282353e+00 -8.97801876e-01 -6.64352298e-01
1.14709228e-01 1.26604244e-01 -4.26754594e-01 4.56053257e-01
-7.57453084e-01 7.45848477e-01 -4.75461513e-01 5.37515581e-01
5.47620952e-01 -3.20192933e-01 5.35457313e-01 -1.93100214e-01
2.52612442e-01 -8.72532800e-02 -7.70805955e-01 1.80002308e+00
-4.88841444e-01 7.23499715e-01 -2.60473583e-02 -9.96642053e-01
6.26717508e-01 1.93514585e-01 -1.94032326e-01 -9.92994487e-01
1.76941916e-01 1.06288418e-01 -7.35275596e-02 -6.65306151e-01
3.26371968e-01 9.65844318e-02 -2.01452374e-01 7.10617006e-01
-2.08554313e-01 -1.76567331e-01 1.39037166e-02 4.86519039e-01
7.40389407e-01 4.90847766e-01 2.63849616e-01 -7.15430453e-03
6.92767739e-01 1.18746810e-01 5.94122171e-01 7.48372972e-01
-1.89507511e-02 4.44158524e-01 7.42621839e-01 -1.04764231e-01
-1.05406308e+00 -7.18153298e-01 5.15434027e-01 8.78819346e-01
5.46909213e-01 -6.97829947e-02 -1.05432475e+00 -1.12910354e+00
-3.59355867e-01 1.04841650e+00 -6.24799669e-01 -1.69742376e-01
-6.20945096e-01 -5.08336201e-02 6.29415691e-01 2.60494292e-01
8.00751805e-01 -1.17091012e+00 -6.19566143e-01 -3.89289975e-01
-3.31356227e-01 -1.17928946e+00 -9.68712687e-01 -1.86970979e-01
-5.05966246e-01 -1.32077849e+00 -2.18967199e-01 -8.58313382e-01
1.25076258e+00 3.15952033e-01 7.93452501e-01 5.15077949e-01
2.56139040e-02 4.26440954e-01 -1.36372775e-01 -2.44458526e-01
-7.91645467e-01 -4.09321710e-02 -4.50831443e-01 1.04334943e-01
-8.46869946e-02 -1.69354290e-01 -5.71735024e-01 3.60834837e-01
-1.30540073e+00 5.91660857e-01 7.23987520e-01 1.00768876e+00
8.87131333e-01 1.99942514e-01 1.47391513e-01 -1.46887541e+00
1.38345495e-01 -1.38699085e-01 -9.93535995e-01 5.77853739e-01
-7.70507395e-01 2.32576787e-01 7.11189985e-01 -4.12701607e-01
-1.27066433e+00 1.30763382e-01 2.76012242e-01 -7.36483216e-01
-6.05510585e-02 2.88115114e-01 -7.99338937e-01 1.75615177e-01
4.34477925e-01 5.67660391e-01 4.66585532e-03 1.50646165e-01
5.75788438e-01 6.21713400e-01 9.53273714e-01 -8.51675034e-01
9.79894936e-01 6.74003243e-01 -1.77173540e-01 -1.99557200e-01
-7.91221261e-01 9.63689014e-02 -4.48236793e-01 -2.49285460e-01
9.82210577e-01 -1.06427538e+00 -8.20093155e-01 5.06184161e-01
-1.19790530e+00 -8.85533631e-01 -3.41247618e-01 9.79323462e-02
-7.16266930e-01 2.07007915e-01 -3.08979124e-01 -4.12961423e-01
-2.36489609e-01 -1.43652117e+00 1.01655149e+00 1.78989336e-01
3.96165282e-01 -8.83712053e-01 -4.68221426e-01 6.71344161e-01
-1.63858011e-02 3.21682006e-01 8.49393249e-01 -2.50672042e-01
-1.28646839e+00 -7.04760551e-02 -2.38698646e-01 3.61493319e-01
-2.21586786e-02 3.71791385e-02 -8.13282430e-01 -4.13009256e-01
3.03344522e-02 -3.61086071e-01 5.97452939e-01 6.97160140e-02
1.41511559e+00 -6.06276333e-01 -1.49080589e-01 7.30104029e-01
1.42715514e+00 2.10794762e-01 8.78176630e-01 3.65771472e-01
9.67224121e-01 3.37872684e-01 9.63913143e-01 3.78717780e-02
5.72015464e-01 5.26356161e-01 7.22780228e-01 1.21684177e-02
-2.76881456e-01 -7.47709394e-01 6.95277035e-01 4.22922462e-01
2.39900097e-01 -5.11235595e-01 -6.26270354e-01 4.12050992e-01
-1.82984459e+00 -8.84645820e-01 1.63518917e-02 1.92181468e+00
9.78226185e-01 -3.06584723e-02 -4.50711131e-01 -2.15726554e-01
9.70011055e-01 3.41050208e-01 -8.90182614e-01 -1.20930195e-01
-1.85436886e-02 -1.73836991e-01 7.45451212e-01 4.36506778e-01
-6.80469453e-01 1.10952950e+00 4.44711161e+00 8.69311392e-01
-1.35519505e+00 4.49697971e-01 8.70154977e-01 4.79211062e-02
-4.86277550e-01 4.44539547e-01 -4.72474754e-01 5.83166778e-01
6.81022763e-01 -3.58444542e-01 8.31432402e-01 7.08923340e-01
2.44260460e-01 -2.31114000e-01 -1.28888512e+00 8.58907223e-01
2.63666153e-01 -1.45349908e+00 1.62848666e-01 5.35101816e-02
1.07394803e+00 -3.50625068e-01 2.02266708e-01 1.37277797e-01
3.14908177e-01 -9.12755251e-01 1.10870659e+00 4.93234962e-01
1.50289285e+00 -5.65552533e-01 4.05538231e-01 4.58273172e-01
-7.24281430e-01 6.25122190e-02 -1.22700326e-01 2.25395545e-01
-1.61424235e-01 3.94326687e-01 -6.95855498e-01 5.58471322e-01
3.51351529e-01 6.44578338e-01 -4.77226615e-01 2.21387893e-01
-1.00432885e+00 6.67529702e-01 1.56795189e-01 3.38418007e-01
5.73589699e-03 -2.19692051e-01 3.55192006e-01 8.23141038e-01
2.98652440e-01 1.03272207e-01 2.74599474e-02 1.23203850e+00
-6.92836046e-01 -3.68906677e-01 -5.42139709e-01 -1.30968586e-01
4.18576241e-01 1.10196996e+00 -6.22201264e-01 -4.35115159e-01
-2.06646994e-01 1.33776724e+00 9.99689549e-02 5.82953990e-01
-1.23743105e+00 -2.52047956e-01 2.70695210e-01 7.35520348e-02
4.32996005e-01 4.42006849e-02 -2.20926151e-01 -1.43011558e+00
2.85477877e-01 -1.30787778e+00 4.04583603e-01 -1.13661122e+00
-5.99593937e-01 3.48832160e-01 -1.31577358e-01 -1.17140996e+00
-1.50128856e-01 -4.53470767e-01 -8.32072139e-01 6.04960263e-01
-1.56726503e+00 -1.48386550e+00 -4.99055475e-01 7.00460196e-01
4.43006247e-01 -3.99961025e-02 3.29501063e-01 9.56728905e-02
-7.06000626e-01 8.14163864e-01 -1.76301375e-01 3.16846281e-01
6.88875616e-01 -1.08493721e+00 3.21028829e-01 1.33360827e+00
6.26252145e-02 4.00193334e-01 5.16786873e-01 -6.61713362e-01
-1.81937337e+00 -1.58351398e+00 4.31353241e-01 -3.69075656e-01
5.87127209e-01 -4.67281431e-01 -6.48049116e-01 1.26471722e+00
3.44129711e-01 1.67998254e-01 -4.23352830e-02 -6.99019074e-01
-5.44281125e-01 -1.38461858e-01 -1.00060976e+00 7.56915987e-01
1.15896964e+00 -7.56972909e-01 -2.76400298e-01 4.54620689e-01
1.06649542e+00 -8.81334782e-01 -5.42963803e-01 -7.45556653e-02
1.35608688e-01 -7.13466644e-01 5.34127772e-01 -4.45352256e-01
9.64851022e-01 -7.42009938e-01 -1.03938848e-01 -1.11948729e+00
3.49745631e-01 -7.47120202e-01 1.61590632e-02 1.39310145e+00
4.08807248e-01 -6.53843284e-01 6.13786876e-01 4.53935176e-01
-1.64876789e-01 -3.49796772e-01 -7.84647644e-01 -5.64297378e-01
-3.97838727e-02 -4.92384821e-01 8.74466598e-01 9.25920010e-01
-2.62015313e-01 7.37664774e-02 -6.04075193e-01 4.18546945e-01
5.34360588e-01 4.63929266e-01 1.09641898e+00 -2.46345967e-01
-5.33478558e-01 -1.14266731e-01 -3.11930627e-01 -1.09354830e+00
5.28010428e-01 -1.08896160e+00 3.23250562e-01 -1.12115800e+00
4.15822208e-01 -2.47721568e-01 3.72074656e-02 6.70491338e-01
-1.88283384e-01 1.15414217e-01 3.04196328e-01 2.74123430e-01
-8.10632467e-01 4.05435145e-01 1.73548412e+00 -6.09452367e-01
2.10447535e-01 -2.36826256e-01 -8.67433071e-01 1.98186815e-01
7.02276886e-01 -4.82965112e-01 -5.47140241e-01 -7.29428649e-01
3.32724661e-01 2.95475066e-01 6.60741925e-01 -4.99234349e-01
3.44970584e-01 -4.61019933e-01 -2.34184427e-05 -9.45334509e-02
-3.06421816e-01 -9.31002557e-01 3.54970872e-01 5.61071396e-01
-6.84861839e-01 -1.09492287e-01 -8.63873661e-02 6.35586798e-01
-2.09080547e-01 -1.28706411e-01 7.71500647e-01 -7.54068792e-02
-8.89767051e-01 4.53649938e-01 5.99886626e-02 1.68744206e-01
1.30409384e+00 4.30875346e-02 -6.87860489e-01 -3.56936306e-01
-1.66607156e-01 3.52086067e-01 9.59136903e-01 2.88757384e-01
7.27799892e-01 -1.18925738e+00 -5.83531320e-01 2.47516289e-01
4.30887759e-01 3.85605365e-01 4.09214199e-01 8.38938832e-01
-5.78215301e-01 1.69769645e-01 -7.08586276e-02 -4.97383505e-01
-1.04032981e+00 8.74188840e-01 5.24462581e-01 -1.49505913e-01
-6.37681663e-01 4.16625977e-01 6.06383324e-01 -4.12349999e-01
7.71890283e-02 -2.40088150e-01 4.40985948e-01 -5.92469573e-01
4.20347095e-01 -1.02998808e-01 -2.48427659e-01 -4.33058202e-01
-8.12040642e-02 7.07139671e-02 -1.58131525e-01 -2.60646105e-01
1.04520154e+00 -3.36516023e-01 -3.13333541e-01 9.07744393e-02
1.09474611e+00 -6.72248378e-03 -1.50002003e+00 -2.21245289e-01
-3.54517162e-01 -6.49655461e-01 1.02615282e-01 -9.37560916e-01
-1.53136885e+00 8.76755238e-01 3.89149338e-01 -4.80896384e-01
1.43916619e+00 -1.86837185e-02 6.18003964e-01 3.52893084e-01
2.34386340e-01 -8.72464359e-01 1.36310533e-02 1.33489206e-01
7.36317337e-01 -1.25920022e+00 -9.49378759e-02 -2.62158811e-01
-1.02103078e+00 8.20383251e-01 8.76632750e-01 2.04219997e-01
2.37386655e-02 9.69316810e-02 -6.21317774e-02 -3.01140189e-01
-8.31600070e-01 1.46262139e-01 -7.70659521e-02 4.89636689e-01
-7.50530735e-02 2.66905785e-01 3.41632403e-02 2.80642867e-01
6.75136503e-03 1.10434167e-01 8.56350780e-01 8.91233861e-01
1.49439320e-01 -8.52625668e-01 -2.99080133e-01 2.07985297e-01
-2.41891086e-01 -1.53332308e-01 -2.38516495e-01 8.08252037e-01
1.43924400e-01 6.99760616e-01 2.15009358e-02 -2.75107205e-01
1.70120463e-01 -2.27921113e-01 5.08174539e-01 -5.08342803e-01
-3.31825614e-01 -2.34119877e-01 -1.48897737e-01 -5.84669769e-01
-4.71927255e-01 -5.03153503e-01 -1.45777977e+00 -3.32075626e-01
-2.36195862e-01 1.14721023e-01 7.14696229e-01 9.40884292e-01
4.06794429e-01 5.83346486e-01 8.40829015e-01 -3.80334675e-01
-4.12834585e-01 -5.48074365e-01 -4.41622347e-01 5.47679067e-01
2.54991233e-01 -1.19434766e-01 -3.62203121e-01 7.21175313e-01] | [11.660076141357422, -0.33424249291419983] |
9e9bfa60-8511-47ec-beb9-cf56e13d40bb | compositional-sequence-labeling-models-for | 1607.06153 | null | http://arxiv.org/abs/1607.06153v1 | http://arxiv.org/pdf/1607.06153v1.pdf | Compositional Sequence Labeling Models for Error Detection in Learner Writing | In this paper, we present the first experiments using neural network models
for the task of error detection in learner writing. We perform a systematic
comparison of alternative compositional architectures and propose a framework
for error detection based on bidirectional LSTMs. Experiments on the CoNLL-14
shared task dataset show the model is able to outperform other participants on
detecting errors in learner writing. Finally, the model is integrated with a
publicly deployed self-assessment system, leading to performance comparable to
human annotators. | ['Helen Yannakoudakis', 'Marek Rei'] | 2016-07-20 | compositional-sequence-labeling-models-for-1 | https://aclanthology.org/P16-1112 | https://aclanthology.org/P16-1112.pdf | acl-2016-8 | ['grammatical-error-detection'] | ['natural-language-processing'] | [ 2.48371974e-01 2.03742087e-01 1.48488179e-01 -3.85568142e-01
-6.65040493e-01 -1.24147840e-01 4.54992563e-01 4.40424234e-01
-9.06700909e-01 6.39688492e-01 3.66298914e-01 -4.67195481e-01
-7.00777322e-02 -2.89557666e-01 -5.19746184e-01 4.18898493e-01
5.47227979e-01 5.86961448e-01 -6.72593042e-02 -2.57707506e-01
5.67665279e-01 1.56484842e-01 -1.23912466e+00 7.61366069e-01
1.10960639e+00 8.20864439e-01 3.48673128e-02 9.58599806e-01
-1.14739127e-01 1.48373449e+00 -1.10819328e+00 -9.39510405e-01
-4.77376848e-01 -5.54231286e-01 -1.48273754e+00 -5.30371547e-01
1.13559794e+00 -4.80434120e-01 -8.86362940e-02 9.14304376e-01
5.58575928e-01 2.98904836e-01 4.29922700e-01 -7.29374170e-01
-1.22695673e+00 1.02630055e+00 4.10768121e-01 4.60910201e-01
6.24705672e-01 -1.13557391e-01 9.96229649e-01 -1.24674773e+00
6.19190454e-01 9.62706923e-01 1.13954246e+00 1.06393111e+00
-1.16664135e+00 -5.86607695e-01 2.86583424e-01 5.22370577e-01
-1.02469623e+00 -8.18431318e-01 4.99324709e-01 -3.59620780e-01
1.68783283e+00 3.32990229e-01 5.20183980e-01 1.67982173e+00
-9.08837393e-02 1.07120943e+00 1.29102933e+00 -9.11002159e-01
-1.35669559e-01 2.26349235e-01 6.57227814e-01 1.04655266e+00
3.32555734e-02 3.23909633e-02 -1.50563014e+00 2.03761116e-01
7.76575655e-02 -3.70565921e-01 -5.48713207e-01 3.84907007e-01
-8.70047808e-01 4.32415634e-01 3.77839446e-01 6.00342631e-01
4.97004315e-02 3.42694968e-01 6.58537865e-01 7.82871723e-01
7.12674260e-01 6.39512241e-01 -3.10774595e-01 -5.57774544e-01
-1.15872312e+00 2.06601664e-01 1.06038094e+00 1.00819433e+00
-1.04532868e-01 2.14033365e-01 -6.11059248e-01 1.18410647e+00
5.55630326e-01 -5.30191604e-03 1.19872928e+00 -7.65527546e-01
7.36024022e-01 5.22377253e-01 4.18284489e-03 -6.60278797e-01
-4.49110150e-01 -4.60520387e-01 -5.42583525e-01 1.75778970e-01
4.66526806e-01 1.11677721e-01 -5.68341613e-01 1.61438334e+00
-6.58808351e-01 -2.30241850e-01 -1.24527968e-01 5.43639600e-01
1.16268778e+00 1.84645429e-02 2.78225988e-01 -9.35061425e-02
8.22169423e-01 -1.37343955e+00 -1.28009748e+00 -2.52021402e-01
1.08936882e+00 -6.52918279e-01 1.18983114e+00 6.64477050e-01
-1.62523198e+00 -8.00416529e-01 -1.15338337e+00 -4.30543095e-01
-4.12544489e-01 5.69521546e-01 6.71955794e-02 8.04218054e-01
-1.25638580e+00 9.27203894e-01 -7.47033417e-01 -4.86815631e-01
4.41659421e-01 3.33643742e-02 -3.07366788e-01 3.04793745e-01
-1.12806714e+00 1.53187907e+00 3.23679715e-01 2.80033022e-01
-8.90509784e-01 -6.29584432e-01 -8.05729747e-01 -1.62123144e-01
-3.41309994e-01 -5.00484705e-01 2.03121471e+00 -1.12018490e+00
-1.64083970e+00 1.28705871e+00 -5.04868031e-01 -7.13254333e-01
7.63727665e-01 -6.32582903e-01 -7.12343514e-01 -4.09560263e-01
-8.87307450e-02 3.10578436e-01 4.31943536e-01 -7.37113774e-01
-6.39894605e-01 -2.57306039e-01 -3.67214352e-01 1.50528446e-01
-9.45368230e-01 3.99337381e-01 2.89853215e-01 -8.59583616e-01
-1.36012688e-01 -5.71937978e-01 4.82478589e-01 -8.92028138e-02
-3.50194335e-01 -8.81914437e-01 2.93474406e-01 -1.11290681e+00
1.79366100e+00 -1.75997305e+00 1.96815923e-01 -1.39270470e-01
2.31741905e-01 5.74575245e-01 -5.39123416e-02 2.98304260e-01
-7.10783377e-02 4.84513730e-01 -2.75013775e-01 -1.03222132e+00
2.14652017e-01 -2.33534604e-01 -3.27432126e-01 1.07860759e-01
1.67726696e-01 9.58790660e-01 -8.16897750e-01 -2.45194614e-01
-2.00495362e-01 1.05224207e-01 -1.41281009e-01 4.60021883e-01
-3.58274840e-02 -5.51939122e-02 3.55538130e-01 4.28391665e-01
8.75114352e-02 1.12275807e-02 3.56397331e-02 3.35072041e-01
-9.53843817e-02 1.28995633e+00 -7.14768231e-01 2.07221484e+00
-6.50839508e-01 1.15702057e+00 -3.58238697e-01 -3.71648163e-01
8.95071924e-01 6.26962304e-01 -7.08170414e-01 -9.32452917e-01
-1.41228065e-01 4.95685548e-01 2.00808033e-01 -5.07775128e-01
6.39728248e-01 8.36650282e-02 2.50645041e-01 9.10543799e-01
6.36116564e-01 7.75184259e-02 4.11673099e-01 1.61656216e-01
1.30184376e+00 2.71250010e-01 1.28189161e-01 -1.48285657e-01
8.29371452e-01 -2.64869899e-01 -4.59869988e-02 1.09150374e+00
-4.85420614e-01 3.84097576e-01 -2.06756741e-02 -6.34940863e-01
-7.23884106e-01 -7.40380049e-01 -1.06359839e-01 1.29474437e+00
-5.47768891e-01 -7.95515895e-01 -7.44109809e-01 -9.65284228e-01
-1.13726497e-01 1.07944405e+00 -7.84260988e-01 -3.04715484e-01
-5.38563192e-01 -1.86931603e-02 1.20133686e+00 8.55356038e-01
5.21334112e-01 -1.50873268e+00 -5.39456010e-01 4.02535796e-01
-3.33279312e-01 -8.27677667e-01 -1.91419840e-01 4.33638662e-01
-9.18537378e-01 -8.43389153e-01 -3.53393674e-01 -9.64035034e-01
3.27027529e-01 -3.75563264e-01 1.46038246e+00 7.96567023e-01
-5.57068884e-02 3.20090353e-01 -2.00126573e-01 -7.35274255e-01
-5.55257857e-01 2.90856004e-01 -7.77156949e-02 -8.23477268e-01
7.88157165e-01 -1.34239420e-01 2.28532329e-02 -1.72113746e-01
-3.81506175e-01 -7.07975030e-03 1.82227373e-01 1.05523264e+00
-1.12551659e-01 -4.85938936e-01 4.34157908e-01 -7.80363977e-01
1.48745537e+00 1.15113810e-01 -6.52801618e-02 4.44327056e-01
-1.06705976e+00 1.58921815e-02 4.32351679e-01 -1.22021399e-01
-1.19477141e+00 -5.18145800e-01 -4.61485416e-01 2.09292009e-01
-3.23416561e-01 3.26087594e-01 4.92790431e-01 -3.02805185e-01
8.97944689e-01 6.67863190e-02 -2.18099713e-01 -6.47623241e-01
-6.45737126e-02 7.29840040e-01 4.05690640e-01 -6.23099387e-01
-2.78107543e-02 -2.85436928e-01 -6.03366554e-01 -4.47965771e-01
-1.23853707e+00 6.51733652e-02 -6.06991410e-01 -3.05326611e-01
4.05863792e-01 -8.96512687e-01 -7.25255370e-01 6.66869581e-01
-1.63519526e+00 -6.31734431e-01 -3.31318706e-01 3.27811658e-01
-4.08863783e-01 3.38479467e-02 -1.28035796e+00 -8.84936750e-01
-5.86176634e-01 -9.15329933e-01 4.72999662e-01 -1.16484659e-03
-1.15082037e+00 -1.29714143e+00 3.61670196e-01 4.78876024e-01
6.37983143e-01 -5.75320780e-01 7.33636975e-01 -1.13327265e+00
1.25019988e-02 7.73809627e-02 -1.00097500e-01 6.49545491e-01
-6.62434042e-01 -3.58845145e-02 -1.15640819e+00 6.48818389e-02
-4.54159044e-02 -9.59061384e-01 1.24486673e+00 -1.53061241e-01
1.39449143e+00 -3.71979415e-01 7.34088793e-02 5.11614144e-01
8.74061108e-01 -1.28585011e-01 2.26927996e-01 6.59636497e-01
7.28558898e-01 5.49706221e-01 6.01701140e-02 1.33389160e-01
5.75385749e-01 5.12844265e-01 2.66250549e-03 5.20906448e-01
-5.73400021e-01 -6.16731942e-01 4.82121885e-01 1.13873386e+00
-2.25642286e-02 -5.24210334e-01 -1.14894176e+00 6.90705419e-01
-1.73243749e+00 -8.40553820e-01 -5.49450338e-01 1.90662301e+00
1.18245018e+00 6.26050830e-02 -1.42446831e-01 3.88386160e-01
4.91607189e-01 -7.43482560e-02 -7.04455301e-02 -1.28146327e+00
2.01676190e-02 8.34272861e-01 1.08793139e-01 8.18766713e-01
-8.21265578e-01 1.07808769e+00 7.90198231e+00 4.09726202e-01
-6.78997993e-01 6.05186403e-01 2.24280030e-01 -2.75783658e-01
-1.80449918e-01 -4.59896952e-01 -9.92096782e-01 6.20544851e-01
1.43378389e+00 1.30939886e-01 2.68273205e-01 3.86642694e-01
-2.96452433e-01 7.07814544e-02 -1.41010809e+00 6.96826756e-01
5.42981505e-01 -1.00355446e+00 -6.74894378e-02 -5.73182285e-01
7.05688953e-01 1.06458701e-01 -1.52288631e-01 5.45404375e-01
5.85526586e-01 -1.41407406e+00 1.14485526e+00 9.02237713e-01
5.06873131e-01 -6.05162084e-01 7.81101763e-01 3.69562358e-01
-1.62320286e-01 -4.54387963e-02 -1.50833324e-01 -7.50297248e-01
-1.24539763e-01 2.16128141e-01 -7.27352560e-01 -4.11106870e-02
9.76226032e-01 8.24026823e-01 -1.01766157e+00 1.01158714e+00
-9.15580571e-01 8.89219761e-01 1.57821074e-01 -4.18005854e-01
-1.61115900e-01 3.98310602e-01 3.27225834e-01 1.61489749e+00
4.40390676e-01 -4.35840577e-01 -3.33072454e-01 1.19422102e+00
-7.24022865e-01 1.19589962e-01 -5.67899406e-01 -5.21703549e-02
5.16140521e-01 9.62449849e-01 1.21727847e-01 -4.66338396e-01
-2.03346893e-01 1.38895118e+00 1.28852677e+00 2.23212510e-01
-2.03660890e-01 -5.31019628e-01 2.49071255e-01 -4.43541378e-01
-1.09458916e-01 -1.24860644e-01 -7.72575378e-01 -9.37693238e-01
3.25670421e-01 -1.06785893e+00 4.01979417e-01 -9.99829888e-01
-1.41129160e+00 7.33475566e-01 -7.25995302e-01 -6.12757385e-01
-2.79813081e-01 -1.00308931e+00 -6.09439015e-01 1.23230934e+00
-1.43845618e+00 -8.81159544e-01 -4.69322681e-01 3.51215512e-01
6.80691481e-01 -4.86078173e-01 1.27316284e+00 2.56463081e-01
-8.91906738e-01 9.95869219e-01 -2.23504663e-01 2.72368520e-01
1.06233728e+00 -1.60403836e+00 5.44358909e-01 1.02569377e+00
3.80947113e-01 8.54189575e-01 3.78873408e-01 -6.69066072e-01
-4.74782348e-01 -7.56091774e-01 1.76607037e+00 -9.63220537e-01
6.95542991e-01 -1.50683641e-01 -1.05007207e+00 1.10206795e+00
7.52296627e-01 -1.22948430e-01 9.80263829e-01 5.45303524e-01
-5.77341378e-01 1.91325426e-01 -8.61163259e-01 2.42707118e-01
1.34120739e+00 -1.04345620e+00 -1.07546234e+00 5.16159654e-01
4.64837104e-01 -7.05083549e-01 -8.59020650e-01 3.58662158e-01
5.91344237e-01 -1.14193416e+00 4.20447439e-01 -9.99225140e-01
1.02480781e+00 8.85513276e-02 4.19871181e-01 -1.72144055e+00
-3.61677051e-01 -4.23696637e-01 -6.77909553e-01 1.34399569e+00
6.66097343e-01 -2.48120666e-01 2.89011866e-01 7.06104517e-01
-4.40398306e-01 -3.73399347e-01 -1.01345766e+00 -6.33666694e-01
3.13611090e-01 -7.93461442e-01 8.43039900e-02 9.24904466e-01
3.98011893e-01 2.70166010e-01 -1.17539652e-01 -5.17145216e-01
1.53217092e-01 -6.25683367e-01 1.95042685e-01 -1.29580104e+00
6.22884370e-04 -1.08892977e+00 -1.03144370e-01 -8.02599669e-01
6.98711157e-01 -1.29065263e+00 2.49674276e-01 -1.57669997e+00
1.59728974e-02 -1.16692036e-01 -4.88531500e-01 5.47224998e-01
-2.92346358e-01 4.20424104e-01 1.24215119e-01 7.71164894e-02
-8.74152362e-01 3.72990310e-01 5.63543618e-01 -2.06829220e-01
1.64858058e-01 -2.53371477e-01 -6.92409217e-01 7.76524603e-01
7.58503497e-01 -5.35705090e-01 2.10727036e-01 -1.15361547e+00
7.57957876e-01 -4.44937348e-01 3.07349235e-01 -1.39428365e+00
5.29811740e-01 6.89175189e-01 6.95426404e-01 -3.82936835e-01
-8.63497779e-02 -5.95924795e-01 -5.34494996e-01 5.10666132e-01
-1.30334413e+00 4.74741250e-01 4.70347345e-01 4.30008732e-02
-3.27587724e-01 -9.82462883e-01 3.73687118e-01 -1.64713800e-01
-4.78342146e-01 -4.72522974e-01 -7.16946363e-01 2.01110721e-01
4.62079734e-01 2.74513245e-01 -6.78110838e-01 -1.96735680e-01
-8.05538237e-01 1.29641280e-01 1.30304694e-01 6.39717281e-01
7.22913325e-01 -1.47865796e+00 -7.09531367e-01 2.81958044e-01
5.27116299e-01 -4.87662166e-01 -1.13526225e-01 7.76660085e-01
-5.83649635e-01 6.62668526e-01 -1.70690998e-01 -1.97571933e-01
-1.47426951e+00 -6.32091016e-02 6.82846367e-01 -4.07558322e-01
-2.96667308e-01 1.48850965e+00 -8.17880154e-01 -7.29512095e-01
7.63282001e-01 -3.97048205e-01 -2.12927550e-01 -2.01088060e-02
9.37086463e-01 7.06234813e-01 6.82820857e-01 -3.29200238e-01
-1.48423746e-01 -1.28509879e-01 -3.52487773e-01 -3.49408805e-01
1.03638422e+00 -4.73026745e-02 -3.13823462e-01 1.14085174e+00
7.97795177e-01 -4.23872732e-02 -4.06514853e-01 -4.27054733e-01
6.57301128e-01 -2.26824388e-01 8.49129260e-02 -1.60249949e+00
-4.56678897e-01 1.19478083e+00 7.16828108e-01 4.25711349e-02
5.86379051e-01 -5.50737798e-01 5.72851956e-01 1.00211477e+00
-1.36571392e-01 -1.70293736e+00 1.59017831e-01 1.15671337e+00
1.10393357e+00 -1.28871882e+00 -2.17971414e-01 3.47262621e-01
-4.60844845e-01 1.51512694e+00 1.15133345e+00 -3.65548171e-02
8.62409174e-03 1.34463593e-01 9.87608582e-02 -2.02982768e-01
-1.05597758e+00 -2.54901480e-02 7.98641443e-01 3.62977564e-01
1.58943772e+00 1.15709864e-02 -6.10746861e-01 6.88175917e-01
-4.13394570e-01 2.05635190e-01 5.27958333e-01 8.26710582e-01
-2.17150867e-01 -1.14601171e+00 -1.17953882e-01 5.00907838e-01
-7.13069975e-01 -4.21422064e-01 -1.00309873e+00 4.96776700e-01
5.63801564e-02 1.02733195e+00 2.31153026e-01 -2.06548378e-01
6.31405592e-01 1.01468813e+00 7.88247228e-01 -7.01489329e-01
-1.38100231e+00 -8.91189098e-01 5.57801902e-01 -6.73068166e-01
-3.05528790e-01 -5.84853649e-01 -8.99169207e-01 -7.51739666e-02
-3.79000604e-01 -1.50835007e-01 7.53348053e-01 1.13084829e+00
1.53645739e-01 9.27509546e-01 -3.00666094e-01 -1.92976505e-01
-5.67658782e-01 -1.67109644e+00 -1.60306484e-01 5.41803241e-01
3.92203748e-01 -2.03101978e-01 -3.53075057e-01 -1.34544343e-01] | [11.234726905822754, 9.417295455932617] |
0469325b-d5e0-4594-bf0c-b1517faf9f52 | jitter-does-matter-adapting-gaze-estimation | 2210.02082 | null | https://arxiv.org/abs/2210.02082v1 | https://arxiv.org/pdf/2210.02082v1.pdf | Jitter Does Matter: Adapting Gaze Estimation to New Domains | Deep neural networks have demonstrated superior performance on appearance-based gaze estimation tasks. However, due to variations in person, illuminations, and background, performance degrades dramatically when applying the model to a new domain. In this paper, we discover an interesting gaze jitter phenomenon in cross-domain gaze estimation, i.e., the gaze predictions of two similar images can be severely deviated in target domain. This is closely related to cross-domain gaze estimation tasks, but surprisingly, it has not been noticed yet previously. Therefore, we innovatively propose to utilize the gaze jitter to analyze and optimize the gaze domain adaptation task. We find that the high-frequency component (HFC) is an important factor that leads to jitter. Based on this discovery, we add high-frequency components to input images using the adversarial attack and employ contrastive learning to encourage the model to obtain similar representations between original and perturbed data, which reduces the impacts of HFC. We evaluate the proposed method on four cross-domain gaze estimation tasks, and experimental results demonstrate that it significantly reduces the gaze jitter and improves the gaze estimation performance in target domains. | ['Feng Lu', 'Yunfei Liu', 'Haofei Wang', 'Mingjie Xu', 'Yiwei Bao', 'Ruicong Liu'] | 2022-10-05 | null | null | null | null | ['gaze-estimation'] | ['computer-vision'] | [ 2.93495536e-01 -2.94555902e-01 1.66351199e-01 -3.99133772e-01
-1.37968376e-01 -3.91808569e-01 1.76307008e-01 -5.03616273e-01
-1.40339255e-01 7.37861633e-01 -1.30966812e-01 1.44470096e-01
-1.17314406e-01 5.17516804e-04 -8.68967414e-01 -8.75990033e-01
2.14842856e-01 -5.49218774e-01 1.96861386e-01 -1.10423580e-01
6.90164924e-01 6.14612103e-02 -1.83335805e+00 -1.61177859e-01
1.22148418e+00 1.08998394e+00 4.23520952e-02 2.92343348e-01
1.54112682e-01 4.40986425e-01 -1.06482673e+00 -3.48692417e-01
3.95768397e-02 -6.65592551e-01 -4.58033860e-01 -1.39960989e-01
7.22241759e-01 -2.67720401e-01 1.74763501e-01 1.34670162e+00
6.59202337e-01 1.82724774e-01 6.91062450e-01 -1.54540694e+00
-1.01387537e+00 -1.39993235e-01 -1.24679756e+00 4.29083049e-01
4.42005903e-01 2.99588412e-01 3.91484380e-01 -6.32955194e-01
3.66194129e-01 1.25180137e+00 6.25037014e-01 8.74036133e-01
-1.18253648e+00 -1.27884018e+00 4.32918936e-01 5.45867383e-01
-1.33280611e+00 -5.90162516e-01 1.13848507e+00 -3.05322707e-01
2.67165095e-01 6.90357089e-02 3.43012363e-01 1.43331182e+00
4.09699947e-01 7.33104229e-01 1.43384540e+00 -4.95300591e-01
-1.39519379e-01 3.15918863e-01 -1.72019396e-02 2.54550666e-01
1.09456114e-01 3.67870241e-01 -7.64007986e-01 2.20062301e-01
3.51722658e-01 -1.27314657e-01 -7.87919700e-01 -1.81747630e-01
-7.39018679e-01 3.40411395e-01 6.37351334e-01 2.98340525e-02
-8.38300884e-02 -9.23348665e-02 7.24705309e-02 2.50088364e-01
5.73484302e-01 3.64116311e-01 -2.16181323e-01 -1.43196970e-01
-6.62650108e-01 6.25214279e-02 2.95739681e-01 8.29829991e-01
6.63400888e-01 8.03746656e-02 -2.03221440e-01 7.48932242e-01
4.28631127e-01 7.37385690e-01 6.60699546e-01 -6.31139636e-01
2.04356775e-01 3.66112918e-01 1.93892524e-01 -1.10550070e+00
-3.79111677e-01 -3.20516646e-01 -6.06475592e-01 5.82134187e-01
6.46500647e-01 -2.62149066e-01 -6.61830783e-01 2.19887781e+00
4.57600504e-01 3.63425136e-01 -1.51750892e-01 1.33047962e+00
7.39649832e-01 2.04725847e-01 1.96344212e-01 -4.87171561e-01
1.27464116e+00 -6.36803329e-01 -1.11826289e+00 -1.13349944e-01
2.38374807e-02 -9.10423040e-01 1.41230059e+00 2.46919930e-01
-8.53883862e-01 -7.49928415e-01 -1.09883928e+00 1.59404114e-01
-1.08658157e-01 -7.27728233e-02 1.29369631e-01 8.07907104e-01
-8.45919013e-01 4.79823798e-01 -4.30822998e-01 -4.48041975e-01
4.86460984e-01 3.29522759e-01 -1.00446545e-01 2.47621134e-01
-1.23920536e+00 8.28124344e-01 3.31777483e-02 1.35868743e-01
-3.39439988e-01 -8.56806695e-01 -5.95706820e-01 -6.84401914e-02
2.98242599e-01 -5.31294823e-01 1.24805307e+00 -1.69181347e+00
-1.68967509e+00 6.18125558e-01 -6.42760336e-01 -1.32214660e-02
2.79746145e-01 -2.14121386e-01 -8.15597296e-01 -1.08400509e-01
-1.22883745e-01 5.81733942e-01 1.60583019e+00 -1.37654245e+00
-5.05789399e-01 -5.88025153e-01 -3.37413177e-02 2.31437013e-01
-4.26395684e-01 1.27859205e-01 -2.16648147e-01 -4.94334370e-01
-3.05893034e-01 -9.95367646e-01 7.05319226e-01 1.40635092e-02
-3.37461144e-01 -2.32490808e-01 1.12553108e+00 -4.93862063e-01
1.38783228e+00 -2.33180261e+00 -1.16828702e-01 -9.27283429e-03
4.60030884e-01 2.97251821e-01 1.27299251e-02 -1.66114002e-01
-2.99024373e-01 6.68449402e-02 8.36073384e-02 -2.52147019e-01
-9.73252729e-02 -1.74716577e-01 -2.88725883e-01 3.55710268e-01
4.45258707e-01 7.32297122e-01 -6.88097894e-01 -4.05025274e-01
-1.70645118e-01 4.34589356e-01 -2.14724690e-01 2.66887456e-01
5.22658937e-02 6.57927513e-01 -2.16654196e-01 5.83756268e-01
1.19572055e+00 -3.68905306e-01 -2.52737939e-01 -4.48287427e-01
-3.38109746e-03 -1.39666945e-01 -6.13573313e-01 1.34076464e+00
-2.48318762e-01 1.13838100e+00 -3.95692945e-01 -3.75131875e-01
9.20399725e-01 -9.36051756e-02 1.88744441e-02 -1.01215947e+00
4.51880664e-01 -5.63646480e-02 4.01932836e-01 -6.63563132e-01
3.24743092e-01 -3.50423753e-02 3.54858726e-01 4.55324501e-01
-9.13052633e-02 4.03761029e-01 -2.21026823e-01 -3.21520507e-01
4.46123600e-01 2.81134546e-01 1.63222596e-01 -7.93536529e-02
6.14762008e-01 -5.66026270e-01 4.18672323e-01 2.21937001e-01
-8.06466103e-01 5.84728062e-01 6.49216175e-01 -9.12139118e-02
-5.28467655e-01 -9.51751888e-01 -1.66108385e-01 1.20371497e+00
7.78888166e-01 1.12013049e-01 -1.18562853e+00 -7.55460262e-01
-1.13670200e-01 6.64572358e-01 -1.03613949e+00 -5.48892379e-01
-2.57867903e-01 -5.18770039e-01 4.68121916e-01 3.01332176e-01
5.77642560e-01 -9.74524319e-01 -6.10630631e-01 -4.13670927e-01
-1.30145416e-01 -8.72490048e-01 -9.48962808e-01 -3.88689280e-01
-4.32831287e-01 -1.21672380e+00 -9.93445277e-01 -6.09799504e-01
6.78841650e-01 4.99808073e-01 8.98418367e-01 -1.62125006e-01
-7.95148239e-02 1.88474223e-01 -1.63375363e-01 -7.52295792e-01
-8.75135735e-02 -7.30990544e-02 2.40809083e-01 4.31457549e-01
7.93262899e-01 -4.09115344e-01 -9.15700972e-01 6.14292085e-01
-5.46086967e-01 -2.95571387e-01 5.11038840e-01 7.66584098e-01
2.31786862e-01 -1.38697878e-01 6.15800083e-01 -7.07679272e-01
9.60412800e-01 -3.02159160e-01 -6.45899057e-01 2.51099497e-01
-8.71678174e-01 4.09965552e-02 3.65069509e-01 -8.76707673e-01
-1.38120997e+00 -4.79274541e-01 3.93410057e-01 -8.79602611e-01
-2.72157043e-01 6.89886734e-02 -1.48839489e-01 -4.58009601e-01
8.69042635e-01 7.20233321e-02 2.63049334e-01 -2.39643291e-01
-8.33762512e-02 6.69149160e-01 2.78991461e-01 -1.37476012e-01
7.76035786e-01 1.89991966e-01 -4.99673113e-02 -7.19015718e-01
-8.75113547e-01 -7.06504658e-02 -2.17023209e-01 -5.08285344e-01
8.04786682e-01 -6.94575667e-01 -1.32423079e+00 8.04460168e-01
-1.12848735e+00 -7.80809373e-02 3.24541330e-01 3.94938827e-01
-3.01015496e-01 2.30266824e-01 1.29322618e-01 -8.83486152e-01
-3.12617719e-01 -1.10183489e+00 8.95392835e-01 1.00273025e+00
-2.68276721e-01 -7.68337309e-01 1.16559148e-01 -4.03374247e-02
4.60811406e-01 1.15353838e-01 6.10281050e-01 -2.80887991e-01
-4.93235350e-01 1.85471252e-01 -5.58984935e-01 1.88686401e-01
2.81951725e-01 2.20237032e-01 -1.30709708e+00 -2.40832686e-01
1.96427494e-01 -1.61326095e-01 4.41129893e-01 5.27224541e-01
1.13554573e+00 -1.34997144e-01 -3.41373771e-01 8.00498962e-01
1.11866832e+00 3.36111307e-01 6.84919119e-01 3.41106892e-01
5.88350177e-01 8.21467459e-01 7.56439030e-01 2.45284572e-01
1.75289214e-01 7.54102230e-01 5.98295510e-01 -2.82750148e-02
-1.49656460e-01 -9.13265496e-02 2.51558304e-01 1.94446445e-01
-1.30865857e-01 -4.08070683e-01 -7.23806143e-01 2.26599306e-01
-1.47008705e+00 -1.00036061e+00 1.72764994e-02 2.09004927e+00
8.99210572e-01 -3.50499004e-02 1.19826265e-01 -1.99030697e-01
9.86129165e-01 1.44918486e-01 -1.03278232e+00 -1.94913700e-01
-1.68337151e-02 1.25610888e-01 2.62583852e-01 5.26918285e-02
-9.05654132e-01 7.07878649e-01 6.26244164e+00 6.11801565e-01
-1.70296395e+00 5.61083257e-02 4.25155640e-01 -4.46934402e-01
-9.38421860e-02 -4.98707891e-01 -7.11728573e-01 1.04813910e+00
6.77473724e-01 -3.45771521e-01 5.08028805e-01 6.07385576e-01
1.36768669e-01 -1.11245535e-01 -8.71183753e-01 1.18598986e+00
3.99644554e-01 -5.45539618e-01 -4.62557554e-01 -3.74390855e-02
5.65887034e-01 -3.55570883e-01 9.10352349e-01 1.98292211e-01
-3.16290647e-01 -1.02276218e+00 4.76301283e-01 7.34763443e-01
9.80373561e-01 -8.13893557e-01 6.31948411e-01 6.23522252e-02
-8.04223716e-01 -8.57242644e-02 -1.87501192e-01 2.02633604e-01
-2.31373265e-01 -1.33849367e-01 -6.61009133e-01 2.33161464e-01
1.03016877e+00 7.11716771e-01 -7.34769702e-01 1.07305586e+00
-1.97072878e-01 3.94069284e-01 3.04657733e-03 -1.44297585e-01
-2.56903976e-01 -2.63169501e-02 6.70213401e-01 4.79925394e-01
3.36008519e-01 -1.98459893e-01 -6.92735910e-01 1.19177759e+00
-2.95650885e-02 -2.64341801e-01 -4.45196062e-01 3.30108345e-01
5.14215291e-01 9.68788147e-01 -6.58611804e-02 3.70770365e-01
-3.26597184e-01 1.06633508e+00 1.94291234e-01 8.40284050e-01
-1.22040439e+00 -6.37703478e-01 1.03863275e+00 1.93679720e-01
3.10648441e-01 3.29125524e-01 -1.89815745e-01 -8.20585907e-01
2.72596359e-01 -8.51036489e-01 -6.17053658e-02 -1.18015909e+00
-1.37065446e+00 6.69847310e-01 -8.32125321e-02 -1.54666007e+00
-2.45708868e-01 -3.38857979e-01 -6.97946370e-01 1.28148067e+00
-1.86510968e+00 -9.84713554e-01 -6.89834356e-01 8.73668134e-01
2.19716936e-01 -1.58202723e-01 5.28385043e-01 1.84500679e-01
-7.71176517e-01 1.40433037e+00 1.76793113e-01 -1.72871500e-01
1.47136164e+00 -1.03932309e+00 7.66021311e-02 8.46591055e-01
-3.34794283e-01 7.97739387e-01 7.74994850e-01 -3.00549656e-01
-8.74253631e-01 -8.82975578e-01 4.19340044e-01 -4.99529719e-01
4.38954175e-01 -1.26475498e-01 -1.20901513e+00 4.06031817e-01
5.22326112e-01 -3.76600996e-02 7.05372989e-01 1.33515820e-01
-4.15444434e-01 -3.93279731e-01 -1.06644332e+00 7.48079956e-01
9.49337423e-01 -6.26580417e-01 -4.41143781e-01 -3.03164542e-01
7.88196743e-01 -5.67541718e-01 -5.03156900e-01 2.22654447e-01
8.81725907e-01 -1.29025054e+00 7.25489140e-01 -4.32288647e-01
3.00123304e-01 -3.83959442e-01 3.60988051e-01 -1.50748861e+00
-1.57248333e-01 -8.27535808e-01 -2.98106760e-01 1.32425654e+00
3.69323418e-02 -9.48369205e-01 5.54341257e-01 5.88015318e-01
1.47131652e-01 -4.67665881e-01 -7.40164518e-01 -6.62602484e-01
-1.19964719e-01 1.72456503e-01 8.69557559e-01 7.64423192e-01
-1.95897326e-01 2.87480593e-01 -4.86357331e-01 4.02272791e-01
6.35945141e-01 5.33474796e-02 1.02638495e+00 -1.15344334e+00
6.09968267e-02 -5.52001357e-01 -3.79021853e-01 -9.88267720e-01
2.98910260e-01 -1.36965262e-02 1.67583466e-01 -5.52489102e-01
-1.28147691e-01 -9.46651772e-02 -5.61731517e-01 -6.10950626e-02
-7.04148114e-01 2.95936942e-01 2.17774972e-01 4.46616203e-01
-5.25808394e-01 6.26289845e-01 1.53369486e+00 2.98929634e-03
-2.88844198e-01 2.75777638e-01 -9.62551475e-01 5.56543469e-01
5.94792426e-01 -4.06834394e-01 -5.70886970e-01 -4.58162010e-01
9.80335250e-02 -2.71223724e-01 4.80636001e-01 -1.04112160e+00
2.37828374e-01 -1.05028905e-01 5.37731171e-01 -4.76416379e-01
3.64894390e-01 -9.57677543e-01 -1.13520771e-02 7.15532303e-02
-1.99240685e-01 -1.82265863e-01 5.16000032e-01 8.46299648e-01
-3.00089002e-01 -1.28165781e-01 9.77868259e-01 5.07139921e-01
-6.94778681e-01 2.22196147e-01 1.36192262e-01 1.85295537e-01
1.15870774e+00 -6.38169765e-01 -6.37774348e-01 -4.04117256e-01
-4.08270627e-01 -5.05010486e-02 6.84175372e-01 6.34131074e-01
4.15305078e-01 -1.30116057e+00 -3.29139471e-01 5.73341131e-01
3.67042869e-01 -3.15137833e-01 6.32368386e-01 1.13880396e+00
4.43069302e-02 1.14736825e-01 -5.44025898e-01 -9.50829148e-01
-1.49763596e+00 6.39045358e-01 4.33753073e-01 3.23755026e-01
1.26389325e-01 1.14492846e+00 6.14475608e-01 2.42262304e-01
1.97647750e-01 -1.61062196e-01 -5.19737601e-01 8.38749036e-02
7.44203091e-01 2.61272103e-01 -2.25210384e-01 -6.98730588e-01
-3.85720223e-01 1.05235922e+00 -3.07449579e-01 3.58682841e-01
7.48302102e-01 -6.10547900e-01 1.55101016e-01 4.82432187e-01
1.20811307e+00 2.34115664e-02 -1.55840290e+00 -2.94286370e-01
-3.79284680e-01 -9.11999762e-01 -1.61952943e-01 -8.13006222e-01
-1.00906348e+00 1.00979519e+00 1.17910874e+00 3.43468636e-01
1.67765999e+00 -2.53881097e-01 7.31147826e-01 -1.52449995e-01
-1.08427905e-01 -7.95228660e-01 2.91225314e-01 3.26144278e-01
8.10507774e-01 -1.59440184e+00 -2.32985079e-01 -1.48423687e-01
-8.87914062e-01 8.41929972e-01 1.04059863e+00 6.86367527e-02
6.67568803e-01 -2.32407287e-01 3.38055998e-01 -2.19214465e-02
-5.50879836e-01 -2.29375005e-01 6.16005182e-01 1.06566715e+00
1.42855272e-01 -3.32525969e-01 1.23431653e-01 5.77201843e-01
-1.76729500e-01 7.41039887e-02 2.51222372e-01 4.41414505e-01
-1.09233521e-01 -6.50873125e-01 -6.09928310e-01 1.80584744e-01
-5.02334833e-01 -2.43144501e-02 -3.63496691e-01 9.25678432e-01
2.09660232e-01 8.46089303e-01 3.09683383e-01 -5.95582902e-01
3.82366896e-01 9.63469893e-02 5.33130050e-01 -2.69345213e-02
-3.68663579e-01 -9.04059336e-02 -4.82293636e-01 -5.71572542e-01
-6.16945863e-01 -5.72779894e-01 -7.27935731e-01 -5.03670216e-01
-6.55094326e-01 -2.32191518e-01 5.71546614e-01 7.63972223e-01
7.08868206e-01 7.46436596e-01 7.69642770e-01 -7.79862225e-01
-5.00982344e-01 -1.21701193e+00 -5.58657885e-01 6.03855610e-01
8.22795630e-01 -1.24413717e+00 -6.95747495e-01 1.82447851e-01] | [14.077929496765137, 0.055266935378313065] |
44c386c1-4fd2-411d-bebb-10a7d0157154 | dkt-stdrl-spatial-and-temporal-representation | 2302.11569 | null | https://arxiv.org/abs/2302.11569v1 | https://arxiv.org/pdf/2302.11569v1.pdf | DKT-STDRL: Spatial and Temporal Representation Learning Enhanced Deep Knowledge Tracing for Learning Performance Prediction | Knowledge tracing (KT) serves as a primary part of intelligent education systems. Most current KTs either rely on expert judgments or only exploit a single network structure, which affects the full expression of learning features. To adequately mine features of students' learning process, Deep Knowledge Tracing Based on Spatial and Temporal Deep Representation Learning for Learning Performance Prediction (DKT-STDRL) is proposed in this paper. DKT-STDRL extracts spatial features from students' learning history sequence, and then further extracts temporal features to extract deeper hidden information. Specifically, firstly, the DKT-STDRL model uses CNN to extract the spatial feature information of students' exercise sequences. Then, the spatial features are connected with the original students' exercise features as joint learning features. Then, the joint features are input into the BiLSTM part. Finally, the BiLSTM part extracts the temporal features from the joint learning features to obtain the prediction information of whether the students answer correctly at the next time step. Experiments on the public education datasets ASSISTment2009, ASSISTment2015, Synthetic-5, ASSISTchall, and Statics2011 prove that DKT-STDRL can achieve better prediction effects than DKT and CKT. | ['Ya Li', 'Zexue Yang', 'Haihong Yun', 'Zhifeng Wang', 'Liting Lyu'] | 2023-02-15 | null | null | null | null | ['knowledge-tracing'] | ['miscellaneous'] | [-3.22990566e-01 -3.33221197e-01 -4.36063230e-01 -2.85323292e-01
-2.24360406e-01 -3.87367994e-01 1.70328885e-01 2.05626041e-01
-5.37023842e-01 4.80104476e-01 8.06129798e-02 -4.51840460e-01
-4.96562332e-01 -1.15998840e+00 -4.86116827e-01 -4.81262684e-01
2.16913730e-01 -9.29372683e-02 7.13574409e-01 -3.26804459e-01
2.82233447e-01 3.00340682e-01 -1.65486443e+00 5.53799748e-01
1.05031347e+00 1.22049558e+00 2.12868348e-01 7.59331763e-01
-5.05609274e-01 1.41207147e+00 -5.88228583e-01 -5.41200936e-02
-4.03412074e-01 -5.03357768e-01 -1.08023608e+00 -4.25078750e-01
-1.43293533e-02 -2.74968684e-01 -7.78071523e-01 8.01396430e-01
4.73732591e-01 6.52570903e-01 4.14257109e-01 -9.84326482e-01
-8.84017646e-01 5.51057279e-01 -3.55935007e-01 5.35934985e-01
3.18026632e-01 2.25377142e-01 4.95591909e-01 -7.69866824e-01
2.57930726e-01 1.00558066e+00 6.94158256e-01 3.72497499e-01
-5.43583691e-01 -6.65600240e-01 2.00203836e-01 9.19603109e-01
-1.17351997e+00 1.14038043e-01 7.78057039e-01 -4.55675304e-01
4.91761267e-01 2.66811065e-02 1.18734062e+00 8.91237617e-01
2.83167213e-02 1.55285335e+00 1.08330977e+00 -4.21996474e-01
-2.25000709e-01 -1.23104125e-01 6.11225843e-01 1.04350841e+00
-3.21269244e-01 2.94132978e-01 -8.75716388e-01 4.53678250e-01
7.03415692e-01 4.29158270e-01 -1.73099488e-01 1.52701080e-01
-1.10240448e+00 6.18194044e-01 4.92484957e-01 6.09024286e-01
-1.28009662e-01 -8.81913677e-03 3.18801314e-01 7.36176789e-01
1.42215446e-01 -6.77835122e-02 -7.44744420e-01 -6.53592706e-01
-7.87600696e-01 1.26926601e-01 3.51242632e-01 7.54919708e-01
6.94982409e-01 9.63403657e-02 -5.41879475e-01 6.92256033e-01
2.41906673e-01 2.11742163e-01 1.14614332e+00 -6.49875164e-01
4.41023409e-01 1.11976504e+00 -2.91150391e-01 -9.66022015e-01
-3.55573446e-01 -5.99802494e-01 -5.41133642e-01 -1.27646118e-01
2.14166492e-01 -3.53989393e-01 -8.30244243e-01 1.49537206e+00
3.38905722e-01 7.57282794e-01 2.06171572e-01 5.65913081e-01
1.35057104e+00 6.31327212e-01 3.36691318e-03 -1.21666186e-01
1.35883725e+00 -8.78667057e-01 -1.00153208e+00 2.23126084e-01
1.12750244e+00 -7.17629731e-01 9.24696505e-01 5.32160759e-01
-9.83246922e-01 -9.80517566e-01 -7.92330801e-01 -3.80098782e-02
-7.09464014e-01 5.18670082e-01 3.02760690e-01 2.77309299e-01
-7.93340981e-01 8.18456352e-01 -6.65993094e-01 3.21590215e-01
5.44186056e-01 1.90289095e-01 2.19541013e-01 6.37340620e-02
-1.57631207e+00 7.52075255e-01 5.01515150e-01 3.15455556e-01
-7.68791616e-01 -9.54152763e-01 -8.39053333e-01 1.94027871e-01
2.37810209e-01 -1.56548172e-01 1.61903381e+00 -6.87335789e-01
-1.84274542e+00 3.27128500e-01 7.05190226e-02 -2.31923848e-01
6.34044111e-01 -4.15676266e-01 -5.37221551e-01 6.70960248e-02
-2.06344321e-01 3.10740978e-01 2.89496869e-01 -5.08637726e-01
-8.47778738e-01 -2.55623609e-01 -3.04971039e-01 2.62038767e-01
-7.76509702e-01 -5.81710815e-01 -4.54881549e-01 -5.96954107e-01
1.39609426e-01 -6.67374134e-01 1.47486692e-02 -4.50236410e-01
-2.48041436e-01 -8.39993834e-01 1.12549055e+00 -6.87534750e-01
1.60195911e+00 -2.11853933e+00 -3.28733176e-02 1.69850230e-01
1.65726706e-01 7.02278614e-01 -2.08161801e-01 5.34109734e-02
-1.50142638e-02 -2.67912269e-01 3.66517812e-01 2.36054718e-01
-1.54067621e-01 1.45412251e-01 -3.17422032e-01 1.73472390e-02
-1.63920671e-01 1.23699164e+00 -1.12445915e+00 -7.55062342e-01
3.92688990e-01 1.58933133e-01 -3.09858948e-01 2.71289647e-01
-1.78830668e-01 2.26359800e-01 -9.96371031e-01 4.28524435e-01
5.29810153e-02 -3.00601814e-02 -1.99958876e-01 -1.32019613e-02
-4.07115698e-01 4.49878007e-01 -9.15416598e-01 1.65274370e+00
-5.28906047e-01 7.98859775e-01 -6.35460675e-01 -1.16079831e+00
1.10058963e+00 5.41588366e-01 4.95632619e-01 -1.14372468e+00
6.05958402e-02 2.02815719e-02 2.25088522e-02 -9.19963777e-01
4.68720675e-01 2.35290512e-01 2.28147823e-02 4.04963851e-01
1.53966337e-01 3.13747764e-01 -2.09485833e-02 2.69017834e-02
1.08145845e+00 5.12126744e-01 -1.69727996e-01 9.27458480e-02
6.69297397e-01 -4.21205200e-02 7.80619502e-01 3.13235104e-01
-4.82841849e-01 -6.99462295e-02 2.86156535e-01 -6.06005311e-01
-1.73722625e-01 -9.18277442e-01 2.42303357e-01 1.27908874e+00
1.09762080e-01 -1.39886171e-01 -5.34475505e-01 -1.05799508e+00
-2.07764637e-02 5.65260589e-01 -8.10015619e-01 -8.06386352e-01
-6.69499755e-01 -2.32919008e-01 7.19141960e-01 8.30898702e-01
1.04437816e+00 -1.38092959e+00 -6.46829247e-01 3.70227009e-01
-2.16499329e-01 -7.75429130e-01 -5.49190938e-01 3.00421100e-02
-9.28711176e-01 -1.18607748e+00 -5.99248588e-01 -1.09836674e+00
4.11852449e-01 1.41121641e-01 7.33882606e-01 2.70308226e-01
-4.04163122e-01 3.47365588e-01 -2.62916058e-01 -3.62623602e-01
2.29129240e-01 2.77620684e-02 2.48766169e-02 -6.28702343e-02
7.11989224e-01 -5.75654685e-01 -6.89485252e-01 1.23236299e-01
-4.55807120e-01 8.86397138e-02 7.72013307e-01 6.30302846e-01
5.66465259e-01 4.43000406e-01 9.05636668e-01 -7.08951831e-01
6.91764057e-01 -3.42450738e-01 -3.45499635e-01 3.85339260e-01
-6.85939074e-01 2.23612368e-01 5.15601516e-01 -8.87947619e-01
-1.29096413e+00 -3.23231995e-01 -1.31730065e-01 -8.34120810e-01
-8.83253440e-02 9.76314902e-01 -6.09172285e-02 9.94162634e-02
3.15595776e-01 8.75592709e-01 -3.02551270e-01 -7.49559641e-01
6.69608638e-02 3.93132567e-01 5.28156698e-01 -6.88324749e-01
5.12693465e-01 -2.23008931e-01 -2.13776171e-01 -5.55091083e-01
-1.39525771e+00 -1.60063326e-01 -7.82858431e-01 -4.73700404e-01
7.90458381e-01 -7.45219588e-01 -1.49405301e+00 7.88829625e-01
-8.63813281e-01 -6.53465688e-01 -4.90218252e-01 8.48949611e-01
-2.67594934e-01 1.15759708e-01 -1.09419096e+00 -6.60807490e-01
-1.16694957e-01 -1.10857308e+00 2.52748013e-01 8.08745980e-01
-3.61928041e-03 -1.30215812e+00 1.78604633e-01 5.67558408e-01
2.07045063e-01 -6.16681129e-02 1.01882136e+00 -7.34514713e-01
-3.36385161e-01 5.08038253e-02 1.42953962e-01 3.91112059e-01
1.27203479e-01 1.13541529e-01 -8.33395660e-01 9.69139785e-02
-1.18488833e-01 -4.19600010e-01 9.69591856e-01 3.42498839e-01
1.72190702e+00 -3.63564909e-01 -1.72362149e-01 4.10008013e-01
9.50194418e-01 5.55645525e-01 4.20851916e-01 5.18494964e-01
9.09295261e-01 5.65173209e-01 8.42947185e-01 1.05018683e-01
8.04537117e-01 2.39657342e-01 2.13534683e-01 2.49871343e-01
-7.48107880e-02 -4.19935375e-01 5.52465320e-01 1.74461591e+00
-2.05194607e-01 2.59827971e-01 -9.88631606e-01 6.98505819e-01
-1.85877812e+00 -1.10005701e+00 -2.93167412e-01 1.77424479e+00
1.25818527e+00 1.74012199e-01 -1.63718715e-01 4.40017074e-01
4.82437581e-01 5.56917004e-02 -6.43334329e-01 -2.59445965e-01
3.22173148e-01 2.90779501e-01 1.32860497e-01 2.05523953e-01
-7.82979846e-01 9.04249609e-01 4.98980093e+00 1.06607676e+00
-1.01557553e+00 -5.08550666e-02 3.84447813e-01 -7.18903244e-02
-8.91835801e-03 -2.60827869e-01 -9.72340822e-01 6.21195078e-01
1.07904780e+00 -9.96649414e-02 -5.14140427e-02 6.45876050e-01
1.01027347e-01 5.86637966e-02 -8.41592491e-01 8.82276177e-01
-3.42497349e-01 -1.22046626e+00 2.98773237e-02 -3.54930349e-02
7.94648468e-01 -5.22698581e-01 5.27136326e-01 9.95262623e-01
6.28659606e-01 -1.11323869e+00 4.27996516e-01 1.21951342e+00
3.40066880e-01 -1.25863826e+00 7.55917490e-01 6.74766481e-01
-1.44839048e+00 -4.73646857e-02 -1.10869378e-01 -5.72761670e-02
-5.31911016e-01 4.98973727e-01 -8.41850519e-01 7.56414831e-01
7.78116584e-01 1.16025066e+00 -7.18282044e-01 8.92544687e-01
-5.99388957e-01 1.03533804e+00 2.62349278e-01 -2.97886610e-01
4.04672414e-01 -1.52862012e-01 -1.18981175e-01 1.07066846e+00
6.05199635e-01 4.17401403e-01 2.34330133e-01 4.43515867e-01
-2.70888120e-01 -1.27965316e-01 -2.22904399e-01 -2.16953397e-01
5.71098685e-01 9.88004863e-01 -3.86550248e-01 -4.05588597e-01
-4.13412541e-01 7.22701848e-01 4.52825606e-01 4.73328978e-01
-6.29718840e-01 -7.36701369e-01 5.08112133e-01 -2.34100476e-01
3.75342578e-01 3.95858139e-02 -1.22121871e-01 -9.74057853e-01
-2.93782979e-01 -6.39228582e-01 6.27692342e-01 -9.23730671e-01
-9.01546597e-01 1.10724136e-01 -3.25781405e-01 -1.33396041e+00
-1.77334711e-01 -4.18388814e-01 -1.07853734e+00 9.33757067e-01
-1.70279086e+00 -7.18097985e-01 -3.53796422e-01 9.49322879e-01
6.64169312e-01 -4.62383956e-01 5.21516681e-01 1.39448568e-01
-1.07283735e+00 6.44254982e-01 1.10100560e-01 7.87269294e-01
3.77766520e-01 -1.25924206e+00 -1.39886409e-01 4.92874235e-01
-2.71609556e-02 5.37632883e-01 2.35690817e-01 -5.85845113e-01
-1.22357988e+00 -1.39744902e+00 7.25977421e-01 -2.42902204e-01
4.82613355e-01 4.71322507e-01 -1.23592043e+00 8.62937033e-01
5.77140003e-02 1.59289673e-01 9.09676671e-01 -4.00030799e-02
-8.88007432e-02 -1.47920206e-01 -6.60997510e-01 5.53410113e-01
6.34684920e-01 -7.28340864e-01 -1.14317942e+00 3.27529609e-02
9.38380301e-01 -5.93692482e-01 -1.40780044e+00 4.04252201e-01
5.00980020e-01 -7.43369818e-01 8.85421395e-01 -7.20434248e-01
6.09816074e-01 -1.04921594e-01 4.10003424e-01 -1.41583776e+00
-1.80177316e-01 -1.13899894e-01 -5.58649182e-01 1.27403235e+00
5.06831007e-03 -2.49066234e-01 9.33729470e-01 1.97361540e-02
-4.73911524e-01 -1.26117420e+00 -5.59168816e-01 -5.76488435e-01
8.54096785e-02 -4.87873435e-01 7.22966135e-01 1.37783730e+00
2.17900556e-02 2.73011386e-01 1.19862981e-01 3.09145451e-02
3.72151524e-01 4.54616368e-01 2.07055986e-01 -1.43246210e+00
3.02122477e-02 -5.70360303e-01 -2.07761794e-01 -1.21388018e+00
2.67010987e-01 -8.11787724e-01 -2.35672712e-01 -1.46456230e+00
-6.39489293e-02 -3.07300508e-01 -7.34454811e-01 8.11080575e-01
-3.42010021e-01 -2.52495617e-01 -2.55832434e-01 -1.27280101e-01
-8.07976663e-01 8.11383426e-01 2.03728795e+00 -5.05013578e-02
-3.96623462e-01 1.64970905e-01 -2.96083659e-01 9.41676915e-01
7.57340908e-01 -1.60298020e-01 -6.81188881e-01 -2.55753547e-01
1.33826986e-01 3.64190400e-01 2.93808043e-01 -1.14814329e+00
7.17249393e-01 -4.74108189e-01 1.13291836e+00 -1.11955595e+00
2.24609196e-01 -4.20115501e-01 -6.67959154e-01 9.18567121e-01
-6.36029005e-01 -4.71988171e-02 3.47534716e-01 4.04736251e-01
-5.04294217e-01 -2.74524391e-01 5.35814106e-01 -8.58871862e-02
-9.83186901e-01 5.87422371e-01 -3.79266351e-01 1.65445045e-01
9.45954978e-01 -8.33759457e-02 -3.41537833e-01 -4.42459695e-02
-1.00545561e+00 7.58537352e-01 -4.97994274e-01 5.27873278e-01
1.21358895e+00 -1.70887554e+00 -3.98270875e-01 2.03148767e-01
-1.93246737e-01 3.44630361e-01 6.86841846e-01 8.85302186e-01
-2.73163259e-01 6.00704312e-01 -1.26000687e-01 -5.60924590e-01
-1.38237870e+00 4.01984602e-01 4.41963315e-01 -5.51051259e-01
-5.73379278e-01 8.53417218e-01 -6.90954626e-02 -7.04025209e-01
2.58604318e-01 -4.96245712e-01 -8.75515521e-01 3.07391703e-01
6.56069994e-01 5.06808877e-01 5.59955351e-02 -2.33301044e-01
-6.69117570e-02 4.38217223e-01 -2.90474653e-01 -4.86319922e-02
1.41682482e+00 1.35252804e-01 3.55814964e-01 6.85502946e-01
1.11833739e+00 -2.02493608e-01 -1.18826091e+00 -7.58303165e-01
2.79359043e-01 -2.14002915e-02 2.33151644e-01 -8.12099516e-01
-1.38912165e+00 1.29717767e+00 5.42116046e-01 -1.10777900e-01
1.09641576e+00 -4.72828299e-01 1.05847573e+00 6.01369381e-01
-6.17616884e-02 -1.15992963e+00 6.73609018e-01 9.38534081e-01
3.59487176e-01 -8.00502956e-01 -2.75913686e-01 1.73339874e-01
-5.03439188e-01 1.26504016e+00 1.16955590e+00 7.31414706e-02
8.34560215e-01 -6.84438720e-02 -1.01437673e-01 -1.28189430e-01
-9.77147341e-01 -1.41466513e-01 5.92709780e-01 1.48999229e-01
5.81160486e-01 -8.13436136e-02 9.15485173e-02 1.14338768e+00
-4.06724513e-01 1.61844879e-01 3.27949762e-01 1.26208198e+00
-8.04693401e-01 -8.27321053e-01 -2.02100351e-01 4.56517518e-01
-1.99309632e-01 4.74540144e-02 -1.46863937e-01 7.21606493e-01
3.91829312e-01 7.14972556e-01 -1.15632758e-01 -8.01681995e-01
6.12119913e-01 4.23929751e-01 4.58678633e-01 -6.25074148e-01
-9.33776796e-01 -3.81496906e-01 -4.63582635e-01 -5.21197379e-01
-2.60038912e-01 -5.13377845e-01 -1.78043783e+00 -3.07112843e-01
-2.08963260e-01 5.91394365e-01 2.11541727e-01 1.44885528e+00
-4.18487750e-02 1.44032061e+00 5.11524081e-01 -4.28344756e-02
-4.61201280e-01 -1.02647054e+00 -4.05938119e-01 8.72546658e-02
4.28497910e-01 -6.02266073e-01 -1.30053088e-02 3.84210423e-02] | [10.14400577545166, 7.0787224769592285] |
4d3bbbdb-2210-4e17-983a-0ae8bec73b87 | birl-benchmark-on-image-registration-methods | 1912.13452 | null | https://arxiv.org/abs/1912.13452v2 | https://arxiv.org/pdf/1912.13452v2.pdf | BIRL: Benchmark on Image Registration methods with Landmark validation | This report presents a generic image registration benchmark with automatic evaluation using landmark annotations. The key features of the BIRL framework are: easily extendable, performance evaluation, parallel experimentation, simple visualisations, experiment's time-out limit, resuming unfinished experiments. From the research practice, we identified and focused on these two main use-cases: (a) comparison of user's (newly developed) method with some State-of-the-Art (SOTA) methods on a common dataset and (b) experimenting SOTA methods on user's custom dataset (which should contain landmark annotation). Moreover, we present an integration of several standard image registration methods aiming at biomedical imaging into the BIRL framework. This report also contains experimental results of these SOTA methods on the CIMA dataset, which is a dataset of Whole Slice Imaging (WSI) from histology/pathology containing several multi-stain tissue samples from three tissue kinds. Source and results: https://borda.github.io/BIRL | ['Jiri Borovec'] | 2019-12-31 | null | null | null | null | ['birl-cima'] | ['medical'] | [ 1.72145039e-01 1.12143010e-01 5.79286851e-02 -1.91388130e-01
-9.49388504e-01 -4.43893135e-01 5.78395009e-01 4.36094910e-01
-6.51589811e-01 5.54492772e-01 -6.24185540e-02 -3.54015738e-01
-3.95663410e-01 -3.30716133e-01 -1.59616351e-01 -8.25183153e-01
-5.06373167e-01 7.81782568e-01 7.41719127e-01 -1.40260935e-01
2.81452715e-01 8.59649181e-01 -1.25256133e+00 3.24771285e-01
1.69676378e-01 7.54673898e-01 3.09637308e-01 6.69087589e-01
1.13474682e-01 2.84224242e-01 -4.69244272e-01 5.48048727e-02
2.94788599e-01 -2.81722516e-01 -1.29392707e+00 -2.64046222e-01
4.76607502e-01 1.84652045e-01 2.28348345e-01 7.93425679e-01
7.63106346e-01 -2.04597950e-01 4.52104300e-01 -1.38479316e+00
-4.33611907e-02 4.30551887e-01 -4.92200136e-01 7.16623366e-01
3.71902317e-01 6.46739304e-02 4.49270040e-01 -5.39820194e-01
1.38854980e+00 7.00739622e-01 9.66803193e-01 4.60607886e-01
-1.34804392e+00 -3.25245589e-01 -3.93411994e-01 5.80759458e-02
-1.53004539e+00 -2.55259991e-01 2.67803878e-01 -6.46298528e-01
9.17782485e-01 8.11134040e-01 4.50399637e-01 7.49806881e-01
3.69346082e-01 4.34822112e-01 1.59212351e+00 -3.99838597e-01
2.43521929e-01 8.93816501e-02 2.38089129e-01 6.82973802e-01
5.38666658e-02 -1.03150226e-01 -1.32259920e-01 -4.15069878e-01
6.91466570e-01 -1.29830122e-01 -4.50032562e-01 -4.40968752e-01
-1.85125613e+00 2.56872982e-01 2.74749368e-01 9.60252643e-01
-3.64750884e-02 -2.36664265e-01 8.27819705e-01 4.40845728e-01
2.96558946e-01 2.19753072e-01 -3.76451999e-01 7.16958195e-02
-1.08858669e+00 6.32177740e-02 6.93756819e-01 7.12333202e-01
5.48159957e-01 -7.24126816e-01 -1.71726853e-01 6.44846141e-01
3.20971787e-01 -1.29604414e-01 1.04030704e+00 -7.28187978e-01
-3.29159647e-01 5.96888483e-01 -2.95047700e-01 -6.36703491e-01
-9.28563893e-01 -2.06646875e-01 -9.66085792e-01 6.23471677e-01
5.93472898e-01 4.45789814e-01 -8.11994970e-01 1.33056951e+00
4.03586805e-01 1.69205219e-01 -1.07996732e-01 9.07596588e-01
1.19953537e+00 3.92150413e-03 -1.22701726e-03 -3.03710073e-01
1.66574788e+00 -8.85669410e-01 -7.13039219e-01 4.80367720e-01
1.17681396e+00 -1.15007949e+00 1.00934052e+00 4.66870189e-01
-8.45406771e-01 -2.82479197e-01 -9.65177655e-01 -3.13913263e-02
-1.01829720e+00 3.05750877e-01 7.83404052e-01 6.04456127e-01
-1.62499952e+00 5.80461800e-01 -9.68153238e-01 -9.68142152e-01
3.92369002e-01 6.99898005e-01 -9.75202382e-01 2.84671187e-01
-6.20077729e-01 9.69694376e-01 2.13125736e-01 -7.71930218e-02
-5.67759097e-01 -8.79258871e-01 -5.42547524e-01 -7.03355789e-01
2.40915995e-02 -5.58181822e-01 8.32837343e-01 -5.84013700e-01
-1.32817805e+00 1.77446389e+00 2.78017759e-01 -3.29099476e-01
8.27798784e-01 3.79710823e-01 -2.90431321e-01 2.21785456e-01
3.34691674e-01 7.75156558e-01 -1.60471067e-01 -1.03277540e+00
-2.67350763e-01 -4.96966392e-01 -2.70103574e-01 -1.51378214e-01
1.07336074e-01 2.63449401e-01 -5.34810185e-01 -5.94885945e-01
1.25143444e-03 -9.97014821e-01 -3.13212723e-01 1.49331614e-01
-4.80883092e-01 -9.05338824e-02 8.45856786e-01 -6.72988653e-01
9.77669954e-01 -1.96762395e+00 4.89525404e-03 2.17722312e-01
1.95930198e-01 2.25874141e-01 -6.30971491e-02 4.67644274e-01
-5.55098116e-01 1.74913704e-01 -2.18703493e-01 -3.80255401e-01
-2.20320478e-01 1.49417641e-02 3.65057766e-01 8.98670673e-01
-6.55304909e-01 6.35824025e-01 -7.77320385e-01 -9.05978322e-01
3.52771997e-01 3.53917480e-01 -9.07997508e-03 6.46090135e-02
3.77753735e-01 8.49626064e-01 -1.93955675e-01 8.20623696e-01
7.93320060e-01 -2.12442100e-01 1.86529994e-01 -6.18825853e-01
-3.92476290e-01 -1.68115020e-01 -1.20474553e+00 2.23304033e+00
-3.01552415e-01 3.56686294e-01 9.45580155e-02 -6.55887961e-01
6.73397303e-01 5.71665108e-01 9.14393902e-01 -3.01889181e-01
2.25171626e-01 4.19815779e-01 -2.11793324e-03 -5.02362907e-01
-1.10098533e-01 2.60109723e-01 2.96364278e-01 5.92293084e-01
3.98513436e-01 -6.86327070e-02 5.79474390e-01 9.83010754e-02
1.31723619e+00 2.76346236e-01 5.78968108e-01 -8.99780571e-01
8.11661720e-01 1.11334413e-01 2.84809589e-01 4.69215900e-01
-5.56064546e-01 8.99136662e-01 4.27433133e-01 -5.18889785e-01
-1.00156796e+00 -9.16103184e-01 -6.64737523e-01 6.37128353e-01
-3.65420245e-03 -5.71691871e-01 -8.78150761e-01 -7.28365600e-01
-4.80824322e-01 -2.13697869e-02 -1.20589840e+00 5.18149376e-01
-3.45483333e-01 -1.09081793e+00 6.76738620e-01 -5.09911738e-02
4.10990804e-01 -9.31787610e-01 -9.79575694e-01 -8.72692242e-02
1.69442192e-01 -1.06611288e+00 -3.07498097e-01 1.33777549e-02
-8.38895679e-01 -1.32079065e+00 -8.76888752e-01 -7.66724706e-01
9.35090780e-01 -1.15267858e-01 1.21125758e+00 4.30163473e-01
-1.06496561e+00 5.51912606e-01 -3.24897110e-01 -1.44266590e-01
-5.13771594e-01 2.25482762e-01 -2.41567031e-01 -3.32014591e-01
-1.02280512e-01 -6.96654439e-01 -7.78304815e-01 6.90686285e-01
-1.22387064e+00 1.54731989e-01 5.10529518e-01 8.46186936e-01
9.87791181e-01 -3.83627921e-01 2.07149684e-01 -1.18995416e+00
4.29949909e-01 -2.27651849e-01 -7.07229376e-01 4.83897865e-01
-6.12708867e-01 -3.17722380e-01 1.78435549e-01 -1.54337496e-01
-6.60430908e-01 -1.32503733e-02 -3.06534231e-01 -6.46646693e-02
-5.31050444e-01 2.74189651e-01 5.74328899e-02 -5.22810757e-01
7.71938980e-01 -1.32553533e-01 5.13408482e-01 -5.40296912e-01
-3.79095748e-02 4.40336496e-01 5.07937133e-01 -4.58684742e-01
2.42545202e-01 8.93580735e-01 4.15984243e-01 -4.43550646e-01
-3.12280416e-01 -8.76401365e-01 -1.12211239e+00 -2.90103853e-01
6.77322328e-01 -2.77901828e-01 -7.11352110e-01 6.01196289e-01
-9.24177170e-01 -5.88951588e-01 -2.82316655e-01 4.58010793e-01
-6.02894366e-01 4.09098804e-01 -6.27069414e-01 -1.04912065e-01
-8.03987622e-01 -1.68843591e+00 1.19757533e+00 1.10241599e-01
-4.09020394e-01 -1.37827790e+00 6.45982504e-01 1.02127478e-01
5.18107235e-01 7.26320267e-01 4.75566745e-01 -8.09134901e-01
-1.96160525e-01 -3.55714798e-01 -6.53080940e-02 -1.92222802e-03
1.83140367e-01 1.94088891e-01 -9.27288771e-01 -3.96645963e-01
-2.23243386e-01 1.15773510e-02 4.45065856e-01 2.72082061e-01
1.08169818e+00 1.66578174e-01 -6.98993683e-01 7.45467365e-01
1.65985179e+00 -7.03288540e-02 8.62411916e-01 8.23434293e-01
1.24660224e-01 5.09254575e-01 6.61909342e-01 7.79131800e-02
-1.15141738e-02 1.08468390e+00 5.00599623e-01 -5.27308762e-01
-4.31461900e-01 5.52436650e-01 -2.35858545e-01 8.02754879e-01
-5.67691982e-01 2.04367965e-01 -1.20139182e+00 5.06165922e-01
-1.70267999e+00 -5.66484332e-01 -4.68706667e-01 2.36846161e+00
8.28662932e-01 -1.45415261e-01 2.18070626e-01 2.46972106e-02
5.61774135e-01 -1.78929970e-01 1.55243263e-01 -1.90003768e-01
-1.90099850e-02 3.72734040e-01 5.58503509e-01 3.16218108e-01
-1.30612636e+00 5.79332232e-01 6.39463711e+00 1.15128160e+00
-1.33425117e+00 6.37405992e-01 8.35847974e-01 2.10460454e-01
1.35091260e-01 -1.13351583e-01 -5.07366896e-01 2.11764842e-01
9.35959101e-01 -3.39674145e-01 4.41549160e-02 5.06847203e-01
9.54670757e-02 -3.91703367e-01 -1.01921701e+00 7.03321457e-01
1.14015624e-01 -1.66291773e+00 -2.24828675e-01 2.74168789e-01
3.00629169e-01 3.25227976e-01 -1.62404135e-01 -1.63657725e-01
-9.27218571e-02 -1.00678992e+00 2.60099024e-01 4.94758129e-01
1.07552063e+00 -6.31739125e-02 1.37535465e+00 -1.74435899e-01
-8.56841624e-01 7.93629408e-01 -1.86406538e-01 5.78950644e-01
-6.40022010e-02 6.85820580e-01 -9.17500973e-01 1.12387764e+00
9.24216211e-01 6.78540170e-01 -1.04699051e+00 1.48549974e+00
1.63039595e-01 3.43238294e-01 -4.73047823e-01 4.48346734e-01
1.61238402e-01 -2.70422250e-01 5.57156324e-01 1.66147482e+00
3.05838346e-01 -2.69614369e-01 2.19254456e-02 4.90552515e-01
3.72256130e-01 6.25682175e-01 -6.58468962e-01 6.79360092e-01
6.56200722e-02 2.11061168e+00 -1.58793008e+00 -8.42914507e-02
-1.78568706e-01 8.27938437e-01 -5.34479842e-02 -1.46053374e-01
-6.42874420e-01 -1.43053696e-01 2.15438724e-01 3.45418304e-01
-2.27097094e-01 1.77201405e-01 -8.82337540e-02 -6.54872417e-01
-3.36316377e-01 -6.27825499e-01 7.79611409e-01 -9.03593838e-01
-1.15608978e+00 1.00627959e+00 3.21988910e-01 -1.42631686e+00
-4.38763667e-03 -8.44086647e-01 -5.58723509e-01 7.27611005e-01
-1.21436810e+00 -1.45086467e+00 -4.58678424e-01 5.52992642e-01
3.93245965e-02 -1.31441802e-01 1.34143591e+00 6.14393771e-01
-4.02454287e-01 5.38852394e-01 -3.35777178e-02 1.37317374e-01
9.38386023e-01 -1.34890091e+00 7.52373738e-03 3.96676540e-01
-9.53767598e-02 7.23241389e-01 6.01067901e-01 -2.48107359e-01
-1.05041313e+00 -8.45105231e-01 6.01932406e-01 -3.45526546e-01
9.19490516e-01 -1.13347434e-01 -7.36047745e-01 8.36479783e-01
5.76667309e-01 5.53961098e-01 1.06896174e+00 -2.41491258e-01
-3.99289466e-02 -8.42950121e-02 -1.61843133e+00 5.38805902e-01
7.24857986e-01 3.98194864e-02 -2.69392103e-01 7.17149258e-01
-9.39169675e-02 -8.50722194e-01 -1.43118668e+00 4.68968481e-01
5.09360313e-01 -1.28407753e+00 9.76455033e-01 -7.02304691e-02
-2.32435800e-02 -5.59344769e-01 2.06077397e-01 -1.05896235e+00
9.00360048e-02 -7.30984986e-01 5.77566028e-01 1.16341650e+00
2.44073942e-01 -1.01209331e+00 3.21748972e-01 2.12504506e-01
-4.42516416e-01 -9.48234141e-01 -1.51935530e+00 -7.87018836e-01
2.16837116e-02 -9.42028090e-02 4.05221432e-01 1.09672558e+00
2.43820935e-01 -1.86460927e-01 3.22607756e-01 -2.21572265e-01
6.80604160e-01 -7.13671073e-02 8.39186430e-01 -1.17297149e+00
-2.47401167e-02 -5.80505133e-01 -1.14720261e+00 5.14222756e-02
-1.10598192e-01 -1.29161751e+00 -3.00143033e-01 -1.47493422e+00
2.46401146e-01 -6.26395166e-01 -4.06068742e-01 7.93838561e-01
4.18057442e-01 8.23741019e-01 -6.39671460e-02 5.59048593e-01
-6.79414690e-01 -5.42471945e-01 1.22608626e+00 5.49021512e-02
2.90088207e-01 -1.39821604e-01 -2.23112151e-01 6.27040088e-01
7.83919215e-01 -5.40512860e-01 -9.80777293e-02 1.51197746e-01
6.14632890e-02 -1.74924821e-01 6.53588176e-01 -1.28949630e+00
2.99598992e-01 2.38189310e-01 -9.95313153e-02 -4.95957494e-01
-5.83089925e-02 -8.29213977e-01 8.24910343e-01 6.74285591e-01
-1.36736080e-01 1.19192109e-01 4.00678933e-01 -1.46342739e-01
-3.05339545e-01 -4.26637590e-01 1.04530585e+00 -3.51016700e-01
-4.88373160e-01 2.20094815e-01 -2.40715325e-01 -2.54663140e-01
1.37616491e+00 -3.84562641e-01 -5.62650144e-01 3.51442665e-01
-1.08994710e+00 -9.56575125e-02 7.87575722e-01 -3.05197705e-02
6.50550872e-02 -1.15654504e+00 -8.44745636e-01 5.99452667e-02
2.51461238e-01 -6.15795664e-02 3.91987711e-01 1.74688661e+00
-1.19005632e+00 3.95887882e-01 -5.55374920e-01 -8.97348404e-01
-1.68005157e+00 3.47020596e-01 6.53959751e-01 -8.70416641e-01
-7.74783731e-01 5.55539727e-01 -3.48904319e-02 -7.20355690e-01
-2.40604401e-01 -1.90147415e-01 -2.84047008e-01 -1.39799546e-02
4.46652234e-01 3.13249946e-01 7.72206128e-01 -6.45820320e-01
-6.80283785e-01 5.71835756e-01 -8.84397030e-02 -5.41403554e-02
1.46621001e+00 3.26425694e-02 -7.01242924e-01 6.53462708e-01
1.24872208e+00 1.18484735e-01 -4.32330072e-01 1.61715075e-01
1.77240714e-01 -2.32221618e-01 1.16833992e-01 -1.00412929e+00
-1.29317904e+00 4.58884865e-01 1.18529546e+00 2.59268969e-01
1.07150340e+00 1.46712095e-01 2.74662584e-01 -1.57752261e-01
6.87272608e-01 -7.06771612e-01 -5.53054392e-01 1.10466801e-01
1.07350397e+00 -9.91367340e-01 4.64492738e-01 -6.29897654e-01
-2.54977256e-01 1.33019590e+00 1.96837768e-01 -2.73890905e-02
9.45517659e-01 7.48429179e-01 5.13417244e-01 -5.07021725e-01
-4.62261289e-01 -1.32163540e-01 1.91319957e-01 7.41430461e-01
8.50023568e-01 -1.14189819e-01 -8.45322788e-01 2.78728724e-01
-9.98332426e-02 2.61435330e-01 4.98986006e-01 1.03496265e+00
3.05538327e-01 -1.23231959e+00 -4.34621722e-01 3.66174757e-01
-7.70075262e-01 -1.73055045e-02 -1.64062023e-01 1.36486697e+00
1.51618391e-01 3.39269936e-01 -1.16689138e-01 1.42102867e-01
1.90117851e-01 -2.53519177e-01 6.06423795e-01 -4.88151938e-01
-1.04149628e+00 4.23676744e-02 -1.38327017e-01 -6.97803795e-01
-8.77897382e-01 -9.13018882e-01 -1.29264283e+00 -6.76249564e-02
-1.48754895e-01 1.63940221e-01 8.53798389e-01 6.96110666e-01
3.03946078e-01 5.60481787e-01 1.58254772e-01 -1.02902937e+00
3.54545772e-01 -1.14076567e+00 -7.43878782e-01 6.42500341e-01
-3.50534290e-01 -6.74384058e-01 -4.47671801e-01 4.05378401e-01] | [14.831329345703125, -2.9791982173919678] |
a3a8bf42-e9cd-46d6-87fa-38db009922d9 | dont-discuss-investigating-semantic-and | null | null | https://aclanthology.org/2021.ranlp-main.168 | https://aclanthology.org/2021.ranlp-main.168.pdf | “Don’t discuss”: Investigating Semantic and Argumentative Features for Supervised Propagandist Message Detection and Classification | One of the mechanisms through which disinformation is spreading online, in particular through social media, is by employing propaganda techniques. These include specific rhetorical and psychological strategies, ranging from leveraging on emotions to exploiting logical fallacies. In this paper, our goal is to push forward research on propaganda detection based on text analysis, given the crucial role these methods may play to address this main societal issue. More precisely, we propose a supervised approach to classify textual snippets both as propaganda messages and according to the precise applied propaganda technique, as well as a detailed linguistic analysis of the features characterising propaganda information in text (e.g., semantic, sentiment and argumentation features). Extensive experiments conducted on two available propagandist resources (i.e., NLP4IF’19 and SemEval’20-Task 11 datasets) show that the proposed approach, leveraging different language models and the investigated linguistic features, achieves very promising results on propaganda classification, both at sentence- and at fragment-level. | ['Serena Villata', 'Elena Cabrio', 'Vorakit Vorakitphan'] | null | null | https://aclanthology.org/2021.ranlp-1.168 | https://aclanthology.org/2021.ranlp-1.168.pdf | ranlp-2021-9 | ['logical-fallacies', 'propaganda-detection'] | ['miscellaneous', 'natural-language-processing'] | [ 2.20852476e-02 2.90434003e-01 -5.00064135e-01 -5.34269176e-02
-2.64134735e-01 -6.35950327e-01 1.56301475e+00 1.01520789e+00
-4.99491274e-01 7.68784285e-01 8.70117188e-01 -5.75009644e-01
-5.38674742e-02 -9.09720361e-01 -1.03318304e-01 -7.13302612e-01
1.60808474e-01 4.11161743e-02 -4.98590544e-02 -6.84610069e-01
9.69985723e-01 2.39005238e-01 -1.33325136e+00 4.32986349e-01
1.07725942e+00 6.53409839e-01 -3.86337996e-01 3.90143186e-01
-8.82743418e-01 1.61397469e+00 -9.90250587e-01 -7.55076587e-01
-6.10604167e-01 -5.75354576e-01 -1.07616889e+00 3.98992840e-03
-2.49710400e-02 3.25712442e-01 3.66565555e-01 1.05850315e+00
2.10299432e-01 -3.63319516e-01 1.02593875e+00 -4.70709383e-01
-4.58667457e-01 1.20737326e+00 -6.06586814e-01 2.85486072e-01
7.81890631e-01 -2.98262894e-01 8.66504431e-01 -6.63832307e-01
9.26614225e-01 1.52462304e+00 3.75129700e-01 2.69685417e-01
-1.08513737e+00 -3.07619393e-01 -1.31462529e-01 2.77693361e-01
-7.58154809e-01 -3.41427773e-01 1.24385822e+00 -9.90198076e-01
6.40642762e-01 2.32092842e-01 7.38745213e-01 1.56192565e+00
4.55338091e-01 5.78250110e-01 1.60837758e+00 -9.05012667e-01
3.35543662e-01 5.43271482e-01 5.74475884e-01 6.37823999e-01
1.94696337e-01 -4.80975419e-01 -8.35657656e-01 -5.64333022e-01
-1.30419821e-01 -6.76904082e-01 -1.16279433e-02 1.59836680e-01
-7.17603862e-01 1.46583855e+00 2.31433943e-01 1.11374414e+00
-5.53467155e-01 -2.23203838e-01 1.02251208e+00 2.72873551e-01
1.24979842e+00 3.11329544e-01 -2.88618058e-02 -1.02001838e-01
-8.36095273e-01 1.65516436e-01 1.08623016e+00 1.15449890e-01
5.02488673e-01 -3.13118130e-01 -1.54426128e-01 7.56086230e-01
5.03624439e-01 5.57027280e-01 6.55675769e-01 5.33048660e-02
6.67548180e-01 7.21787274e-01 -1.56022087e-01 -1.90211153e+00
-6.30295098e-01 -2.70765096e-01 -7.06349671e-01 7.67213199e-03
3.13296616e-01 -3.37775379e-01 -4.07994866e-01 1.40295053e+00
5.50527036e-01 -5.46620488e-01 -4.31778692e-02 5.92903018e-01
1.11286867e+00 6.69807374e-01 4.81297165e-01 -6.10070765e-01
1.58683431e+00 -5.73868096e-01 -9.91659403e-01 -6.08905852e-02
1.03587842e+00 -1.19225216e+00 6.24901235e-01 3.17162097e-01
-8.46995652e-01 3.36330906e-02 -8.93611670e-01 1.75854519e-01
-7.56508231e-01 1.30216062e-01 4.71872658e-01 8.45516980e-01
-3.33020926e-01 4.57480967e-01 -2.93215275e-01 -5.49058259e-01
5.11777580e-01 -5.36272824e-01 -4.90284413e-02 5.51141262e-01
-1.43780923e+00 1.17868042e+00 6.22766435e-01 3.59634422e-02
-3.92382026e-01 -3.06667536e-01 -6.82205677e-01 -3.50240886e-01
5.01311004e-01 -2.00214282e-01 6.52524769e-01 -1.20502830e+00
-1.44488585e+00 1.24023104e+00 2.21737742e-01 -8.55755091e-01
6.12144887e-01 -3.23274136e-01 -5.36235511e-01 2.38176405e-01
1.90344930e-01 -1.52515963e-01 1.32132816e+00 -9.53798234e-01
-2.18417063e-01 -3.30433995e-01 -3.95509414e-02 -1.92318648e-01
-8.26562703e-01 8.24146032e-01 7.97251701e-01 -6.40424073e-01
-1.99843839e-01 -4.60783631e-01 1.45305276e-01 -6.77448750e-01
-9.20582891e-01 -6.47925854e-01 6.57541811e-01 -1.02029312e+00
1.49897337e+00 -1.89136231e+00 2.25867108e-01 5.22137344e-01
4.93421078e-01 5.66744268e-01 4.23622698e-01 1.06531119e+00
1.36349156e-01 4.45456386e-01 5.29507510e-02 4.08137254e-02
-1.00257263e-01 -1.65054575e-01 -4.75869447e-01 7.33967602e-01
3.45659032e-02 6.08567894e-01 -8.98281574e-01 -6.85764670e-01
1.32122681e-01 4.19110417e-01 -2.03407004e-01 -1.49895638e-01
-3.15578252e-01 4.07439530e-01 -1.01724672e+00 1.81169689e-01
4.84284937e-01 -1.57589912e-01 4.55754697e-01 5.19473329e-02
-7.14205265e-01 7.83798337e-01 -4.41488296e-01 1.09188282e+00
-4.95280504e-01 7.07691669e-01 7.64881819e-02 -1.09548628e+00
9.34250832e-01 2.02092975e-01 1.48706242e-01 -5.61442316e-01
7.69152939e-01 3.04690808e-01 -1.89780429e-01 -5.35779417e-01
3.76215816e-01 -1.60258353e-01 -3.52340907e-01 7.97312737e-01
-1.80065259e-01 2.31661305e-01 4.08265918e-01 6.13166809e-01
8.68498027e-01 -2.56758243e-01 9.72571075e-01 -5.70212185e-01
1.13743162e+00 3.28879148e-01 -2.97772318e-01 5.74794769e-01
-4.20044586e-02 -4.22894686e-01 8.39234114e-01 -3.65850240e-01
-5.01061857e-01 -5.34799159e-01 -2.24441245e-01 1.07110476e+00
-1.30290568e-01 -6.24457538e-01 -7.64842331e-01 -9.36080813e-01
-7.29750246e-02 8.55906427e-01 -8.42582107e-01 1.29262870e-02
-3.63719106e-01 -6.72788382e-01 6.74231052e-01 -3.39467585e-01
6.98198855e-01 -1.30436504e+00 -9.75462556e-01 4.68913913e-01
-2.25791186e-01 -6.35210156e-01 3.67744058e-01 4.08060364e-02
-5.20772338e-01 -1.21257114e+00 -3.67056370e-01 -3.98128003e-01
2.23758504e-01 -1.24935610e-02 1.06581914e+00 3.16714853e-01
-2.90186048e-01 4.13561575e-02 -8.08074296e-01 -5.17544806e-01
-1.17682457e+00 1.50997236e-01 -3.06343734e-01 1.35545462e-01
1.35785133e-01 -3.98239553e-01 -2.19940282e-02 -3.93000543e-01
-8.80252361e-01 1.59544960e-01 5.13552368e-01 7.39145219e-01
-3.27060878e-01 -6.91902488e-02 8.23373973e-01 -1.39399886e+00
1.25797129e+00 -7.65053570e-01 -2.91787237e-02 1.44754484e-01
-5.94119549e-01 -6.07356578e-02 6.08649552e-01 -2.74602413e-01
-1.49166584e+00 -7.12526858e-01 -3.95885408e-01 5.47839224e-01
-3.28004003e-01 1.07636666e+00 5.04549325e-01 1.34923607e-01
1.22342980e+00 1.50344819e-02 9.95027795e-02 -5.59075534e-01
5.61115742e-01 8.53691399e-01 -1.83607966e-01 -3.69393021e-01
8.82593274e-01 6.27892315e-01 -1.80902913e-01 -1.07594848e+00
-1.36396730e+00 -5.64241171e-01 -3.34886938e-01 -4.76750791e-01
8.50116730e-01 -4.03573155e-01 -5.99623203e-01 5.14520168e-01
-1.45294213e+00 2.40688577e-01 1.75697118e-01 2.23733470e-01
-1.75722733e-01 6.15094900e-01 -8.68785918e-01 -1.20143712e+00
-6.54896557e-01 -1.73594400e-01 4.99571055e-01 -4.09054495e-02
-5.06458402e-01 -1.42805886e+00 2.60507524e-01 4.56080675e-01
5.01812279e-01 5.82944810e-01 1.12163711e+00 -9.76811171e-01
4.02259529e-01 -1.95828546e-02 -1.83497429e-01 2.14410976e-01
2.24770442e-01 -6.88440427e-02 -6.71885073e-01 2.74282873e-01
3.56269330e-01 -6.82775378e-01 9.97837663e-01 1.33630484e-01
4.38070267e-01 -9.06970024e-01 -1.03757046e-01 -5.74090064e-01
1.17650163e+00 -3.81136626e-01 4.85274732e-01 6.48264170e-01
3.38649064e-01 9.74008322e-01 7.58366287e-01 8.84690166e-01
5.74763454e-02 4.99509484e-01 4.97182816e-01 4.59215418e-02
-7.20622530e-03 -2.65524238e-01 5.86044073e-01 8.71218562e-01
-2.58935422e-01 -1.16390049e-01 -8.40538025e-01 1.90800071e-01
-1.79199278e+00 -1.44512296e+00 -6.87044680e-01 1.61382854e+00
9.60188389e-01 2.45046213e-01 2.18218893e-01 4.61192727e-01
6.92684650e-01 8.64311635e-01 4.84915107e-01 -1.00191510e+00
-3.64699960e-01 1.58337474e-01 1.05417527e-01 5.83001316e-01
-1.24985719e+00 8.83289576e-01 4.45380688e+00 1.23211038e+00
-1.12496865e+00 2.85480410e-01 3.41479361e-01 4.11730647e-01
-3.60735595e-01 -2.14675933e-01 -6.18171811e-01 5.04912853e-01
8.40985835e-01 -3.11335713e-01 5.77031523e-02 7.82051742e-01
4.91345972e-01 -4.45658207e-01 -2.73708045e-01 3.89751077e-01
5.73625684e-01 -1.59755099e+00 -2.89675854e-02 4.18936089e-03
4.26834404e-01 -2.67301023e-01 -1.65760145e-01 1.22808896e-01
-2.21953154e-01 -6.05198324e-01 1.17697227e+00 1.07665978e-01
-8.04684907e-02 -4.77564037e-01 8.77570748e-01 6.95694447e-01
-4.58291978e-01 1.43929154e-01 1.54099479e-01 -3.55166763e-01
5.61166942e-01 1.07323682e+00 -7.16413260e-01 6.95002258e-01
2.23999053e-01 1.03238285e+00 -7.49721467e-01 2.92227030e-01
-8.13035131e-01 1.24039948e+00 3.08298916e-02 -8.39473248e-01
4.22510982e-01 -1.68128654e-01 8.41701925e-01 1.63238204e+00
-6.53196275e-02 -3.12060237e-01 -1.14308141e-01 8.07984829e-01
3.61399174e-01 9.76253986e-01 -6.10693574e-01 -5.83454013e-01
3.98228429e-02 1.47025716e+00 -7.95700908e-01 -3.49816203e-01
2.31427643e-02 5.07479906e-01 4.83420283e-01 -8.14938918e-02
-9.25699890e-01 -1.19300321e-01 -8.49572569e-02 2.07610652e-01
-7.78207630e-02 -1.26577049e-01 -2.02751309e-01 -1.06834757e+00
-1.32458091e-01 -6.31174207e-01 4.69278812e-01 -2.25399733e-01
-1.53665864e+00 4.94148463e-01 -2.36254372e-02 -5.40348291e-01
-3.48817348e-01 -4.79256094e-01 -6.97054088e-01 5.84337115e-01
-1.46999156e+00 -1.32851958e+00 5.15011558e-03 3.50525081e-01
4.00463164e-01 2.80955248e-02 6.83538735e-01 1.66785508e-01
-1.92838416e-01 -1.25049487e-01 -3.90454203e-01 1.66230816e-02
4.80769038e-01 -7.73592889e-01 -3.49590093e-01 4.92501378e-01
2.76588142e-01 5.17420113e-01 1.16162956e+00 -7.13701904e-01
-1.03302431e+00 -5.23866415e-01 1.53731287e+00 -2.48711348e-01
1.45310628e+00 -2.45128453e-01 -7.39860952e-01 5.26775829e-02
6.01029515e-01 -8.93183887e-01 6.99339807e-01 4.25311357e-01
-5.20202219e-01 7.52634108e-01 -1.18042946e+00 5.45433044e-01
9.35790181e-01 -3.06133568e-01 -9.43354964e-01 7.63413131e-01
6.53072521e-02 1.17007762e-01 -5.67912102e-01 -9.42075625e-02
3.86468768e-01 -1.07521868e+00 6.44895434e-01 -9.13273811e-01
1.16920674e+00 1.51736856e-01 3.14678907e-01 -1.36656213e+00
-1.23851568e-01 -4.60826010e-01 -1.18999444e-01 1.64749444e+00
4.94728893e-01 -8.42577934e-01 3.78017604e-01 -2.68773079e-01
1.62286222e-01 -4.91099298e-01 -8.99022818e-01 -4.46070284e-01
1.69100761e-01 -4.74509269e-01 -2.23857537e-01 1.25083983e+00
4.22440797e-01 7.73369074e-01 -5.36232889e-01 -4.90467131e-01
5.18518746e-01 1.02564581e-01 5.74170530e-01 -1.17399538e+00
-1.07455701e-02 -7.56584644e-01 -4.22913313e-01 -4.42329973e-01
4.64927286e-01 -1.00167906e+00 -2.80093580e-01 -1.24010658e+00
2.26484016e-01 -1.94033027e-01 2.18960702e-01 5.49937375e-02
9.88178551e-02 4.45481054e-02 1.02436714e-01 2.02011511e-01
-4.38356698e-01 4.51568037e-01 1.00991285e+00 -2.29813978e-01
-2.07373023e-01 -3.07796419e-01 -6.79702282e-01 1.14432657e+00
9.64240313e-01 -6.14335716e-01 2.35665902e-01 1.70115992e-01
7.25370467e-01 -3.19856107e-01 5.20465374e-01 -4.10126120e-01
-6.50445893e-02 -4.01680797e-01 -2.52966404e-01 -5.04640162e-01
5.93574420e-02 -3.11379880e-01 -2.82106847e-01 8.93073678e-01
-5.86043358e-01 -5.08621812e-01 -1.70234665e-01 6.87208593e-01
-1.89441934e-01 -6.24506891e-01 6.36262238e-01 2.21118078e-01
-3.27508032e-01 -6.05754137e-01 -1.00543296e+00 3.06666523e-01
9.21461165e-01 3.23953867e-01 -9.46487844e-01 -3.51213753e-01
-7.11636543e-01 -3.03896129e-01 2.29310486e-02 2.44056419e-01
1.44805521e-01 -1.08954179e+00 -1.11575985e+00 -3.56668353e-01
1.56483561e-01 -1.03668058e+00 2.39075541e-01 1.35407877e+00
-4.78761554e-01 4.01057452e-01 5.45706525e-02 -2.83799469e-01
-1.60784745e+00 5.15940011e-01 -1.38139322e-01 -5.95038772e-01
-6.16258323e-01 5.19420564e-01 -3.52668792e-01 1.24915138e-01
-1.99275702e-01 1.45827457e-01 -8.56028259e-01 7.27406383e-01
6.32692456e-01 4.89260197e-01 -5.75381145e-02 -9.69476938e-01
-4.37490076e-01 2.15786844e-01 4.26371489e-03 -6.44401163e-02
1.12542546e+00 -1.15353800e-01 -6.55132890e-01 5.25720417e-01
9.02344048e-01 7.50154436e-01 6.08892813e-02 -1.67751163e-01
5.70566058e-01 -4.34898317e-01 -3.50157470e-02 -9.56488550e-01
-3.81893456e-01 7.42435873e-01 -7.67280459e-02 1.30598879e+00
4.27494764e-01 4.66490418e-01 4.43276614e-01 2.48701885e-01
2.67834634e-01 -1.43146670e+00 2.34766722e-01 6.15818381e-01
1.34695983e+00 -9.25172031e-01 2.32627794e-01 -9.97065723e-01
-4.76372629e-01 1.22754681e+00 -9.20876041e-02 -1.70964181e-01
8.04014683e-01 -1.48718625e-01 1.73254192e-01 -7.71791697e-01
-5.40483415e-01 -1.20468609e-01 2.27444738e-01 8.72057676e-02
9.01465356e-01 2.43534088e-01 -1.68610466e+00 4.99824554e-01
-1.12748705e-01 -2.42970750e-01 6.09813571e-01 8.28798950e-01
-6.94829404e-01 -9.55899239e-01 -5.70342779e-01 4.30696726e-01
-8.23104441e-01 -1.37807743e-03 -1.07406104e+00 6.29495561e-01
1.06244303e-01 1.34404349e+00 -5.41475236e-01 -1.80417120e-01
-1.72133520e-01 -1.06133111e-01 3.93289983e-01 -5.93843579e-01
-1.24871325e+00 -1.26984790e-01 1.11156893e+00 -2.57614493e-01
-1.18168581e+00 -6.94103599e-01 -1.11690748e+00 -5.77196896e-01
-6.22493446e-01 4.56926376e-01 9.71144915e-01 1.40381575e+00
-4.67976816e-02 2.83478409e-01 7.52885699e-01 -5.81120610e-01
-7.82871485e-01 -1.27956784e+00 -4.65372711e-01 6.15896285e-01
-2.64525503e-01 -6.87250555e-01 -7.64008641e-01 -2.21245781e-01] | [8.548059463500977, 10.473657608032227] |
bee38dc0-a99a-4ebb-b99c-93c4627fee44 | panoptic-partformer-learning-a-unified-model | 2204.04655 | null | https://arxiv.org/abs/2204.04655v2 | https://arxiv.org/pdf/2204.04655v2.pdf | Panoptic-PartFormer: Learning a Unified Model for Panoptic Part Segmentation | Panoptic Part Segmentation (PPS) aims to unify panoptic segmentation and part segmentation into one task. Previous work mainly utilizes separated approaches to handle thing, stuff, and part predictions individually without performing any shared computation and task association. In this work, we aim to unify these tasks at the architectural level, designing the first end-to-end unified method named Panoptic-PartFormer. In particular, motivated by the recent progress in Vision Transformer, we model things, stuff, and part as object queries and directly learn to optimize the all three predictions as unified mask prediction and classification problem. We design a decoupled decoder to generate part feature and thing/stuff feature respectively. Then we propose to utilize all the queries and corresponding features to perform reasoning jointly and iteratively. The final mask can be obtained via inner product between queries and the corresponding features. The extensive ablation studies and analysis prove the effectiveness of our framework. Our Panoptic-PartFormer achieves the new state-of-the-art results on both Cityscapes PPS and Pascal Context PPS datasets with at least 70% GFlops and 50% parameters decrease. In particular, we get 3.4% relative improvements with ResNet50 backbone and 10% improvements after adopting Swin Transformer on Pascal Context PPS dataset. To the best of our knowledge, we are the first to solve the PPS problem via \textit{a unified and end-to-end transformer model. Given its effectiveness and conceptual simplicity, we hope our Panoptic-PartFormer can serve as a good baseline and aid future unified research for PPS. Our code and models are available at https://github.com/lxtGH/Panoptic-PartFormer. | ['DaCheng Tao', 'Yunhai Tong', 'Guangliang Cheng', 'Yibo Yang', 'Shilin Xu', 'Xiangtai Li'] | 2022-04-10 | null | null | null | null | ['part-level-panoptic-segmentation'] | ['computer-vision'] | [ 1.31115362e-01 2.08013281e-01 -1.69106528e-01 -5.10472298e-01
-9.02261972e-01 -4.86017644e-01 4.01725322e-01 -4.57844019e-01
-8.96799862e-02 2.94110030e-01 1.47416770e-01 -2.40142211e-01
2.28277713e-01 -6.34825170e-01 -9.87256587e-01 -5.98740876e-01
3.86101067e-01 5.60796916e-01 5.02622962e-01 -5.02136871e-02
-6.41452000e-02 2.13385418e-01 -1.53733325e+00 4.40991640e-01
9.41312373e-01 1.24580061e+00 7.08114862e-01 6.77760959e-01
-1.24422356e-01 7.42919207e-01 -2.13015378e-01 -7.04470515e-01
7.48990297e-01 1.24612570e-01 -1.06589210e+00 3.03532898e-01
8.44963551e-01 -2.57869571e-01 -3.82598579e-01 8.55755806e-01
4.58395660e-01 -1.81543455e-01 2.89515913e-01 -1.15574074e+00
-5.92749953e-01 7.79818892e-01 -6.68101132e-01 1.85345456e-01
-2.06076056e-01 4.57648277e-01 1.44966626e+00 -9.02446628e-01
3.12373459e-01 1.18290615e+00 7.03024626e-01 4.33349341e-01
-1.07889974e+00 -4.52696532e-01 5.92303574e-01 1.08149134e-01
-1.34615111e+00 -4.96190310e-01 3.52270007e-01 -2.06181094e-01
1.12885928e+00 4.96111333e-01 6.63110971e-01 6.99252129e-01
1.63270935e-01 1.44881201e+00 7.22625971e-01 6.57267496e-03
-1.54586732e-01 4.61883694e-02 2.47133106e-01 8.66342247e-01
-2.60712709e-02 -1.90502629e-01 -2.76446670e-01 3.16350788e-01
5.12260556e-01 1.51164457e-01 -4.20667857e-01 -9.75590423e-02
-1.15464759e+00 4.50747252e-01 6.61715090e-01 4.43738587e-02
-3.64187717e-01 5.63643217e-01 9.39467400e-02 6.89978376e-02
5.57913125e-01 1.91270664e-01 -1.01688004e+00 -5.37244081e-02
-1.03108680e+00 2.14740947e-01 9.66255784e-01 1.38417149e+00
8.48602116e-01 -1.92625433e-01 -4.35110986e-01 1.10487008e+00
5.77510893e-01 7.17898667e-01 2.40902975e-01 -1.01246727e+00
6.72800303e-01 3.80428731e-01 -2.08686199e-02 -4.28293645e-01
-2.58882880e-01 -7.38082170e-01 -5.43874383e-01 -3.23116153e-01
-1.34121049e-02 -2.06836984e-01 -1.48981774e+00 1.53275669e+00
4.05109435e-01 3.55897874e-01 -8.52784365e-02 9.41145480e-01
8.48673522e-01 7.27724850e-01 1.50625974e-01 3.16204876e-01
1.78865647e+00 -1.76483536e+00 -9.56985280e-02 -5.28205395e-01
4.85001743e-01 -1.17016196e+00 1.14588046e+00 1.68385461e-01
-1.04889572e+00 -7.60577023e-01 -8.62759113e-01 -2.98964858e-01
-5.17923310e-02 4.04135048e-01 5.55063009e-01 4.20453310e-01
-1.09207225e+00 4.10852909e-01 -1.04537511e+00 -3.80362153e-01
6.16043329e-01 5.31654119e-01 4.24380861e-02 -9.69824661e-03
-7.65622377e-01 6.00481212e-01 2.40129918e-01 1.57253653e-01
-7.19956040e-01 -9.01132822e-01 -6.29962683e-01 2.98473656e-01
5.43845356e-01 -1.26026559e+00 1.83885777e+00 -7.02850103e-01
-1.53254652e+00 7.84215093e-01 -1.22733027e-01 -7.68692791e-01
3.41989130e-01 -6.94958389e-01 -9.34593454e-02 1.03043646e-01
2.68246710e-01 1.23468626e+00 7.32395709e-01 -1.02438939e+00
-1.10904527e+00 -9.83134806e-02 7.95600489e-02 4.15810168e-01
-2.28940416e-02 -2.35401914e-01 -1.36535406e+00 -6.19879782e-01
3.02157253e-01 -1.12480223e+00 -2.85291135e-01 1.72769465e-02
-6.37488723e-01 -1.38046756e-01 7.31848121e-01 -4.76037323e-01
8.89771044e-01 -2.21698046e+00 -1.62293226e-01 -2.96658576e-01
3.58892739e-01 2.96143353e-01 -2.54575431e-01 1.38410851e-01
2.27032557e-01 7.01633692e-02 -4.86584485e-01 -8.27730596e-01
2.47227088e-01 3.15924734e-01 -6.55130804e-01 2.89741069e-01
1.67088583e-01 1.06422293e+00 -5.26266634e-01 -5.21789670e-01
3.75422806e-01 3.64783853e-01 -7.58897364e-01 1.01523250e-01
-5.31211436e-01 1.22230597e-01 -6.27584040e-01 9.62635994e-01
8.79098594e-01 -4.59969789e-01 -2.29410097e-01 -3.58236670e-01
-1.32959589e-01 5.60622990e-01 -7.68124282e-01 1.76598632e+00
-5.08370101e-01 5.10590136e-01 2.69076675e-01 -8.91604006e-01
5.51495552e-01 1.09577596e-01 4.34799790e-01 -6.14880562e-01
2.26577166e-02 2.45890856e-01 -2.43966103e-01 -4.02552545e-01
6.49020970e-01 -2.47662142e-03 -7.25540295e-02 -7.96380639e-02
1.39830634e-01 -3.80875766e-01 1.46772727e-01 4.19063605e-02
1.04094660e+00 3.27802420e-01 -2.86591891e-02 -2.27635458e-01
4.26174283e-01 1.53565854e-01 6.06886744e-01 7.90523827e-01
-1.98505387e-01 1.04407823e+00 2.87322551e-01 -4.02141422e-01
-7.14822948e-01 -1.24053919e+00 -2.32736245e-01 1.14082050e+00
3.72509927e-01 -4.90321994e-01 -6.18405223e-01 -8.93775344e-01
-9.01565105e-02 6.92619562e-01 -3.30710679e-01 3.59974891e-01
-7.63154805e-01 -9.04826283e-01 5.21506250e-01 5.95211208e-01
9.05353129e-01 -8.96714211e-01 -4.15140301e-01 2.62766957e-01
-3.17409307e-01 -1.46544540e+00 -7.98602819e-01 1.79247200e-01
-7.67329037e-01 -8.97387981e-01 -8.07513058e-01 -8.47601295e-01
3.23640227e-01 5.10725796e-01 1.25608397e+00 -1.50046647e-01
-2.17212722e-01 3.50758493e-01 -3.27908009e-01 -2.01517284e-01
9.17468667e-02 2.70648122e-01 -4.83418941e-01 3.28187197e-02
1.36523947e-01 -7.80934811e-01 -1.00826335e+00 5.03238082e-01
-7.07712948e-01 5.88575959e-01 1.02140784e+00 3.64411771e-01
9.60446417e-01 -3.00126672e-01 -3.50736305e-02 -8.10948491e-01
7.15330690e-02 -5.00907958e-01 -6.36477232e-01 3.11215729e-01
-5.85341692e-01 8.51341709e-02 5.53713024e-01 1.66911650e-02
-1.01802111e+00 2.51794666e-01 -4.92198229e-01 -5.46089232e-01
-2.15353921e-01 1.35980293e-01 -2.88085401e-01 3.04963857e-01
1.85975730e-01 4.59470034e-01 -5.06723464e-01 -7.56117642e-01
5.90594888e-01 7.27103829e-01 6.08860314e-01 -5.14978230e-01
5.84799945e-01 5.58481336e-01 -3.23891133e-01 -6.14484966e-01
-1.07784462e+00 -8.77778828e-01 -2.90224165e-01 1.47517875e-01
1.07148647e+00 -1.34689212e+00 -5.70816755e-01 5.31611502e-01
-1.12580860e+00 -4.70868886e-01 -2.78442502e-01 1.58144221e-01
-6.37699604e-01 4.34540838e-01 -5.86133957e-01 -3.60434234e-01
-9.07250345e-01 -1.24355125e+00 1.49822199e+00 2.38074094e-01
1.46749645e-01 -4.26604360e-01 -9.83845517e-02 7.38463461e-01
2.06555739e-01 -3.65109354e-01 3.52497309e-01 -5.35337269e-01
-1.05732560e+00 1.05990015e-01 -5.46036303e-01 3.96145374e-01
-8.17237645e-02 -3.32997777e-02 -1.03459024e+00 -8.61154776e-03
1.08208507e-01 -1.69523098e-02 1.28089011e+00 3.28195602e-01
1.27656591e+00 -2.96289504e-01 -4.56894696e-01 1.00458157e+00
1.49702430e+00 9.23343061e-04 5.34875572e-01 5.42372689e-02
9.14573073e-01 2.82231301e-01 5.99541187e-01 3.73410940e-01
7.26693094e-01 7.52459466e-01 4.48926032e-01 3.15309763e-02
-5.47606707e-01 -1.53261080e-01 3.70365560e-01 7.71449149e-01
1.91798791e-01 -6.15178287e-01 -1.04002893e+00 7.36477375e-01
-2.05160618e+00 -6.89728141e-01 -1.82964280e-01 1.65985215e+00
6.12428069e-01 2.31941789e-01 -1.94334894e-01 -4.37253058e-01
5.33329368e-01 3.61873567e-01 -4.95344013e-01 8.07029847e-03
2.02565044e-01 2.88596183e-01 6.61153018e-01 4.12462771e-01
-1.41708601e+00 1.42517447e+00 4.98669529e+00 1.00055635e+00
-1.17202067e+00 1.94251060e-01 8.40457141e-01 -3.15869957e-01
-1.39928311e-01 1.43905267e-01 -1.23709178e+00 4.21476871e-01
7.55193412e-01 3.78203988e-01 4.22516346e-01 1.02729809e+00
2.08404642e-02 -4.66290563e-02 -9.83360231e-01 9.99989212e-01
-1.48537576e-01 -1.27981853e+00 9.85254794e-02 -4.26868424e-02
6.02379501e-01 9.22449768e-01 7.88969547e-02 4.93586689e-01
3.58713597e-01 -6.55903399e-01 1.02085114e+00 3.28886926e-01
4.86496419e-01 -2.60845184e-01 5.77284873e-01 2.17563197e-01
-1.46175539e+00 -1.13196727e-02 -3.25197279e-01 1.96629241e-01
3.05864155e-01 5.70673227e-01 -8.27347398e-01 5.79406917e-01
9.13643539e-01 7.59422779e-01 -3.96552503e-01 1.29999065e+00
-2.94232994e-01 8.86560500e-01 -6.86660469e-01 2.64330178e-01
5.23242176e-01 -2.45876834e-02 5.84374368e-01 1.34971046e+00
2.81417072e-01 5.39608672e-02 4.09219354e-01 7.81330466e-01
-3.19885135e-01 -3.93732600e-02 4.16187160e-02 1.18527099e-01
2.16083899e-01 1.34871984e+00 -9.68572497e-01 -5.89351952e-01
-6.43164039e-01 1.17370033e+00 2.96938926e-01 2.78124571e-01
-1.14986396e+00 -8.04492831e-02 9.37927663e-01 -3.10443202e-03
1.09201610e+00 1.85546037e-02 -6.79946423e-01 -1.41659367e+00
1.53832361e-01 -6.63525343e-01 2.04648256e-01 -7.10936904e-01
-1.14040470e+00 8.04098248e-01 2.45215278e-02 -1.18667138e+00
1.76082745e-01 -5.09131491e-01 -5.04988492e-01 6.07440472e-01
-1.57517433e+00 -1.35705900e+00 -2.87860334e-01 4.26188350e-01
1.16942728e+00 1.72014698e-01 4.31550890e-01 3.29287678e-01
-7.55772889e-01 4.86145616e-01 -1.49637446e-01 2.34360714e-02
4.17281747e-01 -1.28037822e+00 9.02319312e-01 9.59753513e-01
4.75704879e-01 3.38432968e-01 6.12998366e-01 -4.08136696e-01
-1.15348732e+00 -1.33671927e+00 7.74635553e-01 -3.94755781e-01
5.57389677e-01 -4.11938727e-01 -3.96798581e-01 8.47209632e-01
2.36521557e-01 2.01055646e-01 2.90379941e-01 -1.05852000e-02
-1.62475556e-01 -4.03437763e-01 -8.30778182e-01 8.23818386e-01
1.27151990e+00 -2.22363755e-01 -6.09188735e-01 3.87733757e-01
1.26891983e+00 -5.40517867e-01 -5.96141696e-01 6.82607055e-01
4.58326578e-01 -1.02529335e+00 9.43338215e-01 9.88835022e-02
4.44213092e-01 -4.04861420e-01 -5.76983571e-01 -7.63515472e-01
-3.84471178e-01 -7.43984222e-01 -1.72718659e-01 1.05313492e+00
6.25950813e-01 -5.99820673e-01 8.80725622e-01 4.06989932e-01
-8.13711345e-01 -1.18336833e+00 -7.43924201e-01 -4.92680132e-01
-1.43182367e-01 -8.45924258e-01 7.14327097e-01 2.23215982e-01
-6.53035820e-01 4.89701092e-01 -2.81392008e-01 4.39472049e-01
2.35294968e-01 6.81958139e-01 8.70033681e-01 -6.80113375e-01
-7.87982702e-01 -5.26386738e-01 -6.11033067e-02 -1.95493090e+00
-1.53414518e-01 -9.28559065e-01 4.47298318e-01 -1.76939929e+00
2.13775456e-01 -5.20844579e-01 -3.34146947e-01 8.26223254e-01
-1.41925260e-01 4.19368714e-01 6.44455194e-01 4.73189086e-01
-6.83828831e-01 7.25987136e-01 1.40363181e+00 -3.18859130e-01
-2.84428269e-01 4.41154867e-01 -9.94905651e-01 7.16726243e-01
1.01679420e+00 -4.57571626e-01 -5.36827087e-01 -8.74573648e-01
-2.67969102e-01 -2.41325766e-01 3.27361107e-01 -1.30548060e+00
6.98122755e-02 -1.08288107e-02 5.90512529e-02 -8.41969609e-01
6.36681020e-01 -8.20215821e-01 -7.19631789e-03 4.02521551e-01
1.60909832e-01 -2.16496587e-01 3.62509131e-01 5.49221456e-01
-1.05084501e-01 4.29844446e-02 5.64610481e-01 -1.99715361e-01
-9.09952581e-01 5.86017132e-01 -1.68816492e-01 1.01634748e-02
8.96899760e-01 -1.27333522e-01 -4.01041269e-01 -2.79533882e-02
-7.64795721e-01 5.55533767e-01 2.14628115e-01 4.98560190e-01
4.63407427e-01 -6.88245475e-01 -7.06687629e-01 9.23243314e-02
-1.41561776e-01 3.99159282e-01 3.50260586e-01 1.14406335e+00
-7.40757167e-01 7.36942470e-01 1.55871227e-01 -7.60348976e-01
-1.21755493e+00 4.42048550e-01 4.55546975e-01 -3.89460057e-01
-8.32764030e-01 1.27266622e+00 7.41318107e-01 -5.42432487e-01
1.87744483e-01 -8.77291799e-01 1.59521073e-01 -2.59419739e-01
1.49740651e-01 8.41416270e-02 -3.02306097e-02 -5.17333925e-01
-3.62923533e-01 6.16333783e-01 -3.39662224e-01 -4.18657772e-02
1.44824398e+00 -1.69351235e-01 1.17513246e-03 -5.66355251e-02
1.09383392e+00 8.99920147e-03 -1.59793139e+00 -1.95752472e-01
-2.17174679e-01 -1.27265602e-01 -9.91066173e-02 -1.01377296e+00
-1.34297740e+00 7.58351505e-01 4.30694461e-01 9.59693268e-03
1.26975238e+00 3.35281193e-01 1.31145513e+00 3.48461986e-01
2.36646906e-01 -7.78679490e-01 -3.09275299e-01 5.97973347e-01
8.88880432e-01 -9.54632580e-01 -3.80557813e-02 -9.30966198e-01
-5.94833374e-01 6.38037503e-01 8.55627537e-01 -1.98138773e-01
7.01434016e-01 2.43244454e-01 6.89155236e-02 -1.91172600e-01
-9.46561635e-01 -5.81224561e-01 4.80621785e-01 1.44884005e-01
1.89396143e-01 2.02928051e-01 -1.68794930e-01 5.47914863e-01
-4.81755376e-01 -1.47283226e-01 -7.74557069e-02 7.34273672e-01
-5.74387431e-01 -1.13725197e+00 -1.38762638e-01 5.44857085e-01
-5.44222414e-01 -5.08486748e-01 -1.00446954e-01 6.67976439e-01
4.56216127e-01 7.96357930e-01 1.05031103e-01 -4.78354156e-01
2.33530223e-01 1.37218507e-02 3.36676121e-01 -7.32056558e-01
-7.20279217e-01 3.41130883e-01 2.41806462e-01 -7.13569224e-01
-2.06439555e-01 -4.39006448e-01 -1.40097260e+00 -1.04496993e-01
-2.19994426e-01 -1.61975726e-01 4.74166185e-01 9.34629798e-01
7.56874979e-01 5.21742105e-01 3.50649387e-01 -7.99503624e-01
-4.97321963e-01 -9.84875143e-01 -1.32383049e-01 2.67075095e-02
1.32222176e-01 -1.91363573e-01 -7.09997416e-02 5.67885600e-02] | [9.540616989135742, 0.2583373785018921] |
5965b15b-ceb2-456c-ab5d-47ab6fee8750 | data-summarization-via-bilevel-optimization | 2109.12534 | null | https://arxiv.org/abs/2109.12534v1 | https://arxiv.org/pdf/2109.12534v1.pdf | Data Summarization via Bilevel Optimization | The increasing availability of massive data sets poses a series of challenges for machine learning. Prominent among these is the need to learn models under hardware or human resource constraints. In such resource-constrained settings, a simple yet powerful approach is to operate on small subsets of the data. Coresets are weighted subsets of the data that provide approximation guarantees for the optimization objective. However, existing coreset constructions are highly model-specific and are limited to simple models such as linear regression, logistic regression, and $k$-means. In this work, we propose a generic coreset construction framework that formulates the coreset selection as a cardinality-constrained bilevel optimization problem. In contrast to existing approaches, our framework does not require model-specific adaptations and applies to any twice differentiable model, including neural networks. We show the effectiveness of our framework for a wide range of models in various settings, including training non-convex models online and batch active learning. | ['Andreas Krause', 'Marco Tagliasacchi', 'Mojmír Mutný', 'Zalán Borsos'] | 2021-09-26 | null | null | null | null | ['data-summarization'] | ['miscellaneous'] | [ 1.17327891e-01 -3.92753594e-02 -8.09047997e-01 -4.62014169e-01
-7.16970325e-01 -2.72891134e-01 4.95757647e-02 1.08482823e-01
-5.43979466e-01 6.64001226e-01 -1.15738235e-01 -1.33883357e-01
-5.42832196e-01 -6.15187585e-01 -8.89005303e-01 -6.82734430e-01
-1.69092566e-01 7.75636673e-01 -1.46430895e-01 2.43214890e-02
2.21766546e-01 2.89337516e-01 -1.41130435e+00 -1.16753906e-01
9.55047131e-01 1.21336222e+00 3.66646856e-01 4.75391716e-01
3.76273021e-02 5.59734285e-01 -2.28899509e-01 -7.26382956e-02
4.98030543e-01 3.61151487e-01 -6.77192450e-01 1.66102797e-01
2.69434810e-01 -1.16449445e-01 -1.75282285e-01 7.28896320e-01
3.08915585e-01 2.41523042e-01 2.51297206e-01 -1.51101685e+00
-2.49003112e-01 6.63062871e-01 -5.03104925e-01 8.16553012e-02
-2.71127850e-01 -2.31236547e-01 1.08903444e+00 -1.02558243e+00
5.55989183e-02 9.08881724e-01 6.20166600e-01 5.06801546e-01
-1.37713075e+00 -6.20905578e-01 4.76843417e-01 2.37923771e-01
-1.24719238e+00 -6.52660310e-01 5.86737692e-01 -2.66882002e-01
1.12434459e+00 4.47443217e-01 4.68016922e-01 5.92671335e-01
-3.26719493e-01 1.15308642e+00 9.82792079e-01 -5.31178832e-01
4.86994624e-01 2.14092255e-01 5.37001073e-01 4.77844208e-01
5.14477789e-01 -8.82420614e-02 -7.31869578e-01 -6.21435523e-01
4.92354155e-01 4.22935754e-01 6.46223873e-02 -6.18197262e-01
-9.94566202e-01 8.94394457e-01 1.50309548e-01 -1.61295503e-01
-2.54952848e-01 3.00560534e-01 3.26292753e-01 2.21925601e-01
6.48742974e-01 4.69303042e-01 -9.17378783e-01 -1.84955075e-01
-1.14911997e+00 1.51725411e-01 9.92897570e-01 1.26499212e+00
8.71119916e-01 7.83860758e-02 8.01570266e-02 8.43075097e-01
3.62812996e-01 2.78465837e-01 2.95754541e-02 -6.33192897e-01
5.12104928e-01 6.17268205e-01 1.77743390e-01 -7.51333237e-01
-4.37273294e-01 -4.71907288e-01 -8.10153127e-01 -2.48450115e-01
1.70012504e-01 -2.41813138e-01 -6.12163842e-01 1.67589116e+00
4.45466906e-01 2.55487651e-01 -5.12605190e-01 7.34885633e-01
3.43494624e-01 4.68062639e-01 5.53216040e-02 -4.91163760e-01
7.81099916e-01 -1.22530866e+00 -2.70560563e-01 -4.09256935e-01
5.93093514e-01 -5.05611122e-01 1.00612605e+00 4.34152931e-01
-1.24414337e+00 -1.99973032e-01 -8.84544790e-01 1.38640538e-01
-2.36542031e-01 1.44545898e-01 1.17393434e+00 6.43394411e-01
-9.14899170e-01 3.06988180e-01 -9.94003594e-01 -8.82059708e-02
5.56391299e-01 8.17998528e-01 8.08340386e-02 -2.63238728e-01
-6.82402611e-01 6.39391482e-01 4.67889965e-01 1.99059978e-01
-7.94360697e-01 -9.40040290e-01 -6.71175122e-01 3.47567797e-01
7.57574797e-01 -7.89723098e-01 1.20439565e+00 -1.06992269e+00
-1.39808178e+00 7.01879740e-01 -1.82812378e-01 -4.53447759e-01
3.59553277e-01 -2.29710877e-01 1.08501375e-01 -3.04273665e-01
-3.14745963e-01 1.51842952e-01 8.70054543e-01 -9.92152750e-01
-6.99009120e-01 -3.41396481e-01 3.20764452e-01 2.29052648e-01
-9.87406373e-01 2.30564758e-01 -7.67390668e-01 -4.77818608e-01
-2.54134804e-01 -1.10926449e+00 -7.78601229e-01 1.02406695e-01
-3.70047718e-01 -3.26738536e-01 8.14141273e-01 -4.28885706e-02
1.56180644e+00 -2.01504564e+00 2.37098172e-01 4.62244093e-01
4.54012990e-01 6.14008047e-02 -1.82585493e-02 2.27344275e-01
1.46592438e-01 5.42278308e-03 -1.07529491e-01 -6.66592836e-01
3.59179616e-01 3.97904932e-01 -2.59413540e-01 4.35324222e-01
7.93117806e-02 7.27845550e-01 -6.34030938e-01 -4.27666098e-01
1.85649730e-02 2.87867021e-02 -7.97328591e-01 3.61325473e-01
-4.90292430e-01 4.60501574e-03 -5.81717372e-01 6.38710260e-01
5.64350426e-01 -7.36857176e-01 2.13427648e-01 2.30547637e-01
1.91941485e-02 4.11314927e-02 -1.18134224e+00 1.46427286e+00
-6.05172336e-01 2.56874263e-01 5.39257646e-01 -1.59133136e+00
6.10683024e-01 1.00706480e-02 8.28999817e-01 -4.06443894e-01
-6.54200390e-02 1.64728120e-01 -2.23956004e-01 -3.89751583e-01
4.36866373e-01 5.89762330e-02 -2.71885246e-01 5.65428436e-01
-2.46040642e-01 4.31605697e-01 3.20818216e-01 1.44544438e-01
1.08723593e+00 -2.29297712e-01 2.31662109e-01 -3.98305833e-01
1.66521192e-01 2.14589551e-01 8.82597268e-01 7.68907905e-01
9.73259583e-02 2.90511370e-01 3.01835030e-01 -4.56009001e-01
-1.01217806e+00 -8.58672917e-01 -2.23249301e-01 1.71286845e+00
-3.76808569e-02 -5.39620280e-01 -5.60979068e-01 -5.23484945e-01
2.62241602e-01 2.27144420e-01 -3.39251101e-01 -5.99151663e-03
-6.62201703e-01 -1.18891251e+00 -9.86854136e-02 6.04740202e-01
9.26356465e-02 -5.45024693e-01 -5.26573360e-01 2.43465841e-01
2.19264567e-01 -1.11483204e+00 -6.59473002e-01 5.41078687e-01
-1.11724031e+00 -1.11720371e+00 -4.18548912e-01 -8.02445114e-01
1.01427901e+00 3.96088839e-01 1.42353225e+00 2.90973365e-01
-4.34893072e-01 5.57585299e-01 -1.63652211e-01 -5.27069986e-01
2.50266492e-01 3.30990851e-01 2.71639884e-01 -8.70877132e-02
3.51199508e-01 -3.13362598e-01 -6.73257768e-01 3.03568721e-01
-7.60215342e-01 1.64504245e-01 6.81443393e-01 9.70605016e-01
8.49875629e-01 9.73626822e-02 7.08703756e-01 -1.31614304e+00
5.53366065e-01 -7.23155797e-01 -7.77353168e-01 4.11114782e-01
-8.61678123e-01 -4.64520492e-02 1.02173042e+00 -6.58280849e-01
-6.81221426e-01 4.34126437e-01 3.36077124e-01 -5.03429234e-01
2.11000919e-01 6.77087724e-01 -2.33059123e-01 -1.16824239e-01
3.89144361e-01 7.89913461e-02 -2.71156698e-01 -4.32615012e-01
1.18168481e-01 6.43910766e-01 -3.28208730e-02 -9.08798754e-01
8.67400050e-01 4.25951302e-01 -6.20469153e-02 -9.49576795e-01
-9.69973922e-01 -7.04317451e-01 -4.43874896e-01 -1.62099290e-03
1.63582057e-01 -8.71350527e-01 -8.40297818e-01 1.20539606e-01
-6.40661597e-01 -6.08783484e-01 -2.07800671e-01 3.99717063e-01
-6.95573688e-01 3.43860358e-01 -3.41487348e-01 -8.86827946e-01
-7.91547418e-01 -1.08993983e+00 7.79831231e-01 1.50619209e-01
3.73810646e-03 -9.97115016e-01 4.38275002e-02 6.09670877e-01
6.40480816e-01 -6.83472008e-02 1.05596912e+00 -5.30012071e-01
-8.15138638e-01 -3.48826528e-01 -4.96490188e-02 2.90211469e-01
-1.16077393e-01 -1.15341157e-01 -6.85752869e-01 -8.61500323e-01
-2.18750447e-01 -6.85381532e-01 8.51120651e-01 5.61448693e-01
1.92473161e+00 -5.38523197e-01 -5.07580698e-01 8.86557639e-01
1.44470859e+00 -1.05476759e-01 2.08163172e-01 2.44946450e-01
5.17751455e-01 2.16090277e-01 7.30674267e-01 8.72576118e-01
3.90939265e-01 5.51832914e-01 4.76151317e-01 -3.44075978e-01
7.54525363e-01 2.02079445e-01 1.42119482e-01 9.35080826e-01
-2.54979432e-02 -1.10072888e-01 -8.75422835e-01 3.95018816e-01
-2.21947384e+00 -7.50063062e-01 2.93499559e-01 2.52725434e+00
9.32152033e-01 2.46371582e-01 3.02975476e-01 -1.63391661e-02
5.74572980e-01 -6.52953163e-02 -8.90464246e-01 -3.01398903e-01
1.46472186e-01 3.01885456e-01 7.12191641e-01 -4.22229916e-02
-1.24472332e+00 6.71853185e-01 7.21450090e+00 7.82412946e-01
-1.09716117e+00 1.42966196e-01 7.91844010e-01 -7.78202474e-01
-1.00009397e-01 6.52857423e-02 -9.79766667e-01 6.58593237e-01
9.37192500e-01 -3.97873819e-01 7.86936522e-01 1.41685069e+00
3.96336257e-01 6.47690892e-02 -1.28795207e+00 1.17305040e+00
-1.59375638e-01 -1.38636672e+00 -4.96242672e-01 1.36634290e-01
8.99782181e-01 2.43877366e-01 9.38536078e-02 5.43690085e-01
4.94460911e-01 -1.24099028e+00 4.56167310e-01 2.90547013e-01
8.24734986e-01 -7.25521982e-01 4.91410047e-02 6.69896185e-01
-1.16960824e+00 -4.60805953e-01 -5.78860343e-01 -1.67007208e-01
-9.10360441e-02 7.86363363e-01 -5.67472696e-01 -9.41560417e-03
8.28720152e-01 4.95124698e-01 -2.32013822e-01 1.27148223e+00
4.15298313e-01 7.25774646e-01 -6.78326726e-01 -1.83127970e-01
-5.66744544e-02 -1.27695009e-01 8.06526169e-02 1.27109933e+00
2.35821959e-02 1.14486367e-02 5.50560236e-01 6.00207865e-01
-3.14311087e-01 1.08581118e-01 -2.57297277e-01 -1.20000035e-01
7.56976962e-01 1.48761535e+00 -4.48598415e-01 -2.04666600e-01
-8.28600407e-01 3.81178617e-01 8.43292296e-01 2.05467343e-01
-6.84387147e-01 -3.17508504e-02 6.26209199e-01 1.47826105e-01
1.71041310e-01 -4.54875737e-01 -5.21178424e-01 -1.35475719e+00
1.20446868e-01 -9.22799528e-01 6.76470459e-01 -1.58553451e-01
-1.50848997e+00 1.66570172e-01 7.96242133e-02 -1.06837964e+00
-1.67199284e-01 -7.84080088e-01 -7.23635077e-01 5.70678532e-01
-1.57602406e+00 -9.56343472e-01 -3.43844593e-01 7.14247465e-01
5.63731551e-01 -2.15989932e-01 6.76710665e-01 4.20260310e-01
-9.10705328e-01 7.42699742e-01 5.29763222e-01 -1.21754043e-01
5.10430753e-01 -1.25239348e+00 9.72060040e-02 6.34064615e-01
-1.73841521e-01 8.40490580e-01 6.02354884e-01 -2.62449861e-01
-2.12273526e+00 -1.11529088e+00 5.22269487e-01 -1.19184718e-01
6.51851833e-01 -5.96220970e-01 -6.31486416e-01 6.44463062e-01
-2.87830263e-01 3.61439437e-01 1.04821181e+00 6.24930859e-01
-2.01037034e-01 -3.76392007e-01 -8.77020717e-01 3.58809978e-01
1.03033710e+00 -1.56620532e-01 2.92257965e-02 5.71458280e-01
4.85546887e-01 -3.41793418e-01 -9.57232356e-01 4.23506409e-01
3.71310055e-01 -3.73956144e-01 1.05545235e+00 -1.16954470e+00
2.89616078e-01 1.42807364e-01 -1.98840708e-01 -1.06252396e+00
-3.79377425e-01 -7.59171963e-01 -7.36304522e-01 7.57149696e-01
5.98200679e-01 -4.47932273e-01 1.26446474e+00 1.24594080e+00
-1.86679676e-01 -1.26153898e+00 -8.78094316e-01 -7.10410118e-01
6.30973428e-02 -3.83005053e-01 3.47762167e-01 1.06020737e+00
2.32409805e-01 1.88684538e-01 -5.70314825e-01 -8.66490752e-02
1.01251125e+00 3.34229857e-01 9.22865927e-01 -1.30318463e+00
-8.10584307e-01 -2.74971157e-01 2.07263846e-02 -1.28808701e+00
2.49436945e-01 -7.93879926e-01 -1.66500676e-02 -1.25139821e+00
6.39445782e-01 -1.14031875e+00 -4.29778010e-01 6.20417774e-01
-1.14814736e-01 -1.43817961e-01 1.65382311e-01 3.61658752e-01
-1.00239909e+00 3.78061563e-01 8.21293533e-01 -1.74340680e-01
-1.43211171e-01 3.95234853e-01 -7.77732849e-01 5.76637089e-01
7.52312422e-01 -4.01870340e-01 -4.91973042e-01 -6.05503440e-01
5.88643312e-01 5.76939471e-02 3.29825208e-02 -5.60925424e-01
5.50059378e-01 -6.76438689e-01 2.15025082e-01 -3.71954590e-01
2.97990531e-01 -1.08378911e+00 3.57033387e-02 3.20255011e-01
-3.59263688e-01 1.05719961e-01 3.76142636e-02 6.20244443e-01
7.19361156e-02 -1.32291257e-01 8.34396839e-01 -1.24970630e-01
-5.59968054e-01 8.52691650e-01 -6.36531562e-02 3.04074675e-01
1.06192291e+00 -2.02567518e-01 -4.03712690e-01 -1.56106725e-02
-5.63460469e-01 6.90156400e-01 4.04650509e-01 3.26636821e-01
4.71958935e-01 -1.08829093e+00 -5.07511795e-01 4.36210781e-02
2.03131825e-01 4.60287154e-01 -5.82751585e-03 1.00297999e+00
-1.95401490e-01 5.41470647e-01 1.91035748e-01 -4.66063172e-01
-1.24635696e+00 7.33579457e-01 1.55780464e-01 -3.36110443e-01
-2.42328778e-01 6.17478788e-01 1.30031168e-01 -3.18631232e-01
4.83926386e-01 -1.15532696e-01 1.74919367e-01 -2.98375487e-01
4.03907418e-01 4.66614485e-01 1.48658246e-01 -2.20657125e-01
-2.77631789e-01 1.85312331e-01 -3.16656053e-01 5.05978465e-01
1.76020861e+00 4.82030660e-02 -1.53024629e-01 3.01832944e-01
1.18282282e+00 -3.61293435e-01 -1.33655560e+00 -6.75379395e-01
1.37875393e-01 -4.54764664e-01 2.43756071e-01 -3.71897936e-01
-1.26748300e+00 4.88464683e-01 3.28573853e-01 1.13848753e-01
1.24994099e+00 -2.49346681e-02 6.70945108e-01 7.43064940e-01
5.85994005e-01 -1.68402970e+00 3.05100940e-02 5.75153589e-01
3.48150551e-01 -1.44355464e+00 1.54682443e-01 -3.82921726e-01
-2.26979524e-01 1.08489692e+00 8.11643243e-01 -1.15286022e-01
9.25216913e-01 6.24071360e-01 -4.70792502e-01 -4.10653986e-02
-1.27900660e+00 -8.29295348e-03 1.58054307e-01 2.34715790e-01
3.66217762e-01 1.33083344e-01 -3.04683357e-01 7.70412743e-01
7.27614481e-03 1.38731584e-01 1.82335421e-01 1.22845459e+00
-5.63631773e-01 -1.01125300e+00 -2.79511541e-01 1.06572592e+00
-3.20675701e-01 -1.17670037e-01 -6.10450841e-02 4.07529324e-01
-1.87682420e-01 8.64421725e-01 3.00881237e-01 -3.10075581e-01
1.03322119e-01 -2.68310815e-01 3.55559260e-01 -8.48272383e-01
-5.52271008e-01 -2.06406340e-02 6.06404468e-02 -3.50200802e-01
-2.97789127e-01 -5.45570731e-01 -1.01468647e+00 -6.69916511e-01
-6.59248888e-01 6.72870725e-02 7.04308867e-01 9.48945403e-01
4.44330752e-01 1.23706259e-01 9.95090187e-01 -7.28689253e-01
-1.03088415e+00 -5.84795833e-01 -6.59440279e-01 2.10940227e-01
1.93440095e-02 -5.82683027e-01 -3.60637397e-01 -1.07523046e-01] | [8.41159439086914, 4.12514591217041] |
71489772-2f3d-4740-95f5-53c3514fb12e | applicaai-at-semeval-2020-task-11-on-roberta | 2005.07934 | null | https://arxiv.org/abs/2005.07934v2 | https://arxiv.org/pdf/2005.07934v2.pdf | ApplicaAI at SemEval-2020 Task 11: On RoBERTa-CRF, Span CLS and Whether Self-Training Helps Them | This paper presents the winning system for the propaganda Technique Classification (TC) task and the second-placed system for the propaganda Span Identification (SI) task. The purpose of TC task was to identify an applied propaganda technique given propaganda text fragment. The goal of SI task was to find specific text fragments which contain at least one propaganda technique. Both of the developed solutions used semi-supervised learning technique of self-training. Interestingly, although CRF is barely used with transformer-based language models, the SI task was approached with RoBERTa-CRF architecture. An ensemble of RoBERTa-based models was proposed for the TC task, with one of them making use of Span CLS layers we introduce in the present paper. In addition to describing the submitted systems, an impact of architectural decisions and training schemes is investigated along with remarks regarding training models of the same or better quality with lower computational budget. Finally, the results of error analysis are presented. | ['Filip Graliński', 'Łukasz Borchmann', 'Dawid Jurkiewicz', 'Izabela Kosmala'] | 2020-05-16 | null | https://aclanthology.org/2020.semeval-1.187 | https://aclanthology.org/2020.semeval-1.187.pdf | semeval-2020 | ['propaganda-span-identification'] | ['natural-language-processing'] | [ 3.12548161e-01 2.10097402e-01 -1.13205798e-01 -1.54112056e-01
-6.88220084e-01 -4.66809154e-01 1.25257444e+00 2.54573345e-01
-4.38168883e-01 6.62982166e-01 4.11182195e-01 -5.81513941e-01
-9.06172693e-02 -7.83184409e-01 -2.31867388e-01 -6.38461649e-01
7.79134631e-02 5.99116802e-01 6.60899878e-02 -3.53389353e-01
7.37631083e-01 6.08621657e-01 -1.12765169e+00 7.49053240e-01
7.55182505e-01 8.54104042e-01 2.32853204e-01 5.81635892e-01
-4.50745255e-01 1.47260952e+00 -9.46408629e-01 -1.79117441e-01
-8.06192309e-02 -5.12663841e-01 -1.25959265e+00 -1.24627072e-02
5.49083769e-01 5.67756146e-02 2.07955809e-03 5.29039502e-01
2.32408077e-01 -1.76112637e-01 1.22059596e+00 -5.87327361e-01
-2.85305858e-01 1.32500494e+00 -2.34510794e-01 2.59708494e-01
3.95587265e-01 -4.51396346e-01 6.46041870e-01 -7.35666037e-01
9.70437765e-01 1.01529968e+00 6.38694644e-01 3.70814919e-01
-1.18327987e+00 -2.30811402e-01 -4.36976045e-01 3.47466290e-01
-1.27809250e+00 -5.39384127e-01 8.86161089e-01 -8.68829012e-01
1.41052067e+00 1.32437140e-01 5.01297474e-01 1.30328035e+00
3.21854979e-01 6.27657890e-01 1.61163199e+00 -1.15519178e+00
2.03758001e-01 5.98787308e-01 5.58315098e-01 8.78306925e-01
-1.48386195e-01 -1.39909433e-02 -5.04668534e-01 -3.80433619e-01
2.30789319e-01 -7.01418996e-01 3.49341214e-01 2.29818836e-01
-8.54112506e-01 9.55266237e-01 1.08640343e-01 1.12595105e+00
-2.50555724e-01 -9.79221985e-02 8.94700050e-01 2.07651898e-01
5.84677279e-01 4.00174290e-01 -1.40811488e-01 -2.44011879e-02
-1.56818068e+00 3.25856149e-01 9.62767482e-01 1.00755584e+00
3.47801358e-01 4.74504918e-01 -3.86119634e-01 6.77887261e-01
-1.80259161e-02 5.18721282e-01 1.42040521e-01 -2.81256847e-02
6.67459786e-01 4.56977695e-01 -2.87681580e-01 -9.35589254e-01
-5.67799389e-01 -5.36839485e-01 -6.74284160e-01 1.27782226e-01
4.57135379e-01 1.03965081e-01 -9.18043315e-01 1.20263052e+00
-2.48562302e-02 -5.51406085e-01 -1.65492699e-01 2.53133923e-01
7.11141825e-01 5.00376105e-01 3.82725656e-01 -4.84492600e-01
1.63601649e+00 -8.63095999e-01 -6.76185131e-01 1.75602719e-01
9.84360635e-01 -1.16570246e+00 4.84437644e-01 4.15582925e-01
-8.45184326e-01 -6.44148469e-01 -8.71822298e-01 -3.74210067e-02
-4.07615274e-01 6.15312576e-01 3.02337795e-01 9.56129551e-01
-6.18779123e-01 6.36504710e-01 -3.77342880e-01 -5.51418066e-01
2.22160846e-01 -1.98586062e-01 -1.22982919e-01 5.86815953e-01
-1.17369521e+00 1.67776430e+00 5.96256196e-01 -9.43144485e-02
-1.08554363e+00 -3.60164315e-01 -7.31300831e-01 -2.83471476e-02
1.71666056e-01 -2.20049232e-01 1.02375293e+00 -8.62834096e-01
-1.39140177e+00 1.21903861e+00 5.82003826e-03 -8.02590728e-01
4.40917730e-01 -3.11696291e-01 -8.63750875e-01 2.62439400e-01
9.70161408e-02 -1.44747142e-02 1.16685724e+00 -1.00081444e+00
-4.17753786e-01 -1.96848214e-01 -3.00027132e-01 -3.88415337e-01
7.99436681e-03 6.70744121e-01 5.17086565e-01 -6.64293110e-01
-2.45794773e-01 -5.55314422e-01 2.92750210e-01 -8.09310615e-01
-7.39402354e-01 -5.24204433e-01 6.58434451e-01 -1.25171554e+00
1.82384777e+00 -1.80965137e+00 1.18621759e-01 3.49843830e-01
-9.99597646e-03 4.36955959e-01 1.22372046e-01 1.09006047e+00
-7.08411187e-02 2.32452840e-01 -1.37353301e-01 -1.54243246e-01
-9.74452719e-02 -1.55300081e-01 -7.80583262e-01 5.73199689e-01
2.23918870e-01 3.83832544e-01 -4.37301964e-01 -1.09897423e+00
2.54409999e-01 2.02860206e-01 2.59743957e-03 1.40684515e-01
-3.71857345e-01 2.50022054e-01 -5.31666934e-01 5.21172225e-01
5.20399630e-01 1.68645874e-01 3.52502882e-01 -1.30507246e-01
-6.30451500e-01 7.74506688e-01 -6.73596084e-01 1.73640978e+00
-7.84502625e-01 6.58596218e-01 -8.66350085e-02 -9.07810628e-01
1.24287450e+00 4.09512788e-01 2.90134758e-01 -6.94244564e-01
3.55909377e-01 3.21463466e-01 -6.12113662e-02 -6.88015103e-01
8.49766970e-01 -1.88070789e-01 -2.97473401e-01 5.25713742e-01
4.69551235e-01 1.13786146e-01 5.29859245e-01 3.68692309e-01
1.15558863e+00 4.73971725e-01 8.92765701e-01 -5.50385237e-01
9.17071760e-01 5.36131799e-01 9.97027010e-02 6.98862612e-01
1.98331505e-01 7.46156350e-02 4.36760098e-01 -5.41102171e-01
-1.23772204e+00 -9.44599807e-01 -5.74100852e-01 1.05744600e+00
-5.40329874e-01 -7.78218389e-01 -6.96239889e-01 -1.02899992e+00
-6.38271987e-01 1.26520920e+00 -7.53212512e-01 4.46026951e-01
-1.06135905e+00 -3.99466842e-01 1.12549210e+00 1.26637906e-01
4.53661382e-01 -1.38153958e+00 -9.53463674e-01 5.09319603e-01
-1.14007004e-01 -6.46956146e-01 2.21039578e-01 4.08688039e-01
-6.01811051e-01 -1.06642640e+00 -5.08144200e-01 -8.94279480e-01
2.23827809e-01 -1.89607516e-01 1.16141641e+00 6.67163059e-02
-3.60738099e-01 -3.47269885e-02 -8.55233788e-01 -3.09475571e-01
-1.18390715e+00 4.18287754e-01 -2.49665841e-01 -5.10745466e-01
2.37308979e-01 -4.36293244e-01 1.71334460e-01 -1.54397175e-01
-6.79084957e-01 2.13357612e-01 3.70496839e-01 6.59086823e-01
-7.13775307e-02 -6.55208752e-02 4.90189373e-01 -1.06281102e+00
5.86388290e-01 -3.13379198e-01 -3.61648232e-01 4.98490036e-01
-5.98340392e-01 2.42267311e-01 8.41433883e-01 -1.56095952e-01
-1.32959628e+00 3.03226691e-02 -4.02392834e-01 2.53266156e-01
-3.83147418e-01 2.87617356e-01 1.34879453e-02 1.83212042e-01
1.29249215e+00 1.85066283e-01 -4.09134448e-01 -8.24506283e-01
3.51396471e-01 8.89799058e-01 8.06014091e-02 -6.50003552e-01
5.48080266e-01 1.48189873e-01 -5.98693527e-02 -1.02139843e+00
-7.83992171e-01 -6.23833477e-01 -8.96665037e-01 -3.74108613e-01
8.63645077e-01 -4.58313316e-01 -1.83888793e-01 3.28276008e-01
-1.54988408e+00 -4.60615046e-02 -2.18688294e-01 2.00418904e-01
-7.37916470e-01 4.87401426e-01 -6.29297256e-01 -1.30765045e+00
-8.05018842e-01 -7.23860383e-01 8.27605307e-01 -1.17273219e-01
-4.07642633e-01 -9.92799282e-01 4.13102955e-01 2.89307654e-01
3.24589252e-01 1.57802388e-01 1.32104552e+00 -9.23761129e-01
2.58918107e-01 -3.43164988e-02 -7.82289505e-02 2.58642852e-01
-1.74687967e-01 -8.83959234e-02 -9.10138369e-01 2.21157938e-01
2.40426034e-01 -3.99188250e-01 1.01339221e+00 1.27163962e-01
5.98172426e-01 -4.98660147e-01 -2.25733772e-01 4.01405692e-02
1.55724216e+00 1.29967213e-01 7.71747887e-01 6.56241477e-01
3.77942234e-01 8.26500654e-01 7.33897328e-01 3.54111046e-01
-1.39917612e-01 9.70364928e-01 -1.28517570e-02 2.33869135e-01
-4.70658362e-01 -3.40039164e-01 6.87467575e-01 6.03495955e-01
-3.11259359e-01 -2.27902114e-01 -1.04154551e+00 5.25673866e-01
-1.50865185e+00 -1.60429347e+00 -7.52286613e-01 1.85372126e+00
6.91653788e-01 1.92469895e-01 3.99014503e-01 4.58730161e-01
6.19582891e-01 4.16255921e-01 6.14590347e-01 -1.17007113e+00
-1.57790527e-01 5.43027639e-01 2.00844198e-01 5.02171397e-01
-1.30957115e+00 9.64089155e-01 6.25390816e+00 1.59246075e+00
-9.41094637e-01 4.20681626e-01 4.69444972e-03 2.47634932e-01
5.11963740e-02 2.85043120e-01 -1.04930031e+00 3.69018525e-01
1.22132730e+00 2.58538812e-01 2.63022512e-01 9.23370659e-01
3.06416988e-01 -4.50475395e-01 -7.87517130e-01 3.27479243e-01
4.05708641e-01 -1.58245885e+00 8.25673714e-02 9.91125312e-03
3.27033460e-01 -1.64020762e-01 -4.93103802e-01 4.47983325e-01
7.83000812e-02 -9.58072603e-01 1.30207896e+00 4.73365933e-01
6.84767902e-01 -4.80641723e-01 5.53097486e-01 5.90987682e-01
-1.10749853e+00 2.07962543e-01 -2.41548866e-01 -7.92090297e-02
6.49502754e-01 7.38388717e-01 -1.11013806e+00 1.02516019e+00
-1.17107271e-03 3.06748211e-01 -7.73106277e-01 7.12012529e-01
-6.41320050e-01 1.14254820e+00 -1.37655765e-01 -2.59035707e-01
3.20262939e-01 -1.18831666e-02 7.53753006e-01 2.00843787e+00
1.35864928e-01 -4.32272911e-01 1.30118072e-01 9.13373828e-01
5.75493574e-01 5.52421749e-01 -5.23505926e-01 -8.72716904e-02
2.82454133e-01 1.35076690e+00 -7.11600542e-01 -4.15589005e-01
3.49517375e-01 6.38766110e-01 4.84571934e-01 -2.34611034e-01
-9.24221694e-01 -1.70710489e-01 -5.10349214e-01 5.42862952e-01
4.05679084e-02 -2.73576707e-01 -5.02829611e-01 -8.60086977e-01
-1.70973197e-01 -4.80911642e-01 5.15470505e-01 -4.67343003e-01
-1.16273189e+00 1.01282561e+00 3.50765735e-01 -1.06047809e+00
-6.36284351e-01 -5.92928171e-01 -7.29378641e-01 8.37561846e-01
-7.76442409e-01 -1.85328186e+00 2.71501511e-01 2.35051498e-01
7.22535610e-01 -5.47045767e-01 1.09833205e+00 4.31666337e-02
-7.01987594e-02 1.28239349e-01 -2.62557678e-02 1.35135233e-01
5.20331144e-01 -1.07649493e+00 -9.71799940e-02 8.73471677e-01
4.84102458e-01 5.50542176e-01 1.05308533e+00 -8.26624513e-01
-6.53899491e-01 -9.51238275e-01 1.68798912e+00 -3.02040011e-01
9.06723797e-01 -4.69716340e-01 -6.69427633e-01 2.44597301e-01
5.77860594e-01 -8.56976092e-01 4.53900218e-01 4.18891370e-01
-7.52233446e-01 4.86348569e-01 -1.18953276e+00 1.84281990e-02
9.57324147e-01 -5.91432631e-01 -1.01756430e+00 7.07009852e-01
1.36061758e-01 -1.97327379e-02 -6.68232620e-01 -1.52214840e-01
3.30668181e-01 -8.36693466e-01 6.00507021e-01 -6.01906836e-01
9.20882642e-01 -1.30919561e-01 -3.38064916e-02 -8.87006164e-01
-4.38380927e-01 -2.60812730e-01 -1.46563977e-01 1.59071600e+00
3.54592830e-01 -7.13175833e-02 3.59394670e-01 -4.77078736e-01
-1.43844962e-01 -3.11635435e-01 -1.22826767e+00 -1.03013146e+00
2.45998934e-01 -4.27992016e-01 -9.06264335e-02 8.01108897e-01
3.93960774e-01 9.85618353e-01 -8.24601769e-01 -3.71810913e-01
5.99520087e-01 2.36501127e-01 4.12636340e-01 -8.01029027e-01
-4.02787149e-01 -4.12859887e-01 -2.50062853e-01 -4.34166580e-01
2.78416991e-01 -1.32348132e+00 -3.44820544e-02 -1.56142879e+00
2.40355909e-01 -5.28544843e-01 2.54955769e-01 3.86472553e-01
4.87218946e-01 -6.27502277e-02 3.35383475e-01 2.29092762e-01
-2.47616008e-01 2.17476994e-01 4.35953051e-01 -2.16775641e-01
-1.91058874e-01 1.03913978e-01 -8.45426992e-02 5.75266898e-01
1.08772337e+00 -8.80146444e-01 -6.57533878e-04 -2.36976862e-01
2.58502185e-01 -1.21655585e-02 4.97895777e-01 -1.05844605e+00
1.39121994e-01 -3.96454856e-02 1.60703734e-01 -1.09785032e+00
-3.94758657e-02 -5.38860023e-01 2.29729414e-01 7.05209017e-01
-3.89395952e-01 -1.25895336e-01 4.16167192e-02 1.82733968e-01
7.33281225e-02 -1.07634032e+00 8.38583231e-01 -8.05482864e-02
-5.05550086e-01 -5.42337477e-01 -1.04696608e+00 -2.24683747e-01
8.70726347e-01 -7.06104338e-02 -5.23311913e-01 2.72850573e-01
-8.24982643e-01 -4.65503097e-01 1.55179948e-01 9.20702145e-02
2.07739264e-01 -9.18695569e-01 -1.01788116e+00 -3.41451168e-01
1.85037792e-01 -6.64821029e-01 3.35382521e-01 9.58449662e-01
-7.25322604e-01 6.51345074e-01 -4.13391590e-01 -1.89754307e-01
-1.57815504e+00 7.30324984e-01 -1.11972064e-01 -1.11482155e+00
-7.12489188e-01 3.52141112e-01 -4.67842191e-01 -5.07461429e-02
-1.83186293e-01 2.49098092e-01 -5.85933745e-01 2.33783677e-01
5.86647272e-01 6.97963536e-01 4.33433622e-01 -8.76641810e-01
-4.87233102e-01 6.60246536e-02 -1.99729800e-01 -4.34725016e-01
1.35847175e+00 2.72218466e-01 -3.07764649e-01 3.15377355e-01
7.59640515e-01 4.97477591e-01 -1.48044914e-01 -1.09446026e-01
4.89006996e-01 1.33779883e-01 4.79504392e-02 -1.23542011e+00
-1.92589402e-01 1.03426969e+00 3.61971438e-01 3.90143752e-01
8.45206976e-01 -8.07352364e-02 3.11155289e-01 1.59648091e-01
6.73891068e-01 -1.39621806e+00 -3.74745071e-01 7.83277631e-01
1.12907553e+00 -4.16908890e-01 3.56381685e-01 -4.32097048e-01
-4.91738379e-01 1.44878840e+00 7.12026060e-02 -3.21700275e-01
2.86348522e-01 4.96908754e-01 -3.87275308e-01 -4.62922335e-01
-5.02209008e-01 -3.16557825e-01 3.54638636e-01 8.25347245e-01
6.38904691e-01 1.52867585e-01 -1.24577463e+00 3.82470667e-01
-3.37509990e-01 -1.33956969e-01 2.70320892e-01 9.57675874e-01
-6.90151632e-01 -1.25374579e+00 -5.57912230e-01 5.32256186e-01
-4.64844942e-01 -5.12073755e-01 -8.11897218e-01 8.79325867e-01
4.38210875e-01 1.21346796e+00 -1.85014069e-01 -5.61549366e-01
5.78774326e-02 2.95424014e-01 9.17933583e-01 -7.31841981e-01
-1.65136683e+00 1.76778033e-01 9.41570461e-01 2.51108781e-02
-6.26570761e-01 -6.52442396e-01 -8.52596164e-01 -4.57626462e-01
-4.60888028e-01 3.99464995e-01 7.73907244e-01 1.17101300e+00
-3.66239011e-01 4.03137833e-01 4.54406738e-01 -5.10939538e-01
-7.56270409e-01 -1.57833672e+00 -7.87224472e-01 3.75623144e-02
-3.79514039e-01 -4.46630895e-01 -3.63571048e-02 3.18570524e-01] | [8.454607963562012, 10.767108917236328] |
8075ae2e-5524-4e47-935f-9e965e548b22 | mr4mr-mixed-reality-for-melody-reincarnation | 2209.07023 | null | https://arxiv.org/abs/2209.07023v1 | https://arxiv.org/pdf/2209.07023v1.pdf | MR4MR: Mixed Reality for Melody Reincarnation | There is a long history of an effort made to explore musical elements with the entities and spaces around us, such as musique concr\`ete and ambient music. In the context of computer music and digital art, interactive experiences that concentrate on the surrounding objects and physical spaces have also been designed. In recent years, with the development and popularization of devices, an increasing number of works have been designed in Extended Reality to create such musical experiences. In this paper, we describe MR4MR, a sound installation work that allows users to experience melodies produced from interactions with their surrounding space in the context of Mixed Reality (MR). Using HoloLens, an MR head-mounted display, users can bump virtual objects that emit sound against real objects in their surroundings. Then, by continuously creating a melody following the sound made by the object and re-generating randomly and gradually changing melody using music generation machine learning models, users can feel their ambient melody "reincarnating". | ['Nao Tokui', 'Shoma Sawa', 'Keisuke Okazaki', 'Takumi Inoue', 'Ryuku Nobusue', 'Ryogo Ishino', 'Atsuya Kobayashi'] | 2022-09-15 | null | null | null | null | ['music-generation', 'music-generation'] | ['audio', 'music'] | [ 1.28702432e-01 -1.05364099e-01 5.79507411e-01 1.56213209e-01
-5.44318974e-01 -6.18064165e-01 6.69493675e-01 -4.27919179e-01
1.94150373e-01 4.35338080e-01 6.04700208e-01 3.58688354e-01
1.04639810e-02 -9.38213825e-01 -3.26131970e-01 -3.82143915e-01
5.86081259e-02 -2.04131395e-01 1.91172078e-01 -6.28897309e-01
1.83388159e-01 2.43116796e-01 -2.11483598e+00 6.35089278e-01
2.87635535e-01 6.24328732e-01 6.38065696e-01 1.14632857e+00
-3.12739313e-01 1.06838465e+00 -9.74743783e-01 2.02542543e-02
-6.78566247e-02 -7.10322261e-01 -3.98357034e-01 -4.04288381e-01
1.22201540e-01 2.46230941e-02 -2.87291288e-01 8.79361212e-01
1.02572668e+00 5.08717358e-01 -4.20257151e-02 -9.14227724e-01
-8.22016180e-01 1.06698835e+00 -2.89375901e-01 -3.79178196e-01
1.20944679e+00 -1.07300222e-01 9.20473099e-01 -6.93348348e-01
6.97937906e-01 9.02398229e-01 8.48769486e-01 8.01483154e-01
-1.40073633e+00 -1.03456628e+00 -3.60400468e-01 -2.57812768e-01
-1.51834297e+00 -3.69968235e-01 1.20839894e+00 -2.75294304e-01
6.85885072e-01 1.20265090e+00 1.21112251e+00 1.14014709e+00
-2.23089587e-02 6.18257284e-01 8.67390990e-01 -8.88369977e-01
2.59487659e-01 3.66656125e-01 -5.23648620e-01 -1.32752836e-01
-7.78090000e-01 1.81002975e-01 -6.99247420e-01 -4.19828773e-01
1.07723594e+00 -2.68253058e-01 -3.90715301e-01 -1.13811068e-01
-1.36980283e+00 1.34863019e-01 3.00862670e-01 9.89601314e-01
-1.61351234e-01 4.34447199e-01 2.03280285e-01 -2.79206410e-03
2.19264075e-01 9.15808976e-01 -6.24880120e-02 -8.21028233e-01
-8.45896304e-01 5.70652485e-01 7.38688111e-01 6.63353443e-01
-1.80733427e-02 1.87612578e-01 -3.36085856e-02 1.13593626e+00
3.09749633e-01 4.15580630e-01 5.06908536e-01 -9.72308517e-01
-1.29468441e-01 2.66619474e-01 3.77271295e-01 -1.09358263e+00
-3.78382474e-01 -1.46898106e-01 -5.06683886e-01 5.81081688e-01
-9.54732001e-02 -1.87825650e-01 -2.16541260e-01 1.68689930e+00
4.12011415e-01 6.96759760e-01 -5.41167185e-02 8.57820690e-01
1.10223734e+00 7.10765481e-01 -3.06805633e-02 -8.33176170e-03
1.12404251e+00 -4.60734189e-01 -1.29579854e+00 4.22367513e-01
3.49882990e-01 -1.10912132e+00 1.58018899e+00 8.11711729e-01
-1.22141778e+00 -1.10690951e+00 -1.12026596e+00 1.03343926e-01
-7.30358288e-02 -3.30865622e-01 6.48366034e-01 8.20616126e-01
-8.04420412e-01 7.77028024e-01 -5.08594096e-01 -9.96435806e-02
-2.33667359e-01 -1.10254791e-02 2.64742915e-02 9.37365413e-01
-9.77501631e-01 4.77992684e-01 -6.07747138e-02 8.10573027e-02
-2.94156194e-01 -9.04553950e-01 -4.26586092e-01 -1.99618086e-01
-1.29827738e-01 -5.40320635e-01 1.44037759e+00 -1.02243948e+00
-2.09479022e+00 7.91283131e-01 5.24692297e-01 1.63501725e-01
3.73894662e-01 -7.37599075e-01 -1.38833570e+00 -4.09351707e-01
-9.03790370e-02 4.01818186e-01 2.40404338e-01 -1.52910888e+00
-3.78969580e-01 -2.26636529e-01 1.31137982e-01 3.70784134e-01
-1.22049205e-01 3.41320366e-01 -1.60953924e-02 -1.07266235e+00
1.86320752e-01 -8.75509560e-01 -2.04273954e-01 -4.16078448e-01
-5.11146665e-01 2.39996076e-01 8.56133759e-01 -3.30952823e-01
1.93501973e+00 -2.65192962e+00 1.42274201e-01 2.84947604e-01
-9.62717533e-02 7.28774965e-02 2.79001892e-01 7.02927589e-01
-1.66836366e-01 -9.78704542e-03 2.76643157e-01 -3.95399153e-01
2.06670806e-01 -1.52226701e-01 -5.85616529e-01 -2.58774191e-01
-7.63334990e-01 6.56336784e-01 -1.08753169e+00 -1.77259311e-01
4.37166154e-01 8.13072562e-01 -5.91762185e-01 1.75820589e-01
-2.51740068e-01 9.41174924e-01 -2.07516670e-01 1.79614022e-01
4.93718207e-01 2.52338648e-01 1.59581125e-01 1.21963985e-01
-5.53325832e-01 1.92975834e-01 -1.78139412e+00 2.40051603e+00
-7.96497464e-01 4.77672666e-01 1.62117898e-01 2.66505450e-01
1.16269183e+00 7.47572899e-01 4.17330623e-01 -9.08701479e-01
9.60671976e-02 3.01319718e-01 -4.61665481e-01 -9.74385977e-01
1.16642475e+00 -4.44727629e-01 -2.74444550e-01 6.76597476e-01
-6.76766157e-01 -5.08604467e-01 -6.79202914e-01 -1.46319672e-01
9.99687970e-01 7.91112661e-01 1.62671939e-01 2.64025092e-01
2.39070579e-01 -4.70526010e-01 -5.75780310e-02 5.20633519e-01
3.38093460e-01 8.90490174e-01 -4.38396096e-01 -4.65447366e-01
-7.63005197e-01 -1.32954240e+00 9.80808865e-03 1.48673248e+00
3.12507957e-01 -7.74340808e-01 -5.60321808e-01 2.62671173e-01
-3.99635971e-01 9.93082166e-01 -4.43495899e-01 1.13877080e-01
-5.54385364e-01 -1.44221514e-01 7.06949830e-01 3.25297028e-01
2.53192931e-01 -1.98892379e+00 -1.04959726e+00 4.90148097e-01
-2.61423409e-01 -4.38171238e-01 -4.05082136e-01 -1.59100801e-01
-6.18888915e-01 -3.36682796e-01 -5.46530545e-01 -7.63409853e-01
-2.15211973e-01 1.77326307e-01 1.04774547e+00 -1.11070246e-01
-6.43165410e-01 5.65215409e-01 -7.01118588e-01 -5.61298132e-01
-4.27770287e-01 -3.87302339e-01 1.10244183e-02 6.12650998e-02
-2.92713102e-02 -1.33945966e+00 -5.52857995e-01 1.70440823e-01
-1.05392683e+00 4.94457126e-01 -1.54033080e-01 8.69760811e-02
4.81575787e-01 5.85085945e-03 5.52616000e-01 -5.80060482e-01
8.01481426e-01 -3.09645981e-01 1.50395960e-01 -1.19731799e-01
3.50203454e-01 -4.23049778e-01 2.69714952e-01 -9.70154047e-01
-1.28634572e+00 -1.66019365e-01 -3.55375737e-01 -2.76004523e-01
-4.07919765e-01 1.88816875e-01 -3.29702616e-01 3.81905228e-01
1.12596893e+00 -1.15679055e-01 -3.89901131e-01 -6.30559385e-01
7.32885420e-01 1.11783719e+00 9.62260187e-01 -4.45845664e-01
5.06099880e-01 4.53902900e-01 -5.74710667e-01 -8.44355285e-01
-2.57788360e-01 -5.11014275e-02 -2.22519681e-01 -6.69912100e-01
8.45391035e-01 -4.77273554e-01 -7.95301437e-01 2.98894823e-01
-8.35533202e-01 -5.92427611e-01 -9.63790357e-01 7.06382036e-01
-8.15987468e-01 -2.18711365e-02 -4.12734032e-01 -1.26122749e+00
-1.84310362e-01 -6.57879770e-01 8.81636918e-01 4.22393054e-01
-1.13346791e+00 -6.09744966e-01 8.58170211e-01 2.63215125e-01
4.41960841e-01 6.81534171e-01 4.82607931e-01 2.56880522e-01
-2.93570668e-01 -1.55764937e-01 5.71854293e-01 -1.41027018e-01
5.13975739e-01 5.16702607e-02 -1.37128580e+00 3.05263609e-01
-1.21155880e-01 -5.63145280e-02 -2.35772774e-01 2.16900855e-01
9.84566510e-01 -2.04641759e-01 -1.95832491e-01 4.78947461e-01
1.12582672e+00 8.24657142e-01 1.17043006e+00 3.56492996e-01
4.02409256e-01 3.71541739e-01 4.47986424e-01 7.09740758e-01
-6.37141336e-03 1.07597697e+00 1.41147465e-01 -3.35491821e-02
-3.17269087e-01 -7.69985437e-01 2.34689534e-01 7.06178248e-01
-5.19343495e-01 3.07056271e-02 -3.53315949e-01 5.33788837e-02
-1.63966596e+00 -1.48794067e+00 -1.93351924e-01 2.19540238e+00
8.02349329e-01 -1.86563000e-01 7.48759583e-02 5.31650126e-01
6.84832633e-01 1.32234678e-01 -1.27134472e-01 -5.83061337e-01
-8.19125548e-02 6.71322703e-01 -6.35520399e-01 2.78821468e-01
-8.68945777e-01 6.23797536e-01 6.25631094e+00 7.12319851e-01
-1.19435847e+00 2.33752877e-02 -1.27595207e-02 -4.69521046e-01
-6.25663280e-01 -2.93951243e-01 1.71938315e-02 3.26875925e-01
3.93221855e-01 1.20059691e-01 9.88741815e-01 7.71735489e-01
4.31395978e-01 -5.16787022e-02 -8.29055548e-01 1.30808234e+00
-2.12519653e-02 -1.31801820e+00 -4.98053730e-01 -1.14011854e-01
7.72153735e-01 -5.36354959e-01 4.53865707e-01 3.22261095e-01
1.15131266e-01 -1.05809116e+00 1.11281335e+00 8.41059148e-01
1.05493224e+00 -7.66757369e-01 -8.33399668e-02 3.08081776e-01
-1.32022643e+00 -1.40932132e-03 2.82073915e-01 -4.87976581e-01
6.36093736e-01 -3.66778020e-03 -5.28044164e-01 1.12976685e-01
9.62777793e-01 -1.51833231e-02 1.37996394e-02 1.14482462e+00
5.50762489e-02 3.74019861e-01 -1.74455628e-01 -1.89774260e-01
-4.53743130e-01 -2.25566208e-01 8.73143792e-01 1.15248609e+00
7.01374769e-01 4.02271092e-01 -1.84370056e-01 1.08749080e+00
3.19783866e-01 5.39320648e-01 -7.35444605e-01 8.98062810e-02
3.58406872e-01 1.38725436e+00 -6.32535040e-01 4.90987860e-02
-5.39113060e-02 1.13902748e+00 -4.33664203e-01 2.22798750e-01
-1.06556571e+00 -5.80642879e-01 3.52452189e-01 4.56180841e-01
-2.41083130e-01 -2.88922668e-01 -3.20372939e-01 -6.59731328e-01
-1.47601381e-01 -7.16854036e-01 -1.25050843e-01 -1.59189808e+00
-9.39433157e-01 7.74455607e-01 -3.87575001e-01 -1.58536613e+00
-5.29520437e-02 2.48428896e-01 -8.99005175e-01 6.80594981e-01
-9.73750055e-02 -1.15662181e+00 -3.48577946e-01 3.03251714e-01
3.50535989e-01 1.96479131e-02 1.44534528e+00 5.25467873e-01
2.48394042e-01 2.68333852e-01 -3.91487181e-02 -2.28289157e-01
5.93778789e-01 -1.00331950e+00 3.43994617e-01 -8.83498937e-02
7.88233221e-01 5.51848710e-01 1.06997609e+00 -5.78475296e-01
-1.15174556e+00 -3.31201643e-01 6.05934203e-01 -5.83954871e-01
5.23738265e-01 -6.02677822e-01 -6.88382804e-01 4.42639410e-01
1.55021027e-01 -2.80995160e-01 1.25095963e+00 2.43325204e-01
-2.65332669e-01 6.76983520e-02 -1.15895724e+00 1.05272079e+00
1.38822985e+00 -7.88286328e-01 -7.34457791e-01 -3.22283536e-01
7.76065230e-01 -5.89723825e-01 -5.84488213e-01 2.69431859e-01
1.24715257e+00 -1.20601773e+00 9.35399354e-01 -2.12363079e-01
9.78645608e-02 -5.99897265e-01 -3.86734217e-01 -1.19437182e+00
-2.84988374e-01 -1.22801781e+00 7.80275390e-02 1.28188848e+00
1.62812695e-01 -4.67411429e-02 5.99444211e-01 4.18065786e-01
-1.90537021e-01 -1.79395467e-01 -5.94172299e-01 -2.27817774e-01
-5.72782695e-01 -1.13290811e+00 1.13165236e+00 1.09003234e+00
6.96456850e-01 1.02800570e-01 -5.30406952e-01 -5.27941883e-02
-1.43326759e-01 2.23143935e-01 1.12225115e+00 -1.26305604e+00
-8.57618809e-01 -3.76282841e-01 -3.60989690e-01 -8.41688216e-01
-5.10368764e-01 -6.50150776e-01 5.14077395e-02 -1.26559913e+00
-1.23400114e-01 -8.32839727e-01 -3.36346596e-01 -4.85240705e-02
2.62775213e-01 6.67009890e-01 4.81125444e-01 9.04633999e-02
-5.26523888e-01 5.22935450e-01 1.48701918e+00 2.27282792e-01
-1.23238039e+00 2.03111246e-01 -5.72881877e-01 8.80279601e-01
5.79589963e-01 -4.03537508e-03 -4.36767280e-01 -5.50962500e-02
9.54851031e-01 4.95999455e-01 3.04250807e-01 -1.59481573e+00
4.43928428e-02 -6.60276935e-02 3.53319049e-01 -5.82057297e-01
9.19542789e-01 -8.44389558e-01 1.41279626e+00 8.78509209e-02
-5.40585995e-01 -1.86540261e-01 1.85756445e-01 1.62329584e-01
9.06092972e-02 -4.55039889e-02 2.68002033e-01 -1.74937665e-01
-2.62071609e-01 -4.77642745e-01 -5.11120558e-01 -3.84861410e-01
9.88902926e-01 -5.50135314e-01 9.49990973e-02 -6.21019959e-01
-1.62286913e+00 -5.97585559e-01 3.26314092e-01 8.98201406e-01
7.07602739e-01 -1.80447197e+00 -1.65713221e-01 2.30502948e-01
-1.27705466e-02 -6.92835301e-02 8.36617529e-01 1.59354508e-01
-5.10022521e-01 -9.58944559e-02 -2.27953643e-01 -2.13894278e-01
-1.49209726e+00 4.23728466e-01 2.60518163e-01 2.38023967e-01
-1.00328314e+00 6.64594114e-01 6.31002784e-02 -5.70567966e-01
2.25447387e-01 -2.31326595e-01 -5.77808321e-01 2.25651134e-02
9.57940280e-01 6.85736537e-01 -4.45030481e-01 -4.90515947e-01
2.75973767e-01 5.39014816e-01 8.18629563e-01 -8.08091462e-01
1.31934321e+00 -1.40941873e-01 7.27931261e-02 1.14163399e+00
5.85263133e-01 9.09351110e-01 -7.75211036e-01 1.66322470e-01
-2.59116113e-01 -5.85266650e-01 -2.47062907e-01 -1.01877916e+00
-5.53133845e-01 4.82079417e-01 8.72613251e-01 5.15493572e-01
1.16190457e+00 -1.05588539e-02 7.86208451e-01 -1.36466905e-01
9.61871624e-01 -1.10463941e+00 4.60924536e-01 1.21742621e-01
1.19344211e+00 -1.13789327e-01 -5.44836700e-01 -9.36728865e-02
-7.09363937e-01 8.40984523e-01 4.86732364e-01 8.98087323e-02
8.54192734e-01 7.56291151e-01 2.54607230e-01 -1.38848990e-01
-3.21684927e-01 2.96948224e-01 5.12627587e-02 6.67060673e-01
7.23677456e-01 5.76681674e-01 -9.06555355e-03 1.20916343e+00
-9.33235645e-01 3.98485333e-01 3.68136436e-01 9.26551044e-01
-5.82138419e-01 -1.05682540e+00 -8.49047542e-01 -2.22387716e-01
-4.25459117e-01 1.96542174e-01 -3.10990900e-01 4.09713566e-01
8.06554794e-01 9.35799181e-01 2.20477730e-01 -8.16506684e-01
5.73314786e-01 1.19830379e-02 7.45760441e-01 -5.76928973e-01
-1.00506008e+00 3.12329441e-01 -9.83427316e-02 -4.76286352e-01
-4.66830730e-01 -4.15865928e-01 -1.63904464e+00 -1.00934468e-01
-2.96570152e-01 1.85305193e-01 7.80089200e-01 2.29476705e-01
3.45124483e-01 9.39370036e-01 5.90273678e-01 -1.22817183e+00
2.25256652e-01 -1.08729088e+00 -1.06133640e+00 3.15183342e-01
-1.17520869e-01 -2.03098729e-01 -9.32517350e-02 -1.94425173e-02] | [16.03032112121582, 5.468379974365234] |
1d6d8213-8b86-4fec-ad85-9960df63d441 | efficient-data-driven-encoding-of-scene | 2103.02743 | null | https://arxiv.org/abs/2103.02743v1 | https://arxiv.org/pdf/2103.02743v1.pdf | Efficient data-driven encoding of scene motion using Eccentricity | This paper presents a novel approach of representing dynamic visual scenes with static maps generated from video/image streams. Such representation allows easy visual assessment of motion in dynamic environments. These maps are 2D matrices calculated recursively, in a pixel-wise manner, that is based on the recently introduced concept of Eccentricity data analysis. Eccentricity works as a metric of a discrepancy between a particular pixel of an image and its normality model, calculated in terms of mean and variance of past readings of the same spatial region of the image. While Eccentricity maps carry temporal information about the scene, actual images do not need to be stored nor processed in batches. Rather, all the calculations are done recursively, based on a small amount of statistical information stored in memory, thus resulting in a very computationally efficient (processor- and memory-wise) method. The list of potential applications includes video-based activity recognition, intent recognition, object tracking, video description, and so on. | ['Dimitar Filev', 'Gint Puskorius', 'Mostafa Parchami', 'Enrique Corona', 'Bruno Costa'] | 2021-03-03 | null | null | null | null | ['video-description', 'intent-recognition'] | ['computer-vision', 'natural-language-processing'] | [ 5.06344914e-01 -4.62884039e-01 2.01316625e-02 -1.09605610e-01
2.13222522e-02 -3.43072295e-01 7.52083719e-01 4.55806583e-01
-6.28846109e-01 3.67085069e-01 1.40911475e-01 -1.43716812e-01
-2.49857083e-01 -7.03032017e-01 -4.53297526e-01 -8.00825357e-01
-5.07843673e-01 2.34809723e-02 6.55708194e-01 1.45064950e-01
5.27144909e-01 8.70415211e-01 -1.92323554e+00 3.79979372e-01
1.40921324e-02 1.14055407e+00 2.71602750e-01 1.21982956e+00
5.00913151e-02 1.18149555e+00 -6.55530751e-01 7.04835430e-02
9.85505581e-02 -3.16551477e-01 -3.51063013e-01 4.37829643e-01
3.22299272e-01 -3.08968931e-01 -3.51017386e-01 9.74669397e-01
5.48219346e-02 1.69245124e-01 7.44148731e-01 -1.26164484e+00
8.67259726e-02 -1.57382920e-01 -5.61643183e-01 8.05493534e-01
7.89045990e-01 4.12780084e-02 3.35338503e-01 -7.11031854e-01
7.07197964e-01 8.82468402e-01 6.83002770e-01 -1.75287917e-01
-1.28636456e+00 -1.02034353e-01 1.25174941e-02 6.28078818e-01
-1.50957870e+00 -4.03765529e-01 7.42686331e-01 -6.99555397e-01
1.00511634e+00 6.96992397e-01 9.31067526e-01 4.90203232e-01
4.38835293e-01 5.58666945e-01 7.78440475e-01 -4.63413328e-01
4.12454396e-01 1.27310315e-02 -1.28954530e-01 5.01211643e-01
2.48834595e-01 -2.83372343e-01 -6.48471653e-01 -6.32596835e-02
6.60675406e-01 9.92774963e-02 -3.00586730e-01 -7.88581669e-01
-1.38523352e+00 2.22210526e-01 -5.10314517e-02 4.43238139e-01
-7.56208658e-01 2.28119209e-01 5.75231314e-01 3.15972060e-01
3.52841526e-01 -2.15173736e-01 -7.54289776e-02 -6.54911458e-01
-1.27697527e+00 -8.16176366e-03 5.41680992e-01 5.40250838e-01
6.95978642e-01 -2.13976321e-03 -4.54907632e-03 2.81453550e-01
3.03930968e-01 5.22548854e-01 6.48239791e-01 -7.00764596e-01
2.81468660e-01 6.26806736e-01 -3.47664990e-02 -1.72259653e+00
-4.31366295e-01 1.64376229e-01 -7.50370085e-01 5.20780742e-01
4.75131363e-01 2.80180484e-01 -4.91509110e-01 1.29692686e+00
3.54958922e-01 2.82136798e-01 -6.69521838e-02 4.75114256e-01
2.77308404e-01 8.28686059e-01 -1.09499758e-02 -6.94562197e-01
1.32006669e+00 -3.36804390e-01 -7.59044826e-01 1.08962037e-01
3.04314762e-01 -5.72136223e-01 5.34243762e-01 6.99933767e-01
-1.07937169e+00 -6.31584048e-01 -1.14392006e+00 4.41267163e-01
-4.09631491e-01 2.12887246e-02 3.15474719e-01 6.91823006e-01
-1.15286672e+00 4.26035017e-01 -8.61893356e-01 -4.73281056e-01
1.35280147e-01 2.02655315e-01 -4.16812390e-01 1.99754760e-01
-5.81438005e-01 6.25342429e-01 5.15609086e-01 2.07655013e-01
-7.74441481e-01 -4.18915778e-01 -9.46932852e-01 5.62103279e-03
2.30229855e-01 -7.02504739e-02 8.63751590e-01 -9.07703698e-01
-1.05387235e+00 9.86248732e-01 -4.81784254e-01 -5.16423106e-01
7.35030353e-01 1.40986010e-01 -6.13462925e-01 4.08103019e-01
-1.68629050e-01 1.24614745e-01 1.13929784e+00 -9.82038319e-01
-7.59226501e-01 -4.98418599e-01 -1.43750265e-01 2.10407808e-01
-2.29786098e-01 1.92181449e-02 -6.77559376e-01 -5.07729232e-01
3.58299345e-01 -6.91738129e-01 -1.85445741e-01 1.88151106e-01
-6.35120347e-02 3.61391269e-02 9.86655891e-01 -8.21914196e-01
1.66912746e+00 -2.42203903e+00 -2.77715713e-01 6.11373365e-01
9.62756053e-02 3.01153392e-01 3.51900786e-01 4.55489010e-01
-2.16006443e-01 -2.50582308e-01 5.26294857e-02 1.96670011e-01
-4.37356472e-01 -1.77589759e-01 -1.72993660e-01 7.78274238e-01
5.43667516e-03 3.91519308e-01 -9.35139894e-01 -6.92579925e-01
6.47553504e-01 3.61569256e-01 -2.07348689e-01 -1.18422069e-01
5.39017431e-02 1.62257329e-01 -2.37442508e-01 3.86991113e-01
7.31438160e-01 4.66478132e-02 3.91009957e-01 -3.10728669e-01
-4.08163816e-01 -1.15910672e-01 -1.69966829e+00 1.26295745e+00
-2.72488147e-01 1.21074307e+00 -3.35045338e-01 -1.02227533e+00
9.02552843e-01 3.56069177e-01 8.88020515e-01 -8.64468575e-01
2.53771502e-03 -1.93256721e-01 -3.69164616e-01 -6.38316870e-01
8.32663834e-01 4.98677880e-01 1.21077798e-01 6.02603316e-01
-4.61841732e-01 3.17888916e-01 7.78389931e-01 9.93934348e-02
1.24554598e+00 1.47585988e-01 7.17089713e-01 -1.23135485e-01
8.11442554e-01 1.41523212e-01 1.33539483e-01 6.83482647e-01
-2.93053836e-01 4.49232578e-01 5.63426197e-01 -4.68219906e-01
-1.15357673e+00 -1.19243634e+00 -1.57744363e-01 7.32556283e-01
3.18654209e-01 -6.05886221e-01 -7.10161448e-01 -1.49592623e-01
-1.81828544e-01 2.63170272e-01 -6.69219315e-01 3.40874977e-02
-5.79077184e-01 -6.52880788e-01 1.14144146e-01 4.95896697e-01
4.79419440e-01 -9.54101264e-01 -1.52175486e+00 2.99130231e-01
-9.77473706e-03 -9.29185867e-01 -1.89559460e-01 -6.75107539e-02
-1.12646258e+00 -1.10403192e+00 -4.62652713e-01 -2.93791771e-01
8.30610514e-01 6.17453635e-01 8.39527547e-01 9.07419920e-02
-6.24983728e-01 7.65913248e-01 -2.03299403e-01 -8.95279273e-02
-3.61838430e-01 -7.75446713e-01 2.10176274e-01 3.59182924e-01
3.74188334e-01 -4.81550843e-01 -6.16138756e-01 4.43492740e-01
-9.46305513e-01 -1.07158482e-01 2.66007453e-01 3.25582236e-01
7.39874184e-01 5.17252147e-01 -2.26775035e-01 -3.25292557e-01
4.45125014e-01 -3.84306759e-01 -7.99492300e-01 4.72824425e-01
-2.63564974e-01 -5.83391786e-02 4.13600832e-01 -5.24189591e-01
-8.72101784e-01 2.33923629e-01 4.87257719e-01 -4.00286436e-01
-1.86139032e-01 3.34009260e-01 1.94151215e-02 1.31794438e-01
5.77076375e-01 5.86064875e-01 1.42311022e-01 -2.14032292e-01
3.32871340e-02 6.58707678e-01 8.19007933e-01 1.55196607e-01
5.46126902e-01 8.25397015e-01 2.16254398e-01 -1.21577609e+00
9.17914733e-02 -8.54498565e-01 -8.04358900e-01 -8.50248158e-01
7.82520115e-01 -5.56469858e-01 -9.96912241e-01 5.62507927e-01
-8.53905857e-01 -3.81080769e-02 -2.65571326e-01 6.01065278e-01
-7.45093048e-01 6.44400179e-01 1.28522618e-02 -1.15586281e+00
1.12511761e-01 -7.31353104e-01 8.52885485e-01 1.30737334e-01
-5.37571490e-01 -9.09763277e-01 2.47026265e-01 3.90404500e-02
1.18399143e-01 2.92179257e-01 4.98188406e-01 -2.38940105e-01
-6.01252317e-01 -4.79637504e-01 -5.12322970e-02 1.92624524e-01
1.06920063e-01 4.10484046e-01 -9.17591274e-01 -1.51522994e-01
2.70210594e-01 2.84500986e-01 4.01630789e-01 4.04236227e-01
1.02166831e+00 -1.79364339e-01 -4.02997881e-01 3.41327637e-01
1.49158239e+00 6.62538230e-01 9.62134063e-01 5.30654669e-01
3.55342537e-01 4.06987429e-01 8.54961693e-01 1.03103006e+00
1.76299382e-02 8.02810252e-01 3.46924931e-01 2.07662702e-01
-1.15040189e-03 6.06922172e-02 4.78454381e-01 3.84826273e-01
-2.48822525e-01 -1.42303512e-01 -9.66203749e-01 5.22755921e-01
-1.76524675e+00 -1.56214190e+00 -4.39187527e-01 2.69836354e+00
1.71262756e-01 1.00392342e-01 2.45849133e-01 6.46475732e-01
9.29916799e-01 3.05962145e-01 -3.52897316e-01 -5.01458168e-01
6.74370229e-02 -3.80634397e-01 5.70280492e-01 2.19551906e-01
-1.08899319e+00 1.87349871e-01 6.99424601e+00 7.67457426e-01
-1.15041769e+00 -2.57990569e-01 4.67768103e-01 -8.40662979e-04
1.18650496e-01 -1.81512401e-01 -4.09973770e-01 7.12572336e-01
9.30124760e-01 -4.40694332e-01 2.14026853e-01 9.33019459e-01
4.58064556e-01 -1.05115688e+00 -1.04762971e+00 1.40379810e+00
3.44019979e-01 -1.16053081e+00 -2.05888107e-01 2.28803143e-01
3.78292710e-01 -3.11191767e-01 3.64170074e-02 -4.05469149e-01
-4.34960008e-01 -5.40863693e-01 9.60942090e-01 8.93528759e-01
5.56666791e-01 -6.57866716e-01 4.63716477e-01 3.30542147e-01
-1.39792383e+00 -1.14302672e-01 -3.03852320e-01 -1.25480309e-01
2.57151842e-01 5.22705972e-01 -8.12249005e-01 3.37654591e-01
9.47236001e-01 6.33152544e-01 -7.61676013e-01 1.23175919e+00
4.26211268e-01 1.70201913e-01 -3.38358790e-01 -1.03236675e-01
-3.68280932e-02 -3.56848866e-01 5.94420075e-01 1.25495183e+00
4.07305807e-01 -2.52071142e-01 -1.75297499e-01 2.83538252e-01
7.33832836e-01 4.32686098e-02 -7.68071175e-01 3.93605120e-02
1.31707326e-01 1.05484915e+00 -1.19443572e+00 -5.03420293e-01
-4.36021805e-01 9.85877693e-01 -2.42770553e-01 7.56153166e-02
-7.65707672e-01 -3.21253061e-01 5.64768434e-01 2.42029980e-01
5.15089929e-01 -4.86474901e-01 -5.37880100e-02 -8.04163039e-01
2.89984852e-01 -6.44743681e-01 3.83258343e-01 -9.38165009e-01
-6.17014945e-01 5.23635685e-01 3.52494955e-01 -1.83519936e+00
-8.30308735e-01 -7.56564677e-01 -5.84452689e-01 3.41104835e-01
-8.57499838e-01 -2.71999180e-01 -5.32319725e-01 8.67469966e-01
2.91845769e-01 -1.27088770e-01 6.37005925e-01 1.92476913e-01
-3.21717769e-01 6.71307743e-02 4.16890174e-01 -3.66037637e-02
2.41657093e-01 -1.04156291e+00 1.96742192e-01 1.10132515e+00
4.18886811e-01 2.87501484e-01 9.58851814e-01 -6.30137861e-01
-1.25609052e+00 -5.63373089e-01 8.26335132e-01 -5.17919004e-01
6.86468184e-01 -1.96374342e-01 -7.72416592e-01 2.59019345e-01
-1.15801036e-01 -4.66165431e-02 5.96556425e-01 -3.17787349e-01
6.31593913e-02 -3.57627600e-01 -1.00840425e+00 6.59996688e-01
9.13047373e-01 -5.91537237e-01 -3.69447351e-01 1.90877989e-01
-2.19077945e-01 -1.79418609e-01 -7.09165573e-01 1.92952201e-01
7.21026599e-01 -1.48930824e+00 9.85791385e-01 -1.68292344e-01
-9.87355784e-02 -6.36163771e-01 -1.72340855e-01 -6.85166538e-01
-2.55936265e-01 -4.78121907e-01 -2.85500646e-01 8.13019514e-01
2.37533897e-02 -9.80024561e-02 8.87958050e-01 5.04321516e-01
3.10397238e-01 -4.32184011e-01 -1.02758873e+00 -8.56248319e-01
-1.01672697e+00 -9.06658113e-01 2.25920722e-01 4.64760780e-01
1.84315324e-01 -2.99824744e-01 -3.83622319e-01 -1.27758579e-02
5.12826622e-01 -6.30499199e-02 8.96857321e-01 -1.13688791e+00
-1.04009829e-01 -3.94417793e-01 -1.38423479e+00 -8.48561108e-01
-3.85759622e-01 -2.32631087e-01 -1.71823967e-02 -1.42019475e+00
2.15189293e-01 -4.36984487e-02 -8.46986547e-02 -8.09180513e-02
3.31267446e-01 6.19222701e-01 1.36063501e-01 3.23694110e-01
-7.04208553e-01 -1.59164127e-02 6.79619133e-01 -1.04445778e-02
-1.98087215e-01 1.73830345e-01 1.23106040e-01 8.02282512e-01
5.53172767e-01 -3.16756159e-01 -6.37733817e-01 -3.68824638e-02
3.37214947e-01 -2.64597242e-03 5.18317580e-01 -1.49387503e+00
4.09101009e-01 -2.14278087e-01 5.95989764e-01 -9.77381229e-01
4.80000019e-01 -1.23672402e+00 6.98285818e-01 6.88813746e-01
5.95684946e-02 4.53563929e-01 4.24383730e-02 7.86549807e-01
-3.35842162e-01 -4.43960518e-01 5.09668589e-01 -3.98064889e-02
-1.35280561e+00 3.90715785e-02 -9.41714942e-01 -4.18189555e-01
1.56466472e+00 -1.10573304e+00 9.30161700e-02 -4.50608104e-01
-7.37954915e-01 -3.01808238e-01 5.78744411e-01 2.03519210e-01
6.49564326e-01 -1.29428530e+00 -2.05929950e-01 4.32183176e-01
1.74455523e-01 -5.42221129e-01 6.10409975e-01 9.61543858e-01
-1.03401387e+00 4.06211048e-01 -4.75028187e-01 -1.02974021e+00
-1.67650223e+00 8.22301924e-01 5.26510477e-02 -1.90150261e-01
-6.70969307e-01 2.61736780e-01 -6.60954118e-02 5.56546509e-01
2.94457525e-01 -3.11494380e-01 -4.48918104e-01 3.83220583e-01
1.04156876e+00 7.23153055e-01 1.69196531e-01 -9.16663587e-01
-5.82500994e-01 7.60037243e-01 2.38370344e-01 -3.35265785e-01
9.65678990e-01 -3.81474346e-01 -1.63950827e-02 7.12237418e-01
1.08831263e+00 7.11273551e-02 -1.40559995e+00 -1.38304994e-01
3.06719035e-01 -8.94789279e-01 -2.33755410e-02 -2.06154749e-01
-6.64953351e-01 6.02171004e-01 9.43781376e-01 5.53241968e-01
1.44814193e+00 -1.93054676e-01 5.33644855e-02 4.27585512e-01
2.94013798e-01 -1.43603289e+00 2.00296059e-01 3.85569215e-01
7.20021546e-01 -8.59370768e-01 3.06216717e-01 -2.11099148e-01
-7.63372660e-01 1.29014242e+00 1.88504860e-01 2.04685569e-01
7.35680103e-01 2.81027138e-01 -4.92061228e-02 -8.65517259e-02
-4.88256216e-01 -1.75572947e-01 1.31746471e-01 8.73556554e-01
2.14252234e-01 -2.85402507e-01 -7.83643052e-02 -2.64240801e-01
1.19675517e-01 9.62563325e-03 5.23268223e-01 1.23550570e+00
-5.77775896e-01 -5.86465418e-01 -7.17056334e-01 3.39620382e-01
-1.35333374e-01 1.97105706e-01 4.18300480e-02 6.87163353e-01
-1.17266178e-01 7.80126095e-01 8.02181721e-01 -3.18875700e-01
2.68492311e-01 9.57309920e-03 5.04859805e-01 -1.38549194e-01
-1.31708264e-01 -1.06839098e-01 -1.25683993e-01 -8.24093580e-01
-7.37062395e-01 -1.06242716e+00 -9.92366135e-01 -4.01440591e-01
1.39120072e-01 -1.17310666e-01 9.82118845e-01 6.96039855e-01
1.61587939e-01 3.35122436e-01 7.36952245e-01 -1.02908874e+00
-1.12716399e-01 -6.21929646e-01 -6.72841370e-01 5.80774009e-01
2.20350221e-01 -4.67011809e-01 -1.94469929e-01 4.57513005e-01] | [8.208168029785156, 0.11335645616054535] |
358304e5-d4a5-4f66-99a8-668689c08fc1 | elda-using-edges-to-have-an-edge-on-semantic | 2211.08888 | null | https://arxiv.org/abs/2211.08888v1 | https://arxiv.org/pdf/2211.08888v1.pdf | ELDA: Using Edges to Have an Edge on Semantic Segmentation Based UDA | Many unsupervised domain adaptation (UDA) methods have been proposed to bridge the domain gap by utilizing domain invariant information. Most approaches have chosen depth as such information and achieved remarkable success. Despite their effectiveness, using depth as domain invariant information in UDA tasks may lead to multiple issues, such as excessively high extraction costs and difficulties in achieving a reliable prediction quality. As a result, we introduce Edge Learning based Domain Adaptation (ELDA), a framework which incorporates edge information into its training process to serve as a type of domain invariant information. In our experiments, we quantitatively and qualitatively demonstrate that the incorporation of edge information is indeed beneficial and effective and enables ELDA to outperform the contemporary state-of-the-art methods on two commonly adopted benchmarks for semantic segmentation based UDA tasks. In addition, we show that ELDA is able to better separate the feature distributions of different classes. We further provide an ablation analysis to justify our design decisions. | ['Chun-Yi Lee', 'Yi-Chen Lo', 'Chia-Che Chang', 'Chen-Hao Chao', 'Bo-Wun Cheng', 'Hsu-Shen Liu', 'Li-Yuan Tsao', 'Jie-En Yao', 'Shan-Ya Yang', 'Huang-Ru Liao', 'Ting-Hsuan Liao'] | 2022-11-16 | null | null | null | null | ['synthetic-to-real-translation'] | ['computer-vision'] | [ 2.94797659e-01 5.65921143e-02 -4.57472235e-01 -4.49551076e-01
-9.81497109e-01 -5.67346811e-01 8.16642404e-01 2.61156052e-01
-4.65530485e-01 6.78181887e-01 4.45530638e-02 -1.78981915e-01
-4.26266119e-02 -7.48871088e-01 -4.84352618e-01 -6.01748884e-01
3.02014530e-01 4.90870208e-01 5.49730957e-01 -6.74337745e-02
2.96198756e-01 3.19804043e-01 -1.52841830e+00 1.76791742e-01
1.03548968e+00 1.04236960e+00 6.82067201e-02 4.97494452e-02
-3.56882066e-01 2.87687808e-01 -4.23966736e-01 -3.15130502e-01
3.75225276e-01 -4.14135277e-01 -9.90764260e-01 1.06310301e-01
3.81723344e-01 -2.18681619e-01 -1.93134733e-02 8.71362627e-01
5.52839935e-01 2.05199793e-01 1.12123477e+00 -1.19471467e+00
-4.85035449e-01 3.15901250e-01 -6.42209888e-01 1.16829760e-01
2.09668398e-01 -4.58660908e-02 1.18880606e+00 -8.86747181e-01
1.09099674e+00 9.49331045e-01 8.45122576e-01 6.60607934e-01
-1.31729138e+00 -4.65030849e-01 2.36696512e-01 4.06001151e-01
-1.14447927e+00 -3.94782454e-01 1.27604103e+00 -4.97530162e-01
9.59007442e-01 -7.96510503e-02 4.24471080e-01 1.23215961e+00
-3.66901644e-02 1.18756187e+00 1.27314413e+00 -7.26883352e-01
5.32696843e-01 1.01164013e-01 2.47515336e-01 3.25011492e-01
3.12477559e-01 -1.07960869e-03 -5.23524165e-01 1.20567508e-01
5.89534223e-01 -3.76232922e-01 2.71541290e-02 -1.01718569e+00
-9.11832690e-01 1.02625203e+00 4.54424888e-01 2.27661937e-01
-1.90733060e-01 -3.10978591e-01 6.51760936e-01 2.37731457e-01
6.52435124e-01 6.00016236e-01 -6.28016949e-01 -1.95514217e-01
-7.97683060e-01 1.75008759e-01 5.17036080e-01 9.99347091e-01
7.66395628e-01 -8.40039253e-02 -1.41241089e-01 1.13186836e+00
2.22413391e-01 1.23974472e-01 4.37384874e-01 -8.37514997e-01
1.50159791e-01 8.54533315e-01 4.40536812e-02 -6.03369653e-01
-4.60918397e-01 -3.76245439e-01 -4.22827035e-01 2.75246322e-01
6.58286214e-01 1.04529448e-01 -1.28386366e+00 1.72738755e+00
3.66235554e-01 -7.15334341e-02 1.09491020e-01 8.39518607e-01
7.49973178e-01 3.23954262e-02 3.22339684e-01 1.51047215e-01
1.20484066e+00 -9.63153780e-01 -4.67242122e-01 -5.43468714e-01
6.96115136e-01 -7.35623658e-01 1.35967731e+00 2.35930234e-01
-8.21610749e-01 -5.44530988e-01 -1.12738359e+00 -1.71564430e-01
-6.03345394e-01 7.47451708e-02 8.09365809e-01 6.97525442e-01
-6.90479398e-01 5.67318201e-01 -8.03577125e-01 -7.04795003e-01
8.38603735e-01 1.86056852e-01 -3.69450271e-01 -2.03782722e-01
-1.05512798e+00 8.00005674e-01 4.10179734e-01 -5.85475087e-01
-5.75980961e-01 -7.86527455e-01 -1.09386492e+00 -1.64365903e-01
2.96746910e-01 -7.34035492e-01 1.31540382e+00 -7.06405222e-01
-1.50501955e+00 1.33965945e+00 -8.73488784e-02 -6.88683152e-01
5.32640934e-01 -1.75006226e-01 -1.80429667e-01 1.74671277e-01
3.51142168e-01 8.70822132e-01 8.45725954e-01 -1.28839850e+00
-7.75576115e-01 -5.68429291e-01 4.44460427e-03 8.82841349e-02
-4.62714583e-01 -2.33690590e-01 -3.96156043e-01 -8.80760849e-01
2.37937361e-01 -8.53667021e-01 7.34043075e-04 4.60910685e-02
-6.30059540e-02 -3.93989086e-01 7.58013844e-01 -4.49499637e-01
1.17217660e+00 -2.20922232e+00 9.71222743e-02 7.14741834e-03
1.85514554e-01 3.53076637e-01 1.03922915e-02 1.83012918e-01
1.75031647e-01 -2.11094543e-02 -6.82595313e-01 -4.20108587e-01
9.88470167e-02 3.29462349e-01 1.37299104e-02 4.09317255e-01
3.86051804e-01 8.92669857e-01 -9.36222911e-01 -5.98614931e-01
3.29427928e-01 4.76436883e-01 -5.82958996e-01 3.13274637e-02
-1.82242274e-01 4.87690866e-01 -5.44628382e-01 7.86939204e-01
6.83478653e-01 -2.23784477e-01 7.43087158e-02 -1.51196018e-01
1.40841618e-01 4.51196551e-01 -9.32643533e-01 2.05242038e+00
-4.49155211e-01 7.29409873e-01 -1.27951309e-01 -1.35724819e+00
9.90044653e-01 1.77970771e-02 6.13246143e-01 -8.92592967e-01
3.14596713e-01 4.40563112e-01 -9.42401290e-02 -1.67141482e-01
4.58221436e-01 -2.99062312e-01 -1.43296838e-01 7.41953999e-02
3.56880665e-01 -2.04932734e-01 1.38003230e-01 3.07804625e-02
1.09052920e+00 4.98227090e-01 5.09682953e-01 -5.01142561e-01
4.74509269e-01 3.35360914e-01 7.45762706e-01 6.31162703e-01
-6.04647279e-01 8.51211011e-01 1.60728812e-01 -2.21337646e-01
-1.12142193e+00 -1.14348769e+00 -4.91424948e-01 1.10577393e+00
4.28243279e-01 -1.77886650e-01 -7.65231729e-01 -1.10985911e+00
4.75173704e-02 7.94511974e-01 -7.36306250e-01 -2.19366267e-01
-2.78163701e-01 -7.66677141e-01 3.59651566e-01 8.74993920e-01
7.65781283e-01 -9.67764258e-01 -4.89310950e-01 2.69624174e-01
-1.00076243e-01 -1.32402766e+00 -1.99214816e-01 5.08427382e-01
-1.09471440e+00 -8.78496528e-01 -8.96100879e-01 -8.33027601e-01
6.31498098e-01 2.49980032e-01 1.32278824e+00 -4.17366356e-01
-1.55550718e-01 3.55303377e-01 -6.82858586e-01 -4.64534491e-01
-4.23038572e-01 3.87271345e-01 -1.65167481e-01 -2.23784402e-01
7.72202432e-01 -6.31698310e-01 -6.37939870e-01 4.61373925e-01
-7.70497799e-01 -7.28353262e-02 5.44248104e-01 9.21881258e-01
7.03701854e-01 -1.70462370e-01 9.37388957e-01 -1.25553811e+00
6.58408225e-01 -2.61277765e-01 -3.98643374e-01 1.00669324e-01
-1.03341091e+00 2.36622617e-01 4.26437944e-01 -1.69800609e-01
-1.20355725e+00 2.01840281e-01 -1.66780680e-01 -1.92996562e-01
-5.34432173e-01 3.68795902e-01 -2.33412236e-01 -2.44887352e-01
1.00279248e+00 1.61877707e-01 -1.18686091e-02 -6.32835209e-01
3.48858565e-01 6.42609298e-01 5.42888641e-01 -6.84092581e-01
6.25605702e-01 6.49090171e-01 -1.28591642e-01 -5.87498486e-01
-8.63938272e-01 -1.01895559e+00 -9.79046226e-01 -3.59603427e-02
7.24690676e-01 -8.98116112e-01 2.05996245e-01 6.20505154e-01
-8.21914375e-01 -4.17312473e-01 -4.50364828e-01 2.02883542e-01
-7.90137172e-01 4.36304390e-01 -3.83792371e-01 -4.88080174e-01
-1.19248159e-01 -9.80582058e-01 1.08601654e+00 3.16715747e-01
-3.26853722e-01 -1.26702583e+00 8.01337510e-02 4.25772607e-01
3.26540083e-01 3.97298247e-01 7.28106260e-01 -7.53664374e-01
-1.13604151e-01 -1.88773587e-01 -3.83118153e-01 3.35600883e-01
3.14810306e-01 -2.61539876e-01 -1.17448938e+00 -2.51683056e-01
-2.59188175e-01 -1.25286758e-01 1.05418813e+00 5.21587491e-01
8.85046661e-01 3.36824656e-01 -5.56023598e-01 5.35291433e-01
1.38731730e+00 2.73749381e-01 5.66594362e-01 8.04407716e-01
5.55721521e-01 4.87133712e-01 8.92236650e-01 2.71548539e-01
4.26383346e-01 7.43793428e-01 1.41917259e-01 -3.23006630e-01
-5.23067176e-01 -1.84193075e-01 1.20801672e-01 5.72935104e-01
-4.87381108e-02 -4.77029905e-02 -1.08520174e+00 9.87470150e-01
-1.90650177e+00 -4.80291247e-01 1.34907767e-01 2.05155873e+00
8.03087175e-01 4.99769121e-01 3.80419433e-01 3.38148594e-01
4.53712702e-01 -7.68882781e-02 -7.62332618e-01 -4.79469091e-01
-1.60611182e-01 2.54411280e-01 5.89455307e-01 1.49855360e-01
-1.40404570e+00 1.08430970e+00 6.59793806e+00 9.57043231e-01
-9.16209579e-01 1.01631232e-01 4.42945659e-01 4.84024674e-01
-3.31071675e-01 -8.68900269e-02 -5.89954436e-01 3.01440001e-01
4.73592103e-01 -5.26290946e-02 1.01741634e-01 1.27216148e+00
-2.24881351e-01 -1.45133898e-01 -1.20688784e+00 9.04828787e-01
-2.23530442e-01 -1.22487378e+00 -4.80841734e-02 -2.37961467e-02
9.70060170e-01 9.30632427e-02 -2.21787151e-02 5.18836379e-01
2.63074726e-01 -7.46873856e-01 5.04918575e-01 -1.02555722e-01
7.00511813e-01 -6.32455230e-01 7.55186319e-01 1.49985626e-01
-1.16129351e+00 4.91032228e-02 -2.85205990e-01 -2.12621480e-01
-3.13974768e-02 5.72136462e-01 -9.78998423e-01 7.83600152e-01
7.76110590e-01 9.39048707e-01 -5.99797487e-01 1.04430723e+00
-2.03524500e-01 7.29562879e-01 -2.72754997e-01 2.79974163e-01
3.60905826e-01 -8.71668682e-02 5.42041659e-01 1.26409101e+00
3.04360259e-02 -2.43220299e-01 -2.46949289e-02 7.80785143e-01
-8.60549137e-02 1.96398675e-01 -7.61619389e-01 1.49577679e-02
4.18616772e-01 1.04874754e+00 -1.14020538e+00 -2.69368291e-01
-7.30098844e-01 1.06672394e+00 3.27929974e-01 2.53679723e-01
-7.06419647e-01 -3.90251756e-01 7.60116637e-01 2.22029522e-01
5.05345285e-01 -2.52963513e-01 -8.74434233e-01 -1.13466811e+00
-1.72685906e-02 -6.67698145e-01 6.40451014e-01 -4.58982825e-01
-1.58236587e+00 4.98256266e-01 8.55819508e-02 -1.48904359e+00
-1.76799521e-01 -8.58621538e-01 -3.63653570e-01 5.05608618e-01
-1.92001021e+00 -1.25432348e+00 -3.06449324e-01 3.80077809e-01
7.93457687e-01 -2.74552077e-01 7.95639932e-01 4.45660084e-01
-4.37637776e-01 8.28911364e-01 2.14888364e-01 1.57549828e-01
9.81018007e-01 -1.50012946e+00 5.52377522e-01 8.25463414e-01
2.83891797e-01 3.03256541e-01 6.09552860e-01 -4.37756389e-01
-9.29817379e-01 -9.02472079e-01 5.48837304e-01 -6.30468369e-01
4.25193191e-01 -2.54892349e-01 -9.65759575e-01 5.32874823e-01
-1.15622506e-01 4.45147902e-02 5.42349517e-01 3.38371485e-01
-4.22107637e-01 1.66650452e-02 -1.31698430e+00 4.97655183e-01
1.17707729e+00 -5.11013925e-01 -7.54733086e-01 -1.17717087e-01
4.53722715e-01 -3.32890421e-01 -1.00165546e+00 8.02994430e-01
3.87599438e-01 -1.07708180e+00 1.09166908e+00 -3.79260838e-01
3.99287105e-01 -1.40579343e-01 -6.12799153e-02 -1.27809274e+00
-3.09909999e-01 -2.47164935e-01 -1.29347704e-02 1.34887314e+00
4.47045326e-01 -6.80739701e-01 9.50984597e-01 4.93767858e-01
-1.57357886e-01 -6.61090374e-01 -9.59846199e-01 -1.12308002e+00
4.69108611e-01 -4.34305251e-01 4.09579307e-01 1.15027201e+00
-4.94015291e-02 4.89459217e-01 6.82525989e-03 -1.31713822e-01
6.06151462e-01 2.75604129e-01 6.41783416e-01 -1.55468273e+00
1.86069191e-01 -6.65721059e-01 -5.15671432e-01 -1.08996534e+00
1.18743986e-01 -9.31787193e-01 7.29181916e-02 -1.67003489e+00
1.02447212e-01 -6.42706513e-01 -5.41596830e-01 4.44704264e-01
-2.19106629e-01 3.07025462e-01 -7.44418725e-02 2.54536539e-01
-5.99849820e-01 5.06197393e-01 1.29195869e+00 -1.65545925e-01
-2.46704087e-01 -1.40191644e-01 -7.45242000e-01 8.77629220e-01
9.47986484e-01 -4.37991410e-01 -5.08426607e-01 -3.74900848e-01
-1.69982210e-01 -4.40539211e-01 1.42695665e-01 -1.02930200e+00
-1.29749492e-01 -1.43606737e-02 2.19544962e-01 -2.22722724e-01
9.34993997e-02 -8.20826173e-01 -4.59269464e-01 1.45350605e-01
-1.27893582e-01 -5.10986030e-01 4.18143719e-01 6.62310839e-01
-4.89232004e-01 -1.51234120e-01 9.06096280e-01 2.62078512e-02
-1.45798075e+00 1.33572310e-01 -1.49192229e-01 2.68365502e-01
1.03845489e+00 -5.81356466e-01 1.06950529e-01 -1.01772211e-01
-6.47482097e-01 6.54062107e-02 8.05845797e-01 4.43681568e-01
2.16808617e-01 -1.21898913e+00 -4.82166380e-01 2.52580583e-01
5.96956909e-01 2.41718948e-01 1.15163483e-01 5.92680275e-01
-2.97261536e-01 3.96311969e-01 -4.23864633e-01 -7.50463367e-01
-1.00918663e+00 4.19415623e-01 2.86822785e-02 -3.90197188e-01
-7.16579020e-01 8.09741139e-01 1.99825183e-01 -5.38414538e-01
2.28334516e-01 -3.67543966e-01 -1.93427101e-01 1.55248269e-01
-3.41648571e-02 3.26038927e-01 2.37765893e-01 -3.74362886e-01
-6.07027948e-01 6.67869329e-01 -1.03706472e-01 -1.56552047e-02
1.37002099e+00 -2.33678013e-01 5.27596593e-01 2.10874170e-01
9.52498198e-01 -1.47804365e-01 -1.61383796e+00 -4.05841053e-01
3.88063312e-01 -3.53942037e-01 1.78865641e-01 -1.00969410e+00
-8.21257353e-01 1.02612853e+00 6.39001787e-01 2.27007255e-01
1.41974807e+00 1.36080533e-01 9.06103790e-01 3.64969037e-02
4.85770732e-01 -1.44830775e+00 -3.05366907e-02 3.38237971e-01
4.86158192e-01 -1.80162644e+00 -3.13909687e-02 -6.02069080e-01
-8.15987468e-01 9.47242618e-01 6.56062961e-01 -1.37336388e-01
5.65248191e-01 1.43256590e-01 2.16927797e-01 -2.10329425e-02
-1.95401311e-01 -5.57505727e-01 4.75683331e-01 1.10115814e+00
3.62296671e-01 1.70250945e-02 -5.52502036e-01 7.99467504e-01
7.47715980e-02 1.17850021e-01 1.10663712e-01 9.72792804e-01
-2.73295969e-01 -1.44291461e+00 -1.87037662e-02 1.94202483e-01
-4.59861785e-01 1.90740258e-01 -4.56386894e-01 1.03749168e+00
9.38288681e-03 6.07337534e-01 -7.80647919e-02 -2.67524928e-01
4.32919860e-01 3.09583813e-01 6.45766735e-01 -6.06773496e-01
-1.77670747e-01 -5.07861376e-02 7.34785274e-02 -4.04040277e-01
-6.03433967e-01 -7.33692825e-01 -1.35252345e+00 9.30546597e-02
-2.01662213e-01 -1.93821087e-01 5.19582093e-01 9.89874184e-01
4.83103901e-01 5.65147161e-01 3.40351731e-01 -6.50126219e-01
-4.22674596e-01 -8.81742120e-01 -5.06051421e-01 7.34800339e-01
1.11409314e-01 -1.19525492e+00 -2.43914917e-01 1.66481704e-01] | [9.733539581298828, 1.4484225511550903] |
7cc8ef85-633e-4a3d-964c-f0fac2674ff7 | ae-textspotter-learning-visual-and-linguistic | 2008.00714 | null | https://arxiv.org/abs/2008.00714v5 | https://arxiv.org/pdf/2008.00714v5.pdf | AE TextSpotter: Learning Visual and Linguistic Representation for Ambiguous Text Spotting | Scene text spotting aims to detect and recognize the entire word or sentence with multiple characters in natural images. It is still challenging because ambiguity often occurs when the spacing between characters is large or the characters are evenly spread in multiple rows and columns, making many visually plausible groupings of the characters (e.g. "BERLIN" is incorrectly detected as "BERL" and "IN" in Fig. 1(c)). Unlike previous works that merely employed visual features for text detection, this work proposes a novel text spotter, named Ambiguity Eliminating Text Spotter (AE TextSpotter), which learns both visual and linguistic features to significantly reduce ambiguity in text detection. The proposed AE TextSpotter has three important benefits. 1) The linguistic representation is learned together with the visual representation in a framework. To our knowledge, it is the first time to improve text detection by using a language model. 2) A carefully designed language module is utilized to reduce the detection confidence of incorrect text lines, making them easily pruned in the detection stage. 3) Extensive experiments show that AE TextSpotter outperforms other state-of-the-art methods by a large margin. For example, we carefully select a validation set of extremely ambiguous samples from the IC19-ReCTS dataset, where our approach surpasses other methods by more than 4%. The code has been released at https://github.com/whai362/AE_TextSpotter. The image list and evaluation scripts of the validation set have been released at https://github.com/whai362/TDA-ReCTS. | ['Tong Lu', 'Chunhua Shen', 'Zhibo Yang', 'Ping Luo', 'Xiaozhong Ji', 'Wenhai Wang', 'Ding Liang', 'Xuebo Liu', 'Enze Xie'] | 2020-08-03 | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/2183_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123590443.pdf | eccv-2020-8 | ['text-spotting'] | ['computer-vision'] | [ 2.06833750e-01 -2.63422906e-01 6.98453560e-02 -1.49991676e-01
-6.48551047e-01 -6.24024868e-01 5.58156848e-01 2.16288000e-01
-4.49782312e-01 4.66322541e-01 6.55655488e-02 -5.31599879e-01
3.78659487e-01 -4.51121211e-01 -6.19184554e-01 -5.53307235e-01
5.14697969e-01 3.79679173e-01 4.37603861e-01 -1.21872611e-01
5.88129640e-01 2.68438041e-01 -1.34463143e+00 5.87821484e-01
1.06016433e+00 6.33482993e-01 6.68405294e-01 5.87285578e-01
-3.89096648e-01 4.59798932e-01 -8.49868536e-01 -5.01673162e-01
2.33564824e-01 -4.76059109e-01 -4.93544221e-01 3.58794957e-01
6.05715513e-01 -4.38667417e-01 -1.47628486e-01 1.20701408e+00
4.00586069e-01 -1.79205820e-01 6.45052969e-01 -1.02234948e+00
-6.40844941e-01 5.71197152e-01 -1.13075697e+00 8.82693306e-02
3.66372943e-01 1.28170878e-01 9.45984423e-01 -1.43185401e+00
3.89370501e-01 1.19613087e+00 4.29375142e-01 3.38756055e-01
-9.83618319e-01 -8.47873271e-01 2.71527410e-01 2.00259179e-01
-1.71324551e+00 -3.40509593e-01 4.96862054e-01 -5.47673583e-01
7.29616344e-01 5.14914751e-01 5.58187902e-01 1.13051224e+00
8.78181010e-02 1.15018487e+00 9.76263344e-01 -8.54576886e-01
8.89417622e-03 2.97013700e-01 1.11624092e-01 8.25681925e-01
5.09465039e-01 -3.27965587e-01 -4.33827192e-01 1.90608934e-01
6.24572575e-01 -6.97300211e-02 -2.77128339e-01 6.58318121e-03
-1.24043918e+00 7.36347795e-01 2.58074075e-01 4.85789746e-01
1.81667015e-01 -1.74770042e-01 2.98147470e-01 -1.07459296e-02
2.58893788e-01 2.53545046e-01 7.58672059e-02 -2.03250740e-02
-1.04998410e+00 -7.31231272e-02 4.44388807e-01 9.52045679e-01
7.16542006e-01 1.11852676e-01 -2.14197487e-01 1.07036531e+00
1.93590149e-01 6.48140609e-01 6.11826301e-01 -2.25240752e-01
8.90637815e-01 7.29690254e-01 1.31440580e-01 -9.67779815e-01
-3.88234138e-01 -1.69211015e-01 -7.19163060e-01 4.01548117e-01
6.09374166e-01 -2.17399850e-01 -1.08537483e+00 1.11031020e+00
-3.13572376e-03 -2.48779029e-01 -2.68028468e-01 9.80899632e-01
8.16924214e-01 8.25533807e-01 -1.57039970e-01 1.38228506e-01
1.65519679e+00 -1.00603032e+00 -7.22458124e-01 -7.22641408e-01
6.08246505e-01 -1.31546307e+00 1.44378066e+00 4.12128687e-01
-7.81319141e-01 -5.05066931e-01 -1.20527756e+00 -1.47365421e-01
-4.86731678e-01 8.17964733e-01 2.13823140e-01 7.32373238e-01
-9.16899741e-01 -5.99134713e-02 -6.58709288e-01 -6.39135957e-01
2.59038806e-01 7.40491748e-02 -2.14426249e-01 -1.46807516e-02
-8.52339029e-01 7.34119415e-01 4.66579258e-01 9.96651575e-02
-3.92041206e-01 -4.15552966e-02 -9.07354355e-01 -3.44895534e-02
6.11590087e-01 -2.53128588e-01 9.46197331e-01 -1.19136584e+00
-1.14495099e+00 8.83036852e-01 -3.92017871e-01 -6.08686432e-02
8.22374940e-01 -3.55300277e-01 -5.12759507e-01 2.80292958e-01
3.02606463e-01 6.86570108e-01 1.09509695e+00 -1.19670033e+00
-7.12936223e-01 -2.04765439e-01 -3.61249417e-01 2.77896583e-01
-4.19626296e-01 1.85784742e-01 -8.52665663e-01 -1.08128512e+00
6.75333366e-02 -6.78423584e-01 2.04321176e-01 1.65922195e-01
-9.34443116e-01 -1.43538445e-01 9.23401594e-01 -7.05381751e-01
1.24671936e+00 -2.15643406e+00 -3.25449944e-01 9.89267379e-02
2.41642624e-01 4.79393810e-01 -1.58114016e-01 5.23558199e-01
-5.68016954e-02 4.02283341e-01 -1.56780556e-01 -4.63074267e-01
-2.63868812e-02 -1.80913836e-01 -3.00894409e-01 4.35393006e-01
2.30780259e-01 6.67385101e-01 -5.87993979e-01 -6.78672850e-01
4.88737792e-01 3.06688815e-01 -7.33829588e-02 -1.66153893e-01
-9.52995196e-02 4.29866090e-02 -4.14635241e-01 8.05577874e-01
8.55038583e-01 -3.40392619e-01 -2.82681710e-03 -7.08880424e-02
-4.79165047e-01 3.09692789e-02 -1.35846806e+00 1.17436361e+00
3.53845730e-02 1.04442954e+00 -2.16849491e-01 -5.50765395e-01
8.48055184e-01 1.02277659e-01 -7.08943829e-02 -7.69663990e-01
1.30235597e-01 1.33670256e-01 -1.47556320e-01 -4.97679383e-01
7.46668279e-01 3.79787445e-01 -2.48076431e-02 2.32270107e-01
-4.53942329e-01 -4.29337993e-02 6.60743952e-01 3.50181192e-01
6.98332310e-01 -1.50036272e-02 2.83079207e-01 -5.89428656e-03
6.16968811e-01 1.27427459e-01 5.35250187e-01 8.96156609e-01
-2.43775055e-01 9.37090874e-01 5.98216772e-01 -2.64460832e-01
-1.16632104e+00 -9.24103975e-01 -2.39707485e-01 8.57891917e-01
3.40511143e-01 -6.67166650e-01 -7.66965151e-01 -6.36572063e-01
-4.59542163e-02 8.93601120e-01 -5.11168540e-01 3.18338364e-01
-4.64505792e-01 -5.01326859e-01 6.63628995e-01 5.30953288e-01
5.85394084e-01 -9.84028578e-01 -4.11404550e-01 -2.27813378e-01
-2.95833796e-01 -1.23696125e+00 -8.15868020e-01 5.22047514e-03
-4.10419166e-01 -1.04418635e+00 -8.06793571e-01 -9.58995283e-01
9.37545717e-01 5.82721114e-01 5.81531107e-01 3.12788635e-01
-5.64469993e-01 8.08899701e-02 -6.33430421e-01 -3.18318307e-01
-3.99944514e-01 -1.60130158e-01 -2.39497527e-01 7.75437504e-02
4.16177779e-01 2.38395482e-01 -6.25062466e-01 4.99535829e-01
-8.35417151e-01 5.49301982e-01 7.76532054e-01 9.04826880e-01
5.16968310e-01 1.74285427e-01 1.50491312e-01 -6.61863208e-01
4.45572436e-01 -7.55794942e-02 -7.34577537e-01 3.14852804e-01
-4.38349664e-01 -1.75730944e-01 7.35336185e-01 -3.94712448e-01
-9.42482948e-01 1.12947024e-01 -3.66804190e-02 -2.27761790e-01
-4.48525131e-01 1.97364017e-01 -2.34476924e-02 9.85124931e-02
4.85131294e-01 4.25832570e-01 -2.48715535e-01 -3.92239183e-01
1.18826188e-01 9.42214489e-01 2.76124805e-01 -1.84628725e-01
8.27232420e-01 5.30925810e-01 -5.04835784e-01 -1.18336022e+00
-5.57170093e-01 -5.53567827e-01 -5.80575168e-01 -2.27534160e-01
7.03440905e-01 -8.03802848e-01 -3.27219754e-01 6.17052972e-01
-9.92664099e-01 -3.52276862e-01 2.01510683e-01 4.75486010e-01
-8.97654667e-02 8.55753243e-01 -5.32537639e-01 -8.36472332e-01
-2.81161606e-01 -1.14879930e+00 1.20940435e+00 4.37677085e-01
-2.39943430e-01 -6.57808304e-01 -3.40622604e-01 4.36441004e-01
3.03703956e-02 -5.93658574e-02 7.42899358e-01 -6.07190132e-01
-3.99031132e-01 -2.36504242e-01 -4.83951002e-01 8.66315514e-02
1.96899384e-01 4.15893972e-01 -7.75409043e-01 -3.44221890e-01
-4.05034840e-01 -5.79164773e-02 9.56340671e-01 2.74468184e-01
1.08693516e+00 -1.90165162e-01 -4.39754337e-01 3.82891655e-01
1.19925058e+00 3.19465786e-01 7.21272945e-01 6.48244083e-01
8.23244750e-01 5.72537243e-01 7.00988770e-01 4.99760240e-01
1.67163000e-01 6.42753363e-01 2.41161451e-01 -3.59646559e-01
-3.11930597e-01 -2.55991966e-01 4.88349944e-01 4.19727385e-01
4.64139432e-01 -6.70314312e-01 -1.06527424e+00 4.95632887e-01
-1.81804824e+00 -7.94310808e-01 -5.10057569e-01 2.17956805e+00
6.26217902e-01 2.15444088e-01 6.77096546e-02 3.23233724e-01
1.22335732e+00 2.19467252e-01 -3.45569074e-01 -3.50265175e-01
-3.59951735e-01 -1.50251061e-01 4.72367167e-01 5.48069656e-01
-1.21537542e+00 1.19622862e+00 5.07386589e+00 1.09381688e+00
-1.19054186e+00 -2.77005821e-01 7.60700405e-01 4.48281085e-03
-3.44747142e-03 1.93264727e-02 -1.12621951e+00 7.17322409e-01
1.99590683e-01 3.19339186e-02 1.68116942e-01 4.89161789e-01
4.48766261e-01 -3.92083883e-01 -6.95397854e-01 1.05258811e+00
4.46237952e-01 -9.94158685e-01 1.72224417e-01 -1.97047219e-01
4.43078429e-01 -1.52149901e-01 1.64128929e-01 -8.71360675e-02
-3.85752395e-02 -9.48749423e-01 1.05613470e+00 1.90920278e-01
9.71071005e-01 -5.25930524e-01 5.07982552e-01 1.56777039e-01
-1.20480740e+00 4.42600250e-02 -4.11878705e-01 1.51802510e-01
4.44483571e-02 6.80543244e-01 -1.07700610e+00 3.34924281e-01
7.53938794e-01 6.24254763e-01 -1.11670482e+00 1.19218290e+00
-5.93790174e-01 7.01932907e-01 -2.39082605e-01 -1.89940169e-01
1.69159681e-01 -2.15027496e-01 7.46725619e-01 1.46627283e+00
4.35312659e-01 -3.35098594e-01 1.74110681e-01 9.21351373e-01
1.33028314e-01 5.03560543e-01 -5.09122014e-01 -1.95096105e-01
4.19963330e-01 1.40898359e+00 -1.18316436e+00 -3.05258602e-01
-5.88874996e-01 1.19450057e+00 1.27081141e-01 4.40903306e-01
-7.62638509e-01 -7.77481556e-01 3.67047377e-02 1.60688221e-01
5.04088998e-01 -1.90276697e-01 -4.69752342e-01 -1.26863074e+00
2.54614860e-01 -8.76497626e-01 4.93805856e-01 -1.04965842e+00
-1.11610949e+00 7.00669944e-01 -3.08236778e-01 -1.37115479e+00
4.65804636e-02 -8.28481257e-01 -6.04149938e-01 8.22348714e-01
-1.19851172e+00 -1.15639567e+00 -5.45283616e-01 5.40271521e-01
9.49236214e-01 -2.32258111e-01 4.91561621e-01 4.65694889e-02
-1.04424679e+00 8.91919315e-01 3.02048743e-01 5.17066598e-01
1.05033815e+00 -1.22329736e+00 5.30849993e-01 1.35871840e+00
3.00355703e-01 3.91932338e-01 6.07258260e-01 -8.77071023e-01
-1.01067436e+00 -9.16636884e-01 8.24136555e-01 -1.98148310e-01
7.39281774e-01 -7.65115559e-01 -9.68793571e-01 7.05252588e-01
4.16235775e-01 -3.25737715e-01 3.87638569e-01 -2.95482337e-01
-2.94979572e-01 1.51363984e-01 -8.23009610e-01 1.08382475e+00
5.31146646e-01 -2.38534495e-01 -5.27667880e-01 4.17672873e-01
2.52119482e-01 -4.84593272e-01 -1.43356636e-01 5.25454730e-02
3.98976207e-01 -1.02037072e+00 6.44063354e-01 2.42005900e-01
3.36941451e-01 -7.00482190e-01 1.01056457e-01 -8.67459893e-01
-1.21704720e-01 -4.97355729e-01 2.61321992e-01 1.40700209e+00
4.92013782e-01 -7.61957526e-01 6.13365054e-01 1.35396183e-01
-9.78521332e-02 -4.18043047e-01 -8.60730052e-01 -6.83538377e-01
3.50644737e-02 -2.84598291e-01 1.62500307e-01 8.76787722e-01
1.00720055e-01 2.81783730e-01 -4.33086336e-01 1.49657935e-01
3.26415360e-01 8.60054120e-02 6.59114242e-01 -7.90381789e-01
-1.64093241e-01 -7.96109617e-01 -2.66440690e-01 -1.20748186e+00
-2.12944552e-01 -6.54363394e-01 1.14818744e-01 -1.53834701e+00
3.12947035e-01 -2.51878023e-01 1.50142431e-01 6.90341413e-01
-5.69056749e-01 3.24343771e-01 3.50773543e-01 3.20891619e-01
-5.91055095e-01 3.43268305e-01 1.11156917e+00 -2.35593468e-01
-1.46036953e-01 -1.26022786e-01 -6.98981881e-01 8.40408623e-01
1.04089451e+00 -3.25461686e-01 -6.34371340e-02 -5.12697041e-01
1.45191237e-01 -2.29581803e-01 3.88842732e-01 -8.13055873e-01
3.64583760e-01 -6.51297942e-02 7.63632357e-01 -9.93124008e-01
1.22347616e-01 -5.79811752e-01 -3.49061191e-01 3.47531229e-01
-1.17851898e-01 2.17330888e-01 4.94993925e-01 3.37994307e-01
-8.76488388e-02 -5.72485745e-01 7.22454488e-01 1.60888046e-01
-8.52923930e-01 -2.45853886e-01 -7.85079539e-01 -6.36141300e-02
9.49089468e-01 -5.22273898e-01 -7.37442076e-01 -3.99548978e-01
-2.13853046e-01 3.47382575e-01 6.99628949e-01 4.96170908e-01
7.23922849e-01 -9.76910293e-01 -7.98351407e-01 3.71240646e-01
3.47348422e-01 -1.25193268e-01 7.95154423e-02 8.96129668e-01
-8.66578162e-01 3.28859329e-01 -3.27877030e-02 -6.97798014e-01
-1.49828291e+00 4.24469829e-01 1.78596199e-01 2.21204273e-02
-9.77674961e-01 6.04223430e-01 5.47910392e-01 5.95971867e-02
3.44920278e-01 -1.89014956e-01 -1.57533020e-01 -5.08833602e-02
7.05132306e-01 2.57074922e-01 -1.23963818e-01 -7.46633708e-01
-4.02371079e-01 7.36273706e-01 -5.31110883e-01 -1.29403263e-01
8.23101878e-01 -3.44576478e-01 5.49607649e-02 3.90551716e-01
7.48405814e-01 5.84531426e-01 -1.10424507e+00 -2.35798396e-02
2.37000957e-02 -7.58621633e-01 -1.52405038e-01 -9.64869440e-01
-8.21426630e-01 8.61425042e-01 6.64562821e-01 1.82514310e-01
1.06481647e+00 8.07857234e-03 5.50887525e-01 2.12736994e-01
-1.07597239e-01 -1.16453743e+00 3.19983780e-01 5.32380521e-01
1.04865766e+00 -1.38601625e+00 1.04423538e-01 -5.71301639e-01
-1.00612533e+00 1.28550148e+00 8.42727900e-01 9.13654268e-02
1.95471640e-03 4.18668568e-01 3.14702660e-01 1.95240416e-02
-3.55634034e-01 -3.26622784e-01 3.26801986e-01 3.67762238e-01
5.20257473e-01 -1.25844656e-02 -3.65307808e-01 3.21573287e-01
-1.63509622e-01 -5.96350670e-01 6.33360744e-01 8.22648346e-01
-5.50560355e-01 -1.11605704e+00 -7.73649275e-01 3.26683044e-01
-3.86737704e-01 -3.44830811e-01 -8.10690820e-01 9.85532463e-01
-4.94865701e-02 1.08590198e+00 1.83566034e-01 -1.97005585e-01
1.27363831e-01 -5.04457392e-02 1.27425194e-01 -4.61129487e-01
-3.42539281e-01 4.89193350e-01 -6.13952279e-02 -1.91079617e-01
4.26760875e-02 -6.14155710e-01 -1.43704963e+00 -2.13382170e-01
-4.91058171e-01 -1.83499515e-01 5.51168978e-01 7.04597294e-01
2.44246870e-01 4.52068895e-01 4.30909485e-01 -5.20189524e-01
-5.83828427e-02 -9.46844757e-01 -5.55329740e-01 3.26135069e-01
2.34169394e-01 -4.85684901e-01 -5.30113041e-01 9.40478221e-02] | [11.923871040344238, 2.236082077026367] |
288a0d9e-ae2f-4a21-8da5-1bdfd07e726a | lexis-an-optimization-framework-for | 1602.05561 | null | http://arxiv.org/abs/1602.05561v3 | http://arxiv.org/pdf/1602.05561v3.pdf | Lexis: An Optimization Framework for Discovering the Hierarchical Structure of Sequential Data | Data represented as strings abounds in biology, linguistics, document mining,
web search and many other fields. Such data often have a hierarchical
structure, either because they were artificially designed and composed in a
hierarchical manner or because there is an underlying evolutionary process that
creates repeatedly more complex strings from simpler substrings. We propose a
framework, referred to as "Lexis", that produces an optimized hierarchical
representation of a given set of "target" strings. The resulting hierarchy,
"Lexis-DAG", shows how to construct each target through the concatenation of
intermediate substrings, minimizing the total number of such concatenations or
DAG edges. The Lexis optimization problem is related to the smallest grammar
problem. After we prove its NP-Hardness for two cost formulations, we propose
an efficient greedy algorithm for the construction of Lexis-DAGs. We also
consider the problem of identifying the set of intermediate nodes (substrings)
that collectively form the "core" of a Lexis-DAG, which is important in the
analysis of Lexis-DAGs. We show that the Lexis framework can be applied in
diverse applications such as optimized synthesis of DNA fragments in genomic
libraries, hierarchical structure discovery in protein sequences,
dictionary-based text compression, and feature extraction from a set of
documents. | ['Payam Siyari', 'Constantine Dovrolis', 'Bistra Dilkina'] | 2016-02-17 | null | null | null | null | ['text-compression'] | ['natural-language-processing'] | [ 9.16575074e-01 1.57861814e-01 -2.54837483e-01 -2.95259506e-01
-3.50319922e-01 -8.83239090e-01 3.58643681e-02 8.84400487e-01
-2.38271877e-02 6.98534369e-01 2.74494559e-01 -4.81592000e-01
-5.52277386e-01 -1.22113800e+00 -9.59283769e-01 -7.83272326e-01
-4.24930304e-01 6.36884987e-01 5.72361536e-02 -5.87799549e-02
5.80508769e-01 4.38902050e-01 -1.89634049e+00 3.44157785e-01
8.27228606e-01 4.90995318e-01 4.32370454e-01 5.73484480e-01
-5.92905045e-01 -8.51875171e-03 -6.27813756e-01 -5.48008502e-01
1.99013278e-01 -7.24641025e-01 -9.28134084e-01 2.77026147e-01
2.55951017e-01 3.30665439e-01 -1.57859281e-01 1.31479931e+00
-2.05728244e-02 -5.71902022e-02 5.52372336e-01 -9.98480976e-01
-2.22398445e-01 1.03626037e+00 -4.14599150e-01 -1.51114494e-01
4.64932173e-01 -4.84977603e-01 1.26379132e+00 -3.53578627e-01
8.59783947e-01 1.24403799e+00 3.00986439e-01 2.37807170e-01
-1.28857493e+00 -9.93660316e-02 -1.45885944e-01 -7.36660808e-02
-1.29192102e+00 -5.58545850e-02 3.90983045e-01 -5.14669716e-01
1.04020715e+00 7.09103703e-01 7.96483755e-01 4.55830127e-01
3.98227960e-01 5.21807671e-01 5.28915763e-01 -7.80670524e-01
1.90517128e-01 -2.57948637e-01 6.07879162e-01 9.05316472e-01
8.13718259e-01 1.56376755e-03 -4.70088571e-01 -5.31921923e-01
2.10164383e-01 -1.30283743e-01 4.10578884e-02 -2.14119807e-01
-9.65353906e-01 9.62095916e-01 -1.17052555e-01 5.09845376e-01
-3.63213615e-03 -1.39591724e-01 4.69647110e-01 3.38206023e-01
-1.49318129e-01 6.80481672e-01 -2.34251797e-01 1.15643792e-01
-6.01809025e-01 2.49497205e-01 1.03205395e+00 1.12214434e+00
7.26822734e-01 -2.90809631e-01 5.61543107e-02 7.30418146e-01
-8.17295015e-02 9.27523971e-02 5.98908067e-01 -4.70473677e-01
4.28635538e-01 9.63225603e-01 -2.78703958e-01 -1.03777838e+00
-6.11421704e-01 -3.58552754e-01 -8.12192321e-01 -5.52297592e-01
3.19577098e-01 2.04244033e-01 -6.98278129e-01 1.94054210e+00
3.99654955e-01 -2.12480918e-01 -7.14117289e-02 3.89634311e-01
6.20126784e-01 9.98661220e-01 -2.42740273e-01 -6.53291106e-01
1.47657824e+00 -4.60073650e-01 -4.20006990e-01 1.40936691e-02
8.83926332e-01 -4.80998725e-01 8.70715022e-01 5.04923224e-01
-1.20154500e+00 -5.80341555e-02 -1.09013832e+00 -7.15409517e-02
-3.88563871e-01 -2.52006084e-01 5.81341863e-01 6.75341368e-01
-7.20851004e-01 7.77363598e-01 -2.91810364e-01 -4.43141311e-01
-1.29293755e-01 5.10207057e-01 -3.11445355e-01 1.97107628e-01
-8.92498732e-01 4.92776603e-01 8.03585052e-01 -3.77522618e-01
-5.20647168e-01 -3.13746661e-01 -8.38780940e-01 2.81090498e-01
4.82687622e-01 -5.79634666e-01 7.15817690e-01 -7.73375332e-01
-9.88357425e-01 1.04610872e+00 -4.37788248e-01 -5.28972089e-01
-2.08624959e-01 4.48694676e-01 -1.45509586e-01 -1.59295872e-01
-5.62546365e-02 1.22415997e-01 6.02673173e-01 -9.60498989e-01
-8.19534183e-01 -6.42220020e-01 -6.57488629e-02 3.21748555e-02
-4.86943036e-01 4.76691797e-02 -5.79966128e-01 -8.54861259e-01
4.09910411e-01 -1.10809398e+00 -5.07327676e-01 -6.86545551e-01
-7.49288380e-01 -3.89877349e-01 3.38398129e-01 -4.26077962e-01
1.71099436e+00 -2.02528310e+00 7.13847697e-01 8.28915298e-01
2.56278753e-01 1.18325744e-02 -3.03093106e-01 8.95695329e-01
-1.11138500e-01 4.41712230e-01 -5.93556285e-01 3.38003337e-01
-7.72065744e-02 5.44955492e-01 -3.13483387e-01 2.02042863e-01
-1.66016012e-01 5.86682439e-01 -7.86771715e-01 -5.35294771e-01
-2.94963926e-01 -3.82715911e-01 -7.36065447e-01 9.47875381e-02
-5.46685278e-01 -3.35048549e-02 -3.54217112e-01 4.82467145e-01
4.87347066e-01 -2.71460652e-01 8.76841962e-01 -1.45511329e-01
-2.51811266e-01 3.73433679e-01 -1.24132645e+00 1.47541869e+00
2.53456868e-02 2.59301364e-01 -3.11129093e-01 -1.32973981e+00
1.04007661e+00 -2.17572734e-01 6.94382071e-01 -4.36664373e-01
1.52666673e-01 2.74508774e-01 1.99681386e-01 -6.36831820e-01
7.16927707e-01 -8.48936588e-02 -5.39382160e-01 3.55081856e-01
-1.16669469e-01 8.08235351e-03 1.00866520e+00 4.21654880e-01
1.20205832e+00 -4.57826376e-01 7.39898562e-01 -4.37359899e-01
5.33834040e-01 5.16729765e-02 7.46755898e-01 8.96174967e-01
6.05842054e-01 2.19594553e-01 8.78019691e-01 -4.12690043e-01
-1.24542117e+00 -7.71039128e-01 -2.41736561e-01 1.12677979e+00
1.67324558e-01 -1.02174497e+00 -9.04145479e-01 -5.15991211e-01
1.16766348e-01 4.13421094e-01 -4.32086080e-01 -1.71599448e-01
-6.86148942e-01 -7.96386719e-01 5.56326449e-01 4.68933303e-03
-2.09646031e-01 -8.57608855e-01 -5.50151825e-01 5.89738488e-01
-1.64937750e-01 -7.93120265e-01 -6.06997252e-01 5.90503871e-01
-1.05564821e+00 -9.58036780e-01 -1.18337218e-02 -9.06256258e-01
8.52289736e-01 -3.00642420e-02 1.04371512e+00 3.56641084e-01
-4.15851086e-01 -2.64725327e-01 -5.70902288e-01 -4.96027738e-01
-7.06076682e-01 3.68204176e-01 -7.95835443e-03 -1.07915163e-01
1.78180218e-01 -6.65391266e-01 -7.51046911e-02 1.05701633e-01
-1.47463608e+00 -2.37709247e-02 5.36798894e-01 8.47456217e-01
9.71655786e-01 3.75929475e-01 4.65930611e-01 -1.20968723e+00
6.87045336e-01 -6.34035826e-01 -8.71873140e-01 6.11557364e-01
-5.40664792e-01 7.11538970e-01 7.24816084e-01 -3.06546420e-01
-4.04562712e-01 3.64055812e-01 -1.23926915e-01 4.31754887e-01
1.90986708e-01 1.06521940e+00 -5.16184211e-01 8.01890641e-02
4.88413751e-01 4.95968610e-01 -7.55601153e-02 -5.02778888e-01
2.00157523e-01 7.06033111e-01 6.37348890e-01 -8.80770922e-01
3.44702303e-01 2.19669521e-01 6.06971145e-01 -1.05664265e+00
-6.07753694e-01 -3.38340223e-01 -4.76659060e-01 2.16979548e-01
3.85057986e-01 -1.10907882e-01 -7.74371147e-01 9.94934142e-02
-1.16819763e+00 1.86822951e-01 -2.65050977e-01 9.49188508e-03
-5.91791630e-01 7.41992950e-01 -3.23650122e-01 -6.05620742e-01
-3.57445538e-01 -1.13245702e+00 8.56260359e-01 -5.44938026e-03
-5.52966237e-01 -5.27317464e-01 1.45563439e-01 -9.42887664e-02
-4.59111929e-01 3.11348587e-01 1.98961270e+00 -8.63065362e-01
-4.86676663e-01 -1.58363163e-01 2.25510448e-01 -5.71695045e-02
9.97432619e-02 1.05097011e-01 4.73297015e-02 -2.75074273e-01
-2.35485345e-01 5.00662848e-02 6.54436827e-01 1.40815377e-01
1.45085442e+00 -9.08610761e-01 -6.25171602e-01 7.00680017e-01
1.47514474e+00 6.58344448e-01 7.01792896e-01 1.73505560e-01
3.77749681e-01 1.05651426e+00 3.87482435e-01 6.39462113e-01
-1.38319060e-01 7.36482441e-01 3.04671407e-01 3.89843464e-01
3.25948954e-01 -3.49728495e-01 2.08195239e-01 1.10321116e+00
2.96246618e-01 -7.02929616e-01 -7.19913661e-01 6.11470163e-01
-1.72166276e+00 -8.44828546e-01 -4.28747892e-01 2.40681458e+00
1.12735963e+00 6.49120100e-03 3.02835345e-01 5.75622201e-01
6.61566913e-01 -3.62878382e-01 -4.55304831e-01 -1.06168258e+00
-3.39353472e-01 4.67801601e-01 6.68978870e-01 4.81788874e-01
-7.97235787e-01 7.59445548e-01 6.33465290e+00 8.80224407e-01
-6.37311697e-01 -3.32509995e-01 1.90706864e-01 6.47148192e-02
-6.00543737e-01 1.19700648e-01 -1.03675461e+00 6.00486934e-01
9.55226123e-01 -6.24269664e-01 3.46581995e-01 5.80238700e-01
-1.05822839e-01 -1.01770617e-01 -1.23685515e+00 7.31102407e-01
-1.29083276e-01 -1.56004989e+00 5.83212674e-01 4.02224630e-01
7.76070774e-01 -5.40302575e-01 -8.45597237e-02 -3.01003635e-01
3.67366105e-01 -9.44580257e-01 7.14878857e-01 1.20414637e-01
7.70976722e-01 -8.00947547e-01 1.76870912e-01 4.75371242e-01
-1.18709636e+00 -2.81949818e-01 -4.86707509e-01 1.79930687e-01
1.28337413e-01 9.53216076e-01 -6.23582840e-01 6.54600739e-01
2.89581746e-01 3.07915270e-01 -2.06947997e-01 1.14869690e+00
1.80276260e-01 4.45775419e-01 -4.75585878e-01 -4.41031307e-01
2.07137585e-01 -6.34003103e-01 7.53418863e-01 1.56947649e+00
5.76979995e-01 4.54968721e-01 1.34775922e-01 3.68318379e-01
-2.48562336e-01 4.87964392e-01 -8.33917797e-01 -3.38099211e-01
7.28711247e-01 8.40657532e-01 -9.34667766e-01 -4.00615573e-01
-1.22726798e-01 7.17314303e-01 2.97332287e-01 -1.04208760e-01
-4.54726338e-01 -7.37301707e-01 7.13897169e-01 1.89167723e-01
1.30924627e-01 -3.10283631e-01 -4.79935199e-01 -8.59308541e-01
-4.77255955e-02 -1.24031544e+00 7.21587896e-01 -1.67359740e-01
-8.70597064e-01 5.22308767e-01 7.95226693e-02 -8.62321019e-01
-3.60316724e-01 -5.66894054e-01 -2.39122495e-01 4.01763737e-01
-5.91332376e-01 -4.23846096e-01 1.33950472e-01 3.92194301e-01
4.69874650e-01 3.50188091e-03 8.09441864e-01 1.52962849e-01
-5.91590703e-01 6.87033474e-01 5.21387398e-01 -2.94149131e-01
-1.53486303e-03 -1.06729817e+00 3.24802309e-01 9.31858897e-01
3.26907545e-01 7.70507574e-01 8.14071059e-01 -8.10881376e-01
-1.70863116e+00 -1.05077946e+00 1.28613663e+00 1.40868584e-02
4.04868543e-01 -4.35518265e-01 -8.84735465e-01 5.02417505e-01
-2.06292823e-01 -7.70513475e-01 7.71955967e-01 -4.12139669e-02
-2.80518651e-01 -9.85957868e-03 -9.91840661e-01 7.06126869e-01
1.53637290e+00 1.67667996e-02 -3.91404867e-01 5.09147108e-01
7.08245814e-01 -2.85380095e-01 -8.52766514e-01 3.74885291e-01
6.36249721e-01 -5.41949511e-01 8.45996082e-01 -1.13843286e+00
5.12305617e-01 -2.52435595e-01 -2.30902076e-01 -1.00729966e+00
-4.72611457e-01 -8.85082364e-01 -1.51781186e-01 1.03594661e+00
5.33986151e-01 -3.11367184e-01 9.68112886e-01 9.31482203e-03
-6.09162338e-02 -8.23412836e-01 -8.19644034e-01 -1.14258921e+00
-7.14542866e-02 -5.87909482e-02 1.00368166e+00 7.19896138e-01
4.08130139e-01 4.38478589e-01 -2.41694376e-01 -8.76395032e-02
7.58570731e-01 4.53010559e-01 6.29542708e-01 -1.41323924e+00
-3.55288595e-01 -6.26390874e-01 -3.78258914e-01 -1.07391238e+00
8.04218799e-02 -1.41823375e+00 8.37182775e-02 -1.34549594e+00
3.06555063e-01 -5.83734810e-01 9.45326611e-02 4.35640275e-01
1.04060933e-01 -3.90319109e-01 9.66453254e-02 1.58408642e-01
-1.98084429e-01 1.14036895e-01 8.58877718e-01 -1.90996259e-01
-3.91623974e-01 -2.22116843e-01 -8.03577423e-01 5.83976448e-01
6.71012819e-01 -8.92082393e-01 -2.66070276e-01 -2.49596149e-01
6.72523320e-01 2.56778628e-01 -4.15352166e-01 -4.75557208e-01
2.31282175e-01 -5.29236138e-01 -3.35169315e-01 -5.24422050e-01
-1.60257027e-01 -6.88835919e-01 6.18817449e-01 7.80013263e-01
-6.23419464e-01 2.10500583e-01 -2.99592037e-02 3.84455293e-01
-5.36655374e-02 -9.66195643e-01 7.12098598e-01 8.63957964e-03
-3.03271890e-01 2.32230812e-01 -6.28281415e-01 -1.49757519e-01
1.03954601e+00 -4.31744367e-01 -1.88703433e-01 9.52541549e-03
-5.59303701e-01 -3.55778635e-03 2.52349854e-01 2.52006978e-01
6.16357625e-01 -1.05441785e+00 -7.22934544e-01 4.14927185e-01
2.15791300e-01 -1.42594069e-01 -2.73499519e-01 3.32312524e-01
-5.26271939e-01 5.26905239e-01 -2.56249215e-02 -2.84079462e-01
-1.63560402e+00 7.29151189e-01 -2.35953927e-01 -6.03948295e-01
-2.98573107e-01 8.25004399e-01 3.79237056e-01 -1.29407093e-01
2.67532170e-01 -4.03602690e-01 -8.00506920e-02 2.80699879e-02
3.60901356e-01 3.86373520e-01 3.28651577e-01 -4.72453117e-01
-1.24368221e-01 5.74214756e-01 -9.53805745e-02 7.67279044e-02
1.19974577e+00 3.16174030e-02 -9.41271901e-01 9.59165674e-03
1.32359290e+00 2.28260942e-02 -1.94260582e-01 -1.61785245e-01
6.10276699e-01 -3.74949634e-01 -7.10337341e-01 -2.85609215e-01
-6.58287942e-01 2.61148959e-01 -3.76517922e-02 5.41720033e-01
1.17755628e+00 2.76831865e-01 8.41020584e-01 7.63497293e-01
3.56905282e-01 -8.55895281e-01 -1.14457138e-01 8.12896490e-01
7.85201907e-01 -2.06089944e-01 -2.71257877e-01 -7.29703903e-01
-6.99026734e-02 1.19496453e+00 1.48146063e-01 1.64111003e-01
4.66945171e-01 7.23138690e-01 -8.03349376e-01 -3.44258547e-02
-7.31428504e-01 -3.23708862e-01 1.12244129e-01 3.10256153e-01
2.99609244e-01 1.94885045e-01 -1.29857683e+00 8.65327120e-01
-5.83527207e-01 -1.81315079e-01 4.34975803e-01 1.04745400e+00
-8.53486538e-01 -1.71496248e+00 -4.50255483e-01 8.22812259e-01
-4.95274723e-01 -1.73622131e-01 -7.25108147e-01 2.76316613e-01
4.45679188e-01 8.02932262e-01 8.12290981e-02 -5.94991028e-01
2.85705328e-01 -4.76905555e-02 7.07919359e-01 -8.12444568e-01
-4.95433778e-01 -9.09826439e-03 1.55044258e-01 -1.07068576e-01
-1.72403660e-02 -7.07397282e-01 -1.44216585e+00 -3.88404548e-01
-3.11626822e-01 5.35240889e-01 7.31831729e-01 7.41902769e-01
3.48540068e-01 2.62191266e-01 7.52816081e-01 -4.42475528e-02
-3.66723597e-01 -4.69749451e-01 -6.46333158e-01 4.44152415e-01
-1.48900852e-01 -2.21388623e-01 1.23997211e-01 3.18432838e-01] | [7.287760257720947, 5.615289211273193] |
51cc5f84-074f-44bb-9cd9-9cafc4b4162a | zero-shot-style-transfer-for-gesture | 2208.01917 | null | https://arxiv.org/abs/2208.01917v1 | https://arxiv.org/pdf/2208.01917v1.pdf | Zero-Shot Style Transfer for Gesture Animation driven by Text and Speech using Adversarial Disentanglement of Multimodal Style Encoding | Modeling virtual agents with behavior style is one factor for personalizing human agent interaction. We propose an efficient yet effective machine learning approach to synthesize gestures driven by prosodic features and text in the style of different speakers including those unseen during training. Our model performs zero shot multimodal style transfer driven by multimodal data from the PATS database containing videos of various speakers. We view style as being pervasive while speaking, it colors the communicative behaviors expressivity while speech content is carried by multimodal signals and text. This disentanglement scheme of content and style allows us to directly infer the style embedding even of speaker whose data are not part of the training phase, without requiring any further training or fine tuning. The first goal of our model is to generate the gestures of a source speaker based on the content of two audio and text modalities. The second goal is to condition the source speaker predicted gestures on the multimodal behavior style embedding of a target speaker. The third goal is to allow zero shot style transfer of speakers unseen during training without retraining the model. Our system consists of: (1) a speaker style encoder network that learns to generate a fixed dimensional speaker embedding style from a target speaker multimodal data and (2) a sequence to sequence synthesis network that synthesizes gestures based on the content of the input modalities of a source speaker and conditioned on the speaker style embedding. We evaluate that our model can synthesize gestures of a source speaker and transfer the knowledge of target speaker style variability to the gesture generation task in a zero shot setup. We convert the 2D gestures to 3D poses and produce 3D animations. We conduct objective and subjective evaluations to validate our approach and compare it with a baseline. | ['Nicolas Obin', 'Catherine Pelachaud', 'Michele Grimaldi', 'Mireille Fares'] | 2022-08-03 | null | null | null | null | ['gesture-generation'] | ['robots'] | [ 3.74199152e-01 3.35934788e-01 3.88777032e-02 -5.99655271e-01
-4.94174808e-01 -7.93388486e-01 1.10357046e+00 -7.59295702e-01
-4.07962918e-01 3.18297356e-01 6.01611137e-01 1.40545264e-01
5.77406049e-01 -4.20891404e-01 -7.33127236e-01 -7.70706296e-01
1.60910115e-01 9.52433050e-01 -7.72856735e-03 -5.08126199e-01
-1.40164837e-01 5.41988611e-01 -1.60290134e+00 5.55739880e-01
1.56386435e-01 7.08994031e-01 2.44315982e-01 1.25187004e+00
-2.69589424e-01 7.39078939e-01 -9.63892996e-01 5.64261712e-02
1.83619291e-01 -7.26122260e-01 -8.03749025e-01 4.40958738e-01
2.61313379e-01 -6.52233958e-01 -2.67951876e-01 7.10533082e-01
4.24219966e-01 3.62231880e-01 1.05792177e+00 -1.45174086e+00
-3.11637700e-01 8.09810281e-01 -1.03397809e-01 -4.74781096e-01
7.37692237e-01 6.49959862e-01 8.07861328e-01 -5.00204921e-01
1.17555678e+00 1.85030067e+00 3.37003097e-02 1.29349172e+00
-1.41975796e+00 -4.61009234e-01 2.61613667e-01 -3.12066134e-02
-1.08575118e+00 -6.19004667e-01 9.63391900e-01 -6.35874391e-01
7.19270885e-01 3.07068586e-01 6.91298246e-01 2.12315559e+00
-2.85353929e-01 7.59860814e-01 6.80057049e-01 -3.57817024e-01
4.36379284e-01 3.28547746e-01 -9.79922116e-02 5.00813603e-01
-9.22412217e-01 3.93798202e-01 -6.98257148e-01 -2.83882827e-01
7.59376287e-01 -2.73103744e-01 -3.10072035e-01 -2.37527907e-01
-1.34829915e+00 8.89793456e-01 7.80599266e-02 2.28361994e-01
-2.55694717e-01 2.15036556e-01 4.43371892e-01 6.29817009e-01
6.25681970e-03 2.96325356e-01 -4.79097128e-01 -3.63940895e-01
-4.88292813e-01 3.55401218e-01 1.14539278e+00 1.12031901e+00
4.91782427e-01 2.96760887e-01 -3.15601498e-01 7.02204585e-01
5.53827286e-01 9.21163321e-01 7.43466556e-01 -1.13440073e+00
2.06422523e-01 4.55247790e-01 1.26198173e-01 -4.28022861e-01
-2.91117787e-01 4.86564279e-01 -3.26635778e-01 5.77215731e-01
5.15270412e-01 -6.05646908e-01 -1.11934805e+00 2.21154714e+00
3.49753022e-01 2.11509749e-01 3.48449647e-01 9.56060827e-01
1.04180241e+00 9.76130664e-01 4.78679687e-02 -2.21226931e-01
1.39873910e+00 -7.76743352e-01 -7.08857954e-01 -9.22039151e-02
3.92446250e-01 -6.72553241e-01 1.42813599e+00 9.39226523e-02
-8.47099900e-01 -6.41212940e-01 -9.11217868e-01 1.70059666e-01
-1.68699905e-01 -8.27299505e-02 7.18021169e-02 5.10185659e-01
-9.40464020e-01 3.32745701e-01 -1.00006878e+00 -5.76945603e-01
-1.70496657e-01 4.57174897e-01 -2.88436562e-01 5.84628761e-01
-9.47752357e-01 7.23871887e-01 2.42283657e-01 -4.39952016e-01
-1.22687316e+00 -4.69179600e-01 -1.15470791e+00 -7.63729066e-02
2.80893236e-01 -8.17181945e-01 1.65242624e+00 -1.52923119e+00
-2.57606173e+00 6.24051809e-01 -3.38404447e-01 -1.09230697e-01
7.41605222e-01 2.23002508e-01 -1.89087495e-01 1.82963639e-01
-2.63520241e-01 1.08614790e+00 1.24424243e+00 -1.54691398e+00
-5.87992907e-01 -3.11468601e-01 -6.17314279e-02 5.31990886e-01
-1.56971142e-01 -7.13456934e-03 -3.21726233e-01 -5.23382902e-01
-3.61503154e-01 -1.27850914e+00 1.85654730e-01 -8.76600370e-02
-4.62843597e-01 -1.61909789e-01 1.16675639e+00 -4.66545820e-01
4.56228852e-01 -2.21743369e+00 1.01895177e+00 2.58118659e-01
-1.79526716e-01 -1.39957830e-01 -5.78180194e-01 3.55498940e-01
1.83446348e-01 -7.41472319e-02 -9.78816524e-02 -6.69115663e-01
2.70071208e-01 3.33573282e-01 -5.07659435e-01 2.35479593e-01
1.64848477e-01 6.61129177e-01 -7.56101608e-01 -3.14825118e-01
2.96573430e-01 7.68965423e-01 -6.77272081e-01 1.07490969e+00
-7.08630264e-01 1.24247408e+00 -3.87067080e-01 1.74240455e-01
-8.96631479e-02 1.19622871e-01 3.98039788e-01 -2.10821226e-01
9.33209732e-02 2.15740129e-01 -1.18390167e+00 1.89461517e+00
-5.08012652e-01 5.66669047e-01 2.75655866e-01 -3.80330652e-01
8.25105369e-01 7.84693420e-01 2.70438582e-01 -1.21957473e-01
4.78497237e-01 -9.59939212e-02 8.60561728e-02 -8.80237103e-01
2.16690242e-01 -4.44359630e-01 -4.55904096e-01 7.79051840e-01
4.46684808e-01 -5.86645901e-01 -9.93765965e-02 8.79107565e-02
8.09279203e-01 3.91473114e-01 -3.90337966e-02 2.27054954e-01
4.50468928e-01 -4.17576075e-01 1.93846524e-01 2.74159998e-01
1.68578066e-02 3.40144873e-01 4.93543446e-01 -2.65779406e-01
-9.67126966e-01 -1.18899298e+00 4.71182287e-01 1.65335703e+00
-1.79117788e-02 -6.70527294e-03 -9.22822475e-01 -5.73426723e-01
-3.50971907e-01 9.61120069e-01 -8.77396643e-01 -5.15608042e-02
-7.89264560e-01 -1.36420429e-02 6.38107598e-01 3.04206908e-01
1.56963989e-02 -1.47228086e+00 -7.50391245e-01 -4.22489196e-02
-3.27971190e-01 -1.22969019e+00 -8.38844240e-01 -2.88205054e-02
-4.46548998e-01 -6.17932081e-01 -6.99509501e-01 -8.73450220e-01
4.34521556e-01 -5.48692524e-01 7.99034059e-01 -2.88038045e-01
5.00446483e-02 6.75668299e-01 -3.49886507e-01 -3.90181005e-01
-1.21862090e+00 6.29822910e-02 2.70342141e-01 2.99513102e-01
1.51503280e-01 -5.43785214e-01 -1.39525130e-01 3.63921791e-01
-9.05002952e-01 1.31186500e-01 1.34967834e-01 7.77050555e-01
-1.73155256e-02 -7.46913314e-01 2.63991088e-01 -5.70952833e-01
5.20348728e-01 -1.59508139e-01 -3.77854377e-01 1.03440762e-01
3.01231414e-01 3.39131832e-01 4.45691168e-01 -1.10605216e+00
-1.23251891e+00 5.67255080e-01 -1.32141814e-01 -5.96272588e-01
-7.40384340e-01 -4.31996472e-02 -6.28854156e-01 3.36327404e-01
6.98577106e-01 1.18058220e-01 4.43111718e-01 -2.56172240e-01
8.21315765e-01 8.51734221e-01 5.65728903e-01 -7.91283607e-01
8.03529084e-01 4.04552132e-01 -3.33232731e-01 -1.32641637e+00
5.17950282e-02 -3.47643066e-03 -6.76157832e-01 -4.40415502e-01
1.15314054e+00 -6.09509826e-01 -1.13792408e+00 4.78912085e-01
-1.42331266e+00 -8.07200074e-01 -2.22005501e-01 4.01637644e-01
-1.10372472e+00 2.02011429e-02 -6.31004274e-01 -8.03727746e-01
-9.50539708e-02 -1.65265203e+00 1.38215113e+00 8.21697563e-02
-8.58984232e-01 -1.00365651e+00 2.40488395e-01 1.44924462e-01
1.04288429e-01 3.70471328e-01 9.07555640e-01 -7.88169920e-01
-2.10556552e-01 2.18641877e-01 4.14638817e-01 -4.51763272e-02
4.22788322e-01 2.22288460e-01 -1.27242100e+00 -1.47085026e-01
2.60101119e-03 -4.71349150e-01 2.25953013e-01 3.41697246e-01
3.76014650e-01 -5.44997334e-01 -7.73715749e-02 6.51372313e-01
7.36070275e-01 5.50359309e-01 2.24811167e-01 -1.60364702e-01
7.77835906e-01 1.05839372e+00 1.55672148e-01 2.79963642e-01
2.08977699e-01 1.07776225e+00 3.14045101e-01 2.54735589e-01
-9.22990292e-02 -3.78837496e-01 9.12091315e-01 5.84961236e-01
-1.43456027e-01 -2.49915853e-01 -5.85460782e-01 2.74910420e-01
-1.64768708e+00 -1.20887983e+00 5.39646387e-01 1.87536001e+00
9.74295557e-01 -2.12595060e-01 6.63287580e-01 -2.57701188e-01
5.93687654e-01 7.61077255e-02 -6.95317805e-01 -7.41145372e-01
1.54067218e-01 -8.78739655e-02 -1.55921921e-01 9.20633376e-01
-8.71768594e-01 1.23045063e+00 5.17042303e+00 2.23836228e-01
-1.53836036e+00 -1.61276326e-01 3.16576004e-01 -3.79845291e-01
-4.19399291e-01 -4.96482939e-01 -6.83497429e-01 3.58134538e-01
1.03730297e+00 -2.08332445e-02 8.46183538e-01 6.18994415e-01
3.25792611e-01 6.03513300e-01 -1.94106257e+00 8.72586429e-01
3.26997548e-01 -1.02854133e+00 3.34715366e-01 2.88487803e-02
4.28555459e-01 -2.23041713e-01 2.80688733e-01 5.77058971e-01
3.89531106e-01 -1.20281911e+00 1.07014489e+00 3.71480286e-01
1.01589274e+00 -2.92871565e-01 2.78637230e-01 4.40430313e-01
-1.02838409e+00 7.08028451e-02 3.80453378e-01 1.35627165e-01
4.94261891e-01 -7.79488444e-01 -1.40127206e+00 -8.24548583e-03
2.03240156e-01 3.34273070e-01 1.31731197e-01 1.51943583e-02
-2.15011328e-01 3.48950624e-01 -3.17382812e-01 -1.62623495e-01
1.23722561e-01 2.27389392e-02 1.05776024e+00 1.19303346e+00
1.86787561e-01 1.57587975e-01 3.53926092e-01 1.00688195e+00
4.06953692e-02 -3.97630297e-02 -7.04684675e-01 -2.26999134e-01
4.68781590e-01 7.68608868e-01 -2.76529670e-01 -5.12419999e-01
-1.32958949e-01 1.27907968e+00 -3.07528824e-02 7.01999366e-01
-6.24991715e-01 -3.60798016e-02 9.95471597e-01 -2.27577612e-01
2.84065545e-01 -1.38854891e-01 -2.01104045e-01 -1.19079387e+00
-4.96456981e-01 -1.28966033e+00 1.06978431e-01 -6.04924381e-01
-1.14253771e+00 1.06236839e+00 1.36505991e-01 -1.17646503e+00
-1.22833538e+00 -5.99162161e-01 -6.81824028e-01 9.75996375e-01
-6.90477252e-01 -1.45765293e+00 -2.13663861e-01 7.14022458e-01
1.06917822e+00 -6.16802037e-01 1.15077055e+00 -3.23763013e-01
-1.61617771e-01 6.08555377e-01 -6.29283190e-01 2.54536539e-01
6.61943793e-01 -1.11815715e+00 2.06207231e-01 1.57312974e-01
1.11425027e-01 4.79102939e-01 1.15491450e+00 -5.30320287e-01
-1.33598387e+00 -7.79700816e-01 5.74511826e-01 -5.13728082e-01
5.72793841e-01 -3.82014662e-01 -7.30722666e-01 9.77064490e-01
4.43614125e-01 -6.03357434e-01 7.39211202e-01 -2.10358337e-01
-4.35679227e-01 3.00472587e-01 -1.10201907e+00 9.25140738e-01
8.07495117e-01 -6.01348042e-01 -9.24782097e-01 -1.83701426e-01
8.75089228e-01 -5.29825032e-01 -5.96077979e-01 2.14059055e-01
8.46194088e-01 -4.91578400e-01 8.12610447e-01 -8.32851946e-01
4.04723972e-01 -1.53331488e-01 -4.17755574e-01 -1.51426399e+00
2.71314591e-01 -7.23493874e-01 9.78219658e-02 1.16591406e+00
5.21276534e-01 -3.44076365e-01 6.98969305e-01 7.56296158e-01
2.26024352e-02 -9.62860510e-02 -7.87692904e-01 -5.35139501e-01
1.37301609e-01 -2.89590776e-01 8.88362467e-01 7.83166170e-01
1.73996881e-01 7.56357968e-01 -4.93229151e-01 3.33920270e-01
4.35382366e-01 1.28006428e-01 1.30795312e+00 -1.03609347e+00
-7.29834676e-01 -5.75910747e-01 -2.10135192e-01 -9.88449574e-01
6.75772667e-01 -7.57296383e-01 3.72599721e-01 -9.20013070e-01
1.20958025e-02 6.46464825e-02 3.37292373e-01 3.70554358e-01
8.07222128e-02 -6.02379069e-02 5.02676785e-01 1.12106055e-01
1.01048201e-02 8.71213317e-01 1.37455475e+00 -4.28227901e-01
-7.79916048e-01 9.79172736e-02 -4.82486226e-02 9.97110844e-01
5.41674674e-01 -1.73124760e-01 -7.03037977e-01 -4.28386986e-01
-3.03186744e-01 5.91607809e-01 2.39173174e-01 -4.25414056e-01
-2.04967067e-01 -4.79424745e-01 1.72666952e-01 -6.23343028e-02
1.02919030e+00 -9.12748814e-01 1.94772050e-01 4.66607720e-01
-7.48648584e-01 -2.35892758e-01 2.33507186e-01 3.64515394e-01
9.54893231e-02 -6.43044803e-03 6.54020190e-01 -7.05245063e-02
-5.94979823e-01 1.36112332e-01 -7.10460365e-01 -2.22493783e-01
8.65337670e-01 -4.27478254e-02 -2.46241651e-02 -8.78598213e-01
-1.23162174e+00 -2.89601665e-02 5.85278988e-01 6.87407494e-01
5.50922394e-01 -1.33783901e+00 -7.32791722e-01 4.77295280e-01
1.22049086e-01 -2.05946133e-01 1.79214124e-02 1.64065823e-01
-2.25862175e-01 1.09695479e-01 -4.44105744e-01 -9.06951368e-01
-1.60267735e+00 3.42921019e-01 3.59929293e-01 1.93718731e-01
-4.97911304e-01 8.75140846e-01 5.52726388e-01 -6.91002369e-01
5.24643660e-01 -3.82845014e-01 -3.33931297e-01 7.51433223e-02
7.45247304e-01 5.55651747e-02 -6.24873877e-01 -1.14367747e+00
-1.60962418e-01 5.14451146e-01 1.14195853e-01 -1.05447602e+00
1.19429064e+00 -2.32404564e-02 2.32268840e-01 1.07442117e+00
1.33264446e+00 -1.14956081e-01 -1.48585927e+00 -9.56076905e-02
-3.63053322e-01 -1.20694041e-01 -5.45202851e-01 -7.35301971e-01
-7.24758089e-01 8.41848433e-01 6.24301255e-01 -3.27433459e-02
7.90666938e-01 2.73458004e-01 6.83388412e-01 2.96901762e-01
3.45481813e-01 -9.05630052e-01 5.13031006e-01 6.82558477e-01
1.22219849e+00 -1.16066253e+00 -7.87697494e-01 -1.39235228e-01
-1.16610289e+00 1.06874657e+00 5.89335501e-01 3.32361199e-02
4.58137363e-01 3.70199889e-01 6.40554011e-01 1.17197670e-02
-8.57202113e-01 4.76958491e-02 4.40269709e-01 8.25380325e-01
3.28397840e-01 4.00147289e-01 2.23658234e-01 5.12664676e-01
-6.90719724e-01 -2.32569367e-01 3.79269332e-01 5.91792405e-01
-2.15541497e-01 -1.22381020e+00 -4.80272740e-01 -1.56797051e-01
6.98423535e-02 4.04782981e-01 -6.96378171e-01 4.16288167e-01
-8.28872398e-02 1.03288138e+00 1.84028044e-01 -7.02935576e-01
4.12125587e-01 5.04671276e-01 5.34279644e-01 -8.24801683e-01
-6.23567522e-01 3.60789061e-01 1.35610700e-01 -4.71678346e-01
-2.39347920e-01 -9.05326724e-01 -1.39373612e+00 -2.33737379e-01
3.09719414e-01 -1.21465050e-01 7.96555936e-01 1.10847187e+00
-2.72981916e-02 6.03492796e-01 7.50994682e-01 -1.64688635e+00
-6.64649665e-01 -1.24788773e+00 -2.28771925e-01 8.72563899e-01
6.19838059e-01 -5.31565905e-01 -4.31252658e-01 6.83048308e-01] | [5.615528106689453, -0.1146547868847847] |
86f7720f-0569-47c7-b067-221b97556bd4 | neural-networks-and-spelling-features-for | null | null | https://aclanthology.org/W17-5025 | https://aclanthology.org/W17-5025.pdf | Neural Networks and Spelling Features for Native Language Identification | We present the RUG-SU team{'}s submission at the Native Language Identification Shared Task 2017. We combine several approaches into an ensemble, based on spelling error features, a simple neural network using word representations, a deep residual network using word and character features, and a system based on a recurrent neural network. Our best system is an ensemble of neural networks, reaching an F1 score of 0.8323. Although our system is not the highest ranking one, we do outperform the baseline by far. | ['Barbara Plank', 'Robert {\\"O}stling', 'Gintar{\\.e} Grigonyt{\\.e}', 'Johannes Bjerva'] | 2017-09-01 | null | null | null | ws-2017-9 | ['native-language-identification'] | ['natural-language-processing'] | [ 2.04148605e-01 -1.39204264e-01 -3.24059755e-01 -2.97826733e-02
-1.02168179e+00 -6.64835632e-01 7.94624627e-01 -7.08886981e-02
-9.89551187e-01 6.25242829e-01 5.47495484e-01 -8.27821970e-01
1.07224248e-01 -2.25441054e-01 -3.81393284e-01 -2.60761201e-01
3.05252641e-01 5.10943532e-01 1.18895277e-01 -4.38650966e-01
4.48775738e-01 4.23916072e-01 -1.13942289e+00 2.58184582e-01
8.76304626e-01 8.21442783e-01 -1.35521814e-01 7.97513723e-01
-2.36309335e-01 8.75984550e-01 -6.28988385e-01 -4.66056436e-01
3.13362747e-01 -1.64349034e-01 -8.94141078e-01 -8.50854218e-01
3.87055099e-01 -2.40032777e-01 -5.37989676e-01 1.00753736e+00
7.44691372e-01 2.67586052e-01 7.41340637e-01 -6.80940568e-01
-1.08890378e+00 9.82354164e-01 -2.96442471e-02 4.68077809e-01
5.31606019e-01 2.01609954e-01 1.07889509e+00 -1.15162790e+00
4.30734336e-01 1.14183247e+00 1.02240205e+00 6.87314153e-01
-1.09218371e+00 -8.18871677e-01 1.75340638e-01 1.13230020e-01
-1.64649153e+00 -6.54250979e-01 1.75012067e-01 -4.10733700e-01
1.63575387e+00 5.33991084e-02 5.09497449e-02 1.53519154e+00
-2.11245269e-01 8.85867953e-01 1.24327409e+00 -7.62296617e-01
-6.97158976e-03 7.49645978e-02 6.78653359e-01 6.66982055e-01
3.24412942e-01 3.16915542e-01 -4.85492855e-01 -3.22729498e-01
2.38085866e-01 2.46900171e-02 -2.63112873e-01 3.69831145e-01
-1.21025658e+00 9.33583260e-01 2.13073701e-01 6.15888894e-01
-1.81263670e-01 2.81560719e-02 4.80411351e-01 2.55028099e-01
5.44963956e-01 8.37450862e-01 -7.45902717e-01 -1.55177787e-01
-1.27876306e+00 1.94301933e-01 9.43249583e-01 5.63748538e-01
3.86454791e-01 1.49162024e-01 -5.83224833e-01 1.14179504e+00
2.21135572e-01 4.22243148e-01 1.20310330e+00 -3.73428494e-01
4.16138798e-01 3.58827025e-01 -5.68287075e-02 -4.72550690e-01
-5.09630084e-01 -4.51788455e-01 -6.86591089e-01 -5.34189604e-02
3.67236227e-01 -4.73863572e-01 -1.01311505e+00 1.71384561e+00
-7.50651538e-01 2.10110456e-01 2.67135445e-02 5.12280583e-01
1.08177257e+00 6.28982544e-01 1.72919512e-01 8.78456160e-02
1.08455217e+00 -1.43585300e+00 -5.75295806e-01 -2.27252990e-01
9.99979615e-01 -6.82193339e-01 5.27540922e-01 4.84591365e-01
-9.93189692e-01 -6.51289582e-01 -1.08647013e+00 -1.38204349e-02
-1.05228817e+00 4.56846178e-01 4.40976709e-01 9.93852139e-01
-1.58195758e+00 7.94731557e-01 -1.85124382e-01 -6.03687048e-01
-8.35024267e-02 5.13011277e-01 -4.43719566e-01 1.73245221e-01
-1.50798297e+00 1.15422642e+00 5.21085322e-01 -6.30412847e-02
-4.51482087e-01 -6.70029521e-01 -7.52661347e-01 -1.26900315e-01
2.53558122e-02 -5.56055367e-01 1.32399762e+00 -8.14658999e-01
-1.78371704e+00 1.04760134e+00 -1.84513375e-01 -8.54818642e-01
4.62044120e-01 -4.54250664e-01 -7.23364949e-01 -3.22448879e-01
-1.97336767e-02 4.73774016e-01 3.19349468e-01 -6.81817114e-01
-3.97339165e-01 -1.12748936e-01 -3.28826398e-01 -9.54388604e-02
-4.39808041e-01 6.19305909e-01 -1.18294030e-01 -8.32925975e-01
-4.49105233e-01 -1.04403222e+00 -3.65130335e-01 -7.16026366e-01
-6.06685162e-01 -8.10749471e-01 1.11601777e-01 -1.05804765e+00
1.57455015e+00 -1.94804716e+00 -3.33754197e-02 1.07718870e-01
1.02657713e-01 9.53608513e-01 -4.80208546e-01 4.52385962e-01
-3.81760538e-01 7.04012990e-01 1.01745546e-01 -5.82847655e-01
2.14061767e-01 -3.74939978e-01 -3.25158328e-01 3.21324557e-01
1.49355352e-01 1.15019572e+00 -8.67375731e-01 6.14104494e-02
4.45524184e-03 5.20600617e-01 -1.58262596e-01 2.56351948e-01
5.42841963e-02 -2.65388101e-01 -1.41757857e-02 3.32017064e-01
3.95443022e-01 -1.07308336e-01 -1.10074744e-01 1.85241178e-01
-4.22512472e-01 9.32687879e-01 -9.93340790e-01 1.54026508e+00
-4.47339833e-01 7.59775102e-01 -3.32578808e-01 -6.35116696e-01
1.16108024e+00 3.73099118e-01 7.88356550e-03 -3.92459661e-01
5.99414259e-02 6.49633944e-01 9.99479741e-02 -1.10612139e-02
6.46358252e-01 3.03431243e-01 -1.45796269e-01 4.79828686e-01
3.05704743e-01 3.06422204e-01 -3.81764360e-02 2.37720922e-01
1.27912533e+00 7.87210558e-03 2.33341157e-01 -5.20101964e-01
6.95147216e-01 -4.84905601e-01 3.76543581e-01 1.14586687e+00
-4.03731763e-01 9.98027742e-01 2.25390181e-01 -6.19057178e-01
-9.14053142e-01 -7.61936188e-01 6.24733865e-02 1.32733572e+00
-4.83293205e-01 -7.72720993e-01 -6.84506714e-01 -6.74171567e-01
-1.09972827e-01 7.40919948e-01 -6.12088501e-01 -2.41723433e-01
-5.85311949e-01 -3.93361568e-01 1.10049176e+00 6.28964901e-01
2.97547728e-01 -1.24742734e+00 1.09105073e-01 1.62356317e-01
1.71836913e-02 -1.09913266e+00 -8.18645120e-01 4.48440850e-01
-1.06941022e-01 -5.87639570e-01 -1.10752988e+00 -8.81898403e-01
5.30137867e-02 -9.10824165e-02 1.02680981e+00 1.38024792e-01
-1.59362182e-01 2.83805430e-02 -4.71967846e-01 -4.18797553e-01
-2.92750835e-01 7.22002983e-01 4.30279583e-01 -1.50075763e-01
6.81750476e-01 -4.52413410e-02 -3.72257799e-01 -1.80871859e-01
-3.55565190e-01 -2.79642969e-01 6.66502297e-01 9.24124956e-01
-2.03708317e-02 -6.26878560e-01 4.57494617e-01 -8.34732413e-01
6.76541746e-01 -4.05831605e-01 -2.25778461e-01 3.37116122e-01
-7.38023877e-01 1.18798599e-01 7.06418216e-01 -4.00310278e-01
-7.33290672e-01 1.01781413e-01 -6.26885056e-01 -1.76159844e-01
-2.56337732e-01 9.52229053e-02 1.74751997e-01 -2.02559784e-01
6.74921989e-01 4.61234659e-01 -2.95447916e-01 -9.16496277e-01
3.21940958e-01 9.49044883e-01 5.79372168e-01 -2.02260360e-01
6.43528759e-01 -4.39193606e-01 -5.42857051e-01 -7.98078418e-01
-8.43496263e-01 -6.28437102e-01 -7.31333017e-01 1.97069138e-01
9.48435307e-01 -9.26759422e-01 -5.60837269e-01 6.25640869e-01
-1.50085020e+00 -1.83001205e-01 -3.39043327e-02 5.29476762e-01
-6.97151795e-02 3.58246863e-01 -9.32969630e-01 -8.85694206e-01
-6.77608609e-01 -1.10778677e+00 9.28793490e-01 9.61609781e-02
-5.38631499e-01 -8.70299578e-01 3.32717687e-01 1.83126718e-01
7.78779685e-01 -1.97351262e-01 5.32527685e-01 -1.50959229e+00
5.29911332e-02 -3.56030881e-01 -3.53878528e-01 4.03808415e-01
-1.29674494e-01 1.37396622e-02 -9.00019109e-01 -3.65293175e-01
-6.70570970e-01 -4.18651700e-01 1.57240808e+00 2.69998670e-01
1.26432061e+00 -3.14820498e-01 -3.63133401e-01 7.33614504e-01
1.34965265e+00 -5.88220265e-03 5.25287032e-01 4.51172113e-01
7.01867580e-01 2.30267689e-01 -3.74294788e-01 4.76105288e-02
5.36426306e-01 7.57579744e-01 1.53825553e-02 8.04394856e-03
-3.55461985e-01 -5.15991688e-01 6.70102537e-01 1.02111864e+00
-2.88741320e-01 -3.27149361e-01 -1.34351540e+00 5.37995577e-01
-1.68470132e+00 -9.50598359e-01 -1.71141028e-01 2.17904711e+00
4.96390015e-01 2.19017550e-01 4.18388575e-01 -5.00524640e-02
8.17139268e-01 2.20312327e-01 -2.38081753e-01 -1.08516490e+00
-4.59220380e-01 4.97769922e-01 7.31762052e-01 6.10439718e-01
-1.33036208e+00 1.43295622e+00 7.91269684e+00 9.46944892e-01
-9.72401261e-01 8.03055763e-02 5.51291108e-01 2.77341515e-01
-2.42459491e-01 -2.41977632e-01 -1.31656110e+00 5.58476806e-01
1.61857975e+00 -2.50891626e-01 4.42259759e-01 6.99806094e-01
-4.32216138e-01 3.95284414e-01 -6.21357441e-01 9.60983872e-01
6.41362309e-01 -1.08478260e+00 2.63535917e-01 1.79503441e-01
6.28677309e-01 7.37713695e-01 8.49138126e-02 7.18419909e-01
7.57328033e-01 -1.62603295e+00 3.95764619e-01 7.08566368e-01
7.96419680e-01 -6.96710646e-01 8.48368168e-01 2.74390638e-01
-9.36257243e-01 5.37307560e-02 -1.03086390e-01 -1.85964152e-01
-2.59516507e-01 3.78184259e-01 -5.79161584e-01 4.41232562e-01
4.37066346e-01 7.15719402e-01 -8.90309691e-01 1.04856360e+00
-1.16784237e-01 9.50948298e-01 -1.27244979e-01 -5.02608895e-01
3.89648318e-01 2.41791233e-01 5.66158950e-01 1.94608903e+00
2.11252540e-01 -8.20059478e-02 2.77589083e-01 4.90588218e-01
-3.14637512e-01 3.91325682e-01 -9.59800959e-01 -1.61841735e-01
4.47189361e-01 1.30012846e+00 -2.03525126e-01 -3.76440942e-01
-3.76103312e-01 1.33740437e+00 6.20500803e-01 3.36155802e-01
-3.74120235e-01 -7.81235397e-01 6.53518796e-01 -5.03876150e-01
2.43969575e-01 -2.36641511e-01 -2.08795995e-01 -1.30348694e+00
-2.33385339e-01 -7.84209251e-01 3.85165662e-01 -4.33278918e-01
-1.52497578e+00 1.14443624e+00 -4.96592641e-01 -7.88849175e-01
-6.73897207e-01 -1.24168050e+00 -4.71827656e-01 1.24029839e+00
-1.46499550e+00 -9.44003820e-01 3.84086579e-01 2.24993691e-01
5.03497481e-01 -9.11798060e-01 1.25484204e+00 2.23150745e-01
-9.06486511e-01 1.17935550e+00 2.50651509e-01 5.69972932e-01
8.24032187e-01 -1.31516480e+00 8.95310521e-01 7.87410080e-01
2.67653793e-01 8.68170023e-01 2.22631395e-01 -3.53873104e-01
-1.14990914e+00 -9.79405105e-01 1.61376357e+00 -6.55278504e-01
8.98840547e-01 -4.98593867e-01 -7.92702734e-01 7.47158527e-01
4.47686523e-01 9.61171612e-02 8.04255128e-01 3.24342698e-01
-8.18109214e-01 2.84190983e-01 -7.60569394e-01 4.53109890e-01
1.23861480e+00 -7.54410446e-01 -8.52896214e-01 3.99024308e-01
8.41569126e-01 -2.05778684e-02 -5.48439801e-01 4.62392330e-01
7.05642462e-01 -7.29126811e-01 1.00861609e+00 -8.68153214e-01
2.60742515e-01 1.30589202e-01 -1.38201982e-01 -1.63297749e+00
-7.91910052e-01 -7.39362299e-01 1.08056493e-01 1.47191060e+00
9.33570147e-01 -9.68321025e-01 5.32467842e-01 4.10700887e-01
1.71574447e-02 -6.66271150e-01 -8.65194380e-01 -1.14844167e+00
7.51857638e-01 -3.73163611e-01 5.33747673e-01 9.01425064e-01
4.35245000e-02 5.96566975e-01 -5.54174960e-01 -4.16380107e-01
1.62882805e-01 -4.24063265e-01 3.67801845e-01 -1.26260090e+00
-2.66216487e-01 -1.12147057e+00 -3.30364466e-01 -1.02471292e+00
8.66738379e-01 -1.32780707e+00 -7.47628361e-02 -1.27341068e+00
3.65683466e-01 -1.14409827e-01 -7.85667002e-01 5.59002280e-01
-3.80266160e-01 5.27607739e-01 3.19737196e-01 1.37977395e-02
-8.03482175e-01 1.86787486e-01 3.38318020e-01 -1.80522561e-01
-6.20807568e-03 -8.91396031e-02 -8.51279557e-01 6.90947413e-01
1.02830219e+00 -1.08629540e-01 4.09920245e-01 -6.03651822e-01
1.52902067e-01 -5.31500340e-01 -2.41035782e-02 -9.53383803e-01
3.63237619e-01 3.42025518e-01 4.15134817e-01 -4.96127367e-01
1.89304948e-01 -9.94636267e-02 -2.41460368e-01 6.78392589e-01
-6.23791814e-01 4.44061935e-01 2.15001911e-01 3.93853858e-02
3.34073491e-02 -3.68347138e-01 4.82009768e-01 -2.35441089e-01
-7.39021540e-01 3.66853684e-01 -7.70537317e-01 -5.57063408e-02
4.16622519e-01 -4.11358699e-02 -3.88003051e-01 -1.87250316e-01
-3.99458408e-01 6.41765585e-03 6.98015764e-02 9.10103142e-01
2.74201453e-01 -1.40362704e+00 -1.08588052e+00 1.64856315e-01
1.65160194e-01 -9.63787854e-01 -2.80755520e-01 3.27972621e-01
-3.13982129e-01 9.64272141e-01 8.73941928e-02 -9.76623893e-02
-9.52811182e-01 3.82891536e-01 4.00685519e-01 -6.95826828e-01
-3.99670005e-01 1.08891106e+00 -3.79473120e-01 -1.03953147e+00
2.68799543e-01 1.36383280e-01 -6.16593242e-01 -5.29917739e-02
1.00698996e+00 5.48271239e-01 2.70986259e-01 -9.57445085e-01
-7.81863809e-01 6.16877139e-01 -1.84859619e-01 -3.50000143e-01
1.04609072e+00 3.24822694e-01 -1.26411319e-01 6.10489607e-01
1.54981089e+00 3.51612642e-02 -2.15415195e-01 -4.68149573e-01
2.69139379e-01 -5.80385327e-02 2.48306364e-01 -1.08901572e+00
-5.47206759e-01 8.10321033e-01 5.64534307e-01 8.20806250e-02
6.19054914e-01 -3.73585790e-01 9.47581172e-01 5.50400078e-01
1.04374655e-01 -1.01895845e+00 -3.19011331e-01 1.32976925e+00
7.51655996e-01 -1.23607099e+00 -3.50258917e-01 1.38507798e-01
-3.75205010e-01 1.25727272e+00 4.33376521e-01 -3.69038343e-01
7.49267995e-01 2.66739964e-01 8.81253183e-02 5.10672450e-01
-8.40530992e-01 -5.40075779e-01 6.54622614e-01 5.19344687e-01
8.28096807e-01 2.02207223e-01 -6.17231846e-01 8.16519201e-01
-5.66417694e-01 -1.07254751e-01 1.52490184e-01 1.25997156e-01
-3.21819097e-01 -1.17563093e+00 -2.12542206e-01 6.06837451e-01
-8.02450120e-01 -5.95275462e-01 -7.64082491e-01 4.02800858e-01
1.12459652e-01 9.46205974e-01 -7.28258491e-02 -7.17242241e-01
2.56029516e-01 4.55536604e-01 2.24998981e-01 -7.84056902e-01
-1.12296975e+00 -5.59678555e-01 3.04426402e-01 -4.52780098e-01
9.07563642e-02 -8.04052770e-01 -6.49210751e-01 -3.07978064e-01
-1.43516257e-01 1.15795337e-01 8.06712329e-01 1.04494631e+00
2.11920887e-01 2.22964600e-01 6.85315549e-01 -8.09091687e-01
-6.72133684e-01 -1.16856313e+00 -2.68897593e-01 2.01234400e-01
5.86341083e-01 -9.48717222e-02 -5.02270758e-01 -6.36328042e-01] | [10.438088417053223, 10.529435157775879] |
321fed62-e942-4b4d-9486-cfded59862ac | clique-graph-matching-by-preserving-global | null | null | http://openaccess.thecvf.com/content_cvpr_2015/html/Nie_Clique-Graph_Matching_by_2015_CVPR_paper.html | http://openaccess.thecvf.com/content_cvpr_2015/papers/Nie_Clique-Graph_Matching_by_2015_CVPR_paper.pdf | Clique-Graph Matching by Preserving Global & Local Structure | This paper originally proposes the clique-graph and further presents a clique-graph matching method by preserving global and local structures. Especially, we formulate the objective function of clique-graph matching with respective to two latent variables, the clique information in the original graph and the pairwise clique correspondence constrained by the one-to-one matching. Since the objective function is not jointly convex to both latent variables, we decompose it into two consecutive steps for optimization: 1) clique-to-clique similarity measure by preserving local unary and pairwise correspondences; 2) graph-to-graph similarity measure by preserving global clique-to-clique correspondence. Extensive experiments on the synthetic data and real images show that the proposed method can outperform representative methods especially when both noise and outliers exist. | ['Yu-Ting Su', 'An-An Liu', 'Wei-Zhi Nie', 'Zan Gao'] | 2015-06-01 | null | null | null | cvpr-2015-6 | ['graph-similarity'] | ['graphs'] | [ 2.77161986e-01 4.98395443e-01 -2.37313539e-01 -3.04282248e-01
-8.57440412e-01 -3.44005495e-01 3.47909629e-01 2.74240345e-01
1.55321077e-01 5.38827002e-01 -4.99903709e-02 2.79655367e-01
-5.14040649e-01 -8.89597893e-01 -5.56542993e-01 -5.96231222e-01
-3.25312763e-01 5.72631061e-01 1.91400036e-01 1.88754722e-01
2.15382397e-01 3.05151135e-01 -9.45242882e-01 -6.52234852e-02
9.46046114e-01 3.83337587e-01 3.39913517e-02 4.00065660e-01
1.32691920e-01 4.73147124e-01 -1.49960130e-01 -2.59680867e-01
7.47622013e-01 -6.63070261e-01 -6.40874803e-01 7.34188616e-01
6.23540699e-01 3.22014540e-01 -3.19861501e-01 1.50885129e+00
1.59996152e-01 1.96860373e-01 3.47863227e-01 -1.65796196e+00
-4.75356251e-01 2.42441759e-01 -1.00137973e+00 -1.92376852e-01
6.66094542e-01 -1.80523649e-01 1.33829987e+00 -7.06085980e-01
8.43147099e-01 1.47044039e+00 5.38202941e-01 -2.17125043e-01
-1.54361212e+00 -4.52103555e-01 1.47567511e-01 1.54713869e-01
-1.56261766e+00 -5.14233895e-02 1.06450224e+00 -5.03023863e-01
6.57985985e-01 3.80952328e-01 6.48206830e-01 3.18577349e-01
7.39550292e-02 3.72260213e-01 1.27013481e+00 -4.52616185e-01
6.69959113e-02 -2.42081836e-01 1.86471507e-01 8.52264881e-01
4.45336044e-01 1.37313277e-01 -3.74210268e-01 -5.85532904e-01
9.27839100e-01 -6.47894144e-02 -1.78501308e-01 -7.46821165e-01
-1.11360395e+00 8.90222669e-01 4.83164370e-01 1.70596883e-01
-7.06047863e-02 7.71334991e-02 -4.68685403e-02 4.55318421e-01
3.62468988e-01 3.39574605e-01 -2.12807581e-02 5.15224099e-01
-9.58839595e-01 4.02541868e-02 8.43025684e-01 1.49694729e+00
1.37557590e+00 -2.39328653e-01 1.55254066e-01 6.21862411e-01
6.98410630e-01 3.93266708e-01 -4.48988676e-02 -8.25802326e-01
7.25841284e-01 9.55153108e-01 -3.20919633e-01 -1.86513102e+00
-3.71863067e-01 -4.26203519e-01 -1.01588976e+00 -1.01370424e-01
2.09705874e-01 3.54246974e-01 -7.28836238e-01 1.68008959e+00
2.65698940e-01 6.86583757e-01 -3.44797075e-01 8.25661540e-01
8.25888813e-01 3.22588414e-01 -2.19404534e-01 -7.36657202e-01
8.99133444e-01 -1.18284881e+00 -9.37838078e-01 -3.74462873e-01
3.19517374e-01 -9.98675883e-01 5.08603275e-01 -3.37163776e-01
-1.22751915e+00 -5.56713045e-01 -1.11632776e+00 1.97025016e-01
-1.14216832e-02 -1.97560385e-01 3.74987245e-01 4.78559792e-01
-9.34753597e-01 6.07185245e-01 -5.22577703e-01 -3.92048448e-01
-1.91555656e-02 6.70311034e-01 -7.08939791e-01 -9.93678123e-02
-8.33147168e-01 4.93528783e-01 5.56493819e-01 5.51075898e-02
-3.46240610e-01 -2.79843152e-01 -9.18455839e-01 2.06910655e-01
6.63555324e-01 -8.19231510e-01 3.95955920e-01 -8.24612439e-01
-1.10369790e+00 1.15827942e+00 -3.41239542e-01 1.51389971e-01
3.83538336e-01 3.59096378e-01 -3.64925921e-01 1.58542529e-01
3.37040603e-01 7.75272995e-02 8.03253949e-01 -1.47855783e+00
-1.22453175e-01 -5.17264783e-01 -8.68135840e-02 3.52384567e-01
9.62961391e-02 -9.76507738e-02 -8.91903758e-01 -6.25269771e-01
1.02199519e+00 -9.69208539e-01 -4.22537297e-01 -1.47499740e-01
-5.49574554e-01 -1.95955615e-02 8.91042471e-01 -6.89659059e-01
1.32487750e+00 -1.95319605e+00 3.65713477e-01 9.21804070e-01
7.06212223e-01 -3.64843041e-01 -5.13407469e-01 8.22142780e-01
-4.57267702e-01 -2.67198123e-03 -3.60587388e-01 -3.73216182e-01
-7.39033148e-02 2.61285871e-01 3.18116993e-01 1.04549217e+00
-2.70450450e-02 9.04037952e-01 -1.06176484e+00 -8.84566784e-01
2.28254408e-01 -2.42244840e-01 -4.65896517e-01 3.46616685e-01
2.00895041e-01 4.34885859e-01 -3.86599094e-01 7.77918994e-01
1.22251987e+00 -5.14695704e-01 8.38358462e-01 -2.89803505e-01
2.11353928e-01 -3.77730429e-01 -1.75345552e+00 1.46905696e+00
3.08314145e-01 2.70396352e-01 5.21813512e-01 -1.09619737e+00
1.21839118e+00 6.81189150e-02 9.67630923e-01 -4.21294719e-01
-3.68155763e-02 2.24084839e-01 -1.27261341e-01 -2.89928883e-01
1.17512189e-01 -1.40048862e-02 -1.04445077e-01 1.92253217e-01
-4.52959985e-02 1.45065207e-02 3.84097517e-01 6.08810246e-01
1.06047714e+00 -1.43421218e-01 7.30285943e-01 -6.32493854e-01
6.84861243e-01 -6.41064718e-02 8.85389626e-01 5.95842242e-01
-1.45947903e-01 7.38982439e-01 6.64041877e-01 -1.11020207e-01
-9.63267684e-01 -1.22512293e+00 3.93191651e-02 4.61185426e-01
7.95921624e-01 -6.10643446e-01 -6.00124955e-01 -5.20999968e-01
2.05482423e-01 -6.17546141e-02 -4.70675290e-01 3.09672449e-02
-5.91623664e-01 -4.53666598e-01 -7.61306137e-02 1.55965120e-01
3.03726465e-01 -5.20786226e-01 4.51219052e-01 1.76884443e-01
-2.33961061e-01 -1.31104183e+00 -1.04866219e+00 -3.55451673e-01
-1.09077597e+00 -1.68209064e+00 -2.61915863e-01 -1.18727660e+00
1.29764330e+00 5.19380391e-01 1.12113941e+00 5.62156856e-01
-2.65113980e-01 2.32211217e-01 -2.80325294e-01 3.92635256e-01
-2.49327719e-01 -2.48587266e-01 -1.20256864e-01 1.79439202e-01
5.51698841e-02 -8.02448869e-01 -1.53244570e-01 5.86670876e-01
-6.34257674e-01 1.50359511e-01 4.60288495e-01 9.52779055e-01
9.37781036e-01 4.47124764e-02 5.76273538e-02 -1.12768829e+00
7.37077773e-01 -2.78381318e-01 -9.74923134e-01 4.24332649e-01
-7.31241226e-01 -1.86310545e-01 2.29313865e-01 -1.19042292e-01
-6.01581872e-01 4.98309076e-01 5.61641335e-01 -7.07997620e-01
2.29749069e-01 7.13646591e-01 -5.76177776e-01 -4.72855419e-01
7.22028166e-02 -5.71020991e-02 1.98814888e-02 -4.29410994e-01
5.54222584e-01 1.30048186e-01 6.37320936e-01 -4.61053699e-01
1.37781227e+00 5.52979648e-01 4.95242655e-01 -6.07068479e-01
-4.66840148e-01 -1.13576376e+00 -1.20330477e+00 -1.99400976e-01
7.05325484e-01 -8.53022277e-01 -8.66058528e-01 3.18672091e-01
-1.01744175e+00 4.55980480e-01 -1.46101013e-01 4.26075846e-01
-6.41962409e-01 1.10180700e+00 -4.37011123e-01 -4.86632407e-01
2.31730081e-02 -1.00648999e+00 9.46268916e-01 1.49614215e-01
1.57207817e-01 -1.35850012e+00 4.61082250e-01 3.45080465e-01
-2.35640332e-01 8.35976481e-01 9.05820131e-01 -4.99238044e-01
-1.07336938e+00 -3.60889345e-01 -5.28121769e-01 -1.07737713e-01
4.11160588e-01 6.57001212e-02 -4.09234136e-01 -8.15245688e-01
-1.05642527e-01 1.21370524e-01 5.48947990e-01 3.79802465e-01
5.91083527e-01 -4.19571519e-01 -6.21461809e-01 7.63427377e-01
1.76892292e+00 -1.83800995e-01 5.63935220e-01 7.37898350e-02
8.40232491e-01 4.10465479e-01 6.48690760e-01 2.43241847e-01
2.17453599e-01 8.38665903e-01 4.76224512e-01 -5.47752082e-01
-1.67535171e-02 -6.17788672e-01 -6.30614832e-02 1.52767086e+00
-7.99383298e-02 -1.11194707e-01 -5.34248888e-01 5.36045611e-01
-2.43894076e+00 -1.01964724e+00 -7.98138678e-01 2.26764202e+00
2.51067191e-01 -6.39647320e-02 6.36472777e-02 -2.65319467e-01
1.25436449e+00 4.12642151e-01 -2.24462867e-01 -9.38706249e-02
-4.34277356e-01 -1.46515891e-01 4.84743595e-01 8.58721793e-01
-1.13254285e+00 7.87352920e-01 7.07321405e+00 8.23557913e-01
-3.27412158e-01 1.07101865e-01 1.62155613e-01 4.50203896e-01
-5.16421556e-01 8.75563502e-01 -3.05873543e-01 2.91959912e-01
3.23267616e-02 -4.59072143e-01 5.20281911e-01 6.78752482e-01
7.69516975e-02 -2.20004725e-03 -9.38357174e-01 9.93822277e-01
5.46378233e-02 -1.26912808e+00 4.24925424e-02 3.36770505e-01
1.17542779e+00 -4.12454128e-01 -4.88568068e-01 -2.14671671e-01
1.14755757e-01 -9.60772038e-01 1.81845665e-01 7.07798898e-01
7.17287600e-01 -5.55385530e-01 5.60800910e-01 3.75432342e-01
-1.89346552e+00 3.71912390e-01 -5.63917816e-01 5.27210813e-03
2.10601613e-01 5.84577024e-01 -5.96901774e-01 1.02846825e+00
2.36091569e-01 8.90074193e-01 -4.49928045e-01 1.23217332e+00
-1.60356954e-01 7.68885165e-02 -2.70358413e-01 3.94078940e-01
3.97852808e-02 -1.20529950e+00 9.81004775e-01 1.00146198e+00
-5.22735976e-02 1.92789957e-01 9.66033936e-01 1.05141473e+00
-1.86155841e-01 2.71696895e-01 -4.54566717e-01 1.71975121e-02
6.41639888e-01 1.27975965e+00 -8.29091609e-01 -1.56220332e-01
-4.89311665e-01 9.85094547e-01 4.56222832e-01 3.84007663e-01
-6.04622126e-01 -3.15042496e-01 3.86645377e-01 1.58631906e-01
4.38882150e-02 -4.10308808e-01 -2.83482611e-01 -1.10493135e+00
6.46352321e-02 -8.17117453e-01 6.27497494e-01 -5.15635252e-01
-1.49788773e+00 3.53512406e-01 1.73290327e-01 -1.13536406e+00
8.43013972e-02 -3.04101557e-01 -9.16790724e-01 9.20462608e-01
-1.27904749e+00 -1.42201233e+00 -5.42584658e-01 8.99496555e-01
1.09218247e-02 -6.09622896e-02 6.50024831e-01 3.23636770e-01
-3.96602482e-01 6.47637844e-01 3.76840711e-01 9.84975621e-02
5.78108132e-01 -1.15058160e+00 3.25795144e-01 1.00091159e+00
2.76977688e-01 7.10153043e-01 4.65930104e-01 -1.04276240e+00
-1.30947685e+00 -8.47194493e-01 9.96719658e-01 -1.97363734e-01
7.08164632e-01 -3.98912162e-01 -7.48172939e-01 8.76327991e-01
1.07875079e-01 1.33067640e-02 2.98755169e-01 1.76486731e-01
-2.12673411e-01 -1.61163472e-02 -1.19570124e+00 2.61238486e-01
1.15524626e+00 -5.97084582e-01 -4.15983766e-01 7.40291059e-01
6.75298929e-01 -5.22918522e-01 -1.15378845e+00 5.44960797e-01
2.86232769e-01 -9.02750731e-01 9.89955723e-01 -5.93610466e-01
-1.62699327e-01 -5.48505902e-01 -2.24535257e-01 -8.42764676e-01
-9.17528331e-01 -9.34930921e-01 1.76420867e-01 1.15808213e+00
1.68040380e-01 -6.72056139e-01 1.09484529e+00 3.04370046e-01
2.22589541e-02 -5.51513553e-01 -1.14997900e+00 -1.03444839e+00
-4.04677153e-01 2.31902510e-01 5.78027785e-01 1.47464323e+00
8.15651491e-02 3.88712198e-01 -5.98974764e-01 5.61095655e-01
1.20714796e+00 6.10810816e-01 9.97493267e-01 -1.30607116e+00
-5.61608016e-01 -2.11971238e-01 -7.46096849e-01 -1.05048561e+00
5.96063659e-02 -1.14943373e+00 -1.95755273e-01 -1.52816331e+00
9.44428265e-01 -2.70175904e-01 -3.26979570e-02 2.73494542e-01
-2.44197667e-01 9.14019495e-02 4.58834976e-01 6.27949476e-01
-5.73510051e-01 5.27295172e-01 1.49748337e+00 -1.96594656e-01
-2.05723375e-01 9.86330360e-02 -2.65089989e-01 6.56587124e-01
2.74273962e-01 -7.55689561e-01 -2.78972387e-01 1.77607927e-02
-5.26919961e-02 5.74032545e-01 1.87097639e-01 -6.65263712e-01
5.43091834e-01 -5.76191366e-01 -2.58137435e-01 -6.75244212e-01
4.78758574e-01 -9.57206190e-01 7.37294376e-01 5.18182576e-01
4.54361923e-02 1.86512724e-01 -2.87544459e-01 9.48827028e-01
-4.77927566e-01 -2.35223114e-01 9.26663995e-01 -8.19688961e-02
-4.81843233e-01 5.25001943e-01 2.37854704e-01 8.15226883e-02
1.17711747e+00 -7.70775914e-01 -3.75459999e-01 -5.77629447e-01
-1.11972380e+00 4.02591139e-01 6.52489305e-01 1.96458161e-01
6.61170423e-01 -1.67078161e+00 -1.00885761e+00 2.81148314e-01
1.73996583e-01 -2.64774233e-01 3.96877324e-04 9.00278151e-01
-4.50793296e-01 3.07949096e-01 -6.70246705e-02 -8.17176759e-01
-1.86876881e+00 5.78793526e-01 2.87664562e-01 -5.53073645e-01
-3.73508215e-01 7.64216483e-01 3.45088184e-01 -7.47467101e-01
-4.32176404e-02 5.48380949e-02 2.26532832e-01 -2.92389303e-01
-3.12165290e-01 6.79799497e-01 -3.05239946e-01 -1.00535357e+00
-5.51675260e-01 1.02075648e+00 1.36446297e-01 2.60325789e-01
1.04585719e+00 -2.58930326e-01 -8.18848670e-01 -5.23916259e-02
1.46013844e+00 3.20575863e-01 -8.08687925e-01 -4.16613698e-01
4.89983223e-02 -1.10880947e+00 -3.31792682e-01 -1.98418975e-01
-1.04578161e+00 3.67938459e-01 3.14001143e-01 3.34006339e-01
1.05794024e+00 1.72776934e-02 4.64837074e-01 1.44468665e-01
3.29845101e-01 -1.07564604e+00 2.95341313e-01 2.63285518e-01
9.35271800e-01 -1.10001445e+00 6.24946475e-01 -1.25789285e+00
-1.61810860e-01 9.79984641e-01 5.44223547e-01 -4.78647530e-01
9.39146638e-01 -1.69995844e-01 -2.56500304e-01 -4.93781805e-01
-4.03450847e-01 -1.80143818e-01 8.73008192e-01 5.29227674e-01
1.03858091e-01 2.28327885e-01 -5.13512313e-01 1.07502118e-01
1.14115909e-01 -5.40331900e-01 3.33769709e-01 5.51964700e-01
-5.19913495e-01 -1.41327095e+00 -4.78893816e-01 3.65259290e-01
1.68457001e-01 1.09084174e-02 -7.73836434e-01 1.00145233e+00
8.64812583e-02 1.17884314e+00 -1.73547864e-02 -3.89714956e-01
3.95130098e-01 -6.93546712e-01 4.83450234e-01 -7.21944451e-01
-3.94841343e-01 7.37545848e-01 1.16489612e-01 -7.92792737e-01
-6.43989503e-01 -5.91810822e-01 -1.14326882e+00 -3.69643033e-01
-9.92324710e-01 2.11199522e-01 2.00137123e-01 8.32044780e-01
2.27339700e-01 1.45773888e-01 8.93402576e-01 -4.75292385e-01
-4.30735260e-01 -6.73529446e-01 -1.01782584e+00 8.04387569e-01
-4.44016382e-02 -4.09075320e-01 -5.09998083e-01 -4.11432497e-02] | [7.362357139587402, 5.0211591720581055] |
5e02c106-31be-45e1-9cc9-8b4632ff0b09 | generating-followup-questions-for | 2002.12344 | null | https://arxiv.org/abs/2002.12344v1 | https://arxiv.org/pdf/2002.12344v1.pdf | Generating Followup Questions for Interpretable Multi-hop Question Answering | We propose a framework for answering open domain multi-hop questions in which partial information is read and used to generate followup questions, to finally be answered by a pretrained single-hop answer extractor. This framework makes each hop interpretable, and makes the retrieval associated with later hops as flexible and specific as for the first hop. As a first instantiation of this framework, we train a pointer-generator network to predict followup questions based on the question and partial information. This provides a novel application of a neural question generation network, which is applied to give weak ground truth single-hop followup questions based on the final answers and their supporting facts. Learning to generate followup questions that select the relevant answer spans against downstream supporting facts, while avoiding distracting premises, poses an exciting semantic challenge for text generation. We present an evaluation using the two-hop bridge questions of HotpotQA. | ['Christopher Malon', 'Bing Bai'] | 2020-02-27 | null | null | null | null | ['multi-hop-question-answering'] | ['knowledge-base'] | [ 4.36769396e-01 1.20731413e+00 -5.44040315e-02 -4.33616072e-01
-1.62617624e+00 -8.89505267e-01 7.44237840e-01 1.38322309e-01
-1.12857431e-01 1.25018167e+00 7.07036436e-01 -5.71035683e-01
-2.72073984e-01 -1.35697174e+00 -9.38758731e-01 1.63802147e-01
3.76773089e-01 1.28564095e+00 7.36850262e-01 -9.97889221e-01
2.12986946e-01 -1.74191594e-01 -1.58952200e+00 1.12629414e+00
1.35150087e+00 7.22326815e-01 1.49984777e-01 8.94996703e-01
-8.13224316e-01 1.05462003e+00 -8.88341963e-01 -8.33892643e-01
-1.83341373e-02 -7.15068340e-01 -1.77336848e+00 -4.29028451e-01
5.48050284e-01 -4.09736603e-01 -1.48042977e-01 4.20939565e-01
3.70925486e-01 3.13453555e-01 6.16756916e-01 -1.11306250e+00
-1.08107805e+00 1.21788573e+00 3.46072823e-01 2.97357410e-01
9.97642398e-01 2.28842184e-01 1.67090321e+00 -7.21375346e-01
9.21884596e-01 1.41909599e+00 5.12918532e-01 9.36905622e-01
-1.00500131e+00 -1.09445840e-01 1.28102601e-01 3.97126198e-01
-5.96595824e-01 -2.42811576e-01 5.09210348e-01 -3.75985652e-02
9.09439266e-01 4.42428321e-01 3.11183214e-01 1.37013614e+00
-1.20152704e-01 7.14840710e-01 7.41925776e-01 -4.55645323e-01
1.21851839e-01 5.07785119e-02 4.27824110e-01 7.49462485e-01
8.12716931e-02 -8.16363841e-02 -3.94513756e-01 -2.65881091e-01
8.27473328e-02 -6.07836902e-01 -3.28233063e-01 2.86458284e-01
-1.10008717e+00 1.24799848e+00 7.12809682e-01 7.31379166e-02
-2.34469220e-01 -7.29107931e-02 2.17667427e-02 4.92619306e-01
2.94198006e-01 1.07483304e+00 -6.90982521e-01 3.48588169e-01
-8.51722836e-01 9.17833567e-01 1.27145123e+00 9.62750793e-01
9.20147896e-01 -5.69901764e-01 -1.05497861e+00 5.41951895e-01
2.11586639e-01 3.83038193e-01 3.17021012e-01 -1.19169867e+00
8.39764595e-01 7.01692164e-01 3.01449150e-01 -8.87644291e-01
-3.21331561e-01 -2.48556584e-01 -2.25071192e-01 -3.33800405e-01
5.79075277e-01 -4.25439209e-01 -6.95561230e-01 1.79157770e+00
5.26814163e-01 -7.12457895e-02 4.96018410e-01 7.41149545e-01
1.36979651e+00 9.96701837e-01 8.58294591e-03 2.57164121e-01
1.58565366e+00 -1.18351841e+00 -5.13717949e-01 -4.72919405e-01
6.28478646e-01 -5.29518187e-01 1.39758384e+00 -1.42612204e-01
-1.17998707e+00 -6.33242130e-01 -7.98150122e-01 -6.97610617e-01
-4.53456342e-01 -2.61885047e-01 2.64555931e-01 2.94955540e-02
-1.12612617e+00 4.31413740e-01 7.88701475e-02 -2.63061136e-01
1.20105788e-01 1.56340823e-02 8.93119350e-02 -2.15657160e-01
-1.98551774e+00 1.00475419e+00 5.29496729e-01 -3.10505718e-01
-6.45687699e-01 -1.12589741e+00 -7.79621363e-01 1.97790071e-01
5.69396257e-01 -1.55396438e+00 1.66636097e+00 -4.14312065e-01
-1.34332585e+00 7.83925176e-01 -3.10694039e-01 -7.33069301e-01
3.54923844e-01 -8.32263902e-02 -3.90309811e-01 5.99538386e-01
8.25039804e-01 1.16150832e+00 5.60169458e-01 -1.18745053e+00
-7.12944806e-01 -1.50501221e-01 7.14179516e-01 1.59644514e-01
2.03274056e-01 -3.33239168e-01 2.19425559e-02 -3.68857533e-01
-3.70271504e-01 -6.06987774e-01 -1.66162878e-01 -3.77938956e-01
-8.21452022e-01 -5.88837206e-01 6.36247694e-01 -8.65565062e-01
1.12687862e+00 -1.30185854e+00 -1.30592324e-02 4.55118269e-02
1.33245021e-01 -8.77601206e-02 -5.89091539e-01 6.40752137e-01
1.26227692e-01 2.08356217e-01 -3.95797551e-01 2.38411516e-01
2.87721485e-01 2.17873394e-01 -1.10418534e+00 -8.01092863e-01
6.51629150e-01 1.42220986e+00 -1.17752469e+00 -4.74136174e-01
-6.35810554e-01 -1.47650987e-01 -8.61175358e-01 5.78052461e-01
-1.17016947e+00 3.19797367e-01 -5.96208990e-01 1.43343851e-01
2.89649308e-01 -5.54570317e-01 -2.63240248e-01 -1.34327620e-01
3.40318233e-01 1.06132293e+00 -6.68442547e-01 1.62164724e+00
-6.06555581e-01 3.27126980e-01 -2.25850448e-01 -5.69795251e-01
1.04156578e+00 3.77650887e-01 -1.06635176e-01 -7.26101458e-01
-3.18640381e-01 3.42091739e-01 -4.85610098e-01 -8.62044156e-01
8.21335614e-01 -2.58051246e-01 -4.68441635e-01 7.72899926e-01
2.71386713e-01 -6.14881694e-01 6.85005248e-01 7.11194038e-01
1.32570755e+00 2.14911118e-01 -1.06512964e-01 -2.42235325e-02
6.37446940e-01 5.35236359e-01 3.03364284e-02 1.04378223e+00
3.12705874e-01 8.04826975e-01 6.30222380e-01 -4.29068774e-01
-7.19218552e-01 -1.31955528e+00 3.74787122e-01 1.05693996e+00
5.17496914e-02 -3.64140451e-01 -8.69259953e-01 -1.36172795e+00
-1.64740905e-03 1.24871731e+00 -5.37890077e-01 -3.76948684e-01
-7.62446344e-01 -9.60767418e-02 7.44299948e-01 3.68267596e-01
4.00131136e-01 -1.30557561e+00 -4.39760029e-01 1.81086779e-01
-1.20786452e+00 -1.12386286e+00 -3.78807247e-01 -1.86557397e-02
-5.90341747e-01 -1.36991787e+00 -5.58328688e-01 -8.86550963e-01
3.76320243e-01 -1.16008803e-01 1.84333384e+00 3.09991747e-01
1.50345802e-01 4.04883593e-01 -4.03329104e-01 -3.11141551e-01
-7.30566919e-01 6.62197471e-01 -7.16603398e-01 -2.38418013e-01
1.29402295e-01 -3.21933776e-01 -6.58938408e-01 -8.03685933e-02
-9.30985093e-01 5.37666455e-02 4.30706292e-01 8.47614348e-01
4.51284021e-01 -5.41648269e-01 1.22293854e+00 -1.32258213e+00
1.26622522e+00 -1.01781929e+00 -2.75466174e-01 5.90304732e-01
-2.61963964e-01 5.70460558e-01 9.57914770e-01 -1.14519922e-02
-1.60680270e+00 -3.83556396e-01 -5.33602774e-01 4.12806273e-01
-1.12628892e-01 5.24323523e-01 -1.19032897e-01 6.34750903e-01
1.07026410e+00 -1.78966820e-01 -3.58106464e-01 -2.64385194e-01
1.12399995e+00 2.76042730e-01 7.92657197e-01 -9.85446215e-01
1.00959802e+00 2.33869731e-01 -2.76184440e-01 -2.89842755e-01
-1.76214015e+00 -8.14844370e-02 -3.13554317e-01 7.01979250e-02
9.87754166e-01 -4.98369545e-01 -3.57062638e-01 -3.11500549e-01
-1.76911163e+00 -3.45185846e-01 -8.00622284e-01 -2.89153844e-01
-5.80842614e-01 1.04362272e-01 -5.11230707e-01 -2.02446550e-01
-5.30605316e-01 -5.97606599e-01 1.19964182e+00 3.29302609e-01
-6.61160946e-01 -1.03522050e+00 1.93897530e-01 9.29385424e-01
2.39546031e-01 8.55302289e-02 1.51180220e+00 -1.12388158e+00
-9.28602338e-01 5.06288074e-02 -1.31075174e-01 -6.34119613e-03
-1.96401179e-01 -3.02873611e-01 -9.32384312e-01 4.00966585e-01
-1.52362719e-01 -7.57648945e-01 9.91218150e-01 -2.99127176e-02
9.28327203e-01 -8.31509888e-01 -1.88320398e-01 1.22688204e-01
9.56399977e-01 -3.73527616e-01 6.06173933e-01 2.27771968e-01
2.57701457e-01 1.14042521e+00 4.68737036e-01 -2.31989957e-02
9.27667499e-01 3.31183642e-01 3.41297805e-01 1.60761431e-01
-4.38909024e-01 -9.01763380e-01 9.24280584e-02 4.33370948e-01
5.09300172e-01 -4.46592599e-01 -6.49730504e-01 8.19392681e-01
-1.49158406e+00 -1.31343210e+00 -4.57259327e-01 1.54894197e+00
1.25114977e+00 2.71303430e-02 7.96030313e-02 -1.74903333e-01
5.38728535e-01 1.79674834e-01 -2.25614429e-01 -4.18136954e-01
-3.03345293e-01 8.23956668e-01 -3.06409627e-01 8.72611761e-01
-6.03073120e-01 9.92479146e-01 6.31029606e+00 5.78414798e-01
-4.40833569e-01 7.44071752e-02 5.79365671e-01 2.02113152e-01
-1.26016438e+00 4.17212367e-01 -9.68925595e-01 3.65218431e-01
1.11825240e+00 -4.23905283e-01 1.50394306e-01 5.37910581e-01
-1.27839059e-01 4.71054912e-02 -1.39797926e+00 1.00732610e-01
4.11301143e-02 -1.89554787e+00 8.38213563e-01 -5.41811407e-01
7.13935375e-01 -3.83349061e-01 -1.86669417e-02 7.85849988e-01
6.70794964e-01 -1.00095379e+00 5.90762198e-01 6.44941986e-01
4.97584373e-01 -6.42109573e-01 5.89218855e-01 4.76555735e-01
-8.03829014e-01 -3.24772030e-01 -2.63147116e-01 -6.42143115e-02
5.88175535e-01 5.34836709e-01 -1.14863074e+00 7.01612592e-01
4.01118964e-01 -3.44309844e-02 -6.14810288e-01 5.69038630e-01
-9.18220222e-01 5.28017104e-01 -1.70300871e-01 -2.15893805e-01
4.24856067e-01 1.85164452e-01 5.08011401e-01 1.00864410e+00
4.75223094e-01 3.39940548e-01 -1.54187912e-02 1.43744040e+00
-5.46151459e-01 -1.57795727e-01 -6.33977592e-01 1.82598919e-01
6.58916175e-01 1.12957788e+00 -2.04526410e-01 -4.99152064e-01
-1.62424296e-01 1.00400460e+00 5.92026651e-01 4.18188095e-01
-6.60477638e-01 -6.49801075e-01 2.38960475e-01 2.84388095e-01
1.66787520e-01 4.66243595e-01 -1.97931021e-01 -1.24756253e+00
2.08426133e-01 -9.26900208e-01 9.65519011e-01 -1.09151101e+00
-1.51105154e+00 7.80947864e-01 -3.48071307e-02 -6.00837350e-01
-9.40368831e-01 -3.25156599e-01 -1.01337314e+00 1.02284765e+00
-1.88878191e+00 -1.07013035e+00 -2.51601517e-01 5.34040928e-01
6.62681282e-01 2.08286092e-01 9.17259276e-01 -5.27826417e-03
-7.49858608e-03 4.64515269e-01 -6.29026175e-01 2.04773340e-02
5.92875719e-01 -1.48120070e+00 6.48457706e-01 7.07663655e-01
2.51099497e-01 5.32951415e-01 7.34295189e-01 -6.77870572e-01
-8.40031326e-01 -1.34880722e+00 1.68960643e+00 -8.69757175e-01
6.15880311e-01 -9.66279730e-02 -1.11477363e+00 6.60473645e-01
5.15028179e-01 -4.72477078e-01 5.74918568e-01 6.36134818e-02
-4.65564698e-01 1.94598865e-02 -1.09226286e+00 7.14503586e-01
1.05558836e+00 -5.72471380e-01 -1.52211571e+00 5.23757875e-01
1.45544541e+00 -3.67891461e-01 -4.35602129e-01 3.57541621e-01
-5.93636185e-02 -9.49198723e-01 1.06637716e+00 -1.00282657e+00
1.04071975e+00 -2.34024182e-01 1.08416691e-01 -1.44445837e+00
-1.30247504e-01 -8.19656193e-01 -3.55943382e-01 1.27345109e+00
1.19336092e+00 -3.64387304e-01 6.37700856e-01 5.19364238e-01
-5.17361104e-01 -7.90748894e-01 -8.11595678e-01 -3.63025904e-01
2.75957733e-01 -1.66672438e-01 9.32482719e-01 5.70412695e-01
-4.58236188e-02 1.05880165e+00 7.13558272e-02 7.10532069e-02
3.44512850e-01 5.27671039e-01 6.39332771e-01 -1.17335021e+00
-4.53743339e-01 -1.18676037e-01 5.46625972e-01 -1.49457955e+00
4.31292355e-01 -1.30497730e+00 1.11151300e-01 -2.16144872e+00
-1.01597212e-01 -2.95620948e-01 3.29838395e-01 2.72020549e-01
-6.44118249e-01 9.50390548e-02 1.22740224e-01 -2.76338965e-01
-5.76774538e-01 5.89906275e-01 1.40429997e+00 -2.65321583e-01
1.05347909e-01 8.31006467e-02 -1.38917649e+00 4.97650653e-01
5.53409874e-01 -3.12746286e-01 -7.83942640e-01 -7.84308910e-01
6.56982481e-01 3.64170909e-01 5.39968491e-01 -7.31718659e-01
2.73749113e-01 -1.73677340e-01 1.74789250e-01 -6.67537868e-01
2.85777241e-01 -2.69233257e-01 -4.96576071e-01 1.37021109e-01
-1.01496267e+00 7.60069415e-02 -8.18256289e-02 4.73542392e-01
-3.41254532e-01 -4.74615663e-01 3.77492875e-01 -4.15420234e-01
-5.62349677e-01 1.07200108e-01 -1.24083504e-01 1.03791225e+00
6.27885520e-01 5.15884794e-02 -8.58975887e-01 -7.74744153e-01
-7.20248878e-01 7.22586155e-01 -2.01976597e-01 4.36099380e-01
7.65728951e-01 -1.22126687e+00 -1.07298756e+00 -1.95940092e-01
1.96194023e-01 2.23273069e-01 3.83862764e-01 1.53470635e-01
-2.38203794e-01 6.61390483e-01 1.15659088e-01 -2.95922775e-02
-5.74799716e-01 4.81762141e-01 3.01525682e-01 -8.18359375e-01
-4.42041785e-01 1.06379986e+00 5.42851351e-02 -8.85988235e-01
-2.67002702e-01 -4.22670752e-01 -3.59043360e-01 2.09309205e-01
6.12714708e-01 2.61768192e-01 1.94423482e-01 -3.37069258e-02
1.06685661e-01 2.06759483e-01 1.02439806e-01 -4.49032098e-01
9.68375266e-01 -1.93701282e-01 -2.32523814e-01 -9.98764765e-03
8.59111547e-01 1.68933719e-01 -7.58887351e-01 -2.06865326e-01
4.57255661e-01 -7.13003054e-02 -7.37188637e-01 -1.29867744e+00
-2.97692209e-01 7.73432136e-01 -4.63087052e-01 4.47428286e-01
7.83642411e-01 5.87290704e-01 1.50313902e+00 7.81956971e-01
6.85141087e-02 -9.02019680e-01 5.04626453e-01 1.09332442e+00
1.26301277e+00 -9.45887864e-01 -8.45336676e-01 -4.59444046e-01
-5.45325160e-01 1.14493310e+00 9.86855805e-01 3.37000145e-03
5.96834579e-03 -5.39337844e-02 6.13457561e-02 -4.05394554e-01
-1.13675594e+00 -2.20098853e-01 4.42906708e-01 8.11895847e-01
3.06033105e-01 -3.43880802e-01 -1.42235786e-01 9.66036797e-01
-9.27465737e-01 -1.65815562e-01 5.65188766e-01 5.30238926e-01
-6.86327934e-01 -1.09809518e+00 -2.52206802e-01 5.43817997e-01
-3.35343301e-01 -3.95290911e-01 -8.81864071e-01 4.97870058e-01
9.63037983e-02 1.39704871e+00 -8.95562321e-02 -5.07973507e-02
5.36126137e-01 6.20522201e-01 3.32857519e-01 -1.07125688e+00
-7.69751012e-01 -1.00804031e+00 7.05427051e-01 -3.67176026e-01
9.62051749e-02 1.82063505e-02 -1.55558217e+00 -1.49503827e-01
-1.17976144e-02 8.53758574e-01 1.20498940e-01 1.13919485e+00
7.30978727e-01 6.18096888e-01 4.37954664e-01 -8.12729374e-02
-8.97612154e-01 -7.42948532e-01 1.60199791e-01 6.13488913e-01
5.08581519e-01 -1.87757552e-01 -2.38601640e-01 5.20941615e-02] | [11.31721305847168, 8.080440521240234] |
6e5ec2f8-f6cf-4718-8b63-5c1b05e73eb2 | a-crowd-annotated-spanish-corpus-for-humor | 1710.00477 | null | http://arxiv.org/abs/1710.00477v4 | http://arxiv.org/pdf/1710.00477v4.pdf | A Crowd-Annotated Spanish Corpus for Humor Analysis | Computational Humor involves several tasks, such as humor recognition, humor
generation, and humor scoring, for which it is useful to have human-curated
data. In this work we present a corpus of 27,000 tweets written in Spanish and
crowd-annotated by their humor value and funniness score, with about four
annotations per tweet, tagged by 1,300 people over the Internet. It is equally
divided between tweets coming from humorous and non-humorous accounts. The
inter-annotator agreement Krippendorff's alpha value is 0.5710. The dataset is
available for general use and can serve as a basis for humor detection and as a
first step to tackle subjectivity. | ['Guillermo Moncecchi', 'Aiala Rosá', 'Luis Chiruzzo', 'Diego Garat', 'Santiago Castro'] | 2017-10-02 | a-crowd-annotated-spanish-corpus-for-humor-1 | https://aclanthology.org/W18-3502 | https://aclanthology.org/W18-3502.pdf | ws-2018-7 | ['humor-detection'] | ['natural-language-processing'] | [-6.22247756e-01 3.54723871e-01 -4.06647511e-02 1.10005289e-01
-1.93669885e-01 -5.66824317e-01 8.25301051e-01 6.01036370e-01
-3.97028416e-01 1.10510170e+00 8.33635569e-01 -9.46681127e-02
6.15780413e-01 -6.17701411e-01 3.34088318e-03 -3.61312389e-01
2.90726990e-01 4.87030298e-01 1.99857354e-01 -5.77298701e-01
6.00240707e-01 1.88276365e-01 -1.38924372e+00 5.48182487e-01
8.75079751e-01 4.72064704e-01 -5.99528365e-02 7.05599904e-01
-6.45729750e-02 1.93599451e+00 -8.29020560e-01 -1.06704855e+00
-3.31003398e-01 -7.21299827e-01 -1.21599102e+00 -1.67651772e-01
1.41375974e-01 -2.03283635e-05 -3.90683860e-01 9.70160306e-01
4.33501482e-01 1.61857828e-01 5.59589803e-01 -8.87403369e-01
-3.35810155e-01 7.43661523e-01 -1.71009362e-01 1.06971636e-01
7.26254225e-01 2.30628073e-01 1.19808865e+00 -1.14329362e+00
9.11035776e-01 1.04875505e+00 6.85749710e-01 6.65635347e-01
-8.23954344e-01 -3.93172860e-01 -1.08441496e+00 7.21821785e-01
-1.02841389e+00 -2.79683828e-01 7.31455028e-01 -1.04194248e+00
5.38949907e-01 3.39784354e-01 6.59367681e-01 1.06670690e+00
3.91300693e-02 8.09291601e-01 1.16550207e+00 -4.65825498e-01
2.09066257e-01 6.53606296e-01 2.47941434e-01 5.69389701e-01
1.30088955e-01 -5.35349905e-01 -7.27564156e-01 -4.64784294e-01
1.43605277e-01 -3.54010910e-01 -4.07134235e-01 4.83141392e-01
-9.28986609e-01 1.22048748e+00 5.05061150e-01 7.25456655e-01
-3.71599853e-01 -3.84249061e-01 9.28018808e-01 2.85145283e-01
4.28031981e-01 7.94314444e-01 1.23508632e-01 -4.16566670e-01
-1.01456761e+00 6.14015281e-01 1.33141601e+00 7.34479070e-01
6.59759104e-01 5.80406189e-02 -1.29053220e-01 9.39521968e-01
-1.67651400e-01 4.91182864e-01 7.95216203e-01 -7.94705093e-01
8.42379555e-02 8.93971026e-01 4.73746240e-01 -1.61808550e+00
-6.32789135e-01 -7.97666162e-02 -7.67285466e-01 2.31303070e-02
5.88275731e-01 -2.48902924e-02 -4.54818532e-02 1.13141775e+00
1.39305182e-02 -5.86133838e-01 -3.91073972e-01 1.21862340e+00
1.41261637e+00 6.17882192e-01 -8.09939802e-02 -4.81709749e-01
1.48955286e+00 -9.66356099e-01 -9.64961767e-01 3.05166561e-02
4.87835199e-01 -1.07388592e+00 1.21836150e+00 5.21306157e-01
-1.27580500e+00 -1.66605353e-01 -6.59333110e-01 -3.93824100e-01
-3.59687716e-01 -2.03405753e-01 3.98713052e-02 5.08421540e-01
-4.05312002e-01 4.46451217e-01 -4.23366912e-02 -2.75535852e-01
2.42898360e-01 -3.40339065e-01 -3.48802090e-01 4.42458391e-01
-1.46003485e+00 1.36774755e+00 3.87308389e-01 -5.23065925e-01
-7.91385174e-01 -2.25850984e-01 -5.37989497e-01 -2.00793400e-01
-2.23409664e-03 -2.82253921e-01 1.33907127e+00 -1.18411148e+00
-1.12438881e+00 1.56760943e+00 -2.53918290e-01 -6.33961976e-01
7.48503447e-01 -5.14670871e-02 -3.64248186e-01 1.95725307e-01
2.36726359e-01 -3.61453742e-01 7.38064051e-01 -1.12680459e+00
-3.83932382e-01 -1.53959006e-01 -3.69839311e-01 -2.20694378e-01
-5.99410295e-01 6.22511506e-01 1.47998914e-01 -4.69136655e-01
-4.53356475e-01 -5.86731017e-01 1.38981208e-01 -9.15623784e-01
-3.88656467e-01 -3.02486688e-01 4.38153386e-01 -1.18224382e+00
1.89854491e+00 -1.77642357e+00 1.15941400e-02 3.30376066e-02
8.29470813e-01 3.24058682e-01 4.07323748e-01 7.65129507e-01
2.15221971e-01 3.95856872e-02 -8.17149580e-02 -2.24917024e-01
9.17300954e-02 -1.12825722e-01 -2.44514719e-01 6.13294303e-01
-4.64918971e-01 8.45447958e-01 -1.31661522e+00 -6.77784145e-01
1.51349366e-01 8.24068636e-02 -1.72139972e-01 5.35365701e-01
1.67476218e-02 4.07203108e-01 -7.10844547e-02 3.95239800e-01
3.03271592e-01 -2.57205874e-01 9.08209607e-02 2.84446001e-01
-4.41960663e-01 6.53306305e-01 -3.27167243e-01 5.50532937e-01
-4.37119454e-01 1.22984135e+00 -9.98809189e-02 7.29494467e-02
1.29983842e+00 3.50244492e-01 1.87425554e-01 -6.46338880e-01
6.09331131e-01 6.53679430e-01 -1.76907480e-01 -5.87830782e-01
1.12295747e+00 -5.90094090e-01 -3.04199070e-01 4.82916474e-01
-1.45821616e-01 -1.61225274e-01 5.97030222e-01 3.75304729e-01
1.10710871e+00 -6.88717067e-01 1.04457152e+00 -3.59404385e-01
1.02521694e+00 2.91933715e-01 2.86158383e-01 4.59861547e-01
-3.00091356e-01 5.40104926e-01 9.15179431e-01 -8.01517665e-01
-1.59167075e+00 -4.77058709e-01 -3.50352794e-01 1.22320080e+00
-2.31546506e-01 -6.77056491e-01 -8.82931709e-01 -2.49103978e-01
-1.19018786e-01 1.07223678e+00 -5.95558345e-01 2.12531030e-01
-5.40082514e-01 -3.75053138e-01 5.74013233e-01 -2.06071530e-02
4.56246078e-01 -1.45221090e+00 -7.36621916e-01 2.18350828e-01
-8.20848584e-01 -8.58403444e-01 -1.00731365e-01 9.42154787e-03
-1.92238957e-01 -1.29103553e+00 -6.61874473e-01 -4.51622605e-01
2.08522722e-01 2.40067437e-01 1.54575169e+00 5.45032144e-01
1.00993611e-01 -1.36260122e-01 -8.98514330e-01 -2.94110626e-01
-8.48439217e-01 1.95698544e-01 -4.01207283e-02 -3.09242845e-01
7.83418059e-01 -3.31062168e-01 -2.47046918e-01 4.01698291e-01
-4.32780802e-01 -1.61061421e-01 -2.97431558e-01 1.11494339e+00
-3.18747044e-01 -3.72651279e-01 4.85289365e-01 -1.06641269e+00
9.42200303e-01 -8.47160876e-01 -1.57372713e-01 -4.38981831e-01
-3.79758209e-01 -6.91951811e-01 1.19684577e+00 1.15281269e-01
-7.80529737e-01 -3.26822519e-01 -2.35846229e-02 1.10941291e-01
1.76875740e-02 3.46960813e-01 3.59885812e-01 1.17057674e-01
1.40474808e+00 -2.26854429e-01 -1.14290200e-01 -4.15689826e-01
2.65099674e-01 1.24819314e+00 8.26985955e-01 -2.21854657e-01
6.29523635e-01 1.51135370e-01 -2.35103756e-01 -1.11506617e+00
-1.13757908e+00 -1.06759298e+00 -6.03374183e-01 -6.59350395e-01
5.80202937e-01 -8.33856165e-01 -1.20389485e+00 4.22689170e-01
-1.62406993e+00 -1.96797863e-01 -7.42469952e-02 -7.31640905e-02
-5.34264684e-01 5.52044392e-01 -8.83684874e-01 -1.12788212e+00
-6.31622434e-01 -1.75734147e-01 1.47458196e-01 1.86298415e-01
-1.11041677e+00 -1.11332405e+00 4.04174268e-01 8.17940652e-01
1.39656797e-01 3.77496809e-01 5.70724428e-01 -8.27199221e-01
2.92861015e-01 -5.68807483e-01 -3.13903511e-01 1.26057014e-01
-4.54482377e-01 -2.03531951e-01 -9.29793894e-01 1.29097193e-01
2.14364275e-01 -8.27449560e-01 8.22279036e-01 -2.47391045e-01
5.82621336e-01 -1.02602971e+00 3.65333915e-01 -2.93333799e-01
1.12318838e+00 -4.83275771e-01 9.63472366e-01 8.75815332e-01
5.31663299e-01 8.80419791e-01 4.11843926e-01 1.26829326e+00
3.90250683e-01 4.73109126e-01 5.19077003e-01 3.75157386e-01
6.57083765e-02 -4.21802074e-01 3.40837210e-01 9.47949529e-01
-4.33668971e-01 -1.79584712e-01 -1.26169443e+00 7.34018683e-01
-2.08524418e+00 -1.64643335e+00 -1.29564941e+00 2.06854987e+00
9.52883124e-01 -1.10870771e-01 8.95451665e-01 4.11419749e-01
7.05448627e-01 2.83004314e-01 3.32857430e-01 -7.62086987e-01
-4.27953094e-01 -1.96450368e-01 2.51659721e-01 9.08256233e-01
-6.27197444e-01 9.27299201e-01 6.13042879e+00 6.71689093e-01
-5.44119656e-01 4.48878586e-01 2.26325423e-01 -4.95517552e-02
-3.91401231e-01 -1.76268443e-01 -3.95691007e-01 9.02432263e-01
8.21119726e-01 -4.56078708e-01 5.07558823e-01 1.08263242e+00
6.74502671e-01 -3.85985196e-01 -5.99314809e-01 1.07159245e+00
6.04304731e-01 -1.10849798e+00 -4.01620775e-01 -2.52224356e-01
9.83896375e-01 -1.70567289e-01 -4.51187909e-01 2.39977777e-01
2.54113495e-01 -1.24488127e+00 9.49922562e-01 5.58163047e-01
3.49527240e-01 -6.44425213e-01 1.29227090e+00 6.88091040e-01
-5.45726418e-01 -2.19710916e-01 -4.70056564e-01 -7.35683382e-01
3.99895787e-01 9.82296348e-01 -1.12398148e+00 -2.05409616e-01
4.50871795e-01 7.05403030e-01 -5.96186817e-01 8.99640083e-01
-6.31925941e-01 7.97801435e-01 1.56216919e-01 -8.55767846e-01
9.84830707e-02 5.02482392e-02 8.10501993e-01 1.80944097e+00
-5.82026094e-02 1.25394240e-01 -6.85195923e-02 8.94453943e-01
-3.03368747e-01 5.11918724e-01 -4.11605239e-01 6.76666126e-02
5.97182810e-01 1.50025833e+00 -4.57831204e-01 -3.51262957e-01
3.54500078e-02 8.58847260e-01 4.19290274e-01 -1.56087443e-01
-6.83938980e-01 -4.36027616e-01 -2.51532178e-02 6.69514358e-01
-4.51283783e-01 -1.45333961e-01 -1.07144487e+00 -1.03806400e+00
-3.44301552e-01 -9.56910312e-01 3.05720121e-01 -8.21778297e-01
-1.48897314e+00 6.15043640e-01 -4.85532641e-01 -9.70741391e-01
-3.39917243e-01 -6.36204839e-01 -8.85976374e-01 8.86967599e-01
-8.68190944e-01 -9.88630593e-01 -9.00032878e-01 3.82614791e-01
2.55746275e-01 -4.07891273e-01 4.88990843e-01 3.61249030e-01
-1.84532106e-01 1.76679000e-01 -1.61974952e-01 1.75612330e-01
6.48855805e-01 -1.42614520e+00 -7.82608911e-02 2.96136826e-01
-4.28059548e-01 3.38011533e-01 1.50552654e+00 -5.55377185e-01
-5.97938299e-01 -6.24854624e-01 1.94082236e+00 -9.11163032e-01
1.18964875e+00 1.05731145e-01 -1.06856477e+00 1.74922228e-01
4.34702754e-01 -7.45673895e-01 8.25761259e-01 2.38107651e-01
-5.22866309e-01 4.91767615e-01 -1.06060827e+00 4.16352332e-01
4.52804774e-01 -7.11310208e-01 -7.70752788e-01 6.10076785e-01
-4.14132215e-02 -2.38162667e-01 -6.33810759e-01 -2.55126864e-01
4.71599162e-01 -1.42557108e+00 2.60108858e-01 -6.26858950e-01
1.34859574e+00 -1.74404100e-01 -9.67627577e-03 -1.01173687e+00
-6.74271822e-01 -8.10019851e-01 -1.59814376e-02 1.11109519e+00
2.56634504e-01 -1.18597895e-01 5.76601684e-01 4.72210795e-01
-1.02859713e-01 -1.93584219e-01 -7.85088181e-01 -5.24373293e-01
2.76840627e-01 -2.42524385e-01 1.08157933e-01 1.28602076e+00
1.20862603e+00 7.42186189e-01 -1.05286360e+00 -8.68652165e-01
4.49624836e-01 -1.80156767e-01 1.24376392e+00 -1.10783565e+00
6.19472899e-02 -8.31546664e-01 -4.22539532e-01 -5.63983917e-01
3.22051734e-01 -9.09956455e-01 9.72918943e-02 -1.17719328e+00
8.33193064e-01 3.37290555e-01 4.29483652e-01 4.75366712e-01
-1.28703773e-01 7.29009271e-01 3.50134581e-01 6.22965217e-01
-7.51708090e-01 5.01614153e-01 1.01699698e+00 1.10105820e-01
-2.58798867e-01 -3.22647423e-01 -3.30300719e-01 9.75106001e-01
8.22232842e-01 -4.45247203e-01 5.40473521e-01 4.00821596e-01
9.31489050e-01 -1.15507953e-01 5.74911475e-01 -6.82248712e-01
2.26337209e-01 -3.11833590e-01 -9.19130966e-02 -4.99846995e-01
-3.05785686e-02 -3.70745748e-01 -2.32300032e-02 4.80975181e-01
-1.76947281e-01 -3.27576816e-01 -3.28014046e-01 5.46018407e-02
-3.68243396e-01 -1.04773724e+00 1.18219161e+00 -3.84538382e-01
-2.81119347e-01 -4.51922059e-01 -7.28439212e-01 3.22575927e-01
7.61678994e-01 -7.73447305e-02 -6.09268785e-01 -9.02717471e-01
-4.11159784e-01 -6.48315847e-02 8.83534849e-01 1.72920506e-02
3.08535874e-01 -1.34072280e+00 -1.25750756e+00 -5.34247160e-01
2.99912512e-01 -8.29778671e-01 -5.79792820e-03 9.93575871e-01
-8.43880415e-01 2.56755948e-01 -4.28267151e-01 2.38724090e-02
-1.27019954e+00 4.82956946e-01 1.36857569e-01 -2.61200756e-01
-3.53935510e-01 3.12270820e-01 -3.21952939e-01 -1.17334366e-01
-1.85267657e-01 9.14351404e-01 -5.21825373e-01 5.59896767e-01
9.69643950e-01 1.32750452e+00 -1.48141056e-01 -1.19760990e+00
-5.88112324e-02 -4.40014675e-02 2.35369414e-01 -6.36330321e-02
9.35203016e-01 -2.66138822e-01 -8.88776660e-01 9.90840852e-01
1.00859165e+00 3.79248232e-01 -1.72284558e-01 8.08192194e-02
2.78690934e-01 -8.36184323e-01 -2.41680503e-01 -6.94505990e-01
-1.83426023e-01 7.53507912e-01 -3.25464725e-01 1.05519211e+00
6.08212471e-01 5.48442192e-02 9.93706584e-01 2.35831127e-01
3.44376683e-01 -1.47598159e+00 4.44822282e-01 1.34337521e+00
1.25645173e+00 -1.31441796e+00 1.53078049e-01 -1.38858870e-01
-9.78259087e-01 1.46474659e+00 4.39707935e-01 -5.75802736e-02
2.25046486e-01 -2.59876847e-01 -5.68024702e-02 -3.19426775e-01
-6.25359297e-01 -2.34049678e-01 4.29071039e-01 2.60511309e-01
1.16506374e+00 3.50806773e-01 -1.37381387e+00 9.89205420e-01
-8.51968706e-01 -2.21306155e-03 9.99521673e-01 5.62679283e-02
-1.21261907e+00 -4.82004642e-01 -6.22850537e-01 3.36112469e-01
-5.58756471e-01 -1.82170421e-02 -1.14424753e+00 3.85805160e-01
-2.80047487e-02 1.18830287e+00 -4.26865846e-01 -6.24154508e-01
2.29736473e-02 -1.07668176e-01 -9.06953216e-03 -3.80669504e-01
-1.26833522e+00 -4.28399801e-01 7.81620383e-01 -1.07479148e-01
-2.96303064e-01 -5.67587912e-01 -1.16574121e+00 -1.32160830e+00
-1.12807386e-01 3.69456470e-01 2.52261370e-01 8.51802230e-01
-3.39330256e-01 -2.28048712e-01 9.38443899e-01 -6.66618288e-01
-2.47497708e-01 -1.24289525e+00 -6.43769324e-01 9.17479873e-01
1.77929863e-01 -1.71220049e-01 -8.60616684e-01 2.07078099e-01] | [8.894294738769531, 11.009246826171875] |
b44a45f3-28d3-4b86-a198-7657961f640a | chameleon-plug-and-play-compositional | 2304.09842 | null | https://arxiv.org/abs/2304.09842v2 | https://arxiv.org/pdf/2304.09842v2.pdf | Chameleon: Plug-and-Play Compositional Reasoning with Large Language Models | Large language models (LLMs) have achieved remarkable progress in solving various natural language processing tasks due to emergent reasoning abilities. However, LLMs have inherent limitations as they are incapable of accessing up-to-date information (stored on the Web or in task-specific knowledge bases), using external tools, and performing precise mathematical and logical reasoning. In this paper, we present Chameleon, an AI system that mitigates these limitations by augmenting LLMs with plug-and-play modules for compositional reasoning. Chameleon synthesizes programs by composing various tools (e.g., LLMs, off-the-shelf vision models, web search engines, Python functions, and heuristic-based modules) for accomplishing complex reasoning tasks. At the heart of Chameleon is an LLM-based planner that assembles a sequence of tools to execute to generate the final response. We showcase the effectiveness of Chameleon on two multi-modal knowledge-intensive reasoning tasks: ScienceQA and TabMWP. Chameleon, powered by GPT-4, achieves an 86.54% overall accuracy on ScienceQA, improving the best published few-shot result by 11.37%. On TabMWP, GPT-4-powered Chameleon improves the accuracy by 17.0%, lifting the state of the art to 98.78%. Our analysis also shows that the GPT-4-powered planner exhibits more consistent and rational tool selection via inferring potential constraints from instructions, compared to a ChatGPT-powered planner. | ['Jianfeng Gao', 'Song-Chun Zhu', 'Ying Nian Wu', 'Kai-Wei Chang', 'Michel Galley', 'Hao Cheng', 'Baolin Peng', 'Pan Lu'] | 2023-04-19 | null | null | null | null | ['mathematical-reasoning', 'logical-reasoning'] | ['natural-language-processing', 'reasoning'] | [-1.46332771e-01 4.48048443e-01 -1.33698970e-01 2.97774673e-02
-8.62368464e-01 -7.49275208e-01 7.14107156e-01 -1.02525972e-01
-2.56625891e-01 4.63282585e-01 1.42282516e-01 -5.01610160e-01
-3.53369445e-01 -9.95110035e-01 -7.07363188e-01 -8.71188045e-02
2.35775188e-01 1.16991937e+00 3.57375681e-01 -5.47830939e-01
4.09861892e-01 3.09490621e-01 -1.74773192e+00 5.69391847e-01
1.16304612e+00 5.62371731e-01 3.89414608e-01 8.08964908e-01
-4.19277668e-01 1.40533853e+00 -5.26752174e-01 -3.06486547e-01
1.87040061e-01 -3.86225075e-01 -1.11692119e+00 -6.30869627e-01
8.66723210e-02 -3.05339038e-01 -3.56107466e-02 6.30502284e-01
1.66576713e-01 -3.05076572e-03 3.55634332e-01 -1.54483545e+00
-4.25301015e-01 8.47804666e-01 -1.06791094e-01 -2.18951672e-01
8.42478752e-01 7.16142058e-01 8.02246749e-01 -4.40736771e-01
9.11432385e-01 1.54821754e+00 5.28723717e-01 7.61629224e-01
-9.25561547e-01 -5.14439464e-01 -2.77198553e-01 4.68974590e-01
-1.18603659e+00 -5.19346535e-01 4.44771588e-01 -4.39974725e-01
1.70280766e+00 4.41550881e-01 6.50781333e-01 9.31976616e-01
3.76242131e-01 8.80957007e-01 1.04226232e+00 -6.31019890e-01
4.50687498e-01 -1.57109618e-01 -6.12742975e-02 1.15857041e+00
1.60687268e-02 -1.00083090e-02 -1.03331172e+00 -3.68698359e-01
6.43566966e-01 -5.85178971e-01 3.90667617e-02 -1.96738854e-01
-1.33546090e+00 4.84895229e-01 2.40413234e-01 7.14636594e-02
-5.76011300e-01 3.35355997e-01 3.59594196e-01 2.83050507e-01
-1.77862868e-01 1.15726924e+00 -5.05739570e-01 -6.07551098e-01
-5.95529556e-01 5.24499357e-01 1.31402254e+00 1.07732391e+00
4.77536976e-01 -1.82822913e-01 -1.82187364e-01 4.90857691e-01
2.07340866e-01 5.66688657e-01 4.61135298e-01 -1.67294884e+00
4.86275971e-01 1.04787993e+00 1.93059444e-01 -6.86995208e-01
-6.08552337e-01 -5.88951521e-02 -1.26016289e-01 4.48244959e-01
4.37432200e-01 -1.03619158e-01 -6.90992236e-01 1.61372948e+00
2.66627014e-01 -4.61287826e-01 3.58438343e-01 9.33265686e-01
6.37291431e-01 7.30239630e-01 1.66890875e-01 1.20425664e-01
1.41823196e+00 -1.24586248e+00 -3.86131614e-01 -5.55257440e-01
5.70019901e-01 -4.31871027e-01 1.40090823e+00 7.24663258e-01
-1.40981388e+00 -5.93956113e-02 -7.89904773e-01 -1.44765019e-01
-3.14069927e-01 -5.60150370e-02 1.04449630e+00 6.71218410e-02
-1.11681640e+00 3.98849219e-01 -8.15816760e-01 -6.24746680e-01
2.30336860e-01 1.94915846e-01 -1.29169255e-01 -2.55243808e-01
-9.74803269e-01 1.37035203e+00 5.16258955e-01 -2.25056484e-01
-1.02406037e+00 -8.84107590e-01 -7.03001142e-01 1.10596344e-01
9.24715400e-01 -1.18081033e+00 1.83410108e+00 -7.76699543e-01
-1.86816728e+00 6.25537872e-01 6.55141473e-02 -4.78081852e-01
7.20252156e-01 -1.35853201e-01 -1.44778460e-01 3.76864046e-01
2.12115720e-01 8.36777329e-01 3.87192965e-01 -8.69449973e-01
-4.58910763e-01 -1.26120299e-01 4.72829044e-01 2.73312330e-01
1.26736432e-01 1.33443981e-01 -4.97609466e-01 3.05854324e-02
-1.72834486e-01 -9.89401042e-01 -1.35992467e-02 -3.08200028e-02
-1.48139074e-01 -4.52537954e-01 4.43838924e-01 -6.89833164e-01
1.01694620e+00 -1.61495781e+00 4.38894272e-01 -1.30097121e-01
1.54959053e-01 2.64508665e-01 -2.88360596e-01 8.58711123e-01
4.93903816e-01 6.62254691e-02 -1.39165163e-01 1.86125547e-01
4.37135875e-01 3.16934049e-01 -2.06948325e-01 -2.80898541e-01
1.28652304e-01 1.25623739e+00 -1.15500402e+00 -7.50627756e-01
2.74910033e-01 -7.28630275e-03 -6.90599144e-01 3.57824191e-02
-9.64243889e-01 2.89226715e-02 -5.95428586e-01 8.33029568e-01
9.37516019e-02 -2.32730538e-01 4.27527428e-01 2.94730842e-01
-2.68548250e-01 3.99465978e-01 -7.96289563e-01 2.07806420e+00
-9.44703698e-01 5.24811864e-01 1.54716328e-01 -3.13560069e-01
5.58869183e-01 1.81784421e-01 -5.68236113e-02 -7.93026209e-01
2.05747373e-02 3.01854730e-01 6.34350926e-02 -7.64098048e-01
2.93774217e-01 2.20813006e-01 -2.18959734e-01 5.93073428e-01
-7.70872682e-02 -8.16748619e-01 5.62726259e-01 2.98368216e-01
1.66763186e+00 4.97140229e-01 4.30184454e-01 -2.63855308e-01
3.35302353e-01 8.68390679e-01 2.95876861e-01 9.08949256e-01
8.34391937e-02 -2.81496644e-02 5.33843517e-01 -5.77785611e-01
-9.45133209e-01 -8.14643264e-01 4.83952135e-01 1.15417373e+00
-1.90319568e-01 -8.12631130e-01 -9.94442344e-01 -2.75311679e-01
-1.06761254e-01 1.47581172e+00 -2.52720118e-01 -2.77886808e-01
-3.84361982e-01 -2.07679555e-01 8.66124928e-01 3.84669781e-01
8.29205692e-01 -1.52634311e+00 -1.55436301e+00 1.86328813e-01
-4.47292984e-01 -1.02358079e+00 -5.58783524e-02 -8.04032758e-02
-5.83616138e-01 -1.23197722e+00 1.16522409e-01 -3.88870239e-01
3.69707882e-01 1.02733217e-01 1.48859608e+00 1.81289464e-01
-3.46368879e-01 6.29142761e-01 -3.73160064e-01 -4.20182377e-01
-7.64852166e-01 1.93378609e-02 -1.92989975e-01 -8.17388415e-01
2.67637789e-01 -5.66471457e-01 -2.06698012e-02 9.68092158e-02
-7.30824947e-01 8.79809856e-01 7.49526381e-01 6.15147591e-01
1.76944822e-01 5.42207584e-02 1.98923379e-01 -5.12708247e-01
8.58438313e-01 -2.20265225e-01 -7.36110330e-01 5.11001110e-01
-7.60665596e-01 3.46684843e-01 6.48036480e-01 -2.26785570e-01
-1.25143552e+00 8.85953605e-02 3.24004263e-01 -2.59704649e-01
-1.94075741e-02 6.82475030e-01 -1.67184025e-02 2.72343913e-03
9.55798507e-01 3.19723427e-01 7.26398751e-02 -1.81560948e-01
4.48169410e-01 5.57721376e-01 7.12452531e-01 -1.16817975e+00
5.90146184e-01 6.53784722e-02 -6.87749833e-02 -4.68960851e-01
-5.48506260e-01 1.40347704e-01 -9.95253474e-02 -1.88600823e-01
6.31623864e-01 -4.99634176e-01 -1.34642339e+00 3.03937197e-01
-1.22091198e+00 -9.54299331e-01 -1.98463410e-01 7.60202631e-02
-9.98042405e-01 -1.02587342e-01 -6.77578926e-01 -9.83057022e-01
-6.59284770e-01 -1.17229784e+00 9.15853143e-01 3.12356025e-01
-8.56281877e-01 -3.18108737e-01 -1.41095873e-02 8.51272464e-01
5.09615242e-01 3.95922698e-02 1.37174332e+00 -5.52391529e-01
-7.73363590e-01 -1.87797658e-02 -1.08350143e-01 -2.43124053e-01
-4.68916506e-01 1.38877273e-01 -6.54588580e-01 4.18114841e-01
-4.66600329e-01 -6.15202248e-01 1.69405088e-01 4.34954055e-02
8.30829263e-01 -3.75329405e-01 -2.63266861e-01 2.91438431e-01
1.12168860e+00 4.00036424e-01 4.54399586e-01 8.25157166e-01
2.64004529e-01 6.01417542e-01 7.06268966e-01 4.06525284e-01
6.38873041e-01 6.45879626e-01 5.66758215e-01 7.24147499e-01
-1.79694980e-01 -4.50166404e-01 4.33984190e-01 4.18462187e-01
-1.92296594e-01 -1.43505586e-02 -1.68667722e+00 4.21358973e-01
-2.31365323e+00 -8.92394543e-01 -4.81194146e-02 1.72585905e+00
1.16491342e+00 8.47908556e-02 -1.10454895e-01 -3.41168731e-01
2.08552718e-01 -3.28585297e-01 -8.91382337e-01 -6.04808807e-01
2.12130904e-01 1.50672361e-01 4.80876639e-02 5.60266376e-01
-4.79424298e-01 1.41401994e+00 6.19300079e+00 8.39946091e-01
-6.31371498e-01 -1.18473127e-01 -3.99404392e-02 -2.93841094e-01
-2.02811912e-01 1.61858514e-01 -3.25433910e-01 6.78226948e-02
1.17938292e+00 -5.17837703e-01 1.29583740e+00 1.07585073e+00
1.87522501e-01 -5.76214015e-01 -1.13296008e+00 7.76659966e-01
5.07969558e-02 -1.50743616e+00 6.38966337e-02 -3.04226249e-01
3.32574755e-01 -1.38917686e-02 -3.73570025e-01 6.91540837e-01
8.86501491e-01 -1.06819236e+00 1.05392528e+00 6.63376927e-01
3.57952118e-01 -7.02488422e-01 3.95476967e-01 8.97109807e-01
-6.52126014e-01 -1.50106803e-01 1.41953006e-01 -5.59803784e-01
3.37779894e-02 1.52986243e-01 -1.08473039e+00 4.90591139e-01
7.96899140e-01 6.48176819e-02 -4.85711098e-01 6.66981876e-01
-5.89217067e-01 2.34116822e-01 -4.30288136e-01 -3.06260318e-01
4.20125164e-02 -4.24901396e-03 4.95491743e-01 9.52426732e-01
-1.56304296e-02 3.44081819e-01 2.53687799e-01 1.19566417e+00
1.46805808e-01 -2.21567631e-01 -3.73738229e-01 -3.76109838e-01
7.08762169e-01 1.30159128e+00 -4.81777430e-01 -5.48398972e-01
-4.58106659e-02 8.32400322e-01 4.74266142e-01 1.48089498e-01
-9.40911293e-01 -4.27736044e-01 4.95836109e-01 -3.23462561e-02
-2.68846810e-01 -4.06862170e-01 -3.37731749e-01 -9.47376430e-01
-6.78662285e-02 -1.51746213e+00 1.38382509e-01 -1.66923833e+00
-8.64129186e-01 6.22788072e-01 4.22863185e-01 -4.37388599e-01
-6.79473758e-01 -4.81812179e-01 -3.38908732e-01 7.01389313e-01
-1.11065292e+00 -1.24892581e+00 -5.38721740e-01 3.65932077e-01
7.52142727e-01 -2.18656212e-01 1.02760851e+00 -3.12445402e-01
-3.91313493e-01 -3.67959812e-02 -5.99819839e-01 -2.17776507e-01
3.26778561e-01 -1.07614470e+00 5.39584517e-01 5.52454352e-01
-2.52256453e-01 8.11100960e-01 8.22809398e-01 -8.43117595e-01
-2.18507695e+00 -6.46518767e-01 8.13984990e-01 -7.21174359e-01
9.56689060e-01 -1.88359961e-01 -4.84537512e-01 7.51745880e-01
2.64106631e-01 -5.36571860e-01 1.61976337e-01 -6.11127838e-02
-4.87269312e-01 2.02485055e-01 -1.20142531e+00 1.02899289e+00
1.39496517e+00 -5.47456741e-01 -1.02879500e+00 6.26821458e-01
7.79703438e-01 -7.65400946e-01 -6.37502730e-01 2.06969380e-02
7.30156839e-01 -9.89034176e-01 7.76693344e-01 -5.83207309e-01
8.99986386e-01 -3.59114200e-01 -1.63151652e-01 -1.23607767e+00
-1.57536060e-01 -8.66272628e-01 -2.37344071e-01 9.37596619e-01
3.80530506e-01 -7.73099840e-01 4.61318672e-01 1.16416621e+00
-2.27815479e-01 -7.58667767e-01 -5.35232008e-01 -6.83377862e-01
-2.79494524e-01 -7.39777625e-01 6.27679944e-01 5.66858590e-01
4.36056226e-01 4.72942740e-01 1.71263888e-01 5.99524304e-02
3.88579309e-01 2.82709897e-01 8.49609554e-01 -1.05577314e+00
-5.54868996e-01 -5.60270250e-01 2.66456991e-01 -4.92634296e-01
4.14326131e-01 -8.76853585e-01 2.33008653e-01 -1.90420115e+00
2.57557482e-01 -2.46600240e-01 3.55499327e-01 1.17713332e+00
3.19996178e-01 -3.75376999e-01 5.22773504e-01 3.10025513e-01
-7.31691957e-01 2.49524876e-01 1.15849340e+00 -2.33885139e-01
-3.07299286e-01 -5.54980457e-01 -7.47223556e-01 9.42496777e-01
9.14066315e-01 -1.92356139e-01 -4.72476780e-01 -5.91262579e-01
7.42043436e-01 3.02390963e-01 5.39045453e-01 -1.12238443e+00
5.86229801e-01 -6.49144769e-01 -6.66864812e-02 -2.73198783e-01
3.55219305e-01 -6.81292415e-01 4.55015182e-01 6.08805776e-01
-3.91838461e-01 2.21713185e-01 5.02141297e-01 6.03424087e-02
2.24479750e-01 -3.78793001e-01 1.98763669e-01 -6.50888979e-01
-1.07630408e+00 -4.37100112e-01 -7.25987911e-01 7.72478953e-02
1.07444310e+00 3.10061164e-02 -8.61927211e-01 -2.64277399e-01
-1.96526885e-01 4.59580243e-01 6.60322607e-01 4.28763866e-01
3.62225652e-01 -7.25176632e-01 -3.22645664e-01 -7.38336295e-02
1.64060980e-01 3.44555564e-02 -1.65944695e-02 6.91115260e-01
-8.80937219e-01 6.38850808e-01 -4.83221382e-01 -1.10796250e-01
-1.03391862e+00 5.95653892e-01 3.73021424e-01 -3.13939571e-01
-5.05222082e-01 6.44063711e-01 -3.35424364e-01 -7.82456875e-01
-1.29362166e-01 -4.68901068e-01 2.44501874e-01 -3.24490607e-01
5.12413085e-01 4.92079496e-01 4.23628502e-02 2.45539516e-01
-7.58326888e-01 3.46883595e-01 3.34048539e-01 -5.19913554e-01
1.32897568e+00 4.69356835e-01 -4.24506992e-01 2.76608765e-01
1.48322254e-01 -1.70458287e-01 -1.04385757e+00 1.57547846e-01
1.49514750e-01 -4.29954752e-02 1.57110482e-01 -1.64593244e+00
-4.49986607e-01 4.00139242e-01 -2.54600435e-01 7.18993694e-03
1.03972661e+00 6.92271814e-02 6.44824862e-01 9.15800095e-01
1.04433084e+00 -1.12770772e+00 -8.31457786e-03 7.03732610e-01
1.15076959e+00 -9.15248752e-01 6.21994734e-02 -2.47442722e-01
-8.10026765e-01 1.20434260e+00 8.63990188e-01 1.49179131e-01
-2.92084754e-01 4.73104179e-01 -9.84967947e-02 -3.49639565e-01
-1.43899989e+00 -1.69532094e-02 -1.27080560e-01 6.19027913e-01
8.18408430e-02 1.52584225e-01 -1.62384599e-01 6.74025893e-01
-5.14997303e-01 5.33912301e-01 4.42700088e-01 1.34904349e+00
-4.32158649e-01 -8.65115285e-01 -5.20228922e-01 2.27236554e-01
1.85148790e-01 -2.12331206e-01 -6.74533963e-01 9.17217433e-01
-1.73970059e-01 1.14315927e+00 -1.50509387e-01 -3.52045059e-01
3.71417761e-01 3.88253033e-01 7.84317970e-01 -5.64355254e-01
-8.08211505e-01 -5.80804706e-01 6.83084190e-01 -9.95948315e-01
3.11009772e-02 -5.98511815e-01 -1.72749543e+00 -6.11986756e-01
3.19005758e-01 6.11328073e-02 7.70437658e-01 1.01673257e+00
7.47781277e-01 6.98445916e-01 -4.78762686e-01 -8.99030089e-01
-5.84607005e-01 -6.66547418e-01 1.63374767e-01 -8.16279650e-02
-2.74529994e-01 -5.00396729e-01 2.51990743e-02 -1.31459059e-02] | [9.010852813720703, 7.192659378051758] |
d23077b5-801a-4cc3-80a7-75478aa5b750 | a-layer-based-sequential-framework-for-scene | 1902.00671 | null | http://arxiv.org/abs/1902.00671v1 | http://arxiv.org/pdf/1902.00671v1.pdf | A Layer-Based Sequential Framework for Scene Generation with GANs | The visual world we sense, interpret and interact everyday is a complex
composition of interleaved physical entities. Therefore, it is a very
challenging task to generate vivid scenes of similar complexity using
computers. In this work, we present a scene generation framework based on
Generative Adversarial Networks (GANs) to sequentially compose a scene,
breaking down the underlying problem into smaller ones. Different than the
existing approaches, our framework offers an explicit control over the elements
of a scene through separate background and foreground generators. Starting with
an initially generated background, foreground objects then populate the scene
one-by-one in a sequential manner. Via quantitative and qualitative experiments
on a subset of the MS-COCO dataset, we show that our proposed framework
produces not only more diverse images but also copes better with affine
transformations and occlusion artifacts of foreground objects than its
counterparts. | ['Mehmet Ozgur Turkoglu', 'Berkay Kicanaoglu', 'William Thong', 'Luuk Spreeuwers'] | 2019-02-02 | null | null | null | null | ['scene-generation'] | ['computer-vision'] | [ 6.20210171e-01 1.32479191e-01 5.13973594e-01 -2.06036232e-02
-3.12772930e-01 -9.16384399e-01 1.00287771e+00 -3.20017576e-01
4.61010747e-02 8.31659257e-01 1.17528357e-01 -1.91102698e-01
3.82736832e-01 -1.07686937e+00 -8.56702924e-01 -6.70469522e-01
2.47164994e-01 2.31267422e-01 1.76587269e-01 -2.46542946e-01
-9.37259048e-02 5.54954946e-01 -1.40062261e+00 4.50008720e-01
1.00599194e+00 6.33999527e-01 1.68553486e-01 7.71640718e-01
-1.05768993e-01 1.23019123e+00 -9.14964080e-01 -7.19726741e-01
3.61293226e-01 -8.58261883e-01 -5.98983169e-01 5.50699413e-01
4.37915653e-01 -2.09956124e-01 -4.01948541e-01 9.74548876e-01
3.03029299e-01 1.64394021e-01 6.33181512e-01 -1.44160974e+00
-7.11775303e-01 5.19727051e-01 -6.63993895e-01 -8.74599889e-02
5.20954490e-01 5.27598262e-01 5.88484347e-01 -7.44275033e-01
8.49340498e-01 1.31116438e+00 3.46132725e-01 5.92149615e-01
-1.68205249e+00 -6.07086957e-01 3.81662279e-01 -1.70076847e-01
-1.31507182e+00 -3.03754270e-01 1.03907466e+00 -4.10152048e-01
2.78769702e-01 5.96573591e-01 8.94630134e-01 1.53510678e+00
-8.50667711e-04 6.45215690e-01 1.36970747e+00 -3.05949926e-01
3.60948086e-01 7.22677559e-02 -3.54753047e-01 4.12472188e-01
3.33453655e-01 -2.43902709e-02 -2.44822204e-01 7.04849064e-02
9.93343771e-01 -9.94237959e-02 -2.82321244e-01 -5.05657256e-01
-1.39865839e+00 5.85424304e-01 5.00842154e-01 1.57394215e-01
-4.42311734e-01 2.90479243e-01 -7.86403641e-02 -1.60123006e-01
1.58803061e-01 4.73178893e-01 2.95174956e-01 1.65500909e-01
-8.89823794e-01 5.11471331e-01 7.85691798e-01 1.04345489e+00
6.06622219e-01 3.88373345e-01 -5.22813380e-01 4.07718897e-01
-8.84103775e-02 6.52988613e-01 6.33750930e-02 -7.96299696e-01
3.63176614e-01 4.63350296e-01 2.86160707e-01 -1.32484138e+00
1.78900063e-02 -2.69761115e-01 -1.16146350e+00 5.04946828e-01
3.40244889e-01 -3.55910420e-01 -1.19746864e+00 1.90919125e+00
4.97756064e-01 4.25462335e-01 2.25946978e-02 8.35904956e-01
9.26663935e-01 8.62906456e-01 4.16554004e-01 1.09182328e-01
1.24336231e+00 -1.05752087e+00 -8.70668709e-01 -3.60472172e-01
-2.94621259e-01 -7.49527216e-01 1.08667195e+00 3.39180738e-01
-1.34703469e+00 -6.88311636e-01 -1.11183536e+00 -2.19731033e-02
-2.38881648e-01 -5.04771508e-02 7.66775906e-01 6.81400955e-01
-9.40243840e-01 2.94995099e-01 -5.90767205e-01 -3.39986160e-02
5.76365948e-01 -1.37541503e-01 -2.71360785e-01 3.71279791e-02
-8.96005809e-01 7.15304196e-01 4.74709094e-01 9.47171897e-02
-1.23816752e+00 -6.11860454e-01 -9.78590608e-01 -1.95020717e-02
4.01294738e-01 -9.96326566e-01 9.80319738e-01 -1.36401403e+00
-1.63926780e+00 8.04552913e-01 -2.48688143e-02 -2.14536548e-01
9.05621290e-01 -1.97118908e-01 -2.75921017e-01 -6.85751736e-02
-1.03374138e-01 6.58963382e-01 1.02660692e+00 -2.01140380e+00
-3.59910637e-01 1.40254170e-01 4.65374589e-01 3.08670700e-01
1.68454200e-01 -1.27931103e-01 -5.60056388e-01 -9.64426935e-01
-2.41332144e-01 -8.68721843e-01 -5.54144740e-01 -7.00559095e-02
-9.16611850e-01 5.20225942e-01 8.00666213e-01 -6.36811256e-01
8.90510976e-01 -2.07055545e+00 4.77183014e-01 1.07242092e-01
5.88139296e-01 1.75017014e-01 -1.92474872e-01 4.43554878e-01
-1.64894089e-01 1.64656803e-01 -2.91808635e-01 -4.88057017e-01
1.10607192e-01 1.95774943e-01 -4.05438423e-01 2.85001665e-01
3.89092743e-01 1.10952771e+00 -1.04214668e+00 -4.34905082e-01
4.99098569e-01 6.80948913e-01 -5.30326366e-01 4.48066533e-01
-3.74812603e-01 1.00772059e+00 -3.39103818e-01 4.36609179e-01
8.68775547e-01 -2.82685339e-01 2.05943689e-01 -2.07229123e-01
1.21668488e-01 -2.15727404e-01 -1.38368881e+00 1.59543681e+00
-4.44777191e-01 5.96656144e-01 7.54828230e-02 -7.91005433e-01
7.45534539e-01 1.40681654e-01 1.58303112e-01 -6.20934844e-01
1.34034097e-01 -2.98805118e-01 -5.92797324e-02 -3.54842752e-01
4.96194571e-01 -2.82381177e-01 -2.83845514e-01 3.26432824e-01
-1.35317191e-01 -4.17200536e-01 2.73034602e-01 2.88433492e-01
1.16751575e+00 2.95440465e-01 3.26492339e-01 3.40623483e-02
5.16092241e-01 -7.23496079e-02 4.97937709e-01 7.35433519e-01
1.60832554e-01 1.04359472e+00 5.52993000e-01 -4.32855606e-01
-1.30960047e+00 -1.54518390e+00 2.28365004e-01 4.83870745e-01
5.42218149e-01 -2.10476562e-01 -9.38879907e-01 -4.43300396e-01
-2.30575874e-01 9.03053880e-01 -7.71061599e-01 -5.63549921e-02
-5.31396925e-01 -7.26830482e-01 4.13702369e-01 3.11139703e-01
6.72837436e-01 -1.49219429e+00 -7.63477623e-01 7.34767616e-02
-2.56799787e-01 -1.39845800e+00 -3.55885893e-01 -2.72597998e-01
-3.11479181e-01 -7.63073683e-01 -7.95447290e-01 -5.84152997e-01
9.64464605e-01 2.75234431e-01 1.49044108e+00 -1.34966642e-01
-5.03490329e-01 1.81732938e-01 -3.84903073e-01 -4.45631027e-01
-8.33032608e-01 -3.01402301e-01 -2.98931092e-01 5.45952439e-01
-4.16141570e-01 -8.27908158e-01 -7.70361483e-01 1.46488175e-01
-1.34254277e+00 9.63785112e-01 5.18617749e-01 5.67438602e-01
6.11112297e-01 2.22130641e-01 2.19729826e-01 -1.16455567e+00
4.41061825e-01 -4.92533982e-01 -5.88444948e-01 2.97617406e-01
2.02534169e-01 -1.56728283e-01 8.98083091e-01 -6.14939213e-01
-1.42315483e+00 1.02211066e-01 2.56838232e-01 -4.50265884e-01
-4.44462240e-01 -3.94061618e-02 -5.71941316e-01 5.82191274e-02
5.83052993e-01 2.75069386e-01 -4.58100528e-01 -2.17214882e-01
8.90491664e-01 1.49568155e-01 8.50212514e-01 -6.82109654e-01
1.45599985e+00 6.57578290e-01 5.27031720e-03 -6.23367846e-01
-5.85150123e-01 2.96786517e-01 -6.74670279e-01 -3.92921180e-01
1.20470536e+00 -8.05958331e-01 -5.17284811e-01 6.30718589e-01
-1.26864791e+00 -6.99659228e-01 -6.02181613e-01 -1.85105056e-01
-4.53459084e-01 1.00838631e-01 -3.21079791e-01 -6.99505985e-01
-1.21916473e-01 -1.01069736e+00 1.09278488e+00 4.65334296e-01
-2.75247544e-01 -6.97866023e-01 5.51002473e-02 3.46788108e-01
4.61279273e-01 1.17418575e+00 7.05226421e-01 -4.47063893e-02
-1.02157581e+00 -4.71013336e-04 -2.49712899e-01 1.74979314e-01
4.01353210e-01 3.10033441e-01 -8.98530722e-01 -1.56924069e-01
-1.27788931e-01 -9.72075015e-02 3.71278584e-01 -3.02973427e-02
1.18391192e+00 -5.35531044e-01 -1.88997775e-01 8.58895481e-01
1.56425166e+00 4.30420399e-01 1.09797263e+00 1.86689258e-01
1.11636710e+00 4.78705525e-01 1.84731439e-01 3.68846029e-01
2.23030582e-01 5.77014387e-01 3.04486334e-01 -7.24523544e-01
-3.74578655e-01 -5.01667142e-01 1.21273376e-01 3.38972539e-01
-1.42087579e-01 -6.07401967e-01 -6.94621146e-01 4.13658589e-01
-1.60168850e+00 -1.17327511e+00 -7.04972595e-02 1.88483310e+00
8.48627388e-01 -2.55804118e-02 5.11043184e-02 5.43681793e-02
7.33100474e-01 3.03175420e-01 -4.68700111e-01 -2.91105151e-01
-3.49912643e-01 4.56159621e-01 1.25405133e-01 2.28701100e-01
-1.05754530e+00 1.02914345e+00 6.27382565e+00 7.06508100e-01
-1.10760188e+00 -9.67634767e-02 9.98059750e-01 -1.52226567e-01
-4.98371691e-01 -2.69931220e-02 -2.31765285e-01 6.19985998e-01
3.94201726e-01 -2.32122928e-01 6.62010431e-01 6.60223901e-01
7.59000480e-02 -5.40967956e-02 -9.90587890e-01 9.22355711e-01
1.13690704e-01 -1.45700002e+00 4.07485247e-01 -1.58219993e-01
1.14007032e+00 -6.39049888e-01 2.13711157e-01 7.97982886e-02
6.58334672e-01 -1.45746112e+00 1.19896102e+00 5.16588390e-01
8.44656348e-01 -6.49619043e-01 2.57470906e-01 1.49857521e-01
-1.11549044e+00 2.49574408e-01 1.45056995e-03 -2.19011726e-03
4.70155239e-01 5.40167451e-01 -4.67054933e-01 6.15519583e-01
5.20934463e-01 2.57254273e-01 -7.74638593e-01 8.80929351e-01
-5.03628731e-01 2.85727203e-01 4.36762944e-02 2.02984676e-01
2.27417089e-02 -4.19681489e-01 5.71206450e-01 1.23570669e+00
1.82851046e-01 2.94646502e-01 1.69963896e-01 1.45219493e+00
-1.50530800e-01 -5.26348241e-02 -9.06887650e-01 9.84592885e-02
2.54941791e-01 1.51524091e+00 -9.63814199e-01 -4.75410432e-01
-3.48550260e-01 1.29401684e+00 2.71878153e-01 6.52087092e-01
-1.24316812e+00 -1.86039820e-01 5.26408374e-01 7.38592148e-02
3.01293790e-01 -1.58610255e-01 -4.30061460e-01 -1.36112738e+00
7.74961412e-02 -1.37649608e+00 -1.00410216e-01 -1.10256696e+00
-1.22915828e+00 8.77941906e-01 1.68052111e-02 -1.14974391e+00
-3.57328467e-02 -1.90755919e-01 -8.60875368e-01 9.05270517e-01
-1.04390037e+00 -1.43069351e+00 -8.82844865e-01 5.81632137e-01
4.25739050e-01 2.88963765e-01 6.96293652e-01 1.25086874e-01
-6.76344872e-01 3.34158123e-01 -1.65216297e-01 2.49317214e-01
2.34245211e-01 -1.35207176e+00 8.12158585e-01 1.23721087e+00
4.31050450e-01 3.95876646e-01 8.86430442e-01 -6.33784771e-01
-1.23429358e+00 -1.15771258e+00 1.45092115e-01 -4.73318487e-01
3.91907185e-01 -8.54731977e-01 -7.88693428e-01 6.53852880e-01
6.13832057e-01 -6.35886192e-02 3.36781979e-01 -5.05696416e-01
-1.95563331e-01 2.74477452e-02 -1.27135575e+00 1.15215838e+00
1.26304150e+00 -2.68451184e-01 -4.20026630e-01 1.17029287e-01
7.48980939e-01 -7.12309480e-01 -4.63001639e-01 4.71082807e-01
3.01090866e-01 -1.19562387e+00 1.23404109e+00 -4.53026086e-01
7.67898619e-01 -6.44760072e-01 5.23113348e-02 -1.35700297e+00
-2.66253322e-01 -9.82858241e-01 4.02778946e-02 1.41789114e+00
7.35982805e-02 -3.70189220e-01 7.30729640e-01 6.79545105e-01
1.40858784e-01 -3.22267890e-01 -5.53443491e-01 -4.91186976e-01
-1.26938954e-01 -2.00376078e-01 7.73440540e-01 8.37140560e-01
-6.09596848e-01 3.64082158e-01 -4.67245221e-01 2.16989622e-01
7.73934603e-01 3.68497938e-01 1.30406475e+00 -7.43171155e-01
-5.09811640e-01 -3.08250725e-01 -4.03930128e-01 -6.58495605e-01
-7.94769302e-02 -5.35121560e-01 1.36868477e-01 -1.42069840e+00
2.80644566e-01 -5.13103962e-01 -3.92383290e-03 2.11590454e-01
-5.06340981e-01 5.44042289e-01 5.41359961e-01 -2.18805104e-01
-5.03285229e-01 5.47237039e-01 1.63360608e+00 -1.88336417e-01
-1.46205306e-01 -3.43281686e-01 -7.15326071e-01 7.34915793e-01
5.56381464e-01 -1.62815303e-01 -6.18518770e-01 -2.77459741e-01
4.08948436e-02 1.52270496e-01 7.18873084e-01 -1.26204669e+00
-2.42066234e-01 -4.72529709e-01 7.16139674e-01 -3.94792825e-01
3.51503193e-01 -6.80424809e-01 1.02276158e+00 1.89162210e-01
-9.88619402e-02 3.06545831e-02 4.60907876e-01 5.39924860e-01
-2.72525437e-02 1.28976777e-01 8.36198211e-01 -2.57174671e-01
-5.29044151e-01 2.11139187e-01 -3.01699191e-01 2.55081654e-01
1.38062501e+00 -9.55885276e-02 -2.86147505e-01 -5.60172558e-01
-7.38023520e-01 -1.18968733e-01 8.40133309e-01 4.77274299e-01
3.78438950e-01 -1.49438953e+00 -6.64009571e-01 2.14684516e-01
-6.93924800e-02 3.55184853e-01 4.64850843e-01 1.78432524e-01
-9.04995024e-01 2.35838536e-02 -5.11571884e-01 -3.85970175e-01
-1.15821016e+00 8.27462554e-01 2.65723139e-01 -3.89235079e-01
-7.12649822e-01 7.20883548e-01 1.06787634e+00 -3.72706503e-02
-5.71649447e-02 -1.71668321e-01 -4.20471430e-02 -2.36952707e-01
5.51525831e-01 1.34052590e-01 -4.56096679e-01 -7.89292693e-01
-4.39783745e-02 3.85378391e-01 3.39902610e-01 -3.43467712e-01
1.21726501e+00 -7.59613216e-02 -7.48462975e-02 3.48883629e-01
7.63427019e-01 2.53575832e-01 -1.47815955e+00 4.12986875e-02
-5.05440116e-01 -6.78971827e-01 -4.31642681e-01 -8.41859877e-01
-1.29874885e+00 5.82437515e-01 3.16831499e-01 4.20533657e-01
1.29037440e+00 -8.63160044e-02 7.17720866e-01 -1.71781421e-01
3.75350505e-01 -5.05610168e-01 3.72615755e-01 -8.78307298e-02
1.11911249e+00 -9.45125937e-01 -4.94840518e-02 -6.71750247e-01
-8.83202612e-01 6.10519171e-01 7.41201103e-01 -3.58233750e-01
1.59226179e-01 3.74414533e-01 7.81370327e-02 -1.85904935e-01
-4.33946788e-01 -1.50410235e-01 4.31815296e-01 8.63500595e-01
8.41431245e-02 1.52485177e-01 7.85629228e-02 5.08801222e-01
-3.72476280e-01 -1.55538499e-01 5.90308905e-01 7.08481014e-01
2.55308211e-01 -1.12602389e+00 -5.87637126e-01 7.33795688e-02
-3.37307751e-01 -9.12300274e-02 -5.22668600e-01 9.67874467e-01
3.02162141e-01 8.80256712e-01 9.35559273e-02 -3.12257946e-01
3.66476297e-01 -1.76713496e-01 5.88983774e-01 -5.96769035e-01
-5.00318646e-01 -8.11244324e-02 2.00620983e-02 -6.47489607e-01
-3.77431422e-01 -6.02579057e-01 -9.61288869e-01 -3.94738585e-01
1.11486405e-01 -3.43145758e-01 3.55582625e-01 6.81569219e-01
3.39553893e-01 1.09919620e+00 6.71970308e-01 -1.16432703e+00
9.13415626e-02 -6.68381214e-01 -3.98492098e-01 1.01380658e+00
1.34468138e-01 -6.72377586e-01 -2.97718167e-01 4.99412924e-01] | [11.519639015197754, -0.47070378065109253] |
393c4d64-9d14-4a51-8832-fb44a724c0fd | deepnag-deep-non-adversarial-gesture | 2011.09149 | null | https://arxiv.org/abs/2011.09149v1 | https://arxiv.org/pdf/2011.09149v1.pdf | DeepNAG: Deep Non-Adversarial Gesture Generation | Synthetic data generation to improve classification performance (data augmentation) is a well-studied problem. Recently, generative adversarial networks (GAN) have shown superior image data augmentation performance, but their suitability in gesture synthesis has received inadequate attention. Further, GANs prohibitively require simultaneous generator and discriminator network training. We tackle both issues in this work. We first discuss a novel, device-agnostic GAN model for gesture synthesis called DeepGAN. Thereafter, we formulate DeepNAG by introducing a new differentiable loss function based on dynamic time warping and the average Hausdorff distance, which allows us to train DeepGAN's generator without requiring a discriminator. Through evaluations, we compare the utility of DeepGAN and DeepNAG against two alternative techniques for training five recognizers using data augmentation over six datasets. We further investigate the perceived quality of synthesized samples via an Amazon Mechanical Turk user study based on the HYPE benchmark. We find that DeepNAG outperforms DeepGAN in accuracy, training time (up to 17x faster), and realism, thereby opening the door to a new line of research in generator network design and training for gesture synthesis. Our source code is available at https://www.deepnag.com. | ['Joseph J. LaViola Jr', 'Eugene M. Taranta II', 'Mehran Maghoumi'] | 2020-11-18 | null | null | null | null | ['gesture-generation'] | ['robots'] | [ 3.77617896e-01 2.08408043e-01 -7.95019493e-02 -1.99089631e-01
-8.74688089e-01 -9.32952106e-01 9.44673061e-01 -6.48942232e-01
-4.34924006e-01 8.33990455e-01 2.21020788e-01 -3.34417045e-01
2.77664572e-01 -8.26107442e-01 -7.16833711e-01 -8.23171079e-01
1.40881419e-01 2.32122749e-01 -4.38727796e-01 -2.15629727e-01
4.22790684e-02 5.40544748e-01 -1.16291058e+00 -4.75888327e-02
8.10973883e-01 8.24596226e-01 -4.63434637e-01 9.84477162e-01
4.61439669e-01 6.56752646e-01 -1.11399126e+00 -7.54180729e-01
6.45046771e-01 -1.02238274e+00 -6.17946327e-01 -2.53046416e-02
2.61820942e-01 -6.48170292e-01 -5.50441802e-01 6.46819055e-01
9.41622257e-01 2.48616904e-01 5.85258007e-01 -1.54320812e+00
-9.62246537e-01 4.37462181e-01 -2.14576110e-01 -2.74185687e-02
2.57260412e-01 7.18922913e-01 6.92025900e-01 -6.10724628e-01
6.15457773e-01 9.92599070e-01 6.28870368e-01 1.12463725e+00
-1.13022494e+00 -7.87206173e-01 -3.44109476e-01 -4.14998621e-01
-1.16239953e+00 -4.28795576e-01 6.62317574e-01 -2.33185038e-01
5.60769320e-01 5.28361380e-01 6.97986603e-01 1.82817793e+00
-4.14522551e-02 8.48420382e-01 1.34990323e+00 -2.75666356e-01
3.81397069e-01 -8.51341113e-02 -6.40501559e-01 5.78864157e-01
4.28986996e-02 5.89449406e-01 -4.02006656e-01 1.72884107e-01
1.33714247e+00 -3.17042381e-01 -1.25031397e-01 -3.77699584e-02
-1.10221756e+00 1.01720476e+00 5.48466444e-01 1.50754780e-01
-3.25794965e-01 5.36799610e-01 3.67560893e-01 5.12256026e-01
4.04315680e-01 8.42643201e-01 2.19745357e-02 -5.60717285e-01
-6.78690910e-01 5.14060080e-01 7.39947796e-01 9.70902443e-01
1.20791167e-01 5.30057073e-01 -2.47038245e-01 4.92564410e-01
-6.08077459e-02 6.38694823e-01 7.50248671e-01 -9.87616599e-01
4.26512122e-01 2.07308680e-01 -7.11199641e-02 -8.36766005e-01
-1.16812617e-01 -3.73713404e-01 -9.87069666e-01 4.79037672e-01
5.88822186e-01 -5.44521213e-01 -1.12172556e+00 1.80432081e+00
8.08489174e-02 2.17666820e-01 1.99182436e-01 1.08064294e+00
7.69545734e-01 4.82775271e-01 -1.63065270e-03 1.64755076e-01
9.50323582e-01 -1.04784131e+00 -5.61061561e-01 3.99224795e-02
5.01536489e-01 -5.90442419e-01 1.46763444e+00 3.47756326e-01
-1.28417122e+00 -4.97989595e-01 -1.11702633e+00 2.74114423e-02
-3.19293827e-01 7.31744664e-03 9.42928553e-01 1.17932582e+00
-1.02274895e+00 6.08018160e-01 -9.57953572e-01 -2.94096768e-01
5.56340754e-01 4.36967820e-01 -2.34579861e-01 1.33792922e-01
-1.05837846e+00 7.85248041e-01 1.67034209e-01 1.23476222e-01
-9.25228417e-01 -5.46001673e-01 -7.82124162e-01 -3.25401753e-01
3.99193354e-02 -7.72918522e-01 1.37484360e+00 -1.08186400e+00
-2.14947081e+00 5.77715218e-01 3.92608196e-01 -5.59108973e-01
1.06824172e+00 -1.82973817e-01 -2.92963684e-01 -3.11739091e-02
-1.59072071e-01 9.00338173e-01 8.48889887e-01 -1.16299820e+00
-1.14134140e-01 8.33289418e-03 2.16052026e-01 2.91523725e-01
-2.54940510e-01 -1.63675547e-01 -1.82628244e-01 -1.18136728e+00
-3.48090053e-01 -1.15029848e+00 -1.87145248e-01 -2.59194404e-01
-5.16797185e-01 8.84821936e-02 7.68210948e-01 -6.66277051e-01
8.58821809e-01 -2.08437347e+00 6.84500411e-02 2.63133913e-01
1.77130163e-01 4.66912329e-01 -4.91380215e-01 2.98522562e-01
1.30134374e-01 4.58104908e-01 -3.72868031e-01 -2.66138464e-01
1.04102038e-01 2.72380859e-01 -3.85855108e-01 2.81615764e-01
3.78840744e-01 1.45600736e+00 -9.78640020e-01 -5.99978007e-02
1.81969181e-01 6.69053495e-01 -6.42487645e-01 3.66004050e-01
-1.66044205e-01 9.41152155e-01 -2.21037909e-01 7.22820520e-01
3.64089221e-01 -7.88016468e-02 -4.03310880e-02 1.43804029e-01
1.91921100e-01 7.77603313e-02 -6.73097074e-01 1.73971248e+00
-5.32359242e-01 9.74423409e-01 -4.80897129e-01 -6.48212552e-01
1.01347554e+00 3.63937318e-01 2.40655378e-01 -8.39150429e-01
3.87378752e-01 2.47156024e-01 7.91322663e-02 -4.46778834e-01
4.00023550e-01 -2.93581158e-01 -2.22662538e-01 6.53334260e-01
-1.01100557e-01 -4.20672119e-01 1.44595858e-02 -1.45335585e-01
1.23729551e+00 3.54168802e-01 -1.27714455e-01 1.80269778e-01
-3.62121165e-02 -1.12527609e-01 6.76196814e-02 7.50586629e-01
-2.73514763e-02 7.90032506e-01 4.31152016e-01 -3.45165342e-01
-1.33910012e+00 -1.09080756e+00 3.40770304e-01 9.60650742e-01
-6.02421351e-02 -1.37740999e-01 -1.02349007e+00 -8.26367795e-01
-1.25400886e-01 6.69452965e-01 -9.77449775e-01 -2.68883526e-01
-8.40845585e-01 -7.27484882e-01 1.19811130e+00 8.65774751e-01
8.26849043e-01 -1.25715029e+00 -7.66053855e-01 -7.07892627e-02
-1.15504272e-01 -9.48821366e-01 -5.39713502e-01 -1.21398360e-01
-8.57349873e-01 -7.45219827e-01 -1.02665627e+00 -5.62259793e-01
7.54005551e-01 -2.35662013e-01 9.23432767e-01 6.03273883e-02
-1.72199398e-01 3.80589426e-01 -5.86168587e-01 -5.66173077e-01
-6.52657449e-01 3.63935977e-01 2.33320873e-02 -2.95508534e-01
1.29040226e-01 -6.90036595e-01 -7.95269012e-01 4.90060151e-01
-1.06841314e+00 1.91403985e-01 6.95656180e-01 9.33664858e-01
2.59915859e-01 -2.88704187e-01 5.97303987e-01 -6.73958063e-01
1.01328242e+00 -2.04037264e-01 -4.85268116e-01 -9.49628875e-02
-5.24918377e-01 8.37079138e-02 6.08393192e-01 -7.28704274e-01
-7.99096942e-01 -1.17239766e-01 -2.94186473e-01 -3.96074384e-01
-2.95944232e-02 3.35435152e-01 -1.32373944e-01 -3.26695412e-01
8.61512899e-01 1.61187768e-01 3.01981956e-01 -1.94621086e-01
5.76433182e-01 4.86511827e-01 7.01652944e-01 -4.81364757e-01
9.90285337e-01 3.17420065e-01 1.23206722e-02 -6.00291967e-01
-3.43834519e-01 3.66523743e-01 -4.61863726e-01 -1.19939826e-01
7.81195700e-01 -6.45958126e-01 -7.63004899e-01 5.75752854e-01
-8.97406459e-01 -1.01728690e+00 -7.77989089e-01 4.59810972e-01
-8.15467596e-01 -8.93674865e-02 -7.13048100e-01 -6.15223765e-01
-4.49831545e-01 -1.05793941e+00 9.61513579e-01 2.08301559e-01
-3.34043771e-01 -1.06331933e+00 1.27216980e-01 2.45026171e-01
6.68107033e-01 1.03438616e+00 2.91444868e-01 -6.07870936e-01
-5.08529365e-01 -4.07075375e-01 -2.67983023e-02 5.37666917e-01
2.41595939e-01 -1.30241960e-01 -9.57347631e-01 -3.81440490e-01
-8.18831921e-02 -4.97276217e-01 5.30878067e-01 1.59323409e-01
1.14163208e+00 -5.65838635e-01 1.55271217e-01 9.27522123e-01
1.09050322e+00 4.01291370e-01 1.00876009e+00 4.05943304e-01
9.28374946e-01 3.24579567e-01 3.03117424e-01 3.54902625e-01
1.03890866e-01 6.08845115e-01 3.73215586e-01 -3.50111753e-01
-3.28761667e-01 -4.72290844e-01 3.80957723e-01 5.80320895e-01
-6.33756280e-01 -5.88607252e-01 -6.50652826e-01 2.09056377e-01
-1.45667446e+00 -8.72906148e-01 2.85014451e-01 1.92393720e+00
7.38946617e-01 -1.27707511e-01 4.77224499e-01 2.26778671e-01
4.12336588e-01 -2.82273684e-02 -7.18641877e-01 -4.86935496e-01
-1.41020671e-01 8.23442340e-01 6.23306453e-01 2.82021791e-01
-1.00927079e+00 1.04947329e+00 6.48879766e+00 6.09099567e-01
-1.23325872e+00 7.40211606e-02 7.78123677e-01 -1.69042230e-01
-2.45212719e-01 -3.32676381e-01 -1.77538887e-01 5.07521152e-01
9.08582926e-01 -4.43256535e-02 6.22751892e-01 6.93449557e-01
9.38985944e-02 2.04807892e-01 -1.06376767e+00 1.00221550e+00
1.20005935e-01 -1.31069183e+00 1.72862574e-01 4.05778468e-01
9.34443891e-01 -3.56887996e-01 5.60183764e-01 3.31307262e-01
3.82612526e-01 -1.50917482e+00 5.70122778e-01 3.26233357e-01
1.24371111e+00 -7.90684938e-01 7.84727275e-01 -3.68691981e-02
-8.44978690e-01 1.93514198e-01 1.05688579e-01 -1.10605218e-01
5.57908043e-02 -8.88156667e-02 -8.53438675e-01 3.86147976e-01
2.25174353e-01 2.83874273e-01 -5.42450488e-01 6.53805315e-01
-4.59889203e-01 7.68600166e-01 -1.01275496e-01 -1.03181906e-01
3.39845955e-01 -1.04050197e-01 4.07957166e-01 1.03312027e+00
4.20192182e-01 3.09814066e-01 -2.57358491e-01 8.26054931e-01
-4.49073374e-01 -1.75401062e-01 -6.29645824e-01 -4.23013002e-01
2.92494863e-01 8.69302809e-01 -7.31406033e-01 -2.98515439e-01
8.10447410e-02 1.43554735e+00 -1.66411638e-01 3.49685937e-01
-1.05017364e+00 -5.76338887e-01 5.95799804e-01 -3.98471653e-02
-8.54079425e-02 -4.09912497e-01 -6.40363872e-01 -1.06654465e+00
3.51554863e-02 -1.21740818e+00 1.15641333e-01 -6.24508977e-01
-1.08187699e+00 7.86543965e-01 -1.65782467e-01 -1.32346511e+00
-7.19200671e-01 -5.54427445e-01 -6.20991588e-01 8.49202514e-01
-1.09503424e+00 -1.20827937e+00 -7.07250476e-01 5.15906096e-01
5.32348990e-01 -2.88928598e-01 9.23605740e-01 2.84153283e-01
-4.03014064e-01 1.21401513e+00 -1.88769698e-01 5.37177265e-01
4.63459402e-01 -1.17743754e+00 1.11278403e+00 8.66165102e-01
1.21318758e-01 3.66358042e-01 6.60574377e-01 -6.11160517e-01
-1.42939997e+00 -1.08735955e+00 3.46979260e-01 -6.61250889e-01
4.75798368e-01 -3.81905556e-01 -5.72254896e-01 7.28985250e-01
2.57297903e-01 -1.09491967e-01 6.75772130e-01 -4.84972447e-01
-2.69084811e-01 2.68291235e-01 -1.42222106e+00 8.15559626e-01
1.30462396e+00 -3.97023469e-01 -7.27166831e-02 2.17377007e-01
6.49033785e-01 -7.97109306e-01 -8.57372522e-01 3.79201353e-01
7.85891712e-01 -8.99728537e-01 7.41341770e-01 -6.74960494e-01
6.88591957e-01 5.48888855e-02 -3.63451205e-02 -1.52564037e+00
4.78081442e-02 -1.07240736e+00 -2.24126711e-01 1.04570711e+00
4.04690236e-01 -7.77749419e-01 1.19201338e+00 7.05384612e-01
-1.10066906e-02 -7.64444888e-01 -5.39221704e-01 -1.09747517e+00
2.39430860e-01 -2.79123306e-01 8.59758794e-01 9.57880378e-01
-2.03598961e-01 3.17469276e-02 -6.56651080e-01 -2.07417309e-01
3.60921234e-01 -3.24038327e-01 1.26594806e+00 -7.53655553e-01
-4.02610987e-01 -5.55397511e-01 -5.43318391e-01 -1.02810585e+00
2.13858411e-02 -8.26774418e-01 5.69265634e-02 -1.33557880e+00
-2.60249496e-01 -5.17681599e-01 2.24652335e-01 5.70824564e-01
7.33107552e-02 8.89756560e-01 2.85014987e-01 9.02485922e-02
7.28005618e-02 6.08359456e-01 1.56552923e+00 8.97738896e-03
-4.08804983e-01 1.08889639e-02 -7.36798227e-01 4.23885047e-01
1.29432142e+00 -1.23715632e-01 -5.04484355e-01 -5.96116960e-01
-6.08887710e-02 -2.40237024e-02 6.23294771e-01 -1.17478919e+00
-1.68438748e-01 -1.53649122e-01 5.40371895e-01 1.93265736e-01
4.25284177e-01 -5.12094319e-01 2.78208017e-01 5.44138491e-01
-4.91564810e-01 3.65981042e-01 2.09734455e-01 1.76099792e-01
-2.60895751e-02 1.76778242e-01 6.54045403e-01 4.74612266e-02
-3.25814456e-01 3.97938758e-01 -1.86005592e-01 9.12984088e-02
9.77213919e-01 -3.20234448e-01 -2.50794888e-01 -7.55568027e-01
-6.31095529e-01 -1.94419384e-01 4.55016166e-01 5.34946024e-01
8.45569491e-01 -1.62331069e+00 -7.69271910e-01 3.78236681e-01
-1.81800067e-01 -9.99180600e-02 -6.45431131e-02 5.76037228e-01
-7.83128619e-01 6.65736049e-02 -4.58060801e-01 -3.19787651e-01
-1.00765586e+00 3.22977155e-01 2.90006757e-01 -3.74669768e-02
-6.04582608e-01 9.74600673e-01 -1.68610305e-01 -3.35237920e-01
2.05632165e-01 -3.12684596e-01 3.83501858e-01 -2.61386424e-01
3.98328692e-01 5.46451390e-01 -1.54094204e-01 -4.34188724e-01
8.05074442e-03 2.80484617e-01 2.83401608e-01 -6.30858719e-01
1.19113791e+00 3.75558883e-01 3.89609158e-01 1.22389629e-01
1.10753727e+00 -2.87118796e-02 -1.53404415e+00 2.03275889e-01
-3.07367980e-01 -6.01725936e-01 -3.88988793e-01 -9.43198681e-01
-1.22644365e+00 7.02607810e-01 6.31702542e-01 2.89893210e-01
1.20007443e+00 -2.92100102e-01 1.00438440e+00 8.07313696e-02
2.14081064e-01 -7.55910397e-01 5.09534836e-01 3.09978992e-01
1.03999615e+00 -1.14626050e+00 -3.57351631e-01 -7.82774314e-02
-8.63424122e-01 8.88174891e-01 6.46675766e-01 -4.59462762e-01
2.07564309e-01 4.24749643e-01 3.54192317e-01 -5.07285297e-02
-2.49319583e-01 -3.79445851e-02 3.46864671e-01 7.68451452e-01
4.21708792e-01 2.82708853e-01 -2.07690343e-01 2.84373283e-01
-9.50647831e-01 1.31294914e-02 4.72560018e-01 9.24747050e-01
2.45004267e-01 -1.29267371e+00 -2.09197626e-01 3.36348623e-01
-2.97364503e-01 -4.80652153e-02 -5.64441442e-01 1.09928393e+00
2.13221796e-02 8.59887719e-01 1.21561565e-01 -8.20651531e-01
2.72438198e-01 -6.23802356e-02 7.51528144e-01 -2.99026430e-01
-6.99931800e-01 -2.72236168e-01 -1.28905280e-02 -4.57567424e-01
-4.18903410e-01 -5.53696752e-01 -9.34133470e-01 -7.24470794e-01
-1.49128884e-01 2.10702531e-02 8.39766145e-01 7.54442632e-01
2.70908475e-01 5.09050429e-01 7.14467704e-01 -9.93248999e-01
-4.71365184e-01 -9.64544177e-01 -2.84001499e-01 4.56633329e-01
2.23692313e-01 -3.73703808e-01 -3.13780785e-01 1.73225626e-01] | [11.639032363891602, -0.2714634835720062] |
b3b7b4a9-3a11-4aa0-a36f-a6500cda8f2a | pwoc-3d-deep-occlusion-aware-end-to-end-scene | 1904.06116 | null | http://arxiv.org/abs/1904.06116v1 | http://arxiv.org/pdf/1904.06116v1.pdf | PWOC-3D: Deep Occlusion-Aware End-to-End Scene Flow Estimation | In the last few years, convolutional neural networks (CNNs) have demonstrated
increasing success at learning many computer vision tasks including dense
estimation problems such as optical flow and stereo matching. However, the
joint prediction of these tasks, called scene flow, has traditionally been
tackled using slow classical methods based on primitive assumptions which fail
to generalize. The work presented in this paper overcomes these drawbacks
efficiently (in terms of speed and accuracy) by proposing PWOC-3D, a compact
CNN architecture to predict scene flow from stereo image sequences in an
end-to-end supervised setting. Further, large motion and occlusions are
well-known problems in scene flow estimation. PWOC-3D employs specialized
design decisions to explicitly model these challenges. In this regard, we
propose a novel self-supervised strategy to predict occlusions from images
(learned without any labeled occlusion data). Leveraging several such
constructs, our network achieves competitive results on the KITTI benchmark and
the challenging FlyingThings3D dataset. Especially on KITTI, PWOC-3D achieves
the second place among end-to-end deep learning methods with 48 times fewer
parameters than the top-performing method. | ['Oliver Wasenmüller', 'René Schuster', 'Rohan Saxena', 'Didier Stricker'] | 2019-04-12 | null | null | null | null | ['stereo-matching', 'scene-flow-estimation'] | ['computer-vision', 'computer-vision'] | [-9.50228646e-02 -2.98768818e-01 -2.43553832e-01 -4.60678935e-01
-2.72667319e-01 -2.40557730e-01 5.25400400e-01 -1.76740676e-01
-6.08159423e-01 6.85416698e-01 2.37224728e-01 -1.38709828e-01
3.89459878e-02 -5.27857006e-01 -8.36747050e-01 -4.82390732e-01
-1.09404489e-01 3.07456404e-01 1.97632253e-01 -8.82637277e-02
2.59426624e-01 3.01871538e-01 -1.60476649e+00 -9.12867188e-02
8.80576253e-01 1.06985497e+00 3.34591448e-01 7.39700019e-01
-1.41996657e-02 1.32869697e+00 -9.05923322e-02 -2.42926493e-01
5.58661759e-01 -1.91494152e-01 -1.01569736e+00 4.61487994e-02
1.14690340e+00 -7.67349422e-01 -8.52895737e-01 7.11472809e-01
5.15137017e-01 2.92580903e-01 2.54491061e-01 -1.32883692e+00
-2.73391396e-01 6.72830939e-02 -5.51716030e-01 4.30751264e-01
1.83793172e-01 5.53013027e-01 9.50613141e-01 -7.81317413e-01
6.53453112e-01 1.24671042e+00 7.36813426e-01 5.10049045e-01
-1.22056818e+00 -5.21757782e-01 1.99925885e-01 4.26316470e-01
-1.01818490e+00 -4.30386126e-01 7.59427249e-01 -5.77466130e-01
9.86910522e-01 -1.93258613e-01 7.19320118e-01 1.02489007e+00
1.24896571e-01 1.11725271e+00 8.61033142e-01 4.50402200e-02
1.63481861e-01 -3.27516675e-01 3.86851989e-02 9.58376229e-01
1.38529569e-01 2.39401132e-01 -6.94377840e-01 2.89119691e-01
6.70104146e-01 -5.04257381e-02 -5.37353218e-01 -6.83800817e-01
-1.35911036e+00 7.57978618e-01 8.08327198e-01 -7.26908594e-02
-4.96444553e-02 3.95876020e-01 5.72900712e-01 1.51797876e-01
6.73336744e-01 3.86153132e-01 -5.48388541e-01 -3.12146991e-01
-8.42693686e-01 5.38882434e-01 8.44711304e-01 9.15123403e-01
1.02837026e+00 2.18875706e-01 -8.35276544e-02 6.29013538e-01
1.91765115e-01 3.40479136e-01 3.12594742e-01 -1.21293080e+00
6.98737442e-01 4.05852705e-01 2.20282227e-01 -1.20084381e+00
-5.04455864e-01 -6.22973084e-01 -1.07931066e+00 2.11536020e-01
6.75986409e-01 -1.48199245e-01 -7.78306305e-01 1.86874545e+00
4.77255791e-01 6.53386950e-01 -4.47548032e-02 1.27076054e+00
8.69330764e-01 7.54135847e-01 2.54294313e-02 1.73236772e-01
8.29122007e-01 -1.44133985e+00 -5.12080431e-01 -5.22690475e-01
7.38897681e-01 -6.16035938e-01 9.81760561e-01 2.19524786e-01
-9.21572149e-01 -8.61464202e-01 -1.08203697e+00 -6.29300296e-01
-1.42677069e-01 1.02561064e-01 9.54858959e-01 1.74417675e-01
-1.12893927e+00 7.14943647e-01 -9.25734997e-01 -2.79642045e-01
7.38518894e-01 3.48792881e-01 -4.87970471e-01 -3.51234078e-01
-8.80742192e-01 7.16430187e-01 3.01214486e-01 3.31494302e-01
-1.17442226e+00 -1.05510867e+00 -1.15041041e+00 -3.00853271e-02
3.12287927e-01 -1.21669877e+00 1.09285557e+00 -7.50578821e-01
-1.62478197e+00 9.31060672e-01 -1.41863525e-01 -7.39892125e-01
7.78783262e-01 -6.94233716e-01 2.09607333e-01 2.93065868e-02
1.36443615e-01 9.60732877e-01 8.00015748e-01 -7.42716491e-01
-6.97236717e-01 -2.04575643e-01 4.08071339e-01 2.26765364e-01
-1.69111609e-01 -4.06009942e-01 -3.87104958e-01 -2.72799730e-01
-1.04871131e-01 -9.28450942e-01 -4.75426376e-01 4.99240994e-01
-3.27818841e-01 -2.96208829e-01 9.06914771e-01 -2.95247138e-01
6.67067051e-01 -1.94504964e+00 3.76095444e-01 -5.63788950e-01
4.78715420e-01 5.74682832e-01 -1.16075136e-01 1.83597475e-01
1.44728184e-01 -4.43462878e-01 -4.39905405e-01 -8.80646884e-01
6.94240406e-02 3.18590850e-01 -4.62915897e-01 7.30512798e-01
3.54369253e-01 9.25199628e-01 -1.11617708e+00 -4.97101247e-01
7.57848680e-01 5.30767620e-01 -8.96749377e-01 4.76152986e-01
-2.18069762e-01 8.33474398e-01 -2.14596629e-01 2.17995733e-01
7.99658179e-01 -3.63599181e-01 -3.03566724e-01 -1.62902236e-01
-2.85185456e-01 3.40642601e-01 -9.15950596e-01 2.41496921e+00
-7.59369969e-01 1.07211125e+00 -7.09709451e-02 -1.21723390e+00
7.79330611e-01 -8.06752965e-02 5.88930130e-01 -5.78287244e-01
9.87360850e-02 1.53260767e-01 -2.42580354e-01 -7.00823009e-01
4.43170935e-01 1.75513282e-01 1.39434338e-01 1.10380560e-01
3.65455747e-01 -3.41465086e-01 3.52040619e-01 2.15376452e-01
9.98834431e-01 5.93442440e-01 2.20417067e-01 -4.33521360e-01
7.45225132e-01 1.60006419e-01 8.23016822e-01 7.50193536e-01
-5.21555960e-01 7.74507940e-01 3.67753863e-01 -1.10667121e+00
-1.01779556e+00 -6.73008621e-01 2.65297554e-02 7.19579875e-01
5.28912365e-01 -4.65180725e-01 -4.98176485e-01 -6.53019905e-01
8.89591798e-02 9.77049693e-02 -5.78499138e-01 -4.42427658e-02
-7.94498920e-01 -5.33567667e-01 3.55888903e-01 4.29037333e-01
9.42134321e-01 -9.00428116e-01 -9.00920033e-01 3.83943737e-01
-3.37444186e-01 -1.63213933e+00 -5.08751154e-01 -5.58700971e-02
-7.66597569e-01 -1.15465486e+00 -7.65183926e-01 -8.32710624e-01
3.81516665e-01 6.33656800e-01 1.22402322e+00 2.33861636e-02
-6.40449107e-01 1.79024413e-01 -7.23216012e-02 -9.50538367e-02
1.49352148e-01 3.01477760e-01 -1.63390458e-01 2.72639543e-01
2.05672696e-01 -7.75587022e-01 -1.04046524e+00 2.08496809e-01
-7.42350399e-01 2.97786415e-01 2.96203613e-01 8.32115591e-01
4.26994801e-01 -4.03012872e-01 1.47989139e-01 -6.15999281e-01
-8.89696702e-02 -3.82887125e-01 -8.63904119e-01 -1.86013013e-01
-1.19320154e-01 1.57009885e-01 9.63342726e-01 -1.07377924e-01
-1.04548812e+00 2.97754645e-01 -2.08141953e-01 -6.70582175e-01
-3.82401377e-01 2.21578643e-01 1.13981431e-02 -3.48597437e-01
4.76931423e-01 1.00820802e-01 -2.15319041e-02 -4.26809698e-01
2.46201947e-01 2.41982251e-01 8.06439400e-01 -4.09439266e-01
9.26914871e-01 9.23757732e-01 3.31789494e-01 -8.34626138e-01
-1.44905388e+00 -6.52603984e-01 -7.92520642e-01 -2.92851806e-01
1.05840325e+00 -1.24581492e+00 -9.49413419e-01 8.35562170e-01
-1.23078573e+00 -5.83622515e-01 -6.95558563e-02 7.44063497e-01
-7.30059028e-01 4.42634314e-01 -5.16019762e-01 -5.22277236e-01
-3.14809173e-01 -1.29417622e+00 1.30608642e+00 2.64392793e-01
6.50991574e-02 -1.10786998e+00 2.69985795e-01 4.62603152e-01
5.31121969e-01 5.09292543e-01 4.10840988e-01 5.79529144e-02
-9.82218325e-01 8.77508149e-02 -4.27166194e-01 3.46987754e-01
5.96000962e-02 -1.22794047e-01 -1.04987001e+00 -4.61227477e-01
-8.69117379e-02 -6.92361057e-01 1.23516178e+00 3.86563331e-01
1.36622238e+00 9.83494893e-02 1.96000393e-02 1.41582751e+00
1.54181170e+00 -2.25296304e-01 4.72125411e-01 3.13684136e-01
1.12605548e+00 6.79959953e-01 4.59778249e-01 4.15857017e-01
5.93712866e-01 7.56529927e-01 8.62369120e-01 -1.74284860e-01
-3.68098646e-01 -3.95798624e-01 2.16543540e-01 6.16609514e-01
1.01952076e-01 -2.05624640e-01 -7.99107254e-01 6.08660936e-01
-2.03363347e+00 -7.83349812e-01 -2.48692706e-01 2.03062987e+00
5.08338571e-01 1.18303433e-01 -4.98214215e-02 -1.85459631e-03
3.81999493e-01 6.92672670e-01 -7.13115215e-01 6.77474216e-02
-2.49106437e-01 1.22556902e-01 5.06231308e-01 6.01009727e-01
-1.54764962e+00 1.15819848e+00 4.97589779e+00 4.65789169e-01
-1.29346919e+00 -9.28677469e-02 6.41137600e-01 -5.11667989e-02
1.60036594e-01 1.49318099e-01 -9.59396362e-01 3.61808360e-01
4.20782536e-01 1.07432619e-01 2.26099327e-01 7.79294252e-01
3.35834682e-01 -1.73959911e-01 -1.14759815e+00 1.25368321e+00
7.41007850e-02 -1.46364856e+00 -1.72642350e-01 3.13615017e-02
8.91350746e-01 3.40840220e-01 1.40493289e-01 1.85862124e-01
3.30770701e-01 -8.75723720e-01 6.82452083e-01 2.49152303e-01
5.39167643e-01 -4.86502081e-01 6.21654809e-01 3.17924976e-01
-1.09967554e+00 -3.37924473e-02 -6.02669835e-01 -4.03312594e-01
4.05948818e-01 7.17048407e-01 -4.38451320e-01 5.77368677e-01
8.46024394e-01 1.54171968e+00 -4.57089633e-01 1.39058530e+00
-2.16926396e-01 6.15882337e-01 -3.47590923e-01 2.82498121e-01
7.91204810e-01 -2.09632233e-01 5.83146751e-01 1.07476771e+00
1.39494985e-01 -2.93846369e-01 3.20005536e-01 8.48966241e-01
-1.46746472e-01 -6.63104057e-02 -7.73876309e-01 4.57418710e-01
-1.02539435e-01 1.18716753e+00 -4.96318489e-01 -2.50435710e-01
-4.33773667e-01 8.22108924e-01 5.70108712e-01 2.94828743e-01
-7.98346281e-01 -2.44345099e-01 1.03277338e+00 -1.34480596e-01
3.41302663e-01 -4.25513536e-01 -2.19457671e-02 -1.56858265e+00
1.06379226e-01 -4.17586327e-01 1.03521004e-01 -6.13040090e-01
-1.17435443e+00 5.30799687e-01 -1.47005901e-01 -1.38445890e+00
-1.23537399e-01 -7.45087326e-01 -5.13781905e-01 6.35052800e-01
-2.14521623e+00 -9.45448458e-01 -8.58909488e-01 6.98970139e-01
7.19672501e-01 1.19445205e-01 3.72682720e-01 4.71630514e-01
-7.39450037e-01 1.81516811e-01 -7.20572174e-02 2.47837350e-01
8.16372156e-01 -1.33936453e+00 6.47756815e-01 9.94292676e-01
1.16440743e-01 9.99877527e-02 6.55507267e-01 -7.89799094e-02
-1.35316825e+00 -1.36938369e+00 9.95207787e-01 -2.19938397e-01
5.67423940e-01 -5.09361088e-01 -7.41137385e-01 5.26965976e-01
2.77899444e-01 6.75090075e-01 1.87164649e-01 -2.15110645e-01
-4.24771875e-01 -4.51701224e-01 -6.54207945e-01 4.86588418e-01
1.53727090e+00 -4.78108466e-01 -2.79694438e-01 5.14935732e-01
8.05879831e-01 -6.29909754e-01 -5.87746024e-01 4.54602003e-01
2.84430772e-01 -1.37944114e+00 1.10047674e+00 -7.57032454e-01
8.75977457e-01 -3.82182211e-01 2.81356853e-02 -1.25181806e+00
-6.98567480e-02 -9.93465483e-01 -1.55663475e-01 8.04896951e-01
-1.23242021e-01 -5.09244561e-01 9.33728695e-01 3.44648480e-01
-5.09597540e-01 -5.32674193e-01 -8.27192783e-01 -6.97049916e-01
-2.87192184e-02 -3.29816937e-01 3.49087030e-01 8.73900771e-01
-5.23789048e-01 4.07539606e-01 -9.13408995e-01 -7.99273252e-02
8.71113241e-01 3.74820739e-01 1.30299163e+00 -1.16953230e+00
-3.29220176e-01 -3.09959322e-01 -6.98041856e-01 -1.87890208e+00
5.38720727e-01 -7.68502474e-01 2.17618585e-01 -1.31495559e+00
-1.39357582e-01 -3.27725500e-01 -2.35029422e-02 2.80345976e-01
-3.03847432e-01 3.07230741e-01 4.14178252e-01 1.74648970e-01
-8.66839707e-01 8.27205598e-01 1.52613175e+00 -2.26459250e-01
-1.63333043e-01 1.18826819e-03 -1.66237921e-01 6.72456741e-01
6.34750664e-01 -2.52905101e-01 -5.38596392e-01 -1.01313853e+00
4.48103100e-02 2.39065692e-01 7.12273717e-01 -1.31372190e+00
3.86300594e-01 6.08336776e-02 8.70375633e-02 -7.08122551e-01
4.49708998e-01 -6.25506818e-01 -2.24282771e-01 5.70037186e-01
-4.83864158e-01 -1.51221361e-03 7.68719837e-02 6.51679158e-01
-4.42160875e-01 6.94853812e-02 8.23458374e-01 -1.06191047e-01
-1.14360392e+00 8.36550415e-01 -1.10276446e-01 4.44500446e-01
8.47802937e-01 5.45351617e-02 -3.39692235e-01 -3.66398364e-01
-4.00120437e-01 4.05789584e-01 1.81027144e-01 5.53507090e-01
5.25653541e-01 -1.05565882e+00 -7.61156082e-01 1.90034375e-01
2.10714594e-01 6.07193768e-01 4.10425723e-01 9.10299063e-01
-1.07981575e+00 6.58923984e-01 -3.45063210e-01 -1.03137386e+00
-7.94792652e-01 4.14124489e-01 4.59361851e-01 -3.90506923e-01
-8.73986363e-01 1.05356252e+00 5.20427465e-01 -3.86432320e-01
4.01997894e-01 -3.43087107e-01 -4.50758822e-02 -2.78107405e-01
4.43227738e-01 2.49691322e-01 -3.24163027e-02 -6.33450747e-01
-1.87798485e-01 7.20679522e-01 -9.15459916e-02 2.97738731e-01
1.48192072e+00 -3.00120354e-01 3.08336187e-02 2.57201165e-01
1.65568852e+00 -6.45922899e-01 -2.05677557e+00 -4.46606755e-01
-1.38587549e-01 -8.20087314e-01 2.67777145e-01 -2.90840268e-01
-1.52423453e+00 1.29423344e+00 2.78789818e-01 -2.69140810e-01
9.43128347e-01 -3.43446761e-01 1.13005888e+00 4.98138487e-01
4.31747109e-01 -6.55305505e-01 1.12353310e-01 6.99241936e-01
5.14492989e-01 -1.65691340e+00 -2.20274821e-01 -3.21680844e-01
-2.58945793e-01 1.10127926e+00 9.31776047e-01 -5.50934851e-01
5.80911815e-01 -6.64482191e-02 -4.56307717e-02 6.44197688e-02
-9.00099814e-01 -4.69431549e-01 2.08657593e-01 5.29911160e-01
8.35551620e-02 -3.50729316e-01 5.62296323e-02 -2.69291818e-01
-1.27684861e-01 1.78230286e-01 4.74681646e-01 8.17949772e-01
-1.65694833e-01 -8.04070771e-01 1.08869538e-01 1.30971342e-01
-3.09297591e-01 3.68592329e-02 -7.38181621e-02 8.43941331e-01
1.04147360e-01 8.44246328e-01 1.99163899e-01 -2.13970974e-01
2.15026900e-01 -4.93096054e-01 3.69224519e-01 -2.64794260e-01
-5.00374973e-01 -2.50501245e-01 -1.03738651e-01 -1.10707080e+00
-9.37741816e-01 -5.84577858e-01 -7.21142352e-01 -4.19113427e-01
1.56860724e-01 -9.39782709e-02 4.64867175e-01 9.95431125e-01
3.25541943e-01 3.51377219e-01 8.07266712e-01 -1.15009832e+00
-3.24018896e-01 -6.19273245e-01 -2.20962003e-01 6.51653349e-01
7.52155781e-01 -7.28332341e-01 -4.68291789e-01 1.54628843e-01] | [8.678929328918457, -1.914023756980896] |
8b8d7ad0-879a-456c-9591-8ebae7e352bf | knowledge-driven-robot-program-synthesis-from | 2306.02739 | null | https://arxiv.org/abs/2306.02739v2 | https://arxiv.org/pdf/2306.02739v2.pdf | Knowledge-Driven Robot Program Synthesis from Human VR Demonstrations | Aging societies, labor shortages and increasing wage costs call for assistance robots capable of autonomously performing a wide array of real-world tasks. Such open-ended robotic manipulation requires not only powerful knowledge representations and reasoning (KR&R) algorithms, but also methods for humans to instruct robots what tasks to perform and how to perform them. In this paper, we present a system for automatically generating executable robot control programs from human task demonstrations in virtual reality (VR). We leverage common-sense knowledge and game engine-based physics to semantically interpret human VR demonstrations, as well as an expressive and general task representation and automatic path planning and code generation, embedded into a state-of-the-art cognitive architecture. We demonstrate our approach in the context of force-sensitive fetch-and-place for a robotic shopping assistant. The source code is available at https://github.com/ease-crc/vr-program-synthesis. | ['Michael Beetz', 'Rainer Jäkel', 'Darko Katic', 'Andrei Haidu', 'Franklin Kenghagho Kenfack', 'Benjamin Alt'] | 2023-06-05 | null | null | null | null | ['code-generation', 'program-synthesis', 'common-sense-reasoning'] | ['computer-code', 'computer-code', 'reasoning'] | [-3.20528559e-02 4.96280640e-01 1.40669674e-01 -3.38512063e-01
-6.20724261e-02 -8.07430208e-01 4.84885871e-01 -1.87147886e-01
-1.19635843e-01 7.50080943e-01 -3.33398208e-02 -5.76468885e-01
-2.90514201e-01 -1.01279259e+00 -7.98985839e-01 1.11861818e-03
5.11750802e-02 7.93065667e-01 1.54674485e-01 -9.21503961e-01
3.36774588e-01 5.57992816e-01 -1.85374105e+00 -1.02541827e-01
1.06292927e+00 3.04645807e-01 1.12399113e+00 9.56048489e-01
6.61677271e-02 1.20197248e+00 -1.99839741e-01 -2.38708064e-01
1.43323496e-01 -1.79704547e-01 -9.64642584e-01 -3.26149285e-01
-3.39349777e-01 -3.71453971e-01 -6.94898665e-01 9.40111756e-01
6.63820505e-02 6.63765550e-01 6.73544765e-01 -1.40595746e+00
-1.11473858e+00 6.62443757e-01 1.64539605e-01 -4.48912412e-01
9.63006318e-01 4.96944845e-01 7.44023681e-01 -4.11045313e-01
1.15528607e+00 1.26507938e+00 3.02270323e-01 9.25919354e-01
-9.90524054e-01 -3.76227617e-01 -2.31496930e-01 3.91001552e-01
-1.40424919e+00 -2.34509364e-01 5.90781569e-01 -6.18614674e-01
1.28095484e+00 1.08318925e-01 9.95537102e-01 1.18649101e+00
5.34153402e-01 5.22085249e-01 6.01163924e-01 -3.64476353e-01
4.35347497e-01 -5.64513840e-02 1.58604432e-03 1.11678445e+00
3.49654377e-01 4.52630877e-01 -1.86597586e-01 -4.83000763e-02
1.32971811e+00 6.28564283e-02 -1.77880619e-02 -8.67256641e-01
-1.36488223e+00 7.56257892e-01 6.00339055e-01 1.97094515e-01
-8.02957356e-01 7.66190350e-01 3.97863388e-01 7.46903718e-02
-5.18637657e-01 9.73102570e-01 -5.93913317e-01 -3.42674851e-01
-3.63594331e-02 1.03240287e+00 9.53321874e-01 1.35350239e+00
5.57111144e-01 7.83425495e-02 3.07592135e-02 5.03447950e-01
1.91181183e-01 5.71761012e-01 3.57957482e-01 -1.70231152e+00
3.08071524e-01 6.60280645e-01 5.72480083e-01 -9.78915632e-01
-7.82546043e-01 2.28802696e-01 -1.41030297e-01 5.87853730e-01
2.02725664e-01 -2.73016340e-04 -8.58986616e-01 1.52434242e+00
4.12194222e-01 -3.06565702e-01 3.94637018e-01 1.22381759e+00
7.26977766e-01 5.74782670e-01 2.13876963e-01 5.31176805e-01
1.65560091e+00 -9.46291566e-01 -6.13272190e-01 -4.26109821e-01
8.77891779e-01 -8.11941102e-02 1.46001637e+00 4.10314798e-01
-1.04546523e+00 -6.01396143e-01 -1.12054610e+00 -5.98607659e-01
-4.63256240e-01 1.71393633e-01 1.24251318e+00 2.73829818e-01
-1.03617513e+00 5.95835388e-01 -9.79648888e-01 -6.21780336e-01
3.16581637e-01 2.36437306e-01 -4.48822498e-01 -2.25881457e-01
-9.27196920e-01 1.20604181e+00 6.64397597e-01 -8.80288109e-02
-9.66151774e-01 -4.60222900e-01 -1.28780627e+00 -5.51928878e-02
5.07558882e-01 -1.10018802e+00 1.68652403e+00 -3.97189558e-01
-1.73947549e+00 8.80569160e-01 2.62231797e-01 -3.57090265e-01
2.13332191e-01 -3.36511970e-01 -1.40625268e-01 1.38575435e-01
2.20999777e-01 8.48568380e-01 2.69258142e-01 -1.22639275e+00
-5.25123179e-01 -5.49606562e-01 6.32899165e-01 4.12349463e-01
5.08864403e-01 -2.03407675e-01 6.56368211e-02 -2.57240564e-01
2.17559889e-01 -1.26814520e+00 -5.70066929e-01 1.52454168e-01
-2.66371846e-01 -2.56169468e-01 2.91547477e-01 -7.94678211e-01
5.06492078e-01 -1.94963789e+00 7.26611674e-01 -4.85847034e-02
2.04948738e-01 -2.67941386e-01 -6.09754734e-02 3.78984243e-01
2.88433880e-01 -2.93855011e-01 2.14018095e-02 5.06377041e-01
5.61082125e-01 5.17648637e-01 -3.92751902e-01 3.64577509e-02
-8.23278651e-02 1.18763816e+00 -1.31084454e+00 -2.73276001e-01
4.99311239e-01 3.00157487e-01 -7.30529070e-01 1.23891480e-01
-6.20368421e-01 7.01942027e-01 -7.74370313e-01 4.37173575e-01
1.69914290e-01 -5.17215729e-02 4.84096378e-01 2.07569510e-01
2.17965711e-02 3.28743845e-01 -9.36534166e-01 2.30989599e+00
-7.13308990e-01 2.89317846e-01 -4.43995837e-03 -5.97998559e-01
8.45166802e-01 3.25048529e-02 7.53805973e-03 -7.94050097e-01
1.95569441e-01 2.73371845e-01 5.94851039e-02 -8.24034214e-01
1.01420486e+00 2.39032507e-01 -6.54485583e-01 3.65704745e-01
-4.72420752e-02 -9.62972760e-01 2.63856411e-01 2.38439128e-01
1.47991264e+00 1.16127360e+00 6.07720613e-01 -1.40323564e-01
1.92990735e-01 7.97460973e-01 1.82827353e-01 6.84281409e-01
-1.61418468e-01 2.98474394e-02 1.26160100e-01 -4.84701067e-01
-1.30732846e+00 -1.27209306e+00 4.58440423e-01 1.24997604e+00
4.35510099e-01 -2.73406625e-01 -6.81999207e-01 -4.65741642e-02
-9.31540690e-03 1.40635431e+00 -3.91325504e-01 -1.50830820e-01
-8.44661534e-01 1.64101928e-01 5.37081242e-01 5.59071004e-01
2.71670699e-01 -1.77968752e+00 -1.64757371e+00 1.26376808e-01
-2.05368087e-01 -1.10082746e+00 1.85627624e-01 8.67530480e-02
-6.87720180e-01 -1.17639470e+00 -3.90026510e-01 -8.62668574e-01
6.23974144e-01 3.00875127e-01 1.12069130e+00 2.05205381e-02
-5.84502697e-01 6.17908955e-01 -5.09802759e-01 -4.22895640e-01
-7.75658607e-01 -4.87967208e-02 1.33937731e-01 -1.08036709e+00
-4.01609391e-03 -6.42818332e-01 -3.36788028e-01 1.28880456e-01
-3.83729130e-01 5.93973279e-01 4.14015174e-01 5.07498622e-01
3.01054716e-01 -9.08335894e-02 3.05039346e-01 -3.67823035e-01
8.10938001e-01 -3.27064425e-01 -6.84548497e-01 1.36422873e-01
-1.22733772e-01 6.36750460e-02 4.54679608e-01 -3.63281965e-01
-1.11057889e+00 2.95762450e-01 1.82141870e-01 -1.47766471e-01
-4.12315488e-01 3.68155122e-01 -2.33224537e-02 2.13364780e-01
1.07398462e+00 3.02997231e-01 -1.68059871e-01 1.02406912e-01
1.07545877e+00 4.91681129e-01 9.85701263e-01 -1.04381192e+00
7.94787943e-01 2.95138180e-01 -2.76072719e-03 -5.98074019e-01
-8.83956179e-02 5.70847727e-02 -6.03694618e-01 -1.06976442e-01
1.02148902e+00 -7.36121178e-01 -1.31300390e+00 7.96548501e-02
-1.25159895e+00 -1.11298060e+00 -5.25056064e-01 3.25242788e-01
-1.42798221e+00 -1.11627504e-01 -5.90012193e-01 -8.89814496e-01
-5.41131794e-02 -1.17785430e+00 1.04320693e+00 3.49706233e-01
-7.31664360e-01 -3.37587148e-01 -1.03219829e-01 3.46145242e-01
1.46648929e-01 2.63698220e-01 1.04702234e+00 -2.26127818e-01
-6.88051105e-01 1.10386983e-01 -1.86127156e-01 -3.53394806e-01
-1.47966042e-01 -3.66343200e-01 -3.62921178e-01 2.82548368e-01
-3.94596875e-01 -4.95256692e-01 -1.19944498e-01 2.83379406e-01
9.27142680e-01 -1.38317004e-01 -4.83679473e-01 3.00468326e-01
1.14426160e+00 6.46417201e-01 1.05775404e+00 3.66847694e-01
7.48797953e-01 7.96510220e-01 1.02159357e+00 3.33851576e-01
9.29213643e-01 8.58152807e-01 3.73206377e-01 5.64585268e-01
1.10775225e-01 -3.32573831e-01 2.35637471e-01 1.70046359e-01
-4.40260977e-01 1.67765319e-01 -1.16061318e+00 4.38286930e-01
-2.17591572e+00 -1.00628722e+00 -2.79627442e-01 1.66091323e+00
5.76695919e-01 -1.43507704e-01 -2.88885441e-02 -6.99052066e-02
3.72512966e-01 -4.56689537e-01 -6.17108643e-01 -7.06799626e-01
5.47160745e-01 3.48018110e-01 3.39370221e-01 3.50959688e-01
-6.90160811e-01 1.45305550e+00 5.46474266e+00 4.71144915e-01
-4.09855664e-01 1.16153993e-01 -9.06444266e-02 1.94489926e-01
-2.27158759e-02 -3.05909626e-02 -1.57294616e-01 -6.41403720e-02
8.53112757e-01 -1.50946662e-01 1.08498693e+00 1.26255894e+00
2.06510976e-01 -2.18732446e-01 -1.20650983e+00 1.02107644e+00
-2.86821753e-01 -1.26572788e+00 -2.57138819e-01 -1.00510448e-01
2.92487383e-01 -1.66835472e-01 -3.25231075e-01 7.71743000e-01
1.07519090e+00 -1.05672693e+00 1.31754816e+00 6.16883337e-01
6.26970828e-01 -6.64797723e-01 1.01405844e-01 4.29960787e-01
-1.18938720e+00 -5.72763801e-01 -3.18472147e-01 -4.67572689e-01
2.97988087e-01 -1.21035408e-02 -8.31342578e-01 5.82852125e-01
6.47832572e-01 2.12045714e-01 1.07262135e-02 4.41040337e-01
-6.53913021e-01 -4.14277613e-01 8.01433176e-02 -3.82103115e-01
-1.90953448e-01 -3.13784808e-01 5.26713789e-01 6.50251210e-01
5.57559013e-01 6.03754163e-01 1.31881088e-01 1.39431441e+00
3.95839095e-01 -2.55885988e-01 -1.06309187e+00 -2.60196179e-01
5.38294256e-01 1.09820521e+00 -7.95657694e-01 -3.88972998e-01
3.07370871e-02 1.34308255e+00 3.07487458e-01 3.30147952e-01
-1.07013357e+00 -7.56634712e-01 9.22841728e-01 -3.57129946e-02
4.64871675e-02 -8.70156884e-01 -3.43648642e-01 -9.19797242e-01
3.63192409e-02 -7.71054327e-01 -1.40517533e-01 -1.67916429e+00
-5.44214368e-01 2.63305783e-01 1.77061483e-01 -8.69154692e-01
-7.00084627e-01 -8.78025889e-01 -1.71093494e-01 5.34433305e-01
-1.04356873e+00 -1.22669458e+00 -7.07220256e-01 3.60870272e-01
6.05090141e-01 -1.20942838e-01 1.08543694e+00 -2.59411633e-01
2.04440430e-02 -2.19101444e-01 -5.43520033e-01 -1.50901556e-01
4.58047725e-03 -9.18395162e-01 8.03331017e-01 3.46428454e-01
-3.68453741e-01 7.22613275e-01 9.18875456e-01 -8.65990222e-01
-2.03345251e+00 -7.44559288e-01 2.73276389e-01 -8.12751234e-01
6.35373652e-01 -5.01982391e-01 -5.26992679e-01 1.04904544e+00
-4.15947884e-01 -2.37866506e-01 -1.01100672e-02 -1.46782160e-01
-2.31546715e-01 5.36053419e-01 -1.45391726e+00 1.26556039e+00
1.91328621e+00 -3.73156577e-01 -9.85197723e-01 4.09311533e-01
1.09644079e+00 -9.61475074e-01 -6.42230272e-01 5.52045777e-02
8.74981165e-01 -6.49339378e-01 1.29848242e+00 -5.28019726e-01
6.76599860e-01 -5.67583203e-01 -4.26486224e-01 -1.08120990e+00
-5.61438918e-01 -4.70510006e-01 -2.00918853e-01 4.74201560e-01
1.44344062e-01 -7.48725891e-01 4.67223704e-01 7.28765547e-01
-4.58545893e-01 -3.25946450e-01 -5.67148387e-01 -8.20241868e-01
-3.08240801e-01 -6.86199129e-01 8.03057790e-01 7.77583361e-01
5.48223555e-01 1.36512205e-01 2.31272563e-01 3.77585739e-01
3.38929504e-01 1.79605540e-02 9.97242928e-01 -1.41672194e+00
-3.28337759e-01 -2.77940392e-01 -4.85046476e-01 -7.38277137e-01
5.34467638e-01 -9.84264135e-01 4.56557304e-01 -2.03918767e+00
-7.21615553e-02 -5.94729125e-01 5.01870513e-01 5.20781755e-01
4.59830076e-01 -2.40415066e-01 3.50379318e-01 1.34142593e-01
-5.00113487e-01 4.19374853e-01 1.44145679e+00 2.89223697e-02
-6.09053731e-01 -4.82645184e-01 -8.14289033e-01 7.86349714e-01
1.05550349e+00 -2.08102793e-01 -6.50459766e-01 -2.68985003e-01
6.54509425e-01 4.80725497e-01 8.55665088e-01 -1.15881801e+00
9.42021757e-02 -5.58266282e-01 1.58423975e-01 -1.30321935e-01
6.26115382e-01 -8.21007371e-01 4.92285073e-01 9.54896331e-01
-1.42142400e-01 2.03424305e-01 1.38620302e-01 3.61744553e-01
7.14482367e-01 -3.12388122e-01 2.57518619e-01 -5.59930921e-01
-1.31587613e+00 -3.54191035e-01 -7.04183698e-01 -4.68379050e-01
1.31217337e+00 -3.04920107e-01 -6.44092143e-01 -3.63083035e-01
-8.08677495e-01 3.27246815e-01 8.27142060e-01 7.19344795e-01
8.60072494e-01 -1.07277966e+00 -3.52501839e-01 -1.06792040e-02
3.18455905e-01 3.37114483e-02 2.54255652e-01 2.30762079e-01
-1.19845009e+00 5.52016139e-01 -7.12831676e-01 -1.21186450e-01
-7.70826817e-01 6.50111437e-01 1.68920383e-01 -4.59934175e-02
-9.85240340e-01 4.74387199e-01 2.15058967e-01 -1.27168977e+00
-2.87710875e-01 -7.06690133e-01 -1.11969694e-01 -8.99408817e-01
3.82541120e-01 4.71931875e-01 -4.66882616e-01 -4.43424970e-01
-4.00634766e-01 2.94700503e-01 5.30637503e-01 -4.63953793e-01
1.10923398e+00 -8.38050246e-02 -4.49149571e-02 2.09976405e-01
2.77113706e-01 -5.95859468e-01 -9.66290951e-01 4.59054679e-01
-1.65147390e-02 -2.98372358e-01 -3.73777986e-01 -8.74630988e-01
-3.26308608e-01 4.49324697e-01 3.08144420e-01 -6.77828193e-02
5.36539495e-01 1.84119284e-01 6.48029029e-01 1.23450005e+00
1.39384258e+00 -1.12108171e+00 2.95199096e-01 7.31206596e-01
1.18556249e+00 -7.60481298e-01 -3.63818184e-02 -6.44924343e-01
-8.81827712e-01 9.51649904e-01 7.69698918e-01 -3.42106164e-01
-8.11858103e-03 2.30151370e-01 -2.73231715e-01 -5.05851567e-01
-4.10344571e-01 -2.11648896e-01 -3.17037344e-01 1.33379102e+00
1.07971884e-01 4.29827005e-01 -6.67161196e-02 6.85293913e-01
-7.95284212e-01 5.09690404e-01 8.67787361e-01 1.45297396e+00
-5.21730006e-01 -6.32707179e-01 -3.51678371e-01 1.98984504e-01
2.46416524e-01 3.24030638e-01 1.77272968e-03 8.85736287e-01
-7.11570382e-02 8.44558418e-01 1.24598257e-01 -4.29611713e-01
5.18335819e-01 1.26075700e-01 1.15445757e+00 -7.41208732e-01
-1.91255242e-01 -9.02862370e-01 3.56887907e-01 -9.84347880e-01
6.32115081e-02 -5.48244536e-01 -2.23393774e+00 -2.73582250e-01
2.01199129e-01 -3.64084005e-01 1.08908641e+00 6.45190656e-01
6.14326537e-01 8.95696998e-01 -2.75905281e-01 -1.53522503e+00
-3.55884284e-01 -5.92065215e-01 -4.57382441e-01 2.37867206e-01
-1.53226957e-01 -9.91797686e-01 2.46127144e-01 1.36423171e-01] | [4.48160457611084, 0.8017279505729675] |
009e0f14-9cb5-43a8-85ef-54edb3677a15 | clustering-tree-structured-data-on-manifold | 1507.05532 | null | http://arxiv.org/abs/1507.05532v2 | http://arxiv.org/pdf/1507.05532v2.pdf | Clustering Tree-structured Data on Manifold | Tree-structured data usually contain both topological and geometrical
information, and are necessarily considered on manifold instead of Euclidean
space for appropriate data parameterization and analysis. In this study, we
propose a novel tree-structured data parameterization, called
Topology-Attribute matrix (T-A matrix), so the data clustering task can be
conducted on matrix manifold. We incorporate the structure constraints embedded
in data into the negative matrix factorization method to determine meta-trees
from the T-A matrix, and the signature vector of each single tree can then be
extracted by meta-tree decomposition. The meta-tree space turns out to be a
cone space, in which we explore the distance metric and implement the
clustering algorithm based on the concepts like Fr\'echet mean. Finally, the
T-A matrix based clustering (TAMBAC) framework is evaluated and compared using
both simulated data and real retinal images to illustrate its efficiency and
accuracy. | ['Na Lu', 'Hongyu Miao'] | 2015-07-20 | null | null | null | null | ['tree-decomposition'] | ['graphs'] | [ 1.95785612e-01 -1.21419981e-01 -6.80304875e-05 -3.25708568e-01
-1.32143199e-01 -5.11784673e-01 2.09175855e-01 1.86613835e-02
-2.17546746e-01 1.49033576e-01 1.48671851e-01 -3.20127606e-01
-8.79790783e-01 -4.17026252e-01 -1.42377630e-01 -1.00465417e+00
-2.80929595e-01 2.93530285e-01 -3.04462537e-02 1.83168277e-01
2.90870249e-01 4.91617113e-01 -1.43714845e+00 2.10412398e-01
1.29420614e+00 7.94933498e-01 3.28230679e-01 6.36485517e-01
-2.46337891e-01 3.80443603e-01 -1.79888219e-01 -9.31062624e-02
4.03115571e-01 -2.85786390e-01 -6.41366065e-01 5.96261084e-01
2.02770710e-01 2.58490205e-01 3.51555012e-02 1.16804779e+00
4.01739746e-01 -6.17833324e-02 8.27407479e-01 -1.42592049e+00
-2.29391217e-01 3.01628947e-01 -8.85828078e-01 9.37946290e-02
-7.24485219e-02 -2.29973957e-01 6.75313413e-01 -1.01690316e+00
7.71358371e-01 1.38621342e+00 4.24675763e-01 5.61685413e-02
-1.40733838e+00 -4.01551366e-01 6.54385835e-02 3.61596048e-01
-1.65922821e+00 -1.82143897e-01 9.94873106e-01 -8.93315673e-01
8.44853818e-02 4.94761616e-01 9.17837620e-01 3.69886994e-01
4.17004883e-01 5.72105765e-01 1.29866993e+00 -2.93496609e-01
3.47310811e-01 -7.01145679e-02 1.78349137e-01 7.11098611e-01
2.56995738e-01 -1.59992263e-01 -2.30182871e-01 -3.23498130e-01
9.13170457e-01 -8.51049200e-02 -2.02118814e-01 -1.10880375e+00
-1.50434566e+00 6.64572775e-01 3.58951300e-01 2.22499907e-01
-1.61586985e-01 -3.37461442e-01 2.42666870e-01 1.87463567e-01
1.34363353e-01 1.35927379e-01 -1.05496101e-01 7.07230866e-02
-5.85681319e-01 -1.35022938e-01 3.45521718e-01 1.10444546e+00
8.60178411e-01 -4.46245670e-01 1.07733570e-01 8.39763999e-01
5.54695785e-01 3.94330233e-01 4.16876853e-01 -1.10341060e+00
4.78208721e-01 1.21370554e+00 -2.82883883e-01 -1.47625864e+00
-8.23273063e-01 -4.39198583e-01 -1.41326594e+00 1.20667920e-01
1.67940259e-01 1.19713262e-01 -7.99535573e-01 1.43185282e+00
7.40602970e-01 2.65893877e-01 -1.16335154e-01 8.78792942e-01
7.18512118e-01 2.61422783e-01 -3.62036288e-01 -8.07186961e-01
1.21932495e+00 -4.12257761e-01 -6.93411171e-01 3.93395007e-01
9.98543262e-01 -7.85420597e-01 7.43634045e-01 4.55486000e-01
-6.43425882e-01 -4.08769280e-01 -9.07338083e-01 1.34553984e-01
-2.03906707e-02 3.82124960e-01 4.54054475e-01 5.95790148e-01
-9.84269440e-01 2.29343504e-01 -9.27725494e-01 -5.41731775e-01
2.39556313e-01 4.20696944e-01 -5.71010828e-01 6.91202804e-02
-3.83086056e-01 3.35311413e-01 5.28072357e-01 3.72364104e-01
-1.77244365e-01 -3.77657413e-01 -5.19412041e-01 -5.28406203e-01
1.93433881e-01 -9.40449834e-01 3.90871823e-01 -3.99406940e-01
-1.17571104e+00 8.06659460e-01 -3.24174762e-01 -3.37885208e-02
3.74363512e-01 1.45401001e-01 -5.05383611e-01 3.84793222e-01
2.41182908e-01 3.73001605e-01 1.00829995e+00 -1.41829455e+00
-4.46302176e-01 -7.61604190e-01 -2.88067013e-01 3.35561603e-01
-1.93715811e-01 -5.31919207e-03 -4.50498998e-01 -5.27317345e-01
1.04885018e+00 -1.09190404e+00 -5.71627975e-01 -4.70294133e-02
-6.66801631e-01 -9.53434184e-02 1.03410029e+00 -3.80509913e-01
1.74257076e+00 -2.28268123e+00 6.42952681e-01 8.33929360e-01
8.87265742e-01 -7.46408030e-02 -1.07053760e-02 3.99026692e-01
-3.33862394e-01 1.56278238e-01 -3.31331372e-01 8.36051852e-02
-6.56568468e-01 2.30349943e-01 8.44758973e-02 6.52959764e-01
-2.18103975e-01 4.13629353e-01 -6.61099672e-01 -1.03335178e+00
4.43475783e-01 9.57202092e-02 -5.65783620e-01 -7.01925233e-02
2.44166687e-01 6.35721743e-01 -7.86742151e-01 4.26414251e-01
9.98701870e-01 -3.11853409e-01 2.25737289e-01 -6.14318192e-01
-3.52223068e-01 -5.56301475e-01 -1.59591126e+00 1.69446325e+00
-6.44790828e-02 3.24624151e-01 3.60787958e-01 -8.76915157e-01
9.90533888e-01 8.87108073e-02 1.12238967e+00 -1.92978382e-01
2.60753393e-01 5.58141917e-02 4.56998140e-01 -5.99592566e-01
1.38329580e-01 3.23944330e-01 3.25816065e-01 2.55938441e-01
-4.11118805e-01 2.01819018e-01 1.39945343e-01 4.21638608e-01
8.43207359e-01 -1.35948718e-01 2.06651732e-01 -8.36351573e-01
8.77859771e-01 -1.16152726e-01 7.26940095e-01 9.08244634e-04
4.82708402e-02 4.83920574e-01 4.92370963e-01 -3.14489335e-01
-9.33105111e-01 -1.06027865e+00 -7.09926367e-01 3.17714989e-01
4.34778810e-01 -9.64346588e-01 -7.75657773e-01 -3.86260092e-01
-6.79693297e-02 1.23064652e-01 -6.26536608e-01 -2.35619191e-02
-4.22465265e-01 -1.03576481e+00 2.20843218e-02 7.93252066e-02
3.54873151e-01 -1.58811674e-01 -5.71466208e-01 1.23732321e-01
-2.67280996e-01 -8.50117028e-01 -5.43360889e-01 -8.13553482e-02
-1.31451142e+00 -1.37434494e+00 -1.91690937e-01 -6.39696300e-01
9.99181330e-01 4.84234482e-01 5.79153776e-01 5.98342754e-02
-4.95890349e-01 4.91322875e-01 -4.90492195e-01 -5.68814315e-02
-1.98820263e-01 -2.03316525e-01 5.78244269e-01 6.47185683e-01
1.54006377e-01 -9.58958983e-01 -8.08638096e-01 8.04438055e-01
-7.87954152e-01 3.15552205e-01 6.64819956e-01 6.84347093e-01
9.47336614e-01 6.05537236e-01 2.21333250e-01 -5.70201457e-01
4.49329615e-01 -1.63670734e-01 -5.32653749e-01 2.55513608e-01
-6.92193925e-01 -2.73375325e-02 3.86039197e-01 -3.57717514e-01
-5.41736603e-01 3.17516536e-01 4.98904258e-01 -7.95489728e-01
2.60884911e-01 5.63774288e-01 -5.98599076e-01 -3.41846228e-01
5.91287374e-01 9.87782478e-02 3.79522055e-01 -7.05697536e-01
5.86158931e-01 7.24423110e-01 2.55035043e-01 -5.54533780e-01
1.00663745e+00 8.43881130e-01 4.96601045e-01 -1.17331648e+00
-4.15599734e-01 -7.40060985e-01 -1.32179689e+00 -3.54904264e-01
9.19300139e-01 -5.50755918e-01 -8.01127195e-01 3.94121647e-01
-8.66661251e-01 3.24996591e-01 -2.77221985e-02 8.77318621e-01
-7.27918386e-01 6.80786967e-01 -2.79515892e-01 -7.51258671e-01
3.06912372e-03 -1.03028595e+00 8.30382168e-01 -2.98434585e-01
1.64603263e-01 -9.45638776e-01 7.47916326e-02 3.30312788e-01
-2.90828526e-01 5.43350697e-01 1.30618620e+00 -1.25863999e-01
-7.27001429e-01 8.69690850e-02 -2.52156407e-01 1.02684498e-01
2.32254982e-01 4.36757445e-01 -4.86442447e-01 -4.46174741e-01
2.03739524e-01 5.03643572e-01 5.70006430e-01 6.08205259e-01
1.15711915e+00 -2.93307751e-01 -5.99610507e-01 9.04429913e-01
1.27963030e+00 1.27784908e-01 5.37844300e-01 8.82822797e-02
9.97283757e-01 7.93049395e-01 8.02401066e-01 4.28114742e-01
3.73951107e-01 6.72660232e-01 3.72173667e-01 1.52090881e-02
7.87679031e-02 -1.73657477e-01 4.42725271e-02 1.63645482e+00
-1.69119164e-01 3.23453337e-01 -9.54091668e-01 3.84745091e-01
-2.11789536e+00 -6.46083593e-01 -6.78116024e-01 2.35951757e+00
4.03411657e-01 -1.84931591e-01 3.06458950e-01 2.39893049e-01
1.06355464e+00 -2.20834553e-01 -6.49023712e-01 4.66728434e-02
-6.56645149e-02 -4.53271568e-01 3.23276103e-01 3.73891801e-01
-1.03874528e+00 4.87620503e-01 6.35078239e+00 9.99762774e-01
-9.43885744e-01 -2.22293958e-01 2.55850673e-01 4.14417274e-02
-2.84553826e-01 2.20992059e-01 -4.73614573e-01 2.76265502e-01
4.97525275e-01 -3.71482432e-01 1.76540360e-01 5.69496751e-01
3.74236315e-01 -8.63098055e-02 -1.06332195e+00 1.51809919e+00
-1.87173754e-01 -1.07038176e+00 3.33278149e-01 7.76550114e-01
3.43781114e-01 -2.46927649e-01 1.68960959e-01 -4.23200727e-01
-1.87367480e-02 -8.98348272e-01 2.58319616e-01 5.90540290e-01
9.58524466e-01 -6.86030149e-01 3.30223888e-01 5.07446051e-01
-1.64390445e+00 -1.99926466e-01 -6.68764174e-01 2.32629925e-01
-3.42590213e-02 9.52899754e-01 -7.44125724e-01 9.29789960e-01
6.66484296e-01 1.15075731e+00 -8.91861856e-01 1.27176297e+00
4.60358977e-01 3.63865793e-01 -4.94718015e-01 2.52900571e-01
-1.07561335e-01 -1.17983758e+00 9.51772809e-01 6.16310596e-01
4.02662784e-01 2.62341797e-01 3.11055601e-01 7.36293495e-01
4.35709149e-01 6.29343152e-01 -7.11575329e-01 9.87282246e-02
4.41753298e-01 1.62889504e+00 -9.70388889e-01 1.68094516e-01
-2.92263597e-01 5.34256339e-01 1.04956031e-01 2.88794100e-01
-3.83778960e-01 -2.65738040e-01 6.23162210e-01 2.70810485e-01
-7.81879574e-02 -4.93547469e-01 -3.78822684e-01 -1.25134170e+00
2.40169793e-01 -6.25266552e-01 3.70460331e-01 -6.90723240e-01
-1.18516505e+00 5.27071655e-01 1.24943271e-01 -1.98407364e+00
1.94754735e-01 -6.60516143e-01 -3.01131845e-01 4.28797662e-01
-6.60113454e-01 -9.82166350e-01 -2.61995435e-01 9.46385562e-01
-1.45928681e-01 -2.78410912e-01 4.86624926e-01 1.85580119e-01
-6.97106600e-01 2.88488746e-01 5.20984113e-01 1.08704686e-01
3.01891834e-01 -1.23956501e+00 -2.31746674e-01 6.14922285e-01
1.60082862e-01 9.52448070e-01 3.83417875e-01 -6.24752820e-01
-1.69777012e+00 -1.23861766e+00 1.59422562e-01 -7.38760650e-01
7.45993853e-01 -4.62035596e-01 -6.72963619e-01 5.05609453e-01
-2.44034678e-01 -1.09374121e-01 9.56179559e-01 2.55742759e-01
-2.44642600e-01 -3.53972495e-01 -1.13008666e+00 6.73589706e-01
1.45848525e+00 -3.26188296e-01 -4.05696899e-01 3.78976524e-01
7.09267020e-01 -1.04491927e-01 -1.44606423e+00 6.24082685e-01
3.94301265e-01 -9.58719134e-01 8.89221013e-01 -4.05234337e-01
-3.52501087e-02 -9.18907940e-01 -2.93236136e-01 -1.11305034e+00
-7.36296058e-01 -7.56142616e-01 1.00650415e-01 1.23979259e+00
1.01969227e-01 -5.38526654e-01 8.95047069e-01 1.47076696e-01
2.69683624e-05 -9.76001501e-01 -1.35963857e+00 -8.16445112e-01
-1.09511688e-02 -4.84169006e-01 5.61240256e-01 1.03677964e+00
2.43225452e-02 4.99719143e-01 -3.28267440e-02 3.30524892e-01
1.04787982e+00 4.62029278e-01 8.38983476e-01 -1.72088921e+00
1.00904569e-01 -1.97651401e-01 -1.05048096e+00 -9.19820428e-01
-1.62639931e-01 -1.17697501e+00 -4.89840955e-01 -1.42096901e+00
1.55210331e-01 -8.92795384e-01 -8.21995661e-02 -1.21118464e-01
1.39059186e-01 -2.04067409e-01 6.42871708e-02 6.34547532e-01
-5.43537617e-01 8.01262379e-01 1.58998001e+00 1.58819661e-01
-4.20151412e-01 5.46941385e-02 -4.42607582e-01 8.21115196e-01
4.80099410e-01 -9.98291224e-02 -7.50705302e-01 -1.57872722e-01
1.62240088e-01 2.58470505e-01 2.52654582e-01 -1.04354012e+00
2.74869949e-01 -2.55268484e-01 2.32168317e-01 -9.23898399e-01
3.45128417e-01 -1.27904880e+00 3.77697766e-01 2.30074972e-01
4.26493920e-02 1.13923689e-02 -2.23680273e-01 8.90337229e-01
-2.17530414e-01 2.49078393e-01 7.97206819e-01 2.38223881e-01
-3.46543103e-01 6.83562756e-01 -1.51780128e-01 -1.51261203e-02
1.34919548e+00 -8.39233458e-01 -1.62422761e-01 -5.97633831e-02
-9.48119402e-01 4.78295326e-01 6.46576226e-01 2.22726405e-01
8.84935200e-01 -1.82616496e+00 -6.41819119e-01 4.65283036e-01
5.18767178e-01 1.78187132e-01 4.49626982e-01 1.28660655e+00
-4.28769737e-01 2.24641219e-01 -1.53324395e-01 -1.42546654e+00
-1.56654966e+00 9.29320574e-01 2.79943854e-01 1.98264331e-01
-8.32700908e-01 3.81964862e-01 5.62624633e-01 -5.96583784e-01
-9.73527953e-02 -3.82671952e-01 -4.40847367e-01 2.21011832e-01
1.87029496e-01 5.55397034e-01 -1.77886918e-01 -1.06758368e+00
-3.73200446e-01 1.38125467e+00 1.45498455e-01 -5.36670052e-02
9.65097487e-01 -5.26444733e-01 -7.63967872e-01 5.76532662e-01
1.35116029e+00 9.56244394e-02 -7.52349854e-01 -3.42778146e-01
1.01130009e-01 -7.04591990e-01 -8.54056254e-02 -1.47188365e-01
-1.04496753e+00 8.77113163e-01 7.86596835e-01 1.80274546e-01
1.30506909e+00 -2.77749263e-03 3.00116986e-01 3.46849680e-01
5.99748671e-01 -1.10830998e+00 -7.56304637e-02 -2.79477034e-02
9.98492658e-01 -6.76506221e-01 1.75640330e-01 -1.08682060e+00
-5.76540411e-01 1.08030176e+00 4.82605875e-01 -9.71887261e-03
1.28162062e+00 -1.18361436e-01 1.10234238e-01 -3.14301610e-01
-7.31018066e-01 -1.47456571e-01 4.36765283e-01 8.20756912e-01
8.59569833e-02 2.89387316e-01 -3.06066155e-01 2.84324199e-01
-5.72900474e-01 -1.93576381e-01 5.41577578e-01 5.57355523e-01
-4.44930613e-01 -1.13830173e+00 -5.88489473e-01 8.75658691e-01
1.55846089e-01 2.35581636e-01 -2.92446852e-01 5.21518588e-01
2.74805397e-01 1.04689050e+00 3.18405852e-02 -9.24718261e-01
1.54280812e-01 -2.26267487e-01 3.67073804e-01 -5.62973857e-01
-2.71712374e-02 4.23764050e-01 -1.72821134e-01 -6.76590502e-01
-6.51734233e-01 -9.11074102e-01 -1.13423133e+00 -2.81777173e-01
-3.92384619e-01 4.67260152e-01 5.93408048e-01 4.88090754e-01
6.12605453e-01 1.94219023e-01 1.04056001e+00 -2.26070419e-01
-1.52172506e-01 -6.07038736e-01 -1.14624703e+00 3.12970996e-01
2.42225766e-01 -9.49496031e-01 -4.13984895e-01 9.47302580e-03] | [7.966707706451416, 4.559545993804932] |
e876b44f-2231-4a65-a4e4-4750e4886280 | the-reactor-a-fast-and-sample-efficient-actor | 1704.04651 | null | http://arxiv.org/abs/1704.04651v2 | http://arxiv.org/pdf/1704.04651v2.pdf | The Reactor: A fast and sample-efficient Actor-Critic agent for Reinforcement Learning | In this work we present a new agent architecture, called Reactor, which
combines multiple algorithmic and architectural contributions to produce an
agent with higher sample-efficiency than Prioritized Dueling DQN (Wang et al.,
2016) and Categorical DQN (Bellemare et al., 2017), while giving better
run-time performance than A3C (Mnih et al., 2016). Our first contribution is a
new policy evaluation algorithm called Distributional Retrace, which brings
multi-step off-policy updates to the distributional reinforcement learning
setting. The same approach can be used to convert several classes of multi-step
policy evaluation algorithms designed for expected value evaluation into
distributional ones. Next, we introduce the \b{eta}-leave-one-out policy
gradient algorithm which improves the trade-off between variance and bias by
using action values as a baseline. Our final algorithmic contribution is a new
prioritized replay algorithm for sequences, which exploits the temporal
locality of neighboring observations for more efficient replay prioritization.
Using the Atari 2600 benchmarks, we show that each of these innovations
contribute to both the sample efficiency and final agent performance. Finally,
we demonstrate that Reactor reaches state-of-the-art performance after 200
million frames and less than a day of training. | ['Remi Munos', 'Will Dabney', 'Audrunas Gruslys', 'Mohammad Gheshlaghi Azar', 'Marc Bellemare', 'Bilal Piot'] | 2017-04-15 | the-reactor-a-fast-and-sample-efficient-actor-1 | https://openreview.net/forum?id=rkHVZWZAZ | https://openreview.net/pdf?id=rkHVZWZAZ | iclr-2018-1 | ['distributional-reinforcement-learning'] | ['methodology'] | [-1.58516746e-02 -1.16388761e-01 -5.40128529e-01 -2.28169665e-01
-9.94840860e-01 -7.29817033e-01 9.97744620e-01 1.17478855e-01
-1.02928805e+00 9.87753510e-01 4.04684484e-01 -3.72152716e-01
-2.29739860e-01 -6.75248504e-01 -7.55015433e-01 -9.35731113e-01
-4.71615821e-01 5.88460267e-01 6.30412936e-01 -4.27761763e-01
1.65756166e-01 4.01690602e-01 -1.42509174e+00 2.76678234e-01
6.39764071e-01 9.25009608e-01 1.67398289e-01 8.37053835e-01
3.04819673e-01 1.08586931e+00 -6.64027750e-01 -2.15548292e-01
5.30971467e-01 -4.61964220e-01 -6.05394900e-01 -1.06671847e-01
4.27472383e-01 -8.19671810e-01 -2.90069163e-01 9.18749034e-01
9.16618168e-01 5.37813425e-01 6.01281404e-01 -1.05007541e+00
-4.21593040e-01 8.84937823e-01 -6.63992882e-01 3.74780715e-01
8.36075768e-02 6.98806584e-01 9.85842168e-01 -4.60604221e-01
5.61723292e-01 1.48891318e+00 4.37105298e-01 6.64145947e-01
-1.21498454e+00 -3.69342148e-01 5.63086569e-01 3.19566816e-01
-6.82350636e-01 -3.36839348e-01 3.52013797e-01 -2.34835938e-01
1.16204584e+00 -3.80393350e-03 8.23008239e-01 1.40806723e+00
8.54524076e-02 1.27030551e+00 1.13291276e+00 -4.17638749e-01
7.23429024e-01 -3.83696675e-01 -1.94388315e-01 6.29398525e-01
-2.15006620e-02 7.59435296e-01 -3.29287618e-01 -3.53880852e-01
5.41677594e-01 -1.60292000e-01 -9.78911072e-02 -4.12779123e-01
-1.36595690e+00 9.82444167e-01 2.84508109e-01 -2.63226539e-01
-6.75536692e-01 5.17184913e-01 7.08934247e-01 3.13907862e-01
4.41479623e-01 4.20174539e-01 -5.44591904e-01 -7.57145703e-01
-8.06266606e-01 6.83339179e-01 6.29545808e-01 7.28935301e-01
4.94870752e-01 2.76335955e-01 -6.59856975e-01 6.49830163e-01
1.03947729e-01 8.88338745e-01 5.40688515e-01 -1.58540440e+00
2.55744070e-01 -1.20826915e-01 3.68873298e-01 -2.72406638e-01
-3.76852036e-01 -3.86820346e-01 -4.57523197e-01 6.17660940e-01
4.33170438e-01 -4.71647561e-01 -7.54165411e-01 2.03656840e+00
4.91949171e-01 3.11716050e-01 3.21781278e-01 9.64425027e-01
-3.81028242e-02 6.91198468e-01 1.58872813e-01 -4.09661144e-01
1.18868315e+00 -1.26164162e+00 -3.61929417e-01 -5.44728525e-02
5.69362104e-01 -5.17106414e-01 1.25643003e+00 4.67264920e-01
-1.08406794e+00 -2.75375992e-01 -9.50413048e-01 3.74672174e-01
4.34003957e-02 -1.58542380e-01 3.74966681e-01 3.20807755e-01
-1.10126817e+00 6.88642502e-01 -1.08962429e+00 -2.45998994e-01
4.05722350e-01 6.09584153e-02 2.69276470e-01 2.71115512e-01
-1.13730657e+00 1.03444731e+00 4.08045739e-01 -5.16211152e-01
-1.38815498e+00 -8.56593311e-01 -6.73639357e-01 -1.66666090e-01
9.48326349e-01 -5.42965412e-01 1.94502532e+00 -9.05595839e-01
-2.18971705e+00 2.51202166e-01 9.76500884e-02 -1.14684987e+00
7.38248825e-01 -5.06772816e-01 -6.09149598e-02 2.08792135e-01
1.42967282e-02 8.93421829e-01 6.60499156e-01 -1.04934800e+00
-1.15861368e+00 -1.21039137e-01 2.35309988e-01 4.31804597e-01
-1.23124987e-01 -1.18328415e-01 -1.40432462e-01 -8.82978022e-01
-9.11160529e-01 -9.97247100e-01 -3.07578146e-01 -2.92967290e-01
1.31699577e-01 -4.76573974e-01 6.44808352e-01 -3.15240264e-01
1.05403030e+00 -1.97580349e+00 3.70756000e-01 -3.48186195e-02
7.68413884e-04 4.14442390e-01 -3.84802163e-01 4.33129817e-01
3.79018754e-01 -1.90064713e-01 -3.65472704e-01 -3.72498035e-01
3.58599156e-01 3.93309116e-01 -5.66070139e-01 6.33526385e-01
1.44092795e-02 9.23923314e-01 -1.36172509e+00 -2.71421045e-01
2.63038933e-01 1.52239025e-01 -7.61628687e-01 8.37828219e-02
-7.88365722e-01 3.46399873e-01 -2.17670679e-01 2.52565891e-01
3.17321658e-01 7.98489973e-02 1.70215622e-01 2.46043399e-01
-2.61302173e-01 2.22195804e-01 -9.85859454e-01 1.58178663e+00
-3.46612841e-01 4.12499726e-01 -5.34716770e-02 -7.56504714e-01
5.90776265e-01 7.93851912e-02 7.96734214e-01 -9.76128578e-01
4.29039560e-02 2.62456447e-01 -6.79417849e-02 -3.65704268e-01
7.50822425e-01 1.07671343e-01 -6.62578866e-02 6.26722336e-01
-1.99579671e-01 1.76031560e-01 7.59371221e-01 2.27329776e-01
1.05625987e+00 7.09847569e-01 3.18881124e-01 -5.26047528e-01
2.44560972e-01 9.88701433e-02 8.04020762e-01 9.87131894e-01
-5.28791547e-01 1.02753401e-01 6.06687129e-01 -5.85204363e-01
-1.26864398e+00 -1.08154762e+00 2.62112141e-01 1.54408085e+00
-6.02074787e-02 -3.21139544e-01 -7.60029554e-01 -9.21800375e-01
2.91666895e-01 9.59956825e-01 -7.67260611e-01 -2.76154220e-01
-8.49704444e-01 -6.22709155e-01 6.69156492e-01 5.06091177e-01
4.53285456e-01 -1.36308753e+00 -1.37098622e+00 3.40464860e-01
3.77656310e-03 -7.58847952e-01 -7.67895997e-01 1.23797990e-01
-5.74839294e-01 -8.18848372e-01 -8.63698006e-01 -2.99347609e-01
3.78592871e-02 2.93901041e-02 1.27615619e+00 -4.36371148e-01
2.72165507e-01 5.52412748e-01 -6.20192885e-01 -5.40779829e-01
-5.55357277e-01 1.88634545e-01 4.11701083e-01 -2.05632344e-01
1.11392714e-01 -4.15415257e-01 -7.63747215e-01 1.87194556e-01
-8.41762066e-01 -2.65029788e-01 6.76550448e-01 9.08106208e-01
6.94708109e-01 -4.84384835e-01 5.57579041e-01 -5.22477150e-01
8.02371740e-01 -5.21960557e-01 -1.01887822e+00 8.23902413e-02
-7.43552983e-01 5.80363393e-01 8.76662254e-01 -7.43706226e-01
-9.72750306e-01 -2.25550249e-01 -7.80619532e-02 -4.69827503e-01
1.12051345e-01 9.04028341e-02 2.99171448e-01 4.13114697e-01
8.06239307e-01 4.55066979e-01 3.76716226e-01 -2.72065163e-01
6.76760495e-01 4.01640177e-01 4.66488302e-01 -1.05332720e+00
2.48193651e-01 5.69868684e-01 -2.38300651e-01 -4.89444733e-01
-8.29302132e-01 -2.86069304e-01 1.25281243e-02 -2.81186670e-01
7.55290926e-01 -7.76322961e-01 -1.06292224e+00 4.17280763e-01
-8.32505167e-01 -1.22421730e+00 -8.97343576e-01 5.95936716e-01
-1.02465212e+00 3.32942307e-01 -5.57075977e-01 -8.78206611e-01
-3.42630178e-01 -1.32905984e+00 1.14143407e+00 1.60299823e-01
1.19328029e-01 -7.48018801e-01 6.96272135e-01 -3.77586126e-01
4.27832097e-01 9.83726084e-02 5.55460930e-01 -6.63624585e-01
-2.18011633e-01 6.15199447e-01 1.33414343e-01 3.78251314e-01
-8.96833837e-02 -1.40888065e-01 -5.44044733e-01 -7.98517764e-01
-2.00029284e-01 -5.03215075e-01 1.04628038e+00 6.49706781e-01
8.34016323e-01 -4.98211503e-01 9.00478214e-02 5.31551480e-01
1.26447725e+00 4.15683508e-01 5.09770274e-01 8.89421403e-01
3.16327006e-01 2.84357548e-01 8.60119045e-01 8.09622407e-01
4.77906674e-01 8.69856358e-01 5.68679690e-01 1.60793975e-01
-2.15322003e-01 -9.07837525e-02 1.00186408e+00 4.37829018e-01
-3.54663990e-02 -4.27410096e-01 -6.54151380e-01 6.00025892e-01
-2.37975526e+00 -1.46688426e+00 3.78922403e-01 2.34956717e+00
9.25775170e-01 8.63261446e-02 8.48514318e-01 -3.79334450e-01
4.76538450e-01 3.27364236e-01 -9.33909059e-01 -4.19540405e-01
-1.56195104e-01 5.44526316e-02 8.10872734e-01 8.04986417e-01
-1.05586267e+00 1.11318564e+00 5.98791742e+00 1.09342051e+00
-9.36006784e-01 1.78282842e-01 5.31555057e-01 -6.21501088e-01
-8.94142613e-02 -2.38285333e-01 -8.48826826e-01 6.86516643e-01
1.23843718e+00 -9.60183516e-02 9.17185843e-01 8.89536679e-01
3.74246746e-01 -1.86485291e-01 -9.33831513e-01 7.29959846e-01
-3.94826531e-02 -1.30540657e+00 -6.85414821e-02 2.40592659e-01
7.85246372e-01 5.08032084e-01 1.54863358e-01 6.97925508e-01
1.21140230e+00 -5.10093570e-01 1.20396543e+00 3.39738071e-01
4.48790789e-01 -1.01564908e+00 4.96568888e-01 3.08384001e-01
-1.08196974e+00 -3.39658946e-01 -4.32222635e-01 -4.42735180e-02
2.66928256e-01 3.55475426e-01 -5.75113118e-01 4.89042521e-01
8.40463936e-01 6.19757950e-01 -1.89214572e-01 8.30601215e-01
-2.47210220e-01 5.35862982e-01 -4.93059546e-01 -2.54653156e-01
7.53696084e-01 -2.67906785e-01 8.47473562e-01 1.14253020e+00
1.14808023e-01 -8.92402232e-02 6.72924578e-01 3.43078196e-01
7.21814781e-02 -1.68226168e-01 -4.71644372e-01 4.63737309e-01
6.56512141e-01 1.02811074e+00 -4.41515714e-01 -6.33095026e-01
-1.92607552e-01 7.73803949e-01 5.00738382e-01 4.85717356e-01
-1.17222524e+00 -3.73365469e-02 8.64484727e-01 -2.41284817e-01
7.18769133e-01 -3.44338447e-01 2.74694711e-01 -8.42166126e-01
-2.11051200e-02 -1.22068346e+00 5.83531320e-01 -1.99243128e-01
-1.19536996e+00 5.00381351e-01 3.25942695e-01 -1.27863538e+00
-6.85936034e-01 -4.37024117e-01 -2.91239321e-01 3.83778453e-01
-1.70605803e+00 -7.37459838e-01 1.02944814e-01 4.84339803e-01
8.73800874e-01 -3.63296986e-01 5.34006655e-01 1.66793168e-01
-4.71825927e-01 4.28872854e-01 5.57398379e-01 -1.51868761e-01
8.11145544e-01 -1.52142668e+00 5.20108998e-01 9.03450608e-01
-5.48314899e-02 8.59855041e-02 7.19132185e-01 -4.64961022e-01
-1.38325691e+00 -1.33543229e+00 8.51724446e-02 -2.61063904e-01
7.54305899e-01 -4.19215970e-02 -4.35341537e-01 6.89992785e-01
6.27361715e-01 -5.16337119e-02 2.54404753e-01 -1.64861411e-01
-2.82625139e-01 -1.54996097e-01 -9.89793062e-01 9.03947890e-01
1.00155830e+00 -6.69229254e-02 -5.21342814e-01 1.19234435e-01
9.30816352e-01 -3.92471850e-01 -6.96485043e-01 2.81185478e-01
5.40797293e-01 -1.24104500e+00 8.67439389e-01 -6.95594311e-01
2.45424271e-01 -3.29214364e-01 -3.35955858e-01 -1.65509713e+00
-3.01112533e-01 -1.09206259e+00 -5.20505905e-01 8.85906577e-01
3.42680365e-01 -7.71393538e-01 5.99021912e-01 -1.57740220e-01
-2.41415709e-01 -9.96874630e-01 -8.14615488e-01 -1.29120600e+00
3.43386143e-01 -3.19852442e-01 7.17518866e-01 4.38437372e-01
-1.92530513e-01 1.79488748e-01 -5.96164644e-01 -1.90184593e-01
7.93412387e-01 4.44242321e-02 7.70863593e-01 -6.09418452e-01
-6.86462760e-01 -8.14751863e-01 3.06854732e-02 -1.48082042e+00
1.85620770e-01 -7.17273533e-01 2.51760989e-01 -1.24184120e+00
5.05758598e-02 -3.31740052e-01 -4.17775154e-01 4.15032268e-01
-2.09454283e-01 -9.34713259e-02 5.83954871e-01 1.91402376e-01
-1.08193684e+00 1.12878764e+00 1.15107262e+00 -2.63020042e-02
-3.17856938e-01 -1.80581778e-01 -3.75726581e-01 5.19379437e-01
9.66793060e-01 -5.39842725e-01 -6.13184750e-01 -6.34656847e-01
-2.99976394e-02 -3.43127325e-02 3.07347566e-01 -9.01030481e-01
5.15992604e-02 -4.56603110e-01 -6.05921596e-02 -3.84110302e-01
1.76328018e-01 -4.86918062e-01 -4.04637426e-01 7.95674980e-01
-5.50956726e-01 6.39213026e-01 2.76298493e-01 6.87870562e-01
1.35695666e-01 -9.63724032e-03 9.52964783e-01 -8.91080424e-02
-8.70774806e-01 4.29955125e-01 -4.61369425e-01 4.23599541e-01
1.17752087e+00 2.19812497e-01 -6.45410836e-01 -3.76107931e-01
-3.60743612e-01 6.80147588e-01 4.09429789e-01 2.84146070e-01
3.02926391e-01 -1.45542836e+00 -1.10686588e+00 -2.49838203e-01
-1.83093548e-01 -8.71161520e-02 3.07545625e-02 9.13543105e-01
-4.34389561e-01 9.89587605e-02 -3.14099282e-01 -7.45825112e-01
-8.01819980e-01 5.92480838e-01 3.06833744e-01 -6.55374467e-01
-7.36443639e-01 6.19310141e-01 -1.71767008e-02 -4.74941254e-01
3.45546186e-01 -1.83995783e-01 7.59421811e-02 -4.88919951e-02
7.48859823e-01 7.25347340e-01 -2.66528130e-01 -2.05248937e-01
-3.01774532e-01 2.72895157e-01 -2.34965801e-01 -8.16052198e-01
1.36703479e+00 9.52957347e-02 3.53838354e-01 2.94816941e-01
9.06112254e-01 -1.29388003e-02 -2.19367290e+00 -2.49290764e-01
-7.13236351e-03 -3.37090075e-01 -1.78908240e-02 -9.64902639e-01
-1.02365661e+00 4.49622214e-01 6.95268810e-01 2.91931540e-01
1.00879371e+00 -2.13926077e-01 8.35772574e-01 4.34494764e-01
4.71813887e-01 -1.29384148e+00 2.03718886e-01 7.81996489e-01
7.54554152e-01 -1.13273668e+00 8.53872299e-03 5.76694310e-01
-9.96100664e-01 8.80895615e-01 4.76503462e-01 -2.43041337e-01
1.38044536e-01 3.87998372e-01 -1.32255077e-01 7.38103017e-02
-1.14411807e+00 -5.31380177e-01 -1.50980920e-01 6.83861911e-01
-4.62123826e-02 2.17159495e-01 -3.61391813e-01 -5.33696227e-02
-1.19039424e-01 -1.65671945e-01 3.57249737e-01 1.11506486e+00
-5.95784068e-01 -1.16812599e+00 -8.91707167e-02 2.48498246e-01
-3.41338754e-01 -4.60813902e-02 1.05430670e-01 8.65925491e-01
-2.94160277e-01 7.91257203e-01 1.99806914e-01 -2.18510792e-01
3.92568767e-01 -1.51002273e-01 6.20001197e-01 -1.74924627e-01
-6.40092134e-01 1.48925602e-01 1.77192748e-01 -7.92270541e-01
-4.51494306e-01 -8.66197646e-01 -1.33852458e+00 -5.10331154e-01
1.81254372e-01 2.06727684e-01 3.46537143e-01 8.72650504e-01
5.96905231e-01 6.45265102e-01 5.91189623e-01 -9.61775422e-01
-1.26832867e+00 -8.76282990e-01 -1.81198433e-01 3.36850792e-01
5.44510543e-01 -8.22986901e-01 -6.59103990e-02 -9.18234587e-02] | [4.041455268859863, 2.0666840076446533] |
3e2ab14c-4135-494c-9a59-9f6b1a0980c6 | topics-in-contextualised-attention-embeddings | 2301.04339 | null | https://arxiv.org/abs/2301.04339v1 | https://arxiv.org/pdf/2301.04339v1.pdf | Topics in Contextualised Attention Embeddings | Contextualised word vectors obtained via pre-trained language models encode a variety of knowledge that has already been exploited in applications. Complementary to these language models are probabilistic topic models that learn thematic patterns from the text. Recent work has demonstrated that conducting clustering on the word-level contextual representations from a language model emulates word clusters that are discovered in latent topics of words from Latent Dirichlet Allocation. The important question is how such topical word clusters are automatically formed, through clustering, in the language model when it has not been explicitly designed to model latent topics. To address this question, we design different probe experiments. Using BERT and DistilBERT, we find that the attention framework plays a key role in modelling such word topic clusters. We strongly believe that our work paves way for further research into the relationships between probabilistic topic models and pre-trained language models. | ['Shoaib Jameel', 'Alba Garcia Seco de Herrera', 'Mozhgan Talebpour'] | 2023-01-11 | null | null | null | null | ['topic-models'] | ['natural-language-processing'] | [-1.03450008e-01 4.65450257e-01 -5.21203995e-01 -4.45385903e-01
-6.29032731e-01 -4.53876823e-01 1.13740277e+00 4.45697278e-01
-4.16287839e-01 2.49205291e-01 9.35319304e-01 -3.98191571e-01
-7.23656788e-02 -9.65517998e-01 -4.56861168e-01 -7.42401719e-01
-2.70732224e-01 8.10053945e-01 2.47769222e-01 9.59035754e-03
5.01632452e-01 -2.43529994e-02 -1.35510242e+00 4.35729861e-01
4.67479587e-01 8.85472298e-02 5.51039517e-01 5.12117445e-01
-7.20383823e-01 5.40323436e-01 -6.04260206e-01 -1.42623484e-01
-3.89807642e-01 -1.96549088e-01 -8.73339713e-01 3.82302433e-01
-2.01301202e-01 1.20219149e-01 -6.20421916e-02 7.67971635e-01
-1.08526900e-01 3.07273865e-01 1.14998281e+00 -9.12231088e-01
-5.24048269e-01 9.98387635e-01 -5.67540109e-01 2.54422545e-01
-1.34844288e-01 -2.43901998e-01 1.43324435e+00 -9.07607853e-01
5.58920085e-01 1.66981053e+00 4.97552514e-01 1.86750427e-01
-1.56965697e+00 -4.01683152e-01 6.63049698e-01 -1.48179252e-02
-1.33394837e+00 3.84667553e-02 9.55207705e-01 -7.78259456e-01
1.12726665e+00 -1.41765893e-01 6.31922603e-01 1.22682452e+00
4.50478762e-01 7.64900148e-01 7.55505145e-01 -7.35927284e-01
3.37686002e-01 5.69639921e-01 5.81699014e-01 2.53136575e-01
3.24719936e-01 -3.81221175e-01 -4.57021356e-01 -6.56587064e-01
5.05899429e-01 -2.41179075e-02 -7.96241686e-02 -7.55271077e-01
-9.35504019e-01 1.60177791e+00 6.38948157e-02 8.61605287e-01
-6.74231708e-01 3.14756304e-01 3.29377681e-01 -2.91913390e-01
1.10719216e+00 5.35092652e-01 -5.40276766e-01 -1.59170911e-01
-1.13280845e+00 1.41871169e-01 8.02715003e-01 7.16318846e-01
1.02619052e+00 -3.49976242e-01 7.46134371e-02 8.14877510e-01
9.77751255e-01 9.76497829e-02 6.05769753e-01 -6.81919575e-01
-2.05722894e-03 3.90409321e-01 1.64797008e-01 -1.11560392e+00
-3.43331248e-02 -1.37047723e-01 -1.00010216e-01 -3.81119788e-01
2.31344122e-02 -6.38056397e-02 -8.56078863e-01 1.70706129e+00
1.03523366e-01 1.69004828e-01 1.90084934e-01 3.95698339e-01
2.22255290e-01 1.21511257e+00 7.95352221e-01 -2.21748307e-01
1.53222442e+00 -5.03789425e-01 -7.94360936e-01 -2.90626496e-01
7.38918185e-01 -8.74937296e-01 9.29014146e-01 2.56752580e-01
-8.79912078e-01 -2.23956376e-01 -7.07007587e-01 2.77070776e-02
-4.84019041e-01 -2.71383733e-01 9.65787828e-01 7.99988389e-01
-1.16482079e+00 1.01197861e-01 -1.16072845e+00 -9.99132395e-01
3.12331229e-01 -5.69086187e-02 1.37001857e-01 -8.60058516e-03
-1.25469089e+00 7.18967795e-01 5.26054084e-01 -3.25241834e-01
-1.03034425e+00 -5.15605271e-01 -8.69817376e-01 1.94414079e-01
1.83621004e-01 -2.93333620e-01 1.24618721e+00 -7.27586567e-01
-9.96405542e-01 8.67068648e-01 -7.86676764e-01 -4.38502967e-01
-3.08364600e-01 -1.46674469e-01 -9.17918459e-02 2.21812382e-01
2.98317581e-01 8.59299302e-01 7.53717244e-01 -1.67783809e+00
-6.81152463e-01 -3.31149578e-01 -2.48006180e-01 1.72044024e-01
-5.88074028e-01 2.81218410e-01 -4.71141785e-01 -6.55562103e-01
9.73502249e-02 -8.21088493e-01 -4.07111734e-01 -5.03090680e-01
-2.14566633e-01 -9.81636703e-01 8.20098877e-01 -3.19985092e-01
1.17587590e+00 -1.99608946e+00 5.80422804e-02 4.18219060e-01
1.69206545e-01 -1.59294367e-01 8.63627642e-02 8.55050266e-01
-9.66674387e-02 4.44894791e-01 1.95348053e-03 -5.19483864e-01
2.02801660e-01 4.94357556e-01 -9.73782182e-01 4.10381377e-01
1.22118965e-01 6.53810441e-01 -7.10704327e-01 -4.34906006e-01
1.28246754e-01 7.48640120e-01 -8.33030641e-01 -7.10863795e-04
-6.58281386e-01 -1.01692125e-01 -6.96281731e-01 -5.39941108e-03
3.48137826e-01 -3.24582875e-01 6.17884159e-01 2.80801684e-01
-2.19668165e-01 7.54799485e-01 -8.50457430e-01 1.68285227e+00
-3.42915982e-01 9.10288930e-01 -1.92142949e-01 -1.33580649e+00
7.78138638e-01 7.27451444e-01 4.49084133e-01 9.92577001e-02
9.04990882e-02 -3.27991039e-01 -1.66760027e-01 -3.42510730e-01
8.72955501e-01 -6.74924910e-01 -1.68227941e-01 9.27891076e-01
2.43403167e-01 4.35075685e-02 1.54664228e-02 6.58600807e-01
5.82990229e-01 -1.20072931e-01 6.98355213e-02 -6.99916601e-01
-1.86506689e-01 4.49754447e-01 2.46506140e-01 8.04297090e-01
6.59730956e-02 2.82863468e-01 7.91133046e-01 -1.31013736e-01
-1.10092163e+00 -1.13564980e+00 -5.63636720e-01 1.54421413e+00
-2.41733655e-01 -8.27228546e-01 -4.55660462e-01 -3.52686614e-01
-2.53384918e-01 1.21947157e+00 -7.89984286e-01 3.51009029e-03
-2.80435860e-01 -9.28529680e-01 1.45011708e-01 4.73115176e-01
-4.83124852e-01 -9.95086133e-01 -4.59455401e-01 5.19572079e-01
-8.07881802e-02 -7.65001535e-01 8.04570969e-03 3.52016509e-01
-1.04374146e+00 -7.39965856e-01 -6.90501332e-01 -7.90408134e-01
6.47648931e-01 3.55608821e-01 1.27588999e+00 -2.63745278e-01
-1.27054587e-01 9.15148556e-01 -4.87404257e-01 -6.03632808e-01
-3.11496139e-01 2.76190013e-01 -4.12269160e-02 -2.37889320e-01
1.23270667e+00 -6.22818589e-01 -3.06616127e-01 2.14849725e-01
-1.12561023e+00 -1.36812493e-01 1.92337617e-01 5.56589305e-01
1.30316457e-02 2.95915663e-01 4.81479138e-01 -9.69934583e-01
7.91792870e-01 -1.03925431e+00 -2.83864856e-01 -1.52141340e-02
-4.90858465e-01 1.53499231e-01 -2.33992547e-01 -5.13512790e-01
-1.23709643e+00 -4.61270064e-01 5.68244457e-02 -3.17350067e-02
-6.00847304e-01 7.15535164e-01 -4.90691559e-03 9.15152848e-01
5.94524920e-01 1.24647297e-01 -2.07349852e-01 -4.26881969e-01
8.05121720e-01 7.46643186e-01 -2.06750125e-01 -1.02040017e+00
5.13571918e-01 7.77691483e-01 -7.45315731e-01 -1.23374426e+00
-7.57764578e-01 -1.00873661e+00 -5.04375398e-01 1.57150418e-01
1.29668105e+00 -1.13237941e+00 -8.10887143e-02 -1.62899330e-01
-1.39817142e+00 -3.39861423e-01 -3.27663630e-01 7.01281428e-01
-4.75973934e-01 3.09054822e-01 -3.74211758e-01 -1.09277821e+00
1.78558260e-01 -8.47768486e-01 1.17570722e+00 -1.58167973e-01
-8.94216061e-01 -1.66635120e+00 4.11424547e-01 -5.79279475e-02
3.19443941e-01 -2.54249007e-01 1.39953697e+00 -8.82581055e-01
-5.72686553e-01 -3.77320535e-02 1.35630831e-01 -1.04915842e-01
2.51284033e-01 2.94918455e-02 -1.12555504e+00 -1.78207681e-01
2.43853971e-01 -2.23783433e-01 1.04602671e+00 6.58764303e-01
6.84730828e-01 -1.79081902e-01 -8.00296724e-01 -8.98400769e-02
1.42254865e+00 2.36873075e-01 5.45330286e-01 1.84456766e-01
3.99509639e-01 1.00284278e+00 2.73045450e-01 2.17332870e-01
5.06742060e-01 4.61684108e-01 -1.85641736e-01 1.22576430e-02
5.51687777e-01 -3.41110110e-01 2.84316272e-01 7.13746190e-01
3.33146662e-01 -3.32112551e-01 -1.22265637e+00 1.35900116e+00
-1.84795952e+00 -1.03721285e+00 -3.01663368e-03 1.60382652e+00
9.57331240e-01 2.19644293e-01 -6.07259087e-02 -1.59004986e-01
8.02267253e-01 4.47541296e-01 2.18532085e-02 -5.00929832e-01
2.72586375e-01 1.47397548e-01 4.37532216e-02 7.52165616e-01
-1.05936658e+00 1.39580321e+00 6.29184818e+00 6.52794838e-01
-7.85117507e-01 2.99823850e-01 5.80726326e-01 2.06188977e-01
-8.66326392e-01 3.32494557e-01 -1.00554049e+00 2.69932061e-01
1.14027846e+00 -2.45736390e-01 -2.72934854e-01 9.61318970e-01
5.21506488e-01 -3.16677570e-01 -1.05544543e+00 1.61289066e-01
2.08662018e-01 -1.08542633e+00 1.71050191e-01 5.13110340e-01
8.91090989e-01 5.05121015e-02 1.84556901e-01 4.78204995e-01
9.76319075e-01 -9.04237509e-01 3.19536567e-01 2.53890157e-01
2.88740043e-02 -6.44876420e-01 4.33771044e-01 4.48498905e-01
-8.51906776e-01 3.33559394e-01 -8.22582066e-01 -9.66700092e-02
3.40404540e-01 6.83260739e-01 -1.32577562e+00 -1.66728180e-02
4.84741747e-01 6.69402838e-01 -1.53995067e-01 7.00763166e-01
-3.09867829e-01 1.38160110e+00 -2.73458213e-01 -3.52627225e-02
7.28094995e-01 2.49064285e-02 5.55630445e-01 1.35253084e+00
1.20932184e-01 -1.57270595e-01 2.28031844e-01 1.03213692e+00
4.35297549e-01 2.79952973e-01 -7.64676392e-01 -3.47439915e-01
2.35058114e-01 1.04439902e+00 -1.23134542e+00 -5.07012963e-01
-4.20751572e-01 4.40402836e-01 9.42218527e-02 5.75464368e-01
-4.52702582e-01 1.27112210e-01 8.70089173e-01 2.89919585e-01
7.04767764e-01 -6.90434217e-01 -3.37152362e-01 -1.13943636e+00
-5.70025146e-01 -3.08598846e-01 1.63998097e-01 -6.67625487e-01
-1.44372094e+00 3.19487244e-01 5.15414894e-01 -3.56300652e-01
-6.21285081e-01 -3.34738135e-01 -8.62814188e-01 1.04110980e+00
-1.22118318e+00 -9.98843014e-01 3.39702129e-01 3.01407218e-01
9.37591732e-01 -2.42498126e-02 1.04235590e+00 -4.34049070e-01
-1.44785658e-01 -1.28800616e-01 1.90798402e-01 3.41533720e-02
5.44570029e-01 -1.31317437e+00 5.40061116e-01 5.18458486e-01
5.72461486e-01 1.25832951e+00 9.84921813e-01 -7.93030500e-01
-8.06507826e-01 -9.67909276e-01 1.35303545e+00 -8.67052376e-01
8.73496413e-01 -7.20891416e-01 -1.16888607e+00 1.04048610e+00
7.00565934e-01 -7.93223083e-01 1.32625294e+00 8.31832290e-01
-3.27461958e-01 6.43441916e-01 -5.19372642e-01 3.78928930e-01
2.74788439e-01 -5.80491900e-01 -1.26476896e+00 4.49246675e-01
1.11260509e+00 4.84148532e-01 -5.28358400e-01 -2.74027854e-01
1.87251806e-01 -4.35182720e-01 8.68850708e-01 -9.47327375e-01
4.86816943e-01 2.07129121e-02 -3.34614426e-01 -1.44383240e+00
-2.88205504e-01 -1.60039440e-01 2.66469359e-01 1.53751767e+00
5.71164072e-01 -4.81858671e-01 9.29038227e-01 6.56459093e-01
9.72672179e-02 -4.67658818e-01 -5.40790439e-01 -3.63085151e-01
6.66927814e-01 -9.64555681e-01 1.94503605e-01 1.20542943e+00
2.44385794e-01 5.36096573e-01 2.80973706e-02 2.08797917e-01
5.83817899e-01 -8.33782926e-02 6.12616360e-01 -1.54435456e+00
1.47155067e-02 -2.59282529e-01 -2.65707493e-01 -1.13923609e+00
5.99607587e-01 -7.57376969e-01 2.49470353e-01 -1.76876354e+00
3.05253267e-01 -5.47539175e-01 -1.99830487e-01 3.03588390e-01
-1.66639522e-01 -1.56635225e-01 -3.26669030e-02 3.78226608e-01
-6.33063316e-01 6.39125824e-01 5.28494120e-01 1.33781672e-01
-2.95388013e-01 -1.76301330e-01 -1.16492891e+00 1.02363360e+00
8.09177339e-01 -8.92856717e-01 -4.64213192e-01 -3.27164233e-01
3.08509469e-01 -4.39735025e-01 9.14068595e-02 -3.42612654e-01
2.49594465e-01 -1.90704614e-01 9.43578407e-02 -7.66461849e-01
4.08732712e-01 -7.82620788e-01 -4.34154481e-01 1.46282592e-03
-5.84668636e-01 -1.65609524e-01 1.90541834e-01 8.64824772e-01
-1.63034827e-01 -3.00369263e-01 4.86915320e-01 -3.98526460e-01
-6.96681798e-01 -1.33763179e-01 -1.39581418e+00 1.05366036e-02
8.67063999e-01 -2.06174012e-02 1.02961771e-01 -5.80633521e-01
-9.02899146e-01 2.10257545e-01 3.73112261e-01 4.14593339e-01
2.86170512e-01 -1.08747733e+00 -4.79591459e-01 1.59214232e-02
1.13349579e-01 5.97569086e-02 -4.70398515e-02 3.13362032e-01
9.09988135e-02 9.71006632e-01 4.58642870e-01 -8.29819679e-01
-8.35528970e-01 6.53706491e-01 -1.45739421e-01 -2.52212733e-01
-6.46904111e-01 7.77068496e-01 6.65599108e-01 -2.20472619e-01
2.53411829e-01 -3.88392121e-01 -4.07221258e-01 4.90948588e-01
2.98197567e-01 -1.25633284e-01 -4.49305087e-01 -6.57629073e-01
-7.06259757e-02 4.32917446e-01 -3.63389194e-01 -6.16160989e-01
1.60181379e+00 -2.80507326e-01 -1.91849649e-01 1.08913207e+00
1.18974078e+00 -8.03095251e-02 -9.85244751e-01 -3.49751204e-01
4.98883426e-01 -1.95949599e-01 1.83901519e-01 -2.57368535e-01
-2.52513587e-01 1.23659277e+00 2.75325418e-01 5.95833302e-01
3.97813946e-01 7.33442605e-01 3.70086104e-01 2.50383675e-01
1.23177513e-01 -8.56734693e-01 2.13474736e-01 4.80115384e-01
6.57715380e-01 -1.03167474e+00 -2.79660206e-02 -3.40770304e-01
-6.47112310e-01 8.02015185e-01 2.62953699e-01 -3.09931904e-01
1.08386815e+00 1.58881508e-02 -2.30168886e-02 -6.45209968e-01
-1.09713829e+00 -2.71097004e-01 1.82084978e-01 5.98969102e-01
7.71070182e-01 1.42926857e-01 -2.85822332e-01 6.41428292e-01
-2.57162809e-01 -4.93493229e-01 4.71441269e-01 8.50468159e-01
-1.03804362e+00 -1.21342111e+00 -6.08281314e-01 2.12026522e-01
-6.50862515e-01 -3.07343513e-01 -3.32233280e-01 7.76613832e-01
-7.61201531e-02 1.00405157e+00 7.58641660e-01 1.77024454e-01
-3.70195568e-01 7.30344474e-01 -9.12327543e-02 -1.33379960e+00
-1.27019107e-01 5.87291062e-01 -7.51233175e-02 -1.01388572e-02
-4.87919986e-01 -8.87335598e-01 -1.04796112e+00 2.66487002e-01
-5.06702960e-01 9.25629377e-01 9.62444842e-01 1.08830822e+00
5.88676706e-02 4.27687585e-01 3.00408930e-01 -7.91955173e-01
-3.01377736e-02 -1.29921937e+00 -6.38783872e-01 3.83417569e-02
-1.08380996e-01 -6.63002670e-01 -4.30928111e-01 3.25246006e-01] | [10.389795303344727, 6.973628997802734] |
90fb3c25-0100-4c25-b766-449a281f345c | ezcoref-towards-unifying-annotation | 2210.07188 | null | https://arxiv.org/abs/2210.07188v1 | https://arxiv.org/pdf/2210.07188v1.pdf | ezCoref: Towards Unifying Annotation Guidelines for Coreference Resolution | Large-scale, high-quality corpora are critical for advancing research in coreference resolution. However, existing datasets vary in their definition of coreferences and have been collected via complex and lengthy guidelines that are curated for linguistic experts. These concerns have sparked a growing interest among researchers to curate a unified set of guidelines suitable for annotators with various backgrounds. In this work, we develop a crowdsourcing-friendly coreference annotation methodology, ezCoref, consisting of an annotation tool and an interactive tutorial. We use ezCoref to re-annotate 240 passages from seven existing English coreference datasets (spanning fiction, news, and multiple other domains) while teaching annotators only cases that are treated similarly across these datasets. Surprisingly, we find that reasonable quality annotations were already achievable (>90% agreement between the crowd and expert annotations) even without extensive training. On carefully analyzing the remaining disagreements, we identify the presence of linguistic cases that our annotators unanimously agree upon but lack unified treatments (e.g., generic pronouns, appositives) in existing datasets. We propose the research community should revisit these phenomena when curating future unified annotation guidelines. | ["Brendan O'Connor", 'Mohit Iyyer', 'Luke Yeh', 'Jack Merullo', 'Kalpesh Krishna', 'Wenlong Zhao', 'Marzena Karpinska', 'Ankita Gupta'] | 2022-10-13 | null | null | null | null | ['coreference-resolution'] | ['natural-language-processing'] | [ 4.3529116e-02 3.7873384e-01 -1.4807767e-01 -3.9694321e-01
-1.3313395e+00 -1.3587763e+00 5.1815784e-01 3.9698246e-01
-5.6401461e-01 9.5843065e-01 8.5671979e-01 -3.1739664e-01
-2.0116642e-01 -4.8119009e-02 -4.5789739e-01 -3.5326114e-01
7.0182759e-01 9.1319972e-01 4.0307438e-01 -4.5474681e-01
2.8592470e-01 9.8995760e-02 -1.4463899e+00 5.1763809e-01
1.2087460e+00 2.0143628e-01 1.7480196e-01 3.6132732e-01
-1.4158413e-01 8.2638323e-01 -7.7316439e-01 -1.0937459e+00
-6.4587966e-02 -1.9815846e-01 -1.5448337e+00 -4.4785684e-01
7.0539844e-01 1.4376931e-01 -7.7490509e-02 1.1099513e+00
6.8025750e-01 2.6704338e-01 2.0384112e-01 -1.0179455e+00
-6.7982376e-01 9.9853593e-01 -2.3997369e-01 3.4976688e-01
8.2593364e-01 -1.7403373e-01 1.2947395e+00 -5.1864612e-01
1.1650069e+00 1.2620708e+00 9.0947717e-01 8.5007638e-01
-1.0307745e+00 -4.8756504e-01 -9.5403671e-02 2.6707333e-01
-1.2108400e+00 -6.2088984e-01 4.2771348e-01 -6.2944520e-01
8.0776656e-01 5.4593229e-01 1.5871777e-01 1.4080355e+00
-4.6085429e-01 4.9205688e-01 9.3263078e-01 -6.0804021e-01
-5.6009721e-02 2.8693178e-01 2.7816966e-01 6.4463675e-02
3.8144058e-01 -2.4900085e-01 -7.4520236e-01 -5.4637939e-01
3.2257271e-01 -5.8982790e-01 -7.1848971e-01 -2.0975436e-01
-9.8720109e-01 7.1948177e-01 7.1837790e-02 6.6242605e-01
3.5524685e-02 -5.0735521e-01 5.8815879e-01 8.5414872e-02
1.7462847e-01 8.6065704e-01 -5.2175164e-01 -5.5551672e-01
-5.4601109e-01 4.2162275e-01 1.2226672e+00 1.1279678e+00
3.9219952e-01 -6.7918801e-01 -1.8077530e-01 1.0713915e+00
1.0649389e-01 2.9500815e-01 2.3708561e-01 -1.6746200e+00
7.1577829e-01 7.2601205e-01 9.7441977e-01 -1.1122837e+00
-3.9389288e-01 -2.0005416e-02 -4.5735493e-02 -3.3711261e-01
8.6352175e-01 -3.4100130e-01 8.5722364e-02 1.7264509e+00
3.3968872e-01 -2.0530225e-01 2.0315276e-01 1.1125646e+00
1.1869280e+00 -6.3397847e-02 3.4062257e-01 -2.8193903e-01
1.6175015e+00 -8.1411839e-01 -9.1923332e-01 -6.9216691e-02
8.7253022e-01 -1.2079072e+00 1.1523829e+00 2.1368296e-01
-1.1078845e+00 -1.7970227e-01 -6.7981148e-01 -2.8971383e-01
-7.5324938e-02 -1.5373695e-01 3.6090612e-01 4.5593810e-01
-8.2920200e-01 5.4966503e-01 -4.7956443e-01 -7.7506721e-01
2.5284654e-02 2.2278485e-01 -5.3201973e-01 1.2999707e-01
-1.3739209e+00 1.2247854e+00 2.3846842e-01 -2.6043165e-01
-4.1941971e-01 -6.5979540e-01 -7.0013785e-01 -2.8869373e-01
6.0883111e-01 -3.2931113e-01 1.7738463e+00 -9.8210979e-01
-1.1279237e+00 1.4282922e+00 -3.1736919e-01 -1.5388729e-01
4.4077170e-01 -4.2476121e-01 -6.8255138e-01 1.7066766e-01
6.9104052e-01 3.7592024e-01 -1.0316294e-02 -1.4450908e+00
-1.0789279e+00 -2.2868156e-03 4.1538322e-01 4.3505299e-01
-1.7988864e-01 6.2374997e-01 -3.9039686e-01 -3.9404982e-01
-1.7811771e-01 -9.7489983e-01 2.3874266e-01 -4.1406849e-01
-2.1135570e-01 -6.2769586e-01 5.5506456e-01 -5.4437953e-01
1.3449419e+00 -2.0270345e+00 8.8331103e-02 -3.9636645e-01
3.0468452e-01 4.3328038e-01 5.0224230e-02 5.4079515e-01
7.5313024e-02 4.4246301e-01 -2.3724211e-02 -1.4320526e-01
1.3349742e-01 3.5829616e-01 -3.8812289e-01 3.0673471e-01
-2.0064035e-01 5.3696001e-01 -1.5712258e+00 -8.0860734e-01
-1.8396077e-01 2.8060761e-01 -6.2675047e-01 2.8766978e-01
-7.4879318e-02 6.5843248e-01 -4.5117074e-01 5.3633696e-01
2.8292251e-01 -3.0074245e-01 6.0923171e-01 -2.7408591e-01
-4.1689432e-01 7.3465014e-01 -1.0131092e+00 1.9063886e+00
-2.5645202e-01 7.0510626e-01 3.9656946e-01 -5.0668180e-01
7.7645761e-01 7.8370708e-01 3.1749704e-01 -2.6230079e-01
1.9962183e-01 5.0736946e-01 1.5720534e-01 -8.2602602e-01
7.4442762e-01 1.2567359e-02 -3.2254234e-01 6.1279023e-01
1.8558097e-01 -1.3689512e-01 4.7741985e-01 3.5346711e-01
1.0206188e+00 8.5780658e-02 1.9014242e-01 -4.9366766e-01
5.6987554e-01 6.2335455e-01 9.8146915e-01 7.8555769e-01
-7.0669454e-01 4.6870497e-01 4.5224258e-01 -3.3625060e-01
-8.4281170e-01 -6.6003710e-01 -3.6340398e-01 1.3683265e+00
1.3248034e-01 -6.9038951e-01 -8.2316643e-01 -7.8359771e-01
-3.4123230e-01 6.4660597e-01 -4.2425743e-01 2.9548010e-01
-6.6106629e-01 -3.1081983e-01 8.6087710e-01 4.1329676e-01
2.9941171e-01 -9.5404351e-01 -7.9113734e-01 3.4941566e-01
-1.0115497e+00 -1.4620306e+00 -4.9900618e-01 -3.6087804e-02
-1.4178950e-01 -1.6331964e+00 -2.8265598e-01 -7.5775015e-01
2.8539971e-01 2.9032421e-01 1.5133439e+00 3.2338607e-01
1.6452990e-01 5.7617062e-01 -7.4149019e-01 -3.9330232e-01
-5.8512658e-01 1.6499451e-01 1.2018995e-01 -6.2876040e-01
1.0950273e+00 -3.6979297e-01 -2.1428172e-01 5.4732382e-01
-4.9137545e-01 -3.0550715e-01 -8.7986998e-02 6.4012575e-01
1.5075324e-01 -6.4549237e-01 5.1136935e-01 -1.0473576e+00
9.4503206e-01 -4.4444081e-01 -3.5195228e-01 5.2711028e-01
-2.8588513e-01 -2.6331398e-01 2.7999517e-01 -3.0003226e-01
-1.3179374e+00 -4.2661133e-01 -5.1849138e-02 4.6911784e-02
-3.9948758e-01 2.8398755e-01 -7.7043854e-02 4.3731026e-02
9.8082352e-01 -6.6068727e-01 -4.9929416e-01 -5.5965787e-01
2.9330838e-01 9.6483302e-01 1.0663160e+00 -1.2583308e+00
4.8568320e-01 2.5268087e-01 -7.3645616e-01 -3.5469106e-01
-1.3919644e+00 -6.1811650e-01 -8.1489855e-01 -1.4206614e-01
8.7363964e-01 -1.0228066e+00 -8.0894160e-01 -2.7197841e-01
-1.5855660e+00 -2.3060860e-01 -2.4425136e-01 5.2527761e-01
-3.4306914e-01 5.5508500e-01 -6.6414195e-01 -4.9262157e-01
-3.8683191e-01 -1.1111263e+00 8.1072909e-01 5.0145549e-01
-1.2871861e+00 -1.0068808e+00 4.5519283e-01 9.0959102e-01
1.3352628e-01 2.0698780e-01 6.2065876e-01 -1.0991005e+00
1.4749098e-01 4.2481363e-02 -1.4790750e-01 -2.5207767e-01
8.8145278e-02 2.3326322e-01 -9.6627539e-01 2.0474041e-02
-2.7372628e-01 -6.7625332e-01 1.4623211e-01 -2.7621317e-01
3.6182526e-01 -1.8962406e-01 -4.4338533e-01 2.0246279e-01
9.9625170e-01 2.5796804e-01 1.6688225e-01 8.4150434e-01
3.7926370e-01 8.7096792e-01 7.2350961e-01 3.2608667e-01
5.8659470e-01 8.1575787e-01 4.3313134e-02 4.4137454e-01
1.8310916e-01 -2.1957539e-01 3.2836270e-02 8.0506021e-01
-2.5794387e-01 -2.3689401e-01 -1.3022935e+00 9.1549158e-01
-2.1157277e+00 -1.0342803e+00 -4.4926015e-01 1.7816195e+00
1.3747280e+00 -6.0259935e-02 -7.0494883e-02 -9.1137707e-02
1.0157102e+00 -3.6331012e-03 -1.4294681e-01 -4.4537973e-01
-3.9003524e-01 9.7931921e-02 -6.5791912e-02 7.3694956e-01
-1.0754579e+00 1.1940863e+00 6.6234784e+00 3.5873047e-01
-5.6952077e-01 2.8453794e-01 -1.0814522e-01 2.4284966e-02
-5.4901856e-01 2.9631522e-01 -9.1716725e-01 2.9135981e-01
7.7717346e-01 -2.6906058e-01 2.9225987e-01 6.3842714e-01
1.3395563e-01 5.6549445e-02 -1.0148125e+00 7.3775786e-01
1.3443160e-01 -1.2683221e+00 -2.2694260e-01 -3.8890046e-01
7.1459800e-01 1.2840058e-01 -5.8445138e-01 2.3673180e-01
1.1075895e+00 -7.4977648e-01 6.5885276e-01 2.2006005e-01
6.1410409e-01 -3.4887725e-01 9.2131370e-01 1.7922567e-01
-7.5249505e-01 8.4089868e-02 -2.9613009e-01 -1.7971952e-01
3.8314041e-01 1.5209998e-01 -4.2866179e-01 5.0592107e-01
9.9875015e-01 4.3205404e-01 -2.9850611e-01 8.4995860e-01
-6.0019737e-01 4.2514622e-01 -1.8389873e-01 3.5650272e-02
7.2558239e-02 9.8504357e-02 6.7825311e-01 1.2175394e+00
-3.9181460e-02 6.2079477e-01 2.1040960e-01 6.2482834e-01
-2.9261839e-01 1.8395145e-01 -2.7620396e-01 5.3959887e-02
1.2735310e+00 1.5251040e+00 -4.2823958e-01 -1.3283566e-01
-6.4245379e-01 5.9137750e-01 7.0795435e-01 2.8770602e-01
-6.8726951e-01 -3.2823271e-01 6.3542879e-01 -7.1168221e-03
-8.9064054e-02 5.8423970e-02 -1.2544046e-01 -1.2017145e+00
-1.0066123e-01 -1.2910774e+00 7.4125725e-01 -8.5798836e-01
-1.3235142e+00 7.0253235e-01 6.1849136e-02 -9.8813939e-01
-1.7356022e-01 -3.6778015e-01 -6.2797040e-01 7.0895809e-01
-1.2180290e+00 -8.1729710e-01 -5.9014809e-01 5.9816211e-01
3.1948832e-01 -7.0660606e-02 9.6043277e-01 3.8331929e-01
-5.7341295e-01 5.5612785e-01 -3.1551403e-01 4.4758227e-01
1.5136136e+00 -1.1269841e+00 -2.2108292e-02 6.0549134e-01
7.1414419e-02 1.0215328e+00 1.1748648e+00 -6.4374900e-01
-8.1388628e-01 -5.3604007e-01 1.4105990e+00 -9.4940937e-01
1.0190258e+00 4.5247018e-02 -1.0218927e+00 8.8175172e-01
6.8787086e-01 -3.0038178e-01 1.1387979e+00 3.5041735e-01
-4.5749485e-01 4.4096413e-01 -1.0663160e+00 5.7717276e-01
1.2621384e+00 -6.5556812e-01 -1.4779274e+00 3.6812633e-01
6.3978845e-01 -6.8279052e-01 -1.0881016e+00 4.1861090e-01
2.7135968e-01 -8.6425835e-01 4.6603182e-01 -8.2587457e-01
2.5673267e-01 -2.9349038e-01 -2.6887146e-01 -1.0784045e+00
-2.7252123e-01 -9.2154431e-01 1.6152740e-01 1.8162302e+00
3.9838243e-01 -2.6827464e-01 3.8688242e-01 1.2954872e+00
-3.5540178e-01 -1.3347448e-01 -8.7371778e-01 -4.7711250e-01
2.4548122e-01 -2.0661008e-01 4.7958475e-01 1.4726624e+00
8.7105262e-01 6.0656905e-01 2.8477177e-02 4.8552532e-02
3.2518172e-01 7.5531162e-02 8.8375896e-01 -1.6023285e+00
9.3678497e-02 -3.2405832e-01 3.0093816e-01 -8.8998038e-01
5.2294707e-01 -7.3307151e-01 3.7828986e-02 -1.4288282e+00
3.6889440e-01 -4.8149475e-01 1.8107970e-01 5.5434221e-01
-3.7272850e-01 4.8247006e-02 7.2419904e-02 5.6943506e-01
-9.5705217e-01 7.5569622e-02 1.0100622e+00 9.9821590e-02
-1.9042240e-01 -5.2470112e-01 -1.2604370e+00 1.1470394e+00
5.1819795e-01 -5.3557003e-01 3.6339737e-02 -8.2883751e-01
4.8085499e-01 -2.1912989e-01 8.7593518e-02 -7.3887318e-01
6.0998422e-01 -3.3505449e-01 -2.1570590e-01 -2.1006159e-01
-7.7328630e-02 -6.1427337e-01 4.8984751e-02 -9.6460760e-02
-4.4494349e-01 -2.5315506e-02 2.4642505e-01 1.1721841e-01
-3.0587086e-01 -6.6018826e-01 3.7868908e-01 -2.8157908e-01
-7.3488426e-01 -2.3179881e-01 -2.4253854e-01 9.3032825e-01
7.8850120e-01 3.8547467e-02 -8.5671824e-01 -3.5565734e-01
-7.5369400e-01 4.0691355e-01 7.1232373e-01 5.6332159e-01
-8.5140176e-02 -1.0792446e+00 -8.4553844e-01 -5.7821226e-01
2.6895496e-01 5.7581175e-02 1.0166292e-01 7.0284957e-01
-2.8020379e-01 5.4921240e-01 -7.5938851e-02 -2.8701991e-01
-1.1507514e+00 3.3993313e-01 2.4894829e-01 -2.1380173e-01
-4.3955925e-01 7.4833262e-01 -2.4801159e-01 -6.5455097e-01
3.4188911e-01 7.4022949e-02 -5.9963101e-01 4.0283361e-01
5.7845312e-01 3.6260787e-01 -1.2685686e-01 -1.0214082e+00
-5.0364029e-01 4.7478214e-01 -6.0845107e-02 -2.3488402e-01
1.1194026e+00 -3.8119459e-01 -1.2695219e-01 5.3911394e-01
6.2752599e-01 5.4841584e-01 -9.0064961e-01 -3.3366659e-01
5.3944343e-01 -2.9291213e-01 -4.8913151e-01 -9.1525245e-01
-4.3891519e-01 5.0068998e-01 1.5183238e-02 3.3197060e-01
5.0148118e-01 3.0995408e-01 6.4774060e-01 4.2558819e-01
5.5954576e-01 -1.2637724e+00 -2.1962290e-01 8.5328358e-01
9.3954331e-01 -1.2647489e+00 -9.8592006e-02 -4.8606417e-01
-8.2462633e-01 9.9629653e-01 8.8751644e-01 2.8819799e-01
1.6382344e-01 2.8983635e-01 6.0099381e-01 -3.6685875e-01
-7.8960609e-01 -1.9689517e-01 1.2499964e-01 7.1274239e-01
9.8609048e-01 -1.0090821e-01 -8.4812832e-01 1.1974725e+00
-4.1814625e-01 -2.0064136e-01 6.5019870e-01 7.7607185e-01
-3.9785695e-01 -1.2639663e+00 -5.2104831e-01 -1.0018741e-02
-7.7332664e-01 -1.0760381e-01 -8.8074940e-01 8.5394299e-01
1.8636477e-01 1.4340653e+00 1.1340065e-02 -1.7912520e-01
6.0448098e-01 1.2865365e-02 2.0897532e-01 -7.8121334e-01
-9.9177569e-01 -2.7345687e-01 7.3374647e-01 -2.8941116e-01
-9.8649132e-01 -8.6826843e-01 -1.1693707e+00 -3.3488590e-01
-3.9368826e-01 8.4848446e-01 3.9586734e-02 1.3347484e+00
2.2463793e-01 -7.2853364e-02 -4.4377379e-02 -3.9991337e-01
-3.2649767e-01 -9.8581702e-01 -3.8316470e-02 9.2590231e-01
-1.3591288e-01 -8.3850831e-01 -4.3004885e-01 1.3227427e-01] | [9.36636734008789, 9.349306106567383] |
c12653de-e131-445b-9a66-916279433008 | interactive-natural-language-based-person | 2002.08434 | null | https://arxiv.org/abs/2002.08434v1 | https://arxiv.org/pdf/2002.08434v1.pdf | Interactive Natural Language-based Person Search | In this work, we consider the problem of searching people in an unconstrained environment, with natural language descriptions. Specifically, we study how to systematically design an algorithm to effectively acquire descriptions from humans. An algorithm is proposed by adapting models, used for visual and language understanding, to search a person of interest (POI) in a principled way, achieving promising results without the need to re-design another complicated model. We then investigate an iterative question-answering (QA) strategy that enable robots to request additional information about the POI's appearance from the user. To this end, we introduce a greedy algorithm to rank questions in terms of their significance, and equip the algorithm with the capability to dynamically adjust the length of human-robot interaction according to model's uncertainty. Our approach is validated not only on benchmark datasets but on a mobile robot, moving in a dynamic and crowded environment. | ['Wei-Lun Chao', 'Vikram Shree', 'Mark Campbell'] | 2020-02-19 | null | null | null | null | ['person-search'] | ['computer-vision'] | [ 1.57032624e-01 5.27092576e-01 3.32941979e-01 -3.94463211e-01
-3.77927423e-01 -4.38203752e-01 7.69559503e-01 3.02505612e-01
-6.19298458e-01 5.34143507e-01 6.10969588e-02 -2.01535691e-03
-3.03234577e-01 -6.04143441e-01 -3.26713592e-01 -2.75486648e-01
5.28381616e-02 1.24194741e+00 3.60287100e-01 -2.89063752e-01
4.07265335e-01 5.46906054e-01 -1.89497530e+00 -1.31329382e-03
7.33956337e-01 4.91126299e-01 5.93737245e-01 5.74694753e-01
-7.87880421e-02 6.70163214e-01 -4.65301722e-01 -2.11044461e-01
-1.38283772e-02 -4.70946789e-01 -1.30331373e+00 5.60380816e-01
-4.48707752e-02 -2.77126674e-02 8.41831118e-02 8.81140411e-01
4.01070774e-01 4.02782828e-01 8.76805961e-01 -1.19893646e+00
-3.10030222e-01 2.61031389e-01 5.25871180e-02 -4.01125252e-02
1.09780490e+00 3.62434804e-01 8.83017123e-01 -8.46634090e-01
9.06962454e-01 1.57017326e+00 1.82251737e-01 6.22665226e-01
-9.63563502e-01 2.79331684e-01 1.66166484e-01 6.56980753e-01
-1.58344805e+00 -3.57799083e-01 6.42562032e-01 -5.34195781e-01
6.50736570e-01 1.84966877e-01 7.15210319e-01 8.30372095e-01
-1.66485086e-01 7.78748393e-01 7.15276837e-01 -7.79492438e-01
6.20036960e-01 5.08107960e-01 1.16238356e-01 7.52481520e-01
1.66853905e-01 -1.45606279e-01 -6.49990499e-01 -2.80501902e-01
3.88944685e-01 -4.95309561e-01 -1.89601287e-01 -9.45151627e-01
-1.11439049e+00 6.22797132e-01 2.75728732e-01 4.62325156e-01
-6.69375718e-01 -3.25867683e-02 6.26052096e-02 1.95554763e-01
-7.13586435e-02 7.72133946e-01 -2.60722071e-01 -8.92890096e-02
-1.35982022e-01 4.94211435e-01 1.13895285e+00 9.89862740e-01
1.00022507e+00 -7.09541082e-01 -1.83955386e-01 5.85500300e-01
6.34829164e-01 3.99532348e-01 4.64220166e-01 -1.06587160e+00
1.25011146e-01 8.20790350e-01 7.12787628e-01 -1.18550944e+00
-6.60339534e-01 2.27594767e-02 -3.14717233e-01 7.64183924e-02
4.52030927e-01 4.65276875e-02 -5.06169796e-01 1.44578600e+00
6.43443346e-01 -5.13904691e-01 1.99282438e-01 1.08819485e+00
5.35178661e-01 4.47075427e-01 8.06298181e-02 -6.16192780e-02
1.77127767e+00 -8.39758873e-01 -5.01647711e-01 -5.01909912e-01
5.30103207e-01 -2.32448637e-01 1.26231241e+00 4.87202883e-01
-9.13819313e-01 -5.92426419e-01 -6.15823984e-01 -1.36525676e-01
-4.01305050e-01 3.85896385e-01 1.72963649e-01 3.10422182e-01
-1.10560560e+00 1.41839668e-01 -3.82682502e-01 -1.05024457e+00
-2.00956687e-01 3.71421725e-01 -3.47415060e-01 -5.97537905e-02
-1.08867347e+00 1.04255116e+00 5.27503490e-01 6.23897053e-02
-9.65143025e-01 1.36914760e-01 -8.79402459e-01 8.52497593e-02
5.91685891e-01 -1.17906904e+00 1.44771516e+00 -8.99667799e-01
-1.62255085e+00 1.02477098e+00 -3.43358934e-01 -4.95729357e-01
5.70570171e-01 -6.88954890e-02 1.21527784e-01 5.01268268e-01
2.70044863e-01 8.61926198e-01 7.67381787e-01 -1.73755658e+00
-6.74081087e-01 -5.95918000e-01 4.85173613e-01 6.51374519e-01
-5.28374016e-02 -1.19800279e-02 -8.17730248e-01 1.35077775e-01
5.12722619e-02 -9.87506986e-01 -5.56183934e-01 1.66912396e-02
-2.02261344e-01 -5.92292845e-01 3.61464530e-01 -4.09845084e-01
7.76995778e-01 -1.87147570e+00 4.30297941e-01 3.54862273e-01
1.21311396e-01 9.84185040e-02 -2.77540851e-02 5.02819479e-01
5.85646927e-01 -1.92704424e-01 -3.34648043e-01 -6.16970479e-01
2.40706846e-01 2.87654966e-01 2.31796518e-01 1.85079917e-01
6.02998538e-03 7.52211273e-01 -1.06135750e+00 -8.49877656e-01
2.46624604e-01 1.16471626e-01 -3.74212533e-01 5.51885307e-01
-4.65188086e-01 5.84080040e-01 -7.59119034e-01 4.69056994e-01
2.55225956e-01 -3.29280317e-01 2.96581209e-01 3.28019291e-01
-5.15608974e-02 -1.13027386e-01 -1.15815485e+00 1.41354406e+00
-4.90061462e-01 3.20879310e-01 2.29986459e-01 -7.90392697e-01
1.13725996e+00 8.97216797e-02 9.63177457e-02 -6.10552430e-01
1.95065364e-01 1.67427078e-01 -3.68070662e-01 -9.69201684e-01
5.42924225e-01 7.46439174e-02 -3.07236582e-01 3.74153674e-01
-2.04598904e-01 -1.54790670e-01 2.95090765e-01 -3.03842537e-02
1.09659505e+00 -4.75984029e-02 6.45140111e-01 -1.41852051e-01
9.98759568e-01 3.02860588e-01 -3.71898301e-02 1.05992591e+00
-3.55295032e-01 3.20285738e-01 3.24339330e-01 -6.43224418e-01
-7.80306458e-01 -9.22817111e-01 4.57370013e-01 1.11308050e+00
5.42132914e-01 -1.67822912e-01 -8.24230909e-01 -5.37195623e-01
-3.53399932e-01 8.25565219e-01 -6.07865691e-01 -4.98288590e-03
-4.06178862e-01 -1.86567947e-01 1.33229479e-01 -1.24177091e-01
4.42589343e-01 -1.51900184e+00 -1.49653876e+00 1.27283767e-01
-5.05746126e-01 -9.51183021e-01 -9.97545049e-02 -8.40660185e-02
-3.45106453e-01 -9.36638951e-01 -6.63124800e-01 -9.06808734e-01
1.00881362e+00 1.95382267e-01 1.20379102e+00 2.30819091e-01
-2.18758047e-01 1.22302127e+00 -6.38585746e-01 -2.93246180e-01
-5.34058630e-01 1.45499334e-01 -9.33056406e-04 8.62272307e-02
5.08403897e-01 -2.57169724e-01 -4.86612141e-01 5.67782760e-01
-7.43543267e-01 -6.18681833e-02 6.29049063e-01 5.01718819e-01
3.81789744e-01 -4.74184789e-02 3.28989595e-01 -4.86474812e-01
8.85403037e-01 -5.25896847e-01 -5.64616323e-01 5.41252851e-01
-5.24049044e-01 3.29222709e-01 4.25369292e-01 -3.17236423e-01
-1.17287147e+00 5.18150210e-01 5.72315753e-02 5.99424914e-02
-6.32441878e-01 3.29134852e-01 -3.53261590e-01 5.74861839e-03
8.07818830e-01 1.65985301e-01 -5.78978732e-02 -3.73976767e-01
4.73933905e-01 6.81863785e-01 5.93759000e-01 -4.89478886e-01
6.84500396e-01 5.30450225e-01 -1.23858817e-01 -9.65038180e-01
-3.72422308e-01 -8.36606205e-01 -8.24752212e-01 -4.78556812e-01
7.85519063e-01 -4.97256279e-01 -1.19033539e+00 2.07184553e-01
-1.50013721e+00 -3.22136849e-01 -5.12594357e-02 2.38112301e-01
-9.19938803e-01 5.78622997e-01 1.07181007e-02 -1.13829911e+00
-8.84162188e-02 -8.27411115e-01 1.09817719e+00 2.93042243e-01
-3.40633422e-01 -7.86773860e-01 8.66544098e-02 4.74742353e-01
2.15389907e-01 -1.76777944e-01 6.92488968e-01 -8.27821612e-01
-6.31053984e-01 1.58835463e-02 -1.68291122e-01 -2.58332908e-01
-1.14338093e-01 -5.58438838e-01 -6.08626306e-01 3.30666676e-02
-1.35238422e-02 -1.92990869e-01 3.64303976e-01 9.39288549e-03
5.91015875e-01 -3.10392886e-01 -4.65041190e-01 -9.48447362e-02
1.06497574e+00 2.21264377e-01 4.64021206e-01 5.85217535e-01
2.27845654e-01 1.25370252e+00 1.07295752e+00 7.38256156e-01
8.64299536e-01 9.48539972e-01 5.90551913e-01 2.47148320e-01
3.11465353e-01 -3.00059170e-01 1.25685528e-01 -1.34663805e-01
1.46522624e-02 -3.89686167e-01 -1.16083872e+00 6.94633961e-01
-2.03437781e+00 -5.96537769e-01 -1.62712522e-02 1.98574042e+00
3.26618433e-01 -1.41535193e-01 2.31843174e-01 -8.31863061e-02
6.36595190e-01 -1.26985103e-01 -3.18932682e-01 -2.19567567e-01
3.29674274e-01 -5.93054891e-01 4.49129380e-02 7.44413018e-01
-8.39881659e-01 1.07724535e+00 5.83790255e+00 2.33779818e-01
-4.13593739e-01 -9.60548744e-02 3.29401225e-01 4.92551386e-01
-2.71647900e-01 -3.58029567e-02 -6.66531503e-01 9.17561576e-02
7.02778161e-01 -1.45973086e-01 5.85123956e-01 8.84039342e-01
3.82796913e-01 -6.05566859e-01 -1.21147823e+00 9.71117973e-01
3.47594619e-01 -5.97420990e-01 6.40563741e-02 -1.06463857e-01
-1.04139084e-02 -5.35190105e-01 -1.53883368e-01 1.43219709e-01
1.39125660e-01 -8.64141703e-01 8.58174205e-01 8.74057829e-01
1.28121495e-01 -5.37776113e-01 8.04740787e-01 9.14796948e-01
-9.03394818e-01 -2.93656796e-01 -2.75033146e-01 -1.39573246e-01
3.91599536e-01 -5.19937128e-02 -1.52922237e+00 4.89074618e-01
7.47801542e-01 2.16918420e-02 -8.45692337e-01 1.08197927e+00
-2.92664468e-01 -9.56172645e-02 -3.03707212e-01 -6.37075722e-01
1.34011194e-01 -1.59658149e-01 8.03547800e-01 1.04684806e+00
3.71447831e-01 3.29210222e-01 1.15102939e-01 8.72294903e-01
4.63275552e-01 2.76288837e-01 -8.03812623e-01 3.61982048e-01
4.38587040e-01 9.40695822e-01 -7.33439267e-01 -2.43137255e-01
1.51965739e-02 1.25333095e+00 3.35753262e-01 2.72020698e-01
-4.45702493e-01 -1.62880361e-01 6.99404851e-02 2.08386630e-01
1.65482685e-01 -3.68193060e-01 1.50877357e-01 -6.69026077e-01
1.70799181e-01 -7.58529246e-01 4.13885713e-01 -1.14400458e+00
-9.03268397e-01 7.87546575e-01 3.29834968e-01 -1.00125635e+00
-8.05947006e-01 -5.77858508e-01 -1.54492199e-01 4.67516661e-01
-1.43930697e+00 -1.02052557e+00 -6.08610272e-01 5.57408869e-01
6.35862470e-01 -2.70862971e-02 6.34020984e-01 -1.25970468e-01
-1.68283135e-02 6.22197092e-02 -4.91227299e-01 -3.80346149e-01
4.27832216e-01 -1.17121649e+00 1.96426660e-01 4.90872324e-01
7.24395216e-02 5.85412025e-01 1.27723014e+00 -5.09396613e-01
-1.28921068e+00 -5.72646022e-01 1.25993729e+00 -6.33617163e-01
2.60516584e-01 -2.17415258e-01 -7.92056859e-01 4.71049517e-01
4.56419308e-03 -3.46236944e-01 3.75607401e-01 -6.36628196e-02
2.29865655e-01 1.93653554e-01 -1.31210887e+00 7.47283101e-01
1.15008962e+00 -2.41701409e-01 -9.92316127e-01 2.43835390e-01
5.80962181e-01 -1.38382733e-01 -2.51157939e-01 1.43615678e-01
2.96048611e-01 -1.03778899e+00 9.52667058e-01 -3.42404157e-01
-1.83898628e-01 -6.16385460e-01 2.76874229e-02 -1.11602104e+00
-1.81009874e-01 -6.46728694e-01 2.35014051e-01 9.22552526e-01
3.34204704e-01 -5.03383756e-01 6.52747393e-01 8.88629913e-01
1.59026027e-01 -2.59635031e-01 -9.06356752e-01 -4.09510016e-01
-7.70075917e-01 -2.10022077e-01 4.50885147e-01 1.80867627e-01
2.51999378e-01 4.44471538e-01 -4.21756059e-01 5.53111434e-01
3.20256770e-01 -1.98466450e-01 1.00932527e+00 -1.48220575e+00
-2.67746598e-01 -1.07040584e-01 -4.08475071e-01 -1.22510910e+00
2.65262932e-01 -2.94739395e-01 6.95635617e-01 -1.76315057e+00
2.10955113e-01 -1.62476391e-01 3.04813117e-01 1.67730182e-01
-4.09104601e-02 -2.83484727e-01 2.98469216e-01 3.37237865e-01
-1.26996207e+00 6.62657619e-01 8.01836193e-01 -5.92999645e-02
-4.40423459e-01 2.43414983e-01 -4.35122490e-01 9.18083727e-01
7.06810713e-01 -3.61220717e-01 -4.79884267e-01 -3.10623646e-01
4.87646461e-01 1.17583573e-01 7.40520477e-01 -1.16321421e+00
6.17458522e-01 -9.39095616e-02 -3.87879312e-02 -4.83487368e-01
4.62061942e-01 -1.07173574e+00 1.51087493e-01 5.36780596e-01
-4.06366497e-01 2.96883192e-02 -1.87873021e-01 7.28348434e-01
-5.75028732e-02 -8.89263451e-01 4.44075525e-01 -3.58494043e-01
-1.13441825e+00 -8.67769942e-02 -9.16678071e-01 -4.06626076e-01
1.12559617e+00 -1.46301508e-01 1.34963185e-01 -7.65868306e-01
-9.65806305e-01 6.66335583e-01 6.72561228e-01 3.16722035e-01
7.26155341e-01 -9.37970936e-01 -3.88799250e-01 -1.05091661e-01
5.57109833e-01 -1.31189868e-01 1.96467489e-01 3.06531370e-01
-6.29544914e-01 6.09086335e-01 -1.84823461e-02 -4.74876195e-01
-1.26679575e+00 8.79788756e-01 3.48668098e-01 -3.29490632e-01
-2.44574174e-01 4.67957258e-01 1.19262271e-01 -5.69482565e-01
4.26661551e-01 -1.85142100e-01 -8.48587871e-01 -3.67003120e-02
4.49100465e-01 3.24957907e-01 -3.98433834e-01 -8.54196191e-01
-4.63955194e-01 6.64496064e-01 3.63972276e-01 -5.45436919e-01
8.34168971e-01 -8.32630873e-01 -8.26382637e-02 4.29544985e-01
7.25301445e-01 -2.02524856e-01 -9.76410925e-01 -2.97866404e-01
5.68241537e-01 -4.42124009e-01 -6.84805870e-01 -6.01236522e-01
-8.49940032e-02 6.07321024e-01 6.07655764e-01 5.48776388e-01
9.25019085e-01 6.72521532e-01 2.41426587e-01 1.11016285e+00
8.58363688e-01 -1.04153800e+00 5.17409027e-01 5.21522999e-01
1.03668332e+00 -1.26192236e+00 -3.10915619e-01 -3.44776183e-01
-9.35133934e-01 1.09096313e+00 6.27686739e-01 4.01501685e-01
2.52920508e-01 -3.95409852e-01 2.17449553e-02 -5.37919760e-01
-7.14368641e-01 -6.86059952e-01 4.35726419e-02 8.89509141e-01
-2.82215506e-01 -4.46372591e-02 -4.32165414e-01 4.59820688e-01
-1.20378569e-01 2.53468771e-02 5.18351853e-01 8.20003688e-01
-8.67955863e-01 -8.98928821e-01 -6.29438281e-01 -2.33805910e-01
2.55783260e-01 5.09158254e-01 -7.99748480e-01 7.65964568e-01
1.59852639e-01 1.31451213e+00 -8.04335326e-02 -9.30621326e-02
5.98320484e-01 4.31416407e-02 1.26523167e-01 -6.44185662e-01
-2.03415379e-01 -4.48753774e-01 2.68291444e-01 -4.83287454e-01
-4.69638199e-01 -8.82962465e-01 -1.32628095e+00 4.83000875e-01
1.77878380e-01 4.26534235e-01 6.67581022e-01 1.03349352e+00
3.03504229e-01 -7.54573718e-02 6.02318883e-01 -9.29285526e-01
-4.34115261e-01 -7.98183322e-01 -2.29683772e-01 4.05979216e-01
2.28117377e-01 -7.27128863e-01 -4.43884671e-01 -1.02943303e-02] | [4.617668628692627, 0.8738678693771362] |
03271607-40ce-4e38-9c2a-6ed6ad48fe8c | differentially-private-hierarchical | 2302.00037 | null | https://arxiv.org/abs/2302.00037v2 | https://arxiv.org/pdf/2302.00037v2.pdf | Differentially-Private Hierarchical Clustering with Provable Approximation Guarantees | Hierarchical Clustering is a popular unsupervised machine learning method with decades of history and numerous applications. We initiate the study of differentially private approximation algorithms for hierarchical clustering under the rigorous framework introduced by (Dasgupta, 2016). We show strong lower bounds for the problem: that any $\epsilon$-DP algorithm must exhibit $O(|V|^2/ \epsilon)$-additive error for an input dataset $V$. Then, we exhibit a polynomial-time approximation algorithm with $O(|V|^{2.5}/ \epsilon)$-additive error, and an exponential-time algorithm that meets the lower bound. To overcome the lower bound, we focus on the stochastic block model, a popular model of graphs, and, with a separation assumption on the blocks, propose a private $1+o(1)$ approximation algorithm which also recovers the blocks exactly. Finally, we perform an empirical study of our algorithms and validate their performance. | ['Vahab Mirrokni', 'Vincent Cohen-Addad', 'Mohammad Mahdian', 'Alessandro Epasto', 'Jacob Imola'] | 2023-01-31 | null | null | null | null | ['stochastic-block-model'] | ['graphs'] | [ 1.17238089e-01 2.30589151e-01 -1.86646327e-01 -1.09913163e-01
-1.04768908e+00 -7.92157888e-01 -2.80060112e-01 5.82769990e-01
-5.14288008e-01 5.00955760e-01 -2.08215311e-01 -3.18743438e-01
-2.85876274e-01 -9.90635276e-01 -1.08093095e+00 -1.25827944e+00
-4.91245449e-01 6.04284406e-01 1.01040035e-01 3.38747889e-01
1.48394674e-01 3.93825114e-01 -1.46575630e+00 -2.88723465e-02
6.37063026e-01 7.88297653e-01 -3.57510239e-01 7.79806793e-01
1.61984324e-01 3.31349909e-01 -3.27268958e-01 -7.48780906e-01
5.82668424e-01 -6.56059980e-01 -1.00869751e+00 1.62874132e-01
9.93045885e-03 2.69465167e-02 -7.11839855e-01 1.47272909e+00
4.29154485e-01 -4.23667878e-02 7.71754086e-01 -1.59548926e+00
-3.85377049e-01 1.24373972e+00 -1.12644470e+00 -3.83534133e-02
1.15551818e-02 -1.05766259e-01 1.23350525e+00 -9.70318466e-02
4.10211205e-01 8.83354604e-01 6.21441722e-01 4.89044040e-01
-1.61243248e+00 -8.82044256e-01 -1.46975681e-01 -1.37713561e-02
-1.94707751e+00 -6.62167221e-02 6.42514288e-01 -3.79484445e-01
4.30087119e-01 4.79587585e-01 3.34065080e-01 4.84812170e-01
-2.22457215e-01 7.76225090e-01 1.30348563e+00 -5.78913331e-01
5.96217573e-01 -4.05355580e-02 2.11581573e-01 8.13744843e-01
6.29627287e-01 -1.70308605e-01 -2.16413781e-01 -6.25108659e-01
1.86381295e-01 2.71520108e-01 -3.47464561e-01 -4.85378385e-01
-6.11308098e-01 1.01975024e+00 3.61831725e-01 1.84246659e-01
6.80628493e-02 8.32596123e-01 3.72820467e-01 2.32384354e-01
3.42008859e-01 -1.47917271e-01 -3.90490472e-01 -4.06048223e-02
-1.00860131e+00 1.57911107e-02 9.11585271e-01 1.18504024e+00
9.14372146e-01 -4.75134909e-01 -5.08930944e-02 4.57250983e-01
1.33605942e-01 5.10299325e-01 -7.07581490e-02 -1.15249217e+00
1.81909859e-01 3.67582589e-01 3.56990248e-02 -1.12571299e+00
1.37175936e-02 -2.90486306e-01 -1.14757490e+00 -2.25666970e-01
4.82106924e-01 2.01609870e-03 -4.81487155e-01 1.99466264e+00
4.17695731e-01 -1.18160322e-01 -3.00977916e-01 4.83370066e-01
-6.14157654e-02 6.06437743e-01 -1.51118830e-01 -6.61582887e-01
1.29731131e+00 -6.59171224e-01 -1.66940108e-01 2.09300473e-01
9.93373990e-01 -2.44313568e-01 8.75629187e-01 4.47440952e-01
-1.19375408e+00 8.84375051e-02 -8.40348780e-01 4.23188023e-02
-9.98589396e-02 -2.65642107e-01 6.16821051e-01 1.19409990e+00
-1.12547493e+00 6.76421404e-01 -7.49505818e-01 -9.79748666e-02
7.63896823e-01 3.72319937e-01 -2.46671990e-01 -3.31852734e-01
-5.78126550e-01 -1.65392280e-01 1.36565909e-01 -2.40274966e-01
-8.69520545e-01 -5.71294308e-01 -5.58370352e-01 2.17728928e-01
3.78180116e-01 -5.04583955e-01 7.38048434e-01 -6.54628515e-01
-8.26084852e-01 1.05887711e+00 -1.66150033e-01 -7.59194851e-01
5.85921824e-01 4.65073735e-01 3.24043870e-01 5.76904178e-01
-8.51221234e-02 1.05996899e-01 6.11836433e-01 -1.30661130e+00
-8.39726508e-01 -9.35403347e-01 1.66491866e-01 -3.47557873e-01
-5.40570796e-01 -1.74026236e-01 -3.69012117e-01 -4.04272169e-01
2.49896035e-01 -1.22083735e+00 -6.56962097e-01 2.97148083e-03
-3.51209849e-01 -1.54474810e-01 2.40519807e-01 -3.75598282e-01
1.19142926e+00 -2.53581262e+00 2.38479137e-01 6.80318594e-01
8.21755111e-01 -2.22059712e-01 2.46468306e-01 7.01265335e-01
2.65450895e-01 5.41129172e-01 -7.06284523e-01 -5.53656816e-01
5.09127259e-01 1.19957730e-01 5.41601609e-03 1.07976842e+00
-8.62422109e-01 3.81932110e-01 -7.38352954e-01 -3.02951634e-01
-4.01242375e-01 1.31210238e-01 -6.34764612e-01 -2.04203665e-01
-6.86138645e-02 2.34506819e-02 -2.96364695e-01 3.19537252e-01
1.37216556e+00 -5.25156200e-01 6.66099787e-01 1.15843363e-01
3.27352732e-01 -3.61801296e-01 -1.31180060e+00 1.29305077e+00
-1.40219823e-01 3.56200725e-01 6.28121018e-01 -1.32479203e+00
4.19342935e-01 1.15635008e-01 6.94421589e-01 -9.68969241e-02
3.76884252e-01 3.89961749e-01 -2.86714166e-01 -5.24567068e-02
8.64677280e-02 -4.37714070e-01 -4.55200553e-01 8.62733245e-01
-2.99897581e-01 3.98953497e-01 -1.34617150e-01 5.63240409e-01
1.70670033e+00 -8.06657016e-01 -6.73544928e-02 -4.34422165e-01
1.04417264e-01 -3.53433073e-01 6.20078266e-01 1.03823185e+00
-4.51401412e-01 4.56680894e-01 9.63827312e-01 -1.90212931e-02
-1.01818824e+00 -7.54147291e-01 1.46301351e-02 1.09822917e+00
2.01974899e-01 -5.88404536e-01 -1.14714801e+00 -8.53959739e-01
1.18307613e-01 2.29156867e-01 -9.60169733e-01 -1.39787629e-01
3.78595404e-02 -8.59881818e-01 7.12857008e-01 5.41946948e-01
5.70069194e-01 -4.97418374e-01 -4.49412882e-01 -1.43204913e-01
-2.09615869e-03 -8.59699726e-01 -5.57716966e-01 2.20911771e-01
-7.79617906e-01 -1.01892054e+00 -7.02900589e-01 -7.64686525e-01
9.38400984e-01 3.88514370e-01 7.06963837e-01 3.24219495e-01
-2.58893251e-01 6.69124544e-01 -4.43850249e-01 1.24914870e-02
-2.57962465e-01 1.29779831e-01 -1.45084202e-01 3.64278182e-02
6.29219890e-01 -9.51344967e-01 -8.61143708e-01 -8.51890519e-02
-1.11981630e+00 -3.10950816e-01 4.01649117e-01 5.51275730e-01
9.05174136e-01 6.30730808e-01 6.43282458e-02 -1.18784368e+00
2.28085533e-01 -5.56124151e-01 -1.07532012e+00 -2.90700402e-02
-7.43203938e-01 1.31220981e-01 8.16388845e-01 -6.84078485e-02
-2.82844126e-01 4.29941088e-01 -1.31203607e-01 -5.42181492e-01
6.62548691e-02 3.37604493e-01 -2.77525693e-01 -1.29241645e-01
4.42815751e-01 4.31305736e-01 -3.95601280e-02 -4.20444071e-01
4.32692409e-01 7.21604824e-01 3.35270971e-01 -7.39691556e-01
9.07505691e-01 1.04344714e+00 4.15373564e-01 -7.39197791e-01
-4.73579615e-01 -4.19130504e-01 -3.36897641e-01 2.79287606e-01
4.50230509e-01 -7.89756536e-01 -1.34490669e+00 1.85709283e-01
-6.49592936e-01 -4.04858828e-01 -1.98967114e-01 8.41141492e-02
-6.02431953e-01 7.65657961e-01 -6.82920277e-01 -1.22573924e+00
-2.66919076e-01 -9.53775823e-01 6.73595965e-01 -1.46044672e-01
2.44132593e-01 -5.80110848e-01 -2.21094508e-02 3.95869941e-01
-1.00311749e-02 4.11267072e-01 1.05267382e+00 -6.93734169e-01
-6.19326234e-01 -3.03263813e-01 -2.72260487e-01 2.47139946e-01
-4.11183655e-01 -1.54688716e-01 -8.30846548e-01 -6.50778592e-01
-5.00788167e-02 -6.33274689e-02 1.05035841e+00 2.59530306e-01
1.86345780e+00 -8.02579343e-01 -5.68230093e-01 7.17584729e-01
1.69538188e+00 -8.01555961e-02 5.83365202e-01 -1.89572230e-01
5.69678664e-01 4.21496093e-01 1.96704715e-01 8.12184334e-01
4.80608374e-01 3.17492448e-02 3.46776634e-01 2.48324022e-01
3.99664879e-01 -2.79372364e-01 -1.20023172e-02 6.89675689e-01
2.10710660e-01 -2.85432339e-01 -6.88509166e-01 9.00451243e-01
-1.72107720e+00 -7.22530007e-01 -3.61882031e-01 2.66289067e+00
8.70669305e-01 1.78267080e-02 4.23808634e-01 4.73806888e-01
9.03190672e-01 -3.39265913e-02 -3.75014305e-01 -4.42048758e-01
-1.34019837e-01 5.79773962e-01 1.04850674e+00 3.51447135e-01
-8.64860773e-01 7.14480221e-01 5.45673609e+00 1.26880062e+00
-4.53333706e-01 2.80615717e-01 8.94761860e-01 -1.24233961e-01
-5.39322317e-01 1.23761661e-01 -3.01730663e-01 8.11772048e-01
1.17743421e+00 -5.54517448e-01 5.81494331e-01 1.05487323e+00
-3.14654291e-01 -8.25791955e-02 -1.27122903e+00 1.18526125e+00
-1.15059122e-01 -1.26406121e+00 -2.95185417e-01 7.53476560e-01
8.68338346e-01 -3.68928045e-01 1.94205537e-01 2.99298037e-02
6.15450025e-01 -8.85767937e-01 2.61353880e-01 -2.81635195e-01
8.13781142e-01 -1.45847571e+00 5.55665016e-01 4.66912657e-01
-1.32145500e+00 -4.45092946e-01 -5.55835187e-01 2.11122006e-01
-2.56568402e-01 8.12658966e-01 -9.90461409e-02 4.74436343e-01
1.10364330e+00 3.64781357e-03 -2.66037643e-01 8.51610184e-01
1.13392346e-01 7.69419372e-01 -6.66597545e-01 -1.52337894e-01
5.61033376e-02 -2.74114221e-01 2.68559545e-01 1.18802130e+00
4.53503877e-01 7.58471429e-01 1.49063468e-01 3.62741888e-01
-8.59968841e-01 2.52185881e-01 -6.24926507e-01 -2.56684572e-01
9.62407470e-01 9.45286632e-01 -1.05674422e+00 -2.74170637e-01
6.65481910e-02 1.36475646e+00 4.20315385e-01 7.53569007e-02
-9.70231712e-01 -7.37581253e-01 5.31159878e-01 1.79913953e-01
8.68722141e-01 -2.42610037e-01 -2.52831131e-01 -8.96586359e-01
9.42435414e-02 -8.58000159e-01 7.87413538e-01 -5.16061522e-02
-1.38776612e+00 1.42263338e-01 -1.97827831e-01 -6.98944092e-01
4.25627917e-01 -2.53098398e-01 -2.70228833e-01 3.49034876e-01
-8.40832472e-01 -6.41085446e-01 6.67954236e-02 6.19200826e-01
-3.44063103e-01 4.29128379e-01 6.39889419e-01 3.94385338e-01
-4.52856779e-01 1.21663451e+00 7.78282106e-01 9.12062451e-02
3.32572788e-01 -1.23184860e+00 -9.26616639e-02 8.52867126e-01
-4.20332048e-03 3.58993858e-01 8.44516635e-01 -2.05890492e-01
-1.80410028e+00 -1.08922863e+00 7.45479107e-01 -1.24419212e-01
4.75466907e-01 -6.32243752e-01 -7.36451626e-01 7.76970327e-01
-5.89310192e-02 1.84437871e-01 1.11069357e+00 -9.76553038e-02
-6.86938405e-01 -3.04877222e-01 -1.74713087e+00 6.14672482e-01
1.38411510e+00 -6.50003672e-01 2.20639318e-01 4.32448536e-01
8.57764065e-01 -2.98522152e-02 -9.53391314e-01 1.26145557e-01
3.25561076e-01 -1.01760185e+00 6.86636448e-01 -3.28637451e-01
2.26201370e-01 -8.46298262e-02 -4.51534718e-01 -7.29676247e-01
-2.89698243e-01 -1.11256301e+00 -4.47863668e-01 1.03039408e+00
3.87759596e-01 -6.47666872e-01 1.24699175e+00 5.27119398e-01
5.15735447e-01 -6.77729607e-01 -1.27656305e+00 -7.64197767e-01
5.05256891e-01 -3.35703939e-01 3.89553010e-01 8.60106587e-01
2.47225523e-01 4.07154635e-02 -2.64100552e-01 3.86305273e-01
1.08373988e+00 2.43242353e-01 8.35097253e-01 -1.09996915e+00
-7.05362260e-01 -4.77014631e-01 -4.46133137e-01 -1.01968884e+00
3.03194255e-01 -9.74228323e-01 1.79409772e-01 -1.07976615e+00
7.22835958e-01 -5.43352902e-01 -2.96731234e-01 2.74866045e-01
-9.63455252e-03 3.14492375e-01 2.44809035e-02 -5.08801714e-02
-7.57958591e-01 5.72991073e-01 3.93469661e-01 -7.27233440e-02
1.68656170e-01 3.16474363e-02 -9.41938996e-01 4.80923355e-01
6.84182525e-01 -8.29599500e-01 -2.17138603e-01 -2.73657709e-01
3.56978089e-01 4.15204354e-02 8.09480995e-02 -8.31833243e-01
3.08609039e-01 1.36608824e-01 -8.06723386e-02 -5.41406810e-01
6.08931519e-02 -9.39685702e-01 2.08333373e-01 8.42635870e-01
-3.99696410e-01 4.15212400e-02 -1.61342487e-01 1.16437757e+00
3.26438785e-01 -3.09092373e-01 1.04277885e+00 1.90887135e-02
2.31737927e-01 8.20132554e-01 -4.31495667e-01 1.99464858e-01
1.34456718e+00 -1.93514809e-01 -1.49475724e-01 -4.69820887e-01
-5.66794217e-01 3.10314834e-01 8.31232607e-01 -6.88576639e-01
2.69816548e-01 -9.64249492e-01 -6.38947427e-01 -9.84464660e-02
4.59423251e-02 -2.45642830e-02 5.55893302e-01 9.59784925e-01
-6.08477533e-01 4.80453633e-02 2.60893494e-01 -4.73675936e-01
-1.21935892e+00 1.04500866e+00 -1.65599473e-02 -1.11881994e-01
-4.90322620e-01 1.16676223e+00 3.42948027e-02 -4.69219163e-02
4.60242927e-01 2.22904578e-01 7.79909909e-01 -1.43943265e-01
4.28773195e-01 6.21308506e-01 -1.96532428e-01 -3.69426906e-01
-3.32317293e-01 1.86237112e-01 -9.16032866e-02 -5.25705636e-01
1.20846570e+00 -2.37403616e-01 -4.78838474e-01 1.81372702e-01
1.68627846e+00 3.15967381e-01 -1.10006309e+00 -2.39548936e-01
-6.97814301e-02 -6.41787410e-01 -2.23249599e-01 8.33324343e-02
-1.24996591e+00 9.78432536e-01 6.46034539e-01 7.17747748e-01
1.25253010e+00 2.99710184e-01 8.97939146e-01 2.60046810e-01
7.90832520e-01 -1.06531680e+00 -2.45162785e-01 3.51366289e-02
3.82036641e-02 -7.46900678e-01 -4.61633354e-02 -3.50576460e-01
-1.90019310e-01 6.81962848e-01 1.32978261e-01 -2.29941428e-01
9.98423159e-01 2.03895181e-01 -6.89105928e-01 -1.11629799e-01
-6.75350070e-01 -3.46743800e-02 -4.91896093e-01 4.66872871e-01
1.10635571e-01 2.45034933e-01 -7.49198973e-01 7.85481870e-01
-9.81031731e-02 -2.00676516e-01 4.56838518e-01 1.05644953e+00
-5.87829053e-01 -1.08547950e+00 -7.19094425e-02 3.99576396e-01
-7.76218891e-01 -9.57293287e-02 -4.29613769e-01 4.66874212e-01
-1.45253167e-01 8.96334171e-01 4.00561653e-02 -4.47152376e-01
-2.74897963e-01 -2.50262674e-02 6.01420760e-01 -2.93743342e-01
-4.83644754e-01 8.50659162e-02 -4.51105207e-01 -2.88897514e-01
-1.46305397e-01 -3.79199922e-01 -1.36204576e+00 -1.11571014e+00
-6.56906664e-02 6.89407885e-01 4.53211457e-01 7.19202876e-01
7.06870675e-01 -1.50036290e-01 1.25741637e+00 -2.79083461e-01
-6.41201437e-01 -6.22834563e-01 -1.02907014e+00 2.21831292e-01
-4.18843701e-02 -8.54930505e-02 -1.08526969e+00 -8.35223272e-02] | [6.704582691192627, 5.098775386810303] |
bf86c80d-5d36-472c-ba05-0b40aacae988 | fovea-foveated-image-magnification-for | 2108.12102 | null | https://arxiv.org/abs/2108.12102v2 | https://arxiv.org/pdf/2108.12102v2.pdf | FOVEA: Foveated Image Magnification for Autonomous Navigation | Efficient processing of high-res video streams is safety-critical for many robotics applications such as autonomous driving. To maintain real-time performance, many practical systems downsample the video stream. But this can hurt downstream tasks such as (small) object detection. Instead, we take inspiration from biological vision systems that allocate more foveal "pixels" to salient parts of the scene. We introduce FOVEA, an approach for intelligent downsampling that ensures salient image regions remain "magnified" in the downsampled output. Given a high-res image, FOVEA applies a differentiable resampling layer that outputs a small fixed-size image canvas, which is then processed with a differentiable vision module (e.g., object detection network), whose output is then differentiably backward mapped onto the original image size. The key idea is to resample such that background pixels can make room for salient pixels of interest. In order to ensure the overall pipeline remains efficient, FOVEA makes use of cheap and readily available cues for saliency, including dataset-specific spatial priors or temporal priors computed from object predictions in the recent past. On the autonomous driving datasets Argoverse-HD and BDD100K, our proposed method boosts the detection AP over standard Faster R-CNN, both with and without finetuning. Without any noticeable increase in compute, we improve accuracy on small objects by over 2x without degrading performance on large objects. Finally, FOVEA sets a new record for streaming AP (from 17.8 to 23.0 on a GTX 1080 Ti GPU), a metric designed to capture both accuracy and latency. | ['Deva Ramanan', 'Nicolas Cebron', 'Mengtian Li', 'Chittesh Thavamani'] | 2021-08-27 | null | http://openaccess.thecvf.com//content/ICCV2021/html/Thavamani_FOVEA_Foveated_Image_Magnification_for_Autonomous_Navigation_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Thavamani_FOVEA_Foveated_Image_Magnification_for_Autonomous_Navigation_ICCV_2021_paper.pdf | iccv-2021-1 | ['real-time-object-detection', 'small-object-detection'] | ['computer-vision', 'computer-vision'] | [ 3.98685485e-01 8.42692032e-02 4.53028753e-02 -3.08582872e-01
-4.23740536e-01 -2.94480592e-01 4.09799367e-01 1.39567316e-01
-7.91490197e-01 4.19668853e-01 -3.43417712e-02 -2.18029141e-01
4.84723330e-01 -8.61625969e-01 -1.14592266e+00 -5.02005816e-01
-1.09406695e-01 -2.28902251e-01 1.00684190e+00 -2.96989948e-01
4.28371549e-01 5.62058806e-01 -2.21378398e+00 4.03152287e-01
4.06895667e-01 1.18602133e+00 7.44887114e-01 9.50838029e-01
2.75488079e-01 7.86900282e-01 -4.81709421e-01 -7.41193537e-04
5.94794691e-01 1.68852378e-02 -6.92283690e-01 -1.22206762e-01
5.95797360e-01 -6.79097235e-01 -5.37167013e-01 1.03626406e+00
4.06747818e-01 1.77714497e-01 1.85451612e-01 -1.12335169e+00
-3.45932156e-01 6.47339821e-01 -1.01161849e+00 7.67335534e-01
-6.92866594e-02 5.87309420e-01 5.38665235e-01 -9.94971097e-01
5.53528368e-01 1.39416540e+00 3.91835153e-01 4.57638741e-01
-9.41969275e-01 -6.79062009e-01 2.33310282e-01 2.37669557e-01
-9.93904412e-01 -6.83892429e-01 3.26398849e-01 -1.77985549e-01
1.02830958e+00 2.25615650e-01 6.99325860e-01 6.22096121e-01
4.35001343e-01 6.51258707e-01 6.99872673e-01 2.56271064e-02
3.79823059e-01 -1.95255235e-01 1.51913380e-03 4.60487783e-01
2.41653815e-01 6.85565695e-02 -8.23378146e-01 3.52677137e-01
9.35180128e-01 1.83961168e-01 -2.08420441e-01 7.10886484e-03
-1.50658846e+00 3.64257216e-01 7.87508428e-01 -2.32333094e-01
-6.42172635e-01 5.01007557e-01 4.33626771e-01 1.30081609e-01
3.41301173e-01 4.12316710e-01 -4.28367943e-01 -1.60061672e-01
-9.76955175e-01 1.51455387e-01 2.89516866e-01 1.12476909e+00
1.08398569e+00 7.88417086e-02 -1.71393245e-01 5.17607331e-01
4.87740934e-02 6.19792163e-01 3.23318154e-01 -1.27208292e+00
2.79162586e-01 4.48069900e-01 1.25860095e-01 -8.18224251e-01
-4.92795914e-01 -3.55855525e-01 -6.75114393e-01 6.32695138e-01
4.06365693e-01 -1.27736613e-01 -1.01536882e+00 1.43164861e+00
5.20445466e-01 2.99632549e-01 1.14014156e-01 1.27318323e+00
6.25285089e-01 7.87309051e-01 8.90427306e-02 1.35012358e-01
1.62349701e+00 -1.06876373e+00 -1.77461892e-01 -7.02313125e-01
3.96813512e-01 -7.11573362e-01 1.21379185e+00 2.74977773e-01
-1.27083468e+00 -6.67435884e-01 -1.18031371e+00 -4.58386511e-01
-1.76094607e-01 -1.89362988e-01 4.76169527e-01 1.41584277e-01
-1.26427412e+00 5.78054607e-01 -8.80500197e-01 -3.66174996e-01
7.45705724e-01 3.67035985e-01 -1.68045476e-01 -1.40779838e-02
-8.61321211e-01 7.43573904e-01 3.65715802e-01 -1.29953157e-02
-1.15905666e+00 -9.86191988e-01 -8.28925312e-01 1.89071104e-01
3.03813010e-01 -7.38124728e-01 1.34321511e+00 -1.10228920e+00
-1.38561511e+00 7.33580291e-01 -2.27759391e-01 -9.76334810e-01
5.57867050e-01 -3.59639078e-01 3.18130441e-02 3.79562646e-01
1.47662386e-01 1.37353110e+00 1.32013583e+00 -8.02426219e-01
-1.21978855e+00 -3.65563631e-01 2.37890899e-01 3.69408339e-01
-2.36111730e-01 1.30850017e-01 -7.13965952e-01 -4.40346003e-01
-8.40837955e-02 -8.88990998e-01 -4.69976068e-01 3.12030435e-01
-3.30572424e-04 1.03846624e-01 1.12861252e+00 -3.99659067e-01
6.69098854e-01 -2.28265262e+00 -1.96855307e-01 -2.04535887e-01
3.50025892e-01 1.98792413e-01 -1.70311615e-01 -2.25966275e-01
-4.63923533e-03 -2.31069118e-01 -2.64936555e-02 -1.80653542e-01
-5.19187808e-01 -6.49503469e-02 -5.25094092e-01 6.16046369e-01
5.05868256e-01 9.11546826e-01 -1.08421278e+00 -3.81681830e-01
4.88093972e-01 3.90085906e-01 -7.89312840e-01 6.87309504e-02
-2.50279248e-01 1.17884316e-01 -2.17781067e-01 4.86232817e-01
7.69129217e-01 -3.21832597e-01 -3.96388561e-01 -7.35906214e-02
-5.72023869e-01 2.50430405e-01 -9.33887005e-01 1.61047673e+00
-3.87237072e-01 1.06326735e+00 1.23381734e-01 -6.31663918e-01
7.78492153e-01 -3.60587418e-01 2.18240127e-01 -7.87565768e-01
2.26546139e-01 -5.28535768e-02 -1.30597830e-01 -3.01361144e-01
1.12408066e+00 3.11737716e-01 7.42001310e-02 1.42686620e-01
-4.33592081e-01 -2.61175018e-02 1.76204473e-01 2.82920271e-01
1.33446836e+00 6.48000091e-02 1.56243704e-02 -3.53839844e-01
1.85817912e-01 4.98455167e-01 3.94024104e-01 6.60677493e-01
-3.56811821e-01 7.79136717e-01 4.10283536e-01 -5.78403711e-01
-1.28031921e+00 -8.57680082e-01 -3.35875899e-02 1.47909665e+00
6.75290823e-01 -1.78631529e-01 -8.52508545e-01 -1.76344454e-01
3.86505853e-03 5.30259669e-01 -5.26928782e-01 -2.95426160e-01
-8.20272207e-01 -4.51893002e-01 3.00305277e-01 6.02843225e-01
5.91258824e-01 -1.22721505e+00 -1.75482547e+00 2.15133592e-01
-2.55073085e-02 -1.03284323e+00 -4.44869250e-01 2.45976344e-01
-8.65333974e-01 -7.98057437e-01 -7.25709140e-01 -7.83176780e-01
6.73703194e-01 1.05617106e+00 9.26485717e-01 3.79999839e-02
-3.51961106e-01 -1.06460422e-01 -2.12513641e-01 -5.37970960e-01
-2.04092920e-01 8.85407850e-02 -6.19773418e-02 -2.41760492e-01
2.00061381e-01 -2.54742324e-01 -9.76005018e-01 3.10622841e-01
-8.69076431e-01 4.75915402e-01 6.60133600e-01 5.63313603e-01
6.93145216e-01 -2.41753727e-01 4.93995816e-01 -4.41111118e-01
3.48220319e-02 -4.89643544e-01 -9.26734984e-01 -4.41334665e-01
-2.16808289e-01 7.75045389e-03 8.19613099e-01 -4.76403773e-01
-7.82125592e-01 2.02672526e-01 1.42660186e-01 -8.21906745e-01
1.72018353e-02 -2.93157976e-02 2.38145903e-01 1.53312251e-01
8.95807087e-01 3.28156114e-01 1.14796676e-01 4.63308468e-02
4.54969406e-01 6.15493000e-01 9.35998142e-01 -2.36485489e-02
4.63411838e-01 8.97245765e-01 -1.26332089e-01 -8.95267010e-01
-5.43053150e-01 -3.77983332e-01 -2.20315486e-01 -3.88595223e-01
8.38382244e-01 -1.27715814e+00 -1.01355815e+00 3.73294085e-01
-1.04411530e+00 -6.82527363e-01 -4.18698996e-01 2.25862682e-01
-6.64318323e-01 -2.43923627e-02 -6.03926063e-01 -5.35792649e-01
-4.57644552e-01 -1.15671253e+00 1.32242823e+00 5.67835987e-01
-1.97680704e-02 -6.14614561e-02 -5.98827600e-01 5.50525030e-03
5.22295475e-01 -1.18826032e-02 3.37333411e-01 5.68199679e-02
-8.95602226e-01 1.14547201e-01 -6.92753196e-01 8.98398906e-02
-2.22579196e-01 -4.96498717e-04 -1.13771439e+00 -3.55942696e-01
-1.75620988e-01 -6.70746416e-02 1.14467621e+00 5.00222862e-01
1.29337335e+00 -1.18037254e-01 -4.12776589e-01 7.48363495e-01
1.25847125e+00 9.85561013e-02 7.05179811e-01 6.17996037e-01
5.41343987e-01 6.06268406e-01 1.13028252e+00 6.28098011e-01
4.34172660e-01 2.61370212e-01 8.63270521e-01 -1.55456841e-01
-3.95489186e-01 -1.33833617e-01 5.88220239e-01 1.18641421e-01
2.05294147e-01 -3.62101868e-02 -7.86553204e-01 7.88168788e-01
-1.79397523e+00 -9.73422885e-01 -1.00968353e-01 2.25848007e+00
6.78138793e-01 4.31167960e-01 1.47133976e-01 -1.28551424e-01
9.04107749e-01 2.21549705e-01 -9.89821076e-01 -3.16891253e-01
-1.05419785e-01 2.02139001e-03 1.08985674e+00 2.77147293e-01
-1.00292563e+00 1.32440698e+00 5.66168737e+00 5.87619781e-01
-1.58660960e+00 -1.12819120e-01 8.67198765e-01 -6.04501843e-01
5.93126528e-02 -5.65181971e-02 -1.00259507e+00 5.35690784e-01
1.09042478e+00 -3.10086071e-01 5.40070057e-01 1.22255993e+00
5.41373789e-01 -4.81898457e-01 -1.12957954e+00 9.63589013e-01
-1.48732781e-01 -1.66638780e+00 -1.89119518e-01 -8.91693309e-02
5.97754419e-01 6.42712414e-01 2.35554591e-01 1.45113975e-01
2.23553613e-01 -7.23829091e-01 1.18166363e+00 2.77938932e-01
8.65575254e-01 -7.89727509e-01 4.34403479e-01 3.61688226e-01
-1.05934739e+00 -2.11394519e-01 -8.48604262e-01 -2.78042197e-01
1.38899639e-01 6.41871393e-01 -1.03660190e+00 -2.18262777e-01
1.27756882e+00 7.97122300e-01 -5.44989586e-01 9.71369684e-01
-7.35370023e-03 2.67095357e-01 -5.23222625e-01 -5.05089993e-03
2.61117697e-01 2.50900000e-01 6.08909726e-01 1.21242809e+00
3.04561913e-01 8.14872608e-02 -1.63057268e-01 7.64883399e-01
-2.10120142e-01 -2.50513911e-01 -5.28089881e-01 4.90214020e-01
6.60147130e-01 1.44385946e+00 -9.44804788e-01 -7.75672555e-01
-3.57790023e-01 9.62578356e-01 2.44590551e-01 1.75870091e-01
-1.02483857e+00 -4.93271649e-01 1.03947389e+00 4.26350623e-01
6.19006097e-01 -1.98018998e-01 -5.36048889e-01 -9.27578270e-01
1.54849803e-02 -5.93140185e-01 7.14768618e-02 -1.01962388e+00
-4.79746044e-01 5.79436302e-01 -2.39282936e-01 -1.30883884e+00
-1.11108467e-01 -3.58705223e-01 -3.86943996e-01 6.93013072e-01
-1.83278310e+00 -7.24769831e-01 -7.14407802e-01 4.20714319e-01
1.03512120e+00 3.55714083e-01 1.38806310e-02 1.30966887e-01
-6.10907376e-01 2.96765238e-01 -1.89177126e-01 -1.69944435e-01
8.72800171e-01 -1.02055299e+00 9.73358870e-01 1.17835605e+00
-4.48959410e-01 4.17568505e-01 8.84424567e-01 -5.92736423e-01
-1.63971758e+00 -1.66947675e+00 3.26047421e-01 -3.47120702e-01
5.36185324e-01 -2.79294580e-01 -8.20647061e-01 4.99667853e-01
1.20084055e-01 4.26874340e-01 -2.30968356e-01 -4.88135457e-01
-1.62045583e-01 -3.13602716e-01 -1.01876795e+00 9.30193901e-01
1.07666123e+00 -1.78343318e-02 -3.71471345e-01 6.48219287e-02
9.79755282e-01 -6.61579132e-01 -5.24265289e-01 1.43329009e-01
5.13886213e-01 -1.05822837e+00 9.14545953e-01 -3.04133505e-01
6.96976006e-01 -6.99541032e-01 1.10847913e-02 -1.01072073e+00
-2.99624324e-01 -7.06096470e-01 1.61158685e-02 6.38251722e-01
2.83926308e-01 -5.27619243e-01 8.45651031e-01 5.55768311e-01
-4.93296802e-01 -6.13536894e-01 -6.81289792e-01 -4.63172913e-01
-4.19157743e-01 -6.34626091e-01 5.91202497e-01 4.39693749e-01
-4.64139655e-02 3.10193628e-01 -8.78064483e-02 4.58262593e-01
5.41277230e-01 2.08516255e-01 1.03380477e+00 -7.50597656e-01
8.94541666e-02 -4.09306467e-01 -4.94143575e-01 -1.48433340e+00
-2.51742572e-01 -5.78210056e-01 3.72806698e-01 -1.08339942e+00
4.93041314e-02 -3.35230052e-01 -2.37763658e-01 5.10181308e-01
-2.37170875e-01 6.06352627e-01 3.25257689e-01 2.37061515e-01
-7.90081501e-01 2.77394742e-01 1.20403254e+00 3.70381400e-02
-2.75068432e-01 -3.30064267e-01 -7.75561571e-01 7.95566082e-01
8.10849845e-01 -3.88354063e-01 -2.69411087e-01 -6.21819794e-01
5.85385077e-02 1.49551090e-02 5.50110996e-01 -1.28271711e+00
3.86788189e-01 -1.50876760e-01 5.73312402e-01 -8.08204114e-01
3.99806201e-01 -5.26187778e-01 -1.59868270e-01 5.57034671e-01
-2.87003636e-01 1.80137262e-01 4.31113511e-01 4.96510923e-01
8.32005143e-02 2.02209592e-01 1.23092782e+00 3.69042112e-03
-1.17536628e+00 2.48187482e-01 -5.15257657e-01 -3.11530028e-02
1.26058578e+00 -4.70665634e-01 -5.87433040e-01 -1.39532104e-01
-1.91469163e-01 4.07083213e-01 9.63725865e-01 7.30014861e-01
7.88078427e-01 -8.34393978e-01 -6.39437139e-01 3.95164847e-01
4.79331613e-02 4.93592292e-01 3.37928712e-01 8.31832767e-01
-8.70297432e-01 1.65973365e-01 -4.37898934e-01 -9.14972007e-01
-1.14854133e+00 8.50029767e-01 -2.25972920e-03 5.32870591e-01
-9.45744753e-01 9.69508111e-01 6.44014180e-01 4.02402848e-01
1.18355989e-01 -7.97783852e-01 -9.88641828e-02 -1.38875112e-01
1.05076742e+00 4.65791196e-01 -4.79138866e-02 -5.89298606e-01
-4.15213972e-01 2.41422012e-01 -2.55774170e-01 2.51199659e-02
1.35806918e+00 -3.90179217e-01 -3.48523818e-02 8.58133063e-02
1.02130032e+00 -2.87365407e-01 -2.05125690e+00 -1.30828083e-01
-2.43042782e-01 -6.40186906e-01 3.10214698e-01 -1.87357664e-01
-1.14889610e+00 7.78606176e-01 5.60543656e-01 1.41683836e-02
1.23949242e+00 1.22221075e-01 9.54046249e-01 2.49848887e-01
5.07356942e-01 -1.03580904e+00 -2.28085611e-02 5.17805576e-01
8.21181595e-01 -1.24449348e+00 -8.22649524e-02 -3.42304170e-01
-6.01385236e-01 9.99157786e-01 9.43720102e-01 -4.21180218e-01
1.61094025e-01 5.40979564e-01 -5.27879409e-02 2.54905093e-02
-1.04277992e+00 -1.84130192e-01 -2.41316661e-01 4.25056189e-01
-7.43223503e-02 -1.16944917e-01 1.44010827e-01 1.87280506e-01
-4.84791338e-01 7.52208009e-02 9.30169404e-01 8.59688401e-01
-9.80436027e-01 -1.43185258e-01 -3.81091416e-01 5.30959070e-01
-2.73846477e-01 -1.84086502e-01 2.26898894e-01 4.39796925e-01
-2.57869288e-02 8.15218568e-01 5.62310696e-01 -2.82439291e-01
3.18794519e-01 -6.87828362e-01 4.48732264e-02 -5.67843318e-01
-4.98415619e-01 1.18877960e-03 -1.95150942e-01 -9.36894298e-01
-1.54834986e-01 -6.51061594e-01 -1.61763585e+00 -4.36604410e-01
-2.00184494e-01 -3.69166613e-01 7.36838043e-01 5.55001795e-01
6.84408784e-01 6.62573993e-01 4.34701771e-01 -1.39614141e+00
-3.55754644e-01 -7.05077827e-01 -1.86124593e-01 -5.51028084e-03
6.15788341e-01 -3.33246410e-01 -1.66975737e-01 3.68953317e-01] | [9.583905220031738, -0.47187942266464233] |
db374ece-d588-4fe6-b2e1-9020a184b6e9 | scalable-graph-neural-networks-via | 2010.15421 | null | https://arxiv.org/abs/2010.15421v3 | https://arxiv.org/pdf/2010.15421v3.pdf | Scalable Graph Neural Networks via Bidirectional Propagation | Graph Neural Networks (GNN) is an emerging field for learning on non-Euclidean data. Recently, there has been increased interest in designing GNN that scales to large graphs. Most existing methods use "graph sampling" or "layer-wise sampling" techniques to reduce training time. However, these methods still suffer from degrading performance and scalability problems when applying to graphs with billions of edges. This paper presents GBP, a scalable GNN that utilizes a localized bidirectional propagation process from both the feature vectors and the training/testing nodes. Theoretical analysis shows that GBP is the first method that achieves sub-linear time complexity for both the precomputation and the training phases. An extensive empirical study demonstrates that GBP achieves state-of-the-art performance with significantly less training/testing time. Most notably, GBP can deliver superior performance on a graph with over 60 million nodes and 1.8 billion edges in less than half an hour on a single machine. The codes of GBP can be found at https://github.com/chennnM/GBP . | ['Ji-Rong Wen', 'Xiaoyong Du', 'Ye Yuan', 'Yaliang Li', 'Bolin Ding', 'Zhewei Wei', 'Ming Chen'] | 2020-10-29 | null | http://proceedings.neurips.cc/paper/2020/hash/a7789ef88d599b8df86bbee632b2994d-Abstract.html | http://proceedings.neurips.cc/paper/2020/file/a7789ef88d599b8df86bbee632b2994d-Paper.pdf | neurips-2020-12 | ['graph-sampling'] | ['graphs'] | [-1.91213921e-01 2.01124221e-01 -3.92663419e-01 -2.24403456e-01
-4.33746189e-01 -3.19851011e-01 3.00034732e-01 2.30983570e-01
-2.87422806e-01 6.63298488e-01 -3.33996356e-01 -6.80236757e-01
-2.00678438e-01 -1.18727183e+00 -9.79556680e-01 -4.71466005e-01
-7.19200134e-01 4.67111588e-01 2.60956287e-01 -8.41371194e-02
5.68791814e-02 4.17088091e-01 -9.38658297e-01 -3.21660452e-02
6.88675702e-01 9.57035422e-01 1.09056700e-02 9.19657111e-01
-3.85309719e-02 6.87340677e-01 -2.79291719e-01 -5.70692182e-01
4.32864666e-01 -1.58295080e-01 -7.54783213e-01 -3.59747559e-01
6.54331744e-01 -2.56531239e-01 -9.95884776e-01 1.20293236e+00
5.28830945e-01 -1.69789139e-02 2.56989568e-01 -1.56562829e+00
-8.61265182e-01 9.28294718e-01 -5.69321215e-01 9.96424481e-02
-5.90873994e-02 -4.08636928e-02 1.17378199e+00 -8.97926867e-01
4.69429344e-01 1.00029683e+00 1.01026654e+00 4.75411147e-01
-1.27060449e+00 -7.91433036e-01 1.29667863e-01 2.44929656e-01
-1.52077842e+00 -1.09442966e-02 7.28560150e-01 -7.41266506e-03
1.02811491e+00 4.33778763e-02 8.77839208e-01 8.39070618e-01
2.05863267e-01 7.31727779e-01 7.41661072e-01 -3.25821429e-01
9.70156640e-02 -1.28702447e-01 1.98915973e-01 1.20687485e+00
6.57301128e-01 -7.29572680e-03 -5.35198510e-01 -1.03998542e-01
1.00319111e+00 -1.24826826e-01 -1.95490435e-01 -4.25991625e-01
-8.52406919e-01 1.07737744e+00 1.03865564e+00 -1.36924544e-02
-3.47233564e-01 7.45432198e-01 3.89365524e-01 5.10989010e-01
5.64899981e-01 2.65271962e-01 -4.37708110e-01 -8.21326226e-02
-7.83467174e-01 -9.40883309e-02 1.22491157e+00 1.23319983e+00
9.50675547e-01 1.31216988e-01 3.98023218e-01 7.46080518e-01
1.29389390e-01 4.40198123e-01 7.03565702e-02 -6.04175031e-01
7.10986614e-01 7.78093874e-01 -5.25128722e-01 -1.42807364e+00
-5.35214722e-01 -6.26177907e-01 -1.31496799e+00 -1.16280928e-01
3.00911009e-01 -4.19248521e-01 -1.02793610e+00 1.51098347e+00
3.96264255e-01 2.28738055e-01 -3.09890121e-01 4.85379517e-01
8.97058368e-01 6.44401550e-01 -3.26305687e-01 3.60331327e-01
8.17896008e-01 -1.40886879e+00 -1.06429256e-01 -4.65174407e-01
8.82959485e-01 -2.64303118e-01 8.69288445e-01 3.15317094e-01
-9.95614529e-01 -2.24244222e-01 -1.15154421e+00 -6.70674965e-02
-6.53361976e-01 -1.26682119e-02 1.08750093e+00 6.45266533e-01
-1.59392798e+00 9.56777632e-01 -9.44782376e-01 -4.17586178e-01
5.96501946e-01 6.40899658e-01 -4.36914444e-01 -4.52171952e-01
-9.51925099e-01 4.29709166e-01 3.75884861e-01 5.38050756e-02
-8.64973903e-01 -5.80202103e-01 -8.09588969e-01 1.73216000e-01
3.29952866e-01 -5.41240215e-01 8.99676979e-01 -5.88096499e-01
-1.13937032e+00 2.38120869e-01 5.97682223e-02 -6.75817609e-01
3.29140604e-01 -8.66580382e-02 -2.83294708e-01 1.09518014e-01
-7.82474279e-02 6.17010057e-01 4.23432767e-01 -8.16163182e-01
-3.16056460e-01 -2.72369087e-01 5.63968681e-02 8.39616582e-02
-7.06281781e-01 -4.35488284e-01 -8.29528689e-01 -4.48806614e-01
2.47076496e-01 -1.09302270e+00 -3.78733397e-01 -4.28026393e-02
-7.50857830e-01 -3.55114520e-01 7.60176539e-01 -5.46533406e-01
1.11605036e+00 -1.92445838e+00 -1.27035186e-01 6.89740300e-01
7.18706250e-01 2.14410439e-01 -4.13996428e-01 8.27504039e-01
1.35329977e-01 2.07344085e-01 4.95263711e-02 -2.27730542e-01
5.89668937e-02 9.80204195e-02 2.08670452e-01 5.48448861e-01
-6.90872744e-02 1.11776257e+00 -9.85802352e-01 -2.37366527e-01
-7.52784386e-02 6.71770155e-01 -5.17443299e-01 -8.91082883e-02
-2.23143194e-02 2.85487510e-02 -1.47519216e-01 6.29713714e-01
6.30392551e-01 -1.02893662e+00 4.86512125e-01 4.10740674e-02
4.47641134e-01 3.56605589e-01 -1.11051047e+00 1.48282921e+00
-2.23125190e-01 8.25199783e-01 7.31784478e-02 -1.09427714e+00
9.09135699e-01 -7.91416541e-02 2.87183672e-01 -6.32562518e-01
1.12767041e-01 3.53973806e-01 -6.46481151e-03 -4.03412580e-02
4.19262618e-01 5.43271005e-01 2.51479089e-01 4.57555562e-01
3.03041458e-01 2.89814532e-01 6.26967430e-01 4.71568137e-01
1.64749932e+00 -3.59147310e-01 1.49110526e-01 -2.88403928e-01
2.16380525e-02 -2.00574264e-01 3.12359154e-01 8.70555341e-01
6.78021982e-02 2.18325227e-01 6.86968029e-01 -7.43389010e-01
-9.91351545e-01 -8.41280460e-01 3.25800329e-01 1.07662022e+00
2.41722781e-02 -7.51363456e-01 -7.23503470e-01 -8.00741613e-01
1.16601951e-01 4.11379486e-01 -5.99617004e-01 -1.10888042e-01
-4.71902996e-01 -7.95540452e-01 5.91461599e-01 6.21112823e-01
6.35842800e-01 -8.70513439e-01 -5.45030162e-02 1.99987143e-01
3.16549867e-01 -1.05954039e+00 -4.34220374e-01 2.77078778e-01
-1.16034091e+00 -1.22278059e+00 -6.18678093e-01 -1.11980689e+00
1.10834503e+00 3.46482724e-01 1.23284125e+00 4.68629628e-01
-1.54206038e-01 1.40601754e-01 -2.62753367e-01 -1.03105985e-01
-1.05131857e-01 6.31759644e-01 -9.18822810e-02 -3.81401360e-01
4.11887944e-01 -9.88454700e-01 -6.13246381e-01 5.96814565e-02
-4.53426421e-01 2.92434245e-01 8.61151814e-01 8.61436248e-01
4.49420780e-01 1.38206437e-01 4.56631452e-01 -1.20805633e+00
8.10707092e-01 -6.51731431e-01 -7.78220713e-01 1.38809234e-01
-1.16610587e+00 4.24826071e-02 7.97730148e-01 -4.48602378e-01
-1.13228068e-01 -2.76255518e-01 -4.23559286e-02 -3.39827329e-01
2.01929733e-01 8.30781102e-01 3.31773013e-01 -5.68787694e-01
7.34559536e-01 2.06770897e-01 1.89355500e-02 -3.40712041e-01
3.42667729e-01 2.31194034e-01 4.42350321e-02 -4.43149239e-01
8.82166326e-01 1.98150903e-01 3.44531447e-01 -7.62252629e-01
-4.64962095e-01 -3.08509618e-01 -3.97033542e-01 -2.38745719e-01
9.46877524e-02 -7.48144984e-01 -7.19533324e-01 4.87550050e-01
-8.06545675e-01 -8.82183433e-01 -4.16918099e-03 4.15995419e-01
-1.75830707e-01 2.51853228e-01 -1.02131367e+00 -3.68426591e-01
-7.95375943e-01 -7.42996037e-01 5.47015488e-01 2.17145577e-01
2.52784967e-01 -1.28217137e+00 -7.65448576e-03 -1.23495340e-01
7.65494227e-01 1.26799807e-01 8.30679774e-01 -7.51613617e-01
-7.88714826e-01 -6.55031025e-01 -6.34323418e-01 3.71487290e-01
-7.00131506e-02 -1.85945004e-01 -5.25331020e-01 -7.63636947e-01
-7.15725780e-01 -3.77871245e-01 7.66358078e-01 3.32096338e-01
1.23043144e+00 -5.90367317e-01 -4.39375490e-01 9.82219398e-01
1.77782059e+00 -1.45023063e-01 4.29642767e-01 1.16476350e-01
1.17206025e+00 1.77375689e-01 -8.15272331e-04 1.59425154e-01
5.54825187e-01 1.25439957e-01 5.70211589e-01 -2.27141991e-01
-2.57314622e-01 -5.58317304e-01 1.55629128e-01 1.22160614e+00
-6.32078052e-02 -4.69731957e-01 -1.24859869e+00 5.84479749e-01
-1.79781818e+00 -6.03185117e-01 -5.57890497e-02 2.15345621e+00
5.71989179e-01 2.46068001e-01 4.94181700e-02 -1.59903355e-02
7.53291428e-01 2.80404836e-01 -6.24496043e-01 -2.11896524e-01
1.82023183e-01 2.39562035e-01 1.00317752e+00 5.04571855e-01
-9.39572275e-01 9.35004532e-01 5.94281054e+00 7.74237275e-01
-1.18281531e+00 -9.18162838e-02 6.72636509e-01 -3.36036906e-02
-1.11360632e-01 2.78884862e-02 -8.06967616e-01 3.13200027e-01
1.23754692e+00 -3.35879385e-01 8.86674583e-01 1.14157188e+00
-3.14369231e-01 3.06756467e-01 -1.00044453e+00 9.34526861e-01
-2.19336450e-02 -1.52468300e+00 5.67922592e-02 3.29108387e-01
8.94435346e-01 8.18596780e-01 -5.81002943e-02 3.97641271e-01
7.16737926e-01 -1.19549394e+00 1.42091051e-01 1.19522385e-01
8.99332643e-01 -8.74650121e-01 7.13400424e-01 2.36446366e-01
-1.46156371e+00 1.74348168e-02 -6.12704813e-01 -1.04638942e-01
-1.05289377e-01 8.46079946e-01 -1.11209023e+00 5.17824233e-01
8.42943609e-01 7.83958018e-01 -7.73333251e-01 1.15619791e+00
-3.39661211e-01 8.39724243e-01 -6.12049937e-01 -6.18511617e-01
4.33211505e-01 -2.10757270e-01 2.51083016e-01 1.16207016e+00
4.83967006e-01 -3.07483673e-01 2.56221056e-01 5.17013252e-01
-6.72292233e-01 2.26220638e-01 -9.42539096e-01 -3.56976569e-01
7.29218483e-01 1.31561983e+00 -8.71360421e-01 -2.90266991e-01
-5.37702143e-01 8.46524656e-01 9.45362091e-01 5.66641986e-01
-8.00596118e-01 -9.57636952e-01 1.95827499e-01 1.57799691e-01
5.23170173e-01 -4.78892773e-01 -4.79242206e-02 -8.01001966e-01
2.09731534e-01 -6.90517306e-01 3.35868388e-01 -3.83350343e-01
-1.23225129e+00 7.61584699e-01 -2.77072638e-01 -7.41635919e-01
-1.59647148e-02 -8.96117151e-01 -5.72419703e-01 6.88179970e-01
-1.31330645e+00 -1.08613658e+00 -5.41670442e-01 4.53341067e-01
8.12119916e-02 -1.15151800e-01 8.69104028e-01 4.94746894e-01
-5.03620148e-01 8.85962069e-01 2.22623453e-01 6.31823123e-01
2.17992112e-01 -1.41275692e+00 1.02778733e+00 9.79791701e-01
2.81787485e-01 5.63423574e-01 3.44562829e-01 -7.07723916e-01
-1.62394273e+00 -1.25874031e+00 8.36184859e-01 7.30658919e-02
7.32646406e-01 -4.70967561e-01 -7.47801483e-01 9.78706062e-01
1.68658942e-01 3.67691845e-01 5.09733677e-01 5.56969941e-01
-5.29742897e-01 -2.70537615e-01 -9.49860156e-01 7.39320338e-01
1.33044648e+00 -4.46119517e-01 3.13436836e-01 6.23096824e-01
6.18259549e-01 -5.73798776e-01 -1.01560628e+00 2.63824642e-01
3.86828870e-01 -7.99669564e-01 7.69585073e-01 -4.11231458e-01
2.51655757e-01 1.20636048e-02 1.25080645e-02 -1.37278819e+00
-4.82490003e-01 -7.37783015e-01 -5.89151919e-01 7.70815372e-01
7.37031341e-01 -1.12971914e+00 1.20882452e+00 2.49200702e-01
7.79353380e-02 -1.29823196e+00 -7.08496571e-01 -9.31209505e-01
-4.74696560e-03 -3.16108704e-01 5.77166677e-01 9.34516788e-01
-4.38180231e-02 4.67499524e-01 -3.08801860e-01 2.03348666e-01
7.31677830e-01 -1.23261213e-01 9.41759706e-01 -1.24872231e+00
-3.58409017e-01 -4.01278228e-01 -6.06051326e-01 -1.29594910e+00
-1.26922876e-01 -1.23943317e+00 -1.27208114e-01 -1.93567121e+00
4.59876060e-02 -5.77906191e-01 -2.89270580e-01 6.96292400e-01
1.10943131e-01 2.61374712e-01 -2.02685334e-02 -9.68202949e-02
-6.96167409e-01 3.59240204e-01 1.13118303e+00 -1.31138995e-01
-1.67741627e-03 -1.30456463e-01 -7.25296617e-01 5.18842876e-01
1.26484406e+00 -5.64445555e-01 -5.91025531e-01 -6.60747945e-01
6.51953757e-01 -2.58203864e-01 3.00498635e-01 -1.32602417e+00
6.84576154e-01 2.08227322e-01 3.41820568e-01 -5.04997134e-01
2.39952832e-01 -4.69181120e-01 5.79283237e-02 5.67588985e-01
-7.85940811e-02 5.94541490e-01 1.62788600e-01 7.73292124e-01
1.38617950e-02 1.02967381e-01 5.07692397e-01 -7.43433312e-02
-5.23248315e-01 8.59655678e-01 1.45874858e-01 9.63657051e-02
7.66461968e-01 3.29108015e-02 -6.58287525e-01 -5.03473222e-01
-3.98913801e-01 4.06414360e-01 3.40051413e-01 1.04483657e-01
7.08984792e-01 -1.37068808e+00 -6.55654788e-01 2.62073189e-01
-2.23302335e-01 2.78615952e-01 3.18487398e-02 9.27816808e-01
-1.00286233e+00 3.73271495e-01 6.62866456e-04 -3.59668970e-01
-1.18480074e+00 4.28065300e-01 3.35382193e-01 -4.92432594e-01
-9.21318471e-01 1.30280185e+00 -2.79873401e-01 -6.61447227e-01
4.78856117e-01 -2.12734137e-02 2.62699157e-01 -4.38136369e-01
3.14099044e-01 6.37453616e-01 3.56753357e-02 -1.12650990e-01
-2.07181767e-01 2.13210210e-01 -3.43485445e-01 2.50615329e-01
1.44923770e+00 2.84641027e-01 -2.55322903e-01 2.58376777e-01
1.41817963e+00 -1.26192406e-01 -1.20668161e+00 -3.20293725e-01
-3.08109820e-02 -2.70814031e-01 2.60392316e-02 -5.67234218e-01
-1.49966431e+00 8.18419397e-01 3.18325460e-01 6.53235137e-01
1.00225329e+00 -5.92834651e-02 9.69018459e-01 7.17126250e-01
6.70273125e-01 -9.74473894e-01 7.07513839e-02 7.23107874e-01
6.54424012e-01 -1.03507102e+00 1.54177040e-01 -3.98847103e-01
-4.07298505e-02 1.13434851e+00 6.96344733e-01 -5.36216855e-01
9.86401141e-01 2.02097520e-01 -1.79703102e-01 -5.19079089e-01
-8.87220800e-01 2.08130568e-01 1.72977850e-01 5.11648536e-01
3.45693827e-01 4.08303201e-01 -6.13945983e-02 9.79336649e-02
-3.41109902e-01 -2.73385674e-01 4.29374069e-01 7.49854207e-01
-2.62926221e-01 -1.06408048e+00 3.24604303e-01 9.15381432e-01
-1.96156129e-01 -4.96654958e-01 -3.04265261e-01 1.00004804e+00
-4.93437380e-01 6.58175468e-01 -8.31995010e-02 -7.92926610e-01
-2.05569901e-02 -1.10489897e-01 4.31846052e-01 -5.30949235e-01
-2.90448248e-01 -4.03592914e-01 2.54357189e-01 -9.35594618e-01
2.30674352e-02 -7.22110122e-02 -1.16497827e+00 -9.92920697e-01
-3.45947921e-01 1.79669291e-01 7.50184059e-01 3.16103786e-01
7.51412392e-01 5.08962452e-01 4.43939239e-01 -7.52468944e-01
-5.58846712e-01 -1.05907094e+00 -7.06353068e-01 1.85894612e-02
2.23627597e-01 -2.14728385e-01 -5.19674242e-01 -6.11730397e-01] | [6.997605323791504, 6.007991790771484] |
33ca33f7-a7fb-4372-8ec1-26f274efc917 | ensemble-model-with-batch-spectral | 2006.04323 | null | https://arxiv.org/abs/2006.04323v2 | https://arxiv.org/pdf/2006.04323v2.pdf | Ensemble Model with Batch Spectral Regularization and Data Blending for Cross-Domain Few-Shot Learning with Unlabeled Data | In this paper, we present our proposed ensemble model with batch spectral regularization and data blending mechanisms for the Track 2 problem of the cross-domain few-shot learning (CD-FSL) challenge. We build a multi-branch ensemble framework by using diverse feature transformation matrices, while deploying batch spectral feature regularization on each branch to improve the model's transferability. Moreover, we propose a data blending method to exploit the unlabeled data and augment the sparse support set in the target domain. Our proposed model demonstrates effective performance on the CD-FSL benchmark tasks. | ['Yuhong Guo', 'Jieping Ye', 'Zhen Zhao', 'Bingyu Liu'] | 2020-06-08 | null | null | null | null | ['cross-domain-few-shot', 'cross-domain-few-shot-learning'] | ['computer-vision', 'computer-vision'] | [ 1.41563594e-01 -3.76313776e-01 -3.21751326e-01 -4.52154100e-01
-9.01791871e-01 -4.53455865e-01 5.62137425e-01 -2.65578091e-01
-2.66474932e-01 7.44875133e-01 3.06341439e-01 5.78029361e-03
-2.16222093e-01 -4.02266026e-01 -4.06269610e-01 -6.09839618e-01
1.52731940e-01 3.80497903e-01 1.30855739e-01 -3.92058074e-01
-1.42746314e-01 1.50504932e-01 -1.36852336e+00 2.97624439e-01
1.12494338e+00 9.79950488e-01 -6.61590546e-02 1.81851268e-01
-1.99138820e-01 8.51017535e-01 -1.38285309e-01 -2.14918315e-01
5.28860509e-01 -4.00412112e-01 -5.32687545e-01 3.98291945e-01
4.68855709e-01 -7.01844767e-02 -4.77708340e-01 8.91738892e-01
4.76785958e-01 6.99207187e-01 8.52144837e-01 -1.55598986e+00
-6.08573437e-01 4.65447724e-01 -7.71434665e-01 2.90946692e-01
-3.15065607e-02 2.80682951e-01 1.21234953e+00 -1.38207948e+00
7.33668149e-01 1.04025984e+00 6.23750985e-01 6.20197773e-01
-1.41141534e+00 -9.04138029e-01 3.59535754e-01 4.04667050e-01
-1.34831488e+00 -8.33283246e-01 9.03539419e-01 -6.70278847e-01
8.49857271e-01 -1.44617736e-01 3.26817572e-01 1.36988199e+00
-3.61291379e-01 9.91505742e-01 9.59599972e-01 -4.84038711e-01
5.11079907e-01 1.02564089e-01 5.82743645e-01 6.09688878e-01
-5.70435412e-02 1.55927315e-01 -9.06978846e-01 -4.30400759e-01
4.64328915e-01 1.80206969e-01 -9.64882225e-02 -7.70893991e-01
-8.38780284e-01 9.35344934e-01 9.34690312e-02 2.59855151e-01
-2.74347216e-01 -2.11452842e-01 6.28195763e-01 4.23138559e-01
1.03544939e+00 1.10296100e-01 -4.57690567e-01 1.41986050e-02
-9.53916550e-01 4.63436730e-02 6.33698106e-01 1.19989383e+00
7.39085674e-01 1.52925774e-01 -5.07066846e-01 1.26790214e+00
1.13956369e-01 3.19707990e-01 4.92758393e-01 -9.06358063e-01
6.17479503e-01 5.64036727e-01 1.28098773e-02 -5.10770857e-01
-1.64308384e-01 -5.51640987e-01 -8.50534320e-01 -1.60772741e-01
2.34527811e-02 -3.06046218e-01 -9.04774547e-01 1.88188040e+00
5.87974250e-01 1.03710902e+00 2.50520080e-01 7.12299705e-01
7.07895577e-01 4.94733185e-01 1.52538493e-01 -3.61860871e-01
9.72912490e-01 -1.30545998e+00 -7.55111933e-01 -2.79616892e-01
9.69052672e-01 -2.84176350e-01 1.01228738e+00 2.36754030e-01
-5.32918453e-01 -5.04864037e-01 -1.19111204e+00 9.12716612e-02
-1.14439003e-01 2.44444951e-01 3.81827146e-01 4.05782282e-01
-4.97893959e-01 5.66790342e-01 -8.64864945e-01 -3.31752151e-01
6.98620260e-01 -7.40700290e-02 -4.50742394e-01 -5.12482584e-01
-1.11267626e+00 6.21811152e-01 6.95113182e-01 -3.21969002e-01
-7.54924595e-01 -1.09642220e+00 -8.85333896e-01 6.50314894e-03
5.35800636e-01 -6.29947245e-01 8.60602498e-01 -8.37480724e-01
-1.47473121e+00 6.22143090e-01 -1.45020410e-01 -5.07948756e-01
3.41288179e-01 -5.07925570e-01 -6.63907290e-01 -1.43755600e-01
1.29589975e-01 2.11125076e-01 8.80896032e-01 -1.03933668e+00
-5.59075654e-01 -4.87798184e-01 -4.65208828e-01 1.60808414e-01
-6.76870465e-01 -1.00222118e-01 -4.22508121e-01 -8.79972279e-01
-1.17635190e-01 -9.13325250e-01 -4.52757329e-01 -1.68586567e-01
-1.00736484e-01 -3.32800418e-01 9.27900851e-01 -5.71707845e-01
1.50393045e+00 -2.60649705e+00 3.00878048e-01 2.68017590e-01
2.18644559e-01 3.87374938e-01 -5.72832406e-01 3.95325541e-01
-1.16447248e-01 -4.38157380e-01 -5.18369675e-01 -6.03021204e-01
-1.38530182e-02 2.42104769e-01 -5.01259089e-01 4.54018444e-01
2.83383161e-01 7.04517066e-01 -8.89895320e-01 -3.17333132e-01
1.27072126e-01 3.22298139e-01 -6.58055604e-01 2.65273750e-01
-2.89725929e-01 3.86765003e-01 -4.17037368e-01 4.84966695e-01
7.20978022e-01 -4.64920044e-01 2.59926051e-01 -3.06192249e-01
1.40248612e-01 1.24055268e-02 -1.38173735e+00 2.22086525e+00
-4.20572728e-01 1.39115185e-01 -3.19600366e-02 -1.05142522e+00
9.75169241e-01 1.05412349e-01 7.47732460e-01 -3.95439357e-01
1.38536179e-02 1.67997435e-01 -1.22881562e-01 -1.44009262e-01
1.53892562e-01 -2.72342831e-01 -6.56761676e-02 5.73885560e-01
7.92622805e-01 2.74784386e-01 3.38165253e-01 4.77308333e-01
1.16066694e+00 2.89308786e-01 5.04043996e-01 -2.58755207e-01
4.88516778e-01 -5.75418472e-02 8.71228337e-01 6.62646711e-01
-3.06838185e-01 4.07534242e-01 1.23726964e-01 -2.38900244e-01
-9.98211205e-01 -9.16011691e-01 2.20927931e-02 1.66205120e+00
-1.46153048e-02 -6.61990047e-01 -3.13429683e-01 -1.01602948e+00
1.40696123e-01 9.64148760e-01 -6.00672245e-01 -5.13346195e-01
-1.34159192e-01 -6.38292491e-01 4.26023304e-01 5.95200241e-01
3.89476299e-01 -5.07803559e-01 -1.43857658e-01 3.70546728e-01
-2.21946239e-01 -1.05335379e+00 -7.58359432e-01 4.33341414e-01
-7.13700771e-01 -9.77522910e-01 -5.13106048e-01 -6.21979177e-01
2.67608374e-01 4.57830906e-01 9.34581161e-01 -5.08172154e-01
-2.91881084e-01 4.19145167e-01 -6.08258247e-01 -1.36164784e-01
-6.38687313e-02 2.17995390e-01 3.05210680e-01 4.13463861e-01
6.03910446e-01 -6.62894487e-01 -2.30408207e-01 3.09340000e-01
-7.32440412e-01 -6.96350485e-02 1.21296115e-01 1.35875797e+00
5.52768290e-01 -4.27779615e-01 7.96856761e-01 -1.07878804e+00
4.11531657e-01 -8.58009517e-01 -4.04521197e-01 5.68518996e-01
-6.07954562e-01 1.37012377e-01 4.86093432e-01 -4.73779619e-01
-1.29247046e+00 1.94440767e-01 2.77540416e-01 -1.02798247e+00
1.84270162e-02 5.06196499e-01 -1.86634898e-01 2.44162083e-02
9.31253672e-01 2.84351587e-01 -1.57647226e-02 -7.66608477e-01
6.13009632e-01 7.84573913e-01 2.89055556e-01 -7.08143473e-01
8.20802271e-01 4.49741393e-01 -3.08680534e-01 -6.91485405e-01
-1.41264427e+00 -9.28497732e-01 -7.77616382e-01 -5.80266491e-02
4.86868382e-01 -1.47482455e+00 -6.91344077e-03 4.45802182e-01
-8.13834965e-01 -1.90031365e-01 -4.57231283e-01 4.91010100e-01
-5.46726167e-01 2.93723673e-01 -4.07078356e-01 -7.17295408e-01
-4.25432593e-01 -6.60604537e-01 9.11148131e-01 -4.54441383e-02
-3.04227434e-02 -9.24859345e-01 6.47329569e-01 3.49961102e-01
3.65469724e-01 -1.21058345e-01 6.25086069e-01 -1.42728436e+00
-1.48372995e-02 -7.28677679e-03 -2.08184510e-01 5.32050908e-01
3.07258397e-01 -2.95136154e-01 -1.02640843e+00 -6.34685040e-01
-1.32807747e-01 -7.23754168e-01 1.15073645e+00 1.80061191e-01
1.05633652e+00 1.11429542e-01 -4.02620494e-01 7.60961831e-01
1.20140374e+00 -1.15432009e-01 8.89808759e-02 1.04056284e-01
7.22152948e-01 4.90760028e-01 7.81351984e-01 9.00457978e-01
2.29591757e-01 5.73998988e-01 -1.60224259e-01 1.72363967e-01
-8.61784816e-02 -2.42474765e-01 4.41074401e-01 1.09724939e+00
1.43160760e-01 -3.12919803e-02 -9.33054447e-01 4.49872434e-01
-2.23635149e+00 -9.97205019e-01 1.28902063e-01 1.90568781e+00
7.43886232e-01 -7.52077848e-02 2.43251115e-01 -3.12971860e-01
7.15269804e-01 2.96824813e-01 -8.52799594e-01 2.22489044e-01
-2.93736219e-01 3.10329020e-01 2.65723765e-01 2.32352033e-01
-1.43744063e+00 9.79033530e-01 6.24912024e+00 1.28342116e+00
-7.83788204e-01 4.11037207e-01 3.49683285e-01 -4.76110280e-01
-1.91990227e-01 2.71957763e-03 -9.47050691e-01 4.03975159e-01
9.21786070e-01 -5.55262923e-01 4.52611893e-01 8.40963900e-01
-1.61157265e-01 2.58213937e-01 -1.09371483e+00 1.00448418e+00
2.28884444e-01 -1.34223378e+00 -1.58625364e-01 -1.40718058e-01
9.46241796e-01 3.54654342e-01 -5.56099564e-02 9.03324068e-01
6.77579284e-01 -5.72642386e-01 2.74774849e-01 3.50499868e-01
8.37492347e-01 -7.11067498e-01 2.99858481e-01 5.74093282e-01
-1.24804163e+00 -2.86803484e-01 -2.76705623e-01 2.49033123e-01
1.04767278e-01 6.33609653e-01 -6.19503438e-01 9.67536628e-01
5.16328990e-01 1.26374388e+00 -6.03796661e-01 9.43891585e-01
2.69008309e-01 8.85353804e-01 -3.80029738e-01 4.16628003e-01
2.29238987e-01 -5.54339588e-01 6.31571710e-01 1.08912146e+00
3.62832576e-01 1.10083848e-01 5.46418428e-01 7.02281177e-01
-1.76195070e-01 2.64824688e-01 -7.15353012e-01 -1.16120480e-01
6.70363545e-01 1.11050117e+00 -5.49668893e-02 -4.29007888e-01
-7.53738880e-01 9.71577048e-01 7.12200463e-01 4.09076124e-01
-8.65828037e-01 -1.78670809e-01 6.80032849e-01 -3.29733133e-01
4.88044262e-01 -2.41892319e-02 -1.30101711e-01 -1.73044705e+00
-2.54576474e-01 -9.65106845e-01 8.24916720e-01 -3.20881009e-01
-2.05121517e+00 4.15884495e-01 -1.84732392e-01 -1.41968715e+00
1.49534701e-03 -3.55245769e-01 -5.75001061e-01 8.53545189e-01
-1.31591976e+00 -1.28692269e+00 -1.11723706e-01 9.53484535e-01
7.04182804e-01 -7.44707942e-01 7.14116752e-01 4.55863923e-01
-8.56246531e-01 7.22206116e-01 5.93202651e-01 -1.35069853e-02
1.11753631e+00 -8.14756393e-01 2.92936563e-01 7.24813581e-01
3.72877628e-01 5.55980384e-01 4.55892831e-01 -6.97328091e-01
-1.15640128e+00 -1.46115386e+00 2.35967517e-01 -2.91553259e-01
1.02437782e+00 -4.19508308e-01 -1.19492722e+00 9.08105254e-01
1.40255958e-01 4.20408666e-01 1.27209556e+00 4.71138269e-01
-8.72352123e-01 -1.04115352e-01 -9.97322500e-01 2.44079873e-01
1.18590438e+00 -6.01139843e-01 -7.59719849e-01 2.78499514e-01
7.75778115e-01 1.68497805e-02 -8.09193432e-01 5.17258883e-01
3.64332348e-01 -5.27668178e-01 8.13938379e-01 -1.26662290e+00
2.89317340e-01 -1.21616445e-01 -5.66977680e-01 -1.70992148e+00
-5.98554015e-01 -3.95018488e-01 -4.02575046e-01 1.41060281e+00
3.20256621e-01 -4.18922216e-01 6.54327631e-01 3.74792427e-01
-1.31200373e-01 -4.18475091e-01 -1.17500556e+00 -1.02874994e+00
-4.84646205e-03 -3.93939734e-01 2.75940418e-01 1.32574129e+00
1.07473351e-01 6.99640989e-01 -7.37458169e-01 -1.58021346e-01
1.03290772e+00 1.25533834e-01 7.88883030e-01 -1.40095186e+00
-5.72618425e-01 -1.22599021e-01 -3.62467803e-02 -7.91893184e-01
5.46628177e-01 -1.28676701e+00 -1.65432245e-01 -8.79295528e-01
4.76796091e-01 -3.09203804e-01 -8.36495697e-01 5.31056225e-01
-3.30660760e-01 -1.29200280e-01 3.77504915e-01 1.67777985e-01
-1.06887710e+00 1.02326679e+00 6.94184303e-01 -1.56313345e-01
-2.47502103e-01 -1.48969293e-01 -5.71413398e-01 4.86167073e-01
5.01761913e-01 -3.78556609e-01 -6.00431681e-01 -2.31657535e-01
-2.71670192e-01 -1.84332058e-01 6.90148026e-02 -9.17748094e-01
2.63884038e-01 -3.25573981e-01 2.09355906e-01 -5.68172216e-01
4.23330069e-01 -8.32964242e-01 8.93770531e-02 1.91996157e-01
-5.06254911e-01 -5.53897023e-01 2.20375165e-01 1.09932590e+00
-2.02230975e-01 -3.31260785e-02 1.01358843e+00 2.46355385e-01
-8.21059108e-01 5.58279335e-01 5.13866544e-02 4.17462081e-01
1.34347975e+00 1.97557002e-01 -3.07397008e-01 1.31250530e-01
-1.06896925e+00 6.50713265e-01 3.65837872e-01 4.55983192e-01
5.18586636e-01 -1.72079575e+00 -8.98521721e-01 4.50968504e-01
7.49481022e-01 -3.27661663e-01 6.70958042e-01 9.42898214e-01
2.28784442e-01 5.96836582e-02 -2.46365324e-01 -4.85991180e-01
-1.04319119e+00 5.87032855e-01 1.74537346e-01 -4.80885923e-01
-6.05565906e-01 9.19145644e-01 2.41860032e-01 -6.52378917e-01
1.55259863e-01 5.52135706e-01 -9.54667851e-02 2.81354278e-01
5.39275229e-01 7.02436805e-01 -5.20629734e-02 -3.31232458e-01
-4.59853053e-01 2.32108533e-01 -3.49954903e-01 -3.06701530e-02
1.72331560e+00 -1.01942979e-01 2.10020736e-01 7.56493330e-01
1.06934547e+00 -2.76070684e-01 -1.37893307e+00 -1.03091121e+00
3.05179536e-01 -3.80243272e-01 1.27504170e-01 -7.65345097e-01
-8.13139319e-01 6.76599205e-01 5.03956020e-01 -3.47076625e-01
1.11228955e+00 -5.03008142e-02 6.49737835e-01 3.51285607e-01
3.40558946e-01 -1.39707172e+00 1.11320920e-01 7.70403624e-01
6.20988250e-01 -1.38980103e+00 -1.31900772e-01 -3.83135796e-01
-9.98564601e-01 7.79706061e-01 7.99333036e-01 -3.77172679e-01
8.75306666e-01 2.37411082e-01 -2.64111072e-01 -1.67785764e-01
-1.34375250e+00 -3.29614043e-01 4.73007321e-01 5.10890782e-01
2.12424293e-01 -8.02419260e-02 -1.03287436e-01 1.06119716e+00
6.25809014e-01 2.86148310e-01 1.15973214e-02 9.33103681e-01
-5.66533267e-01 -1.10458219e+00 -1.07022345e-01 6.93259716e-01
5.40290326e-02 -1.99253038e-01 -2.97314435e-01 3.29018801e-01
-7.11154342e-02 8.07901740e-01 -1.03629269e-01 -5.26695848e-01
3.87589425e-01 6.87609792e-01 3.72507751e-01 -8.98663104e-01
-3.94500196e-01 3.38185698e-01 1.00417450e-01 -5.98499119e-01
-5.09079635e-01 -6.32975280e-01 -7.51709223e-01 -1.89693999e-02
-3.16232413e-01 2.12147281e-01 1.18725553e-01 1.11716902e+00
7.36354411e-01 4.65170503e-01 1.00802851e+00 -4.98276860e-01
-1.09135425e+00 -1.19339347e+00 -1.06358743e+00 4.72269714e-01
2.42045209e-01 -1.01387846e+00 -5.24686396e-01 9.99550447e-02] | [10.014922142028809, 3.03242826461792] |
cc98dc40-1320-4db5-9053-db4b220114da | one-shot-video-inpainting | 2302.14362 | null | https://arxiv.org/abs/2302.14362v1 | https://arxiv.org/pdf/2302.14362v1.pdf | One-Shot Video Inpainting | Recently, removing objects from videos and filling in the erased regions using deep video inpainting (VI) algorithms has attracted considerable attention. Usually, a video sequence and object segmentation masks for all frames are required as the input for this task. However, in real-world applications, providing segmentation masks for all frames is quite difficult and inefficient. Therefore, we deal with VI in a one-shot manner, which only takes the initial frame's object mask as its input. Although we can achieve that using naive combinations of video object segmentation (VOS) and VI methods, they are sub-optimal and generally cause critical errors. To address that, we propose a unified pipeline for one-shot video inpainting (OSVI). By jointly learning mask prediction and video completion in an end-to-end manner, the results can be optimal for the entire task instead of each separate module. Additionally, unlike the two stage methods that use the predicted masks as ground truth cues, our method is more reliable because the predicted masks can be used as the network's internal guidance. On the synthesized datasets for OSVI, our proposed method outperforms all others both quantitatively and qualitatively. | ['Sangyoun Lee', 'Suhwan Cho', 'Sangjin Lee'] | 2023-02-28 | null | null | null | null | ['video-inpainting', 'video-object-segmentation', 'video-semantic-segmentation'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 3.53685975e-01 8.82351473e-02 -2.04235032e-01 -2.66661972e-01
-5.92520237e-01 -4.76940453e-01 2.18456343e-01 -2.88737267e-01
-3.70723009e-01 5.74092507e-01 -7.67990872e-02 -5.42516671e-02
4.22646224e-01 -5.79498827e-01 -9.84223187e-01 -4.85957295e-01
4.20864165e-01 2.06135616e-01 6.94231510e-01 -2.95912251e-02
1.33071840e-01 2.87968159e-01 -1.33935905e+00 3.32726002e-01
1.01741958e+00 1.02176321e+00 5.91280282e-01 5.02666891e-01
-3.28853577e-01 9.85759914e-01 -5.02897978e-01 -3.85849625e-01
5.33271968e-01 -5.74445009e-01 -7.93936312e-01 5.52076340e-01
5.54584205e-01 -7.27989435e-01 -5.45006514e-01 1.09093952e+00
2.10603978e-02 1.92709953e-01 3.43195438e-01 -1.20900655e+00
-2.29484245e-01 4.89423007e-01 -8.98615539e-01 2.62230355e-02
3.12877774e-01 4.27499682e-01 7.80128241e-01 -9.62560833e-01
6.73125207e-01 1.12686551e+00 3.96526158e-01 5.80837131e-01
-1.08446503e+00 -5.75276732e-01 5.17984211e-01 1.49079099e-01
-1.29940641e+00 -6.33430064e-01 9.20624018e-01 -4.93514389e-01
4.82710958e-01 1.77711174e-01 7.86686838e-01 6.56744838e-01
1.27289817e-02 1.31925011e+00 5.08180022e-01 -2.07718194e-01
1.20730594e-01 -3.56349871e-02 -2.23735541e-01 6.92094982e-01
7.12357685e-02 -3.54078263e-02 -4.01767939e-01 2.98321694e-01
1.03873968e+00 3.72304112e-01 -4.91494119e-01 -2.84393579e-01
-1.25043499e+00 4.55595016e-01 3.14587206e-01 5.43803722e-02
-5.98591805e-01 2.99406618e-01 1.92243308e-01 1.03191070e-01
5.63527524e-01 1.54664040e-01 -4.24525499e-01 2.84554530e-02
-1.60061693e+00 3.71034503e-01 5.27779222e-01 1.06672978e+00
9.70258176e-01 1.87947676e-01 -4.19603348e-01 6.97398365e-01
1.92936197e-01 1.36281475e-02 1.19774295e-02 -1.23758793e+00
5.18254995e-01 4.79528517e-01 5.00751615e-01 -1.08523476e+00
5.59548810e-02 -2.65403301e-01 -7.00063407e-01 1.28334954e-01
4.92960930e-01 -1.96389496e-01 -1.25564861e+00 1.53383493e+00
4.96215492e-01 5.75985074e-01 -2.00433359e-01 1.24247396e+00
8.60834420e-01 8.86467576e-01 -5.99845964e-03 -5.02949297e-01
1.11144555e+00 -1.53396666e+00 -9.51477706e-01 -6.43813968e-01
3.06993216e-01 -7.71637201e-01 8.58802080e-01 4.05693591e-01
-1.34071672e+00 -6.16698146e-01 -9.56396401e-01 -1.78590044e-01
2.51015216e-01 1.00353867e-01 3.88500512e-01 1.37833506e-01
-1.05300295e+00 6.05631709e-01 -9.69572008e-01 -1.44031700e-02
7.20300257e-01 2.19552904e-01 -2.66471863e-01 -2.60733247e-01
-9.00911808e-01 5.12116015e-01 3.39313537e-01 3.36175561e-01
-1.12320685e+00 -4.98635173e-01 -9.01293397e-01 2.17615291e-01
7.89007723e-01 -5.08745193e-01 1.26656353e+00 -1.24230909e+00
-1.34202015e+00 4.79423225e-01 -3.97849053e-01 -3.02130848e-01
8.22489321e-01 -2.88273245e-01 1.33849293e-01 4.39982206e-01
1.21129550e-01 1.21033466e+00 1.07511783e+00 -1.43778527e+00
-6.90980852e-01 6.09290078e-02 1.59704685e-01 1.65243238e-01
-1.97593629e-01 8.18281919e-02 -1.25541425e+00 -9.08296525e-01
2.36677036e-01 -6.38982952e-01 -4.22266930e-01 3.92465055e-01
-3.48788917e-01 -1.80308297e-02 1.03337777e+00 -1.16465914e+00
1.24172711e+00 -2.20361543e+00 3.14633876e-01 -2.55875677e-01
3.40269923e-01 4.00980681e-01 -1.48046523e-01 1.13827892e-01
1.09029286e-01 1.86859109e-02 -5.79808593e-01 -7.60111809e-01
-3.45631629e-01 2.13794038e-01 -2.37434700e-01 3.35960180e-01
4.87530500e-01 9.86189961e-01 -9.17442560e-01 -7.58242667e-01
4.50847238e-01 2.67491549e-01 -8.13870549e-01 4.10712749e-01
-5.13265431e-01 6.42605066e-01 -1.96372181e-01 8.04866493e-01
7.70607233e-01 -1.90565884e-01 2.77194697e-02 -2.78084338e-01
7.85831653e-04 -1.75848573e-01 -1.29522455e+00 1.80174351e+00
-1.80822462e-01 7.34447420e-01 4.00501549e-01 -1.02561617e+00
6.26715660e-01 3.74170035e-01 6.56940103e-01 -5.09841502e-01
1.36633188e-01 6.88941702e-02 -1.02109842e-01 -6.44469917e-01
5.44538736e-01 6.48039579e-02 2.84991860e-01 3.22242320e-01
4.16386090e-02 2.24461332e-02 4.88912851e-01 2.03048840e-01
9.41966414e-01 5.24291635e-01 3.97268794e-02 2.09559128e-01
4.12077516e-01 9.16815251e-02 1.13181770e+00 5.44281900e-01
-4.30548698e-01 1.20198691e+00 5.35524607e-01 -2.74020791e-01
-9.92744744e-01 -6.85157657e-01 2.86570936e-01 7.44090855e-01
6.01948142e-01 -3.92699569e-01 -1.08812869e+00 -7.88571656e-01
-3.11696380e-01 4.79402274e-01 -3.33239108e-01 -3.61772749e-04
-7.38766730e-01 -1.56940028e-01 1.66161060e-01 5.41176856e-01
5.50805628e-01 -1.22870636e+00 -5.66855431e-01 4.14947331e-01
-4.34609801e-01 -1.39597058e+00 -8.36643338e-01 -1.43157229e-01
-8.74894202e-01 -1.04841328e+00 -8.86424959e-01 -8.82360697e-01
8.25177848e-01 7.30230749e-01 8.61505032e-01 5.51377833e-01
-1.69028923e-01 -4.82588485e-02 -3.40365291e-01 -5.77633083e-02
-3.44069719e-01 -2.28915513e-01 -2.12843344e-01 2.42124736e-01
-9.31796134e-02 -4.02873874e-01 -8.66149187e-01 3.50416899e-01
-1.21093225e+00 5.47373652e-01 5.52818894e-01 6.50934577e-01
6.90359831e-01 -2.04942692e-02 3.67835701e-01 -8.45169902e-01
5.08456714e-02 -3.70827734e-01 -6.35576904e-01 1.29466012e-01
-2.37865746e-01 -2.80429214e-01 7.72582233e-01 -3.65838408e-01
-1.07868648e+00 3.39336038e-01 -1.56154737e-01 -1.04585361e+00
-1.18010469e-01 2.44441047e-01 -2.71843761e-01 3.26573290e-02
2.10439608e-01 2.92759329e-01 7.20113143e-02 -5.58391392e-01
2.39781857e-01 5.61358213e-01 5.60783744e-01 -2.14831933e-01
7.61479318e-01 4.36245918e-01 -4.45066065e-01 -6.16889179e-01
-9.02965665e-01 -3.66013467e-01 -6.25702202e-01 -4.53004539e-01
1.00441206e+00 -1.05806339e+00 -3.43658745e-01 5.69135547e-01
-1.44065011e+00 -4.20635372e-01 -9.05787945e-02 1.91992864e-01
-4.37975913e-01 6.12745881e-01 -6.91806495e-01 -5.84808648e-01
-1.58766672e-01 -1.61829376e+00 1.16940784e+00 3.77394348e-01
-1.47385858e-02 -6.74017131e-01 -4.70748425e-01 4.22523111e-01
1.43097237e-01 1.41921192e-01 4.70214307e-01 -1.93754122e-01
-1.07989717e+00 -2.81435736e-02 -3.82484287e-01 5.17229915e-01
1.81491718e-01 1.32014006e-01 -8.30623627e-01 -2.42339015e-01
2.06177328e-02 -1.38081983e-01 1.02900183e+00 4.57502663e-01
1.42954242e+00 -2.65004575e-01 -3.17609191e-01 7.48553455e-01
1.41504085e+00 3.32580864e-01 8.21559012e-01 9.67250168e-02
1.01872241e+00 6.55411422e-01 9.58863974e-01 3.06915373e-01
2.56256253e-01 6.61795497e-01 5.32548428e-01 -3.59903604e-01
-2.73073375e-01 -3.10280442e-01 5.00158966e-01 6.01597726e-01
1.31474733e-01 -4.18870926e-01 -6.51150644e-01 6.12407088e-01
-2.02173138e+00 -1.00677979e+00 -8.62851143e-02 2.10569739e+00
8.76310289e-01 4.38164212e-02 3.05098109e-02 -1.95454545e-02
9.04661179e-01 4.12551492e-01 -6.98020279e-01 8.05462971e-02
1.49150357e-01 -8.63821581e-02 2.53193021e-01 5.85170209e-01
-1.01999986e+00 1.32545626e+00 6.05518389e+00 7.63752162e-01
-1.06013370e+00 1.73024222e-01 9.01003182e-01 -2.00904667e-01
-2.31778249e-01 2.96605766e-01 -5.24524033e-01 7.46082902e-01
3.88391674e-01 3.05307031e-01 5.82744956e-01 6.29626811e-01
5.33685923e-01 -4.50987875e-01 -1.07230091e+00 1.19660294e+00
-3.91740054e-02 -1.57173669e+00 9.09651536e-03 -2.65375406e-01
8.11836302e-01 -2.84898669e-01 -3.22301596e-01 1.70253858e-01
-1.14659071e-01 -7.59861648e-01 1.17751706e+00 5.03149569e-01
7.74424016e-01 -6.86096311e-01 5.22865176e-01 4.42600459e-01
-1.18801892e+00 -2.53115091e-02 -3.91644686e-01 -6.30601868e-02
5.73062658e-01 6.25079155e-01 -3.96302581e-01 4.97513086e-01
5.72061300e-01 7.92127132e-01 -4.36747909e-01 1.21995461e+00
-2.93596447e-01 6.35857344e-01 -1.04835019e-01 5.13001263e-01
2.26427495e-01 -4.54588324e-01 4.82729256e-01 9.37687576e-01
1.74019590e-01 2.74822623e-01 5.15084565e-01 1.00522745e+00
-1.22695297e-01 -1.41040519e-01 -3.62923026e-01 -1.46349028e-01
4.04675215e-01 1.29676080e+00 -1.10557294e+00 -5.81681907e-01
-5.80642581e-01 1.26487243e+00 2.72122175e-01 5.79332411e-01
-1.06877136e+00 -3.40116560e-01 7.41855681e-01 1.28630251e-01
5.03768086e-01 -1.73199013e-01 -2.66390204e-01 -1.34186101e+00
2.82102615e-01 -8.42077792e-01 -3.96507941e-02 -8.02423894e-01
-8.12978506e-01 4.55805600e-01 -1.78487003e-01 -1.52854085e+00
-1.48488835e-01 -2.64200449e-01 -6.82086885e-01 6.41026080e-01
-1.45149744e+00 -9.49130177e-01 -4.95024621e-01 3.48043352e-01
1.03643203e+00 1.28486663e-01 1.17862105e-01 5.91284990e-01
-8.66991341e-01 4.82309431e-01 -1.75255775e-01 2.89616108e-01
6.69081390e-01 -7.97289729e-01 4.12986547e-01 1.36632764e+00
3.01115382e-02 2.71084577e-01 7.98569024e-01 -8.57126892e-01
-1.33439350e+00 -1.34648430e+00 3.95333827e-01 1.48326838e-02
1.87953576e-01 -2.85060257e-01 -1.03891551e+00 6.65904224e-01
3.13686877e-01 2.32224941e-01 3.70356813e-02 -6.23503387e-01
2.61307120e-01 -5.03979810e-02 -9.39284086e-01 6.99763775e-01
1.25490963e+00 -9.45056528e-02 -2.66113788e-01 3.95805061e-01
1.06784463e+00 -6.66909933e-01 -4.81041491e-01 2.82428116e-01
2.12146357e-01 -1.06015575e+00 8.39875937e-01 -3.47609103e-01
8.23173642e-01 -8.24712932e-01 6.22693002e-02 -1.06009471e+00
-9.56271216e-02 -7.66879797e-01 -2.76202828e-01 1.28331864e+00
1.77426904e-01 -1.02838665e-01 8.43147933e-01 7.15789616e-01
-3.91700447e-01 -8.49893987e-01 -5.06868184e-01 -4.50583935e-01
-5.37372768e-01 -6.22862339e-01 5.10478139e-01 7.16487229e-01
-5.44960737e-01 5.08320183e-02 -7.76752412e-01 1.40413687e-01
5.64423740e-01 2.82068908e-01 8.89414489e-01 -8.25542867e-01
-3.19991618e-01 -2.82483667e-01 -1.81323320e-01 -1.52984750e+00
1.24402106e-01 -6.81775868e-01 4.57439095e-01 -1.84971654e+00
1.91060886e-01 -3.41397434e-01 -2.75385957e-02 5.78008652e-01
-5.35388410e-01 4.71229136e-01 4.52181816e-01 3.51767749e-01
-6.28998697e-01 6.34670496e-01 1.51066637e+00 -1.48019060e-01
-3.54944259e-01 -9.38253328e-02 -6.51741803e-01 7.48683929e-01
5.06288767e-01 -4.98815507e-01 -4.36353803e-01 -6.02355838e-01
-1.68209776e-01 4.79745716e-01 4.58131194e-01 -8.73402596e-01
2.52721608e-01 -4.98105675e-01 4.91934925e-01 -7.69517958e-01
4.20947641e-01 -7.55294621e-01 2.23246321e-01 3.14791650e-01
3.82677652e-02 -1.69021159e-01 1.11249676e-02 6.70602798e-01
-4.78837341e-01 -3.14014405e-01 8.47614646e-01 -2.11075410e-01
-9.35822666e-01 7.95704961e-01 -2.34637573e-01 -4.26247194e-02
1.12116218e+00 -5.52780986e-01 -2.65694060e-03 -3.81782323e-01
-5.97408414e-01 5.33539057e-01 8.07587981e-01 4.01011199e-01
8.38387311e-01 -1.06728065e+00 -5.67257822e-01 2.12717339e-01
-2.28396207e-01 6.57210886e-01 4.19125885e-01 1.06575501e+00
-8.08202684e-01 1.16880040e-03 -6.28411397e-02 -7.74427295e-01
-9.56111491e-01 6.34036422e-01 2.22755477e-01 6.53843209e-02
-6.93803489e-01 8.78786564e-01 6.82034254e-01 4.99556921e-02
3.66847396e-01 -2.41506949e-01 -1.49075627e-01 2.91571226e-02
6.84774876e-01 4.16925512e-02 -1.93737864e-01 -5.29965878e-01
-9.74122137e-02 3.30199808e-01 -2.81221986e-01 -8.06963965e-02
1.40229595e+00 -3.44981432e-01 -1.62270576e-01 2.14087263e-01
9.91209030e-01 -9.84461606e-02 -2.04728460e+00 -1.81839973e-01
-2.67217308e-01 -9.02287900e-01 3.83709930e-02 -3.32105309e-01
-1.67611051e+00 8.61097515e-01 1.91408321e-01 -1.04323290e-01
1.21261585e+00 -1.32281348e-01 1.23207736e+00 -1.09196715e-01
2.81253546e-01 -1.06693852e+00 1.05560742e-01 2.32390970e-01
7.45356500e-01 -1.20702565e+00 -5.46710864e-02 -7.38749206e-01
-7.54933953e-01 1.01860785e+00 9.58821654e-01 -1.95714682e-01
2.83296019e-01 6.81710169e-02 8.77966508e-02 1.58113137e-01
-7.00460494e-01 -1.17666349e-01 2.10110143e-01 3.36099774e-01
2.84493059e-01 -3.32134604e-01 -3.75859857e-01 4.80200261e-01
2.88988054e-01 5.86272366e-02 5.84163785e-01 9.35031593e-01
-4.59090918e-01 -1.07299304e+00 -3.94787937e-01 5.18990755e-01
-3.85035098e-01 -6.18279120e-03 -1.81148767e-01 5.52618384e-01
5.91551736e-02 9.29227591e-01 -1.94860324e-02 -3.69091064e-01
1.28768384e-01 -1.17587820e-01 3.63208354e-01 -7.40592301e-01
-2.81174242e-01 3.27491164e-01 -2.43343621e-01 -7.58205414e-01
-3.95674229e-01 -5.65849543e-01 -1.42551029e+00 -1.38647363e-01
-3.47469240e-01 -2.02006906e-01 2.98508644e-01 1.20378494e+00
3.48866642e-01 6.54027700e-01 6.09770834e-01 -1.26971197e+00
-1.57260463e-01 -7.96193063e-01 -2.77246445e-01 4.58605617e-01
3.07312131e-01 -7.36882329e-01 -1.53304279e-01 3.78596663e-01] | [9.369644165039062, -0.23502781987190247] |
52957f4e-a624-4387-ae53-bb13655805c0 | constructing-high-frequency-economic | 2303.01863 | null | https://arxiv.org/abs/2303.01863v1 | https://arxiv.org/pdf/2303.01863v1.pdf | Constructing High Frequency Economic Indicators by Imputation | Monthly and weekly economic indicators are often taken to be the largest common factor estimated from high and low frequency data, either separately or jointly. To incorporate mixed frequency information without directly modeling them, we target a low frequency diffusion index that is already available, and treat high frequency values as missing. We impute these values using multiple factors estimated from the high frequency data. In the empirical examples considered, static matrix completion that does not account for serial correlation in the idiosyncratic errors yields imprecise estimates of the missing values irrespective of how the factors are estimated. Single equation and systems-based dynamic procedures yield imputed values that are closer to the observed ones. This is the case in the counterfactual exercise that imputes the monthly values of consumer sentiment series before 1978 when the data was released only on a quarterly basis. This is also the case for a weekly version of the CFNAI index of economic activity that is imputed using seasonally unadjusted data. The imputed series reveals episodes of increased variability of weekly economic information that are masked by the monthly data, notably around the 2014-15 collapse in oil prices. | ['Susannah Scanlan', 'Serena Ng'] | 2023-03-03 | null | null | null | null | ['matrix-completion'] | ['methodology'] | [-5.58183268e-02 -6.39097989e-02 -4.90455776e-01 -4.90591079e-02
-7.06597269e-01 -7.62265205e-01 8.95372629e-01 9.18892026e-02
-6.46306276e-01 1.04013801e+00 9.71484900e-01 -5.95843315e-01
-3.82613838e-01 -9.05955434e-01 -5.25530040e-01 -6.61181569e-01
4.85095754e-02 2.18180954e-01 -6.62715733e-01 -4.27372232e-02
2.50129431e-01 -7.63976499e-02 -1.00292146e+00 -2.84597315e-02
6.35595500e-01 4.69650030e-01 1.16284825e-01 5.16856313e-01
1.01644350e-02 1.00871468e+00 -5.28881907e-01 -6.51892781e-01
4.10490751e-01 -2.07489785e-02 -2.46542722e-01 1.60119995e-01
-4.47755277e-01 -3.92839938e-01 -1.32272229e-01 1.00472832e+00
-2.93717477e-02 -1.47349000e-01 8.91681314e-01 -1.00185013e+00
-8.18835139e-01 1.03558147e+00 -9.78725016e-01 4.06624198e-01
6.13951921e-01 9.56721976e-02 8.62746656e-01 -8.21137905e-01
5.67654073e-01 1.03387940e+00 8.94694865e-01 -5.42083323e-01
-2.01610661e+00 -5.58537245e-01 -6.41010478e-02 -3.57254118e-01
-1.12961543e+00 -4.09453630e-01 5.06930590e-01 -9.17414069e-01
9.70031142e-01 1.67662114e-01 4.61075574e-01 1.33760667e+00
6.15085781e-01 1.08961586e-03 1.23186684e+00 -4.54324067e-01
-3.77607346e-02 1.00274414e-01 1.50132880e-01 -3.03296894e-01
8.91549766e-01 4.97072339e-01 -3.08927238e-01 -7.95085788e-01
7.60199308e-01 3.58685821e-01 -1.10201538e-01 2.87343830e-01
-1.92645979e+00 9.61590052e-01 -2.42692515e-01 4.11811769e-01
-1.18465173e+00 -2.10341439e-01 4.19908911e-01 6.74685478e-01
7.59039879e-01 2.21195117e-01 -9.78088021e-01 -3.68976593e-01
-9.28715885e-01 3.09475243e-01 1.24509978e+00 3.36698920e-01
6.42122030e-01 3.66298944e-01 -3.04679312e-02 6.65052652e-01
6.13924153e-02 1.04430199e+00 6.26943529e-01 -1.00900805e+00
4.70014274e-01 9.50525329e-02 8.09947133e-01 -1.28873324e+00
-3.54201734e-01 -4.60246235e-01 -9.43649650e-01 -2.74111684e-02
9.70240295e-01 -7.47086167e-01 -6.04400694e-01 1.93603754e+00
7.12373406e-02 1.00703754e-01 2.13334784e-01 4.78342414e-01
-4.80673127e-02 7.39506602e-01 5.97915985e-02 -8.43808413e-01
1.31026077e+00 -3.82304996e-01 -1.04089856e+00 -3.80317159e-02
6.23352051e-01 -9.27680612e-01 6.03454888e-01 4.50574160e-01
-6.75645292e-01 -3.39401096e-01 -5.18997610e-01 6.45347536e-01
-2.54183322e-01 -1.09833673e-01 7.74702013e-01 5.67106903e-01
-6.31827295e-01 3.79373014e-01 -9.26181138e-01 2.49435797e-01
-4.68925267e-01 1.08342715e-01 -1.65729076e-01 3.54133099e-01
-1.38415885e+00 8.77297521e-01 1.09386690e-01 -1.50346113e-02
-3.42380226e-01 -9.93120372e-01 -7.02385724e-01 1.61221549e-01
1.19120382e-01 -4.76182610e-01 1.09384584e+00 -1.07450843e+00
-9.88457561e-01 2.05125615e-01 -1.00522101e-01 -5.25806963e-01
4.42402363e-01 3.39975566e-01 -8.21071744e-01 -4.44911599e-01
6.86334312e-01 -3.88374746e-01 6.21002793e-01 -7.38083184e-01
-5.17153680e-01 -4.67133254e-01 -2.80250967e-01 -1.78888857e-01
7.28209466e-02 6.61753565e-02 3.43813956e-01 -1.35769868e+00
-3.66146527e-02 -8.30665708e-01 -3.41438115e-01 -1.04961360e+00
-3.24591488e-01 2.65179843e-01 -9.11469609e-02 -1.17890859e+00
1.56854677e+00 -1.93247533e+00 -7.41949305e-02 4.25899804e-01
-4.15413789e-02 -8.16338480e-01 3.41531426e-01 6.74900591e-01
-6.82727039e-01 1.11808397e-01 -3.45738530e-01 -2.75082678e-01
1.96920574e-01 2.70834595e-01 -5.17423391e-01 7.88164556e-01
-7.42061213e-02 5.27243674e-01 -4.75324929e-01 4.55329865e-01
1.08256020e-01 7.41312280e-02 -4.27199870e-01 -3.25120509e-01
3.93795043e-01 2.81707048e-01 8.40665773e-02 3.08530509e-01
7.91868806e-01 -2.22854584e-01 2.28345782e-01 8.27496350e-02
-5.67155361e-01 3.89860421e-01 -1.44116068e+00 1.38755739e+00
-4.79805678e-01 4.69942302e-01 -1.55647583e-02 -8.44623029e-01
6.05663598e-01 7.21857429e-01 7.43841410e-01 -5.83495677e-01
-2.41227090e-01 2.44574293e-01 2.45814130e-01 -2.80151963e-01
7.07277894e-01 -4.12117481e-01 -3.62921834e-01 7.96983063e-01
-1.55447245e-01 2.12781161e-01 3.68782938e-01 9.18338373e-02
1.18406308e+00 -1.14709593e-01 4.36006904e-01 -5.19254506e-01
1.19554743e-01 1.17897294e-01 7.79108763e-01 7.32816279e-01
5.76196969e-01 4.13130879e-01 7.77538717e-01 -2.22043142e-01
-1.38235581e+00 -9.68729675e-01 -6.56303763e-01 6.69855952e-01
-7.87484288e-01 -2.19420522e-01 -2.65269637e-01 -3.43782380e-02
4.94314849e-01 8.87076557e-01 -9.51632798e-01 1.25374839e-01
1.75351948e-01 -1.57354498e+00 1.35139868e-01 8.06737691e-02
-4.19093035e-02 -5.49326122e-01 -4.94825512e-01 6.32817328e-01
-2.72949517e-01 -6.29678309e-01 -2.31475383e-01 2.22195134e-01
-5.43645859e-01 -7.67965019e-01 -7.82530665e-01 1.09221697e-01
4.66525823e-01 -1.22599438e-01 1.14720809e+00 -4.04197484e-01
6.04113877e-01 1.20567873e-01 -1.24172613e-01 -6.45666122e-01
-4.21603948e-01 4.33299355e-02 2.88543910e-01 4.07739103e-01
6.16070867e-01 -6.70710027e-01 -3.53901297e-01 -1.18905060e-01
-9.77751851e-01 -1.70874491e-01 2.11751461e-01 9.93227422e-01
2.92436071e-02 1.80210739e-01 1.00980330e+00 -9.20268714e-01
7.22414970e-01 -1.17626786e+00 -6.91239715e-01 -8.05641264e-02
-7.81348228e-01 3.25280875e-02 2.21814230e-01 -8.56419027e-01
-1.29913664e+00 -2.67479569e-01 2.40914091e-01 -7.34049156e-02
-8.57478157e-02 1.27467322e+00 4.57824856e-01 5.33888102e-01
5.61686277e-01 -1.80812865e-01 2.24909820e-02 -6.72291100e-01
-3.38669777e-01 6.00364327e-01 5.56377113e-01 -4.12146270e-01
6.43974781e-01 5.94244659e-01 -3.33858490e-01 -3.29097390e-01
-5.95494270e-01 -4.26124662e-01 -3.63546848e-01 3.09808761e-01
6.57823026e-01 -1.57140970e+00 -6.11364365e-01 4.55236793e-01
-9.39143181e-01 -1.16567627e-01 -5.52761614e-01 1.18687642e+00
-3.91164273e-01 -1.42470598e-01 -6.93454087e-01 -1.20260429e+00
1.23692669e-01 -1.12861896e+00 6.11231506e-01 -2.32839391e-01
-6.93248689e-01 -1.34000957e+00 5.63488722e-01 -1.99971139e-01
5.44803619e-01 8.17873597e-01 7.14591146e-01 -6.22718394e-01
1.39889345e-01 -4.60029066e-01 8.81938413e-02 -7.28438422e-02
5.64528286e-01 2.32403845e-01 -7.08082378e-01 -2.81739861e-01
4.99537349e-01 3.64587396e-01 6.34202063e-01 7.22529233e-01
1.90618187e-01 -8.56671095e-01 1.74580924e-02 2.91674584e-01
1.63941360e+00 3.55426639e-01 3.32171232e-01 5.86112082e-01
3.16243052e-01 4.98480827e-01 8.10193270e-02 1.09872186e+00
6.59723759e-01 4.63388473e-01 -9.84480008e-02 -1.50465593e-01
6.83198094e-01 -1.40307322e-01 5.15258789e-01 8.80752265e-01
-1.37615547e-01 1.95126846e-01 -9.80435073e-01 9.83539879e-01
-1.73576236e+00 -1.36733079e+00 -5.83806038e-01 2.40744758e+00
9.08280790e-01 2.55369246e-01 4.88251090e-01 1.87352121e-01
5.32125235e-01 1.15294330e-01 -1.77956939e-01 -4.54597831e-01
-3.31193566e-01 -4.19098139e-02 1.00572312e+00 7.78233767e-01
-6.63781941e-01 -3.80220339e-02 7.08635092e+00 3.45136344e-01
-6.09701872e-01 1.26342073e-01 9.58102345e-01 -4.40731347e-02
-7.88138926e-01 3.93511593e-01 -4.93893385e-01 9.07199919e-01
1.65379763e+00 -8.53166401e-01 3.95740449e-01 3.48937690e-01
5.90772629e-01 -6.52176321e-01 -6.80571139e-01 6.67348802e-01
-5.46176970e-01 -1.08553576e+00 -4.59320784e-01 8.90283823e-01
1.09787905e+00 1.67764332e-02 2.52483070e-01 2.16553524e-01
9.20256853e-01 -7.51421392e-01 1.00043631e+00 9.28916752e-01
6.25599682e-01 -1.18580484e+00 8.28442574e-01 5.51814079e-01
-6.83741808e-01 -3.30332756e-01 -4.51032072e-01 -7.91229844e-01
3.10240179e-01 1.31175137e+00 -3.22329760e-01 4.64140534e-01
5.29501021e-01 7.84287989e-01 -1.72785968e-01 5.11561394e-01
2.78827310e-01 9.23944831e-01 -5.78911662e-01 7.70603359e-01
9.58846435e-02 -9.08921659e-01 5.56225479e-01 9.31613207e-01
6.25541866e-01 -1.63400322e-01 -2.96758264e-01 7.50146925e-01
2.31745765e-01 -2.22504288e-01 -1.05906081e+00 1.18681245e-01
2.96841741e-01 9.58934486e-01 -2.96428144e-01 -3.99714381e-01
-1.30691147e+00 3.61098558e-01 -8.56806785e-02 7.24084616e-01
-5.69124937e-01 -5.35016693e-02 6.26424551e-01 1.32629827e-01
1.10474609e-01 -5.10217905e-01 -4.18181926e-01 -1.69354689e+00
-1.19661435e-01 -9.87532496e-01 3.72760326e-01 -5.70499480e-01
-1.43789375e+00 -1.10691950e-01 2.10335299e-01 -9.13221598e-01
-1.13149238e+00 -9.03873425e-03 -5.11413395e-01 1.50086558e+00
-1.08179963e+00 -2.64363408e-01 6.74283922e-01 4.46335435e-01
2.41092160e-01 2.26089768e-02 9.15743232e-01 7.18889683e-02
-5.29712975e-01 8.36010464e-03 8.38877082e-01 1.29188523e-01
7.38862693e-01 -1.31879568e+00 6.42269790e-01 1.01931381e+00
2.69314617e-01 7.99940050e-01 9.90450859e-01 -1.08532512e+00
-1.09103012e+00 -5.68318605e-01 1.17343879e+00 -6.90860689e-01
1.32921624e+00 -1.01367936e-01 -8.63562226e-01 9.64897215e-01
3.03334832e-01 -5.96859336e-01 7.33269989e-01 1.06894888e-01
-4.82591353e-02 5.93242757e-02 -1.05638814e+00 3.13929528e-01
1.74093246e-01 -5.98243535e-01 -7.30090022e-01 2.31213078e-01
7.45304406e-01 -5.37655205e-02 -1.64681506e+00 2.35220999e-01
7.62784243e-01 -8.28149796e-01 8.49131465e-01 -4.33797747e-01
2.18786612e-01 -1.86594546e-01 -4.45462823e-01 -1.63016117e+00
-8.06038082e-01 -6.27231300e-01 1.19314440e-01 1.54456401e+00
8.47724438e-01 -1.01951349e+00 3.25111181e-01 9.40863371e-01
5.95657527e-01 4.08731326e-02 -1.01068926e+00 -5.92720509e-01
1.30669206e-01 -4.11829203e-01 9.33594465e-01 1.61123550e+00
5.81952371e-02 -2.43703313e-02 -5.16211390e-01 2.40471624e-02
7.70976663e-01 1.45899802e-01 5.99807203e-01 -1.30081809e+00
-6.13907695e-01 3.47321965e-02 1.42775446e-01 -3.40125889e-01
1.33269668e-01 -4.11131531e-01 -7.46638000e-01 -8.76855671e-01
2.64600784e-01 3.39112580e-02 2.29958650e-02 9.17824656e-02
-1.43959047e-02 1.37817599e-02 7.14526325e-02 3.01543713e-01
5.79921007e-01 2.73355782e-01 7.45406032e-01 4.97531518e-02
-2.34730184e-01 8.85033235e-02 -9.45031643e-01 7.51933992e-01
7.23551512e-01 -6.72739267e-01 -7.67332017e-02 -3.32911640e-01
6.66093230e-01 8.40245903e-01 3.70205134e-01 -5.49413383e-01
1.37100145e-01 -6.05524719e-01 6.78265810e-01 -4.67909187e-01
-7.10530020e-03 -1.08783150e+00 1.24707913e+00 3.25096220e-01
-2.11034730e-01 9.17582035e-01 2.69665390e-01 6.53127611e-01
-2.96489060e-01 -5.17016239e-02 1.49545908e-01 -2.32506543e-01
1.12699576e-01 -1.79545134e-02 -1.04262197e+00 -2.16553420e-01
5.98547161e-01 -4.96867672e-02 -1.58931062e-01 -6.73317432e-01
-1.10502434e+00 -8.67459774e-02 6.38907552e-01 1.67512462e-01
-4.55313548e-02 -1.46706462e+00 -1.11018014e+00 2.62873113e-01
-2.99132109e-01 -5.70121646e-01 9.93742868e-02 1.28945780e+00
9.28066596e-02 4.70843464e-01 1.06915571e-01 -1.32222980e-01
-3.95155370e-01 7.18482077e-01 1.05286352e-01 -5.81921577e-01
-3.67533445e-01 -2.21160442e-01 2.46431872e-01 -1.49261892e-01
-3.99902344e-01 -4.56536442e-01 -1.02547519e-01 7.03677118e-01
6.17980957e-01 4.85098362e-01 -2.73891781e-02 -8.34859908e-01
7.10058538e-03 2.58814812e-01 1.35099992e-01 -6.51547432e-01
1.65410686e+00 -8.12242210e-01 -2.88941979e-01 1.14522672e+00
9.37193573e-01 2.28884384e-01 -1.27048469e+00 -1.91703171e-01
2.54597574e-01 -3.84674817e-01 -3.13422047e-02 -5.44910491e-01
-9.35164154e-01 1.05584718e-01 1.11930013e-01 7.25762606e-01
9.05949593e-01 -5.37405252e-01 1.53421775e-01 -1.76829040e-01
3.45123291e-01 -7.92568982e-01 -8.89569402e-01 2.42261037e-01
8.14565957e-01 -8.64583910e-01 1.38443992e-01 2.90226191e-01
-4.38312888e-01 6.30796373e-01 -2.49648586e-01 -3.15688103e-01
1.03594553e+00 5.30133903e-01 -2.75584698e-01 1.30103096e-01
-1.03569031e+00 1.54020235e-01 1.16166146e-02 4.32833225e-01
4.07014877e-01 2.77548462e-01 -7.81370282e-01 7.70366907e-01
-4.98447806e-01 -3.00454676e-01 1.20067716e+00 6.43433332e-01
2.41311923e-01 -7.25731909e-01 -9.57953930e-01 9.66294289e-01
-1.31463623e+00 -3.90505642e-01 1.64536759e-01 8.01488221e-01
-4.23727125e-01 1.04643297e+00 5.12361109e-01 -1.71651430e-02
2.28833884e-01 2.73063809e-01 -3.51294637e-01 -2.05445752e-01
-5.11111796e-01 5.23124993e-01 2.54441410e-01 -1.49038419e-01
-6.50416970e-01 -1.17894137e+00 -6.71511948e-01 -8.02355230e-01
-2.74484575e-01 2.60958225e-01 4.23750311e-01 1.00055170e+00
7.92814717e-02 3.24539244e-01 1.07186389e+00 -1.05654800e+00
-7.48777330e-01 -1.26721072e+00 -9.80372131e-01 4.34161186e-01
7.39888549e-01 -5.34443915e-01 -1.05909848e+00 2.06184834e-01] | [6.029023170471191, 4.037415981292725] |
eaee78d4-f0e1-4e75-b165-9eb3f6b17d98 | fakeswarm-improving-fake-news-detection-with | 2305.19194 | null | https://arxiv.org/abs/2305.19194v1 | https://arxiv.org/pdf/2305.19194v1.pdf | FakeSwarm: Improving Fake News Detection with Swarming Characteristics | The proliferation of fake news poses a serious threat to society, as it can misinform and manipulate the public, erode trust in institutions, and undermine democratic processes. To address this issue, we present FakeSwarm, a fake news identification system that leverages the swarming characteristics of fake news. To extract the swarm behavior, we propose a novel concept of fake news swarming characteristics and design three types of swarm features, including principal component analysis, metric representation, and position encoding. We evaluate our system on a public dataset and demonstrate the effectiveness of incorporating swarm features in fake news identification, achieving an f1-score and accuracy of over 97% by combining all three types of swarm features. Furthermore, we design an online learning pipeline based on the hypothesis of the temporal distribution pattern of fake news emergence, validated on a topic with early emerging fake news and a shortage of text samples, showing that swarm features can significantly improve recall rates in such cases. Our work provides a new perspective and approach to fake news detection and highlights the importance of considering swarming characteristics in detecting fake news. | ['Xuesong Ye', 'Jun Wu'] | 2023-05-30 | null | null | null | null | ['fake-news-detection'] | ['natural-language-processing'] | [-2.33961463e-01 -7.62075558e-02 -1.50464147e-01 1.59129947e-01
-1.27202898e-01 -8.81648123e-01 1.26458383e+00 5.43931603e-01
-1.38593405e-01 4.98136044e-01 4.53948855e-01 -3.31904858e-01
-3.97678912e-02 -8.59227717e-01 -6.33898735e-01 -4.51629072e-01
-2.44265184e-01 4.34449881e-01 2.76739150e-01 -7.65332103e-01
8.58959615e-01 3.44726622e-01 -1.28621531e+00 5.61021268e-01
7.28442669e-01 6.92222178e-01 -4.72195357e-01 5.91776490e-01
3.15048277e-01 8.92536521e-01 -1.62376940e+00 -5.65955520e-01
1.37324899e-01 -4.07819569e-01 -5.79649568e-01 -2.18850542e-02
1.17733270e-01 -4.04042214e-01 -7.28383958e-01 9.71656203e-01
2.21899211e-01 -4.68313426e-01 5.28769553e-01 -1.28668320e+00
-9.19885814e-01 7.05195248e-01 -3.34038347e-01 4.91250724e-01
4.23751146e-01 2.07446173e-01 8.71548653e-01 -4.88051236e-01
7.83170998e-01 1.15976155e+00 9.43045557e-01 1.58808436e-02
-6.38965607e-01 -5.88362455e-01 -2.85371065e-01 -6.96538016e-02
-8.64245057e-01 -3.63738477e-01 6.32178962e-01 -5.73728979e-01
6.20839357e-01 4.87335593e-01 1.28155327e+00 1.64588654e+00
5.21587074e-01 9.38573837e-01 1.24949634e+00 -2.98308134e-02
2.40684673e-01 2.43203908e-01 -3.18806130e-03 8.28049481e-01
9.83858228e-01 3.47363263e-01 -8.71945024e-01 -9.12425220e-01
2.32707351e-01 -5.28122522e-02 -2.06746235e-01 3.60884994e-01
-1.72386992e+00 9.50900555e-01 3.58802974e-01 4.32396889e-01
-9.78141353e-02 1.30074695e-01 5.14855325e-01 6.75325513e-01
8.53577435e-01 1.29285681e+00 -3.85043025e-01 -4.51411426e-01
-7.06709027e-01 7.92409852e-02 1.02351725e+00 3.73629838e-01
1.84483618e-01 -1.42034605e-01 -2.26489361e-02 3.66199672e-01
2.60784086e-02 9.54897523e-01 5.87610066e-01 -4.73591596e-01
1.74219191e-01 8.60227644e-01 5.43795764e-01 -1.92658854e+00
-5.07672429e-01 -6.32009029e-01 -3.83023649e-01 -5.16070545e-01
3.11573952e-01 -9.94151831e-02 -2.48436153e-01 1.16573894e+00
4.84737426e-01 4.67803091e-01 4.65867482e-02 9.14091468e-01
7.54831910e-01 6.07906044e-01 -7.95119107e-01 2.28242278e-02
1.36229336e+00 -9.25064683e-01 -7.84258723e-01 -2.41660878e-01
9.83239591e-01 -7.66596079e-01 7.06213295e-01 2.76154369e-01
-3.30502450e-01 2.70554453e-01 -1.12414396e+00 5.20374835e-01
-9.89375889e-01 2.91972935e-01 9.75467920e-01 8.87077153e-01
-6.86414838e-01 6.15877748e-01 -7.79186547e-01 -5.02244294e-01
6.34770751e-01 8.87678415e-02 -2.28702009e-01 1.96664765e-01
-1.22408330e+00 7.50054598e-01 -2.57669892e-02 -1.84919998e-01
-4.76539433e-01 -3.34569097e-01 -3.15080196e-01 -1.16884656e-01
2.08431542e-01 -4.29626137e-01 6.25373662e-01 -6.20891750e-01
-1.39853716e+00 6.37323916e-01 3.93015355e-01 -4.82646137e-01
6.63124800e-01 -2.19113529e-02 -6.66256964e-01 3.77025247e-01
2.72771269e-01 -1.30233079e-01 1.04129279e+00 -1.16615236e+00
-3.41734409e-01 -2.37956092e-01 -2.28925243e-01 -3.34940672e-01
-8.70091319e-01 -5.28506748e-02 3.24614733e-01 -1.03428304e+00
1.58245996e-01 -1.14912009e+00 2.60554224e-01 -3.00762892e-01
-6.03868842e-01 4.19049375e-02 1.25636113e+00 -6.29472733e-01
9.90003049e-01 -2.09162211e+00 -2.48528391e-01 1.24095000e-01
8.08826983e-01 4.11820233e-01 -1.21790186e-01 6.68592811e-01
7.30854392e-01 3.32576632e-01 6.29270896e-02 -1.12682350e-01
-2.36647859e-01 -1.61318585e-01 -6.18266523e-01 1.09680164e+00
1.60791755e-01 8.89592946e-01 -1.52454543e+00 1.54099271e-01
-2.19507828e-01 1.55141994e-01 -5.22731602e-01 -1.82354286e-01
-9.41761285e-02 4.25030798e-01 -6.51078045e-01 8.59719217e-01
5.10960281e-01 -7.51508534e-01 2.43044436e-01 1.18954003e-01
2.40535568e-03 4.23727125e-01 -4.90365654e-01 7.41915822e-01
7.50517175e-02 1.07222319e+00 -1.21676341e-01 -7.17109740e-01
8.76447380e-01 -3.66889499e-02 5.20360827e-01 -3.42854232e-01
6.45365238e-01 4.99084204e-01 -1.30660132e-01 -5.66251516e-01
8.73448014e-01 6.05903864e-01 -4.56127763e-01 1.10577667e+00
-2.07548812e-01 -2.68114001e-01 -2.21622020e-01 5.25930703e-01
1.58249938e+00 -6.60677493e-01 1.86533242e-01 -1.52146384e-01
-4.90510315e-02 4.09054399e-01 1.53478384e-01 1.25516295e+00
-3.77638906e-01 9.81324241e-02 7.45563805e-01 -7.80127645e-01
-7.73427486e-01 -3.95819575e-01 1.00196891e-01 7.31786072e-01
5.97133696e-01 -7.36120880e-01 -3.14667135e-01 -9.77450550e-01
5.73494613e-01 3.00273806e-01 -1.01475906e+00 -6.62345767e-01
-2.80325741e-01 -1.38332200e+00 1.06258988e+00 -3.90080422e-01
5.06945431e-01 -5.69378257e-01 -4.50750619e-01 1.07376330e-01
-3.94213974e-01 -1.28094387e+00 -1.91802129e-01 -4.39429462e-01
-4.55642611e-01 -1.41635513e+00 -3.09242997e-02 -2.82942802e-01
5.71938276e-01 8.82784665e-01 8.00553143e-01 7.22951591e-01
-1.89894602e-01 2.34985754e-01 -7.47475564e-01 -3.48359942e-01
-1.08580315e+00 8.09727460e-02 4.79215145e-01 1.53583273e-01
5.74045740e-02 -1.68602362e-01 -4.72084403e-01 4.58531797e-01
-7.50475168e-01 -2.01430693e-01 4.81338173e-01 9.59164977e-01
-4.89716023e-01 1.16435103e-01 6.82698905e-01 -8.04035008e-01
8.58523965e-01 -8.84592235e-01 -6.85357511e-01 1.42401978e-01
-4.29558158e-01 -3.26560885e-01 7.40335286e-01 -7.51221478e-01
-2.38457888e-01 -5.77010512e-01 6.06295466e-01 5.20629156e-03
4.92085367e-01 4.97517169e-01 6.37635529e-01 -5.61685860e-01
8.23911488e-01 3.77359718e-01 3.09109688e-01 -1.55331969e-01
1.63117260e-01 1.07369566e+00 5.30251376e-02 -1.71336114e-01
9.02778983e-01 1.03159940e+00 -2.54441321e-01 -1.12927616e+00
-8.41054082e-01 -3.25195312e-01 -2.48535145e-02 -1.37853280e-01
1.38461113e-01 -7.88388431e-01 -1.00453568e+00 8.50377738e-01
-1.22157264e+00 2.35097874e-02 8.68153349e-02 5.06259680e-01
3.61315794e-02 6.07266605e-01 -8.99434805e-01 -5.99967599e-01
-9.69988257e-02 -9.33958769e-01 1.16872919e+00 -3.11084986e-01
-8.71778727e-02 -9.10025477e-01 7.18239397e-02 7.24173546e-01
7.12393701e-01 5.03341794e-01 3.71958077e-01 -1.23617542e+00
-5.06606400e-01 -5.57995737e-01 -1.81282982e-01 -9.47842225e-02
2.21283793e-01 1.03168003e-01 -6.63980782e-01 -5.73730707e-01
5.59652932e-02 -1.68310642e-01 8.46803904e-01 -1.64473131e-01
4.67735887e-01 -8.78943741e-01 -6.01812243e-01 3.94667983e-01
6.90593958e-01 -3.63739133e-01 4.40730006e-02 6.03454053e-01
5.75253904e-01 5.10058761e-01 4.97830600e-01 8.84507120e-01
4.41147596e-01 4.19625908e-01 3.87798190e-01 2.19442189e-01
1.19094163e-01 -1.78091675e-01 5.59404671e-01 1.13192213e+00
3.22408020e-01 -5.75368583e-01 -8.60788524e-01 4.13661003e-01
-1.63365293e+00 -1.00056732e+00 -2.84944564e-01 1.82223165e+00
3.29146743e-01 3.40981744e-02 2.15963602e-01 -2.95409076e-02
8.44184995e-01 4.18847769e-01 -3.07566881e-01 -5.49455248e-02
-6.24925613e-01 -5.34911931e-01 8.05783987e-01 3.29294205e-01
-1.26856482e+00 1.25724983e+00 6.78641939e+00 5.01576364e-01
-1.48831320e+00 1.80901110e-01 2.82748967e-01 5.74463606e-02
-2.41830572e-01 -5.39574251e-02 -3.98582965e-01 9.60286379e-01
7.59227693e-01 -2.68069096e-02 5.85153878e-01 4.58295345e-01
3.16979080e-01 -1.74140818e-02 -3.43595684e-01 7.21572816e-01
3.16556513e-01 -1.98595536e+00 -2.11005863e-02 5.33641934e-01
9.35387373e-01 4.08364356e-01 1.73495620e-01 -1.66106634e-02
5.42573392e-01 -6.74874544e-01 7.61475801e-01 1.75975785e-01
3.16166252e-01 -3.94778013e-01 8.64901483e-01 4.33415771e-01
-3.70637983e-01 -3.92094880e-01 -2.42677808e-01 -3.39063227e-01
-1.82199091e-01 8.98240328e-01 -1.39810562e+00 3.08857113e-01
5.11529267e-01 1.18355942e+00 -9.79614377e-01 8.48134041e-01
-2.74405000e-03 8.45682263e-01 -5.95074892e-01 -6.56717896e-01
2.00272441e-01 -3.61537896e-02 1.14083135e+00 9.96760488e-01
4.47820336e-01 -2.03140855e-01 9.08997580e-02 5.68759441e-01
-1.65452376e-01 -1.50874019e-01 -7.73857951e-01 -7.94652939e-01
9.63432908e-01 9.49036419e-01 -1.01000738e+00 -5.00066102e-01
3.37076671e-02 1.01190770e+00 -6.74568722e-03 -8.56823549e-02
-8.46634150e-01 -2.17477188e-01 5.31856716e-01 2.13507004e-02
1.47927135e-01 -2.98984319e-01 -1.65450647e-01 -1.67530441e+00
-2.08825633e-01 -1.23344493e+00 3.12745608e-02 -2.89379865e-01
-1.51245904e+00 4.45751846e-01 -5.36406457e-01 -1.36302471e+00
3.04691643e-01 -3.84567261e-01 -4.52241063e-01 -1.65127233e-01
-1.15437376e+00 -1.27960086e+00 -3.58485788e-01 9.96678248e-02
-8.71770009e-02 -6.31162167e-01 6.18986845e-01 -5.92957027e-02
-4.86390591e-01 3.90569568e-01 7.31715500e-01 2.72900403e-01
8.06186914e-01 -7.82899439e-01 7.58129478e-01 6.15268886e-01
2.22367406e-01 5.61828375e-01 8.53408337e-01 -9.64400530e-01
-1.87547696e+00 -1.02517569e+00 7.42782056e-01 -9.68453884e-01
1.33361769e+00 -7.10054338e-01 -4.79726642e-01 3.95922720e-01
-3.34953666e-01 1.03371114e-01 7.06011534e-01 -1.16121531e-01
-6.40604019e-01 3.06636959e-01 -1.41499031e+00 4.15814489e-01
9.06033456e-01 -5.82932472e-01 -4.66667622e-01 1.03819299e+00
8.09851229e-01 -1.83267236e-01 -7.87419975e-01 1.85064338e-02
6.16540313e-01 -1.11157906e+00 7.97825336e-01 -6.45776451e-01
3.30303609e-01 -2.03412920e-01 1.47439852e-01 -1.48959267e+00
-1.30300701e-01 -1.00820231e+00 -3.59876513e-01 6.73466563e-01
1.83534980e-01 -1.27081776e+00 7.42650568e-01 -5.61481297e-01
2.00269490e-01 -5.47080457e-01 -9.43282545e-01 -8.77890646e-01
-2.82218367e-01 2.12886572e-01 7.53400385e-01 1.71227574e+00
2.39341795e-01 3.36879902e-02 -6.83410525e-01 4.65269059e-01
4.45230305e-01 2.34951928e-01 1.07080579e+00 -1.36301601e+00
-2.54581332e-01 -4.43503499e-01 -6.57992840e-01 -8.45062494e-01
8.17650855e-02 -7.32783973e-01 -3.69383305e-01 -9.58709180e-01
-1.46448493e-01 -5.70617914e-01 2.24423617e-01 1.89189762e-01
3.45045216e-02 5.72542548e-01 -9.65441242e-02 7.87519813e-01
-6.90343857e-01 5.89260340e-01 1.32526779e+00 -3.65510106e-01
-3.06729227e-02 -9.51921716e-02 -6.38340712e-01 5.76275826e-01
7.41940141e-01 -8.39375198e-01 2.32149437e-01 -2.46242985e-01
5.89272201e-01 -1.11674562e-01 5.39381206e-01 -9.04996395e-01
3.04731339e-01 -1.10103033e-01 1.35693967e-01 -2.49750599e-01
4.92902607e-01 -3.65005434e-01 -3.46986741e-01 9.84176993e-01
1.85959786e-02 1.27642989e-01 -1.91822246e-01 1.06331134e+00
5.26599027e-02 3.89247030e-01 5.10953844e-01 8.32016468e-02
2.54915077e-02 -1.41246647e-01 -8.74264717e-01 3.37964781e-02
8.78386140e-01 1.32059737e-03 -1.25045013e+00 -4.27474856e-01
-4.47013043e-02 -1.07140355e-01 8.03076625e-01 7.35551834e-01
5.07840097e-01 -9.10144985e-01 -9.80878592e-01 3.16269040e-01
3.10764372e-01 -9.10516560e-01 -3.25545788e-01 1.09406459e+00
-6.95969284e-01 1.42226979e-01 7.58969858e-02 -4.80052978e-01
-1.01655257e+00 3.71920526e-01 1.48141474e-01 -6.42131865e-02
-5.20806730e-01 6.40999794e-01 -7.18669057e-01 -5.66352546e-01
-4.38939333e-01 -5.92513680e-02 -4.03421782e-02 2.56893814e-01
3.83304060e-01 5.34225643e-01 7.91832283e-02 -9.07405257e-01
-4.94120121e-01 -3.20604183e-02 -2.93945540e-02 2.29014844e-01
1.49785066e+00 2.66663432e-02 -6.42125547e-01 1.70014858e-01
1.14619613e+00 7.82659054e-01 -5.88986635e-01 -7.60753900e-02
2.41127640e-01 -9.57716525e-01 -5.23007996e-02 -9.00526226e-01
-7.24958658e-01 -1.48694608e-02 1.12977438e-01 9.79362369e-01
1.38005689e-01 1.35123983e-01 1.12697756e+00 8.26869011e-01
5.74542403e-01 -9.53252137e-01 6.61273599e-01 7.09424555e-01
5.03175914e-01 -1.48486674e+00 3.18338633e-01 -4.67623860e-01
-2.21698657e-01 1.01091599e+00 1.24066599e-01 -2.97944993e-01
5.67810714e-01 1.38531417e-01 5.32294512e-02 -7.68286049e-01
-5.75824797e-01 4.57095087e-01 -1.90304257e-02 3.05934399e-01
-1.48385286e-01 2.38630489e-01 -3.51627588e-01 3.49284023e-01
-5.64197361e-01 -5.17643273e-01 1.13939500e+00 1.07798207e+00
-5.74808955e-01 -7.47985959e-01 -5.55604815e-01 6.06456757e-01
-4.55503851e-01 -6.19390374e-03 -1.21301949e+00 5.62336862e-01
-1.28129959e-01 1.24376404e+00 -8.57629552e-02 -9.02353525e-01
-2.42565528e-01 -6.97927594e-01 1.11088380e-01 -3.29480112e-01
-9.63280559e-01 -4.54552740e-01 3.47651243e-01 -4.13441181e-01
-2.37167105e-01 -4.15258914e-01 -5.89364946e-01 -8.35468113e-01
-7.86478221e-01 5.32076895e-01 8.40811968e-01 1.01730394e+00
9.93930757e-01 -4.48244624e-02 1.11105669e+00 -6.20634496e-01
-7.49672234e-01 -9.99773502e-01 -4.10681576e-01 4.74876761e-01
7.05032289e-01 -7.34139502e-01 -1.05965686e+00 -5.10626197e-01] | [8.131845474243164, 10.267253875732422] |
0cf6868d-4fa9-4c3e-9035-f2dadb48c4b5 | asl-skeleton3d-and-asl-phono-two-novel | 2201.02065 | null | https://arxiv.org/abs/2201.02065v1 | https://arxiv.org/pdf/2201.02065v1.pdf | ASL-Skeleton3D and ASL-Phono: Two Novel Datasets for the American Sign Language | Sign language is an essential resource enabling access to communication and proper socioemotional development for individuals suffering from disabling hearing loss. As this population is expected to reach 700 million by 2050, the importance of the language becomes even more essential as it plays a critical role to ensure the inclusion of such individuals in society. The Sign Language Recognition field aims to bridge the gap between users and non-users of sign languages. However, the scarcity in quantity and quality of datasets is one of the main challenges limiting the exploration of novel approaches that could lead to significant advancements in this research area. Thus, this paper contributes by introducing two new datasets for the American Sign Language: the first is composed of the three-dimensional representation of the signers and, the second, by an unprecedented linguistics-based representation containing a set of phonological attributes of the signs. | ['Cleber Zanchettin', 'Cleison Correia de Amorim'] | 2022-01-06 | null | null | null | null | ['sign-language-recognition'] | ['computer-vision'] | [ 2.55017728e-02 -4.55104373e-02 -1.72801644e-01 -1.12085775e-01
-5.49782097e-01 -2.02154487e-01 4.59934980e-01 -9.46649164e-02
-7.00192392e-01 6.50289059e-01 7.64222145e-01 -1.10205457e-01
-2.22006887e-01 -5.58714390e-01 7.49990046e-02 -6.57603323e-01
2.19816014e-01 3.10631782e-01 -1.48469564e-02 -4.46844101e-01
3.00408304e-01 7.40635335e-01 -2.16723919e+00 -3.09636798e-02
9.04746950e-01 8.33022475e-01 1.90843701e-01 2.87555784e-01
-3.83141905e-01 3.96300107e-01 -2.85611451e-01 -1.77549660e-01
1.12877809e-01 -5.25945663e-01 -4.03943062e-01 -1.49097517e-01
5.81283629e-01 -5.95801055e-01 -1.04508698e-01 7.89220095e-01
8.53170633e-01 -1.64187662e-02 5.13155222e-01 -6.66409850e-01
-4.76333827e-01 2.78025925e-01 -2.62681037e-01 -2.99818844e-01
5.39910913e-01 9.47307125e-02 1.04476345e+00 -6.60331845e-01
8.18383813e-01 9.62912917e-01 1.94142386e-01 7.95558810e-01
-6.79122984e-01 -4.43436146e-01 1.63100883e-01 3.32562804e-01
-1.11557949e+00 -6.54121816e-01 6.32372200e-01 -7.04838336e-01
5.70196927e-01 1.24103881e-01 1.28318882e+00 7.44830191e-01
-6.08822703e-01 5.72001398e-01 1.27582228e+00 -7.50212550e-01
1.37127042e-02 -8.95995796e-02 1.86869666e-01 5.32953084e-01
4.69085991e-01 1.06062494e-01 -8.53870511e-01 1.76215414e-02
6.23858511e-01 -3.85492057e-01 -4.58058715e-01 -2.60581523e-01
-9.40678298e-01 3.94832939e-01 2.37799034e-01 6.75277233e-01
-3.65018129e-01 -6.81190193e-02 1.68098122e-01 2.24324852e-01
-3.35603282e-02 1.49364829e-01 -1.35146126e-01 -8.03267002e-01
-6.74152493e-01 8.97318274e-02 6.34570122e-01 2.00879619e-01
8.75604749e-02 8.62893909e-02 2.12522835e-01 1.06755412e+00
4.82402742e-01 7.98292220e-01 5.56841969e-01 -5.81961215e-01
1.36503428e-01 9.40401316e-01 -1.26224458e-01 -5.18298745e-01
-3.18492234e-01 -1.74871638e-01 -4.32081699e-01 5.59142768e-01
8.09772313e-01 2.13574678e-01 -1.06744874e+00 1.68234336e+00
1.63709179e-01 -3.88577819e-01 -2.19864264e-01 9.01284099e-01
9.36732948e-01 -1.01129875e-01 1.82121277e-01 1.12295851e-01
1.42139065e+00 -2.56973863e-01 -5.36263108e-01 -1.25523239e-01
3.11219662e-01 -7.09558606e-01 1.25524449e+00 3.61343533e-01
-9.16190147e-01 -1.70072779e-01 -7.94835687e-01 -5.97574562e-02
-4.64117587e-01 2.94772714e-01 7.33665407e-01 8.86580706e-01
-1.02836668e+00 -6.73812181e-02 -5.27058721e-01 -7.47134447e-01
6.16671264e-01 3.25707108e-01 -6.53480589e-01 -2.71359235e-01
-8.36121738e-01 1.03142154e+00 -4.98187542e-02 2.98235834e-01
1.53695986e-01 -5.04312277e-01 -8.13755929e-01 -2.64537215e-01
-2.58208662e-02 -4.31092769e-01 9.00036752e-01 -5.81516266e-01
-1.39850128e+00 1.26113808e+00 -2.74709165e-01 6.06941469e-02
7.38631368e-01 -3.12568098e-01 -4.10165250e-01 1.37285115e-02
-7.30263889e-02 3.82866293e-01 6.26709700e-01 -1.13788819e+00
-8.62465143e-01 -6.62555397e-01 -3.32323253e-01 7.16721043e-02
-4.79342729e-01 5.23540378e-01 -3.71540695e-01 -4.15635765e-01
4.05075341e-01 -6.34358108e-01 2.26035446e-01 5.06717861e-01
-1.83196291e-01 -5.63317463e-02 5.51428616e-01 -1.07821214e+00
1.05264795e+00 -2.13166928e+00 -5.66676958e-03 4.15928662e-01
3.32826942e-01 5.60974360e-01 -1.11006256e-02 4.13321882e-01
2.98260421e-01 3.70234028e-02 -3.20125252e-01 -1.08135104e-01
-1.44908288e-02 2.02976823e-01 -1.42936990e-01 3.38638335e-01
-3.62759596e-03 7.15590000e-01 -8.65017176e-01 -2.76239306e-01
3.93133610e-01 7.60711670e-01 -3.31368446e-01 -1.68651506e-01
1.21284775e-01 6.35843992e-01 -3.62302780e-01 1.00943124e+00
3.72841179e-01 3.23035628e-01 1.64522976e-01 2.44991928e-01
-5.75845182e-01 1.87693834e-01 -9.71413970e-01 1.05983102e+00
-2.94117808e-01 9.19555306e-01 2.02558219e-01 -6.51820004e-01
8.22820663e-01 3.50732058e-01 6.54540420e-01 -1.03937531e+00
2.57380933e-01 6.77812278e-01 3.11940253e-01 -7.58440137e-01
2.35878348e-01 -3.14024568e-01 1.78727239e-01 3.65246385e-01
-5.10779381e-01 5.06125018e-03 2.85345733e-01 -2.28073344e-01
6.58508241e-01 8.65053758e-02 4.30743158e-01 1.09067440e-01
5.09191871e-01 -4.21021581e-01 3.80940348e-01 2.22666711e-01
-3.63599181e-01 4.72200811e-01 2.95062095e-01 -2.67233461e-01
-6.43987894e-01 -1.22745800e+00 -2.49634147e-01 6.10845387e-01
-3.42886120e-01 2.61649758e-01 -4.07823563e-01 2.54672742e-03
2.32615605e-01 2.71622807e-01 -3.57899159e-01 1.09423269e-02
-6.20816052e-01 -4.13114190e-01 4.65339780e-01 4.47329104e-01
6.84200823e-01 -1.27971876e+00 -8.27837944e-01 -3.00182737e-02
-2.89323807e-01 -1.08412302e+00 -3.84335294e-02 -6.02371991e-01
-5.44225097e-01 -1.22768188e+00 -1.08253205e+00 -9.98838365e-01
8.00968826e-01 -1.25537366e-01 4.79420751e-01 4.49076772e-01
-5.54107249e-01 5.87070286e-01 -5.71105778e-01 -4.93745059e-01
-4.80851740e-01 -1.85147896e-01 1.12092860e-01 4.93682101e-02
4.20397222e-01 -6.26440525e-01 -4.47260499e-01 -4.50150631e-02
-5.92515886e-01 -6.43856227e-02 5.76678157e-01 4.82429385e-01
3.00413221e-01 -3.71150225e-01 4.64413375e-01 -1.47929460e-01
5.91100097e-01 6.75197393e-02 -4.64127302e-01 2.96219140e-01
-2.82750815e-01 -1.03569634e-01 3.76204289e-02 -2.97835201e-01
-9.13523436e-01 7.96732977e-02 -2.49492243e-01 7.00417042e-01
-1.96431652e-01 4.59243298e-01 -2.99673826e-01 -2.83096820e-01
1.96105272e-01 2.60204107e-01 2.66709656e-01 -7.64813185e-01
3.61651629e-01 1.16573799e+00 7.11268008e-01 -3.98115903e-01
5.04382670e-01 6.62190497e-01 1.55732468e-01 -1.63077569e+00
-1.18378200e-01 -5.66903651e-01 -7.67845392e-01 -6.88473463e-01
5.63022375e-01 -4.45645690e-01 -8.98287594e-01 9.92224693e-01
-8.11658204e-01 -2.22110659e-01 -3.81047040e-01 7.45342851e-01
-4.00435567e-01 4.19983357e-01 5.22796288e-02 -1.04723895e+00
-1.33603379e-01 -8.61881018e-01 6.92284644e-01 1.85939804e-01
-3.54077876e-01 -5.97329199e-01 1.13363646e-01 6.37196124e-01
3.12875628e-01 5.50759733e-01 1.11838710e+00 -9.41972658e-02
-3.71130019e-01 -5.90385199e-01 -4.26362544e-01 5.04957318e-01
4.13350016e-01 1.58266991e-03 -1.04770446e+00 -2.23705452e-02
-3.86194587e-01 -2.41197169e-01 9.30146217e-01 5.31530619e-01
3.93751949e-01 1.92927092e-01 -4.55892086e-02 1.47716239e-01
1.09850740e+00 2.31080592e-01 6.89147472e-01 1.67023391e-01
4.12748456e-01 9.40053642e-01 1.50938094e-01 4.40911889e-01
6.40781939e-01 7.09949017e-01 1.39708951e-01 2.25024987e-02
-8.22058141e-01 -4.55090433e-01 1.25575606e-02 9.54897225e-01
-5.87708473e-01 1.22110724e-01 -1.24498534e+00 8.24567497e-01
-1.45465565e+00 -9.09992397e-01 -1.18800581e-01 2.52493000e+00
3.89558405e-01 -2.72146255e-01 2.80254871e-01 7.77362227e-01
4.03318882e-01 -2.46074438e-01 -4.88544166e-01 -8.40411708e-02
-4.87277061e-01 3.50772232e-01 6.17130473e-02 5.54286957e-01
-6.58664644e-01 8.82350981e-01 6.22098207e+00 1.13459602e-01
-1.40473676e+00 -3.89958441e-01 -9.00393948e-02 1.05590336e-02
-2.37387344e-01 -1.78864196e-01 -6.48627460e-01 3.71830165e-01
3.93175125e-01 4.60224971e-02 3.28207314e-01 2.41315156e-01
6.42810822e-01 -5.02747774e-01 -6.72172487e-01 1.04374075e+00
2.44691417e-01 -7.87842691e-01 -4.08317819e-02 3.42722118e-01
3.44244003e-01 -8.52593109e-02 1.23169042e-01 -3.04776371e-01
-1.79445669e-01 -8.82974565e-01 8.32098007e-01 9.10891175e-01
1.08778036e+00 -2.01082826e-01 5.97194374e-01 1.29837975e-01
-1.13552845e+00 -2.66851187e-01 2.26626351e-01 -4.08115268e-01
5.72579741e-01 2.89164692e-01 -4.77820247e-01 2.74393521e-02
5.16307294e-01 3.36637020e-01 -3.60139579e-01 1.54714000e+00
-5.47285736e-01 4.70152169e-01 -5.27910531e-01 -3.00699711e-01
-2.73883194e-01 -2.11720139e-01 6.45176113e-01 5.47374904e-01
5.48185587e-01 6.58354862e-03 -2.64122427e-01 3.30797017e-01
2.15711713e-01 4.42897558e-01 -5.52245498e-01 -3.12584192e-01
4.61412579e-01 5.83565176e-01 -6.26252711e-01 1.67437851e-01
-5.26835501e-01 7.79754460e-01 2.20381290e-01 2.83767909e-01
5.27011082e-02 -2.57425308e-01 9.48925734e-01 4.24761772e-01
1.44432802e-02 -4.53987151e-01 -4.33769614e-01 -8.03842664e-01
5.03317595e-01 -8.22841823e-01 2.58319020e-01 -4.28622425e-01
-9.27383363e-01 2.69088954e-01 -2.32045799e-01 -1.04905140e+00
-2.09235251e-01 -9.73708868e-01 -2.63546139e-01 1.00264060e+00
-1.56955862e+00 -1.53986824e+00 -4.18692559e-01 1.32969663e-01
-1.63222384e-02 -1.40786037e-01 9.80150342e-01 5.25100470e-01
-2.11643890e-01 5.34719467e-01 7.51694739e-02 2.14952663e-01
4.72322702e-01 -7.81656444e-01 9.43235457e-02 5.60220540e-01
9.10946950e-02 5.55892527e-01 2.23945841e-01 -6.05292737e-01
-8.68241608e-01 -1.41530514e-01 1.54110360e+00 -2.98303723e-01
3.83424670e-01 -1.09623916e-01 -4.07801121e-01 1.78364649e-01
-5.09815097e-01 -3.72806758e-01 8.80423546e-01 6.22317009e-02
-2.79729515e-01 -2.67179497e-02 -1.24753177e+00 7.88337409e-01
1.26556575e+00 -7.57470012e-01 -4.32285339e-01 -1.20666496e-01
-5.54428808e-02 -5.89341223e-02 -5.85433364e-01 3.09294671e-01
1.43946171e+00 -9.97307658e-01 8.72111738e-01 -4.91202950e-01
1.65499356e-02 -1.59090906e-01 -1.91139668e-01 -8.70405436e-01
3.27745587e-01 -4.00075823e-01 1.59218073e-01 1.27015126e+00
1.84545130e-01 -9.13833618e-01 7.90365279e-01 8.75410676e-01
5.31988852e-02 -7.56049275e-01 -1.29533935e+00 -7.40142822e-01
3.57320905e-02 -7.08515346e-01 6.60272121e-01 4.68310267e-01
3.08622755e-02 -5.01566470e-01 -2.17577532e-01 -3.16547632e-01
5.38284719e-01 -1.51967078e-01 6.53276920e-01 -1.66326761e+00
1.52993485e-01 -1.02565205e+00 -8.60410094e-01 -7.91502655e-01
-8.33317265e-02 -6.60871089e-01 -1.62994325e-01 -1.97741091e+00
-6.92235082e-02 -3.38466436e-01 -5.31866290e-02 4.81669962e-01
2.34159902e-01 3.36888283e-01 2.87117779e-01 8.59378651e-02
2.98328519e-01 3.33316684e-01 1.36978567e+00 1.89445674e-01
-4.20372725e-01 2.42293984e-01 -7.90965021e-01 8.59659314e-01
8.26644659e-01 8.65888819e-02 -1.32268876e-01 -2.88391113e-01
2.05907404e-01 -2.22772926e-01 4.70807791e-01 -9.39295650e-01
1.73309371e-02 -1.45091549e-01 -6.85527772e-02 -3.89079541e-01
5.52594483e-01 -6.59166217e-01 -3.23154658e-01 6.44509494e-01
7.23107457e-02 -4.02705550e-01 -5.43008335e-02 -2.24635471e-02
-2.22134233e-01 -7.03220908e-03 7.70529330e-01 2.33191680e-02
-8.82152736e-01 -1.89372115e-02 -5.03467858e-01 1.15895867e-01
7.09114909e-01 -7.26513267e-01 -2.02085868e-01 -5.84479451e-01
-7.36673117e-01 -7.50666633e-02 4.45011437e-01 4.76391971e-01
6.82398558e-01 -1.03059840e+00 -5.81862628e-01 4.96507704e-01
3.74397099e-01 -3.27973485e-01 2.09661096e-01 6.97219491e-01
-7.32912302e-01 4.29562837e-01 -5.22561908e-01 4.19195816e-02
-1.61727798e+00 -4.03727204e-01 1.26175463e-01 2.64097244e-01
-6.62360489e-01 8.63540590e-01 -5.02168596e-01 -2.53069282e-01
4.66290206e-01 -3.40505809e-01 -4.55163598e-01 2.41868153e-01
5.60809553e-01 8.02060485e-01 -2.80600250e-01 -1.19557166e+00
-3.76902431e-01 9.90881383e-01 3.34223360e-01 -4.84671623e-01
1.17508578e+00 1.35014564e-01 -3.42064470e-01 4.70970452e-01
6.56144083e-01 2.72584945e-01 -8.07624161e-01 8.74468498e-03
1.61196589e-01 -5.90186298e-01 -7.19245002e-02 -1.03789568e+00
-6.55613065e-01 9.82937098e-01 7.80085444e-01 -2.89209515e-01
1.08754444e+00 1.98148161e-01 7.16817498e-01 2.22242787e-01
5.63686430e-01 -1.19875026e+00 -1.69217736e-01 3.70719731e-01
9.94521737e-01 -1.17230368e+00 -7.84864053e-02 -3.20797622e-01
-1.76854372e-01 7.37630248e-01 1.05205178e-01 4.47279364e-01
7.82925785e-01 -5.22000603e-02 7.66540349e-01 -1.02095217e-01
1.99040323e-01 -9.67996061e-01 5.09154618e-01 1.00530887e+00
6.36896133e-01 3.41542393e-01 -9.76262867e-01 3.99832308e-01
-5.14194489e-01 1.82171300e-01 1.25714511e-01 9.51929808e-01
-5.18005133e-01 -1.61936140e+00 -3.41367185e-01 6.64668143e-01
5.10472357e-02 1.41077459e-01 -6.60618186e-01 8.75579476e-01
3.92469704e-01 8.55787158e-01 -1.75190240e-01 -8.96485522e-03
6.16394937e-01 2.70846039e-01 7.40795076e-01 -3.69151860e-01
-7.39030838e-02 -4.54485893e-01 2.71543503e-01 -1.18400477e-01
-3.27396959e-01 -9.93334770e-01 -1.31731522e+00 -3.16920988e-02
5.38905412e-02 -3.91902894e-01 1.08132362e+00 9.23989534e-01
4.78313863e-02 3.83577123e-02 8.18719119e-02 -4.37096745e-01
-3.88324857e-01 -8.08013141e-01 -8.81977975e-01 3.91807348e-01
3.70503694e-01 -7.86616206e-01 -1.94617689e-01 -1.62161112e-01] | [9.035935401916504, -6.342881679534912] |
aa23a1d8-89e5-473e-9bcc-a0d3879b0776 | center-feature-fusion-selective-multi-sensor | 2209.12880 | null | https://arxiv.org/abs/2209.12880v2 | https://arxiv.org/pdf/2209.12880v2.pdf | Center Feature Fusion: Selective Multi-Sensor Fusion of Center-based Objects | Leveraging multi-modal fusion, especially between camera and LiDAR, has become essential for building accurate and robust 3D object detection systems for autonomous vehicles. Until recently, point decorating approaches, in which point clouds are augmented with camera features, have been the dominant approach in the field. However, these approaches fail to utilize the higher resolution images from cameras. Recent works projecting camera features to the bird's-eye-view (BEV) space for fusion have also been proposed, however they require projecting millions of pixels, most of which only contain background information. In this work, we propose a novel approach Center Feature Fusion (CFF), in which we leverage center-based detection networks in both the camera and LiDAR streams to identify relevant object locations. We then use the center-based detection to identify the locations of pixel features relevant to object locations, a small fraction of the total number in the image. These are then projected and fused in the BEV frame. On the nuScenes dataset, we outperform the LiDAR-only baseline by 4.9% mAP while fusing up to 100x fewer features than other fusion methods. | ['Ming C. Wu', 'Masayoshi Tomizuka', 'Wei Zhan', 'Yiyang Zhou', 'Philip Jacobson'] | 2022-09-26 | null | null | null | null | ['robust-3d-object-detection'] | ['computer-vision'] | [ 1.86817031e-02 -6.25772715e-01 4.38494943e-02 -3.76226634e-01
-8.50325167e-01 -9.70367610e-01 7.75769770e-01 8.18063319e-02
-4.96257126e-01 2.59224921e-01 -2.45389357e-01 -6.76532760e-02
3.28459859e-01 -1.01764703e+00 -9.04500484e-01 -6.18031979e-01
4.65344846e-01 2.51120836e-01 8.61049891e-01 -1.40008181e-01
1.96548894e-01 9.59313452e-01 -2.02284575e+00 1.67601302e-01
6.81950033e-01 1.09082782e+00 4.52376455e-01 5.57656169e-01
-2.89928883e-01 1.50320545e-01 -2.28916049e-01 -1.50869459e-01
5.48075378e-01 3.08760136e-01 2.39761844e-01 3.38750482e-02
1.18974090e+00 -7.18562543e-01 -2.93726921e-01 1.25825000e+00
1.92100316e-01 1.41980778e-02 5.02563834e-01 -1.40000117e+00
-2.38577068e-01 -1.24098457e-01 -1.14571881e+00 1.92317352e-01
2.06625015e-01 7.36849755e-02 7.60434210e-01 -1.52353358e+00
3.76429319e-01 1.51591587e+00 6.38976693e-01 1.82753101e-01
-9.83405411e-01 -1.17487621e+00 1.86433345e-01 7.49182031e-02
-1.59072745e+00 -4.95129794e-01 6.43692374e-01 -4.65338051e-01
9.01194811e-01 2.69938320e-01 6.78644598e-01 5.05011559e-01
1.84427109e-02 5.34370124e-01 9.52173412e-01 -1.89089358e-01
-1.15830317e-01 2.78532475e-01 1.41181275e-01 4.95567411e-01
7.94045031e-01 3.10527593e-01 -7.12336481e-01 -2.15176895e-01
6.02588058e-01 7.26033092e-01 -6.85692281e-02 -4.98914897e-01
-1.15543365e+00 8.74645293e-01 6.82286322e-01 -2.30931744e-01
-4.59808558e-01 2.57567197e-01 -8.60323906e-02 -2.93710589e-01
6.40593290e-01 -2.44718820e-01 -1.58193767e-01 4.12139386e-01
-9.44080591e-01 4.84917551e-01 2.22489506e-01 1.20278084e+00
1.31229210e+00 -3.91924411e-01 2.56050434e-02 4.45843101e-01
6.28417373e-01 1.29936886e+00 -3.61617625e-01 -1.14316142e+00
7.40793526e-01 1.10002863e+00 3.09338659e-01 -1.08326638e+00
-2.14607388e-01 -8.02947134e-02 -5.58681726e-01 5.03852248e-01
2.78545413e-02 -2.21269742e-01 -9.78648305e-01 1.28765547e+00
7.24766970e-01 3.72933835e-01 1.17776103e-01 9.45201278e-01
7.93255806e-01 7.64297247e-01 -2.90343821e-01 2.11906746e-01
1.39864600e+00 -5.29249966e-01 -3.89802754e-01 -3.77416313e-01
3.59061837e-01 -8.97604108e-01 3.46777588e-01 -3.41017311e-03
-7.68355250e-01 -5.11005998e-01 -1.05741990e+00 -8.31401721e-02
-5.53994179e-01 2.77088791e-01 4.49122995e-01 6.01716220e-01
-1.07349217e+00 -1.67174190e-02 -8.21072638e-01 -5.55651307e-01
5.81503570e-01 2.51126140e-01 -3.94551218e-01 -5.09275496e-01
-6.03245437e-01 9.68642056e-01 2.87953854e-01 -1.50236979e-01
-8.48698616e-01 -8.17249298e-01 -8.96719635e-01 -2.61871312e-02
5.38026690e-01 -6.34633660e-01 8.08300376e-01 -2.40988821e-01
-7.49357045e-01 5.78511715e-01 -5.45470536e-01 -4.53137159e-01
3.79221827e-01 -4.22866553e-01 -1.96724668e-01 2.35853091e-01
2.88557827e-01 1.19534898e+00 8.22171986e-01 -1.40578163e+00
-1.44511807e+00 -6.66934311e-01 3.02406829e-02 3.31216395e-01
-7.12163001e-02 1.79337740e-01 -6.71669185e-01 -2.62192022e-02
4.90594685e-01 -1.07833219e+00 -2.32195497e-01 5.20554006e-01
-7.30461180e-02 -2.38037020e-01 1.52856290e+00 -8.11351612e-02
6.96228385e-01 -2.33010507e+00 -3.11123401e-01 6.52510151e-02
4.21646506e-01 1.73497304e-01 -1.59705859e-02 3.15470695e-01
3.97237450e-01 -4.24820483e-02 1.73887014e-02 -5.50064743e-01
-2.26409510e-01 2.42299959e-01 -6.04605198e-01 5.94257295e-01
4.95122463e-01 8.20774615e-01 -7.48979330e-01 -5.35319149e-01
7.90395975e-01 6.82947993e-01 -5.56548297e-01 -1.15777351e-01
-9.35603753e-02 1.24988250e-01 -4.67549831e-01 8.94303203e-01
1.31062031e+00 1.32388443e-01 -5.09793282e-01 -3.76395941e-01
-6.40554070e-01 -2.45858893e-01 -1.15378034e+00 1.40390527e+00
-8.32978114e-02 7.43767262e-01 2.39862844e-01 -2.80379027e-01
9.27176476e-01 -1.37219802e-01 5.89182258e-01 -2.06725821e-01
-7.27820918e-02 6.94223270e-02 -3.43285620e-01 2.05539935e-03
8.60546410e-01 1.66930586e-01 5.83115704e-02 4.62785773e-02
-1.05695337e-01 -3.91680896e-01 2.32769579e-01 3.93457532e-01
9.74613190e-01 5.65956980e-02 1.21761940e-01 1.73458412e-01
4.33031231e-01 4.02715683e-01 5.94839454e-01 6.72518671e-01
-2.47690707e-01 7.83194780e-01 -1.34094939e-01 -2.19426930e-01
-9.64709461e-01 -1.00521934e+00 -2.26314142e-01 6.39040112e-01
5.71771920e-01 -4.88278925e-01 -3.95930976e-01 -5.92006266e-01
5.33682287e-01 5.86421311e-01 -2.67425537e-01 -2.41042627e-03
-4.09890115e-01 -4.05379981e-01 2.35430539e-01 5.47050595e-01
6.23052001e-01 -1.07895873e-01 -1.05668342e+00 -1.93374362e-02
-9.70826764e-03 -1.42695439e+00 -2.71511555e-01 -4.42269631e-02
-7.71896124e-01 -1.00799775e+00 -4.90291178e-01 -1.99598625e-01
6.11488640e-01 1.42344332e+00 6.60989106e-01 -1.29905418e-01
-2.07465664e-01 4.98560309e-01 -3.08682561e-01 -8.38211894e-01
1.32718250e-01 -4.02116686e-01 2.37991467e-01 1.04037844e-01
8.12145114e-01 -2.41215795e-01 -5.95499814e-01 3.72128516e-01
-5.41363597e-01 8.53377283e-02 7.50786662e-01 2.77314633e-01
7.19414532e-01 -2.47374266e-01 2.60298662e-02 -3.98879200e-01
-2.17907697e-01 -5.27817488e-01 -1.19555163e+00 -5.51565960e-02
-2.16134667e-01 -3.72359991e-01 -1.19711362e-01 -9.70431268e-02
-8.62067342e-01 6.14082575e-01 2.32731313e-01 -1.14649045e+00
-2.72871941e-01 1.38884977e-01 -1.53729737e-01 -4.33770835e-01
4.97427404e-01 7.24875033e-02 -9.99193862e-02 -3.99213374e-01
6.11433983e-01 7.94168591e-01 4.87363786e-01 -1.63064897e-01
1.12552977e+00 1.06205797e+00 2.24762246e-01 -9.62687194e-01
-7.76645660e-01 -1.06672895e+00 -8.72767270e-01 -3.60561728e-01
8.86551261e-01 -1.45311546e+00 -4.74721253e-01 1.22461498e-01
-1.53295290e+00 5.47883987e-01 -8.01928192e-02 6.14050865e-01
-1.24122910e-01 2.36800581e-01 4.16955166e-02 -1.11123455e+00
-8.02629516e-02 -1.19308019e+00 1.69625175e+00 3.96757543e-01
4.33590382e-01 -2.21254393e-01 -1.42072082e-01 2.53637701e-01
2.49093845e-01 2.36963540e-01 2.98982799e-01 -1.94103152e-01
-1.11158073e+00 -5.06659925e-01 -7.26985514e-01 1.06687143e-01
5.34384660e-02 2.97667593e-01 -1.12264442e+00 -9.61683989e-02
-1.38154000e-01 2.26982847e-01 1.19469774e+00 3.47632974e-01
5.37627757e-01 3.99396956e-01 -8.79673839e-01 6.06542528e-01
1.56130910e+00 1.96620356e-02 2.02963829e-01 -6.51070178e-02
8.14616024e-01 5.07507682e-01 9.40497577e-01 4.09918517e-01
5.99050105e-01 7.03483164e-01 8.62785459e-01 -1.42499954e-01
-2.33444273e-01 -1.51311636e-01 5.56028962e-01 1.80348292e-01
1.02845673e-03 -1.07386298e-02 -1.05414939e+00 5.38372874e-01
-1.84611833e+00 -9.30852056e-01 -4.43204224e-01 2.14876795e+00
1.78480253e-01 2.34803632e-02 -7.78366029e-02 -3.02980751e-01
9.20395255e-01 9.90446098e-03 -6.29469812e-01 4.92088407e-01
-2.68956125e-01 -2.56918401e-01 1.15433681e+00 2.98425049e-01
-1.29404533e+00 9.68788087e-01 5.64570522e+00 4.05802548e-01
-1.04129815e+00 2.16782123e-01 -1.56825967e-03 -3.47251594e-01
-9.48974714e-02 2.48300076e-01 -1.64461207e+00 3.47676665e-01
7.91575074e-01 -1.49467308e-02 -2.92557548e-03 8.31992686e-01
2.17248887e-01 -5.65450311e-01 -9.14360285e-01 1.27765596e+00
3.65149438e-01 -1.35502779e+00 2.90024988e-02 3.59865159e-01
8.19916010e-01 6.91056013e-01 -3.76863517e-02 6.35468110e-04
3.84621888e-01 -5.17326415e-01 8.85086656e-01 4.92176801e-01
6.91011786e-01 -6.87715232e-01 5.94572186e-01 5.64349830e-01
-1.63263392e+00 -1.83132917e-01 -7.02282369e-01 1.09171368e-01
3.92493963e-01 7.02225327e-01 -7.13792086e-01 4.49834049e-01
1.08337486e+00 9.09940839e-01 -8.65640163e-01 1.18884802e+00
8.67869854e-02 3.02862346e-01 -8.88893306e-01 1.49920940e-01
2.65135467e-01 -2.78896302e-01 7.16936052e-01 8.67855728e-01
8.24876904e-01 1.44313395e-01 4.68948662e-01 9.96772945e-01
-3.13922837e-02 -3.43991488e-01 -1.04291236e+00 3.80004287e-01
8.30184221e-01 1.52587640e+00 -5.92521608e-01 -4.87539172e-01
-9.38086987e-01 5.17631412e-01 -6.97585717e-02 2.36371964e-01
-8.66303921e-01 -1.94992095e-01 9.62680161e-01 2.45201215e-01
6.27010047e-01 -5.79819322e-01 -1.55363679e-01 -9.83669221e-01
-2.65202280e-02 -1.09418385e-01 -3.74251418e-03 -1.00128746e+00
-1.03797817e+00 1.12708449e-01 2.93433011e-01 -1.55377805e+00
5.62664196e-02 -6.31031275e-01 -4.23989177e-01 1.08818054e+00
-1.83728123e+00 -1.50809526e+00 -7.66077697e-01 6.48012340e-01
4.77390230e-01 -4.26222198e-02 3.27087194e-01 3.68797421e-01
-4.51553941e-01 -1.64534286e-01 1.33400917e-01 -1.65146440e-01
7.14575171e-01 -9.00872588e-01 4.04268920e-01 1.07020009e+00
2.58128047e-01 4.40311104e-01 3.66218150e-01 -8.97046924e-01
-1.73652422e+00 -1.50668049e+00 5.66505909e-01 -8.73159409e-01
4.00464416e-01 -5.13578296e-01 -6.51671946e-01 6.08012915e-01
7.07687810e-02 4.30237681e-01 2.59239078e-01 -3.70189220e-01
-2.59543538e-01 -4.08653915e-01 -1.08011508e+00 4.64222848e-01
9.19291019e-01 -3.03675950e-01 -5.10386169e-01 2.08375737e-01
8.99574578e-01 -4.70649213e-01 -3.90134066e-01 5.27869463e-01
3.51211578e-01 -8.80447030e-01 1.13199317e+00 7.98581615e-02
-4.78308462e-02 -1.02184105e+00 -6.47428036e-01 -9.58614469e-01
-2.06632242e-01 -1.86778679e-02 -7.56477565e-02 1.33908010e+00
1.09689608e-01 -6.33279800e-01 7.45697021e-01 3.62134874e-01
-3.18390429e-01 -1.53227681e-02 -9.86770272e-01 -5.20813584e-01
-3.77236962e-01 -6.04407966e-01 7.21470714e-01 4.37398136e-01
-7.15844095e-01 3.87399793e-01 -3.10966372e-02 6.85436010e-01
8.43430579e-01 2.97879010e-01 1.15370929e+00 -1.57284200e+00
5.27851701e-01 -2.31326565e-01 -7.25960732e-01 -9.44120884e-01
-8.78830105e-02 -7.64746785e-01 1.62426114e-01 -1.53137946e+00
3.09758037e-01 -4.14042383e-01 -4.72329259e-02 3.57335538e-01
-2.37937182e-01 4.94413763e-01 5.04562736e-01 4.12325948e-01
-6.14417076e-01 3.82648051e-01 5.78282595e-01 -1.61591321e-01
1.68188792e-02 -4.93978523e-02 -3.65750402e-01 8.45276535e-01
4.65389371e-01 -4.60739404e-01 -5.49659505e-02 -6.43892229e-01
5.14236502e-02 -3.44029099e-01 6.58735394e-01 -1.42925751e+00
6.78186655e-01 -1.97769374e-01 7.06724226e-01 -1.78604591e+00
8.71952891e-01 -1.35490286e+00 1.69204757e-01 2.27235109e-01
4.05893028e-01 1.92657351e-01 3.69503409e-01 8.92931879e-01
-1.98400259e-01 -2.92195976e-02 7.19696999e-01 -9.84852687e-02
-9.59673882e-01 4.79609281e-01 -1.16431192e-01 -5.84483147e-01
1.52153456e+00 -3.95918936e-01 -4.03638840e-01 2.06134859e-02
1.65920891e-02 4.83780086e-01 7.31438220e-01 3.69099140e-01
9.20119822e-01 -1.39858365e+00 -8.09542477e-01 4.16942924e-01
4.99868691e-01 4.49850678e-01 2.71281749e-01 8.94508541e-01
-5.01066267e-01 6.61919475e-01 -1.19141331e-02 -1.37436724e+00
-1.52982616e+00 6.28103197e-01 6.06211796e-02 5.69496214e-01
-6.33006573e-01 6.32426202e-01 4.10216093e-01 -2.35943913e-01
5.80169223e-02 -4.73945051e-01 -1.59157380e-01 3.00764263e-01
6.77597463e-01 3.36270511e-01 8.17718729e-02 -1.16402674e+00
-5.86748242e-01 1.16209686e+00 9.25905332e-02 -1.87706634e-01
1.24446058e+00 -4.22587723e-01 -4.16258946e-02 2.94304103e-01
8.23310554e-01 1.37476861e-01 -1.48472190e+00 -4.92238432e-01
-1.83307678e-01 -8.48539472e-01 4.30838376e-01 -2.67040521e-01
-1.04566944e+00 1.13360274e+00 9.27198291e-01 -1.20244622e-01
7.81304359e-01 2.12489933e-01 4.73035604e-01 3.79311293e-01
6.57185853e-01 -7.18189597e-01 -3.01407516e-01 4.79404300e-01
6.23023689e-01 -1.67007399e+00 2.49544322e-01 -6.59417212e-01
-3.83814335e-01 1.03467882e+00 7.24696517e-01 -1.25304013e-01
6.23747706e-01 3.08246970e-01 -7.56668299e-02 -2.24351466e-01
-7.72903562e-01 -7.30950356e-01 2.01332912e-01 7.12981761e-01
-2.51462340e-01 8.68128054e-03 3.66681904e-01 1.31564051e-01
1.87230065e-01 -2.47416571e-01 3.32180798e-01 1.02408803e+00
-1.01116049e+00 -6.76234901e-01 -9.65992987e-01 7.64052868e-01
9.83396024e-02 -2.90583409e-02 -4.25457656e-01 6.40439034e-01
3.78495574e-01 1.09224474e+00 4.09575433e-01 -4.35599715e-01
4.63093340e-01 -1.66745499e-01 2.51566648e-01 -8.06443930e-01
-1.55605301e-01 2.96368957e-01 -3.42914999e-01 -7.22931266e-01
-4.93111312e-01 -9.60915744e-01 -1.03594637e+00 -2.27579474e-01
-7.34625041e-01 -3.45750928e-01 1.00598025e+00 6.19705081e-01
5.64120233e-01 6.49939403e-02 5.31380057e-01 -1.47958863e+00
-3.08207840e-01 -7.25163817e-01 -4.35911506e-01 -1.01173714e-01
5.87590396e-01 -1.07451928e+00 -3.15059483e-01 -7.80237392e-02] | [7.721103191375732, -2.4248061180114746] |
a9e90d2d-cfa7-4da1-a015-d5ac0b7a51cc | a-functional-regression-approach-to-facial | 1612.02203 | null | http://arxiv.org/abs/1612.02203v2 | http://arxiv.org/pdf/1612.02203v2.pdf | A Functional Regression approach to Facial Landmark Tracking | Linear regression is a fundamental building block in many face detection and
tracking algorithms, typically used to predict shape displacements from image
features through a linear mapping. This paper presents a Functional Regression
solution to the least squares problem, which we coin Continuous Regression,
resulting in the first real-time incremental face tracker. Contrary to prior
work in Functional Regression, in which B-splines or Fourier series were used,
we propose to approximate the input space by its first-order Taylor expansion,
yielding a closed-form solution for the continuous domain of displacements. We
then extend the continuous least squares problem to correlated variables, and
demonstrate the generalisation of our approach. We incorporate Continuous
Regression into the cascaded regression framework, and show its computational
benefits for both training and testing. We then present a fast approach for
incremental learning within Cascaded Continuous Regression, coined iCCR, and
show that its complexity allows real-time face tracking, being 20 times faster
than the state of the art. To the best of our knowledge, this is the first
incremental face tracker that is shown to operate in real-time. We show that
iCCR achieves state-of-the-art performance on the 300-VW dataset, the most
recent, large-scale benchmark for face tracking. | ['Fernando de la Torre', 'Enrique Sánchez-Lozano', 'Michel Valstar', 'Georgios Tzimiropoulos', 'Brais Martinez'] | 2016-12-07 | null | null | null | null | ['landmark-tracking'] | ['computer-vision'] | [ 6.47248998e-02 -1.85581505e-01 -2.28693619e-01 -1.27828270e-01
-8.82481098e-01 -3.24939996e-01 5.48373640e-01 -5.64982593e-01
-2.06751168e-01 5.13640881e-01 -2.94025630e-01 -3.40561807e-01
-2.90054344e-02 -1.04105525e-01 -1.04606700e+00 -6.97547615e-01
-3.22030187e-01 3.86973798e-01 2.22185656e-01 -9.57794785e-02
-1.51930912e-03 8.60567868e-01 -1.59520543e+00 -1.51171699e-01
2.18194261e-01 9.82066810e-01 -4.01286811e-01 7.42378175e-01
3.18227261e-01 3.84936064e-01 -4.37772781e-01 -2.22292632e-01
4.18134332e-01 -3.70137542e-01 -4.50936466e-01 2.11795911e-01
1.12215698e+00 -3.50021005e-01 -1.48914099e-01 5.97947598e-01
7.35458255e-01 -5.98713271e-02 6.05367005e-01 -1.55525494e+00
-7.51754284e-01 9.03514996e-02 -1.04889083e+00 3.85182537e-02
5.52896738e-01 -2.09167033e-01 5.31767130e-01 -1.31947279e+00
5.28393686e-01 1.51380885e+00 1.35115099e+00 1.05782259e+00
-1.65531838e+00 -9.22161043e-01 3.75336587e-01 -4.00291771e-01
-1.40288317e+00 -8.93562436e-01 6.27914548e-01 -6.22025728e-01
8.64876390e-01 2.27022350e-01 5.32222092e-01 8.87214780e-01
-9.08722356e-02 6.65929735e-01 1.14073837e+00 -6.25623941e-01
-1.72317520e-01 6.79872409e-02 -1.04553193e-01 1.04303563e+00
1.11853406e-02 5.92255831e-01 -4.73847598e-01 -3.45658004e-01
9.62954104e-01 -1.26025647e-01 -2.46301100e-01 -5.86835980e-01
-7.15750337e-01 8.29169750e-01 3.17465544e-01 1.58642143e-01
-7.08004385e-02 5.80679178e-01 2.05795839e-01 4.30346638e-01
7.56801426e-01 -2.06673339e-01 -7.37336934e-01 7.90931880e-02
-1.27052987e+00 1.10159047e-01 8.40100825e-01 8.81709635e-01
4.51062799e-01 3.47338974e-01 -2.85477508e-02 7.53479779e-01
7.10665524e-01 7.79339373e-01 2.34764278e-01 -9.31317031e-01
-7.34464154e-02 1.70312062e-01 9.83451977e-02 -5.95708966e-01
-3.57912093e-01 -2.68422067e-01 -3.86255413e-01 7.88863778e-01
7.50655651e-01 -3.86595130e-01 -5.94119489e-01 1.69189012e+00
7.30257988e-01 8.61658633e-01 -3.90103430e-01 4.68770593e-01
6.59071445e-01 3.36540759e-01 -2.56053150e-01 -8.00775051e-01
1.11178923e+00 -7.75814533e-01 -4.80808914e-01 1.91579461e-01
3.73929352e-01 -8.10281396e-01 3.39556873e-01 4.79370952e-01
-1.00816262e+00 -5.83644927e-01 -7.49133646e-01 1.99097261e-01
-9.84909013e-02 1.83242038e-01 5.86122155e-01 1.01489186e+00
-1.54886472e+00 6.91011727e-01 -1.04101527e+00 -3.42381239e-01
5.63378751e-01 8.39866281e-01 -3.78929108e-01 3.57929945e-01
-5.65825820e-01 9.14931059e-01 -4.67982799e-01 2.46433154e-01
-6.19170725e-01 -1.22329807e+00 -7.67875075e-01 -5.03913343e-01
4.05268818e-01 -3.17695796e-01 1.41763735e+00 -9.34048355e-01
-1.66719091e+00 9.55779791e-01 -7.71022439e-01 -4.08307612e-01
6.77448928e-01 -3.91936481e-01 -3.00737053e-01 -3.19250911e-01
-2.75095373e-01 8.08540344e-01 1.52723563e+00 -1.21338618e+00
-3.17547113e-01 -4.87490028e-01 -2.98477292e-01 -2.78577119e-01
-3.81351233e-01 5.48384666e-01 -5.84605873e-01 -5.93993902e-01
-2.16526315e-01 -1.13898504e+00 -1.76451504e-02 6.45783067e-01
2.29867697e-01 -6.37108326e-01 1.25139844e+00 -7.35989809e-01
1.17389119e+00 -1.94940090e+00 5.36108054e-02 -3.76361259e-03
1.81316644e-01 2.83004284e-01 6.19765222e-02 -1.73863545e-01
-4.19297189e-01 -2.83654124e-01 -1.67545006e-01 -7.92699933e-01
-6.60342425e-02 -6.80828020e-02 -4.56035107e-01 1.06437683e+00
3.04971725e-01 1.05650759e+00 -7.72543490e-01 -5.39694667e-01
7.10459650e-02 1.08598995e+00 -4.43602085e-01 -3.96706648e-02
-1.25681430e-01 5.70468485e-01 -3.69148031e-02 6.96314633e-01
7.44028866e-01 1.60657048e-01 -1.88725576e-01 -1.29090533e-01
-4.04819965e-01 -4.11410511e-01 -1.13533223e+00 1.45737934e+00
-4.85183418e-01 1.08726001e+00 3.00032407e-01 -1.06009853e+00
1.04848981e+00 6.06892645e-01 1.01554787e+00 -3.81386250e-01
-9.16967094e-02 2.13723019e-01 -4.09064084e-01 -7.35137612e-03
1.06410220e-01 -1.58351272e-01 2.87259430e-01 6.75026178e-02
2.26657763e-02 9.16647986e-02 -1.78014815e-01 -1.66179672e-01
7.94255316e-01 9.66697931e-01 2.07610145e-01 -1.19736686e-01
8.09716225e-01 -1.84113234e-01 3.21390241e-01 4.15931344e-01
-1.96635857e-01 5.33382595e-01 3.66290510e-01 -4.26420599e-01
-7.20748901e-01 -7.05643773e-01 -5.97587347e-01 1.29815102e+00
-4.48266894e-01 -1.96124166e-01 -7.16435611e-01 -6.54891372e-01
5.11737227e-01 9.62320641e-02 -8.33344638e-01 4.08369511e-01
-1.13164794e+00 -5.20195246e-01 5.74077070e-01 6.48501694e-01
-1.56093985e-01 -1.07166862e+00 -3.89059871e-01 1.70648158e-01
6.95313334e-01 -1.03212726e+00 -8.35995972e-01 6.27707392e-02
-9.56421912e-01 -1.24953711e+00 -7.15359271e-01 -8.20106864e-01
7.63651609e-01 2.59952396e-02 9.83535349e-01 3.40729296e-01
-6.74090326e-01 8.88326883e-01 1.15282409e-01 -1.85270831e-01
-2.99042583e-01 -4.68557507e-01 2.70049244e-01 -6.36267886e-02
2.16815457e-01 -1.88197464e-01 -4.75979060e-01 4.34830248e-01
-2.01057732e-01 -5.28661907e-01 2.27335066e-01 8.73687685e-01
5.45649529e-01 -2.45754302e-01 4.27007020e-01 -8.09732318e-01
2.70357102e-01 -2.49334350e-01 -1.25753856e+00 1.56049520e-01
-6.54338062e-01 5.59224412e-02 5.83824813e-01 -8.76030385e-01
-9.14703131e-01 6.41389132e-01 -2.52014250e-01 -8.61420572e-01
2.08921731e-01 -1.01893544e-01 2.99003005e-01 -8.79768014e-01
7.30200827e-01 9.76638794e-02 3.08870047e-01 -5.17593026e-01
5.53100228e-01 4.10790265e-01 6.18076503e-01 -5.62934279e-01
1.17251325e+00 7.07049847e-01 5.33531785e-01 -9.35049534e-01
-5.22403419e-01 -6.22536063e-01 -9.35627580e-01 -4.15596843e-01
4.51518804e-01 -8.73135746e-01 -1.26094568e+00 3.84805113e-01
-1.09431577e+00 -4.69705075e-01 -2.15754107e-01 8.96315649e-02
-6.86227202e-01 2.44736597e-01 -3.51538569e-01 -1.17725277e+00
-3.78947109e-01 -9.22861159e-01 1.42698765e+00 -4.21796106e-02
-2.11918876e-02 -1.22056973e+00 3.56080920e-01 -1.11948863e-01
3.49166870e-01 5.31262755e-01 -5.76490983e-02 -1.77791774e-01
-1.33334130e-01 -1.12506039e-01 -1.08984023e-01 1.22431926e-01
1.82345450e-01 2.97495037e-01 -9.80065763e-01 -9.31384921e-01
1.98330823e-02 -1.63618460e-01 8.61434460e-01 6.44189715e-01
8.66402268e-01 -1.91581264e-01 -8.09984326e-01 7.65451372e-01
1.40096903e+00 -3.30058928e-03 3.55131775e-01 -5.90385422e-02
7.83860385e-01 4.03897971e-01 6.09672368e-01 1.49928182e-01
4.39549610e-02 1.16606307e+00 9.00847837e-02 -2.14675546e-01
-4.96723235e-01 -1.48279294e-01 8.68157148e-01 4.35524017e-01
-1.19475372e-01 5.43667734e-01 -6.20480597e-01 3.37496936e-01
-1.79494178e+00 -1.13772023e+00 -2.26836994e-01 2.43152928e+00
8.20751727e-01 -2.30333418e-01 5.89734137e-01 -4.19830456e-02
6.72897577e-01 -2.16204092e-01 -6.74393713e-01 -3.06257278e-01
4.04256374e-01 6.33277774e-01 5.77383995e-01 7.99210668e-01
-1.45989752e+00 1.04690516e+00 7.21498680e+00 7.14190900e-01
-1.26535845e+00 3.39564174e-01 3.38628262e-01 -2.58367658e-01
3.90695959e-01 -1.52128369e-01 -1.47544563e+00 3.19285661e-01
1.02887714e+00 2.46283099e-01 7.18444765e-01 9.96634364e-01
6.27708063e-02 3.58805239e-01 -1.23631358e+00 9.89017904e-01
4.19694036e-01 -8.84479821e-01 -7.63306141e-01 1.96842462e-01
8.54801238e-01 -1.09645620e-01 2.62225837e-01 4.91502643e-01
3.70991647e-01 -1.26347280e+00 5.39170027e-01 5.33767283e-01
1.19781280e+00 -4.01853144e-01 -3.20940502e-02 1.68898299e-01
-1.70590198e+00 -4.43985090e-02 -3.49828571e-01 3.57631654e-01
-5.19442782e-02 1.44301921e-01 -7.77201891e-01 3.88352014e-02
5.18065572e-01 9.05848742e-01 -5.97595453e-01 1.11570084e+00
2.21290030e-02 5.58015943e-01 -5.62688053e-01 1.40687615e-01
-3.79947543e-01 8.10437948e-02 5.05810022e-01 1.32643223e+00
3.99499029e-01 -2.11693823e-01 1.87135890e-01 5.30914426e-01
-2.01645613e-01 -9.64949094e-03 -5.39415240e-01 3.93766612e-01
2.74664462e-01 1.28936601e+00 -5.53190172e-01 9.69498530e-02
-8.65126252e-01 7.66944885e-01 5.26557684e-01 1.46046087e-01
-9.72341359e-01 1.23902224e-01 5.18570900e-01 2.33510956e-01
8.37919116e-01 -2.41351664e-01 7.74061121e-03 -8.80938768e-01
6.03401996e-02 -8.76975656e-01 3.02269757e-01 -3.55032116e-01
-1.16781616e+00 5.31059504e-01 6.27693608e-02 -1.19028306e+00
-5.08487999e-01 -8.67630959e-01 -4.15171981e-01 8.40611041e-01
-1.45435584e+00 -1.35233808e+00 -1.08642898e-01 7.16304839e-01
5.61217964e-01 -1.84063748e-01 8.47490191e-01 2.58291453e-01
-5.02448916e-01 8.97619069e-01 1.11731090e-01 1.13624223e-02
8.98218572e-01 -1.24543154e+00 5.41745126e-01 6.18492365e-01
3.01356703e-01 7.48971224e-01 6.99765980e-01 -7.37920880e-01
-1.82002831e+00 -1.01757276e+00 4.84111190e-01 -7.73047268e-01
7.81468570e-01 -5.28608024e-01 -8.89513075e-01 9.78808224e-01
-1.16608515e-01 7.54125297e-01 2.58185089e-01 2.22692788e-01
-4.91684437e-01 -2.71905839e-01 -1.22418404e+00 3.85744542e-01
1.22584856e+00 -3.54793906e-01 -3.70139062e-01 3.34248185e-01
4.85464573e-01 -6.99690819e-01 -9.85669494e-01 4.21044230e-01
1.03951502e+00 -5.77049255e-01 1.33122957e+00 -3.28683227e-01
-3.22985739e-01 -3.71442258e-01 2.57003367e-01 -9.74935651e-01
-2.44494557e-01 -1.14721203e+00 -8.71096790e-01 1.25985003e+00
2.74436593e-01 -6.55914307e-01 9.41716254e-01 3.29922974e-01
2.51740605e-01 -1.03248906e+00 -1.26743841e+00 -8.65237057e-01
2.86799520e-01 -4.96182263e-01 3.64618987e-01 6.39109552e-01
-3.13097894e-01 -5.88104725e-02 -5.74429333e-01 1.16510451e-01
9.36030567e-01 -1.87610257e-02 8.83305311e-01 -1.49366009e+00
-3.16104949e-01 -4.25714850e-01 -4.22049940e-01 -1.36231470e+00
4.84388679e-01 -7.09767938e-01 7.82198533e-02 -8.80523801e-01
-9.45989192e-02 -4.38104123e-01 2.85741948e-02 5.75632632e-01
-1.04775324e-01 7.72084534e-01 7.48173967e-02 2.28946447e-01
-4.28871095e-01 1.85339943e-01 9.78699565e-01 3.09597794e-02
-3.93763393e-01 3.08970332e-01 -5.97131610e-01 6.32252991e-01
3.30609620e-01 -5.61776638e-01 -4.22302410e-02 1.52474353e-02
-9.38594639e-02 1.10738851e-01 4.38970953e-01 -7.13741362e-01
6.05347216e-01 1.31582111e-01 7.81843126e-01 -5.10848522e-01
5.80668390e-01 -9.78881657e-01 1.64563924e-01 5.05107939e-01
5.54509945e-02 -3.34127247e-02 5.53774774e-01 3.12992752e-01
3.20126474e-01 -2.94219889e-03 9.46441650e-01 3.65547270e-01
-3.79509598e-01 5.85376203e-01 1.63220584e-01 -1.81616351e-01
1.20336449e+00 -4.40958053e-01 -3.58361229e-02 -1.36636868e-01
-7.40003884e-01 2.05448754e-02 3.05162996e-01 4.18815374e-01
6.63482547e-01 -1.45200062e+00 -1.03867185e+00 3.75188559e-01
-4.85315114e-01 -3.87371808e-01 -3.01204205e-01 1.04523456e+00
-1.47130713e-01 4.40067828e-01 2.80764014e-01 -9.32083786e-01
-1.84043777e+00 7.60110915e-01 5.72204709e-01 -3.60514931e-02
-5.77248573e-01 8.53988707e-01 -2.05385670e-01 -3.46483678e-01
3.40103686e-01 4.27122638e-02 -2.12017879e-01 6.95361868e-02
6.11783922e-01 3.53402168e-01 -1.16824347e-03 -1.09093130e+00
-5.78761399e-01 1.25698209e+00 8.34440142e-02 -8.70962590e-02
1.37373686e+00 1.88663274e-01 -2.84714013e-01 2.09462762e-01
1.32184970e+00 2.76472420e-01 -1.81456947e+00 -2.33776122e-01
-5.77926636e-04 -5.53548038e-01 6.81588128e-02 -3.86711657e-01
-1.24395120e+00 5.32562494e-01 9.16223824e-01 4.08650748e-02
9.57536876e-01 -4.13562171e-03 3.53823215e-01 1.48617387e-01
4.02606726e-01 -5.57216227e-01 9.44976211e-02 2.88148046e-01
1.12967217e+00 -1.19736743e+00 4.79097962e-01 -4.71315354e-01
-1.40779108e-01 1.21168351e+00 6.14675641e-01 -4.92951781e-01
8.76443505e-01 8.13988030e-01 -1.64258465e-01 5.19475937e-02
-7.92423368e-01 -1.05642594e-01 7.21901596e-01 7.87696242e-01
7.81284332e-01 -3.09788346e-01 -3.04814689e-02 2.92868968e-02
-8.26606750e-02 7.70994946e-02 -2.59520829e-01 7.68271446e-01
-2.13901684e-01 -1.27469075e+00 -9.33030486e-01 1.99675739e-01
-4.71703738e-01 1.79814324e-01 -1.97269708e-01 9.27836478e-01
4.43943404e-02 8.66196632e-01 9.10213888e-02 1.47152290e-01
3.35080713e-01 2.66682446e-01 1.07946706e+00 -3.31516445e-01
-6.86235607e-01 3.15058768e-01 -3.62876981e-01 -5.88581085e-01
-3.86458188e-01 -1.14112329e+00 -1.09481061e+00 -3.07850659e-01
-6.85386419e-01 -1.26444325e-01 8.35111916e-01 7.84717500e-01
3.02502774e-02 3.32226574e-01 8.33044529e-01 -1.15341473e+00
-5.77896118e-01 -8.85897517e-01 -1.22750625e-01 1.86796024e-01
7.96391547e-01 -9.11104500e-01 -3.86704504e-01 2.12641522e-01] | [13.459795951843262, 0.18885652720928192] |
3b0f7bc0-019e-4258-9dba-c73f083e167a | stan-stage-adaptive-network-for-multi-task | 2306.12232 | null | https://arxiv.org/abs/2306.12232v1 | https://arxiv.org/pdf/2306.12232v1.pdf | STAN: Stage-Adaptive Network for Multi-Task Recommendation by Learning User Lifecycle-Based Representation | Recommendation systems play a vital role in many online platforms, with their primary objective being to satisfy and retain users. As directly optimizing user retention is challenging, multiple evaluation metrics are often employed. Existing methods generally formulate the optimization of these evaluation metrics as a multitask learning problem, but often overlook the fact that user preferences for different tasks are personalized and change over time. Identifying and tracking the evolution of user preferences can lead to better user retention. To address this issue, we introduce the concept of "user lifecycle", consisting of multiple stages characterized by users' varying preferences for different tasks. We propose a novel Stage-Adaptive Network (STAN) framework for modeling user lifecycle stages. STAN first identifies latent user lifecycle stages based on learned user preferences, and then employs the stage representation to enhance multi-task learning performance. Our experimental results using both public and industrial datasets demonstrate that the proposed model significantly improves multi-task prediction performance compared to state-of-the-art methods, highlighting the importance of considering user lifecycle stages in recommendation systems. Furthermore, online A/B testing reveals that our model outperforms the existing model, achieving a significant improvement of 3.05% in staytime per user and 0.88% in CVR. These results indicate that our approach effectively improves the overall efficiency of the multi-task recommendation system. | ['Suhang Wang', 'Xuanji Xiao', 'Wenhao Zheng', 'Wanda Li'] | 2023-06-21 | null | null | null | null | ['multi-task-learning'] | ['methodology'] | [ 1.53179482e-01 -6.94157720e-01 -7.12027192e-01 -4.95781928e-01
-4.87891704e-01 -4.09254253e-01 1.77007914e-01 1.73852280e-01
-3.94859999e-01 3.53210419e-01 2.63151079e-01 -4.23951060e-01
-6.65981710e-01 -3.48164737e-01 -2.18112901e-01 -4.62854117e-01
6.49579838e-02 3.84142667e-01 9.28983688e-02 -2.96321929e-01
5.28706312e-01 -1.12769477e-01 -1.79449880e+00 5.26079595e-01
1.32132578e+00 1.15093327e+00 5.11633098e-01 3.56105775e-01
3.81766632e-02 4.76836860e-01 -2.29636818e-01 -3.32539558e-01
-3.75943743e-02 1.37176499e-01 -6.83782160e-01 -1.43620539e-02
1.47371933e-01 -6.51859939e-02 4.33701724e-02 6.07732594e-01
5.21128356e-01 5.26982725e-01 5.28624356e-01 -1.08836377e+00
-7.24739790e-01 5.44387460e-01 -7.10814059e-01 2.76911587e-01
1.88313454e-01 -6.29341185e-01 1.35584986e+00 -7.97727644e-01
-9.42083523e-02 9.45687771e-01 5.87654591e-01 3.58897924e-01
-1.24175501e+00 -9.04419243e-01 5.95620036e-01 2.65908599e-01
-1.07969439e+00 -3.97807747e-01 5.72867036e-01 -7.71865726e-01
5.78194976e-01 3.59644502e-01 1.52513146e-01 9.73189294e-01
3.38290453e-01 1.07947075e+00 8.18550110e-01 -3.24180275e-01
-8.74090791e-02 2.90117979e-01 8.60265255e-01 3.39634448e-01
2.02744991e-01 -2.67255545e-01 -6.09334111e-01 -1.51474819e-01
2.89851934e-01 8.01812232e-01 1.04531780e-01 -3.34064752e-01
-7.68306971e-01 7.42949665e-01 -6.80301785e-02 3.20366919e-01
-3.06707770e-01 -5.08789539e-01 2.90044099e-01 3.89257789e-01
6.15441382e-01 3.50497842e-01 -6.34069443e-01 -1.42125085e-01
-8.78902555e-01 1.65865839e-01 5.24874806e-01 9.34539855e-01
6.13375068e-01 -4.03678566e-01 -7.40171850e-01 1.28335166e+00
4.40162271e-01 1.73152983e-01 8.88562202e-01 -6.09005868e-01
2.64218569e-01 8.22418571e-01 3.36206436e-01 -9.42285299e-01
-5.37753344e-01 -9.43401098e-01 -7.41392076e-01 -4.94700372e-01
1.92639828e-01 -5.96838482e-02 -5.77817321e-01 1.48163939e+00
-4.73122001e-02 7.75918886e-02 -2.02603444e-01 4.37972426e-01
4.57310706e-01 5.24409473e-01 3.48207653e-01 -6.06770992e-01
1.39813221e+00 -1.09934509e+00 -7.63142765e-01 -3.37447852e-01
5.22055805e-01 -8.46974611e-01 1.30943859e+00 7.12462902e-01
-8.37143719e-01 -8.96568358e-01 -6.94149554e-01 2.51897365e-01
8.07108656e-02 3.75799179e-01 8.05229366e-01 9.09141719e-01
-6.58054352e-01 6.06548667e-01 -5.23420095e-01 -1.87136307e-01
2.04121709e-01 5.26770890e-01 3.46786439e-01 -9.37972367e-02
-1.09782040e+00 4.15980637e-01 1.06448650e-01 -9.85213816e-02
-5.18804669e-01 -9.14743066e-01 -3.81761581e-01 2.07387522e-01
4.12307054e-01 -4.52142984e-01 1.54471350e+00 -7.21527457e-01
-1.24992812e+00 2.96749741e-01 -4.50400144e-01 -4.72223796e-02
1.71630070e-01 -7.41888523e-01 -7.31039286e-01 -6.93930864e-01
-3.89737561e-02 -3.07562739e-01 7.81484365e-01 -1.17368329e+00
-1.39427459e+00 -5.76486647e-01 -6.93817288e-02 2.85906792e-01
-1.19589591e+00 1.37034625e-01 -9.35164273e-01 -4.23442990e-01
-1.07246354e-01 -9.24578428e-01 -2.33685941e-01 -9.07342136e-01
1.84598677e-02 -6.27767205e-01 6.95775568e-01 -6.12242222e-01
2.16543055e+00 -1.96850955e+00 5.17576262e-02 2.68568337e-01
2.78900027e-01 2.32381389e-01 -8.47315008e-04 1.31931737e-01
2.76623607e-01 1.88637450e-01 3.46503645e-01 -6.82773173e-01
-2.66912188e-02 -5.58572300e-02 -9.27646831e-02 3.45414653e-02
-4.90014225e-01 6.42937899e-01 -7.90888667e-01 -2.98765838e-01
-8.77555311e-02 2.62888223e-01 -7.03165412e-01 4.20844465e-01
8.81600901e-02 5.22126555e-01 -7.54118264e-01 7.22143173e-01
3.27634007e-01 -6.22054577e-01 5.70905149e-01 -5.80906123e-02
-2.44792193e-01 1.05291538e-01 -1.00065577e+00 1.52756560e+00
-7.62371540e-01 1.05286904e-01 -2.39869684e-01 -9.53718364e-01
1.00107384e+00 1.90258086e-01 9.60209906e-01 -1.09917891e+00
2.14640629e-02 5.77496476e-02 -3.48113067e-02 -5.37878036e-01
9.15179253e-01 2.61648834e-01 -2.10793391e-01 5.73936462e-01
-2.86126941e-01 8.76077473e-01 1.72654197e-01 1.83878653e-02
8.84449780e-01 7.79185891e-02 1.35816559e-01 -2.82647222e-01
4.80737031e-01 -4.53519940e-01 8.29574466e-01 8.76699448e-01
-1.72615528e-01 -2.97257863e-03 1.98108256e-01 -2.60670096e-01
-5.98267198e-01 -4.90031242e-01 -1.11891799e-01 2.17682672e+00
1.87958255e-01 -4.91284579e-01 -2.06134915e-01 -6.19160950e-01
1.98621675e-01 6.36624157e-01 -5.46458542e-01 -2.30995789e-01
-4.34679061e-01 -1.02934325e+00 -7.94699714e-02 3.67187738e-01
1.07092932e-02 -9.64497745e-01 -2.24077150e-01 3.86236459e-01
-4.33682382e-01 -9.47807252e-01 -7.26959586e-01 2.46322956e-02
-1.17536628e+00 -8.14805746e-01 -6.46609008e-01 -7.69305348e-01
4.73917812e-01 8.30076396e-01 1.11162245e+00 7.13244006e-02
2.10033506e-01 2.91657656e-01 -6.78929925e-01 -1.58778310e-01
6.49104491e-02 6.54599071e-01 3.69770378e-01 4.04632866e-01
4.20399159e-01 -5.82827389e-01 -8.15820098e-01 8.57518554e-01
-7.40071774e-01 5.84979989e-02 7.21769869e-01 7.01868534e-01
5.55949211e-01 3.19486797e-01 9.23042834e-01 -1.42842853e+00
1.05493188e+00 -8.30984235e-01 -9.87318382e-02 7.48384535e-01
-1.63516283e+00 -1.10357784e-01 5.72929323e-01 -6.90305233e-01
-1.33002675e+00 -1.66069463e-01 1.35878369e-01 -1.47588119e-01
2.22425073e-01 7.97193587e-01 6.72081411e-02 2.74834812e-01
4.31475222e-01 7.91798234e-02 -3.93044859e-01 -8.14956188e-01
2.05970407e-02 1.03016472e+00 2.11283237e-01 -8.49512279e-01
2.63685882e-01 -2.33082902e-02 -4.06628788e-01 -5.58218956e-01
-1.32015193e+00 -1.29357910e+00 -5.48746467e-01 -3.70010495e-01
3.68938535e-01 -9.87784684e-01 -1.05660391e+00 2.64751494e-01
-6.02548420e-01 -3.11618418e-01 3.54715377e-01 3.39721859e-01
-1.95177406e-01 2.82675624e-01 -5.22817194e-01 -1.07613289e+00
-6.06499732e-01 -1.15804386e+00 7.34647453e-01 4.83708203e-01
-1.20910242e-01 -1.06948459e+00 2.03888081e-02 7.74926543e-01
5.87322235e-01 -3.86154234e-01 1.02324128e+00 -6.68624759e-01
-3.74475420e-02 -3.20995092e-01 -1.65136158e-01 1.21723890e-01
3.52972686e-01 -3.87939513e-01 -7.96395779e-01 -6.87599301e-01
-2.58181959e-01 -1.40157744e-01 7.94801950e-01 4.98287588e-01
1.61200321e+00 -1.20260142e-01 -5.78001320e-01 2.35666767e-01
1.13153517e+00 4.83150631e-01 4.80084211e-01 4.17860806e-01
7.39616454e-01 7.54029810e-01 1.03749371e+00 7.85865605e-01
5.66239357e-01 9.10898209e-01 2.21747696e-01 7.11957291e-02
4.80089635e-01 -1.15485989e-01 5.48569202e-01 1.01602530e+00
-3.35468471e-01 -2.89989114e-01 -7.43939161e-01 3.51084083e-01
-2.45877934e+00 -7.56117105e-01 -1.87041178e-01 2.46680260e+00
6.83686614e-01 2.50093222e-01 4.15088713e-01 5.50642386e-02
6.25455678e-01 -9.84339416e-02 -5.66543102e-01 -9.40398499e-02
4.53177840e-01 1.81803599e-01 3.76652002e-01 2.78703421e-01
-1.20963943e+00 7.73918331e-01 5.53970337e+00 7.57458329e-01
-8.67433488e-01 3.45233887e-01 6.11382902e-01 -1.76588327e-01
-1.09021172e-01 -2.39706233e-01 -1.31354761e+00 5.65389514e-01
9.82255161e-01 -3.96099478e-01 3.24256212e-01 7.37589538e-01
5.95312715e-01 2.73298323e-01 -9.08946991e-01 8.89879465e-01
1.44472450e-01 -7.47693658e-01 -1.16299339e-01 2.81924665e-01
8.29452038e-01 -3.95099103e-01 3.86798710e-01 9.48637784e-01
2.40965873e-01 -6.84631109e-01 4.49421912e-01 8.27374578e-01
4.08637553e-01 -8.65421832e-01 4.89364624e-01 5.73881328e-01
-1.34188879e+00 -7.25185394e-01 -3.69714737e-01 -4.12603281e-02
1.47844076e-01 6.47769392e-01 -5.16355693e-01 6.32324219e-01
8.59482348e-01 9.30927634e-01 -6.87119246e-01 1.17597294e+00
3.60417545e-01 8.26484025e-01 3.94915074e-01 6.25560731e-02
-2.12780938e-01 -3.44872355e-01 -6.33126050e-02 1.24414980e+00
6.41651750e-01 -9.59456116e-02 5.09999692e-01 8.54017511e-02
-8.56374130e-02 4.74899411e-01 -1.05102569e-01 -2.85317618e-02
3.39377493e-01 1.52638185e+00 -3.45435292e-01 4.89787050e-02
-5.22678733e-01 8.29129755e-01 2.12429270e-01 3.96553218e-01
-7.22255051e-01 -1.37184694e-01 7.07221270e-01 2.44605735e-01
1.96276039e-01 -1.29064068e-01 -3.46118689e-01 -9.27295148e-01
-1.75062716e-01 -8.47426057e-01 5.68891585e-01 -2.89449915e-02
-1.15597689e+00 4.48256135e-01 -3.34785759e-01 -1.36032045e+00
8.94477963e-02 -4.21823561e-01 -4.27663654e-01 7.19501019e-01
-1.47452390e+00 -1.23621094e+00 -4.17258173e-01 4.88504410e-01
9.24025774e-01 -3.62460434e-01 6.19680941e-01 9.84369397e-01
-1.03154027e+00 9.19255674e-01 4.51212257e-01 -6.06308162e-01
1.06690240e+00 -1.01349103e+00 -6.00825995e-02 6.29338801e-01
-1.26188785e-01 9.63008344e-01 4.34469461e-01 -5.66692173e-01
-1.30274999e+00 -1.29100406e+00 1.10387182e+00 -5.05841613e-01
4.28355783e-01 -1.29294008e-01 -8.32816482e-01 5.82882524e-01
-1.28657341e-01 -5.77135384e-01 1.23031473e+00 1.12009358e+00
-1.13789424e-01 -3.91179651e-01 -5.09582639e-01 3.89645845e-01
1.03482878e+00 -2.49489486e-01 -1.06940925e-01 2.06959367e-01
6.05426371e-01 -2.89788153e-02 -1.13783348e+00 4.57956046e-01
1.06637669e+00 -7.44431317e-01 9.79535699e-01 -8.51005971e-01
3.11442345e-01 7.50770494e-02 -1.09463014e-01 -1.16995192e+00
-1.02248573e+00 -5.41237652e-01 -6.25404418e-01 1.29933476e+00
4.86802816e-01 -2.55878061e-01 8.03278744e-01 8.47718716e-01
-2.20532030e-01 -9.26522315e-01 -2.08571658e-01 -4.99813497e-01
-3.28685552e-01 -3.77243429e-01 3.92712802e-01 9.71335530e-01
9.52924136e-03 7.27909386e-01 -9.63530362e-01 2.34800298e-02
5.40351748e-01 3.75631630e-01 6.37545228e-01 -2.02516484e+00
-3.28552246e-01 -6.14971876e-01 5.27404070e-01 -1.36061716e+00
1.52846083e-01 -8.86136234e-01 -5.05076461e-02 -1.44831014e+00
5.91032684e-01 -8.91861320e-01 -1.16788173e+00 4.00284618e-01
-4.80332822e-01 -1.36092782e-01 -1.93107665e-01 5.66314936e-01
-1.17470717e+00 3.29938680e-01 1.13872600e+00 2.24624667e-02
-6.91958070e-01 9.45470572e-01 -1.31243503e+00 4.94797677e-01
7.45441735e-01 -5.21215022e-01 -5.96626878e-01 -3.08987081e-01
4.11264420e-01 2.45310068e-02 -3.76963496e-01 -8.01642299e-01
3.88573676e-01 -4.59154129e-01 1.92333311e-01 -4.93159920e-01
9.35268998e-02 -7.46753812e-01 9.13876295e-02 2.40655944e-01
-4.99510586e-01 1.84054732e-01 3.25382836e-02 7.81803727e-01
1.50268286e-01 -1.49585649e-01 3.88810515e-01 3.68300021e-01
-8.37561488e-01 6.18243814e-01 -1.59725785e-01 -4.15199995e-01
7.31535435e-01 -2.67134160e-02 -9.03821811e-02 -2.21329823e-01
-7.64260292e-01 5.89157581e-01 -7.25149289e-02 9.69605625e-01
4.06209558e-01 -1.23528290e+00 -4.69954252e-01 4.81062271e-02
4.75123882e-01 -4.37882274e-01 5.20391762e-01 8.63474011e-01
4.40179467e-01 5.10622382e-01 -2.55334582e-02 -4.89831269e-01
-1.68188655e+00 2.79326081e-01 3.84214222e-02 -8.19554508e-01
-5.18567972e-02 5.54653406e-01 1.83095470e-01 -3.66487563e-01
5.80740452e-01 -8.39361362e-03 -9.90034044e-01 5.35964847e-01
4.24673885e-01 6.86509967e-01 1.43051639e-01 -3.93118501e-01
-3.24640907e-02 2.69764960e-01 -6.36737049e-01 3.88858318e-01
1.38928676e+00 -3.82116169e-01 1.13323465e-01 6.22294128e-01
7.77684033e-01 7.50601739e-02 -9.03973103e-01 -7.20176280e-01
4.16912526e-01 -4.65873897e-01 1.92373887e-01 -7.07596779e-01
-9.06935930e-01 5.94319761e-01 9.17693079e-01 3.36673111e-01
1.16005945e+00 -3.02917242e-01 9.02983189e-01 3.96279454e-01
3.18130612e-01 -1.25470245e+00 4.63737220e-01 7.07575738e-01
4.48424548e-01 -1.33512914e+00 -1.14257142e-01 -1.21508747e-01
-8.43785524e-01 7.22835839e-01 8.64328802e-01 5.34303010e-01
8.43420088e-01 -4.27725554e-01 -1.19732074e-01 5.94393946e-02
-7.29073763e-01 -1.05141886e-01 7.47042298e-01 1.03960233e-02
9.47520375e-01 3.24240208e-01 -4.45355117e-01 1.29409492e+00
2.44987711e-01 1.16493568e-01 -5.74216619e-02 8.50322723e-01
-6.82573557e-01 -1.74633646e+00 -6.88275769e-02 9.41422045e-01
-6.71393335e-01 3.55237573e-02 1.75644070e-01 1.81914374e-01
1.14823408e-01 1.21454155e+00 -2.00530738e-01 -9.21168447e-01
6.87815964e-01 1.46326914e-01 1.34826675e-01 -8.43499541e-01
-7.43407965e-01 3.20220590e-01 -1.32002339e-01 -3.51572067e-01
-4.09297734e-01 -7.07982957e-01 -7.32116699e-01 -2.95498431e-01
-4.92540300e-01 3.32226187e-01 5.19869328e-01 9.14568543e-01
5.74510098e-01 9.32967544e-01 1.03947353e+00 -5.56705832e-01
-6.10039949e-01 -1.07961738e+00 -6.65435076e-01 4.70624834e-01
9.89998803e-02 -9.60333824e-01 8.92540663e-02 3.33198085e-02] | [10.13988971710205, 5.610818386077881] |
da6863d2-169a-4f41-80d9-f23c8a5d97f4 | minimax-risk-classifiers-with-0-1-loss | 2201.06487 | null | https://arxiv.org/abs/2201.06487v5 | https://arxiv.org/pdf/2201.06487v5.pdf | Minimax risk classifiers with 0-1 loss | Supervised classification techniques use training samples to learn a classification rule with small expected 0-1 loss (error probability). Conventional methods enable tractable learning and provide out-of-sample generalization by using surrogate losses instead of the 0-1 loss and considering specific families of rules (hypothesis classes). This paper presents minimax risk classifiers (MRCs) that minize the worst-case 0-1 loss with respect to uncertainty sets of distributions that can include the underlying distribution, with a tunable confidence. We show that MRCs can provide tight performance guarantees at learning and are strongly universally consistent using feature mappings given by characteristic kernels. The paper also proposes efficient optimization techniques for MRC learning and shows that the methods presented can provide accurate classification together with tight performance guarantees in practice. | ['Peter Grünwald', 'Mauricio Romero', 'Santiago Mazuelas'] | 2022-01-17 | null | null | null | null | ['classification'] | ['methodology'] | [ 9.73262936e-02 5.31516790e-01 -7.52759457e-01 -7.73183346e-01
-1.35366738e+00 -5.22012413e-01 2.51749784e-01 4.78569478e-01
-5.09931386e-01 1.17327845e+00 -7.02859104e-01 -4.59397107e-01
-7.21741140e-01 -9.01147902e-01 -1.01927924e+00 -7.19413042e-01
-5.20658731e-01 4.57893968e-01 8.49943012e-02 3.74955326e-01
6.35991991e-02 5.98633885e-01 -1.54619670e+00 2.55568355e-01
1.02195680e+00 1.50751960e+00 -5.74188352e-01 7.32425928e-01
2.88686693e-01 5.81093013e-01 -7.33544230e-01 -6.14135027e-01
4.41822946e-01 -1.05674518e-02 -4.55974132e-01 -3.07134569e-01
8.49793792e-01 -1.48993969e-01 -2.10719300e-03 1.08940578e+00
1.64098546e-01 4.05085266e-01 1.18125057e+00 -1.48757124e+00
-7.06634939e-01 4.05814528e-01 -1.92579508e-01 -3.49288173e-02
1.41660556e-01 -3.34029794e-01 1.09812582e+00 -6.86865330e-01
-2.78175529e-02 1.02288675e+00 1.14365375e+00 6.58606768e-01
-1.42690837e+00 -7.84762979e-01 7.19353259e-02 -1.96418405e-01
-1.45959616e+00 -2.10225165e-01 1.55818537e-01 -3.16871434e-01
8.78964126e-01 5.11802614e-01 9.11135301e-02 9.00172949e-01
4.36610818e-01 7.80124784e-01 9.99843359e-01 -5.41561842e-01
5.90162933e-01 7.28253782e-01 3.46076697e-01 7.96722293e-01
5.27749181e-01 4.55814451e-01 -2.37131238e-01 -5.88156879e-01
3.86428684e-01 -9.24719870e-02 -2.21645162e-01 -6.21357024e-01
-5.42074800e-01 1.16110134e+00 2.98128933e-01 -3.38832021e-01
1.33487254e-01 2.39410132e-01 3.18835914e-01 5.71314156e-01
4.65592235e-01 2.58657783e-01 -6.27130330e-01 2.74770707e-01
-8.51653636e-01 3.16077799e-01 1.11026621e+00 1.21165526e+00
4.06279385e-01 -8.38015154e-02 -2.72486299e-01 8.15779150e-01
1.08309746e-01 6.59867167e-01 1.23664804e-01 -7.99072266e-01
5.26883781e-01 -3.21910903e-02 3.39898765e-01 -4.79162842e-01
-1.90042809e-01 -8.79416108e-01 -5.01989424e-01 6.31778419e-01
4.86278147e-01 -1.32183626e-01 -6.86705768e-01 1.70747530e+00
2.92939782e-01 -1.53822091e-03 2.13819325e-01 3.81370902e-01
-8.76543894e-02 4.14840728e-01 1.38432264e-01 -4.59197700e-01
7.15543807e-01 -4.00430501e-01 -3.61047119e-01 2.80685484e-01
6.26790404e-01 -3.74529690e-01 1.11918497e+00 7.55545497e-01
-8.49848151e-01 -2.60801286e-01 -1.30530500e+00 3.05457652e-01
-3.08981121e-01 -6.16445243e-02 5.07357299e-01 1.22261608e+00
-5.70241332e-01 9.27186430e-01 -6.61848426e-01 2.48573110e-01
8.24399948e-01 5.36490560e-01 -3.65184128e-01 1.77349404e-01
-9.44496334e-01 9.85942304e-01 4.15742457e-01 -1.64088458e-01
-8.52247357e-01 -1.19037700e+00 -7.36937642e-01 -1.29661664e-01
2.27737576e-01 -2.96647489e-01 1.29162681e+00 -9.59577978e-01
-1.33592570e+00 6.05977833e-01 2.60375917e-01 -9.48879361e-01
8.70182574e-01 -3.50965858e-01 -3.80874336e-01 9.06737223e-02
-6.98477551e-02 1.89479128e-01 9.98319566e-01 -9.69432473e-01
-9.68522310e-01 -1.38183042e-01 -6.94221556e-02 -2.20996007e-01
-4.56403255e-01 -3.35465521e-01 6.33137584e-01 -7.15337396e-01
-3.80968362e-01 -4.59634960e-01 -2.98483646e-03 5.63168406e-01
-3.51113528e-01 -3.32392544e-01 6.86957300e-01 -3.52393448e-01
1.14004755e+00 -1.95330298e+00 -4.49574709e-01 5.97657621e-01
-2.45167255e-01 1.46578461e-01 2.29308888e-01 2.55224332e-02
5.23794889e-02 1.87491804e-01 -5.85591972e-01 -1.71178266e-01
3.95303875e-01 2.82509446e-01 -6.82000816e-01 7.08542883e-01
1.67448267e-01 4.68604743e-01 -8.13538074e-01 -2.86584139e-01
8.94621536e-02 3.72774571e-01 -4.63132918e-01 1.05629832e-01
-6.46982417e-02 -4.13491368e-01 -3.50948423e-01 4.75660354e-01
6.41373754e-01 -4.20886874e-02 -1.54542208e-01 7.39761591e-02
4.76663470e-01 -1.17130630e-01 -1.14884627e+00 7.75345862e-01
-6.81604743e-01 2.28648648e-01 -3.04013137e-02 -1.37665176e+00
1.07573557e+00 9.13902521e-02 1.25662163e-01 1.99108899e-01
-1.00833468e-01 4.94909793e-01 -4.08420712e-01 3.85536589e-02
-3.61002058e-01 -7.21322179e-01 -1.64626375e-01 2.74576992e-01
2.97784775e-01 4.54753172e-03 -4.22458500e-01 -2.95025229e-01
8.36862624e-01 -2.20338225e-01 4.60156739e-01 -6.24839962e-01
4.13799286e-01 -3.53629947e-01 4.90949661e-01 1.14719951e+00
-1.93027020e-01 4.10952389e-01 4.67061788e-01 -2.84308612e-01
-9.02405381e-01 -1.66925204e+00 -9.99157667e-01 6.82115257e-01
-1.20163284e-01 1.65286317e-01 -3.86587471e-01 -1.19220722e+00
7.94978678e-01 1.08727157e+00 -8.32607627e-01 -4.69841748e-01
1.03353607e-02 -8.75919878e-01 7.33296275e-01 7.64875829e-01
2.36868829e-01 -4.24876660e-01 -3.70920092e-01 5.20785861e-02
7.00602829e-01 -7.91396141e-01 -4.39432800e-01 7.27133989e-01
-9.38478172e-01 -1.28419578e+00 -4.80959207e-01 -7.45156348e-01
7.21109807e-01 -3.79050195e-01 8.28543246e-01 -3.92405331e-01
-4.10294563e-01 5.56319654e-01 -3.45699154e-02 -4.54228580e-01
-3.56334150e-01 -1.64848998e-01 4.06174153e-01 -2.25618538e-02
3.28434139e-01 -4.49202091e-01 -2.42046461e-01 3.14277261e-01
-5.74905515e-01 -7.62501717e-01 2.75288820e-01 1.13114953e+00
8.24245453e-01 3.99266988e-01 1.07295251e+00 -1.09884810e+00
6.43882275e-01 -6.14629447e-01 -9.91605401e-01 7.99946547e-01
-1.12847543e+00 2.66286224e-01 9.79828894e-01 -5.24019897e-01
-1.05243647e+00 -1.15968853e-01 2.82169789e-01 -6.03118598e-01
-4.11089472e-02 -2.32999050e-03 -1.58426389e-01 -4.25828069e-01
1.03891313e+00 -2.39970777e-02 7.03359768e-02 -5.19858897e-01
1.99323326e-01 9.42953944e-01 5.34886956e-01 -1.21030438e+00
7.09919810e-01 3.78873408e-01 4.46613073e-01 -5.50318837e-01
-1.48914587e+00 -1.31973594e-01 -4.58815426e-01 1.33417279e-01
2.76827693e-01 -6.45456076e-01 -7.72818089e-01 -5.20108454e-02
-4.43042696e-01 -3.58674616e-01 -7.50330567e-01 7.10433662e-01
-9.77249563e-01 2.02650934e-01 -5.04373431e-01 -1.38002455e+00
-3.57442915e-01 -3.93579602e-01 8.51092875e-01 7.49480650e-02
-6.78126588e-02 -1.27186430e+00 -2.25418285e-01 -4.92660068e-02
2.51518607e-01 4.95525002e-01 1.01553071e+00 -8.97072852e-01
-1.41148463e-01 -6.99447215e-01 -7.93717118e-05 1.08713698e+00
1.65762514e-01 5.16250823e-03 -1.15724158e+00 -4.65501666e-01
-6.15729168e-02 -6.36534929e-01 9.48146105e-01 4.42039996e-01
1.64413929e+00 -7.77130246e-01 -1.65810108e-01 8.26334059e-01
1.37067199e+00 6.23963848e-02 2.55208254e-01 1.39933720e-01
1.19603887e-01 5.59198320e-01 7.80100644e-01 4.35864747e-01
-2.33076945e-01 3.39736879e-01 -2.64973957e-02 5.81856072e-01
2.79406279e-01 -2.92870760e-01 4.34873790e-01 5.46568423e-04
4.46794480e-01 8.25956911e-02 -5.04798591e-01 2.50206500e-01
-1.83892941e+00 -9.76137638e-01 2.88563401e-01 2.93313241e+00
1.16175807e+00 5.25171041e-01 1.61181748e-01 3.49020928e-01
7.37586200e-01 -4.50963497e-01 -8.25460672e-01 -6.24871731e-01
-1.22440100e-01 6.08144581e-01 7.91746676e-01 1.04626083e+00
-1.32510829e+00 2.64310032e-01 7.61948633e+00 1.30995035e+00
-7.69801199e-01 -2.85917446e-02 8.21383655e-01 -3.41494858e-01
-2.35599622e-01 -2.17518121e-01 -1.09065485e+00 4.68181372e-01
1.08309853e+00 -2.95854598e-01 2.51033425e-01 1.19531441e+00
-4.37296987e-01 2.08772033e-01 -1.41537094e+00 6.94396496e-01
-2.70259172e-01 -1.12691081e+00 -3.52774486e-02 -1.22357123e-02
8.29469621e-01 -3.41643929e-01 3.82922709e-01 4.26093102e-01
3.05426151e-01 -1.23954213e+00 6.93765044e-01 7.86737144e-01
1.26761699e+00 -1.21290791e+00 7.24857569e-01 3.47009122e-01
-7.43993163e-01 -3.87849838e-01 -7.87126422e-01 2.26361677e-01
-3.70440602e-01 9.74148512e-01 -5.59081435e-01 3.26388210e-01
3.41957927e-01 5.39481342e-01 -1.27310351e-01 1.30018556e+00
-1.19470835e-01 8.13981771e-01 -7.56174982e-01 -1.10091023e-01
-3.76179293e-02 4.24557291e-02 4.56481427e-01 1.26577079e+00
2.59167284e-01 -1.89916521e-01 3.19627970e-01 6.42452121e-01
-1.67590469e-01 -1.94620900e-02 -4.30058122e-01 3.65894824e-01
7.66651273e-01 7.39206314e-01 -2.73014873e-01 -9.93924662e-02
-2.09586471e-01 6.20860934e-01 4.90240514e-01 3.53491634e-01
-5.55485189e-01 -9.49768186e-01 7.40138590e-01 5.04442900e-02
4.26881760e-01 3.52699339e-01 -3.51981461e-01 -9.29196358e-01
3.95050764e-01 -5.92485249e-01 9.33515251e-01 -8.61172453e-02
-2.08164811e+00 3.47813033e-02 2.42083162e-01 -1.15761447e+00
-7.70149305e-02 -1.11481667e+00 -4.66274649e-01 7.62023568e-01
-1.36966896e+00 -7.84598589e-01 4.43238497e-01 4.34923440e-01
-3.63547727e-02 -3.44172895e-01 1.06467998e+00 1.29447943e-02
-2.88220525e-01 1.50771296e+00 5.79060733e-01 -9.98824313e-02
4.52983409e-01 -1.68019044e+00 -3.66403043e-01 2.44653434e-01
5.13220020e-02 4.88141149e-01 6.58503592e-01 -3.14904630e-01
-7.82921016e-01 -1.32462931e+00 5.80900788e-01 -5.80100477e-01
6.49413884e-01 -2.05142349e-01 -9.99025226e-01 5.74129283e-01
-6.70371830e-01 7.17100680e-01 9.91031528e-01 3.12709600e-01
-9.58863199e-01 -7.03404844e-01 -1.95998156e+00 3.18376757e-02
8.30832720e-01 -7.96801805e-01 -5.71630239e-01 4.65332687e-01
4.23442096e-01 -1.80495739e-01 -1.23554337e+00 5.96163034e-01
8.90819430e-01 -6.52497470e-01 9.34843600e-01 -1.19003093e+00
-1.42183527e-01 -6.99842442e-03 -4.64955479e-01 -1.03505206e+00
4.56853360e-02 -6.97922766e-01 -5.90400755e-01 8.18082869e-01
6.78936005e-01 -1.00401580e+00 6.79924369e-01 6.49020135e-01
1.13364816e-01 -1.21911168e+00 -1.30288231e+00 -1.47844958e+00
7.19003558e-01 -6.06428266e-01 3.79997700e-01 5.75442612e-01
2.24532336e-01 -3.34964961e-01 -1.21620268e-01 3.07360977e-01
1.35950875e+00 2.05414385e-01 1.50908187e-01 -1.34121382e+00
-5.80953717e-01 -5.09462416e-01 -8.32400858e-01 -6.08094871e-01
4.87111419e-01 -1.09226525e+00 -9.86202620e-03 -7.17191696e-01
5.88778816e-02 -9.09227729e-01 -8.08619916e-01 5.69910586e-01
-2.32010260e-02 -1.10033020e-01 -2.89084136e-01 -1.93781942e-01
-5.05233347e-01 6.45038843e-01 6.61257505e-01 5.49537502e-03
1.99966326e-01 5.42625844e-01 -6.65991724e-01 7.01833129e-01
8.58670712e-01 -6.22946024e-01 -5.99428475e-01 2.68359393e-01
-1.14478938e-01 -7.33411089e-02 4.86099452e-01 -9.73383904e-01
-8.52841213e-02 -2.86796480e-01 6.36068046e-01 -1.43263578e-01
1.11177310e-01 -6.40626013e-01 -2.65188694e-01 5.43402135e-01
-1.11672127e+00 -7.21026421e-01 5.14779724e-02 1.09630728e+00
7.27030486e-02 -5.11526525e-01 1.23819149e+00 3.86106908e-01
5.29847443e-02 2.78983235e-01 4.20064293e-02 3.73659045e-01
1.30474377e+00 -3.25544458e-03 -1.61333472e-01 -3.20413768e-01
-9.74458575e-01 2.97175676e-01 1.07670896e-01 -3.08121890e-02
5.94802499e-01 -1.29275620e+00 -7.30373919e-01 1.56388909e-01
1.73823252e-01 -1.45533174e-01 -3.35846186e-01 2.78226644e-01
-1.91967949e-01 4.14073527e-01 1.70377105e-01 -3.52534741e-01
-8.03149462e-01 4.27824348e-01 9.09073055e-01 6.97864294e-02
-4.88154233e-01 1.18755507e+00 1.16323300e-01 -6.74738944e-01
6.19277000e-01 -2.93981910e-01 4.88120586e-01 -3.28919291e-01
7.13669240e-01 4.73161727e-01 1.67647824e-01 8.64612162e-02
-3.10584724e-01 3.69758308e-01 -1.05246074e-01 -6.71751648e-02
9.99515116e-01 2.56980836e-01 4.10567284e-01 6.46666229e-01
1.53594589e+00 -2.05547124e-01 -1.69029415e+00 -1.76640630e-01
2.50411808e-01 -6.86687291e-01 -7.22014531e-02 -1.02074635e+00
-6.71664953e-01 9.07622874e-01 7.71208704e-01 1.12749621e-01
7.63996005e-01 -1.28682762e-01 3.84180874e-01 8.19690645e-01
6.21301651e-01 -1.26266038e+00 -3.43454391e-01 5.17771952e-02
8.27145517e-01 -1.13481522e+00 -3.98686621e-03 -4.18722421e-01
-2.84374028e-01 1.35949731e+00 3.54297847e-01 -5.41367292e-01
1.21671903e+00 4.08609748e-01 -2.41738454e-01 4.66784745e-01
-9.10376489e-01 1.83644980e-01 6.05746806e-01 1.06088734e+00
1.32759467e-01 4.24271822e-01 -2.27155060e-01 9.70232904e-01
-2.80241966e-01 -2.21825466e-01 5.45175038e-02 7.01902568e-01
-6.80148542e-01 -8.88433695e-01 -1.64209247e-01 1.06400681e+00
-6.29354417e-01 2.44180113e-01 1.45556942e-01 8.20908248e-01
-8.44525993e-02 8.38458359e-01 -5.36326692e-02 -2.33720958e-01
2.47181579e-01 1.38860449e-01 8.06776702e-01 -4.54855382e-01
-5.16607687e-02 -5.56793392e-01 6.26533180e-02 -3.71996045e-01
1.18256405e-01 -6.93907022e-01 -1.24956381e+00 -2.20146969e-01
-4.83851254e-01 4.69806135e-01 2.12677553e-01 8.15002382e-01
-1.23680104e-02 -1.11156367e-01 9.25329626e-01 -8.58369768e-02
-2.00519705e+00 -5.43123126e-01 -9.08597648e-01 1.15959935e-01
6.75901175e-01 -7.38444149e-01 -9.78429794e-01 -3.00059378e-01] | [7.998234272003174, 4.198162078857422] |
571a2f57-d036-4637-b0dc-6dc6b2f52838 | dna-teq-an-adaptive-exponential-quantization | 2306.16430 | null | https://arxiv.org/abs/2306.16430v1 | https://arxiv.org/pdf/2306.16430v1.pdf | DNA-TEQ: An Adaptive Exponential Quantization of Tensors for DNN Inference | Quantization is commonly used in Deep Neural Networks (DNNs) to reduce the storage and computational complexity by decreasing the arithmetical precision of activations and weights, a.k.a. tensors. Efficient hardware architectures employ linear quantization to enable the deployment of recent DNNs onto embedded systems and mobile devices. However, linear uniform quantization cannot usually reduce the numerical precision to less than 8 bits without sacrificing high performance in terms of model accuracy. The performance loss is due to the fact that tensors do not follow uniform distributions. In this paper, we show that a significant amount of tensors fit into an exponential distribution. Then, we propose DNA-TEQ to exponentially quantize DNN tensors with an adaptive scheme that achieves the best trade-off between numerical precision and accuracy loss. The experimental results show that DNA-TEQ provides a much lower quantization bit-width compared to previous proposals, resulting in an average compression ratio of 40% over the linear INT8 baseline, with negligible accuracy loss and without retraining the DNNs. Besides, DNA-TEQ leads the way in performing dot-product operations in the exponential domain, which saves 66% of energy consumption on average for a set of widely used DNNs. | ['Antonio González', 'Marc Riera', 'Bahareh Khabbazan'] | 2023-06-28 | null | null | null | null | ['quantization'] | ['methodology'] | [ 7.12323980e-03 -2.06407338e-01 -3.95148620e-02 -4.62514251e-01
-3.68425965e-01 -3.43703091e-01 1.89491555e-01 2.82691509e-01
-1.14260495e+00 4.83604014e-01 -2.75856525e-01 -5.51545560e-01
4.36538495e-02 -9.70951259e-01 -7.42939711e-01 -7.97830522e-01
-4.00010403e-03 1.50779530e-01 5.19948006e-01 5.47959178e-04
1.76932245e-01 6.29422247e-01 -1.64735734e+00 1.15333781e-01
5.81627369e-01 1.60720348e+00 1.54264539e-01 5.89024127e-01
-1.84923798e-01 7.69837439e-01 -7.80308783e-01 -1.01057208e+00
4.33797121e-01 1.46715567e-01 -4.60460246e-01 -4.76418853e-01
8.06790113e-01 -9.42541301e-01 -8.05989861e-01 1.51521373e+00
4.41714346e-01 -1.24810398e-01 4.24955457e-01 -1.11965823e+00
-4.97461915e-01 8.08159649e-01 -1.84155226e-01 4.22980428e-01
-6.83029890e-01 -5.23218103e-02 9.71974730e-01 -7.86309063e-01
1.82062075e-01 1.12531805e+00 7.35059023e-01 5.82565844e-01
-1.09700453e+00 -7.86036134e-01 -1.44187719e-01 4.76823956e-01
-1.72369802e+00 -5.61120868e-01 3.57881069e-01 3.53992246e-02
1.28652418e+00 1.67950809e-01 7.34449506e-01 4.86118197e-01
1.88009962e-01 5.55042565e-01 3.61702889e-01 -9.88026112e-02
6.64533377e-01 -3.00156802e-01 2.44617775e-01 7.23444939e-01
6.31977379e-01 -2.03077376e-01 -5.35158932e-01 1.56557307e-01
8.33995640e-01 6.47572577e-02 -1.79661214e-01 -2.30228808e-02
-1.09515679e+00 6.45279646e-01 7.37787127e-01 2.31759846e-01
-3.38959128e-01 9.18061137e-01 9.21948433e-01 1.46781951e-01
4.28114496e-02 2.69256663e-02 -4.89786297e-01 -6.32580876e-01
-1.08907282e+00 1.94851384e-01 3.56749058e-01 1.09181559e+00
6.20508909e-01 4.59905326e-01 -3.44914496e-02 7.73838997e-01
2.52250820e-01 6.43876255e-01 6.48280501e-01 -1.10316145e+00
8.11446428e-01 3.68707925e-01 -2.95705199e-01 -8.51344466e-01
-1.17984466e-01 -4.05567855e-01 -1.25536883e+00 1.88991025e-01
5.20981729e-01 -1.15849443e-01 -9.11373377e-01 1.58975148e+00
3.19014713e-02 -1.48630336e-01 -2.56072711e-02 8.83301377e-01
5.49475431e-01 7.41068602e-01 -1.06044382e-01 -8.69142264e-03
1.49606442e+00 -7.74626911e-01 -9.71310914e-01 -4.17841747e-02
8.34356368e-01 -4.88322705e-01 9.73887205e-01 6.11110151e-01
-1.26617670e+00 -4.50436801e-01 -1.44917130e+00 -6.32112324e-01
-2.98922360e-01 3.33931714e-01 5.60483158e-01 9.71032321e-01
-1.40992177e+00 8.50537300e-01 -1.27478826e+00 2.41308644e-01
7.04128444e-01 6.87573195e-01 -2.53137536e-02 1.12331741e-01
-1.09738433e+00 5.65065682e-01 7.44282365e-01 1.59659848e-01
-3.64977330e-01 -9.35466051e-01 -5.61760187e-01 5.19910336e-01
-1.67158365e-01 -3.67232770e-01 1.34843564e+00 -4.28340137e-01
-1.66590583e+00 3.04106146e-01 -9.84028801e-02 -1.05610752e+00
3.67815703e-01 -1.28214210e-01 -2.03256533e-01 2.32233390e-01
-5.66469371e-01 1.00965512e+00 7.49219060e-01 -3.55058491e-01
-7.21505821e-01 -4.92473602e-01 3.11065540e-02 -1.36016876e-01
-1.09890020e+00 -2.87884325e-01 -5.18465281e-01 -7.56553054e-01
2.82495797e-01 -8.15364778e-01 -9.97335091e-02 5.53122878e-01
-1.98400728e-02 -3.54150116e-01 7.58606195e-01 -2.43753091e-01
1.47447181e+00 -2.23137331e+00 -2.90384144e-01 2.31497176e-02
5.63171029e-01 6.74654126e-01 1.75886139e-01 -1.35944754e-01
2.61634320e-01 7.08134472e-02 -7.26306364e-02 -5.21355033e-01
2.80814141e-01 5.36089838e-01 -4.45604235e-01 3.97097528e-01
1.15524143e-01 7.40362942e-01 -6.93358898e-01 -3.48948359e-01
-5.53172529e-02 7.37953424e-01 -8.30890417e-01 -3.36439252e-01
-5.25918156e-02 -7.25640476e-01 1.22158021e-01 3.26106161e-01
9.47316885e-01 -2.02217385e-01 2.84337372e-01 -6.45765066e-01
-6.01028763e-02 7.48458564e-01 -9.09827471e-01 1.54287529e+00
-1.62742019e-01 9.90005851e-01 -1.26593798e-01 -8.65272343e-01
9.39072132e-01 1.25093609e-01 2.03839511e-01 -7.65341938e-01
5.15910745e-01 4.84012336e-01 1.29261374e-01 2.19720155e-01
8.43011320e-01 6.75541833e-02 2.42361948e-01 1.01453729e-01
1.85716659e-01 1.38648778e-01 3.72867525e-01 -4.95954417e-02
1.02551711e+00 -6.64194524e-01 -2.18583032e-01 -3.40378463e-01
1.52994290e-01 -3.04943264e-01 5.10599554e-01 3.75826150e-01
-4.92635280e-01 1.84359759e-01 5.18116176e-01 -5.44999421e-01
-1.39242053e+00 -1.07452250e+00 -3.48819226e-01 1.04604149e+00
-1.83624953e-01 -6.61745489e-01 -1.02182353e+00 -2.18606293e-01
-2.24408150e-01 4.08698648e-01 -2.43044719e-01 -3.24345380e-01
-6.02418602e-01 -6.96372211e-01 1.17604494e+00 7.32282758e-01
9.06702876e-01 -4.31779832e-01 -1.07822990e+00 2.02834755e-01
5.39725937e-04 -1.46272671e+00 -4.98074234e-01 6.03617847e-01
-1.50789607e+00 -2.55649418e-01 -7.28869677e-01 -6.99380577e-01
4.93444830e-01 -2.04241611e-02 7.66863406e-01 -1.01281382e-01
9.36199725e-02 -5.62550724e-01 -8.07013288e-02 -2.34680638e-01
-9.09613818e-02 2.73483932e-01 3.17378402e-01 -3.18545163e-01
6.76545322e-01 -6.55462384e-01 -7.20896125e-01 -3.53786163e-02
-1.15388775e+00 -2.19621852e-01 6.19778216e-01 7.59371221e-01
5.74560046e-01 3.40799898e-01 1.18100993e-01 -1.57834783e-01
3.64185452e-01 1.44531056e-01 -9.07947421e-01 -5.88989332e-02
-7.18748927e-01 5.85432768e-01 9.43351269e-01 -6.34185553e-01
-2.98795789e-01 -3.06160003e-02 -4.49011654e-01 -8.11794341e-01
4.42300916e-01 2.88444549e-01 2.33443249e-02 -1.84989810e-01
5.76925337e-01 2.61341393e-01 -2.63067503e-02 -3.61323267e-01
2.68019378e-01 5.56561232e-01 6.80661798e-01 -2.32950091e-01
4.00640577e-01 2.96107024e-01 2.41552934e-01 -6.37748182e-01
-4.27726984e-01 2.69199703e-02 -3.79893452e-01 1.28607735e-01
6.06109381e-01 -8.21428120e-01 -1.06532001e+00 5.27540863e-01
-1.25799835e+00 -3.07104468e-01 -1.85383037e-01 6.48203731e-01
-2.00343907e-01 4.43387836e-01 -9.76718128e-01 -5.38520932e-01
-7.14550257e-01 -1.37090898e+00 8.04702282e-01 -3.55706587e-02
1.88205838e-01 -6.42701685e-01 -5.96265852e-01 -7.67475516e-02
5.54133892e-01 -3.01928252e-01 9.53227818e-01 -2.67848790e-01
-6.60393655e-01 3.84432301e-02 -6.38207912e-01 6.10408306e-01
-1.09626651e-01 -2.66076047e-02 -9.31255460e-01 -2.37673402e-01
-7.19200224e-02 -2.09888637e-01 8.63468230e-01 1.80490345e-01
1.44629395e+00 -5.85947633e-01 1.40923306e-01 9.14200068e-01
1.43759418e+00 2.90937513e-01 8.69793653e-01 1.89698115e-01
6.34492099e-01 -1.23039439e-01 1.91451758e-01 6.97796047e-01
3.01628113e-01 5.96825480e-01 6.12618387e-01 3.31170768e-01
-1.98807225e-01 -4.98634391e-02 3.18098843e-01 1.38894618e+00
5.01082055e-02 -8.74572322e-02 -9.81460452e-01 5.64456046e-01
-1.42653465e+00 -7.57564843e-01 -5.39280027e-02 2.23895288e+00
1.27260435e+00 5.88775456e-01 -1.41504914e-01 7.60116935e-01
3.70105743e-01 -2.91753095e-02 -6.06990099e-01 -6.63778007e-01
7.42718577e-02 3.87439221e-01 1.27694106e+00 3.26803863e-01
-9.66683328e-01 7.95289934e-01 6.32305098e+00 1.20328784e+00
-1.44479394e+00 1.13166884e-01 6.26854956e-01 -2.67590642e-01
1.74382944e-02 -6.44808173e-01 -1.19491100e+00 6.15706086e-01
1.49212968e+00 -2.18145400e-02 4.86836195e-01 1.00036132e+00
-1.57455891e-01 3.80896896e-01 -1.20659876e+00 1.50990272e+00
-3.61748606e-01 -1.51187694e+00 3.80883276e-01 2.09281549e-01
5.15696645e-01 3.00073177e-01 3.08438420e-01 1.79423779e-01
-2.90083904e-02 -8.52238655e-01 1.16093624e+00 1.20633215e-01
1.16136765e+00 -1.06967163e+00 8.35700750e-01 1.54012740e-01
-1.19664049e+00 -1.36411607e-01 -9.96073425e-01 -2.77085006e-02
-1.30767941e-01 8.82569194e-01 -8.55997324e-01 -1.60428315e-01
1.00117373e+00 3.15034747e-01 -2.94707954e-01 8.90728176e-01
1.70778558e-01 5.74503601e-01 -6.88211381e-01 -5.16509235e-01
5.05633950e-01 -1.67564169e-01 -4.11162376e-02 1.21086740e+00
5.07590234e-01 8.25248510e-02 -6.02263272e-01 6.67732596e-01
-5.38840771e-01 -1.86396152e-01 -5.21774217e-02 -1.06742360e-01
8.61229062e-01 7.63278663e-01 -6.52682483e-01 -6.61466658e-01
-1.03407346e-01 1.11122870e+00 4.33488369e-01 1.92190241e-02
-9.69048798e-01 -9.74481285e-01 9.74886179e-01 -9.83840376e-02
7.63008416e-01 -5.34900725e-01 -6.94676995e-01 -8.83109510e-01
3.79114419e-01 -5.75862169e-01 -2.16332167e-01 -4.04120713e-01
-7.89875448e-01 8.41766119e-01 -2.58967578e-01 -1.04008317e+00
-5.62519170e-02 -1.06068468e+00 3.03712755e-01 5.55204690e-01
-1.37696636e+00 -3.43842357e-01 -1.90101400e-01 3.72397244e-01
2.52023727e-01 4.49919179e-02 6.34668589e-01 1.01603842e+00
-5.21462798e-01 1.37902009e+00 4.81902212e-01 3.46127898e-01
2.42373601e-01 -1.08050442e+00 7.59315729e-01 7.41938531e-01
3.36934067e-02 8.00846040e-01 5.52962005e-01 -9.30183828e-02
-1.53623343e+00 -1.28005302e+00 9.37182367e-01 2.43820623e-01
7.52059340e-01 -5.15299141e-01 -1.00972342e+00 3.63328874e-01
-1.90249860e-01 4.15900499e-01 4.63397563e-01 -4.59691167e-01
-6.63554072e-01 -6.98196590e-01 -1.19731295e+00 8.09129894e-01
8.95054162e-01 -5.95755041e-01 -1.16339140e-01 9.16167647e-02
9.42723930e-01 -4.67676759e-01 -1.23057652e+00 1.81250677e-01
6.53281868e-01 -8.83133531e-01 9.83175695e-01 -8.86625499e-02
4.00281310e-01 -3.66209060e-01 -6.23841822e-01 -9.91615653e-01
-1.84220105e-01 -3.79770935e-01 -6.05284274e-01 8.36367607e-01
2.19591886e-01 -5.20473838e-01 1.02511108e+00 5.64778924e-01
-1.90800682e-01 -9.41428721e-01 -1.31641102e+00 -9.88050520e-01
2.04030633e-01 -7.81915188e-01 8.44378710e-01 4.70483392e-01
-1.30801603e-01 9.83311534e-02 1.28656194e-01 1.61633000e-01
7.32916951e-01 -6.04536533e-01 1.97264582e-01 -9.72808719e-01
-5.97433187e-03 -7.56959379e-01 -9.93734300e-01 -1.80859601e+00
-2.51113743e-01 -7.65670002e-01 1.91135798e-02 -1.12072945e+00
-2.68488824e-01 -5.82670510e-01 -3.36313844e-01 4.57391053e-01
2.54103839e-01 7.36420035e-01 2.61700541e-01 6.84777200e-02
-5.51111639e-01 7.88933337e-01 9.58130956e-01 -2.22699031e-01
1.24883167e-01 -5.37338436e-01 -2.15566829e-01 6.59453928e-01
7.61476457e-01 -3.39491993e-01 -5.95655739e-01 -1.11311316e+00
2.31534988e-01 -3.02220255e-01 2.33346790e-01 -1.43026042e+00
5.51885724e-01 2.75602490e-01 2.04733476e-01 -9.17408705e-01
5.62359035e-01 -8.76039922e-01 -1.11060828e-01 6.53229654e-01
-2.01458797e-01 4.17107612e-01 4.14699823e-01 2.17827469e-01
-2.18909904e-01 -3.61068845e-01 8.91203821e-01 4.26894248e-01
-4.96356159e-01 4.58616674e-01 -4.68498528e-01 -1.15639098e-01
5.73220432e-01 -2.19951361e-01 -3.59472185e-01 7.93327093e-02
-2.61454672e-01 -3.09013933e-01 2.00099438e-01 -7.67494589e-02
7.53377080e-01 -1.56605625e+00 -2.90308297e-01 2.14221597e-01
-2.28697047e-01 2.13585153e-01 6.67284206e-02 5.63658416e-01
-1.22007036e+00 8.72824132e-01 -2.80958444e-01 -7.37007737e-01
-1.09093022e+00 3.49276841e-01 3.81251782e-01 -1.44909859e-01
-5.44554770e-01 1.15344191e+00 -2.85010785e-01 -3.01087479e-04
7.60539174e-01 -1.29701400e+00 2.88476020e-01 -1.12776116e-01
9.11242008e-01 5.07488012e-01 6.91302538e-01 -3.32694322e-01
-3.54918420e-01 3.90272290e-01 -1.55761093e-01 -1.14219509e-01
1.14064300e+00 1.66808963e-01 -2.13848665e-01 2.41947398e-01
1.66638851e+00 -5.17610788e-01 -1.33388412e+00 -2.72213876e-01
-7.13780299e-02 -3.24364126e-01 6.08152032e-01 -3.06081027e-01
-1.62457943e+00 1.25499153e+00 9.18376625e-01 1.47239536e-01
1.38836491e+00 -4.98625964e-01 1.35385203e+00 9.39661145e-01
4.49456632e-01 -9.96898234e-01 -8.41370896e-02 7.87195146e-01
3.95027310e-01 -7.36856401e-01 -4.55966219e-02 -1.17120922e-01
-2.00263128e-01 1.35492373e+00 4.13745314e-01 -1.64231658e-01
7.90989339e-01 7.20798194e-01 -1.35325402e-01 1.16795480e-01
-8.77981365e-01 2.66348511e-01 1.06247544e-01 3.17245841e-01
2.33748794e-01 9.04039592e-02 -2.59559393e-01 3.54421347e-01
-8.04772139e-01 1.11322895e-01 4.08820868e-01 6.71739399e-01
-5.51539898e-01 -7.95342326e-01 -2.75872111e-01 4.87858385e-01
-6.55008614e-01 -4.76159185e-01 3.31863523e-01 2.94652909e-01
1.93585396e-01 6.70214832e-01 6.57474697e-01 -7.09673882e-01
1.29768610e-01 -1.36270806e-01 4.35126126e-01 3.30871679e-02
-4.10004169e-01 -1.64075926e-01 -2.88352728e-01 -4.60183203e-01
-2.61717346e-02 -2.56969243e-01 -1.51333725e+00 -9.45407033e-01
-2.67059386e-01 -1.81199744e-01 1.00968635e+00 5.67053914e-01
4.13076341e-01 6.92757845e-01 1.83767378e-01 -6.08905971e-01
-9.71529484e-01 -8.48602176e-01 -4.31434333e-01 -6.80662915e-02
5.38036764e-01 -4.73558426e-01 -1.77723929e-01 6.06934950e-02] | [8.543625831604004, 3.0354163646698] |
3041b162-efe5-49bc-a6e4-abd967290c34 | procontext-exploring-progressive-context | 2210.15511 | null | https://arxiv.org/abs/2210.15511v4 | https://arxiv.org/pdf/2210.15511v4.pdf | ProContEXT: Exploring Progressive Context Transformer for Tracking | Existing Visual Object Tracking (VOT) only takes the target area in the first frame as a template. This causes tracking to inevitably fail in fast-changing and crowded scenes, as it cannot account for changes in object appearance between frames. To this end, we revamped the tracking framework with Progressive Context Encoding Transformer Tracker (ProContEXT), which coherently exploits spatial and temporal contexts to predict object motion trajectories. Specifically, ProContEXT leverages a context-aware self-attention module to encode the spatial and temporal context, refining and updating the multi-scale static and dynamic templates to progressively perform accurately tracking. It explores the complementary between spatial and temporal context, raising a new pathway to multi-context modeling for transformer-based trackers. In addition, ProContEXT revised the token pruning technique to reduce computational complexity. Extensive experiments on popular benchmark datasets such as GOT-10k and TrackingNet demonstrate that the proposed ProContEXT achieves state-of-the-art performance. | ['Xuansong Xie', 'Yifeng Geng', 'Wangmeng Xiang', 'Xu Bao', 'Bin Luo', 'Chenyang Li', 'Jun-Yan He', 'Zhi-Qi Cheng', 'Jin-Peng Lan'] | 2022-10-27 | null | null | null | null | ['visual-object-tracking'] | ['computer-vision'] | [-1.10372566e-01 -5.46678007e-01 -4.37357187e-01 -1.99867919e-01
-2.83596992e-01 -7.01626360e-01 6.54737830e-01 -1.62071347e-01
-3.80721092e-01 4.89395231e-01 2.44985506e-01 -1.75122947e-01
2.94580221e-01 -3.20412904e-01 -7.12106645e-01 -4.59404558e-01
-1.98998541e-01 6.12843968e-03 1.07827353e+00 1.93989798e-01
-6.88395426e-02 4.72529769e-01 -1.63006043e+00 2.39845201e-01
5.89682162e-01 1.06568909e+00 3.26168209e-01 8.10878515e-01
-1.38003752e-01 8.67041647e-01 -3.80255580e-01 -5.22150278e-01
3.67926240e-01 5.88119142e-02 -4.76412743e-01 -1.06914509e-02
1.06976259e+00 -4.13684070e-01 -7.99345016e-01 9.95751977e-01
2.96950847e-01 3.41398939e-02 1.26257196e-01 -1.35699308e+00
-7.78712034e-01 3.42220068e-01 -7.15240180e-01 7.45736718e-01
1.25898346e-01 8.52073491e-01 7.76817679e-01 -1.01668370e+00
7.11669743e-01 1.35879910e+00 9.00552750e-01 5.77249885e-01
-9.54906404e-01 -7.97719061e-01 8.36806238e-01 3.11504900e-01
-1.14674938e+00 -4.60533053e-01 4.56539065e-01 -4.51924324e-01
9.90204394e-01 1.68212593e-01 1.13237607e+00 1.21819091e+00
4.34171975e-01 1.09121752e+00 7.76814878e-01 5.45178540e-02
-6.09277710e-02 -5.23590147e-01 2.13072047e-01 5.06208062e-01
4.60927248e-01 5.65814734e-01 -6.72050953e-01 -4.61226255e-02
6.36524796e-01 1.30418435e-01 -2.09741205e-01 -4.73101705e-01
-1.43672633e+00 4.28948402e-01 5.48922300e-01 8.53242800e-02
-1.47301584e-01 6.37721181e-01 4.37206030e-01 -6.67009205e-02
1.99634328e-01 6.72685727e-02 -5.86156368e-01 -1.89490974e-01
-1.03441000e+00 2.78326541e-01 2.96561092e-01 1.19419003e+00
4.66929823e-01 2.49909863e-01 -8.04212213e-01 2.06464484e-01
4.87974703e-01 7.39246070e-01 1.54321149e-01 -1.06544209e+00
1.86953947e-01 6.06648207e-01 1.99600711e-01 -7.41706669e-01
-2.80886054e-01 -6.20867550e-01 -1.96245253e-01 1.36779219e-01
5.22301197e-01 8.19167793e-02 -1.14928150e+00 1.79440105e+00
8.65049839e-01 1.04158306e+00 -3.81032586e-01 9.95472491e-01
8.91005576e-01 3.22022617e-01 5.55686235e-01 -1.79513246e-01
1.47988224e+00 -1.23175633e+00 -7.78778851e-01 -4.33795899e-01
2.63317615e-01 -7.02828646e-01 8.11857581e-01 3.74023169e-02
-1.01380777e+00 -8.22264075e-01 -8.84039938e-01 -5.53274192e-02
-2.28013501e-01 4.53460589e-02 6.53546393e-01 5.33163786e-01
-8.81277263e-01 3.54810357e-01 -1.15715063e+00 -2.63115168e-01
7.18537569e-01 3.52671206e-01 -1.04355745e-01 -1.06951408e-01
-8.37304235e-01 5.49914837e-01 4.00594831e-01 2.79795676e-01
-1.03380430e+00 -1.08162534e+00 -7.75693536e-01 -2.18518868e-01
7.33718574e-01 -8.66709888e-01 1.36285639e+00 -7.11426079e-01
-1.20971417e+00 5.19454598e-01 -4.88193661e-01 -6.05318129e-01
5.48986435e-01 -5.74852288e-01 -5.31387985e-01 -6.19938485e-02
3.00644916e-02 9.07514036e-01 1.04928005e+00 -1.10530818e+00
-1.04395533e+00 -1.83008879e-01 -1.42852902e-01 -6.21007644e-02
-1.32271767e-01 2.67744571e-01 -1.30533600e+00 -8.95817995e-01
-2.52343059e-01 -9.36537504e-01 -1.44264132e-01 4.86650646e-01
-1.92606404e-01 -2.39948675e-01 1.62183797e+00 -3.70017469e-01
1.46646357e+00 -1.93455625e+00 -1.34799495e-01 -2.12366402e-01
5.32960415e-01 6.67043090e-01 -2.17202812e-01 -1.75863996e-01
3.30949962e-01 -2.76357472e-01 1.23134919e-01 -6.16591632e-01
-1.13569282e-01 3.82364631e-01 -3.16007525e-01 4.20148730e-01
4.10719603e-01 1.51016736e+00 -1.21216321e+00 -8.02726150e-01
3.38619947e-01 6.59063101e-01 -6.17171645e-01 -3.96639220e-02
-6.55947089e-01 4.23308969e-01 -4.27587807e-01 9.58200395e-01
6.28396213e-01 -3.34508747e-01 1.46239623e-01 -2.91305661e-01
-3.97261202e-01 1.05323866e-01 -9.83780265e-01 1.69198978e+00
7.33773634e-02 8.40323329e-01 -1.16430588e-01 -1.38951167e-01
5.37440538e-01 -5.00387512e-02 7.04082847e-01 -7.83857882e-01
1.55950055e-01 -3.34448785e-01 -1.20152041e-01 -3.22251737e-01
8.51820588e-01 5.36941171e-01 1.84598729e-01 2.33739093e-02
-8.34677815e-02 7.84278393e-01 1.56085059e-01 3.00248384e-01
1.19861686e+00 6.25708759e-01 9.22828093e-02 -9.22279954e-02
3.82013321e-01 4.38104160e-02 1.07449436e+00 7.18510270e-01
-8.04008484e-01 3.11330557e-01 8.54291171e-02 -7.21420467e-01
-6.67001247e-01 -1.11410904e+00 1.36572123e-01 1.27099276e+00
4.81871754e-01 -5.50207078e-01 -4.13902551e-01 -9.85561311e-01
3.40963453e-01 3.27807933e-01 -8.90482306e-01 -4.69469577e-02
-9.40207124e-01 -3.47741544e-01 4.16738123e-01 9.09967661e-01
3.23301196e-01 -9.54136133e-01 -1.08585429e+00 3.93395513e-01
-3.63569852e-04 -1.46921813e+00 -1.12417817e+00 -1.31299838e-01
-7.99804688e-01 -1.23685777e+00 -3.73345941e-01 -5.01416445e-01
2.51186669e-01 5.69063485e-01 1.14904356e+00 2.59008110e-01
-3.71269345e-01 5.83765626e-01 -2.77647823e-01 -4.24604654e-01
-9.93427262e-03 -1.59147069e-01 2.37935179e-04 4.43390794e-02
4.67628926e-01 -2.92706907e-01 -8.56468678e-01 5.03931940e-01
-4.99595135e-01 -4.79731858e-02 3.79805386e-01 5.95234811e-01
8.17140937e-01 -4.15012866e-01 2.43909240e-01 -3.58992994e-01
-1.01128943e-01 -3.52157742e-01 -9.41728652e-01 4.97994781e-01
-3.45306933e-01 -1.63757026e-01 1.97382629e-01 -9.85679030e-01
-1.00310433e+00 3.15309256e-01 2.95307964e-01 -1.02666950e+00
1.98089272e-01 -1.65522188e-01 -2.01081470e-01 -3.76688451e-01
1.95489481e-01 2.81402946e-01 -3.65521848e-01 -4.17079329e-01
4.28208590e-01 -6.27620518e-02 9.72545624e-01 -4.88719374e-01
9.75715458e-01 7.36007512e-01 -1.82646945e-01 -4.55363691e-01
-7.80874968e-01 -4.50533986e-01 -6.29245460e-01 -5.77307224e-01
1.03845513e+00 -9.90756512e-01 -1.09136653e+00 4.22780156e-01
-1.13196838e+00 -4.86662954e-01 -3.96476895e-01 3.74663085e-01
-1.55488402e-01 4.12260145e-01 -5.03687203e-01 -7.78771579e-01
-3.49767268e-01 -1.23593187e+00 1.33515859e+00 4.78358507e-01
4.98752631e-02 -8.65933061e-01 5.21845147e-02 -6.16869926e-02
4.10148054e-01 2.61006385e-01 1.04644977e-01 -1.60832107e-01
-1.26763844e+00 1.77470952e-01 -4.38926578e-01 -2.41876826e-01
-1.42915547e-02 1.84111878e-01 -9.33956742e-01 -5.09916782e-01
-4.47985739e-01 2.34397035e-02 9.40474927e-01 5.40088713e-01
1.07939327e+00 -3.10926158e-02 -7.28399336e-01 8.78784955e-01
1.15352750e+00 2.03306273e-01 4.30016339e-01 2.99784303e-01
1.00520849e+00 -9.14216507e-03 9.99073267e-01 3.20782542e-01
6.86746836e-01 1.12716007e+00 5.60185373e-01 1.77294448e-01
-6.45945728e-01 -3.07498187e-01 5.39011180e-01 5.14713109e-01
1.19486697e-01 -8.58553499e-02 -7.85050035e-01 7.85505652e-01
-1.97767162e+00 -1.27582622e+00 -1.91033319e-01 2.01648712e+00
6.88236833e-01 2.84021825e-01 4.48189080e-01 -5.09837270e-01
7.44082928e-01 3.12606841e-01 -9.10739005e-01 2.70688444e-01
-1.26932099e-01 -8.52917284e-02 7.04963923e-01 3.36497337e-01
-1.33721375e+00 1.32857931e+00 6.35278082e+00 6.04467750e-01
-1.16410875e+00 3.64009082e-01 1.10315189e-01 -5.27223289e-01
5.51899634e-02 8.56222212e-02 -1.29952562e+00 7.49647677e-01
6.95719004e-01 2.83402894e-02 2.47793570e-02 7.92824864e-01
1.80070803e-01 -2.21209340e-02 -1.00692940e+00 8.15052748e-01
-2.65420943e-01 -1.39890993e+00 -1.86039850e-01 2.85818670e-02
6.77796781e-01 4.09907937e-01 1.84224620e-01 4.16041970e-01
5.11236727e-01 -6.72058105e-01 1.03007305e+00 5.77022731e-01
5.60612619e-01 -3.58879775e-01 2.25293234e-01 -5.68388551e-02
-2.14471412e+00 -2.26566732e-01 -2.19090596e-01 2.95475006e-01
3.28241736e-01 2.39392892e-01 -4.72463369e-01 5.45103014e-01
9.43176091e-01 9.98662889e-01 -9.61102307e-01 1.48513973e+00
4.77691414e-03 6.44025624e-01 -4.85160977e-01 3.75625849e-01
2.29210675e-01 4.05893564e-01 7.80068994e-01 1.37754476e+00
2.24609450e-02 7.94129297e-02 5.78953147e-01 6.97373986e-01
1.13543503e-01 -5.24147332e-01 -2.38276467e-01 3.21706802e-01
7.91828990e-01 1.08549118e+00 -7.13000178e-01 -4.76846814e-01
-6.53594375e-01 5.81558585e-01 3.02135557e-01 3.47774416e-01
-1.40663588e+00 2.86241502e-01 9.74072337e-01 2.96751447e-02
1.04931521e+00 -3.86420161e-01 -1.48418039e-01 -1.18132925e+00
6.07740209e-02 -7.11607873e-01 5.56613088e-01 -5.38507819e-01
-1.03089166e+00 5.26815116e-01 1.32777533e-02 -1.39367139e+00
2.67722040e-01 -4.23138767e-01 -6.57791615e-01 4.06326443e-01
-1.77295148e+00 -1.54539609e+00 -5.27493536e-01 7.60159612e-01
7.98316956e-01 1.49094254e-01 2.18357388e-02 4.40894485e-01
-7.61183619e-01 8.32361937e-01 -3.52917165e-01 1.75368801e-01
8.44338357e-01 -1.05154812e+00 9.50513005e-01 1.19152761e+00
1.23305142e-01 6.84179246e-01 5.35622537e-01 -1.08714843e+00
-1.67549837e+00 -1.65754938e+00 5.53906918e-01 -8.53450418e-01
7.09836423e-01 -2.78427571e-01 -1.02058887e+00 8.30537200e-01
1.61119804e-01 8.24378490e-01 2.51216710e-01 -1.93439633e-01
-5.92336535e-01 3.52322939e-03 -7.02387869e-01 7.49893963e-01
1.68339229e+00 -3.22230160e-01 -5.28337121e-01 -4.04618494e-02
1.12230921e+00 -7.61492968e-01 -8.14446390e-01 3.98632973e-01
8.77070546e-01 -8.06875050e-01 1.27353227e+00 -5.70982695e-01
-1.98029935e-01 -8.93186092e-01 -6.07906729e-02 -5.50404072e-01
-4.88343626e-01 -8.83391559e-01 -1.00005460e+00 1.11115825e+00
-5.62932305e-02 -3.05192947e-01 9.11129296e-01 5.67281246e-01
-1.97431207e-01 -9.13154125e-01 -9.33143795e-01 -1.01830781e+00
-2.80550480e-01 -7.19114780e-01 7.82865822e-01 7.61224449e-01
-6.06264651e-01 -1.39846250e-01 -5.45206904e-01 3.86308312e-01
9.06468987e-01 1.82855144e-01 9.45460498e-01 -1.16971493e+00
-2.65900016e-01 -3.59457880e-01 -4.99285966e-01 -1.48329866e+00
1.05379954e-01 -6.37831211e-01 6.15879074e-02 -1.10922170e+00
2.04524502e-01 -4.42929685e-01 -3.80823433e-01 6.49652660e-01
-6.46659374e-01 3.58774871e-01 6.84210956e-01 2.96166509e-01
-1.31679440e+00 6.39248073e-01 1.30645299e+00 -1.82778955e-01
-2.44976044e-01 -4.37036008e-02 -5.12148440e-01 6.50409162e-01
3.72862786e-01 -6.06388390e-01 -2.95080274e-01 -4.97486204e-01
-2.10366175e-01 -2.02984452e-01 8.01301181e-01 -1.15630865e+00
5.81143081e-01 -3.21930617e-01 5.98743618e-01 -1.11426020e+00
2.60181606e-01 -9.85622585e-01 1.95889130e-01 5.03383934e-01
-9.84884799e-02 4.01797056e-01 5.52003324e-01 9.92055655e-01
1.38034388e-01 4.51210171e-01 7.17644215e-01 3.11580479e-01
-1.26479685e+00 7.61632025e-01 -4.77954298e-02 1.72774270e-01
1.20357311e+00 -6.28281534e-01 -5.24883270e-01 1.93693548e-01
-5.60511827e-01 7.20888197e-01 6.33071423e-01 8.57071936e-01
6.04227662e-01 -1.49877656e+00 -3.03175241e-01 -6.46283617e-03
1.27430350e-01 -1.12122737e-01 3.58934939e-01 9.69307840e-01
-2.73907874e-02 3.89326066e-01 -4.62572761e-02 -1.08475423e+00
-1.56607211e+00 8.76501977e-01 3.42249185e-01 -2.65283585e-01
-1.00335312e+00 7.74020553e-01 3.72991472e-01 1.62178189e-01
4.54471141e-01 -7.05790102e-01 5.74447075e-03 -3.06254357e-01
8.24269235e-01 2.03449771e-01 -2.77305722e-01 -7.79765010e-01
-5.48915505e-01 7.58486927e-01 -3.14872980e-01 1.73794702e-01
7.96441197e-01 -2.86263973e-01 4.26663578e-01 4.50526699e-02
8.05076599e-01 4.93863225e-03 -2.17315006e+00 -4.89434838e-01
1.77700847e-01 -8.66599143e-01 1.06400147e-01 -7.56630123e-01
-1.37148523e+00 4.63443220e-01 5.82591236e-01 -2.28171572e-01
1.12333941e+00 -1.52735859e-01 1.11381960e+00 2.48249277e-01
4.55711633e-01 -8.30514371e-01 2.21779227e-01 7.18741596e-01
4.76234078e-01 -1.13777256e+00 5.87431230e-02 -3.83336306e-01
-6.37150645e-01 7.83024609e-01 1.07316208e+00 9.60072130e-02
5.04294336e-01 5.66363394e-01 -1.55988604e-01 -1.75822109e-01
-1.11896014e+00 -4.89447564e-01 5.73981166e-01 8.10523510e-01
1.79375917e-01 -2.84375966e-01 9.56650227e-02 3.65503281e-01
2.79471546e-01 1.93510458e-01 -2.80218944e-02 1.20287848e+00
-3.83500695e-01 -1.02265775e+00 -5.73568642e-01 1.59224600e-01
-3.49759668e-01 1.08931921e-01 -2.65778720e-01 9.42020774e-01
4.44870114e-01 7.33234704e-01 8.02548602e-02 -3.99124652e-01
3.69402707e-01 -1.71931729e-01 4.94581312e-01 -1.75551116e-01
-9.85662818e-01 2.86886930e-01 -2.14226633e-01 -1.14600861e+00
-5.72078705e-01 -9.06195402e-01 -1.20642471e+00 -2.17486307e-01
-2.37020984e-01 -3.45517725e-01 3.55201185e-01 7.89322257e-01
7.01849520e-01 8.38250399e-01 9.20667276e-02 -9.85961318e-01
-1.52680054e-01 -4.59475875e-01 2.04919800e-01 2.46341407e-01
6.23973370e-01 -1.05256855e+00 2.72099048e-01 1.36471331e-01] | [6.304574489593506, -2.0792808532714844] |
50e98043-6c83-44ae-ae28-042390b1bbfd | sim-to-real-transfer-for-miniature-autonomous | 2011.05617 | null | https://arxiv.org/abs/2011.05617v1 | https://arxiv.org/pdf/2011.05617v1.pdf | Sim-To-Real Transfer for Miniature Autonomous Car Racing | Sim-to-real, a term that describes where a model is trained in a simulator then transferred to the real world, is a technique that enables faster deep reinforcement learning (DRL) training. However, differences between the simulator and the real world often cause the model to perform poorly in the real world. Domain randomization is a way to bridge the sim-to-real gap by exposing the model to a wide range of scenarios so that it can generalize to real-world situations. However, following domain randomization to train an autonomous car racing model with DRL can lead to undesirable outcomes. Namely, a model trained with randomization tends to run slower; a higher completion rate on the testing track comes at the expense of longer lap times. This paper aims to boost the robustness of a trained race car model without compromising racing lap times. For a training track and a testing track having the same shape (and same optimal paths), but with different lighting, background, etc., we first train a model (teacher model) that overfits the training track, moving along a near optimal path. We then use this model to teach a student model the correct actions along with randomization. With our method, a model with 18.4\% completion rate on the testing track is able to help teach a student model with 52\% completion. Moreover, over an average of 50 trials, the student is able to finish a lap 0.23 seconds faster than the teacher. This 0.23 second gap is significant in tight races, with lap times of about 10 to 12 seconds. | ['I-Chen Wu', 'Yuan-Hao Chen', 'Jin-Bo Huang', 'Ting-Han Wei', 'Yeong-Jia Roger Chu'] | 2020-11-11 | null | null | null | null | ['carracing-v0'] | ['playing-games'] | [-6.37848228e-02 3.95760238e-02 -1.73557267e-01 -1.92292854e-01
-7.26187408e-01 -8.89204979e-01 2.55346000e-01 -3.05849351e-02
-5.26534677e-01 8.38212013e-01 -4.83583927e-01 -8.26241374e-01
1.24042109e-01 -9.59133148e-01 -1.28864896e+00 -3.90551865e-01
7.27662668e-02 3.38378757e-01 4.53478903e-01 -6.15251660e-01
7.17890859e-02 7.11241841e-01 -1.86129689e+00 1.91281887e-03
1.04905415e+00 3.81215662e-01 2.25887269e-01 8.22857559e-01
6.73637912e-02 9.53970313e-01 -9.16132569e-01 1.98658314e-02
3.55797470e-01 -6.04060531e-01 -8.57942939e-01 -2.68961847e-01
5.86968541e-01 -2.35899478e-01 -4.55246031e-01 8.17280948e-01
2.61485249e-01 5.25600553e-01 1.98925763e-01 -1.44641209e+00
1.45322410e-02 3.83347750e-01 -4.92233843e-01 -1.21958151e-01
4.13666964e-01 5.59274316e-01 3.04805130e-01 -8.48250687e-02
5.34080684e-01 9.90671813e-01 4.80415910e-01 7.39618480e-01
-1.36162066e+00 -1.03126979e+00 4.79781814e-02 3.70894372e-02
-1.22478747e+00 -4.56629060e-02 5.26878059e-01 -2.55093575e-01
4.57441986e-01 8.57192129e-02 6.69017494e-01 1.21350741e+00
4.24681574e-01 5.65575480e-01 1.40025198e+00 -1.39120683e-01
4.21536356e-01 2.76122987e-01 -1.35256618e-01 7.03659654e-01
1.66870788e-01 7.87473023e-01 -3.81485403e-01 2.88468450e-01
5.67481160e-01 -1.56705618e-01 -1.34875625e-01 -5.99668860e-01
-9.86228287e-01 6.54003739e-01 6.08434737e-01 -7.47097936e-03
-1.55470610e-01 3.68000776e-01 3.48266065e-01 5.07155716e-01
-4.00239229e-01 9.08780932e-01 -4.22276676e-01 -5.65321207e-01
-6.91061258e-01 7.85533726e-01 7.32248306e-01 8.59191120e-01
8.78579736e-01 2.24040598e-01 1.04471587e-01 3.40251237e-01
-2.57806808e-01 7.37022400e-01 2.94536650e-01 -1.16080451e+00
4.49210644e-01 5.17075360e-01 3.40233326e-01 -6.49731040e-01
-3.45200628e-01 -2.33797073e-01 -2.92135477e-01 9.53190446e-01
9.07793820e-01 -4.34127629e-01 -8.34652603e-01 2.00331068e+00
4.73789781e-01 3.55807215e-01 2.11610362e-01 8.51587892e-01
3.28258723e-01 7.56411135e-01 8.43367204e-02 3.50083292e-01
1.10697353e+00 -8.34812760e-01 -2.04521969e-01 -4.17803943e-01
1.01838815e+00 -7.50664532e-01 1.59121835e+00 5.19222200e-01
-1.05505466e+00 -1.08563483e+00 -1.33526433e+00 4.75468636e-01
-5.13341069e-01 -1.63969040e-01 3.44967574e-01 8.72261703e-01
-8.12100589e-01 1.02501810e+00 -7.71309257e-01 -1.52743757e-01
5.00766337e-02 3.27218443e-01 -4.13966656e-01 -3.35773647e-01
-1.09765315e+00 1.12202454e+00 4.13530409e-01 -3.27600777e-01
-1.47592163e+00 -1.03867471e+00 -1.01110387e+00 -8.57900679e-02
4.88390058e-01 -2.02634856e-01 1.53487039e+00 -8.65301907e-01
-1.86637974e+00 5.05349159e-01 8.68363529e-02 -2.98157275e-01
6.97731972e-01 -3.22841674e-01 -4.88484383e-01 -2.14247420e-01
1.46400005e-01 6.36454701e-01 5.03843665e-01 -1.37080574e+00
-6.78531528e-01 1.28895789e-02 4.77056086e-01 1.35405168e-01
3.75734389e-01 -3.93962562e-01 -1.02876425e-01 -1.93976685e-01
-2.26330161e-01 -1.28871047e+00 -3.48205715e-01 -2.85005480e-01
3.14150341e-02 2.92420350e-02 8.60531509e-01 -2.71010339e-01
8.27686012e-01 -2.32404065e+00 -1.95056394e-01 1.81507617e-01
-1.46090895e-01 3.22319329e-01 -3.78367960e-01 3.49871099e-01
-3.51735741e-01 -9.20314789e-02 -1.20484307e-01 1.27321392e-01
-1.55318826e-01 4.09572989e-01 -4.29876417e-01 4.78586972e-01
8.90983492e-02 6.26211464e-01 -1.14460337e+00 -1.11801244e-01
2.41426393e-01 6.08117096e-02 -6.36528492e-01 5.34270763e-01
-4.54533212e-02 7.75701940e-01 -1.90101534e-01 1.29342824e-01
7.40934730e-01 2.94140488e-01 1.04785658e-01 3.29346299e-01
-2.40843341e-01 1.81252450e-01 -1.42242372e+00 1.67698824e+00
-7.76580751e-01 7.21894622e-01 -9.57798138e-02 -8.76992881e-01
1.02255559e+00 1.14197075e-01 1.37718841e-01 -1.06924343e+00
1.13738224e-01 2.13020653e-01 2.19994158e-01 -5.96215129e-01
5.87496936e-01 -2.34840304e-01 -2.59796470e-01 4.77048784e-01
-1.79391831e-01 -5.05519509e-01 3.36980641e-01 7.57611170e-02
1.03298557e+00 5.43392718e-01 -3.01465183e-01 -2.96704441e-01
3.44553143e-01 5.12896895e-01 5.36351323e-01 8.79637361e-01
-3.89547646e-01 1.59250706e-01 5.37092030e-01 -6.31405592e-01
-1.10626149e+00 -1.19067466e+00 2.62354761e-01 1.07938790e+00
5.05839527e-01 -1.92411512e-01 -7.39614964e-01 -7.98163533e-01
-5.04023507e-02 1.21621335e+00 -5.85678935e-01 -7.13479400e-01
-9.71443951e-01 -6.76271915e-02 8.23795497e-01 6.96743369e-01
5.79103649e-01 -9.52566922e-01 -1.02158499e+00 1.56423301e-01
-1.06554553e-01 -1.03576338e+00 -3.29581559e-01 2.11006731e-01
-8.28937173e-01 -1.36341512e+00 -3.37578267e-01 -6.82465374e-01
7.37966895e-01 2.96664059e-01 1.17460442e+00 3.32669586e-01
-2.39070609e-01 -3.72371031e-03 -2.74076402e-01 -3.39945257e-01
-8.30639184e-01 -1.57972246e-01 1.36848524e-01 -5.69533169e-01
1.62839249e-01 -2.02158883e-01 -3.76770020e-01 6.25050068e-01
-9.39151168e-01 1.89292893e-01 2.47001082e-01 9.16533649e-01
2.07242087e-01 2.85020024e-01 2.05147341e-01 -6.74476922e-01
4.41715717e-01 -2.34031841e-01 -9.18896377e-01 5.97441122e-02
-5.64713120e-01 3.61594588e-01 1.02924979e+00 -7.84718037e-01
-8.76242280e-01 2.26934887e-02 8.76853149e-03 -3.04982841e-01
-5.05696058e-01 1.54062554e-01 2.78307255e-02 -7.45275989e-02
1.14292049e+00 6.13185251e-03 1.60252601e-01 -1.33784443e-01
2.62126148e-01 1.91052668e-02 5.13969839e-01 -1.05874157e+00
1.20622563e+00 8.30835253e-02 5.10745402e-03 -3.81005704e-01
-7.41535366e-01 1.40544668e-01 -3.88010114e-01 -4.24562037e-01
6.57834649e-01 -7.25839615e-01 -1.03024840e+00 3.30583334e-01
-5.81760347e-01 -1.17493403e+00 -5.81684709e-01 7.02465534e-01
-5.61897397e-01 -2.24344894e-01 -2.77357727e-01 -6.20434225e-01
5.10706544e-01 -1.45816290e+00 5.35021126e-01 7.62289166e-01
-1.57443941e-01 -6.92606151e-01 2.86298126e-01 3.00157964e-01
4.26493168e-01 3.65310073e-01 8.26773643e-01 -2.84045875e-01
-3.46936464e-01 -2.48186469e-01 1.25774994e-01 2.65101373e-01
-9.73720774e-02 2.86566708e-02 -9.26914513e-01 -4.89394903e-01
-4.09923553e-01 -5.50441623e-01 2.91932821e-01 7.22128898e-02
1.26208806e+00 3.07988346e-01 -2.80405104e-01 3.86631906e-01
1.32773805e+00 5.20520806e-01 6.65078461e-01 5.38772702e-01
3.50804389e-01 5.34633219e-01 1.04600608e+00 -1.56820357e-01
1.89960033e-01 5.17548859e-01 4.39578950e-01 -3.55063081e-01
-1.74813718e-02 -7.10154235e-01 6.85578108e-01 2.29291052e-01
1.67953461e-01 8.09174031e-02 -1.06644833e+00 4.97200578e-01
-1.40815866e+00 -9.75732625e-01 5.78126982e-02 2.49743819e+00
7.36168325e-01 4.88973558e-01 3.44095647e-01 2.51793563e-01
4.46939081e-01 -1.63555890e-01 -5.57737470e-01 -8.31512928e-01
3.69371891e-01 4.96811062e-01 8.05225730e-01 6.43837869e-01
-7.05686569e-01 1.12898600e+00 6.51215076e+00 6.78678691e-01
-1.40666795e+00 -3.84858131e-01 4.19604212e-01 1.89408034e-01
-4.49086241e-02 3.04605871e-01 -6.69169903e-01 2.73852050e-01
1.34754777e+00 -1.75498605e-01 7.48984456e-01 1.04688752e+00
4.14754957e-01 -4.82850641e-01 -1.22048688e+00 5.97881973e-01
-2.33947590e-01 -1.22378480e+00 -4.48853880e-01 -3.24659087e-02
7.51781344e-01 -2.18953371e-01 3.12684849e-02 1.11045587e+00
8.94760966e-01 -1.31923699e+00 7.65935421e-01 1.02769181e-01
8.19304764e-01 -1.16856599e+00 6.36130273e-01 6.94576383e-01
-1.03118110e+00 -1.01299964e-01 -8.71844739e-02 -2.68110693e-01
-7.30200857e-02 -1.85417920e-01 -1.02596378e+00 4.00077432e-01
7.34701753e-01 1.07544675e-01 -5.55106163e-01 9.54656422e-01
-4.12973702e-01 7.64671445e-01 -1.19182803e-01 -7.62922913e-02
3.94446433e-01 -1.47717878e-01 2.37697482e-01 7.89255917e-01
2.08945841e-01 1.36133209e-02 3.89534235e-01 8.32857788e-01
1.64022133e-01 -2.20076203e-01 -8.44217181e-01 1.71910688e-01
3.75953197e-01 8.71735215e-01 -5.30536950e-01 -4.04649556e-01
4.64101061e-02 7.33352542e-01 2.44389534e-01 4.99190152e-01
-1.40555894e+00 -6.58259749e-01 8.61906052e-01 2.15751916e-01
5.95634282e-02 -4.25121695e-01 -1.76599637e-01 -5.67736447e-01
-2.13025257e-01 -1.17989385e+00 4.07281071e-02 -7.58791029e-01
-6.34503007e-01 4.85834658e-01 2.03747466e-01 -1.41016042e+00
-6.11856766e-02 -5.78594685e-01 -8.72579575e-01 8.30696404e-01
-1.44272101e+00 -4.79826450e-01 -6.01068854e-01 5.76450706e-01
5.25239646e-01 -4.84997593e-02 6.86017811e-01 2.15688229e-01
-6.92803979e-01 7.77632356e-01 -5.94388023e-02 1.12713717e-01
7.92218089e-01 -1.34782755e+00 5.33464253e-01 7.30081737e-01
-1.62266418e-01 5.17076135e-01 1.05565906e+00 -3.83592010e-01
-1.37107205e+00 -1.01410460e+00 3.13742399e-01 -4.90676314e-01
5.16861081e-01 -3.83311696e-02 -9.94349837e-01 5.05863965e-01
-4.01185229e-02 -1.25606865e-01 4.65772718e-01 -1.91740766e-02
-3.78332764e-01 -2.98768699e-01 -1.20554614e+00 1.00325215e+00
6.16366744e-01 -3.30486804e-01 -4.50069726e-01 1.95363220e-02
6.25217140e-01 -9.68426347e-01 -7.01782107e-01 2.18190238e-01
5.00279486e-01 -8.80342484e-01 6.90719724e-01 -9.49968994e-01
4.01859283e-01 -6.36382520e-01 1.17397726e-01 -1.72226596e+00
2.90865302e-02 -6.49809957e-01 2.89030790e-01 6.73399270e-01
4.69670951e-01 -5.44237912e-01 8.69939983e-01 5.10770082e-01
-1.43222168e-01 -5.80528319e-01 -6.14805222e-01 -1.04055190e+00
5.08677185e-01 -5.98746419e-01 8.23241591e-01 6.92154825e-01
6.01333287e-03 5.04183061e-02 -1.43097192e-01 2.86168903e-01
3.98171633e-01 7.77662545e-02 1.43730664e+00 -9.54859436e-01
-2.11005300e-01 -2.66669273e-01 -1.24263883e-01 -1.14606953e+00
2.48033285e-01 -6.12892509e-01 3.60980749e-01 -9.26583946e-01
-2.12154418e-01 -7.74464011e-01 1.21262250e-02 3.41124296e-01
-1.35018811e-01 -1.14242315e-01 3.63284200e-01 -4.09527302e-01
-2.09375665e-01 3.23171556e-01 1.56483042e+00 1.15376495e-01
-4.32697058e-01 1.24050520e-01 -4.36945826e-01 5.60243487e-01
1.12917483e+00 -7.01707542e-01 -5.55392861e-01 -3.04576308e-01
6.80036992e-02 2.77284086e-01 3.57551187e-01 -1.51028168e+00
1.04209498e-01 -5.79504192e-01 3.78005505e-01 -2.99781889e-01
2.35719651e-01 -8.72748733e-01 7.40760341e-02 7.86319613e-01
-3.57368231e-01 4.13277715e-01 7.95223594e-01 3.43180776e-01
-4.34802845e-02 -2.71638542e-01 1.03315496e+00 -4.01539654e-02
-6.64049983e-01 -1.01557299e-01 -6.01651788e-01 3.48024905e-01
1.39251423e+00 -4.34453577e-01 -2.64744043e-01 -3.59050125e-01
-5.50396264e-01 4.67988789e-01 6.36260450e-01 6.12856627e-01
2.65629768e-01 -1.16265655e+00 -4.33446795e-01 4.40269530e-01
4.71406728e-02 6.32017776e-02 3.47855389e-01 6.34521365e-01
-8.97820950e-01 5.60836904e-02 -5.20821214e-01 -6.80773139e-01
-1.05548060e+00 6.10260367e-01 6.35402858e-01 -2.65424192e-01
-6.51552677e-01 5.18516541e-01 -4.35583368e-02 -8.83863568e-01
2.81870216e-01 -3.37289155e-01 1.23263560e-01 -5.29902816e-01
5.38661242e-01 3.77238303e-01 -2.55615972e-02 -2.52111256e-01
-1.61027819e-01 5.22085011e-01 7.24152699e-02 -2.93142647e-01
8.38463187e-01 4.19290364e-01 6.53928280e-01 3.43320847e-01
8.60602677e-01 5.67929968e-02 -1.62738764e+00 1.17831945e-01
-1.66854113e-01 -5.94325483e-01 -1.33994073e-01 -7.96276391e-01
-9.05804992e-01 8.74460816e-01 7.70501137e-01 -2.22368985e-02
7.69038856e-01 -4.21336561e-01 7.20845222e-01 3.97419274e-01
3.80051255e-01 -1.25190854e+00 2.39397004e-01 5.75857580e-01
4.94001329e-01 -1.26520026e+00 -2.50822216e-01 1.61217544e-02
-9.13203418e-01 1.10671270e+00 1.27795088e+00 -3.37649494e-01
2.65323043e-01 3.93332303e-01 2.81720102e-01 1.76071224e-03
-4.37505811e-01 2.73841284e-02 -1.87489241e-01 6.61890388e-01
1.16319455e-01 3.20236862e-01 2.19358653e-01 1.37997437e-02
-6.20201409e-01 -1.71724316e-02 8.03803802e-01 1.05442369e+00
-5.17490149e-01 -1.16040015e+00 -7.04756618e-01 -1.71341226e-01
1.29574254e-01 6.22089505e-01 1.13651022e-01 1.20421720e+00
1.51605070e-01 9.83163238e-01 2.67077565e-01 -5.82504630e-01
8.01078975e-01 6.07856587e-02 3.84942710e-01 -4.89343256e-01
-7.95293391e-01 -5.20614266e-01 -5.76775074e-02 -8.77347946e-01
6.80356473e-02 -3.00708026e-01 -1.54317927e+00 -8.62684369e-01
7.78329838e-03 3.51904839e-01 6.00610733e-01 8.46781373e-01
8.15588385e-02 7.70957947e-01 7.88872242e-01 -6.77832782e-01
-6.68018579e-01 -5.14642894e-01 -4.02493089e-01 3.59620899e-01
4.04871702e-01 -7.27001369e-01 -2.29752377e-01 -1.74184367e-01] | [4.840635299682617, 1.272264838218689] |
bdffecfe-5044-4a6f-9363-4e954b426360 | accurate-3d-facial-geometry-prediction-by | 2104.08403 | null | https://arxiv.org/abs/2104.08403v1 | https://arxiv.org/pdf/2104.08403v1.pdf | Accurate 3D Facial Geometry Prediction by Multi-Task, Multi-Modal, and Multi-Representation Landmark Refinement Network | This work focuses on complete 3D facial geometry prediction, including 3D facial alignment via 3D face modeling and face orientation estimation using the proposed multi-task, multi-modal, and multi-representation landmark refinement network (M$^3$-LRN). Our focus is on the important facial attributes, 3D landmarks, and we fully utilize their embedded information to guide 3D facial geometry learning. We first propose a multi-modal and multi-representation feature aggregation for landmark refinement. Next, we are the first to study 3DMM regression from sparse 3D landmarks and utilize multi-representation advantage to attain better geometry prediction. We attain the state of the art from extensive experiments on all tasks of learning 3D facial geometry. We closely validate contributions of each modality and representation. Our results are robust across cropped faces, underwater scenarios, and extreme poses. Specially we adopt only simple and widely used network operations in M$^3$-LRN and attain a near 20\% improvement on face orientation estimation over the current best performance. See our project page here. | ['Ulrich Neumann', 'Qiangeng Xu', 'Cho-Ying Wu'] | 2021-04-16 | null | null | null | null | ['3d-face-modeling'] | ['computer-vision'] | [-2.33890638e-01 3.22997510e-01 -6.15412556e-02 -5.76265633e-01
-9.62852180e-01 -2.09330454e-01 1.42729536e-01 -4.06785578e-01
-3.03207114e-02 7.73176625e-02 2.87805021e-01 5.36238179e-02
-1.60493925e-01 -4.29086626e-01 -7.78240621e-01 -4.17746216e-01
-3.73533010e-01 5.67026317e-01 -4.62494999e-01 -2.65298605e-01
3.22359055e-01 1.17331648e+00 -1.59214354e+00 -1.08753733e-01
3.55307549e-01 1.18159676e+00 -3.72554034e-01 3.80680799e-01
4.54576239e-02 1.21716082e-01 1.95593629e-02 -4.40259039e-01
6.96922243e-01 5.73694818e-02 -6.80035710e-01 1.60299182e-01
1.14933515e+00 -4.20585096e-01 1.18966410e-02 7.09832013e-01
9.67648327e-01 3.00612655e-02 9.47068989e-01 -1.31106639e+00
-3.62669230e-01 1.15675390e-01 -1.24715638e+00 -3.22129548e-01
5.21837354e-01 -4.40710127e-01 7.42727339e-01 -1.66978717e+00
6.11391366e-01 1.63349652e+00 1.11315811e+00 6.64572775e-01
-8.00875247e-01 -1.05389583e+00 4.67009693e-01 3.38719087e-03
-2.01667976e+00 -1.09043908e+00 9.08394754e-01 -2.92775691e-01
8.29328656e-01 -1.71574056e-01 3.85742664e-01 5.10649443e-01
-1.24991670e-01 4.44151372e-01 7.93285072e-01 -3.62227529e-01
-2.73575395e-01 -3.25314522e-01 -4.53529775e-01 1.57364738e+00
-6.04033619e-02 -1.41039982e-01 -7.93836594e-01 -1.76824629e-01
1.08936787e+00 3.11685167e-02 2.73791887e-02 -2.98556805e-01
-5.26400149e-01 6.18031561e-01 2.43299752e-01 -2.38169357e-01
-3.16421151e-01 2.06076831e-01 -1.04639217e-01 1.00221254e-01
7.62635648e-01 -9.58957672e-02 -6.76530957e-01 6.66687936e-02
-9.55721855e-01 2.56704926e-01 5.48719168e-01 1.29477310e+00
1.16478086e+00 2.91194022e-01 3.57237101e-01 1.09588683e+00
8.20989072e-01 9.71627414e-01 -5.12409247e-02 -1.33517468e+00
3.04470390e-01 4.90602463e-01 -3.10732007e-01 -1.21619725e+00
-7.50719726e-01 -2.93798018e-02 -7.37275004e-01 1.60256028e-01
9.43130404e-02 -3.41157198e-01 -9.56426144e-01 1.69096565e+00
7.61012316e-01 3.70739847e-01 -1.55770868e-01 7.08446681e-01
1.22003233e+00 2.00237915e-01 -1.36192381e-01 -2.14318320e-01
1.24434018e+00 -7.17228234e-01 -2.18804911e-01 1.59518085e-02
7.51359761e-01 -9.03403044e-01 4.13086295e-01 6.22832514e-02
-1.27450097e+00 -4.66258645e-01 -5.34472942e-01 -1.59538820e-01
-4.92957793e-03 1.74114734e-01 6.68542743e-01 6.08320355e-01
-1.28504956e+00 3.58120501e-01 -6.83460534e-01 -2.66682416e-01
7.44731903e-01 7.94933975e-01 -8.75987172e-01 -3.56334597e-01
-5.36844254e-01 8.66503417e-01 -4.37860668e-01 3.16119969e-01
-9.64511693e-01 -9.22635436e-01 -1.24932170e+00 -4.54552859e-01
2.94115514e-01 -5.48134923e-01 9.52587306e-01 -3.84694964e-01
-1.44903076e+00 1.02444518e+00 -7.57754087e-01 3.20763499e-01
7.48016685e-02 -2.47961402e-01 -1.86637357e-01 1.91515505e-01
4.57866816e-03 9.27197337e-01 9.05194402e-01 -1.33267617e+00
-3.46296906e-01 -1.00025952e+00 -2.27146134e-01 4.53501791e-01
-1.80597857e-01 9.38363001e-02 -7.75915861e-01 -4.48794633e-01
7.25138485e-01 -9.36204731e-01 -2.74473369e-01 3.78792048e-01
-1.71185344e-01 -2.76987106e-01 7.10541904e-01 -6.78448260e-01
7.11382389e-01 -1.93794072e+00 1.84923112e-01 4.90098894e-01
2.60023713e-01 -1.31683245e-01 -4.61219907e-01 5.41981123e-02
4.56541777e-02 2.28276089e-01 1.46020502e-01 -1.08971286e+00
-4.48650122e-02 -5.15141301e-02 2.26277739e-01 8.48679781e-01
2.18897715e-01 7.04513490e-01 -3.49801898e-01 -7.95684814e-01
2.30054334e-02 9.04025972e-01 -9.99843419e-01 3.48376222e-02
1.62706926e-01 4.32241201e-01 -6.22229218e-01 1.46200085e+00
9.96330023e-01 -1.18626773e-01 -1.85569599e-02 -6.25348866e-01
2.19666250e-02 -4.51398045e-01 -1.34382570e+00 2.07057667e+00
-6.90270424e-01 9.44846720e-02 3.80887866e-01 -6.08510613e-01
1.17154503e+00 3.32926512e-01 9.05712962e-01 -5.32982647e-01
1.03675991e-01 2.18257040e-01 -5.45496404e-01 -3.56928825e-01
3.38115811e-01 -8.17624182e-02 2.21637502e-01 4.61285383e-01
3.89966726e-01 -1.54912412e-01 -4.48604703e-01 -1.15441278e-01
4.64112103e-01 6.17637038e-01 2.37750784e-01 -2.89895296e-01
4.83896136e-01 -6.39815986e-01 6.72335505e-01 1.13786377e-01
-7.84852505e-02 8.59080791e-01 3.68234575e-01 -5.21742284e-01
-8.26247633e-01 -7.75946617e-01 -3.38824958e-01 1.34674013e+00
-2.75813392e-03 -5.01085877e-01 -4.58012134e-01 -6.59483969e-01
3.45080584e-01 -1.12600960e-01 -6.94637656e-01 2.16307625e-01
-7.30856538e-01 -7.86147654e-01 5.97974241e-01 5.35232663e-01
2.50275701e-01 -4.66188639e-01 1.14618398e-01 -1.73352763e-01
5.84470704e-02 -1.25580609e+00 -5.48491538e-01 -3.07163417e-01
-9.51121271e-01 -1.11197650e+00 -7.04364538e-01 -8.98046911e-01
1.08043170e+00 2.48429060e-01 7.87301540e-01 2.33382210e-01
-3.18479419e-01 7.28067875e-01 -1.70142710e-01 -4.52619106e-01
3.14166963e-01 -7.78011307e-02 5.91871798e-01 1.51286259e-01
3.05434495e-01 -8.65379035e-01 -6.36348665e-01 5.41671634e-01
-1.10762991e-01 -2.99987584e-01 6.05447888e-01 2.82695472e-01
7.99122155e-01 -4.22751725e-01 3.80755067e-01 -6.41765535e-01
-7.94586912e-02 -3.34060937e-01 -4.12743032e-01 2.12663040e-01
-5.05573928e-01 -4.55369130e-02 1.90663695e-01 8.78266897e-03
-8.27809155e-01 3.72398317e-01 -5.10300875e-01 -6.57207906e-01
-2.16811478e-01 1.61333263e-01 -4.09612715e-01 -8.80134225e-01
3.54870647e-01 1.49370739e-02 2.93434113e-01 -6.26700461e-01
4.49656665e-01 4.49041069e-01 4.19211052e-02 -8.32635880e-01
9.07963932e-01 6.76401436e-01 6.28978848e-01 -1.14356244e+00
-8.89525294e-01 -2.06697747e-01 -1.05112338e+00 -3.11502337e-01
5.18474936e-01 -1.38163638e+00 -1.10448444e+00 4.64739531e-01
-1.10231483e+00 2.58219391e-02 1.93951979e-01 2.89074004e-01
-5.56853950e-01 4.93674397e-01 -3.55939686e-01 -9.49416637e-01
-4.32076812e-01 -1.17351961e+00 1.67866838e+00 1.43659562e-01
5.48153259e-02 -9.47741151e-01 -1.17783837e-01 4.44079161e-01
1.39936000e-01 3.34465295e-01 6.73613131e-01 -1.94232196e-01
-4.12281901e-01 -3.30008864e-02 -2.46141449e-01 -2.21024677e-02
1.80849567e-01 1.64690360e-01 -1.11055470e+00 -3.19988549e-01
-4.27057445e-01 -4.74587649e-01 6.75414741e-01 5.64597905e-01
1.25627768e+00 -6.39258772e-02 -2.49608949e-01 1.31614876e+00
1.18010581e+00 -1.39054075e-01 2.80916154e-01 -1.12324908e-01
1.10265446e+00 9.14107800e-01 5.61264455e-01 7.61101723e-01
9.32359517e-01 6.25219464e-01 5.65772891e-01 -1.99297413e-01
-8.36971775e-02 -3.12339395e-01 2.60144621e-01 7.87226558e-01
-7.63336301e-01 3.22166562e-01 -8.31773400e-01 2.19693452e-01
-1.42643297e+00 -6.02757990e-01 2.94031769e-01 1.88750899e+00
4.47642475e-01 -4.76863891e-01 2.03421891e-01 -2.94592291e-01
5.38035035e-01 2.92912275e-01 -3.95649254e-01 -8.56558383e-02
-3.95879894e-02 3.97725344e-01 3.91299754e-01 7.30437696e-01
-1.06305945e+00 1.22059906e+00 6.62334347e+00 7.31181443e-01
-8.65119815e-01 -4.95208055e-02 6.73314035e-01 -2.34389290e-01
-3.54889482e-01 -2.51066625e-01 -1.46897578e+00 -1.80065215e-01
5.38951576e-01 2.50504017e-01 1.56886026e-01 8.32268834e-01
1.38332844e-01 2.28967428e-01 -1.03953743e+00 1.48519325e+00
7.17094660e-01 -1.13350642e+00 3.15048009e-01 3.25961798e-01
7.02854931e-01 -9.60094705e-02 1.89879477e-01 1.29833296e-02
7.47975931e-02 -1.37982583e+00 5.29324949e-01 5.99037349e-01
1.29294407e+00 -9.38069105e-01 4.30422634e-01 -8.55685398e-02
-1.56486738e+00 8.48223194e-02 -4.65330511e-01 2.14507401e-01
2.56242931e-01 3.52492094e-01 -5.38911581e-01 6.06127858e-01
7.51996756e-01 9.48025644e-01 -4.76073861e-01 6.75442934e-01
1.61381140e-01 -4.47672419e-02 -5.93554497e-01 3.68871778e-01
-1.21610880e-01 -1.74394429e-01 2.23693252e-01 8.79536390e-01
7.59161115e-01 3.99304897e-01 2.21626922e-01 2.27676362e-01
-3.02155644e-01 3.78073961e-01 -8.08650792e-01 4.32188243e-01
5.16148090e-01 1.31173110e+00 -4.53347981e-01 1.74519509e-01
-5.37901103e-01 6.20395482e-01 4.03244436e-01 2.08397880e-01
-5.92030168e-01 5.26056252e-02 1.13194335e+00 7.72399232e-02
2.31365442e-01 -4.72124547e-01 -1.55433372e-01 -9.39862370e-01
-7.04296827e-02 -6.45291686e-01 1.24682419e-01 -4.67732966e-01
-1.07699442e+00 6.04313672e-01 2.46565267e-02 -9.76154149e-01
-1.20898433e-01 -7.88615346e-01 -2.16017604e-01 8.25584650e-01
-1.85465515e+00 -1.74845791e+00 -3.76880199e-01 8.94980431e-01
3.14310223e-01 -3.98890674e-01 9.76852834e-01 4.48578686e-01
-5.88325620e-01 1.07587564e+00 -3.25381666e-01 2.83780843e-01
7.80822635e-01 -7.29221284e-01 2.79974580e-01 4.54406261e-01
2.14529023e-01 4.83423620e-01 6.93118423e-02 -6.52381659e-01
-1.88750076e+00 -1.15128410e+00 7.09867835e-01 -5.55838764e-01
4.71230932e-02 -4.06697452e-01 -3.79026383e-01 8.82938862e-01
-4.71094489e-01 3.65008295e-01 9.84704435e-01 5.14892697e-01
-5.48585296e-01 -3.17569971e-01 -1.26912403e+00 5.14635324e-01
1.70462418e+00 -4.52784091e-01 7.95182511e-02 3.31971079e-01
3.93936098e-01 -6.01430535e-01 -1.32449603e+00 7.60247469e-01
9.49001789e-01 -7.50765920e-01 1.16817057e+00 -5.58379471e-01
2.20576420e-01 -1.30592331e-01 -5.74664056e-01 -9.65225995e-01
-2.57379025e-01 -5.80550551e-01 -1.77825034e-01 1.07942927e+00
5.37125826e-01 -3.12957615e-01 1.27686274e+00 4.30021256e-01
-2.10308865e-01 -9.89155412e-01 -1.11988449e+00 -2.43428335e-01
2.62918472e-02 -5.26796222e-01 8.99739742e-01 8.51379395e-01
-4.12783623e-01 -7.64918886e-03 -6.39916837e-01 5.99395037e-01
1.01955116e+00 2.60903716e-01 9.53379869e-01 -1.27513361e+00
3.24534655e-01 -1.92720696e-01 -5.48662007e-01 -1.43880510e+00
5.54697156e-01 -8.50235462e-01 -1.80335164e-01 -1.12020314e+00
2.49179211e-02 -5.78823864e-01 9.10613537e-02 7.43023396e-01
1.58843040e-01 8.20903361e-01 1.15422823e-01 -3.47508416e-02
-5.98275542e-01 8.28166127e-01 1.45503426e+00 1.98946029e-01
-8.80429894e-02 -4.04427946e-02 -9.38971698e-01 1.00990880e+00
3.54837507e-01 -7.55567402e-02 -2.04738215e-01 -8.04543436e-01
7.54753873e-02 1.07575022e-01 1.59394845e-01 -7.08420694e-01
1.98336840e-01 -2.23854706e-01 7.82097638e-01 -5.66265643e-01
8.27697635e-01 -8.92675519e-01 -1.51305944e-01 -1.12305157e-01
1.25981033e-01 2.17898995e-01 1.99203923e-01 2.89779425e-01
1.13367988e-02 1.59522519e-01 8.61108243e-01 -2.78956205e-01
-9.16111469e-01 1.17464221e+00 3.09736639e-01 -9.19222459e-02
9.32165504e-01 -5.28256476e-01 8.29251111e-02 -4.17024225e-01
-1.07295930e+00 3.50518733e-01 4.60756332e-01 3.41527194e-01
1.11206663e+00 -1.52210402e+00 -9.44418490e-01 6.17418706e-01
3.15863416e-02 3.07514310e-01 4.72904712e-01 7.85832644e-01
-5.53704798e-01 1.18659467e-01 -2.37723887e-01 -7.49414325e-01
-1.41091490e+00 -1.34764403e-01 6.25409365e-01 4.50770259e-01
-3.57478201e-01 1.23045731e+00 2.67728139e-02 -7.76296258e-01
3.50703388e-01 1.72122106e-01 -4.52817708e-01 2.17385188e-01
4.49467868e-01 4.72049147e-01 1.18256763e-01 -1.29032278e+00
-5.92412829e-01 1.77274072e+00 -3.44195752e-03 2.00782999e-01
1.65777445e+00 -3.55951726e-01 -1.81925595e-01 -8.44223797e-02
1.45864129e+00 9.72546637e-02 -1.33474040e+00 -1.84887543e-01
-5.35691679e-01 -5.36124527e-01 -1.00372151e-01 -2.39316031e-01
-1.49886680e+00 7.49049723e-01 5.58659256e-01 -7.91958749e-01
1.11528242e+00 1.66993365e-01 4.86435801e-01 3.54459226e-01
7.33091116e-01 -7.94198096e-01 6.87365979e-02 6.76033676e-01
9.83421147e-01 -1.37626195e+00 3.53045315e-01 -6.93410099e-01
-3.73877496e-01 1.12518442e+00 9.18840706e-01 -8.76249894e-02
1.39142823e+00 1.53554901e-01 2.60724813e-01 -5.26774764e-01
-3.48698795e-01 -2.15797931e-01 4.23815548e-01 7.25674689e-01
5.35865963e-01 -3.97339761e-01 3.69465172e-01 1.12294182e-01
-2.66410291e-01 -3.85117680e-01 5.64174987e-02 7.40357935e-01
-3.08854252e-01 -1.07861841e+00 -4.19020861e-01 3.94800156e-01
-4.58516657e-01 5.14314473e-02 -1.04726695e-01 8.44421744e-01
1.22009084e-01 6.13287807e-01 1.03905000e-01 -5.26517212e-01
4.03878659e-01 8.98411646e-02 8.50779057e-01 -6.56333327e-01
-9.38416366e-03 2.52646625e-01 -1.38590723e-01 -9.32915092e-01
-7.19169676e-01 -8.21708918e-01 -1.19437814e+00 -5.43121457e-01
-1.96095109e-01 -2.58908439e-02 8.23207080e-01 7.92413414e-01
6.52069986e-01 -1.33046776e-01 9.79011238e-01 -1.46910143e+00
-2.04317212e-01 -8.71994138e-01 -8.95780921e-01 -2.70508993e-02
4.00352150e-01 -1.00615323e+00 -3.22018862e-01 -1.54446676e-01] | [13.318603515625, 0.1651419848203659] |
527a7cbe-9561-471e-8710-4a69ab96c3a3 | a-plug-and-play-priors-framework-for | 2012.13074 | null | https://arxiv.org/abs/2012.13074v3 | https://arxiv.org/pdf/2012.13074v3.pdf | A Plug-and-Play Priors Framework for Hyperspectral Unmixing | Spectral unmixing is a widely used technique in hyperspectral image processing and analysis. It aims to separate mixed pixels into the component materials and their corresponding abundances. Early solutions to spectral unmixing are performed independently on each pixel. Nowadays, investigating proper priors into the unmixing problem has been popular as it can significantly enhance the unmixing performance. However, it is non-trivial to handcraft a powerful regularizer, and complex regularizers may introduce extra difficulties in solving optimization problems in which they are involved. To address this issue, we present a plug-and-play (PnP) priors framework for hyperspectral unmixing. More specifically, we use the alternating direction method of multipliers (ADMM) to decompose the optimization problem into two iterative subproblems. One is a regular optimization problem depending on the forward model, and the other is a proximity operator related to the prior model and can be regarded as an image denoising problem. Our framework is flexible and extendable which allows a wide range of denoisers to replace prior models and avoids handcrafting regularizers. Experiments conducted on both synthetic data and real airborne data illustrate the superiority of the proposed strategy compared with other state-of-the-art hyperspectral unmixing methods. | ['Wei Chen', 'Jie Chen', 'Xiuheng Wang', 'Min Zhao'] | 2020-12-24 | null | null | null | null | ['hyperspectral-unmixing'] | ['computer-vision'] | [ 7.93213129e-01 -5.44232905e-01 7.39304423e-02 7.23096728e-02
-4.15074527e-01 -4.05200094e-01 4.42326993e-01 -2.15115711e-01
-3.00217986e-01 7.14274704e-01 -4.05535363e-02 -1.34391516e-01
-3.73542905e-01 -7.37203419e-01 -3.56841981e-01 -1.42695868e+00
5.46896219e-01 1.24912463e-01 -3.43505740e-01 -2.75505781e-01
-1.00113906e-01 4.80184913e-01 -1.48836207e+00 -7.00378269e-02
1.60144532e+00 8.40337694e-01 6.12235606e-01 2.76909232e-01
-9.30984542e-02 5.14458597e-01 -2.46360898e-01 1.22968834e-02
6.13058388e-01 -5.43335795e-01 -4.89991486e-01 8.21395278e-01
2.26051524e-01 -8.18163902e-02 -1.90230086e-01 1.64457798e+00
4.00160909e-01 3.15967888e-01 5.17830372e-01 -1.27381420e+00
-3.97968233e-01 3.40213418e-01 -1.14016151e+00 -8.71425122e-02
-3.22342306e-01 1.36574525e-02 7.67370284e-01 -6.90205932e-01
4.81001586e-02 9.76713717e-01 5.84900677e-01 -1.20426342e-02
-1.37997842e+00 -5.80197096e-01 2.00565070e-01 1.85359538e-01
-1.44540095e+00 -2.87526846e-01 9.36034918e-01 -5.95515072e-01
3.10987651e-01 5.77796280e-01 7.63083041e-01 5.85964739e-01
-2.42560655e-01 6.37148619e-01 1.37833798e+00 -4.23414111e-01
1.04529172e-01 -2.35895932e-01 1.29990742e-01 2.47937292e-01
5.57310343e-01 -2.91055869e-02 -1.29830554e-01 -1.99303851e-01
7.71589756e-01 2.36674428e-01 -8.01763833e-01 -3.34311187e-01
-1.16756237e+00 6.81948543e-01 4.34308439e-01 1.74658313e-01
-8.46553028e-01 -2.65942991e-01 -2.57914275e-01 9.72144678e-02
6.56726599e-01 3.43039095e-01 -1.95309401e-01 4.73700285e-01
-1.11442697e+00 1.86756626e-01 3.94793063e-01 5.38750589e-01
1.03081524e+00 2.50366122e-01 5.07477298e-02 1.21729410e+00
5.88849187e-01 7.06385314e-01 4.96009737e-01 -8.02882373e-01
3.80013883e-01 4.11979854e-01 3.55849385e-01 -1.02867520e+00
-1.04952417e-01 -7.42617786e-01 -1.37059379e+00 2.39531264e-01
2.48438850e-01 -2.10476369e-01 -8.90964329e-01 1.51638639e+00
4.73221391e-01 5.59328318e-01 -7.28758844e-03 1.07632899e+00
4.71883565e-01 1.20340991e+00 1.53063489e-02 -8.17935824e-01
1.31299925e+00 -1.08851171e+00 -1.04630792e+00 -6.56295300e-01
-2.04829909e-02 -1.13163304e+00 4.59074557e-01 5.13583660e-01
-9.40536618e-01 -4.80484843e-01 -1.25996006e+00 1.98987976e-01
-7.39060938e-02 3.15371454e-01 4.70968604e-01 4.98942018e-01
-7.32604623e-01 5.55372834e-01 -7.78306961e-01 -1.70120507e-01
1.21498615e-01 2.44733617e-01 -2.01396585e-01 -8.46694112e-02
-9.79383767e-01 8.35569918e-01 5.75284123e-01 7.43175983e-01
-6.31671846e-01 -4.76976842e-01 -6.86266720e-01 -3.86114344e-02
5.67966223e-01 -8.17398190e-01 8.87747228e-01 -1.45733106e+00
-1.68816972e+00 6.44075632e-01 -3.03981334e-01 8.63036886e-02
3.37341070e-01 -2.13151485e-01 -6.69809997e-01 -2.55185105e-02
9.80667323e-02 1.48404300e-01 1.44466805e+00 -1.37754846e+00
-5.75321078e-01 -2.36178577e-01 -3.33827943e-01 4.52912986e-01
-4.39681858e-01 -1.42466262e-01 -3.95313799e-01 -1.02641308e+00
6.88439369e-01 -9.62216735e-01 -3.97224247e-01 -1.08467191e-01
-2.90288419e-01 4.25965190e-01 7.07366586e-01 -1.11898088e+00
1.17683542e+00 -2.22502160e+00 7.13489175e-01 3.02229941e-01
4.82403226e-02 5.29487967e-01 -2.32331485e-01 3.57608587e-01
-5.92675865e-01 -2.42916077e-01 -8.25082719e-01 -1.45386711e-01
-1.64786413e-01 2.63332456e-01 -1.23881601e-01 7.49512553e-01
9.39276814e-03 3.41868907e-01 -8.47923219e-01 -7.51013607e-02
2.60332465e-01 5.30163348e-01 -2.27500856e-01 3.97465080e-01
-1.28906995e-01 6.11124694e-01 -2.00830564e-01 6.10949874e-01
1.23852134e+00 -1.58992067e-01 3.37347776e-01 -6.68147862e-01
-4.17467058e-01 -2.15262398e-01 -1.82701945e+00 1.45704997e+00
-2.24369884e-01 2.18909621e-01 8.98670793e-01 -1.33337641e+00
6.58350289e-01 3.73897403e-01 5.03153026e-01 -7.71894827e-02
4.40623425e-02 4.84326422e-01 -6.52514398e-02 -5.74263692e-01
3.83011222e-01 -7.06678867e-01 7.00677931e-01 1.57090500e-01
-2.11991623e-01 -4.21096623e-01 2.83631057e-01 -4.13596988e-01
3.68492872e-01 7.84051940e-02 5.48748851e-01 -4.73354667e-01
8.93236458e-01 -3.79942283e-02 7.72192657e-01 3.31943482e-01
1.05169863e-01 4.70500827e-01 6.01535030e-02 -8.67213830e-02
-7.05422699e-01 -7.04461515e-01 -2.29668170e-01 6.89507902e-01
2.78536797e-01 4.47377302e-02 -5.58463752e-01 -8.27727690e-02
-2.46882230e-01 4.05974209e-01 -2.23956496e-01 1.96826935e-01
-2.36198485e-01 -1.66067564e+00 -2.39529237e-02 7.59094730e-02
9.27991450e-01 -6.03980780e-01 6.12603594e-03 2.99639225e-01
-5.91438472e-01 -8.25162828e-01 -3.66234303e-01 1.82007760e-01
-1.05903828e+00 -9.60561037e-01 -1.05673766e+00 -7.44356155e-01
8.30689192e-01 8.74742150e-01 6.36911035e-01 -1.41547859e-01
7.60156736e-02 -1.75781492e-02 -2.72346318e-01 -3.17012131e-01
-3.53003561e-01 -2.43021414e-01 -3.83277647e-02 8.25777888e-01
-1.92070100e-02 -7.69927084e-01 -4.44315016e-01 3.76442850e-01
-1.47010803e+00 2.36029774e-01 7.44160533e-01 8.31304610e-01
7.53720403e-01 6.84505999e-01 5.94591536e-02 -7.01687038e-01
5.28143048e-01 -5.56373358e-01 -8.04976583e-01 4.05303776e-01
-5.01741469e-01 -1.86824575e-02 4.04151320e-01 -4.23909783e-01
-1.21024513e+00 3.29352379e-01 -2.42187618e-03 -3.35933685e-01
2.17782427e-02 9.00176764e-01 -7.29415238e-01 -3.14254105e-01
3.85540813e-01 4.17895347e-01 2.81982392e-01 -6.77802026e-01
4.65558171e-01 7.66149819e-01 5.58092833e-01 -4.44375336e-01
1.24859214e+00 5.84681213e-01 1.20292969e-01 -1.33084416e+00
-8.70209634e-01 -6.87293172e-01 -4.02030081e-01 -1.04240626e-01
8.11416745e-01 -1.11370182e+00 -2.09202796e-01 9.91255522e-01
-9.39098299e-01 -1.31762892e-01 -1.11205809e-01 5.14535427e-01
-1.51990995e-01 9.03298438e-01 -3.86316061e-01 -5.80306470e-01
-1.74194649e-01 -1.35082150e+00 6.35688484e-01 3.48552316e-01
2.77768344e-01 -9.66718078e-01 9.77395922e-02 5.07569909e-01
4.19577688e-01 2.90843427e-01 6.50860727e-01 6.31514117e-02
-4.50074971e-01 -8.81528035e-02 -2.57275194e-01 8.18489492e-01
5.71671844e-01 2.47852188e-02 -1.05719423e+00 -3.62847209e-01
6.52116776e-01 1.83353052e-01 9.72515821e-01 5.71907222e-01
9.73723173e-01 -3.87160480e-01 -1.30964130e-01 9.37745035e-01
1.61340773e+00 6.63013533e-02 8.05431306e-01 4.46957767e-01
8.61615777e-01 5.24823070e-01 3.98934007e-01 4.44846690e-01
-1.00723326e-01 5.20476699e-01 6.60487294e-01 -3.81252736e-01
2.87216127e-01 8.84635597e-02 2.03084469e-01 1.00396538e+00
-4.19713467e-01 -2.65822411e-01 -6.27378166e-01 2.43558168e-01
-1.99734592e+00 -9.27206874e-01 -5.05168617e-01 2.05087948e+00
8.99288476e-01 -5.85261762e-01 -1.97908938e-01 4.24069732e-01
9.94920969e-01 6.02523029e-01 -4.73243475e-01 3.62448782e-01
-4.39995646e-01 1.52496994e-01 5.60434043e-01 7.24491835e-01
-1.40025270e+00 5.67065835e-01 5.28961086e+00 9.81599808e-01
-1.31297863e+00 1.90913036e-01 2.92003840e-01 2.94056535e-01
-2.94190526e-01 1.73223764e-01 -2.77947605e-01 4.30864036e-01
3.06476444e-01 1.06779635e-01 9.55770910e-01 3.77915800e-01
4.33644354e-01 -9.20097083e-02 -4.90805477e-01 1.24576676e+00
-3.28118876e-02 -8.90839517e-01 1.01146422e-01 5.00652827e-02
9.13806081e-01 -2.52210915e-01 -9.81085002e-02 -2.76257455e-01
1.69340465e-02 -8.71693313e-01 6.08291030e-01 4.93979096e-01
3.77688408e-01 -4.39119726e-01 6.49690390e-01 5.03094375e-01
-1.11491644e+00 4.87372093e-02 -4.58463579e-01 -2.72460014e-01
2.55188346e-01 1.09831512e+00 -1.38525376e-02 9.44459438e-01
3.75780791e-01 9.65884805e-01 -1.49111152e-01 1.18530250e+00
-5.85904419e-01 4.61382657e-01 -3.44413877e-01 6.63345337e-01
2.06658334e-01 -1.09028161e+00 6.88377559e-01 7.77501643e-01
6.60242021e-01 5.30083954e-01 2.03162998e-01 7.41100669e-01
1.04192391e-01 2.47343302e-01 -1.43214434e-01 -2.31321320e-01
1.33544475e-01 1.48778844e+00 -6.14359379e-01 -2.43502319e-01
-3.85610998e-01 1.01946723e+00 -1.85874134e-01 7.41123855e-01
-6.36603892e-01 -2.08034217e-02 8.38185012e-01 6.82001486e-02
3.10135275e-01 -1.59557223e-01 -1.51443392e-01 -1.58015823e+00
1.97440516e-02 -1.41766572e+00 3.25313210e-01 -8.31737697e-01
-1.36836660e+00 4.52212632e-01 -1.08319543e-01 -1.47642374e+00
3.68451208e-01 -7.32981205e-01 -5.80699086e-01 1.21598840e+00
-1.90474844e+00 -1.10803258e+00 -7.82184482e-01 5.78658760e-01
1.88357621e-01 1.55544579e-01 6.28387570e-01 4.22422081e-01
-1.03781509e+00 -3.21541190e-01 5.09504795e-01 -3.80758017e-01
6.56220317e-01 -9.32581425e-01 -4.61938798e-01 1.52601218e+00
-2.10123733e-01 5.08603990e-01 8.11634958e-01 -5.86858332e-01
-1.15187979e+00 -1.14551473e+00 5.45706987e-01 3.53806734e-01
7.67993808e-01 3.39197010e-01 -9.94490981e-01 6.15885198e-01
4.45635349e-01 -9.20545030e-03 7.98104227e-01 -4.22071368e-01
7.51059651e-02 -4.12769586e-01 -9.09223437e-01 4.81949687e-01
5.99584997e-01 -2.13837281e-01 -4.45564955e-01 4.83793557e-01
2.03460917e-01 -3.17174226e-01 -7.44776189e-01 5.68292558e-01
1.77553311e-01 -8.68495166e-01 1.13017774e+00 -1.30198404e-01
3.10761988e-01 -9.35656309e-01 -2.12729752e-01 -1.61317992e+00
-6.18982494e-01 -9.63984489e-01 4.74600047e-02 1.27798676e+00
9.60056707e-02 -8.85755181e-01 2.99059302e-01 4.15468544e-01
-1.51081726e-01 -2.76080102e-01 -4.67202544e-01 -6.65621877e-01
-2.74015367e-01 -1.72089949e-01 6.33032858e-01 1.29345644e+00
-3.71100962e-01 1.93539783e-01 -7.19273388e-01 8.27849448e-01
9.18970525e-01 3.03933233e-01 5.03015459e-01 -1.24335599e+00
-4.98926878e-01 -5.10193408e-01 1.73745647e-01 -1.17857146e+00
1.34443000e-01 -7.59217203e-01 2.13146776e-01 -1.57729244e+00
-1.10431686e-02 -2.87672073e-01 -2.03302965e-01 4.04668719e-01
-5.70533395e-01 1.47478804e-01 3.99826020e-02 5.73784411e-01
2.10567966e-01 6.98107660e-01 1.26323748e+00 -5.68132281e-01
-3.30351591e-01 1.03727497e-01 -8.39151561e-01 8.29278648e-01
7.84165919e-01 -3.47323149e-01 -4.25087869e-01 -6.40494406e-01
2.62400955e-01 4.37871506e-03 1.99159175e-01 -8.96181047e-01
1.46266043e-01 -6.14479363e-01 2.13144589e-02 -2.89003640e-01
3.36556435e-01 -1.11415243e+00 8.36384356e-01 2.60536849e-01
2.74722308e-01 -3.43345046e-01 2.05262095e-01 4.36965317e-01
-5.60236871e-01 -6.20141208e-01 1.02520084e+00 -2.18271151e-01
-6.47650778e-01 3.26170564e-01 -3.64135146e-01 -3.21205437e-01
8.46481800e-01 -5.10692261e-02 -9.04093757e-02 -1.68485120e-01
-7.23847210e-01 8.51042494e-02 5.32819510e-01 -1.49264723e-01
3.46863508e-01 -1.18373621e+00 -8.56619179e-01 3.73305738e-01
-9.99320745e-02 1.23846352e-01 4.45534706e-01 1.08167791e+00
-8.96547556e-01 -5.73339313e-02 -1.20926626e-01 -4.72930789e-01
-1.11983275e+00 5.39724767e-01 5.62355518e-01 -2.53188848e-01
-1.19367287e-01 8.52672100e-01 2.07509294e-01 -2.28588432e-01
-8.56559202e-02 -3.31599206e-01 -2.80960649e-01 3.95815820e-01
4.73644316e-01 6.67925358e-01 4.16827612e-02 -8.12776744e-01
-1.56887844e-01 6.99675560e-01 5.08902252e-01 -5.30378474e-03
1.48509538e+00 -2.25257993e-01 -1.01612175e+00 9.62977111e-02
8.00156355e-01 3.65298726e-02 -1.25629961e+00 -3.10215265e-01
-4.15636390e-01 -4.16769654e-01 5.48184454e-01 -4.23493505e-01
-1.26030457e+00 7.86911249e-01 5.37330210e-01 4.36924666e-01
1.61978531e+00 -7.14052916e-01 5.94643593e-01 1.62846893e-01
-5.92947565e-02 -9.59131837e-01 -3.32831562e-01 2.37048075e-01
1.00699019e+00 -1.17920578e+00 4.15917039e-01 -8.97182524e-01
-2.47696728e-01 1.06905210e+00 2.03593865e-01 1.18066430e-01
7.40369439e-01 -1.20986529e-01 2.69689411e-01 8.86690393e-02
2.39585400e-01 -3.33961576e-01 4.37212348e-01 2.33519390e-01
2.38859296e-01 1.64815620e-01 -4.52617019e-01 4.00678873e-01
2.16075122e-01 -2.46075187e-02 3.23966831e-01 8.15670490e-01
-4.79267448e-01 -1.32596779e+00 -1.02621925e+00 2.05058649e-01
-2.34749049e-01 -2.53248394e-01 9.74846929e-02 4.05753881e-01
4.70577449e-01 1.21690106e+00 -4.19294149e-01 -8.77562258e-03
1.88846797e-01 6.38299361e-02 2.63367593e-01 -5.39210916e-01
-1.55491844e-01 5.54479182e-01 -1.89716369e-01 -1.18532479e-01
-1.09875345e+00 -6.60397351e-01 -6.59794867e-01 -9.27981287e-02
-7.15872884e-01 2.84437895e-01 4.63523716e-01 9.45032299e-01
3.13144363e-02 4.41208482e-01 6.33543730e-01 -1.23452032e+00
-6.19055331e-01 -9.80853438e-01 -1.05969238e+00 3.03882152e-01
3.74287903e-01 -6.01575017e-01 -5.59618235e-01 2.92768836e-01] | [10.10165786743164, -2.0761470794677734] |
c8d2c5f6-964c-4ac4-add4-4c4c4627f824 | towards-carbon-neutral-edge-computing | 2304.11374 | null | https://arxiv.org/abs/2304.11374v1 | https://arxiv.org/pdf/2304.11374v1.pdf | Towards Carbon-Neutral Edge Computing: Greening Edge AI by Harnessing Spot and Future Carbon Markets | Provisioning dynamic machine learning (ML) inference as a service for artificial intelligence (AI) applications of edge devices faces many challenges, including the trade-off among accuracy loss, carbon emission, and unknown future costs. Besides, many governments are launching carbon emission rights (CER) for operators to reduce carbon emissions further to reverse climate change. Facing these challenges, to achieve carbon-aware ML task offloading under limited carbon emission rights thus to achieve green edge AI, we establish a joint ML task offloading and CER purchasing problem, intending to minimize the accuracy loss under the long-term time-averaged cost budget of purchasing the required CER. However, considering the uncertainty of the resource prices, the CER purchasing prices, the carbon intensity of sites, and ML tasks' arrivals, it is hard to decide the optimal policy online over a long-running period time. To overcome this difficulty, we leverage the two-timescale Lyapunov optimization technique, of which the $T$-slot drift-plus-penalty methodology inspires us to propose an online algorithm that purchases CER in multiple timescales (on-preserved in carbon future market and on-demanded in the carbon spot market) and makes decisions about where to offload ML tasks. Considering the NP-hardness of the $T$-slot problems, we further propose the resource-restricted randomized dependent rounding algorithm to help to gain the near-optimal solution with no help of any future information. Our theoretical analysis and extensive simulation results driven by the real carbon intensity trace show the superior performance of the proposed algorithms. | ['Xu Chen', 'Xiaoxi Zhang', 'Zhi Zhou', 'Huirong Ma'] | 2023-04-22 | null | null | null | null | ['edge-computing'] | ['time-series'] | [ 2.44174123e-01 8.37868527e-02 -4.29386914e-01 3.77890207e-02
-6.59828067e-01 -5.10781527e-01 -1.95124701e-01 -1.42912254e-01
-3.31951052e-01 1.00331628e+00 -4.94850546e-01 -6.06496453e-01
-5.05700707e-01 -6.22007787e-01 -7.84591794e-01 -8.72995794e-01
-4.99233156e-02 6.03038192e-01 -4.17880625e-01 3.07925850e-01
-5.47366142e-02 9.70291793e-02 -9.94044185e-01 -4.00004864e-01
1.24091399e+00 1.55560207e+00 4.60170984e-01 2.84511656e-01
1.80802658e-01 1.42487381e-02 -1.51398763e-01 -2.88879126e-01
7.51147687e-01 -1.69820059e-02 -1.77068546e-01 -1.35314286e-01
-4.61495578e-01 -5.85490763e-01 2.88647324e-01 1.24890292e+00
3.47774386e-01 -1.82449311e-01 3.37158501e-01 -1.74811256e+00
-3.01484317e-01 6.18590653e-01 -8.69586766e-01 4.34162980e-03
-6.19256318e-01 2.23822966e-01 9.15115833e-01 -3.12119454e-01
2.47409329e-01 6.48795009e-01 3.63569498e-01 6.43394470e-01
-7.64897943e-01 -7.69500911e-01 5.49368501e-01 2.32110471e-01
-1.04710603e+00 -2.86650300e-01 6.32154942e-01 -3.13105792e-01
7.94954419e-01 3.50592434e-01 6.01122916e-01 4.27325219e-01
3.84650588e-01 5.66391945e-01 1.06146622e+00 -3.01558465e-01
6.20832920e-01 8.97275656e-02 -2.81127453e-01 2.08094448e-01
6.32867455e-01 1.07004963e-01 -2.87209064e-01 -1.46987393e-01
2.12107584e-01 1.93887465e-02 -2.44693071e-01 -8.86437073e-02
-9.81297612e-01 1.90629303e-01 3.13111812e-01 -1.45961896e-01
-7.36499310e-01 5.14261246e-01 5.20942986e-01 1.49557829e-01
6.99162185e-01 -1.44525245e-01 -9.24907565e-01 -7.27085024e-02
-8.40875387e-01 -3.44675444e-02 6.68911636e-01 1.09556496e+00
3.11592430e-01 9.89521593e-02 -1.22468948e-01 3.01416442e-02
1.32901564e-01 9.64693487e-01 -1.40674129e-01 -8.43922138e-01
9.11691904e-01 1.10403158e-01 7.49701798e-01 -6.12274408e-01
-2.86056280e-01 -6.55152321e-01 -8.49557757e-01 -3.64330448e-02
3.40780646e-01 -5.76939523e-01 -4.50132221e-01 1.83731115e+00
4.25626069e-01 2.13978261e-01 -3.20085198e-01 1.10144567e+00
-6.08248591e-01 7.47311294e-01 8.00067857e-02 -1.14112723e+00
1.23300469e+00 -6.12831235e-01 -8.52939308e-01 -3.55717428e-02
5.29663861e-01 -1.10039771e-01 8.58913004e-01 2.33160242e-01
-1.01383770e+00 2.57678539e-01 -9.98336613e-01 5.78568876e-01
-3.19729298e-01 8.57094601e-02 6.81967616e-01 8.71307194e-01
-5.84930003e-01 2.40985945e-01 -8.53113830e-01 1.05621584e-01
5.64461589e-01 3.98040205e-01 5.93598962e-01 -7.58651346e-02
-1.06785583e+00 5.16529560e-01 6.78878948e-02 5.64363122e-01
-1.07093096e+00 -1.07950127e+00 7.77076036e-02 1.98800161e-01
1.00300813e+00 -8.47997904e-01 1.15409887e+00 -8.10122013e-01
-1.18347299e+00 2.21616983e-01 1.14928797e-01 -5.70664406e-01
1.00525415e+00 -4.47281189e-02 -2.99728096e-01 -2.17450753e-01
2.81338934e-02 2.36762807e-01 8.31422269e-01 -8.33272278e-01
-9.60630655e-01 -8.53591740e-01 -2.73897350e-02 4.98693794e-01
-6.77488744e-01 -3.66886377e-01 -2.24094801e-02 -3.28000098e-01
-4.04718131e-01 -1.14725792e+00 -4.12999809e-01 -1.57072991e-01
-4.48438972e-01 -1.31730452e-01 7.56755531e-01 -6.79249108e-01
1.06748450e+00 -1.88109672e+00 -7.72705451e-02 -5.84777668e-02
-2.78698146e-01 -2.83661515e-01 1.93944424e-01 -2.99456753e-02
3.67692828e-01 2.87537575e-01 -2.71339446e-01 -1.80617601e-01
1.87845245e-01 1.07678011e-01 -3.43055964e-01 3.81896734e-01
-1.83578566e-01 5.75989187e-01 -7.46823549e-01 7.51867518e-02
-1.01781785e-01 -5.54181933e-02 -1.51225239e-01 -1.46251068e-01
-8.32154691e-01 2.21933261e-01 -7.72770584e-01 9.55762625e-01
8.26980710e-01 -1.51114613e-01 3.09016973e-01 -1.33535624e-01
-1.41580001e-01 -2.77307451e-01 -1.09485602e+00 1.22939336e+00
-7.41717041e-01 1.34259850e-01 6.68841898e-01 -1.03318918e+00
2.43087888e-01 1.01847596e-01 8.66250575e-01 -7.35614359e-01
-2.23923456e-02 5.30460477e-01 -3.90516222e-01 -3.74271721e-01
2.03099608e-01 -3.43431205e-01 -2.03961343e-01 5.49021900e-01
-9.94481802e-01 3.08715582e-01 -2.27715865e-01 2.68093869e-02
1.00889826e+00 4.24392410e-02 -1.79547727e-01 -5.20834506e-01
1.67390600e-01 7.34131336e-02 9.56976175e-01 2.31343299e-01
-3.54949802e-01 -3.33778441e-01 4.63654101e-01 -2.10089639e-01
-9.13555026e-01 -6.24106884e-01 5.23022152e-02 8.82497728e-01
2.66273290e-01 3.89347851e-01 -8.79308343e-01 -6.90958261e-01
2.04201967e-01 8.21078300e-01 -2.63449609e-01 8.36487338e-02
-3.29597950e-01 -1.13516283e+00 -1.94529638e-01 2.92936355e-01
4.66610342e-01 -4.47643071e-01 -1.16766906e+00 1.91675574e-01
-1.70155242e-01 -1.10766757e+00 -7.54135489e-01 3.59699517e-01
-7.34902620e-01 -7.57493615e-01 -7.35054433e-01 -3.71546447e-02
8.49339485e-01 3.39347243e-01 6.14249289e-01 -3.70986611e-01
-3.21212292e-01 2.11767957e-01 1.73644274e-01 -7.63064444e-01
2.22019002e-01 1.30149871e-01 3.13659877e-01 -7.91093241e-03
-8.16609561e-02 -2.89336950e-01 -6.87920749e-01 2.51318157e-01
-7.27647841e-01 5.32408990e-02 4.19855863e-01 5.05701482e-01
1.07964325e+00 9.02225375e-01 1.26937497e+00 -7.78067350e-01
5.88256061e-01 -6.27766430e-01 -1.21979880e+00 6.88678026e-01
-1.35147834e+00 -8.23324248e-02 8.40662718e-01 -2.33703300e-01
-9.92667198e-01 8.29571709e-02 7.62242019e-01 -4.43733811e-01
5.64949214e-01 4.04808491e-01 -6.54817820e-01 1.69689149e-01
-1.75819740e-01 -3.80976386e-02 -3.03391427e-01 -1.70606926e-01
4.56384122e-01 7.12226987e-01 4.59355235e-01 -7.29897618e-01
8.38691831e-01 4.29056793e-01 3.09315622e-01 -1.81622759e-01
-6.16269827e-01 -1.89290252e-02 1.34006431e-02 -6.39042556e-01
5.68798304e-01 -1.09164727e+00 -1.23254097e+00 3.06432933e-01
-8.30691218e-01 -2.37875491e-01 -1.48969606e-01 2.67537326e-01
-6.97453737e-01 1.45410329e-01 -4.37583998e-02 -1.43685770e+00
-7.41116822e-01 -8.21912706e-01 5.48028111e-01 2.38549128e-01
6.31821215e-01 -5.15971959e-01 -4.48614866e-01 5.56951582e-01
7.23454714e-01 3.89678091e-01 1.12622988e+00 1.34776443e-01
-1.09472406e+00 -1.27814868e-02 -4.25314397e-01 2.33656168e-01
5.99917248e-02 -2.70386487e-01 -4.92418438e-01 -7.03429759e-01
1.88248321e-01 6.05033189e-02 4.31261450e-01 7.27823019e-01
1.46398902e+00 -7.27616608e-01 -5.76478660e-01 4.64524627e-01
1.89360619e+00 3.65534097e-01 2.30416566e-01 3.29688698e-01
3.66688460e-01 5.29767156e-01 1.11815274e+00 9.32880282e-01
5.90688586e-01 5.48247218e-01 1.03461087e+00 3.62220615e-01
5.11590302e-01 -6.02295026e-02 3.65688264e-01 5.68409860e-01
4.14849445e-02 -4.49760735e-01 -7.18343318e-01 6.58960998e-01
-1.80950344e+00 -5.22448123e-01 1.39722824e-01 2.69217420e+00
6.19990289e-01 2.44411021e-01 2.30667233e-01 -5.38263209e-02
8.88440967e-01 -1.88156933e-01 -1.42631495e+00 -4.02838618e-01
1.08221821e-01 -5.18148541e-01 1.38127077e+00 -6.60874918e-02
-5.10981858e-01 3.84270787e-01 4.99187708e+00 9.71678495e-01
-1.07515955e+00 4.57774818e-01 1.10252202e+00 -5.07837653e-01
-6.56447470e-01 1.51494756e-01 -6.89112723e-01 9.73982215e-01
1.22208488e+00 -6.40049517e-01 1.15856004e+00 6.41366601e-01
8.59485447e-01 -2.51987010e-01 -8.60055983e-01 8.04105222e-01
-4.46471006e-01 -1.16501904e+00 -6.78493917e-01 3.10567528e-01
7.19316721e-01 1.89272299e-01 2.41027340e-01 8.26100335e-02
6.83292374e-02 -4.63500679e-01 8.93614769e-01 4.68056500e-01
1.10214138e+00 -1.13311636e+00 5.06079018e-01 6.49491370e-01
-1.36651576e+00 -7.86810279e-01 -2.32553974e-01 8.26057568e-02
5.22894084e-01 1.40734291e+00 -5.18279850e-01 7.38871455e-01
7.29997993e-01 1.65809050e-01 2.90347248e-01 7.16322541e-01
1.72860116e-01 2.92620540e-01 -5.96677125e-01 -6.10282719e-02
-1.66005582e-01 -5.04560590e-01 4.83955920e-01 5.52553594e-01
3.96592408e-01 2.65278339e-01 1.71794385e-01 1.01998627e+00
-3.02977532e-01 -8.26515555e-02 -2.44037315e-01 -5.50265312e-01
8.34414780e-01 1.29553926e+00 -8.25344503e-01 3.06636859e-02
-9.24205035e-02 7.67036319e-01 -2.50024885e-01 4.16179448e-01
-1.31068027e+00 -2.73074716e-01 5.11971474e-01 9.81796607e-02
7.41035789e-02 -2.49212384e-01 -8.80623043e-01 -7.94076800e-01
5.54421663e-01 -3.51851344e-01 4.05419171e-01 -5.95076263e-01
-1.04718864e+00 3.13853309e-03 -2.02212483e-01 -1.03290474e+00
2.08008647e-01 -2.12031633e-01 -4.76552337e-01 7.00471699e-01
-2.06985283e+00 -8.58553350e-01 3.16249672e-03 1.55847400e-01
5.46126544e-01 2.68563479e-01 1.70846015e-01 5.02319038e-01
-9.47716355e-01 2.62797266e-01 4.57295150e-01 -7.85072029e-01
2.43444294e-01 -8.62454534e-01 -1.21911563e-01 6.35872066e-01
-6.55476034e-01 -1.48434397e-02 5.73365688e-01 -8.49896133e-01
-1.91569221e+00 -1.26756859e+00 5.38931727e-01 2.15658978e-01
7.09054351e-01 -1.71476007e-01 -2.05749318e-01 2.63408542e-01
-8.08239356e-02 6.54202029e-02 2.06968814e-01 -3.05600047e-01
2.13176191e-01 -6.53249443e-01 -1.48025584e+00 3.95277262e-01
1.12093341e+00 -2.49883354e-01 5.13396800e-01 6.07641459e-01
9.54009533e-01 9.22805220e-02 -7.05512047e-01 6.97685957e-01
4.32583243e-01 -1.92778140e-01 3.95175040e-01 -4.64132637e-01
-1.91890135e-01 -2.79420644e-01 -3.52833688e-01 -1.15070128e+00
2.17389479e-01 -8.75193059e-01 -2.05630541e-01 1.17650902e+00
6.60828292e-01 -6.93842053e-01 8.12265277e-01 1.14831340e+00
-1.41949626e-02 -7.77761996e-01 -1.46248853e+00 -1.19483924e+00
3.26394364e-02 -4.06439424e-01 6.39060318e-01 7.93028831e-01
-8.87909532e-02 -2.74374068e-01 -5.02122819e-01 3.28759372e-01
1.05198741e+00 3.02179247e-01 1.55557662e-01 -9.35944974e-01
-1.01306096e-01 -9.63795111e-02 4.52956229e-01 -6.33562982e-01
1.46624669e-01 -7.33494163e-01 2.76858687e-01 -1.19350505e+00
4.74612981e-01 -1.06715107e+00 -6.04749143e-01 5.54984629e-01
1.28830746e-01 -5.92001379e-01 3.74763548e-01 2.57727295e-01
-6.46880269e-01 6.44113481e-01 9.14969146e-01 -3.23224962e-01
-7.68507272e-02 1.87079206e-01 -5.38856387e-01 2.12065622e-01
6.90047622e-01 -8.18853736e-01 -6.35607421e-01 -6.14941478e-01
6.37906134e-01 4.65995491e-01 5.38036712e-02 -7.80940056e-01
2.85230726e-01 -7.78100312e-01 -3.32982391e-01 -6.97970033e-01
-7.70389140e-02 -1.46416008e+00 7.73557007e-01 8.21134925e-01
-3.22501995e-02 -8.79873056e-03 -2.33166963e-01 1.21119487e+00
6.43447220e-01 -1.53171271e-02 5.70184946e-01 -5.38133979e-02
-1.99438468e-01 6.36656702e-01 -8.52285177e-02 -1.19128145e-01
1.48385155e+00 6.37615249e-02 -5.88484168e-01 -1.23072468e-01
-4.37456071e-01 9.33700979e-01 3.34971875e-01 3.69976580e-01
-2.87921224e-02 -1.04570341e+00 -3.56503069e-01 -3.63935381e-01
-3.57698888e-01 -1.91712808e-02 6.25179708e-01 1.08765376e+00
-6.96362834e-03 5.85331261e-01 -4.19748249e-03 -3.84903550e-01
-5.34277499e-01 7.43863940e-01 3.99072349e-01 -3.25530738e-01
-9.69761014e-02 4.90622789e-01 -1.20377466e-01 6.45070896e-02
3.00699234e-01 -2.96133786e-01 6.81690872e-01 8.01954344e-02
2.57545024e-01 8.88034344e-01 2.90619671e-01 2.39872426e-01
-4.47264850e-01 2.64177650e-01 3.56862694e-01 6.43932912e-03
1.28473175e+00 -5.50126016e-01 -8.30264986e-02 1.97154000e-01
5.92791021e-01 -2.53316253e-01 -1.39331162e+00 -1.45388991e-02
2.08360404e-01 -3.02335232e-01 4.18678284e-01 -1.13490868e+00
-1.41839266e+00 5.10364711e-01 7.68791318e-01 2.95344710e-01
1.50345397e+00 -4.62186426e-01 1.13984668e+00 4.93501686e-02
7.48257935e-01 -1.72068596e+00 -4.55834508e-01 -1.90248385e-01
4.39874023e-01 -9.14834261e-01 -6.38490319e-02 -3.95897537e-01
-4.12619948e-01 7.35839665e-01 5.59872031e-01 5.54648995e-01
6.95723474e-01 4.64110106e-01 -4.90323007e-01 2.29524508e-01
-1.12971032e+00 1.07901424e-01 -4.71663147e-01 2.73831427e-01
-4.21497554e-01 9.32527959e-01 -6.43515885e-01 8.51757526e-01
5.58104336e-01 3.13760668e-01 4.27027673e-01 6.60628378e-01
-4.15753782e-01 -8.03234339e-01 -1.35805696e-01 6.43603861e-01
-3.97999048e-01 1.19499326e-01 6.35440499e-02 2.20765531e-01
3.54053289e-01 7.32707977e-01 -1.34409681e-01 -8.96121264e-02
1.17278494e-01 -9.78725478e-02 8.82531330e-02 -7.66317323e-02
9.68853291e-03 8.79000798e-02 8.86367559e-02 -1.97127536e-01
-2.41706654e-01 -5.51219523e-01 -1.44068682e+00 -3.93354356e-01
-7.71201968e-01 3.03813498e-02 1.41834390e+00 1.09947908e+00
5.87704539e-01 4.78901982e-01 1.25589609e+00 -5.16396463e-01
-1.22431517e+00 -3.90688419e-01 -7.87993789e-01 -3.14299107e-01
1.77152921e-02 -4.38697577e-01 -5.95332921e-01 -3.24641705e-01] | [5.818693161010742, 1.814211130142212] |
2c670ef7-4b1b-4bb9-a448-3c1c19100cf0 | gesture-recognition-in-robotic-surgery-a | 2102.00027 | null | https://arxiv.org/abs/2102.00027v1 | https://arxiv.org/pdf/2102.00027v1.pdf | Gesture Recognition in Robotic Surgery: a Review | Objective: Surgical activity recognition is a fundamental step in computer-assisted interventions. This paper reviews the state-of-the-art in methods for automatic recognition of fine-grained gestures in robotic surgery focusing on recent data-driven approaches and outlines the open questions and future research directions. Methods: An article search was performed on 5 bibliographic databases with the following search terms: robotic, robot-assisted, JIGSAWS, surgery, surgical, gesture, fine-grained, surgeme, action, trajectory, segmentation, recognition, parsing. Selected articles were classified based on the level of supervision required for training and divided into different groups representing major frameworks for time series analysis and data modelling. Results: A total of 52 articles were reviewed. The research field is showing rapid expansion, with the majority of articles published in the last 4 years. Deep-learning-based temporal models with discriminative feature extraction and multi-modal data integration have demonstrated promising results on small surgical datasets. Currently, unsupervised methods perform significantly less well than the supervised approaches. Conclusion: The development of large and diverse open-source datasets of annotated demonstrations is essential for development and validation of robust solutions for surgical gesture recognition. While new strategies for discriminative feature extraction and knowledge transfer, or unsupervised and semi-supervised approaches, can mitigate the need for data and labels, they have not yet been demonstrated to achieve comparable performance. Important future research directions include detection and forecast of gesture-specific errors and anomalies. Significance: This paper is a comprehensive and structured analysis of surgical gesture recognition methods aiming to summarize the status of this rapidly evolving field. | ['Danail Stoyanov', 'Matthew J. Clarkson', 'Beatrice van Amsterdam'] | 2021-01-29 | null | null | null | null | ['surgical-gesture-recognition'] | ['medical'] | [ 2.58650392e-01 -1.07391894e-01 -9.97515976e-01 -2.96056122e-01
-9.60793495e-01 -5.53397179e-01 3.73173058e-01 1.29582956e-01
-8.12582254e-01 3.04616779e-01 6.54336989e-01 -2.75337040e-01
-7.41193652e-01 1.32217675e-01 -3.03645153e-02 -1.11841667e+00
-3.96483153e-01 4.75695282e-01 6.02418110e-02 1.57402307e-01
6.29736185e-01 8.46300185e-01 -1.41680741e+00 4.18130130e-01
2.37685546e-01 7.39073873e-01 2.20314473e-01 8.24035048e-01
-1.22076482e-01 7.18001604e-01 -4.98233974e-01 3.64859313e-01
7.46430308e-02 -5.62553883e-01 -9.09779012e-01 -2.36590117e-01
-2.17218757e-01 -5.20344973e-01 -4.95027572e-01 3.65170300e-01
1.00336611e+00 -9.25711170e-02 7.14430988e-01 -6.70309603e-01
1.00355983e-01 4.95646119e-01 4.19520475e-02 4.68053430e-01
3.64818811e-01 2.32382759e-01 1.31450132e-01 -6.99914992e-01
1.07768273e+00 6.17633879e-01 6.61417544e-01 8.57111096e-01
-6.53372467e-01 -6.04824066e-01 -2.33663872e-01 6.20204583e-03
-8.03172052e-01 -2.13464037e-01 3.98425311e-01 -7.93612182e-01
1.14070427e+00 5.15668988e-01 9.54129159e-01 1.42741895e+00
7.91179240e-01 8.70279849e-01 1.00922823e+00 -6.54740691e-01
-1.56069230e-02 -2.51095653e-01 1.08560450e-01 8.49058509e-01
1.33690149e-01 8.05249929e-01 -8.47147226e-01 3.94865833e-02
9.38379705e-01 2.87805825e-01 -1.94940701e-01 -5.22961140e-01
-1.83560419e+00 5.14697909e-01 2.20515355e-01 1.01909888e+00
-6.57982111e-01 1.80862188e-01 8.65481079e-01 3.30264270e-01
-1.65162701e-02 5.03390908e-01 -4.22273159e-01 -1.02469313e+00
-1.05646431e+00 -3.01283419e-01 9.07405496e-01 1.06245017e+00
-2.86786973e-01 -3.07811260e-01 -3.26004654e-01 5.86190701e-01
3.13481957e-01 -5.72353713e-02 1.30915713e+00 -7.51816809e-01
-2.34750565e-03 5.63050330e-01 -1.90649033e-01 -5.22550881e-01
-1.16787183e+00 7.10293278e-02 -5.08400321e-01 3.19300860e-01
3.71184140e-01 -2.06907898e-01 -1.36608684e+00 9.23941076e-01
-1.92710832e-02 -3.62390488e-01 6.55960217e-02 9.56217229e-01
1.20590186e+00 -2.11338833e-01 4.38202769e-01 -4.20265973e-01
1.22367454e+00 -9.97491300e-01 -9.80390429e-01 4.77458984e-02
1.10105896e+00 -8.80692244e-01 4.72584248e-01 4.21169490e-01
-7.28492618e-01 -4.65608500e-02 -7.34888017e-01 1.96970835e-01
-3.71708423e-01 4.76283938e-01 1.11261439e+00 6.81940913e-01
-6.76058292e-01 6.30722880e-01 -1.82213879e+00 -8.76661241e-01
3.69706661e-01 8.51131856e-01 -5.86638153e-01 1.55164838e-01
-5.69515169e-01 1.38524210e+00 3.87048304e-01 3.63375932e-01
-9.44912553e-01 -4.57905799e-01 -7.37430096e-01 -7.87993193e-01
4.33576107e-02 -3.74107152e-01 1.44934034e+00 -6.43326819e-01
-1.69407201e+00 1.31458402e+00 -7.04381913e-02 -2.84786075e-01
5.43824017e-01 -3.58174652e-01 -2.55197823e-01 3.38161349e-01
-2.80002773e-01 3.46320391e-01 5.43602765e-01 -6.85969532e-01
-6.42261505e-01 -4.66249019e-01 -5.93488753e-01 3.29747021e-01
-1.32471576e-01 4.76577580e-01 -4.28022116e-01 -8.38986814e-01
2.73931950e-01 -1.29859948e+00 -4.63235199e-01 4.12726067e-02
-2.47175153e-02 -2.29618862e-01 3.40004653e-01 -6.97984874e-01
1.27178538e+00 -2.08659959e+00 3.31864685e-01 2.90049370e-02
4.51443978e-02 8.45828578e-02 2.59866029e-01 7.04875648e-01
-1.52469516e-01 -1.76280871e-01 -4.71960865e-02 1.01868637e-01
-3.48136872e-01 6.80137396e-01 2.59369284e-01 7.62201011e-01
-3.26928020e-01 1.05923951e+00 -1.08645511e+00 -5.57594955e-01
8.12792420e-01 1.63053334e-01 -1.75839320e-01 1.61130190e-01
5.25664985e-01 1.00384450e+00 -5.78016996e-01 1.02298188e+00
-3.70268345e-01 1.98157817e-01 2.56231248e-01 -1.54388160e-01
-4.49524969e-01 9.71938744e-02 -6.03977263e-01 2.42075253e+00
-4.90917116e-01 8.04091990e-01 -2.80127954e-02 -1.03237057e+00
7.21863031e-01 7.15735435e-01 1.45613778e+00 -4.67671722e-01
6.36008978e-01 8.11377048e-01 2.63057709e-01 -1.24355769e+00
3.42558436e-02 -2.16713279e-01 -1.72643170e-01 -1.21813409e-01
3.40793639e-01 -1.58779234e-01 -2.74909791e-02 -3.67195666e-01
1.27492666e+00 5.83302557e-01 6.23400271e-01 -1.53178964e-02
6.44501625e-03 7.00198472e-01 3.64446014e-01 8.02314937e-01
-7.52206922e-01 5.15069366e-01 1.51476562e-01 -4.08024788e-01
-4.31383133e-01 -8.05949867e-01 -1.71916977e-01 9.18703854e-01
9.14133340e-02 -1.41720355e-01 -3.80919427e-01 -7.00525105e-01
-1.32161990e-01 1.28337994e-01 -8.02518010e-01 -1.15687661e-01
-9.21292961e-01 -7.44676828e-01 5.89656055e-01 8.10513973e-01
-2.08677247e-01 -1.49166822e+00 -1.35780108e+00 3.82882893e-01
-2.88041221e-04 -9.10488427e-01 -3.57615277e-02 5.97316206e-01
-1.51630878e+00 -1.46439493e+00 -1.09696066e+00 -1.24412620e+00
6.22468293e-01 -1.81085020e-01 4.92892832e-01 -1.35101289e-01
-1.02564096e+00 1.11045480e+00 -6.80841446e-01 -6.66523397e-01
-2.95261621e-01 -5.75146712e-02 -9.34419259e-02 -8.14664841e-01
6.72225475e-01 -1.58474296e-01 -8.14419210e-01 3.21637779e-01
-7.75442600e-01 -3.56951237e-01 1.29028535e+00 1.23156369e+00
5.81555605e-01 -7.47296393e-01 7.68221170e-02 -3.57166648e-01
5.21625340e-01 -2.93521911e-01 -9.42406058e-02 1.14807203e-01
-8.31255257e-01 -1.16372015e-02 -2.92735666e-01 -7.21523762e-01
-7.21364141e-01 4.02864873e-01 3.62001285e-02 -4.94122744e-01
-6.34575963e-01 8.20200384e-01 7.13954091e-01 -4.47559714e-01
6.85325623e-01 1.86716959e-01 6.03364944e-01 -4.31133479e-01
-2.43098452e-03 5.75078905e-01 6.29061043e-01 -3.35230261e-01
1.18492484e-01 4.74994123e-01 -6.13459647e-02 -7.25465834e-01
-1.13437787e-01 -1.18192053e+00 -1.03799224e+00 -5.31339884e-01
9.65131640e-01 -4.82609838e-01 -5.02964914e-01 6.25439525e-01
-9.22495723e-01 -6.75265610e-01 -4.06777412e-01 1.54347467e+00
-9.73759651e-01 1.02847494e-01 -7.02722669e-01 -4.48909760e-01
-5.53678095e-01 -1.54683673e+00 1.10637653e+00 8.92292708e-02
-6.30119622e-01 -7.94356763e-01 5.01787901e-01 7.37551302e-02
3.88517052e-01 7.37479210e-01 4.58822221e-01 -8.99316251e-01
-1.90001234e-01 -6.89428270e-01 3.15846533e-01 -7.02912509e-02
4.48789418e-01 1.09007815e-02 -2.31968895e-01 -4.38157097e-02
3.09379343e-02 -3.58669497e-02 4.87079978e-01 7.75622606e-01
1.05668056e+00 -3.63703109e-02 -8.08157265e-01 4.99959499e-01
1.16907883e+00 7.61798382e-01 5.86209714e-01 7.09755421e-01
3.57593060e-01 5.97295225e-01 1.16870987e+00 1.91441581e-01
-1.88479960e-01 5.20165384e-01 3.21851343e-01 1.51168063e-01
-2.50663161e-01 4.00039732e-01 9.45211947e-02 9.58480418e-01
-6.74572408e-01 3.89184147e-01 -1.37595737e+00 8.19944739e-01
-1.66718805e+00 -7.88886726e-01 -1.80630580e-01 1.91274858e+00
5.07122695e-01 -1.48295581e-01 -5.45226708e-02 4.33447398e-02
2.47819901e-01 -9.89695713e-02 -4.89639044e-01 -3.06447238e-01
4.95590031e-01 5.33244908e-01 8.87975097e-01 -4.37494181e-02
-1.25073695e+00 7.35110700e-01 6.76495886e+00 3.57125729e-01
-1.45848930e+00 9.34555680e-02 -1.86967820e-01 -2.80229539e-01
5.29191375e-01 -2.95107394e-01 -2.68450975e-01 2.18139932e-01
9.30837452e-01 2.40964945e-02 -9.55148861e-02 8.70161653e-01
4.23759818e-01 -3.70931745e-01 -1.13296986e+00 1.16030014e+00
1.09124132e-01 -1.37145042e+00 -5.85119069e-01 1.55120669e-02
3.56510878e-01 5.03672957e-01 -2.38011137e-01 1.63222820e-01
-1.90710768e-01 -9.71057177e-01 3.85054111e-01 7.72137582e-01
9.67135906e-01 -3.90298367e-02 9.85645890e-01 1.00511327e-01
-9.61934805e-01 -2.37681434e-01 4.57673073e-01 2.28658900e-01
2.73162685e-02 -2.11898848e-01 -9.58946168e-01 7.97288060e-01
8.26220334e-01 9.31203485e-01 -1.08956704e-02 1.38491344e+00
-7.33740255e-02 4.98210490e-01 -3.59600067e-01 -3.07760745e-01
4.34396058e-01 2.70568222e-01 6.08443141e-01 1.44771898e+00
3.01091373e-01 3.67101163e-01 5.64241670e-02 -1.97427392e-01
5.47348619e-01 1.26264378e-01 -6.62240982e-01 -4.65095699e-01
5.56722609e-03 1.04024708e+00 -1.00728548e+00 7.56690279e-02
-6.27686322e-01 8.82779062e-01 -4.70440924e-01 2.23427728e-01
-3.74323249e-01 -4.41946119e-01 3.21161956e-01 -1.56750634e-01
-2.31852934e-01 -6.39565527e-01 -5.27200222e-01 -7.67842412e-01
-1.28663173e-02 -8.35813880e-01 8.64157617e-01 -3.82374048e-01
-6.59189939e-01 3.63018006e-01 2.28044644e-01 -1.89900947e+00
-7.80321419e-01 -1.08247471e+00 -2.62769610e-01 5.01157403e-01
-9.67351615e-01 -9.72138703e-01 -6.33564591e-01 6.21294558e-01
8.82995486e-01 -1.52413860e-01 1.48343766e+00 5.38002923e-02
-1.91210851e-01 2.03779414e-01 3.12281609e-01 4.48601604e-01
9.52516854e-01 -8.91572654e-01 -5.89959145e-01 2.44363710e-01
-2.98454940e-01 8.04935455e-01 4.84827429e-01 -7.24695146e-01
-1.76003182e+00 -3.22414637e-01 4.51502055e-01 -6.24587953e-01
8.02360356e-01 2.98379123e-01 -2.28978336e-01 7.56681085e-01
3.55090722e-02 -5.45290997e-03 1.09417617e+00 -1.52213901e-01
3.46593320e-01 3.28786999e-01 -1.05736446e+00 3.67599070e-01
1.12382054e+00 -1.95848197e-01 -9.92792249e-01 4.18818593e-01
-7.93512538e-02 -1.03366911e+00 -1.37675333e+00 9.00387943e-01
1.24802518e+00 -4.72931147e-01 7.31301844e-01 -8.04833531e-01
4.43744034e-01 3.78792793e-01 3.00414324e-01 -1.00913858e+00
1.53613672e-01 -6.54990196e-01 -4.72375751e-02 1.76089734e-01
1.88637733e-01 -4.69678372e-01 9.41459298e-01 7.11043954e-01
-7.00048327e-01 -9.73999321e-01 -1.17465496e+00 -6.91226304e-01
-9.67032928e-03 -5.08848786e-01 -1.92285180e-01 7.83504367e-01
7.79117227e-01 -6.88151717e-01 1.04669325e-01 -3.32798243e-01
1.28025502e-01 8.91619697e-02 5.32178760e-01 -9.74152863e-01
3.33214968e-01 -9.50795114e-01 -9.44688857e-01 -4.44602311e-01
-2.68041104e-01 -8.01121593e-01 4.13593650e-01 -2.10765815e+00
-1.28952876e-01 -2.50964463e-01 -3.72034222e-01 5.80384672e-01
4.80104536e-01 -1.79539770e-01 -2.36526772e-01 5.98560035e-01
-1.22054629e-01 4.08135355e-02 1.26896930e+00 -1.01568764e-02
-5.38415313e-01 2.59479135e-01 6.55353814e-02 7.49496043e-01
6.98700249e-01 -6.27203345e-01 -3.33810970e-03 -2.57561412e-02
-4.73949879e-01 2.82907784e-01 2.54200339e-01 -7.73306489e-01
5.92485309e-01 -2.00364724e-01 1.93039492e-01 -7.40003645e-01
2.60965489e-02 -8.98127973e-01 8.82357657e-02 1.11865330e+00
-2.98584610e-01 4.76052839e-04 4.23565507e-01 3.91157597e-01
-5.68690121e-01 -1.77404642e-01 5.08593380e-01 -3.87200058e-01
-1.16330302e+00 2.57406831e-01 -8.82413387e-01 -2.11415038e-01
1.17681766e+00 -5.57267070e-01 2.40256816e-01 1.02819696e-01
-1.32756507e+00 -9.80888456e-02 -4.09712940e-02 8.74469876e-01
7.52081394e-01 -1.10560489e+00 -3.58611763e-01 -1.67404488e-01
4.51763123e-01 -4.07705829e-02 3.57001096e-01 1.71287584e+00
-7.11883783e-01 9.52964127e-01 -3.77023816e-01 -7.06851304e-01
-1.43458402e+00 3.59217316e-01 3.72854352e-01 -4.29272383e-01
-1.06898057e+00 5.38471937e-01 -3.62469018e-01 -4.39894348e-01
8.83088946e-01 -6.64169610e-01 -4.18999732e-01 1.46825776e-01
1.58932999e-01 2.31793180e-01 3.30690295e-01 -4.84252274e-01
-7.47011542e-01 6.34895980e-01 2.03120038e-01 -1.56564191e-01
1.39671850e+00 2.04773784e-01 9.39741731e-02 6.39931738e-01
1.12351215e+00 -5.48476458e-01 -5.88795960e-01 1.68171868e-01
4.53438550e-01 -2.26981249e-02 -3.02461889e-02 -1.15688753e+00
-9.18701768e-01 5.20897746e-01 1.19610405e+00 -3.98969322e-01
1.20769978e+00 1.38344005e-01 3.40162337e-01 2.07453936e-01
6.68996394e-01 -1.14164317e+00 2.10575033e-02 3.87522936e-01
1.12596464e+00 -1.36295521e+00 -8.78894255e-02 -1.69234082e-01
-5.88848948e-01 1.58454239e+00 1.09291717e-01 -1.83227330e-01
1.04741180e+00 4.24967200e-01 4.41886455e-01 -6.34423792e-01
-1.00928256e-02 -1.68768600e-01 5.78768909e-01 4.42202985e-01
7.43265390e-01 2.42652565e-01 -1.15944719e+00 5.55234849e-01
4.30613533e-02 6.83812141e-01 7.10812509e-02 1.66635180e+00
4.17257063e-02 -9.33260024e-01 -1.59584776e-01 7.72951066e-01
-8.02267373e-01 1.33549824e-01 -1.37964129e-01 1.35602999e+00
-3.17255855e-01 6.61875129e-01 -3.59771430e-01 -2.85504788e-01
5.90586841e-01 1.91754833e-01 7.96547472e-01 -5.49144983e-01
-1.04026806e+00 2.72540480e-01 2.90920764e-01 -1.14477181e+00
-9.59693968e-01 -9.79014814e-01 -1.53841329e+00 5.46570122e-01
-1.25987485e-01 -1.44561216e-01 1.35180473e+00 1.01180971e+00
2.42136903e-02 9.36887562e-01 -1.41899306e-02 -1.05234683e+00
-3.30897391e-01 -1.12983775e+00 -2.35871926e-01 3.19139957e-02
3.37648273e-01 -9.34378922e-01 -3.60155255e-01 3.94872352e-02] | [14.04604721069336, -3.397076368331909] |
42aee254-c71b-4151-baaf-aff10cb5dff4 | fine-grained-representation-learning-and | 1808.04505 | null | http://arxiv.org/abs/1808.04505v1 | http://arxiv.org/pdf/1808.04505v1.pdf | Fine-Grained Representation Learning and Recognition by Exploiting Hierarchical Semantic Embedding | Object categories inherently form a hierarchy with different levels of
concept abstraction, especially for fine-grained categories. For example, birds
(Aves) can be categorized according to a four-level hierarchy of order, family,
genus, and species. This hierarchy encodes rich correlations among various
categories across different levels, which can effectively regularize the
semantic space and thus make prediction less ambiguous. However, previous
studies of fine-grained image recognition primarily focus on categories of one
certain level and usually overlook this correlation information. In this work,
we investigate simultaneously predicting categories of different levels in the
hierarchy and integrating this structured correlation information into the deep
neural network by developing a novel Hierarchical Semantic Embedding (HSE)
framework. Specifically, the HSE framework sequentially predicts the category
score vector of each level in the hierarchy, from highest to lowest. At each
level, it incorporates the predicted score vector of the higher level as prior
knowledge to learn finer-grained feature representation. During training, the
predicted score vector of the higher level is also employed to regularize label
prediction by using it as soft targets of corresponding sub-categories. To
evaluate the proposed framework, we organize the 200 bird species of the
Caltech-UCSD birds dataset with the four-level category hierarchy and construct
a large-scale butterfly dataset that also covers four level categories.
Extensive experiments on these two and the newly-released VegFru datasets
demonstrate the superiority of our HSE framework over the baseline methods and
existing competitors. | ['Wenxi Wu', 'Liang Lin', 'Yuefang Gao', 'Xiaonan Luo', 'Tianshui Chen', 'Le Dong'] | 2018-08-14 | null | null | null | null | ['fine-grained-image-recognition'] | ['computer-vision'] | [-5.29665835e-02 -2.78521895e-01 -2.12313652e-01 -7.29910553e-01
1.79616407e-01 -4.97026891e-01 2.76409179e-01 3.03284734e-01
-3.14409256e-01 3.40150625e-01 4.68778312e-01 1.33846283e-01
-2.86533892e-01 -1.03036010e+00 -3.22343618e-01 -5.22706866e-01
-1.44568428e-01 1.56079382e-01 4.89374816e-01 -3.36418711e-02
2.27190569e-01 2.54217684e-01 -1.96480358e+00 6.11420393e-01
7.25914717e-01 1.55168843e+00 2.10491523e-01 -3.51083279e-02
-7.76936933e-02 5.57764590e-01 -3.94024074e-01 -2.46010721e-02
2.47653648e-01 -9.10370573e-02 -6.53097928e-01 3.55708003e-01
5.61085522e-01 -1.85875297e-01 -1.88553318e-01 1.15261710e+00
-1.37581304e-01 1.81299970e-01 8.11342478e-01 -1.19398534e+00
-8.91901612e-01 5.78223884e-01 -6.61248207e-01 2.62783587e-01
-2.46659711e-01 -1.03166975e-01 1.49361789e+00 -9.11967754e-01
4.54078875e-02 1.40773594e+00 6.00446999e-01 3.77711475e-01
-1.19314730e+00 -9.85587001e-01 6.92274630e-01 3.38957906e-01
-1.61092567e+00 1.79673925e-01 7.13554382e-01 -7.98158348e-01
7.08685219e-01 -3.36762927e-02 6.41307354e-01 5.61969280e-01
6.06001131e-02 4.80243862e-01 1.35068369e+00 8.36468190e-02
1.53409764e-01 1.48853526e-01 8.39949489e-01 1.00048065e+00
2.53596425e-01 1.83133960e-01 -4.33493763e-01 -2.86125648e-03
6.04089141e-01 5.11678815e-01 3.81172225e-02 -4.93799508e-01
-1.30809009e+00 1.12774456e+00 1.33261752e+00 4.83042061e-01
-2.31734768e-01 3.56033742e-02 2.46962070e-01 -1.45817157e-02
3.03252608e-01 3.49488854e-01 -5.79919279e-01 6.54139042e-01
-7.75583982e-01 -1.00134816e-02 5.30730903e-01 8.58933210e-01
1.08459687e+00 -2.37319127e-01 -4.73630250e-01 1.20996344e+00
6.55416071e-01 -2.62598395e-01 7.73356795e-01 -6.30178869e-01
2.21443459e-01 1.10098016e+00 -2.33722255e-01 -1.30086219e+00
-5.18746495e-01 -9.38719153e-01 -1.00607121e+00 -2.47390196e-02
-6.37276173e-02 3.10091048e-01 -1.13321400e+00 1.78841841e+00
3.93494457e-01 4.40397441e-01 -1.97019711e-01 1.00883901e+00
1.16963184e+00 8.40307236e-01 3.08485001e-01 1.82847291e-01
1.54391217e+00 -1.32198703e+00 -1.23587765e-01 -2.04239205e-01
2.53842205e-01 -2.38098964e-01 1.17068231e+00 2.02979788e-01
-2.40222245e-01 -1.01837850e+00 -1.26749659e+00 -9.39136818e-02
-6.11339748e-01 3.07444513e-01 8.57268631e-01 2.16587290e-01
-9.63052213e-01 5.09853780e-01 -5.57171822e-01 -1.04687341e-01
5.72426856e-01 2.72863209e-01 -2.09845692e-01 -8.94540027e-02
-1.39808404e+00 3.44888240e-01 6.70476317e-01 -1.34965062e-01
-9.20209646e-01 -8.91021013e-01 -9.10487354e-01 4.96001482e-01
9.96213332e-02 -6.62972033e-01 7.46950209e-01 -6.80863082e-01
-1.00205946e+00 9.46841836e-01 8.49730987e-03 -3.21153462e-01
-2.42386043e-01 1.49818927e-01 -3.52091789e-01 -8.83761570e-02
6.17152631e-01 1.18588996e+00 8.39058042e-01 -1.26392555e+00
-1.20096087e+00 -4.62067276e-01 3.29485804e-01 1.70780227e-01
-5.76285601e-01 -2.50220329e-01 -3.15913618e-01 -9.62557793e-01
3.64375442e-01 -8.07868600e-01 -2.82997578e-01 -1.20663993e-01
-3.13211620e-01 -7.44940639e-01 5.14591277e-01 -2.61226803e-01
1.39599419e+00 -2.29444599e+00 1.79356173e-01 8.87964517e-02
4.48185772e-01 -1.93580493e-01 -9.93003622e-02 -9.71921235e-02
-1.02814166e-02 6.50137737e-02 -1.93954542e-01 6.90644160e-02
2.22012773e-01 2.99100101e-01 -3.07394177e-01 1.35463759e-01
1.71302170e-01 6.76614523e-01 -8.34800422e-01 -4.08394456e-01
1.66187420e-01 3.20084840e-01 -9.83751774e-01 1.83124438e-01
5.64998537e-02 1.47461340e-01 -4.91858363e-01 6.86907411e-01
3.41857553e-01 -5.40503204e-01 -9.27871615e-02 -6.61205709e-01
-1.42339617e-01 1.90101475e-01 -1.10010815e+00 1.29263198e+00
-5.29627085e-01 1.19522661e-01 -3.54273200e-01 -1.07497513e+00
9.68344569e-01 -2.84994394e-01 1.99417293e-01 -3.29110384e-01
7.42411017e-02 -2.86555029e-02 2.29318663e-01 1.02138393e-01
3.66514444e-01 -4.19679999e-01 -6.04784131e-01 -1.03320852e-02
3.64402920e-01 1.15292735e-01 1.01475209e-01 -2.73503624e-02
6.18548632e-01 -2.85393536e-01 4.55457538e-01 -5.05949855e-01
6.51907086e-01 -7.96809569e-02 8.44930351e-01 6.75475359e-01
-3.75414819e-01 1.87331870e-01 2.92623788e-01 -6.47077501e-01
-6.68593168e-01 -9.94586110e-01 -5.36606193e-01 1.55636978e+00
5.07985890e-01 -4.78057086e-01 -5.85386038e-01 -8.91794801e-01
3.80909234e-01 4.08218622e-01 -1.00191832e+00 -2.69082338e-01
-4.33107745e-03 -7.09053755e-01 1.61743730e-01 6.69621527e-01
8.05712819e-01 -9.95543003e-01 -3.13345015e-01 1.65384680e-01
-1.15843825e-02 -9.31568623e-01 -5.85657299e-01 2.98756182e-01
-7.45656371e-01 -8.46888423e-01 -2.31400356e-01 -1.04766512e+00
6.81362569e-01 3.92591804e-01 1.17992282e+00 2.96500444e-01
-1.56495005e-01 -9.30554047e-02 -4.77726460e-01 1.00316899e-02
2.40530789e-01 -2.13330146e-02 2.43100286e-01 3.96053731e-01
6.17669821e-01 -4.99279767e-01 -7.28576124e-01 6.83804870e-01
-7.45849073e-01 -1.31013691e-02 6.03027284e-01 1.17074382e+00
8.28232646e-01 6.69106901e-01 5.53488195e-01 -7.94584870e-01
2.43915781e-01 -6.46753192e-01 -6.62338853e-01 4.25517233e-03
-4.70053345e-01 6.85234368e-02 7.73174822e-01 -2.27129504e-01
-9.25180912e-01 5.29846065e-02 -1.81351215e-01 -3.08100253e-01
-4.77549195e-01 6.03615761e-01 -3.73937458e-01 4.80577797e-02
2.89066136e-01 8.38790834e-02 -6.50348127e-01 -5.75618446e-01
2.96951085e-01 7.16257811e-01 4.29695398e-01 -5.24207413e-01
7.83826590e-01 2.77701974e-01 -1.40592158e-01 -5.35997212e-01
-1.57687998e+00 -6.61518812e-01 -9.53715503e-01 6.27932996e-02
9.34813201e-01 -1.15194404e+00 -5.71185350e-01 2.64820367e-01
-7.69432366e-01 5.44201247e-02 -2.38112155e-02 3.85547876e-01
-1.57963946e-01 1.43693775e-01 -6.08404338e-01 -4.16167229e-02
5.85001111e-02 -1.17110479e+00 1.26740777e+00 4.70336199e-01
-3.19893844e-03 -9.26973045e-01 -2.74804890e-01 4.83429104e-01
1.32983118e-01 9.73365605e-02 1.15797269e+00 -4.84351486e-01
-5.38792908e-01 3.51305567e-02 -5.82114637e-01 4.49491858e-01
3.49362105e-01 -3.58274817e-01 -8.08760643e-01 -4.41250473e-01
-1.03599057e-01 -3.94426346e-01 1.41906202e+00 1.23264655e-01
1.63570750e+00 -3.58152390e-01 -3.84891063e-01 7.26507068e-01
1.48170471e+00 2.42894776e-02 -7.20383972e-02 3.31928045e-01
9.86146033e-01 6.82818651e-01 7.51192868e-01 3.32763255e-01
5.86806357e-01 7.31254339e-01 3.67302954e-01 1.37169912e-01
-1.32469818e-01 -4.47550058e-01 -1.43237673e-02 9.64663625e-01
1.70202747e-01 2.32505754e-01 -6.70276403e-01 4.63949680e-01
-1.52274990e+00 -7.88372993e-01 1.49130523e-01 1.95354402e+00
9.53280687e-01 1.89633623e-01 -2.88850293e-02 1.88652298e-03
8.99377167e-01 2.35934854e-01 -5.93475640e-01 -7.32036158e-02
2.19133139e-01 4.00386266e-02 2.88178056e-01 3.65878433e-01
-1.59760916e+00 1.25160742e+00 5.51605892e+00 9.70281184e-01
-1.09244657e+00 1.10540733e-01 4.86800224e-01 2.14058280e-01
1.68960523e-02 -1.81486949e-01 -1.22364080e+00 6.36415184e-01
5.14023125e-01 -1.14130981e-01 2.69082665e-01 1.17104781e+00
-3.80448788e-01 3.01659644e-01 -1.14656770e+00 9.10533607e-01
-3.77131775e-02 -1.09243369e+00 4.42336231e-01 4.81950790e-02
8.32005382e-01 -5.21853156e-02 1.56038061e-01 7.61534274e-01
4.77312535e-01 -1.07899976e+00 6.72761619e-01 2.09453270e-01
7.44533658e-01 -7.53999293e-01 6.43915772e-01 4.14439291e-01
-1.73501754e+00 -5.22296488e-01 -9.34825003e-01 -1.46454886e-01
-3.42658460e-01 4.42348361e-01 -4.77476090e-01 3.26678842e-01
9.82376993e-01 1.14894569e+00 -8.47360492e-01 8.99755716e-01
-3.12576354e-01 7.54566014e-01 -1.18161812e-01 5.45910969e-02
7.44155049e-01 -4.55390625e-02 5.77894635e-02 1.15429664e+00
1.96931958e-01 4.20142800e-01 7.17564166e-01 6.54425383e-01
-2.65994579e-01 -4.85214442e-02 -1.77027375e-01 1.35476694e-01
5.09007215e-01 1.40966582e+00 -8.73491108e-01 -4.28245276e-01
-5.97161710e-01 7.47370243e-01 6.01741016e-01 -1.53186079e-02
-7.50513554e-01 -4.40925688e-01 9.10432160e-01 -3.74837480e-02
5.72308779e-01 -9.74168107e-02 -2.50479281e-01 -1.25123298e+00
-4.52770621e-01 -5.53084671e-01 8.97109091e-01 -4.41500425e-01
-1.78314292e+00 7.40384698e-01 -1.46008402e-01 -1.34480321e+00
2.27341622e-01 -8.12817335e-01 -2.19484597e-01 7.18585253e-01
-1.58714509e+00 -1.33470762e+00 -4.53367203e-01 6.54501259e-01
6.80776000e-01 -3.45278889e-01 7.61643052e-01 3.60303938e-01
-3.86194527e-01 7.23171890e-01 1.89834610e-02 3.14559072e-01
3.15025419e-01 -1.37277186e+00 -4.27796878e-02 5.86774647e-01
2.67083496e-01 8.18066776e-01 2.49874294e-01 -5.19631505e-01
-5.94435394e-01 -1.71194577e+00 6.82234526e-01 -3.90418917e-01
9.47157383e-01 -4.77126092e-01 -1.00010383e+00 5.51102400e-01
-3.14005494e-01 2.11848751e-01 9.99191165e-01 3.32699955e-01
-8.52518737e-01 -4.84609306e-01 -9.79807496e-01 1.57805368e-01
1.09623456e+00 -5.43318093e-01 -1.18664038e+00 1.51697859e-01
9.09548342e-01 1.04343250e-01 -1.00432873e+00 7.26127863e-01
5.85393488e-01 -8.80346000e-01 1.03479004e+00 -5.87654769e-01
6.26711249e-01 -6.20663226e-01 -4.77902859e-01 -1.64392805e+00
-1.10748458e+00 4.82361197e-01 1.57696992e-01 1.18012595e+00
9.07219723e-02 -5.20872772e-01 6.74873531e-01 3.75223719e-02
-3.46528769e-01 -8.15947235e-01 -6.66451037e-01 -7.31264234e-01
1.14482060e-01 -1.25041246e-01 7.77538955e-01 1.02487242e+00
-2.65967458e-01 7.56682992e-01 -3.06960285e-01 4.72681135e-01
9.52773333e-01 8.11717093e-01 3.01786393e-01 -1.63163018e+00
-2.79571354e-01 -5.70465982e-01 -8.34988773e-01 -1.22183192e+00
3.97715509e-01 -1.27464414e+00 2.98684150e-01 -1.48056543e+00
4.29224521e-01 -6.72647357e-01 -9.11461830e-01 6.97022796e-01
-4.33088839e-01 7.45178521e-01 1.68392509e-01 3.72523069e-01
-7.04122066e-01 5.91227591e-01 1.06000817e+00 -4.68337923e-01
1.23050489e-01 -1.66957840e-01 -1.02298832e+00 9.57266986e-01
4.74833429e-01 -4.23184037e-01 -5.06082475e-01 -3.37340474e-01
-2.85762936e-01 -4.12375897e-01 4.35205340e-01 -1.27211404e+00
2.14278594e-01 -1.50950223e-01 7.13726699e-01 -6.99921787e-01
1.32933363e-01 -9.16687310e-01 -5.55025786e-02 4.29343969e-01
-5.05682945e-01 -6.23526216e-01 4.23021279e-02 6.03937089e-01
-6.71890020e-01 -3.43491621e-02 1.11798263e+00 -2.30302271e-02
-1.18718088e+00 8.47440898e-01 6.46442249e-02 -1.15992524e-01
9.50375557e-01 -1.39511183e-01 -1.20672569e-01 3.18890810e-01
-9.08449709e-01 4.85585064e-01 2.95664012e-01 7.33352423e-01
5.72521627e-01 -1.64622998e+00 -4.87685353e-01 3.48917812e-01
5.67918956e-01 -1.39881954e-01 3.94186437e-01 1.80980504e-01
-8.75771511e-03 6.57769501e-01 -6.19167268e-01 -7.69919515e-01
-1.15835404e+00 8.84744227e-01 2.67447621e-01 -3.33884567e-01
-2.56418467e-01 1.18555748e+00 1.22108734e+00 -5.69036901e-01
1.63565218e-01 -4.03100461e-01 -9.26263928e-01 2.03059480e-01
5.04950821e-01 -8.41073319e-02 -1.83955371e-01 -1.07389557e+00
-4.74607527e-01 9.37291920e-01 -1.62846684e-01 5.18307149e-01
1.27713788e+00 -9.62004885e-02 -3.24084610e-01 4.40231413e-01
1.33262110e+00 -3.09812725e-01 -1.21554458e+00 -5.12607932e-01
2.01293811e-01 -5.43946385e-01 2.31857166e-01 -7.54878461e-01
-1.10856783e+00 1.00165558e+00 5.67520559e-01 2.16418326e-01
1.20467448e+00 3.44605058e-01 6.31123781e-01 1.88699901e-01
6.56409562e-01 -8.28739107e-01 1.49208814e-01 7.05383182e-01
8.70860994e-01 -1.10133612e+00 -2.05562472e-01 -5.88053524e-01
-5.89511156e-01 8.53998601e-01 8.91731441e-01 -3.33635092e-01
1.05572593e+00 -2.60641783e-01 -2.32347071e-01 -1.76406652e-01
-6.86027586e-01 -3.98787647e-01 6.98668540e-01 4.12326574e-01
3.49320084e-01 4.36309725e-01 -1.86655566e-01 1.27560663e+00
-3.12297702e-01 -8.03587139e-02 -3.83937508e-02 4.49914962e-01
-8.88841748e-01 -8.06828856e-01 -1.94716737e-01 7.85492837e-01
-2.35698625e-01 -2.38901839e-01 -1.37384832e-01 3.80295247e-01
8.31735730e-01 8.58386278e-01 2.85801202e-01 -6.94631577e-01
4.27839249e-01 -2.18147740e-01 2.31021315e-01 -1.14733350e+00
-5.11568606e-01 -1.76975757e-01 -2.00446025e-01 -5.69711030e-01
-3.57040554e-01 -4.83711511e-01 -1.30862772e+00 6.98004803e-03
5.87122999e-02 4.06945646e-01 9.89001393e-02 8.33496392e-01
2.17190787e-01 5.51451802e-01 9.96987641e-01 -7.68926620e-01
-4.76437658e-01 -8.68442714e-01 -9.24435616e-01 5.66803694e-01
2.89453328e-01 -1.31883276e+00 -5.04797995e-01 -1.30332913e-02] | [9.673028945922852, 2.060192346572876] |
b616b533-52bc-453c-9b8c-a817ee959373 | unifying-neural-learning-and-symbolic | 2004.13577 | null | https://arxiv.org/abs/2004.13577v1 | https://arxiv.org/pdf/2004.13577v1.pdf | Unifying Neural Learning and Symbolic Reasoning for Spinal Medical Report Generation | Automated medical report generation in spine radiology, i.e., given spinal medical images and directly create radiologist-level diagnosis reports to support clinical decision making, is a novel yet fundamental study in the domain of artificial intelligence in healthcare. However, it is incredibly challenging because it is an extremely complicated task that involves visual perception and high-level reasoning processes. In this paper, we propose the neural-symbolic learning (NSL) framework that performs human-like learning by unifying deep neural learning and symbolic logical reasoning for the spinal medical report generation. Generally speaking, the NSL framework firstly employs deep neural learning to imitate human visual perception for detecting abnormalities of target spinal structures. Concretely, we design an adversarial graph network that interpolates a symbolic graph reasoning module into a generative adversarial network through embedding prior domain knowledge, achieving semantic segmentation of spinal structures with high complexity and variability. NSL secondly conducts human-like symbolic logical reasoning that realizes unsupervised causal effect analysis of detected entities of abnormalities through meta-interpretive learning. NSL finally fills these discoveries of target diseases into a unified template, successfully achieving a comprehensive medical report generation. When it employed in a real-world clinical dataset, a series of empirical studies demonstrate its capacity on spinal medical report generation as well as show that our algorithm remarkably exceeds existing methods in the detection of spinal structures. These indicate its potential as a clinical tool that contributes to computer-aided diagnosis. | ['Benzheng Wei', 'Zhongyi Han', 'Yilong Yin', 'Shuo Li'] | 2020-04-28 | null | null | null | null | ['medical-report-generation'] | ['medical'] | [ 5.80758691e-01 9.80286896e-01 -1.76002875e-01 -2.14240640e-01
-8.09147179e-01 -3.00560117e-01 5.28160214e-01 3.12639862e-01
1.91625252e-01 7.23604798e-01 3.16874772e-01 -7.70401895e-01
-2.54639059e-01 -1.09813631e+00 -9.08675671e-01 -5.75699210e-01
9.87565368e-02 7.68162847e-01 4.25560512e-02 -1.35801420e-01
-1.30837902e-01 6.28475428e-01 -1.02313292e+00 5.09658217e-01
9.39431787e-01 6.94968343e-01 -3.66770141e-02 7.39244103e-01
-2.36841798e-01 1.60224450e+00 -6.15778685e-01 -7.20870793e-01
-1.94036216e-01 -8.58675659e-01 -9.78768170e-01 2.24961489e-01
1.00032575e-01 -2.07532197e-01 -4.54159915e-01 1.13870347e+00
3.60392272e-01 -3.41842443e-01 9.16491270e-01 -1.22662783e+00
-1.05593038e+00 7.83211589e-01 -3.58242422e-01 1.05213836e-01
4.73956913e-01 4.01176184e-01 6.34387910e-01 -5.30450940e-01
8.31928909e-01 1.07200134e+00 4.37750399e-01 6.23810530e-01
-1.19998169e+00 -5.42536199e-01 -1.72767639e-01 1.84238590e-02
-1.00282276e+00 2.30157956e-01 8.63804698e-01 -6.66642427e-01
5.70942104e-01 3.75764489e-01 7.94447243e-01 1.21652234e+00
7.93549538e-01 7.76440918e-01 1.02189970e+00 -1.81878611e-01
3.46726358e-01 -1.80228144e-01 -8.58683884e-02 1.20399415e+00
3.57026905e-01 1.66283786e-01 -3.10421586e-01 -5.72260842e-02
1.04387701e+00 3.36184740e-01 -1.93123743e-01 -6.58648238e-02
-1.44351232e+00 1.02544522e+00 9.70326543e-01 2.54776299e-01
-5.86919010e-01 2.79280245e-01 3.45833510e-01 1.33889699e-02
9.26644951e-02 3.84877235e-01 3.00050497e-01 6.86732292e-01
-8.96014929e-01 3.51943046e-01 6.67838454e-01 7.51150489e-01
3.94854648e-03 2.77834505e-01 -5.76336145e-01 2.95159698e-01
3.06924343e-01 6.84604943e-01 5.91658235e-01 -8.75039399e-01
1.30259827e-01 1.00549197e+00 -3.96135777e-01 -1.09130859e+00
-5.84248185e-01 -5.65589428e-01 -1.21897280e+00 2.48904914e-01
7.00333640e-02 8.01216140e-02 -1.20419383e+00 1.40309215e+00
3.92039955e-01 3.13516527e-01 2.74406552e-01 9.56972063e-01
1.35614026e+00 3.72906804e-01 3.01221520e-01 -3.77851985e-02
1.59750402e+00 -8.78995299e-01 -7.73142993e-01 9.56421066e-03
4.50800449e-01 -2.85996050e-01 7.66639113e-01 3.15759569e-01
-1.06374991e+00 -4.43555981e-01 -9.97772098e-01 -1.49289846e-01
-2.22574189e-01 1.61124960e-01 9.16409671e-01 9.56432670e-02
-1.04633236e+00 1.14479527e-01 -8.38876724e-01 -1.40817642e-01
7.15821683e-01 4.03599679e-01 -4.42963898e-01 -6.27345815e-02
-1.18229568e+00 7.49158621e-01 5.89433551e-01 -1.86560024e-02
-9.96320307e-01 -6.85548425e-01 -1.15252376e+00 -3.74640077e-02
5.97699881e-01 -1.52891302e+00 1.31174517e+00 -6.11401677e-01
-1.25257754e+00 1.23816478e+00 9.23538730e-02 -7.47820735e-01
5.07407725e-01 2.50805527e-01 -5.07429242e-01 5.34400523e-01
2.69012332e-01 6.58032358e-01 9.99412954e-01 -1.34803462e+00
-2.61572838e-01 -3.74765337e-01 7.54125714e-02 -3.08072306e-02
2.48255789e-01 -4.37350631e-01 -2.83898354e-01 -1.09561467e+00
1.99076712e-01 -8.81125152e-01 -6.55986190e-01 5.36145195e-02
-6.62344217e-01 -2.19139278e-01 4.48514014e-01 -6.64913774e-01
1.03892004e+00 -1.93001068e+00 2.14479253e-01 3.38184327e-01
8.54510367e-01 5.30994050e-02 2.87105590e-01 1.35417208e-01
-3.18707734e-01 -2.40869448e-02 -6.54779673e-01 6.15068339e-02
-3.32492143e-01 3.95473927e-01 -4.43832040e-01 1.87396348e-01
5.98215520e-01 1.53136313e+00 -1.18440557e+00 -8.00015926e-01
2.51663297e-01 2.51310855e-01 -6.82426929e-01 3.73880684e-01
-3.05149853e-01 7.80006111e-01 -6.17705524e-01 7.18070388e-01
2.34984025e-01 -7.70108998e-01 2.50312001e-01 -2.25325033e-01
5.70592582e-01 5.99984778e-03 -4.92063582e-01 1.80261266e+00
-3.84581894e-01 1.03772245e-01 -2.64509767e-01 -9.86866772e-01
8.31800580e-01 3.35511684e-01 5.05530238e-01 -6.36920035e-01
2.75764227e-01 1.05998412e-01 -1.08328667e-02 -7.98580945e-01
6.34832457e-02 -7.70707726e-01 -3.02847743e-01 3.77939105e-01
9.34406146e-02 -5.83413064e-01 -2.17237771e-01 4.80559617e-01
1.48948455e+00 -1.09984688e-01 5.30905843e-01 3.16441983e-01
6.98657095e-01 3.98045540e-01 1.56436384e-01 6.46470785e-01
-8.48017707e-02 6.12880826e-01 6.26538336e-01 -3.50439757e-01
-8.85285974e-01 -1.81657076e+00 8.26615915e-02 3.68899971e-01
1.11806758e-01 9.30236951e-02 -7.03637481e-01 -1.04995298e+00
8.31510648e-02 8.66209626e-01 -7.64514923e-01 -5.78698158e-01
-4.80282336e-01 -5.10527194e-01 5.62618971e-01 7.51172364e-01
1.44475847e-01 -1.52046120e+00 -3.83373857e-01 1.69370845e-01
-3.69714387e-02 -1.13330424e+00 -1.78133637e-01 -5.99826872e-03
-8.69716108e-01 -1.33848882e+00 -8.21873009e-01 -7.94721127e-01
1.05707300e+00 -2.77151495e-01 1.20280540e+00 1.41889825e-01
-6.74677253e-01 4.71025676e-01 -1.93411767e-01 -5.34477949e-01
-1.08331263e+00 -2.92428434e-01 -2.49422178e-01 -1.39676094e-01
-2.09302250e-02 -3.71545851e-01 -6.93156183e-01 -3.57169598e-01
-1.28412700e+00 2.70018965e-01 9.39134479e-01 1.13057411e+00
8.71551394e-01 -7.41449930e-03 5.84255695e-01 -1.33404505e+00
8.24686050e-01 -7.47602463e-01 -1.29223019e-01 2.36557633e-01
-5.64757049e-01 1.73903644e-01 7.91152358e-01 9.03482884e-02
-1.00974691e+00 -3.83375050e-03 -2.96348989e-01 -6.82133853e-01
-2.57306725e-01 5.57387352e-01 3.36470932e-01 3.41655910e-01
8.10140014e-01 5.53843737e-01 2.78951675e-01 1.93208337e-01
4.40140903e-01 2.36225113e-01 1.17185962e+00 -1.99092522e-01
9.02660906e-01 7.43575454e-01 4.91166443e-01 -1.32905841e-01
-1.04126370e+00 -3.44101228e-02 -4.37845826e-01 -2.32189342e-01
1.37390804e+00 -7.62900412e-01 -8.76567900e-01 2.71686614e-02
-1.11336792e+00 8.09216797e-02 -5.54833114e-01 2.87685066e-01
-8.60244155e-01 1.56967491e-01 -7.63280869e-01 -3.16320240e-01
-6.17977619e-01 -1.32872236e+00 1.24809635e+00 9.32537541e-02
-3.41922760e-01 -1.15556550e+00 -9.48536247e-02 4.93463606e-01
-2.55681396e-01 8.64315271e-01 1.25858343e+00 -4.87382889e-01
-6.86058044e-01 -3.46435532e-02 -2.02034250e-01 2.18249664e-01
1.55300394e-01 -1.76800206e-01 -6.67409778e-01 -7.63539644e-03
2.60307211e-02 -1.78006485e-01 7.48144507e-01 4.32361364e-01
1.47980630e+00 -4.27119464e-01 -3.51894528e-01 6.73166335e-01
1.20090711e+00 3.34496260e-01 4.57892358e-01 9.75246653e-02
8.99145722e-01 4.90787417e-01 4.73361731e-01 1.47928759e-01
5.29433608e-01 8.84701759e-02 6.45535946e-01 -5.90456247e-01
-3.89360428e-01 -5.66268027e-01 -2.46396344e-02 7.60880589e-01
-6.45556599e-02 -7.82779381e-02 -1.12578487e+00 3.08431178e-01
-1.72077727e+00 -8.95675421e-01 -2.51190662e-01 1.54844701e+00
9.68958139e-01 2.31888518e-01 -1.07888073e-01 1.50637582e-01
4.62426871e-01 -1.37571409e-01 -5.76137483e-01 -4.32262212e-01
1.78308576e-01 6.25355959e-01 2.63319582e-01 1.60608634e-01
-8.77112865e-01 8.47478449e-01 5.89784765e+00 5.45755804e-01
-1.18480098e+00 2.25403920e-01 6.36009395e-01 2.94486970e-01
-6.21769071e-01 -5.31030059e-01 -2.48495921e-01 3.35006058e-01
7.25686431e-01 -2.20979080e-01 8.43605846e-02 8.15848768e-01
1.67550609e-01 1.09262034e-01 -1.08301091e+00 8.98984969e-01
8.06868598e-02 -1.94489062e+00 7.00745046e-01 -1.18753463e-01
8.03961873e-01 -4.14239734e-01 1.10574238e-01 1.62816480e-01
7.14160860e-01 -1.31491923e+00 5.70909381e-01 9.15877819e-01
1.00345933e+00 -7.58123696e-01 7.90934682e-01 1.84939504e-01
-7.07397997e-01 -1.51791703e-02 2.03641325e-01 3.82822603e-01
2.91287035e-01 3.85180235e-01 -1.46165919e+00 8.59662592e-01
4.07804072e-01 7.05629468e-01 -5.09483814e-01 7.06800461e-01
-7.13668704e-01 7.12605298e-01 2.22576201e-01 2.59750575e-01
3.15185815e-01 1.48515806e-01 5.34737647e-01 1.01710725e+00
7.74597824e-02 2.66879380e-01 1.88063517e-01 1.32077074e+00
-1.33049116e-01 -6.08590022e-02 -9.33409095e-01 -7.32141137e-02
-1.04978777e-01 1.10818923e+00 -1.08457553e+00 -5.90056837e-01
-1.02177463e-01 9.79315102e-01 1.65656552e-01 2.93630868e-01
-1.18520772e+00 -8.63328800e-02 1.37164846e-01 3.05268317e-01
-1.04366869e-01 1.03130676e-01 -7.16782808e-01 -9.18441713e-01
-2.34030724e-01 -8.46053660e-01 4.22120422e-01 -1.01513124e+00
-1.26162195e+00 7.41880476e-01 -3.05032700e-01 -1.31467402e+00
-6.18133307e-01 -6.99397385e-01 -5.79869151e-01 5.77728570e-01
-1.29889631e+00 -1.45653284e+00 -3.76643032e-01 7.40745246e-01
4.06157374e-01 -1.63181245e-01 8.91303837e-01 -1.94980053e-03
-1.76086500e-01 5.86136937e-01 -4.49882388e-01 3.49813551e-01
2.78770357e-01 -1.38483620e+00 1.91029504e-01 7.13954091e-01
-2.06511747e-02 5.81914783e-01 4.26113158e-01 -8.96535277e-01
-1.07103980e+00 -1.45467710e+00 4.77592915e-01 -3.71369392e-01
7.92219520e-01 6.40451312e-02 -9.06879485e-01 7.31644750e-01
-4.09999192e-02 6.89981207e-02 7.48569965e-01 -6.56744421e-01
-9.36312899e-02 1.49648175e-01 -1.26756918e+00 8.30164373e-01
1.12705469e+00 -5.77738464e-01 -8.68862748e-01 5.16965985e-01
1.21220040e+00 -6.94947720e-01 -1.17158520e+00 7.28400171e-01
1.28505394e-01 -7.22472012e-01 1.15893567e+00 -8.96664977e-01
1.18261206e+00 -4.54928011e-01 2.12475166e-01 -1.03615582e+00
-2.36771926e-01 -2.56916046e-01 -1.71269640e-01 6.70138597e-01
3.10535461e-01 -5.70459068e-01 5.93639076e-01 1.56614110e-01
-4.28661227e-01 -1.20492756e+00 -6.56743944e-01 -2.84083366e-01
-8.78841951e-02 -4.49199408e-01 4.90460455e-01 1.12544274e+00
-3.61120224e-01 1.80650219e-01 2.38569155e-01 3.69499713e-01
6.79056227e-01 1.39086366e-01 4.69995826e-01 -8.52263033e-01
-4.57479805e-01 -5.36401272e-01 -6.53911769e-01 -3.34414661e-01
2.33508945e-01 -1.59581244e+00 -2.44847238e-01 -2.03813267e+00
1.59534574e-01 -1.16797313e-01 -2.92941481e-01 6.78065836e-01
-2.88932621e-01 3.40670317e-01 1.09863050e-01 1.06650516e-01
-2.11175382e-01 2.53554225e-01 2.01479793e+00 -6.25733793e-01
8.10125545e-02 4.71431799e-02 -7.60352552e-01 8.88323665e-01
4.03781682e-01 -5.95192611e-01 -4.70433474e-01 9.21221599e-02
2.67051786e-01 5.96179843e-01 9.79277313e-01 -8.91224921e-01
1.54356182e-01 -2.19817802e-01 4.23414290e-01 -6.40795410e-01
-2.11338297e-01 -7.96004176e-01 1.95659369e-01 1.05554283e+00
-3.85752618e-01 -1.33098975e-01 7.34471977e-02 6.75179601e-01
-4.56523776e-01 -8.39751447e-04 5.25676608e-01 -4.90544677e-01
-7.19725788e-01 3.04217935e-01 -2.84628719e-01 1.97155662e-02
1.31283867e+00 -1.69434920e-01 -1.91854730e-01 -1.59284487e-01
-1.26852596e+00 1.03419609e-01 8.56172889e-02 4.31994170e-01
9.86264586e-01 -1.24115753e+00 -6.73544884e-01 2.13457286e-01
1.76634729e-01 5.53925216e-01 3.95542324e-01 9.29500461e-01
-9.22974586e-01 2.87954390e-01 -4.96472679e-02 -9.36585724e-01
-9.50246513e-01 7.43335366e-01 3.01509112e-01 -6.91181839e-01
-8.33357990e-01 8.03251863e-01 4.77201611e-01 -4.70847905e-01
-1.06332891e-01 -6.35353625e-01 -6.34378567e-02 -3.72109920e-01
2.71146446e-01 -2.36885101e-02 -5.47878407e-02 -3.16123039e-01
-2.24867061e-01 2.90002584e-01 -2.86978055e-02 1.82057962e-01
1.27332270e+00 2.56685227e-01 -3.16285282e-01 1.92652479e-01
8.98847461e-01 -1.64623588e-01 -6.89874709e-01 6.36600470e-03
-9.63531435e-02 2.93752491e-01 -1.70125365e-01 -1.01192296e+00
-1.13917637e+00 7.05111206e-01 2.95018911e-01 2.14363057e-02
1.21916997e+00 4.77556765e-01 8.13701034e-01 2.59105086e-01
3.83228213e-01 -4.41780120e-01 3.85696322e-01 9.83853340e-02
1.25431740e+00 -1.30572999e+00 -3.11333388e-02 -6.48731530e-01
-7.90417075e-01 1.07614946e+00 5.50312519e-01 -2.66601920e-01
5.54066181e-01 4.49008137e-01 1.18444175e-01 -6.89962745e-01
-4.00575191e-01 -1.59200281e-01 4.45179582e-01 5.34695566e-01
2.58007586e-01 3.96441609e-01 -3.01325738e-01 8.05730581e-01
-5.18579662e-01 3.70258987e-01 3.97850066e-01 8.60233724e-01
-1.02377892e-01 -6.57280982e-01 -5.24553001e-01 5.58014095e-01
-5.13898611e-01 -1.47788152e-01 -3.24382007e-01 7.47965455e-01
3.78523886e-01 4.67356741e-01 -9.32446495e-02 -4.62021559e-01
3.37342799e-01 -7.69157782e-02 5.28635442e-01 -8.48773360e-01
-6.39614224e-01 -3.33684593e-01 -3.06818157e-01 -6.16316378e-01
-2.81060278e-01 -2.06992179e-01 -2.14138603e+00 3.54432352e-02
3.18426579e-01 -2.35139936e-01 4.06265944e-01 9.00236249e-01
2.88636297e-01 1.44996953e+00 2.78562278e-01 -2.47928500e-01
-4.46554571e-01 -5.14116764e-01 -4.58285362e-01 7.10670233e-01
4.03162718e-01 -6.83979928e-01 3.75747941e-02 4.57740188e-01] | [15.026083946228027, -1.4274706840515137] |
35c9b6a6-1ddc-41cf-83d0-c3d52d9b0aed | out-of-distribution-detection-with-energy | 2302.12002 | null | https://arxiv.org/abs/2302.12002v2 | https://arxiv.org/pdf/2302.12002v2.pdf | Master's Thesis: Out-of-distribution Detection with Energy-based Models | Today, deep learning is increasingly applied in security-critical situations such as autonomous driving and medical diagnosis. Despite its success, the behavior and robustness of deep networks are not fully understood yet, posing a significant risk. In particular, researchers recently found that neural networks are overly confident in their predictions, even on data they have never seen before. To tackle this issue, one can differentiate two approaches in the literature. One accounts for uncertainty in the predictions, while the second estimates the underlying density of the training data to decide whether a given input is close to the training data, and thus the network is able to perform as expected.In this thesis, we investigate the capabilities of EBMs at the task of fitting the training data distribution to perform detection of out-of-distribution (OOD) inputs. We find that on most datasets, EBMs do not inherently outperform other density estimators at detecting OOD data despite their flexibility. Thus, we additionally investigate the effects of supervision, dimensionality reduction, and architectural modifications on the performance of EBMs. Further, we propose Energy-Prior Network (EPN) which enables estimation of various uncertainties within an EBM for classification, bridging the gap between two approaches for tackling the OOD detection problem. We identify a connection between the concentration parameters of the Dirichlet distribution and the joint energy in an EBM. Additionally, this allows optimization without a held-out OOD dataset, which might not be available or costly to collect in some applications. Finally, we empirically demonstrate that Energy-Prior Network (EPN) is able to detect OOD inputs, datasets shifts, and adversarial examples. Theoretically, EPN offers favorable properties for the asymptotic case when inputs are far from the training data. | ['Sven Elflein'] | 2023-01-28 | null | null | null | null | ['medical-diagnosis'] | ['medical'] | [ 5.15034422e-02 2.36555219e-01 -2.16795560e-02 -2.07116246e-01
-5.14991581e-01 -4.14973617e-01 4.93865609e-01 2.26065382e-01
-4.83344138e-01 8.40191066e-01 -2.46162310e-01 -4.99995142e-01
-1.89966023e-01 -8.87102783e-01 -9.83756065e-01 -1.01612663e+00
1.16038471e-02 4.09386396e-01 2.64414430e-01 2.47942865e-01
2.33575944e-02 7.69128859e-01 -1.59120870e+00 -1.19417086e-01
6.81923866e-01 1.24390352e+00 -3.22784856e-02 5.13453066e-01
1.13533631e-01 2.72876441e-01 -9.78938758e-01 -2.46125773e-01
1.93481594e-01 -2.20351815e-02 -3.37148845e-01 -4.08345819e-01
1.26872743e-02 -4.40083861e-01 -2.70735949e-01 1.17836988e+00
6.87503695e-01 5.11675775e-02 1.20198739e+00 -1.46373713e+00
-1.92045286e-01 1.51831076e-01 -3.86513114e-01 2.68046916e-01
-7.86057934e-02 1.58021659e-01 4.07815486e-01 -6.81893051e-01
3.38459373e-01 9.66821969e-01 6.79617286e-01 6.91647947e-01
-1.10669947e+00 -6.75847352e-01 -2.38690138e-01 -7.83272311e-02
-1.38177872e+00 -3.53840947e-01 6.02601230e-01 -5.15193760e-01
7.56356776e-01 -4.81811911e-02 2.22280547e-01 1.50756955e+00
5.76465964e-01 4.05860186e-01 9.33624089e-01 -3.92546296e-01
6.69535220e-01 5.99570870e-01 -5.72816730e-02 3.57156038e-01
4.54817027e-01 4.12660539e-01 -4.31215227e-01 -3.18522096e-01
4.24195170e-01 -1.28188565e-01 -2.75986463e-01 -1.69360861e-01
-4.80158061e-01 6.97457373e-01 1.91553682e-01 2.24979043e-01
-3.07532549e-01 2.90887561e-02 2.85381287e-01 1.52766304e-02
5.37183642e-01 1.26556873e-01 -2.68784583e-01 -9.29221883e-02
-9.69536126e-01 -8.96068215e-02 8.83780301e-01 6.00048184e-01
5.40069401e-01 3.31130736e-02 1.03323450e-02 5.04161477e-01
3.56263012e-01 4.79141027e-01 3.11203331e-01 -5.74211061e-01
2.61772066e-01 3.11098844e-01 9.84664634e-02 -8.99571717e-01
-4.09754872e-01 -5.29870868e-01 -1.10831606e+00 4.17629123e-01
5.89422166e-01 -4.90761995e-01 -9.50394332e-01 1.73457682e+00
4.67168599e-01 1.47203937e-01 2.58582056e-01 7.25576758e-01
4.74745184e-01 5.98240852e-01 -1.40468497e-02 -6.32543042e-02
1.14034069e+00 -1.05814368e-01 -6.03504539e-01 -2.75171012e-01
4.93261904e-01 -2.88013160e-01 7.11749673e-01 5.46064734e-01
-7.19048560e-01 -4.36757088e-01 -1.25407827e+00 3.89889270e-01
-5.53158164e-01 -3.91021743e-02 2.54182607e-01 9.95521545e-01
-7.21338809e-01 8.56410861e-01 -9.02163982e-01 -1.72157168e-01
4.90940511e-01 4.69904065e-01 -7.92088956e-02 1.03752628e-01
-1.48448765e+00 9.26362753e-01 4.00955498e-01 1.58946350e-01
-1.15259731e+00 -6.49561107e-01 -7.03501165e-01 3.64023931e-02
2.07726926e-01 -3.70342433e-01 8.56625617e-01 -6.71535552e-01
-1.27630234e+00 6.58867180e-01 7.35786483e-02 -8.34876478e-01
8.03850174e-01 -2.95794681e-02 -4.40180987e-01 -1.39544755e-02
-2.37407297e-01 7.04416037e-01 1.14431775e+00 -1.15843213e+00
-4.33946311e-01 -3.67502630e-01 -1.02947190e-01 -1.14371136e-01
-3.99844617e-01 -4.16014880e-01 -1.23159498e-01 -3.60248685e-01
-2.45022580e-01 -8.12912583e-01 -1.80106554e-02 -9.53782722e-02
-6.13258541e-01 -2.24260420e-01 8.32606256e-01 -4.25055474e-01
1.00406730e+00 -2.51184893e+00 -3.95104557e-01 5.23892701e-01
3.54390703e-02 2.78124213e-01 3.71740878e-01 1.82531193e-01
6.01176545e-02 2.17812091e-01 -3.44006419e-01 -3.69818926e-01
2.81157106e-01 4.89542186e-01 -4.98709083e-01 7.80993819e-01
2.90089965e-01 4.04494584e-01 -5.41322470e-01 -2.80261874e-01
1.18301675e-01 8.35165441e-01 -2.35226482e-01 1.37547016e-01
-1.14871211e-01 4.37269121e-01 -2.29663208e-01 4.89480883e-01
7.95964003e-01 -1.12477146e-01 -1.09943368e-01 -1.33097656e-02
2.27747664e-01 1.41086966e-01 -1.19456577e+00 1.09370768e+00
-3.89331639e-01 7.23574281e-01 1.48336694e-01 -1.05382371e+00
9.31090772e-01 3.48208547e-01 2.14704305e-01 -6.31448090e-01
3.98734331e-01 2.89275348e-01 -2.10441872e-02 -3.65641654e-01
3.89216542e-01 -3.42656821e-01 -1.95020884e-02 3.60368997e-01
-9.10743512e-03 1.06851690e-01 -1.99437276e-01 -9.30629820e-02
1.09277046e+00 -2.66344994e-01 -5.96962422e-02 -3.45089465e-01
1.80463880e-01 -3.00794482e-01 5.89373767e-01 8.25290203e-01
-2.87325323e-01 4.48889405e-01 7.61408269e-01 -1.68757349e-01
-9.15427983e-01 -1.12086093e+00 -6.90526724e-01 5.42558074e-01
2.27609456e-01 1.28322244e-01 -1.05553889e+00 -7.48425305e-01
1.88468382e-01 8.72773528e-01 -4.91790742e-01 -4.60248530e-01
-7.60877430e-02 -9.87900078e-01 8.42828810e-01 5.59281588e-01
3.52963954e-01 -6.50151253e-01 -9.24156129e-01 -3.27399746e-03
1.64405048e-01 -1.17666018e+00 2.87289871e-03 6.66974902e-01
-6.64871573e-01 -9.64601874e-01 -5.45703888e-01 -2.65052944e-01
7.04832256e-01 -3.99585098e-01 9.66779053e-01 -9.74272266e-02
-3.84949476e-01 4.28950965e-01 1.34033188e-02 -8.16390872e-01
-8.47934067e-01 -2.30970569e-02 4.22090590e-01 -4.89437580e-02
5.32760739e-01 -4.39313561e-01 -6.57208800e-01 4.17919695e-01
-1.05632210e+00 -6.01743937e-01 4.55778927e-01 6.80460513e-01
4.12836373e-01 6.31513119e-01 8.14112961e-01 -6.58751547e-01
5.98893762e-01 -7.49662638e-01 -5.96739411e-01 7.81283621e-03
-6.68551207e-01 2.02503189e-01 7.32164383e-01 -4.64110285e-01
-7.71402359e-01 -2.62147337e-01 -4.04518455e-01 -6.69154406e-01
-5.72733879e-01 1.57351390e-01 -2.52664953e-01 1.10189073e-01
8.04674387e-01 -9.21815559e-02 1.17793076e-01 -2.86046952e-01
-1.82753533e-01 8.90325427e-01 4.44946378e-01 -5.81717551e-01
6.35982394e-01 5.19062579e-01 2.74757415e-01 -9.26065981e-01
-7.27993608e-01 -1.81266353e-01 -3.59785974e-01 -2.23105878e-01
8.82602632e-01 -6.59453392e-01 -6.79203153e-01 5.34122407e-01
-9.77453351e-01 -2.82528162e-01 -4.17141706e-01 4.82265085e-01
-2.21880913e-01 3.01572114e-01 -3.91398519e-01 -1.16754186e+00
-1.22709543e-01 -1.19791687e+00 9.02551949e-01 2.47872695e-01
-1.72763839e-01 -1.34383893e+00 -2.19603226e-01 8.52221549e-02
3.38053107e-01 3.89285892e-01 1.07770324e+00 -1.08313239e+00
-1.55939981e-01 -4.63601440e-01 8.34582821e-02 6.40470922e-01
-5.59346601e-02 -6.05200119e-02 -1.46096051e+00 -4.19191450e-01
3.10687274e-01 -2.16049433e-01 7.30436921e-01 5.85988820e-01
1.22896862e+00 -9.18947011e-02 -4.37567145e-01 4.12918448e-01
1.17174470e+00 1.79588765e-01 5.69479644e-01 1.25667557e-01
1.60993949e-01 7.32290685e-01 4.06528056e-01 4.89759743e-01
-9.80511960e-03 3.57820630e-01 8.18486333e-01 6.84525445e-03
2.47691229e-01 -2.40502849e-01 4.33625668e-01 6.19786121e-02
4.16135818e-01 -6.39642119e-01 -9.73182678e-01 4.82636362e-01
-1.33514512e+00 -5.54896116e-01 3.50914635e-02 2.42026544e+00
6.05780423e-01 5.51874459e-01 -5.59797278e-03 3.57034564e-01
7.94194877e-01 -1.71790078e-01 -9.62366641e-01 -4.23197359e-01
1.13169625e-01 1.46749556e-01 7.81201303e-01 2.60860026e-01
-1.08807361e+00 2.24596441e-01 6.38588333e+00 1.06269872e+00
-1.10123301e+00 1.03575654e-01 9.70012128e-01 3.35242935e-02
-1.93431303e-01 -3.44444245e-01 -1.12686801e+00 7.84104109e-01
1.18213260e+00 2.94993013e-01 3.47904190e-02 8.17008555e-01
-4.56755608e-03 -4.65778172e-01 -1.27049172e+00 8.90097082e-01
-1.02077320e-01 -8.68610740e-01 -3.44642013e-01 4.85800743e-01
4.83861834e-01 -1.13461003e-01 3.54936481e-01 1.92812160e-01
1.39892817e-01 -1.22769284e+00 5.35091698e-01 5.88420153e-01
7.38080323e-01 -8.46201003e-01 1.06442523e+00 8.78124535e-01
-6.07917488e-01 -7.39322901e-02 -4.89228874e-01 2.11647227e-01
-4.56689335e-02 1.12257481e+00 -1.03711438e+00 3.53917271e-01
6.65330052e-01 7.08584487e-02 -1.67595953e-01 8.30414951e-01
-1.69604495e-01 6.38691604e-01 -7.48584032e-01 -3.09262634e-03
1.29440695e-01 9.67407003e-02 6.70912087e-01 9.91130233e-01
5.27697563e-01 -3.92468959e-01 -1.81303799e-01 1.10795927e+00
-1.24861531e-01 -5.08407176e-01 -6.85457587e-01 1.11402273e-01
6.18276834e-01 9.70659077e-01 -7.56072581e-01 -2.99794357e-02
7.50959094e-04 6.64177954e-01 -8.66493136e-02 3.26143235e-01
-9.33224857e-01 -2.30359137e-01 6.77961707e-01 3.18610847e-01
5.77560402e-02 5.58799803e-02 -2.01775894e-01 -6.97460115e-01
1.17270462e-01 -5.50054491e-01 3.49154115e-01 -4.58331436e-01
-1.49578846e+00 3.48458052e-01 1.58313766e-01 -1.00821567e+00
-2.58181125e-01 -9.44444180e-01 -5.50239265e-01 8.86116982e-01
-1.42169785e+00 -3.45166951e-01 -1.13261774e-01 3.44357729e-01
1.29103824e-01 -6.66089281e-02 6.79826379e-01 3.74668956e-01
-7.02337027e-01 6.77244425e-01 3.22677881e-01 1.46558851e-01
6.36156321e-01 -1.23768234e+00 1.60806865e-01 7.70857930e-01
-7.03737512e-02 3.31939340e-01 8.81602228e-01 -5.69708467e-01
-1.06468964e+00 -1.05797470e+00 2.40173250e-01 -4.29252774e-01
4.04423058e-01 -4.41352129e-01 -1.02023828e+00 1.43844977e-01
-9.40951779e-02 1.10167421e-01 6.69527769e-01 -1.78591818e-01
1.31107569e-02 -1.50609553e-01 -1.56679630e+00 1.28919020e-01
4.83091742e-01 -5.86003602e-01 -4.04084384e-01 2.98631698e-01
2.54223347e-01 -5.53391397e-01 -8.60403180e-01 5.44081688e-01
3.33482832e-01 -1.17085016e+00 8.07681620e-01 -2.67540038e-01
8.65351930e-02 -1.16911016e-01 -1.23516180e-01 -1.24775136e+00
3.40376318e-01 -4.11220431e-01 -3.79648477e-01 1.39274108e+00
3.24276030e-01 -8.07227194e-01 9.24093246e-01 7.73469150e-01
1.12248491e-02 -8.26060295e-01 -1.31150019e+00 -9.64188457e-01
2.80231267e-01 -6.66791320e-01 5.18981159e-01 6.06118083e-01
-3.55715722e-01 -1.11288473e-01 -1.94877863e-01 6.41722083e-01
6.14276946e-01 -4.88626033e-01 4.91915464e-01 -1.21498108e+00
-3.59603941e-01 -2.80645281e-01 -5.09339571e-01 -8.85936558e-01
2.31553212e-01 -6.64661586e-01 1.96719676e-01 -1.31046450e+00
-1.60794228e-01 -5.35522521e-01 -2.97932535e-01 2.09828109e-01
2.01571316e-01 8.17574188e-02 -2.67847121e-01 9.56051648e-02
-6.92642257e-02 3.50933969e-01 6.59589708e-01 -8.07805136e-02
-3.54324840e-02 2.40134761e-01 -2.78272599e-01 9.37249541e-01
8.52067411e-01 -7.59433925e-01 -4.64080334e-01 -4.36534211e-02
3.35886061e-01 -1.08604193e-01 6.83650553e-01 -1.25005352e+00
3.45202982e-01 1.89760879e-01 6.02047265e-01 -6.40075207e-01
3.07096928e-01 -1.08558130e+00 1.50781870e-01 3.72617185e-01
-1.86610967e-02 -2.55300224e-01 3.95384252e-01 8.31769288e-01
-5.76711632e-02 -7.12516963e-01 9.27446485e-01 1.94228098e-01
-2.99303293e-01 1.64981410e-01 -5.41686356e-01 1.77794740e-01
1.21706963e+00 -3.44109267e-01 -2.80976474e-01 -2.45979175e-01
-5.80784500e-01 1.96503848e-01 3.26821506e-01 8.87480080e-02
4.38767433e-01 -9.47230816e-01 -1.78161308e-01 4.53178644e-01
-2.29494765e-01 3.60465884e-01 3.22440326e-01 7.34445810e-01
-2.30824068e-01 2.31719539e-01 1.58976376e-01 -9.27556694e-01
-8.96600485e-01 4.76538599e-01 7.00683653e-01 -1.83116972e-01
-3.80052000e-01 7.67893791e-01 1.64457232e-01 -3.05472523e-01
6.72157168e-01 -3.58262628e-01 4.53068130e-03 1.94743395e-01
2.41981134e-01 2.72959083e-01 2.84915984e-01 -1.75873354e-01
-3.49889398e-01 2.75053650e-01 4.98777069e-02 1.06941685e-01
1.09871852e+00 8.03861842e-02 2.69003838e-01 5.94443798e-01
1.03787589e+00 -2.07812160e-01 -1.55580175e+00 8.33056346e-02
-8.86104554e-02 -1.18888646e-01 1.78945065e-01 -6.44283175e-01
-8.58889341e-01 1.25850558e+00 8.80301297e-01 6.24355018e-01
1.07688177e+00 -8.96033794e-02 7.25077868e-01 2.68720597e-01
1.77468628e-01 -1.32208741e+00 -5.47174178e-02 3.16967994e-01
3.57429832e-01 -1.19102609e+00 -1.07360587e-01 2.85833701e-02
-4.60942149e-01 1.11858034e+00 3.91424239e-01 3.06450799e-02
9.46342647e-01 6.92516625e-01 -2.17017278e-01 -1.32187933e-01
-6.49841309e-01 2.67577678e-01 2.41143107e-01 6.09610915e-01
-9.17272270e-02 -7.48273805e-02 4.37362373e-01 5.88826656e-01
-2.05449894e-01 -3.44015002e-01 4.07872498e-01 7.65986741e-01
-4.11064506e-01 -7.21837461e-01 -4.79867011e-01 4.39967453e-01
-7.92753756e-01 1.98782369e-01 -1.92383036e-01 8.13847959e-01
2.61852682e-01 8.11905324e-01 2.81247020e-01 -2.91750968e-01
2.54401714e-01 3.07726949e-01 8.12406242e-02 -3.47210854e-01
-2.69191921e-01 -1.20913073e-01 -1.46359220e-01 -2.02217579e-01
-2.67377138e-01 -5.48352480e-01 -1.25509858e+00 -2.08451942e-01
-5.17467856e-01 1.62062109e-01 9.66258049e-01 1.11774695e+00
4.08587098e-01 6.80845082e-01 5.20574808e-01 -6.85267806e-01
-9.46428418e-01 -8.15671742e-01 -7.70864069e-01 5.51261343e-02
5.36269426e-01 -8.60007226e-01 -9.33989346e-01 -2.86694258e-01] | [7.585549354553223, 3.6221511363983154] |
8a3e3f74-7f10-42ea-a30a-c4736be9c947 | model-based-monitoring-and-state-estimation | 2305.00252 | null | https://arxiv.org/abs/2305.00252v1 | https://arxiv.org/pdf/2305.00252v1.pdf | Model-Based Monitoring and State Estimation for Digital Twins: The Kalman Filter | A digital twin (DT) monitors states of the physical twin (PT) counterpart and provides a number of benefits such as advanced visualizations, fault detection capabilities, and reduced maintenance cost. It is the ability to be able to detect the states inside the DT that enable such benefits. In order to estimate the desired states of a PT, we propose the use of a Kalman Filter (KF). In this tutorial, we provide an introduction and detailed derivation of the KF. We demonstrate the use of KF to monitor and anomaly detection through an incubator system. Our experimental result shows that KF successfully can detect the anomaly during monitoring. | ['Peter Gorm Larsen', 'Cláudio Gomes', 'Hao Feng'] | 2023-04-29 | null | null | null | null | ['fault-detection'] | ['miscellaneous'] | [ 1.13679200e-01 -1.52772695e-01 2.86125720e-01 2.93565065e-01
3.59466672e-01 -3.77685249e-01 2.51743168e-01 1.94599037e-03
3.52766126e-01 7.00120747e-01 -6.69379056e-01 -5.09641469e-01
-1.98065504e-01 -7.00257003e-01 -5.07329583e-01 -7.69514441e-01
-3.93538445e-01 -4.64989960e-01 5.76337218e-01 6.90228716e-02
2.72877157e-01 8.94645810e-01 -1.56953228e+00 -3.58477756e-02
7.47296035e-01 1.33884251e+00 4.38280344e-01 9.49326634e-01
2.03596085e-01 6.81597710e-01 -1.16781056e+00 4.04102534e-01
4.97605093e-02 -1.53801456e-01 -3.27488393e-01 -2.29097709e-01
-2.11356953e-01 -5.18328488e-01 -2.86732972e-01 4.76378381e-01
3.72423559e-01 -2.09864199e-01 2.38154814e-01 -1.66329908e+00
1.37292504e-01 -1.32196039e-01 -1.03027567e-01 2.34702617e-01
7.38697469e-01 2.68175334e-01 -1.99492887e-01 -5.02249897e-01
2.50711858e-01 7.83510268e-01 8.15746963e-01 2.19338164e-01
-1.19641960e+00 -4.62336391e-01 -1.90906346e-01 -9.58947763e-02
-1.30395782e+00 -6.82834327e-01 2.50251353e-01 -3.60503972e-01
1.37287831e+00 4.44169998e-01 7.13379443e-01 7.69116223e-01
1.08350921e+00 3.31568450e-01 1.06463182e+00 -5.68892241e-01
3.81494373e-01 6.28537089e-02 -6.42599538e-03 7.20885992e-01
5.90145707e-01 2.16830239e-01 -7.21308887e-01 -4.09198821e-01
1.32258654e+00 -3.91136110e-02 -5.45364320e-01 -3.49780619e-01
-9.91174996e-01 -3.07234198e-01 -3.85432094e-01 3.56687307e-01
-5.16636848e-01 4.38455611e-01 3.17104697e-01 4.39114124e-01
8.63365233e-02 5.54979205e-01 -5.68368196e-01 -7.07501709e-01
-2.76362121e-01 -2.96604991e-01 1.03638065e+00 8.75503838e-01
3.08799952e-01 3.73008609e-01 6.47638738e-02 -6.48504123e-02
3.72783303e-01 5.10339856e-01 4.11503434e-01 -1.14152348e+00
-3.01848769e-01 7.46280909e-01 5.59477210e-01 -5.85370958e-01
-4.53853846e-01 -2.35031262e-01 -4.85841662e-01 5.35837710e-01
4.41254638e-02 -3.82586837e-01 -7.91479349e-01 1.09429061e+00
5.09041965e-01 8.26220810e-01 2.20028862e-01 1.89189147e-02
5.21542788e-01 3.92456949e-01 -6.81225479e-01 -7.13535488e-01
1.03216791e+00 -1.90788180e-01 -1.46007478e+00 1.17992483e-01
4.83178109e-01 -5.82100868e-01 4.21135813e-01 6.90863490e-01
-7.67867863e-01 -3.74071360e-01 -1.29337299e+00 7.36641288e-01
-4.76015389e-01 -9.45217349e-03 5.93265653e-01 8.21577370e-01
-1.20371711e+00 1.01171279e+00 -1.48864579e+00 -7.35791326e-01
-2.35454097e-01 3.99121106e-01 -3.43039066e-01 5.00443220e-01
-9.71613824e-01 1.43422306e+00 8.36219043e-02 2.35232100e-01
-8.10750723e-01 -6.29569054e-01 -8.56244981e-01 -5.22736609e-02
3.61633599e-02 -3.58802736e-01 1.35970032e+00 -1.47534743e-01
-1.99978590e+00 1.65757284e-01 -1.03369981e-01 -1.55946150e-01
1.95341676e-01 -4.21975195e-01 -8.38665664e-01 4.65437979e-01
-3.26573581e-01 -5.16406834e-01 7.79646158e-01 -7.96387255e-01
-6.29947364e-01 -1.17378533e-01 -2.75146604e-01 -3.62050742e-01
-4.45788771e-01 6.75139204e-02 -5.20123765e-02 -2.07331050e-02
2.41341367e-01 -5.71034849e-01 9.77228656e-02 -1.25122353e-01
-4.24324423e-02 1.53321624e-01 1.48005259e+00 -5.67577541e-01
1.41907299e+00 -2.08052230e+00 -6.26123071e-01 5.22718355e-02
1.14400901e-01 5.52093863e-01 4.80814993e-01 1.09766114e+00
4.63940948e-02 -2.92170286e-01 4.34304684e-01 -1.25975490e-01
-3.01977217e-01 3.44367683e-01 -2.73258388e-01 8.71271551e-01
1.56468496e-01 5.00454843e-01 -8.20691347e-01 7.36459196e-02
4.51235801e-01 3.53173882e-01 4.84693974e-01 4.49331850e-01
3.08396280e-01 5.09135664e-01 -4.39030915e-01 7.01316714e-01
5.37369788e-01 -2.26558849e-01 3.04340780e-01 -1.70925468e-01
-7.74924517e-01 3.59766752e-01 -1.21953011e+00 1.15230131e+00
-1.26115724e-01 6.38451099e-01 3.92036766e-01 -5.59551299e-01
1.01213515e+00 1.05705380e+00 4.88879740e-01 -4.62344408e-01
-1.46758119e-02 1.73711002e-01 -2.56103814e-01 -1.04777288e+00
2.82087624e-01 2.23553881e-01 3.27978373e-01 5.71308494e-01
1.19991740e-02 -1.70381367e-01 -1.61929205e-01 1.76814422e-01
1.64651072e+00 3.51831526e-01 4.14195299e-01 -3.67729634e-01
3.70872766e-01 -3.52796674e-01 8.46981645e-01 4.41613644e-01
-3.13516736e-01 -3.00669316e-02 5.46183884e-01 -2.75713623e-01
-3.42596203e-01 -7.66716063e-01 -1.51231706e-01 1.46537900e-01
3.65140468e-01 -5.69167733e-01 -3.07362080e-01 -2.21918091e-01
3.61916453e-01 7.07351089e-01 -4.62484032e-01 -4.21116859e-01
-3.19336087e-01 -2.27505535e-01 5.03051877e-01 6.81328416e-01
2.82363802e-01 -4.59155262e-01 -1.42377532e+00 2.83974677e-01
2.41645873e-01 -9.24357533e-01 2.23410189e-01 5.73091686e-01
-1.00480449e+00 -1.01714385e+00 7.36861769e-03 -3.88410121e-01
6.51860416e-01 3.62168878e-01 5.44977367e-01 1.86699033e-01
-4.61250067e-01 7.76923358e-01 -2.97421545e-01 -7.53717303e-01
-5.48720479e-01 -8.61752808e-01 6.27017558e-01 -2.23898664e-01
2.93810479e-02 -5.77905238e-01 -4.69537199e-01 4.73715037e-01
-4.74230915e-01 -8.10725614e-02 7.76173621e-02 4.16397572e-01
5.32030798e-02 2.70305544e-01 5.79163909e-01 -2.63034999e-01
4.97684449e-01 -1.52572721e-01 -1.10195279e+00 4.62943912e-01
-1.07569814e+00 5.96290408e-03 3.90030950e-01 -4.40234035e-01
-1.21188235e+00 3.74906026e-02 3.85867149e-01 -5.25106966e-01
-4.80111092e-02 3.78660589e-01 5.11027202e-02 -4.75789964e-01
3.50843400e-01 1.73581138e-01 4.72525656e-01 -6.31035566e-01
-1.75856531e-01 7.92293072e-01 6.37910366e-01 -1.73539475e-01
3.96238893e-01 4.08513874e-01 2.74712890e-01 -6.88506961e-01
5.62112220e-02 -4.21045810e-01 -5.17382205e-01 -7.46774197e-01
2.85611838e-01 -5.59392273e-01 -1.08736503e+00 7.51872838e-01
-1.03916383e+00 -4.93193656e-01 -8.40625465e-02 5.81199646e-01
-3.50767225e-01 4.08564329e-01 -7.78299093e-01 -1.27740121e+00
-3.78464192e-01 -7.51252770e-01 7.15859115e-01 4.65197653e-01
-2.40497366e-01 -9.86290693e-01 2.65803397e-01 -3.42033178e-01
6.98839009e-01 6.65077090e-01 1.59528852e-01 -1.22426577e-01
-4.64114308e-01 -5.82725644e-01 4.29438174e-01 1.71973571e-01
7.72789359e-01 7.31629193e-01 -1.36129737e+00 -7.47316122e-01
4.24275666e-01 2.97178924e-01 1.04203254e-01 1.89158574e-01
5.59377491e-01 6.35448424e-03 -9.22500134e-01 5.43144941e-01
1.27680719e+00 5.89855015e-01 6.54436767e-01 1.86613172e-01
4.55181226e-02 1.16056263e-01 1.01097035e+00 7.08952904e-01
-5.68992235e-02 4.42695022e-01 5.54699838e-01 -4.44623567e-02
3.51657569e-01 8.71283039e-02 7.09342897e-01 5.08851647e-01
3.66564915e-02 -9.06598866e-02 -7.45636225e-01 1.99563950e-01
-1.81298137e+00 -5.65183103e-01 -3.84271622e-01 2.49757648e+00
4.06672627e-01 3.48073505e-02 -2.88353443e-01 3.91194850e-01
6.74686432e-01 -5.54947555e-01 -5.24224281e-01 -5.58023870e-01
4.59767491e-01 9.53198969e-02 6.72612488e-01 8.37819427e-02
-8.24315369e-01 2.52429843e-01 8.12443829e+00 1.84472930e-02
-1.23355389e+00 -1.97851419e-01 -1.76002726e-01 2.20458925e-01
4.09747779e-01 -1.52094215e-02 -4.32763398e-01 4.01068777e-01
1.42591977e+00 -4.54153985e-01 -1.09769879e-02 4.36670512e-01
6.54952645e-01 -7.61779428e-01 -1.18188906e+00 6.89822018e-01
-3.33350539e-01 -9.78824139e-01 -6.67740345e-01 -1.41394613e-02
3.73927891e-01 -3.63832593e-01 -3.91061574e-01 -1.71121210e-01
1.73591092e-01 -5.86519122e-01 4.89101350e-01 8.88908505e-01
8.31441522e-01 -4.47240919e-01 6.99681997e-01 6.08282626e-01
-1.27061903e+00 -2.91256666e-01 -6.94117025e-02 -6.31805003e-01
2.48429269e-01 7.89253235e-01 -1.04164433e+00 8.51670504e-01
6.81552470e-01 2.24675164e-01 -6.39072284e-02 1.27781975e+00
-1.16617590e-01 3.10675234e-01 -5.08787274e-01 1.84552208e-01
-6.76648617e-01 -1.06535286e-01 4.50424552e-01 8.31259429e-01
6.92209423e-01 1.99412033e-01 -2.56634474e-01 6.57536447e-01
6.94439113e-01 -6.32503450e-01 -8.19915414e-01 -2.16262460e-01
6.81799650e-01 1.22679400e+00 -6.19163454e-01 -4.65161830e-01
-4.29288864e-01 9.95741844e-01 -2.48954922e-01 2.06128240e-01
-4.76104349e-01 -6.56973481e-01 9.27576482e-01 -1.21743985e-01
2.14091957e-01 -5.81690788e-01 -2.16527447e-01 -7.90486753e-01
-7.47486129e-02 -4.55016851e-01 2.31220171e-01 -8.84577155e-01
-5.12718201e-01 4.24220674e-02 -2.28207454e-01 -1.45351613e+00
-1.81853458e-01 -4.69885290e-01 -8.68312418e-01 9.41790044e-01
-9.98744071e-01 -4.92343664e-01 -5.53274453e-01 2.96840847e-01
-1.85906529e-01 3.75181168e-01 1.05253530e+00 6.02420308e-02
-8.78997445e-01 3.42658907e-02 4.30545688e-01 -4.10575628e-01
7.26107657e-01 -1.24145734e+00 4.73656863e-01 1.27208340e+00
-8.33203375e-01 8.66654873e-01 1.06340468e+00 -9.80072856e-01
-2.09536552e+00 -6.79308474e-01 4.55440551e-01 -4.84661043e-01
6.60416007e-01 -1.55646503e-01 -8.14933062e-01 8.34593892e-01
1.27783135e-01 1.21770322e-01 5.35755873e-01 -4.30506259e-01
4.11676079e-01 -4.68117923e-01 -1.29612005e+00 1.90787032e-01
4.38527852e-01 -4.93225694e-01 -2.45637402e-01 2.28056252e-01
5.73564947e-01 -7.05246091e-01 -1.29332626e+00 5.06846905e-01
8.26573431e-01 -9.09125209e-01 4.74470288e-01 -3.69658619e-02
-7.34910250e-01 -8.06239903e-01 5.27283967e-01 -1.29882109e+00
-2.75004357e-01 -1.20710754e+00 -7.71020651e-01 1.21763587e+00
2.06105951e-02 -1.44819009e+00 1.88750103e-01 9.65556800e-01
-1.75581500e-01 -4.54129547e-01 -8.74777853e-01 -1.23130572e+00
-9.03828382e-01 -2.47762948e-01 6.21516883e-01 7.57116199e-01
1.05272079e+00 -3.28297883e-01 -1.10178463e-01 8.56696069e-01
4.06057149e-01 -1.60456002e-02 5.88537037e-01 -1.14558625e+00
-1.65583745e-01 3.37346762e-01 -7.12311029e-01 -6.45116746e-01
-7.85093606e-01 2.01849729e-01 1.20858386e-01 -1.58088994e+00
-3.82261485e-01 -1.53271914e-01 -5.16179383e-01 2.63201326e-01
-6.46269023e-02 -3.27738196e-01 -2.98531562e-01 3.67646627e-02
-2.77885139e-01 1.81822315e-01 8.43098521e-01 6.08264148e-01
-3.32364827e-01 1.87787086e-01 9.99542512e-03 3.87117118e-01
9.37848449e-01 -1.72474414e-01 -3.40312153e-01 -1.32900327e-01
1.45400375e-01 5.23916602e-01 2.62202591e-01 -1.22580659e+00
4.23053682e-01 -1.06078535e-01 4.83009517e-01 -6.06033325e-01
4.16755378e-01 -8.96447539e-01 8.65538239e-01 9.50224519e-01
4.47164148e-01 3.46376806e-01 4.04437035e-01 5.59004843e-01
1.24970213e-01 3.23344320e-02 5.02498269e-01 1.42285183e-01
-3.34853381e-01 -3.03479880e-01 -1.03951085e+00 -1.14824569e+00
1.46674287e+00 -6.32226229e-01 -8.26807141e-01 -2.50780582e-01
-4.08986390e-01 4.47914749e-01 8.34643424e-01 2.54769567e-02
5.95611036e-01 -8.34270597e-01 1.79877311e-01 7.04586625e-01
-1.38618767e-01 -2.05594286e-01 -8.83660093e-02 1.40999234e+00
-7.57452786e-01 4.10667449e-01 -2.99646229e-01 -5.83500445e-01
-1.53198993e+00 5.25655329e-01 5.89778841e-01 3.39663804e-01
-7.08907366e-01 6.03289127e-01 -1.99355826e-01 1.39655247e-01
1.32569656e-01 -6.26021445e-01 -1.26915798e-01 -4.33268785e-01
8.50684166e-01 8.87141287e-01 3.68052214e-01 1.50248900e-01
-5.29589117e-01 1.14892624e-01 5.05903900e-01 1.94959059e-01
1.19585788e+00 -3.52098823e-01 -3.61661822e-01 8.06190968e-01
4.44620550e-01 -2.09764183e-01 -1.43710017e+00 1.40810773e-01
1.73697453e-02 -5.47675908e-01 1.82646830e-02 -6.56097651e-01
-6.93643987e-01 5.33695817e-01 8.72409523e-01 7.16299653e-01
1.11314523e+00 -2.63483018e-01 4.85414773e-01 2.11926177e-01
6.71890378e-01 -1.11514544e+00 -2.13544816e-01 1.49434283e-01
6.07724965e-01 -2.59219408e-01 6.25088513e-02 -6.32645607e-01
-9.73335132e-02 1.54662168e+00 7.18115270e-01 2.51686811e-01
6.16250396e-01 1.09497106e+00 2.60233998e-01 -3.28912288e-01
-1.37111080e+00 2.31855720e-01 -5.82831502e-01 8.04119408e-01
2.44910300e-01 1.86259359e-01 -3.70007530e-02 1.96096543e-02
4.93459821e-01 4.88898039e-01 8.74767482e-01 1.79267585e+00
-8.15280557e-01 -8.10240030e-01 -7.86571801e-01 6.36898279e-01
-3.50074410e-01 5.80780149e-01 -6.93077967e-02 2.29945675e-01
-1.58899203e-01 1.31730378e+00 6.97542876e-02 -5.58599114e-01
4.51497644e-01 1.83923349e-01 4.61719871e-01 -2.90786088e-01
-3.66566002e-01 -8.00054967e-02 3.86351526e-01 -1.21363246e+00
-1.42471343e-01 -7.22200036e-01 -1.42476952e+00 -2.42482319e-01
-8.60391915e-01 3.90751734e-02 8.52069318e-01 8.05357337e-01
6.15628123e-01 7.80149162e-01 7.76454270e-01 -5.43184698e-01
-4.83202994e-01 -9.73592937e-01 -1.07672870e+00 -3.43029559e-01
6.39090538e-01 -9.06278193e-01 -5.95367908e-01 -9.86992419e-02] | [6.483114719390869, 2.489853858947754] |
9514edcf-a530-495e-911e-c7f215c9622d | the-whole-is-greater-than-the-sum-of-its-3 | 2305.07855 | null | https://arxiv.org/abs/2305.07855v1 | https://arxiv.org/pdf/2305.07855v1.pdf | The Whole Is Greater than the Sum of Its Parts: Improving DNN-based Music Source Separation | This paper presents the crossing scheme (X-scheme) for improving the performance of deep neural network (DNN)-based music source separation (MSS) without increasing calculation cost. It consists of three components: (i) multi-domain loss (MDL), (ii) bridging operation, which couples the individual instrument networks, and (iii) combination loss (CL). MDL enables the taking advantage of the frequency- and time-domain representations of audio signals. We modify the target network, i.e., the network architecture of the original DNN-based MSS, by adding bridging paths for each output instrument to share their information. MDL is then applied to the combinations of the output sources as well as each independent source, hence we called it CL. MDL and CL can easily be applied to many DNN-based separation methods as they are merely loss functions that are only used during training and do not affect the inference step. Bridging operation does not increase the number of learnable parameters in the network. Experimental results showed that the validity of Open-Unmix (UMX) and densely connected dilated DenseNet (D3Net) extended with our X-scheme, respectively called X-UMX and X-D3Net, by comparing them with their original versions. We also verified the effectiveness of X-scheme in a large-scale data regime, showing its generality with respect to data size. X-UMX Large (X-UMXL), which was trained on large-scale internal data and used in our experiments, is newly available at https://github.com/asteroid-team/asteroid/tree/master/egs/musdb18/X-UMX. | ['Yuki Mitsufuji', 'Shusuke Takahashi', 'Stefan Uhlich', 'Naoya Takahashi', 'Ryosuke Sawata'] | 2023-05-13 | null | null | null | null | ['music-source-separation'] | ['music'] | [-1.50093019e-01 -2.10427731e-01 1.67237431e-01 4.78172041e-02
-8.00030410e-01 -6.12128913e-01 3.18187982e-01 -2.54869133e-01
-4.99961585e-01 7.78540909e-01 4.92028892e-02 -1.69311136e-01
-6.19019687e-01 -4.92329478e-01 -7.04847217e-01 -8.31620336e-01
-1.60155222e-01 2.44223684e-01 1.53481558e-01 -1.63314655e-01
-3.52267057e-01 4.27652121e-01 -1.45165312e+00 1.10410914e-01
9.47189987e-01 1.21866286e+00 2.91461259e-01 4.38574910e-01
6.74875304e-02 7.26082325e-01 -6.07700050e-01 -2.73298413e-01
4.69431370e-01 -5.72522163e-01 -5.57787895e-01 -3.74589175e-01
1.73426747e-01 -9.77166593e-02 -4.54508483e-01 1.01517510e+00
9.14292097e-01 3.28640103e-01 2.80654073e-01 -1.45936227e+00
-1.54694989e-01 1.07152736e+00 -5.69321275e-01 2.73108572e-01
-3.12584847e-01 -7.84505159e-02 9.78334427e-01 -8.80426109e-01
2.88866639e-01 1.20576632e+00 1.01463974e+00 2.95064181e-01
-1.12634206e+00 -1.07985950e+00 -8.20413157e-02 2.00825691e-01
-1.41362441e+00 -4.61095452e-01 9.30048048e-01 -2.51401693e-01
4.80054885e-01 3.32386285e-01 5.10227084e-01 1.20278406e+00
-4.38989222e-01 7.83543050e-01 1.00213242e+00 -4.14652169e-01
2.19640553e-01 -1.20413937e-01 2.89036572e-01 4.87334430e-01
4.47934866e-02 2.34281898e-01 -6.89404488e-01 -1.73767284e-01
8.83324504e-01 -3.08954179e-01 -5.52166641e-01 -5.58650270e-02
-1.30364561e+00 6.91177249e-01 5.54481268e-01 4.58581030e-01
-2.26477444e-01 1.43506035e-01 5.87946594e-01 4.69736367e-01
3.93869579e-01 4.25030857e-01 -6.12149239e-01 -1.43314764e-01
-1.07573307e+00 2.52138317e-01 7.41653800e-01 7.25002587e-01
5.28820217e-01 4.46842670e-01 -5.74662797e-02 1.19069088e+00
8.21379125e-02 4.58851129e-01 8.19078565e-01 -1.11599088e+00
5.13615727e-01 2.85813183e-01 -2.86938488e-01 -8.86742055e-01
-5.37774205e-01 -1.01311004e+00 -1.10165620e+00 1.44410893e-01
5.31572759e-01 -5.85179865e-01 -6.69288814e-01 2.25608611e+00
2.65047967e-01 5.77819288e-01 1.78219214e-01 1.00829446e+00
9.43996549e-01 5.29931784e-01 -3.43942165e-01 4.88483049e-02
1.15830195e+00 -1.03643167e+00 -6.08142316e-01 -4.08349521e-02
4.73153979e-01 -7.88562000e-01 9.28038001e-01 5.50240338e-01
-1.12396252e+00 -8.40491772e-01 -9.50372398e-01 4.84139547e-02
-2.10851550e-01 6.56280279e-01 4.09282207e-01 2.32398883e-01
-9.53130662e-01 9.58055854e-01 -7.64389873e-01 -1.37508184e-01
4.22480136e-01 4.08523381e-01 -3.17743063e-01 3.52728099e-01
-1.37920845e+00 5.12233317e-01 5.23217440e-01 2.67205626e-01
-8.88904512e-01 -8.99536371e-01 -4.94786620e-01 2.65359253e-01
4.80641961e-01 -6.89280033e-01 1.06314921e+00 -1.18568492e+00
-1.65055537e+00 4.45835024e-01 2.38272697e-01 -5.88761449e-01
5.23566246e-01 -5.63924551e-01 -5.38823366e-01 1.77091420e-01
3.36300880e-02 6.91055715e-01 9.94243324e-01 -1.26770604e+00
-4.57283735e-01 -2.44027540e-01 -1.71657681e-01 1.23949163e-01
-2.87480325e-01 -1.64871782e-01 -3.81690830e-01 -1.12620199e+00
7.97210708e-02 -1.03057039e+00 1.31183058e-01 -3.11006624e-02
-6.95249975e-01 4.83072624e-02 8.38109314e-01 -9.64650989e-01
1.11069214e+00 -2.49271941e+00 3.67119581e-01 2.49666870e-01
2.44035423e-01 6.87188268e-01 -4.68925446e-01 4.03010696e-01
-3.71812403e-01 -1.13102444e-01 -4.41921741e-01 -5.82181334e-01
1.36601359e-01 1.34404331e-01 -4.87618536e-01 2.30154768e-01
-4.34428686e-03 6.82060301e-01 -5.74013829e-01 -2.82150686e-01
4.57335226e-02 5.21202564e-01 -4.91918564e-01 7.24468082e-02
-5.63899353e-02 4.36922282e-01 -1.27510697e-01 3.94841760e-01
7.49890745e-01 6.43176213e-02 6.96922466e-02 -4.00775909e-01
-1.08148180e-01 3.39914083e-01 -1.45451844e+00 1.87191701e+00
-4.74967718e-01 5.51900923e-01 5.30063272e-01 -9.97251689e-01
9.14776206e-01 4.17129815e-01 4.40176100e-01 -4.50262189e-01
2.55201370e-01 4.54862386e-01 1.65652633e-01 -1.94379762e-01
1.84866413e-01 -1.40728474e-01 8.28034356e-02 3.17004919e-01
5.05576730e-01 3.52734119e-01 1.92960322e-01 1.48991615e-01
8.55770648e-01 7.36526921e-02 3.86298150e-02 -3.77135634e-01
5.09848654e-01 -4.33225900e-01 9.11573648e-01 6.70060456e-01
-4.06051949e-02 5.94558477e-01 6.30619526e-01 2.07658499e-01
-6.83880091e-01 -1.09212303e+00 -2.01821819e-01 8.69651973e-01
9.47490185e-02 -4.47134823e-01 -6.28168583e-01 -4.10056174e-01
2.55715032e-03 5.13439715e-01 -3.39407444e-01 -1.52480960e-01
-5.44863522e-01 -7.36521363e-01 1.09185565e+00 4.81010586e-01
7.64554024e-01 -9.66651082e-01 -2.67653942e-01 2.20166311e-01
-4.18190509e-01 -1.03061473e+00 -4.43217605e-01 5.52663803e-01
-8.85868967e-01 -9.82566655e-01 -7.00865865e-01 -4.23244178e-01
2.46304408e-01 1.21732146e-01 7.67238200e-01 -8.45376700e-02
1.08720131e-01 7.63507560e-02 -4.00777936e-01 -4.40963924e-01
-2.02791199e-01 2.51767218e-01 3.22874159e-01 3.35165083e-01
-1.34703755e-01 -1.29042006e+00 -4.60235775e-01 3.63894343e-01
-8.68612826e-01 -9.10026729e-02 6.10134602e-01 8.51332605e-01
4.08172011e-01 2.23861590e-01 8.81319821e-01 -4.94247705e-01
5.39209127e-01 -4.58522737e-01 -5.64934850e-01 -8.01340267e-02
-2.94314116e-01 -6.64807186e-02 5.67423403e-01 -6.35886848e-01
-8.47746849e-01 -2.27625504e-01 -4.98631537e-01 -8.99028778e-01
3.43879424e-02 5.37390590e-01 -4.23677891e-01 7.56000057e-02
3.13492090e-01 2.48172134e-02 3.81082930e-02 -1.04217887e+00
2.73232102e-01 6.98614955e-01 8.01510930e-01 -5.74843109e-01
8.75996351e-01 4.05438185e-01 -7.18610883e-02 -8.63421202e-01
-8.40496063e-01 -3.95539522e-01 -5.33260226e-01 8.81937295e-02
6.60140932e-01 -1.08989286e+00 -6.43020928e-01 7.69359648e-01
-1.01424336e+00 -4.10783052e-01 -4.97933567e-01 8.46331358e-01
-1.77077547e-01 1.73823863e-01 -6.84082150e-01 -5.28809667e-01
-3.82885814e-01 -8.18093419e-01 8.19238961e-01 9.85027403e-02
-1.60927549e-01 -1.04283154e+00 1.34678379e-01 3.35882425e-01
2.70700812e-01 1.66578680e-01 9.31288540e-01 -8.67283762e-01
-1.86303109e-01 1.76302567e-01 -1.75723821e-01 1.01248121e+00
-1.13930523e-01 -1.86018601e-01 -1.37425506e+00 -2.46165320e-01
2.33148187e-01 -2.59959728e-01 1.08550990e+00 5.07849455e-01
1.10608149e+00 -2.14660928e-01 -5.24040647e-02 9.08790290e-01
1.37824118e+00 5.84342983e-03 7.13580668e-01 2.35507503e-01
8.44305277e-01 3.38984907e-01 2.31905520e-01 3.04615855e-01
-5.08178324e-02 8.99554610e-01 4.03833359e-01 -3.84833962e-01
-3.85126084e-01 -1.52098507e-01 5.28992653e-01 1.28501904e+00
-1.60150945e-01 -1.67269260e-01 -6.35178208e-01 4.40326542e-01
-1.87639177e+00 -8.63459527e-01 -1.56727970e-01 2.23912740e+00
9.43248332e-01 5.07832803e-02 3.27624589e-01 6.16166830e-01
6.46314621e-01 7.64216706e-02 -5.13182461e-01 -3.15295672e-03
-3.62062484e-01 4.72592086e-01 2.09073484e-01 4.44831014e-01
-1.05701685e+00 5.44083297e-01 4.61155987e+00 1.38330591e+00
-1.29384398e+00 5.06008506e-01 9.61574391e-02 -3.43362302e-01
1.07720424e-03 -5.72492853e-02 -6.88060880e-01 6.19853139e-01
8.67824852e-01 1.24495566e-01 5.32891929e-01 5.01017809e-01
1.84473351e-01 2.34806716e-01 -8.87630820e-01 9.86277342e-01
-2.43660852e-01 -1.11099815e+00 -6.73490018e-02 -1.26145780e-01
3.95669222e-01 2.76012450e-01 -1.66921183e-01 5.44126511e-01
2.64274031e-02 -5.71598411e-01 1.03433406e+00 4.19884682e-01
7.11307049e-01 -7.75573194e-01 7.24488437e-01 3.76763523e-01
-1.14580202e+00 -2.27736011e-01 -2.06891507e-01 4.72956374e-02
6.79281950e-02 9.81927931e-01 -1.27522275e-01 1.10291624e+00
7.68513858e-01 7.31315911e-01 -3.83343071e-01 9.65982795e-01
-2.39614353e-01 9.63188231e-01 -5.25557518e-01 5.62531173e-01
1.54416651e-01 -3.38702351e-01 1.07619929e+00 1.06432843e+00
4.17103380e-01 -3.74158144e-01 -5.77423312e-02 9.48347211e-01
-2.43805408e-01 -2.35180214e-01 -1.18562512e-01 1.47514820e-01
6.41842246e-01 1.35565066e+00 -5.66692770e-01 -3.50538641e-02
-1.25418887e-01 6.58002973e-01 1.61390185e-01 4.67562646e-01
-1.09967148e+00 -6.53658569e-01 8.09258580e-01 8.09306651e-02
6.39140785e-01 -1.77961383e-02 -6.65304214e-02 -1.06499970e+00
5.70165403e-02 -1.07581818e+00 2.74818987e-01 -8.02695155e-01
-1.27762842e+00 8.81221294e-01 -9.89731699e-02 -1.62474215e+00
1.06238641e-01 -4.15102839e-01 -6.44452274e-01 9.27744150e-01
-1.52248740e+00 -1.10176516e+00 -2.13260397e-01 8.13116431e-01
1.47537053e-01 -3.23398054e-01 7.10047662e-01 8.73101056e-01
-8.29115152e-01 6.83549821e-01 3.62680227e-01 2.50816762e-01
7.52047658e-01 -1.04173553e+00 -9.37236100e-02 8.55172515e-01
4.46177900e-01 4.97916996e-01 3.40987861e-01 -2.84858972e-01
-9.04306591e-01 -1.23432243e+00 5.71199298e-01 2.00784709e-02
6.99473143e-01 -5.81112564e-01 -9.88691986e-01 6.03052974e-01
1.09267354e-01 -1.50468349e-01 6.45261288e-01 7.66786486e-02
-3.21477801e-01 -5.22245049e-01 -7.60046899e-01 3.14882338e-01
9.51865852e-01 -5.36674559e-01 -5.40233552e-01 1.38790682e-01
8.34391534e-01 -3.91600668e-01 -8.14129293e-01 5.10562360e-01
4.13713992e-01 -1.00119746e+00 1.08908534e+00 -3.12482208e-01
3.35332185e-01 -4.16165620e-01 -2.46458754e-01 -1.32746625e+00
-1.97802931e-01 -8.45090806e-01 -1.90167025e-01 1.73983979e+00
3.04668665e-01 -7.90600061e-01 9.44956169e-02 -2.48677731e-01
-4.53364700e-01 -7.14979649e-01 -1.11765647e+00 -1.25295079e+00
1.14765219e-01 -5.92079699e-01 6.73641503e-01 1.06241488e+00
-3.63918334e-01 5.04053175e-01 -3.92127573e-01 2.88057119e-01
5.35250962e-01 1.83019787e-01 6.31570637e-01 -1.31974554e+00
-6.79483533e-01 -5.29968798e-01 -9.95471179e-02 -9.36668694e-01
1.06545642e-01 -1.19395912e+00 -3.18361670e-01 -1.11679840e+00
-2.78626293e-01 -7.01700091e-01 -4.97367918e-01 6.39330566e-01
-5.90020604e-02 2.92588621e-01 5.55691481e-01 5.36464036e-01
-3.23698878e-01 8.09628248e-01 1.03459084e+00 1.12543538e-01
-4.09371763e-01 2.32200295e-01 -6.98486745e-01 9.32945728e-01
8.45852196e-01 -6.61324263e-01 -4.93460864e-01 -4.01860118e-01
-1.70342829e-02 -9.89013631e-03 7.25550115e-01 -1.38041389e+00
6.99333847e-02 5.14213979e-01 1.00612134e-01 -5.13049901e-01
5.08084953e-01 -7.08654165e-01 3.28145355e-01 3.52070212e-01
-2.41769046e-01 -3.09830099e-01 4.75905001e-01 3.45126182e-01
-5.19908607e-01 -2.18532935e-01 8.50304484e-01 3.04716736e-01
-3.67327362e-01 5.37152514e-02 1.21382639e-01 2.02296078e-01
6.63417161e-01 5.41241281e-02 -3.26876283e-01 -3.03892910e-01
-8.89748752e-01 1.28372595e-01 -1.59448698e-01 1.94710955e-01
2.80743092e-01 -1.51814055e+00 -7.37465382e-01 1.76005006e-01
-3.37078601e-01 3.44329141e-02 3.50187451e-01 1.27827144e+00
-2.40733191e-01 1.60088032e-01 -2.00762331e-01 -4.54609007e-01
-1.18631029e+00 3.31504285e-01 5.01817584e-01 -4.36525762e-01
-7.17680514e-01 1.01227164e+00 2.73566186e-01 -6.95034385e-01
5.89063823e-01 -4.41716313e-01 -2.68686432e-02 2.37072274e-01
2.45653525e-01 6.44023955e-01 1.11716077e-01 -4.31598932e-01
-3.60619009e-01 5.18931627e-01 2.84639448e-01 -1.78223968e-01
1.40818679e+00 -1.69259179e-02 -2.99697012e-01 6.56943262e-01
1.19274151e+00 1.67136937e-01 -1.17518520e+00 -4.13546234e-01
-3.38150799e-01 3.63032855e-02 2.90314466e-01 -8.84105384e-01
-1.53518844e+00 1.00232315e+00 5.43620288e-01 8.34244564e-02
1.49922144e+00 -1.20326318e-01 1.06117988e+00 1.60488695e-01
8.36802572e-02 -8.53667617e-01 -1.08865805e-01 4.91418898e-01
9.89999592e-01 -7.18752921e-01 -1.47929892e-01 -2.19992876e-01
-4.62340713e-01 9.22359645e-01 3.85632843e-01 -5.97256608e-02
8.66961777e-01 4.47147667e-01 2.01950982e-01 -3.35991420e-02
-5.90105355e-01 -2.83112317e-01 2.51924574e-01 3.44751269e-01
5.40623665e-02 -1.64908066e-01 -1.25176966e-01 1.26427341e+00
-3.21810693e-01 4.02439609e-02 1.32967129e-01 6.64383292e-01
-1.53642828e-02 -1.12508368e+00 -5.23792028e-01 3.12684953e-01
-3.57434899e-01 -2.21040562e-01 -3.64466757e-01 9.18823421e-01
5.74864805e-01 7.66898990e-01 -1.24299943e-01 -5.60305893e-01
3.08524072e-01 4.91141826e-02 2.54705876e-01 -1.77738130e-01
-7.74326026e-01 2.84713000e-01 8.08065012e-03 -4.20171678e-01
-4.90517795e-01 -6.34982586e-01 -1.24818671e+00 -3.82678986e-01
-5.00721514e-01 2.91452527e-01 4.89878446e-01 7.61595488e-01
4.88145947e-01 8.95426810e-01 4.67777729e-01 -9.47319329e-01
-7.30491817e-01 -1.07842195e+00 -1.04920185e+00 2.41534919e-01
4.25320387e-01 -6.18712664e-01 -7.16273785e-01 8.04041475e-02] | [15.468769073486328, 5.468409538269043] |
28d0e110-9383-4e12-9aa4-b41c98c0409f | exploring-the-robustness-of-distributional-1 | null | null | https://openreview.net/forum?id=z2zmSDKONK | https://openreview.net/pdf?id=z2zmSDKONK | Exploring the Robustness of Distributional Reinforcement Learning against Noisy State Observations | In real scenarios, state observations that an agent observes may contain measurement errors or adversarial noises, misleading the agent to take suboptimal actions or even collapse while training. In this paper, we study the training robustness of distributional Reinforcement Learning~(RL), a class of state-of-the-art methods that estimate the whole distribution, as opposed to only the expectation, of the total return. Firstly, we propose State-Noisy Markov Decision Process~(SN-MDP) in the tabular case to incorporate both random and adversarial state observation noises, in which the contraction of both expectation-based and distributional Bellman operators is derived. Beyond SN-MDP with the function approximation, we theoretically characterize the bounded gradient norm of histogram-based distributional loss, accounting for the better training robustness of distribution RL. We also provide stricter convergence conditions of the Temporal-Difference~(TD) learning under more flexible state noises, as well as the sensitivity analysis by the leverage of influence function. Finally, extensive experiments on the suite of games show that distributional RL enjoys better training robustness compared with its expectation-based counterpart across various state observation noises. | ['Linglong Kong', 'Shangling Jui', 'Hengshuai Yao', 'Yingnan Zhao', 'Yi Liu', 'Ke Sun'] | 2021-09-29 | null | null | null | null | ['distributional-reinforcement-learning'] | ['methodology'] | [-7.74977803e-02 8.70564431e-02 -7.34805241e-02 -1.07998222e-01
-1.13889968e+00 -8.92659783e-01 4.48559076e-01 -8.45232308e-02
-7.66184092e-01 9.76410568e-01 -3.89526375e-02 -7.35235989e-01
-3.68414432e-01 -6.19087815e-01 -8.14729512e-01 -1.28345609e+00
-3.82897198e-01 2.71854281e-01 -2.71762572e-02 -2.59958468e-02
-1.60076898e-02 4.26480204e-01 -1.06117356e+00 -5.59237003e-01
8.54753971e-01 1.15289557e+00 -1.56878740e-01 7.51323640e-01
3.93224269e-01 1.22928488e+00 -9.15245295e-01 -4.81401652e-01
5.66286564e-01 -3.87124151e-01 -3.89828265e-01 -2.27473732e-02
1.66927963e-01 -6.94702923e-01 -8.08816195e-01 1.75730050e+00
8.07324708e-01 6.12366796e-01 5.34145415e-01 -1.39733732e+00
-3.78571838e-01 8.78973603e-01 -4.18592334e-01 1.71730727e-01
1.14006385e-01 7.47148991e-01 7.38375127e-01 -1.27193213e-01
2.68861473e-01 1.34104872e+00 5.95213950e-01 7.48971522e-01
-1.31959629e+00 -7.76498795e-01 2.51699954e-01 -3.32196727e-02
-1.14685488e+00 -1.41990528e-01 4.29217100e-01 -4.07111526e-01
5.99833786e-01 3.75144631e-02 2.49727607e-01 1.35212171e+00
4.49407458e-01 8.89582932e-01 1.53774726e+00 -1.34443212e-02
6.31972373e-01 7.30593279e-02 -1.33817881e-01 6.64929628e-01
2.77426481e-01 8.29803884e-01 3.18662524e-02 -2.64268905e-01
7.57014215e-01 7.59758949e-02 -3.61252546e-01 -5.28165638e-01
-8.81533265e-01 7.82378733e-01 -3.75415129e-03 -3.42835158e-01
-5.15639722e-01 5.88567793e-01 5.89154184e-01 6.67158782e-01
2.47555003e-01 2.69939840e-01 -3.27010185e-01 -5.93516886e-01
-5.69009066e-01 3.65589410e-01 8.79399002e-01 1.03927195e+00
5.08636713e-01 5.55457056e-01 -6.52689815e-01 9.83572602e-02
8.87872577e-02 1.07239556e+00 2.48478293e-01 -1.20804036e+00
7.74998724e-01 -6.55835420e-02 5.62042773e-01 -4.34695005e-01
-3.13758731e-01 -6.76338613e-01 -1.03043449e+00 4.44856107e-01
7.18802750e-01 -9.62841094e-01 -7.09063888e-01 2.40684271e+00
3.60966057e-01 4.20912713e-01 2.35979736e-01 5.94159603e-01
-1.19972698e-01 4.19711798e-01 -6.39173687e-02 -5.53648353e-01
6.47369623e-01 -4.07964140e-01 -8.71053874e-01 -3.23432051e-02
5.23108184e-01 -1.47856712e-01 1.05474257e+00 2.83304244e-01
-1.15557265e+00 -3.28953177e-01 -8.55280995e-01 7.57624686e-01
8.89393538e-02 -3.29693794e-01 1.92308947e-01 8.55864406e-01
-9.25360203e-01 1.01693940e+00 -9.98345852e-01 1.39585450e-01
3.32375735e-01 2.58073241e-01 -6.75649475e-03 9.40745324e-02
-1.32455897e+00 1.07996380e+00 3.83860230e-01 -1.30487522e-02
-1.65133011e+00 -8.62092614e-01 -9.42875147e-01 1.06822113e-02
9.70115900e-01 -5.45539260e-01 1.64823866e+00 -4.50493127e-01
-1.98275983e+00 1.23254135e-01 3.96800220e-01 -8.66836548e-01
9.62768197e-01 -2.62237489e-01 -2.04211533e-01 1.23861142e-01
-9.15242881e-02 -3.77400592e-02 1.06738186e+00 -1.06695318e+00
-6.09316111e-01 -3.52764577e-01 2.11890683e-01 2.17568457e-01
3.76273170e-02 -3.86624157e-01 1.22367799e-01 -5.26370227e-01
-5.58263600e-01 -9.79116678e-01 -6.41136348e-01 -4.87183869e-01
-5.77574372e-01 -1.03999555e-01 6.04963779e-01 -4.66250390e-01
1.42773128e+00 -2.25638223e+00 1.39572665e-01 3.81680906e-01
-6.74824566e-02 9.13907960e-02 -4.27730046e-02 5.95760703e-01
3.44610848e-02 -2.88152155e-02 -4.15179759e-01 -4.83063936e-01
6.50491595e-01 4.34167027e-01 -7.57290721e-01 9.80346918e-01
-1.83536336e-02 7.17641890e-01 -1.02501285e+00 -1.65581986e-01
2.07551464e-01 -8.35271999e-02 -3.46777827e-01 2.66346365e-01
-2.59255201e-01 6.20924830e-01 -6.07864261e-01 2.31318533e-01
5.40910423e-01 1.36611134e-01 1.69382859e-02 3.80317241e-01
7.30504747e-04 -3.44362259e-02 -1.40461576e+00 1.33023846e+00
-3.97130609e-01 2.55336285e-01 2.80716807e-01 -1.10344589e+00
4.52553988e-01 2.70894736e-01 4.55721825e-01 -2.88680017e-01
4.65209901e-01 -1.11942865e-01 -7.49748200e-02 -3.28073323e-01
2.60225862e-01 -6.35889411e-01 -5.66999316e-01 5.54612100e-01
1.69296682e-01 -2.19216153e-01 -1.24959312e-02 1.98935494e-01
1.14080274e+00 1.16035797e-01 2.52281785e-01 -2.45978013e-01
1.85558483e-01 -4.99748915e-01 5.33407032e-01 1.24911392e+00
-6.54003680e-01 -1.22530080e-01 8.80229831e-01 2.64221638e-01
-9.01045859e-01 -1.34423459e+00 5.71812466e-02 9.66678083e-01
7.24420473e-02 5.36326617e-02 -7.66069770e-01 -9.37354624e-01
2.53186226e-01 1.28570783e+00 -7.73758769e-01 -7.10739255e-01
-2.17181832e-01 -6.30604029e-01 8.94221842e-01 5.34334600e-01
5.70740938e-01 -7.36095071e-01 -6.42555952e-01 1.55083716e-01
2.25088865e-01 -8.60592544e-01 -7.46928632e-01 4.67878461e-01
-5.19826651e-01 -8.44799459e-01 -5.32349527e-01 -1.91138059e-01
3.39051604e-01 -4.10870045e-01 9.07852292e-01 -7.39763081e-01
3.87653410e-01 9.04158652e-01 -3.46021913e-02 -6.77832782e-01
-6.22472525e-01 -2.50526220e-01 5.60804844e-01 4.49397638e-02
-9.09553096e-02 -4.81800467e-01 -4.82252777e-01 1.88422680e-01
-9.47601855e-01 -7.38795817e-01 5.53613722e-01 8.76298308e-01
5.89284003e-01 3.67697775e-01 5.51480412e-01 -8.43879759e-01
9.85893369e-01 -4.07196403e-01 -1.16466975e+00 1.89210892e-01
-5.05546153e-01 3.77162367e-01 8.20421040e-01 -7.36616373e-01
-1.09080279e+00 -2.09962413e-01 5.76117188e-02 -9.26737964e-01
4.63339128e-02 3.42277795e-01 -1.65923312e-01 1.37886882e-01
5.43001711e-01 5.94515920e-01 2.30814606e-01 -2.86983326e-02
4.42279071e-01 2.70882308e-01 6.66746616e-01 -9.20332849e-01
1.11144853e+00 3.34249854e-01 3.08377355e-01 -5.06526172e-01
-1.05937099e+00 -2.98482776e-01 -7.53219053e-02 -1.16809934e-01
6.35492384e-01 -6.97709858e-01 -1.25823629e+00 4.94401753e-01
-5.91152966e-01 -8.48933101e-01 -1.07638884e+00 6.48446977e-01
-1.10076404e+00 4.34912741e-01 -7.00219929e-01 -1.50229514e+00
-6.81945980e-02 -1.23339391e+00 7.29287863e-01 5.97000532e-02
4.65466857e-01 -1.20950174e+00 3.82819921e-01 -3.13614666e-01
3.04780215e-01 5.42727828e-01 6.56891406e-01 -8.02039444e-01
-2.57546991e-01 -2.62186915e-01 2.94933647e-01 8.58771503e-01
-2.07519501e-01 -3.11744750e-01 -6.50876284e-01 -8.17095518e-01
4.22064155e-01 -4.04063821e-01 7.54822671e-01 6.84756517e-01
1.18456089e+00 -6.95831716e-01 1.14796765e-01 4.61952865e-01
1.48909509e+00 3.87896538e-01 5.32359421e-01 1.22236252e-01
4.00270134e-01 2.58988827e-01 8.34153771e-01 1.01555812e+00
2.00902298e-02 1.03754036e-01 7.16833770e-01 4.14625823e-01
5.75476885e-01 -5.05677164e-01 1.07241774e+00 5.82143188e-01
2.18212903e-01 -1.44206539e-01 -5.02428293e-01 2.68023998e-01
-1.76903176e+00 -1.32581103e+00 3.56431544e-01 2.70505238e+00
8.61856341e-01 2.71686196e-01 4.00089830e-01 -1.11948207e-01
6.74302340e-01 1.72570005e-01 -1.16877246e+00 -2.48325124e-01
4.72652875e-02 2.25019440e-01 1.19412565e+00 6.10463679e-01
-1.06864440e+00 6.67592943e-01 6.69641590e+00 1.13417912e+00
-8.64816844e-01 1.74990833e-01 4.26790833e-01 -1.17272206e-01
-8.80382359e-02 -2.49770150e-01 -7.61505604e-01 5.63412070e-01
1.28503919e+00 -4.04952019e-01 6.87997878e-01 9.71699178e-01
2.75029689e-01 5.57142496e-02 -1.09177554e+00 7.61107206e-01
-4.15976614e-01 -8.21407974e-01 -3.60015839e-01 3.75760317e-01
9.35152113e-01 1.69822693e-01 5.34979343e-01 8.81949484e-01
1.06856525e+00 -8.79687726e-01 8.11397672e-01 6.81919634e-01
9.33776200e-01 -1.17849803e+00 9.29556608e-01 5.38690388e-01
-8.65730286e-01 -4.92202491e-01 -4.47292209e-01 3.29127349e-02
8.14091507e-03 2.94636548e-01 -5.31343460e-01 5.22751272e-01
2.59908468e-01 4.77684259e-01 -7.09481165e-02 8.05282354e-01
-3.55009139e-01 8.87648225e-01 -3.48998368e-01 -1.25339329e-01
5.32117307e-01 -4.57761407e-01 8.82380068e-01 9.30801570e-01
2.73224562e-01 -3.81826870e-02 5.57586253e-01 6.94178641e-01
-9.84072983e-02 -4.61665183e-01 -7.38505423e-01 -9.67685282e-02
5.43812573e-01 8.35048676e-01 -1.03923716e-01 -3.39875847e-01
-1.89677224e-01 7.29970396e-01 2.53571600e-01 7.08535790e-01
-1.11168361e+00 -3.93987656e-01 1.04101181e+00 -4.07536745e-01
4.13867176e-01 -1.54513344e-01 8.50205421e-02 -1.03384900e+00
-9.16517675e-02 -9.06618118e-01 4.70994949e-01 -6.88642636e-02
-1.58986735e+00 1.82219878e-01 1.74441144e-01 -1.40078866e+00
-4.84808326e-01 -4.56718445e-01 -6.82630062e-01 7.20312893e-01
-1.23833811e+00 -3.85622412e-01 4.89164293e-01 9.58380342e-01
1.92987829e-01 -1.43787518e-01 4.22622442e-01 2.37326380e-02
-9.76538241e-01 8.50692034e-01 6.91306353e-01 1.28007561e-01
5.00128925e-01 -1.81943130e+00 -1.68598033e-02 9.96695578e-01
-2.67466158e-01 4.59473401e-01 1.09947407e+00 -4.77736890e-01
-1.50603640e+00 -1.34229767e+00 -3.65845621e-01 -4.99882191e-01
1.14919865e+00 7.13023636e-03 -6.22206509e-01 1.07545006e+00
-5.31917717e-03 -4.50079106e-02 3.48651946e-01 -2.99156934e-01
-1.01086482e-01 -9.59732533e-02 -1.39300573e+00 7.38730311e-01
7.05649495e-01 -6.68634593e-01 -4.04513329e-01 8.88298228e-02
9.86732960e-01 -4.89927679e-01 -1.07056391e+00 1.52694106e-01
1.72388673e-01 -7.75308967e-01 6.90241039e-01 -1.01302421e+00
8.68592225e-03 -2.45010436e-01 -2.88639516e-01 -1.75059986e+00
-1.16370693e-01 -1.25058389e+00 -4.64339226e-01 9.33088481e-01
2.12347656e-01 -9.47892904e-01 5.24770200e-01 4.17083651e-01
-2.49502242e-01 -7.82264292e-01 -1.23681879e+00 -1.25271499e+00
5.35235167e-01 -6.83817744e-01 3.90702218e-01 5.14144063e-01
5.43428548e-02 4.58695069e-02 -5.16780317e-01 5.06480753e-01
9.86629367e-01 -9.71273333e-02 6.65762782e-01 -5.95608294e-01
-7.31886387e-01 -5.38188100e-01 -2.28609547e-01 -1.16168547e+00
3.89782637e-01 -5.58509707e-01 4.47229326e-01 -9.67024863e-01
-1.80052835e-02 -1.56220138e-01 -6.97876871e-01 1.23186596e-01
-3.21110219e-01 -5.54635942e-01 3.50828826e-01 -1.75839126e-01
-1.00545537e+00 9.69198763e-01 1.48173773e+00 -2.09499262e-02
-9.56483632e-02 6.21046007e-01 -5.59586763e-01 7.04519629e-01
8.45980108e-01 -5.45198679e-01 -6.19393170e-01 1.99228093e-01
-2.00900584e-02 4.59431469e-01 2.51104325e-01 -9.87396121e-01
1.47191316e-01 -4.72807050e-01 -4.40427773e-02 -4.80629027e-01
1.42994776e-01 -7.50354350e-01 -9.73432958e-02 8.58271182e-01
-6.48591638e-01 1.64198317e-02 9.34111476e-02 1.17143166e+00
4.91725504e-02 -2.17858419e-01 9.59190190e-01 -4.36899327e-02
-2.41660327e-01 6.67849004e-01 -5.47422647e-01 6.09693587e-01
1.05106997e+00 3.14741582e-01 -2.47076333e-01 -8.46307814e-01
-6.24259293e-01 4.61506009e-01 8.08615014e-02 -2.99515098e-01
5.00084341e-01 -1.26644039e+00 -6.71849787e-01 -2.13106647e-01
-1.46875158e-01 1.20302578e-02 4.14977819e-01 1.05866015e+00
1.05971009e-01 8.69347081e-02 1.18794061e-01 -2.56615847e-01
-6.18946970e-01 9.30168509e-01 5.81832469e-01 -7.05500364e-01
-4.15744483e-01 7.62270451e-01 1.86272740e-01 -3.84688079e-01
5.82975030e-01 -6.37540102e-01 2.66766697e-01 -1.77325830e-01
5.64982831e-01 6.08232558e-01 -3.39407355e-01 -1.19489424e-01
-2.48912033e-02 -5.31099876e-03 1.81633100e-01 -5.22175372e-01
9.04978991e-01 -2.30949119e-01 4.11324382e-01 6.58167064e-01
1.00243521e+00 -1.92909986e-01 -2.13861156e+00 -3.92441183e-01
-1.59741610e-01 -2.37810224e-01 -1.06469005e-01 -7.76925325e-01
-1.00309169e+00 8.36868942e-01 6.71623230e-01 4.37810212e-01
9.80527937e-01 -2.75019586e-01 6.34195685e-01 5.25628865e-01
5.57632983e-01 -1.24542999e+00 -5.61376438e-02 6.92941666e-01
6.06754243e-01 -9.51230943e-01 -2.41203874e-01 3.18745047e-01
-1.00827861e+00 6.13141596e-01 3.54301780e-01 -3.27315927e-01
6.65327787e-01 3.03429425e-01 -2.36285985e-01 1.79675952e-01
-9.09698904e-01 -4.56341833e-01 -1.20035060e-01 8.08887899e-01
-3.92772913e-01 3.83985072e-01 1.58532590e-01 6.23037398e-01
-3.00604075e-01 -3.16965193e-01 6.50683165e-01 9.98221457e-01
-2.68120021e-01 -7.24404037e-01 -2.98330516e-01 4.76011813e-01
-6.63045943e-01 3.28042805e-02 1.73811525e-01 8.12365234e-01
-3.41974258e-01 1.02248561e+00 -2.39345822e-02 -3.86532426e-01
5.73966324e-01 2.50372719e-02 4.72124130e-01 -3.08876961e-01
-4.23038274e-01 -7.60153979e-02 -1.15234785e-01 -6.47726119e-01
-7.60426223e-02 -7.97638297e-01 -1.06253695e+00 -6.15493357e-01
-3.20613354e-01 1.24519087e-01 4.04087007e-01 1.02465653e+00
-3.49681750e-02 6.93736911e-01 9.74916101e-01 -5.85978687e-01
-2.08718991e+00 -1.04735863e+00 -1.09314442e+00 4.93502855e-01
6.38749659e-01 -6.90550029e-01 -7.12286532e-01 -5.78551590e-01] | [4.269867897033691, 2.5972611904144287] |
c09bff26-7e79-43b6-885f-9a20bd28ebfe | fairness-in-streaming-submodular-maximization-1 | 2305.15118 | null | https://arxiv.org/abs/2305.15118v1 | https://arxiv.org/pdf/2305.15118v1.pdf | Fairness in Streaming Submodular Maximization over a Matroid Constraint | Streaming submodular maximization is a natural model for the task of selecting a representative subset from a large-scale dataset. If datapoints have sensitive attributes such as gender or race, it becomes important to enforce fairness to avoid bias and discrimination. This has spurred significant interest in developing fair machine learning algorithms. Recently, such algorithms have been developed for monotone submodular maximization under a cardinality constraint. In this paper, we study the natural generalization of this problem to a matroid constraint. We give streaming algorithms as well as impossibility results that provide trade-offs between efficiency, quality and fairness. We validate our findings empirically on a range of well-known real-world applications: exemplar-based clustering, movie recommendation, and maximum coverage in social networks. | ['Jakub Tarnawski', 'Jakab Tardos', 'Ashkan Norouzi-Fard', 'Federico Fusco', 'Marwa El Halabi'] | 2023-05-24 | null | null | null | null | ['movie-recommendation'] | ['miscellaneous'] | [ 3.20415229e-01 2.97592372e-01 -7.90764570e-01 -8.20712864e-01
-7.01524675e-01 -5.35192788e-01 2.13347599e-01 5.67001224e-01
-4.95699197e-01 1.01876581e+00 1.42752126e-01 9.67610180e-02
-4.43131775e-01 -8.02952409e-01 -6.36228085e-01 -6.31475031e-01
-3.96245271e-01 6.43033504e-01 -2.44799197e-01 -5.05179055e-02
3.94477367e-01 3.90058368e-01 -1.30401754e+00 6.81193545e-02
1.04190660e+00 1.00289047e+00 -3.34554166e-01 1.99151829e-01
1.73253030e-01 4.82886344e-01 -3.91903698e-01 -8.25810373e-01
7.45728791e-01 -3.55867833e-01 -6.82521880e-01 4.50466186e-01
5.52941024e-01 -2.46049136e-01 2.05064211e-02 1.31165349e+00
4.86513793e-01 3.51504803e-01 6.77691042e-01 -1.85027504e+00
-3.55148375e-01 8.86569321e-01 -9.53192353e-01 2.63432711e-02
1.74465090e-01 -2.62857378e-01 1.47088063e+00 -4.70621824e-01
6.56715214e-01 1.19775093e+00 2.51975507e-01 6.07985437e-01
-1.31610847e+00 -4.69211906e-01 7.28371888e-02 6.87181875e-02
-1.43842828e+00 -5.31277597e-01 5.92495620e-01 -2.33993188e-01
1.13214508e-01 8.16290259e-01 4.34000820e-01 5.08680224e-01
-5.76167554e-02 9.63258803e-01 9.52057362e-01 -1.32087290e-01
6.28891885e-01 4.44749266e-01 -2.22576950e-02 2.25285411e-01
7.66715825e-01 -2.94973522e-01 -7.12808430e-01 -7.09037602e-01
5.25590241e-01 1.03757031e-01 -2.66049773e-01 -7.70216823e-01
-8.81494939e-01 1.20732594e+00 1.73822284e-01 -2.84105718e-01
-3.24226290e-01 1.42703295e-01 1.54114902e-01 4.31796908e-01
6.21950626e-01 6.21659994e-01 -1.42863885e-01 2.46452108e-01
-1.25831234e+00 6.08018994e-01 9.67502296e-01 1.11233354e+00
6.61498487e-01 -2.04907060e-01 -9.71183404e-02 8.66329491e-01
9.50222388e-02 3.72471720e-01 -1.98780857e-02 -1.56179929e+00
4.52064872e-01 4.79417384e-01 4.15174067e-01 -1.13306129e+00
-2.68371046e-01 -2.43997753e-01 -8.31199408e-01 -3.01952921e-02
5.41800857e-01 -2.76106745e-01 -3.05250585e-01 1.97886288e+00
7.34068274e-01 -1.69766545e-01 -1.82508931e-01 1.00148392e+00
2.81386048e-01 4.53090876e-01 -2.50694603e-01 -9.08397079e-01
7.64858305e-01 -4.48423088e-01 -5.62422097e-01 1.71613201e-01
5.21258593e-01 -3.51999998e-01 7.15869546e-01 5.35511911e-01
-1.31382954e+00 2.78657824e-01 -7.03321099e-01 5.23702130e-02
1.82466209e-01 -5.56746244e-01 8.43546212e-01 9.86604571e-01
-7.08960772e-01 6.27243698e-01 -3.19208562e-01 -4.26301032e-01
7.41248250e-01 4.91783082e-01 -1.77237064e-01 -1.65791079e-01
-8.76720905e-01 3.48618746e-01 2.28401870e-02 -2.41514236e-01
-7.45476186e-01 -9.04052556e-01 -7.16280222e-01 1.62691265e-01
7.71525860e-01 -6.33596778e-01 9.95671928e-01 -1.01132989e+00
-9.48231339e-01 1.07340479e+00 -4.16950434e-02 -5.42822540e-01
8.39123428e-01 2.78476238e-01 -1.57794252e-01 -5.05941594e-03
1.53367966e-01 5.64890087e-01 6.77314341e-01 -1.15633428e+00
-7.83065856e-01 -7.95827866e-01 4.01559561e-01 4.29944038e-01
-7.88110435e-01 2.53029406e-01 1.41836300e-01 -4.83731568e-01
-1.34755686e-01 -9.14956868e-01 -6.78012967e-01 2.57332474e-01
-5.57158053e-01 -2.18059286e-01 2.52671063e-01 -1.42510340e-01
1.22473538e+00 -1.96440470e+00 2.59922326e-01 5.13031781e-01
2.07395881e-01 -3.75061959e-01 -6.11419277e-03 3.81569058e-01
4.55664217e-01 2.21930847e-01 -5.65059662e-01 -2.85841465e-01
2.48630136e-01 1.14097953e-01 -1.13951847e-01 9.85541165e-01
-2.63493538e-01 5.35682797e-01 -7.83235312e-01 -6.94923043e-01
-4.23425198e-01 -3.77933204e-01 -7.46000528e-01 -1.09694868e-01
-5.12055755e-01 -1.48821756e-01 -2.43571371e-01 7.50870824e-01
8.40825617e-01 -2.63949841e-01 2.94349074e-01 4.38035816e-01
2.19158098e-01 -1.83444560e-01 -1.55903482e+00 1.45380628e+00
-5.30777089e-02 2.98363626e-01 6.18852854e-01 -1.01759744e+00
6.13686323e-01 4.29629050e-02 8.77906084e-01 -1.12887576e-01
2.48255089e-01 1.25300258e-01 -2.13756636e-01 -3.22880059e-01
8.11183214e-01 -5.16619980e-01 -2.49798730e-01 8.25917423e-01
-5.25965393e-01 -2.13575870e-01 3.13328296e-01 4.90740865e-01
6.72400534e-01 -6.70906186e-01 3.48561883e-01 -7.44616330e-01
1.62968427e-01 2.03957275e-01 7.93033838e-01 6.81029141e-01
-1.81024313e-01 7.48759091e-01 7.34804809e-01 -1.99500456e-01
-8.94096911e-01 -9.70489323e-01 -3.15886825e-01 1.15935230e+00
2.48865232e-01 -2.04353154e-01 -6.42926514e-01 -6.15698576e-01
5.15839398e-01 6.42656803e-01 -7.40360320e-01 -2.83850487e-02
-1.41811132e-01 -1.04078293e+00 1.38471097e-01 2.13859782e-01
1.07202433e-01 -4.57164347e-01 -3.76318008e-01 -9.79310796e-02
-1.61934420e-01 -1.04839611e+00 -9.14186358e-01 -3.22593480e-01
-9.22749996e-01 -1.04219711e+00 -7.23019540e-01 -4.89735752e-01
8.39264989e-01 3.25306445e-01 1.02273190e+00 -6.94371015e-02
-1.80060565e-01 5.62842071e-01 -2.29417738e-02 -5.97678244e-01
-1.07021563e-01 1.21980011e-01 3.15711737e-01 3.86987507e-01
3.53917241e-01 -4.74126011e-01 -5.50441802e-01 3.27575296e-01
-1.23339677e+00 -2.98325598e-01 1.22304717e-02 4.52789247e-01
7.04425514e-01 5.73469959e-02 9.35542822e-01 -1.16095090e+00
7.71367252e-01 -9.18624580e-01 -8.18654597e-01 3.00638020e-01
-7.47838676e-01 -3.29191953e-01 4.86183524e-01 -4.01257992e-01
-8.21548164e-01 -3.49543616e-02 5.60145438e-01 -1.58527493e-01
2.79375404e-01 5.67784965e-01 -4.91705328e-01 3.59673649e-02
6.08289719e-01 7.67776184e-03 1.24625929e-01 -1.37330353e-01
4.32911634e-01 9.25694108e-01 1.56055629e-01 -8.15852702e-01
5.66974282e-01 9.27560627e-01 3.70672613e-01 -9.92780507e-01
-9.18447495e-01 -3.87626588e-01 -2.81555504e-01 -7.51853660e-02
3.23095471e-01 -7.58324564e-01 -6.53194427e-01 1.23883083e-01
-6.76243424e-01 -1.38931975e-01 -6.90978408e-01 2.68974304e-01
-7.09454834e-01 4.35422540e-01 -2.42881551e-01 -1.04494643e+00
-2.20552728e-01 -6.27011478e-01 5.46038687e-01 2.74963140e-01
-1.66251898e-01 -8.01302016e-01 9.89535302e-02 5.71898341e-01
1.75230235e-01 3.86250347e-01 6.05830431e-01 -6.48872674e-01
-8.28501821e-01 -1.90195635e-01 -2.06501354e-02 2.19978377e-01
-3.77535857e-02 -1.82831720e-01 -5.54419219e-01 -5.66142142e-01
-2.72831358e-02 -4.98859048e-01 8.22833538e-01 6.03106916e-01
1.36027670e+00 -5.77228367e-01 -4.03643847e-02 6.46202505e-01
1.38548982e+00 -2.79352844e-01 2.94082344e-01 -8.23916961e-03
4.15231913e-01 9.38221395e-01 8.16645563e-01 1.22449565e+00
5.62202811e-01 4.56154406e-01 4.29284811e-01 3.55040640e-01
6.33396387e-01 -9.57070887e-02 4.91503216e-02 1.77308679e-01
1.20457262e-01 -3.83837134e-01 -4.67280269e-01 8.48494232e-01
-2.06568193e+00 -1.16897762e+00 3.91470492e-02 2.63633275e+00
9.11823869e-01 -2.36323208e-01 7.77118921e-01 3.55989374e-02
9.06057119e-01 2.46718422e-01 -7.30060816e-01 -4.86127704e-01
-4.44621235e-01 -3.08258563e-01 6.94696665e-01 4.81808871e-01
-9.26869631e-01 4.13603753e-01 6.31147242e+00 8.73341024e-01
-7.14327812e-01 -4.80591543e-02 1.24514115e+00 -9.37089860e-01
-9.16542768e-01 -1.50058955e-01 -6.53705895e-01 3.87512743e-01
5.09328544e-01 -7.63045251e-01 5.80159724e-01 8.18363309e-01
3.46995026e-01 -2.50559986e-01 -1.30655599e+00 9.59561527e-01
1.58320665e-01 -1.23885059e+00 -8.86038244e-02 3.76336634e-01
1.44810951e+00 -3.41555595e-01 1.89347669e-01 -2.70686775e-01
3.87451410e-01 -1.22778618e+00 4.26395684e-01 3.18690273e-03
9.10733700e-01 -1.16215587e+00 2.55917221e-01 4.42649484e-01
-6.09793782e-01 -2.57477909e-01 -7.24792063e-01 -2.41582364e-01
3.30998391e-01 9.69764173e-01 -3.71740609e-01 1.63008884e-01
4.34489429e-01 4.37182307e-01 7.95331150e-02 1.32419336e+00
6.05154224e-02 4.09173816e-01 -6.43562138e-01 -1.37269914e-01
-9.90217403e-02 -3.71210814e-01 6.18194640e-01 8.48452091e-01
8.29532892e-02 3.19566578e-01 8.73427391e-02 8.18883181e-01
-6.64122522e-01 5.42246103e-01 -4.44021493e-01 -2.44229324e-02
6.78409398e-01 1.34515178e+00 -6.12640381e-01 -8.19387212e-02
-3.41666788e-01 5.51325023e-01 1.73491433e-01 2.54460156e-01
-6.78452015e-01 -7.52267167e-02 1.13230300e+00 4.96029466e-01
3.26617993e-02 -4.21072729e-02 -7.33737946e-01 -1.24031663e+00
6.82096332e-02 -8.76504600e-01 8.99640203e-01 5.04861213e-02
-1.46368325e+00 -1.66809112e-01 2.34468669e-01 -8.36044014e-01
-3.01334006e-03 -2.20198575e-02 -5.76315522e-01 5.09792984e-01
-1.34574676e+00 -6.81847036e-01 5.06124983e-04 6.85154259e-01
1.65676236e-01 -4.04873565e-02 2.65060931e-01 2.45156914e-01
-5.01774430e-01 7.14799404e-01 3.67818385e-01 -3.80000561e-01
5.91871977e-01 -1.21096671e+00 -1.74019173e-01 8.11845541e-01
1.35754839e-01 3.48677069e-01 8.04681838e-01 -4.88781750e-01
-1.46084332e+00 -1.03086150e+00 8.40626359e-01 -2.30103642e-01
3.30382943e-01 -3.83323789e-01 -5.02617776e-01 5.46459496e-01
-1.78977817e-01 1.12110287e-01 9.87562001e-01 1.70144558e-01
-1.51501104e-01 -4.64144409e-01 -1.81098258e+00 5.07401526e-01
1.21442020e+00 -5.33870198e-02 1.52181946e-02 5.37219524e-01
4.58652020e-01 -1.65610731e-01 -9.37794924e-01 1.49293810e-01
4.28440601e-01 -8.39047253e-01 7.94852674e-01 -8.56295705e-01
6.64033413e-01 9.18500125e-02 -5.52757502e-01 -1.18200600e+00
-8.80949721e-02 -1.00914729e+00 -1.66228920e-01 1.23717892e+00
4.23071086e-01 -4.27454978e-01 1.31930602e+00 1.31339777e+00
4.60244745e-01 -9.31813478e-01 -9.47997749e-01 -9.16123152e-01
3.39943439e-01 -2.12184191e-01 8.03022325e-01 1.23819947e+00
3.09906900e-01 -2.59428080e-02 -7.64468849e-01 7.31342658e-02
1.14808869e+00 6.09102488e-01 8.33069861e-01 -1.30072963e+00
-3.53980482e-01 -3.76308084e-01 -1.94126382e-01 -9.50756550e-01
5.65425381e-02 -8.84963870e-01 -3.13521981e-01 -1.23326886e+00
7.39531755e-01 -6.17380202e-01 -1.01263493e-01 3.26171331e-02
-8.89625102e-02 1.27971470e-01 3.25092137e-01 7.87628591e-02
-9.43347752e-01 5.47372639e-01 1.06713271e+00 -8.28678384e-02
-2.08049357e-01 4.58546311e-01 -1.30997503e+00 3.76630336e-01
8.85553598e-01 -6.47006452e-01 -5.30492067e-01 -2.85001904e-01
5.52818656e-01 1.72456712e-01 4.34744433e-02 -3.05624366e-01
2.09736362e-01 -1.05759096e+00 -2.81951930e-02 -2.49995917e-01
3.44466716e-01 -8.62690866e-01 4.81498279e-02 3.10898632e-01
-6.74087048e-01 -2.72590697e-01 -4.12471503e-01 6.08128428e-01
-3.24659422e-02 -3.45368207e-01 1.06233644e+00 -1.34086177e-01
-2.04009756e-01 7.42668688e-01 3.21833454e-02 8.23417783e-01
1.40033889e+00 -1.96101084e-01 -2.16884583e-01 -8.64526391e-01
-3.43433499e-01 7.75292397e-01 7.74409235e-01 1.15341730e-01
7.05654502e-01 -1.24512935e+00 -1.18079066e+00 -1.69536561e-01
1.49962261e-01 3.49360853e-02 1.98815376e-01 8.51852834e-01
-1.27491966e-01 4.47203927e-02 -2.64886860e-02 -3.25596929e-01
-1.28329670e+00 6.61105275e-01 1.50918156e-01 1.43407241e-01
4.98249233e-02 9.74848986e-01 8.79403725e-02 -2.53330797e-01
3.47595125e-01 9.81637985e-02 1.44013286e-01 3.17369431e-01
5.37178159e-01 8.79771352e-01 -3.93740445e-01 -4.86378819e-01
-3.59413087e-01 2.27634422e-02 1.53371438e-01 -2.61194229e-01
1.47957587e+00 -3.60862583e-01 -3.24861050e-01 2.46818170e-01
1.03475654e+00 1.98702201e-01 -1.01305401e+00 -3.79359126e-01
-1.08650647e-01 -1.02432048e+00 -1.55676365e-01 -3.10546309e-01
-1.18778789e+00 3.88066322e-01 -4.65342551e-02 5.50534248e-01
1.06280279e+00 1.50707021e-01 8.33098352e-01 2.04794466e-01
5.89125097e-01 -1.37917948e+00 -2.43286252e-01 -6.26353025e-02
6.36978924e-01 -1.48405230e+00 4.30553168e-01 -3.81072402e-01
-6.79310977e-01 6.54484212e-01 4.93391961e-01 -1.09277286e-01
8.09773922e-01 6.49375394e-02 -2.52367944e-01 -5.86254932e-02
-7.65687585e-01 1.76532567e-01 2.17551619e-01 3.66264671e-01
1.48057580e-01 4.72307265e-01 -7.71116495e-01 8.20790768e-01
-1.81721643e-01 -1.18439585e-01 9.38265681e-01 8.26762319e-01
-6.78998590e-01 -9.22428250e-01 -4.78049159e-01 1.03584301e+00
-6.81760013e-01 1.10725366e-01 -5.20636857e-01 3.08943808e-01
-1.16135873e-01 1.10368884e+00 -3.04150023e-02 5.15802354e-02
-7.38264471e-02 -4.87893969e-01 5.72029650e-01 -6.20512247e-01
-3.46935451e-01 -2.75363177e-01 1.02591850e-01 -4.60374981e-01
-4.69388247e-01 -1.10327756e+00 -9.38010931e-01 -8.06577444e-01
-3.51564825e-01 2.02437490e-01 4.38687474e-01 6.45609498e-01
1.10011399e-01 -3.93542916e-01 9.80066240e-01 -3.40265185e-01
-1.26935291e+00 -3.52039218e-01 -1.28505266e+00 5.50076783e-01
9.14539099e-02 -1.72936812e-01 -4.12033916e-01 -3.43333691e-01] | [6.609245300292969, 4.972165584564209] |
e5bb81b0-6f64-41aa-88f3-1bc3120b0cac | towards-massively-multi-domain-multilingual | 2305.14463 | null | https://arxiv.org/abs/2305.14463v1 | https://arxiv.org/pdf/2305.14463v1.pdf | Towards Massively Multi-domain Multilingual Readability Assessment | We present ReadMe++, a massively multi-domain multilingual dataset for automatic readability assessment. Prior work on readability assessment has been mostly restricted to the English language and one or two text domains. Additionally, the readability levels of sentences used in many previous datasets are assumed on the document-level other than sentence-level, which raises doubt about the quality of previous evaluations. We address those gaps in the literature by providing an annotated dataset of 6,330 sentences in Arabic, English, and Hindi collected from 64 different domains of text. Unlike previous datasets, ReadMe++ offers more domain and language diversity and is manually annotated at a sentence level using the Common European Framework of Reference for Languages (CEFR) and through a Rank-and-Rate annotation framework that reduces subjectivity in annotation. Our experiments demonstrate that models fine-tuned using ReadMe++ achieve strong cross-lingual transfer capabilities and generalization to unseen domains. ReadMe++ will be made publicly available to the research community. | ['Wei Xu', 'Mohit Chandra', 'Michael J. Ryan', 'Tarek Naous'] | 2023-05-23 | null | null | null | null | ['cross-lingual-transfer'] | ['natural-language-processing'] | [-3.24817118e-03 2.91168034e-01 -3.49427247e-03 -3.84114563e-01
-1.42682183e+00 -9.83660936e-01 5.62238753e-01 5.71568727e-01
-6.40490592e-01 9.27903771e-01 8.41299057e-01 -2.11713284e-01
-2.20401451e-01 -7.08055794e-01 -3.95374447e-01 4.32922207e-02
5.86934745e-01 8.06170166e-01 6.28967807e-02 -6.25552416e-01
3.22122723e-01 -8.27764198e-02 -1.40595007e+00 5.31110406e-01
1.53861856e+00 2.45426834e-01 1.58849955e-01 8.46250772e-01
1.56768858e-01 9.25550878e-01 -6.97338641e-01 -8.94478559e-01
2.71620322e-02 -5.12046814e-01 -1.63168001e+00 -8.78177769e-03
9.13953364e-01 -2.49996692e-01 -2.07252666e-01 9.56818044e-01
8.03089201e-01 2.05765106e-02 6.87665880e-01 -6.50957346e-01
-1.42974150e+00 7.75395691e-01 -5.91011671e-03 2.33909369e-01
1.00982285e+00 -3.63403335e-02 1.15566552e+00 -9.74780917e-01
8.12029719e-01 1.21827519e+00 5.90064704e-01 4.78376597e-01
-9.07877266e-01 -1.62771404e-01 -1.52752653e-01 1.68204159e-01
-1.20515764e+00 -4.92844582e-01 1.93039089e-01 -6.15315139e-01
1.34083641e+00 2.99340427e-01 1.44851133e-01 1.06940472e+00
-4.36569378e-02 5.92229903e-01 1.51309419e+00 -9.42920566e-01
-2.53457755e-01 8.91872197e-02 4.00629014e-01 5.58890879e-01
9.58443210e-02 -3.80304605e-01 -8.02481413e-01 2.19952330e-01
2.60610849e-01 -8.37845445e-01 -4.60920215e-01 1.84434712e-01
-1.45212388e+00 6.57048106e-01 -2.19089799e-02 2.47999117e-01
1.95853442e-01 -7.64749527e-01 7.24580109e-01 1.06962466e+00
7.59853780e-01 8.46066773e-01 -7.40564585e-01 -6.18472993e-01
-8.77415597e-01 1.67803764e-01 1.02406609e+00 1.26065004e+00
4.42621082e-01 -4.11824971e-01 -4.68048215e-01 1.27699041e+00
1.32343873e-01 8.01874280e-01 7.31737018e-01 -7.32820094e-01
1.14253044e+00 6.82619691e-01 1.36431128e-01 -7.62429476e-01
-5.01851618e-01 -4.18192238e-01 -4.57107753e-01 -1.17838822e-01
6.84405982e-01 -2.89492793e-02 -4.11691934e-01 1.51773190e+00
-2.57819206e-01 -9.21421170e-01 2.83975959e-01 5.95802486e-01
1.28910303e+00 4.86256927e-01 1.00807540e-01 1.11786000e-01
1.39644599e+00 -1.02048194e+00 -6.58191144e-01 -2.89887667e-01
1.08054233e+00 -1.07583141e+00 1.84423244e+00 4.44789678e-01
-1.31477296e+00 -6.68668091e-01 -9.22401011e-01 -6.30662620e-01
-5.95506966e-01 4.21866924e-01 -1.61889344e-01 9.28857803e-01
-1.28040314e+00 1.02434121e-01 -4.03767109e-01 -6.39915168e-01
2.04648487e-02 -1.68554947e-01 -6.31574690e-01 -3.73690724e-01
-1.45156538e+00 1.40874684e+00 4.66581851e-01 -2.93233812e-01
-5.52180827e-01 -4.78068411e-01 -1.14568591e+00 -1.52197808e-01
-2.94037253e-01 -3.01692903e-01 1.22656190e+00 -9.64722872e-01
-1.26431215e+00 1.56685305e+00 2.59736758e-02 -3.14226002e-02
5.89974642e-01 -4.50022250e-01 -7.98102617e-01 -1.86496392e-01
4.06747460e-01 4.24409419e-01 2.49655470e-02 -5.51249087e-01
-6.01668239e-01 -9.95186493e-02 2.75439650e-01 7.82523751e-01
-8.71114075e-01 4.72261161e-01 -2.42641523e-01 -6.69314623e-01
-4.23530400e-01 -6.96917832e-01 4.54990238e-01 -4.52347696e-01
-9.58344191e-02 -5.13592422e-01 2.14263797e-02 -1.50749207e+00
1.67916369e+00 -1.76767933e+00 3.06375355e-01 -2.36067519e-01
1.33343220e-01 2.01009721e-01 -5.17618120e-01 7.71217227e-01
1.90739885e-01 1.25749618e-01 -6.57132361e-03 -1.19980082e-01
3.00865620e-01 -2.05817282e-01 2.63505399e-01 2.78554082e-01
2.38523588e-01 8.51091504e-01 -1.22338092e+00 -4.07418460e-01
-1.27416169e-02 8.01402610e-03 -4.35765058e-01 3.67479399e-02
2.61664689e-01 3.03779423e-01 -6.04765024e-03 4.86233920e-01
4.53824669e-01 -1.20858684e-01 -5.31812832e-02 2.20659733e-01
-3.10373098e-01 8.54364693e-01 -8.40927362e-01 1.94105291e+00
-8.28520417e-01 8.97693992e-01 -3.43194187e-01 -3.16534102e-01
9.27869678e-01 3.31732482e-01 -3.05457264e-01 -1.14761984e+00
-8.50449409e-03 5.26926875e-01 4.26919982e-02 -6.91179156e-01
9.37752187e-01 4.45044726e-01 -4.55478787e-01 4.48295742e-01
2.66637355e-01 -3.20934504e-01 8.32136154e-01 1.06647201e-01
9.63406146e-01 1.44339457e-01 3.60077411e-01 -8.66865873e-01
7.93393373e-01 1.88192442e-01 1.80694029e-01 6.29328609e-01
-2.17562526e-01 6.33569181e-01 1.67935684e-01 -1.88611560e-02
-1.29945838e+00 -1.09288824e+00 -5.58067203e-01 1.60234976e+00
-3.50040734e-01 -4.25423265e-01 -9.83904243e-01 -6.78056479e-01
-3.63296807e-01 7.16898263e-01 -4.80186194e-01 -4.04770188e-02
-2.32953623e-01 -7.89318740e-01 6.57110870e-01 4.68392223e-01
6.18121445e-01 -1.05325007e+00 -2.39784896e-01 2.99231470e-01
-7.74256468e-01 -1.07508218e+00 -5.04359603e-01 -1.96890935e-01
-3.55059505e-01 -1.00081408e+00 -8.87342691e-01 -1.30819845e+00
3.84491265e-01 -2.84940690e-01 2.04461074e+00 -1.93337917e-01
3.79560031e-02 5.83982348e-01 -9.78381991e-01 -4.30213988e-01
-9.54680085e-01 6.30235136e-01 -1.08654685e-01 -7.01088965e-01
8.99340630e-01 1.44857913e-01 -1.36990607e-01 1.96988121e-01
-6.51988983e-01 -1.70533285e-02 3.69409114e-01 8.34202170e-01
1.13025598e-01 -1.20418563e-01 1.01709175e+00 -7.87941873e-01
1.17141259e+00 -3.32437664e-01 -2.80027002e-01 6.84668541e-01
-4.33486640e-01 -2.63025790e-01 3.54845583e-01 -4.40758877e-02
-1.16690850e+00 -3.84156227e-01 -3.81440967e-01 8.85917664e-01
-2.85506457e-01 7.84656465e-01 -3.73926669e-01 2.87560690e-02
1.05947375e+00 1.14679761e-01 -3.19118351e-01 -3.33348274e-01
3.31852466e-01 1.25265825e+00 7.43542671e-01 -8.32920074e-01
6.03042364e-01 -4.27706569e-01 -6.61071539e-01 -7.74790227e-01
-1.06460810e+00 -2.85836220e-01 -1.13543701e+00 -3.47676158e-01
7.54206002e-01 -1.49405539e+00 5.11594699e-04 7.83455610e-01
-8.64533782e-01 -5.20132065e-01 -1.88204437e-01 2.39537492e-01
-4.05443609e-01 4.75407898e-01 -8.81160975e-01 -1.94865808e-01
-4.15694892e-01 -9.65369940e-01 9.43576396e-01 1.57528013e-01
-7.41959810e-01 -1.40348637e+00 3.05269390e-01 8.36111188e-01
4.27721649e-01 6.06394969e-02 9.42705750e-01 -6.55480146e-01
2.05831870e-01 -2.85281599e-01 -2.37662762e-01 4.32717770e-01
2.01843321e-01 -1.29461184e-01 -8.64159167e-01 -5.33263862e-01
-3.70402217e-01 -1.15461373e+00 3.70464444e-01 -3.78813185e-02
7.44096816e-01 -2.55899876e-01 4.52126831e-01 -1.08189294e-02
1.26674628e+00 -3.31653804e-01 7.98881590e-01 8.85453582e-01
5.48004508e-01 7.38446772e-01 6.18813157e-01 3.00307602e-01
1.03454125e+00 4.64245051e-01 -2.56921411e-01 -1.08611293e-01
-3.67463052e-01 -6.93853796e-02 4.93962437e-01 1.52611375e+00
-6.87647015e-02 -4.83942479e-01 -1.38551068e+00 9.62750316e-01
-1.27779090e+00 -8.32183778e-01 -3.18523109e-01 2.28854203e+00
1.43960357e+00 -5.88159589e-03 1.73833549e-01 1.77961484e-01
5.81261098e-01 -3.04431438e-01 -6.29639700e-02 -7.20854223e-01
-6.69417977e-01 1.10102102e-01 2.57396907e-01 7.01226771e-01
-1.12192857e+00 8.64552557e-01 6.61848354e+00 7.57352829e-01
-4.92572635e-01 1.94483653e-01 5.04321218e-01 2.63450772e-01
-4.45658386e-01 -3.89390320e-01 -9.19132531e-01 4.31448787e-01
1.08495319e+00 -4.37803477e-01 2.80889273e-01 3.58601362e-01
2.32537761e-02 -1.20071173e-02 -1.23559809e+00 6.87504828e-01
3.91212791e-01 -5.23600996e-01 1.89758055e-02 -2.89369345e-01
9.51175570e-01 3.19118649e-01 1.81124985e-01 5.81396818e-01
6.01876974e-01 -1.20674634e+00 7.94293642e-01 4.62955028e-01
1.26154888e+00 -7.61745393e-01 9.45618391e-01 2.72650510e-01
-8.36978316e-01 9.83148441e-02 -4.42619056e-01 -4.47561800e-01
-6.91920072e-02 3.75380546e-01 -4.55149472e-01 5.71437597e-01
8.27658772e-01 7.79951513e-01 -1.39842629e+00 7.60508180e-01
-2.90329635e-01 6.25370085e-01 1.55305564e-01 -1.68111891e-01
5.94386198e-02 9.44360942e-02 1.78420365e-01 1.71935856e+00
3.14489424e-01 -2.86342770e-01 2.20844522e-01 2.79011071e-01
-1.49928004e-01 5.69059014e-01 -3.41513216e-01 -2.49137506e-02
6.56669736e-01 1.05049241e+00 -3.20631534e-01 -2.19281465e-01
-1.00985122e+00 1.21668017e+00 8.23085904e-01 -7.77170900e-03
-3.20078790e-01 -9.24596071e-01 5.67793921e-02 4.19215895e-02
-3.97619188e-01 -1.55430332e-01 -3.63174647e-01 -1.44583046e+00
1.95872203e-01 -1.37677300e+00 6.07797384e-01 -6.11403286e-01
-1.80289137e+00 7.42184579e-01 -2.90024459e-01 -1.44466639e+00
-1.21316068e-01 -9.10853922e-01 4.37239483e-02 1.30285847e+00
-1.52371156e+00 -1.15362167e+00 -5.28539658e-01 5.77533960e-01
8.57550681e-01 -3.86928409e-01 1.05993867e+00 4.32136029e-01
-3.70500892e-01 9.17575240e-01 4.40336674e-01 5.52497208e-01
1.42297864e+00 -1.75652432e+00 6.04403734e-01 9.04870331e-01
-1.44941747e-01 4.50459898e-01 5.45611143e-01 -5.84139287e-01
-5.66455424e-01 -1.07801235e+00 1.86365628e+00 -1.28753269e+00
8.84610236e-01 -3.24823737e-01 -8.85626256e-01 6.44375265e-01
8.77221346e-01 -6.84971273e-01 1.02420938e+00 3.86655301e-01
-3.70409369e-01 2.10392073e-01 -1.14269447e+00 4.04767156e-01
9.20509756e-01 -9.98352647e-01 -1.03143954e+00 6.24862790e-01
3.96283448e-01 -7.26469398e-01 -1.49884117e+00 7.70050585e-02
3.83783132e-01 -6.70467675e-01 6.06160402e-01 -5.06418705e-01
8.57851446e-01 -8.66152123e-02 -1.63943231e-01 -1.87729442e+00
-6.66515589e-01 -1.45803675e-01 3.52396339e-01 1.73694885e+00
8.03425789e-01 -4.04561788e-01 8.80784541e-02 6.64569259e-01
-4.54953998e-01 -7.58914556e-03 -8.94885182e-01 -8.21665406e-01
9.67380524e-01 -2.42858613e-03 5.58076262e-01 1.32003987e+00
4.73120242e-01 5.13201356e-01 -7.32818693e-02 5.70600219e-02
2.58540571e-01 -3.77288610e-01 4.27492201e-01 -1.34422100e+00
-6.81547821e-02 -6.45487905e-01 -2.92715997e-01 -6.80980802e-01
3.17131639e-01 -1.26642048e+00 1.96696389e-02 -1.80864704e+00
4.88452256e-01 -1.26429796e-01 -1.53639109e-03 4.89524990e-01
-5.77880204e-01 3.14374655e-01 -1.41835243e-01 -4.23287712e-02
-7.33447433e-01 3.93789083e-01 1.20603824e+00 -1.74471304e-01
-3.96492407e-02 -5.46490133e-01 -7.48270571e-01 3.69436711e-01
7.96994328e-01 -8.96136928e-03 -4.29512829e-01 -1.25881720e+00
7.12398469e-01 -3.03497255e-01 -6.90287873e-02 -1.22314751e+00
-2.04298180e-02 1.30001113e-01 4.97972637e-01 -3.59245278e-02
-2.64145076e-01 -3.01558852e-01 -4.71413732e-01 1.96377501e-01
-7.02890813e-01 7.13817298e-01 3.13569814e-01 -7.29286522e-02
-3.71956170e-01 -3.79973203e-01 7.34610915e-01 -1.51518896e-01
-6.14190519e-01 -1.15870491e-01 -6.20105028e-01 8.95007789e-01
5.42338073e-01 6.22479338e-03 -7.99781501e-01 -2.98963964e-01
-5.23510993e-01 4.17364866e-01 7.26516604e-01 8.51092279e-01
7.36681819e-02 -1.40247893e+00 -1.61953473e+00 -1.22614183e-01
9.11333084e-01 -4.03639853e-01 2.62245357e-01 4.18288529e-01
-7.79351354e-01 5.91882765e-01 -5.31392634e-01 -2.26155102e-01
-1.21517575e+00 1.48112476e-01 2.64248759e-01 -3.71704996e-01
-2.07453698e-01 6.90513313e-01 -3.54478985e-01 -1.18600535e+00
-7.52189308e-02 7.94110596e-02 -6.45107985e-01 3.50800067e-01
7.07961619e-01 7.70931125e-01 5.24144709e-01 -9.80479777e-01
-1.20234452e-01 4.97579455e-01 -2.09246531e-01 -3.29780936e-01
1.02316046e+00 -5.64110994e-01 -3.67632866e-01 5.26516259e-01
9.48325574e-01 2.73514271e-01 -6.46374047e-01 -3.14093411e-01
3.70740086e-01 -2.10625947e-01 -1.27011582e-01 -1.44815302e+00
-3.04359823e-01 6.72229528e-01 5.62765956e-01 -7.69305751e-02
1.22985435e+00 -4.82963584e-02 4.02706981e-01 5.16763985e-01
2.82860488e-01 -1.77884209e+00 -2.27453876e-02 1.09332907e+00
1.11022818e+00 -1.33612561e+00 -2.95600414e-01 -2.01068893e-01
-7.51182556e-01 1.11519706e+00 8.23316693e-01 4.58285183e-01
2.15285048e-01 -4.96670604e-04 4.12618667e-01 1.81585640e-01
-4.23083723e-01 -7.24856034e-02 8.04684699e-01 8.78409743e-01
1.39809477e+00 2.43336380e-01 -7.03092992e-01 5.74052989e-01
-8.92822623e-01 -2.04473823e-01 9.72498536e-01 8.08124661e-01
-4.52562273e-01 -1.18752861e+00 -2.83604413e-01 4.75556552e-01
-5.66001594e-01 -5.37658453e-01 -4.98113483e-01 6.57250524e-01
-1.27958953e-01 1.42818213e+00 -1.26189953e-02 -5.83101884e-02
4.27826941e-01 7.01402128e-02 5.59939146e-01 -8.77983510e-01
-6.61542237e-01 -5.92297554e-01 6.61970377e-01 -8.54741633e-02
-2.81543136e-01 -8.38234186e-01 -5.09432435e-01 -2.59926468e-01
-6.48832917e-02 1.72973692e-01 1.04465872e-01 9.16033030e-01
1.82230733e-02 3.88180792e-01 3.37992728e-01 -1.83300182e-01
-4.28912789e-01 -1.44860077e+00 -4.74776208e-01 5.69004238e-01
2.08177984e-01 -9.14920866e-02 -1.98404714e-01 2.49016255e-01] | [11.045701026916504, 10.160038948059082] |
7505f62c-2704-4ba4-87d7-b3e393cf254d | composite-force-learning-of-chaotic-echo | 2207.02420 | null | https://arxiv.org/abs/2207.02420v1 | https://arxiv.org/pdf/2207.02420v1.pdf | Composite FORCE learning of chaotic echo state networks for time-series prediction | Echo state network (ESN), a kind of recurrent neural networks, consists of a fixed reservoir in which neurons are connected randomly and recursively and obtains the desired output only by training output connection weights. First-order reduced and controlled error (FORCE) learning is an online supervised training approach that can change the chaotic activity of ESNs into specified activity patterns. This paper proposes a composite FORCE learning method based on recursive least squares to train ESNs whose initial activity is spontaneously chaotic, where a composite learning technique featured by dynamic regressor extension and memory data exploitation is applied to enhance parameter convergence. The proposed method is applied to a benchmark problem about predicting chaotic time series generated by the Mackey-Glass system, and numerical results have shown that it significantly improves learning and prediction performances compared with existing methods. | ['Yongping Pan', 'Kohei Nakajima', 'Kai Hu', 'Yansong Li'] | 2022-07-06 | null | null | null | null | ['time-series-prediction'] | ['time-series'] | [ 1.51740298e-01 -1.30995408e-01 -2.18949784e-02 2.77864575e-01
2.69564331e-01 -1.14836104e-01 6.03880048e-01 -5.57943225e-01
-4.25594956e-01 1.13868546e+00 -2.17227817e-01 -1.05014615e-01
-2.81967580e-01 -5.07845223e-01 -4.74354804e-01 -1.19600892e+00
-3.89860153e-01 2.55968928e-01 3.87436777e-01 -4.89549071e-01
7.29923725e-01 4.80660051e-01 -1.37186217e+00 -1.32689655e-01
8.60192597e-01 8.76564145e-01 1.87483743e-01 7.74854958e-01
9.70701873e-02 1.28777361e+00 -3.44928384e-01 6.46557570e-01
2.39934698e-01 -8.07791829e-01 -2.28640214e-01 -1.55074418e-01
-6.29766345e-01 3.91961932e-02 -5.00395656e-01 8.44540417e-01
4.10802186e-01 3.33760649e-01 7.61006773e-01 -8.09917748e-01
-7.24494934e-01 7.11224079e-01 2.95044836e-02 7.01042175e-01
-1.93852589e-01 2.06159607e-01 1.46601111e-01 -9.99180019e-01
1.91235825e-01 6.50221109e-01 8.86669457e-01 8.47257137e-01
-1.36760473e+00 -1.03240347e+00 -1.75824434e-01 5.53950109e-03
-1.12123728e+00 -2.26839453e-01 1.08868158e+00 -4.23560947e-01
1.35474837e+00 1.33507043e-01 9.71672475e-01 8.14573467e-01
9.16529655e-01 2.39913270e-01 1.34177577e+00 -3.91855359e-01
3.38460475e-01 2.88907766e-01 2.70959556e-01 6.62386596e-01
3.23680520e-01 8.89895976e-01 -3.91907275e-01 -2.41962984e-01
8.55773449e-01 1.61247268e-01 -2.67983675e-01 -1.52988270e-01
-9.99619961e-01 7.81453967e-01 4.29428607e-01 6.25072122e-01
-5.94650447e-01 1.10385388e-01 1.01239741e-01 7.23467469e-01
1.91454887e-01 7.25603163e-01 -4.33537394e-01 -3.60462107e-02
-9.99813974e-01 -8.69348347e-02 1.06184101e+00 9.69646722e-02
5.20078123e-01 9.89345074e-01 3.60922098e-01 6.98039830e-01
3.78425539e-01 8.86945367e-01 1.44651353e+00 -7.95305073e-01
-7.76906535e-02 7.79348731e-01 5.43339066e-02 -1.06056070e+00
-5.23819923e-01 -5.30732512e-01 -1.15718806e+00 5.38175106e-01
-6.02430850e-03 -6.05766654e-01 -7.37631202e-01 1.43344283e+00
-7.83922151e-02 8.29461694e-01 6.20177686e-01 6.87716067e-01
4.85491365e-01 1.24866390e+00 -3.28421861e-01 -8.44226539e-01
4.99545604e-01 -1.10792387e+00 -8.30000103e-01 -8.25397819e-02
3.55474234e-01 -3.10647964e-01 5.91578364e-01 5.22515535e-01
-1.02087104e+00 -5.39956152e-01 -1.32235885e+00 8.08478296e-01
-6.23003066e-01 1.69727635e-02 2.39725769e-01 2.33309925e-01
-8.46133113e-01 1.10500479e+00 -1.10263968e+00 1.40663758e-01
-4.04363453e-01 6.64331138e-01 8.83654207e-02 1.01712036e+00
-1.44548786e+00 1.08999658e+00 6.89800978e-01 5.75334132e-01
-7.78994143e-01 -3.25300366e-01 -5.40406704e-01 -1.66156366e-01
-1.52347252e-01 -1.56458855e-01 8.59193802e-01 -1.11747944e+00
-2.14366126e+00 9.71440002e-02 1.48831382e-01 -7.24526465e-01
1.57545492e-01 3.15857232e-01 -7.37310767e-01 -1.75416887e-01
-5.31594515e-01 1.49126723e-01 1.21983778e+00 -8.31910789e-01
-1.12119257e-01 1.08500175e-01 -1.19054139e+00 -1.10659376e-02
-5.28738201e-01 -3.40000510e-01 6.49035454e-01 -5.68760276e-01
3.71592730e-01 -1.20341027e+00 -4.61073548e-01 -7.47664630e-01
2.14422122e-03 -2.16073453e-01 1.06766081e+00 -3.38129997e-01
1.64717507e+00 -1.82663536e+00 2.44419217e-01 6.83365345e-01
-9.29198787e-02 4.31106389e-01 3.78830656e-02 5.56005359e-01
-1.77081570e-01 -2.40610257e-01 -5.16362250e-01 3.55085552e-01
-5.01531303e-01 1.79972589e-01 -4.83710825e-01 2.47768328e-01
1.40688956e-01 8.89276922e-01 -7.98445642e-01 -1.30989283e-01
7.77192414e-02 5.36720157e-01 -5.02042361e-02 2.02874422e-01
9.19683501e-02 4.98221725e-01 -3.80634815e-01 2.01076865e-01
2.66237795e-01 -5.25256813e-01 5.60934171e-02 3.29131365e-01
-5.44766724e-01 -1.47763789e-01 -1.30165982e+00 6.19223893e-01
-2.94202685e-01 7.40283847e-01 -5.40698171e-01 -1.09186590e+00
1.74771929e+00 3.80580783e-01 4.74469304e-01 -6.34597421e-01
2.32795626e-01 7.15483487e-01 2.97385454e-01 -5.58790088e-01
2.47393668e-01 2.94826999e-02 1.25411794e-01 7.09887087e-01
-1.78159371e-01 -1.11147836e-01 -7.12733120e-02 -2.94828147e-01
8.06954145e-01 -1.39865503e-01 2.28969648e-01 -5.01319766e-01
9.47128952e-01 6.16743788e-03 5.60745895e-01 7.01867223e-01
2.96098571e-02 -4.65680100e-02 1.81470960e-01 -8.05988371e-01
-1.52887940e+00 -5.63334286e-01 -3.47018689e-01 5.15899301e-01
1.52176186e-01 4.26237017e-01 -3.88650835e-01 2.71532647e-02
-6.89697713e-02 3.34414691e-01 -8.00873399e-01 -6.66851699e-01
-1.11295199e+00 -8.66555929e-01 4.70442176e-01 2.98316032e-01
7.21745610e-01 -1.96651256e+00 -7.15472758e-01 7.28541970e-01
4.19442117e-01 -4.88961458e-01 -2.94583589e-01 4.53087956e-01
-1.40142667e+00 -8.95152032e-01 -9.25353408e-01 -1.29409218e+00
7.28260517e-01 -2.41533741e-01 5.53101182e-01 3.40955555e-01
1.10090412e-01 -4.26681824e-02 -1.26226515e-01 -3.06410436e-02
-8.81181240e-01 1.92371652e-01 4.71028388e-01 1.04192771e-01
2.52440989e-01 -8.13358068e-01 -4.91057009e-01 3.08914006e-01
-6.57066762e-01 7.42729614e-03 7.33896375e-01 1.37575936e+00
5.00511527e-01 4.51843292e-02 1.04482293e+00 -6.49347723e-01
1.04160476e+00 -5.14631450e-01 -9.05306041e-01 3.47360402e-01
-1.48136890e+00 4.17217374e-01 1.02951157e+00 -1.03760576e+00
-8.32137406e-01 -1.45987310e-02 4.25460339e-01 -5.91771424e-01
6.07710719e-01 6.10271871e-01 1.09033823e+00 -5.23779988e-01
8.21265757e-01 1.15766346e+00 6.44083560e-01 -1.63249597e-01
-3.21768612e-01 6.44056678e-01 5.75150669e-01 -6.42950013e-02
5.49498022e-01 -1.16202243e-01 1.47443846e-01 -6.41694725e-01
-5.26667051e-02 2.48138178e-02 -3.51475298e-01 -3.53195578e-01
3.58549327e-01 -6.65133953e-01 -1.00430119e+00 9.47540283e-01
-6.94543004e-01 -7.70438015e-01 -2.34080434e-01 8.38151813e-01
-7.18668938e-01 -7.81338662e-02 -1.00021386e+00 -1.10218859e+00
-7.12148845e-01 -5.80222487e-01 -5.57092242e-02 7.45391011e-01
2.56826460e-01 -1.12651968e+00 8.16595018e-01 -6.22896254e-01
7.77294874e-01 1.93492323e-01 4.31214243e-01 -6.14292622e-01
-3.14897269e-01 -1.96877778e-01 4.21961635e-01 6.23044729e-01
-5.40352240e-02 1.81965336e-01 -3.13803881e-01 -6.02484405e-01
4.36261564e-01 -2.41268083e-01 8.97756040e-01 3.07149529e-01
4.21144098e-01 -5.92288077e-01 -3.64851803e-01 3.76708418e-01
1.55660915e+00 7.52527952e-01 6.63686037e-01 5.28150558e-01
1.58314943e-01 5.31781167e-02 3.03217899e-02 4.01604027e-01
-1.52519941e-01 -3.70445028e-02 -4.25891653e-02 2.32538328e-01
3.07991385e-01 -1.77993000e-01 6.67002261e-01 1.68323100e+00
-4.04937923e-01 2.52688915e-01 -7.62774944e-01 2.07360074e-01
-1.89553213e+00 -1.37431586e+00 -2.08990183e-02 1.77266550e+00
1.02123737e+00 3.66134942e-01 -3.12877633e-02 5.00498414e-01
9.01986241e-01 -1.05119370e-01 -1.12845337e+00 -6.26110494e-01
-3.73245060e-01 1.22928895e-01 5.55073500e-01 5.47738194e-01
-5.34178793e-01 7.78788507e-01 7.04799843e+00 4.09285307e-01
-1.72406626e+00 -2.38396883e-01 3.61926317e-01 1.39586166e-01
3.73266526e-02 -1.23410501e-01 -8.09021294e-01 9.05234754e-01
1.46303165e+00 -6.55709878e-02 8.81347537e-01 5.76974988e-01
3.43070418e-01 4.42275479e-02 -2.65558660e-01 9.13100183e-01
-6.53086370e-03 -1.35338211e+00 -3.47522378e-01 -1.41165212e-01
1.33623266e+00 6.76724762e-02 3.21820289e-01 5.51092327e-01
7.85795152e-02 -8.35431397e-01 3.36741358e-01 1.55611050e+00
3.03400546e-01 -6.59826517e-01 6.22104406e-01 6.17534578e-01
-1.03080535e+00 -7.13666260e-01 -4.54637051e-01 -3.10162753e-01
-6.00212216e-02 2.08522841e-01 -4.63925660e-01 -4.00419950e-01
5.44979513e-01 8.26695800e-01 -3.99992228e-01 1.04131687e+00
5.37464172e-02 9.52265859e-01 -4.93908286e-01 -9.39378083e-01
3.24887395e-01 -4.66931254e-01 7.11053848e-01 8.89005661e-01
4.71013993e-01 4.36727524e-01 -1.92678064e-01 9.03236508e-01
3.21810693e-01 -7.04225451e-02 -6.55550659e-01 -5.71137108e-02
5.85753381e-01 1.02031934e+00 -8.49805355e-01 -3.80050272e-01
8.62692893e-02 4.83862162e-01 7.62075260e-02 4.22080189e-01
-7.08288372e-01 -7.21752167e-01 -2.60709584e-01 -2.24898294e-01
4.96419907e-01 -3.37995738e-01 -1.08655848e-01 -1.06267548e+00
-2.70794362e-01 -5.59697568e-01 -2.60249395e-02 -8.26090276e-01
-1.07765496e+00 9.09256637e-01 -4.03257191e-01 -1.55557668e+00
-8.58789504e-01 -4.44431752e-01 -8.62705231e-01 6.76144719e-01
-1.40228629e+00 -3.63493711e-01 -5.49938809e-03 7.46608734e-01
4.08421338e-01 -8.12924981e-01 7.44062960e-01 -1.16531678e-01
-7.05978990e-01 5.00587821e-01 7.80326664e-01 -4.12918404e-02
1.47256069e-02 -8.07826519e-01 -1.92134082e-01 6.11271262e-01
-2.16002449e-01 4.61667925e-01 8.12498391e-01 -9.01343286e-01
-1.00640893e+00 -7.43112385e-01 1.02359319e+00 8.81521925e-02
7.96541572e-01 -1.19006559e-02 -1.25099802e+00 3.85237366e-01
1.38025373e-01 2.47248709e-01 1.13413274e-01 -6.14492595e-01
2.78000355e-01 -1.11916654e-01 -9.92199361e-01 4.44154829e-01
3.48356277e-01 -3.58618438e-01 -6.98290646e-01 -5.17350323e-02
4.03181940e-01 -2.75145233e-01 -8.50114167e-01 5.23386896e-01
6.71522439e-01 -6.68697178e-01 5.21778822e-01 -6.63442552e-01
9.22471955e-02 -3.23125184e-01 3.89668822e-01 -1.32207859e+00
-5.28464019e-01 -9.33732331e-01 -7.53732085e-01 5.26436865e-01
7.20541716e-01 -1.11501908e+00 9.37903523e-01 4.79474604e-01
-2.99613010e-02 -1.01045537e+00 -8.21347654e-01 -1.05482626e+00
7.53361136e-02 4.25026655e-01 3.27949107e-01 9.31822121e-01
2.64549613e-01 1.62489489e-01 -5.49898207e-01 -8.87101665e-02
2.89434642e-01 -3.53262536e-02 1.64586395e-01 -1.24467516e+00
-2.19329074e-01 -6.19264901e-01 -2.62556314e-01 -7.17731237e-01
1.85451239e-01 -6.97314203e-01 1.27823174e-01 -7.69579589e-01
-2.49876454e-01 -5.68987429e-01 -7.88806796e-01 2.35680103e-01
7.75199234e-02 2.52097845e-01 -2.41199080e-02 9.29903209e-01
-2.49483004e-01 6.35185122e-01 1.27498269e+00 5.85883111e-02
-9.29446280e-01 4.17760342e-01 -2.97993198e-02 4.16123152e-01
1.06818306e+00 -8.05947423e-01 -4.17351425e-01 3.68604004e-01
4.43168730e-01 5.42290807e-01 2.15023875e-01 -1.40414310e+00
8.03647995e-01 -1.13245994e-01 4.87311035e-01 -6.32703304e-01
3.06433626e-02 -7.05519378e-01 4.08047467e-01 1.34827495e+00
-5.99430859e-01 3.26882362e-01 -1.74157217e-01 5.46022296e-01
-2.38404557e-01 -5.73327303e-01 9.08634901e-01 1.64960492e-02
-4.16690797e-01 4.49502468e-02 -8.18838954e-01 -2.74604172e-01
1.00660050e+00 -5.26777387e-01 -7.04654455e-02 -1.44531161e-01
-1.03752673e+00 4.76541137e-03 -2.62925833e-01 1.66466549e-01
8.41362894e-01 -1.56967652e+00 -5.84728003e-01 6.66605115e-01
-6.11930430e-01 -6.81879580e-01 1.98486038e-02 9.60026443e-01
-4.51637655e-01 6.39987171e-01 -4.23877686e-01 -4.76709068e-01
-5.53211689e-01 4.68208313e-01 1.03113186e+00 -4.08948720e-01
-4.58458662e-01 4.94820148e-01 -9.23001289e-01 -5.38267255e-01
-5.03427312e-02 -9.82191935e-02 -8.23531926e-01 -2.59066433e-01
3.81513953e-01 5.01984358e-01 -8.48841965e-02 -3.96920919e-01
7.56343454e-02 7.69134760e-01 1.13066487e-01 -1.63456231e-01
1.49378622e+00 9.34034586e-02 -3.26130420e-01 8.76829088e-01
1.12647092e+00 -4.32586193e-01 -1.42990541e+00 -3.83260936e-01
2.04731017e-01 3.69163901e-01 -1.26302466e-01 -7.82595634e-01
-9.66123998e-01 4.10770625e-01 1.04746616e+00 6.94038928e-01
9.95083809e-01 -5.78638911e-01 1.00081956e+00 9.72342849e-01
2.89245509e-02 -1.37007546e+00 3.54512811e-01 9.69830453e-01
8.55502486e-01 -1.14886475e+00 -3.05538118e-01 6.61775947e-01
-3.23289871e-01 1.71486616e+00 7.04244971e-01 -1.23860145e+00
1.22926354e+00 2.16343686e-01 -2.03146845e-01 1.05509974e-01
-1.20428050e+00 4.80377495e-01 7.71546438e-02 1.13774836e-01
5.62005714e-02 -3.06842446e-01 -7.06496179e-01 5.31690180e-01
-2.64754415e-01 2.58649826e-01 5.98663986e-01 8.33552361e-01
-1.00403619e+00 -5.51724195e-01 -3.92156094e-01 5.08780718e-01
-1.39569804e-01 -1.08668851e-02 2.06469253e-01 7.61667073e-01
-2.63828129e-01 4.99638140e-01 2.77748615e-01 -4.97532755e-01
5.70700392e-02 4.04519849e-02 1.65341228e-01 -3.46862525e-02
-8.14613879e-01 -1.69917401e-02 -6.34374976e-01 -7.08557069e-02
-3.31908435e-01 -6.30454600e-01 -1.67264044e+00 -1.00776508e-01
-7.39421546e-01 6.58846438e-01 7.98680067e-01 8.42506766e-01
9.50316116e-02 4.71459329e-01 1.46717763e+00 -7.64836967e-01
-9.22547340e-01 -1.03621769e+00 -5.95251739e-01 -1.95576146e-01
6.51085436e-01 -5.33142686e-01 -1.07271433e+00 1.84337690e-03] | [6.620084762573242, 3.364825963973999] |
d5acf4fb-db0c-42b9-8718-b168246b2e5f | m2l-at-semeval-2016-task-8-amr-parsing-with | null | null | https://aclanthology.org/S16-1178 | https://aclanthology.org/S16-1178.pdf | M2L at SemEval-2016 Task 8: AMR Parsing with Neural Networks | null | ['Yevgeniy Puzikov', 'Sadao Kurohashi', 'Daisuke Kawahara'] | 2016-06-01 | null | null | null | semeval-2016-6 | ['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] | [-7.187536716461182, 3.8045973777770996] |
4eccee47-b10f-4a5d-9345-6fecc60ccff3 | deep-outdoor-illumination-estimation | 1611.06403 | null | http://arxiv.org/abs/1611.06403v3 | http://arxiv.org/pdf/1611.06403v3.pdf | Deep Outdoor Illumination Estimation | We present a CNN-based technique to estimate high-dynamic range outdoor
illumination from a single low dynamic range image. To train the CNN, we
leverage a large dataset of outdoor panoramas. We fit a low-dimensional
physically-based outdoor illumination model to the skies in these panoramas
giving us a compact set of parameters (including sun position, atmospheric
conditions, and camera parameters). We extract limited field-of-view images
from the panoramas, and train a CNN with this large set of input image--output
lighting parameter pairs. Given a test image, this network can be used to infer
illumination parameters that can, in turn, be used to reconstruct an outdoor
illumination environment map. We demonstrate that our approach allows the
recovery of plausible illumination conditions and enables photorealistic
virtual object insertion from a single image. An extensive evaluation on both
the panorama dataset and captured HDR environment maps shows that our technique
significantly outperforms previous solutions to this problem. | ['Jean-François Lalonde', 'Kalyan Sunkavalli', 'Yannick Hold-Geoffroy', 'Sunil Hadap', 'Emiliano Gambaretto'] | 2016-11-19 | deep-outdoor-illumination-estimation-1 | http://openaccess.thecvf.com/content_cvpr_2017/html/Hold-Geoffroy_Deep_Outdoor_Illumination_CVPR_2017_paper.html | http://openaccess.thecvf.com/content_cvpr_2017/papers/Hold-Geoffroy_Deep_Outdoor_Illumination_CVPR_2017_paper.pdf | cvpr-2017-7 | ['light-source-estimation'] | ['computer-vision'] | [ 6.37120962e-01 -2.84668058e-01 2.95323879e-01 -5.98550916e-01
-6.34942651e-01 -1.08414578e+00 4.25688088e-01 -6.37908757e-01
-1.18507870e-01 4.71239507e-01 2.90982611e-02 -2.36378908e-01
1.13352261e-01 -9.52203631e-01 -1.20207322e+00 -6.32080674e-01
3.15504104e-01 1.67037115e-01 2.00726360e-01 -2.48474538e-01
1.39147818e-01 8.52542579e-01 -1.95242262e+00 -9.86185148e-02
5.28372228e-01 1.05988455e+00 2.82582432e-01 1.19014645e+00
4.69975263e-01 4.86651480e-01 -4.68847245e-01 6.51988238e-02
9.40742671e-01 -3.56635004e-01 -2.95086741e-01 5.24179995e-01
1.25818753e+00 -1.02629912e+00 -5.30940533e-01 6.63943589e-01
2.70351171e-01 3.15019161e-01 2.17004627e-01 -7.91614592e-01
-2.21936390e-01 -3.34182560e-01 -3.50407660e-01 -1.16877429e-01
3.84810984e-01 4.21079934e-01 6.18586898e-01 -4.07025576e-01
6.57903492e-01 1.07016325e+00 6.21844411e-01 -2.64708921e-02
-1.20774257e+00 -5.22130549e-01 -1.98157236e-01 -2.01134101e-01
-1.25337029e+00 -5.07752836e-01 7.20830739e-01 -3.13074626e-02
8.49619389e-01 3.67505789e-01 1.07552576e+00 9.04134095e-01
2.41737157e-01 1.59502283e-01 1.37701559e+00 -5.97631693e-01
1.34864196e-01 -4.78928396e-03 -4.22481030e-01 6.58339381e-01
-2.16190591e-01 4.45402324e-01 -5.09193659e-01 -1.57425702e-02
1.17846119e+00 -6.45460859e-02 -5.46508431e-01 -3.69135737e-01
-1.04982889e+00 4.25183237e-01 8.23755503e-01 -4.95308578e-01
-1.46441907e-01 4.15295690e-01 -4.22911763e-01 3.91805172e-01
2.27173358e-01 4.75914001e-01 -4.97995317e-01 1.06641792e-01
-6.65670991e-01 1.76691428e-01 6.71042442e-01 7.10064769e-01
1.26000011e+00 -4.89080586e-02 6.23425543e-01 7.49561965e-01
1.50645301e-01 1.27898979e+00 -2.06484631e-01 -1.55512798e+00
3.06411475e-01 2.55462885e-01 2.89334536e-01 -9.34985399e-01
-2.96412736e-01 4.34487797e-02 -3.88512820e-01 5.65403283e-01
4.07089621e-01 -1.76160023e-01 -1.25499022e+00 1.43521702e+00
4.93620008e-01 8.42565000e-02 5.85816689e-02 1.08517361e+00
2.76013166e-01 7.76156783e-01 -9.45632219e-01 6.15223013e-02
9.17481303e-01 -6.58560872e-01 -2.81396955e-01 -6.66286528e-01
-1.12642385e-01 -9.08548236e-01 1.21873915e+00 4.68860269e-01
-8.70869339e-01 -4.14876431e-01 -1.21665251e+00 -3.78818929e-01
-4.06081796e-01 1.48589253e-01 5.58033705e-01 6.42047048e-01
-1.28944075e+00 4.13882941e-01 -6.03215873e-01 -5.28954029e-01
-1.55051071e-02 3.27520788e-01 -1.75678730e-01 -3.81762266e-01
-8.89860213e-01 8.42445910e-01 1.71802312e-01 4.30185735e-01
-1.24422598e+00 -5.04694939e-01 -9.71151710e-01 1.93266734e-03
4.02785152e-01 -4.52129006e-01 1.02193308e+00 -1.15334511e+00
-1.76888204e+00 6.88936710e-01 9.72539708e-02 -2.29419664e-01
3.44036222e-01 -3.12771082e-01 -3.31408650e-01 3.04700792e-01
-4.23605382e-01 9.01385963e-01 8.48098576e-01 -1.59568834e+00
-3.99628699e-01 -1.94590092e-01 3.59424859e-01 6.02144778e-01
-8.04108530e-02 3.72111276e-02 -7.70477593e-01 3.84090133e-02
2.76450843e-01 -1.16114199e+00 -1.43258825e-01 5.66036105e-01
-5.25568426e-01 1.00741136e+00 8.00956964e-01 -6.36233866e-01
2.89838105e-01 -1.86041057e+00 -3.91198732e-02 4.13749605e-01
-1.67738929e-01 -1.80080831e-01 -9.45510641e-02 7.00161830e-02
1.63148105e-01 -3.41714650e-01 -3.28972012e-01 -2.19266877e-01
-3.16368222e-01 3.41359109e-01 -5.54555178e-01 5.72798789e-01
-1.18775494e-01 6.75648212e-01 -5.19976735e-01 -2.51008153e-01
8.43719006e-01 8.63978207e-01 -2.67163515e-01 5.48350453e-01
-6.53192818e-01 4.98996466e-01 2.62091048e-02 7.02368021e-01
9.32659745e-01 -2.53394470e-02 1.98484540e-01 -2.16681138e-01
-1.42663658e-01 -2.63797008e-02 -1.12118280e+00 1.68624139e+00
-9.01936710e-01 1.23266721e+00 -1.55110494e-03 4.36302740e-03
1.08328938e+00 -2.99892455e-01 1.62723869e-01 -8.56509864e-01
3.03738117e-01 -2.16948017e-02 -4.55552220e-01 -5.57698250e-01
6.82028592e-01 4.47702110e-02 1.68600343e-02 6.00040257e-01
-2.09166408e-01 -8.30477297e-01 -4.00163293e-01 -6.91254511e-02
7.13037610e-01 5.44339895e-01 9.49732438e-02 1.01726204e-01
3.10658902e-01 -2.36240625e-01 2.87743688e-01 6.35608315e-01
3.93445730e-01 1.19606495e+00 -1.15961628e-02 -7.63815582e-01
-1.40471303e+00 -1.22218025e+00 -6.78319186e-02 7.43969858e-01
2.87417531e-01 9.04675946e-02 -6.37948871e-01 -1.95740595e-01
-2.24190861e-01 4.47327435e-01 -7.85512447e-01 2.29849577e-01
-5.55012882e-01 -6.54563367e-01 2.80573368e-01 2.87089676e-01
8.07057381e-01 -8.57164264e-01 -1.13884640e+00 -2.60274917e-01
-2.86082804e-01 -1.29803252e+00 -1.80091202e-01 3.13799322e-01
-6.59902036e-01 -1.12016821e+00 -2.41378129e-01 -3.36526185e-01
5.36411047e-01 6.06967807e-01 1.25183952e+00 -7.46663511e-02
-4.85141337e-01 3.71030867e-01 -3.78983794e-03 -3.76955509e-01
-3.68580014e-01 -2.90957183e-01 -5.75752258e-02 2.27859672e-02
-1.03471331e-01 -7.47673094e-01 -7.17350721e-01 5.43048501e-01
-9.95915115e-01 3.64823610e-01 2.81836182e-01 2.85871863e-01
7.68319786e-01 -6.75462037e-02 -2.88855493e-01 -5.16338468e-01
-2.68894702e-01 -1.20051801e-01 -1.37026393e+00 3.02014142e-01
-2.13584304e-01 -2.43291602e-01 5.31607449e-01 -3.54949802e-01
-1.45582843e+00 5.13255477e-01 3.45523000e-01 -7.08081126e-01
-3.00028861e-01 -2.04756886e-01 -3.51774067e-01 -5.02903819e-01
9.36759710e-01 7.20815063e-02 -2.59539902e-01 -2.32455343e-01
6.67714417e-01 5.97978294e-01 1.05443823e+00 -4.97225642e-01
1.05044222e+00 1.10927427e+00 1.37104049e-01 -1.03786314e+00
-9.93816018e-01 -1.56036794e-01 -7.29739249e-01 -3.74781609e-01
8.63855541e-01 -1.06870770e+00 -4.07025337e-01 6.76925182e-01
-8.27911198e-01 -9.96408224e-01 -1.79115757e-01 3.08513075e-01
-6.82389081e-01 4.60594930e-02 -1.41457409e-01 -6.90869272e-01
-4.72061448e-02 -9.84029889e-01 1.44110942e+00 5.56203306e-01
4.08926994e-01 -9.69286978e-01 2.81225711e-01 3.76310945e-01
2.99296051e-01 5.37716448e-01 4.26956147e-01 5.09416997e-01
-1.25978649e+00 -1.00409530e-01 -3.26018453e-01 3.68095964e-01
1.24530457e-01 3.07233900e-01 -1.53919935e+00 -1.94772393e-01
4.67686430e-02 -5.75270593e-01 6.97354555e-01 6.04478776e-01
1.21856201e+00 -3.15348096e-02 5.49281202e-02 1.53036225e+00
1.90097106e+00 -9.60448682e-02 9.32467699e-01 5.78159451e-01
1.03877997e+00 4.13445294e-01 6.94884181e-01 1.94921941e-01
2.78351247e-01 8.58835697e-01 6.71301186e-01 -5.09760916e-01
-2.05125391e-01 -1.81335568e-01 2.99951643e-01 2.23233730e-01
-1.70162991e-02 -4.39741850e-01 -6.99819684e-01 3.00146848e-01
-1.39530647e+00 -5.74102700e-01 1.56241402e-01 2.40447521e+00
5.83989203e-01 -2.26544335e-01 -2.72399187e-01 -3.90165925e-01
5.43809414e-01 3.79664630e-01 -7.98166156e-01 -3.58584553e-01
-5.51990330e-01 5.44946566e-02 1.10279465e+00 7.82345414e-01
-8.95420730e-01 9.86974061e-01 7.04217291e+00 -1.25666577e-02
-1.38426137e+00 -4.16529715e-01 5.00882387e-01 -2.53356665e-01
-4.47549075e-01 2.32195094e-01 -5.80640137e-01 1.48454174e-01
1.02963161e+00 3.70755345e-01 1.07826757e+00 5.87121189e-01
1.28002077e-01 -5.75528622e-01 -9.04770136e-01 9.86848593e-01
5.30623078e-01 -1.26120138e+00 -3.07230294e-01 3.21542770e-01
1.04761541e+00 4.95903641e-01 5.28191663e-02 -5.17521858e-01
6.45907462e-01 -8.43429267e-01 4.91559416e-01 6.76630795e-01
1.09905744e+00 -6.50580645e-01 2.87967324e-01 5.70025574e-03
-7.93543160e-01 -2.01243475e-01 -5.81700087e-01 -2.03745924e-02
2.04680160e-01 3.32604676e-01 -8.33991289e-01 3.75134349e-01
9.52576756e-01 5.64325869e-01 -7.80702055e-01 9.46248233e-01
-4.12815213e-01 3.21023464e-01 -9.37700689e-01 5.26372731e-01
-1.73382923e-01 -4.53178585e-01 1.34559944e-01 7.22589076e-01
3.47006738e-01 5.44568636e-02 -9.03848037e-02 8.30155969e-01
-2.29455307e-01 -4.00225639e-01 -8.07167590e-01 6.25848889e-01
3.07261258e-01 1.48708916e+00 -6.85534179e-01 -7.90019035e-02
-1.75785720e-01 9.78582680e-01 1.02615379e-01 5.72690547e-01
-7.98559546e-01 -2.96066195e-01 6.38878703e-01 -5.45021594e-02
4.43490684e-01 -2.64835447e-01 -1.32929489e-01 -1.26599109e+00
-1.67829469e-02 -9.12785232e-01 -7.51018524e-02 -1.65833163e+00
-5.82352281e-01 3.80436838e-01 1.90862671e-01 -1.20592928e+00
-1.60074189e-01 -5.85754693e-01 -6.45360172e-01 6.98744953e-01
-1.60619068e+00 -1.48018932e+00 -1.02235091e+00 5.76109231e-01
3.99480730e-01 2.76443273e-01 8.30705225e-01 -1.63572595e-01
-2.16542885e-01 1.93062928e-02 4.48861629e-01 -1.49153918e-01
8.64953876e-01 -1.28372920e+00 6.04364812e-01 1.21947336e+00
1.65820673e-01 3.00585598e-01 7.71478653e-01 -2.64556766e-01
-1.55889738e+00 -1.02286565e+00 8.20243657e-02 -7.54084349e-01
2.71808535e-01 -7.17927516e-01 -6.84377909e-01 9.11205888e-01
1.86518654e-01 1.31584108e-01 2.90025353e-01 -2.48283386e-01
-4.57711965e-01 -4.99680728e-01 -1.06688821e+00 6.13762438e-01
8.09104681e-01 -7.25674272e-01 -2.44639471e-01 5.29810250e-01
7.86778927e-01 -8.99963558e-01 -7.28749454e-01 8.91747102e-02
1.00048423e+00 -1.21381044e+00 1.19856906e+00 1.20371357e-01
4.92859721e-01 -5.67182362e-01 -3.98044437e-01 -1.34251869e+00
3.58545125e-01 -7.51641631e-01 -4.53491230e-03 8.06830525e-01
1.97934374e-01 -5.11107504e-01 5.20569861e-01 6.21381819e-01
6.25101328e-02 -2.73420006e-01 -4.44936275e-01 -4.45436031e-01
-2.42634028e-01 -4.49218988e-01 7.82911658e-01 5.60823083e-01
-1.04685175e+00 1.32599324e-02 -6.79168403e-01 6.51524723e-01
9.32934761e-01 5.61986387e-01 1.46436083e+00 -9.51817751e-01
-4.89613265e-01 3.65513742e-01 -1.72132418e-01 -9.03860152e-01
-1.37564531e-02 -1.01337969e-01 4.92441595e-01 -1.38296187e+00
6.35634214e-02 -4.81328726e-01 9.93396491e-02 4.53792602e-01
1.37829967e-03 7.22263992e-01 2.63239467e-03 1.28218725e-01
-3.24913979e-01 2.35924810e-01 9.59037840e-01 2.71537751e-01
-3.46419811e-01 -2.14139894e-01 -2.07018226e-01 6.69405222e-01
7.72155166e-01 -2.22206622e-01 -4.45841849e-01 -8.41504872e-01
3.84021461e-01 9.33334082e-02 5.66162288e-01 -1.21761513e+00
-3.09327878e-02 -5.68172038e-01 1.02905762e+00 -5.90754569e-01
7.43009150e-01 -7.97640860e-01 3.78660560e-01 -7.39931166e-02
-1.56612113e-01 -3.40049744e-01 1.86099276e-01 4.23830509e-01
3.45731109e-01 2.28373051e-01 8.68264854e-01 -9.16533023e-02
-7.67864823e-01 1.46459326e-01 8.38044509e-02 -2.81495333e-01
5.99991918e-01 -3.83771658e-01 -5.75274944e-01 -6.15930259e-01
-1.03559360e-01 1.05760239e-01 1.36163735e+00 4.75310713e-01
6.31746590e-01 -9.89598334e-01 -3.00211221e-01 6.60371542e-01
1.97581396e-01 5.02816200e-01 3.26150179e-01 2.63564438e-01
-1.22153270e+00 -5.57310134e-02 -3.99747014e-01 -8.46675992e-01
-1.28306377e+00 2.63389111e-01 8.87826622e-01 4.42165017e-01
-7.99089670e-01 5.29054642e-01 3.33237469e-01 -6.82649255e-01
-1.11359172e-01 -4.41666573e-01 3.52221698e-01 -5.78040123e-01
5.35165012e-01 1.04294620e-01 6.54889047e-02 -6.79438353e-01
-9.77579951e-02 9.92703378e-01 3.95687759e-01 -5.15709162e-01
1.45936155e+00 -6.03912175e-01 -7.35991169e-03 3.99580985e-01
1.35326815e+00 1.13011584e-01 -2.09771132e+00 -1.60798475e-01
-8.82043302e-01 -1.07178462e+00 3.65686893e-01 -8.75925541e-01
-1.12142384e+00 8.54502678e-01 8.93737495e-01 -3.08356553e-01
1.49195111e+00 -1.24877304e-01 6.09896362e-01 9.62610781e-01
3.23520184e-01 -1.17392039e+00 -6.16959669e-02 3.91013831e-01
6.55812442e-01 -1.34761989e+00 2.45263547e-01 -8.23448449e-02
-3.40349287e-01 1.41569102e+00 6.11346424e-01 -1.20016038e-01
2.51730323e-01 3.98515493e-01 7.02712417e-01 -1.96966156e-01
-3.44383329e-01 -2.61062700e-02 1.19875975e-01 6.87681496e-01
-2.96439886e-01 -5.15435748e-02 9.57457960e-01 -7.03592837e-01
-6.25372767e-01 -1.80892140e-01 9.38694477e-01 7.97959685e-01
-6.81970775e-01 -5.89710414e-01 -7.58827448e-01 1.00086249e-01
8.60094577e-02 9.53670442e-02 -3.34195673e-01 6.64741039e-01
1.78464025e-01 8.54655087e-01 3.65242839e-01 -2.07328185e-01
1.46687597e-01 -4.06594425e-01 8.02206457e-01 -2.59557247e-01
-2.11770684e-01 1.55335113e-01 2.68990304e-02 -1.00600922e+00
-4.80948567e-01 -4.73120958e-01 -6.76417947e-01 -5.65718636e-02
-4.78294224e-01 -5.29118538e-01 1.21943450e+00 7.81057060e-01
4.02818015e-03 1.45222783e-01 1.15612042e+00 -1.41391420e+00
1.05105266e-02 -5.13223290e-01 -4.58836108e-01 2.18772307e-01
8.03589463e-01 -3.60106587e-01 -7.13741362e-01 2.57031202e-01] | [9.706751823425293, -2.988553285598755] |
7877fc0f-46fb-4b57-90fc-f623cafa87d9 | boosting-synthetic-data-generation-with | 2301.07427 | null | https://arxiv.org/abs/2301.07427v1 | https://arxiv.org/pdf/2301.07427v1.pdf | Boosting Synthetic Data Generation with Effective Nonlinear Causal Discovery | Synthetic data generation has been widely adopted in software testing, data privacy, imbalanced learning, and artificial intelligence explanation. In all such contexts, it is crucial to generate plausible data samples. A common assumption of approaches widely used for data generation is the independence of the features. However, typically, the variables of a dataset depend on one another, and these dependencies are not considered in data generation leading to the creation of implausible records. The main problem is that dependencies among variables are typically unknown. In this paper, we design a synthetic dataset generator for tabular data that can discover nonlinear causalities among the variables and use them at generation time. State-of-the-art methods for nonlinear causal discovery are typically inefficient. We boost them by restricting the causal discovery among the features appearing in the frequent patterns efficiently retrieved by a pattern mining algorithm. We design a framework for generating synthetic datasets with known causalities to validate our proposal. Broad experimentation on many synthetic and real datasets with known causalities shows the effectiveness of the proposed method. | ['Riccardo Guidotti', 'Fosca Giannotti', 'Martina Cinquini'] | 2023-01-18 | null | null | null | null | ['causal-discovery', 'synthetic-data-generation', 'synthetic-data-generation'] | ['knowledge-base', 'medical', 'miscellaneous'] | [ 2.22571135e-01 3.25656593e-01 -3.58000308e-01 -6.05438471e-01
-8.47984478e-02 -3.26896966e-01 4.93851036e-01 3.69546890e-01
1.75947696e-01 1.34230220e+00 5.78645617e-02 -2.16511279e-01
-5.34368694e-01 -1.23381782e+00 -1.00972319e+00 -4.44233745e-01
-1.96618468e-01 5.37795246e-01 -1.24704793e-01 3.73119935e-02
5.17358601e-01 1.93019852e-01 -2.18593311e+00 3.57722133e-01
1.16842735e+00 5.39478540e-01 -1.76471859e-01 2.26810858e-01
-3.49612892e-01 7.94685066e-01 -9.10948813e-01 -4.99433964e-01
1.48682892e-01 -6.88043594e-01 -6.87147439e-01 6.41134707e-03
-8.80123675e-02 -4.39451002e-02 1.69312850e-01 9.24081683e-01
9.12543982e-02 -2.95065999e-01 6.36491895e-01 -1.84353936e+00
-4.19376194e-01 1.04015625e+00 -6.83449447e-01 -2.42199339e-02
3.82416040e-01 -1.26299143e-01 9.57555175e-01 -5.42184174e-01
6.92261696e-01 1.18916249e+00 2.62275368e-01 4.08418536e-01
-1.14548039e+00 -1.04253387e+00 1.78425610e-01 4.48774636e-01
-1.21125174e+00 -2.08653718e-01 7.40289271e-01 -4.04412448e-01
4.54295605e-01 6.09628677e-01 6.69134796e-01 9.94623125e-01
2.53752500e-01 6.79752946e-01 1.02033532e+00 -6.60169005e-01
4.65269893e-01 5.47658324e-01 2.62685806e-01 4.25785422e-01
1.03479755e+00 5.57270870e-02 -7.87583649e-01 -5.80998421e-01
3.27817768e-01 9.25078914e-02 -2.23616019e-01 -2.80242294e-01
-9.51541007e-01 8.65200460e-01 -5.73447570e-02 1.24313317e-01
-2.70854682e-01 -1.18572965e-01 1.50042072e-01 4.62879032e-01
1.12517051e-01 5.31317294e-01 -7.69858599e-01 1.76541395e-02
-5.76965332e-01 7.46921122e-01 8.58475626e-01 1.34796286e+00
6.52375102e-01 -1.86231405e-01 -1.25966251e-01 3.82570118e-01
2.97585666e-01 3.55282784e-01 4.81623799e-01 -2.78731823e-01
3.96798193e-01 1.10975552e+00 1.11022398e-01 -1.15206015e+00
-1.53139263e-01 -2.83998758e-01 -7.95510948e-01 9.73086059e-02
3.56925905e-01 -2.26212248e-01 -6.71011865e-01 1.67785013e+00
5.51498950e-01 1.52805433e-01 1.22057483e-01 5.00144184e-01
4.56170201e-01 2.07905993e-01 -1.33358493e-01 -6.97364688e-01
9.05273318e-01 -3.05484712e-01 -8.60060632e-01 2.42287815e-01
4.64972556e-01 -5.46663702e-01 9.88573730e-01 5.91671586e-01
-5.45047462e-01 -2.80358404e-01 -1.02489424e+00 5.64390302e-01
-1.44295692e-01 -1.85577855e-01 1.06236827e+00 7.56517708e-01
-1.86034575e-01 5.77252150e-01 -4.87145871e-01 3.25318030e-03
4.10421520e-01 4.42937881e-01 -2.64384419e-01 -1.94585770e-01
-1.20592320e+00 2.35912427e-01 5.41515946e-01 -2.00617790e-01
-6.76031113e-01 -8.36049020e-01 -4.79683846e-01 -6.16791248e-02
5.93420446e-01 -7.01276064e-01 7.51945913e-01 -9.28877771e-01
-6.69122458e-01 1.92188352e-01 -2.53855199e-01 -5.63557446e-01
5.87461829e-01 -1.00314826e-01 -6.01960301e-01 -5.96860111e-01
1.11304447e-01 3.39739607e-03 7.26647019e-01 -1.31977439e+00
-7.91664004e-01 -3.93118829e-01 4.53804340e-03 -2.79306024e-01
-3.89579564e-01 -2.62885600e-01 4.48287688e-02 -4.55125511e-01
1.63787026e-02 -7.64102042e-01 -3.68445843e-01 -5.26084304e-01
-9.70933080e-01 -2.00899139e-01 7.39981771e-01 -1.21234342e-01
1.32981002e+00 -1.84212434e+00 -1.24438278e-01 5.49580276e-01
2.00574677e-02 -2.45137036e-01 2.58213252e-01 3.84068996e-01
-4.23086464e-01 5.58417976e-01 -2.09092155e-01 3.07705432e-01
-1.75361812e-01 3.09819877e-01 -5.37821829e-01 1.74024045e-01
3.34500462e-01 4.16164339e-01 -6.50097311e-01 -5.61822474e-01
-1.01362243e-01 -8.33654776e-03 -8.04972291e-01 4.32530552e-01
-5.59496224e-01 3.75438094e-01 -7.25552022e-01 4.80387270e-01
6.71695113e-01 -8.60802010e-02 4.52057511e-01 -1.17752738e-01
-7.21271560e-02 2.43523598e-01 -1.53111577e+00 9.64403033e-01
-7.22087100e-02 1.51519671e-01 -8.23184013e-01 -9.86751616e-01
1.11228538e+00 2.88245261e-01 4.76693690e-01 -4.85229313e-01
-1.52454942e-01 2.47912630e-02 1.82506889e-01 -7.94379175e-01
1.81797847e-01 -1.99056435e-02 3.78810912e-02 5.47736526e-01
-2.86690503e-01 1.42966539e-01 5.08797944e-01 5.84678426e-02
1.24556208e+00 -9.88714173e-02 5.41583538e-01 -1.59957662e-01
5.24051070e-01 3.60504210e-01 1.13539779e+00 5.84929407e-01
7.10219920e-01 5.02486169e-01 1.12663639e+00 -6.38217270e-01
-1.00850344e+00 -6.47609413e-01 -2.20083237e-01 2.65698045e-01
-1.32328898e-01 -5.17612338e-01 -4.40538138e-01 -1.04416060e+00
1.52893603e-01 1.08508396e+00 -9.16028738e-01 -9.78095159e-02
-3.07492137e-01 -1.16504776e+00 3.24926108e-01 1.25324339e-01
1.04257777e-01 -1.05953324e+00 -6.48124635e-01 2.48405561e-01
-6.06490411e-02 -6.52428210e-01 2.83841521e-01 1.68212116e-01
-7.77387321e-01 -1.68441713e+00 1.96356446e-01 -1.10709205e-01
1.12247360e+00 -1.44449353e-01 1.13196421e+00 3.10173094e-01
-4.77738529e-01 -4.62514132e-01 -5.66495538e-01 -7.84407675e-01
-5.52498996e-01 -1.19125694e-01 1.40006214e-01 3.31342630e-02
3.71437222e-01 -7.26443648e-01 -2.81222045e-01 3.81857187e-01
-1.15627158e+00 5.11406511e-02 6.50830925e-01 9.00812030e-01
7.18093991e-01 7.69656122e-01 8.74859691e-01 -1.54103339e+00
6.38638437e-01 -8.31441760e-01 -6.62736773e-01 3.95753890e-01
-9.46296155e-01 5.78200698e-01 9.29599345e-01 -5.01842141e-01
-1.01826656e+00 -1.68352835e-02 3.57809067e-01 -2.11896494e-01
-4.12544072e-01 6.34741127e-01 -7.01564372e-01 4.19756055e-01
7.15099752e-01 -2.62864958e-03 -3.24006677e-01 -3.06770802e-01
1.70939580e-01 4.16483432e-01 -1.44646969e-02 -5.98777473e-01
9.75213766e-01 2.37883314e-01 3.13627392e-01 -3.57319951e-01
-5.79859734e-01 9.10903737e-02 -3.78144801e-01 1.56444103e-01
2.79965788e-01 -4.50384200e-01 -7.18271852e-01 -6.40332187e-03
-9.59256589e-01 2.43618637e-01 -4.61940289e-01 3.05733472e-01
-2.69250393e-01 -1.53408632e-01 3.98608148e-01 -8.59265685e-01
-1.32870302e-01 -1.00904298e+00 4.43947643e-01 1.92679107e-01
-5.07144988e-01 -5.94334900e-01 7.11732805e-02 1.31290615e-01
-3.73825594e-03 6.56438112e-01 1.52420485e+00 -8.87977600e-01
-7.15195417e-01 -2.64720857e-01 1.18053660e-01 -1.65056095e-01
5.71679711e-01 4.87589687e-01 -6.34129167e-01 1.38608620e-01
-2.47411542e-02 -9.78929624e-02 3.34329247e-01 1.24402955e-01
1.40545273e+00 -7.41455376e-01 -4.86548275e-01 2.40481451e-01
1.32844484e+00 3.75991523e-01 5.62029958e-01 1.37369454e-01
5.56340158e-01 9.57984388e-01 1.11701536e+00 7.70501673e-01
2.04417005e-01 5.12758434e-01 4.57017094e-01 1.01342350e-01
4.03653502e-01 -4.34154809e-01 1.88329984e-02 4.20199871e-01
7.97372013e-02 -3.10309410e-01 -7.93633401e-01 7.71797478e-01
-1.92946982e+00 -9.67671037e-01 -6.03650928e-01 2.31678295e+00
1.14930427e+00 2.64757305e-01 -3.39766890e-02 6.76694751e-01
5.92286825e-01 -5.06257951e-01 -5.21536469e-01 -3.08031350e-01
-4.15823199e-02 1.85989588e-01 4.22683090e-01 1.62231073e-01
-6.71322823e-01 3.28007400e-01 5.25338411e+00 5.43518722e-01
-8.18502426e-01 -3.16290885e-01 7.66927838e-01 -1.57001048e-01
-1.02075362e+00 2.03396782e-01 -9.15016413e-01 4.15021688e-01
6.86903894e-01 -8.85389209e-01 3.08118500e-02 9.41395819e-01
3.04813266e-01 -9.58150327e-02 -1.42106104e+00 4.78612930e-01
-1.26947612e-01 -1.15451074e+00 3.83510649e-01 1.67779550e-01
9.97746587e-01 -8.70087862e-01 -8.56915563e-02 -1.40231267e-01
3.70957673e-01 -1.29695654e+00 4.08353448e-01 5.79436660e-01
5.37205637e-01 -1.28006148e+00 9.29438710e-01 4.66094524e-01
-6.05813622e-01 -2.47085363e-01 -3.71769071e-01 -7.74877444e-02
-3.24289501e-01 1.29939914e+00 -1.32921863e+00 7.90099680e-01
5.55083871e-01 4.53053594e-01 -5.58510065e-01 1.04333878e+00
-4.30158913e-01 7.32291937e-01 -2.01697648e-01 -2.40028411e-01
-4.01063800e-01 1.55767407e-02 2.57507354e-01 6.41870260e-01
3.75121653e-01 -5.50757311e-02 -2.62694597e-01 1.11289263e+00
-7.32626989e-02 2.72013545e-01 -9.09923851e-01 1.49636880e-01
6.83912694e-01 8.86599004e-01 -4.48298007e-01 -1.33969396e-01
-2.74326473e-01 1.72677591e-01 5.18865772e-02 4.59087454e-02
-8.88538063e-01 -1.60312831e-01 5.44414163e-01 4.40319002e-01
4.04568277e-02 2.82176018e-01 -6.33560777e-01 -9.39171135e-01
2.83326209e-01 -1.23533630e+00 3.38391751e-01 -1.64877892e-01
-1.26774573e+00 4.88568127e-01 1.62888631e-01 -1.35756540e+00
-4.94050473e-01 -4.13714647e-02 -5.33989787e-01 8.30054581e-01
-1.10517490e+00 -7.63235092e-01 -4.03906137e-01 6.33128226e-01
4.54614699e-01 -2.80295521e-01 9.72234130e-01 9.55681726e-02
-5.85479379e-01 7.37242460e-01 -3.49807054e-01 -2.42133066e-01
5.54998338e-01 -1.11922336e+00 4.20195580e-01 8.97191584e-01
2.01676071e-01 9.08935606e-01 1.00173318e+00 -1.12062824e+00
-1.41341853e+00 -1.22143900e+00 8.24143291e-01 -8.01292360e-02
4.40635830e-01 -2.96813607e-01 -9.10555959e-01 4.96636659e-01
-8.50546882e-02 -8.51624310e-02 9.07434165e-01 2.14999691e-01
-7.70578384e-02 -2.92512953e-01 -1.35007358e+00 5.35204232e-01
8.93578291e-01 2.72970140e-01 -5.00048399e-01 3.15479964e-01
7.19729066e-01 1.39796019e-01 -5.86094737e-01 5.88199258e-01
5.17941892e-01 -1.11770356e+00 4.19680715e-01 -1.00591481e+00
7.46053696e-01 -5.69968045e-01 5.85021712e-02 -1.23225713e+00
6.18415736e-02 -5.25838554e-01 -9.62092429e-02 1.54746616e+00
8.39460671e-01 -5.38414061e-01 9.64822710e-01 8.26133013e-01
4.46510315e-01 -5.76611876e-01 -7.45671809e-01 -4.86745685e-01
-4.22640830e-01 -3.43608707e-01 1.18222892e+00 1.29105484e+00
1.53806340e-02 3.07585180e-01 -5.22295892e-01 2.51970559e-01
9.28249478e-01 5.72397411e-01 1.09818602e+00 -1.45790076e+00
-4.83366370e-01 1.82175711e-01 -4.51942265e-01 -1.53892353e-01
1.49719324e-02 -4.70977336e-01 -2.27890968e-01 -1.15524125e+00
2.99097240e-01 -1.00423896e+00 -3.48140337e-02 4.18451905e-01
-3.96493703e-01 -4.29087162e-01 -4.50951427e-01 -8.38339999e-02
-2.49634665e-02 4.95602727e-01 1.03782666e+00 3.94000188e-02
-4.01911765e-01 2.95420498e-01 -8.19290280e-01 7.34980702e-01
9.01457429e-01 -9.58569765e-01 -8.38530123e-01 -2.64817756e-03
5.58299065e-01 1.40736580e-01 6.94067031e-02 -7.21236348e-01
7.71128237e-02 -7.38864005e-01 3.24005634e-01 -5.90688169e-01
-3.68570805e-01 -1.01759696e+00 9.18677330e-01 3.93342078e-01
-3.46348971e-01 6.84046447e-02 -7.64101967e-02 4.66810822e-01
-3.44428003e-01 -3.31448555e-01 3.27374816e-01 -3.80248688e-02
-2.22525343e-01 1.49753734e-01 8.83208364e-02 7.18738511e-02
1.32724261e+00 6.98963702e-02 -3.35493535e-01 -4.97500151e-02
-3.09280545e-01 1.94284543e-01 3.47415328e-01 5.83212852e-01
7.94220030e-01 -1.17697632e+00 -8.07979643e-01 4.66840297e-01
3.29333365e-01 3.04119617e-01 -1.25292584e-01 5.23916245e-01
-1.08429797e-01 1.87907398e-01 -2.76816815e-01 -2.55568326e-01
-1.38669789e+00 6.80937529e-01 -4.89054844e-02 -2.07833141e-01
-1.85213342e-01 5.16120851e-01 4.82794493e-02 -1.97095707e-01
1.02385192e-03 -2.04084486e-01 -3.50409508e-01 4.35503274e-02
5.34412444e-01 2.87900954e-01 1.27092987e-01 8.96326378e-02
-2.61965394e-01 8.54886398e-02 -2.41734058e-01 1.80161342e-01
1.49968255e+00 2.90715426e-01 -3.89637291e-01 6.19709551e-01
5.42113066e-01 3.90842706e-01 -5.53066194e-01 1.70281246e-01
2.54448533e-01 -8.14155340e-01 -4.76812720e-01 -5.78066111e-01
-9.63764727e-01 3.31164658e-01 3.24889392e-01 5.51592708e-01
1.20874786e+00 -2.34214097e-01 1.25478253e-01 2.16710821e-01
5.83509505e-01 -5.93593001e-01 -2.67010033e-01 -3.16248387e-02
8.47304285e-01 -1.01905620e+00 2.01555982e-01 -8.02767992e-01
-4.74799126e-01 1.09956110e+00 8.05528522e-01 9.73298699e-02
4.46110934e-01 7.05537438e-01 -2.16664821e-01 -1.05535284e-01
-1.25505626e+00 1.75800726e-01 -3.08506116e-02 5.89037299e-01
5.18040061e-01 1.62957430e-01 -8.58428121e-01 8.89830232e-01
-4.06007022e-01 1.16437346e-01 8.24668288e-01 8.19858789e-01
-4.80071381e-02 -1.61738670e+00 -5.44533432e-01 8.19466233e-01
-6.18677735e-01 -7.15102628e-02 -5.15837729e-01 8.51379991e-01
5.97756326e-01 1.04283786e+00 -8.52377340e-02 -4.20867175e-01
5.26935220e-01 1.34452395e-02 1.48505479e-01 -5.80157161e-01
-3.85152668e-01 -4.01577353e-01 2.67479420e-01 -3.64926785e-01
-2.50537366e-01 -8.61318707e-01 -1.33426225e+00 -3.41196865e-01
-4.99733716e-01 5.16276300e-01 5.86500287e-01 8.21929336e-01
2.29331315e-01 7.24363923e-01 9.97998178e-01 2.91675985e-01
-3.61879945e-01 -8.71307611e-01 -4.81261373e-01 6.19906366e-01
-1.90255586e-02 -8.75818372e-01 -3.34499240e-01 3.05319518e-01] | [8.680005073547363, 5.742864608764648] |
c74f8793-12ba-4f28-b81d-b5789f51b02b | probabilistic-appearance-invariant-topometric | 2107.07707 | null | https://arxiv.org/abs/2107.07707v1 | https://arxiv.org/pdf/2107.07707v1.pdf | Probabilistic Appearance-Invariant Topometric Localization with New Place Awareness | Probabilistic state-estimation approaches offer a principled foundation for designing localization systems, because they naturally integrate sequences of imperfect motion and exteroceptive sensor data. Recently, probabilistic localization systems utilizing appearance-invariant visual place recognition (VPR) methods as the primary exteroceptive sensor have demonstrated state-of-the-art performance in the presence of substantial appearance change. However, existing systems 1) do not fully utilize odometry data within the motion models, and 2) are unable to handle route deviations, due to the assumption that query traverses exactly repeat the mapping traverse. To address these shortcomings, we present a new probabilistic topometric localization system which incorporates full 3-dof odometry into the motion model and furthermore, adds an "off-map" state within the state-estimation framework, allowing query traverses which feature significant route detours from the reference map to be successfully localized. We perform extensive evaluation on multiple query traverses from the Oxford RobotCar dataset exhibiting both significant appearance change and deviations from routes previously traversed. In particular, we evaluate performance on two practically relevant localization tasks: loop closure detection and global localization. Our approach achieves major performance improvements over both existing and improved state-of-the-art systems. | ['Michael Milford', 'Niko Sünderhauf', 'Tobias Fischer', 'Ming Xu'] | 2021-07-16 | null | null | null | null | ['loop-closure-detection', 'visual-place-recognition'] | ['computer-vision', 'computer-vision'] | [-7.68980384e-02 -3.25061440e-01 -3.69705379e-01 -4.59051341e-01
-1.17615414e+00 -9.01226103e-01 7.46975482e-01 2.92829812e-01
-6.87239408e-01 6.37888908e-01 -1.67218506e-01 -1.71106592e-01
-7.81350508e-02 -3.98036391e-01 -1.05450809e+00 -3.77831459e-01
-2.36318067e-01 4.88095045e-01 6.13804817e-01 -3.45148742e-01
5.39652228e-01 7.22893596e-01 -1.46919096e+00 -4.13853049e-01
5.22503436e-01 7.85792172e-01 2.76913673e-01 7.95230031e-01
3.46096367e-01 5.85392892e-01 -2.33446792e-01 4.40102816e-01
1.51874065e-01 4.83001880e-02 -4.41120595e-01 -1.75036147e-01
6.46219969e-01 -2.56374121e-01 -6.60305202e-01 9.96294498e-01
3.43358696e-01 3.68992299e-01 5.32801747e-01 -1.40690756e+00
-4.95544463e-01 -1.89199001e-01 -5.85877538e-01 1.98070079e-01
7.58936703e-01 1.22022592e-01 7.04390705e-01 -8.68895590e-01
8.32749426e-01 1.13068914e+00 8.20091426e-01 2.29043230e-01
-1.41592717e+00 -1.64463237e-01 2.90400475e-01 1.63251042e-01
-1.75494158e+00 -6.52763903e-01 7.14801013e-01 -4.52195674e-01
1.08990490e+00 -1.89052504e-02 2.16332182e-01 1.16825092e+00
5.75906456e-01 6.07578874e-01 8.43676865e-01 -1.85605109e-01
5.68669081e-01 -1.11907832e-01 -2.17924416e-01 1.00299346e+00
3.27695608e-01 3.53259236e-01 -8.47568214e-01 -1.84822440e-01
6.31454647e-01 -1.41651496e-01 -1.15857422e-01 -1.20887494e+00
-1.44495189e+00 6.42266273e-01 7.46264756e-01 -1.64100304e-01
-4.40637022e-01 5.86257875e-01 -1.06895808e-02 -5.89419305e-02
-3.49962562e-02 3.09626937e-01 -2.56374538e-01 -3.53147388e-01
-6.55422509e-01 2.89122283e-01 6.72106326e-01 1.39664149e+00
1.23442388e+00 -5.66421822e-02 1.28963962e-01 4.94975001e-01
6.55429482e-01 1.00421941e+00 4.95479643e-01 -1.31643808e+00
4.92021501e-01 4.78164315e-01 6.50596380e-01 -1.13532412e+00
-6.05631828e-01 -1.28562823e-01 -1.24176204e-01 3.99582624e-01
2.61663347e-01 7.03486726e-02 -1.21499991e+00 1.72722208e+00
4.83330309e-01 -1.75566822e-01 1.41388774e-01 1.05755496e+00
3.06958526e-01 2.92106420e-01 -1.61548987e-01 4.70135361e-01
1.12373805e+00 -7.76933074e-01 -5.23222983e-01 -7.04021752e-01
6.80296838e-01 -6.59237027e-01 5.75517058e-01 1.04715176e-01
-5.03328264e-01 -4.23402458e-01 -1.54902947e+00 -2.12781802e-01
-5.03418803e-01 2.17532247e-01 5.29235840e-01 3.40473503e-01
-1.24356759e+00 3.74145836e-01 -1.34523034e+00 -7.55985618e-01
-1.91442966e-01 3.09484065e-01 -8.06339264e-01 -3.63928258e-01
-8.36690187e-01 1.18088031e+00 1.36169493e-01 1.86824128e-01
-1.16965079e+00 -9.90837812e-02 -1.42533100e+00 -5.48525810e-01
2.18572959e-01 -6.05589449e-01 1.35111904e+00 -1.74057797e-01
-1.47199523e+00 5.30111849e-01 -7.15570748e-01 -3.65796655e-01
4.39042926e-01 -1.77488193e-01 -3.34820509e-01 7.61897191e-02
5.08980095e-01 7.84627914e-01 4.81561184e-01 -1.47961879e+00
-8.26835990e-01 -6.01166368e-01 2.98836585e-02 6.24491870e-01
6.59848094e-01 -6.01091623e-01 -6.56946003e-01 -3.11516318e-02
1.06906915e+00 -1.39717245e+00 -4.48208481e-01 2.79532313e-01
-2.46178806e-01 2.76222825e-01 7.59170592e-01 -2.22779140e-01
5.81421614e-01 -2.10993767e+00 -4.01364006e-02 2.22126767e-01
-1.57885626e-01 -4.46452141e-01 -1.00798972e-01 4.80400056e-01
4.74532485e-01 -4.01597828e-01 -1.82379130e-02 -7.52759099e-01
1.57661900e-01 5.70110857e-01 -3.20668817e-01 1.01915026e+00
-7.44517222e-02 7.34717429e-01 -1.25318635e+00 -1.66555807e-01
5.29745102e-01 3.72549951e-01 -3.00078690e-01 -1.69203356e-01
-8.19543228e-02 5.01919270e-01 -5.07645607e-01 8.92856538e-01
4.42055523e-01 9.47476253e-02 2.71504708e-02 8.81779492e-02
-5.65963566e-01 3.85652125e-01 -1.26949453e+00 2.36877251e+00
-2.45605618e-01 7.12374449e-01 6.27792925e-02 -3.79089415e-01
8.59031916e-01 -4.52353843e-02 4.16257173e-01 -9.30304945e-01
-1.53394133e-01 4.28931117e-01 -3.96495253e-01 -1.06594175e-01
1.20606911e+00 2.00871110e-01 -3.28059584e-01 -1.71201870e-01
1.16311591e-02 -1.78894550e-01 -5.54403029e-02 6.57956749e-02
1.48828447e+00 6.24179244e-01 3.47214252e-01 -6.42014667e-02
2.86148190e-01 6.46622896e-01 6.35235369e-01 9.78497267e-01
-6.74543679e-01 7.88218737e-01 6.53770566e-03 -2.62829989e-01
-8.25189471e-01 -1.53333735e+00 7.07189590e-02 8.64930212e-01
1.09814727e+00 -1.68220848e-01 -8.91438201e-02 -4.26004022e-01
2.91045576e-01 5.76283813e-01 -6.28012955e-01 -8.49471763e-02
-4.53604192e-01 -4.67832685e-01 7.03768492e-01 5.94062686e-01
4.15281445e-01 -5.20581841e-01 -9.10053551e-01 3.22151303e-01
-4.03741390e-01 -1.22087109e+00 -2.87480831e-01 3.75542492e-01
-6.64733648e-01 -1.00937724e+00 -3.54177535e-01 -5.85860014e-01
6.92265689e-01 6.97150409e-01 5.27168989e-01 -5.42037547e-01
-1.65583372e-01 8.18756998e-01 -2.12548301e-01 -1.46622479e-01
-2.01456338e-01 -4.73356172e-02 4.61530060e-01 -3.40188146e-01
2.62633413e-01 -2.90893942e-01 -5.96822143e-01 4.88072544e-01
-3.07054400e-01 -4.03303623e-01 6.00678086e-01 5.60950041e-01
9.86963272e-01 -4.90847141e-01 3.35565545e-02 -6.73855022e-02
1.08216926e-01 -4.90272045e-01 -8.92193615e-01 8.67931470e-02
-5.97175181e-01 3.76863182e-01 5.87422177e-02 -2.37999186e-01
-7.62979388e-01 5.26042223e-01 5.75283580e-02 -4.51256871e-01
-2.61124521e-01 4.20805156e-01 1.55179456e-01 -5.50697565e-01
9.78293538e-01 3.78269702e-01 -5.02231494e-02 -2.19489992e-01
4.39971000e-01 3.30097347e-01 8.99618447e-01 -4.41912621e-01
8.00356090e-01 1.05591989e+00 2.87550569e-01 -7.06371009e-01
-2.19030842e-01 -9.47981477e-01 -7.17036605e-01 7.23185856e-03
6.68931723e-01 -1.14291239e+00 -6.90493345e-01 2.66288698e-01
-1.15811193e+00 -2.08352178e-01 9.57270488e-02 8.32596123e-01
-8.42433274e-01 4.06464607e-01 -2.88020879e-01 -9.31505442e-01
2.58298486e-01 -1.31697845e+00 1.34728432e+00 2.03967690e-01
-1.41807243e-01 -7.91097105e-01 5.73191524e-01 -2.64631331e-01
2.49637291e-01 3.95732701e-01 1.75421432e-01 -1.70318097e-01
-9.54502761e-01 -5.38966835e-01 -3.42326537e-02 -3.44861478e-01
8.99093971e-02 -3.32514793e-01 -8.35140467e-01 -3.85626942e-01
-3.94149721e-01 -4.06883620e-02 7.66011000e-01 2.53522038e-01
3.98021489e-02 1.61520466e-01 -8.38674068e-01 5.54675937e-01
1.48898947e+00 2.02013180e-01 4.33569342e-01 7.82872140e-01
5.28721869e-01 4.20833766e-01 9.26930428e-01 2.60328114e-01
1.09252381e+00 9.79660034e-01 7.87460744e-01 1.92150861e-01
9.19255167e-02 -7.69985795e-01 5.16115367e-01 3.59853774e-01
3.13400030e-01 -1.20862797e-01 -9.62782860e-01 7.82850206e-01
-2.10983610e+00 -5.66765428e-01 -1.09081842e-01 2.44277811e+00
2.16778934e-01 1.01851225e-01 -4.98184592e-01 -4.26893234e-01
4.23756003e-01 2.75472760e-01 -6.69720113e-01 -9.36617889e-03
6.31659701e-02 -4.21775490e-01 1.32113266e+00 8.57057154e-01
-1.30680716e+00 1.05046368e+00 6.51730919e+00 1.87266976e-01
-1.04808962e+00 -4.27224301e-03 -3.67637783e-01 9.45814103e-02
5.25334403e-02 3.43070328e-01 -1.00387871e+00 9.23290923e-02
7.81146228e-01 3.08249831e-01 4.14525032e-01 1.16297686e+00
9.59891751e-02 -8.26628268e-01 -1.22622228e+00 1.02532947e+00
9.07351226e-02 -9.31992948e-01 -5.04903674e-01 1.66083336e-01
5.47558963e-01 8.34571719e-01 1.75567776e-01 3.17515731e-01
4.45821732e-01 -6.10484362e-01 1.11353409e+00 4.53288943e-01
5.41771948e-01 -3.89031291e-01 4.94255602e-01 4.50944364e-01
-1.40176880e+00 4.41724807e-02 -3.51890206e-01 -3.47464941e-02
3.05124193e-01 5.99903800e-02 -1.09445059e+00 7.66067445e-01
5.58297217e-01 4.93386418e-01 -6.87651634e-01 1.25684583e+00
-3.71584028e-01 -2.15378031e-01 -7.38643527e-01 7.77612925e-02
4.02559251e-01 1.12349905e-01 9.18447793e-01 7.42949188e-01
3.69093716e-01 -3.20137113e-01 3.89852613e-01 7.24674225e-01
3.66489500e-01 -4.65138435e-01 -9.95457649e-01 4.38112885e-01
7.77448058e-01 9.13659334e-01 -6.21131420e-01 -9.77800116e-02
-1.47248864e-01 1.04443586e+00 3.92298460e-01 4.92211163e-01
-8.39951515e-01 -4.26879078e-01 9.83045101e-01 -1.66459486e-01
2.78034151e-01 -1.00882888e+00 -3.15808132e-02 -1.11776865e+00
1.88109860e-01 -1.48235559e-01 -3.08731981e-02 -8.94751072e-01
-6.01257563e-01 2.65278190e-01 -6.12020120e-02 -1.44756484e+00
-5.31876683e-01 -4.21980709e-01 3.26146930e-02 8.41714859e-01
-1.85158253e+00 -1.09321797e+00 -3.83453548e-01 3.45097601e-01
3.22154880e-01 3.75057787e-01 9.01814342e-01 9.82359573e-02
-4.95038144e-02 3.00900668e-01 3.35171640e-01 -1.21356264e-01
8.81125450e-01 -1.19690573e+00 6.35358930e-01 9.96249199e-01
3.77147436e-01 9.16292191e-01 9.33316350e-01 -8.02380025e-01
-1.81986356e+00 -9.32002485e-01 7.84279287e-01 -1.03292274e+00
5.23205936e-01 -4.24815893e-01 -4.56410885e-01 9.85141933e-01
-4.98083472e-01 2.19790697e-01 2.26440243e-02 1.32000819e-02
-5.06333828e-01 5.91487661e-02 -1.11904204e+00 7.46971905e-01
8.88416231e-01 -7.62710690e-01 -6.78859949e-01 -3.55316214e-02
4.97960448e-01 -9.22416627e-01 -4.65150297e-01 5.00374854e-01
8.08790386e-01 -8.03125560e-01 1.02155304e+00 8.86982307e-02
-4.40304220e-01 -9.45278049e-01 -7.16155708e-01 -1.17123783e+00
-2.11060092e-01 -4.55804884e-01 -1.55500531e-01 7.58323610e-01
3.70471895e-01 -6.30940020e-01 7.38011897e-01 5.90000510e-01
-1.58897772e-01 -2.17454880e-01 -1.34467232e+00 -9.84485924e-01
-6.00852489e-01 -5.50994873e-01 1.23632222e-01 4.93876845e-01
4.98652197e-02 9.04270634e-02 -1.90951020e-01 9.04498398e-01
7.45835781e-01 -1.37183368e-01 9.95332062e-01 -7.92203367e-01
1.39123231e-01 -5.25673702e-02 -9.20444191e-01 -1.54696643e+00
-3.75658795e-02 -5.10389745e-01 1.02485061e+00 -1.65257084e+00
-3.70506763e-01 -4.05277967e-01 -2.49962956e-01 3.26849967e-01
1.97469607e-01 1.78368181e-01 -3.87320220e-02 4.69265282e-01
-1.00442147e+00 7.23591566e-01 5.92457414e-01 -3.73697318e-02
-3.95191878e-01 -1.97846979e-01 -1.72618419e-01 7.71809876e-01
4.36681956e-01 -5.69022596e-01 -3.36250424e-01 -6.46427393e-01
1.19335331e-01 1.76873222e-01 6.33745730e-01 -1.28823388e+00
7.51791358e-01 -9.75267962e-02 4.11726832e-01 -8.99185061e-01
8.58890831e-01 -7.55957127e-01 8.06825981e-02 3.10902596e-01
-6.51799887e-02 4.77250546e-01 2.99255401e-01 1.16192925e+00
-5.05054034e-02 1.34136183e-02 5.13467312e-01 -1.05688870e-02
-1.41937804e+00 7.65014514e-02 -6.16317689e-01 -2.95194179e-01
9.03166354e-01 -4.39408511e-01 -4.00142342e-01 -2.47752070e-01
-5.23048341e-01 3.27332467e-01 1.09697378e+00 7.00610220e-01
6.59039080e-01 -1.29059947e+00 -5.73881082e-02 2.59657204e-01
7.44601309e-01 6.82899281e-02 5.85346408e-02 1.06571531e+00
-7.39516258e-01 7.17695355e-01 -8.78247246e-02 -1.02536833e+00
-7.32122600e-01 4.48921114e-01 2.69640148e-01 1.73010945e-01
-4.53792989e-01 5.12974381e-01 5.13756238e-02 -9.04370725e-01
1.57792717e-01 -4.90828156e-01 2.15799764e-01 -4.49022919e-01
2.17509404e-01 3.55182350e-01 4.62899059e-02 -1.05059469e+00
-8.39313209e-01 6.45841181e-01 2.13175014e-01 -7.34234154e-01
6.42310858e-01 -8.20407033e-01 3.57140601e-01 6.80751860e-01
1.21993279e+00 -6.73328340e-02 -1.58234274e+00 -1.81902573e-01
3.81523997e-01 -4.00303125e-01 -9.52186584e-02 -5.82891762e-01
-1.10232413e-01 5.31385720e-01 7.46318996e-01 -4.70091969e-01
2.44542286e-01 -4.85038608e-02 3.51501465e-01 8.55769992e-01
1.20578754e+00 -9.56069410e-01 -2.69148737e-01 8.74969482e-01
5.55341005e-01 -1.49105346e+00 8.78701806e-02 -3.11794300e-02
-4.73404050e-01 7.17735827e-01 5.04976690e-01 -2.21117169e-01
3.27273190e-01 9.80255008e-02 2.19623700e-01 -1.28119245e-01
-5.52833438e-01 -2.76854515e-01 2.63257563e-01 8.48681092e-01
-1.02005176e-01 1.55880908e-02 2.08769605e-01 -7.30720013e-02
3.50023881e-02 -2.04439357e-01 3.62380981e-01 1.47455084e+00
-6.47974730e-01 -6.71857953e-01 -4.79886323e-01 -2.94195086e-01
9.83717442e-02 1.83894545e-01 -1.05456285e-01 1.00538361e+00
-1.73487306e-01 8.82060289e-01 2.18371823e-02 -4.60956216e-01
5.23826718e-01 -3.86315547e-02 6.32749915e-01 -4.93050128e-01
1.78321958e-01 -1.89924762e-01 3.88000347e-02 -1.19314075e+00
-2.28157401e-01 -7.69022048e-01 -1.55626810e+00 -7.48907253e-02
-1.77128598e-01 -5.99892624e-02 1.31426632e+00 8.63688111e-01
6.88420713e-01 2.70010948e-01 3.07536691e-01 -1.30781662e+00
-6.44942462e-01 -7.41141021e-01 -5.62123299e-01 -8.98963287e-02
8.04279506e-01 -1.07950544e+00 -1.99311748e-01 -1.82660490e-01] | [7.42788028717041, -2.0664968490600586] |
be63e692-d627-44fc-8ec0-cd0637d20176 | composite-biomarker-image-for-advanced | 2304.12423 | null | https://arxiv.org/abs/2304.12423v1 | https://arxiv.org/pdf/2304.12423v1.pdf | Composite Biomarker Image for Advanced Visualization in Histopathology | Immunohistochemistry (IHC) biomarkers are essential tools for reliable cancer diagnosis and subtyping. It requires cross-staining comparison among Whole Slide Images (WSIs) of IHCs and hematoxylin and eosin (H&E) slides. Currently, pathologists examine the visually co-localized areas across IHC and H&E glass slides for a final diagnosis, which is a tedious and challenging task. Moreover, visually inspecting different IHC slides back and forth to analyze local co-expressions is inherently subjective and prone to error, even when carried out by experienced pathologists. Relying on digital pathology, we propose Composite Biomarker Image (CBI) in this work. CBI is a single image that can be composed using different filtered IHC biomarker images for better visualization. We present a CBI image produced in two steps by the proposed solution for better visualization and hence more efficient clinical workflow. In the first step, IHC biomarker images are aligned with the H&E images using one coordinate system and orientation. In the second step, the positive or negative IHC regions from each biomarker image (based on the pathologists recommendation) are filtered and combined into one image using a fuzzy inference system. For evaluation, the resulting CBI images, from the proposed system, were evaluated qualitatively by the expert pathologists. The CBI concept helps the pathologists to identify the suspected target tissues more easily, which could be further assessed by examining the actual WSIs at the same suspected regions. | ['H. R. Tizhoosh', 'Soma Sikdar', 'Adrian Batten', 'Ricardo Gonzalez', 'Morteza Babaie', 'Abubakr Shafique'] | 2023-04-24 | null | null | null | null | ['whole-slide-images'] | ['computer-vision'] | [ 2.48621896e-01 -9.10656974e-02 -6.89600334e-02 1.13073692e-01
-6.12796247e-01 -9.32138324e-01 3.58133428e-02 8.58602226e-01
-4.02442336e-01 5.51054776e-01 -3.81080687e-01 -3.20328534e-01
-9.10804123e-02 -6.60555720e-01 -1.12477936e-01 -1.23099327e+00
1.23611838e-02 3.89226496e-01 4.59003687e-01 4.27319482e-02
1.45894885e-01 1.02734053e+00 -1.45758879e+00 2.62593359e-01
6.81280494e-01 7.04287112e-01 4.06206876e-01 8.76733124e-01
-2.74137527e-01 5.05454481e-01 -3.56066465e-01 -8.25183392e-02
1.30717933e-01 -2.82979310e-01 -4.33336705e-01 1.14678502e-01
5.39772436e-02 1.39507502e-01 4.31747854e-01 1.39204228e+00
1.47676662e-01 -2.87523627e-01 7.79661059e-01 -1.10241294e+00
-2.53000677e-01 -1.01971230e-03 -6.16696239e-01 1.10469386e-01
2.76124001e-01 4.51794505e-01 6.08388841e-01 -8.81607234e-01
1.02823591e+00 9.85063255e-01 3.98157001e-01 3.06394547e-01
-1.14072418e+00 -5.69367468e-01 -2.47114658e-01 6.16526663e-01
-1.68226790e+00 1.71720222e-01 6.52924120e-01 -7.03713596e-01
3.71925563e-01 5.98408222e-01 9.01942134e-01 -8.37446526e-02
3.15409809e-01 4.71593618e-01 1.52134955e+00 -4.23092574e-01
3.31805855e-01 5.42843580e-01 2.55097389e-01 8.67426515e-01
3.94035995e-01 -1.53751373e-01 1.34139806e-01 6.69530332e-02
3.10760856e-01 4.39683348e-01 -5.57362974e-01 -1.33874834e-01
-1.18492126e+00 4.18512851e-01 6.23429179e-01 8.69210064e-01
-3.20896238e-01 -4.81068194e-01 2.31064022e-01 1.02910675e-01
-1.87189162e-01 1.99891984e-01 9.84360203e-02 5.87515175e-01
-9.89565015e-01 -3.32627863e-01 2.19047293e-01 1.57658398e-01
8.12464774e-01 -5.64007163e-01 1.39837146e-01 5.03004730e-01
4.01797473e-01 6.31943285e-01 5.82674265e-01 -3.49743277e-01
-2.32040972e-01 1.13024318e+00 1.59967199e-01 -1.15914440e+00
-6.90897405e-01 -1.93237439e-01 -1.02314007e+00 7.44230151e-01
6.05051160e-01 2.80300260e-01 -9.68975067e-01 8.78942192e-01
5.45401752e-01 -1.92849800e-01 4.79857344e-03 1.20199382e+00
6.16144359e-01 4.39696342e-01 4.56024379e-01 -3.53552431e-01
1.72826242e+00 -5.51256657e-01 -8.95823002e-01 2.98852891e-01
6.17334962e-01 -8.29916716e-01 9.34582591e-01 3.94826353e-01
-8.61758232e-01 -3.34203601e-01 -1.41178823e+00 5.59805334e-02
-7.41861463e-01 7.69420981e-01 2.17497095e-01 2.55519152e-01
-1.12117052e+00 2.34314889e-01 -9.98883367e-01 -7.04016209e-01
2.79180884e-01 4.27069992e-01 -7.14048028e-01 1.34743482e-01
-7.75339723e-01 9.90540087e-01 1.25536442e-01 4.97733295e-01
-4.71025914e-01 -6.55253470e-01 -5.93956888e-01 -6.12492673e-02
-1.36706740e-01 -2.55722046e-01 7.51803517e-01 -8.60571742e-01
-1.29625666e+00 1.22882771e+00 -4.52998489e-01 9.66187567e-03
3.02112818e-01 8.16660762e-01 -3.77739519e-01 7.79702485e-01
-1.04159817e-01 4.55152988e-01 2.62021840e-01 -1.44169700e+00
-1.00392246e+00 -6.01548493e-01 -1.70160219e-01 1.03559740e-01
-1.83289900e-01 -1.31788880e-01 -3.14231247e-01 -3.98582816e-01
1.07296288e-01 -7.50662148e-01 -3.08408024e-04 4.37582582e-01
-4.52884853e-01 -2.61375178e-02 8.21492136e-01 -7.46019423e-01
1.09989893e+00 -2.45378685e+00 -1.28110200e-01 9.18590188e-01
2.43816584e-01 3.88496667e-01 1.39663190e-01 -1.17755070e-01
-1.31806314e-01 3.80524039e-01 2.45971009e-01 1.19861796e-01
-2.08416924e-01 -2.41028965e-01 5.18493116e-01 7.00506508e-01
1.09608389e-01 5.35330772e-01 -8.60148489e-01 -1.16611552e+00
3.92394602e-01 4.76811260e-01 2.48558074e-01 1.43686414e-01
3.53664517e-01 3.77263635e-01 -2.33288199e-01 1.05885780e+00
8.27643692e-01 -2.76015490e-01 5.97370148e-01 -7.59161294e-01
-8.79566073e-02 -4.83937055e-01 -1.13895524e+00 6.66694880e-01
-3.16359967e-01 4.82913852e-01 4.31682408e-01 -7.97632396e-01
9.85877335e-01 5.03127217e-01 2.21762791e-01 -4.22156274e-01
2.21136153e-01 5.13283968e-01 -1.93951558e-02 -7.11256385e-01
-1.79935068e-01 -5.57671368e-01 4.45513308e-01 3.07017833e-01
-3.60779017e-01 2.74575837e-02 5.80157697e-01 -8.62435475e-02
6.73561454e-01 -6.66232347e-01 6.21514082e-01 -1.97157219e-01
1.14330530e+00 4.42255467e-01 4.03446883e-01 1.93403646e-01
-5.05431771e-01 2.78162748e-01 4.35891569e-01 -6.15264237e-01
-8.64442050e-01 -1.19627690e+00 -3.45790029e-01 2.23786324e-01
4.90162909e-01 2.59873718e-01 -5.53017259e-01 -4.54438239e-01
3.97594422e-02 4.48375531e-02 -8.03846955e-01 2.59883910e-01
-2.16853306e-01 -7.03583539e-01 1.04205385e-01 2.31610447e-01
1.66475758e-01 -7.32504427e-01 -8.48808110e-01 9.22779087e-04
6.73879683e-02 -3.56138587e-01 -6.43107891e-02 8.90251249e-02
-7.49470234e-01 -1.35491240e+00 -9.16279435e-01 -1.07091331e+00
1.39313877e+00 4.88850445e-01 5.29059827e-01 6.51745975e-01
-6.17385149e-01 -1.47836238e-01 -3.14176679e-01 -3.20395142e-01
-6.01954818e-01 -4.15212363e-01 -2.33411953e-01 1.35490417e-01
2.79180795e-01 -2.84970790e-01 -8.78276467e-01 3.21070433e-01
-1.07589865e+00 1.13274030e-01 8.91700089e-01 8.56722713e-01
1.17600107e+00 2.54489362e-01 2.33422458e-01 -7.55312502e-01
2.86755919e-01 -2.18841657e-01 -7.66587913e-01 7.67446220e-01
-3.70846123e-01 -3.20402831e-01 8.34480464e-01 -1.74603149e-01
-9.93414581e-01 9.79361311e-03 1.23755157e-01 -3.94495547e-01
-2.93012857e-01 6.90556586e-01 3.04402001e-02 -2.75449604e-01
4.71896589e-01 -1.39322281e-01 5.50550997e-01 4.95663695e-02
-2.10990056e-01 7.78624654e-01 5.98036945e-01 3.55377495e-01
6.33529663e-01 8.06273520e-01 3.61202329e-01 -3.30370545e-01
-9.07554701e-02 -7.61936843e-01 -4.82492983e-01 -6.30181849e-01
8.96567523e-01 -5.04443407e-01 -8.47410679e-01 6.43460155e-02
-8.61151159e-01 -8.56084600e-02 2.39077345e-01 5.40245950e-01
1.00398980e-01 2.77826279e-01 -6.82888091e-01 -6.61796868e-01
-4.09902632e-01 -1.29525900e+00 8.43075037e-01 6.98777080e-01
-3.79840255e-01 -1.09294307e+00 8.21901709e-02 1.79881781e-01
1.93606332e-01 4.09793228e-01 1.32598186e+00 -3.28389823e-01
-2.97888607e-01 -6.56708896e-01 -2.52695888e-01 -8.11448544e-02
3.51889819e-01 8.48219216e-01 -7.52544940e-01 6.01285975e-03
-1.91191688e-01 1.92028865e-01 3.34094763e-01 3.63311261e-01
7.80373693e-01 -1.56972662e-01 -8.93202305e-01 3.57085168e-01
1.91855121e+00 6.11413717e-01 5.82787037e-01 4.43275005e-01
1.88485011e-01 9.14328754e-01 1.01386523e+00 -1.18577676e-02
2.14529976e-01 4.00819361e-01 3.65037501e-01 -9.33260441e-01
-1.28464654e-01 2.38812715e-01 6.56747743e-02 4.63941336e-01
-8.66273940e-02 4.51570228e-02 -9.82708931e-01 9.18935061e-01
-1.34161723e+00 -7.70810544e-01 -4.22040403e-01 1.97918069e+00
7.22974956e-01 -1.51058987e-01 3.22723649e-02 4.70830619e-01
1.08906364e+00 -6.88388824e-01 -3.87792230e-01 -4.07497674e-01
-2.91914105e-01 2.10160822e-01 2.92987734e-01 6.09732628e-01
-8.23178351e-01 1.88318610e-01 5.38587856e+00 7.15022922e-01
-1.90406692e+00 -2.00375989e-01 8.85505438e-01 -7.61837959e-02
-2.94814795e-01 -2.56896377e-01 -3.74910116e-01 4.92358923e-01
2.64588445e-01 -4.16612297e-01 -6.12329170e-02 4.53020453e-01
4.64055419e-01 -8.04825008e-01 -8.41639221e-01 6.70039833e-01
-2.96654224e-01 -1.41399169e+00 2.40914542e-02 -2.25094557e-02
5.27975261e-01 -4.55713719e-01 -2.73014549e-02 -2.81878620e-01
-5.48310876e-02 -8.10916960e-01 3.92892629e-01 6.89553678e-01
9.05766845e-01 -5.61541319e-01 1.59601748e+00 -3.76843661e-02
-1.25532603e+00 1.76304325e-01 -5.14054149e-02 4.39145595e-01
-7.49311224e-02 7.16252983e-01 -1.23097467e+00 6.04960203e-01
7.65603006e-01 3.31495613e-01 -8.20160389e-01 9.86722589e-01
-1.37879118e-01 2.42527083e-01 -2.41227522e-01 -3.21274519e-01
1.29187226e-01 -1.74818948e-01 3.36996078e-01 1.34570241e+00
4.81212854e-01 -4.95425574e-02 -9.68713537e-02 6.60404444e-01
4.74576563e-01 4.40793842e-01 -2.42288888e-01 4.57454622e-02
4.27312762e-01 1.95891559e+00 -1.41733217e+00 -6.31261885e-01
-1.53257772e-01 4.72851545e-01 -8.54109377e-02 4.79001343e-01
-5.35449922e-01 -5.49538672e-01 2.43264109e-01 2.36795619e-01
-5.11729866e-02 5.05315363e-01 -4.10141051e-01 -5.43604374e-01
-1.92262173e-01 -5.78585148e-01 5.96540213e-01 -7.89314210e-01
-1.20011437e+00 5.73039532e-01 -2.32670188e-01 -1.69770908e+00
5.65467887e-02 -8.43776584e-01 -7.67211914e-01 9.98789012e-01
-1.53829837e+00 -9.63008821e-01 -5.52245080e-01 3.99527520e-01
-3.16852331e-01 4.06308591e-01 8.04063499e-01 1.73013046e-01
-7.16153979e-01 4.58855748e-01 1.79245815e-01 -1.61543846e-01
8.10654521e-01 -1.46139717e+00 -8.42518091e-01 6.96356475e-01
-6.65251732e-01 6.69197142e-01 6.69584155e-01 -5.57978153e-01
-9.84227836e-01 -9.73955810e-01 9.32411969e-01 2.94824243e-01
6.45306051e-01 2.40220517e-01 -8.41843605e-01 2.35426188e-01
2.34276995e-01 2.23054111e-01 1.17797816e+00 -5.77139318e-01
1.35081321e-01 -2.62151271e-01 -1.56700861e+00 8.00517738e-01
-6.68300688e-02 -3.57671291e-01 -1.76801115e-01 1.76220164e-01
-2.99511701e-01 -1.35391563e-01 -1.20129061e+00 4.71434414e-01
8.90056133e-01 -9.09155548e-01 5.14271617e-01 6.33981451e-02
1.22806085e-02 -1.25259519e+00 3.91704887e-01 -1.06462753e+00
-2.28336662e-01 1.32042632e-01 8.26881826e-01 9.42878067e-01
4.41318601e-01 -4.87744212e-01 6.69422150e-01 2.08685338e-01
1.78645730e-01 -9.73261058e-01 -7.88952947e-01 -4.32723492e-01
-3.21525097e-01 2.39624739e-01 4.71848816e-01 6.71711445e-01
7.74238050e-01 -3.34840566e-01 6.75473630e-01 3.61726880e-01
5.04387558e-01 1.70211017e-01 5.07582307e-01 -1.08990872e+00
7.12686479e-02 -8.18189979e-01 -8.73909950e-01 1.83587208e-01
-3.44367415e-01 -1.00513721e+00 4.97729406e-02 -1.55697954e+00
3.90512854e-01 -2.88610488e-01 -5.07143676e-01 5.66288471e-01
-4.23835546e-01 8.58212352e-01 2.52881765e-01 4.04210508e-01
-3.04251075e-01 -3.11529577e-01 1.41053212e+00 -5.58822453e-01
-8.36385861e-02 -5.49119055e-01 -5.55842996e-01 6.80741906e-01
6.93139195e-01 -1.99774578e-01 1.30136299e-03 2.87976831e-01
4.98346090e-02 8.92908499e-03 4.35979247e-01 -1.03559768e+00
4.42032784e-01 -2.00442314e-01 5.58551490e-01 -8.13069403e-01
-1.56479314e-01 -1.01194584e+00 6.18672431e-01 8.72780025e-01
-6.07405156e-02 4.31806445e-02 1.65632471e-01 1.88154355e-01
-6.46454036e-01 -2.20101863e-01 9.01445866e-01 -2.10934449e-02
-3.93825650e-01 -2.07061484e-01 -7.53225029e-01 -9.68182266e-01
1.44814122e+00 -6.88491404e-01 -4.31960851e-01 -1.19547196e-01
-8.83281231e-01 3.03333908e-01 8.02936435e-01 -6.31690264e-01
5.64589441e-01 -1.22040963e+00 -3.54266942e-01 -7.44323954e-02
3.98435473e-01 7.28650168e-02 8.34453344e-01 1.46543193e+00
-1.14703417e+00 3.22644264e-01 -5.15401244e-01 -8.27308297e-01
-1.72551382e+00 6.42433524e-01 3.70395631e-01 -7.46155798e-01
-1.41026959e-01 5.19444764e-01 3.84518415e-01 7.96771273e-02
-1.09393373e-02 -5.32082319e-01 -6.83731496e-01 1.86207741e-01
7.63299525e-01 3.32565546e-01 2.41695225e-01 -7.00119376e-01
-5.41500986e-01 9.40835834e-01 -9.67601538e-02 -1.61462780e-02
8.35836530e-01 -1.98644504e-01 -4.36912894e-01 4.21679825e-01
1.15923166e+00 1.69032037e-01 -8.00609946e-01 2.65586019e-01
-6.39682263e-02 -2.49024153e-01 -1.12739027e-01 -1.01806867e+00
-1.14639568e+00 7.56764293e-01 9.06945944e-01 1.97795883e-01
1.52242398e+00 -7.92554989e-02 2.66132087e-01 -2.06644833e-01
1.18992694e-01 -9.61457312e-01 -2.89813548e-01 -4.11302179e-01
6.03894114e-01 -1.04474998e+00 5.50096743e-02 -6.73852384e-01
-5.18436551e-01 1.58409917e+00 2.77549714e-01 3.73714976e-02
5.71241081e-01 4.48746145e-01 7.57984757e-01 -3.13381732e-01
-5.98226845e-01 -7.99361169e-02 3.38661551e-01 5.30865371e-01
3.98886353e-01 2.55719304e-01 -7.59652615e-01 6.17080212e-01
1.89085841e-01 1.86941326e-01 4.79573458e-01 7.82627940e-01
-2.23811463e-01 -8.29101741e-01 -8.10360074e-01 2.66836524e-01
-4.04056728e-01 4.14395005e-01 -4.71073240e-01 9.55761373e-01
4.05329525e-01 9.31915104e-01 1.49079785e-01 -3.46617371e-01
1.84388921e-01 -2.19285801e-01 2.87525564e-01 -1.38428554e-01
-6.66327238e-01 3.70158136e-01 -4.47365791e-01 -1.57344460e-01
-7.83039212e-01 -5.79322040e-01 -1.72874212e+00 -1.43820852e-01
-2.84677655e-01 3.46971750e-01 7.37657070e-01 7.69752443e-01
3.64954062e-02 4.91025627e-01 6.73772752e-01 -5.38784742e-01
3.47914636e-01 -7.34184384e-01 -8.02968144e-01 5.32925606e-01
4.34005797e-01 -6.04059577e-01 -6.82373047e-01 5.19209146e-01] | [15.072463035583496, -3.071242332458496] |
0566ecd2-d682-4075-96df-135f1eff7500 | rrpn-guidance-towards-more-accurate-scene | 2009.13118 | null | https://arxiv.org/abs/2009.13118v1 | https://arxiv.org/pdf/2009.13118v1.pdf | RRPN++: Guidance Towards More Accurate Scene Text Detection | RRPN is among the outstanding scene text detection approaches, but the manually-designed anchor and coarse proposal refinement make the performance still far from perfection. In this paper, we propose RRPN++ to exploit the potential of RRPN-based model by several improvements. Based on RRPN, we propose the Anchor-free Pyramid Proposal Networks (APPN) to generate first-stage proposals, which adopts the anchor-free design to reduce proposal number and accelerate the inference speed. In our second stage, both the detection branch and the recognition branch are incorporated to perform multi-task learning. In inference stage, the detection branch outputs the proposal refinement and the recognition branch predicts the transcript of the refined text region. Further, the recognition branch also helps rescore the proposals and eliminate the false positive proposals by the jointing filtering strategy. With these enhancements, we boost the detection results by $6\%$ of F-measure in ICDAR2015 compared to RRPN. Experiments conducted on other benchmarks also illustrate the superior performance and efficiency of our model. | ['Jianqi Ma'] | 2020-09-28 | null | null | null | null | ['scene-text-detection'] | ['computer-vision'] | [ 3.59987795e-01 4.05908711e-02 -2.79478610e-01 -2.17235938e-01
-8.11019480e-01 -3.34348902e-02 5.46978891e-01 -9.35185403e-02
-3.66775364e-01 3.72340649e-01 3.62142563e-01 -9.77603048e-02
2.13645503e-01 -7.75881112e-01 -4.18122590e-01 -7.78409719e-01
5.11697769e-01 4.76956278e-01 8.85816693e-01 2.33600382e-02
4.75138783e-01 1.73081562e-01 -1.46114957e+00 6.26154780e-01
8.22456062e-01 9.20034051e-01 5.26492596e-01 4.75859493e-01
-3.88031006e-01 5.30299783e-01 -5.07322550e-01 -5.48076093e-01
2.90953636e-01 -9.56003089e-03 -3.80784571e-01 -8.56372342e-02
2.60747373e-01 -5.00100493e-01 -2.62826949e-01 9.67088461e-01
6.48921251e-01 1.25752548e-02 4.85029280e-01 -7.90007770e-01
-2.15281397e-01 1.03203940e+00 -1.28540838e+00 6.26030564e-02
1.44674078e-01 5.20208105e-02 1.32677627e+00 -1.36642182e+00
3.36758256e-01 1.24838662e+00 6.97187126e-01 4.90024328e-01
-8.61320257e-01 -1.01929677e+00 4.78918672e-01 1.67488798e-01
-1.57874203e+00 -4.63750780e-01 6.15779400e-01 -2.53344513e-02
8.97035956e-01 2.69309700e-01 3.83493811e-01 9.75104511e-01
1.68654606e-01 1.46689546e+00 3.94625485e-01 -3.87120843e-01
-3.27020846e-02 -1.65915743e-01 2.60544956e-01 9.80407715e-01
2.83400595e-01 -2.51656532e-01 -6.19281828e-01 -2.50675052e-01
8.01057339e-01 2.35438704e-01 -2.36135066e-01 -1.97443273e-02
-1.03785956e+00 6.68830693e-01 3.07351470e-01 1.81140646e-01
-2.86661059e-01 -2.29656398e-02 4.25108373e-01 -2.44804502e-01
2.97912478e-01 2.50466354e-02 -5.39516032e-01 2.73281425e-01
-1.47690952e+00 1.63347885e-01 3.81163180e-01 1.09649491e+00
5.50291657e-01 -1.85028780e-02 -8.29258382e-01 1.06346023e+00
6.15366042e-01 2.20693603e-01 4.35041904e-01 -4.40341711e-01
6.95133388e-01 9.80876803e-01 -1.88012078e-01 -8.05230916e-01
-4.17707503e-01 -8.34941566e-01 -1.07569313e+00 -4.09727454e-01
1.07544683e-01 -1.55921727e-01 -1.16420639e+00 1.22413588e+00
3.51263314e-01 1.50672436e-01 -2.31142923e-01 6.56378150e-01
8.22639048e-01 8.96109045e-01 1.11445107e-01 -2.69163907e-01
1.48119903e+00 -1.41302633e+00 -7.16341555e-01 -2.44100675e-01
6.07780278e-01 -9.41373706e-01 8.36521327e-01 5.97471416e-01
-9.52835560e-01 -7.27926672e-01 -9.95770693e-01 -9.52925012e-02
2.52788782e-01 8.13236654e-01 4.22251880e-01 3.98669988e-01
-7.72744596e-01 2.71066546e-01 -7.51943111e-01 -1.30921736e-01
6.42720103e-01 4.59972054e-01 1.15194283e-01 -1.94154769e-01
-8.99690866e-01 4.27080244e-01 6.22288942e-01 3.14225286e-01
-6.65282786e-01 -4.41300124e-01 -4.98599976e-01 3.79839897e-01
6.22891903e-01 -4.80399251e-01 1.17473912e+00 -5.77447414e-01
-1.40552127e+00 3.14265460e-01 -3.19616675e-01 -4.18770194e-01
6.18991852e-01 -2.49282300e-01 -1.99797004e-01 1.08497590e-02
1.75534159e-01 1.07055998e+00 1.07067668e+00 -9.67978597e-01
-1.28403902e+00 -1.76337242e-01 -3.06041598e-01 2.75235623e-01
-3.17257792e-01 -5.85626997e-03 -1.25185883e+00 -8.94991219e-01
4.42119986e-01 -6.32116497e-01 -4.08590913e-01 -2.53617503e-02
-5.45469820e-01 -5.05195141e-01 9.95454192e-01 -6.14900827e-01
1.54574823e+00 -2.18267941e+00 -5.16010523e-02 2.21090257e-01
3.18386883e-01 3.71933341e-01 -1.16705380e-01 9.34867263e-02
2.59579629e-01 7.91279152e-02 -8.33426043e-02 -6.24480963e-01
-4.18785959e-02 1.39035866e-01 -3.82717371e-01 2.40031689e-01
3.36274415e-01 7.55759716e-01 -4.89866883e-01 -9.54835415e-01
1.41313508e-01 2.24163249e-01 -6.79670513e-01 -2.64314525e-02
-3.46286923e-01 2.27252301e-02 -7.51585782e-01 6.97061479e-01
9.28930163e-01 -2.56022424e-01 2.01213270e-01 -3.95797372e-01
-1.04203522e-01 3.20024252e-01 -1.59138656e+00 1.44617820e+00
6.25154004e-02 3.36060077e-01 1.71216115e-01 -7.70339727e-01
1.00534844e+00 2.05162376e-01 7.11385757e-02 -5.38549602e-01
9.25417542e-02 -8.86256173e-02 -3.02246436e-02 -1.65769696e-01
8.36288393e-01 1.83897242e-01 6.88757673e-02 1.89096123e-01
-1.13712057e-01 1.92601815e-01 1.59890965e-01 2.33630791e-01
1.08341074e+00 3.14231753e-01 3.37561518e-01 -1.45105198e-01
8.48142505e-01 -2.57648438e-01 1.04385257e+00 9.86439526e-01
-1.95248082e-01 6.26264811e-01 3.50335270e-01 -4.68274295e-01
-7.84578681e-01 -8.47513258e-01 -1.32963195e-01 1.43460953e+00
1.65072680e-02 -6.18882596e-01 -6.12713814e-01 -1.01703250e+00
-2.83736616e-01 5.43602467e-01 -2.95108467e-01 8.57213810e-02
-8.27903628e-01 -1.13296211e+00 7.04583526e-01 6.59511924e-01
7.48332441e-01 -1.01009476e+00 -2.33447939e-01 2.96558499e-01
-4.04238343e-01 -1.02396405e+00 -5.78601182e-01 2.97007561e-01
-9.24913406e-01 -7.22301781e-01 -6.94476128e-01 -9.47395563e-01
6.66368425e-01 2.62649804e-01 6.12292767e-01 2.78510511e-01
-8.36842284e-02 -3.46676111e-01 -5.34448087e-01 -4.11365956e-01
-2.20812693e-01 2.41435170e-01 -1.51548803e-01 -8.39161649e-02
2.45064899e-01 -8.54032636e-02 -5.20107985e-01 5.27531803e-01
-7.13205457e-01 4.36629117e-01 1.04519594e+00 8.50590467e-01
8.80380929e-01 2.33901873e-01 5.24862647e-01 -7.41318345e-01
3.19421083e-01 1.63190216e-02 -6.82280362e-01 2.50074744e-01
-5.17082751e-01 8.89181569e-02 6.09287143e-01 -3.23804617e-01
-1.32908678e+00 4.74487662e-01 -4.60028321e-01 -1.02706030e-01
1.89371422e-01 2.45057285e-01 -3.39239091e-01 2.43106499e-01
3.55568528e-01 4.94199663e-01 -5.37413657e-01 -6.88241184e-01
5.16617149e-02 7.53370106e-01 4.05796558e-01 -2.30902195e-01
6.99926674e-01 4.91966933e-01 -3.06466460e-01 -7.71182179e-01
-9.48364854e-01 -5.84972680e-01 -4.66608107e-01 9.18767303e-02
5.97406209e-01 -1.09269893e+00 -6.20288014e-01 3.32192332e-01
-1.23045933e+00 1.44196833e-02 2.82848123e-02 3.22388828e-01
-6.37273937e-02 6.86921716e-01 -8.01289618e-01 -9.23545063e-01
-9.83870387e-01 -1.24576116e+00 1.43045974e+00 3.69991452e-01
1.55213987e-02 -2.61202246e-01 -2.63791829e-01 4.14205164e-01
6.18433543e-02 -6.27761781e-01 8.58989000e-01 -8.51328790e-01
-7.41079867e-01 -1.86170876e-01 -6.22704744e-01 1.39721539e-02
-2.53913879e-01 1.67806700e-01 -9.61611331e-01 -1.83499262e-01
-1.37809560e-01 9.34602767e-02 1.26787591e+00 4.39027935e-01
1.37194002e+00 -1.51063845e-01 -7.83473194e-01 4.06412631e-01
1.13914597e+00 -2.09206305e-02 6.50110245e-01 1.55606300e-01
6.53391957e-01 2.48284206e-01 8.48860323e-01 5.84795237e-01
3.38870406e-01 6.65612817e-01 3.12968165e-01 -8.56275782e-02
-1.36839613e-01 -3.13867837e-01 4.18786496e-01 9.06442404e-01
1.70271948e-01 -4.70093846e-01 -8.38689148e-01 2.81422645e-01
-2.01867461e+00 -7.80126154e-01 -4.28698480e-01 1.86974430e+00
8.92007291e-01 6.68507159e-01 -1.57649219e-01 1.67218670e-01
9.56487179e-01 1.01282828e-01 -3.89837503e-01 -2.78731305e-02
-8.17924887e-02 2.08408251e-01 4.56685036e-01 2.51312882e-01
-1.32013667e+00 1.25491357e+00 5.80045080e+00 1.42960584e+00
-6.73794270e-01 -7.82739893e-02 4.78736907e-01 2.85252593e-02
8.75844806e-02 -3.08890585e-02 -1.68483675e+00 2.58056045e-01
5.25222659e-01 3.85682613e-01 -1.00480311e-01 7.75646806e-01
2.46438906e-01 -1.65464714e-01 -9.08785105e-01 8.73021364e-01
9.48275551e-02 -1.40708542e+00 4.49743569e-01 -1.94227874e-01
5.03920317e-01 3.51560861e-02 -1.71859294e-01 6.43446088e-01
3.24150324e-01 -5.61077952e-01 6.84240758e-01 3.21481049e-01
4.35707837e-01 -7.74553239e-01 8.92283261e-01 7.45772660e-01
-1.57620108e+00 -2.65513033e-01 -6.45801961e-01 2.06525773e-01
1.43520907e-01 6.68638408e-01 -1.07935238e+00 6.59541011e-01
6.63712502e-01 6.15359008e-01 -6.15690053e-01 1.20154178e+00
-4.25551504e-01 7.48554289e-01 -3.69392931e-01 -1.73624560e-01
2.52751529e-01 3.58998170e-03 5.11356533e-01 1.63602412e+00
1.05495095e-01 -1.06508499e-02 4.07580674e-01 6.47778153e-01
-3.18031698e-01 2.14161515e-01 1.16033204e-01 4.70078588e-01
4.75642592e-01 1.47683156e+00 -1.04841101e+00 -4.89335030e-01
-3.89832675e-01 9.51664388e-01 2.91773528e-01 -2.81739384e-02
-1.01783502e+00 -4.99322653e-01 1.63437232e-01 -1.16976760e-01
8.27620387e-01 1.42017379e-01 -3.78614664e-01 -9.94584501e-01
6.75331429e-02 -8.24186325e-01 4.52845901e-01 -5.47149181e-01
-9.90251362e-01 2.93504119e-01 -3.51907998e-01 -9.12279725e-01
2.98534334e-01 -3.76563907e-01 -6.80412412e-01 6.43635213e-01
-1.38403165e+00 -1.27207923e+00 -1.62713937e-02 2.20881328e-01
1.17425978e+00 1.09589159e-01 4.28007662e-01 3.36712658e-01
-1.17426515e+00 8.74428749e-01 3.27669717e-02 3.51097286e-01
7.60907173e-01 -9.05869007e-01 6.37951612e-01 9.86015201e-01
-3.23332250e-02 4.75775987e-01 4.23760504e-01 -8.96596253e-01
-1.06688118e+00 -1.33469105e+00 7.07373679e-01 -1.85667291e-01
3.61293316e-01 -4.79287833e-01 -9.76415932e-01 4.68299299e-01
-1.89217970e-01 -2.44453430e-01 4.59052354e-01 1.42166568e-02
-2.15546563e-01 -1.42324522e-01 -9.50974286e-01 9.39511895e-01
1.02327788e+00 -1.09975776e-02 -7.82088399e-01 2.40396798e-01
7.87971318e-01 -2.95642823e-01 -4.33648556e-01 5.04524827e-01
8.01811635e-01 -7.49079823e-01 8.80929291e-01 6.38222024e-02
3.13453853e-01 -4.67416286e-01 -1.33867607e-01 -4.41242009e-01
-4.99661714e-01 -4.38768387e-01 -2.42280528e-01 1.45650136e+00
6.31516814e-01 -1.30503953e-01 1.09139490e+00 4.30809557e-02
-3.67033273e-01 -7.69017994e-01 -9.88740027e-01 -3.54476511e-01
-2.90276855e-01 -5.81788838e-01 4.95807379e-01 3.65197092e-01
-1.53003782e-01 5.80392122e-01 -4.65308249e-01 2.77139127e-01
4.82552171e-01 2.30679885e-02 7.51728594e-01 -1.27491355e+00
-4.94777262e-01 -4.53235686e-01 1.22052029e-01 -1.77493095e+00
-1.08100325e-01 -7.36332953e-01 3.52204293e-01 -1.54717171e+00
6.24283910e-01 -4.11306977e-01 -2.60503978e-01 6.71328545e-01
-5.44155955e-01 9.27181169e-02 2.19496459e-01 4.38863188e-01
-1.06208968e+00 5.80293536e-01 1.18066871e+00 -1.96531460e-01
-4.20552999e-01 2.14003101e-01 -6.92277193e-01 1.11579907e+00
6.05759144e-01 -6.33570969e-01 -9.73679721e-02 -5.04074812e-01
5.08033149e-02 -6.93942159e-02 -4.40160632e-02 -9.15663242e-01
4.24392134e-01 1.80826530e-01 7.67797649e-01 -1.36274672e+00
3.18094879e-01 -7.18494117e-01 -3.23121488e-01 7.24457443e-01
-2.57065833e-01 -1.02716476e-01 1.48286819e-01 8.71640444e-01
1.47939965e-01 -2.78182030e-01 7.76690960e-01 9.29620937e-02
-5.31121850e-01 3.23368639e-01 -5.10740876e-01 -2.36565724e-01
7.56897688e-01 -4.12013948e-01 -2.11305663e-01 8.74199644e-02
-3.05380225e-01 5.79836607e-01 1.38125464e-01 2.97155470e-01
7.65928864e-01 -8.15012753e-01 -6.74045682e-01 3.16084683e-01
5.88453934e-02 3.48407865e-01 1.93010673e-01 9.72842515e-01
-2.35786766e-01 4.12684470e-01 2.57260293e-01 -6.78840458e-01
-1.52550685e+00 4.01832163e-01 -4.14069854e-02 -7.97333479e-01
-7.94570982e-01 1.03831875e+00 4.77921009e-01 -4.91047889e-01
5.53017259e-01 -3.98518533e-01 -4.52799827e-01 -1.39090894e-02
6.77999079e-01 4.04796571e-01 -1.14237918e-02 -3.22251141e-01
-4.52563941e-01 4.03854728e-01 -7.58706689e-01 9.58502293e-02
1.12866199e+00 -9.66435373e-02 -3.58044542e-02 -1.29280537e-01
5.54057539e-01 1.21735565e-01 -1.17982292e+00 -3.95681381e-01
1.66598573e-01 -1.87128276e-01 2.55407035e-01 -8.42166245e-01
-8.94428432e-01 7.44219363e-01 4.08730924e-01 -2.82227755e-01
1.16738558e+00 -1.97591662e-01 8.52162540e-01 5.54841638e-01
9.67014059e-02 -1.15006971e+00 1.46988675e-01 6.56764805e-01
5.75277925e-01 -9.33681130e-01 3.04860294e-01 -6.00737035e-01
-5.33545315e-01 1.02844882e+00 8.73651326e-01 6.77813739e-02
3.13237071e-01 4.15314287e-01 -4.20435160e-01 1.48008317e-01
-8.90070260e-01 -5.58008738e-02 1.86931193e-01 2.93388337e-01
3.29839081e-01 -1.55923367e-01 -5.40460885e-01 9.53093648e-01
2.10368946e-01 -6.79198876e-02 3.86482000e-01 8.07212591e-01
-9.89219725e-01 -1.08258998e+00 -5.01753747e-01 9.11015034e-01
-6.60209715e-01 -4.76469576e-01 -3.99910808e-01 6.17209733e-01
2.28151530e-01 9.21629488e-01 -2.56515332e-02 -4.98606414e-01
3.04158986e-01 1.34004448e-02 5.25586344e-02 -7.74136961e-01
-7.65462697e-01 6.17597044e-01 3.12304180e-02 -3.11323434e-01
-5.44761829e-02 -5.99153757e-01 -1.53417242e+00 -1.58778355e-01
-1.03996837e+00 9.62283164e-02 3.62378031e-01 1.00715125e+00
4.17936414e-01 6.20932162e-01 3.78860652e-01 -6.52224243e-01
-6.64179862e-01 -1.12363088e+00 -2.90901661e-01 -4.22369719e-01
1.52351454e-01 -3.93924117e-01 -2.46128682e-02 -1.24649882e-01] | [12.062909126281738, 2.264540910720825] |
5bdffefb-a07a-4002-ae81-166c221c19bf | low-resource-neural-machine-translation-for | 2104.00366 | null | https://arxiv.org/abs/2104.00366v2 | https://arxiv.org/pdf/2104.00366v2.pdf | Low-Resource Neural Machine Translation for Southern African Languages | Low-resource African languages have not fully benefited from the progress in neural machine translation because of a lack of data. Motivated by this challenge we compare zero-shot learning, transfer learning and multilingual learning on three Bantu languages (Shona, isiXhosa and isiZulu) and English. Our main target is English-to-isiZulu translation for which we have just 30,000 sentence pairs, 28% of the average size of our other corpora. We show the importance of language similarity on the performance of English-to-isiZulu transfer learning based on English-to-isiXhosa and English-to-Shona parent models whose BLEU scores differ by 5.2. We then demonstrate that multilingual learning surpasses both transfer learning and zero-shot learning on our dataset, with BLEU score improvements relative to the baseline English-to-isiZulu model of 9.9, 6.1 and 2.0 respectively. Our best model also improves the previous SOTA BLEU score by more than 10. | ['Bruce A. Bassett', 'Evander Nyoni'] | 2021-04-01 | null | null | null | null | ['low-resource-neural-machine-translation'] | ['natural-language-processing'] | [ 1.38103394e-02 -2.94757597e-02 -4.60711002e-01 -1.51079193e-01
-1.67049384e+00 -6.94614589e-01 7.79096305e-01 3.90775315e-03
-6.04589105e-01 1.41525126e+00 3.69276613e-01 -7.24013209e-01
1.92895383e-01 -6.34747207e-01 -8.58331084e-01 -4.03981745e-01
2.09377617e-01 8.94954860e-01 1.93140358e-02 -7.08662689e-01
-7.76657164e-02 -1.55928373e-01 -7.85522342e-01 2.61815131e-01
1.13026333e+00 1.46409467e-01 2.44331822e-01 7.21955419e-01
-3.74160744e-02 4.21125829e-01 -2.75167555e-01 -4.70476747e-01
2.96474010e-01 -8.13542783e-01 -8.68822813e-01 -5.18212080e-01
7.82797694e-01 -3.97406936e-01 -2.75553763e-01 9.23285127e-01
7.98866868e-01 2.70496272e-02 9.20746982e-01 -6.70277119e-01
-1.04793239e+00 8.93733561e-01 -5.23014128e-01 2.68567741e-01
2.08377331e-01 -6.59135878e-02 1.16639602e+00 -1.23316205e+00
1.08271134e+00 1.14636087e+00 7.70389259e-01 6.51960671e-01
-1.05551946e+00 -7.88247108e-01 -7.49921799e-01 1.32779971e-01
-1.09461713e+00 -7.85564184e-01 1.61791012e-01 -4.70237285e-01
1.41507089e+00 -1.03405684e-01 5.56291193e-02 1.22949851e+00
3.72944504e-01 5.29181123e-01 1.31511128e+00 -1.01985645e+00
-5.56800850e-02 1.11754745e-01 -1.82029992e-01 6.21684313e-01
1.22828104e-01 5.16989939e-02 -6.45434916e-01 6.20571263e-02
5.33877254e-01 -4.49566543e-01 1.82783246e-01 -1.05083182e-01
-1.38637650e+00 1.00463974e+00 3.37605998e-02 5.33254087e-01
-1.27209695e-02 -1.01015814e-01 6.24223292e-01 8.71117771e-01
8.27560902e-01 4.35565084e-01 -8.22516084e-01 -4.55508530e-01
-9.16611969e-01 -1.08512389e-02 6.93284810e-01 1.13753200e+00
8.58479738e-01 2.44127616e-01 -5.54443300e-02 1.17512143e+00
-2.29815334e-01 9.21667874e-01 4.39237028e-01 -5.88290572e-01
8.59019041e-01 -1.03425920e-01 2.39574909e-01 -1.68873712e-01
4.89623062e-02 -4.96948175e-02 -5.36826968e-01 -1.87757865e-01
6.23468757e-01 -6.88949227e-01 -1.03901172e+00 1.60186481e+00
-1.09428115e-01 -1.91764861e-01 5.70810735e-01 4.94191766e-01
5.60977697e-01 1.22085583e+00 1.87384393e-02 -2.90242761e-01
1.10410750e+00 -1.18411410e+00 -6.29961431e-01 -3.35013390e-01
1.07101786e+00 -1.20347404e+00 1.19920456e+00 -1.89730987e-01
-1.19130671e+00 -5.75250745e-01 -1.07044339e+00 -1.96044981e-01
-4.74881440e-01 -5.01082242e-02 4.56370771e-01 4.62166280e-01
-8.67249668e-01 7.46618152e-01 -6.61237597e-01 -9.94195282e-01
1.40270501e-01 2.39199772e-01 -5.92766523e-01 -3.19347501e-01
-1.65284121e+00 1.51412165e+00 4.54265505e-01 -6.62497103e-01
-7.08327830e-01 -8.16057622e-01 -9.65148926e-01 -6.00473695e-02
8.60383734e-02 -2.54230648e-01 1.36995685e+00 -7.60332525e-01
-1.41000116e+00 1.00610292e+00 3.41823474e-02 -4.65210497e-01
4.62329000e-01 -2.54847080e-01 -4.48479354e-01 -1.12996377e-01
4.40184057e-01 6.68197870e-01 3.24561894e-01 -7.14970469e-01
-7.70853996e-01 -2.71435440e-01 -2.90605724e-01 2.28494912e-01
-2.36234099e-01 6.14149868e-01 -1.58688501e-01 -5.33834696e-01
-3.90568882e-01 -9.32482839e-01 8.18356350e-02 -5.92033207e-01
3.05021167e-01 -3.52772951e-01 5.43973267e-01 -1.18425214e+00
1.01432538e+00 -1.60907221e+00 3.86832729e-02 -4.28864121e-01
-3.97201329e-01 5.33785999e-01 -5.64818859e-01 9.55510020e-01
2.01062873e-01 -3.08574941e-02 -1.83764473e-01 6.92841932e-02
-1.60237864e-01 3.05609435e-01 -6.40695468e-02 3.33048105e-01
3.83212179e-01 1.13041115e+00 -1.35590637e+00 -3.71024281e-01
4.62458245e-02 2.20636815e-01 -2.39265338e-01 1.47645131e-01
8.45168009e-02 3.33393931e-01 9.22147855e-02 6.13878548e-01
3.03920984e-01 3.36081922e-01 1.53530702e-01 3.56268406e-01
-3.54750812e-01 5.26795208e-01 -2.60631412e-01 1.97575390e+00
-8.17229509e-01 8.35890830e-01 -3.09357792e-01 -6.29890263e-01
1.02028978e+00 5.91191769e-01 2.29556859e-01 -9.50844169e-01
9.24838483e-02 9.36942339e-01 5.51805615e-01 -3.27913642e-01
3.86339217e-01 -4.73663270e-01 -3.26537520e-01 5.15062094e-01
6.59918308e-01 -1.19381137e-01 4.00379151e-01 2.27776747e-02
8.07490587e-01 4.16663498e-01 3.98504853e-01 -6.81411266e-01
1.13276377e-01 1.89950436e-01 5.86632431e-01 5.84462881e-01
-3.72437835e-01 5.02486706e-01 2.83409327e-01 -3.68369609e-01
-1.78226674e+00 -1.30150890e+00 -1.95540816e-01 1.34784079e+00
-4.97187614e-01 -2.15586856e-01 -9.61253524e-01 -7.08835185e-01
-3.46393287e-01 1.03501296e+00 -5.38305044e-01 3.53774838e-02
-6.82882845e-01 -4.97042567e-01 6.96106255e-01 3.13527703e-01
2.13020548e-01 -9.60431337e-01 -2.93747693e-01 2.38087282e-01
-2.59968579e-01 -1.01708174e+00 -8.22878599e-01 2.33727515e-01
-7.29928434e-01 -4.27053720e-01 -1.26029789e+00 -1.11202872e+00
4.94211875e-02 7.35817999e-02 1.18387020e+00 -6.60091817e-01
-2.04925295e-02 -1.62548468e-01 -4.93122220e-01 -4.87568736e-01
-7.86141276e-01 5.27931750e-01 3.84385943e-01 -6.41125917e-01
8.48762870e-01 -4.14202571e-01 1.16864912e-01 -6.46234080e-02
-3.80047500e-01 4.45145033e-02 6.58811808e-01 1.20487535e+00
-5.62139675e-02 -7.64315724e-01 1.02551639e+00 -1.04187191e+00
4.68933672e-01 -6.09381378e-01 -3.47328186e-01 5.52409649e-01
-7.34124839e-01 6.08859025e-03 5.81145108e-01 -3.14069331e-01
-1.13831556e+00 -1.89034402e-01 6.56161532e-02 -2.42827162e-01
-3.23977396e-02 4.80952680e-01 1.55524850e-01 2.05141485e-01
8.91987383e-01 -1.54837787e-01 -1.76128164e-01 -3.01186115e-01
5.93812168e-01 1.20826674e+00 4.74513918e-01 -4.88576233e-01
6.83149993e-01 -2.37436712e-01 -6.03531122e-01 -9.23807025e-01
-6.26559138e-01 -4.71170783e-01 -1.06756687e+00 1.64767474e-01
1.01734805e+00 -1.13056254e+00 1.85902998e-01 3.68487626e-01
-1.44764614e+00 -6.14363730e-01 -1.38465732e-01 7.54357398e-01
-6.24801278e-01 1.05591811e-01 -1.11174512e+00 -5.19548237e-01
-7.51872241e-01 -9.42970514e-01 8.44297051e-01 -6.67453557e-02
-4.21438128e-01 -1.25176442e+00 5.57440460e-01 3.20058435e-01
4.51611847e-01 -8.62308592e-02 1.21777940e+00 -7.85187423e-01
1.33847758e-01 9.25351828e-02 -3.20382804e-01 2.69243062e-01
3.56512398e-01 -2.82949418e-01 -7.01867759e-01 -4.32351887e-01
-4.64279979e-01 -7.82972217e-01 5.31975865e-01 9.39700603e-02
-3.08147490e-01 -5.77744432e-02 1.35277376e-01 1.44089893e-01
1.51126873e+00 4.43466783e-01 5.12472451e-01 3.49351794e-01
5.27578115e-01 5.81461549e-01 6.73552394e-01 1.02640642e-02
3.20022076e-01 5.67972660e-01 -4.16837275e-01 -9.48674530e-02
-4.74217653e-01 -5.41889131e-01 8.20931852e-01 1.72357523e+00
-5.53236343e-02 1.17372625e-01 -1.31312454e+00 1.13260615e+00
-1.74860322e+00 -7.39205658e-01 2.91458927e-02 2.26017928e+00
1.25930798e+00 -2.54785623e-02 9.37049370e-03 -5.91271758e-01
6.75501823e-01 4.18415330e-02 -1.59573674e-01 -1.03129590e+00
-6.95460141e-02 7.21661747e-01 6.34076953e-01 8.21039438e-01
-9.71538186e-01 1.69533730e+00 6.15986252e+00 1.03034282e+00
-1.00792825e+00 4.88119096e-01 5.45553088e-01 5.49467579e-02
-1.69724286e-01 7.14923963e-02 -8.56719851e-01 2.22180128e-01
1.70483601e+00 -4.37780768e-01 6.67980075e-01 4.81665432e-01
1.11339077e-01 2.54324436e-01 -1.10459864e+00 7.09680140e-01
3.07052612e-01 -1.16735053e+00 -3.45668867e-02 -5.21626621e-02
1.48986125e+00 6.93319559e-01 -1.43759921e-01 7.92135477e-01
5.88650048e-01 -9.34940934e-01 3.06946248e-01 4.16600145e-02
1.27252591e+00 -9.54599977e-01 8.09929788e-01 2.94395417e-01
-6.29025102e-01 5.38481474e-01 -8.28801334e-01 -2.29814380e-01
9.94223654e-02 1.02708802e-01 -1.02100635e+00 6.39612198e-01
4.57911372e-01 6.09823465e-01 -2.79188812e-01 5.66254377e-01
-2.52781570e-01 9.16842401e-01 4.44534309e-02 -1.20049104e-01
8.17120552e-01 -4.24129009e-01 2.37454847e-01 1.40872586e+00
7.90760577e-01 -1.74689740e-01 1.25303879e-01 3.23723078e-01
-3.36874515e-01 6.21598780e-01 -1.07715881e+00 -2.90925443e-01
4.03285414e-01 9.76631105e-01 -2.00266972e-01 -5.09859025e-01
-6.99751496e-01 1.32053864e+00 5.07594824e-01 2.66224951e-01
-5.72455823e-01 -9.34802592e-01 5.05476475e-01 -1.51786253e-01
2.67928481e-01 -3.04344952e-01 -1.17398560e-01 -1.30210352e+00
-3.10624897e-01 -1.01010430e+00 8.30946714e-02 -6.22152150e-01
-1.32828200e+00 6.26932740e-01 -1.40598267e-01 -1.06172752e+00
-7.66772568e-01 -5.53580105e-01 -6.57321751e-01 1.30411923e+00
-1.34833968e+00 -1.57123530e+00 6.28003359e-01 1.74264446e-01
1.11525607e+00 -6.22675598e-01 1.26649666e+00 3.99642557e-01
-1.97403342e-01 6.91751420e-01 6.64227307e-01 2.02972054e-01
1.27117848e+00 -1.15577948e+00 9.52319741e-01 8.49044323e-01
3.11396241e-01 5.00833690e-01 6.10460699e-01 -6.54064476e-01
-1.22057939e+00 -9.91410196e-01 1.83312237e+00 -6.68538749e-01
1.22132587e+00 -4.99772280e-01 -6.81623042e-01 8.27847660e-01
1.01564932e+00 -3.92810136e-01 7.25386739e-01 3.84620607e-01
-5.11600494e-01 1.10186689e-01 -8.76025975e-01 8.31558108e-01
7.50077963e-01 -7.79263556e-01 -9.06595409e-01 4.08494473e-01
8.52542996e-01 -2.49529332e-02 -1.09106934e+00 1.56108871e-01
6.46996140e-01 -4.40616459e-01 5.07727265e-01 -1.22704613e+00
8.05898070e-01 2.87447929e-01 -3.91475767e-01 -1.77031267e+00
-3.95958215e-01 -6.96860313e-01 3.61742467e-01 1.25342119e+00
8.88580978e-01 -2.02093795e-01 3.94301802e-01 3.35537642e-02
-5.13692200e-02 -3.86380792e-01 -1.08978653e+00 -1.05395293e+00
1.11223221e+00 -3.52114625e-02 8.15594196e-02 1.39352500e+00
3.24301660e-01 1.15512896e+00 -8.84087145e-01 -3.33766520e-01
5.25785089e-01 -1.67901561e-01 6.24617755e-01 -8.39889646e-01
-3.81478310e-01 -1.23896293e-01 -5.43870777e-03 -4.67023373e-01
2.30964482e-01 -1.25262821e+00 2.30103835e-01 -1.66800106e+00
3.39936405e-01 -1.33019492e-01 -2.25801677e-01 6.46988273e-01
-2.99456835e-01 5.68956852e-01 4.07081634e-01 1.36152729e-01
-2.68462032e-01 3.31508785e-01 1.10933805e+00 -2.37984657e-01
-1.33891255e-01 -3.88760090e-01 -3.08523506e-01 3.76614183e-01
9.67535794e-01 -3.71255755e-01 -1.45922035e-01 -8.42951417e-01
-4.38167565e-02 1.21244036e-01 -6.24440134e-01 -7.50075638e-01
-1.32464007e-01 -1.69415176e-01 6.55946210e-02 -1.54258341e-01
4.03108113e-02 -2.64310122e-01 -3.17462623e-01 5.96430957e-01
-2.96810299e-01 1.60570994e-01 2.24206075e-01 -5.78689277e-02
-9.33324397e-02 -4.08063918e-01 8.40204597e-01 -2.98371613e-01
-7.56246567e-01 1.09242819e-01 -6.03133798e-01 1.86527535e-01
8.89581382e-01 2.54410625e-01 -4.05845225e-01 -3.60211909e-01
-4.35507774e-01 2.66673900e-02 3.29606444e-01 6.75952375e-01
1.79982468e-01 -1.52137935e+00 -1.33820343e+00 1.04849488e-01
3.61123711e-01 -8.26408148e-01 -1.59388185e-01 8.20839882e-01
-6.33703113e-01 7.38364100e-01 -7.31536150e-01 -1.52994931e-01
-1.06747818e+00 3.74353379e-01 -2.07964107e-01 -3.86288017e-01
-1.75397277e-01 5.47406673e-01 -1.75685763e-01 -1.08119786e+00
-1.60970449e-01 8.19687992e-02 2.38820791e-01 1.05295926e-01
2.81613946e-01 6.17929459e-01 8.30072761e-02 -7.08146036e-01
-3.17313820e-02 5.11077464e-01 -3.18301797e-01 -4.73628312e-01
1.25605762e+00 -4.06953134e-02 -9.26886722e-02 8.95349205e-01
1.30862164e+00 2.78333247e-01 -7.80719101e-01 -3.00021261e-01
1.68899223e-01 -1.71845064e-01 -2.64893711e-01 -9.32067335e-01
-2.93329805e-01 1.41207957e+00 5.18951058e-01 -4.71431822e-01
5.50547242e-01 -9.84533802e-02 1.31860328e+00 4.43715036e-01
6.07515097e-01 -1.34935164e+00 -3.23327690e-01 1.11990130e+00
4.00703609e-01 -1.51178396e+00 -4.05392557e-01 5.16778007e-02
-6.34529352e-01 9.98034239e-01 2.74420500e-01 -1.38826802e-01
2.03713536e-01 1.17698722e-01 4.89916503e-01 4.93122846e-01
-8.69369149e-01 -3.62271994e-01 1.80165380e-01 7.22203135e-01
9.01296437e-01 3.02915782e-01 -7.30864346e-01 2.11644679e-01
-1.77386239e-01 -8.65730364e-03 4.39989209e-01 6.37958944e-01
-7.18442857e-01 -1.45515430e+00 -1.40622973e-01 2.27408677e-01
-7.23963916e-01 -6.47107184e-01 -2.75164008e-01 8.12729180e-01
-7.42113367e-02 9.25766468e-01 1.18969612e-01 -2.47618958e-01
8.42021685e-03 4.26079601e-01 6.51647031e-01 -8.74751329e-01
-6.03658140e-01 4.87690032e-01 3.96909505e-01 1.98892727e-01
-7.77576491e-02 -6.13295853e-01 -9.86593723e-01 -3.83586228e-01
-1.86969176e-01 3.86353254e-01 6.11284554e-01 8.96913469e-01
1.63021371e-01 3.96956466e-02 4.16644186e-01 -5.53556323e-01
-7.03195751e-01 -1.45513928e+00 -3.39620501e-01 1.65075973e-01
-5.38191684e-02 -1.29140288e-01 9.75376070e-02 8.91376510e-02] | [11.438811302185059, 10.241499900817871] |
e83d9bd1-95e4-4f83-8f66-e1e12bc40d52 | underwater-image-restoration-via-contrastive | 2106.10718 | null | https://arxiv.org/abs/2106.10718v1 | https://arxiv.org/pdf/2106.10718v1.pdf | Underwater Image Restoration via Contrastive Learning and a Real-world Dataset | Underwater image restoration is of significant importance in unveiling the underwater world. Numerous techniques and algorithms have been developed in the past decades. However, due to fundamental difficulties associated with imaging/sensing, lighting, and refractive geometric distortions, in capturing clear underwater images, no comprehensive evaluations have been conducted of underwater image restoration. To address this gap, we have constructed a large-scale real underwater image dataset, dubbed `HICRD' (Heron Island Coral Reef Dataset), for the purpose of benchmarking existing methods and supporting the development of new deep-learning based methods. We employ accurate water parameter (diffuse attenuation coefficient) in generating reference images. There are 2000 reference restored images and 6003 original underwater images in the unpaired training set. Further, we present a novel method for underwater image restoration based on unsupervised image-to-image translation framework. Our proposed method leveraged contrastive learning and generative adversarial networks to maximize the mutual information between raw and restored images. Extensive experiments with comparisons to recent approaches further demonstrate the superiority of our proposed method. Our code and dataset are publicly available at GitHub. | ['Lars Petersson', 'Hongdong Li', 'Mohammad Ali Armin', 'Ran Wei', 'Saeed Anwar', 'Janet Anstee', 'Elizabeth Botha', 'Tim Malthus', 'Mehrdad Shoeiby', 'Junlin Han'] | 2021-06-20 | null | null | null | null | ['underwater-image-restoration'] | ['computer-vision'] | [ 6.47450447e-01 -3.13055515e-02 8.47575665e-01 -3.91673863e-01
-9.30357456e-01 -2.93131977e-01 3.74429524e-01 -3.87994438e-01
-6.35488331e-01 7.29604721e-01 4.68121767e-01 -1.01281010e-01
-9.92548466e-02 -8.97761047e-01 -8.87492478e-01 -9.87642407e-01
-2.06902891e-01 -2.78016806e-01 -2.46858206e-02 -3.63625526e-01
5.05473256e-01 2.89328158e-01 -1.49705660e+00 -2.04864711e-01
1.11848450e+00 7.39434779e-01 4.61123675e-01 7.80289352e-01
1.79542989e-01 6.36209071e-01 -5.00232816e-01 -3.59803021e-01
6.68693602e-01 -7.12688088e-01 -3.80524904e-01 1.73802972e-01
3.95213336e-01 -8.31831098e-01 -7.19971061e-01 1.28095222e+00
9.88608718e-01 4.91893619e-01 4.88680035e-01 -6.58743858e-01
-1.12292171e+00 3.91204298e-01 -3.89005661e-01 2.03353748e-01
1.30966380e-01 1.34207606e-01 6.90859914e-01 -8.44100654e-01
3.05639774e-01 1.03143728e+00 5.94359398e-01 6.29118383e-01
-7.27161944e-01 -7.14110911e-01 -3.23674947e-01 3.55477817e-02
-9.35984910e-01 -7.00774074e-01 6.77210152e-01 -2.30678722e-01
4.21467721e-01 8.21262226e-02 6.37834430e-01 6.41417921e-01
3.52274418e-01 3.86207104e-01 1.45111549e+00 -4.20912027e-01
1.19284645e-01 -4.43534791e-01 -5.40178478e-01 5.29645860e-01
8.14965814e-02 4.03843731e-01 -3.92137676e-01 1.21759109e-01
8.44485581e-01 2.47243032e-01 -6.80650413e-01 1.59270182e-01
-6.57855690e-01 6.46492183e-01 6.48266494e-01 -3.43964398e-01
-7.46393576e-02 1.66941639e-02 5.06867357e-02 5.40558517e-01
4.01650757e-01 9.77965146e-02 -1.01809703e-01 2.83132970e-01
-5.08834660e-01 1.00761496e-01 5.70875585e-01 5.85239291e-01
9.18339193e-01 4.86328453e-01 5.10567486e-01 1.14993274e+00
7.57815182e-01 8.20134997e-01 5.55784643e-01 -1.27419698e+00
2.75153637e-01 2.18427464e-01 9.03596058e-02 -9.62740302e-01
-7.88028911e-02 -6.39721081e-02 -9.66143489e-01 4.97955263e-01
-8.13797191e-02 -1.58176452e-01 -9.97883141e-01 1.17620289e+00
1.13072731e-01 3.06143105e-01 7.61968195e-01 1.07094204e+00
1.03976369e+00 7.73413002e-01 -2.86834925e-01 -1.53860345e-01
7.55501449e-01 -7.58194387e-01 -6.44555569e-01 -2.92499393e-01
-7.75317773e-02 -8.06575119e-01 8.71554971e-01 1.61832735e-01
-9.74086761e-01 -3.96982998e-01 -1.31945014e+00 -1.11482248e-01
-6.73865303e-02 -3.95186007e-01 3.53792399e-01 5.94952285e-01
-1.18686163e+00 4.49256837e-01 -6.97286606e-01 -7.16464296e-02
3.79332006e-01 1.90688521e-01 -5.46692312e-01 -5.35652757e-01
-1.08267713e+00 6.35500133e-01 -1.14593673e-02 6.19680882e-01
-1.45171428e+00 -3.98393780e-01 -1.16080582e+00 -3.37945312e-01
-2.67890185e-01 -3.39767456e-01 1.02181256e+00 -8.43499541e-01
-1.53284943e+00 6.34442627e-01 2.44433284e-01 -2.77735710e-01
3.92072886e-01 -3.63998145e-01 -2.96558440e-01 2.21124128e-01
4.05724458e-02 3.91937643e-01 6.06621742e-01 -1.78927791e+00
-4.47432011e-01 -3.84367049e-01 9.72615033e-02 4.83270645e-01
-1.74908400e-01 -9.01435763e-02 -3.54324758e-01 -6.64033353e-01
4.63299543e-01 -6.88634753e-01 -4.26423997e-01 2.51385689e-01
-2.04507902e-01 6.07900918e-01 6.70234263e-01 -9.59506571e-01
3.99459302e-01 -2.06487799e+00 -1.37572573e-03 -1.89473242e-01
-1.28738835e-01 1.97943881e-01 -4.68399763e-01 6.99891627e-01
1.22721739e-01 4.87744510e-02 -8.46979737e-01 -5.22992551e-01
-3.24088365e-01 5.87594867e-01 -4.46054876e-01 9.38316405e-01
-1.76524431e-01 5.09909451e-01 -9.66832161e-01 -3.42193395e-01
3.95871580e-01 7.67138839e-01 -4.75590378e-01 6.43988907e-01
2.72184074e-01 8.03405881e-01 -1.48604646e-01 8.89019012e-01
9.86271739e-01 4.31831509e-01 6.83722552e-03 -2.11019069e-01
-2.78563648e-01 -1.38221025e-01 -7.95740962e-01 1.58311117e+00
-5.79926491e-01 7.26466835e-01 3.10867101e-01 -9.48565125e-01
1.03509009e+00 7.82836154e-02 2.51633555e-01 -9.79546964e-01
8.81975889e-02 3.22924882e-01 -2.35571787e-01 -9.08293724e-01
5.97375751e-01 -5.50107837e-01 4.32747453e-01 2.18028098e-01
-1.86383486e-01 -3.49164456e-01 3.00513636e-02 4.96237492e-03
7.91613579e-01 3.70708883e-01 -1.68577477e-01 -1.57721832e-01
1.83749437e-01 -3.08294922e-01 5.81405818e-01 5.91095924e-01
-1.67533979e-01 1.25390542e+00 -2.20188409e-01 -3.86284560e-01
-1.05295527e+00 -1.02855170e+00 -2.62203306e-01 6.77636445e-01
7.15623915e-01 4.23289269e-01 -5.07649839e-01 1.11284174e-01
-3.29704136e-01 1.14254229e-01 -6.87829852e-01 9.83192697e-02
-4.67366964e-01 -1.12245119e+00 6.82480693e-01 2.22591028e-01
8.35034788e-01 -1.18932629e+00 -3.24975759e-01 2.18389392e-01
-2.41578743e-01 -9.76353049e-01 -3.11502367e-01 -1.29043326e-01
-1.10182655e+00 -1.19175875e+00 -8.65338027e-01 -1.05520225e+00
8.41417015e-01 7.00386643e-01 8.30382228e-01 3.41805905e-01
-2.84310818e-01 2.46918201e-01 -6.61960661e-01 -2.06448957e-01
-4.91317302e-01 -8.04319382e-01 1.04014445e-02 -3.83727215e-02
-4.48791653e-01 -8.45528901e-01 -1.17498577e+00 3.16928715e-01
-1.54871142e+00 -1.96785718e-01 6.94362819e-01 8.67998958e-01
4.12553310e-01 -5.35635613e-02 4.08700556e-01 -6.47549987e-01
3.93394858e-01 -5.71527362e-01 -6.03094399e-01 -2.28524189e-02
-4.19625223e-01 -2.46899575e-01 3.51818830e-01 7.62910396e-02
-1.13828862e+00 -1.15373373e-01 -6.22658193e-01 6.72206059e-02
7.26583675e-02 6.58201575e-01 -2.57545095e-02 -5.47547281e-01
4.45658892e-01 6.52720094e-01 3.66389900e-01 -5.18981874e-01
-1.96000026e-03 8.88254285e-01 8.36375892e-01 -2.87983835e-01
1.03968441e+00 8.61214221e-01 -3.23505968e-01 -1.04625726e+00
-7.09972084e-01 -3.11702251e-01 -2.27669373e-01 -3.53244334e-01
6.50338233e-01 -1.35833418e+00 -5.83735347e-01 1.01613903e+00
-7.39230990e-01 -5.61554015e-01 1.22667022e-01 3.60082299e-01
-8.98154899e-02 9.33291376e-01 -8.08781505e-01 -8.48995626e-01
-6.79165959e-01 -1.27943206e+00 8.25548708e-01 5.92097819e-01
8.24546814e-01 -1.00676453e+00 3.91221553e-01 6.22825921e-01
6.89485669e-01 4.68739778e-01 2.60843992e-01 1.34705901e-01
-5.32415032e-01 -1.24928094e-01 -3.90032262e-01 7.77066886e-01
4.89941925e-01 -1.07462421e-01 -9.45880890e-01 -5.10357857e-01
1.34799019e-01 -4.98599738e-01 1.04551017e+00 2.63101816e-01
7.63723493e-01 -2.88574696e-01 2.54742056e-01 9.94225085e-01
1.74697363e+00 2.41686657e-01 1.35979104e+00 6.64131820e-01
5.72823226e-01 5.66823602e-01 3.15836608e-01 4.47248042e-01
5.65885007e-01 1.37140736e-01 9.71823275e-01 -3.56367201e-01
-1.33233055e-01 -1.22886844e-01 4.45495784e-01 9.18351233e-01
-4.35105741e-01 -3.62438262e-01 -5.45193136e-01 7.46733665e-01
-1.15186119e+00 -7.57245362e-01 6.37819897e-03 1.99343503e+00
6.83739662e-01 -3.76025528e-01 -6.88974857e-01 -1.41038420e-02
5.39130807e-01 1.74111739e-01 -5.55849671e-01 2.51162112e-01
-4.36372280e-01 9.43118408e-02 6.14286721e-01 7.23572195e-01
-9.59012687e-01 5.87212861e-01 6.30355024e+00 1.20817110e-01
-1.00008345e+00 -1.10758461e-01 5.31298339e-01 4.96237606e-01
-5.66902518e-01 9.91786923e-03 -3.37852091e-01 3.92505318e-01
5.91202557e-01 3.06392282e-01 5.02661586e-01 4.28546488e-01
5.02356172e-01 -9.42201465e-02 -3.32120389e-01 7.49479353e-01
3.52516204e-01 -1.11164367e+00 1.14596829e-01 8.85135308e-02
1.09413838e+00 3.64894301e-01 -1.27237812e-01 -3.92617397e-02
5.21303058e-01 -9.84148324e-01 5.57765543e-01 5.53475976e-01
7.94242561e-01 -5.22650123e-01 1.06197608e+00 -3.16262133e-02
-8.11685443e-01 -1.26920105e-03 -6.96805418e-01 -2.16936558e-01
1.78100795e-01 2.92920440e-01 -2.85823673e-01 6.30774379e-01
1.03987849e+00 7.90643990e-01 -2.42496759e-01 1.26628280e+00
-3.92669737e-01 6.22678220e-01 -1.49817392e-01 4.72354472e-01
-1.51249319e-02 -7.28736341e-01 3.65177393e-01 9.21188891e-01
5.45798481e-01 4.24605995e-01 -1.64067835e-01 4.36824530e-01
-3.94810647e-01 2.50564944e-02 -5.76979518e-01 2.08385095e-01
2.90797740e-01 1.07586074e+00 -4.41224635e-01 1.02061458e-01
-4.13497627e-01 9.61081088e-01 -1.90056816e-01 4.89648759e-01
-4.81595218e-01 -4.10678089e-01 6.99835598e-01 -2.20953718e-01
5.56065924e-02 -4.21018928e-01 1.80388726e-02 -1.21425068e+00
-2.59678781e-01 -8.13706040e-01 4.63937819e-02 -9.30127084e-01
-1.41500258e+00 7.40311623e-01 -3.29692334e-01 -1.48760164e+00
3.31880540e-01 -4.68724936e-01 -7.23604202e-01 7.92030632e-01
-2.32352805e+00 -1.03510892e+00 -9.50090826e-01 3.49552095e-01
6.73957646e-01 -2.64912974e-02 7.98265755e-01 4.11987633e-01
-3.58006626e-01 2.13061526e-01 8.05999339e-01 3.96829426e-01
6.54588878e-01 -1.05267107e+00 3.06152999e-01 1.25777316e+00
-3.70191991e-01 5.87209404e-01 9.43706512e-01 -4.80496615e-01
-1.79421210e+00 -9.86233294e-01 3.32250744e-02 1.45401821e-01
6.17542922e-01 1.86873898e-01 -1.02216351e+00 5.58549941e-01
4.66437191e-01 1.08301505e-01 8.14171493e-01 -8.75769615e-01
-3.22056152e-02 -3.52117062e-01 -1.14851868e+00 5.32921910e-01
7.28408337e-01 -2.79100806e-01 -3.72298032e-01 2.27856442e-01
6.01099670e-01 -5.47388613e-01 -9.02033329e-01 5.98843932e-01
5.76144874e-01 -1.13731432e+00 1.05161130e+00 2.27005377e-01
9.18899417e-01 -6.44348800e-01 -4.59231049e-01 -1.27694571e+00
2.99423873e-01 -4.51137245e-01 4.90613461e-01 1.13990283e+00
2.05762118e-01 -8.33270550e-01 4.43768412e-01 3.57490689e-01
-7.07570732e-01 -3.08124483e-01 -8.13868165e-01 -2.83949703e-01
1.36770263e-01 -8.97306874e-02 1.98624238e-01 6.56987727e-01
-5.86721539e-01 -1.66827589e-01 -8.79714012e-01 7.12907434e-01
1.12431979e+00 1.99649051e-01 8.64235222e-01 -7.81300724e-01
-1.99986219e-01 1.27524376e-01 -3.07575047e-01 -1.18118095e+00
-1.82754859e-01 -4.56434876e-01 7.17807889e-01 -1.94625878e+00
3.12406063e-01 -3.08588356e-01 -1.95968658e-01 5.05956352e-01
-2.78016418e-01 1.13004184e+00 -2.11377516e-01 5.35518646e-01
-1.41559258e-01 1.01663339e+00 1.48756528e+00 -1.37535661e-01
1.10652126e-01 -2.77065516e-01 -8.48236918e-01 3.90733093e-01
7.78913021e-01 -3.38117808e-01 -6.59414679e-02 -9.57248688e-01
-3.68677080e-04 1.73343286e-01 3.08959156e-01 -9.45476949e-01
2.70363539e-01 -1.57340512e-01 1.98222622e-01 -6.45043179e-02
2.81075686e-01 -6.62700653e-01 2.59644777e-01 5.25725663e-01
-3.97142768e-02 -2.52328068e-01 5.38346097e-02 8.07385325e-01
-5.62596917e-01 -4.16078091e-01 9.71716285e-01 -3.29026312e-01
-1.00481355e+00 2.28713781e-01 -2.22591519e-01 -1.41871482e-01
5.15546739e-01 -3.51011574e-01 -4.63570923e-01 -5.49037278e-01
-1.34948641e-01 1.25738576e-01 7.88628459e-01 2.13180229e-01
1.22758186e+00 -6.37195051e-01 -1.12552059e+00 1.69606194e-01
8.22309125e-03 2.81884044e-01 5.57066321e-01 4.56398726e-01
-1.45711327e+00 -6.28609657e-01 -3.35439205e-01 -1.88449174e-01
-9.94919956e-01 -1.29527211e-01 5.55404365e-01 3.89150053e-01
-9.47161198e-01 7.74897039e-01 2.28060931e-01 -6.75022721e-01
-3.86222564e-02 1.34836912e-01 -2.86668301e-01 -3.75676960e-01
7.42885590e-01 2.86735535e-01 -1.19361721e-01 -6.80354059e-01
8.83487836e-02 7.92278647e-01 2.52741307e-01 3.76362680e-03
1.87562668e+00 -6.84812427e-01 -4.45991695e-01 -5.87782683e-03
1.13406587e+00 5.80574460e-02 -1.73042130e+00 -3.81021458e-03
-6.53226793e-01 -7.21885741e-01 1.61487445e-01 -4.90057945e-01
-1.41692066e+00 7.48312652e-01 9.43250597e-01 -9.47818998e-03
1.33709908e+00 -4.10894185e-01 9.30316329e-01 3.76009613e-01
3.13470513e-01 -5.71681619e-01 1.46999046e-01 4.46022987e-01
8.95804703e-01 -1.65052688e+00 2.12596521e-01 -1.91565529e-01
-3.73479605e-01 1.13800645e+00 3.49236488e-01 -2.81052828e-01
3.85646790e-01 3.75229895e-01 9.15231109e-01 -1.03136236e-02
2.55457647e-02 -3.53615806e-02 -3.25130314e-01 5.88240504e-01
2.50683427e-01 -3.72203678e-01 -1.48738652e-01 9.63961408e-02
-5.61636724e-02 -1.17581315e-01 9.72008824e-01 1.05847657e+00
-6.28850818e-01 -9.45113182e-01 -4.62785870e-01 1.37479424e-01
-7.21853435e-01 -3.30740511e-01 3.50222737e-01 4.41588908e-01
-1.75083190e-01 1.17452371e+00 -1.96185619e-01 -3.01473290e-01
1.02665402e-01 -7.72031665e-01 1.41011596e-01 -3.30456883e-01
-3.34381573e-02 1.51683046e-02 -2.53621012e-01 2.71449815e-02
-1.01723623e+00 -3.97639602e-01 -1.20443726e+00 5.22374511e-02
-2.55215794e-01 3.29124510e-01 8.84019434e-01 8.82669866e-01
-7.62967840e-02 3.97292882e-01 9.41738427e-01 -1.51662421e+00
-2.85434514e-01 -1.19329798e+00 -7.13389397e-01 4.49287146e-01
4.96298224e-01 -4.01766241e-01 -8.98952186e-01 4.87212032e-01] | [10.690375328063965, -3.5187036991119385] |
286a52c7-9dc7-4d00-b375-2a6ec548a555 | integrated-planning-of-multi-energy-grids | 2301.08454 | null | https://arxiv.org/abs/2301.08454v1 | https://arxiv.org/pdf/2301.08454v1.pdf | Integrated Planning of Multi-energy Grids: Concepts and Challenges | In order to meet ever-stricter climate targets and achieve the eventual decarbonization of the energy supply of German industrial metropolises, the focus is on gradually phasing out nuclear power, then coal and gas combined with the increased use of renewable energy sources and employing hydrogen as a clean energy carrier. While complete electrification of the energy supply of households and the transportation sector may be the ultimate goal, a transitional phase is necessary as such massive as well as rapid expansion of the electrical distribution grid is infeasible. Additionally, German industries have expressed their plans to use hydrogen as their primary strategy in meeting carbon targets. This poses challenges to the existing electrical, gas, and heating distribution grids. It becomes necessary to integrate the planning and developing procedures for these grids to maximize efficiencies and guarantee security of supply during the transition. The aim of this paper is thus to highlight those challenges and present novel concepts for the integrated planning of the three grids as one multi-energy grid. | ['Detlef Schulz', 'Christian Becker', 'Arne Speerforck', 'Christian Töbermann', 'Davood Babazadeh', 'Natalia Sanina', 'Alex Povel', 'Johannes Heise', 'Daniela Vorwerk', 'Marwan Mostafa'] | 2023-01-20 | null | null | null | null | ['novel-concepts'] | ['reasoning'] | [-1.47865802e-01 2.91769207e-01 -6.30585849e-02 1.41476300e-02
-1.87213123e-01 -5.81002355e-01 8.94106627e-01 2.65684456e-01
-2.49544173e-01 1.39862633e+00 5.54083101e-02 -6.42378986e-01
-1.12936527e-01 -1.26992249e+00 2.02473193e-01 -1.14792001e+00
3.92190158e-01 6.21937156e-01 -2.71668464e-01 -3.18342894e-01
1.83760568e-01 9.72352803e-01 -1.18076658e+00 -6.04268014e-01
1.21622717e+00 6.16344452e-01 3.28737140e-01 1.36758298e-01
-2.50411779e-02 7.82101788e-03 -3.44356149e-01 2.98908144e-01
3.19066644e-01 -6.81077063e-01 -6.41000688e-01 -6.70991689e-02
-9.81464982e-01 -3.20900530e-01 1.81173190e-01 1.02061498e+00
2.55880207e-01 2.87864447e-01 6.17952943e-01 -1.18156147e+00
-8.11162144e-02 5.47539890e-01 -3.42298388e-01 -1.72384650e-01
-1.80611730e-01 2.90137857e-01 7.87206233e-01 -3.77818853e-01
1.18362486e-01 3.09311926e-01 -4.97801192e-02 1.01867706e-01
-1.15363526e+00 -4.95048910e-01 -2.22371787e-01 1.67332113e-01
-1.33242667e+00 -4.63203639e-01 4.85809535e-01 -6.51375204e-02
1.46901393e+00 5.75462461e-01 1.22888410e+00 1.69339553e-01
2.03380048e-01 -2.13884830e-01 1.31360650e+00 -5.95281780e-01
5.03509820e-01 1.86022550e-01 -4.10086125e-01 -2.60054529e-01
8.89956534e-01 2.08105370e-02 1.91849858e-01 2.20308468e-01
2.50973523e-01 -4.79637414e-01 -4.69597697e-01 -7.28987381e-02
-8.83739114e-01 5.11373818e-01 3.45655859e-01 1.17564404e+00
-5.73893428e-01 -8.40769242e-03 2.95742840e-01 -3.88340689e-02
1.73780307e-01 4.58397895e-01 -1.61497161e-01 -3.30782086e-01
-1.25275898e+00 1.90004945e-01 7.06741095e-01 6.71743751e-01
6.42005920e-01 5.41138589e-01 6.71414554e-01 2.30580702e-01
2.53497452e-01 8.71160030e-01 -1.05316425e-02 -7.97561586e-01
2.18727186e-01 4.63609606e-01 5.40304661e-01 -3.25164050e-01
-3.79464030e-01 -6.28362179e-01 -1.05783474e+00 5.17457902e-01
3.00955981e-01 -4.46844101e-01 -6.82914674e-01 1.15534961e+00
2.77675450e-01 -8.11445177e-01 5.47580384e-02 5.46552181e-01
-2.48168349e-01 1.15355754e+00 1.78754613e-01 -8.43415260e-01
9.76051450e-01 -2.87932694e-01 -1.06119764e+00 2.67585695e-01
3.92846376e-01 -4.33901578e-01 2.46937051e-01 2.74433464e-01
-1.48007154e+00 -2.13987127e-01 -1.04645157e+00 4.03309792e-01
-5.92628121e-01 -2.59297371e-01 1.65994748e-01 7.31500447e-01
-8.77482831e-01 6.17063403e-01 -9.12968218e-01 -5.29675841e-01
5.45873903e-02 2.09748030e-01 -1.63874581e-01 1.60820752e-01
-1.23756862e+00 1.45958591e+00 8.51228416e-01 5.97931564e-01
-3.96746844e-01 -4.97265011e-01 -5.56115806e-01 5.12806654e-01
-7.15055019e-02 -2.75546640e-01 8.41164470e-01 -2.29461476e-01
-1.23537004e+00 2.74221957e-01 2.54429102e-01 -5.03330171e-01
5.42495310e-01 3.53473693e-01 -4.75055277e-01 6.79451972e-02
-1.30046859e-01 2.34083369e-01 -1.64434716e-01 -1.22405136e+00
-1.22186875e+00 -1.92165121e-01 -4.69031602e-01 3.27518612e-01
-3.96511644e-01 -1.53801680e-01 6.52349591e-01 1.09469131e-01
-1.58786699e-01 -9.38178003e-01 -4.47995245e-01 -9.09644067e-01
-3.92863363e-01 -3.83035213e-01 9.13594604e-01 -9.41702425e-01
1.00971639e+00 -1.80676043e+00 1.16851769e-01 7.21267939e-01
-2.55210876e-01 3.78984935e-03 7.17159152e-01 8.67207825e-01
-1.69224590e-01 7.97919407e-02 -3.52476120e-01 2.60720789e-01
3.76681149e-01 5.40532231e-01 2.31384467e-02 5.35339773e-01
5.09568043e-02 5.26633382e-01 -1.00588346e+00 9.52320173e-02
6.31926954e-01 3.11727285e-01 1.84815720e-01 1.57892331e-02
-3.11235767e-02 5.07151067e-01 -1.78724870e-01 6.56156957e-01
6.39627934e-01 2.10495397e-01 6.81105614e-01 -1.56330541e-01
-1.01493239e+00 2.81979501e-01 -1.02901936e+00 1.15374613e+00
-3.20591778e-01 1.04138002e-01 3.14182192e-01 -1.41508329e+00
8.96833062e-01 6.09661341e-01 1.12840819e+00 -1.17412996e+00
-2.53786564e-01 9.15885031e-01 1.30122721e-01 -2.98466980e-01
4.84439284e-01 -5.60878217e-01 1.09754749e-01 3.39060910e-02
-3.63117635e-01 -6.47923350e-01 4.95156169e-01 -1.05863050e-01
5.18517077e-01 4.07171607e-01 3.26508492e-01 -1.01454079e+00
5.85896969e-01 7.55850524e-02 6.46509409e-01 -5.31226814e-01
-5.73556032e-03 -2.66669244e-01 3.18515033e-01 -2.44570840e-02
-1.57397914e+00 -9.37103927e-01 -4.02030796e-01 -3.17887180e-02
-1.03038691e-01 1.57971174e-01 -4.24930066e-01 -2.58680612e-01
-2.40871742e-01 1.48527944e+00 -4.94449101e-02 2.27402076e-01
-6.65078938e-01 -9.41462398e-01 -1.94595560e-01 1.37318864e-01
7.38764763e-01 -5.98042727e-01 -1.16886580e+00 5.42956829e-01
-1.34270728e-01 -6.87376738e-01 2.06348911e-01 8.35320294e-01
-2.62216240e-01 -8.77880752e-01 -7.62016475e-01 -3.20997208e-01
6.42910361e-01 -4.02343459e-02 7.86918819e-01 -1.43118113e-01
2.59179566e-02 -9.26358923e-02 -2.09674612e-01 -4.56411898e-01
-5.33805072e-01 2.54863381e-01 1.14306435e-01 -7.54882514e-01
7.46900681e-03 -9.99377787e-01 -8.15727592e-01 5.83756790e-02
-6.64173365e-01 2.83381552e-01 1.46715701e-01 4.79736656e-01
2.73135424e-01 9.00205791e-01 1.25311279e+00 -2.82147050e-01
2.81982332e-01 -7.10123777e-01 -8.47390592e-01 6.37450665e-02
-1.18319547e+00 -3.29532415e-01 6.18336022e-01 4.70686913e-01
-1.33339906e+00 -4.45006974e-02 -4.88600552e-01 4.62436020e-01
-1.86724141e-01 4.16376770e-01 -6.24675035e-01 1.85978293e-01
-2.03682005e-01 1.17456764e-01 -2.18699455e-01 -3.68645579e-01
8.92960876e-02 5.12795389e-01 5.85038483e-01 -2.90354818e-01
1.29690909e+00 2.20582321e-01 4.13119763e-01 -9.04657364e-01
1.54826999e-01 -3.44384164e-01 -5.08981943e-01 -4.98320520e-01
1.15679765e+00 -6.95482969e-01 -5.36167681e-01 1.45317540e-01
-7.96530783e-01 -3.73719454e-01 -5.57722270e-01 4.64213520e-01
-2.90683180e-01 3.39024276e-01 6.41355440e-02 -1.11192358e+00
-3.42001319e-01 -9.87761319e-01 8.69442225e-02 5.16614556e-01
-2.63374478e-01 -9.72056866e-01 1.07983179e-01 -7.05005899e-02
7.95695424e-01 7.98238516e-01 1.06902862e+00 -1.98259309e-01
-5.62488854e-01 2.53909618e-01 2.37882789e-02 5.18469334e-01
7.76852787e-01 6.73018247e-02 -6.44555330e-01 -6.31528139e-01
7.24847987e-02 -2.84132129e-03 3.21706861e-01 -2.52241660e-02
5.56562722e-01 -2.11093679e-01 -2.99825490e-01 1.77926555e-01
2.17675471e+00 1.06657052e+00 8.54768753e-01 5.90497732e-01
2.99786143e-02 6.14330351e-01 5.82517266e-01 6.37040794e-01
3.92603964e-01 3.80086482e-01 7.26777852e-01 -3.23389798e-01
3.79745811e-01 2.14694932e-01 -1.08135715e-01 9.82666314e-01
-5.23112297e-01 -1.13424808e-01 -9.26917195e-01 8.62350523e-01
-1.45874560e+00 -1.05479288e+00 -3.54899228e-01 2.34640241e+00
7.02861726e-01 9.10456553e-02 9.84147415e-02 6.01424813e-01
3.23973387e-01 -2.22611994e-01 -1.75961509e-01 -5.38145721e-01
-3.35756689e-02 2.04116359e-01 7.04701006e-01 3.54616880e-01
-4.20218170e-01 -4.58164103e-02 5.81360102e+00 2.59941131e-01
-9.46456492e-01 -1.00948736e-01 7.71963954e-01 -2.43877172e-02
-7.95081735e-01 4.27504480e-01 -5.63889742e-01 7.14977026e-01
1.32325447e+00 -8.54898036e-01 4.92686033e-01 1.35573208e-01
8.30401301e-01 -6.02028310e-01 -4.94787633e-01 3.02158624e-01
-7.98303485e-01 -9.87172008e-01 -6.05087221e-01 6.34087324e-01
8.93946350e-01 6.24141246e-02 -6.28111601e-01 -5.70100583e-02
2.64771134e-01 -7.78977633e-01 6.47041023e-01 5.74089885e-01
7.57870078e-01 -1.49823976e+00 6.51241422e-01 3.74458671e-01
-1.42720628e+00 -1.72875643e-01 -9.71746892e-02 -2.04663221e-02
8.32138777e-01 8.25848460e-01 -4.53774571e-01 1.17232025e+00
8.11587453e-01 -5.44780493e-02 1.52242646e-01 7.76076555e-01
-1.32080615e-01 4.22162712e-01 -5.95944762e-01 1.67002782e-01
4.56250250e-01 -8.67381334e-01 1.07973985e-01 7.82066405e-01
6.93834543e-01 1.84730485e-01 9.17363763e-02 7.48922527e-01
3.99977297e-01 7.37332553e-02 -5.27281463e-01 -2.48879328e-01
5.99421382e-01 1.33485663e+00 -8.83112609e-01 -3.24411660e-01
-5.15814483e-01 4.45955336e-01 -3.08555812e-01 3.25727105e-01
-8.91281426e-01 -6.84937060e-01 4.96614635e-01 2.89741009e-01
-7.90743530e-02 -4.92676973e-01 -4.69219953e-01 -7.52417028e-01
-1.53896928e-01 -4.70619708e-01 -4.14359905e-02 -2.93827981e-01
-6.78566277e-01 2.63434738e-01 3.20633173e-01 -7.78259277e-01
-6.80575252e-01 -6.39761314e-02 -6.65769875e-01 1.54474199e+00
-2.01861525e+00 -1.07717514e+00 2.88081989e-02 1.72374412e-01
4.35392857e-01 1.25698254e-01 8.60436738e-01 3.91174674e-01
-4.30036783e-01 -3.47968459e-01 5.50834477e-01 -5.82869828e-01
7.79670780e-04 -1.35543692e+00 1.04145240e-02 1.08421242e+00
-5.74752033e-01 7.18989372e-02 9.05881643e-01 -7.67514646e-01
-1.36207664e+00 -7.63207972e-01 1.17456770e+00 6.29871368e-01
8.29784989e-01 -1.51672125e-01 -7.04749525e-01 3.37068141e-01
1.31147218e+00 -9.01940465e-01 5.14562488e-01 -3.82471830e-01
5.46525180e-01 -2.09696501e-01 -1.29551566e+00 5.03788590e-01
4.43327785e-01 -1.32845417e-01 -1.01557545e-01 4.19804633e-01
1.22766867e-01 1.77279204e-01 -1.25310194e+00 3.68868172e-01
4.26296651e-01 -8.67719173e-01 3.56468409e-01 1.13409229e-01
-8.22067186e-02 -3.75902653e-01 -1.37688026e-01 -1.71463978e+00
-4.69558328e-01 -6.71362340e-01 3.68445128e-01 1.54610872e+00
4.62041795e-01 -1.02233315e+00 6.53906405e-01 8.44350219e-01
-3.31970006e-01 -6.30670249e-01 -1.31739557e+00 -5.19435167e-01
3.23268235e-01 2.13756189e-01 7.96790659e-01 1.14758790e+00
6.90907300e-01 -4.40516733e-02 -1.23290680e-01 1.92767695e-01
9.01775301e-01 1.60466712e-02 2.44708553e-01 -1.00435388e+00
1.23082869e-01 -4.69683647e-01 1.53829038e-01 3.17112878e-02
-2.31625736e-01 -7.42646277e-01 -1.26013935e-01 -2.35256243e+00
-1.31819800e-01 -2.75021225e-01 -2.66486794e-01 5.47187030e-01
3.74968410e-01 1.83242504e-02 4.54810023e-01 -7.56033733e-02
5.17810524e-01 6.81383967e-01 1.01449060e+00 -9.08104107e-02
5.67140728e-02 3.44459489e-02 -4.71247375e-01 3.12670171e-01
1.33052337e+00 1.39883310e-01 -5.59193850e-01 1.32275820e-01
4.65479910e-01 1.87212527e-01 -2.15829939e-01 -1.16159868e+00
7.21704438e-02 -6.86922073e-01 3.37638736e-01 -9.63798583e-01
-2.29679123e-01 -1.64080143e+00 1.25776529e+00 1.03030646e+00
4.16427463e-01 1.52816713e-01 -9.88760442e-02 -1.57268047e-01
-6.44987896e-02 -1.13879308e-01 1.21137822e+00 -1.79496229e-01
-4.21608657e-01 -2.13907555e-01 -7.02434659e-01 -8.16824794e-01
1.57576835e+00 -3.32442015e-01 -4.02774781e-01 8.90206248e-02
-7.51311123e-01 7.93397903e-01 7.16569364e-01 -4.45710123e-02
2.04370469e-01 -1.13060880e+00 -6.73768342e-01 2.09735259e-02
-5.33524275e-01 -4.49475944e-02 2.33841494e-01 7.24034131e-01
-7.56749153e-01 6.07097387e-01 -6.50203884e-01 -5.14236689e-02
-9.10967469e-01 5.01466632e-01 6.99271858e-01 -4.05516326e-01
-6.76820219e-01 -6.93478137e-02 -3.72160316e-01 1.84158638e-01
-2.60890096e-01 -1.88384548e-01 -1.19564608e-01 1.83058545e-01
2.66802490e-01 7.06772327e-01 2.86480159e-01 -2.27752268e-01
-2.45864525e-01 -1.98285189e-02 6.28679335e-01 -5.76722398e-02
1.53896701e+00 -4.48736250e-01 -5.93975306e-01 2.70741791e-01
7.22439647e-01 3.69620323e-03 -9.96143699e-01 7.18649745e-01
5.44147640e-02 -2.49914616e-01 9.23574567e-02 -9.96645689e-01
-9.38814163e-01 3.46721679e-01 2.10541084e-01 8.57564330e-01
1.58711362e+00 -5.45965195e-01 5.49819529e-01 -1.80155635e-02
5.96287310e-01 -1.48713577e+00 -8.48379195e-01 3.73725481e-02
5.83218038e-01 -4.11720157e-01 2.30283946e-01 -9.23681334e-02
4.82812338e-02 1.14781153e+00 2.48381555e-01 2.46009097e-01
4.69481170e-01 6.71331704e-01 -6.06146097e-01 1.40601307e-01
-5.39456010e-01 -3.00437838e-01 -6.75543129e-01 4.78233457e-01
4.76839930e-01 3.37254733e-01 -1.07605577e+00 -2.04610363e-01
8.73322040e-02 -1.53145134e-01 4.30774838e-01 9.86769438e-01
-7.06613421e-01 -1.50501311e+00 -3.85455430e-01 4.69145358e-01
-4.92522806e-01 1.66277826e-01 4.13756609e-01 1.00041306e+00
5.26067793e-01 8.05574119e-01 1.37067335e-02 3.66159499e-01
5.30851424e-01 2.59897232e-01 3.12843651e-01 -3.00751999e-02
-3.80659580e-01 4.43071157e-01 5.27966678e-01 2.76415832e-02
-4.84730422e-01 -8.98369968e-01 -1.63465166e+00 -6.28291845e-01
-2.72453219e-01 1.00883520e+00 1.37199223e+00 9.23302770e-01
-3.67861062e-01 6.76101625e-01 1.10625565e+00 -8.00623894e-01
-8.03720653e-01 -8.06491613e-01 -1.22808242e+00 -1.57265157e-01
-1.23119727e-01 -1.43477887e-01 -3.55553120e-01 -1.68945938e-01] | [5.7118659019470215, 2.5361053943634033] |
94f26c87-5579-4e8c-a66e-bf33f071122a | intraday-trading-strategy-based-on-time | 2103.13507 | null | https://arxiv.org/abs/2103.13507v1 | https://arxiv.org/pdf/2103.13507v1.pdf | Intraday trading strategy based on time series and machine learning for Chinese stock market | This article comes up with an intraday trading strategy under T+1 using Markowitz optimization and Multilayer Perceptron (MLP) with published stock data obtained from the Shenzhen Stock Exchange and Shanghai Stock Exchange. The empirical results reveal the profitability of Markowitz portfolio optimization and validate the intraday stock price prediction using MLP. The findings further combine the Markowitz optimization, an MLP with the trading strategy, to clarify this strategy's feasibility. | ['J. Shen', 'Y. Zhou', 'Q. Wang'] | 2021-03-24 | null | null | null | null | ['stock-price-prediction'] | ['time-series'] | [-1.03834260e+00 -2.82915473e-01 -6.89776540e-01 -3.91418785e-01
-1.43536627e-01 -4.19725388e-01 4.18244779e-01 -9.07524526e-02
-4.36358094e-01 1.15927899e+00 3.53943080e-01 -6.47415102e-01
-3.36401194e-01 -9.99835193e-01 -4.91197109e-01 -4.06185687e-01
-5.78886867e-01 2.02246860e-01 -2.74553269e-01 -4.91387360e-02
8.78908098e-01 6.35166347e-01 -8.08295369e-01 1.35779589e-01
4.79222476e-01 2.00891995e+00 -6.12713620e-02 4.56052065e-01
-3.90855193e-01 1.37439930e+00 -7.01809883e-01 -5.05271912e-01
1.13635802e+00 3.24940354e-01 -2.06968874e-01 -1.28875300e-01
-2.79214323e-01 -9.50344026e-01 -5.41029572e-01 9.73938763e-01
6.15635775e-02 1.73592016e-01 4.77851182e-02 -1.02942264e+00
-1.23691678e+00 1.22754586e+00 -3.13526481e-01 7.06161261e-01
-4.46467668e-01 -4.50518616e-02 1.37641656e+00 -9.95912433e-01
-1.18987627e-01 7.10543752e-01 8.84153247e-01 -4.12492335e-01
-6.08962595e-01 -3.25895280e-01 4.72061746e-02 1.73057839e-02
-6.72266126e-01 2.00457931e-01 8.97833169e-01 -4.23294991e-01
1.12879074e+00 3.31685901e-01 1.12792563e+00 1.12801850e-01
1.03508019e+00 1.02632797e+00 1.30122459e+00 -2.80905694e-01
4.88128453e-01 3.06581616e-01 3.82346600e-01 -1.34727865e-01
6.07290983e-01 1.00019658e+00 1.00783095e-01 -2.80074775e-01
1.13394797e+00 6.00195169e-01 1.98016524e-01 4.33356643e-01
-9.80366409e-01 1.15924799e+00 5.22575855e-01 4.66502547e-01
-1.09557581e+00 -1.06994957e-01 2.54566401e-01 1.15526974e+00
6.71887457e-01 1.01721084e+00 -7.07021773e-01 -1.48281649e-01
-1.15734255e+00 4.02000338e-01 1.49386847e+00 6.21469021e-01
1.90546334e-01 9.50598836e-01 -1.32687509e-01 5.17889857e-01
5.25641024e-01 6.53849363e-01 8.70531619e-01 -8.80414486e-01
2.88459092e-01 1.03953846e-01 5.48972487e-01 -6.43879831e-01
-4.80673909e-01 -8.25961649e-01 -2.74713457e-01 6.40541136e-01
2.13817507e-01 -7.47967184e-01 -1.37619093e-01 5.21424592e-01
-5.30517459e-01 2.03763470e-01 5.23191810e-01 6.82020485e-01
1.33305237e-01 8.81779790e-01 -3.63495708e-01 -4.53161031e-01
1.04945278e+00 -1.53233945e+00 -8.94570947e-01 -1.20965384e-01
7.97454640e-02 -6.36666358e-01 -2.92923837e-03 6.39930904e-01
-1.31574237e+00 -5.41072309e-01 -1.38213575e+00 7.84360647e-01
-4.94283557e-01 6.26959419e-03 7.19237030e-01 4.13812250e-01
-7.94411242e-01 1.21353185e+00 -4.69361186e-01 7.20058084e-01
2.24583521e-02 3.04243624e-01 4.76312101e-01 1.16092432e+00
-1.47611713e+00 1.51199913e+00 1.01147377e+00 4.15587604e-01
-7.83261880e-02 -5.41343033e-01 -4.95008260e-01 1.09829426e-01
-6.51324093e-02 1.24650829e-01 1.52010834e+00 -8.78609240e-01
-2.05150723e+00 2.76887119e-01 6.53710186e-01 -1.41357458e+00
6.04444623e-01 -3.12051177e-01 -9.30439770e-01 -2.02752724e-01
-2.40791813e-01 -3.93478014e-02 2.29700640e-01 -3.56723398e-01
-1.02490747e+00 -2.43673012e-01 -2.89092392e-01 -5.20438291e-02
1.23124018e-01 -1.59279868e-01 5.74223757e-01 -1.47787845e+00
2.17196703e-01 -4.19488221e-01 -1.69249967e-01 -8.66237640e-01
-2.78323650e-01 1.10225432e-01 2.18559474e-01 -1.20817673e+00
1.38787127e+00 -1.70314300e+00 -1.14042592e+00 4.38909352e-01
-1.93435118e-01 -1.61450535e-01 5.07838964e-01 6.63895488e-01
-4.75938320e-01 -1.61969811e-02 2.74496585e-01 1.76975802e-01
4.46950525e-01 1.26137123e-01 -9.81454730e-01 3.84463191e-01
-9.00218934e-02 1.53981733e+00 -2.80210882e-01 1.10910900e-01
3.18892837e-01 -7.08158553e-01 1.05347335e-01 2.15863928e-01
-2.36679241e-01 -4.58990902e-01 -5.25618136e-01 1.17190158e+00
8.08449566e-01 -3.16298217e-01 -3.54624629e-01 3.54576081e-01
-7.89783835e-01 2.83921361e-01 -1.04850078e+00 4.25272763e-01
2.32062072e-01 6.49051428e-01 -3.42703998e-01 -8.17915201e-01
1.53828979e+00 3.47477078e-01 4.83945042e-01 -8.48236263e-01
6.63748607e-02 7.77455568e-01 -2.83260763e-01 -5.21988213e-01
6.99953914e-01 -7.82120168e-01 1.16651379e-01 9.27396417e-01
-3.97732377e-01 3.25995475e-01 -2.98279613e-01 -1.09572685e+00
4.89719719e-01 -2.47565657e-01 4.63146895e-01 -3.67169023e-01
4.04499918e-02 1.06732048e-01 4.47453022e-01 8.75996232e-01
-3.81897151e-01 1.56320989e-01 5.32396197e-01 -1.00735080e+00
-9.79329944e-01 -1.10379088e+00 -4.63811576e-01 8.42190862e-01
-4.35901314e-01 1.01393759e+00 -9.96666178e-02 -2.29354054e-01
9.54754710e-01 1.01330316e+00 -7.55754590e-01 3.59469235e-01
-2.41595030e-01 -8.49719405e-01 4.21761334e-01 7.68224001e-01
8.96462858e-01 -1.38407373e+00 -5.47169685e-01 9.08000946e-01
8.85227442e-01 -5.27952850e-01 -3.22379917e-01 4.07932401e-01
-1.24132371e+00 -6.01157367e-01 -1.18534255e+00 -6.81860745e-01
-3.52875628e-02 -3.17984313e-01 8.53159606e-01 -5.18842041e-01
6.41849101e-01 -1.85271829e-01 9.23868676e-04 -9.45729315e-01
-2.34100759e-01 -2.91605353e-01 -1.20107099e-01 8.81642569e-03
6.54323637e-01 -4.12720233e-01 -6.26995742e-01 2.30430096e-01
-4.97571766e-01 -5.28905749e-01 9.53657508e-01 7.04439044e-01
3.42983425e-01 1.61215946e-01 1.26506484e+00 -5.46631932e-01
9.19603229e-01 -7.59143770e-01 -1.52319133e+00 4.12443876e-01
-1.49294448e+00 9.56925526e-02 2.43376017e-01 -9.60232839e-02
-9.41597819e-01 -7.92818427e-01 1.17318682e-01 -4.09636080e-01
5.01561582e-01 1.00489426e+00 4.96407986e-01 -1.36702836e-01
4.59501008e-03 5.80388486e-01 5.98382890e-01 -7.81751931e-01
-2.83057243e-01 5.87885022e-01 5.41161239e-01 1.04825288e-01
3.40369046e-01 3.14699709e-01 -5.04532635e-01 -3.21654856e-01
-5.71037054e-01 -1.15110122e-01 -3.91848534e-01 -1.92731798e-01
4.88140076e-01 -8.75580013e-01 -9.78397131e-01 8.37328076e-01
-8.27198684e-01 -3.53097245e-02 -6.01493120e-01 1.28867972e+00
-7.10849583e-01 -1.11659884e-01 -1.48803353e+00 -1.29973233e+00
-5.45727730e-01 -6.38817370e-01 8.71470422e-02 4.33973581e-01
-4.84802984e-02 -1.59482181e+00 2.89983034e-01 2.13657513e-01
7.84470975e-01 4.16566044e-01 4.01366293e-01 -1.77757061e+00
-6.14782631e-01 -8.41781855e-01 -3.51394266e-02 8.74772131e-01
9.58299860e-02 -2.33859807e-01 -7.22678959e-01 -8.06635469e-02
9.57381546e-01 1.52129561e-01 5.89872301e-01 7.90479720e-01
2.40287781e-01 -8.89339626e-01 5.76120019e-01 4.35335547e-01
1.63408530e+00 1.09600699e+00 2.88976580e-01 1.36874390e+00
-5.77911921e-02 4.71604407e-01 5.30378699e-01 9.57680106e-01
2.48704270e-01 -4.29227024e-01 8.27542990e-02 4.06775147e-01
1.18823755e+00 -3.60148758e-01 3.26678842e-01 1.01180947e+00
1.43112391e-01 5.73105812e-01 -4.39335704e-01 1.09366722e-01
-1.88712275e+00 -1.18681073e+00 4.83698606e-01 1.95678473e+00
5.37857711e-01 8.00354540e-01 2.18749657e-01 9.53109190e-02
6.23964071e-01 5.74018121e-01 -7.73820400e-01 -5.67684352e-01
-5.61137080e-01 -8.75337720e-02 1.31185985e+00 4.04263347e-01
-1.20015490e+00 3.41108978e-01 7.98044491e+00 4.63319451e-01
-1.46237493e+00 -5.05660892e-01 1.05683982e+00 -4.93389033e-02
-3.84281665e-01 -8.43144730e-02 -8.21560085e-01 8.93558502e-01
1.14784503e+00 -7.55768478e-01 5.33533037e-01 1.18011808e+00
4.95470673e-01 7.94895738e-02 -7.83129871e-01 6.81687057e-01
-5.01114190e-01 -2.09688187e+00 -3.57250422e-01 2.05735266e-01
7.90543258e-01 2.28407443e-01 2.49296531e-01 4.58380282e-01
2.42380485e-01 -9.23546016e-01 1.09750581e+00 1.32545996e+00
-6.68562412e-01 -8.68810356e-01 1.35751808e+00 2.22231910e-01
-6.80034220e-01 -5.28222322e-01 -4.02404875e-01 -6.28487289e-01
8.79118145e-02 6.00370467e-01 -6.80812478e-01 2.07818627e-01
2.23209128e-01 6.09439611e-01 -1.45270810e-01 1.14928043e+00
1.04763776e-01 5.52921653e-01 -2.38493115e-01 -7.41685808e-01
8.47627163e-01 -7.24400818e-01 3.41826379e-01 6.02201343e-01
1.36429921e-01 4.69084755e-02 -1.00190558e-01 1.26098180e+00
2.25954860e-01 -2.30973512e-02 -4.36462402e-01 -2.92580873e-01
6.80847943e-01 3.30960482e-01 -2.92615354e-01 -4.13251787e-01
-7.88650870e-01 3.38106632e-01 -3.47873598e-01 4.72186446e-01
-4.22268510e-01 -4.44656551e-01 2.25076884e-01 -1.03948846e-01
4.74215865e-01 -1.05949804e-01 -1.52263772e+00 -8.10821712e-01
4.09736305e-01 -2.63835400e-01 3.74505460e-01 -5.86349308e-01
-1.74432135e+00 -4.58901050e-03 -1.32142872e-01 -1.42033827e+00
-4.38695431e-01 -9.50014055e-01 -1.21146202e+00 9.41157877e-01
-1.85755849e+00 -2.93557703e-01 9.20331657e-01 -2.17853393e-02
2.00214744e-01 -8.94415140e-01 2.42158890e-01 -3.60270590e-02
-4.36305374e-01 2.15745032e-01 1.16165483e+00 4.72088456e-01
-1.05275512e-01 -1.65193689e+00 8.71132255e-01 3.30297500e-01
-2.34316379e-01 3.01961303e-01 4.82047200e-01 -9.38613653e-01
-1.27051735e+00 -6.44309223e-01 8.79851282e-01 1.33473143e-01
1.48453605e+00 6.45498157e-01 -1.03583968e+00 9.41696584e-01
5.52067876e-01 -3.48076969e-01 7.06935346e-01 -3.99870932e-01
4.36496466e-01 -4.32345748e-01 -1.18726873e+00 2.02984720e-01
-3.05058628e-01 -2.30638906e-01 -1.45607603e+00 7.02247769e-03
6.86342835e-01 -7.67456144e-02 -1.49816895e+00 3.26211810e-01
9.53984559e-01 -1.12053204e+00 5.93856096e-01 -6.41922891e-01
-2.45825499e-02 4.37409282e-01 -2.10440993e-01 -1.16805863e+00
-3.70286226e-01 -8.42007458e-01 -2.12082952e-01 6.58830404e-01
8.81015122e-01 -1.64359391e+00 9.99629200e-01 1.28865910e+00
1.47451693e-02 -9.78485763e-01 -1.16489899e+00 -1.13472319e+00
2.39376366e-01 -2.68871337e-01 1.13405979e+00 1.05234098e+00
3.16297054e-01 -5.51013172e-01 -5.16949773e-01 -1.41767934e-01
5.19186556e-01 7.07963288e-01 6.51430413e-02 -1.09786284e+00
-5.30967176e-01 -1.11931491e+00 -4.97503243e-02 -8.09903562e-01
9.44493862e-04 -4.14245427e-01 -4.76247579e-01 -8.56952608e-01
-5.67136705e-01 -5.73070049e-02 -1.19676638e+00 9.31222364e-02
1.46143690e-01 -6.29468441e-01 2.87663996e-01 7.27454484e-01
6.19250610e-02 4.47924376e-01 1.06611717e+00 -5.21110520e-02
-3.20051521e-01 8.12956929e-01 -6.55946851e-01 8.82070065e-01
1.03538954e+00 1.16921067e-01 2.64023751e-01 -1.05527207e-01
2.17890039e-01 6.81260824e-01 9.53671639e-04 -6.24684215e-01
3.85522813e-01 -5.39525926e-01 8.20260346e-01 -1.11027896e+00
6.41376674e-02 -5.70731223e-01 2.19613865e-01 9.26485181e-01
-3.96629214e-01 6.74335182e-01 3.42021793e-01 2.83271998e-01
-8.23410034e-01 -4.97686297e-01 7.30473638e-01 -2.76202321e-01
-9.04170036e-01 1.09112367e-01 -2.74487734e-01 -1.24169476e-01
1.40400684e+00 -6.73017859e-01 -2.57589698e-01 -3.25991154e-01
-1.03487515e+00 6.50747955e-01 2.94658286e-03 2.29216412e-01
7.39801943e-01 -1.66021454e+00 -8.21952164e-01 5.50551713e-01
-5.42490721e-01 -8.51905465e-01 -2.42768392e-01 5.03308296e-01
-1.05182016e+00 9.60535705e-01 -5.47783971e-01 2.34622359e-01
9.02540237e-02 5.30873001e-01 6.87627316e-01 -4.90305454e-01
-5.98064423e-01 5.82439661e-01 -5.55285931e-01 1.33243024e-01
1.82661518e-01 -5.60937285e-01 -3.30531210e-01 5.49644291e-01
6.52761936e-01 7.99145162e-01 -7.34765977e-02 6.87464997e-02
2.20979556e-01 1.88518941e-01 1.49927884e-01 -5.01770079e-01
1.64102042e+00 1.20390534e-01 -2.66354084e-01 1.10072780e+00
1.21555960e+00 -2.66809136e-01 -1.31760144e+00 -3.14423591e-01
1.10782111e+00 -4.71726567e-01 -7.53779337e-02 -7.94005156e-01
-1.02293599e+00 8.40633661e-02 3.48314971e-01 8.55009913e-01
6.54329836e-01 -5.30704081e-01 1.13277304e+00 7.00526416e-01
1.76312804e-01 -1.71775389e+00 -6.33746326e-01 5.71640551e-01
1.14528430e+00 -1.31890500e+00 -1.97730027e-02 6.46549582e-01
-7.31692493e-01 1.55404544e+00 -1.21667691e-01 -8.88087988e-01
1.59057820e+00 3.90278071e-01 3.57255012e-01 1.30470440e-01
-9.10158396e-01 4.32353139e-01 1.85647205e-01 -2.65257865e-01
-1.03288759e-02 5.29934764e-01 -3.60144168e-01 1.13194990e+00
-5.72224259e-01 2.15567797e-01 2.62473792e-01 1.33203447e+00
-8.19395721e-01 -5.06220102e-01 -7.48567045e-01 9.36032414e-01
-1.02077615e+00 -3.09491605e-01 -1.91699341e-01 9.94725466e-01
-7.82186270e-01 5.34848690e-01 6.82016015e-01 -2.75777131e-01
3.35205674e-01 2.69606858e-01 -3.97249758e-01 -1.18171060e-02
-1.00514221e+00 3.25058758e-01 -3.60024422e-02 -6.22748993e-02
1.76684394e-01 -7.48881519e-01 -1.15173173e+00 -4.83422637e-01
-2.05800802e-01 4.03266788e-01 5.31695127e-01 8.82975876e-01
-9.05169249e-02 6.27802014e-02 1.44771469e+00 -4.42552269e-01
-1.74092710e+00 -1.12449431e+00 -1.64876509e+00 -3.80501390e-01
7.18650222e-01 -1.99145451e-01 -4.90927845e-01 -3.96024913e-01] | [4.489369869232178, 4.219675540924072] |
c4f0e196-4d64-428d-a1e3-235b3533be4a | on-the-stability-plasticity-dilemma-of-class | 2304.01663 | null | https://arxiv.org/abs/2304.01663v1 | https://arxiv.org/pdf/2304.01663v1.pdf | On the Stability-Plasticity Dilemma of Class-Incremental Learning | A primary goal of class-incremental learning is to strike a balance between stability and plasticity, where models should be both stable enough to retain knowledge learned from previously seen classes, and plastic enough to learn concepts from new classes. While previous works demonstrate strong performance on class-incremental benchmarks, it is not clear whether their success comes from the models being stable, plastic, or a mixture of both. This paper aims to shed light on how effectively recent class-incremental learning algorithms address the stability-plasticity trade-off. We establish analytical tools that measure the stability and plasticity of feature representations, and employ such tools to investigate models trained with various algorithms on large-scale class-incremental benchmarks. Surprisingly, we find that the majority of class-incremental learning algorithms heavily favor stability over plasticity, to the extent that the feature extractor of a model trained on the initial set of classes is no less effective than that of the final incremental model. Our observations not only inspire two simple algorithms that highlight the importance of feature representation analysis, but also suggest that class-incremental learning approaches, in general, should strive for better feature representation learning. | ['Bohyung Han', 'Dongwan Kim'] | 2023-04-04 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Kim_On_the_Stability-Plasticity_Dilemma_of_Class-Incremental_Learning_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Kim_On_the_Stability-Plasticity_Dilemma_of_Class-Incremental_Learning_CVPR_2023_paper.pdf | cvpr-2023-1 | ['class-incremental-learning'] | ['computer-vision'] | [ 2.72292495e-01 1.27684981e-01 -1.99229524e-01 -2.41873562e-01
-4.79375482e-01 -6.79271638e-01 8.57195258e-01 4.14337367e-01
-2.18345508e-01 5.03640473e-01 1.66928768e-01 -5.79042472e-02
-3.95077646e-01 -8.22012603e-01 -8.16694975e-01 -8.58044028e-01
-7.62398988e-02 3.88194054e-01 5.76955736e-01 -2.20135361e-01
4.74434137e-01 7.18559206e-01 -1.89419305e+00 4.00429279e-01
5.68082452e-01 9.24701452e-01 -1.92446709e-01 5.58717728e-01
1.60779729e-02 7.72432506e-01 -2.10666791e-01 -1.35342821e-01
1.72508940e-01 -5.03097892e-01 -7.87954748e-01 -1.42438814e-01
4.88152176e-01 1.08356997e-01 -1.87225819e-01 7.34928489e-01
2.19191134e-01 1.66774094e-01 9.56232309e-01 -1.12708271e+00
-7.74695218e-01 6.83723867e-01 -1.40424088e-01 5.97614706e-01
9.07652006e-02 1.80506378e-01 1.17309654e+00 -9.03398573e-01
6.39521718e-01 8.07076991e-01 9.40972805e-01 5.26536822e-01
-1.34139550e+00 -4.71625984e-01 4.51303899e-01 1.68721706e-01
-1.10109794e+00 -6.66773975e-01 6.74085319e-01 -6.51059568e-01
9.66138244e-01 2.33590320e-01 9.51202810e-01 8.24822843e-01
3.68234307e-01 8.30817580e-01 1.21807206e+00 -4.88744974e-01
2.88507044e-01 3.52652133e-01 4.46873605e-01 6.32056296e-01
5.49333572e-01 3.87011349e-01 -8.39505851e-01 -8.16047341e-02
4.43821758e-01 1.16839096e-01 -3.32146198e-01 -7.76794136e-01
-7.40436196e-01 6.73572004e-01 3.59830201e-01 5.04026353e-01
-2.25985557e-01 5.27566262e-02 3.03872943e-01 6.79557383e-01
1.80427641e-01 8.94937754e-01 -5.58096230e-01 -3.05435389e-01
-8.69830728e-01 2.24861428e-01 4.68013048e-01 4.78859305e-01
9.87671256e-01 2.58195251e-02 -2.97346199e-03 7.42185235e-01
1.38892293e-01 1.41719446e-01 8.89259994e-01 -7.82152355e-01
-1.62204728e-01 9.81581092e-01 -3.06570619e-01 -7.95823097e-01
-2.36455217e-01 -6.14193559e-01 -2.52081811e-01 4.06959713e-01
3.70705158e-01 1.48638844e-01 -7.49615371e-01 1.81495023e+00
2.42779143e-02 3.65663785e-03 -3.22601795e-02 7.32836798e-02
4.55777377e-01 5.49350917e-01 -9.34693962e-02 -2.60228604e-01
6.95057631e-01 -8.89832079e-01 1.04158781e-01 -2.77331829e-01
6.70505047e-01 -4.41611946e-01 1.06746638e+00 3.93808633e-01
-1.17462182e+00 -6.16405606e-01 -1.19660997e+00 3.08157772e-01
-4.85734850e-01 -4.69336301e-01 5.77285111e-01 7.21876621e-01
-9.04169977e-01 1.08959699e+00 -8.29308987e-01 -3.86538833e-01
4.27409738e-01 3.37672979e-01 -3.69563907e-01 1.11534014e-01
-6.88929915e-01 1.01442826e+00 3.83954018e-01 -1.94779813e-01
-7.52170086e-01 -7.07790673e-01 -3.89485210e-01 1.96320325e-01
1.36142820e-01 -4.21821952e-01 1.15396512e+00 -1.25861371e+00
-1.23667765e+00 7.43257642e-01 6.08600490e-02 -5.04219174e-01
3.02971572e-01 -9.07164738e-02 -1.42588019e-01 2.74479133e-03
-3.54427993e-01 5.14603198e-01 9.76418257e-01 -1.47650528e+00
-7.26958394e-01 -3.28591377e-01 -3.79201621e-02 1.63109049e-01
-6.11750126e-01 -3.57600302e-01 1.90518603e-01 -3.65746260e-01
2.74108529e-01 -1.13217151e+00 1.70952111e-01 1.50990799e-01
1.83991820e-01 -3.49504173e-01 7.52784252e-01 -3.14217992e-02
1.18741202e+00 -2.21228433e+00 -7.63375079e-03 1.43592209e-01
7.56118298e-02 4.05223757e-01 -1.91887170e-01 4.06781167e-01
-4.16038573e-01 1.26047179e-01 -1.08186364e-01 8.52893293e-02
-1.68191373e-01 2.48698130e-01 -6.70420349e-01 3.00300777e-01
3.38766962e-01 1.08023512e+00 -8.25734794e-01 -3.03782254e-01
3.62963639e-02 1.95591509e-01 -6.22841477e-01 -1.11970693e-01
-1.92837536e-01 -3.51160280e-02 -2.54075617e-01 6.27326071e-01
2.14498073e-01 -2.51882195e-01 1.30481586e-01 -1.73686501e-02
-2.51611054e-01 2.51238436e-01 -9.63199794e-01 1.12797427e+00
-1.74366921e-01 7.24515915e-01 -3.49427283e-01 -1.24746573e+00
8.94750357e-01 7.28803277e-02 5.67173541e-01 -8.56612563e-01
-1.39868520e-02 3.65789831e-01 2.99640089e-01 -2.43407696e-01
4.32102561e-01 -3.52199703e-01 1.74681932e-01 6.86816692e-01
1.68104365e-01 -1.52638301e-01 1.39608234e-01 -4.64588106e-02
1.11538613e+00 1.07797712e-01 2.56542653e-01 -3.07101488e-01
1.39138609e-01 -5.94617007e-03 4.68872756e-01 8.36910069e-01
-3.46544772e-01 5.80481589e-01 5.58561563e-01 -5.57244837e-01
-8.32655191e-01 -1.38189888e+00 -3.13258499e-01 1.52505195e+00
-1.86079532e-01 -3.34529519e-01 -2.02168807e-01 -8.41102123e-01
2.37063482e-01 4.88398314e-01 -9.70161200e-01 -8.09776962e-01
-4.92962033e-01 -8.41219127e-01 1.82185188e-01 7.54401565e-01
-1.80948377e-02 -1.00169027e+00 -1.00102198e+00 2.61241347e-01
5.43115199e-01 -3.49812388e-01 7.47040734e-02 7.05188751e-01
-1.36121953e+00 -1.18005860e+00 -4.38915998e-01 -8.32512558e-01
7.47768342e-01 2.99267203e-01 1.07150304e+00 3.69551510e-01
-1.87055394e-02 8.17334890e-01 -4.62688506e-01 -3.43762010e-01
-4.17024791e-01 2.44628727e-01 1.01885751e-01 -6.55755773e-02
2.14643881e-01 -9.05722857e-01 -3.69391441e-01 2.51296371e-01
-8.96191657e-01 -2.94011325e-01 7.76250482e-01 9.16454554e-01
4.94788885e-01 1.31078601e-01 8.92209351e-01 -9.81337130e-01
4.90300924e-01 -4.55352008e-01 2.94070095e-02 4.96121317e-01
-1.15889847e+00 3.80990744e-01 6.23433769e-01 -7.09778190e-01
-7.38408566e-01 1.35113969e-01 -4.58224081e-02 -1.53963670e-01
3.60480249e-01 6.07485950e-01 1.78716123e-01 -3.00471842e-01
1.04449368e+00 5.98301828e-01 1.18436232e-01 -3.01319063e-01
4.52501267e-01 2.12823674e-01 3.89765263e-01 -9.89338934e-01
8.58286023e-01 3.98320884e-01 -2.82141626e-01 -7.55787611e-01
-7.75481582e-01 -1.37176201e-01 -7.60941267e-01 -1.25032872e-01
1.47963867e-01 -5.28824508e-01 -1.01192102e-01 4.37523723e-01
-4.29196566e-01 -6.35522902e-01 -9.58625913e-01 9.50353034e-03
-5.56471884e-01 1.23596445e-01 -2.90077657e-01 -4.71716225e-01
-1.35445908e-01 -7.97890127e-01 4.67796743e-01 3.52344871e-01
-3.81088763e-01 -1.02490437e+00 3.50209832e-01 -1.62287235e-01
6.31834209e-01 4.07711007e-02 1.37312090e+00 -8.22572708e-01
-3.50731790e-01 -3.40817302e-01 8.07516128e-02 3.88329536e-01
3.26068968e-01 4.15816516e-01 -1.06736004e+00 -5.63007832e-01
-1.72291592e-01 -7.07781971e-01 1.27850914e+00 9.58327129e-02
9.32495236e-01 -2.23355651e-01 -3.84285331e-01 2.99967766e-01
1.28985465e+00 1.96231529e-01 6.11582696e-01 4.40918654e-01
2.37233967e-01 3.97467732e-01 1.26343235e-01 1.51216477e-01
1.26498073e-01 4.74103212e-01 8.72875825e-02 3.95715058e-01
-3.48995179e-01 -2.92170644e-01 4.02995855e-01 8.84947181e-01
-1.28967278e-02 3.04272294e-01 -9.74255383e-01 6.55062020e-01
-1.70465088e+00 -1.07774007e+00 5.62847137e-01 2.38996029e+00
1.02478790e+00 7.45664895e-01 2.33793899e-01 4.56274331e-01
1.71208456e-01 9.43804309e-02 -7.31637299e-01 -4.43593770e-01
-1.20179251e-01 3.66419464e-01 -1.18744723e-01 2.67126381e-01
-8.47296238e-01 6.74619019e-01 7.35359287e+00 5.64910173e-01
-1.49132633e+00 -1.85592756e-01 5.81982315e-01 -3.15528005e-01
-3.72662395e-01 1.00276202e-01 -8.22845042e-01 2.97258347e-01
1.07839060e+00 -5.53620458e-01 1.28155291e-01 1.07468545e+00
-3.90246063e-01 -1.04077786e-01 -1.46154845e+00 5.18749416e-01
2.86797043e-02 -1.33404195e+00 1.76721260e-01 2.09592190e-02
9.26891387e-01 7.55158812e-02 4.75047350e-01 6.36496305e-01
2.43122086e-01 -1.04857230e+00 8.09956431e-01 7.15025365e-01
2.62808263e-01 -5.73089123e-01 2.80651242e-01 2.66149998e-01
-1.22049236e+00 -4.60609525e-01 -3.69886488e-01 -1.89271033e-01
-3.63717854e-01 3.86241883e-01 -6.90045953e-01 3.45897973e-02
4.36545163e-01 7.78987527e-01 -1.22039020e+00 1.13823843e+00
4.48273830e-02 6.53678715e-01 -1.78347319e-01 -9.08148382e-03
5.92055824e-03 2.12593108e-01 2.13108242e-01 1.10192060e+00
1.23868160e-01 -2.86132604e-01 -4.37537022e-02 4.82021302e-01
4.71277498e-02 2.84192339e-02 -6.88314140e-01 -4.00461078e-01
5.46608329e-01 1.01356769e+00 -8.80360186e-01 -1.75363794e-01
-3.35222751e-01 4.67972010e-01 6.95210516e-01 -6.01818785e-02
-3.57439578e-01 -1.02351680e-01 4.37374294e-01 3.52927715e-01
4.83137786e-01 -3.38844597e-01 -4.12490159e-01 -9.87579167e-01
-8.38332176e-02 -8.33697021e-01 4.68850970e-01 -5.40074766e-01
-1.17904580e+00 4.52169806e-01 -1.59144208e-01 -1.00135720e+00
-2.74008751e-01 -7.80825257e-01 -7.65907705e-01 2.02361956e-01
-1.29721165e+00 -8.69544744e-01 -1.27325328e-02 2.66109794e-01
3.68227452e-01 -2.03705236e-01 8.25746894e-01 -1.00578412e-01
-4.74822581e-01 6.83902502e-01 4.36859071e-01 -2.44312957e-01
5.01092017e-01 -1.13261592e+00 1.50422573e-01 6.03134751e-01
5.94386339e-01 9.65594888e-01 5.21395743e-01 -3.05160284e-01
-1.43100166e+00 -7.52220809e-01 7.25963831e-01 -8.53855073e-01
5.64928114e-01 -1.47204131e-01 -1.25423753e+00 6.57145619e-01
-1.88091055e-01 1.12404160e-01 6.09481096e-01 5.62555552e-01
-7.75227427e-01 -3.93247753e-01 -8.21017742e-01 4.12008822e-01
9.46017623e-01 -5.28190494e-01 -9.98855531e-01 -1.06137894e-01
3.64257097e-01 3.92115675e-02 -6.86569154e-01 4.81373280e-01
9.93351877e-01 -1.14021909e+00 8.95318389e-01 -7.59923637e-01
4.31525320e-01 4.88333404e-02 -1.18751898e-01 -1.37986565e+00
-6.17824078e-01 -1.42111421e-01 -4.29081112e-01 1.06519270e+00
6.13195598e-01 -7.08382785e-01 7.77696252e-01 5.74325740e-01
-1.03854947e-01 -1.24244893e+00 -8.37043047e-01 -1.07714128e+00
6.78797245e-01 -2.77084172e-01 2.85160273e-01 8.90905440e-01
9.32942256e-02 4.21022922e-01 1.21669285e-01 -5.19170880e-01
9.77671817e-02 3.36757481e-01 5.62191546e-01 -1.62463033e+00
-3.38111103e-01 -8.51415694e-01 -3.50230873e-01 -7.08441675e-01
4.62937616e-02 -9.20047760e-01 -8.00309181e-02 -1.33651829e+00
4.17183280e-01 -6.94201887e-01 -7.97278702e-01 7.74302483e-01
-1.15873806e-01 1.38370916e-01 4.14068475e-02 5.22418976e-01
-5.91828883e-01 4.53749508e-01 8.88330042e-01 -3.27010453e-02
-4.36397493e-01 1.17509685e-01 -1.18084085e+00 5.92085123e-01
6.28745794e-01 -4.52447832e-01 -7.26100564e-01 -1.86532632e-01
6.15369618e-01 -6.73300743e-01 1.99698776e-01 -1.27629590e+00
2.90335774e-01 -3.03224772e-01 5.70920169e-01 -6.89582601e-02
2.18472302e-01 -4.55737144e-01 -1.46389484e-01 7.10639834e-01
-4.61834639e-01 2.79271752e-02 3.01486254e-01 6.62489653e-01
1.18114159e-01 -4.63245958e-01 1.10626090e+00 -2.82428622e-01
-6.90653741e-01 1.76997110e-01 -4.60808158e-01 2.40734175e-01
8.04381788e-01 -5.54359853e-01 -4.30848479e-01 -7.51618147e-02
-6.48739994e-01 -2.28086621e-01 7.53452539e-01 4.63365138e-01
4.24046308e-01 -1.16065180e+00 -5.12177050e-01 2.96281397e-01
3.01510096e-01 -3.28615278e-01 -1.02781497e-01 6.53070986e-01
-1.47709757e-01 2.69444972e-01 -4.32748228e-01 -4.37347680e-01
-1.02434576e+00 6.02090597e-01 4.59965914e-01 -1.44334018e-01
-5.41198611e-01 9.98789907e-01 -4.99132350e-02 -2.92826951e-01
1.73397422e-01 -2.20871851e-01 -1.64222848e-02 4.38387841e-01
3.74040127e-01 2.02980623e-01 2.54623085e-01 -4.37071741e-01
-3.23580563e-01 5.63080132e-01 -5.38292229e-01 -1.94863137e-02
1.55131364e+00 2.61376172e-01 1.66084930e-01 8.40617716e-01
9.98851359e-01 -4.56630439e-02 -1.43081284e+00 -3.45639139e-01
2.29905456e-01 -2.16793582e-01 -1.71312124e-01 -7.81619072e-01
-7.98452079e-01 9.45678711e-01 6.39046311e-01 3.99197757e-01
1.01002038e+00 3.27336155e-02 3.78365397e-01 6.38074160e-01
3.49895537e-01 -1.20827401e+00 6.84047282e-01 6.36968017e-01
9.02385950e-01 -8.90472531e-01 2.27350205e-01 2.23328114e-01
-4.07395363e-01 1.25366390e+00 6.65813088e-01 -3.51505488e-01
7.29276299e-01 4.99823242e-02 -3.09132218e-01 -7.00194463e-02
-1.14830029e+00 -1.22598954e-01 3.35379571e-01 5.15594006e-01
5.64843535e-01 -1.70567304e-01 -4.04071398e-02 4.71235693e-01
-3.00594836e-01 -2.05655266e-02 4.83018279e-01 1.34500730e+00
-1.03529561e+00 -1.20545602e+00 2.39893943e-02 7.34367251e-01
5.02859615e-02 8.92452970e-02 -7.64557600e-01 8.45906675e-01
1.15627460e-01 4.95919704e-01 6.39078692e-02 -5.45888364e-01
3.19208592e-01 6.56599820e-01 8.60796034e-01 -8.57479155e-01
-5.67187786e-01 -4.33632940e-01 -2.47552946e-01 -1.63706243e-01
-1.77093059e-01 -8.64680529e-01 -8.97837758e-01 -5.16300276e-02
-4.25730199e-01 2.37554908e-01 3.27992171e-01 9.88922656e-01
2.72117138e-01 2.26269960e-01 7.89126217e-01 -6.01363540e-01
-7.69223988e-01 -7.58966327e-01 -4.73088652e-01 3.43254119e-01
3.37954819e-01 -7.80312955e-01 -4.12660688e-01 -4.90866490e-02] | [9.006114959716797, 3.132052183151245] |
c86dfdc2-1d0a-41a7-bac1-d56fa15e2232 | sequence-tagging-with-contextual-and-non | 1906.01569 | null | https://arxiv.org/abs/1906.01569v1 | https://arxiv.org/pdf/1906.01569v1.pdf | Sequence Tagging with Contextual and Non-Contextual Subword Representations: A Multilingual Evaluation | Pretrained contextual and non-contextual subword embeddings have become available in over 250 languages, allowing massively multilingual NLP. However, while there is no dearth of pretrained embeddings, the distinct lack of systematic evaluations makes it difficult for practitioners to choose between them. In this work, we conduct an extensive evaluation comparing non-contextual subword embeddings, namely FastText and BPEmb, and a contextual representation method, namely BERT, on multilingual named entity recognition and part-of-speech tagging. We find that overall, a combination of BERT, BPEmb, and character representations works best across languages and tasks. A more detailed analysis reveals different strengths and weaknesses: Multilingual BERT performs well in medium- to high-resource languages, but is outperformed by non-contextual subword embeddings in a low-resource setting. | ['Michael Strube', 'Benjamin Heinzerling'] | 2019-06-04 | sequence-tagging-with-contextual-and-non-1 | https://aclanthology.org/P19-1027 | https://aclanthology.org/P19-1027.pdf | acl-2019-7 | ['multilingual-named-entity-recognition', 'multilingual-nlp'] | ['natural-language-processing', 'natural-language-processing'] | [-5.28539360e-01 -4.46176052e-01 -5.65503955e-01 -1.90125957e-01
-1.13244510e+00 -1.08130431e+00 7.87994921e-01 4.32042181e-01
-1.08007622e+00 8.24195206e-01 8.38229358e-01 -7.00223863e-01
9.28010866e-02 -2.86463648e-01 -2.65462458e-01 -3.72085810e-01
-1.41498987e-02 4.16347980e-01 -6.00065887e-02 -1.61618590e-01
-1.77278802e-01 4.41869467e-01 -8.85841787e-01 -1.40586793e-01
9.10693228e-01 4.28972840e-01 2.26342008e-01 2.90107697e-01
-5.42356908e-01 3.90648454e-01 -4.21054512e-01 -9.06014800e-01
-2.12369603e-03 1.07980281e-01 -7.42740810e-01 -4.19391066e-01
4.13262427e-01 2.06213206e-01 -1.92931011e-01 9.00582552e-01
8.97690058e-01 1.81793109e-01 5.29165030e-01 -6.45855427e-01
-1.45360565e+00 8.24397743e-01 -2.60269135e-01 5.48247576e-01
3.91809732e-01 -5.54019734e-02 1.69489491e+00 -1.14160383e+00
8.51218879e-01 1.18008506e+00 1.23498762e+00 3.61653507e-01
-1.25973380e+00 -5.21669924e-01 3.07923675e-01 -1.63735434e-01
-1.47019327e+00 -3.29328537e-01 1.74255654e-01 -3.58798116e-01
1.54385638e+00 -1.65683880e-01 4.42661315e-01 1.32663941e+00
2.69800752e-01 5.89447439e-01 1.21488833e+00 -6.37947679e-01
-8.28750208e-02 2.08245307e-01 1.99208841e-01 3.44382703e-01
6.94497466e-01 4.76076901e-02 -3.54822338e-01 -4.98952508e-01
4.62517798e-01 -1.77571159e-02 -5.97986765e-02 -1.09487707e-02
-1.39412951e+00 9.23876107e-01 1.45199239e-01 1.03736877e+00
-4.15637106e-01 -3.76170850e-03 5.24373710e-01 1.69368461e-01
7.87872374e-01 8.41364264e-01 -1.01200318e+00 -4.74605411e-01
-6.55050218e-01 -1.00622982e-01 7.19153047e-01 8.37418497e-01
7.09062457e-01 1.73264176e-01 1.44301616e-02 1.33312690e+00
1.90927073e-01 6.40105009e-01 7.36877024e-01 -3.22543263e-01
5.60320556e-01 2.43557826e-01 -1.70339122e-02 -7.85061479e-01
-4.67056721e-01 -1.74040705e-01 -3.17035139e-01 -3.72973531e-01
3.59510273e-01 -5.76255679e-01 -7.59625733e-01 1.82874155e+00
1.26813665e-01 2.61721257e-02 2.37313390e-01 6.98978901e-01
6.95201457e-01 6.95642591e-01 6.48225009e-01 1.50728032e-01
1.72880352e+00 -9.46019471e-01 -7.97403991e-01 -6.81946576e-01
7.93395817e-01 -1.12008893e+00 1.12443697e+00 -9.19756666e-02
-6.43116653e-01 -4.32358325e-01 -8.69112968e-01 -4.00503427e-01
-8.29936326e-01 1.72778308e-01 1.03594542e+00 8.54308724e-01
-1.01569283e+00 7.31064901e-02 -7.27023065e-01 -6.84575498e-01
3.50916050e-02 -8.63019843e-03 -9.32411432e-01 -5.96712232e-01
-1.33274007e+00 1.24915874e+00 5.59379518e-01 -2.43085429e-01
-5.16787767e-01 -8.18548560e-01 -1.26648474e+00 -1.34991497e-01
-2.81368922e-02 -1.99163243e-01 9.31626797e-01 -5.53057730e-01
-1.05269206e+00 9.43071127e-01 -1.93863753e-02 -1.92067623e-01
-5.86849786e-02 -5.12267292e-01 -9.42119062e-01 -2.09384784e-01
2.43156582e-01 5.78364611e-01 9.14854929e-03 -9.20226097e-01
-6.94145739e-01 -1.45006910e-01 1.59230262e-01 2.27608368e-01
-6.14774048e-01 6.09430611e-01 -4.82957661e-01 -8.19747686e-01
-2.85727352e-01 -8.55070949e-01 -4.13465261e-01 -5.42167544e-01
-2.43163090e-02 -5.39757729e-01 3.22607547e-01 -9.04092014e-01
1.40109837e+00 -2.33409905e+00 -2.50518888e-01 -3.90198767e-01
-2.39578739e-01 5.09414732e-01 -5.08709729e-01 7.86176801e-01
-1.91372246e-01 6.27512276e-01 1.26008401e-02 -4.44549203e-01
2.62019873e-01 5.02072394e-01 -9.04700980e-02 4.44781572e-01
6.13311052e-01 1.02639627e+00 -1.22966027e+00 -4.70218152e-01
1.75070629e-01 9.76243317e-01 -4.30001944e-01 -5.31368367e-02
2.14835122e-01 4.38124798e-02 -3.06620628e-01 8.75773549e-01
4.63722259e-01 1.21164136e-01 6.36394858e-01 1.90742984e-01
-4.63366598e-01 8.24694276e-01 -8.17519724e-01 1.88659406e+00
-9.01261091e-01 7.71124959e-01 -1.16317347e-01 -5.16788900e-01
7.95852721e-01 6.29067719e-01 2.31151775e-01 -7.40578711e-01
-3.83624844e-02 5.17317772e-01 3.06440704e-02 -3.09816003e-01
1.06217337e+00 -3.43725115e-01 -6.49996161e-01 3.70796323e-01
5.75891554e-01 2.95822561e-01 2.93478727e-01 3.57294902e-02
1.14162409e+00 4.11749110e-02 6.68663800e-01 -3.29036593e-01
9.89425182e-02 1.15051650e-01 8.14715981e-01 4.29921687e-01
-4.36945677e-01 5.44636607e-01 3.33182663e-01 -4.31712270e-01
-1.09055960e+00 -1.13346887e+00 -5.32659054e-01 1.59492576e+00
-3.25651675e-01 -5.37924767e-01 -1.61936715e-01 -9.18600321e-01
1.26563907e-01 7.05132306e-01 -4.18041170e-01 4.74487126e-01
-6.39464796e-01 -9.15542006e-01 8.00641060e-01 6.88508749e-01
-2.37976804e-01 -1.14877462e+00 8.64278674e-02 6.13419652e-01
-6.87282765e-03 -1.37642574e+00 -7.44565129e-01 4.72258866e-01
-5.16795576e-01 -7.62495458e-01 -7.77201951e-01 -1.19554162e+00
2.12905735e-01 1.11242503e-01 1.37176073e+00 -3.02309573e-01
-1.32586971e-01 4.26729083e-01 -7.23594069e-01 -2.53077179e-01
-1.54420063e-01 3.60345930e-01 2.28221878e-01 -4.24036533e-01
7.96646833e-01 -3.24252576e-01 -1.72890171e-01 -4.36467305e-02
-8.49909842e-01 -8.06945980e-01 9.26658928e-01 8.99332702e-01
6.45241559e-01 -4.70214039e-01 6.06538117e-01 -9.79460597e-01
7.42686570e-01 -7.36481905e-01 -2.20102280e-01 3.89352173e-01
-6.15264833e-01 1.61313564e-02 6.19739711e-01 -6.18110061e-01
-1.00373077e+00 -3.35603863e-01 -6.01323247e-01 1.11574076e-01
-2.20268071e-01 7.48378396e-01 -9.35838968e-02 1.83088496e-01
5.06979346e-01 -1.64158553e-01 -7.33459353e-01 -8.86238933e-01
7.58869290e-01 8.29474390e-01 2.16563001e-01 -7.43360102e-01
5.55810213e-01 1.44350873e-02 -8.79408121e-01 -9.47089672e-01
-8.02547157e-01 -7.64225841e-01 -6.12868547e-01 3.93584758e-01
1.18089342e+00 -1.17180002e+00 1.00321561e-01 -1.15671456e-01
-1.15551996e+00 -1.96983159e-01 -3.02516013e-01 8.95359516e-01
1.29823849e-01 2.69050092e-01 -8.87095690e-01 -5.31245649e-01
-1.19543567e-01 -1.05316520e+00 8.62523556e-01 -1.01680592e-01
-2.47432292e-01 -1.62433660e+00 5.83932042e-01 1.69964075e-01
4.58643734e-01 1.48447827e-01 1.00497580e+00 -1.28353643e+00
2.80807428e-02 -4.33395833e-01 -2.57167786e-01 3.57332081e-01
2.58840978e-01 -1.09901406e-01 -8.47522199e-01 -3.82255346e-01
-7.16621220e-01 -4.32495564e-01 5.77022731e-01 1.17101245e-01
3.30597758e-01 -6.55803978e-02 -2.36802936e-01 4.89063054e-01
1.67381096e+00 5.37615307e-02 3.99919659e-01 5.18920004e-01
7.02480614e-01 1.47523746e-01 3.46420258e-01 1.37801602e-01
7.76905596e-01 5.47277272e-01 -1.22443192e-01 2.31323000e-02
-3.13309252e-01 -3.93963814e-01 5.32572269e-01 1.60099852e+00
-5.32203801e-02 -2.42404476e-01 -1.29437327e+00 1.08129108e+00
-1.29353070e+00 -6.54527545e-01 4.39127386e-02 2.00935674e+00
1.02428842e+00 -2.79308762e-02 -2.46326506e-01 -5.09595454e-01
6.29151642e-01 5.52663863e-01 1.04220845e-01 -8.13388944e-01
-4.15654242e-01 7.36494005e-01 6.28374219e-01 3.54445755e-01
-1.22271657e+00 1.37095499e+00 6.79819012e+00 8.33996534e-01
-1.02024686e+00 7.88944423e-01 1.24694377e-01 4.34003204e-01
-5.53964376e-01 1.82438731e-01 -1.02920389e+00 2.22051740e-01
1.24158943e+00 -1.51003107e-01 1.09301619e-01 9.48810577e-01
-3.91386569e-01 2.86984146e-01 -9.25427198e-01 7.26920605e-01
1.22617751e-01 -1.08300614e+00 -4.69979458e-02 -7.03911483e-03
1.00370228e+00 7.31915355e-01 -8.56235996e-02 7.73284197e-01
7.87551463e-01 -9.73368883e-01 5.45828521e-01 -7.81803504e-02
1.02198863e+00 -7.91370511e-01 9.84658957e-01 -1.37867734e-01
-1.41976094e+00 8.90776291e-02 -5.69375277e-01 1.59954563e-01
3.80775064e-01 4.83766079e-01 -4.28671777e-01 7.38969386e-01
5.45028329e-01 6.11494660e-01 -4.67132598e-01 9.61756825e-01
-5.41130066e-01 8.29605281e-01 2.59025046e-03 3.39796324e-03
6.03571951e-01 4.64824177e-02 2.89687157e-01 2.01485538e+00
2.42564246e-01 -2.14539796e-01 3.76646340e-01 1.67248785e-01
-2.89750695e-01 5.83288193e-01 -6.01383746e-01 -6.43969953e-01
7.73648441e-01 1.41904342e+00 -3.84943366e-01 -1.50683492e-01
-1.07017922e+00 8.22316051e-01 8.43603671e-01 3.84383708e-01
-6.34638131e-01 -5.40993392e-01 1.23061335e+00 -3.14240783e-01
4.93990928e-01 -7.47156858e-01 3.49650010e-02 -1.28023601e+00
-2.44052827e-01 -8.54283452e-01 5.73963702e-01 -5.01261115e-01
-1.75459409e+00 8.36845160e-01 -2.94677436e-01 -8.82626474e-01
-2.37428434e-02 -9.01858389e-01 -3.00058901e-01 9.33836222e-01
-1.94330549e+00 -1.28697753e+00 3.86459619e-01 1.61891147e-01
3.83125842e-01 -1.68400809e-01 1.19281936e+00 7.58332193e-01
-6.39747918e-01 8.08462739e-01 2.91134149e-01 6.17651761e-01
1.01925552e+00 -1.35345244e+00 6.28183782e-01 7.43828952e-01
6.19540930e-01 9.04181421e-01 3.28289956e-01 -5.95623910e-01
-1.33252919e+00 -1.02573645e+00 1.72688341e+00 -6.12244904e-01
1.21088195e+00 -4.16390181e-01 -8.04063201e-01 9.71864939e-01
6.52361035e-01 2.52562195e-01 1.13322639e+00 8.23859155e-01
-8.68430078e-01 1.61453635e-01 -8.67848456e-01 5.62000275e-01
9.59103346e-01 -9.10159111e-01 -1.06122327e+00 2.55039096e-01
9.88981903e-01 -6.51946384e-03 -1.36201072e+00 4.39691283e-02
4.30080026e-01 -4.30468857e-01 9.52506602e-01 -7.87215054e-01
1.89826429e-01 9.66318399e-02 -4.49889332e-01 -1.60356879e+00
-5.81337571e-01 -3.21705580e-01 5.42773426e-01 1.62811363e+00
8.19385707e-01 -9.91457403e-01 1.80122018e-01 2.27826864e-01
-4.50029373e-01 -5.62348425e-01 -1.06185079e+00 -1.06493211e+00
6.69605136e-01 -6.44865632e-01 5.43187797e-01 1.63295960e+00
-2.54625827e-02 4.01613146e-01 -1.52916148e-01 2.14585379e-01
1.26709998e-01 -2.45240375e-01 3.04381847e-01 -1.01684797e+00
-3.78702730e-02 -3.83515954e-01 -5.46854496e-01 -8.68908107e-01
5.88570952e-01 -1.24966002e+00 3.84073257e-02 -1.84702551e+00
1.22692980e-01 -8.66129041e-01 -7.37275779e-01 6.72620893e-01
-4.32574272e-01 3.69189560e-01 5.34687489e-02 1.81819517e-02
-5.40776908e-01 3.91680807e-01 7.07294106e-01 2.03196872e-02
-2.93280762e-02 -7.72742927e-01 -6.96517706e-01 5.22725105e-01
6.64191484e-01 -7.02088952e-01 1.69523507e-01 -1.00577867e+00
3.45840275e-01 -3.15674901e-01 -2.50126809e-01 -8.23735833e-01
-2.27780342e-01 -8.62012133e-02 3.33163798e-01 -2.63612032e-01
1.95548519e-01 -5.69252014e-01 -2.19190523e-01 -3.56547199e-02
-6.62744641e-02 6.36599839e-01 3.92111033e-01 3.70179355e-01
-4.50328797e-01 -2.57529855e-01 3.50740075e-01 -1.37216851e-01
-1.08796656e+00 3.54985297e-01 -4.48068172e-01 7.83240497e-01
6.59032285e-01 -8.87140259e-03 -3.16229254e-01 1.17147967e-01
-4.81551111e-01 5.66520430e-02 4.31399882e-01 8.11738133e-01
1.40085548e-01 -1.69639611e+00 -6.46943986e-01 -5.10137454e-02
5.36985874e-01 -6.26962185e-01 6.53408319e-02 7.21473932e-01
-4.86536920e-01 6.91746354e-01 -5.77584542e-02 -5.08104414e-02
-7.69668818e-01 4.84273225e-01 -4.36868146e-02 -5.21441638e-01
-3.65348190e-01 8.49439025e-01 -8.21451023e-02 -1.04206645e+00
-1.12931304e-01 -3.70972306e-01 -2.13115603e-01 4.57810909e-01
4.88823950e-01 1.38691708e-01 2.36726422e-02 -1.07915545e+00
-6.91767097e-01 4.14457619e-01 -8.01983401e-02 -2.94510543e-01
1.52428079e+00 6.30116537e-02 3.07407994e-02 6.48830652e-01
1.31013930e+00 6.60224557e-01 -8.02896500e-01 -3.59029531e-01
4.87353593e-01 -3.30892682e-01 -1.08769625e-01 -7.25213289e-01
-7.38540947e-01 8.56497407e-01 3.26077521e-01 7.81048089e-02
6.18174076e-01 1.57780960e-01 9.35151994e-01 2.70811349e-01
5.64341068e-01 -1.11674547e+00 -2.75346994e-01 9.95325625e-01
3.55678439e-01 -1.17961597e+00 -7.26785064e-02 1.09548956e-01
-8.87379348e-01 7.93396175e-01 3.23768228e-01 -6.48346543e-02
8.23929727e-01 3.31514001e-01 4.15438205e-01 1.96404271e-02
-6.87053382e-01 -5.34907103e-01 2.50744075e-01 6.36248708e-01
1.07102692e+00 4.07725245e-01 -5.97806096e-01 5.93900740e-01
-1.71858549e-01 -4.59713012e-01 1.95741460e-01 9.68163908e-01
-2.51668930e-01 -1.53214300e+00 -2.24581137e-01 1.16588250e-01
-1.06633341e+00 -7.20993817e-01 -1.41023725e-01 1.23881280e+00
4.10443395e-01 8.95669043e-01 2.02068686e-01 -1.29350707e-01
3.85266483e-01 4.27511692e-01 2.06675440e-01 -9.53075588e-01
-8.21560502e-01 -4.66856994e-02 6.01490200e-01 -2.90012509e-01
-3.84791136e-01 -7.90907860e-01 -1.03587604e+00 -2.42612362e-01
-3.97459954e-01 3.31830710e-01 7.83529818e-01 8.03357959e-01
3.31838697e-01 1.83678225e-01 1.93407610e-01 -5.65740645e-01
-3.05714577e-01 -1.15316010e+00 -5.80885172e-01 5.11410177e-01
8.67593065e-02 -5.66024184e-01 -1.95584193e-01 -3.73308957e-01] | [10.635122299194336, 9.737406730651855] |
35de5623-ec1d-48f4-83a1-52da0f6b931e | detecting-heads-using-feature-refine-net-and | 1803.09256 | null | http://arxiv.org/abs/1803.09256v4 | http://arxiv.org/pdf/1803.09256v4.pdf | Detecting Heads using Feature Refine Net and Cascaded Multi-Scale Architecture | This paper presents a method that can accurately detect heads especially
small heads under the indoor scene. To achieve this, we propose a novel method,
Feature Refine Net (FRN), and a cascaded multi-scale architecture. FRN exploits
the multi-scale hierarchical features created by deep convolutional neural
networks. The proposed channel weighting method enables FRN to make use of
features alternatively and effectively. To improve the performance of small
head detection, we propose a cascaded multi-scale architecture which has two
detectors. One called global detector is responsible for detecting large
objects and acquiring the global distribution information. The other called
local detector is designed for small objects detection and makes use of the
information provided by global detector. Due to the lack of head detection
datasets, we have collected and labeled a new large dataset named SCUT-HEAD
which includes 4405 images with 111251 heads annotated. Experiments show that
our method has achieved state-of-the-art performance on SCUT-HEAD. | ['Zirong Chen', 'Zikai Sun', 'Lele Xie', 'Lianwen Jin', 'Zirui Cai', 'Dezhi Peng'] | 2018-03-25 | null | null | null | null | ['head-detection'] | ['computer-vision'] | [-3.38710546e-01 6.25483468e-02 1.84044585e-01 -5.41173339e-01
-4.82576042e-01 1.55394375e-01 1.97259158e-01 9.34903473e-02
-5.91628909e-01 5.56494057e-01 4.14645791e-01 3.96763355e-01
2.07591549e-01 -9.94259000e-01 -6.04321659e-01 -6.07007802e-01
-2.58169800e-01 2.44212911e-01 8.74363005e-01 3.79800685e-02
-8.17386433e-02 6.46646440e-01 -1.71234465e+00 3.89985032e-02
5.20840168e-01 9.88931060e-01 5.57808936e-01 6.66305482e-01
1.81692801e-02 6.02619588e-01 -6.59671068e-01 -1.87540382e-01
5.33398911e-02 -1.83264628e-01 -5.08866608e-01 1.74360067e-01
3.95613343e-01 -6.81741118e-01 -5.04143953e-01 8.76074135e-01
1.08776677e+00 -9.32646096e-02 4.57587630e-01 -1.32348466e+00
-2.99066514e-01 7.28827715e-01 -8.05376768e-01 3.89893472e-01
2.55955786e-01 -9.17307809e-02 6.33418560e-01 -9.44101214e-01
2.68688023e-01 1.63861752e+00 6.99306965e-01 6.03971899e-01
-4.52611893e-01 -1.09501028e+00 2.61187479e-02 1.53875753e-01
-1.69694507e+00 -4.37305868e-01 3.75838757e-01 -5.72815910e-02
4.98816639e-01 -1.94815099e-02 6.97943330e-01 5.93799531e-01
1.90518275e-01 1.22581601e+00 9.66114104e-01 -2.53977656e-01
-2.95643136e-03 -1.57704115e-01 3.89339715e-01 1.01630270e+00
3.76006961e-01 -3.91258448e-02 -5.19270718e-01 4.95498739e-02
7.30837345e-01 3.49690050e-01 -1.19662359e-01 -2.51914442e-01
-7.08945990e-01 8.85663450e-01 1.01992679e+00 3.95476133e-01
-3.05847734e-01 5.77942096e-02 3.03085059e-01 -3.74125212e-01
1.04595706e-01 -1.36240289e-01 -3.06810468e-01 4.48111951e-01
-1.02149856e+00 1.22389868e-02 5.73356271e-01 1.10605717e+00
7.60671318e-01 -2.02735290e-01 -5.14939785e-01 8.07163835e-01
5.55671930e-01 7.82123744e-01 6.36441708e-01 -6.69208884e-01
1.64652318e-01 8.58745098e-01 -2.28507876e-01 -6.60935283e-01
-1.02577353e+00 -3.30448419e-01 -8.89069200e-01 -1.24175757e-01
1.55878901e-01 -1.89158469e-01 -1.23813307e+00 1.41641223e+00
6.01616740e-01 -1.34862289e-02 -2.26886213e-01 8.03631544e-01
1.52205896e+00 6.43474281e-01 1.57961011e-01 2.36326605e-02
1.50377488e+00 -1.03912926e+00 -7.19040513e-01 -7.11728632e-02
2.23548651e-01 -6.03660107e-01 4.31314021e-01 1.67824492e-01
-9.75618899e-01 -8.61091077e-01 -1.07154369e+00 -1.68065310e-01
-4.91349697e-01 4.13705856e-01 5.31215847e-01 6.08739197e-01
-1.07720995e+00 1.69429034e-01 -8.45582187e-01 -7.73496747e-01
6.63247764e-01 7.86812961e-01 -2.98591107e-01 -1.67334706e-01
-8.56238842e-01 6.22962058e-01 4.74314362e-01 1.71017811e-01
-1.14950573e+00 -8.12824932e-04 -1.02361965e+00 2.71179050e-01
1.84837073e-01 -5.14645457e-01 1.29798937e+00 -3.32289070e-01
-1.15871894e+00 6.48568511e-01 -2.77512372e-01 -2.81218976e-01
2.02998847e-01 -2.58334160e-01 -2.59374082e-01 7.70568177e-02
4.00278330e-01 1.05478275e+00 7.37364411e-01 -1.00779152e+00
-1.23114014e+00 -7.31529534e-01 -2.85281122e-01 8.49828273e-02
-5.98123550e-01 3.15590173e-01 -9.36580956e-01 -2.10888252e-01
5.55036254e-02 -8.20557237e-01 -8.69348720e-02 -1.11366637e-01
-7.56936193e-01 -5.20354390e-01 1.01715553e+00 -4.51673657e-01
1.38736713e+00 -2.15613937e+00 -2.02886164e-01 1.55179918e-01
5.15935600e-01 2.83519447e-01 9.73561183e-02 -3.03843692e-02
3.13614279e-01 -4.30562735e-01 5.12753911e-02 -5.70768893e-01
-1.56061783e-01 1.48744062e-01 4.24676597e-01 5.22849441e-01
-4.88968194e-02 7.85632133e-01 -6.46663547e-01 -1.01649916e+00
2.95956075e-01 6.11102462e-01 -3.34775746e-01 4.19584751e-01
5.94465852e-01 1.71611980e-01 -7.20418930e-01 1.00940597e+00
1.15121341e+00 -1.40075311e-01 -1.70428351e-01 -1.59528077e-01
-3.08583051e-01 -2.73728877e-01 -1.33002484e+00 1.47476196e+00
-2.64057189e-01 2.70535737e-01 2.96061546e-01 -3.65949661e-01
9.07329500e-01 2.83797458e-03 2.14736864e-01 -5.52595198e-01
5.22960067e-01 1.34021491e-01 -2.14401931e-02 -3.47585589e-01
5.13005674e-01 1.73873648e-01 -2.01278865e-01 -2.04168111e-01
3.69776905e-01 1.52433321e-01 2.11630508e-01 2.78705418e-01
1.34145892e+00 -2.52801746e-01 4.48531181e-01 -3.60055864e-01
6.42758727e-01 -5.72162032e-01 7.26484835e-01 6.59393132e-01
-4.64141041e-01 6.02126658e-01 -2.42115799e-02 -4.99069303e-01
-5.19409955e-01 -9.25919175e-01 -2.78944850e-01 1.62365687e+00
1.35699630e-01 -4.28026497e-01 -8.88348520e-01 -6.92600608e-01
3.49793106e-01 -1.32591762e-02 -8.46600175e-01 -1.92470253e-02
-5.40899873e-01 -8.55577409e-01 4.80076641e-01 1.07824886e+00
9.90739942e-01 -1.34136117e+00 -7.57744670e-01 2.10945711e-01
1.14621632e-01 -1.06440890e+00 -5.45385718e-01 5.13594210e-01
-4.10295069e-01 -1.04933727e+00 -1.03011096e+00 -8.83716464e-01
6.75194860e-01 3.19234341e-01 8.77610922e-01 2.47646987e-01
-4.40937161e-01 4.58859093e-03 -3.65402192e-01 -7.78415501e-01
1.70647860e-01 4.35254693e-01 1.57035738e-02 -3.51488888e-02
4.45045829e-01 -3.11952621e-01 -6.81813657e-01 3.28917921e-01
-6.04736567e-01 -2.97619164e-01 1.09756315e+00 5.59789002e-01
3.61581832e-01 -1.08957522e-01 5.14328301e-01 -8.07010531e-01
3.16141218e-01 -3.89504671e-01 -5.63670158e-01 8.51318054e-03
-2.02118382e-01 3.93217392e-02 4.23196763e-01 -1.61601737e-01
-1.00037193e+00 4.79233086e-01 -3.43273968e-01 -2.61063367e-01
-4.37708229e-01 -1.50554731e-01 -5.11108935e-01 -1.91576540e-01
3.38817477e-01 2.23032743e-01 -4.26569223e-01 -6.85421407e-01
1.53672174e-01 9.57472384e-01 8.53827894e-01 -2.27211237e-01
7.12601483e-01 3.76894474e-01 5.42693622e-02 -9.10870790e-01
-8.78973901e-01 -9.09886062e-01 -8.76293123e-01 -1.14582039e-01
1.15388548e+00 -1.17141509e+00 -8.11337471e-01 9.10456479e-01
-9.34919894e-01 -4.77492884e-02 4.64411825e-02 2.38361225e-01
8.06356221e-02 5.45689352e-02 -9.14658070e-01 -9.21472430e-01
-6.52680159e-01 -9.55011845e-01 1.61534786e+00 8.40912819e-01
3.44026417e-01 -3.73653531e-01 -2.64456838e-01 8.10870528e-03
5.09270549e-01 7.41704330e-02 1.63167402e-01 -5.66592813e-01
-5.12383342e-01 -3.68742824e-01 -5.39947212e-01 -1.46869887e-02
-4.63788919e-02 2.71809250e-02 -1.24144411e+00 -5.30947924e-01
-5.06154776e-01 -4.21855271e-01 1.12515879e+00 6.82467699e-01
1.26876044e+00 2.39527062e-01 -7.76748955e-01 6.09425008e-01
1.16532779e+00 6.48609027e-02 5.40868104e-01 3.70562851e-01
8.78445566e-01 3.01729351e-01 5.41235507e-01 8.06963146e-01
8.54939342e-01 4.90160316e-01 3.70608091e-01 -3.84815395e-01
-3.65854055e-01 -3.79439235e-01 1.97434828e-01 6.14368498e-01
7.72647485e-02 -1.26256227e-01 -6.57518625e-01 6.17234945e-01
-1.69966567e+00 -4.97640014e-01 -6.49906248e-02 1.84353447e+00
3.77325803e-01 3.82332087e-01 4.85352308e-01 1.45880982e-01
9.65115249e-01 9.88799427e-03 -3.91388208e-01 -2.91372966e-02
1.21919751e-01 2.29747236e-01 6.98113322e-01 -5.16968779e-02
-1.52657151e+00 9.15901124e-01 6.06557894e+00 8.59700501e-01
-1.02351749e+00 1.98702693e-01 3.16738635e-01 2.69956112e-01
5.86652219e-01 -6.04110360e-01 -1.57280874e+00 5.86033523e-01
7.50285625e-01 3.66133422e-01 -2.25009307e-01 1.14185619e+00
-1.10730000e-01 -3.58041257e-01 -8.34715962e-01 1.00630116e+00
1.25474706e-01 -6.58843040e-01 -2.80674785e-01 1.28960535e-01
3.50735277e-01 2.61455119e-01 -2.04879716e-01 6.14377558e-01
4.14772421e-01 -7.33511567e-01 6.55747056e-01 3.02036375e-01
6.41835332e-01 -9.04323220e-01 1.07500398e+00 4.20343488e-01
-1.90291584e+00 -4.06460226e-01 -6.18610919e-01 5.94353527e-02
9.56944469e-03 5.19955039e-01 -8.31010461e-01 2.59062618e-01
1.05732799e+00 5.04547477e-01 -1.05912590e+00 1.57366157e+00
-3.72398913e-01 3.13237458e-01 -6.88012183e-01 -1.48246050e-01
3.89240161e-02 6.12388134e-01 -5.45541160e-02 1.46544826e+00
3.49377781e-01 6.47791922e-02 5.17622769e-01 3.09202641e-01
-3.94274920e-01 6.78492412e-02 -4.48106945e-01 6.58430338e-01
7.37461567e-01 1.80131471e+00 -1.00227571e+00 -3.70199353e-01
-5.15638232e-01 8.97036016e-01 2.38138407e-01 -2.34299242e-01
-7.23778784e-01 -5.97364008e-01 6.04308359e-02 2.74918396e-02
6.37363076e-01 1.60444766e-01 2.08044752e-01 -8.18549335e-01
-3.99400473e-01 -2.09276646e-01 4.96896684e-01 -6.22025371e-01
-1.05415750e+00 6.34862244e-01 3.11414152e-02 -8.03425789e-01
1.76048949e-01 -4.98070151e-01 -7.69560158e-01 4.40071493e-01
-1.60433185e+00 -1.75502479e+00 -8.29840899e-01 8.81163180e-01
7.50919998e-01 8.50583091e-02 5.62078953e-01 5.89271724e-01
-7.97062099e-01 8.92417371e-01 -1.50452480e-01 4.13065135e-01
8.06302130e-01 -1.36495268e+00 2.27385283e-01 6.72684968e-01
-4.00572330e-01 3.68014187e-01 2.93989420e-01 -6.30674541e-01
-1.06524873e+00 -1.36195207e+00 6.09285772e-01 -2.55685985e-01
1.77800581e-02 -4.23029125e-01 -5.98650217e-01 5.70673883e-01
-2.94109434e-02 5.21403730e-01 4.11955476e-01 -6.66174814e-02
6.14015497e-02 -3.48816454e-01 -1.44873941e+00 -2.83798784e-01
1.05929339e+00 -6.96997419e-02 -4.64569241e-01 7.48335272e-02
5.89585066e-01 -5.39149642e-01 -5.53354383e-01 5.70787907e-01
6.30228579e-01 -8.45800102e-01 7.00085163e-01 1.29735902e-01
5.76712377e-02 -3.74317586e-01 -2.06164390e-01 -9.99663353e-01
-5.98995507e-01 -2.99031045e-02 -2.10650846e-01 1.46066976e+00
1.28515020e-01 -2.13053152e-01 8.08866739e-01 2.26852074e-01
-3.51275891e-01 -6.26481354e-01 -7.76723206e-01 -5.94318271e-01
-2.45386496e-01 1.72501966e-01 8.51434708e-01 2.70492196e-01
-4.16696757e-01 5.62627733e-01 -3.29352796e-01 2.84556717e-01
6.81021512e-01 8.80862325e-02 8.67817402e-01 -1.37121308e+00
1.46230981e-01 -2.95902253e-03 -6.77785993e-01 -1.21698272e+00
-2.19904855e-02 -4.25018996e-01 3.45318109e-01 -1.51360774e+00
8.40049386e-01 -3.16568986e-02 -4.57869560e-01 6.94814503e-01
-3.59815687e-01 4.79811072e-01 3.55084240e-01 3.68804112e-02
-1.25690567e+00 6.03754997e-01 1.21407020e+00 7.29508623e-02
-1.05203532e-01 2.34761149e-01 -6.21215045e-01 9.87824857e-01
3.64922881e-01 -3.51500541e-01 1.45895645e-01 -2.81054229e-01
-6.88103437e-01 -2.03705072e-01 5.97225688e-02 -1.58739114e+00
4.69787091e-01 4.00474966e-01 1.12910414e+00 -1.07759583e+00
4.20725107e-01 -7.57904828e-01 -4.56207901e-01 6.33307636e-01
1.21309452e-01 2.47725621e-02 1.30183890e-01 4.09980893e-01
-7.16063604e-02 9.71513912e-02 9.77141380e-01 -1.62308797e-01
-1.05279088e+00 4.04644370e-01 -2.22356141e-01 -4.44849193e-01
1.07209337e+00 4.02667858e-02 -2.31729031e-01 -1.07763812e-01
-6.74241781e-01 7.15720773e-01 2.41994336e-01 3.48036647e-01
8.47917318e-01 -1.32220244e+00 -5.14128566e-01 4.77408558e-01
1.27216920e-01 1.80463463e-01 -4.01211856e-03 5.56102574e-01
-4.14037079e-01 4.51738030e-01 -3.54320168e-01 -4.85619754e-01
-1.52610922e+00 4.35255766e-01 2.74054855e-01 -2.39800155e-01
-4.50289369e-01 1.20293987e+00 4.61930871e-01 -4.53389794e-01
4.33927327e-01 -4.96741951e-01 -7.37489045e-01 5.46399914e-02
6.95302486e-01 3.76880229e-01 -4.25339341e-02 -1.15119040e+00
-6.48904681e-01 6.27213478e-01 -1.48451060e-01 4.79551345e-01
1.27819717e+00 -2.61435628e-01 -8.88381377e-02 6.39506662e-03
1.13219833e+00 -3.80778462e-02 -1.05550981e+00 -8.25145990e-02
-2.28582650e-01 -3.64125311e-01 2.31719360e-01 -4.90737379e-01
-1.46162641e+00 8.83599818e-01 1.20831740e+00 -1.78752709e-02
1.29555130e+00 2.41344064e-01 1.21901417e+00 3.51601422e-01
7.39549100e-01 -9.64875937e-01 8.11293796e-02 5.71807683e-01
4.40748572e-01 -1.47203994e+00 -1.11922082e-02 -5.09536862e-01
-3.52986813e-01 9.24131691e-01 1.16324222e+00 -1.61968306e-01
7.61474311e-01 4.75529552e-01 -2.31519982e-01 -4.22987223e-01
-1.53064877e-01 -8.45167041e-01 6.88368380e-02 5.68476081e-01
2.70752937e-01 3.30634296e-01 -1.96210146e-01 8.51157010e-01
-1.93506569e-01 2.19846442e-01 1.99466228e-01 1.03880608e+00
-1.02160239e+00 -8.61332297e-01 -7.11925685e-01 6.69835269e-01
-6.23764992e-01 2.35850051e-01 -2.33424559e-01 7.35948861e-01
6.80431783e-01 8.89947295e-01 -1.77221484e-02 -5.84823430e-01
4.59586263e-01 -2.81638771e-01 3.07216674e-01 -8.03791523e-01
-4.35421288e-01 3.83737922e-01 -2.21494153e-01 -5.93528807e-01
-4.16341215e-01 -4.09475625e-01 -1.50128102e+00 -2.62928665e-01
-8.06637824e-01 1.52270630e-01 5.21937490e-01 9.42773998e-01
8.20759833e-02 7.29386628e-01 6.94412947e-01 -1.23399079e+00
-1.44531891e-01 -1.36461842e+00 -1.16166174e+00 2.22316891e-01
4.44323719e-01 -9.80848551e-01 1.02959558e-01 -2.00265378e-01] | [8.195125579833984, -0.5072635412216187] |
b2f6c148-8b2d-45f6-86dc-8044e06ab87d | covidia-covid-19-interdisciplinary-academic | 2304.07242 | null | https://arxiv.org/abs/2304.07242v1 | https://arxiv.org/pdf/2304.07242v1.pdf | Covidia: COVID-19 Interdisciplinary Academic Knowledge Graph | The pandemic of COVID-19 has inspired extensive works across different research fields. Existing literature and knowledge platforms on COVID-19 only focus on collecting papers on biology and medicine, neglecting the interdisciplinary efforts, which hurdles knowledge sharing and research collaborations between fields to address the problem. Studying interdisciplinary researches requires effective paper category classification and efficient cross-domain knowledge extraction and integration. In this work, we propose Covidia, COVID-19 interdisciplinary academic knowledge graph to bridge the gap between knowledge of COVID-19 on different domains. We design frameworks based on contrastive learning for disciplinary classification, and propose a new academic knowledge graph scheme for entity extraction, relation classification and ontology management in accordance with interdisciplinary researches. Based on Covidia, we also establish knowledge discovery benchmarks for finding COVID-19 research communities and predicting potential links. | ['Chenghu Zhou', 'Xinbing Wang', 'Weinan Zhang', 'Luoyi Fu', 'Jiaxin Ding', 'Cheng Deng'] | 2023-04-14 | null | null | null | null | ['relation-classification'] | ['natural-language-processing'] | [-8.04053128e-01 -1.69393122e-01 -6.59793794e-01 4.18827981e-01
1.62569314e-01 -7.74717808e-01 2.79759973e-01 7.24196076e-01
-5.57651408e-02 1.05461144e+00 8.79551321e-02 -5.98226726e-01
-9.92667913e-01 -1.42979646e+00 -4.64442492e-01 -1.13149114e-01
2.04501182e-01 5.22133827e-01 1.54174015e-01 4.25761417e-02
2.22680584e-01 6.55877888e-01 -1.08543789e+00 5.19408733e-02
1.24040818e+00 3.66124779e-01 1.84776440e-01 2.12504417e-01
-6.18715405e-01 3.71924311e-01 -8.23718369e-01 -5.76806962e-01
-2.00109378e-01 -1.05105348e-01 -1.05011475e+00 -5.54298162e-01
-2.24560592e-02 5.50771058e-01 -4.86116171e-01 9.15672898e-01
3.17355633e-01 -2.58864641e-01 7.99874842e-01 -1.62924874e+00
-1.19017804e+00 7.05378175e-01 -3.83603841e-01 5.35231054e-01
4.24252182e-01 -2.27640525e-01 9.14493680e-01 -5.21640182e-01
1.45515013e+00 1.04310739e+00 8.71689439e-01 3.36567461e-02
-6.97749317e-01 -1.04464114e+00 3.47810537e-02 7.25773096e-01
-1.71096623e+00 5.04867315e-01 4.18559253e-01 -9.23202336e-01
7.21568465e-01 1.76850289e-01 1.07524478e+00 1.07564330e+00
1.05905928e-01 -1.15864702e-01 8.71200562e-01 -2.70011514e-01
1.51090264e-01 4.05974150e-01 7.46647179e-01 6.71054602e-01
1.28688240e+00 -4.92900133e-01 -6.74348354e-01 -1.80037290e-01
6.75986826e-01 3.17814857e-01 -2.98593730e-01 9.34162587e-02
-1.09288967e+00 3.20384800e-01 2.11613536e-01 6.58481658e-01
-3.34195852e-01 -6.84119821e-01 3.36237758e-01 4.97937560e-01
2.11285695e-01 7.98601866e-01 -6.33439302e-01 2.59626091e-01
-6.08704865e-01 1.83009550e-01 1.03327620e+00 1.34574103e+00
5.91210067e-01 -4.90884006e-01 -2.49608353e-01 7.08044648e-01
3.89452428e-01 3.83933306e-01 1.63152143e-01 -5.59070289e-01
1.84480950e-01 1.42589617e+00 -1.69297889e-01 -1.15038478e+00
-5.48191786e-01 -6.61607981e-01 -6.95614278e-01 -6.05782986e-01
2.50024289e-01 -1.69408754e-01 -3.92637938e-01 1.06236565e+00
3.01941872e-01 8.89414489e-01 8.42358842e-02 5.44302344e-01
1.69079697e+00 9.39032659e-02 3.94498050e-01 -2.57052541e-01
1.74725413e+00 -6.29866123e-01 -1.02609754e+00 7.16553390e-01
6.49029195e-01 -8.30970705e-01 2.77677923e-01 2.37386987e-01
-6.94149256e-01 -2.55960584e-01 -6.68872178e-01 1.98894031e-02
-1.34371281e+00 -3.21227536e-02 6.58424854e-01 4.92812991e-01
-9.18524384e-01 1.81868330e-01 -2.31343761e-01 -5.80151439e-01
7.62851477e-01 -2.89935488e-02 -3.26525897e-01 -1.77058339e-01
-1.44022310e+00 1.11241937e+00 4.24662083e-01 -5.75072467e-01
-5.03985643e-01 -1.59263456e+00 -2.76356488e-01 -4.47990820e-02
4.06945676e-01 -1.12683856e+00 1.79722518e-01 4.05060172e-01
-6.75399721e-01 9.83210862e-01 -4.02428545e-02 -8.33960101e-02
1.03132531e-01 1.19533211e-01 -1.14875603e+00 1.99141353e-01
3.29978555e-01 -2.40053479e-02 -6.25551045e-02 -1.05938649e+00
-6.03840768e-01 -4.61173564e-01 -1.48954526e-01 -2.54025787e-01
-8.20684671e-01 4.40710261e-02 -5.94628692e-01 -7.01351583e-01
-3.00593048e-01 -5.58884919e-01 9.33393463e-02 -3.49451602e-01
-3.85736138e-01 -8.53908837e-01 1.07343018e+00 -5.22881627e-01
1.55662847e+00 -1.56680155e+00 1.09618261e-01 4.30041999e-01
9.21534061e-01 4.25208002e-01 4.38233353e-02 4.38762099e-01
8.74134153e-02 5.30923069e-01 5.77900648e-01 6.36000037e-01
-2.50693411e-01 2.30773419e-01 -1.51483566e-01 2.07681641e-01
-5.59982993e-02 1.02097738e+00 -1.16321683e+00 -1.03897285e+00
-6.86028525e-02 3.23525578e-01 -2.85063475e-01 8.66374969e-02
-1.92341909e-01 2.56638676e-01 -1.07688320e+00 1.09867871e+00
8.02501202e-01 -7.48595834e-01 6.07316375e-01 -2.63045728e-01
-3.21260124e-01 -1.09715918e-02 -8.32951188e-01 1.33525205e+00
-1.71905398e-01 4.62599427e-01 -3.83862257e-02 -1.28478253e+00
1.09175694e+00 5.04564524e-01 7.93428183e-01 -1.90484166e-01
1.44299775e-01 2.53311425e-01 -1.24976434e-01 -8.16989422e-01
1.73154354e-01 6.62007451e-01 3.54924619e-01 2.17283711e-01
2.65041828e-01 3.58577192e-01 4.92788374e-01 4.28448379e-01
1.15419590e+00 -2.56085005e-02 1.43070534e-01 -5.50007403e-01
5.76914072e-01 3.60995829e-01 6.33330464e-01 5.60500562e-01
4.78308871e-02 -2.64753461e-01 5.00031948e-01 -3.69727910e-01
-5.47946990e-01 -8.82521689e-01 -7.31472135e-01 6.61485851e-01
3.17363799e-01 -6.96895182e-01 -4.89689589e-01 -6.78744853e-01
4.88576680e-01 1.80035442e-01 -2.61942714e-01 1.03604130e-01
-1.85255408e-01 -8.52080941e-01 8.10998142e-01 1.17947787e-01
3.13775897e-01 -8.45777214e-01 5.53056411e-02 1.31827518e-01
-1.45784646e-01 -1.34329069e+00 2.06391484e-01 -2.17991605e-01
-6.40612781e-01 -1.71564138e+00 -5.54758549e-01 -1.09807217e+00
4.21080261e-01 3.17391366e-01 1.42750072e+00 3.23576599e-01
-6.90926671e-01 3.68028134e-01 -3.11131775e-01 -8.61275673e-01
-5.41772619e-02 4.86720741e-01 4.42602225e-02 -6.57423198e-01
9.88915920e-01 -6.69122815e-01 -2.90393531e-01 3.87019902e-01
-5.72645545e-01 -3.54283214e-01 3.55632484e-01 4.80297565e-01
3.79189163e-01 1.19936563e-01 1.23173475e+00 -6.34485602e-01
6.07583046e-01 -1.23421276e+00 -6.44900322e-01 7.95791090e-01
-1.21112633e+00 -3.40975851e-01 2.44281814e-01 -2.01414809e-01
-8.50768864e-01 -8.58203948e-01 3.96760792e-01 -4.76556629e-01
-2.54849017e-01 7.38523841e-01 -1.96308076e-01 -1.69398293e-01
4.35221881e-01 -1.76994592e-01 -4.19010282e-01 -7.35747933e-01
2.52568901e-01 9.41940546e-01 2.38136068e-01 -8.02991033e-01
5.74661613e-01 3.12761068e-02 3.31002951e-01 -8.39199364e-01
-7.91176617e-01 -6.85826004e-01 -6.34492218e-01 -2.85182238e-01
1.20037591e+00 -1.19617176e+00 -1.40257275e+00 -1.16987094e-01
-1.32512677e+00 5.72818458e-01 9.91684571e-02 7.49056518e-01
3.90223682e-01 2.20688134e-01 -5.65919995e-01 -2.65965700e-01
-2.49140680e-01 -7.93880224e-01 5.13320327e-01 5.94277680e-01
-2.89239913e-01 -1.41777980e+00 3.38651478e-01 7.47487485e-01
1.92158803e-01 1.74941152e-01 1.19657886e+00 -7.21778274e-01
-7.21294641e-01 2.21254021e-01 -6.07268453e-01 -3.27974081e-01
1.35225013e-01 2.60721240e-02 -5.26834965e-01 2.54796773e-01
-8.50106418e-01 1.43453211e-01 7.54440665e-01 1.09528981e-01
1.33169568e+00 -3.27006355e-02 -1.09012854e+00 5.37137270e-01
1.26110625e+00 3.02189291e-01 3.63017261e-01 6.99693859e-01
1.13188255e+00 6.09376669e-01 3.11193764e-01 2.78456062e-01
7.29885638e-01 4.37828332e-01 -7.96168074e-02 -1.32350397e-04
-1.63038626e-01 9.12228972e-02 -3.71101916e-01 1.17675471e+00
-8.02824020e-01 -2.95145333e-01 -1.38649070e+00 8.84061277e-01
-1.81996202e+00 -8.80360425e-01 -8.32126319e-01 1.55237234e+00
1.16564202e+00 -2.77447045e-01 -1.06801659e-01 -3.83659095e-01
6.97029412e-01 -5.26023805e-01 -1.82831451e-01 -7.53006339e-02
-4.55451578e-01 4.63479221e-01 3.70708674e-01 -2.57351976e-02
-7.17321455e-01 8.92276645e-01 6.32415295e+00 8.12953711e-01
-6.77050829e-01 2.35595286e-01 4.71145362e-02 1.89329877e-01
-4.82252002e-01 -1.09894052e-01 -1.11207461e+00 4.13324475e-01
7.21871674e-01 -7.24772811e-01 -1.01780947e-02 5.62336743e-01
8.35699216e-02 5.07567644e-01 -7.94428885e-01 8.53719890e-01
-2.83231437e-01 -2.15910149e+00 2.17723489e-01 3.03571731e-01
1.07820821e+00 -1.10671818e-01 -4.65875834e-01 2.69709796e-01
7.19296098e-01 -9.55318034e-01 -2.53181368e-01 1.04953527e+00
3.94555748e-01 -6.29155278e-01 7.89505422e-01 -7.69340098e-02
-1.25790215e+00 1.94641471e-01 -4.29833233e-01 1.71286061e-01
-3.43691021e-01 1.06115055e+00 -7.68799722e-01 1.48688412e+00
1.03641427e+00 1.28904438e+00 -4.78076398e-01 1.19340026e+00
-1.48416787e-01 7.30504453e-01 3.14994007e-01 -3.37288752e-02
-2.76083797e-01 -5.38525879e-01 4.98421997e-01 1.50535536e+00
4.85007882e-01 1.25488251e-01 3.76515359e-01 1.03477669e+00
-5.92549384e-01 3.46055657e-01 -7.06875443e-01 -4.19718027e-01
1.08484411e+00 1.27750039e+00 -6.86730683e-01 -4.86199290e-01
-5.57694197e-01 9.05957520e-02 3.17910463e-01 4.81623411e-01
-5.45168638e-01 -5.58280706e-01 9.35635805e-01 8.29886794e-02
1.94893405e-01 -3.91284078e-02 -5.16530454e-01 -1.15522349e+00
-4.06022817e-01 -5.48277259e-01 8.17306876e-01 -4.84933048e-01
-1.78512013e+00 1.24114685e-01 -5.49832359e-02 -1.08981836e+00
4.93913591e-01 -6.21973217e-01 -4.30441856e-01 9.07896042e-01
-1.66160572e+00 -1.26476777e+00 -4.17001396e-01 5.96841216e-01
-1.15898363e-01 -6.17217600e-01 9.70029294e-01 6.18420243e-01
-9.69504178e-01 3.37427527e-01 1.63850427e-01 1.77535459e-01
9.39464688e-01 -1.00820613e+00 -2.38057539e-01 2.05378652e-01
-3.15266490e-01 9.94047999e-01 1.40256032e-01 -1.14753306e+00
-1.50618625e+00 -1.43520284e+00 1.03984010e+00 -9.38265026e-01
1.15839970e+00 -1.35314395e-03 -1.28477478e+00 4.91679579e-01
5.21617234e-01 5.22646680e-03 1.22296357e+00 4.84810054e-01
-6.14485264e-01 -2.79343650e-02 -9.11728263e-01 4.55256701e-01
1.65135753e+00 -3.38790447e-01 -8.89966428e-01 6.36858404e-01
8.71794403e-01 6.13450781e-02 -1.95229125e+00 1.72063738e-01
4.27966028e-01 -1.34322010e-02 1.34695590e+00 -1.11817455e+00
4.58370954e-01 -3.54213625e-01 4.51282889e-01 -1.04482687e+00
-5.06843925e-01 -2.62684256e-01 -2.63292760e-01 1.39717090e+00
2.84128398e-01 -4.82033134e-01 5.40026844e-01 -3.88902165e-02
-5.55806756e-02 -5.71127951e-01 -5.19958317e-01 -9.50803220e-01
3.27538401e-01 4.38086130e-02 8.32365990e-01 1.97851741e+00
3.27580452e-01 4.60055172e-01 1.44850612e-01 3.86923820e-01
7.04864442e-01 5.63023090e-01 5.66817105e-01 -2.09032249e+00
1.67401627e-01 -7.68514335e-01 -5.44713616e-01 2.02462859e-02
3.17826450e-01 -1.29002559e+00 -9.90566015e-01 -2.05609512e+00
3.87541145e-01 -6.17484689e-01 -4.42386657e-01 6.45036995e-01
-2.06783161e-01 -7.37169087e-02 -3.34762722e-01 3.01386565e-01
-8.06563258e-01 -7.04706013e-02 1.44819820e+00 -2.10148916e-01
-1.75944477e-01 -7.21933544e-01 -6.91556752e-01 5.44890821e-01
4.76302922e-01 -4.30472881e-01 -2.67007530e-01 -4.27524373e-02
4.38114882e-01 -2.48954594e-01 5.13785183e-01 -7.45690703e-01
4.82589543e-01 -5.06213665e-01 3.20522785e-01 -5.67030966e-01
-5.03041148e-01 -6.32987440e-01 4.46523219e-01 3.81207436e-01
-5.27820252e-02 -8.89352709e-03 2.26027966e-01 5.91022313e-01
-1.19024374e-01 -1.21501707e-01 9.55889598e-02 -3.68866883e-02
-5.83470941e-01 3.40604067e-01 -4.43310410e-01 4.38953787e-01
1.21544027e+00 1.93820074e-01 -9.90691781e-01 5.10786533e-01
-7.36322701e-01 5.35665810e-01 2.32953519e-01 5.61066985e-01
3.19146663e-01 -1.02102768e+00 -9.45635557e-01 -2.18162552e-01
4.22707528e-01 -2.70976603e-01 2.43657842e-01 8.86399627e-01
-4.11558867e-01 7.05900550e-01 -4.55414742e-01 -3.46640497e-01
-1.35397208e+00 5.72835505e-01 -5.12012951e-02 -5.52557170e-01
-4.78336245e-01 5.02102196e-01 -1.66041687e-01 -4.71266508e-01
3.41516972e-01 1.69857919e-01 -1.02102149e+00 4.35591936e-01
3.50738227e-01 8.17786038e-01 2.53821701e-01 -1.71399504e-01
-7.04861283e-01 6.52982533e-01 -8.47988278e-02 6.44192219e-01
1.45208061e+00 1.93380594e-01 -6.24240994e-01 1.41908094e-01
7.87724614e-01 2.92994887e-01 1.09385908e-01 -2.25625291e-01
3.46130461e-01 -2.12686270e-01 3.23789045e-02 -1.06770539e+00
-9.27573562e-01 3.83336991e-01 2.48091117e-01 2.55379349e-01
7.58694351e-01 2.09810585e-01 5.66790104e-01 4.20984298e-01
2.31242701e-01 -1.08386254e+00 -2.52199858e-01 5.74342310e-01
5.89485943e-01 -8.29884648e-01 2.99743026e-01 -1.00155389e+00
6.87729269e-02 1.01210868e+00 6.86091483e-01 4.06587094e-01
1.20091212e+00 3.09702218e-01 -6.27902895e-02 -7.77576029e-01
-6.08074307e-01 -1.04026139e-01 5.52299738e-01 7.16162741e-01
5.45418382e-01 2.43777916e-01 -8.77245665e-01 9.46917415e-01
-6.99046478e-02 4.61516887e-01 3.49541098e-01 7.91789949e-01
-2.41125956e-01 -1.44265854e+00 -4.48172569e-01 6.23916805e-01
-5.18459022e-01 -1.17233410e-01 -7.38912046e-01 8.04246902e-01
8.56319189e-01 8.08784246e-01 -7.60350153e-02 -2.01102376e-01
4.03180689e-01 8.96223038e-02 2.19151974e-01 -4.95576054e-01
-7.55743623e-01 -2.36936942e-01 9.54295620e-02 9.94757712e-02
-6.79109335e-01 -3.22333097e-01 -1.38787806e+00 -5.29698253e-01
-1.69933423e-01 5.09594619e-01 4.26446408e-01 9.33637321e-01
1.06170762e+00 1.08198214e+00 1.44101724e-01 3.68616134e-01
4.34558392e-01 -6.99605644e-01 -5.12558281e-01 2.57591337e-01
-3.35893214e-01 -9.00018036e-01 2.07847163e-01 1.79236382e-01] | [9.326519012451172, 8.20254135131836] |
73119800-393e-4226-8677-5273ba855fff | duplex-diffusion-models-improve-speech-to | 2305.12628 | null | https://arxiv.org/abs/2305.12628v1 | https://arxiv.org/pdf/2305.12628v1.pdf | Duplex Diffusion Models Improve Speech-to-Speech Translation | Speech-to-speech translation is a typical sequence-to-sequence learning task that naturally has two directions. How to effectively leverage bidirectional supervision signals to produce high-fidelity audio for both directions? Existing approaches either train two separate models or a multitask-learned model with low efficiency and inferior performance. In this paper, we propose a duplex diffusion model that applies diffusion probabilistic models to both sides of a reversible duplex Conformer, so that either end can simultaneously input and output a distinct language's speech. Our model enables reversible speech translation by simply flipping the input and output ends. Experiments show that our model achieves the first success of reversible speech translation with significant improvements of ASR-BLEU scores compared with a list of state-of-the-art baselines. | ['Xianchao Wu'] | 2023-05-22 | null | null | null | null | ['speech-to-speech-translation'] | ['speech'] | [ 7.31025517e-01 2.93003976e-01 -3.94274831e-01 -1.82460204e-01
-1.72821498e+00 -7.72341907e-01 8.88523579e-01 -7.69492865e-01
-5.24321459e-02 8.08654130e-01 7.74720252e-01 -8.96536946e-01
6.06672585e-01 -3.94172460e-01 -1.03080821e+00 -7.85003185e-01
3.58593792e-01 6.95304513e-01 2.93697953e-01 -3.30747336e-01
1.73326023e-02 -9.19192731e-02 -8.97198141e-01 8.02853942e-01
7.68560171e-01 3.05317044e-01 3.89122099e-01 1.08579564e+00
-1.90939590e-01 7.76605546e-01 -6.87564433e-01 -3.33179086e-01
2.09739447e-01 -8.09030056e-01 -7.34749198e-01 -2.79195756e-01
1.75655067e-01 -4.06101406e-01 -6.78928614e-01 7.37831473e-01
9.66039360e-01 -2.42793873e-01 9.09865439e-01 -1.07480371e+00
-1.14828885e+00 9.61248696e-01 -4.58473533e-01 1.65231407e-01
4.38732475e-01 2.06988528e-01 1.05640483e+00 -1.09138381e+00
6.42695725e-01 1.26466131e+00 3.29578459e-01 8.25974762e-01
-1.41747963e+00 -9.22453403e-01 -2.54300594e-01 6.71570376e-03
-1.22652221e+00 -1.09812582e+00 3.96928012e-01 -4.56631988e-01
1.42338264e+00 9.61127952e-02 2.54040897e-01 1.72483683e+00
2.44996846e-01 8.25707793e-01 1.07466280e+00 -3.66768569e-01
-7.96411559e-03 -1.20555900e-01 -6.20799124e-01 4.42893147e-01
-5.32321692e-01 3.41729909e-01 -1.11669767e+00 -2.07769834e-02
6.25892818e-01 -3.21870536e-01 -1.55427381e-01 1.10283725e-01
-1.62809157e+00 4.77652818e-01 -1.04350239e-01 1.71308815e-01
-1.45957515e-01 2.93827772e-01 4.26125407e-01 7.73547530e-01
5.40292084e-01 1.76982600e-02 -3.02406609e-01 -5.87744653e-01
-1.12152970e+00 -1.69368744e-01 7.83112705e-01 1.34525192e+00
4.83775109e-01 1.79503113e-01 -4.34417933e-01 7.85086930e-01
3.39136750e-01 8.36100578e-01 4.18903619e-01 -8.10216308e-01
1.00947058e+00 -4.17055160e-01 8.48904029e-02 3.04971077e-02
1.55156225e-01 -4.37657312e-02 -6.44189596e-01 -2.56849557e-01
5.78859895e-02 -4.83082771e-01 -1.14615631e+00 1.67774642e+00
5.97905330e-02 3.95961791e-01 4.41330552e-01 7.02719331e-01
5.64699531e-01 1.30121505e+00 -2.00765520e-01 -2.74649590e-01
1.02358794e+00 -1.52039778e+00 -8.38290572e-01 -3.08733821e-01
5.49103856e-01 -1.34947300e+00 1.25917041e+00 1.48646250e-01
-1.16324651e+00 -3.84856641e-01 -1.15097082e+00 -2.95749903e-01
1.82042316e-01 2.23187953e-01 -9.04689282e-02 7.40927935e-01
-1.39409697e+00 3.27545136e-01 -8.78382564e-01 -1.38220295e-01
-6.02958575e-02 3.36412221e-01 -2.91857541e-01 -5.03281094e-02
-1.43434727e+00 9.04219627e-01 -1.27261654e-01 -1.76739424e-01
-1.51789355e+00 -5.86164415e-01 -4.62118536e-01 5.73160201e-02
4.21527587e-02 -9.64072824e-01 1.71759498e+00 -6.75429106e-01
-2.38412070e+00 6.06078684e-01 -4.58929062e-01 -3.90146255e-01
7.27373838e-01 -2.56201446e-01 -3.79323393e-01 3.28021646e-02
4.15814519e-02 7.35007048e-01 1.10539222e+00 -9.81318116e-01
-7.27559090e-01 1.49824977e-01 -4.58481967e-01 3.76953930e-01
-3.23456645e-01 2.79151261e-01 -4.79392081e-01 -7.02381074e-01
-1.88425988e-01 -1.21713746e+00 1.76429972e-01 -2.76328057e-01
-6.23454809e-01 -1.97635263e-01 8.87474954e-01 -7.75588572e-01
1.08321381e+00 -2.05743098e+00 5.52148402e-01 -2.13656425e-01
-1.73034385e-01 1.73046499e-01 -4.75473583e-01 9.48491991e-01
7.47141242e-02 2.37917706e-01 -1.90043330e-01 -7.28334785e-01
1.07474893e-01 5.93261793e-02 -8.85358214e-01 2.22873718e-01
1.87223211e-01 8.64501834e-01 -8.77816737e-01 -2.61443973e-01
-1.92473665e-01 7.05137074e-01 -2.89128095e-01 5.50047636e-01
-1.77134037e-01 5.43161988e-01 -8.83956850e-02 4.25131589e-01
2.23354802e-01 6.05007634e-04 4.32550043e-01 2.61968762e-01
-1.08180247e-01 1.10186791e+00 -5.42520583e-01 1.98955011e+00
-6.09429240e-01 8.38212490e-01 -9.54966061e-03 -6.41740501e-01
1.03234053e+00 1.00106764e+00 6.29010424e-02 -6.48300946e-01
-2.90570647e-01 5.44501901e-01 -4.38248478e-02 -5.01250267e-01
3.75863492e-01 -4.70363110e-01 -7.34740915e-03 7.76651621e-01
1.33570686e-01 -6.05540991e-01 -2.94043899e-01 2.52295524e-01
1.17822397e+00 3.79292250e-01 -2.61694103e-01 -6.31898493e-02
1.91285372e-01 -2.99393773e-01 5.07686973e-01 6.96169972e-01
-1.74978361e-01 8.17369401e-01 4.53780413e-01 1.66856289e-01
-1.63137543e+00 -1.43878865e+00 5.10100126e-01 1.43750238e+00
-2.04375565e-01 -3.93876374e-01 -8.12973142e-01 -5.08496583e-01
-3.69317472e-01 7.61159837e-01 -1.64718047e-01 -1.82073548e-01
-8.06720853e-01 -3.01643103e-01 1.21655250e+00 4.06072319e-01
1.38173461e-01 -7.90527940e-01 4.33252990e-01 3.17436785e-01
-7.12358713e-01 -1.14615774e+00 -1.26274359e+00 2.08464757e-01
-7.53775537e-01 -1.69982895e-01 -1.08257163e+00 -1.15746117e+00
8.22242349e-02 4.48327512e-01 9.24028933e-01 -3.98070306e-01
5.15798032e-01 -2.44223282e-01 -7.67319053e-02 -1.71016723e-01
-1.14919174e+00 6.19234979e-01 2.56195694e-01 -1.27213925e-01
5.50038889e-02 -7.73319304e-01 -4.43426520e-01 4.32853162e-01
-6.35202587e-01 1.08771257e-01 7.79667377e-01 7.99387217e-01
3.76449913e-01 -7.40275860e-01 7.33399034e-01 -4.68457550e-01
7.95925379e-01 -3.59698266e-01 -1.85402527e-01 3.76535743e-01
-5.79296827e-01 2.90949881e-01 6.34504497e-01 -4.67044413e-01
-1.11472917e+00 2.60075659e-01 -3.22536498e-01 -4.56880748e-01
2.76734352e-01 2.09050030e-01 -1.16268277e-01 4.65430379e-01
6.90102935e-01 7.68012583e-01 9.99101624e-02 -2.19600379e-01
7.85337210e-01 1.23373902e+00 6.28362477e-01 -7.10541964e-01
9.19464886e-01 1.75070807e-01 -5.04202843e-01 -7.83439934e-01
-2.70482212e-01 -2.93327987e-01 -4.04623002e-01 -6.89185932e-02
7.71513522e-01 -1.32881200e+00 -3.91908109e-01 4.74818379e-01
-1.43114614e+00 -7.87796676e-01 1.17504641e-01 3.98250431e-01
-8.85517299e-01 3.09348524e-01 -1.08312774e+00 -6.65502727e-01
-5.41322291e-01 -1.54349995e+00 1.51642799e+00 -2.46175081e-01
-1.77577674e-01 -6.45376563e-01 3.08759570e-01 4.63653028e-01
5.82248390e-01 -6.34120405e-01 8.40964615e-01 -4.59265351e-01
-8.52732062e-01 2.63082623e-01 7.17028901e-02 2.63241917e-01
1.59944445e-01 2.26700455e-01 -9.49250937e-01 -2.99943238e-01
-3.15216750e-01 -4.96945620e-01 1.02934754e+00 1.33276120e-01
3.50385100e-01 -4.04105484e-01 -4.20843512e-01 5.20430684e-01
6.77286685e-01 2.17786118e-01 8.64605188e-01 5.24656698e-02
4.85305101e-01 3.91717762e-01 1.40162304e-01 2.71277223e-02
4.30359662e-01 6.90673411e-01 -1.55409425e-01 -3.74733098e-02
-5.46896756e-01 -8.54168713e-01 1.27896798e+00 1.53468895e+00
2.62414604e-01 -6.96742058e-01 -7.25473881e-01 5.64477205e-01
-1.55551672e+00 -9.98639226e-01 6.21751919e-02 1.99223304e+00
1.13887370e+00 9.40082967e-02 2.67539859e-01 -5.85627817e-02
8.93564403e-01 4.17924196e-01 -2.93745726e-01 -4.57690299e-01
-4.47056517e-02 3.89956869e-02 4.89109755e-01 7.84560204e-01
-6.49257958e-01 1.32084870e+00 7.64759445e+00 1.00312603e+00
-1.30127788e+00 5.38141131e-01 4.48876977e-01 -3.20998013e-01
-7.37222672e-01 1.64567098e-01 -8.15858126e-01 4.39924777e-01
1.43825591e+00 -1.15877114e-01 7.07214773e-01 3.03773791e-01
4.52283919e-01 7.16317534e-01 -1.10509372e+00 6.74966216e-01
-1.10579476e-01 -1.44895756e+00 8.36043209e-02 1.20448239e-01
8.74947309e-01 4.54236358e-01 4.38055247e-01 3.34281385e-01
7.02069759e-01 -9.76788342e-01 1.04424846e+00 2.85955787e-01
1.36959136e+00 -3.93226624e-01 1.18490987e-01 4.79475081e-01
-9.88694489e-01 1.47028923e-01 5.01748873e-03 1.02762140e-01
6.53774083e-01 3.04730803e-01 -1.30441308e+00 3.31900150e-01
2.58126140e-01 4.97096270e-01 4.28779811e-01 4.49072808e-01
-6.87214434e-01 1.19301593e+00 -2.48021170e-01 -1.34721175e-01
2.72566348e-01 -2.12144747e-01 6.00101352e-01 1.45493841e+00
8.41868997e-01 -2.38142267e-01 1.04196958e-01 6.17862701e-01
-6.21315658e-01 -1.25766724e-01 -8.29994619e-01 -2.59409934e-01
8.10059369e-01 6.64546430e-01 -1.50998428e-01 -5.31569481e-01
-4.52620596e-01 1.55467165e+00 2.40069121e-01 6.55058265e-01
-8.08857679e-01 -3.37934911e-01 8.53445649e-01 -9.46795195e-02
4.53098863e-01 -7.13492811e-01 -2.06013575e-01 -1.08688533e+00
-1.33165225e-01 -9.47116554e-01 -3.43260095e-02 -7.79119909e-01
-1.13820577e+00 6.05141759e-01 -6.34601593e-01 -1.04578841e+00
-4.79634672e-01 4.67121787e-02 -6.09379888e-01 1.08699489e+00
-1.45380664e+00 -1.29609275e+00 4.59373027e-01 4.61641490e-01
1.00639856e+00 -3.43923032e-01 7.13787675e-01 4.63643789e-01
-3.34428906e-01 8.52344692e-01 5.81934154e-01 4.84353825e-02
1.31092179e+00 -1.20571876e+00 1.20147729e+00 8.95699441e-01
3.26070666e-01 5.54420710e-01 7.78972626e-01 -6.39387846e-01
-1.59439433e+00 -1.00891852e+00 1.36298776e+00 -6.10748470e-01
7.85524964e-01 -6.97841823e-01 -6.76224351e-01 9.42407966e-01
6.18393362e-01 -3.94908696e-01 6.11136079e-01 -1.06171891e-01
-6.31289721e-01 -1.19049221e-01 -5.73516667e-01 7.98227966e-01
1.29241240e+00 -1.33822727e+00 -6.00514472e-01 4.79638785e-01
1.37707162e+00 -5.20719469e-01 -7.63741374e-01 -1.12309642e-01
6.93763435e-01 -5.96751153e-01 9.35855150e-01 -5.29419303e-01
6.50017858e-01 -4.26010519e-01 -3.21628898e-01 -1.63757730e+00
-8.91475677e-02 -1.50103331e+00 -1.96600795e-01 1.19544375e+00
9.79097784e-01 -4.45028424e-01 5.37112653e-01 -1.14059716e-01
-6.43583179e-01 -2.78789192e-01 -1.29001594e+00 -9.96999919e-01
4.86921966e-01 -2.05952778e-01 6.17749035e-01 6.44424736e-01
3.42972398e-01 9.69731331e-01 -9.55623925e-01 2.49355346e-01
1.04483150e-01 -2.27264702e-01 7.91981399e-01 -3.87459815e-01
-4.91702586e-01 -4.82291460e-01 2.62536973e-01 -1.89162719e+00
1.84470415e-01 -1.18153369e+00 4.94439393e-01 -1.49380684e+00
1.77938744e-01 -1.78607747e-01 8.10601749e-03 3.06463480e-01
-1.29794613e-01 1.99110553e-01 1.42769486e-01 5.19856453e-01
-2.72053897e-01 9.17804301e-01 1.22651970e+00 -3.89412850e-01
-3.33976001e-01 -7.15929940e-02 -5.94918609e-01 -6.37841672e-02
6.07063830e-01 -7.40882456e-01 -5.05694747e-01 -9.04369891e-01
8.86852890e-02 5.33576608e-01 -3.41827869e-01 -5.83706677e-01
4.63307470e-01 -8.75534341e-02 -2.52001196e-01 -3.29265922e-01
4.05365705e-01 -1.93416163e-01 2.04925239e-01 2.69064218e-01
-7.34187543e-01 3.12405020e-01 -2.12457955e-01 7.08213389e-01
-1.48057073e-01 2.70737946e-01 4.99500394e-01 1.92943215e-01
3.43925133e-02 2.35464469e-01 -9.62028444e-01 -1.59928665e-01
7.00774431e-01 2.89041370e-01 -5.58158100e-01 -7.48641551e-01
-5.12557864e-01 1.15944836e-02 3.31842422e-01 8.03948164e-01
4.76215392e-01 -1.30073535e+00 -1.09694457e+00 4.99282107e-02
-5.74907437e-02 -4.79604155e-01 -1.06576435e-01 8.07102740e-01
-2.76773661e-01 5.84855318e-01 2.58425176e-01 -7.53486753e-01
-1.18448400e+00 4.58710104e-01 3.91001910e-01 -1.10853739e-01
-4.83671308e-01 9.50032413e-01 3.72940674e-02 -8.02102745e-01
2.28292763e-01 -1.35955468e-01 5.91888011e-01 -2.87691474e-01
4.78369743e-01 2.01872975e-01 1.64915726e-01 -5.37882030e-01
-8.81646872e-02 2.58069873e-01 -1.56201482e-01 -9.56597567e-01
9.73471701e-01 -3.19192886e-01 1.55858770e-01 5.07569015e-01
1.26587725e+00 1.11796275e-01 -1.47475910e+00 -3.35580051e-01
-1.67200908e-01 -2.56223291e-01 -1.45507500e-01 -8.06504726e-01
-6.39529645e-01 1.25157082e+00 3.71044844e-01 -8.63960385e-02
6.63236558e-01 4.45343815e-02 1.28206503e+00 3.85373116e-01
1.62122026e-01 -1.12016284e+00 3.36037129e-01 6.77544534e-01
8.53448153e-01 -8.66941035e-01 -5.65314889e-01 -2.40926653e-01
-9.28063095e-01 1.03672063e+00 1.11639895e-01 3.43339860e-01
2.78838575e-01 6.23766601e-01 4.61817712e-01 3.96302819e-01
-1.23642480e+00 4.72640097e-02 -2.73544919e-02 8.25301051e-01
7.76662350e-01 1.63034186e-01 -1.14545323e-01 1.68222368e-01
-2.83006936e-01 2.84422580e-02 4.20077831e-01 7.24240124e-01
-4.96982068e-01 -1.43564343e+00 -2.50039786e-01 -8.83592218e-02
-4.52516019e-01 -5.62367260e-01 -5.91693759e-01 6.28940240e-02
-5.44236839e-01 1.32296360e+00 -1.62381679e-01 -7.44911909e-01
1.97874293e-01 4.74156469e-01 2.79077411e-01 -5.03116310e-01
-5.16119421e-01 5.10294020e-01 3.32783937e-01 -1.64729789e-01
-1.35560900e-01 -6.45347714e-01 -1.28175342e+00 -5.03210485e-01
-3.03562880e-01 6.73725009e-02 8.00166011e-01 9.97398257e-01
7.89653003e-01 4.63423699e-01 8.77516627e-01 -7.87352860e-01
-1.00097477e+00 -1.15391362e+00 -2.06289932e-01 -1.68043509e-01
6.57225966e-01 -7.70564973e-02 -2.13283077e-01 2.34425828e-01] | [14.527359008789062, 7.1347270011901855] |
5f6e0dde-d66f-4e63-a4e7-7a64be9e9d2a | a-comparative-analysis-of-the-face | 2211.02952 | null | https://arxiv.org/abs/2211.02952v1 | https://arxiv.org/pdf/2211.02952v1.pdf | A Comparative Analysis of the Face Recognition Methods in Video Surveillance Scenarios | Facial recognition is fundamental for a wide variety of security systems operating in real-time applications. In video surveillance based face recognition, face images are typically captured over multiple frames in uncontrolled conditions; where head pose, illumination, shadowing, motion blur and focus change over the sequence. We can generalize that the three fundamental operations involved in the facial recognition tasks: face detection, face alignment and face recognition. This study presents comparative benchmark tables for the state-of-art face recognition methods by testing them with same backbone architecture in order to focus only on the face recognition solution instead of network architecture. For this purpose, we constructed a video surveillance dataset of face IDs that has high age variance, intra-class variance (face make-up, beard, etc.) with native surveillance facial imagery data for evaluation. On the other hand, this work discovers the best recognition methods for different conditions like non-masked faces, masked faces, and faces with glasses. | ['Bal Murat', 'Eker Onur'] | 2022-11-05 | null | null | null | null | ['face-detection', 'face-alignment'] | ['computer-vision', 'computer-vision'] | [ 4.67428625e-01 -2.77063191e-01 1.06375717e-01 -7.78035760e-01
3.26964468e-01 -4.67380494e-01 5.40910482e-01 -1.07021654e+00
-1.28833652e-01 4.01994377e-01 -1.32660761e-01 -2.88232546e-02
-1.15849353e-01 -2.46327981e-01 -4.70020801e-01 -1.00532150e+00
-4.51637059e-01 3.54476459e-02 -2.75097247e-02 -4.09504175e-02
2.56785959e-01 1.06865978e+00 -2.20068789e+00 4.21956897e-01
-2.57306881e-02 1.19938517e+00 -2.57421225e-01 7.63212979e-01
2.75722444e-01 6.65108263e-01 -7.60393918e-01 -5.44374466e-01
4.83133227e-01 -2.43924156e-01 -4.62285101e-01 6.91098571e-01
9.56707895e-01 -5.94500780e-01 -4.17501569e-01 1.17574215e+00
6.42672300e-01 -9.30900052e-02 5.29608846e-01 -1.69987965e+00
-4.53066051e-01 -1.60942137e-01 -8.76157165e-01 6.01611674e-01
6.00004256e-01 1.52280465e-01 -1.52103171e-01 -8.81557703e-01
4.28404450e-01 1.72478378e+00 5.36093533e-01 1.01110542e+00
-7.00754285e-01 -1.01625228e+00 2.26019219e-01 4.71733958e-01
-1.52403235e+00 -1.22254801e+00 5.52647412e-01 -3.83777112e-01
6.34361386e-01 4.90775555e-01 3.44430476e-01 1.23801565e+00
1.89041972e-01 5.35776801e-02 1.08165634e+00 -4.58503574e-01
-1.68067500e-01 5.18093891e-02 2.10828945e-01 8.81663203e-01
2.70200610e-01 1.83159366e-01 -6.11292183e-01 -3.00309181e-01
6.67820275e-01 3.49359930e-01 -5.60239792e-01 1.51260659e-01
-7.19202280e-01 3.33750516e-01 -3.54135454e-01 1.31469369e-01
-3.16332638e-01 -3.70553255e-01 1.74445987e-01 6.25528514e-01
8.76440108e-02 -4.78389770e-01 -3.83902401e-01 4.08822566e-01
-9.13194597e-01 -8.92476644e-03 8.92084539e-01 1.06808364e+00
5.07096052e-01 2.10078463e-01 -1.61657974e-01 6.03305876e-01
5.43246627e-01 7.85357535e-01 4.96420890e-01 -7.41185248e-01
-5.10803889e-03 3.77495080e-01 -1.19829863e-01 -1.22450793e+00
-3.95713598e-01 1.82510123e-01 -9.13280725e-01 4.70154166e-01
2.08323598e-01 -2.50318468e-01 -1.02257621e+00 1.52206910e+00
5.73098063e-01 7.36099303e-01 3.14718932e-02 7.65232086e-01
1.05981827e+00 2.58703291e-01 -1.78663164e-01 -7.88788140e-01
1.73213851e+00 -8.70826781e-01 -9.65736747e-01 4.52147424e-02
-1.72502175e-01 -1.00361764e+00 3.06710660e-01 4.66955155e-01
-7.11714268e-01 -6.34285271e-01 -7.86319494e-01 5.62816024e-01
-2.75261968e-01 2.03350633e-01 2.32986376e-01 1.41848171e+00
-1.39895737e+00 6.80700392e-02 -4.06858772e-01 -6.84621394e-01
4.16967005e-01 7.68754423e-01 -8.85302424e-01 -1.43402413e-01
-7.39224315e-01 6.93060815e-01 -1.29325241e-01 4.56548303e-01
-1.19785547e+00 -1.40579700e-01 -6.45000935e-01 -3.23337346e-01
3.76005650e-01 -2.82867759e-01 8.09549689e-01 -1.58234870e+00
-1.47930610e+00 1.30492103e+00 -5.56342661e-01 -1.27268061e-02
1.94842607e-01 -3.71176959e-03 -1.01930022e+00 -4.79406752e-02
-4.39477950e-01 2.39647195e-01 1.52590215e+00 -8.32470417e-01
-2.95018137e-01 -1.06710875e+00 -3.44199389e-01 1.87634733e-02
-2.73508191e-01 9.83867168e-01 -2.82368571e-01 -2.13207662e-01
-1.32385209e-01 -9.94122207e-01 4.46090609e-01 2.45197359e-02
-8.93398151e-02 5.51745743e-02 1.52824974e+00 -6.63671613e-01
9.17862952e-01 -2.25768590e+00 -1.66310673e-03 1.14346609e-01
-5.95945343e-02 5.23187578e-01 -1.67380497e-01 -1.36940479e-01
-3.64991575e-01 -6.07065484e-02 1.77868694e-01 -1.96298555e-01
-4.37030107e-01 1.14024065e-01 -3.73871177e-02 8.07194591e-01
1.36643052e-01 1.55071005e-01 -8.17129239e-02 -6.10057116e-01
-5.59353419e-02 6.75329030e-01 -2.90953577e-01 4.95005876e-01
4.01153684e-01 4.48589146e-01 -2.90854573e-01 1.24388325e+00
1.14605045e+00 2.99643934e-01 2.81824231e-01 -4.10964102e-01
5.32915480e-02 -7.23198533e-01 -1.44841266e+00 9.74251568e-01
3.13365132e-01 5.36763787e-01 7.78459489e-01 -8.41697395e-01
9.80312765e-01 6.94101572e-01 3.42488945e-01 -2.22790048e-01
5.19564629e-01 -1.30747646e-01 1.40314683e-01 -1.05839670e+00
6.27232343e-02 2.97687858e-01 6.87456191e-01 2.59664088e-01
2.15563044e-01 6.67003572e-01 1.26928864e-02 -5.11073112e-01
9.52918947e-01 -1.52605683e-01 2.11001918e-01 -4.59135115e-01
1.07732081e+00 -7.46545374e-01 5.23923934e-01 3.19089621e-01
-7.55647779e-01 4.46093708e-01 3.70347172e-01 -7.52313316e-01
-5.65855205e-01 -6.92600369e-01 -1.50837258e-01 1.14055729e+00
-2.12235212e-01 1.22235037e-01 -1.24280834e+00 -6.64450705e-01
-1.62334248e-01 -2.76143640e-01 -6.81905746e-01 -6.77638361e-03
-5.37734032e-01 -1.00272322e+00 8.62121880e-01 -2.72768140e-02
7.37927377e-01 -1.11259329e+00 -5.33850074e-01 -4.11219478e-01
3.43334854e-01 -1.23176992e+00 -4.88886237e-01 -4.79511589e-01
-5.05985081e-01 -1.64436996e+00 -6.16661668e-01 -9.98792171e-01
9.48224783e-01 3.93758714e-01 9.61604059e-01 4.70100522e-01
-5.77183962e-01 5.30814111e-01 -2.01889142e-01 -4.36515689e-01
-1.64052062e-02 -5.14733195e-01 6.45077825e-01 7.71718621e-01
6.20908320e-01 -3.55470896e-01 -4.94908810e-01 7.56154656e-01
-8.26887190e-01 -6.28167689e-01 3.14323783e-01 6.96973443e-01
3.59027623e-03 1.76375270e-01 1.46677524e-01 -6.41972899e-01
3.75568062e-01 -3.69788021e-01 -7.51911521e-01 7.01454580e-01
-1.63985699e-01 -5.46238482e-01 1.31929770e-01 -3.92074823e-01
-1.05699241e+00 2.11886734e-01 3.39890391e-01 -5.47298372e-01
-6.33824766e-01 -3.71819764e-01 -6.21116579e-01 -6.38694406e-01
4.06447023e-01 2.96398610e-01 5.72075665e-01 -3.69556248e-01
-2.60141850e-01 1.07731378e+00 4.53068167e-01 -2.17798576e-01
5.79182386e-01 5.06481826e-01 1.57206520e-01 -1.36552060e+00
-3.00260872e-01 -7.93925300e-02 -8.60718548e-01 -5.90926707e-01
7.72998452e-01 -8.27479362e-01 -1.15281081e+00 1.09075165e+00
-1.24255455e+00 3.85284334e-01 7.85791218e-01 3.87213737e-01
-7.00315237e-02 2.85447299e-01 -3.14534634e-01 -1.17137122e+00
-4.28685963e-01 -1.31541014e+00 9.63613868e-01 4.20815498e-01
2.13954061e-01 -6.04240477e-01 -2.68408239e-01 3.23229522e-01
4.67987448e-01 2.65176713e-01 2.40833476e-01 -5.24848163e-01
-4.56478983e-01 -1.37156332e-02 -2.27760952e-02 4.85792011e-01
6.32348955e-01 5.72414279e-01 -1.38056588e+00 -5.26642740e-01
4.43326294e-01 -4.71760631e-02 4.77186382e-01 3.93935084e-01
1.01739693e+00 -4.58630025e-01 -2.75878400e-01 8.14754486e-01
1.00875187e+00 7.07364202e-01 7.89691210e-01 -4.16736007e-02
4.23377782e-01 1.03763759e+00 2.93367952e-01 4.17658895e-01
-1.24533519e-01 6.25992954e-01 3.16600859e-01 1.35538414e-01
9.16319266e-02 6.31158769e-01 6.24780059e-01 2.59451240e-01
-3.45908195e-01 -3.32237810e-01 -7.59255707e-01 -1.61678597e-01
-1.42696404e+00 -1.34667337e+00 2.86324739e-01 2.17295265e+00
2.54094839e-01 -5.21633744e-01 1.76847011e-01 1.62673742e-01
1.34268773e+00 1.79652914e-01 -3.31153005e-01 -1.38460115e-01
-2.05336154e-01 4.56906073e-02 3.17053020e-01 2.30890587e-01
-1.25344253e+00 6.28187776e-01 7.21497011e+00 2.61833936e-01
-1.43729091e+00 9.57103297e-02 9.87613201e-01 -3.63200396e-01
5.94957471e-01 -5.60154438e-01 -1.26102376e+00 5.18840909e-01
9.17818785e-01 2.92900056e-01 6.32985473e-01 6.90492153e-01
4.85431924e-02 5.32435775e-02 -1.06693792e+00 1.49792159e+00
8.78436387e-01 -6.62403464e-01 3.99195775e-02 -6.94841295e-02
3.12969476e-01 -3.67130488e-01 2.19305694e-01 -1.83223903e-01
-4.31047499e-01 -1.38139820e+00 3.54545593e-01 6.29842758e-01
8.34684849e-01 -5.28676748e-01 7.49742270e-01 -6.14503697e-02
-1.03084731e+00 -1.36988729e-01 -2.77960479e-01 -4.95006293e-02
-1.52397901e-01 4.83506806e-02 -4.39347267e-01 2.42900014e-01
8.58704090e-01 4.26092356e-01 -6.76980793e-01 6.40099227e-01
4.73120183e-01 1.95966810e-01 -2.15799645e-01 1.73115551e-01
-2.75515735e-01 -6.18520975e-02 3.50347251e-01 1.01276636e+00
4.03939366e-01 2.87274063e-01 1.55038625e-01 2.36588672e-01
5.08268401e-02 -1.61442608e-01 -9.77499783e-01 2.88647205e-01
5.43387830e-01 1.37516892e+00 -5.61677754e-01 -9.51035544e-02
-5.92687130e-01 8.30884457e-01 -1.78680718e-01 3.18979055e-01
-8.85759056e-01 1.16610475e-01 1.15252793e+00 1.20235816e-01
1.63246796e-01 -1.23869739e-02 3.33288908e-01 -8.96233439e-01
1.59212694e-01 -1.27287543e+00 3.44902724e-01 -4.76329505e-01
-9.66809988e-01 1.10555792e+00 1.45180479e-01 -8.95546734e-01
-1.25022978e-01 -1.03564405e+00 -5.95597982e-01 7.82525361e-01
-1.38498807e+00 -9.65317011e-01 -7.30802536e-01 1.25017333e+00
5.45556426e-01 -9.83546734e-01 8.30461621e-01 7.19095647e-01
-1.23849869e+00 7.92533040e-01 -3.12612742e-01 3.43196481e-01
8.82203579e-01 -2.00447485e-01 1.01021111e-01 9.96023357e-01
-2.59699523e-01 6.17370963e-01 4.70983624e-01 -4.47422355e-01
-1.67788363e+00 -9.77612257e-01 3.63598764e-01 -4.38098341e-01
8.01877156e-02 -1.63048252e-01 -6.89368188e-01 7.02062607e-01
3.76146674e-01 5.40428519e-01 7.02858508e-01 -3.31695080e-01
-3.64175975e-01 -5.41738391e-01 -1.54350221e+00 2.38628924e-01
1.04623616e+00 -3.50572467e-01 -1.53347060e-01 3.60559583e-01
4.47847471e-02 -1.52536526e-01 -5.79250813e-01 4.97632682e-01
9.13762093e-01 -1.27133381e+00 9.05114472e-01 -7.46826410e-01
-3.57269526e-01 -3.69452596e-01 -4.07881618e-01 -5.13878703e-01
-2.63841510e-01 -8.67446482e-01 3.87452208e-02 1.53571677e+00
9.72973881e-04 -7.69704401e-01 8.09192479e-01 6.57763958e-01
3.44576776e-01 -3.29829901e-01 -1.00965238e+00 -7.06011832e-01
-7.75799453e-01 3.50654632e-01 7.57152319e-01 8.29084873e-01
-6.02972686e-01 -3.80262136e-02 -7.09593773e-01 5.79863548e-01
7.84626007e-01 -4.09238100e-01 7.66879916e-01 -1.34597230e+00
2.40352690e-01 -1.62989333e-01 -9.66590703e-01 -2.07364112e-01
3.83106619e-01 -1.06293872e-01 -2.25747779e-01 -3.75113994e-01
3.05111140e-01 1.44734532e-01 -1.07248910e-01 3.43243599e-01
1.11538440e-01 3.23212117e-01 9.87301767e-02 9.87227857e-02
-3.39961708e-01 5.24615534e-02 7.38070786e-01 1.04985079e-02
3.14059019e-01 4.16677110e-02 -2.48231605e-01 8.64294231e-01
7.52074242e-01 -3.00744832e-01 -4.16935116e-01 -4.76233929e-01
-7.44881272e-01 -4.47646575e-03 4.69884217e-01 -1.15985644e+00
3.66138965e-01 -2.74017185e-01 6.88096285e-01 -1.15658045e-01
4.26970959e-01 -1.16957498e+00 6.28763437e-01 4.99747723e-01
1.55075848e-01 6.88085079e-01 1.04994476e-01 2.90864706e-01
-2.27792412e-01 -2.07050622e-01 1.01743782e+00 -2.75968462e-01
-7.37393498e-01 6.21151328e-01 -2.97147155e-01 -3.11684608e-01
1.43221331e+00 -6.31282210e-01 -4.31136370e-01 -1.11585669e-01
-5.35057962e-01 -2.49289870e-01 3.15232933e-01 6.37453675e-01
7.16983795e-01 -1.16745818e+00 -8.36927354e-01 1.05166686e+00
-3.35845381e-01 -5.72961509e-01 2.46123001e-01 8.66092324e-01
-6.07059956e-01 2.62552023e-01 -8.24423671e-01 -5.57576180e-01
-2.38048267e+00 6.08516216e-01 6.60439253e-01 5.21021843e-01
7.10278675e-02 9.77324903e-01 1.31082624e-01 -5.43875620e-02
5.67055881e-01 3.37369621e-01 -5.75284541e-01 1.15628138e-01
1.17644751e+00 5.38297713e-01 1.58611596e-01 -1.14827728e+00
-6.80057526e-01 8.68272245e-01 -1.21036835e-01 2.96197176e-01
8.03454161e-01 -1.64929420e-01 -5.11039853e-01 -9.24646258e-02
1.10825944e+00 -3.21499288e-01 -9.12226379e-01 2.23916434e-02
-1.42758444e-01 -8.16121697e-01 -2.03475669e-01 -3.69986743e-01
-1.56909966e+00 5.29554009e-01 1.29382706e+00 1.62151158e-01
1.47571623e+00 -4.43522722e-01 5.13070002e-02 5.20089328e-01
4.39489454e-01 -6.98428273e-01 -1.64763570e-01 3.91082972e-01
9.64410007e-01 -1.37187934e+00 1.32953993e-03 -6.04998767e-01
-3.68819445e-01 1.30525804e+00 7.86464572e-01 1.13714933e-01
1.13542104e+00 5.96788466e-01 2.27482930e-01 -2.45727226e-01
-7.98037171e-01 1.67239249e-01 1.73402339e-01 9.27265704e-01
5.40466547e-01 -2.40677312e-01 1.18037485e-01 1.77910164e-01
1.45197380e-02 -1.24356762e-01 2.75666505e-01 8.82082224e-01
-4.66326803e-01 -8.32817137e-01 -1.02897561e+00 2.92435974e-01
-8.38050783e-01 1.98593616e-01 -6.04826629e-01 6.29147708e-01
2.63304055e-01 1.15996861e+00 1.77112564e-01 -4.49739516e-01
2.99969047e-01 2.81860977e-01 7.15173602e-01 -2.30401531e-01
-5.79947233e-01 -1.95685118e-01 -6.78361058e-02 -6.83380425e-01
-1.00876355e+00 -7.65309572e-01 -4.33572173e-01 -6.86441660e-01
-1.21671438e-01 -1.72759056e-01 6.02082551e-01 6.96222186e-01
3.54646921e-01 2.29416236e-01 8.52599382e-01 -9.74609792e-01
-4.32246149e-01 -1.12801719e+00 -7.42247283e-01 2.72781581e-01
6.22827232e-01 -8.33660483e-01 -4.37109470e-01 5.34216821e-01] | [13.282374382019043, 0.8011878132820129] |
43eb7dfd-fa06-4a3a-a18d-cbdced162299 | multi-modal-temporal-convolutional-network | 2107.09504 | null | https://arxiv.org/abs/2107.09504v1 | https://arxiv.org/pdf/2107.09504v1.pdf | Multi-Modal Temporal Convolutional Network for Anticipating Actions in Egocentric Videos | Anticipating human actions is an important task that needs to be addressed for the development of reliable intelligent agents, such as self-driving cars or robot assistants. While the ability to make future predictions with high accuracy is crucial for designing the anticipation approaches, the speed at which the inference is performed is not less important. Methods that are accurate but not sufficiently fast would introduce a high latency into the decision process. Thus, this will increase the reaction time of the underlying system. This poses a problem for domains such as autonomous driving, where the reaction time is crucial. In this work, we propose a simple and effective multi-modal architecture based on temporal convolutions. Our approach stacks a hierarchy of temporal convolutional layers and does not rely on recurrent layers to ensure a fast prediction. We further introduce a multi-modal fusion mechanism that captures the pairwise interactions between RGB, flow, and object modalities. Results on two large-scale datasets of egocentric videos, EPIC-Kitchens-55 and EPIC-Kitchens-100, show that our approach achieves comparable performance to the state-of-the-art approaches while being significantly faster. | ['Juergen Gall', 'Yazan Abu Farha', 'Olga Zatsarynna'] | 2021-07-18 | null | null | null | null | ['action-anticipation'] | ['computer-vision'] | [-6.71323538e-02 -3.13202925e-02 1.24617510e-01 -6.75459027e-01
-2.99557209e-01 -2.33191460e-01 6.11635923e-01 1.59547683e-02
-5.82308710e-01 3.95694613e-01 2.83315009e-03 -1.80771664e-01
1.02164216e-01 -8.66917610e-01 -7.73900628e-01 -5.90284944e-01
-1.35608539e-01 2.35789016e-01 7.97757983e-01 -4.26409602e-01
9.96531695e-02 4.73759562e-01 -1.97051287e+00 4.12360162e-01
4.89380896e-01 1.20451164e+00 2.47980818e-01 9.34583426e-01
3.00813705e-01 1.31145942e+00 4.79057431e-02 -2.34391242e-01
-1.95972808e-03 -1.91767275e-01 -8.07177365e-01 -1.86759442e-01
2.58049257e-02 -4.72303748e-01 -4.65382725e-01 6.20541096e-01
2.79686809e-01 4.87180829e-01 3.44755262e-01 -1.29876208e+00
6.08646385e-02 3.77250969e-01 -2.82580167e-01 1.28313705e-01
1.96689710e-01 5.14620602e-01 7.15144455e-01 -3.61964703e-01
6.49751723e-01 1.22056103e+00 5.52972734e-01 4.40593123e-01
-1.04662895e+00 -4.87993836e-01 3.31894457e-01 7.86104977e-01
-1.15699661e+00 -7.10955799e-01 8.14182639e-01 -4.62951332e-01
1.00140011e+00 -2.41993219e-02 6.52209997e-01 1.13339782e+00
3.60088617e-01 8.76930296e-01 7.33877957e-01 8.21603462e-02
5.05175471e-01 -2.47368403e-02 -2.70158611e-02 6.11325383e-01
-1.48874745e-01 1.51648298e-01 -5.84552824e-01 4.48132426e-01
3.91937226e-01 2.81132348e-02 -9.29248985e-03 -4.34300274e-01
-1.49561727e+00 7.38646150e-01 5.53605318e-01 2.21789137e-01
-6.37522936e-01 5.40056288e-01 6.56998575e-01 9.82362702e-02
2.10822225e-01 1.64072394e-01 -2.70225018e-01 -6.05585277e-01
-7.96286345e-01 2.81309873e-01 6.15797281e-01 7.21880734e-01
7.01954007e-01 -1.57547936e-01 -4.40329574e-02 2.05463737e-01
1.74866587e-01 2.67156512e-01 2.73121387e-01 -1.27153635e+00
3.29857677e-01 6.19609475e-01 4.18358386e-01 -1.04610968e+00
-7.62027860e-01 -8.07257369e-02 -7.11364985e-01 3.57977152e-01
4.43566650e-01 -9.15378332e-02 -6.98984504e-01 1.83724368e+00
4.87177819e-01 3.44534546e-01 1.87617004e-01 1.21259820e+00
4.70292956e-01 6.65155649e-01 2.92870224e-01 -4.51026186e-02
1.28798199e+00 -1.06864178e+00 -7.67702043e-01 -4.53220069e-01
7.85122097e-01 -5.46289206e-01 8.24741960e-01 2.86008924e-01
-1.05637634e+00 -8.33473980e-01 -1.02796817e+00 -2.24321038e-01
-3.99658680e-01 4.53036539e-02 9.75865543e-01 6.99448138e-02
-8.65844369e-01 7.86529362e-01 -1.21241868e+00 -4.97690082e-01
1.80009112e-01 3.27696025e-01 -5.55556297e-01 -9.15156957e-03
-1.25216222e+00 1.23100364e+00 4.29212093e-01 4.47923273e-01
-6.98302984e-01 -3.75031918e-01 -9.98043060e-01 1.32307913e-02
1.96797386e-01 -6.32563651e-01 1.52935922e+00 -9.14476156e-01
-1.68992162e+00 3.66911441e-01 -3.41746271e-01 -7.79020965e-01
8.39381158e-01 -3.28004479e-01 -4.10940081e-01 2.36501560e-01
-7.72669092e-02 9.67973769e-01 7.40045369e-01 -7.19398260e-01
-8.93228054e-01 -2.41019905e-01 3.92376512e-01 1.26580000e-01
-1.45899087e-01 -1.61485791e-01 -3.53048027e-01 9.81412679e-02
7.52913440e-03 -1.28004050e+00 -4.25339997e-01 1.03498735e-02
-1.84840783e-02 -3.98345351e-01 1.02374709e+00 -4.12627459e-01
8.72568429e-01 -2.17591286e+00 5.78012466e-02 -1.83076590e-01
-2.07545096e-03 1.21278770e-01 -4.49839421e-02 2.43084952e-01
1.49200633e-01 -4.52061474e-01 2.76144221e-02 -4.46149945e-01
3.69738117e-02 1.59498721e-01 -3.17612618e-01 5.41038513e-01
4.13750023e-01 8.68034899e-01 -9.71248388e-01 -3.95361513e-01
6.66488886e-01 6.48978829e-01 -5.19404769e-01 3.20355982e-01
-2.83711940e-01 5.07067800e-01 -4.95294929e-01 2.84051925e-01
2.35561728e-01 -2.50985712e-01 1.48263257e-02 -3.45701933e-01
-4.33723122e-01 3.65420699e-01 -9.74084556e-01 1.93281639e+00
-6.66214228e-01 8.10410321e-01 -2.22556472e-01 -1.05662251e+00
6.67933524e-01 3.48273128e-01 5.75510323e-01 -1.03858685e+00
3.17206055e-01 1.64454598e-02 9.27274600e-02 -6.39172614e-01
7.20618486e-01 1.19993396e-01 -1.46606311e-01 3.51418823e-01
-2.52751917e-01 -5.52812591e-02 4.70843464e-01 1.25064626e-01
1.16223812e+00 6.48847520e-01 -2.54899580e-02 3.41168046e-02
6.02789044e-01 2.17981398e-01 7.01387167e-01 3.60289663e-01
-5.79738975e-01 3.03801090e-01 4.78886932e-01 -8.93237054e-01
-9.85987067e-01 -7.05117881e-01 2.04785943e-01 9.01788831e-01
4.35272008e-01 -3.23211759e-01 -6.07747257e-01 -3.45714182e-01
-2.11552501e-01 8.05658638e-01 -6.55176520e-01 -4.24629539e-01
-5.75843155e-01 -4.42952931e-01 2.77617067e-01 7.51139998e-01
7.08978891e-01 -1.09929836e+00 -1.34866118e+00 4.38320041e-01
-5.80682039e-01 -1.66481125e+00 1.43720144e-02 -6.14316612e-02
-7.54161417e-01 -1.02595031e+00 -2.15178102e-01 -2.76709616e-01
3.96976590e-01 2.83146232e-01 9.60075915e-01 -3.80631089e-01
-1.33209050e-01 2.25686431e-01 -2.47411489e-01 -3.45072538e-01
-8.04044977e-02 6.90985471e-02 1.20596841e-01 1.29791647e-01
5.45110047e-01 -5.71269035e-01 -8.93420219e-01 3.47916663e-01
-6.12546444e-01 6.35755301e-01 5.77078581e-01 5.62805414e-01
3.83823395e-01 6.36459514e-02 3.30660075e-01 -3.53705764e-01
5.95037490e-02 -3.54728520e-01 -6.53933764e-01 -6.59207702e-02
-2.46559411e-01 2.52071530e-01 7.85592973e-01 -4.76666898e-01
-1.12529314e+00 5.75139105e-01 -7.07497373e-02 -3.92669022e-01
-2.69131273e-01 2.33828232e-01 1.53427005e-01 1.31169915e-01
3.98851216e-01 -5.66135496e-02 -5.19031547e-02 -2.82769371e-02
4.63090390e-01 4.84701663e-01 5.81244707e-01 -2.22464845e-01
4.19488877e-01 7.89235830e-01 1.76845253e-01 -7.99228609e-01
-6.52572453e-01 -4.46123570e-01 -6.31438017e-01 -6.94047630e-01
1.00086892e+00 -1.01666951e+00 -1.33610296e+00 4.37622070e-01
-1.29247248e+00 -5.30859232e-01 -9.27900597e-02 6.93634391e-01
-8.54325593e-01 7.35098124e-02 -5.89290738e-01 -8.59943151e-01
-2.02676788e-01 -1.32842541e+00 1.00959873e+00 2.84278691e-01
-4.15738225e-01 -6.84101582e-01 -7.19648525e-02 5.44705153e-01
5.55955112e-01 3.10232997e-01 3.28383058e-01 -1.92950532e-01
-9.47283983e-01 -2.53074765e-01 -2.88115621e-01 -2.76908390e-02
-2.03249514e-01 1.48385987e-01 -1.20576990e+00 -6.77946880e-02
-2.38207832e-01 -3.87769133e-01 8.98281336e-01 3.56296450e-01
1.05966032e+00 5.88725600e-03 -2.91371584e-01 3.97895485e-01
1.05527961e+00 2.13105321e-01 7.60148704e-01 2.98786163e-01
5.80921710e-01 9.24905598e-01 8.97625744e-01 4.54842091e-01
1.00976264e+00 6.38004363e-01 6.43032849e-01 1.46670386e-01
1.16395094e-01 9.24619474e-03 5.64059436e-01 6.72296226e-01
-3.28893214e-01 -1.20333903e-01 -8.92429829e-01 7.05901861e-01
-2.40099692e+00 -1.27449346e+00 -1.97409391e-01 2.03841996e+00
3.97951633e-01 2.46087328e-01 -3.90533842e-02 2.82753527e-01
3.98048639e-01 1.76594615e-01 -7.33383715e-01 -5.85192621e-01
2.72588789e-01 -1.74554691e-01 3.03485781e-01 3.76710922e-01
-1.26734746e+00 1.08694887e+00 5.64275074e+00 2.10039601e-01
-1.39222050e+00 -7.19771832e-02 6.08524501e-01 -2.66832232e-01
2.07035437e-01 5.41191697e-02 -5.47874868e-01 4.99213576e-01
1.41764927e+00 1.97770119e-01 4.02115792e-01 9.58938062e-01
6.55646861e-01 -5.42186975e-01 -1.30022824e+00 9.74887967e-01
-2.87668943e-01 -1.13701081e+00 -5.82633317e-01 -2.73271054e-01
4.40977335e-01 2.57157952e-01 -1.85681984e-01 3.10404062e-01
2.97925204e-01 -7.91201055e-01 1.00346041e+00 7.54179358e-01
2.62676805e-01 -8.56218457e-01 7.54393041e-01 3.67360741e-01
-1.36712146e+00 -1.49302766e-01 -1.75980985e-01 -2.96774089e-01
4.17246103e-01 6.95409954e-01 -6.39138460e-01 2.84600109e-01
8.76496136e-01 8.50417435e-01 -3.51492494e-01 6.82644248e-01
-2.67912775e-01 2.53168911e-01 -3.94888163e-01 -1.79607332e-01
2.93366075e-01 -3.24803442e-02 2.06845373e-01 1.07635522e+00
3.09888512e-01 1.54028982e-01 -2.15419661e-03 6.06917799e-01
3.41042578e-01 -2.86908567e-01 -6.31063342e-01 -6.30114526e-02
5.03237545e-02 1.43304491e+00 -6.45597219e-01 -4.37503785e-01
-5.21266103e-01 1.06764996e+00 3.83821279e-01 2.20568344e-01
-1.12878275e+00 -3.45906585e-01 9.54819500e-01 -4.06396985e-02
3.12090993e-01 -5.93744397e-01 -1.19173467e-01 -1.05453670e+00
1.66026920e-01 -4.29849058e-01 1.71324372e-01 -8.88630569e-01
-6.03340268e-01 6.19088531e-01 -3.69267583e-01 -1.18934977e+00
-7.00039089e-01 -5.28082967e-01 -4.28404331e-01 5.36563814e-01
-1.81631100e+00 -1.26229322e+00 -6.92209125e-01 5.67388833e-01
4.53794718e-01 3.51570755e-01 6.85978413e-01 3.48345190e-01
-6.16642177e-01 2.40151267e-02 -4.41628516e-01 -7.61060417e-02
6.34872377e-01 -8.71898472e-01 4.97704357e-01 9.55570698e-01
-3.57555121e-01 3.10213447e-01 9.77480412e-01 -2.52413154e-01
-1.67789960e+00 -1.12889147e+00 9.76745129e-01 -3.69857311e-01
6.09705389e-01 -2.18563125e-01 -6.81515753e-01 5.73320270e-01
1.49547597e-02 2.83525914e-01 2.35731289e-01 -8.95073265e-03
-1.60408258e-01 -3.77635032e-01 -8.68465483e-01 7.47066438e-01
9.71202135e-01 -5.32469034e-01 -3.02215636e-01 1.10803254e-01
6.44564509e-01 -3.50930333e-01 -6.94303095e-01 4.21865821e-01
8.75125170e-01 -1.36901700e+00 7.75490284e-01 -3.16604942e-01
5.26041269e-01 -4.96687502e-01 5.92769310e-03 -1.04648232e+00
-3.07540774e-01 -4.91967052e-01 -4.50345755e-01 6.83636367e-01
2.07196355e-01 -5.36174536e-01 6.80746377e-01 9.45111513e-01
-2.66920142e-02 -5.65469861e-01 -7.90479958e-01 -5.65527320e-01
-5.02422750e-01 -8.50016117e-01 5.64982593e-01 5.89209557e-01
7.29795471e-02 3.45199466e-01 -3.59367490e-01 1.68757737e-01
4.59793687e-01 1.74854904e-01 8.85355473e-01 -1.12810493e+00
1.32329360e-01 -2.55012780e-01 -7.36687124e-01 -8.72573376e-01
2.42883995e-01 -2.62200713e-01 2.72107959e-01 -1.56367683e+00
-1.91642493e-01 -3.28694493e-01 -3.35107923e-01 5.08012116e-01
9.58401859e-02 1.35577828e-01 5.17119542e-02 1.77102443e-02
-9.31519032e-01 7.39463568e-01 1.12283611e+00 -3.62567566e-02
-1.83374971e-01 -1.20009772e-01 -1.95639327e-01 7.69056141e-01
7.92986274e-01 -9.23915356e-02 -6.15506768e-01 -4.18625951e-01
3.93213272e-01 1.39492154e-01 5.36092877e-01 -1.44271755e+00
5.99981368e-01 -1.80713192e-01 2.48748824e-01 -6.76713526e-01
8.46506774e-01 -8.83922994e-01 1.01731889e-01 6.11449420e-01
-2.50586480e-01 1.59787416e-01 2.08614618e-01 5.10367393e-01
-2.62871385e-01 4.16530222e-01 8.50835443e-01 -3.70436721e-02
-1.23789692e+00 4.03141201e-01 -4.71054405e-01 -4.00628000e-01
1.38438201e+00 -6.93756342e-02 -2.24234909e-01 -4.47992980e-01
-4.27590430e-01 4.97140855e-01 3.44908088e-01 8.08608234e-01
5.82991064e-01 -1.19488752e+00 -4.17868048e-01 7.21269995e-02
2.30649143e-01 8.97721797e-02 5.08610308e-01 1.12774956e+00
-4.81527418e-01 5.63798845e-01 -2.82344818e-01 -7.69295514e-01
-9.61913109e-01 4.92770344e-01 2.74141669e-01 -9.76699591e-02
-7.06136346e-01 5.44116020e-01 1.61484122e-01 -2.06330836e-01
9.18576494e-02 -5.02657056e-01 -3.10022086e-01 1.23365805e-01
6.84120417e-01 3.91296417e-01 6.82198256e-02 -8.03419054e-01
-4.90677893e-01 3.86805683e-01 3.02552003e-02 -9.62213054e-02
1.38418400e+00 -2.14487612e-01 -1.23060439e-02 5.69106638e-01
8.95242631e-01 -7.44772255e-01 -1.59790766e+00 4.32914197e-02
-7.23663345e-02 -2.50467539e-01 3.66623729e-01 -4.07129824e-01
-1.00277400e+00 1.05287325e+00 6.61253989e-01 -7.29499944e-03
1.10965872e+00 -3.38696927e-01 1.05534101e+00 5.55953383e-01
5.34841835e-01 -1.32010221e+00 1.88094676e-02 6.11222804e-01
5.97487867e-01 -1.63438118e+00 -2.11362675e-01 -1.26538947e-01
-7.88196743e-01 1.15785670e+00 8.04381549e-01 1.12382174e-02
4.70453620e-01 1.26846030e-01 2.57619773e-03 -6.12149574e-02
-1.09699428e+00 -3.71317804e-01 1.16709732e-01 2.28664398e-01
3.35435003e-01 6.66414425e-02 4.68645021e-02 2.67823428e-01
8.42665380e-04 2.22245306e-01 3.41576904e-01 1.00858819e+00
-2.60008603e-01 -6.35522127e-01 -4.12739068e-02 9.53783393e-02
-8.18528980e-02 3.47819418e-01 9.46193039e-02 5.84460795e-01
7.12887198e-02 1.07813108e+00 3.01004380e-01 -5.74919164e-01
3.03812623e-01 -1.33518325e-02 3.30446571e-01 3.79336849e-02
-2.05520272e-01 -3.72544140e-01 2.44344145e-01 -1.31666720e+00
-7.17012227e-01 -6.74312413e-01 -1.42674553e+00 -7.06125081e-01
-3.98562364e-02 -1.76091582e-01 9.07962739e-01 1.15053785e+00
6.53837383e-01 7.68569291e-01 5.55779278e-01 -1.18486130e+00
-8.13300088e-02 -7.77826369e-01 -7.86982775e-02 3.79519403e-01
4.28042978e-01 -7.45308578e-01 -8.65465328e-02 1.06232911e-01] | [7.191586494445801, 0.4076496660709381] |
9a179d34-20af-4fe5-9b2f-4b5a608acd0d | type-information-utilized-event-detection-via | 2211.08168 | null | https://arxiv.org/abs/2211.08168v1 | https://arxiv.org/pdf/2211.08168v1.pdf | Type Information Utilized Event Detection via Multi-Channel GNNs in Electrical Power Systems | Event detection in power systems aims to identify triggers and event types, which helps relevant personnel respond to emergencies promptly and facilitates the optimization of power supply strategies. However, the limited length of short electrical record texts causes severe information sparsity, and numerous domain-specific terminologies of power systems makes it difficult to transfer knowledge from language models pre-trained on general-domain texts. Traditional event detection approaches primarily focus on the general domain and ignore these two problems in the power system domain. To address the above issues, we propose a Multi-Channel graph neural network utilizing Type information for Event Detection in power systems, named MC-TED, leveraging a semantic channel and a topological channel to enrich information interaction from short texts. Concretely, the semantic channel refines textual representations with semantic similarity, building the semantic information interaction among potential event-related words. The topological channel generates a relation-type-aware graph modeling word dependencies, and a word-type-aware graph integrating part-of-speech tags. To further reduce errors worsened by professional terminologies in type analysis, a type learning mechanism is designed for updating the representations of both the word type and relation type in the topological channel. In this way, the information sparsity and professional term occurrence problems can be alleviated by enabling interaction between topological and semantic information. Furthermore, to address the lack of labeled data in power systems, we built a Chinese event detection dataset based on electrical Power Event texts, named PoE. In experiments, our model achieves compelling results not only on the PoE dataset, but on general-domain event detection datasets including ACE 2005 and MAVEN. | ['Pengtao Xie', 'Shan Xue', 'Qingyun Sun', 'Jiawei Sheng', 'Yiming Hei', 'Cheng Ji', 'Lihong Wang', 'JianXin Li', 'Qian Li'] | 2022-11-15 | null | null | null | null | ['type'] | ['speech'] | [-4.49569263e-02 -4.21137828e-03 -4.91945773e-01 -2.61135370e-01
-3.41414899e-01 -5.85036814e-01 3.01478148e-01 5.13947785e-01
-3.54795381e-02 5.34279108e-01 4.83815789e-01 -5.33031762e-01
-4.75535542e-01 -1.38733518e+00 -3.44661385e-01 -5.72936177e-01
-3.17054063e-01 2.58972019e-01 -4.26777489e-02 -4.53292280e-01
1.57073420e-02 4.08274561e-01 -1.15012074e+00 2.79774189e-01
8.98124874e-01 1.11118686e+00 3.30396563e-01 -1.65707944e-03
-5.76968431e-01 1.07466197e+00 -9.77461219e-01 -9.35978824e-05
-1.67740241e-01 -4.11504447e-01 -6.76237822e-01 -6.75059110e-03
-5.12918949e-01 -1.05055898e-01 -9.18405712e-01 1.04026711e+00
6.00265920e-01 2.12806240e-01 5.67673028e-01 -1.58518934e+00
-5.82689464e-01 1.26383448e+00 -2.50228077e-01 4.88913149e-01
6.01707220e-01 -1.74694464e-01 1.11458445e+00 -6.37682319e-01
3.65895599e-01 9.81462836e-01 7.10405827e-01 7.94832781e-02
-5.46734571e-01 -9.35515165e-01 3.61253560e-01 6.54315531e-01
-1.54630351e+00 1.55074939e-01 1.31219685e+00 -2.55511165e-01
1.51222360e+00 8.13123137e-02 8.23863089e-01 1.20801079e+00
2.91018248e-01 8.18231225e-01 1.80110112e-01 -1.55450016e-01
2.54487455e-01 -1.90007851e-01 2.84213364e-01 5.32741427e-01
1.01536460e-01 -7.02851117e-02 -4.27087098e-01 -1.76253423e-01
6.37602448e-01 4.72648412e-01 -2.56443083e-01 1.33532807e-01
-9.53538418e-01 7.14941025e-01 3.32699895e-01 8.37132990e-01
-4.31284785e-01 -2.34663188e-01 8.94840837e-01 7.10802674e-02
5.13223827e-01 4.67594057e-01 -7.28401303e-01 -2.44587511e-01
-5.72740853e-01 -7.28457645e-02 9.75658834e-01 1.39720762e+00
6.27158403e-01 5.14226437e-01 -4.34190512e-01 9.20665324e-01
9.56095979e-02 5.39820135e-01 5.07010460e-01 -1.36477565e-02
8.44212174e-01 1.15142226e+00 -4.28475022e-01 -1.13938582e+00
-1.04163527e+00 -5.56043625e-01 -9.95576560e-01 -7.44377255e-01
-1.12971582e-01 -4.89273906e-01 -6.97559357e-01 1.67701423e+00
1.08510196e-01 3.19779694e-01 5.02839237e-02 6.14539683e-01
1.17593181e+00 1.05889189e+00 2.80167848e-01 -4.53980982e-01
1.65498412e+00 -4.78025585e-01 -1.39286697e+00 -2.29488403e-01
8.04058015e-01 -2.83828586e-01 6.54757619e-01 2.26368293e-01
-5.41531563e-01 -3.25330645e-01 -1.15975308e+00 1.33826584e-01
-7.59130120e-01 1.10850506e-01 7.42485821e-01 2.27926672e-01
-3.41546774e-01 3.76605392e-01 -4.43303943e-01 -3.75255257e-01
3.81647587e-01 7.86515772e-02 5.73291592e-02 1.18445322e-01
-1.97697961e+00 7.88115621e-01 8.99870694e-01 3.92586663e-02
-8.05651546e-01 -9.40171361e-01 -1.27635705e+00 3.20610017e-01
7.41625845e-01 -2.30627164e-01 8.19083512e-01 -2.01990157e-01
-9.93807256e-01 6.43019006e-02 -5.19364607e-03 -2.12965235e-01
-4.47456837e-01 1.11021943e-01 -1.27348113e+00 1.59452826e-01
2.50752240e-01 -2.32590586e-01 5.54105222e-01 -7.43788600e-01
-8.52971792e-01 -2.64727563e-01 6.86443523e-02 2.50817955e-01
-8.65631759e-01 -5.16214110e-02 -3.36010844e-01 -1.00124645e+00
-1.81956276e-01 -2.57139176e-01 -1.88135535e-01 -7.64792442e-01
-3.55299234e-01 -7.96547711e-01 9.62831438e-01 -7.88051486e-01
1.93684256e+00 -2.25256085e+00 3.48753948e-03 4.06268865e-01
2.24096090e-01 -1.17532268e-01 -1.37846582e-02 7.35462248e-01
-3.78475070e-01 -8.87105167e-02 -7.55395889e-02 3.59004647e-01
2.58072227e-01 5.93188941e-01 -6.01446509e-01 7.12697580e-02
1.40948564e-01 9.23938274e-01 -1.22243464e+00 -4.72606033e-01
3.92967612e-01 1.45309985e-01 -1.89310938e-01 2.63084471e-01
-6.10498302e-02 3.76172625e-02 -9.78535116e-01 5.55258811e-01
3.00862908e-01 -4.31337684e-01 3.11286777e-01 -6.17478013e-01
1.14580892e-01 6.46278381e-01 -1.11615992e+00 1.92738044e+00
-7.45125890e-01 1.55081049e-01 -1.91879243e-01 -1.56046343e+00
9.90523338e-01 6.54824913e-01 1.15653741e+00 -9.66897488e-01
2.18120709e-01 -1.47205114e-01 -4.24454242e-01 -8.01109552e-01
5.63866198e-01 -4.73514162e-02 -6.60124779e-01 1.75758779e-01
4.07405734e-01 -1.27555355e-01 4.73548412e-01 5.10292053e-01
1.38445807e+00 -4.72011566e-01 3.37153316e-01 -3.11783969e-01
3.41959715e-01 6.82833493e-02 8.72407973e-01 3.70350420e-01
2.62869656e-01 1.52216151e-01 4.49010670e-01 -5.10252297e-01
-5.16970992e-01 -1.05543733e+00 -3.12694818e-01 9.17757094e-01
4.62691277e-01 -1.14983702e+00 -3.21756750e-01 -1.07854724e+00
6.09977078e-03 9.53991652e-01 -2.65213490e-01 -5.02998471e-01
-5.76858282e-01 -9.43565726e-01 7.66578019e-01 9.88994062e-01
3.98866713e-01 -1.22334123e+00 6.66184947e-02 5.82058012e-01
-3.59605789e-01 -1.31400406e+00 -4.36514020e-01 6.25467896e-01
-4.64124590e-01 -1.14308143e+00 -1.60024494e-01 -8.33668649e-01
6.12856984e-01 -2.30438277e-01 1.24928999e+00 -6.53778166e-02
-1.90568954e-01 2.46038228e-01 -7.21336782e-01 -4.09025043e-01
-1.61477938e-01 1.74558654e-01 5.58636636e-02 -3.56703550e-01
6.56104922e-01 -5.12604892e-01 -1.90083325e-01 3.08409393e-01
-1.08323455e+00 -2.31830373e-01 3.64551365e-01 7.38941371e-01
1.08180605e-01 9.83318567e-01 1.01447415e+00 -7.61268258e-01
7.61800289e-01 -9.14024115e-01 -4.16889995e-01 2.08788484e-01
-8.00211787e-01 -1.42932922e-01 1.13050246e+00 -5.14448225e-01
-1.09127796e+00 -2.85134047e-01 -2.71652311e-01 -1.91236213e-01
-4.35646847e-02 9.94114697e-01 -6.29109859e-01 3.26205343e-01
4.29917127e-01 2.86054522e-01 -6.87851906e-01 -2.50958920e-01
2.74673194e-01 8.08210015e-01 4.42298740e-01 -6.89904034e-01
7.81059325e-01 2.96797976e-02 -1.43406346e-01 -8.19504082e-01
-6.60515428e-01 -5.09932756e-01 -1.82471484e-01 -8.24864134e-02
8.51521134e-01 -9.25859094e-01 -5.85354209e-01 2.46848732e-01
-1.16090703e+00 -9.63005275e-02 -4.97098356e-01 4.81923878e-01
-4.84687835e-02 4.70145226e-01 -8.12640369e-01 -4.96038139e-01
-2.04517126e-01 -7.98637390e-01 1.00502217e+00 2.00567424e-01
-2.66287625e-01 -1.30501354e+00 -3.99736702e-01 -1.61414057e-01
1.84447259e-01 -7.88964927e-02 1.36900508e+00 -9.67017412e-01
-1.57943100e-01 -2.61126459e-01 -1.43290669e-01 1.86568946e-01
5.90268731e-01 -7.24563003e-01 -5.23239374e-01 -1.66934147e-01
9.80203971e-02 1.33010074e-01 3.31787288e-01 2.90360332e-01
1.59756482e+00 -5.67197084e-01 -6.39534771e-01 3.61485481e-01
1.03177512e+00 7.28765666e-01 4.07374024e-01 -3.26554105e-02
1.04613400e+00 2.81661302e-01 5.03293931e-01 9.52411234e-01
6.52799129e-01 4.23238337e-01 8.19826648e-02 -1.43958658e-01
4.09340076e-02 -5.65946877e-01 1.23742700e-01 1.23035038e+00
3.19555610e-01 -5.65064669e-01 -7.88630962e-01 8.19074988e-01
-1.81972790e+00 -8.80815864e-01 5.28722629e-02 1.67045832e+00
1.01981831e+00 1.12389520e-01 -2.48844445e-01 3.64508539e-01
6.46592259e-01 3.83939862e-01 -2.81933248e-01 8.60134959e-02
-1.63836047e-01 9.45952758e-02 4.94397730e-01 7.43353590e-02
-1.05164063e+00 5.85410833e-01 5.33234930e+00 1.49217069e+00
-8.06397140e-01 1.21485200e-02 1.26859322e-01 1.19329087e-01
-5.91558039e-01 -1.18587799e-01 -8.65301371e-01 6.72322690e-01
8.60303044e-01 -5.50612807e-01 4.44205612e-01 6.69842064e-01
2.24510774e-01 3.44304591e-01 -1.25572145e+00 1.23095977e+00
1.25789955e-01 -1.27672780e+00 8.74863043e-02 -2.38180429e-01
4.18731421e-01 -1.92614198e-01 -4.51416492e-01 6.96860492e-01
2.22495571e-01 -8.02722037e-01 4.42294955e-01 1.01324849e-01
6.03181064e-01 -7.84363806e-01 7.66687274e-01 2.23435178e-01
-1.95425951e+00 -3.25304717e-01 -1.31691933e-01 2.57423683e-03
5.15896320e-01 8.88885021e-01 -7.71825969e-01 1.38303459e+00
7.64652193e-01 1.29460371e+00 -1.47544309e-01 4.84388351e-01
-4.14934069e-01 7.10724115e-01 -4.42681640e-01 -2.72476505e-02
-2.54391655e-02 7.44001567e-03 5.19055307e-01 1.23542655e+00
3.90810490e-01 3.95058066e-01 6.27200902e-01 9.38395917e-01
-1.12660214e-01 -1.03806920e-01 -7.99101412e-01 -3.06286514e-01
8.92670631e-01 1.36320841e+00 -6.29836917e-01 -4.60697234e-01
-6.48416281e-01 4.67641622e-01 -2.03470886e-01 7.16187358e-01
-9.56731677e-01 -8.60121429e-01 3.42175156e-01 -1.17702320e-01
6.13841489e-02 -1.41377196e-01 -3.58000219e-01 -1.21790075e+00
5.82936779e-02 -6.36030674e-01 9.78448987e-01 -5.93516767e-01
-1.71971428e+00 5.45835793e-01 4.63945866e-01 -1.24718130e+00
-3.41818005e-01 -4.32804585e-01 -5.62232435e-01 6.38166070e-01
-1.34447193e+00 -9.50903356e-01 -1.02920204e-01 8.18128109e-01
7.70468533e-01 -3.02899361e-01 7.44445682e-01 8.80146205e-01
-6.60591960e-01 4.88932759e-01 -5.21699667e-01 5.92391849e-01
2.99490184e-01 -1.23345602e+00 1.46964312e-01 8.06066692e-01
1.17489576e-01 4.24651533e-01 3.95223260e-01 -9.36969936e-01
-1.71412456e+00 -1.32150590e+00 8.79782200e-01 -1.77945927e-01
9.94484365e-01 -3.54766160e-01 -1.07037282e+00 5.84949732e-01
9.37151685e-02 -6.15760610e-02 6.33658051e-01 3.40638012e-01
-6.14815354e-02 -1.69308409e-01 -8.09583545e-01 3.78499687e-01
1.28246772e+00 -9.31823552e-01 -9.64075387e-01 5.69124103e-01
8.07874978e-01 -3.81031841e-01 -1.25552714e+00 7.42970169e-01
-2.89925218e-01 1.08277537e-01 9.95050907e-01 -3.54223311e-01
7.90846273e-02 -4.51468080e-01 4.69326302e-02 -1.50444508e+00
-4.03113276e-01 -4.81907755e-01 -3.41931343e-01 1.52108347e+00
4.97915834e-01 -5.16192615e-01 4.24024493e-01 3.22950691e-01
-6.70830667e-01 -4.85332906e-01 -9.02356982e-01 -8.32209885e-01
-2.69474596e-01 -9.65925574e-01 1.10841727e+00 1.35164177e+00
8.65613580e-01 6.37274981e-01 -2.40907207e-01 4.66779083e-01
2.31758624e-01 1.46980986e-01 9.89364460e-02 -1.13270116e+00
1.31296776e-02 -4.61728334e-01 -2.12243229e-01 -9.96444762e-01
5.66345572e-01 -1.23668671e+00 9.76349620e-05 -1.67065525e+00
9.35046002e-03 -4.32285130e-01 -5.38663089e-01 9.95017350e-01
-6.11350723e-02 -3.51408213e-01 -3.25566471e-01 -1.37485489e-01
-6.39965534e-01 7.82994747e-01 1.00918877e+00 -5.34566760e-01
-9.40775424e-02 -3.18518251e-01 -4.75316703e-01 6.76437795e-01
6.03413701e-01 -3.26703697e-01 -6.80141985e-01 -4.13358450e-01
5.92595279e-01 2.87512511e-01 1.14115573e-01 -7.70723581e-01
6.26412272e-01 -1.00726113e-01 2.41805375e-01 -7.74946749e-01
-1.18770272e-01 -1.19255269e+00 4.25908994e-03 1.08912170e-01
-1.22818321e-01 1.14080876e-01 3.82865757e-01 5.63954473e-01
-2.92447954e-01 -1.35832548e-01 -9.68293995e-02 1.54310033e-01
-1.26214790e+00 4.72628444e-01 -6.62219942e-01 1.66493237e-01
6.89899385e-01 1.32423446e-01 -4.19952482e-01 -1.99286222e-01
-5.19916654e-01 6.02646589e-01 -2.55128890e-01 1.01549244e+00
6.10460818e-01 -1.72558820e+00 -3.63387585e-01 5.70086181e-01
3.99143636e-01 8.57464820e-02 3.81410182e-01 4.87533838e-01
8.80818292e-02 2.48316213e-01 1.96734920e-01 -3.73763114e-01
-6.02378130e-01 6.14531636e-01 8.03605765e-02 -3.71808261e-01
-9.62853312e-01 2.89945602e-01 3.28371316e-01 -2.66096026e-01
3.65518361e-01 -6.73783123e-01 -6.80673122e-01 2.52046794e-01
4.66449231e-01 2.56994635e-01 4.66491193e-01 -3.20725083e-01
-5.23466885e-01 3.67618412e-01 2.48667210e-01 3.50849390e-01
1.25193167e+00 -4.79043573e-02 -2.33153418e-01 2.89618015e-01
1.01233959e+00 -1.71315923e-01 -6.39111102e-01 -2.65363544e-01
9.75007713e-02 6.76113442e-02 2.64100522e-01 -9.06163871e-01
-1.48926997e+00 5.41553199e-01 6.22469001e-02 6.54918075e-01
1.51078558e+00 8.06678608e-02 1.11474478e+00 5.01229286e-01
5.58591902e-01 -1.29568756e+00 -1.17874451e-01 7.86901832e-01
6.52208269e-01 -7.50399172e-01 -1.44064113e-01 -5.28674841e-01
-4.50173467e-01 9.76338863e-01 5.12069523e-01 2.86119640e-01
8.27031195e-01 7.84186959e-01 -2.71142811e-01 -5.73142588e-01
-4.92503881e-01 -1.23796560e-01 4.07352835e-01 5.60542405e-01
1.94140151e-01 1.14438444e-01 -2.79231459e-01 1.42250812e+00
-1.80899992e-03 -1.13756210e-01 2.06481189e-01 9.87456977e-01
-9.53858718e-02 -1.11799455e+00 -9.48199555e-02 5.81618011e-01
-4.61891860e-01 -2.14707196e-01 7.76706776e-03 5.68248808e-01
1.71731964e-01 1.33504319e+00 9.04559121e-02 -6.61049843e-01
6.67612791e-01 -3.26507799e-02 4.76713367e-02 -6.80758953e-01
-7.37475812e-01 1.68991759e-02 3.17901999e-01 -5.35032392e-01
1.26919174e-03 -1.71245500e-01 -1.84796965e+00 -5.54068154e-03
-4.38110173e-01 7.35306740e-01 2.61919677e-01 1.29080415e+00
3.81783277e-01 1.35403204e+00 7.07176626e-01 -3.31112534e-01
-2.23210841e-01 -1.05170166e+00 -1.03382170e+00 6.85663819e-01
-5.79455718e-02 -7.50364602e-01 -2.15187237e-01 -1.69135541e-01] | [9.067193984985352, 9.151049613952637] |
6e3bb292-d615-43b7-8b3d-f55c001b33bb | are-facial-attributes-adversarially-robust | 1605.05411 | null | http://arxiv.org/abs/1605.05411v3 | http://arxiv.org/pdf/1605.05411v3.pdf | Are Facial Attributes Adversarially Robust? | Facial attributes are emerging soft biometrics that have the potential to
reject non-matches, for example, based on mismatching gender. To be usable in
stand-alone systems, facial attributes must be extracted from images
automatically and reliably. In this paper, we propose a simple yet effective
solution for automatic facial attribute extraction by training a deep
convolutional neural network (DCNN) for each facial attribute separately,
without using any pre-training or dataset augmentation, and we obtain new
state-of-the-art facial attribute classification results on the CelebA
benchmark. To test the stability of the networks, we generated adversarial
images -- formed by adding imperceptible non-random perturbations to original
inputs which result in classification errors -- via a novel fast flipping
attribute (FFA) technique. We show that FFA generates more adversarial examples
than other related algorithms, and that DCNNs for certain attributes are
generally robust to adversarial inputs, while DCNNs for other attributes are
not. This result is surprising because no DCNNs tested to date have exhibited
robustness to adversarial images without explicit augmentation in the training
procedure to account for adversarial examples. Finally, we introduce the
concept of natural adversarial samples, i.e., images that are misclassified but
can be easily turned into correctly classified images by applying small
perturbations. We demonstrate that natural adversarial samples commonly occur,
even within the training set, and show that many of these images remain
misclassified even with additional training epochs. This phenomenon is
surprising because correcting the misclassification, particularly when guided
by training data, should require only a small adjustment to the DCNN
parameters. | ['Manuel Günther', 'Ethan M. Rudd', 'Andras Rozsa', 'Terrance E. Boult'] | 2016-05-18 | null | null | null | null | ['facial-attribute-classification'] | ['computer-vision'] | [ 7.94326782e-01 3.40639949e-01 3.22909474e-01 -7.61682451e-01
-5.97976565e-01 -9.22440767e-01 6.07028961e-01 -1.15733877e-01
-3.73788685e-01 8.34767580e-01 -4.29963380e-01 -1.35974810e-01
1.05836622e-01 -8.92318189e-01 -1.04743659e+00 -9.74755168e-01
-2.32568458e-01 3.97838920e-01 -2.13698879e-01 -1.81952730e-01
-2.32801199e-01 8.82503629e-01 -1.87920976e+00 3.23045701e-01
7.24769890e-01 1.17719996e+00 -7.50467181e-01 6.95030093e-01
3.73021394e-01 3.06565642e-01 -8.51519525e-01 -9.94020224e-01
7.12540627e-01 -4.83195454e-01 -6.90950274e-01 -2.69712303e-02
1.02037287e+00 -4.23872232e-01 -1.09834008e-01 1.21114421e+00
5.49892545e-01 -1.89041644e-01 8.78035367e-01 -1.62525713e+00
-7.37121522e-01 1.20723687e-01 -9.14623365e-02 -4.43923771e-01
3.56843352e-01 3.46476734e-01 5.14681995e-01 -8.76720786e-01
5.60157895e-01 1.21358359e+00 8.60146523e-01 1.13111770e+00
-1.46573842e+00 -1.00020242e+00 -2.04881847e-01 1.49972755e-02
-1.50655937e+00 -8.09986472e-01 6.75963223e-01 -2.37670243e-01
3.10486436e-01 6.29028678e-01 3.91128212e-01 1.50762904e+00
7.25264773e-02 2.37398237e-01 1.42858768e+00 -3.37474823e-01
1.34275392e-01 1.76245064e-01 -3.90027076e-01 7.04711556e-01
2.82139897e-01 4.23647434e-01 -1.68600991e-01 -2.83045501e-01
3.80744278e-01 -2.52518237e-01 -9.90094692e-02 -1.99187681e-01
-8.58299613e-01 6.37379169e-01 2.89906800e-01 -1.36407137e-01
-2.76919246e-01 1.70686282e-02 3.31610948e-01 6.41578972e-01
3.77114087e-01 5.46193242e-01 -5.20660222e-01 2.37348795e-01
-6.85219109e-01 1.48533165e-01 7.34416842e-01 7.38397300e-01
7.48891771e-01 2.37906903e-01 4.58190851e-02 8.50689292e-01
-2.04771370e-01 8.48939478e-01 2.15918794e-01 -8.10791194e-01
-1.04412511e-01 4.73832726e-01 1.61213782e-02 -9.05131638e-01
-3.35439175e-01 -1.00793160e-01 -9.99292672e-01 7.82854795e-01
8.27663600e-01 -1.70988739e-01 -1.25538862e+00 2.05449867e+00
2.83296317e-01 -6.65767863e-02 1.49294138e-01 7.98518836e-01
6.91708922e-01 1.11799859e-01 2.16366842e-01 -1.86795399e-01
1.28169286e+00 -5.27630448e-02 -6.38380051e-01 -5.79712279e-02
2.26529747e-01 -7.79950142e-01 1.11560464e+00 4.13275272e-01
-8.94063652e-01 -5.02059937e-01 -1.24174929e+00 3.48590374e-01
-7.75719643e-01 -4.69647273e-02 5.80169022e-01 1.20257747e+00
-9.11391497e-01 6.77961648e-01 -3.73867214e-01 -3.09696198e-01
7.87790835e-01 9.38748717e-01 -1.04457068e+00 7.10060075e-02
-1.35367560e+00 7.65127480e-01 1.61534902e-02 1.90530270e-01
-9.38758194e-01 -5.69563508e-01 -8.37828934e-01 -9.64246690e-02
1.51707724e-01 -3.86742413e-01 8.91673386e-01 -1.95873761e+00
-1.41404581e+00 1.26649714e+00 -5.89865483e-02 -3.25235039e-01
6.57021582e-01 3.73343304e-02 -8.56299341e-01 8.48120153e-02
-2.49159098e-01 7.55314529e-01 1.34613991e+00 -1.54149950e+00
-1.32651031e-01 -4.98001039e-01 -5.04813902e-02 -3.64830345e-01
-4.66907233e-01 1.17312498e-01 1.97568566e-01 -8.19256186e-01
-1.03521943e-01 -1.25178659e+00 1.55992001e-01 2.94825852e-01
-5.84155917e-01 1.18886828e-01 8.23806465e-01 -4.09752637e-01
6.14627123e-01 -2.21316576e+00 -3.29702228e-01 6.52575314e-01
1.30487964e-01 5.27793586e-01 -3.68718743e-01 -8.56532156e-02
-5.59342265e-01 4.36682403e-01 -4.33744282e-01 -1.13063321e-01
-2.21937690e-02 3.33473861e-01 -3.16199094e-01 7.48629153e-01
6.62985623e-01 7.86046445e-01 -6.96845055e-01 -2.39157513e-01
-1.59374788e-01 5.24839640e-01 -4.08687681e-01 2.33837903e-01
1.18892267e-02 3.44841510e-01 2.51414776e-02 1.06090999e+00
1.04505575e+00 4.68490124e-01 -8.10592920e-02 -2.83104360e-01
5.12726724e-01 -3.41218799e-01 -9.48013604e-01 8.22997212e-01
-2.44755834e-01 6.78632081e-01 6.99705631e-02 -9.23031330e-01
9.58215475e-01 3.40108693e-01 2.83525854e-01 -5.07749259e-01
2.32597858e-01 4.75641668e-01 2.35940799e-01 -3.28130454e-01
1.93114623e-01 -4.00626957e-01 -6.99583143e-02 3.50422919e-01
1.72785640e-01 1.02246352e-01 -2.26962432e-01 -1.84459060e-01
8.32118154e-01 -7.79570267e-02 8.73652175e-02 -1.89005375e-01
6.73615992e-01 -3.23199213e-01 6.07297182e-01 7.41627216e-01
-3.30629677e-01 8.52085233e-01 5.93461394e-01 -6.60799682e-01
-1.20066798e+00 -1.16625679e+00 -4.10111874e-01 9.69315112e-01
-1.47746578e-01 3.79617549e-02 -1.10114944e+00 -1.07914376e+00
3.00008297e-01 2.72312999e-01 -1.04832327e+00 -5.62309682e-01
-4.71353680e-01 -8.94912183e-01 1.17169046e+00 4.56870556e-01
4.73719627e-01 -1.17307723e+00 -9.77306366e-02 -1.54989079e-01
6.95230216e-02 -1.06628358e+00 -2.12726593e-01 1.02395244e-01
-4.20564741e-01 -1.28556728e+00 -5.14385402e-01 -4.84854251e-01
1.16941118e+00 -4.87594604e-01 1.02476776e+00 3.92646700e-01
-3.52057606e-01 2.54770607e-01 -7.60827735e-02 -7.16079772e-01
-9.05035555e-01 -8.71874988e-02 6.56861305e-01 5.47490120e-01
4.56032276e-01 -6.47626460e-01 -5.70775688e-01 5.46152711e-01
-9.80574250e-01 -4.92114186e-01 5.96738756e-01 1.05015111e+00
3.87452841e-01 -2.40950331e-01 8.02516580e-01 -1.15851676e+00
3.59956175e-01 -2.30529830e-01 -3.92613053e-01 7.88453743e-02
-5.43445408e-01 -2.33912282e-02 1.02890253e+00 -7.36606717e-01
-6.75742924e-01 3.67500275e-01 -3.78105313e-01 -4.17270094e-01
-5.82231581e-01 -4.52433936e-02 -6.12992167e-01 -6.44303083e-01
1.02312267e+00 1.14340924e-01 4.11548138e-01 -5.13890646e-02
1.70316041e-01 7.04169035e-01 8.20859969e-01 -6.55056238e-01
1.11006629e+00 5.69650888e-01 3.26033324e-01 -6.34336352e-01
-5.45367420e-01 2.86815494e-01 -7.44127393e-01 -3.92473489e-01
5.10156214e-01 -5.29839158e-01 -1.13969088e+00 9.29524302e-01
-8.09459865e-01 -1.63697138e-01 -2.32368380e-01 4.94473986e-02
-5.36235094e-01 -8.86421371e-03 -3.28588545e-01 -7.16422081e-01
-1.56825647e-01 -1.09181821e+00 7.81799018e-01 1.35000363e-01
-3.11304331e-01 -6.46863043e-01 -5.09742260e-01 3.93480100e-02
5.44950545e-01 7.27092981e-01 8.91806424e-01 -1.06365299e+00
-2.47040033e-01 -6.12825453e-01 -1.63895879e-02 7.22982943e-01
4.08463031e-01 2.89031535e-01 -1.55010748e+00 -3.51686329e-01
-2.31588542e-01 -4.97859389e-01 6.32048905e-01 -4.50134873e-02
1.45758498e+00 -7.15860128e-01 -8.15746635e-02 9.75232542e-01
1.18117595e+00 1.85623795e-01 9.34660017e-01 2.51058847e-01
5.82188785e-01 7.32907593e-01 2.81203479e-01 9.67028663e-02
-1.55846596e-01 6.14144862e-01 7.39240706e-01 -3.82304817e-01
-3.65351588e-02 -6.03684271e-03 2.96870708e-01 -1.08438000e-01
-3.38870972e-01 -1.41244054e-01 -6.66908622e-01 4.00146663e-01
-1.17721701e+00 -1.00637376e+00 9.38909426e-02 2.58325315e+00
1.03245640e+00 1.07818767e-01 1.49438068e-01 4.50405210e-01
7.01581299e-01 -1.74366251e-01 -6.01964593e-01 -6.46122456e-01
-4.91883814e-01 7.17757821e-01 6.92415953e-01 4.01598215e-01
-1.44685936e+00 8.46273124e-01 6.36092567e+00 5.60681224e-01
-1.38375580e+00 -1.30392015e-01 9.56611454e-01 -1.11282602e-01
-1.51650950e-01 -4.18553054e-01 -3.90153795e-01 4.37019169e-01
8.17868471e-01 9.94991809e-02 4.05260265e-01 7.67314911e-01
-1.77675009e-01 2.22928584e-01 -1.10409355e+00 6.67724967e-01
1.66785941e-01 -9.97010887e-01 2.02105194e-01 6.34710640e-02
6.40089750e-01 -6.50612354e-01 4.93585259e-01 7.56513923e-02
3.17833930e-01 -1.54731846e+00 5.39446950e-01 4.67901468e-01
1.32693851e+00 -1.02149761e+00 9.32272017e-01 -2.45991156e-01
-5.92576504e-01 -1.96168814e-02 -3.17213863e-01 2.12426797e-01
-4.62195337e-01 3.85637105e-01 -6.67971849e-01 2.79736280e-01
6.04409695e-01 7.25206062e-02 -8.09602499e-01 5.11959255e-01
-1.49383664e-01 6.11601591e-01 -4.02782798e-01 2.87755340e-01
-1.99874252e-01 9.30822417e-02 4.65564877e-01 9.74170029e-01
3.77910912e-01 9.91143584e-02 -4.19497430e-01 6.35625124e-01
-4.05636400e-01 -6.06965721e-02 -8.38002205e-01 1.76523864e-01
4.30220813e-01 1.28530443e+00 -3.71273279e-01 -1.69242889e-01
-9.90679637e-02 1.23667300e+00 4.92614172e-02 4.50468659e-01
-6.92087054e-01 -5.37455678e-01 1.12060988e+00 1.15263157e-01
1.98071934e-02 3.10382634e-01 -2.94883579e-01 -8.42291653e-01
1.72701016e-01 -1.28155160e+00 2.48953611e-01 -5.24890423e-01
-1.46515584e+00 9.05418694e-01 -3.35729063e-01 -1.22765064e+00
-3.94276172e-01 -7.69541204e-01 -4.08750594e-01 9.48610604e-01
-1.16585040e+00 -1.39747715e+00 -5.31106889e-01 8.45602632e-01
-2.14268550e-01 -4.02685970e-01 1.27796376e+00 2.08644524e-01
-2.55368590e-01 1.48332775e+00 -1.64979890e-01 4.18939501e-01
1.10129988e+00 -1.10088086e+00 3.46451998e-01 8.19774151e-01
-7.02497810e-02 6.48467481e-01 8.16348553e-01 -4.05867696e-01
-1.09918571e+00 -1.24629855e+00 8.34164143e-01 -5.69774687e-01
2.83780783e-01 -6.81116164e-01 -1.07133114e+00 5.98232090e-01
-8.62853453e-02 6.22284591e-01 8.44268084e-01 -1.80224076e-01
-7.61524737e-01 -4.00608242e-01 -1.74341393e+00 5.71885943e-01
1.08168066e+00 -6.51522577e-01 -1.89891487e-01 2.48062372e-01
2.34008670e-01 -3.37847382e-01 -8.94595265e-01 7.83063352e-01
9.80663776e-01 -8.85438323e-01 1.11311281e+00 -9.66171086e-01
4.35239226e-01 -2.74852425e-01 -7.86768198e-02 -1.31612408e+00
8.46716836e-02 -7.94359803e-01 1.27657205e-01 1.43906975e+00
4.70107555e-01 -7.99743831e-01 9.01265025e-01 7.25975156e-01
1.77366450e-01 -5.54718494e-01 -1.06670475e+00 -9.78315413e-01
2.03782111e-01 -2.28220165e-01 9.84585106e-01 1.23115909e+00
-3.66049647e-01 -4.03166324e-01 -4.57097948e-01 3.35275292e-01
6.04374409e-01 -3.78493994e-01 8.30359578e-01 -1.24941754e+00
1.95291221e-01 -3.84253532e-01 -8.89930367e-01 5.31505235e-03
5.36911488e-01 -6.95268929e-01 3.00835595e-02 -5.03805459e-01
-1.31221667e-01 -7.04155684e-01 -1.91506833e-01 9.10954893e-01
-3.53330344e-01 9.87442136e-01 2.43627792e-03 -1.27080185e-02
2.50972897e-01 2.75180548e-01 1.00523078e+00 -2.34531730e-01
1.97021782e-01 2.83348233e-01 -6.92859292e-01 6.77675545e-01
8.37658942e-01 -5.86584806e-01 -1.19082779e-01 1.44247040e-01
1.17825627e-01 -4.34136033e-01 6.50737464e-01 -9.29314971e-01
-1.14398539e-01 -1.65086221e-02 8.73064339e-01 1.83801875e-01
3.49396914e-01 -1.01719522e+00 2.69291997e-01 2.93188214e-01
-3.00314993e-01 4.63846847e-02 4.43801641e-01 2.06798643e-01
-1.59533992e-01 -4.94317561e-02 1.11923146e+00 1.66635036e-01
-4.28724557e-01 4.57123220e-01 -3.82572085e-01 -1.90819621e-01
1.06340826e+00 -4.67813402e-01 -4.42769319e-01 -4.63972181e-01
-9.29147482e-01 -3.51141065e-01 7.21694767e-01 2.50924140e-01
5.46746016e-01 -1.58134568e+00 -8.25831711e-01 9.08806384e-01
2.74834514e-01 -3.85069698e-01 4.99509741e-03 5.14109552e-01
-4.80951905e-01 -3.08490843e-01 -5.63178062e-01 -4.03856516e-01
-1.59602082e+00 6.65758908e-01 7.18279600e-01 4.38147008e-01
-5.09297736e-02 8.54668498e-01 5.44771031e-02 -5.92374682e-01
8.23072046e-02 2.00122505e-01 8.05219337e-02 9.28111076e-02
4.09341455e-01 -1.16645604e-01 4.02484000e-01 -7.92762756e-01
-4.93642628e-01 5.11660218e-01 -3.72265503e-02 1.71146691e-01
1.09721279e+00 3.21867943e-01 -1.07786827e-01 -1.14155024e-01
1.22759128e+00 1.56263798e-01 -1.09411979e+00 7.68924430e-02
-2.90210664e-01 -6.62266374e-01 -5.39348423e-01 -9.56088543e-01
-1.38936079e+00 7.12225914e-01 9.13915515e-01 3.73561025e-01
1.41918838e+00 -2.24344075e-01 3.49469423e-01 3.14584643e-01
8.97498056e-02 -8.59267175e-01 -2.15478569e-01 1.84755236e-01
9.02279735e-01 -1.26090765e+00 -7.66743794e-02 -5.30083060e-01
-4.55790162e-01 1.16463494e+00 6.83511198e-01 -5.67038655e-02
5.12396574e-01 3.18529129e-01 4.64430004e-01 -4.25367616e-03
-4.10194129e-01 2.23614667e-02 3.47102523e-01 1.03841269e+00
1.85194254e-01 -1.07865399e-02 1.73335038e-02 5.30995727e-01
-6.09661400e-01 -1.49619773e-01 5.84410012e-01 6.44292712e-01
1.74166337e-01 -1.23851585e+00 -5.73527098e-01 5.49443662e-01
-8.35977674e-01 8.42895475e-04 -7.92873681e-01 9.16502297e-01
3.94986033e-01 5.82176089e-01 1.42646194e-01 -7.52438664e-01
3.39476526e-01 3.11753839e-01 6.53701782e-01 -4.57691513e-02
-8.13737094e-01 -5.81761122e-01 1.59908012e-01 -4.77392107e-01
-3.99792880e-01 -6.46644056e-01 -7.55650282e-01 -6.87215507e-01
-2.40110442e-01 -2.48577908e-01 5.43504775e-01 8.60435009e-01
2.74959594e-01 2.56999701e-01 8.71036410e-01 -7.56592095e-01
-5.20338476e-01 -9.01098490e-01 -4.36970025e-01 9.78937447e-01
5.48545837e-01 -6.88173115e-01 -6.70001864e-01 3.50021690e-01] | [12.960293769836426, 1.0183435678482056] |
78849529-2a6c-4867-aaa1-5327088d869b | parametric-representation-for-singing-voice | 2006.04142 | null | https://arxiv.org/abs/2006.04142v1 | https://arxiv.org/pdf/2006.04142v1.pdf | Parametric Representation for Singing Voice Synthesis: a Comparative Evaluation | Various parametric representations have been proposed to model the speech signal. While the performance of such vocoders is well-known in the context of speech processing, their extrapolation to singing voice synthesis might not be straightforward. The goal of this paper is twofold. First, a comparative subjective evaluation is performed across four existing techniques suitable for statistical parametric synthesis: traditional pulse vocoder, Deterministic plus Stochastic Model, Harmonic plus Noise Model and GlottHMM. The behavior of these techniques as a function of the singer type (baritone, counter-tenor and soprano) is studied. Secondly, the artifacts occurring in high-pitched voices are discussed and possible approaches to overcome them are suggested. | ['Daniel Erro', 'Thomas Drugman', 'Onur Babacan', 'Tuomo Raitio', 'Thierry Dutoit'] | 2020-06-07 | null | null | null | null | ['singing-voice-synthesis'] | ['speech'] | [-4.55040112e-02 -2.02358410e-01 1.50368989e-01 6.48471899e-03
-6.71121061e-01 -6.67667449e-01 6.79589212e-01 -2.59407312e-01
-6.35704175e-02 9.31699693e-01 5.39063394e-01 -1.93462700e-01
-4.06457782e-01 -1.52115256e-01 -2.04861034e-02 -8.06556523e-01
2.08336748e-02 1.58897534e-01 2.56228328e-01 -1.97014287e-01
2.49403551e-01 5.87852359e-01 -1.86291897e+00 -1.06534600e-01
5.22950947e-01 5.95557570e-01 2.93733478e-01 1.21971035e+00
6.16052635e-02 3.97528023e-01 -1.39598060e+00 9.93149132e-02
-7.37122744e-02 -8.35225880e-01 -4.74003196e-01 3.01823653e-02
2.48729941e-02 8.66053402e-02 -3.08594465e-01 8.76103401e-01
1.02729309e+00 5.73449314e-01 1.02611804e+00 -6.07801259e-01
-4.78326678e-01 8.04793596e-01 6.04557954e-02 5.94552815e-01
5.04419446e-01 2.46319361e-02 5.69519699e-01 -8.34187806e-01
2.68555373e-01 1.19619143e+00 5.63997507e-01 4.92023438e-01
-1.19099379e+00 -4.32874054e-01 -7.69457579e-01 -5.97094595e-02
-1.48043394e+00 -6.48543537e-01 1.06667042e+00 -3.20116907e-01
8.15950811e-01 6.52106941e-01 4.76128638e-01 9.22971666e-01
8.50959867e-02 2.35018387e-01 1.44330847e+00 -8.03677917e-01
2.54553825e-01 4.62679505e-01 1.47560313e-01 -4.83863428e-02
-1.08089246e-01 5.47334254e-01 -6.62425160e-01 -1.67767435e-01
6.63032651e-01 -8.08395505e-01 -5.76832414e-01 2.91777253e-01
-7.50308156e-01 5.70080280e-01 -1.50286406e-01 1.04514694e+00
-4.30865496e-01 1.25645503e-01 5.05409479e-01 2.56311119e-01
2.68410116e-01 6.29450142e-01 -1.96354613e-01 -4.23149437e-01
-1.45780063e+00 4.45946127e-01 9.56242323e-01 7.47567892e-01
-1.38290480e-01 1.13178670e+00 -2.88533092e-01 1.18362296e+00
3.27768892e-01 3.57302785e-01 8.02379072e-01 -8.26712847e-01
-9.87163335e-02 -5.15166104e-01 2.12621197e-01 -4.97584343e-01
-1.43206120e-01 -2.86577821e-01 -2.25945830e-01 2.27383256e-01
3.66766542e-01 -4.01369125e-01 -8.75413299e-01 1.24881256e+00
1.29710630e-01 1.33223325e-01 2.19991907e-01 6.91136241e-01
1.11671138e+00 1.06959009e+00 -9.11814198e-02 -8.23532760e-01
1.23935699e+00 -7.28686094e-01 -1.52819359e+00 3.61807317e-01
-3.12739015e-01 -1.49222791e+00 1.01950634e+00 6.82847142e-01
-1.23966300e+00 -9.11751390e-01 -9.69957173e-01 1.37319461e-01
-1.29825369e-01 9.53646656e-03 4.86475378e-02 1.25755930e+00
-9.55288529e-01 7.98783541e-01 -3.61707926e-01 -1.37590706e-01
-5.02867997e-01 2.14697197e-01 2.78914809e-01 7.56354809e-01
-1.17535961e+00 8.82241189e-01 3.42338860e-01 1.60177842e-01
-8.79450977e-01 -5.76789260e-01 -5.44265091e-01 -7.56262913e-02
-9.20667735e-05 -1.34264067e-01 1.73422742e+00 -6.26023293e-01
-2.28195047e+00 2.68130690e-01 -3.22247356e-01 -3.19624305e-01
2.30954051e-01 -2.02111885e-01 -1.08789384e+00 6.41551986e-02
-5.32369852e-01 2.28028186e-03 1.22719550e+00 -1.35035908e+00
-2.65442550e-01 1.68612376e-01 -4.86450523e-01 1.63680434e-01
2.57916480e-01 4.38770384e-01 3.76426190e-01 -9.03185129e-01
-1.31675810e-01 -6.06174350e-01 1.12182796e-01 -6.90260947e-01
-4.53694791e-01 -3.41574371e-01 1.04227114e+00 -8.19806695e-01
1.62744689e+00 -2.25983167e+00 3.63933370e-02 1.36460900e-01
-5.89464009e-01 7.07048774e-01 3.88558447e-01 7.97665477e-01
-1.79884434e-01 6.54798001e-02 -1.19962804e-01 -2.13439956e-01
-1.25722000e-02 2.15372369e-01 -4.82235998e-01 1.99599519e-01
-2.31378868e-01 4.40893024e-01 -6.67746902e-01 -4.92158771e-01
4.91790235e-01 8.34955335e-01 -1.21054627e-01 4.18528020e-01
1.86221555e-01 3.24004263e-01 7.13100061e-02 6.82331145e-01
3.59693587e-01 9.17611659e-01 -1.28130242e-01 -3.15583587e-01
-6.88623309e-01 6.96580827e-01 -1.26161265e+00 9.63473499e-01
-5.20620584e-01 6.43531084e-01 3.25553626e-01 -6.37616813e-01
1.25373816e+00 1.32421362e+00 -2.05137371e-03 7.41172135e-02
3.03570300e-01 7.40051329e-01 4.93078470e-01 -6.75789893e-01
7.59605110e-01 -7.14162707e-01 4.13122028e-01 -2.19942793e-01
3.78872871e-01 -9.33692873e-01 2.17932433e-01 -5.18041313e-01
2.92625487e-01 1.88610986e-01 5.15173316e-01 -5.76650918e-01
6.03516519e-01 -3.06908548e-01 2.27076977e-01 2.23850176e-01
-4.04037565e-01 5.82487643e-01 2.70483971e-01 1.98646232e-01
-1.00227046e+00 -1.12595522e+00 -4.20827985e-01 8.15018773e-01
-3.70131493e-01 -1.98548734e-01 -9.02056456e-01 2.20294595e-01
-3.21455419e-01 1.30248702e+00 -1.65564612e-01 4.55940701e-02
-6.22683287e-01 -4.16263700e-01 8.46238315e-01 2.08065093e-01
-5.46900369e-02 -1.52166593e+00 -4.37909484e-01 4.82057065e-01
1.41679987e-01 -6.76126599e-01 -6.49516881e-01 3.74549806e-01
-9.17153001e-01 -4.38325584e-01 -9.11870003e-01 -7.35779941e-01
-9.80178341e-02 -9.20513570e-02 7.37162948e-01 -4.94443029e-01
-1.92582414e-01 5.04341960e-01 -4.23101753e-01 -5.73756874e-01
-9.47470784e-01 -5.09960413e-01 3.24916869e-01 -9.02678370e-02
2.26785868e-01 -9.32141006e-01 -4.50991839e-01 7.99600556e-02
-7.99692333e-01 -8.85946393e-01 1.34370789e-01 7.23789215e-01
5.61262012e-01 1.89085647e-01 7.21771657e-01 -6.06496930e-01
1.23445296e+00 -1.87722147e-01 -2.58783668e-01 -2.76742280e-01
-3.02078843e-01 -3.06977659e-01 8.30003202e-01 -6.44785881e-01
-1.24911678e+00 -1.71444550e-01 -5.26307762e-01 -5.94532132e-01
-3.23627740e-01 2.99303651e-01 -1.00660548e-02 -3.33697796e-02
9.63899672e-01 4.01083589e-01 -1.52965412e-01 -7.73846269e-01
1.85679719e-01 1.01663268e+00 8.83434832e-01 -3.98001909e-01
9.48826373e-01 -2.54383743e-01 -2.48907045e-01 -1.68810987e+00
-2.82277197e-01 -5.81059039e-01 -3.69973481e-01 -4.18454885e-01
7.65726268e-01 -2.96553254e-01 -2.71264464e-01 2.29640856e-01
-1.22803462e+00 -4.66938168e-02 -6.82799578e-01 9.84172225e-01
-8.10964584e-01 4.30494428e-01 -5.96455812e-01 -1.56451440e+00
-5.75196922e-01 -1.10300052e+00 5.10887444e-01 4.86285001e-01
-5.05668163e-01 -9.87393618e-01 2.24336177e-01 2.28922710e-01
5.48691392e-01 1.74752206e-01 7.22092092e-01 -4.55269784e-01
2.84872890e-01 -1.35937437e-01 7.38586426e-01 8.38593602e-01
3.59685183e-01 4.01728094e-01 -1.31784642e+00 1.59718934e-02
5.04529119e-01 -2.00146094e-01 3.87521803e-01 7.30949700e-01
7.75882006e-01 -3.12276334e-01 2.40668297e-01 2.86975950e-01
1.23363233e+00 8.59971106e-01 7.18444467e-01 -3.01871210e-01
-2.83925273e-02 4.77830052e-01 5.55901527e-01 2.82758087e-01
-5.20807743e-01 5.64602077e-01 -1.23786014e-02 2.39397630e-01
-5.89555681e-01 -2.75710344e-01 3.93829137e-01 1.43507075e+00
-5.61352193e-01 -4.40907836e-01 -2.42716655e-01 6.91979289e-01
-8.99208724e-01 -1.02196681e+00 -3.46708149e-01 2.42713571e+00
8.11158478e-01 1.52031090e-02 3.93149883e-01 1.02069676e+00
8.81557822e-01 3.73527318e-01 1.16516083e-01 -1.27059245e+00
-1.57203302e-01 7.23536491e-01 3.18272471e-01 7.89596915e-01
-7.46849000e-01 6.34786069e-01 7.40561104e+00 9.55854297e-01
-1.23518813e+00 2.33914420e-01 1.62506782e-04 9.35090110e-02
-2.31169999e-01 -6.43106252e-02 -6.71120584e-01 4.08245772e-01
1.37286401e+00 -4.00358677e-01 6.37844205e-01 5.60701311e-01
6.32759213e-01 -5.79150096e-02 -6.63283348e-01 8.63071442e-01
-2.28028391e-02 -7.44191766e-01 -3.69039848e-02 -2.38779083e-01
6.80956006e-01 -5.11816263e-01 2.25835577e-01 1.89363018e-01
-3.93159688e-01 -1.08027172e+00 8.94704938e-01 4.90472823e-01
8.25900733e-01 -9.08746898e-01 5.06239951e-01 2.88460910e-01
-1.17025149e+00 2.73208469e-01 -2.97410846e-01 5.34103177e-02
5.26731551e-01 3.22257072e-01 -9.61654782e-01 5.22641957e-01
5.51833324e-02 -2.35311136e-01 4.77650836e-02 1.35839164e+00
-2.43059069e-01 1.36405575e+00 -3.40182036e-01 -3.98004651e-01
1.75603956e-01 -5.27388938e-02 1.14091218e+00 1.37320352e+00
6.33856952e-01 8.01234841e-02 -3.61351877e-01 8.08705449e-01
4.58024144e-01 4.85072911e-01 -4.56828624e-01 -4.79194432e-01
6.71402335e-01 8.76539648e-01 -2.11224094e-01 -1.46029070e-01
-1.74809143e-01 6.44170702e-01 -7.66652942e-01 5.30780792e-01
-6.79308295e-01 -7.12765217e-01 3.69621128e-01 2.69034535e-01
2.19026908e-01 -2.72996336e-01 -1.84225291e-02 -3.86489630e-01
-3.70090246e-01 -8.70369852e-01 -1.26469925e-01 -6.76336527e-01
-7.58270085e-01 8.10050368e-01 3.08476776e-01 -1.18708658e+00
-5.76672554e-01 -4.73794699e-01 -8.93969476e-01 1.28043246e+00
-9.79195476e-01 -5.22074401e-01 2.70506889e-01 2.89188355e-01
1.06314611e+00 -3.85896266e-01 9.71591353e-01 2.80309260e-01
-7.18085915e-02 4.33289587e-01 2.30888933e-01 -5.10321736e-01
4.43898737e-01 -1.26793694e+00 1.15251066e-02 6.00152493e-01
2.87703723e-01 6.38064384e-01 1.53243756e+00 -2.64478475e-01
-1.12141907e+00 -5.92383325e-01 1.08389533e+00 6.79097176e-02
5.64209521e-01 3.11757289e-02 -9.83508587e-01 -3.68108810e-03
5.07478893e-01 -2.67835170e-01 6.16474748e-01 -3.34778845e-01
7.58465976e-02 2.85410043e-02 -1.07925701e+00 4.21098709e-01
2.82125920e-01 -6.57995462e-01 -9.67199028e-01 1.47159576e-01
6.05566442e-01 -2.88446963e-01 -1.03849256e+00 2.48962626e-01
3.56736839e-01 -1.02630949e+00 7.37247705e-01 -1.24842905e-01
6.12900686e-03 -5.01867950e-01 -1.13650210e-01 -1.60679317e+00
-1.11377157e-01 -1.34452164e+00 -5.89636005e-02 1.49663305e+00
4.50055718e-01 -5.19620180e-01 2.96226114e-01 -1.32801622e-01
-5.04714429e-01 -3.56971264e-01 -1.15511608e+00 -1.08268225e+00
1.00482246e-02 -3.40586454e-01 1.35001644e-01 4.90366846e-01
1.14461184e-01 5.19250691e-01 -7.98123658e-01 -2.26770714e-02
2.96189547e-01 -3.10930789e-01 4.22569782e-01 -9.58328545e-01
-7.15252876e-01 -4.47345287e-01 -2.45013356e-01 -6.79360151e-01
-3.15973699e-01 -3.87075126e-01 3.27747822e-01 -1.32502544e+00
-7.08871424e-01 5.72668649e-02 -2.95791566e-01 -3.14074010e-01
-1.92366153e-01 3.54892947e-02 2.09513098e-01 -1.48517445e-01
7.16577828e-01 4.95377451e-01 1.42212212e+00 1.96996033e-01
-6.24709785e-01 6.18360579e-01 -2.24086959e-02 6.42832756e-01
8.83181870e-01 -4.04470623e-01 -7.34410942e-01 2.89848477e-01
-6.43299818e-01 5.80242753e-01 6.78713694e-02 -1.16273940e+00
3.14478837e-02 1.32069856e-01 -1.99227303e-01 -7.94425428e-01
8.22100878e-01 -5.17412305e-01 5.57575285e-01 3.94828290e-01
-2.40618065e-01 -5.42642809e-02 2.16766015e-01 4.09125328e-01
-6.17720068e-01 -8.98576200e-01 1.17180586e+00 -8.46367329e-04
-3.08448583e-01 -3.63846362e-01 -6.32491410e-01 -1.55829594e-01
6.94782674e-01 -5.55845678e-01 1.57615423e-01 -6.02478921e-01
-8.66622269e-01 -6.17876470e-01 -1.83127224e-01 1.68317050e-01
4.69211161e-01 -1.01007104e+00 -5.75015545e-01 -1.07703879e-01
-4.60038841e-01 -5.73079228e-01 3.47495824e-01 7.97976077e-01
-6.22387528e-01 6.54425204e-01 -4.20585871e-02 -8.07347670e-02
-1.49971199e+00 5.49874604e-01 6.33884132e-01 2.75928944e-01
-2.29036704e-01 8.18302751e-01 -1.01231553e-01 -4.94895168e-02
3.84475112e-01 -2.21121624e-01 -6.84570074e-01 -3.12975347e-02
3.40340376e-01 7.63731599e-01 -6.55871704e-02 -8.55453849e-01
-8.61269459e-02 4.40820694e-01 5.09181976e-01 -6.09151959e-01
6.56083047e-01 7.23662302e-02 -2.47797891e-02 1.17953789e+00
8.58459949e-01 5.55478752e-01 -6.42990112e-01 2.09693730e-01
-7.41804689e-02 -3.22280049e-01 1.40353933e-01 -6.81589961e-01
-4.42213178e-01 1.04021740e+00 5.32207489e-01 5.63088059e-01
1.19358027e+00 -3.70220542e-01 7.28715837e-01 -6.57545552e-02
3.34281512e-02 -1.17185247e+00 -2.52178550e-01 4.35404360e-01
1.21556687e+00 -3.52325678e-01 -1.35921717e-01 -6.09811902e-01
-5.65191627e-01 1.32884467e+00 8.69920626e-02 -4.56937373e-01
7.42964745e-01 3.26108575e-01 1.79666638e-01 2.65948236e-01
-6.68423533e-01 -2.36052543e-01 4.59438741e-01 8.47295761e-01
1.15124488e+00 1.90215662e-01 -1.12628567e+00 3.70952755e-01
-7.81655848e-01 -1.92194685e-01 6.69068456e-01 6.24100983e-01
-8.07201624e-01 -1.12436795e+00 -9.04860139e-01 1.73197806e-01
-9.14877355e-01 -6.84657469e-02 -3.93087387e-01 8.06275427e-01
4.02705409e-02 1.30823410e+00 -2.44413123e-01 -3.18273455e-01
7.17355847e-01 5.46747625e-01 4.95447427e-01 -6.63774610e-01
-9.27108765e-01 1.03517842e+00 3.72574508e-01 2.11733982e-01
-3.54959190e-01 -6.83149457e-01 -1.06591773e+00 1.23117216e-01
-4.79139835e-01 7.62975395e-01 8.58158231e-01 6.04655027e-01
-5.79604566e-01 8.30990255e-01 6.31526411e-01 -7.49643922e-01
-8.56233597e-01 -1.51993430e+00 -1.11512661e+00 -3.02448068e-02
3.13773662e-01 -5.52605927e-01 -7.99390078e-01 1.71646774e-01] | [15.016693115234375, 6.0463995933532715] |
695f5f18-e319-4345-8855-69671cd4e3cf | clr-gam-contrastive-point-cloud-learning-with | 2302.14306 | null | https://arxiv.org/abs/2302.14306v1 | https://arxiv.org/pdf/2302.14306v1.pdf | CLR-GAM: Contrastive Point Cloud Learning with Guided Augmentation and Feature Mapping | Point cloud data plays an essential role in robotics and self-driving applications. Yet, annotating point cloud data is time-consuming and nontrivial while they enable learning discriminative 3D representations that empower downstream tasks, such as classification and segmentation. Recently, contrastive learning-based frameworks have shown promising results for learning 3D representations in a self-supervised manner. However, existing contrastive learning methods cannot precisely encode and associate structural features and search the higher dimensional augmentation space efficiently. In this paper, we present CLR-GAM, a novel contrastive learning-based framework with Guided Augmentation (GA) for efficient dynamic exploration strategy and Guided Feature Mapping (GFM) for similar structural feature association between augmented point clouds. We empirically demonstrate that the proposed approach achieves state-of-the-art performance on both simulated and real-world 3D point cloud datasets for three different downstream tasks, i.e., 3D point cloud classification, few-shot learning, and object part segmentation. | ['Yi-Ting Chen', 'Srikanth Malla'] | 2023-02-28 | null | null | null | null | ['3d-point-cloud-classification', 'point-cloud-classification'] | ['computer-vision', 'computer-vision'] | [ 1.33878469e-01 -9.90853980e-02 -4.42131132e-01 -3.56689602e-01
-8.05841982e-01 -4.49520439e-01 6.07761502e-01 3.24126214e-01
-1.25996381e-01 1.44013301e-01 -4.47414905e-01 -2.93260783e-01
-3.24501425e-01 -7.00075388e-01 -8.30738187e-01 -5.89518964e-01
-3.04004908e-01 9.50522542e-01 3.19272786e-01 -1.03603154e-01
5.52573442e-01 1.13388336e+00 -2.20497561e+00 -2.61203080e-01
9.76010561e-01 1.04756367e+00 6.54338419e-01 2.87621439e-01
-7.54619658e-01 2.94589996e-03 -2.65862681e-02 1.74963877e-01
6.15864873e-01 2.95544356e-01 -6.72476232e-01 3.95272344e-01
3.82469684e-01 1.12577312e-01 -1.93876833e-01 9.29265916e-01
2.48884961e-01 4.73387986e-01 8.66846085e-01 -1.66730225e+00
-3.33122522e-01 -1.08825071e-02 -7.67807066e-01 1.04388513e-01
-1.05257826e-02 2.03906655e-01 6.85836613e-01 -1.20406783e+00
5.04972219e-01 1.15354085e+00 5.72293282e-01 4.26972389e-01
-9.95068669e-01 -6.71626925e-01 2.82176912e-01 1.42990932e-01
-1.27253556e+00 -1.74700022e-02 1.10119128e+00 -5.49081564e-01
1.06170201e+00 8.41031298e-02 1.12997675e+00 6.14128947e-01
-5.39431609e-02 1.00515223e+00 8.58092189e-01 -2.93682247e-01
5.70508480e-01 -5.77338226e-02 1.73845217e-01 8.69961858e-01
2.10319445e-01 1.68798655e-01 -6.05015099e-01 -2.11540431e-01
9.80441451e-01 4.68301207e-01 2.53780812e-01 -1.18376195e+00
-1.09522080e+00 8.81977081e-01 5.45840740e-01 -1.13622975e-02
-4.87516671e-01 2.34680027e-01 2.97533542e-01 2.19320789e-01
6.92066431e-01 5.54679215e-01 -7.21939564e-01 -2.84450501e-01
-6.60442173e-01 3.73713851e-01 3.58861715e-01 1.58610654e+00
1.31630754e+00 -1.34286238e-02 5.65821864e-02 7.89897621e-01
4.89460677e-01 5.45201063e-01 2.79541165e-01 -9.46490407e-01
2.80597031e-01 1.03303134e+00 -1.87705886e-02 -6.33124709e-01
-3.63317937e-01 -3.04402798e-01 -7.45733798e-01 6.18507028e-01
-1.31194413e-01 2.78997034e-01 -1.51338434e+00 1.29571283e+00
6.95513248e-01 8.15768123e-01 6.28191754e-02 9.46645439e-01
9.63183701e-01 5.29697061e-01 1.25474513e-01 2.50982791e-02
1.00188112e+00 -8.97069633e-01 -1.59511447e-01 -3.67082387e-01
7.24317849e-01 -4.41128492e-01 1.19741726e+00 -3.90048213e-02
-9.58946884e-01 -6.52422726e-01 -1.09487283e+00 -1.52370974e-01
-4.79782492e-01 -2.05804721e-01 1.15648222e+00 3.04453492e-01
-5.88283360e-01 7.20640123e-01 -1.13283014e+00 -2.87513375e-01
1.03341818e+00 5.47586679e-01 -3.85473400e-01 -2.42964238e-01
-3.68738294e-01 6.81609511e-01 2.90371358e-01 -1.98461905e-01
-9.11218524e-01 -9.82192755e-01 -1.06861806e+00 -2.04264447e-01
4.30656791e-01 -7.00068474e-01 1.09473360e+00 -9.01319608e-02
-1.34721494e+00 1.12544894e+00 -1.47623017e-01 -3.73193204e-01
1.26787603e-01 -5.22509873e-01 1.76633865e-01 -1.23332515e-01
2.87352681e-01 1.07336760e+00 8.95592332e-01 -1.43352044e+00
-7.35613883e-01 -7.52672136e-01 -1.36053577e-01 4.56980020e-01
-4.75674011e-02 -5.53953469e-01 -6.15156591e-01 -2.88968265e-01
1.04172611e+00 -1.03468573e+00 -7.31739461e-01 4.13151115e-01
-1.78928033e-01 -6.21940672e-01 1.30383313e+00 1.65312693e-01
3.42405319e-01 -2.14869380e+00 8.22644532e-02 2.70241469e-01
8.43069553e-02 1.67470917e-01 -1.30390227e-01 3.85867469e-02
1.82500720e-01 -3.47864740e-02 -4.12970781e-01 -5.11584520e-01
4.01761048e-02 6.30831480e-01 -3.45562994e-01 5.71375310e-01
4.49251562e-01 1.20750570e+00 -1.11047268e+00 -6.23270094e-01
8.36230993e-01 3.14136267e-01 -4.95355338e-01 2.69603819e-01
-4.58728731e-01 5.31142116e-01 -6.99993014e-01 1.10612202e+00
8.09265316e-01 -2.12192014e-01 -5.35360217e-01 6.33433810e-04
-1.17227845e-01 9.21058876e-04 -1.08189034e+00 2.51075006e+00
-4.82358605e-01 4.31881905e-01 -2.83083707e-01 -1.01023257e+00
1.39069617e+00 -2.11034030e-01 8.48826826e-01 -2.90259957e-01
8.77812654e-02 2.74678707e-01 -4.11092520e-01 -4.07338500e-01
7.08341002e-01 1.57227188e-01 -1.21233247e-01 3.35029304e-01
3.09132457e-01 -1.02603340e+00 -3.13371241e-01 -9.37852357e-03
8.40945542e-01 5.12309790e-01 1.63815677e-01 -2.07368970e-01
2.24334717e-01 4.87663537e-01 3.20184857e-01 6.92543745e-01
-1.08875819e-01 4.55400616e-01 -1.40096679e-01 -4.91316408e-01
-1.01233065e+00 -1.21063352e+00 -5.80562353e-02 8.70062292e-01
8.17762971e-01 -1.69763327e-01 -1.36894077e-01 -6.37875259e-01
5.17570078e-01 7.01851487e-01 -3.84053916e-01 -2.93623716e-01
-4.51707721e-01 -2.33469188e-01 -3.48680504e-02 7.20307529e-01
2.60529190e-01 -8.77266586e-01 -9.08203602e-01 8.42953399e-02
3.36532712e-01 -1.00211549e+00 5.94097711e-02 5.96543252e-01
-1.50092888e+00 -1.00968838e+00 -3.67936641e-01 -1.00305223e+00
8.34869385e-01 8.72673690e-01 1.12081146e+00 -6.39563501e-02
-5.47126412e-01 6.11726165e-01 -4.41992819e-01 -8.65029633e-01
9.25792530e-02 2.52910018e-01 3.08507215e-02 -4.92761284e-01
7.51146197e-01 -8.74566853e-01 -3.87753844e-01 2.45849162e-01
-5.87835252e-01 3.11348066e-02 7.00158358e-01 6.87874436e-01
1.35385633e+00 -3.61195475e-01 3.58208477e-01 -6.01193368e-01
1.54664293e-01 -5.35732567e-01 -7.63262093e-01 -5.32834791e-02
-4.99110222e-01 9.12089348e-02 -1.06548354e-01 -3.79359096e-01
-5.91920316e-01 5.99797070e-01 -4.73637171e-02 -1.30161476e+00
-4.43425775e-01 2.04561621e-01 -5.09780571e-02 -4.55990970e-01
5.66602767e-01 2.53924251e-01 1.88492715e-01 -6.48968518e-01
6.41166389e-01 6.17758453e-01 6.74292803e-01 -6.48991942e-01
1.21559393e+00 5.72920918e-01 4.29669291e-01 -8.04738402e-01
-7.29443550e-01 -1.02980340e+00 -1.12034142e+00 -2.03077957e-01
6.68520808e-01 -1.00005126e+00 -5.37005246e-01 2.30183497e-01
-1.00243962e+00 -1.31022558e-01 -7.14565337e-01 4.45413977e-01
-1.14490259e+00 2.72007614e-01 5.90898991e-02 -9.77287948e-01
-3.69208306e-01 -1.13183630e+00 1.58480358e+00 2.54063010e-01
1.79344788e-01 -5.90222180e-01 -2.35786662e-02 9.02595446e-02
8.22866615e-03 4.26964074e-01 1.00250947e+00 -5.74612141e-01
-9.31645393e-01 -2.76815087e-01 -1.11796625e-01 -7.91022554e-02
2.48777479e-01 -3.33413929e-01 -9.62318122e-01 -1.82432696e-01
-2.55166411e-01 -4.92949784e-01 6.62629247e-01 2.34582022e-01
1.48345304e+00 3.48844290e-01 -6.24765396e-01 8.78427148e-01
1.30801928e+00 -7.69738527e-03 3.15370768e-01 1.93483904e-01
7.73560643e-01 3.95140529e-01 1.35182714e+00 5.65892220e-01
2.16000661e-01 5.07313430e-01 8.90150130e-01 1.05690649e-02
1.42654166e-01 -4.30777162e-01 -4.46822435e-01 5.21278977e-01
-2.84530763e-02 2.14075238e-01 -1.19666910e+00 7.51296759e-01
-2.13824010e+00 -3.86519969e-01 -5.71388677e-02 1.93886673e+00
2.87258238e-01 1.59421355e-01 -1.41476691e-01 1.32232472e-01
5.60124099e-01 -5.98482825e-02 -1.05389774e+00 6.66954294e-02
1.62306979e-01 4.53572303e-01 4.78782386e-01 -1.45954276e-02
-1.36465061e+00 1.30703783e+00 5.54592466e+00 6.74390614e-01
-9.20852423e-01 1.13598436e-01 2.26739064e-01 -4.58564050e-02
-2.45345548e-01 -1.95024937e-01 -6.55866444e-01 7.93447867e-02
3.15295339e-01 -2.13919461e-01 1.14236332e-01 1.55861211e+00
-1.95505753e-01 7.79508501e-02 -1.22233796e+00 1.57507849e+00
-1.58479840e-01 -1.75422895e+00 6.56562448e-02 5.75974621e-02
8.44433069e-01 4.56373930e-01 7.08801597e-02 4.30572659e-01
2.13745445e-01 -8.31862688e-01 5.58533549e-01 3.32317293e-01
7.61977613e-01 -9.49497700e-01 5.49241126e-01 5.03338933e-01
-1.40690947e+00 -1.03421956e-01 -6.20263577e-01 7.80509859e-02
2.06707984e-01 4.82855827e-01 -8.68958950e-01 5.98082662e-01
8.81415725e-01 9.30532813e-01 -3.27962279e-01 1.39925575e+00
-2.94439029e-02 1.47827297e-01 -5.02049506e-01 -6.00904226e-02
5.78557372e-01 -2.16387615e-01 8.71656716e-01 6.76789045e-01
5.16296923e-01 2.66199887e-01 4.88387406e-01 9.18226719e-01
1.00882173e-01 -6.63988665e-02 -1.01646459e+00 -9.41331834e-02
7.85860598e-01 1.17245412e+00 -9.47685957e-01 -1.85183734e-01
-2.94567198e-01 6.29317760e-01 3.37046623e-01 -3.02301124e-02
-5.34096003e-01 -2.11225465e-01 1.17590165e+00 -7.13743642e-02
4.27533418e-01 -7.92899132e-01 -6.26111329e-01 -7.44361579e-01
-9.45109576e-02 -1.24885224e-01 7.53090084e-02 -6.30223930e-01
-1.32145607e+00 2.96848327e-01 2.22839206e-01 -1.71575141e+00
-2.89930940e-01 -5.40701270e-01 -4.27372485e-01 7.11679459e-01
-1.80984879e+00 -1.25724518e+00 -7.64573574e-01 6.60423756e-01
9.05791402e-01 -3.13453168e-01 7.33698845e-01 -1.53264955e-01
4.37569199e-03 1.98703557e-01 -7.16509148e-02 -2.81123579e-01
8.62850621e-02 -1.16669595e+00 8.29442143e-01 4.59858537e-01
3.69212389e-01 2.14617163e-01 3.80029410e-01 -7.31472254e-01
-2.00233030e+00 -1.57247353e+00 7.67779201e-02 -3.04513991e-01
1.76958367e-01 -5.10347664e-01 -9.76862073e-01 4.44344252e-01
-5.36605060e-01 4.68010455e-01 5.73469162e-01 7.36288503e-02
-1.29043683e-01 1.82134196e-01 -1.30611217e+00 3.71944189e-01
1.77154791e+00 -3.78374070e-01 -6.36273205e-01 5.96676826e-01
1.09417093e+00 -7.10110009e-01 -7.49045670e-01 8.33716393e-01
1.74712673e-01 -5.04157841e-01 1.21508682e+00 -7.76827276e-01
1.71065271e-01 -2.92049825e-01 -4.27781671e-01 -1.11627638e+00
-3.21763933e-01 -4.44706142e-01 -5.66499591e-01 7.55762935e-01
6.73015937e-02 -4.35118645e-01 1.45634985e+00 3.13549489e-01
-8.37190866e-01 -1.11901224e+00 -1.12135553e+00 -9.96542096e-01
-2.24340960e-01 -7.95598030e-01 8.70747864e-01 8.34278047e-01
-4.25503403e-01 6.55090883e-02 1.99736819e-01 4.64433998e-01
8.42694283e-01 6.10532224e-01 1.11298215e+00 -1.77605498e+00
2.30919302e-01 -3.15697461e-01 -1.00033343e+00 -1.27950716e+00
3.97192746e-01 -1.04112947e+00 2.98810750e-01 -1.63433123e+00
-1.34668604e-01 -1.17044568e+00 -2.02403337e-01 5.39869308e-01
2.80498508e-02 1.02503315e-01 1.67045474e-01 3.71497452e-01
-5.48072636e-01 8.09540093e-01 1.11264896e+00 -3.04786384e-01
-6.91359878e-01 1.99025497e-01 -3.83220047e-01 6.34832084e-01
7.32698083e-01 -3.05934995e-01 -6.70369267e-01 -5.07082939e-01
-2.62936085e-01 -1.95617899e-01 4.37123507e-01 -1.12797892e+00
2.68788010e-01 -3.37558240e-01 3.34838599e-01 -1.39342284e+00
8.46811295e-01 -9.12975192e-01 -1.53317422e-01 3.62015545e-01
5.53353168e-02 -1.48373753e-01 4.80511546e-01 8.64326954e-01
3.98893692e-02 -2.07841992e-01 6.07627153e-01 -3.35688055e-01
-1.38568342e+00 9.49734807e-01 1.28620952e-01 -2.34686881e-01
1.27211285e+00 -6.40489817e-01 1.61972001e-01 1.66684449e-01
-7.01307237e-01 4.51816261e-01 5.83733857e-01 7.13531494e-01
1.20122957e+00 -1.37211812e+00 -3.95045668e-01 5.35469770e-01
5.09557962e-01 8.48183155e-01 2.51228720e-01 4.20042574e-01
-2.90505290e-01 2.82354414e-01 -3.70750010e-01 -1.48692679e+00
-1.12636054e+00 5.77048600e-01 -1.26106739e-01 1.59140348e-01
-1.05238473e+00 1.07025290e+00 -1.19378403e-01 -7.55282462e-01
4.06300008e-01 -4.10458386e-01 -1.24758936e-01 -1.95030943e-01
5.39162420e-02 3.66694540e-01 1.61887184e-01 -3.92768741e-01
-4.18053478e-01 9.49018955e-01 -7.96293002e-03 2.24088624e-01
1.56101334e+00 5.07674739e-02 2.01268658e-01 6.08661950e-01
1.08992422e+00 -9.14282739e-01 -1.40965521e+00 -4.88230020e-01
1.37423754e-01 -9.20432866e-01 2.52932519e-01 -2.43437171e-01
-8.40396345e-01 8.76580298e-01 9.40002084e-01 -1.47504061e-01
7.75184453e-01 5.54034293e-01 6.48360074e-01 7.59555459e-01
1.02750826e+00 -8.50426912e-01 2.33065665e-01 5.40546179e-01
7.99579263e-01 -1.50007606e+00 6.03587627e-02 -8.40645254e-01
-3.75879139e-01 8.27604115e-01 8.38926136e-01 -3.25949550e-01
8.24857414e-01 3.31004546e-03 -1.31176487e-01 -5.76040268e-01
-5.68757653e-01 -5.29867113e-01 2.78498679e-01 1.05409038e+00
-1.99277669e-01 -4.43812050e-02 3.57966602e-01 1.68805867e-01
-2.53869444e-01 -2.51909852e-01 -1.60537718e-03 1.50586486e+00
-7.20751405e-01 -8.56061339e-01 -2.31817544e-01 7.36965179e-01
5.30978262e-01 4.12295282e-01 -1.35971859e-01 8.38840067e-01
1.23365872e-01 5.71483970e-01 5.80215633e-01 -5.19061208e-01
2.98623174e-01 4.44837660e-02 6.64209902e-01 -1.05068421e+00
-8.13328996e-02 -2.54281461e-01 -4.26932335e-01 -6.70046866e-01
-5.07873058e-01 -7.55417109e-01 -1.58507073e+00 1.42089665e-01
-4.77602214e-01 1.08637229e-01 1.14397478e+00 9.61389303e-01
7.64346540e-01 3.63512039e-01 7.76423156e-01 -1.64670241e+00
-3.59927177e-01 -7.36113489e-01 -4.57683384e-01 2.56061792e-01
1.97586879e-01 -1.29584241e+00 3.43747344e-03 -3.27340245e-01] | [8.005725860595703, -3.247001886367798] |
a7954252-f700-4027-90a5-7a17e62bcfd4 | youtube-asl-a-large-scale-open-domain | 2306.15162 | null | https://arxiv.org/abs/2306.15162v1 | https://arxiv.org/pdf/2306.15162v1.pdf | YouTube-ASL: A Large-Scale, Open-Domain American Sign Language-English Parallel Corpus | Machine learning for sign languages is bottlenecked by data. In this paper, we present YouTube-ASL, a large-scale, open-domain corpus of American Sign Language (ASL) videos and accompanying English captions drawn from YouTube. With ~1000 hours of videos and >2500 unique signers, YouTube-ASL is ~3x as large and has ~10x as many unique signers as the largest prior ASL dataset. We train baseline models for ASL to English translation on YouTube-ASL and evaluate them on How2Sign, where we achieve a new finetuned state of the art of 12.39 BLEU and, for the first time, report zero-shot results. | ['Manfred Georg', 'Garrett Tanzer', 'David Uthus'] | 2023-06-27 | null | null | null | null | ['sign-language-recognition'] | ['computer-vision'] | [-1.48547903e-01 -2.19716296e-01 -6.75560534e-01 -3.91400456e-01
-1.21509659e+00 -9.79005814e-01 5.63205183e-01 -9.59524333e-01
-8.54786992e-01 6.76065803e-01 8.29897344e-01 -1.97762311e-01
4.67033088e-01 6.22362196e-02 -8.35631728e-01 -1.32110015e-01
6.77514970e-02 3.45826417e-01 3.34791243e-01 -5.92905357e-02
8.20788220e-02 1.41173497e-01 -1.33417988e+00 5.16025364e-01
5.99953830e-01 5.86654961e-01 -2.45058522e-01 1.06628025e+00
-1.33501723e-01 9.32772934e-01 -3.97914529e-01 -7.44092226e-01
4.35308129e-01 -5.34182429e-01 -6.47270501e-01 -3.83094370e-01
1.49985898e+00 -8.94695461e-01 -8.52565825e-01 8.14458489e-01
8.25198054e-01 -3.83248597e-01 4.70810384e-01 -1.46281374e+00
-1.03548419e+00 5.77917635e-01 -1.12982966e-01 1.12090200e-01
5.51642895e-01 7.66397119e-01 1.16660535e+00 -9.01250899e-01
1.40684021e+00 1.10930872e+00 4.79933977e-01 1.25930607e+00
-5.40350616e-01 -9.80849802e-01 1.67858228e-02 2.72213425e-02
-9.15113032e-01 -7.67084658e-01 1.71538666e-01 -4.82613117e-01
1.00798750e+00 -1.63609199e-02 9.27245557e-01 1.68138528e+00
-3.96845400e-01 1.48093116e+00 9.79920268e-01 -2.41043046e-01
-1.51160985e-01 -6.78749025e-01 8.27236474e-03 8.30296755e-01
2.19667003e-01 -8.64904150e-02 -1.00285470e+00 8.20619613e-02
8.06734741e-01 -4.84613866e-01 -2.53453761e-01 -1.48103207e-01
-1.53330207e+00 4.83674824e-01 2.89219320e-01 -1.21206984e-01
-5.16888313e-03 8.46530139e-01 8.67444873e-01 6.61632717e-01
-5.09272143e-02 3.15999597e-01 -3.91211122e-01 -1.08507001e+00
-6.12631440e-01 2.60993570e-01 8.51412654e-01 1.55916262e+00
-1.21019691e-01 1.17517404e-01 -1.07053705e-01 5.88105142e-01
3.15873146e-01 1.27109456e+00 6.43820763e-01 -1.11974442e+00
8.14059675e-01 2.19887242e-01 2.15706229e-01 4.75733131e-02
1.20410815e-01 3.38219374e-01 -1.91497907e-01 1.33651316e-01
7.30673015e-01 -3.10354501e-01 -1.48253071e+00 1.68198216e+00
-2.63119668e-01 1.79328308e-01 -1.10240862e-01 1.09820998e+00
1.06903744e+00 2.72216111e-01 1.69882149e-01 3.62195224e-01
1.02031827e+00 -1.32580590e+00 -4.13266718e-01 -1.02554247e-01
8.43857706e-01 -8.82555544e-01 1.47645736e+00 2.94475883e-01
-9.05937016e-01 -1.83672264e-01 -6.47529781e-01 -3.51585716e-01
-6.05016127e-02 3.09703171e-01 5.83645821e-01 4.96790767e-01
-9.88885283e-01 2.96970278e-01 -1.03697729e+00 -7.99710274e-01
6.46903694e-01 -3.92136686e-02 -5.45785904e-01 -2.44207308e-01
-7.73579538e-01 7.66887486e-01 -1.84170067e-01 -2.45037287e-01
-6.34449184e-01 -6.80810511e-01 -8.03693831e-01 -5.31009912e-01
4.77149896e-02 -1.99853569e-01 1.78010821e+00 -9.52314138e-01
-1.66720414e+00 1.27142620e+00 -1.96091339e-01 -3.20540965e-01
1.10436833e+00 -5.77275872e-01 -6.56441987e-01 1.36108473e-01
9.69793946e-02 9.03284192e-01 5.95728934e-01 -6.77704751e-01
-6.80158079e-01 1.84008107e-01 -2.34716117e-01 -7.96973333e-02
-1.58859342e-01 5.70858181e-01 -7.90845633e-01 -6.62768304e-01
-2.53258944e-01 -1.19669366e+00 1.74355313e-01 7.28072703e-01
4.09088954e-02 -1.36412382e-01 7.36560524e-01 -9.09756899e-01
1.18069518e+00 -2.25771618e+00 -1.44099295e-01 -6.10581450e-02
1.15658782e-01 5.93833923e-01 -9.20902193e-01 3.83074552e-01
4.60086763e-01 1.55119225e-01 -2.31541805e-02 -1.86352819e-01
2.97846794e-01 3.82904857e-01 -2.57281899e-01 3.69328350e-01
-2.67861690e-02 1.30734289e+00 -1.26072371e+00 -5.00426829e-01
9.74360406e-02 3.37843508e-01 -7.36221611e-01 2.45035514e-02
-1.03109777e-01 2.20078230e-01 -2.58082032e-01 1.06487191e+00
4.21192884e-01 -2.09578350e-01 3.61650102e-02 7.01168329e-02
-1.23260476e-01 1.94677398e-01 -5.47147095e-01 2.03554487e+00
-2.72754133e-01 1.42766047e+00 -4.78046667e-03 -1.82882920e-02
6.88998640e-01 2.10344508e-01 4.64298815e-01 -8.39493454e-01
8.45116004e-03 8.27304780e-01 -1.65191328e-03 -7.75220871e-01
2.32133538e-01 2.65899241e-01 -3.35942090e-01 5.08728206e-01
3.59410316e-01 -5.02738655e-02 6.60216033e-01 3.18889260e-01
1.31325114e+00 6.16431117e-01 4.25270088e-02 2.41961181e-01
-9.42553133e-02 3.66267450e-02 3.87504160e-01 7.23739624e-01
-7.16385722e-01 6.16566896e-01 4.09964651e-01 -3.51920396e-01
-1.44573486e+00 -1.32093120e+00 1.26579732e-01 1.08732474e+00
-1.46238804e-01 -6.07578099e-01 -5.88835537e-01 -8.50242913e-01
2.43919268e-01 3.56425405e-01 -1.26973495e-01 4.04451907e-01
-1.10219538e+00 1.11091904e-01 1.19251287e+00 9.08229470e-01
6.18722141e-01 -1.21497679e+00 -6.14084244e-01 2.01744866e-02
-5.50199114e-02 -1.53668022e+00 -1.14874518e+00 -5.87775290e-01
-5.22366703e-01 -1.25778306e+00 -1.21810746e+00 -1.15200210e+00
5.51110327e-01 -1.77757904e-01 1.16207576e+00 -5.88781610e-02
-3.08586240e-01 3.01573247e-01 -6.65093303e-01 -1.68738768e-01
-5.21052957e-01 -1.80559084e-01 4.44754586e-02 -3.37965578e-01
8.17992687e-01 -2.52493888e-01 -5.77405095e-01 2.59186804e-01
-4.95884955e-01 6.18211627e-02 7.22122729e-01 8.32571805e-01
1.17802963e-01 -1.53505290e+00 2.67358243e-01 -3.21812987e-01
4.46137249e-01 -3.33897509e-02 -7.06594706e-01 3.15889895e-01
-1.67809069e-01 9.92132723e-02 4.81770486e-01 -5.92024863e-01
-5.41754961e-01 2.42125452e-01 -3.35478447e-02 -3.86920124e-01
2.62699323e-03 1.55973896e-01 1.41729444e-01 -1.83335900e-01
4.99885768e-01 2.81844139e-01 5.91928102e-02 -6.15153134e-01
7.38016129e-01 1.22046423e+00 8.39604378e-01 -3.92068326e-01
5.55216432e-01 3.27108622e-01 -4.64292288e-01 -6.99602485e-01
-3.34858268e-01 -5.32594383e-01 -6.60684407e-01 -4.99127299e-01
7.43239522e-01 -1.16900229e+00 -8.53152037e-01 8.87407243e-01
-9.99562979e-01 -7.61091948e-01 -3.61186922e-01 6.00228548e-01
-7.64846504e-01 3.12483132e-01 -8.29625070e-01 -3.33555818e-01
-3.90699834e-01 -9.22532260e-01 1.18018723e+00 2.99153272e-02
-4.20262337e-01 -3.00149739e-01 3.44198912e-01 3.67996871e-01
3.67203236e-01 7.97748044e-02 1.48442864e-01 -4.25609142e-01
-7.05930471e-01 -4.68716472e-01 -7.00902522e-01 6.24720991e-01
1.00508019e-01 2.08354890e-01 -5.38576365e-01 -4.35833424e-01
-1.28376973e+00 -8.64739001e-01 9.61295366e-01 1.17916524e-01
5.99394441e-01 -3.34277987e-01 -3.94027244e-04 7.50175178e-01
1.20280051e+00 -5.31654060e-02 4.90250081e-01 2.63586700e-01
7.42800713e-01 -8.54357854e-02 5.56266963e-01 4.35094297e-01
3.68865490e-01 5.73878765e-01 -6.07007109e-02 2.13930607e-01
-8.84734511e-01 -9.38322127e-01 7.38862991e-01 1.18301082e+00
-3.47445488e-01 -3.12689871e-01 -9.31144834e-01 9.95424271e-01
-1.83161259e+00 -1.14909852e+00 4.58280481e-02 2.04861927e+00
8.05363894e-01 -1.23521499e-01 3.98181111e-01 -5.55797040e-01
4.71102089e-01 3.22855234e-01 -7.98577905e-01 -3.83491606e-01
-3.90472233e-01 1.51642963e-01 8.88387561e-01 5.42299688e-01
-1.08886933e+00 1.41479874e+00 7.04539394e+00 4.98091280e-01
-1.32488680e+00 8.79565030e-02 -8.81498232e-02 -8.38832319e-01
-7.82048330e-02 -3.06269843e-02 -7.14591563e-01 8.02483022e-01
8.11717093e-01 -2.63783157e-01 4.67237025e-01 9.70044255e-01
1.87574804e-01 3.25909615e-01 -1.07895887e+00 1.24507427e+00
2.51926601e-01 -1.31253767e+00 2.46721014e-01 1.43526897e-01
1.04540038e+00 1.13453579e+00 -1.87567055e-01 3.19246620e-01
6.42683148e-01 -8.30556691e-01 8.88330162e-01 4.25086558e-01
1.72050536e+00 -1.54628366e-01 4.97685343e-01 -2.41578266e-01
-1.17817163e+00 3.32928984e-03 1.37983188e-01 -7.69786537e-02
8.34739923e-01 -3.26240629e-01 -3.46985310e-01 -3.01648349e-01
6.12265766e-01 1.22236133e+00 -4.34141427e-01 1.39876795e+00
-5.15240252e-01 8.68852556e-01 -7.47934997e-01 -7.17227578e-01
5.77914298e-01 1.46408737e-01 5.33699095e-01 1.56402147e+00
4.86970931e-01 -3.93177345e-02 -5.97749762e-02 1.34503096e-01
-5.32610655e-01 2.02477366e-01 -7.53968716e-01 -5.01988769e-01
4.25560445e-01 3.91452581e-01 -1.91314723e-02 -7.03623414e-01
-6.66261911e-01 1.53588998e+00 1.33989111e-01 5.31286001e-01
-8.41184258e-01 -5.10475338e-01 1.21327353e+00 5.17474636e-02
3.83902401e-01 -4.38631177e-01 2.71727651e-01 -1.68880236e+00
4.62711275e-01 -8.36571515e-01 1.79812133e-01 -1.00407302e+00
-1.39083767e+00 3.80394310e-01 -3.63738924e-01 -1.72941160e+00
-4.59102333e-01 -1.05572605e+00 -3.43725115e-01 2.96073407e-01
-1.55492711e+00 -1.57826579e+00 -3.85706067e-01 5.08459210e-01
7.87742019e-01 -5.19835711e-01 6.17613435e-01 5.61231792e-01
-6.67930301e-03 1.06801045e+00 5.73690832e-01 8.31128657e-01
1.26591170e+00 -1.13987768e+00 1.22690344e+00 8.03157449e-01
3.03090096e-01 2.52796620e-01 3.84252369e-01 -7.09802032e-01
-1.37306070e+00 -8.78298819e-01 1.41898787e+00 -9.56818938e-01
1.20867884e+00 -2.90476978e-01 -2.71437429e-02 9.56863225e-01
9.54174995e-02 3.71560812e-01 5.11205256e-01 -3.54427338e-01
-9.17601824e-01 2.58725375e-01 -9.42359388e-01 9.67257917e-01
1.91646338e+00 -7.88333356e-01 -5.16626000e-01 3.86083215e-01
4.43613023e-01 -5.96572578e-01 -7.28523731e-01 6.03247508e-02
1.41574788e+00 -3.78460467e-01 4.85671580e-01 -1.24403536e+00
7.89663732e-01 -5.99455722e-02 -2.12012574e-01 -1.10717225e+00
3.87336016e-02 -9.53406751e-01 -1.41646966e-01 8.15546989e-01
4.81374770e-01 -4.96796459e-01 1.05148840e+00 6.17870450e-01
-1.29928425e-01 -4.94173884e-01 -1.05077052e+00 -1.23542714e+00
3.77649032e-02 -5.10482609e-01 3.04278463e-01 5.22788942e-01
5.74550368e-02 -1.29546717e-01 -8.57710600e-01 -5.90257466e-01
7.83837676e-01 5.47490194e-02 1.26881623e+00 -7.40600407e-01
-1.70853868e-01 -6.92917466e-01 -6.87330008e-01 -1.80182922e+00
1.18276976e-01 -8.90612185e-01 5.85942343e-02 -1.53990734e+00
1.36001542e-01 -1.73066452e-03 -1.36517361e-01 8.10549140e-01
2.67522395e-01 6.55295014e-01 7.52108753e-01 2.21454546e-01
-8.66509914e-01 2.34379902e-01 1.38152075e+00 -2.24907741e-01
1.04469031e-01 -4.25027758e-01 -1.44638717e-01 5.69692910e-01
4.51562643e-01 -1.58288017e-01 2.21696883e-01 -1.09761798e+00
3.96892317e-02 -4.33999181e-01 1.90058604e-01 -7.92357981e-01
1.36196390e-01 -1.12753652e-01 -2.12491766e-01 -3.99859697e-01
2.95486808e-01 -3.53029370e-01 -4.26864922e-01 5.56250334e-01
-5.07199764e-01 -6.31031394e-02 1.33624852e-01 3.42603356e-01
-3.56265008e-01 2.06391037e-01 5.97420156e-01 3.22587006e-02
-1.40315795e+00 2.75693864e-01 -3.50675017e-01 5.39963901e-01
6.08768940e-01 -2.70619124e-01 -6.34774148e-01 -6.20381296e-01
-4.50565249e-01 4.03424025e-01 6.71341717e-01 9.11901832e-01
6.39559925e-01 -1.59615803e+00 -1.08830237e+00 2.57375747e-01
6.99102402e-01 -5.34213960e-01 -3.11957672e-02 4.40836281e-01
-1.01900530e+00 5.36942065e-01 -5.09748101e-01 -2.17725351e-01
-1.67981541e+00 -2.13871732e-01 -1.78869948e-01 2.29835004e-01
-8.49129677e-01 1.09024799e+00 -5.60531497e-01 -5.00002325e-01
3.23000908e-01 -6.94303811e-01 4.69259530e-01 -2.90277451e-01
6.01096392e-01 2.20996290e-01 -6.49231076e-01 -8.45572531e-01
-4.29574996e-01 1.01515090e+00 1.65917024e-01 -3.38277042e-01
1.13412511e+00 3.24353784e-01 2.99543202e-01 1.40961945e-01
1.42492306e+00 1.83082357e-01 -1.52214789e+00 -1.25871271e-01
-8.65227655e-02 -6.78740442e-01 -4.91123974e-01 -1.01217854e+00
-8.89805079e-01 6.14052653e-01 6.16639912e-01 -9.63529229e-01
7.38953292e-01 2.57564187e-01 1.52394080e+00 7.55021572e-01
5.71253181e-01 -1.31564438e+00 -1.64611787e-02 8.53756726e-01
9.64227080e-01 -1.49988806e+00 -3.97218227e-01 8.73783231e-02
-8.29631448e-01 8.35531592e-01 5.59296966e-01 -4.23173845e-01
4.57411796e-01 2.90450782e-01 7.32553184e-01 1.14428543e-01
-5.47563910e-01 -2.18137667e-01 2.32125774e-01 5.09486616e-01
4.81238902e-01 2.19684213e-01 -6.35970950e-01 2.00477704e-01
-3.68357033e-01 7.22580075e-01 3.47278744e-01 1.08404827e+00
-1.78451538e-01 -1.01947403e+00 1.50074482e-01 5.22768259e-01
-6.81301206e-02 -1.47212699e-01 -6.40098631e-01 8.60812664e-01
-2.62476712e-01 4.30818886e-01 7.75980353e-02 -2.64876008e-01
5.50579846e-01 6.24716841e-02 6.62583411e-01 -1.40369236e-01
-2.99515754e-01 -2.06959769e-01 5.19748509e-01 -1.00807428e+00
-3.08299422e-01 -6.95132315e-01 -1.24995041e+00 -2.88956046e-01
9.57140252e-02 -4.79819745e-01 8.01750660e-01 7.45840430e-01
5.46032190e-01 -2.62417197e-01 8.60390738e-02 -6.24582171e-01
-6.21226728e-01 -9.14146602e-01 -2.67684907e-01 8.18718493e-01
4.15893286e-01 -2.63656914e-01 -4.42785531e-01 2.50993520e-01] | [9.192279815673828, -6.518263339996338] |
36160953-4de5-4078-826e-10913ab2e450 | probabilistic-failure-analysis-in-model | 1611.05083 | null | http://arxiv.org/abs/1611.05083v2 | http://arxiv.org/pdf/1611.05083v2.pdf | Probabilistic Failure Analysis in Model Validation & Verification | Automated fault localization is an important issue in model validation and
verification. It helps the end users in analyzing the origin of failure. In
this work, we show the early experiments with probabilistic analysis approaches
in fault localization. Inspired by the Kullback-Leibler Divergence from
Bayesian probabilistic theory, we propose a suspiciousness factor to compute
the fault contribution for the transitions in the reachability graph of model
checking, using which to rank the potential faulty transitions. To
automatically locate design faults in the simulation model of detailed design,
we propose to use the statistical model Hidden Markov Model (HMM), which
provides statistically identical information to component's real behavior. The
core of this method is a fault localization algorithm that gives out the set of
suspicious ranked faulty components and a backward algorithm that computes the
matching degree between the HMM and the simulation model to evaluate the
confidence degree of the localization conclusion. | ['Xavier Crégut', 'Marc Pantel', 'Ning Ge'] | 2016-11-15 | null | null | null | null | ['fault-localization'] | ['computer-code'] | [-1.15960985e-01 1.08306704e-03 2.28841249e-02 -2.87825823e-01
-6.79074764e-01 -2.54587024e-01 2.93645740e-01 3.04724693e-01
4.12784278e-01 4.27057803e-01 -1.83668166e-01 -7.75327563e-01
-5.80154240e-01 -6.67409718e-01 -6.39353156e-01 -4.10070300e-01
-4.13808227e-01 5.46696365e-01 7.94031799e-01 1.72446474e-01
4.29241180e-01 6.18524194e-01 -1.58348346e+00 3.01285237e-01
4.25691366e-01 9.04363513e-01 -1.16992295e-02 8.11629713e-01
6.01860471e-02 8.22804570e-01 -7.17740059e-01 3.94825369e-01
-3.08036000e-01 -6.78130090e-01 -9.54932094e-01 -6.78401440e-02
-4.29642081e-01 -3.37051690e-01 -1.84988864e-02 1.42582905e+00
-1.45625472e-01 -3.46630096e-01 6.48017526e-01 -1.85652077e+00
2.82139868e-01 9.73941505e-01 -1.59390077e-01 1.86502077e-02
9.13782954e-01 6.89072534e-02 6.06708765e-01 -6.57887518e-01
3.46999407e-01 1.31721842e+00 5.57987809e-01 1.31876603e-01
-9.78537261e-01 -4.38713253e-01 -2.83746183e-01 3.96732807e-01
-1.73534274e+00 -1.25556916e-01 7.89068818e-01 -6.11551762e-01
7.63754308e-01 2.80404836e-01 2.37751514e-01 6.78179324e-01
8.07861269e-01 1.95756525e-01 1.00158381e+00 -8.33277881e-01
7.46007562e-01 3.27012278e-02 5.37682891e-01 9.15664136e-01
3.43681276e-01 3.38288724e-01 -2.26655632e-01 -7.65957177e-01
3.73729527e-01 8.82484838e-02 -3.56754065e-02 -1.35042086e-01
-5.97348630e-01 4.83272523e-01 -1.79923773e-01 4.66406316e-01
-2.77406096e-01 6.05677664e-01 2.41232514e-01 1.37185305e-01
-2.42303044e-01 5.29755140e-04 -3.51993233e-01 -1.29498318e-01
-9.48289692e-01 8.16397592e-02 1.02396476e+00 8.49585891e-01
6.44835830e-01 -1.25188723e-01 4.25983891e-02 -2.33858049e-01
1.13914454e+00 3.15682620e-01 9.08267349e-02 -6.95416808e-01
-2.57553875e-01 8.97319019e-01 6.42891154e-02 -6.77903771e-01
4.55359779e-02 -2.26138234e-01 -1.96493983e-01 5.96810281e-01
1.03299864e-01 1.13005690e-01 -7.88842201e-01 1.29773331e+00
2.28762373e-01 4.06207383e-01 -2.91425586e-01 4.21186417e-01
-1.88359097e-01 4.49870765e-01 -2.83433467e-01 -1.08907916e-01
1.20720279e+00 -4.53353301e-02 -6.24635220e-01 3.84421349e-01
6.69241786e-01 -7.33066976e-01 5.32010853e-01 6.33380055e-01
-5.60744882e-01 -2.92922348e-01 -1.41784036e+00 1.01705325e+00
-1.52741998e-01 -1.07012056e-01 1.68615028e-01 9.41039681e-01
-1.03182864e+00 9.53919709e-01 -1.19173646e+00 -1.83519036e-01
-1.86515167e-01 3.18499118e-01 -1.13405526e-01 -2.18792468e-01
-1.03992248e+00 8.39914799e-01 4.15376157e-01 1.57408386e-01
-1.63738716e+00 -1.64016485e-01 -7.50616550e-01 9.42194462e-02
3.75162333e-01 -8.69270414e-02 1.11140847e+00 -6.38637543e-01
-1.34605515e+00 6.80990443e-02 -1.77324712e-01 -3.15364182e-01
2.29216740e-01 1.13505997e-01 -8.33541155e-01 1.16313286e-02
2.68665627e-02 -3.02341014e-01 8.03012192e-01 -1.38507569e+00
-5.57868600e-01 -2.47388795e-01 -2.95695871e-01 -1.01561260e+00
4.17085797e-01 3.57439488e-01 -1.52268931e-01 7.95388296e-02
4.48442847e-01 -6.82028651e-01 -3.79949242e-01 -5.87327719e-01
-6.47904813e-01 -2.53726989e-01 8.87023866e-01 -5.78048408e-01
1.47161794e+00 -2.37328887e+00 -4.84418660e-01 9.81653512e-01
-9.77447033e-02 -3.41335386e-01 5.74541509e-01 7.82822251e-01
-2.62006849e-01 -3.26451771e-02 -2.79538482e-01 2.32029706e-01
2.13511989e-01 -7.68837184e-02 -4.15234149e-01 8.59940469e-01
2.78869927e-01 1.33112222e-01 -8.52771282e-01 -4.76299495e-01
2.35584393e-01 -2.16835178e-02 -2.07161099e-01 2.02556014e-01
-4.36165988e-01 -5.20484932e-02 -6.91288769e-01 6.37434244e-01
6.03253841e-01 -1.01532906e-01 3.05557132e-01 -7.19501376e-02
-1.60754517e-01 2.95829952e-01 -1.69760621e+00 1.04534125e+00
3.30806971e-02 2.11154759e-01 -3.44990432e-01 -5.23399413e-01
9.46760774e-01 5.33819795e-01 7.87714571e-02 -8.80364627e-02
3.50185990e-01 2.93915629e-01 -1.15699470e-01 -2.74970680e-01
1.35603815e-01 -2.18946231e-03 -3.68951291e-01 5.20988166e-01
5.54055385e-02 5.56461401e-02 -2.56982967e-02 1.61981255e-01
1.61516416e+00 9.95891169e-02 9.11631659e-02 -5.87525189e-01
5.32344997e-01 8.79396796e-02 5.70624650e-01 5.03604829e-01
9.43911299e-02 1.61137179e-01 9.28218961e-01 -5.53772524e-02
-6.21064365e-01 -1.20002782e+00 1.67793959e-01 1.03387907e-01
9.02008042e-02 -8.12972307e-01 -8.13619971e-01 -9.51153576e-01
-2.87090391e-02 8.75813127e-01 -4.20230687e-01 -6.90254211e-01
-1.23963311e-01 -2.18885154e-01 6.32777572e-01 4.48308825e-01
-1.23516448e-01 -5.30976772e-01 -6.61608517e-01 1.14245974e-01
3.29722345e-01 -5.12633622e-01 8.66903514e-02 6.19699299e-01
-9.42401469e-01 -1.44614065e+00 3.06870222e-01 -4.68329728e-01
8.45579147e-01 -5.30761898e-01 7.74671435e-01 2.39097685e-01
-5.08078277e-01 1.16975151e-01 -3.97311151e-01 -1.01723626e-01
-1.14127004e+00 -8.37957084e-01 1.09520905e-01 -2.97105849e-01
7.03348359e-03 -2.68540382e-01 6.81544319e-02 5.95899940e-01
-8.25281084e-01 -6.73828363e-01 4.33739752e-01 4.22856748e-01
4.10525173e-01 1.11740959e+00 1.90555587e-01 -5.37535548e-01
4.02538598e-01 -4.18662935e-01 -1.08972836e+00 3.86513621e-01
-9.01051164e-01 6.64449155e-01 5.24980903e-01 -6.25676364e-02
-6.25662625e-01 2.56760418e-01 -1.75340548e-01 -2.91257530e-01
-3.55713248e-01 7.82036781e-01 -6.35996580e-01 -6.44700276e-03
3.59772116e-01 -2.20051035e-02 -1.61926448e-01 -5.71441293e-01
-2.46900809e-03 5.53627431e-01 2.97341913e-01 -5.09079754e-01
5.86260498e-01 1.94183037e-01 4.03084159e-01 -3.36490244e-01
1.41769480e-02 -3.96808326e-01 -2.49938324e-01 -4.92092013e-01
4.35387433e-01 -2.89143860e-01 -1.12290037e+00 2.91148216e-01
-1.10587478e+00 1.27755878e-02 -5.24924174e-02 6.40311480e-01
-4.13739949e-01 6.73620105e-01 -6.26557171e-01 -1.31186593e+00
7.82859325e-02 -1.41278386e+00 9.48577523e-01 -6.76881373e-02
-4.91136402e-01 -5.04122078e-01 2.82555908e-01 -3.97453696e-01
-4.58037928e-02 3.62157375e-01 1.44767463e+00 -7.41652369e-01
-4.32011843e-01 -8.00059199e-01 2.47673362e-01 4.67283964e-01
-7.29894936e-02 7.11802363e-01 -7.89959013e-01 -2.31934357e-02
3.46334517e-01 4.52688634e-01 2.60761917e-01 1.71506986e-01
6.03373110e-01 -1.18862791e-02 -7.13531137e-01 -1.77084222e-01
1.62669635e+00 6.24099314e-01 8.36768150e-01 -1.65317476e-01
3.15280378e-01 4.31015640e-01 9.38297153e-01 6.84292853e-01
-2.25445911e-01 5.43295324e-01 5.70157468e-01 5.40298700e-01
1.23006910e-01 -4.70640808e-01 8.15569460e-01 4.77342010e-01
6.33220732e-01 -6.11850917e-02 -1.20784581e+00 3.25117141e-01
-1.60163331e+00 -8.03052545e-01 -4.73945260e-01 2.49056244e+00
4.44318652e-01 7.73841083e-01 8.02980810e-02 8.35089147e-01
1.10294688e+00 -6.55497670e-01 8.90266001e-02 -4.17018712e-01
5.61716795e-01 2.83549547e-01 5.63749552e-01 7.76738286e-01
-6.20126963e-01 4.08758879e-01 6.61955786e+00 9.23683763e-01
-6.78150058e-01 -4.59066108e-02 2.32983730e-03 6.99428618e-01
-1.96315333e-01 8.12241495e-01 -7.45689332e-01 6.69691324e-01
1.51038408e+00 -4.48870808e-02 -1.84107050e-01 1.01318443e+00
2.77931511e-01 -7.89388001e-01 -1.46098697e+00 3.27430725e-01
-6.71585202e-02 -7.93329000e-01 -1.87236518e-01 1.48077935e-01
3.35284680e-01 -4.43572998e-01 -4.05736625e-01 -2.15185165e-01
5.93702972e-01 -8.37715924e-01 1.05576706e+00 9.63562906e-01
3.71642292e-01 -1.18067932e+00 1.11519468e+00 2.72276640e-01
-1.08555543e+00 1.52770709e-02 1.14616655e-01 -6.71428144e-02
3.57747585e-01 1.15903783e+00 -1.40740705e+00 7.54219651e-01
5.53451180e-01 -2.67284568e-02 -4.60447848e-01 1.32604134e+00
-4.03915763e-01 7.69583166e-01 -4.06594366e-01 -5.71046360e-02
-1.35546714e-01 2.63288200e-01 4.03434664e-01 7.45016932e-01
4.48742956e-01 -5.59403181e-01 1.17902614e-01 9.76372182e-01
5.94064713e-01 -6.16275370e-01 -3.66148442e-01 -7.91355222e-02
5.43400228e-01 6.56206608e-01 -1.29878366e+00 -1.44546613e-01
9.25820246e-02 7.15146124e-01 -5.29705465e-01 -5.62514365e-02
-8.91514063e-01 -6.20781064e-01 1.61066368e-01 1.78586662e-01
2.24933997e-01 -2.84480751e-01 -9.45375189e-02 -2.02183783e-01
5.00276014e-02 -7.42174685e-01 2.28723988e-01 -6.97419822e-01
-9.26167965e-01 5.87052822e-01 2.92852789e-01 -1.36329198e+00
-4.19188887e-01 -5.84681809e-01 -8.73178124e-01 8.53175879e-01
-6.61122143e-01 -7.08139956e-01 2.00113297e-01 4.91436839e-01
7.85678700e-02 3.19303542e-01 6.92471623e-01 2.00175598e-01
-5.95385313e-01 2.66793311e-01 -2.58492738e-01 -3.15232724e-02
5.05299091e-01 -9.93438184e-01 3.29233408e-01 1.33311915e+00
-2.05728337e-01 5.53611100e-01 1.33914852e+00 -1.35416090e+00
-1.63845479e+00 -7.63479710e-01 8.89016628e-01 -3.10856491e-01
9.63206649e-01 2.44317111e-02 -6.22637987e-01 6.11713469e-01
-2.50911325e-01 -1.75733507e-01 3.26436460e-01 -1.94519833e-01
-1.56448796e-01 4.55695018e-02 -1.25628412e+00 2.92401880e-01
2.04948425e-01 -7.11893559e-01 -6.62273109e-01 4.44942452e-02
6.56857491e-01 1.54972672e-01 -8.60074699e-01 4.25386131e-01
4.60951537e-01 -9.31490123e-01 3.69791150e-01 -2.01945692e-01
-1.38493642e-01 -1.13984561e+00 -2.89318264e-01 -7.71666169e-01
-2.96038717e-01 -4.76786911e-01 -4.03886400e-02 1.36783981e+00
5.60378850e-01 -3.98938239e-01 5.94320178e-01 4.12980050e-01
-2.66570330e-01 -3.74009937e-01 -1.03821504e+00 -8.98830712e-01
-7.94246793e-01 -8.66600633e-01 6.33460104e-01 4.71405566e-01
4.60015208e-01 -2.40302250e-01 1.15226563e-02 8.89734924e-01
6.99197769e-01 -2.16520745e-02 8.46718401e-02 -1.40116036e+00
-5.60065329e-01 -2.33252376e-01 -9.94817913e-01 -2.61074640e-02
1.02034755e-01 -4.02295262e-01 7.87005126e-01 -1.20298398e+00
7.68320560e-02 -1.24135949e-01 -3.60773951e-01 1.51531681e-01
2.16885507e-01 -4.54915613e-01 -4.06278074e-01 -1.07058644e-01
-6.15921676e-01 1.63388774e-01 1.22971170e-01 1.03780679e-01
1.14018537e-01 4.20180500e-01 3.48209031e-02 7.60840416e-01
5.24736583e-01 -1.06683528e+00 -1.83353901e-01 2.95797139e-01
3.53860348e-01 4.88152474e-01 5.70518434e-01 -1.41489649e+00
3.10693502e-01 -7.56391659e-02 -3.48543823e-02 -8.83074939e-01
-2.83483803e-01 -1.19889343e+00 7.48655021e-01 1.08474088e+00
-6.79638833e-02 3.96994770e-01 -1.24854602e-01 6.63613260e-01
-2.70004243e-01 -8.23099434e-01 4.74941790e-01 2.60504484e-01
-7.01996624e-01 -1.84535280e-01 -7.71200061e-01 -7.28465319e-01
1.12993348e+00 6.92634210e-02 -4.65646014e-02 1.06681667e-01
-6.82865262e-01 -5.74858971e-02 5.82231820e-01 -8.35517347e-02
8.47059071e-01 -1.12194490e+00 -1.00274533e-01 3.31141412e-01
1.82755128e-01 -6.37162089e-01 3.49350423e-02 7.37582088e-01
-5.73055446e-01 1.23588428e-01 9.68223736e-02 -5.53266048e-01
-1.18928599e+00 7.65730917e-01 9.02674645e-02 -2.98300125e-02
-1.00040570e-01 5.37547410e-01 -6.64915562e-01 2.06523851e-01
8.18692669e-02 -6.87393248e-01 1.07546926e-01 -5.79106748e-01
5.47208965e-01 7.94375539e-01 3.81623894e-01 -2.92149574e-01
-1.03531861e+00 3.17550033e-01 2.82731295e-01 -5.41347444e-01
8.25692773e-01 1.00148693e-01 -5.77457488e-01 7.55536914e-01
1.01464975e+00 -3.27836499e-02 -8.55037570e-01 3.01616907e-01
8.77015471e-01 -2.54626364e-01 2.27236718e-01 -8.23898315e-01
-4.63391393e-01 5.94134629e-01 9.07507360e-01 4.02418733e-01
9.74596322e-01 2.89213210e-01 2.87150800e-01 1.89309400e-02
8.22375596e-01 -7.23006487e-01 -3.99581909e-01 4.11761135e-01
1.97745755e-01 -4.97836679e-01 -1.22611471e-01 -4.93736714e-01
-2.17095017e-01 1.19309592e+00 1.08485684e-01 -2.71507949e-01
9.13347542e-01 1.03229761e+00 -5.04731178e-01 -2.80708700e-01
-7.68701315e-01 2.49458384e-03 -5.42398216e-03 4.32893038e-01
1.50339603e-01 2.34692231e-01 -2.32576728e-01 9.33764756e-01
2.16758728e-01 1.11047074e-01 5.38634658e-01 1.46742284e+00
-8.62521648e-01 -1.30538630e+00 -7.77418613e-01 1.42529190e-01
-2.09107101e-01 2.27119923e-01 -4.00810689e-01 4.71038848e-01
1.48519665e-01 1.36653841e+00 -2.04881519e-01 -1.05236638e+00
2.95937687e-01 3.60639602e-01 4.17838097e-01 -6.32577300e-01
-2.58933932e-01 7.69137219e-03 9.94151756e-02 -7.93386161e-01
2.51772791e-01 -5.55554867e-01 -1.50529706e+00 -1.00569218e-01
-7.75322199e-01 7.50844061e-01 9.75739121e-01 1.40237248e+00
2.31579885e-01 7.90978611e-01 9.99975443e-01 -1.98677629e-01
-6.01954699e-01 -7.27509320e-01 -1.00390971e+00 -1.83454871e-01
2.57985204e-01 -8.07153285e-01 -6.34512365e-01 7.71544455e-03] | [5.476309299468994, 2.6866700649261475] |
48a86aa5-d71b-4aa9-8e43-fb767231817f | uwsod-toward-fully-supervised-level-capacity | null | null | http://proceedings.neurips.cc/paper/2020/hash/4e0928de075538c593fbdabb0c5ef2c3-Abstract.html | http://proceedings.neurips.cc/paper/2020/file/4e0928de075538c593fbdabb0c5ef2c3-Paper.pdf | UWSOD: Toward Fully-Supervised-Level Capacity Weakly Supervised Object Detection | Weakly supervised object detection (WSOD) has attracted extensive research attention due to its great flexibility of exploiting large-scale dataset with only image-level annotations for detector training. Despite its great advance in recent years, WSOD still suffers limited performance, which is far below that of fully supervised object detection (FSOD). As most WSOD methods depend on object proposal algorithms to generate candidate regions and are also confronted with challenges like low-quality predicted bounding boxes and large scale variation. In this paper, we propose a unified WSOD framework, termed UWSOD, to develop a high-capacity general detection model with only image-level labels, which is self-contained and does not require external modules or additional supervision. To this end, we exploit three important components, i.e., object proposal generation, bounding-box fine-tuning and scale-invariant features. First, we propose an anchor-based self-supervised proposal generator to hypothesize object locations, which is trained end-to-end with supervision created by UWSOD for both objectness classification and regression. Second, we develop a step-wise bounding-box fine-tuning to refine both detection scores and coordinates by progressively select high-confidence object proposals as positive samples, which bootstraps the quality of predicted bounding boxes. Third, we construct a multi-rate resampling pyramid to aggregate multi-scale contextual information, which is the first in-network feature hierarchy to handle scale variation in WSOD. Extensive experiments on PASCAL VOC and MS COCO show that the proposed UWSOD achieves competitive results with the state-of-the-art WSOD methods while not requiring external modules or additional supervision. Moreover, the upper-bound performance of UWSOD with class-agnostic ground-truth bounding boxes approaches Faster R-CNN, which demonstrates UWSOD has fully-supervised-level capacity. | ['Feiyue Huang', 'Yongjian Wu', 'Zhiwei Chen', 'Rongrong Ji', 'Yunhang Shen'] | 2020-12-01 | null | null | null | neurips-2020-12 | ['object-proposal-generation'] | ['computer-vision'] | [ 1.05254008e-02 4.20091301e-02 -4.08351272e-01 -3.31017762e-01
-1.05979371e+00 -5.36177576e-01 4.13473219e-01 -2.12261379e-02
-5.98625779e-01 3.29341441e-01 -1.34033144e-01 8.05438459e-02
2.53011346e-01 -6.82057261e-01 -8.68648648e-01 -6.08331144e-01
1.24871954e-01 2.72398382e-01 1.16378665e+00 -2.30823651e-01
-3.33163254e-02 6.66754127e-01 -1.48460400e+00 1.41117856e-01
7.78122365e-01 1.24334526e+00 4.99663740e-01 5.35398841e-01
6.13964088e-02 2.82817066e-01 -3.93551499e-01 -2.29831591e-01
4.60762739e-01 -1.19731337e-01 -2.93107688e-01 1.16578229e-01
6.89826190e-01 -4.00785327e-01 -2.68064231e-01 1.02308214e+00
6.49705231e-01 -1.89301148e-01 5.37235677e-01 -1.25916338e+00
-5.84244728e-01 3.51597875e-01 -8.27258348e-01 1.71414211e-01
-1.40945405e-01 5.35068214e-01 1.08692539e+00 -1.24593866e+00
5.92522204e-01 1.18974459e+00 7.48861015e-01 6.93146944e-01
-1.57313395e+00 -9.24000740e-01 3.52950126e-01 -2.08491102e-01
-1.65914059e+00 -1.46266565e-01 7.65293419e-01 -4.18433547e-01
7.38764048e-01 6.09947965e-02 4.48824584e-01 9.06572878e-01
-1.95325747e-01 1.02999711e+00 9.82136667e-01 -3.57005179e-01
2.16117442e-01 4.37126637e-01 1.43051520e-03 8.19004416e-01
4.87807393e-01 1.32806554e-01 -4.79343772e-01 2.88437232e-02
9.68724549e-01 1.24058403e-01 -6.82353303e-02 -8.79637837e-01
-1.32652605e+00 8.24308872e-01 1.06678176e+00 4.22119722e-02
-4.39828858e-02 9.39544216e-02 2.88013279e-01 -2.66656756e-01
3.67230833e-01 3.75528783e-01 -4.40572053e-01 5.11304975e-01
-9.45152879e-01 2.22715721e-01 3.12708944e-01 1.12986314e+00
7.21590281e-01 -9.10648927e-02 -5.93414009e-01 8.72303247e-01
2.60476112e-01 6.75091922e-01 3.15974534e-01 -4.21636701e-01
5.17294586e-01 9.02704954e-01 2.32533962e-01 -8.16944301e-01
-5.38172245e-01 -8.05827737e-01 -7.16030657e-01 2.32055277e-01
3.00287306e-01 1.12621397e-01 -1.07828808e+00 1.87962937e+00
6.49890363e-01 2.09145602e-02 -2.55190670e-01 1.15721703e+00
8.78045976e-01 5.54124951e-01 1.35668322e-01 6.30660728e-02
1.41618359e+00 -1.13971686e+00 -1.39270529e-01 -3.43962818e-01
5.57142913e-01 -4.85673249e-01 1.14098585e+00 1.06888309e-01
-8.43204558e-01 -8.88666630e-01 -1.11732304e+00 -5.43067008e-02
-1.87538847e-01 7.32104599e-01 4.66160357e-01 4.46662575e-01
-7.51092851e-01 1.66143075e-01 -8.55507731e-01 -3.17177117e-01
9.04940426e-01 3.05956930e-01 -2.34339014e-01 1.19327232e-01
-7.21427202e-01 6.40417516e-01 6.93099916e-01 7.42229121e-03
-1.03576386e+00 -6.76403046e-01 -8.44007432e-01 -3.30201872e-02
7.79770195e-01 -4.79974657e-01 1.06537247e+00 -6.01054430e-01
-1.18908036e+00 8.38046968e-01 -4.91002053e-02 -4.59154278e-01
7.30709910e-01 -8.61262009e-02 -2.71564901e-01 1.16585016e-01
4.04081106e-01 1.12793243e+00 9.30361569e-01 -1.23665488e+00
-9.97569323e-01 -4.59988892e-01 -1.17491923e-01 3.86768021e-02
-5.08488595e-01 3.10891098e-03 -8.71442914e-01 -7.71364152e-01
3.65924716e-01 -9.34042096e-01 -4.39167470e-01 5.44060230e-01
-5.85199654e-01 -3.93328816e-01 8.08748782e-01 -8.28003213e-02
1.10922825e+00 -2.20170951e+00 -1.25467017e-01 1.38047248e-01
2.92013377e-01 3.49479586e-01 -2.95601755e-01 -4.87912446e-02
1.81465194e-01 -7.67962784e-02 -2.22057492e-01 -4.47942078e-01
7.18721151e-02 -8.99728164e-02 -4.63729203e-01 4.79206324e-01
8.45664620e-01 1.07218874e+00 -8.44248176e-01 -5.39807141e-01
1.65719613e-01 3.51185769e-01 -6.27099097e-01 1.92250878e-01
-3.28716397e-01 2.28911266e-01 -6.51443303e-01 7.50400841e-01
8.24034750e-01 -4.72338825e-01 -2.26667255e-01 -4.54069763e-01
-2.47007921e-01 -6.11913428e-02 -1.54330981e+00 1.60099304e+00
-2.72513300e-01 3.17777038e-01 -6.18280843e-02 -8.03755403e-01
1.15058005e+00 -3.00817728e-01 1.10089809e-01 -5.31172156e-01
-5.09778634e-02 4.34416145e-01 -2.20017985e-01 -8.57644081e-02
3.87278050e-01 2.93131955e-02 -1.43315107e-01 1.00072727e-01
2.17841431e-01 -1.59352422e-01 2.80085027e-01 3.61676902e-01
1.02354252e+00 3.03521842e-01 4.28231239e-01 -1.49316832e-01
6.57786906e-01 1.09215982e-01 8.31942141e-01 8.90017867e-01
-3.26056033e-01 8.78200293e-01 4.43743676e-01 -1.75058648e-01
-1.07714951e+00 -1.17931533e+00 -5.66210985e-01 1.24167538e+00
5.21233737e-01 -2.57438123e-01 -7.65739024e-01 -1.04940772e+00
2.39685997e-01 2.52812296e-01 -6.64500475e-01 -1.11499183e-01
-5.57686865e-01 -7.46093452e-01 5.19822121e-01 9.62700129e-01
6.00517988e-01 -9.95100021e-01 -5.85323095e-01 2.69842833e-01
2.63657242e-01 -1.36876714e+00 -7.06293702e-01 3.54270071e-01
-6.94556594e-01 -9.44830716e-01 -7.93548107e-01 -8.40039670e-01
8.66519034e-01 5.35430431e-01 7.72706747e-01 -1.28289998e-01
-7.37664878e-01 -9.93067771e-03 -3.07775766e-01 -4.98077959e-01
-1.20651826e-01 1.18972436e-01 1.29652098e-01 1.12116644e-02
3.47684503e-01 -3.82009000e-02 -8.86698782e-01 7.58826196e-01
-9.08922672e-01 1.03561260e-01 9.72126007e-01 9.73646402e-01
9.03469622e-01 -2.89982170e-01 7.68712521e-01 -6.22213364e-01
-2.09150650e-02 -1.21324867e-01 -1.01875746e+00 1.71135277e-01
-4.91853446e-01 -5.90777621e-02 5.73408961e-01 -5.49661815e-01
-7.80828416e-01 5.07246852e-01 -3.13138179e-02 -4.02962625e-01
-1.08590543e-01 -1.63868397e-01 -1.85973093e-01 -1.81772396e-01
1.04315233e+00 3.31003159e-01 -1.14039361e-01 -4.00824875e-01
4.51830864e-01 6.09087527e-01 6.66259766e-01 -3.37291092e-01
1.29047251e+00 7.81946838e-01 -2.64879286e-01 -5.15915990e-01
-1.36725032e+00 -8.45761418e-01 -8.32278848e-01 3.62139866e-02
9.20136809e-01 -1.21473908e+00 -5.31533539e-01 3.30108911e-01
-1.04724181e+00 -2.08655700e-01 -5.01637220e-01 3.91776383e-01
-3.52972269e-01 3.96671332e-02 -4.42628801e-01 -7.22283542e-01
-3.74717057e-01 -1.12620068e+00 1.58393526e+00 2.55580962e-01
2.48955309e-01 -3.60336453e-01 -2.85056770e-01 3.59845251e-01
2.46293053e-01 2.29998067e-01 2.75470376e-01 -6.48614645e-01
-9.34890449e-01 -4.13657755e-01 -7.96314240e-01 4.43375915e-01
-1.68187037e-01 -2.46499479e-01 -1.03473139e+00 -5.46150506e-01
-5.06120443e-01 -6.43650293e-01 1.11231899e+00 3.93328011e-01
1.28502691e+00 6.94704503e-02 -5.81781209e-01 7.08007336e-01
1.27875781e+00 -4.23131287e-01 -1.30311958e-02 2.36360088e-01
6.61801338e-01 4.06153023e-01 9.14837837e-01 4.59938258e-01
2.47067168e-01 9.41157281e-01 4.13363814e-01 -4.86729294e-01
-3.13452810e-01 -4.88256812e-01 3.17055732e-01 1.41798913e-01
2.87101805e-01 -7.43732601e-03 -6.78953230e-01 5.09728611e-01
-1.92388213e+00 -7.70012796e-01 -6.40902445e-02 2.13056016e+00
7.90332794e-01 5.65650165e-01 5.31505764e-01 -1.44823223e-01
8.41279864e-01 9.72521827e-02 -9.08826828e-01 4.10742998e-01
-1.40267059e-01 -3.68689448e-02 6.82756007e-01 9.97393578e-03
-1.34924328e+00 1.17975557e+00 4.72616816e+00 1.15767944e+00
-1.13021219e+00 2.48319119e-01 4.88340557e-01 -1.85417235e-01
1.96518108e-01 -1.84698015e-01 -1.55454206e+00 4.00264502e-01
1.85074583e-01 2.39524141e-01 -3.13636400e-02 1.36182463e+00
1.65520042e-01 2.31289025e-02 -9.68643069e-01 7.36361325e-01
5.95631823e-02 -1.40491009e+00 -1.42832965e-01 -1.76993862e-01
7.62397528e-01 2.93213964e-01 -2.82292068e-02 5.26544511e-01
3.27433497e-01 -5.86205184e-01 1.06111586e+00 -6.44456372e-02
1.03890252e+00 -5.21863103e-01 5.96878588e-01 5.72444379e-01
-1.55100060e+00 -4.06655520e-01 -6.68962717e-01 5.04849315e-01
9.96676534e-02 6.02840960e-01 -9.65100110e-01 2.24032477e-01
8.42332661e-01 4.29563493e-01 -8.53648484e-01 1.16328096e+00
-4.12290454e-01 5.86652219e-01 -5.67189038e-01 -1.89646289e-01
3.04327637e-01 2.57881939e-01 4.63035494e-01 1.49759126e+00
1.31760180e-01 -2.48970594e-02 5.91945708e-01 1.17941117e+00
-1.53589651e-01 5.54390661e-02 -1.90111905e-01 4.60414588e-01
6.37201488e-01 1.62104023e+00 -9.55187798e-01 -3.60105038e-01
-3.57941717e-01 8.15601826e-01 5.58412731e-01 1.19967304e-01
-1.02899814e+00 -2.05508634e-01 3.17858040e-01 4.36342269e-01
6.99387670e-01 -1.07453100e-01 -1.73773095e-01 -1.12800121e+00
2.05379531e-01 -4.34902936e-01 4.64562654e-01 -5.32382667e-01
-1.28997350e+00 5.47984123e-01 -1.57877818e-01 -1.44391978e+00
3.67897570e-01 -7.20434248e-01 -5.29927254e-01 6.85653031e-01
-1.75058877e+00 -1.61747050e+00 -7.04134762e-01 3.84633780e-01
6.35603070e-01 -5.17194606e-02 3.07577074e-01 1.48263380e-01
-8.50201905e-01 8.05885494e-01 -6.45781904e-02 4.28774506e-01
8.79107773e-01 -1.31050587e+00 4.59077209e-01 8.85660291e-01
2.86480606e-01 4.61364329e-01 3.11919838e-01 -6.81996942e-01
-1.18821859e+00 -1.61851084e+00 3.87088180e-01 -4.32339966e-01
7.51495779e-01 -8.85119498e-01 -8.73235762e-01 3.10649991e-01
-6.94879830e-01 9.99210000e-01 2.23196924e-01 -6.21495657e-02
-5.46079695e-01 -4.90343630e-01 -1.06193674e+00 4.50092882e-01
1.29566431e+00 -3.67970616e-01 -4.30748701e-01 5.05974710e-01
9.70770597e-01 -4.58717525e-01 -4.34127033e-01 7.03172028e-01
2.73398489e-01 -8.42211068e-01 1.03455889e+00 -4.44293916e-01
1.11765496e-01 -7.70606399e-01 3.42660323e-02 -8.18121493e-01
-4.67842996e-01 -3.65608126e-01 -1.16706364e-01 1.24622738e+00
5.32720327e-01 -4.17612880e-01 1.10270894e+00 2.28197724e-01
-2.33713403e-01 -9.06808794e-01 -8.83687496e-01 -1.07673764e+00
-2.65877396e-01 -5.45452118e-01 4.42362487e-01 5.50262988e-01
-4.31418300e-01 2.94896454e-01 -9.51645002e-02 4.59957480e-01
8.02847445e-01 2.83424556e-01 1.12740529e+00 -1.11695838e+00
-3.20202559e-01 -4.35544133e-01 -5.98609328e-01 -1.42948544e+00
-1.45591304e-01 -8.59501362e-01 3.55771661e-01 -1.23366952e+00
4.66352195e-01 -7.34737515e-01 -4.07507032e-01 5.84659517e-01
-4.52495813e-01 7.96146929e-01 1.69379681e-01 3.35976481e-01
-9.84066427e-01 7.84465551e-01 1.28477681e+00 2.92584859e-02
-3.12710166e-01 4.11712080e-02 -5.41851342e-01 8.03416073e-01
4.04457361e-01 -5.60813308e-01 -1.31516874e-01 -1.91316445e-04
-1.40056849e-01 -2.52454698e-01 7.85225928e-01 -1.09899962e+00
2.67191350e-01 -6.57358095e-02 5.69245040e-01 -8.64165068e-01
2.17603087e-01 -6.45338833e-01 -5.47670722e-01 5.27207494e-01
-3.37407202e-01 -4.45541650e-01 8.21309015e-02 8.28677714e-01
-3.04390900e-02 -4.40639071e-02 1.23904753e+00 2.06483305e-01
-8.74157906e-01 6.07438326e-01 3.32996488e-01 9.14446861e-02
1.31878006e+00 -3.11279237e-01 -2.43708953e-01 2.99452752e-01
-4.43972051e-01 4.42465335e-01 4.08747524e-01 5.11653423e-01
5.08419573e-01 -1.30649090e+00 -7.60789216e-01 2.80721992e-01
6.48693860e-01 6.09272301e-01 1.29005134e-01 7.75174320e-01
-1.70735180e-01 3.45112771e-01 1.94982722e-01 -1.07087493e+00
-1.02748716e+00 5.66133916e-01 1.51082814e-01 -2.30021358e-01
-6.87071443e-01 1.18311560e+00 5.39921343e-01 -5.69746137e-01
2.98612207e-01 -3.45288873e-01 7.17167035e-02 -4.55274954e-02
4.84301358e-01 -6.64621592e-02 -3.29013728e-02 -3.74692976e-01
-4.36201543e-01 6.19726598e-01 -2.63931692e-01 1.98380217e-01
1.24608862e+00 1.43943891e-01 4.18598980e-01 1.10380910e-01
8.72075081e-01 -9.26521197e-02 -1.83730328e+00 -6.00874305e-01
5.28662931e-03 -4.42131519e-01 1.39092669e-01 -7.16802299e-01
-8.86386514e-01 6.87379718e-01 6.74630940e-01 -1.04139604e-01
8.38040709e-01 3.79156679e-01 7.90082932e-01 3.95337671e-01
3.75979424e-01 -1.23697710e+00 5.26538849e-01 2.19027311e-01
9.08779204e-01 -1.65756691e+00 9.46499854e-02 -6.06689036e-01
-6.11462414e-01 9.25204575e-01 1.16025138e+00 -2.33331844e-01
3.38482529e-01 4.11973983e-01 -1.74684957e-01 -1.51027618e-02
-4.45480704e-01 -4.76815581e-01 7.07887053e-01 4.83888209e-01
1.35832295e-01 -6.79709166e-02 -5.60319275e-02 6.88592017e-01
5.48391007e-02 -9.46778283e-02 3.77141759e-02 6.67202294e-01
-8.04783583e-01 -7.23431170e-01 -5.20464480e-01 5.46293974e-01
4.80550574e-04 5.65448366e-02 -2.07329825e-01 8.49425972e-01
2.54554212e-01 5.00232518e-01 -8.23146664e-03 -1.78413793e-01
5.86443305e-01 -4.11511987e-01 1.97025269e-01 -9.40897822e-01
-3.80650818e-01 1.65709257e-01 -3.17213804e-01 -6.70135915e-01
-1.34308964e-01 -4.74344671e-01 -1.42946124e+00 2.89480656e-01
-9.66354191e-01 -1.44675657e-01 5.64703643e-01 6.71834111e-01
4.04829592e-01 4.37854141e-01 6.96574330e-01 -1.13372719e+00
-9.47399318e-01 -9.76325452e-01 -3.89330417e-01 3.62576514e-01
3.00673187e-01 -8.18968356e-01 -1.72536194e-01 -2.44679898e-01] | [9.179481506347656, 0.8976779580116272] |
3e6b24e3-787a-423f-bd9b-d4dede21fd64 | graph-signal-processing-over-multilayer | 2108.13639 | null | https://arxiv.org/abs/2108.13639v3 | https://arxiv.org/pdf/2108.13639v3.pdf | Image Processing via Multilayer Graph Spectra | Graph signal processing (GSP) has become an important tool in image processing because of its ability to reveal underlying data structures. Many real-life multimedia datasets, however, exhibit heterogeneous structures across frames. Multilayer graphs (MLG), instead of traditional single-layer graphs, provide better representation of these datasets such as videos and hyperspectral images. To generalize GSP to multilayer graph models and develop multilayer analysis for image processing, this work introduces a tensor-based framework of multilayer graph signal processing (M-GSP) and present useful M-GSP tools for image processing. We then present guidelines for applying M-GSP in image processing and introduce several applications, including RGB image compression, edge detection and hyperspectral image segmentation. Successful experimental results demonstrate the efficacy and promising futures of M-GSP in image processing. | ['Zhi Ding', 'Qinwen Deng', 'Songyang Zhang'] | 2021-08-31 | null | null | null | null | ['hyperspectral-image-segmentation'] | ['computer-vision'] | [ 6.27325833e-01 -3.70885283e-01 1.65112749e-01 -2.78531071e-02
7.70799145e-02 -2.48459816e-01 6.57674745e-02 2.58718342e-01
2.44278178e-01 1.62409961e-01 -4.41661850e-02 -3.29286367e-01
-4.90140259e-01 -1.01507771e+00 -2.41808042e-01 -7.66876578e-01
-6.23476207e-01 -4.33333069e-01 4.04926419e-01 -2.00742051e-01
2.69484252e-01 7.87400842e-01 -1.45750368e+00 4.96126264e-01
5.03091097e-01 9.05415893e-01 2.36487612e-01 1.15201187e+00
-3.74016404e-01 9.03316021e-01 -2.39715576e-01 -3.45748663e-01
1.40096769e-01 -3.27731460e-01 -7.06244409e-01 8.63144100e-01
2.82948166e-01 4.88566272e-02 -5.04732072e-01 1.47744954e+00
1.57339901e-01 -1.07516926e-02 4.28033024e-01 -1.46011055e+00
-6.26834333e-01 4.67902899e-01 -8.11640024e-01 3.65494460e-01
6.54951155e-01 -1.41520211e-02 8.48925471e-01 -9.05145228e-01
5.77742696e-01 1.38526857e+00 8.14530790e-01 -1.77666992e-01
-1.20478094e+00 -7.82203749e-02 1.88007489e-01 5.34937024e-01
-1.17774820e+00 6.40376508e-02 1.15624988e+00 -5.02203345e-01
6.63469255e-01 8.42956543e-01 9.65123713e-01 2.75302649e-01
3.12250465e-01 8.16606104e-01 1.45676804e+00 -5.39338410e-01
-2.30565555e-02 -5.02984166e-01 5.68757415e-01 9.76595044e-01
4.45831656e-01 -2.78955966e-01 -8.63875449e-01 -1.73999015e-02
8.15140963e-01 -3.74754705e-02 -5.00109553e-01 -8.83129686e-02
-1.43167686e+00 4.36847627e-01 4.08412457e-01 4.17693466e-01
-6.02151573e-01 1.47028238e-01 2.27296591e-01 3.75644743e-01
6.15949094e-01 -3.53884418e-03 1.78080618e-01 4.01485682e-01
-9.38549042e-01 -1.73098326e-01 4.85190749e-01 9.38905001e-01
8.60187650e-01 3.23484212e-01 -5.14574572e-02 6.31414056e-01
5.33528209e-01 3.77538711e-01 1.73665553e-01 -9.02557731e-01
2.53454864e-01 8.38476300e-01 -4.63306278e-01 -1.79904628e+00
-5.99921942e-01 -1.55404374e-01 -1.07777500e+00 9.46969539e-02
2.61531733e-02 2.47557610e-01 -6.93191826e-01 8.46741617e-01
3.21811706e-01 4.65036988e-01 -1.10809170e-01 8.25792849e-01
1.10508406e+00 8.49015653e-01 -1.37423515e-01 -4.20800626e-01
1.46932745e+00 -5.01864195e-01 -8.45098615e-01 2.91355606e-02
3.20015341e-01 -7.54281640e-01 9.02786076e-01 6.49355471e-01
-1.11425710e+00 -4.78684694e-01 -1.06630123e+00 1.07177243e-01
-4.02072668e-01 6.17392920e-02 8.36891592e-01 7.14316487e-01
-1.21592736e+00 5.48998475e-01 -6.94830000e-01 -6.19528770e-01
3.54873389e-01 3.28338176e-01 -4.73355323e-01 -2.76730031e-01
-7.57660389e-01 9.92116183e-02 4.18458492e-01 5.90048432e-01
-2.98104048e-01 -3.30704182e-01 -8.25324237e-01 -9.30354819e-02
2.35151365e-01 -7.32936919e-01 4.28882390e-01 -7.72359490e-01
-1.40406036e+00 8.26268435e-01 -6.92285523e-02 -2.37778306e-01
1.28158361e-01 3.72302532e-01 -6.50111854e-01 7.00419426e-01
-2.71089643e-01 1.19901009e-01 1.33414555e+00 -1.14526224e+00
-2.13219509e-01 -5.79671800e-01 1.46642312e-01 1.90532342e-01
-2.32998192e-01 1.58499733e-01 -3.74228239e-01 -5.78257561e-01
8.64240468e-01 -6.78234994e-01 -2.40110025e-01 -1.29643351e-01
-6.86836898e-01 2.36171111e-01 1.18864548e+00 -7.04095185e-01
1.28704500e+00 -2.16175961e+00 2.87609041e-01 3.16786498e-01
5.87827206e-01 6.85950369e-02 -1.38781697e-01 6.28319860e-01
-1.89849988e-01 1.54430121e-01 -3.07181269e-01 -1.17529750e-01
-2.48600721e-01 2.31776819e-01 -2.92849205e-02 5.65568864e-01
2.05362096e-01 8.77945125e-01 -8.20410252e-01 -4.49259967e-01
4.79788214e-01 2.80835569e-01 -2.98718631e-01 -2.05736130e-01
-4.57878187e-02 4.11337793e-01 -4.68981951e-01 8.60080898e-01
8.92316997e-01 -4.31333244e-01 1.72794536e-01 -8.26015651e-01
-2.52731562e-01 -5.07627726e-01 -1.20612407e+00 1.39117861e+00
2.56515056e-01 8.07313085e-01 3.90568912e-01 -1.35972655e+00
7.22882509e-01 1.91222072e-01 8.58127594e-01 -4.52964693e-01
4.77104150e-02 -1.77780718e-01 -2.87453562e-01 -9.23139930e-01
7.05589235e-01 -1.09176271e-01 3.98418397e-01 9.47642624e-02
1.48879871e-01 -4.37971652e-01 7.11068869e-01 4.42448497e-01
1.01975465e+00 -1.62737355e-01 1.23043150e-01 -3.66208285e-01
9.80508387e-01 -8.87286849e-03 2.09642678e-01 4.45207417e-01
-1.40899971e-01 6.63761318e-01 5.28885782e-01 -2.47064739e-01
-6.23092771e-01 -6.50678277e-01 7.20062256e-02 5.87358892e-01
2.29038328e-01 -7.29177237e-01 -5.65703988e-01 -1.94784150e-01
-1.60523236e-01 -3.60524803e-02 -2.37519726e-01 2.03589290e-01
-1.62750959e-01 -1.36787236e+00 3.97435725e-01 -8.69767517e-02
7.90916324e-01 -7.04052866e-01 -1.08879320e-01 1.99538544e-01
-1.35071412e-01 -1.56928289e+00 -1.21404566e-01 -3.65137756e-01
-1.16846073e+00 -1.45262635e+00 -3.54437470e-01 -7.02964664e-01
9.47931170e-01 1.03862679e+00 9.45186794e-01 2.85474151e-01
-5.48783362e-01 9.64555919e-01 -6.46566987e-01 -2.29956895e-01
-3.25564355e-01 -7.14163482e-01 -1.64402634e-01 6.21774435e-01
-9.94514301e-02 -5.77963829e-01 -5.91431558e-01 3.27033699e-02
-1.33599806e+00 3.89784575e-01 3.47668171e-01 3.83452535e-01
8.09052706e-01 8.01216543e-01 -1.02929987e-01 -1.06164646e+00
9.03467417e-01 -1.66989654e-01 -7.73030281e-01 3.36295903e-01
-3.60849112e-01 -3.60215753e-01 3.45143139e-01 -5.10257222e-02
-7.29663134e-01 -2.78342307e-01 2.97645479e-01 -3.62728268e-01
1.96317270e-01 1.18844664e+00 8.66380632e-02 -7.44575679e-01
2.58177370e-01 2.92352766e-01 6.67420551e-02 -2.59226680e-01
3.26094598e-01 2.43438765e-01 4.53822702e-01 -3.90581816e-01
6.86333656e-01 6.15683317e-01 7.33242691e-01 -1.65559614e+00
-3.57525527e-01 -5.83088338e-01 -6.97698832e-01 -8.70414138e-01
9.29895103e-01 -6.80247664e-01 -8.49159181e-01 6.64159060e-01
-9.98911440e-01 -1.46745309e-01 -7.68805593e-02 5.58789074e-01
-1.92320019e-01 9.64931309e-01 -7.46215820e-01 -7.92675972e-01
-2.76922528e-02 -1.21412337e+00 9.57887769e-01 1.34696782e-01
4.33634371e-01 -1.29452455e+00 -2.19926998e-01 6.13268852e-01
2.21638665e-01 6.55178487e-01 1.05853403e+00 2.18534231e-01
-7.73091137e-01 -2.61105925e-01 -4.20750141e-01 4.99556214e-01
8.23482275e-02 4.29163963e-01 -8.12365115e-01 -2.55631268e-01
2.24617779e-01 2.88423866e-01 5.23413122e-01 6.32582605e-01
1.31743670e+00 -2.49770060e-01 -1.26755713e-02 7.87243962e-01
1.77756655e+00 2.51488034e-02 7.35882878e-01 1.13873623e-01
1.12273765e+00 8.21841240e-01 2.23637015e-01 3.04493546e-01
3.03012729e-01 2.84177870e-01 6.71248853e-01 -2.13175669e-01
-4.46666241e-01 2.05038056e-01 5.60495138e-01 1.68441796e+00
-4.18846190e-01 -3.35049212e-01 -6.81996524e-01 -5.71936630e-02
-1.68573618e+00 -9.80158746e-01 -1.20569849e+00 1.92444098e+00
-4.11984734e-02 -1.74647793e-01 -4.34391461e-02 5.85129201e-01
9.01442468e-01 5.24363101e-01 -1.98896334e-01 -2.22576350e-01
-5.81335723e-01 1.50285825e-01 6.54317319e-01 4.65691298e-01
-1.01199675e+00 6.26288116e-01 6.94506836e+00 5.15884519e-01
-1.36417663e+00 -2.91523449e-02 3.94839704e-01 4.64168847e-01
-2.24892825e-01 2.06816763e-01 -5.35883084e-02 2.63786912e-01
5.39771974e-01 -3.09688985e-01 6.30433679e-01 2.33046487e-01
3.68250221e-01 -7.23098963e-02 -7.21354067e-01 1.34312212e+00
-8.93465430e-02 -1.35858679e+00 4.78470117e-01 1.13848433e-01
6.63577735e-01 -1.14548057e-01 -5.33339344e-02 -3.44925255e-01
-8.61148611e-02 -6.24729693e-01 5.15939772e-01 5.23970187e-01
4.29700643e-01 -2.18421504e-01 3.71680886e-01 2.38717441e-02
-1.59596229e+00 7.16824681e-02 -6.51496708e-01 -1.02479115e-01
1.97636366e-01 1.00052977e+00 -6.57118022e-01 1.40645170e+00
4.90151316e-01 1.16619086e+00 -7.61609614e-01 1.17700648e+00
-6.73574978e-04 5.80635548e-01 -7.79683366e-02 1.58828825e-01
2.76765049e-01 -1.00037420e+00 9.96161282e-01 1.24372876e+00
4.15244967e-01 4.27957177e-01 2.07817644e-01 6.47408426e-01
2.04340816e-01 2.37643510e-01 -8.39643478e-01 -5.28772712e-01
-1.20024391e-01 1.54429948e+00 -1.27710783e+00 -1.86869174e-01
-6.84877157e-01 8.15184832e-01 -4.11353827e-01 7.15185046e-01
-3.52305591e-01 -1.99402809e-01 3.08066726e-01 7.78361270e-03
-5.58076389e-02 -7.74545908e-01 -1.47045488e-02 -1.39954722e+00
-1.49862453e-01 -7.58378386e-01 5.25123596e-01 -1.20963013e+00
-1.21103752e+00 4.30608958e-01 1.37100136e-02 -1.41760778e+00
3.75264525e-01 -1.06586206e+00 -5.43779612e-01 3.66203696e-01
-1.46620464e+00 -1.29056990e+00 -8.86208653e-01 1.00058150e+00
2.07182556e-01 3.30760516e-02 4.26882446e-01 3.29083383e-01
-7.34826386e-01 -3.73031676e-01 -7.80290887e-02 2.10230537e-02
-9.27627012e-02 -1.27680433e+00 3.04907952e-02 1.39779294e+00
4.24924552e-01 2.95161843e-01 6.28254831e-01 -6.00576699e-01
-2.30788970e+00 -1.01193917e+00 6.50067031e-02 1.39907241e-01
9.34727490e-01 -4.87844693e-03 -8.03888559e-01 6.26326799e-01
1.83843568e-01 1.94836617e-01 6.51628613e-01 -3.93414527e-01
-6.90649599e-02 -3.34346920e-01 -8.97333741e-01 6.03137791e-01
1.01535761e+00 -7.34766245e-01 2.41200477e-02 7.90781617e-01
4.64662284e-01 -1.35222271e-01 -1.25587130e+00 4.73371804e-01
2.37562895e-01 -1.20275545e+00 1.18089449e+00 -2.24513933e-01
-1.29762935e-02 -6.66181803e-01 -4.00816262e-01 -1.08273649e+00
-5.40182233e-01 -8.78561318e-01 1.68561816e-01 9.44305956e-01
-8.81710947e-02 -7.00814128e-01 6.83251798e-01 1.63962021e-01
-2.73818105e-01 -3.97009104e-01 -5.55755258e-01 -5.78928530e-01
-8.11175406e-01 -9.08046722e-01 4.65620756e-01 1.07021892e+00
-7.65189994e-03 3.67787592e-02 -1.02295719e-01 7.20331311e-01
1.13332212e+00 1.42609894e-01 5.12766957e-01 -1.26306248e+00
-2.45536298e-01 -5.36571503e-01 -1.04128730e+00 -4.91070598e-01
-2.21395865e-01 -1.24387288e+00 -4.44468558e-01 -1.85249484e+00
-1.38426214e-01 -5.17242178e-02 -2.57944316e-01 -9.86970961e-03
1.66572765e-01 7.58978307e-01 4.26212758e-01 2.46213540e-01
-5.07698357e-01 2.15461433e-01 1.48720944e+00 -3.96938562e-01
-1.72577295e-02 -1.85681462e-01 -2.97860712e-01 6.79424822e-01
4.43526179e-01 -5.95732071e-02 -6.94833279e-01 -3.52915734e-01
6.36018634e-01 2.36334071e-01 4.13050473e-01 -9.75708306e-01
4.31863785e-01 -9.87378657e-02 9.30468589e-02 -5.42768121e-01
3.23138714e-01 -8.83447587e-01 5.87480843e-01 3.89610052e-01
2.45584384e-01 2.32837796e-01 5.66548556e-02 6.27794206e-01
-6.11827433e-01 -1.88240305e-01 3.80872548e-01 -2.46374577e-01
-1.03089297e+00 5.08501172e-01 -4.11332458e-01 -6.39862180e-01
9.50281918e-01 -5.03854811e-01 -5.81117988e-01 -2.70540833e-01
-1.06910646e+00 -4.47168425e-02 1.61431402e-01 -1.44581380e-03
1.06534863e+00 -1.15478170e+00 -5.67557693e-01 3.57559234e-01
2.05141250e-02 -1.31294474e-01 2.85126835e-01 1.23143876e+00
-1.25440311e+00 6.92593008e-02 -2.25276470e-01 -9.61569250e-01
-1.67595363e+00 4.03656632e-01 4.79612052e-02 -5.56518286e-02
-7.66213477e-01 7.14589894e-01 2.95698076e-01 4.69438136e-02
-2.62454540e-01 -7.15298772e-01 -3.69247377e-01 1.97775975e-01
3.59048158e-01 5.98612785e-01 3.05176854e-01 -8.48056734e-01
-2.03240588e-01 8.33743513e-01 4.15950567e-01 6.48853462e-03
1.38436484e+00 -4.83505189e-01 -9.85323310e-01 6.81265175e-01
1.01298356e+00 -1.01244353e-01 -8.77378643e-01 7.59454146e-02
1.54958770e-01 -4.53735948e-01 4.29668278e-01 -8.43587667e-02
-1.30524135e+00 8.34647536e-01 4.00912702e-01 1.11522019e+00
1.54382730e+00 -4.05442715e-01 2.90579468e-01 2.04151183e-01
5.81220269e-01 -6.85718954e-01 3.69552851e-01 2.26714998e-01
8.90581071e-01 -6.49352491e-01 4.28654522e-01 -9.80897665e-01
-3.91613632e-01 1.67863417e+00 -1.77385718e-01 -7.70294964e-02
8.84893358e-01 -4.58879881e-02 -2.54235357e-01 -8.44830155e-01
-2.71658927e-01 -3.57091159e-01 4.38824028e-01 5.50977409e-01
4.31661367e-01 2.93324113e-01 -2.19570592e-01 -1.55103475e-01
6.30031154e-02 -8.28676745e-02 9.10156608e-01 9.97744977e-01
-3.58554244e-01 -9.76958752e-01 -7.21013308e-01 4.95704025e-01
-4.89895403e-01 6.49972409e-02 -4.40617442e-01 5.88992596e-01
-1.77195877e-01 1.15830851e+00 -3.06348622e-01 -5.79263866e-01
1.85485050e-01 -6.73629940e-02 8.29161525e-01 -4.20595914e-01
-3.18181515e-01 1.92974731e-01 -1.48243615e-02 -8.43046248e-01
-1.12106645e+00 -5.28802454e-01 -9.85906363e-01 -3.57988089e-01
-2.89191604e-01 -5.73062338e-02 1.04085040e+00 7.19065785e-01
1.75830320e-01 7.12093353e-01 5.20187557e-01 -9.44637001e-01
4.26774889e-01 -5.64970136e-01 -1.12758482e+00 4.02936041e-01
3.45979244e-01 -2.97201842e-01 -2.55978316e-01 3.50248963e-01] | [10.12912654876709, -1.946967363357544] |
a077d184-2e11-479c-960f-5d43823c3494 | mask-guided-portrait-editing-with-conditional-1 | 1905.10346 | null | https://arxiv.org/abs/1905.10346v1 | https://arxiv.org/pdf/1905.10346v1.pdf | Mask-Guided Portrait Editing with Conditional GANs | Portrait editing is a popular subject in photo manipulation. The Generative Adversarial Network (GAN) advances the generating of realistic faces and allows more face editing. In this paper, we argue about three issues in existing techniques: diversity, quality, and controllability for portrait synthesis and editing. To address these issues, we propose a novel end-to-end learning framework that leverages conditional GANs guided by provided face masks for generating faces. The framework learns feature embeddings for every face component (e.g., mouth, hair, eye), separately, contributing to better correspondences for image translation, and local face editing. With the mask, our network is available to many applications, like face synthesis driven by mask, face Swap+ (including hair in swapping), and local manipulation. It can also boost the performance of face parsing a bit as an option of data augmentation. | ['Jianmin Bao', 'Dong Chen', 'Hao Yang', 'Shuyang Gu', 'Lu Yuan', 'Fang Wen'] | 2019-05-24 | mask-guided-portrait-editing-with-conditional | http://openaccess.thecvf.com/content_CVPR_2019/html/Gu_Mask-Guided_Portrait_Editing_With_Conditional_GANs_CVPR_2019_paper.html | http://openaccess.thecvf.com/content_CVPR_2019/papers/Gu_Mask-Guided_Portrait_Editing_With_Conditional_GANs_CVPR_2019_paper.pdf | cvpr-2019-6 | ['face-parsing'] | ['computer-vision'] | [ 5.65803170e-01 5.95930517e-01 -1.33525178e-01 -5.48932433e-01
-4.72130507e-01 -7.33057559e-01 5.61048567e-01 -9.78124917e-01
3.53045285e-01 7.87060022e-01 3.33367556e-01 5.26446924e-02
3.73652965e-01 -9.46431756e-01 -1.11192334e+00 -7.01387167e-01
4.17123139e-01 2.58090109e-01 -6.46069705e-01 -3.33893269e-01
-2.86286324e-01 7.95079648e-01 -1.53171897e+00 4.11926776e-01
7.83790231e-01 9.05369520e-01 -2.58971632e-01 4.99340355e-01
-2.49742866e-01 5.49826384e-01 -5.93961954e-01 -1.10255289e+00
4.30279166e-01 -6.68414056e-01 -4.53263581e-01 2.85107374e-01
8.12320948e-01 -7.96584427e-01 -3.58572185e-01 8.67929995e-01
5.75438738e-01 -3.28682870e-01 5.28966069e-01 -1.76787555e+00
-1.20904827e+00 7.65635669e-01 -8.02941203e-01 -7.25358844e-01
3.28541249e-01 6.06576860e-01 6.39102042e-01 -1.00301099e+00
8.50039721e-01 1.66753376e+00 5.89457810e-01 1.32571626e+00
-9.99629080e-01 -1.03110516e+00 -2.85047088e-02 -5.39834261e-01
-1.09610474e+00 -8.97990465e-01 9.07962680e-01 -3.16044688e-01
2.53821641e-01 4.35021996e-01 6.99184179e-01 1.61354649e+00
-1.07335538e-01 7.06650436e-01 8.92357051e-01 -2.73169070e-01
-1.34675786e-01 1.67730991e-02 -7.26079524e-01 9.06833112e-01
-2.71482253e-03 1.90746084e-01 -6.09829605e-01 2.68531740e-02
1.26534510e+00 4.74387743e-02 -1.73862904e-01 -7.68173337e-02
-9.94008005e-01 7.05728471e-01 4.44717437e-01 -2.84903556e-01
-1.45865068e-01 4.96001989e-01 9.34115201e-02 1.89931184e-01
4.67687041e-01 3.17024320e-01 -1.89041182e-01 2.72770047e-01
-8.14113915e-01 3.93297255e-01 5.43112099e-01 1.39550889e+00
6.78394079e-01 4.78782564e-01 -6.23636782e-01 8.60435903e-01
2.22131714e-01 6.66228116e-01 -5.37285917e-02 -1.18242002e+00
4.18709964e-01 5.67980647e-01 -2.36039892e-01 -7.25314617e-01
6.97179660e-02 6.45533055e-02 -9.64649320e-01 4.41660881e-01
7.35708233e-03 -4.73401040e-01 -1.22650611e+00 2.12746143e+00
6.35269940e-01 2.77811825e-01 -1.54716030e-01 5.48118532e-01
1.05475664e+00 4.66207653e-01 2.03435961e-02 5.43749779e-02
1.27486491e+00 -1.16943598e+00 -1.02163339e+00 -3.49980295e-01
5.26535027e-02 -7.94699609e-01 1.10333037e+00 -7.12057874e-02
-1.32679951e+00 -5.77656627e-01 -7.74954081e-01 -5.12858391e-01
-1.69564724e-01 2.64568776e-01 8.30433190e-01 7.90517092e-01
-1.21931529e+00 5.01488090e-01 -4.70172554e-01 -1.93044305e-01
1.08062494e+00 5.35346627e-01 -7.44360983e-01 -1.21265173e-01
-8.18336725e-01 4.67132986e-01 -1.52710706e-01 1.20388985e-01
-9.95597064e-01 -9.49641049e-01 -9.49414074e-01 -3.04745976e-02
2.60373890e-01 -1.13011122e+00 9.91835773e-01 -1.30910051e+00
-2.10545492e+00 1.06344855e+00 -8.16898569e-02 -5.68867773e-02
7.72289932e-01 -1.27721861e-01 -2.30992794e-01 -1.39289334e-01
-1.36726322e-02 1.41476369e+00 1.51646841e+00 -1.32360399e+00
-9.97044370e-02 -4.43560928e-01 6.44430518e-02 -6.73388913e-02
-3.03830236e-01 7.02222660e-02 -5.34318030e-01 -9.28078115e-01
-4.23588127e-01 -9.36358929e-01 4.00437079e-02 6.93157732e-01
-6.58403516e-01 1.25330135e-01 1.09467304e+00 -7.93627381e-01
7.75907755e-01 -2.41222167e+00 2.85771221e-01 -4.53997068e-02
1.53733194e-01 1.82874992e-01 -5.71585596e-01 2.74488568e-01
-1.98309779e-01 6.02295101e-01 -3.73835385e-01 -7.44538367e-01
-2.08445042e-02 1.77482560e-01 -4.12892401e-01 1.73441187e-01
9.24402654e-01 1.40751696e+00 -3.83409113e-01 -5.09822190e-01
-2.02791356e-02 8.64556372e-01 -7.93858230e-01 3.18996638e-01
-4.19753075e-01 9.07950282e-01 -3.00412804e-01 1.13055170e+00
1.09991682e+00 2.19377622e-01 7.11573735e-02 -2.39531323e-01
2.39126474e-01 -1.89949334e-01 -7.37563074e-01 1.75331199e+00
-3.69598597e-01 4.04937655e-01 4.28695261e-01 -1.71426088e-01
9.83697236e-01 2.40002722e-01 2.07045004e-01 -3.62342089e-01
2.42721438e-01 -7.38571510e-02 -2.66214520e-01 -2.79224873e-01
2.43425682e-01 8.60468447e-02 1.17611311e-01 3.33052367e-01
1.23742782e-01 -3.83306980e-01 -1.31324604e-01 -4.07979786e-02
7.50830710e-01 5.24558961e-01 -9.96172428e-02 4.75576147e-03
1.74549267e-01 -4.32834744e-01 5.83761454e-01 5.94695620e-02
1.29400343e-01 9.51125920e-01 6.54637873e-01 -3.33874196e-01
-1.32967269e+00 -9.32533264e-01 1.44250780e-01 9.57411885e-01
-2.83817410e-01 -2.40201965e-01 -1.27877688e+00 -7.61427402e-01
1.35032192e-01 4.81780827e-01 -8.03918779e-01 -3.33255291e-01
-9.19102252e-01 -2.46858910e-01 7.74704635e-01 6.34472549e-01
5.44877768e-01 -1.17430866e+00 7.61887282e-02 -1.27020180e-01
1.88103467e-02 -1.06001246e+00 -9.17735159e-01 -4.37645406e-01
-5.79647124e-01 -7.18351424e-01 -7.24777997e-01 -8.53197813e-01
9.60425556e-01 -3.42523642e-02 1.10763180e+00 2.91855544e-01
-3.18306357e-01 2.93588400e-01 -9.90201756e-02 -5.56217730e-01
-5.70473790e-01 -3.92585024e-02 -3.31334434e-02 2.32368931e-01
-3.75369012e-01 -7.86680102e-01 -6.23340964e-01 2.84443438e-01
-1.13101709e+00 1.66334599e-01 5.20667970e-01 8.07671785e-01
5.28048813e-01 -4.56295729e-01 5.16316712e-01 -1.14199829e+00
4.87062991e-01 -2.89014637e-01 -5.92177689e-01 3.50644678e-01
-1.18862763e-01 -1.18052870e-01 5.84106505e-01 -5.51431358e-01
-1.20171750e+00 2.01780364e-01 -3.49263906e-01 -6.28988028e-01
1.29832610e-01 -2.94634253e-01 -8.10354352e-01 -4.62332100e-01
4.45057869e-01 -3.27273943e-02 4.09098148e-01 -2.43390813e-01
9.02260423e-01 3.71469826e-01 5.01593649e-01 -6.36643350e-01
1.13131511e+00 5.58696926e-01 5.59859425e-02 -5.82817018e-01
-4.37746495e-01 7.29741037e-01 -6.16584420e-01 -6.81201369e-02
8.89176607e-01 -1.05631161e+00 -6.94526196e-01 5.44374108e-01
-1.20505798e+00 -3.67869169e-01 -5.21421850e-01 -2.41611063e-01
-5.78647256e-01 -1.04826666e-01 -4.67098951e-01 -6.52838767e-01
-5.30956328e-01 -1.18183029e+00 1.55632460e+00 4.60839063e-01
-1.62412763e-01 -5.78441739e-01 -4.06947136e-01 4.68081295e-01
5.13305843e-01 7.96756268e-01 8.47570062e-01 6.47419319e-02
-8.70801389e-01 3.84303294e-02 -6.63293824e-02 4.11680698e-01
3.55393708e-01 5.44743359e-01 -1.13582504e+00 -2.46289641e-01
-2.39678189e-01 -4.55051631e-01 6.58605635e-01 1.40436441e-01
1.43318021e+00 -6.83187306e-01 -4.25244093e-01 1.23858178e+00
9.94182944e-01 7.61757344e-02 1.06007493e+00 -2.75282472e-01
1.12836075e+00 6.61065936e-01 3.39921474e-01 3.62621754e-01
1.36714578e-01 4.19106364e-01 5.65777123e-01 -3.64584357e-01
-6.32877827e-01 -6.15863740e-01 5.14887393e-01 2.43051797e-01
5.57926309e-04 -4.11356151e-01 -4.10552680e-01 2.62108624e-01
-1.26694715e+00 -7.03404784e-01 3.37616146e-01 1.67965007e+00
9.98045206e-01 -4.99766588e-01 7.77001008e-02 -2.66820371e-01
9.80424345e-01 2.43130565e-01 -7.56623745e-01 -3.61250818e-01
-1.05567932e-01 6.85469508e-01 3.24788094e-01 3.50226432e-01
-6.78222895e-01 1.32776487e+00 6.76381159e+00 6.39930248e-01
-1.22198093e+00 -3.46022774e-03 9.59717572e-01 -1.89292565e-01
-6.84683323e-01 -1.42973930e-01 -9.73569930e-01 5.22630811e-01
3.24434489e-01 1.89387456e-01 8.37558806e-01 8.11832368e-01
-1.36218056e-01 5.64260840e-01 -1.31927776e+00 1.02339697e+00
2.87270546e-01 -1.56760418e+00 5.97516894e-01 1.61845803e-01
9.37790155e-01 -7.36879468e-01 6.10499501e-01 2.01410040e-01
2.88317412e-01 -1.49248266e+00 7.31648028e-01 4.00450498e-01
1.56619501e+00 -9.44298565e-01 1.11294523e-01 -6.29743710e-02
-8.60906601e-01 -1.51443994e-02 -1.06122484e-02 4.12915379e-01
2.05475748e-01 2.75352389e-01 -5.97676218e-01 3.53712976e-01
4.01534110e-01 3.62684995e-01 -3.69099051e-01 2.02633917e-01
-5.10716558e-01 2.70343035e-01 -1.65746883e-01 3.95477980e-01
-2.01566383e-01 -3.23048383e-01 3.48806202e-01 6.62418664e-01
4.94198143e-01 -7.90100768e-02 -1.81522921e-01 1.37523127e+00
-9.32420671e-01 -2.78957188e-01 -8.31959844e-01 -4.23568815e-01
7.27504849e-01 1.43509185e+00 -2.67605633e-01 -4.04416770e-02
-4.14804332e-02 1.25563037e+00 3.58469754e-01 2.11620629e-01
-1.15456736e+00 -1.70124099e-01 1.08619118e+00 3.15902025e-01
2.36361831e-01 -2.18841992e-03 -3.51970553e-01 -9.97964025e-01
1.70013756e-01 -1.30336130e+00 -1.83231726e-01 -7.49139667e-01
-1.38769078e+00 6.44502997e-01 -2.86218584e-01 -8.14166129e-01
-1.10661268e-01 -5.73425829e-01 -9.75170255e-01 8.76933575e-01
-1.31422067e+00 -1.97901416e+00 -4.62260783e-01 6.57143474e-01
4.22272533e-01 -2.77709365e-01 7.82327533e-01 3.32363367e-01
-7.73576617e-01 1.22651005e+00 -3.62023503e-01 2.63504416e-01
9.06551480e-01 -8.12297702e-01 9.79810536e-01 6.59000754e-01
-1.13948949e-01 7.13802636e-01 2.54164696e-01 -6.83144867e-01
-1.72505844e+00 -1.53818822e+00 3.09792548e-01 -6.08953059e-01
5.89715168e-02 -7.48588324e-01 -4.47094858e-01 1.02361274e+00
3.59550655e-01 8.12376589e-02 6.13989234e-01 -3.33817363e-01
-4.32056040e-01 -2.51062900e-01 -1.69668090e+00 9.67014372e-01
1.50191188e+00 -5.06763339e-01 1.69347569e-01 4.79035348e-01
1.08765996e+00 -6.79834127e-01 -7.57247925e-01 5.21924615e-01
6.16578877e-01 -8.12565327e-01 1.06917441e+00 -5.64331472e-01
9.22920227e-01 -9.82353687e-02 1.35023102e-01 -1.25536478e+00
-2.56325513e-01 -1.21564758e+00 -1.00195957e-02 1.92475200e+00
3.81155044e-01 -6.24101400e-01 8.63166571e-01 6.86754823e-01
-1.17071614e-01 -6.36905909e-01 -6.99970782e-01 -4.49464977e-01
2.34524816e-01 1.47925958e-01 1.42933166e+00 9.93269086e-01
-7.64649689e-01 1.57113627e-01 -8.35822642e-01 -6.03402853e-02
4.58103418e-01 -7.09728524e-02 1.27906466e+00 -8.46768796e-01
-1.87562332e-01 -2.99172193e-01 -4.86860909e-02 -8.05194318e-01
3.76729518e-01 -9.06910717e-01 -2.73229957e-01 -1.26330948e+00
-3.19237523e-02 -3.58095497e-01 5.98347187e-01 7.16200411e-01
-1.51432261e-01 4.11369324e-01 3.94055307e-01 -2.04493254e-01
3.02261323e-01 6.64970696e-01 1.94780385e+00 -1.71357185e-01
9.34275091e-02 -2.60288924e-01 -1.24862921e+00 5.83797574e-01
7.85444319e-01 -3.19027543e-01 -3.63568604e-01 -8.81623566e-01
3.10459286e-02 -8.87677521e-02 3.89353871e-01 -6.43007278e-01
-9.75875929e-02 -1.73229799e-01 7.58056164e-01 -8.05340260e-02
6.34441435e-01 -7.74130285e-01 4.42851305e-01 1.20321229e-01
-2.69207329e-01 6.34296685e-02 2.00124517e-01 3.26012194e-01
-2.76437011e-02 1.50190338e-01 1.05790365e+00 -2.68235773e-01
-1.85223252e-01 8.65035772e-01 3.93112779e-01 -1.02301493e-01
1.22916317e+00 -1.95616111e-01 -4.23025519e-01 -3.84647071e-01
-4.94412214e-01 1.17631301e-01 6.53882027e-01 5.78079879e-01
5.90657353e-01 -1.77222133e+00 -9.22753632e-01 6.70648575e-01
-1.26264095e-01 2.98432618e-01 1.88961089e-01 3.53634566e-01
-5.20994425e-01 -2.66187787e-01 -5.22463262e-01 -3.82304519e-01
-1.27836502e+00 4.38662529e-01 2.12365881e-01 1.38167202e-01
-3.40901017e-01 1.16687179e+00 4.16292995e-01 -6.68905675e-01
4.22266945e-02 -1.21681064e-01 1.76821649e-01 -8.36682096e-02
5.49641848e-01 1.72338471e-01 -3.26543689e-01 -5.43438911e-01
-7.76256472e-02 7.53285348e-01 3.92131396e-02 5.12335859e-02
1.33152640e+00 1.45803556e-01 -4.58647937e-01 -1.87026054e-01
1.08279526e+00 3.74257505e-01 -1.52998710e+00 2.05783084e-01
-8.01484823e-01 -7.39572406e-01 -4.09748465e-01 -8.31959248e-01
-1.48455238e+00 7.72440493e-01 4.00414675e-01 -2.52860427e-01
1.10078478e+00 -1.04113922e-01 1.05110300e+00 -7.05064610e-02
2.04628348e-01 -6.76578045e-01 1.17923252e-01 2.40331039e-01
1.38402164e+00 -1.15479875e+00 -2.51281828e-01 -8.68815899e-01
-8.02407384e-01 9.92278278e-01 8.37767720e-01 -1.55924514e-01
3.65592360e-01 6.83226049e-01 -2.63549518e-02 -1.97527613e-02
-4.53687996e-01 2.25966826e-01 1.70158863e-01 8.90531003e-01
3.98629904e-01 5.54053634e-02 1.23303153e-01 3.38349134e-01
-5.03266692e-01 -6.30146712e-02 1.24012582e-01 7.75722623e-01
1.28394574e-01 -1.48197615e+00 -3.04570913e-01 3.00074667e-01
-4.36356276e-01 -1.89841107e-01 -8.87047172e-01 8.26547384e-01
5.29209912e-01 6.04459047e-01 5.26041724e-02 -2.90902972e-01
2.02617899e-01 9.16606262e-02 7.98895717e-01 -6.14556968e-01
-6.50385320e-01 -1.49941653e-01 -4.01777253e-02 -6.24705911e-01
-1.72676057e-01 -2.97808677e-01 -9.87146139e-01 -6.50319815e-01
-1.52873576e-01 -3.86407882e-01 7.41482973e-01 5.73229253e-01
7.35536993e-01 6.91268981e-01 9.23853636e-01 -8.74670029e-01
-2.99154818e-01 -8.35305870e-01 -4.45170134e-01 3.59793067e-01
2.73697287e-01 -3.77804250e-01 -2.55330145e-01 3.55407715e-01] | [12.477517127990723, -0.22139418125152588] |
f69edb0c-b3ae-4b81-b60b-ef193b56da4a | can-language-models-solve-graph-problems-in | 2305.10037 | null | https://arxiv.org/abs/2305.10037v1 | https://arxiv.org/pdf/2305.10037v1.pdf | Can Language Models Solve Graph Problems in Natural Language? | Large language models (LLMs) are increasingly adopted for a variety of tasks with implicit graphical structures, such as planning in robotics, multi-hop question answering or knowledge probing, structured commonsense reasoning, and more. While LLMs have advanced the state-of-the-art on these tasks with structure implications, whether LLMs could explicitly process textual descriptions of graphs and structures, map them to grounded conceptual spaces, and perform structured operations remains underexplored. To this end, we propose NLGraph (Natural Language Graph), a comprehensive benchmark of graph-based problem solving designed in natural language. NLGraph contains 29,370 problems, covering eight graph reasoning tasks with varying complexity from simple tasks such as connectivity and shortest path up to complex problems such as maximum flow and simulating graph neural networks. We evaluate LLMs (GPT-3/4) with various prompting approaches on the NLGraph benchmark and find that 1) language models do demonstrate preliminary graph reasoning abilities, 2) the benefit of advanced prompting and in-context learning diminishes on more complex graph problems, while 3) LLMs are also (un)surprisingly brittle in the face of spurious correlations in graph and problem settings. We then propose Build-a-Graph Prompting and Algorithmic Prompting, two instruction-based approaches to enhance LLMs in solving natural language graph problems. Build-a-Graph and Algorithmic prompting improve the performance of LLMs on NLGraph by 3.07% to 16.85% across multiple tasks and settings, while how to solve the most complicated graph reasoning tasks in our setup with language models remains an open research question. The NLGraph benchmark and evaluation code are available at https://github.com/Arthur-Heng/NLGraph. | ['Yulia Tsvetkov', 'Xiaochuang Han', 'Zhaoxuan Tan', 'Tianxing He', 'Shangbin Feng', 'Heng Wang'] | 2023-05-17 | null | null | null | null | ['multi-hop-question-answering'] | ['knowledge-base'] | [ 2.55401641e-01 7.58383572e-01 -2.22339809e-01 -1.42096192e-01
-4.63699222e-01 -8.69076848e-01 5.75455606e-01 6.63567305e-01
-1.83582664e-01 5.06093085e-01 2.56368488e-01 -9.84498620e-01
-4.91213113e-01 -9.73540723e-01 -8.71155798e-01 -1.62755791e-02
-5.80245316e-01 8.54531944e-01 1.93605170e-01 -3.92211348e-01
4.78811115e-01 4.18523729e-01 -1.07016563e+00 2.06514329e-01
1.09398246e+00 5.46615541e-01 2.64881104e-01 8.71814191e-01
-5.14046013e-01 1.38452637e+00 -2.24975184e-01 -3.58818024e-01
8.29546303e-02 -2.28505552e-01 -1.42887890e+00 -1.44420385e-01
6.98270857e-01 -1.61444899e-02 -4.66089189e-01 1.00844359e+00
8.16026405e-02 4.28258061e-01 3.53750408e-01 -1.62484491e+00
-7.27069318e-01 1.15106022e+00 -3.14254552e-01 1.81959048e-01
8.05946052e-01 6.22631192e-01 1.40999591e+00 -4.38171685e-01
6.93371296e-01 1.67405725e+00 5.45465350e-01 3.95755440e-01
-1.29312003e+00 -3.65937471e-01 6.30149305e-01 3.34668130e-01
-1.19812393e+00 -1.15429401e-01 6.01781547e-01 -3.58477414e-01
1.45444810e+00 3.79945785e-02 6.76016450e-01 7.99830794e-01
2.30535567e-01 6.46224201e-01 1.02107239e+00 -3.14988106e-01
2.11124837e-01 -5.02659678e-01 5.44415176e-01 1.50451887e+00
4.82313454e-01 4.38746484e-03 -4.94859308e-01 -2.03080282e-01
7.52076983e-01 -3.75025839e-01 -2.05362797e-01 -4.08966243e-01
-1.17734730e+00 8.87904584e-01 7.72736192e-01 1.19035289e-01
-1.60614401e-01 7.05574632e-01 2.82242149e-01 4.69500959e-01
-5.27006127e-02 1.11184776e+00 -4.03409749e-01 -1.68043226e-02
-2.86018431e-01 5.43717980e-01 1.39916193e+00 1.23401356e+00
9.50890481e-01 4.58447859e-02 -4.64149527e-02 3.03538859e-01
2.44309276e-01 2.86549389e-01 8.70018732e-03 -1.03883994e+00
8.73073220e-01 9.00498390e-01 -3.80463868e-01 -1.23124182e+00
-9.29089963e-01 -2.80687481e-01 -6.37039125e-01 -3.46671715e-02
7.32937038e-01 -9.99242738e-02 -9.62648928e-01 1.69690609e+00
1.15605250e-01 1.62617773e-01 5.30719273e-02 7.18805015e-01
1.08828163e+00 6.71870053e-01 2.08682448e-01 2.58154482e-01
1.52828157e+00 -1.11943591e+00 -3.32659870e-01 -9.51343417e-01
1.16116714e+00 -1.29724383e-01 1.51142836e+00 4.86716300e-01
-9.32997704e-01 -2.53250450e-01 -7.77017474e-01 -5.06720841e-01
-4.85780537e-01 -6.15387321e-01 1.12080491e+00 2.39632860e-01
-1.42074013e+00 5.71589291e-01 -6.52898908e-01 -4.45689708e-01
4.48598355e-01 4.56698120e-01 -3.41531843e-01 -6.84825480e-01
-1.15281677e+00 9.00267363e-01 6.95894659e-01 -1.81977808e-01
-9.33800161e-01 -8.16251695e-01 -1.25111592e+00 2.72526383e-01
1.19848394e+00 -9.36912239e-01 1.22120512e+00 -2.37195939e-01
-1.18581223e+00 8.03990602e-01 4.02575061e-02 -6.07784688e-01
2.05541611e-01 -1.31381884e-01 -9.54302624e-02 1.11345008e-01
6.78342357e-02 8.16381574e-01 3.34552050e-01 -9.84006405e-01
-7.54827410e-02 -1.86901376e-01 7.43393898e-01 3.50670844e-01
3.00013989e-01 -4.36381817e-01 -3.47559541e-01 -1.83123365e-01
1.83752432e-01 -1.04413247e+00 -5.70375741e-01 -1.18569098e-01
-6.01603746e-01 -5.84121346e-01 3.88516188e-01 -4.16335672e-01
1.00157416e+00 -1.69525492e+00 3.62712413e-01 2.55623132e-01
8.26164901e-01 -1.05780631e-01 -4.35267240e-01 7.70711303e-01
-4.38259132e-02 3.52324128e-01 -2.78590709e-01 9.59472954e-02
2.38988087e-01 4.55377191e-01 -7.08777308e-02 1.12712860e-01
1.94953009e-01 1.59580553e+00 -1.35790133e+00 -5.20548463e-01
1.68064952e-01 -2.85582602e-01 -8.50285888e-01 9.76357013e-02
-9.44507837e-01 1.90078095e-01 -4.34225112e-01 5.71718693e-01
1.89170748e-01 -8.37763965e-01 4.30918247e-01 1.05233185e-01
3.78416806e-01 5.97508907e-01 -9.82072830e-01 2.06271768e+00
-3.27064693e-01 5.86090147e-01 7.81212673e-02 -8.12086344e-01
7.03612030e-01 -1.81800753e-01 -1.11501880e-01 -8.48876119e-01
-1.69146121e-01 -1.49858102e-01 4.17641908e-01 -6.50535703e-01
5.93293071e-01 1.25423372e-02 -2.74780601e-01 5.06390989e-01
9.79763549e-03 -7.93553770e-01 5.48888206e-01 7.87418962e-01
1.58982301e+00 -3.46386023e-02 5.29074848e-01 -4.92683768e-01
1.15201816e-01 3.78162950e-01 -6.01662025e-02 9.92234409e-01
-3.38885821e-02 -6.33661076e-03 1.04197109e+00 -4.03915197e-01
-5.30692935e-01 -9.17225897e-01 6.58718288e-01 1.40542531e+00
1.78034708e-01 -7.87174642e-01 -5.41104317e-01 -7.09938407e-01
1.30613297e-01 1.09821653e+00 -4.19467807e-01 -3.06242496e-01
-6.53417647e-01 -1.90050170e-01 6.50436103e-01 4.10788208e-01
3.20870489e-01 -1.39412510e+00 -2.69105881e-01 1.12262160e-01
-8.16601068e-02 -1.46497941e+00 -3.14477563e-01 3.07345092e-01
-9.80871856e-01 -1.44873214e+00 1.32225186e-01 -7.75706947e-01
6.66504681e-01 3.07403147e-01 1.73377788e+00 5.78813136e-01
-1.51572168e-01 6.43708408e-01 -2.75690645e-01 -2.50650018e-01
-5.20019710e-01 3.96710336e-01 -3.91190678e-01 -8.99425209e-01
5.76393642e-02 -8.03672910e-01 -1.48739547e-01 6.12905435e-02
-8.15584600e-01 3.12742531e-01 5.54174483e-01 4.31849927e-01
2.65374869e-01 2.14683264e-02 3.49488705e-01 -1.26558161e+00
1.13070583e+00 -6.10321224e-01 -8.53489459e-01 2.22040221e-01
-7.44920671e-01 5.90922296e-01 6.57219946e-01 -2.06534237e-01
-5.20750940e-01 -3.27120215e-01 1.52131483e-01 -1.31500617e-01
-1.29617721e-01 1.06956339e+00 1.37032876e-02 -1.83729202e-01
1.03616560e+00 -3.98668796e-02 3.14831063e-02 -1.74076781e-02
8.05646360e-01 -3.73706877e-01 5.45567691e-01 -1.33217633e+00
9.71730888e-01 -7.09436014e-02 3.90379161e-01 -6.90076590e-01
-9.00749683e-01 -3.01819623e-01 -3.65897954e-01 1.92397878e-01
7.32822180e-01 -7.55723178e-01 -1.09033430e+00 1.20479360e-01
-1.12212789e+00 -1.21130133e+00 -1.18831366e-01 -1.76558774e-02
-5.03220260e-01 5.44089377e-01 -7.66691327e-01 -6.25518203e-01
-2.64947265e-01 -1.07551777e+00 9.20670092e-01 2.51647960e-02
-4.80075091e-01 -1.39301789e+00 -1.25260964e-01 3.89037579e-01
4.37757730e-01 3.45941067e-01 1.63140082e+00 -8.14754725e-01
-1.01752651e+00 2.76107281e-01 -3.92502159e-01 -3.45481247e-01
-1.10384874e-01 -4.72680420e-01 -4.82703537e-01 -1.99456096e-01
-4.72257614e-01 -6.47324502e-01 6.04476810e-01 1.23840973e-01
1.33243096e+00 -5.58515429e-01 -2.56789088e-01 4.67719465e-01
1.26181901e+00 -7.74247795e-02 5.89690924e-01 1.06472813e-01
1.11332655e+00 5.94719946e-01 2.58324325e-01 3.17546986e-02
8.66827726e-01 2.11272731e-01 7.65213370e-01 2.10876465e-01
-9.71896872e-02 -7.13825643e-01 2.31866062e-01 6.14102900e-01
1.91333607e-01 -5.48499346e-01 -1.47819543e+00 3.75237763e-01
-1.78482103e+00 -6.30777240e-01 -4.10402864e-01 1.60400832e+00
6.77247107e-01 3.91696930e-01 4.52663191e-02 -1.46688059e-01
3.65034848e-01 3.74435544e-01 -7.88184822e-01 -4.22608018e-01
2.58346528e-01 4.34850425e-01 4.48189557e-01 9.25796270e-01
-6.42774940e-01 1.46107781e+00 5.40869093e+00 6.41523540e-01
-7.94693470e-01 -3.80559385e-01 4.09817845e-01 3.42629433e-01
-4.67830807e-01 3.53889316e-01 -3.94285411e-01 -2.35210627e-01
9.13455963e-01 -3.63185674e-01 1.17446280e+00 8.63860071e-01
-1.84384704e-01 -1.78058550e-01 -1.47119951e+00 9.22745824e-01
-1.48885906e-01 -1.39714646e+00 1.20705731e-01 5.30726500e-02
5.29470205e-01 2.26843745e-01 -2.87158757e-01 8.06428611e-01
9.37911510e-01 -1.48920488e+00 5.71838737e-01 2.35069081e-01
5.22463322e-01 -3.37718695e-01 2.81589717e-01 5.81702709e-01
-1.31086290e+00 -2.08948310e-02 -9.74280909e-02 -6.40592277e-01
5.59292324e-02 3.34716767e-01 -1.21239758e+00 6.52810574e-01
2.13875100e-01 5.31706929e-01 -9.04626429e-01 7.43523717e-01
-7.33370543e-01 6.34603500e-01 -2.80262053e-01 -4.11811084e-01
5.24055898e-01 -8.70497376e-02 4.63066816e-01 9.87096608e-01
-1.32111132e-01 4.64013427e-01 6.11129284e-01 1.16095257e+00
-1.91079676e-01 -1.22642718e-01 -8.39091837e-01 -7.04357624e-01
4.09100682e-01 1.24334669e+00 -9.60446179e-01 -3.44678044e-01
-1.25143737e-01 5.21236777e-01 7.48040080e-01 6.49878740e-01
-6.49222732e-01 -2.00056255e-01 2.76747733e-01 3.30523938e-01
-1.76763445e-01 -7.65761256e-01 -3.45905900e-01 -9.61128950e-01
-1.21563390e-01 -1.11527348e+00 7.31795371e-01 -1.11675155e+00
-1.15246761e+00 4.18660104e-01 4.05463010e-01 -2.89328337e-01
-3.16829264e-01 -9.04838383e-01 -5.89118361e-01 5.64228594e-01
-1.29916060e+00 -8.45272422e-01 -6.59918725e-01 5.97823858e-01
3.17512721e-01 1.22208484e-01 7.14916170e-01 -2.53710896e-01
-1.74090981e-01 1.25825346e-01 -8.55239034e-01 1.17894717e-01
2.44335249e-01 -1.56281722e+00 1.05470252e+00 6.35215878e-01
3.15253377e-01 7.11636245e-01 5.78067958e-01 -7.30250120e-01
-2.06248641e+00 -1.03056169e+00 5.88509500e-01 -5.10688901e-01
1.05962181e+00 -5.42723954e-01 -1.00850391e+00 1.09002054e+00
1.33850932e-01 6.14530519e-02 1.43764853e-01 4.76926833e-01
-5.68856895e-01 4.73413587e-01 -8.60074878e-01 8.55461240e-01
1.75991714e+00 -7.12270617e-01 -7.01814771e-01 7.38741994e-01
1.30676222e+00 -7.81542957e-01 -6.90444708e-01 3.35585810e-02
2.94959862e-02 -6.51152313e-01 8.87968719e-01 -8.16269398e-01
5.52301109e-01 -1.36270285e-01 -1.27687842e-01 -1.35473633e+00
-4.84422088e-01 -9.63259995e-01 -2.05817863e-01 6.50798738e-01
5.42525232e-01 -9.64378417e-01 8.55738282e-01 6.26216471e-01
-3.39484215e-01 -8.15236568e-01 -4.54383492e-01 -5.98914862e-01
1.02344982e-01 -6.36626184e-01 4.73563969e-01 1.03280103e+00
1.89266786e-01 1.02398682e+00 2.37403065e-01 1.57593906e-01
4.51466858e-01 1.77808911e-01 1.01608455e+00 -1.37041593e+00
-5.13857007e-01 -6.26747310e-01 -6.93019778e-02 -1.17017674e+00
5.84482849e-01 -1.42573273e+00 -1.34968176e-01 -2.21423674e+00
-1.09572075e-01 -3.80068630e-01 2.53127962e-01 8.91503453e-01
-2.23581761e-01 -6.80771589e-01 3.01901489e-01 -3.14732909e-01
-9.47985351e-01 2.95732528e-01 1.50181842e+00 -3.01572382e-01
-1.15325063e-01 -3.53130847e-01 -9.89219069e-01 6.03596747e-01
7.21149921e-01 -3.09382737e-01 -9.10794795e-01 -5.11189222e-01
7.17536926e-01 4.78013754e-01 6.54328346e-01 -8.79747331e-01
4.58269984e-01 -5.61928570e-01 -2.56260723e-01 -1.94040477e-01
5.75703047e-02 -5.64996481e-01 -5.19270524e-02 6.81453168e-01
-4.50610667e-01 4.76688147e-01 6.18929863e-01 5.78773320e-01
1.26704901e-01 -1.22796848e-01 2.18181238e-01 -5.54173350e-01
-1.13139486e+00 4.07293975e-01 -1.23760603e-01 7.45622873e-01
6.80161655e-01 -3.86962295e-02 -9.12426174e-01 -7.08735347e-01
-6.30594969e-01 7.24680424e-01 3.42534959e-01 4.03047681e-01
6.35721385e-01 -8.32835078e-01 -5.44461310e-01 -2.76470602e-01
2.68933654e-01 5.48638463e-01 1.08151384e-01 7.67767668e-01
-8.60356152e-01 5.19008994e-01 8.98639411e-02 -4.09974426e-01
-8.59702766e-01 8.65675986e-01 2.11402103e-01 -7.89705217e-01
-6.61926150e-01 8.76387417e-01 4.11707968e-01 -8.14546466e-01
1.11038782e-01 -9.82774675e-01 8.87684301e-02 -4.41118151e-01
6.85372874e-02 2.63377935e-01 -4.23168018e-02 4.50765043e-02
-4.09351736e-01 2.74377942e-01 6.39252290e-02 2.62258589e-01
1.25431168e+00 2.78902858e-01 -3.27323973e-01 1.59006551e-01
8.42167497e-01 -1.58954412e-01 -7.72826612e-01 -3.86978179e-01
3.61346841e-01 4.96213213e-02 -2.95355320e-01 -8.63495469e-01
-6.96797490e-01 7.33920574e-01 -2.77984887e-01 4.04621899e-01
7.57859766e-01 2.63577968e-01 4.94305193e-01 1.08204114e+00
6.97447956e-01 -4.45011228e-01 4.63516831e-01 9.59494591e-01
1.12069571e+00 -1.11218631e+00 2.18691349e-01 -7.10127532e-01
-5.47380507e-01 1.01164877e+00 8.54538083e-01 -2.26871952e-01
3.28440279e-01 9.02140439e-02 -4.81581360e-01 -9.01336670e-01
-1.06640851e+00 -2.22644269e-01 3.52146566e-01 6.16201282e-01
2.33592808e-01 2.15145439e-01 2.20163628e-01 2.16383830e-01
-6.15026236e-01 -2.15810195e-01 5.08596897e-01 8.71111274e-01
-4.30069000e-01 -7.70677686e-01 -2.35852852e-01 5.70311308e-01
1.23270847e-01 -5.43305099e-01 -6.59863770e-01 1.11455989e+00
-4.70991224e-01 1.03238833e+00 -2.49154791e-01 -2.17594504e-01
3.27066869e-01 4.05640863e-02 7.44724452e-01 -1.04918766e+00
-4.64097500e-01 -6.39682889e-01 7.15969563e-01 -8.85080636e-01
8.10382366e-02 -2.42343575e-01 -1.82115304e+00 -5.10391176e-01
-1.04425751e-01 1.12032644e-01 2.59788185e-01 1.08088601e+00
4.42937285e-01 7.60955393e-01 -4.24430519e-01 -6.91370487e-01
-4.91544515e-01 -7.58861899e-01 -2.73205698e-01 1.38475358e-01
1.29409775e-01 -4.85865504e-01 -2.25408480e-01 -2.87697881e-01] | [8.921626091003418, 7.457149505615234] |
bd60b107-1e8a-432b-971a-af0f14d39a93 | securing-visually-aware-recommender-systems | 2306.07992 | null | https://arxiv.org/abs/2306.07992v1 | https://arxiv.org/pdf/2306.07992v1.pdf | Securing Visually-Aware Recommender Systems: An Adversarial Image Reconstruction and Detection Framework | With rich visual data, such as images, becoming readily associated with items, visually-aware recommendation systems (VARS) have been widely used in different applications. Recent studies have shown that VARS are vulnerable to item-image adversarial attacks, which add human-imperceptible perturbations to the clean images associated with those items. Attacks on VARS pose new security challenges to a wide range of applications such as e-Commerce and social networks where VARS are widely used. How to secure VARS from such adversarial attacks becomes a critical problem. Currently, there is still a lack of systematic study on how to design secure defense strategies against visual attacks on VARS. In this paper, we attempt to fill this gap by proposing an adversarial image reconstruction and detection framework to secure VARS. Our proposed method can simultaneously (1) secure VARS from adversarial attacks characterized by local perturbations by image reconstruction based on global vision transformers; and (2) accurately detect adversarial examples using a novel contrastive learning approach. Meanwhile, our framework is designed to be used as both a filter and a detector so that they can be jointly trained to improve the flexibility of our defense strategy to a variety of attacks and VARS models. We have conducted extensive experimental studies with two popular attack methods (FGSM and PGD). Our experimental results on two real-world datasets show that our defense strategy against visual attacks is effective and outperforms existing methods on different attacks. Moreover, our method can detect adversarial examples with high accuracy. | ['Xin Li', 'Neil Zhenqiang Gong', 'Bin Liu', 'Minglei Yin'] | 2023-06-11 | null | null | null | null | ['contrastive-learning', 'image-reconstruction', 'contrastive-learning'] | ['computer-vision', 'computer-vision', 'methodology'] | [ 7.41005689e-02 -6.77161813e-01 1.90204427e-01 2.17318282e-01
-5.26770592e-01 -1.16994369e+00 6.06043398e-01 -2.98727065e-01
-7.70256966e-02 2.07888767e-01 -1.85492523e-02 -3.92480582e-01
3.11599195e-01 -9.82064188e-01 -7.55695343e-01 -7.16156900e-01
-9.02226344e-02 -2.34084934e-01 3.96142691e-01 -4.02849495e-01
2.24112645e-01 6.69637144e-01 -1.01323557e+00 2.94452012e-01
6.22139633e-01 1.01637948e+00 -4.10275124e-02 6.83787942e-01
1.90382987e-01 5.22167385e-01 -7.58528829e-01 -8.10998857e-01
8.46885026e-01 -2.35253558e-01 -1.61635235e-01 9.46120471e-02
4.27556276e-01 -5.44004142e-01 -6.39181435e-01 1.57172394e+00
5.62481999e-01 1.64912373e-03 6.99647427e-01 -1.23828971e+00
-1.44166493e+00 3.10294747e-01 -1.00058270e+00 2.94111997e-01
4.35384572e-01 6.32109642e-01 5.08898318e-01 -8.05352390e-01
2.00691402e-01 1.57883167e+00 4.29547429e-01 6.94859743e-01
-9.40072000e-01 -1.06715727e+00 3.31917793e-01 3.94285321e-01
-1.37060928e+00 -2.59711802e-01 9.54636395e-01 -3.02202880e-01
2.23003253e-01 6.89113200e-01 3.55878294e-01 1.41591179e+00
1.74029812e-01 7.98342824e-01 1.24052656e+00 -2.38712654e-01
6.58160448e-02 5.23242712e-01 -1.94846079e-01 7.00245500e-01
4.03186142e-01 3.98507923e-01 -3.36759649e-02 -3.53640646e-01
9.04921412e-01 1.91899836e-01 -5.27501225e-01 -2.92218357e-01
-7.90548265e-01 9.93286312e-01 5.95595539e-01 -1.14185371e-01
-1.18874975e-01 -1.03530601e-01 3.31075996e-01 3.46917689e-01
1.61344156e-01 3.22312593e-01 2.58784816e-02 7.29351997e-01
-2.11224213e-01 5.36343791e-02 4.96778101e-01 7.39525855e-01
1.80475995e-01 4.21383470e-01 -1.08149603e-01 6.56328917e-01
5.30249596e-01 9.80853379e-01 3.70191872e-01 -3.89465004e-01
4.20129806e-01 2.92458117e-01 2.30349287e-01 -1.79068899e+00
1.18618540e-01 -3.42879206e-01 -1.10633039e+00 6.17007971e-01
4.11380529e-01 -9.91485789e-02 -7.64148057e-01 1.54269040e+00
4.83432204e-01 4.47071850e-01 1.74526125e-01 1.18045473e+00
7.35189080e-01 9.54454482e-01 8.09770152e-02 -2.49251261e-01
1.56019688e+00 -8.10227633e-01 -6.69833839e-01 -2.23111033e-01
-1.00251392e-01 -1.07910180e+00 1.14968836e+00 5.63357711e-01
-9.41590548e-01 -7.70382166e-01 -1.29571247e+00 4.14827943e-01
-3.44410062e-01 -2.69154429e-01 3.79833639e-01 1.13626277e+00
-6.15763903e-01 -6.11548685e-02 -4.28933203e-01 -1.54267758e-01
4.60197568e-01 1.41164213e-01 -3.53912890e-01 -7.84241110e-02
-1.44080353e+00 6.29426956e-01 1.21051073e-01 1.34364873e-01
-1.17397392e+00 -2.62595057e-01 -7.47614324e-01 -9.64253545e-02
5.13028622e-01 -6.34957016e-01 8.06376696e-01 -1.30313766e+00
-1.29591012e+00 7.13064134e-01 4.67914015e-01 -5.34164786e-01
4.99858230e-01 -6.16033450e-02 -1.01908684e+00 2.76095778e-01
-3.50934893e-01 -9.95754227e-02 1.41744006e+00 -1.63678503e+00
-3.53637755e-01 -5.04178822e-01 3.54306787e-01 2.03793988e-01
-7.85723507e-01 4.37660098e-01 -4.31813121e-01 -1.10022974e+00
-2.80640811e-01 -8.75881910e-01 -1.78809524e-01 3.79248708e-01
-4.98814732e-01 3.30937564e-01 1.16110146e+00 -6.18914127e-01
1.07986820e+00 -2.26674390e+00 -1.05308034e-01 3.80172819e-01
2.33276561e-01 1.05173504e+00 -2.56393820e-01 3.81201535e-01
4.21210825e-02 2.47455463e-01 1.47987887e-01 2.01001003e-01
-3.36430334e-02 -3.06187123e-02 -7.85074532e-01 7.51353085e-01
-1.56279624e-01 8.48097682e-01 -6.57222569e-01 -3.04462820e-01
3.75389487e-01 5.09002388e-01 -3.34568530e-01 4.15508091e-01
5.57540506e-02 2.73523539e-01 -6.20155036e-01 6.78865075e-01
1.16250598e+00 -4.55492362e-02 -5.26905954e-02 -3.84897619e-01
2.58536577e-01 -6.04509711e-01 -1.35061157e+00 8.65705252e-01
-2.36179084e-01 1.33355066e-01 3.08708102e-01 -8.43907773e-01
7.84969687e-01 1.71151206e-01 -1.36241049e-01 -6.40320957e-01
1.88793287e-01 -9.04247612e-02 -9.35584679e-02 -5.70442200e-01
1.22310102e-01 6.98442534e-02 -9.42760184e-02 2.46875778e-01
-3.73988479e-01 3.51384759e-01 -3.07151258e-01 4.75348592e-01
9.01477873e-01 -3.80109429e-01 3.46949041e-01 -9.77021409e-04
1.00258982e+00 -3.71934146e-01 4.46058422e-01 9.33727443e-01
-3.37341040e-01 3.50718528e-01 -6.36130944e-02 -6.26477122e-01
-8.32537353e-01 -1.31645513e+00 3.99613418e-02 7.40130186e-01
8.16724539e-01 -1.87714584e-02 -6.12554193e-01 -1.13395429e+00
5.79843745e-02 3.63168240e-01 -4.48806554e-01 -4.23599839e-01
-4.83260214e-01 -6.00933492e-01 5.93347490e-01 3.72379988e-01
7.53646910e-01 -9.91477072e-01 6.59283437e-03 -5.63077815e-02
4.66116369e-02 -1.19629776e+00 -8.55462849e-01 -6.13198161e-01
-3.30608785e-01 -1.25576150e+00 -6.80642486e-01 -8.88879418e-01
7.39484787e-01 9.11158502e-01 6.93799853e-01 3.17854524e-01
-3.32093775e-01 5.14341056e-01 -4.96006697e-01 -5.33638954e-01
-6.74412727e-01 -7.05607712e-01 3.20726573e-01 4.82639998e-01
1.98715404e-01 -3.40283662e-01 -1.00867939e+00 6.68226779e-01
-1.23888135e+00 -2.35511899e-01 5.07600904e-01 7.27755010e-01
3.72072428e-01 1.88478887e-01 5.49622655e-01 -8.67347002e-01
7.36203551e-01 -5.41425288e-01 -8.32001448e-01 3.98585826e-01
-3.34954321e-01 -4.80145961e-01 1.22523713e+00 -8.34519863e-01
-9.72123563e-01 -1.56396657e-01 -1.79832458e-01 -6.50752425e-01
-2.62273431e-01 2.20334291e-01 -4.55655664e-01 -7.42458284e-01
8.18299830e-01 3.55082721e-01 1.30346594e-02 -3.52733642e-01
5.03743231e-01 8.39429140e-01 6.81759477e-01 -2.46987462e-01
1.63544393e+00 6.42325401e-01 -8.77995715e-02 -6.75921679e-01
-5.59583843e-01 -2.48018026e-01 1.34843066e-01 -2.81009614e-01
5.59146464e-01 -8.87696922e-01 -8.42335880e-01 7.33722508e-01
-9.06665027e-01 2.52190262e-01 2.88429022e-01 2.51827955e-01
-5.43781482e-02 1.13954675e+00 -8.10661852e-01 -9.09849644e-01
-6.42531276e-01 -1.10684955e+00 6.10046089e-01 3.62235039e-01
4.65986699e-01 -9.71897304e-01 -7.67972097e-02 4.62570727e-01
4.73363072e-01 3.90139163e-01 6.44757867e-01 -6.40323997e-01
-6.88684285e-01 -3.37724566e-01 -2.19720989e-01 5.96134961e-01
1.89946935e-01 4.22506500e-03 -7.90891290e-01 -7.19276786e-01
3.30945879e-01 -2.28115737e-01 7.09443450e-01 -5.52365743e-02
1.40960455e+00 -8.81793499e-01 -1.88695714e-01 8.33112836e-01
1.60622871e+00 4.10501868e-01 8.49628270e-01 3.83045822e-01
7.58909404e-01 2.17202038e-01 5.69486678e-01 3.47364396e-01
-1.02125145e-01 8.23912084e-01 9.13704515e-01 -3.64292771e-01
-9.61371213e-02 -1.93569481e-01 5.92805564e-01 4.56471682e-01
7.09723160e-02 -6.43444419e-01 -3.46619815e-01 2.09089905e-01
-1.71475184e+00 -1.17745185e+00 -2.27291569e-01 2.21266818e+00
5.57677925e-01 2.75901482e-02 6.16899580e-02 2.82444246e-02
9.46632266e-01 2.67522335e-01 -5.38140833e-01 -3.53189975e-01
-1.76764637e-01 5.00022620e-02 5.24989963e-01 1.49212718e-01
-1.54280281e+00 7.60358453e-01 5.55046749e+00 1.08751881e+00
-1.07125497e+00 -2.26978492e-02 4.80333179e-01 2.01232061e-01
-1.86895683e-01 -3.61157835e-01 -4.96434450e-01 7.73411095e-01
2.85337240e-01 -7.40493760e-02 5.49702883e-01 8.92664790e-01
6.46472126e-02 5.68814754e-01 -4.39920604e-01 9.74285960e-01
3.32007051e-01 -1.12492120e+00 3.92762214e-01 -2.44157501e-02
7.05462158e-01 -6.47361159e-01 7.18940198e-01 1.28782734e-01
4.53725457e-01 -9.57789481e-01 4.72617388e-01 1.90621123e-01
7.77190089e-01 -8.17133605e-01 5.96468806e-01 3.24811339e-01
-1.29635286e+00 -2.54976243e-01 -5.49956501e-01 1.92500159e-01
6.88025430e-02 1.35964409e-01 -3.51822048e-01 6.65421426e-01
5.85724592e-01 3.59927505e-01 -5.79198658e-01 1.00841665e+00
-3.02364707e-01 5.75478256e-01 4.65587676e-02 6.61004633e-02
1.12002015e-01 -1.50288433e-01 7.57245898e-01 1.07241535e+00
1.51835859e-01 2.25227430e-01 3.88020247e-01 5.42056441e-01
-5.10118045e-02 3.32942635e-01 -9.03123200e-01 2.27381468e-01
4.94798064e-01 1.41746104e+00 -4.83555079e-01 -7.98712447e-02
-6.27216280e-01 1.16693366e+00 2.38338653e-02 4.15646464e-01
-1.07077873e+00 -4.29333359e-01 8.07026684e-01 2.82829944e-02
3.44834626e-01 -1.30827010e-01 1.47748619e-01 -1.37946236e+00
-9.15025994e-02 -1.45822072e+00 5.77146113e-01 -6.76065505e-01
-1.85379934e+00 4.78995860e-01 -2.90975362e-01 -1.52103257e+00
2.05748901e-01 -7.51251101e-01 -8.01592290e-01 7.81670034e-01
-1.47488308e+00 -1.36129141e+00 -3.42980951e-01 1.21365225e+00
3.96280646e-01 -6.21096373e-01 7.86236942e-01 2.53369391e-01
-5.80027580e-01 1.09477127e+00 1.24768868e-01 3.95001829e-01
7.64387488e-01 -9.56154644e-01 5.07315218e-01 1.46067214e+00
4.01850015e-01 7.25465775e-01 7.12075472e-01 -5.22383809e-01
-1.75220990e+00 -1.27064574e+00 -2.05189019e-01 -2.36747667e-01
6.44382060e-01 -3.31830263e-01 -9.57548320e-01 5.79842091e-01
2.66512275e-01 3.13966215e-01 5.81585109e-01 -4.63734716e-01
-7.61537075e-01 -2.48887479e-01 -1.44951451e+00 8.37264776e-01
9.41230774e-01 -5.16998470e-01 -4.50484991e-01 4.50212330e-01
5.09609163e-01 -3.46150070e-01 -5.07843137e-01 4.04811978e-01
5.47779918e-01 -9.74798977e-01 1.52688181e+00 -5.53815842e-01
1.12292446e-01 -6.07737482e-01 -3.81170698e-02 -1.14243436e+00
-5.54315984e-01 -7.65009224e-01 -4.63893563e-01 1.21799922e+00
-1.97495177e-01 -8.24024022e-01 2.67693579e-01 1.51740775e-01
1.95839360e-01 -4.17554349e-01 -4.81078595e-01 -8.68163884e-01
-1.54612288e-01 -8.12965631e-02 5.64891756e-01 9.37620997e-01
-4.32738662e-01 -1.17145009e-01 -8.54087949e-01 9.63179886e-01
8.87757063e-01 2.57387646e-02 1.07314944e+00 -7.62081683e-01
-5.52294731e-01 -1.88609689e-01 -6.76414907e-01 -9.41910565e-01
-1.46599546e-01 -5.19227147e-01 -3.84069502e-01 -1.07342112e+00
3.21552902e-01 -1.78814188e-01 -4.77390975e-01 2.76851028e-01
-5.10994196e-01 7.76739597e-01 3.01369846e-01 2.95982003e-01
-4.67165530e-01 2.75902182e-01 1.23526108e+00 -3.20763499e-01
3.06395739e-01 -7.51505047e-03 -1.08303607e+00 9.82583880e-01
7.20262170e-01 -2.57479221e-01 -4.82939065e-01 -3.16783428e-01
-4.81717708e-03 -6.26398772e-02 5.98353267e-01 -7.62077749e-01
3.25610638e-02 -2.99286634e-01 6.11223936e-01 -1.99330017e-01
3.83585058e-02 -1.10228944e+00 1.85940370e-01 6.39964104e-01
-6.23774296e-03 1.99103020e-02 1.71269283e-01 9.41856384e-01
3.21984515e-02 -1.02688923e-01 1.16153383e+00 -1.99377373e-01
-8.31443369e-01 5.04544795e-01 -1.23192847e-01 -1.41538039e-01
1.55072725e+00 -1.31581187e-01 -6.75483763e-01 -6.03674114e-01
-4.52803701e-01 8.05315077e-02 5.50456643e-01 5.58537543e-01
9.40660000e-01 -1.41729569e+00 -7.49553621e-01 4.64426696e-01
3.40098068e-02 -6.46316767e-01 5.74858189e-01 2.08427548e-01
-5.49400508e-01 -4.47441973e-02 -2.45348617e-01 -2.21813485e-01
-1.66674888e+00 1.44233119e+00 1.82691336e-01 -2.52656817e-01
-5.46437144e-01 8.45046341e-01 7.69337654e-01 1.23251840e-01
2.83213794e-01 4.86913860e-01 -3.68947893e-01 -4.93994802e-01
9.45210040e-01 1.09322242e-01 -3.84759665e-01 -7.70273507e-01
-2.98362851e-01 6.01999462e-01 -4.50062633e-01 3.44051212e-01
8.80422533e-01 -2.69000441e-01 2.35615909e-01 -3.60179812e-01
1.03957164e+00 6.65098488e-01 -1.13578677e+00 -3.12006623e-01
-7.14693904e-01 -9.06346738e-01 -8.57478902e-02 -8.00398707e-01
-1.45217490e+00 7.60845959e-01 9.72437918e-01 6.63086534e-01
1.36220264e+00 -3.41742754e-01 1.06735289e+00 -1.20737247e-01
5.10492623e-01 -5.26484609e-01 4.69512939e-01 -1.08382002e-01
9.68146026e-01 -1.23142445e+00 1.80238076e-02 -5.93757510e-01
-7.08535731e-01 9.37745333e-01 7.18491018e-01 -6.25413954e-01
4.87453580e-01 2.16049746e-01 3.49387795e-01 6.40853941e-02
-2.77259380e-01 2.56839227e-02 5.43747067e-01 8.77620280e-01
-3.65059406e-01 -8.83261859e-02 -1.84105650e-01 7.26657271e-01
2.27494508e-01 -4.90941882e-01 4.09369797e-01 6.20470762e-01
-3.49953026e-01 -1.15375912e+00 -8.88175964e-01 5.86297214e-02
-8.48904669e-01 -5.39378934e-02 -2.84521937e-01 6.75478458e-01
5.56879751e-02 1.13652456e+00 -4.28934306e-01 -7.11832464e-01
3.01627934e-01 -5.51609814e-01 3.45407844e-01 -2.87231088e-01
-5.88540375e-01 1.57102644e-02 -3.56036782e-01 -4.25013304e-01
-1.29146039e-01 -3.55622143e-01 -6.20688379e-01 -4.70016539e-01
-4.26401109e-01 -1.26339793e-01 3.14640045e-01 6.02927148e-01
2.66185611e-01 2.24952668e-01 1.30298948e+00 -5.13583302e-01
-1.04086244e+00 -3.53269786e-01 -6.98691726e-01 9.50179577e-01
1.95784122e-01 -5.32564938e-01 -4.91144955e-01 -1.26540318e-01] | [5.534847259521484, 7.919505596160889] |
26158862-1fb1-46b7-86f3-c2c5bfcabe59 | learning-the-predictability-of-the-future | 2101.01600 | null | https://arxiv.org/abs/2101.01600v1 | https://arxiv.org/pdf/2101.01600v1.pdf | Learning the Predictability of the Future | We introduce a framework for learning from unlabeled video what is predictable in the future. Instead of committing up front to features to predict, our approach learns from data which features are predictable. Based on the observation that hyperbolic geometry naturally and compactly encodes hierarchical structure, we propose a predictive model in hyperbolic space. When the model is most confident, it will predict at a concrete level of the hierarchy, but when the model is not confident, it learns to automatically select a higher level of abstraction. Experiments on two established datasets show the key role of hierarchical representations for action prediction. Although our representation is trained with unlabeled video, visualizations show that action hierarchies emerge in the representation. | ['Carl Vondrick', 'Ruoshi Liu', 'Dídac Surís'] | 2021-06-19 | null | http://openaccess.thecvf.com//content/CVPR2021/html/Suris_Learning_the_Predictability_of_the_Future_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Suris_Learning_the_Predictability_of_the_Future_CVPR_2021_paper.pdf | cvpr-2021-1 | ['self-supervised-action-recognition'] | ['computer-vision'] | [ 1.08600609e-01 7.06787467e-01 -4.16530222e-01 -4.81875420e-01
-3.65767717e-01 -4.34916019e-01 3.57686579e-01 2.41782948e-01
1.66493639e-01 6.10990405e-01 9.27977264e-01 -7.11677894e-02
-1.32463336e-01 -7.49232829e-01 -7.53130734e-01 -4.26903576e-01
-6.44809067e-01 3.04404795e-01 3.61876577e-01 -2.30613332e-02
2.85381943e-01 3.51381570e-01 -1.60726821e+00 9.57665861e-01
3.06898475e-01 9.58817422e-01 -1.18289597e-01 8.17865551e-01
3.77427340e-01 1.68755662e+00 1.61718559e-02 -7.77087212e-02
3.48428935e-01 -5.99427283e-01 -1.17690122e+00 5.21179616e-01
3.26789677e-01 -5.11280954e-01 -5.13512194e-01 5.19775271e-01
-4.32657629e-01 2.91645341e-02 8.90789270e-01 -1.31201530e+00
-5.83472669e-01 6.89404905e-01 -1.32507443e-01 1.97219253e-01
6.25467479e-01 6.12201691e-02 1.51560962e+00 -8.84589016e-01
9.81646836e-01 9.88783658e-01 7.08338022e-01 4.17315245e-01
-1.39893186e+00 -2.79131889e-01 4.98019725e-01 4.87121463e-01
-1.23254693e+00 -2.42541954e-01 6.35074794e-01 -8.88291240e-01
6.04010105e-01 1.18171073e-01 9.82516468e-01 1.04654086e+00
2.91871101e-01 8.81926000e-01 9.04138744e-01 -1.59588560e-01
4.59332526e-01 -1.70019209e-01 1.98733300e-01 9.28405046e-01
-7.85194412e-02 3.12023878e-01 -8.75958085e-01 -3.17045972e-02
7.66258657e-01 1.06409103e-01 -2.75541723e-01 -7.87533522e-01
-1.05209804e+00 9.23012972e-01 3.99918348e-01 1.10445268e-01
-3.90161514e-01 1.98485598e-01 2.24082530e-01 2.85480887e-01
8.26557279e-02 5.80987871e-01 -4.35783386e-01 -3.41385782e-01
-9.44889069e-01 1.14704780e-01 6.56102121e-01 1.12675929e+00
9.19371724e-01 -1.66136906e-01 -1.43091187e-01 -3.39296162e-02
2.29791522e-01 -3.31219614e-01 5.88823378e-01 -1.46674442e+00
1.74877122e-02 8.73393238e-01 1.87115058e-01 -9.81949985e-01
-3.73175830e-01 -9.70376469e-03 -4.42330509e-01 4.16151285e-01
3.56186092e-01 8.96031410e-02 -8.52910519e-01 1.62495351e+00
1.14473263e-02 2.98425525e-01 1.60186633e-01 6.40690386e-01
4.39031661e-01 9.17103887e-01 1.30712152e-01 -3.36506963e-01
8.05946946e-01 -7.74860144e-01 -2.83750415e-01 2.88135797e-01
8.63106906e-01 -1.34724259e-01 9.70442355e-01 4.98659641e-01
-9.11249042e-01 -7.42802680e-01 -1.18023479e+00 -1.87714502e-01
-6.89730346e-02 -1.43921226e-01 5.77164650e-01 3.45215872e-02
-1.31582105e+00 1.07338178e+00 -1.01656401e+00 -4.50404882e-01
2.43367359e-01 1.94754586e-01 -4.73094374e-01 1.49704382e-01
-8.78546476e-01 6.65780544e-01 4.80837643e-01 -1.54721245e-01
-1.13394928e+00 -5.05131245e-01 -7.97874570e-01 2.31331974e-01
1.56192496e-01 -3.75250548e-01 1.26010740e+00 -8.69421482e-01
-1.38504660e+00 4.40686554e-01 -5.20395823e-02 -6.45028293e-01
5.64457655e-01 -2.42163375e-01 -6.68717027e-02 6.92111373e-01
-9.22464430e-02 9.93565977e-01 8.87523115e-01 -1.10747397e+00
-8.94098341e-01 -2.23217845e-01 5.36777973e-01 1.76519513e-01
-4.08871919e-01 -2.76653796e-01 8.80567953e-02 -2.78989673e-01
4.35555249e-01 -1.12387085e+00 -2.95406669e-01 2.47408524e-01
-1.40484214e-01 -3.87805253e-01 8.62421155e-01 -3.55778307e-01
1.17270362e+00 -2.02534389e+00 2.71994770e-01 1.11912698e-01
4.70038354e-01 -4.57502753e-01 1.94738820e-01 6.06466830e-01
-1.25425085e-01 2.92464942e-01 1.53111249e-01 1.32609889e-01
-9.33006778e-02 2.98139125e-01 -5.01115680e-01 4.98731732e-01
1.16486788e-01 4.66878325e-01 -8.67244601e-01 -7.47671127e-01
2.21784785e-01 2.69604415e-01 -1.03119564e+00 4.53010738e-01
-3.37520361e-01 6.16368711e-01 -6.19969130e-01 2.95679271e-01
7.43583729e-03 -6.86734498e-01 4.29803401e-01 -7.43121654e-02
-2.73850113e-01 2.69824147e-01 -1.10004985e+00 1.33590603e+00
-9.31593776e-02 7.42692828e-01 -6.37912512e-01 -1.09458876e+00
8.75209868e-01 4.77145374e-01 8.04332078e-01 3.56093049e-02
-3.14290434e-01 -3.54318380e-01 -2.32747614e-01 -6.80039346e-01
3.70213628e-01 -3.39343213e-02 -9.64236856e-02 3.98712605e-01
5.04605323e-02 -1.25748925e-02 1.82893425e-02 6.00101352e-01
1.26190221e+00 5.62834799e-01 4.90325689e-01 -2.99559265e-01
2.84629911e-01 3.24468344e-01 7.38679349e-01 6.06862307e-01
-2.33205676e-01 6.81407690e-01 9.33502197e-01 -9.77712512e-01
-1.02502596e+00 -1.15749466e+00 -1.01302648e-02 1.40188801e+00
9.76479277e-02 -1.12852538e+00 -4.15248245e-01 -7.13610590e-01
-4.67917979e-01 6.15664661e-01 -1.14959812e+00 -1.85814314e-02
-4.72112328e-01 2.12200239e-01 -8.63054544e-02 7.33098805e-01
1.95756450e-01 -9.64826822e-01 -1.05089629e+00 2.11111419e-02
-9.52754989e-02 -1.06692624e+00 -2.62762666e-01 2.46945858e-01
-1.16780782e+00 -1.41988266e+00 -2.92491317e-02 -7.13764369e-01
6.63908601e-01 4.35242094e-02 1.10623169e+00 2.57581979e-01
-1.33262366e-01 8.11416268e-01 -5.72563052e-01 1.43674970e-01
-3.40083003e-01 -1.75492123e-01 7.90140703e-02 -6.84206411e-02
1.53157830e-01 -5.52448511e-01 -6.53636694e-01 1.81969449e-01
-6.19727492e-01 3.43354374e-01 1.51847005e-01 5.77670574e-01
6.41796470e-01 1.05899416e-01 8.59828815e-02 -8.90204966e-01
5.70971332e-02 -6.02346063e-01 -3.76186877e-01 2.29877755e-01
-5.20212412e-01 2.74171114e-01 7.94004202e-01 -2.92854428e-01
-8.53957355e-01 7.63351440e-01 4.02086824e-01 -5.91946125e-01
-3.21287602e-01 5.36324322e-01 1.29050091e-01 4.17209357e-01
7.82934189e-01 1.81722477e-01 -2.52752692e-01 -3.01779985e-01
3.85742456e-01 1.03110299e-01 2.93349028e-01 -6.09148383e-01
5.46437740e-01 4.60767478e-01 1.46069273e-01 -8.30063999e-01
-1.04232407e+00 -9.74872783e-02 -1.24283600e+00 -4.77469623e-01
1.13143802e+00 -9.11715567e-01 -7.49131441e-01 -3.15917969e-01
-8.46661925e-01 -2.18428791e-01 -5.97311974e-01 5.77457666e-01
-1.12914634e+00 1.81910887e-01 -5.80881357e-01 -7.47376263e-01
3.22332948e-01 -1.01979601e+00 6.96122944e-01 -8.26237723e-02
-8.14256310e-01 -8.51779282e-01 1.40828922e-01 2.25156188e-01
-1.61906630e-01 5.13003170e-01 1.04163420e+00 -7.54271686e-01
-9.87081170e-01 -4.85231504e-02 3.04019541e-01 6.81267083e-02
-1.39245968e-02 3.91871452e-01 -6.13233209e-01 -1.20133594e-01
-9.18445140e-02 -7.39099622e-01 5.67727208e-01 2.81172276e-01
1.46694303e+00 -7.12781489e-01 -2.36329138e-01 5.59410512e-01
1.16782308e+00 1.70518011e-01 6.11898363e-01 1.68891340e-01
4.79863405e-01 7.92866528e-01 6.14025176e-01 5.61213017e-01
5.27315021e-01 5.65741599e-01 5.13013959e-01 3.74575317e-01
1.54600292e-01 -8.65176678e-01 4.61988926e-01 6.16885483e-01
-3.11985523e-01 2.85489649e-01 -8.00429285e-01 1.04714572e-01
-1.91794896e+00 -1.31437051e+00 8.14339984e-03 1.89290512e+00
6.98803365e-01 2.11797982e-01 4.61445630e-01 9.22794789e-02
4.55506176e-01 1.41444772e-01 -5.54635465e-01 -3.95424515e-01
3.40519905e-01 -1.80221364e-01 -6.49518892e-02 6.42664135e-01
-1.17717838e+00 9.84192133e-01 7.66877460e+00 2.19640166e-01
-9.08227623e-01 -4.07175392e-01 8.59876871e-01 4.46170494e-02
-1.55254871e-01 3.57601613e-01 -7.61232674e-01 1.53827980e-01
8.25791121e-01 -4.23103154e-01 1.20737836e-01 1.21226203e+00
1.63216129e-01 1.49851516e-01 -1.88478756e+00 6.91943884e-01
-1.47327483e-01 -1.66068566e+00 3.69273901e-01 1.44323543e-01
5.07248700e-01 -3.68500262e-01 4.68005240e-02 3.56588393e-01
5.48564374e-01 -1.28768957e+00 8.01335514e-01 6.64757133e-01
4.59192008e-01 -5.58924496e-01 9.67351869e-02 6.18569970e-01
-1.24408543e+00 -3.85107785e-01 -5.43314219e-01 -4.52410847e-01
-1.24867946e-01 -1.47636801e-01 -9.93815005e-01 -1.16222188e-01
6.65248752e-01 1.07333207e+00 -7.51960337e-01 8.13390374e-01
-5.05213082e-01 5.69417596e-01 5.53017221e-02 1.22141838e-01
3.22701275e-01 -2.04692930e-01 6.32714629e-02 1.09707081e+00
4.51421857e-01 8.01267564e-01 6.57093823e-01 4.24714327e-01
2.59619206e-01 7.51519427e-02 -1.07704163e+00 -6.39015734e-02
2.44880483e-01 9.92174268e-01 -6.86413288e-01 -4.70731109e-01
-4.39721197e-01 7.81540394e-01 5.51567018e-01 3.03636491e-01
-6.17169619e-01 2.64733136e-01 4.33460265e-01 4.51306790e-01
4.74674791e-01 -4.52095658e-01 -1.71700284e-01 -1.14640427e+00
-3.63871336e-01 -5.34282267e-01 7.31959641e-01 -1.05536366e+00
-8.01458001e-01 6.22286320e-01 1.76160276e-01 -1.59313369e+00
-7.78803408e-01 -5.65021992e-01 -4.98633265e-01 9.22559053e-02
-7.56584644e-01 -9.19240057e-01 -1.26140535e-01 8.34202349e-01
6.89883709e-01 5.02087772e-02 1.12947679e+00 -6.24271393e-01
1.48546714e-02 3.60102095e-02 -1.75695837e-01 1.23571701e-01
3.62980701e-02 -1.35792816e+00 -2.04530597e-01 4.89802182e-01
6.27030671e-01 4.71832007e-01 9.73778784e-01 -5.00361621e-01
-9.94275928e-01 -8.20955157e-01 9.21515524e-01 -5.52277744e-01
8.81549358e-01 -2.78458208e-01 -8.49533856e-01 1.22696996e+00
2.61192843e-02 1.67746678e-01 1.12176406e+00 2.04469055e-01
-4.78061855e-01 -4.65759821e-03 -9.51678753e-01 6.06358230e-01
1.26852143e+00 -4.34320807e-01 -8.31601799e-01 5.85932434e-01
7.35008121e-01 -7.34340474e-02 -1.00573754e+00 3.66206557e-01
7.20326841e-01 -1.51533973e+00 7.71948755e-01 -1.26288128e+00
9.99458432e-01 -1.29762873e-01 -4.39786911e-01 -1.22075093e+00
-6.47692323e-01 -6.30364776e-01 -5.35670280e-01 6.16862237e-01
4.94497418e-01 9.63053107e-02 1.33097482e+00 8.96108031e-01
-1.00115247e-01 -1.03989697e+00 -8.14292789e-01 -6.36476576e-01
1.63207754e-01 -1.74288332e-01 2.25215852e-01 7.06590414e-01
6.51104629e-01 3.60953093e-01 -5.02547741e-01 2.45021939e-01
5.94341397e-01 5.17635882e-01 6.38250470e-01 -1.39909470e+00
-2.36681819e-01 1.36438027e-01 -9.16116655e-01 -1.18229485e+00
1.96022898e-01 -6.52107954e-01 2.38400586e-02 -1.51540267e+00
2.72676915e-01 -8.13699365e-02 -2.55709440e-01 4.82034296e-01
2.02808440e-01 2.53508054e-02 2.84729779e-01 5.35653293e-01
-1.03281832e+00 5.96577048e-01 1.10236835e+00 8.06516707e-02
-1.91548467e-01 2.77414359e-02 -5.66259384e-01 1.20681250e+00
7.76912510e-01 -3.11157405e-01 -9.12769079e-01 -1.31411910e-01
-1.85581651e-02 4.19543147e-01 2.04593226e-01 -1.25656056e+00
2.23438382e-01 -4.81674701e-01 8.57174575e-01 -4.62740809e-01
6.14892066e-01 -9.58613992e-01 9.40412879e-02 5.29198110e-01
-9.25318897e-01 9.01835263e-02 -4.04308319e-01 6.60272539e-01
-9.35426131e-02 -1.45241231e-01 7.65723884e-01 -2.33524993e-01
-8.94217193e-01 4.98358220e-01 -7.71571875e-01 3.03434134e-02
1.20595908e+00 -5.75081110e-01 -3.85488831e-02 -9.39211667e-01
-1.51283646e+00 1.16522372e-01 6.37201011e-01 2.78772473e-01
8.62123132e-01 -1.44725084e+00 -3.52267832e-01 2.04461087e-02
1.89203098e-01 -4.22106117e-01 -9.16081145e-02 2.89353162e-01
-2.67868400e-01 3.41856480e-01 -4.23125356e-01 -7.24610507e-01
-1.09855521e+00 8.64034116e-01 1.30807143e-02 -9.54678878e-02
-1.12692547e+00 5.66404819e-01 3.61293554e-01 2.97792673e-01
3.90342951e-01 -4.01094973e-01 -5.73831916e-01 4.94664423e-02
4.90864366e-01 2.42361903e-01 -4.81931418e-01 -8.22675824e-01
-7.01159239e-02 4.64871079e-01 3.11702881e-02 -1.05555296e-01
1.64807189e+00 -2.05633759e-01 1.18366890e-01 9.29408371e-01
1.15368068e+00 -2.76864707e-01 -2.01197052e+00 7.23351613e-02
4.72002149e-01 -5.69632590e-01 -3.94161373e-01 -2.19665349e-01
-7.86225021e-01 8.22198749e-01 2.65798002e-01 4.38904524e-01
8.27698588e-01 2.46254474e-01 1.12769425e-01 7.47311711e-01
5.49753666e-01 -1.15681803e+00 7.05734730e-01 5.82775056e-01
1.08720553e+00 -1.02149475e+00 2.69684494e-01 -4.11950797e-01
-9.87695813e-01 1.56145537e+00 8.96690071e-01 -4.54471350e-01
9.20499265e-01 9.71800014e-02 -2.65707642e-01 -3.86317313e-01
-1.30720091e+00 1.57697260e-01 1.48650452e-01 6.80423498e-01
5.84576309e-01 -1.08489752e-01 1.79541126e-01 3.92722905e-01
-6.00928605e-01 7.96466470e-02 9.63258624e-01 8.87898386e-01
-8.48974228e-01 -6.85144961e-01 -1.48743475e-02 3.85708034e-01
-1.47926167e-01 1.78559989e-01 -4.11356956e-01 8.93842161e-01
9.29519162e-02 7.81369388e-01 2.21321568e-01 -7.28720307e-01
9.93592292e-02 1.05319202e-01 4.71943736e-01 -9.15997505e-01
1.74047630e-02 -1.79906487e-01 6.16290867e-02 -1.04189277e+00
-3.96443456e-01 -8.48734915e-01 -1.44610620e+00 -1.89619809e-02
3.51144612e-01 4.52645063e-01 7.52196759e-02 7.58868396e-01
1.12466477e-01 9.26540494e-02 7.49919534e-01 -8.27167392e-01
-5.31989992e-01 -6.84298098e-01 -7.84474671e-01 5.32105863e-01
3.50104779e-01 -7.31644809e-01 -3.58057737e-01 7.07033992e-01] | [8.4578275680542, 0.6941186189651489] |
5ce6549a-f954-41ed-8c21-3dee78ae21a9 | pseudo-lidar-from-visual-depth-estimation | 1812.07179 | null | https://arxiv.org/abs/1812.07179v6 | https://arxiv.org/pdf/1812.07179v6.pdf | Pseudo-LiDAR from Visual Depth Estimation: Bridging the Gap in 3D Object Detection for Autonomous Driving | 3D object detection is an essential task in autonomous driving. Recent techniques excel with highly accurate detection rates, provided the 3D input data is obtained from precise but expensive LiDAR technology. Approaches based on cheaper monocular or stereo imagery data have, until now, resulted in drastically lower accuracies --- a gap that is commonly attributed to poor image-based depth estimation. However, in this paper we argue that it is not the quality of the data but its representation that accounts for the majority of the difference. Taking the inner workings of convolutional neural networks into consideration, we propose to convert image-based depth maps to pseudo-LiDAR representations --- essentially mimicking the LiDAR signal. With this representation we can apply different existing LiDAR-based detection algorithms. On the popular KITTI benchmark, our approach achieves impressive improvements over the existing state-of-the-art in image-based performance --- raising the detection accuracy of objects within the 30m range from the previous state-of-the-art of 22% to an unprecedented 74%. At the time of submission our algorithm holds the highest entry on the KITTI 3D object detection leaderboard for stereo-image-based approaches. Our code is publicly available at https://github.com/mileyan/pseudo_lidar. | ['Bharath Hariharan', 'Wei-Lun Chao', 'Divyansh Garg', 'Kilian Q. Weinberger', 'Yan Wang', 'Mark Campbell'] | 2018-12-18 | pseudo-lidar-from-visual-depth-estimation-1 | http://openaccess.thecvf.com/content_CVPR_2019/html/Wang_Pseudo-LiDAR_From_Visual_Depth_Estimation_Bridging_the_Gap_in_3D_CVPR_2019_paper.html | http://openaccess.thecvf.com/content_CVPR_2019/papers/Wang_Pseudo-LiDAR_From_Visual_Depth_Estimation_Bridging_the_Gap_in_3D_CVPR_2019_paper.pdf | cvpr-2019-6 | ['3d-object-detection-from-stereo-images'] | ['computer-vision'] | [ 2.90731877e-01 -2.52523333e-01 -1.51885986e-01 -4.20887262e-01
-7.85868824e-01 -4.75269079e-01 5.39277673e-01 7.80642629e-02
-6.92030966e-01 3.41979533e-01 -3.01516503e-01 -5.22845507e-01
7.46723637e-02 -8.90808940e-01 -8.85927200e-01 -3.73235404e-01
1.37694120e-01 7.40618527e-01 6.32665634e-01 -4.07019019e-01
4.82548654e-01 6.78936779e-01 -2.08107758e+00 2.27576494e-02
5.44463933e-01 1.22939873e+00 2.37597227e-01 7.45092750e-01
-2.37813324e-01 4.76709098e-01 -2.04785034e-01 -1.38424128e-01
6.77561522e-01 1.23636842e-01 -2.68991679e-01 -2.68414289e-01
1.02287579e+00 -6.68936014e-01 -4.36064124e-01 9.41568792e-01
5.39760590e-01 -4.90174949e-01 5.67331553e-01 -1.26716185e+00
-1.46676213e-01 1.39007196e-01 -7.34562039e-01 3.21077615e-01
2.35489771e-01 3.66851091e-01 8.47807348e-01 -1.10762715e+00
5.66743672e-01 1.24046624e+00 9.19636190e-01 3.69854271e-01
-1.23175168e+00 -9.37136471e-01 -2.13934794e-01 2.20259666e-01
-1.59938240e+00 -5.57435274e-01 4.74297345e-01 -6.13030553e-01
1.23238719e+00 -9.79492590e-02 6.54875100e-01 6.88915312e-01
2.26691574e-01 6.71547592e-01 1.18680048e+00 -2.28950500e-01
-1.10038579e-01 1.84139177e-01 6.45913780e-02 7.41256654e-01
7.09997237e-01 6.69615209e-01 -5.16770959e-01 2.88079560e-01
6.07499003e-01 -1.49038479e-01 7.07621053e-02 -7.62680709e-01
-9.61832166e-01 9.03901756e-01 7.03373492e-01 -2.01922745e-01
-1.53112158e-01 3.31606895e-01 3.77911389e-01 1.68080628e-01
4.60521936e-01 1.09993100e-01 -2.22908422e-01 -1.68213144e-01
-9.39897895e-01 4.98421431e-01 3.85651618e-01 9.11367655e-01
1.00148690e+00 -1.08625501e-01 3.67628902e-01 3.48691672e-01
2.22296819e-01 6.56751037e-01 1.27787113e-01 -1.00601196e+00
5.16534865e-01 8.92655492e-01 2.19047189e-01 -5.82551420e-01
-4.61693138e-01 -4.18202400e-01 -3.25035870e-01 9.32375193e-01
5.86839616e-01 1.90326780e-01 -9.79218543e-01 1.19829941e+00
4.56558913e-01 -1.17776014e-01 -1.42583877e-01 9.21893895e-01
7.61159182e-01 2.60761082e-01 -3.34887952e-01 3.88085544e-01
1.20259750e+00 -4.25265789e-01 -5.07445410e-02 -7.10495591e-01
5.26836395e-01 -7.50506341e-01 6.95289075e-01 4.36212987e-01
-7.10210502e-01 -6.49401546e-01 -1.43437731e+00 -3.08314234e-01
-6.17668390e-01 -5.39125362e-03 4.03368443e-01 8.40388596e-01
-1.09009778e+00 4.27855194e-01 -6.22461498e-01 -5.72983444e-01
6.22487426e-01 2.90944040e-01 -4.76990789e-01 -2.83519566e-01
-6.60643280e-01 1.26489079e+00 3.86918336e-01 -2.14518785e-01
-8.14536870e-01 -8.97380829e-01 -8.16762090e-01 -4.04508859e-01
4.29385364e-01 -6.18560970e-01 1.48721874e+00 -2.66284943e-01
-8.63479912e-01 1.29845345e+00 -1.72433421e-01 -8.82773459e-01
9.59998846e-01 -3.97582591e-01 2.09102467e-01 -4.60486859e-02
4.33501780e-01 1.17777824e+00 7.01364636e-01 -1.30198622e+00
-1.22416127e+00 -5.22367001e-01 1.45757228e-01 1.23615421e-01
2.14190736e-01 -2.42909208e-01 -2.83536971e-01 1.86736718e-01
2.26966515e-01 -9.16009605e-01 -8.84874910e-02 4.39491689e-01
-1.76220402e-01 -1.98880032e-01 9.47988033e-01 3.00689824e-02
4.66821909e-01 -2.11121392e+00 -3.59096199e-01 -3.63744825e-01
2.72256643e-01 3.52479696e-01 1.89027771e-01 4.07380193e-01
5.74172698e-02 -1.03344642e-01 -2.21480295e-01 -5.86266816e-01
4.46621329e-02 2.39043728e-01 -3.98453057e-01 6.62272453e-01
4.65485185e-01 7.56585479e-01 -8.81233513e-01 -4.84304279e-01
7.97338963e-01 5.55317879e-01 -3.30625087e-01 -4.51330952e-02
-8.42860267e-02 2.76846826e-01 -2.44482458e-01 7.53678203e-01
9.04693067e-01 3.30890775e-01 -4.08463031e-01 -1.75905507e-02
-6.75498068e-01 5.55235028e-01 -1.11244071e+00 1.65599906e+00
-3.02458972e-01 1.08614278e+00 1.25764593e-01 -7.98333049e-01
9.76480484e-01 -2.50337362e-01 3.74178112e-01 -8.63912880e-01
2.01397628e-01 6.33211970e-01 -3.81496362e-03 -1.22713573e-01
8.75180483e-01 -2.07412928e-01 2.65622605e-02 -3.69892642e-02
-1.86408043e-01 -6.98128164e-01 2.26003230e-01 1.00056276e-01
1.04944611e+00 3.28642756e-01 2.22393826e-01 -1.50862157e-01
2.86113471e-01 5.47872126e-01 2.06745625e-01 9.34579611e-01
-3.39600414e-01 8.50425065e-01 2.78529048e-01 -6.31047606e-01
-1.27675200e+00 -1.04377925e+00 -6.11089766e-01 4.67061102e-01
2.25253493e-01 -2.26593971e-01 -4.79747623e-01 -3.42575997e-01
6.66871130e-01 5.21607935e-01 -6.35669291e-01 -4.88817915e-02
-3.94974560e-01 -4.45007056e-01 6.92329705e-01 6.35084689e-01
4.80427891e-01 -5.70817709e-01 -1.30144334e+00 3.27147901e-01
2.23676264e-01 -1.35527921e+00 2.02290982e-01 6.10398293e-01
-9.25401807e-01 -1.04779565e+00 -2.96237260e-01 -3.20104897e-01
1.51996836e-01 8.09049010e-01 1.22350037e+00 -3.00030652e-02
-7.87956178e-01 1.73422590e-01 -1.42368600e-01 -9.71402824e-01
-2.55197555e-01 5.31851463e-02 9.46535990e-02 -3.87107939e-01
6.72364235e-01 -3.91813457e-01 -7.22626090e-01 1.37730449e-01
-5.91040611e-01 -9.79181007e-03 6.32985771e-01 4.16017622e-01
4.53697294e-01 -3.22389603e-01 2.02718209e-02 -4.99848574e-01
5.81837143e-04 -1.57403991e-01 -1.07648301e+00 -5.98830462e-01
-7.03434527e-01 -2.85099130e-02 2.64110863e-02 5.39739579e-02
-4.83052909e-01 4.72147584e-01 -1.99741632e-01 -6.25118136e-01
-4.53197777e-01 -3.40406336e-02 3.07099700e-01 -3.75471056e-01
9.23131049e-01 1.51768357e-01 9.30244327e-02 -4.18679297e-01
2.90861607e-01 7.05008686e-01 6.72621787e-01 -2.65825331e-01
1.03243852e+00 9.90985453e-01 3.65997732e-01 -8.36539090e-01
-9.25895214e-01 -8.18286002e-01 -7.89509356e-01 -2.01434091e-01
7.21358299e-01 -1.27953053e+00 -4.89730090e-01 4.86186981e-01
-1.17569852e+00 -2.28452608e-01 -2.68321574e-01 3.22109103e-01
-4.03176874e-01 8.87520760e-02 -2.25428805e-01 -1.16067123e+00
-4.34430540e-02 -1.24487972e+00 1.48327541e+00 1.57132512e-03
1.09573901e-01 -3.68236750e-01 8.19977000e-02 4.02806789e-01
4.49315876e-01 3.65064621e-01 3.99092883e-01 -9.54918638e-02
-9.46431518e-01 -6.36037230e-01 -7.40923703e-01 2.57884860e-01
-2.19718412e-01 1.34665065e-03 -1.31239128e+00 -4.73595299e-02
-3.74375552e-01 -2.17942894e-01 1.31621897e+00 3.74510199e-01
5.36015391e-01 5.00743508e-01 -5.91555059e-01 5.50811291e-01
1.66893005e+00 -1.82235882e-01 5.22625923e-01 6.43466353e-01
5.39253235e-01 5.82689166e-01 8.92412066e-01 3.54524583e-01
6.16221309e-01 7.03616619e-01 1.02349305e+00 7.12185726e-02
-3.25009853e-01 -2.15550348e-01 2.21802711e-01 4.59352806e-02
3.29063311e-02 1.00711592e-01 -1.20691335e+00 7.70086884e-01
-1.77049601e+00 -8.74243915e-01 -5.53179264e-01 2.31781101e+00
3.56464952e-01 8.10085058e-01 2.52010673e-01 3.10310304e-01
4.33343202e-01 -1.53671373e-02 -5.35698295e-01 -3.38350475e-01
8.22532400e-02 1.99271336e-01 1.15922880e+00 5.22614360e-01
-1.21209037e+00 7.85528660e-01 5.68705654e+00 5.08065224e-01
-1.22689831e+00 5.45829572e-02 1.04229718e-01 -3.32925886e-01
2.09028944e-01 -4.26799469e-02 -1.33592856e+00 1.91169605e-01
9.21513379e-01 -2.85097975e-02 -2.20225379e-02 1.05729151e+00
2.44632483e-01 -5.35069287e-01 -1.07976985e+00 1.13507879e+00
5.41968085e-02 -1.23809588e+00 -2.28450805e-01 4.09010470e-01
4.36743051e-01 9.26227272e-01 1.91818457e-02 2.59182662e-01
2.12119699e-01 -9.80101049e-01 1.34482348e+00 4.88151610e-02
8.27696383e-01 -6.82908297e-01 7.05269277e-01 6.14820004e-01
-1.43976068e+00 -1.22425757e-01 -6.57124221e-01 -4.86084729e-01
-4.55635116e-02 6.96918666e-01 -1.09703863e+00 4.34272856e-01
9.57496643e-01 6.57047212e-01 -7.21544981e-01 1.27126122e+00
-2.14861363e-01 2.11080015e-01 -6.42895043e-01 5.68724424e-02
4.30688232e-01 7.51292799e-03 6.08958364e-01 1.18881285e+00
4.27678049e-01 -2.23892614e-01 1.66772202e-01 9.49617386e-01
6.73822185e-04 -3.08055311e-01 -1.19669378e+00 3.52250189e-01
5.67812562e-01 1.30069077e+00 -5.06940126e-01 -1.75050586e-01
-4.53064442e-01 4.69754726e-01 3.11302543e-01 -3.55261326e-01
-6.59653962e-01 -2.91820586e-01 9.92791414e-01 5.03722429e-01
5.44025123e-01 -5.45581043e-01 -4.96116728e-01 -6.06570125e-01
1.38742149e-01 -3.56321156e-01 -6.92803711e-02 -8.61072958e-01
-7.87402570e-01 2.85013050e-01 6.54664310e-03 -1.27421093e+00
-1.98974088e-01 -1.06675363e+00 -3.40917468e-01 8.18217695e-01
-2.04018497e+00 -1.03561413e+00 -6.14870071e-01 1.08392879e-01
5.02194643e-01 1.29535481e-01 4.09504831e-01 4.38089043e-01
-1.40394241e-01 1.59501106e-01 -5.62274866e-02 -3.73284519e-02
5.87841988e-01 -1.27611685e+00 8.41334343e-01 8.32211792e-01
1.49627343e-01 1.15349911e-01 8.06535840e-01 -4.60504234e-01
-1.58562148e+00 -1.05562592e+00 7.69448459e-01 -1.03356779e+00
5.75258911e-01 -3.77299517e-01 -6.13372386e-01 5.04601717e-01
-5.77533394e-02 9.42049846e-02 -3.65194120e-02 -3.08317244e-01
-4.64451909e-01 -2.95948118e-01 -1.02901769e+00 2.53164619e-01
1.20309556e+00 -5.21685719e-01 -5.30369341e-01 1.51296169e-01
5.04066348e-01 -6.85120523e-01 -3.13791096e-01 6.35156393e-01
7.59348392e-01 -1.48745096e+00 1.16527295e+00 2.12103371e-02
3.88667285e-01 -5.78963935e-01 -2.88885355e-01 -8.04280639e-01
-7.48633072e-02 -1.43378749e-01 2.67141998e-01 7.69724727e-01
2.33514056e-01 -6.12630010e-01 1.00885296e+00 -5.93708344e-02
-3.45255196e-01 -4.21372175e-01 -1.28975666e+00 -9.45176363e-01
1.78477410e-02 -1.01661193e+00 3.96487921e-01 4.06777084e-01
-4.72349584e-01 2.68638998e-01 2.92087849e-02 2.69831866e-01
1.00078166e+00 1.14922099e-01 1.20079374e+00 -1.55449665e+00
1.78725824e-01 -5.99320471e-01 -8.35071623e-01 -9.85002875e-01
-2.23533213e-01 -8.17541301e-01 3.32902849e-01 -1.54529750e+00
-1.03609208e-02 -4.13456917e-01 9.06831250e-02 3.86239916e-01
1.26324728e-01 8.16925287e-01 1.66369021e-01 2.34078944e-01
-3.27274799e-01 1.53579071e-01 8.10377836e-01 -2.27928907e-02
1.20579705e-01 -2.58710265e-01 -3.64845365e-01 6.85534358e-01
8.56869698e-01 -6.18489385e-01 -3.62443179e-02 -5.28220952e-01
2.22927541e-01 -3.63209158e-01 9.16524649e-01 -1.56455076e+00
1.68634906e-01 2.57077992e-01 3.00212592e-01 -1.04633367e+00
7.46573150e-01 -6.74984455e-01 -4.57156785e-02 8.66442978e-01
1.20100789e-01 -4.55936305e-02 4.85590756e-01 3.99956375e-01
3.89005169e-02 -1.55035421e-01 8.68723273e-01 -3.09448600e-01
-1.06264651e+00 1.79931596e-01 -2.85666764e-01 -1.30702347e-01
8.07549000e-01 -6.32052779e-01 -4.24890488e-01 -1.48078904e-01
3.93861420e-02 1.00531485e-02 7.46488035e-01 6.27092123e-01
5.84791720e-01 -1.10445881e+00 -8.80177498e-01 2.24809483e-01
5.06263793e-01 2.02831000e-01 -2.82853842e-01 8.89696956e-01
-7.21054494e-01 6.96057200e-01 -2.21990287e-01 -1.17632794e+00
-1.23125923e+00 3.01765025e-01 4.00295049e-01 2.47491628e-01
-6.85784400e-01 7.92792678e-01 1.44913465e-01 -4.78255481e-01
2.05601186e-01 -4.46680278e-01 2.69225955e-01 8.35748091e-02
3.23580980e-01 2.49431103e-01 3.73077095e-01 -6.75658286e-01
-6.41936123e-01 8.97813499e-01 7.83969760e-02 1.24249859e-02
1.15358698e+00 -2.47428287e-02 2.66614407e-01 4.60617423e-01
9.54427302e-01 -1.48399442e-01 -1.36617470e+00 -1.05184026e-01
-9.17481109e-02 -7.86431313e-01 4.23254639e-01 -4.25725698e-01
-6.71577990e-01 1.26869726e+00 1.04361534e+00 1.02535591e-01
4.20617223e-01 5.38180508e-02 4.80774820e-01 3.53559494e-01
7.15904176e-01 -8.43170285e-01 -2.20424399e-01 6.41663730e-01
6.08896852e-01 -1.73206604e+00 2.98526227e-01 -3.12902659e-01
-2.55493596e-02 1.04285240e+00 5.42586148e-01 -3.80415767e-01
2.75928020e-01 5.21524310e-01 2.37574086e-01 -2.76210725e-01
-4.79663938e-01 -7.72317469e-01 1.31322771e-01 6.44443870e-01
2.45603591e-01 2.67891176e-02 1.85849667e-01 -1.62283808e-01
-2.78110653e-01 9.21323672e-02 3.93978089e-01 1.13784146e+00
-9.98836637e-01 -1.05271327e+00 -5.83113492e-01 4.09788132e-01
1.71658685e-04 -1.98118892e-02 -5.08634984e-01 1.13454866e+00
4.67201710e-01 9.20064926e-01 1.83307111e-01 -4.34926689e-01
6.08098328e-01 2.98117287e-02 6.00109875e-01 -6.86169684e-01
-2.45670110e-01 -2.95034111e-01 1.20518766e-01 -8.11335146e-01
-2.61426508e-01 -8.77950430e-01 -9.76725101e-01 -4.76856977e-01
-3.88304383e-01 -4.36258405e-01 1.06423438e+00 6.89154863e-01
3.12008262e-01 1.55396655e-01 2.59157628e-01 -1.53929770e+00
-6.09819293e-01 -7.85365045e-01 -3.27753425e-01 4.89146449e-02
6.15052104e-01 -8.61378312e-01 -3.72413099e-01 -4.47418720e-01] | [7.778453350067139, -2.585094928741455] |
a6ccf720-681f-497f-b847-eb0f0449223c | language-models-are-greedy-reasoners-a | 2210.01240 | null | https://arxiv.org/abs/2210.01240v4 | https://arxiv.org/pdf/2210.01240v4.pdf | Language Models Are Greedy Reasoners: A Systematic Formal Analysis of Chain-of-Thought | Large language models (LLMs) have shown remarkable reasoning capabilities given chain-of-thought prompts (examples with intermediate reasoning steps). Existing benchmarks measure reasoning ability indirectly, by evaluating accuracy on downstream tasks such as mathematical reasoning. However, it is unclear how these models obtain the answers and whether they rely on simple heuristics rather than the generated chain-of-thought. To enable systematic exploration of the reasoning ability of LLMs, we present a new synthetic question-answering dataset called PrOntoQA, where each example is generated from a synthetic world model represented in first-order logic. This allows us to parse the generated chain-of-thought into symbolic proofs for formal analysis. Our analysis on InstructGPT and GPT-3 shows that LLMs are quite capable of making correct individual deduction steps, and so are generally capable of reasoning, even in fictional contexts. However, they have difficulty with proof planning: When multiple valid deduction steps are available, they are not able to systematically explore the different options. | ['He He', 'Abulhair Saparov'] | 2022-10-03 | null | null | null | null | ['mathematical-reasoning'] | ['natural-language-processing'] | [ 1.31827429e-01 1.04602385e+00 -1.08746327e-02 -1.87717155e-01
-1.01116514e+00 -1.05161011e+00 9.15197849e-01 1.75594270e-01
2.08189547e-01 7.84113586e-01 1.58183053e-01 -1.43095803e+00
-3.66565198e-01 -1.29071295e+00 -9.63169396e-01 2.07254902e-01
-1.06023267e-01 8.04113209e-01 3.44827503e-01 -5.51167905e-01
2.59411603e-01 1.17481247e-01 -1.14116538e+00 7.11165786e-01
1.28299582e+00 2.11877152e-01 -3.22392546e-02 1.16889524e+00
-5.26530445e-01 1.88680947e+00 -5.21218538e-01 -7.22305238e-01
1.01419255e-01 -6.84853077e-01 -1.39822364e+00 -5.26586294e-01
3.57256025e-01 -6.31396472e-01 -9.64120477e-02 1.14512432e+00
-3.14439327e-01 -1.58101127e-01 5.68953395e-01 -1.60437036e+00
-3.85009736e-01 1.46735847e+00 3.04183781e-01 9.43698511e-02
1.10085678e+00 6.44178033e-01 1.31888270e+00 -2.08498716e-01
8.32744360e-01 1.81537986e+00 4.18753475e-01 6.38468802e-01
-1.40817654e+00 -2.34570146e-01 -6.87538385e-02 3.84219319e-01
-9.24660504e-01 -3.31178308e-01 4.56465930e-01 -3.82192433e-01
1.27185261e+00 4.08272475e-01 5.64567626e-01 9.15551960e-01
3.57704997e-01 7.76726544e-01 1.44122326e+00 -7.05934823e-01
3.82688880e-01 1.08371422e-01 2.98746020e-01 1.07759857e+00
2.47425392e-01 3.84287983e-02 -5.59824407e-01 -3.70419413e-01
5.88866770e-01 -6.48151040e-01 -3.73893194e-02 -1.12886034e-01
-1.42378724e+00 8.62909079e-01 1.66340992e-01 2.24961787e-01
-8.11001211e-02 4.70096320e-01 1.89599887e-01 6.46905780e-01
-5.42212009e-01 1.22784245e+00 -5.40818810e-01 -4.59337294e-01
-7.65808523e-01 1.08416247e+00 1.37905645e+00 8.79428267e-01
2.86935329e-01 -2.77807117e-01 -3.60131592e-01 7.23391026e-02
2.69414037e-01 6.07126713e-01 1.94402948e-01 -1.74015915e+00
8.18194866e-01 8.35252523e-01 3.63913506e-01 -7.17960238e-01
-3.95208687e-01 1.13415003e-01 -1.72935277e-02 2.67639637e-01
9.58474398e-01 -1.60718128e-01 -4.44924593e-01 1.77757084e+00
1.25390351e-01 -1.88927099e-01 7.04301476e-01 6.53463006e-01
6.89367235e-01 7.12293386e-01 5.23695722e-02 -9.98286381e-02
1.40810084e+00 -8.36636007e-01 -5.15534043e-01 -3.91989738e-01
1.14278185e+00 -8.88749808e-02 1.32633018e+00 6.56114638e-01
-1.65489244e+00 -7.87567273e-02 -9.46030915e-01 -1.73635960e-01
-2.26909980e-01 -4.24922913e-01 1.11993110e+00 9.50988084e-02
-1.04583359e+00 5.24695218e-01 -6.07527614e-01 -3.73949632e-02
2.04112619e-01 -4.49333377e-02 -1.21095851e-01 -3.83340627e-01
-1.58085465e+00 1.04497921e+00 6.15474463e-01 -2.04258054e-01
-1.15892291e+00 -6.90025389e-01 -1.03573108e+00 3.06692392e-01
8.23756039e-01 -7.65776277e-01 1.93960834e+00 -4.85142589e-01
-1.35232675e+00 6.24989986e-01 -1.45208746e-01 -8.60926092e-01
7.97360122e-01 4.56537195e-02 -2.77827561e-01 4.34616834e-01
3.23673040e-01 8.27861845e-01 3.68070155e-01 -9.72876370e-01
-3.84016722e-01 2.03108042e-03 1.16027415e+00 -7.30875432e-02
6.07111752e-01 1.13910876e-01 2.19369620e-01 -1.73329879e-02
2.42896706e-01 -8.15183043e-01 -1.83676243e-01 -1.87357038e-01
-6.85176730e-01 -3.40949982e-01 2.67155897e-02 -5.46244323e-01
1.05017948e+00 -1.66856241e+00 1.58189833e-02 1.34970129e-01
3.66572767e-01 3.97068970e-02 2.46023899e-03 6.45625293e-01
-3.13624516e-02 6.67640448e-01 -3.62528637e-02 5.60502648e-01
7.56345451e-01 3.18303943e-01 -7.96139181e-01 -3.43749315e-01
3.78098279e-01 1.46820712e+00 -1.29532313e+00 -8.76357615e-01
2.14257669e-02 -5.56727767e-01 -8.65596473e-01 2.43009001e-01
-1.29354882e+00 8.71227589e-03 -6.71455741e-01 5.96111357e-01
2.80228108e-01 -4.45901632e-01 3.61704737e-01 5.09874105e-01
3.15820813e-01 9.14808929e-01 -9.39532876e-01 1.51918805e+00
-4.30964261e-01 5.74280739e-01 -2.83400059e-01 -4.19157833e-01
3.29261184e-01 1.92008480e-01 -4.77459937e-01 -8.09196532e-01
-2.17702881e-01 2.98831910e-01 6.23024464e-01 -6.99250877e-01
1.77481383e-01 -4.71387088e-01 -3.00286859e-01 7.84250617e-01
-3.07121217e-01 -8.71369302e-01 7.66793489e-01 7.38884985e-01
1.51656270e+00 1.67972729e-01 2.55002111e-01 -1.22924700e-01
5.15143275e-01 8.04231346e-01 2.83086360e-01 1.13816607e+00
1.17868513e-01 2.15221252e-02 1.32719505e+00 -4.53894436e-01
-9.33091581e-01 -1.21011674e+00 2.18010321e-01 6.62035048e-01
-1.57662347e-01 -9.35275733e-01 -9.19239223e-01 -7.42893100e-01
1.03716413e-02 1.67746913e+00 -1.60831943e-01 -1.58446878e-01
-6.28196239e-01 6.61047399e-02 1.21036160e+00 4.46417570e-01
5.34793198e-01 -1.23050487e+00 -1.02678478e+00 2.11567014e-01
-5.55700123e-01 -1.17479610e+00 4.89479989e-01 3.75201330e-02
-8.88537109e-01 -1.47004616e+00 1.53151348e-01 -2.00220376e-01
6.25796556e-01 -3.26438308e-01 1.32939982e+00 5.07983983e-01
2.41036654e-01 3.47978085e-01 -1.68998972e-01 -4.66737181e-01
-1.17856908e+00 -1.75230712e-01 -2.70539999e-01 -1.04484177e+00
3.16075683e-01 -4.54816908e-01 2.50063594e-02 -2.12114118e-02
-6.83912516e-01 4.80051011e-01 3.70073915e-01 5.96882701e-01
2.73412955e-03 3.09294045e-01 1.87946007e-01 -9.79603946e-01
1.12370992e+00 -3.03249002e-01 -7.36779749e-01 6.60815537e-01
-4.47521925e-01 6.94306552e-01 1.07284439e+00 -1.24645881e-01
-1.05595851e+00 -6.72707975e-01 2.05825731e-01 2.16158908e-02
-1.94057330e-01 5.21030724e-01 8.79181772e-02 3.66049141e-01
9.61093187e-01 2.77438730e-01 -1.49500296e-01 1.63838819e-01
6.54604256e-01 7.72031620e-02 5.36045432e-01 -1.63840604e+00
1.06690884e+00 -1.09693825e-01 3.29512268e-01 -8.16900209e-02
-8.17800522e-01 5.08013904e-01 -8.28586891e-02 6.63553029e-02
6.49313688e-01 -5.58043718e-01 -1.29312921e+00 -3.15575711e-02
-1.20558405e+00 -1.15214586e+00 -3.67602944e-01 1.15470700e-01
-1.13416719e+00 2.95191139e-01 -8.59021485e-01 -9.96924281e-01
-8.76375958e-02 -1.31689095e+00 8.27330887e-01 1.36306152e-01
-9.12125468e-01 -8.68097126e-01 -1.72021106e-01 4.83861208e-01
8.14850628e-02 7.91498050e-02 1.82222509e+00 -8.81942511e-01
-8.76775026e-01 5.94617315e-02 -9.13659707e-02 2.48513650e-03
-3.13551098e-01 1.29291147e-01 -6.84316814e-01 3.49989742e-01
-2.85878271e-01 -7.61986017e-01 1.40909990e-03 -6.07885756e-02
1.04029703e+00 -7.82502353e-01 -1.29471123e-01 6.83403835e-02
9.03579414e-01 3.73698329e-03 7.81233370e-01 4.95600343e-01
-5.34272306e-02 5.11542618e-01 6.57641590e-01 -2.52340138e-01
7.25054145e-01 2.04918817e-01 1.91316545e-01 6.17830455e-01
1.63849995e-01 -9.38646972e-01 3.48488808e-01 1.01218037e-01
1.30337283e-01 -4.04041819e-02 -1.49842262e+00 3.61274600e-01
-1.78973210e+00 -1.13983405e+00 -2.65712053e-01 1.55595219e+00
1.22184622e+00 6.45140111e-01 -1.45175323e-01 3.43312830e-01
3.66846323e-02 -1.23544894e-01 -3.09164256e-01 -9.89375174e-01
1.10233903e-01 2.39012375e-01 -4.64795679e-02 9.37272131e-01
-2.53029555e-01 1.06999838e+00 7.10894346e+00 6.72902346e-01
-4.82377350e-01 -3.78733307e-01 3.03239584e-01 -2.83790287e-02
-8.01726043e-01 4.93733943e-01 -3.59695584e-01 1.03253298e-01
1.15710568e+00 -2.83425510e-01 8.43203723e-01 6.64324284e-01
5.12225479e-02 -5.71384132e-01 -1.64348388e+00 3.93744498e-01
-3.58846217e-01 -1.43841779e+00 2.42794141e-01 -2.65533477e-01
2.65512049e-01 -6.14454031e-01 -3.27944964e-01 8.59852791e-01
8.71624053e-01 -1.37390816e+00 1.12073004e+00 6.01597130e-01
4.31063652e-01 -5.27683437e-01 5.05274832e-01 1.09300923e+00
-4.19542074e-01 -4.43776906e-01 -5.10140043e-03 -7.23424494e-01
1.08343892e-01 2.26793304e-01 -1.22589993e+00 4.32235241e-01
1.69287696e-01 -1.06780007e-01 -7.69642651e-01 3.31996948e-01
-1.07686555e+00 7.57655919e-01 -3.97398680e-01 -4.31320339e-01
4.16178226e-01 -9.72950086e-02 4.75989312e-01 8.90670121e-01
5.17417155e-02 6.42477453e-01 -7.55803734e-02 1.55067277e+00
3.21818471e-01 -5.41816652e-01 -5.70371568e-01 -5.64856291e-01
3.17100465e-01 9.28751945e-01 -5.85397482e-01 -7.16689110e-01
1.01861946e-01 3.16339612e-01 4.69855189e-01 3.18173617e-01
-8.72327745e-01 -2.49902755e-01 3.19045663e-01 1.32614076e-01
-2.63277918e-01 -2.61268735e-01 -2.72959411e-01 -1.15843475e+00
1.46522492e-01 -1.54823530e+00 3.12697142e-01 -1.74161255e+00
-6.07116818e-01 1.57150641e-01 6.51763082e-01 -3.12448680e-01
-9.25097644e-01 -8.18232894e-01 -7.76403308e-01 7.70917535e-01
-1.20737028e+00 -9.06835556e-01 4.28390540e-02 2.68795758e-01
4.30122733e-01 2.10973486e-01 9.23457265e-01 -5.44418216e-01
-4.26960811e-02 2.61319041e-01 -8.98725450e-01 1.50991648e-01
1.15956016e-01 -1.55158651e+00 5.87484837e-01 8.18432331e-01
1.57483622e-01 1.17475414e+00 1.03389108e+00 -5.77925682e-01
-1.76167834e+00 -5.85728288e-01 1.07846332e+00 -9.87226546e-01
9.62158084e-01 -1.90831855e-01 -8.20973694e-01 1.04372501e+00
5.99734187e-02 -3.15780044e-01 1.49547741e-01 8.37004930e-03
-5.67024887e-01 3.29342633e-01 -1.22196698e+00 1.19964874e+00
1.24858618e+00 -7.19728529e-01 -1.53380525e+00 6.06708884e-01
9.86130118e-01 -6.91630483e-01 -5.85064113e-01 1.56252250e-01
3.31407398e-01 -1.08986533e+00 7.30647743e-01 -1.31649411e+00
1.17383146e+00 -5.44604659e-01 -2.49370053e-01 -1.07147956e+00
-5.50125614e-02 -8.12565625e-01 -4.58232552e-01 9.79306936e-01
7.29398072e-01 -7.01395392e-01 3.84411097e-01 1.10876739e+00
1.25468582e-01 -9.04576421e-01 -5.12152910e-01 -6.07537031e-01
3.77840996e-01 -9.66231763e-01 8.70527208e-01 6.38093829e-01
9.66636300e-01 4.25861090e-01 5.41458845e-01 1.12206906e-01
4.68087465e-01 5.09994864e-01 8.78001511e-01 -8.41304660e-01
-6.45772159e-01 -4.54454601e-01 8.57056305e-02 -9.06730235e-01
5.27868569e-01 -1.14212692e+00 1.32989481e-01 -1.91842341e+00
-7.00949458e-03 -3.73239011e-01 4.91305619e-01 7.81559348e-01
2.07388755e-02 -5.57416677e-01 3.48396420e-01 -2.60445446e-01
-7.19374061e-01 -1.42181948e-01 1.48582006e+00 -2.40994737e-01
2.59607792e-01 -2.46613562e-01 -7.73374975e-01 1.11499918e+00
5.94434977e-01 -2.31375322e-01 -7.98081815e-01 -3.28384250e-01
1.17724526e+00 7.40723550e-01 8.83954167e-01 -8.63829851e-01
2.59449363e-01 -8.30648363e-01 -1.36508033e-01 -1.78143740e-01
-1.08252011e-01 -4.84457016e-01 1.10938311e-01 6.99204028e-01
-7.92787731e-01 2.14485869e-01 2.85255075e-01 5.57656959e-02
-8.01128224e-02 -5.76587498e-01 2.21599966e-01 -6.43594623e-01
-6.61205113e-01 -5.70222199e-01 -6.14417255e-01 4.90184039e-01
8.89502048e-01 5.25955446e-02 -5.17711520e-01 -5.04783511e-01
-5.90441525e-01 5.66791832e-01 5.16502619e-01 -6.03808984e-02
4.40443486e-01 -9.00129080e-01 -5.53600669e-01 -8.12495202e-02
1.45943373e-01 2.85150826e-01 -1.44542143e-01 7.06537962e-01
-9.67296481e-01 7.30414927e-01 8.75305235e-02 -3.12573045e-01
-7.53393173e-01 7.78859913e-01 6.78258896e-01 -5.17036438e-01
-5.87762117e-01 5.49832940e-01 2.16479457e-04 -7.58058250e-01
-2.11843446e-01 -1.21220505e+00 1.25730291e-01 -5.62735975e-01
5.90403557e-01 1.24037758e-01 -1.61820605e-01 2.57758111e-01
-4.37875807e-01 1.69517145e-01 2.69625723e-01 -4.38349456e-01
9.14246023e-01 3.77195358e-01 -2.56587148e-01 3.75862300e-01
2.38693446e-01 1.40795678e-01 -6.50679886e-01 8.23390335e-02
1.48541868e-01 -1.87001660e-01 -5.90067327e-01 -1.19577849e+00
-2.40946049e-03 8.58523607e-01 -6.95241868e-01 6.37972474e-01
5.81033528e-01 3.99814509e-02 6.10745251e-01 1.18782365e+00
8.06557477e-01 -7.04778254e-01 -1.09178960e-01 7.14171350e-01
8.57619047e-01 -9.55177724e-01 -1.50707230e-01 -3.77215683e-01
-4.81168330e-01 1.16026402e+00 7.03210115e-01 7.39738941e-02
-3.15568864e-01 5.26356637e-01 -1.21743426e-01 -3.60543102e-01
-1.31004739e+00 1.77457944e-01 -2.49449760e-01 4.24785346e-01
1.34719640e-01 1.30132005e-01 -6.32142127e-02 6.54744267e-01
-1.01083148e+00 2.39651576e-01 8.14742982e-01 9.98040557e-01
-3.55929106e-01 -8.85471344e-01 -6.70760393e-01 3.83778304e-01
-3.37281227e-01 -1.48608521e-01 -6.26348019e-01 1.01290822e+00
-2.03171507e-01 1.20115173e+00 -4.06862259e-01 5.58774918e-02
1.33788630e-01 5.53322017e-01 1.00916827e+00 -7.15678036e-01
-3.91299039e-01 -8.66982281e-01 8.30554903e-01 -7.87003934e-01
1.28253445e-01 -3.76261979e-01 -1.82322598e+00 -6.95852518e-01
6.75083920e-02 5.15438855e-01 1.83026582e-01 1.33990240e+00
1.29676061e-02 6.12097979e-01 -1.53052375e-01 -9.97029915e-02
-1.00216985e+00 -7.00136185e-01 -1.61191642e-01 2.73576975e-01
2.98919946e-01 -1.99986547e-01 -3.83549780e-01 -9.53396724e-04] | [9.267304420471191, 7.220120429992676] |
911c25e3-09da-435d-bf87-1f1e150e9cfb | metric-learning-based-timing-synchronization | 2307.00217 | null | https://arxiv.org/abs/2307.00217v1 | https://arxiv.org/pdf/2307.00217v1.pdf | Metric Learning-Based Timing Synchronization by Using Lightweight Neural Network | Timing synchronization (TS) is one of the key tasks in orthogonal frequency division multiplexing (OFDM) systems. However, multi-path uncertainty corrupts the TS correctness, making OFDM systems suffer from a severe inter-symbol-interference (ISI). To tackle this issue, we propose a timing-metric learning-based TS method assisted by a lightweight one-dimensional convolutional neural network (1-D CNN). Specifically, the receptive field of 1-D CNN is specifically designed to extract the metric features from the classic synchronizer. Then, to combat the multi-path uncertainty, we employ the varying delays and gains of multi-path (the characteristics of multi-path uncertainty) to design the timing-metric objective, and thus form the training labels. This is typically different from the existing timing-metric objectives with respect to the timing synchronization point. Our method substantively increases the completeness of training data against the multi-path uncertainty due to the complete preservation of metric information. By this mean, the TS correctness is improved against the multi-path uncertainty. Numerical results demonstrate the effectiveness and generalization of the proposed TS method against the multi-path uncertainty. | ['Hui Lin', 'Jiafan Wang', 'Chuangui Rao', 'Shuhai Tang', 'Na Yang', 'Chaojin Qing'] | 2023-07-01 | null | null | null | null | ['metric-learning', 'metric-learning'] | ['computer-vision', 'methodology'] | [-8.86802524e-02 -1.30936280e-01 -5.79017162e-01 -5.23542538e-02
-6.14548147e-01 -3.14168096e-01 2.11577907e-01 -2.97593057e-01
3.72723080e-02 9.11296964e-01 -1.64313093e-01 -8.11787486e-01
-6.20728910e-01 -3.46303850e-01 -5.04509211e-01 -9.92154896e-01
-8.16518784e-01 -2.43319139e-01 -4.31445539e-01 -2.75715917e-01
3.45784932e-01 6.34858072e-01 -5.95743537e-01 -3.75314951e-01
8.86056304e-01 1.44048393e+00 -9.64793041e-02 4.18099642e-01
2.85535127e-01 4.45095479e-01 -1.02398992e+00 2.60393888e-01
4.39765304e-01 -5.08422077e-01 -1.94685772e-01 -1.75636947e-01
-1.52928784e-01 -4.76953834e-01 -9.83232915e-01 9.23581243e-01
5.99632800e-01 -2.03829929e-01 4.26459908e-01 -1.52383065e+00
-2.92972058e-01 7.70792663e-01 -7.45663345e-01 5.42551219e-01
-2.10991964e-01 -5.74304760e-02 5.65778673e-01 -4.94245797e-01
1.43467888e-01 1.11578715e+00 6.32233441e-01 -3.22484877e-03
-6.36604846e-01 -1.20439255e+00 -9.35204700e-02 1.70648396e-01
-1.68581867e+00 -6.16814196e-01 8.29559326e-01 -2.01923326e-01
3.20326537e-01 -5.38030863e-02 1.00488819e-01 7.60212183e-01
8.36425245e-01 5.06963074e-01 5.88098049e-01 -5.07606328e-01
1.54397801e-01 -2.37093419e-01 -1.77336872e-01 2.91936159e-01
3.47845078e-01 8.16561222e-01 -1.56946778e-01 -9.81529430e-02
8.51230621e-01 -3.04543912e-01 -3.01056355e-01 -2.20319659e-01
-1.32151794e+00 2.10318267e-01 4.04956758e-01 3.93617749e-01
1.59545705e-01 5.71473837e-01 3.94773751e-01 8.12711537e-01
5.92576489e-02 5.13321757e-01 -3.17180187e-01 -9.18925554e-02
-1.13231051e+00 4.60708980e-03 5.83031058e-01 1.52791214e+00
5.52715838e-01 8.26636434e-01 -4.26498592e-01 -3.86726595e-02
4.91079390e-01 4.97907460e-01 1.21772692e-01 -5.94251275e-01
8.45310330e-01 -2.74362385e-01 8.22381899e-02 -1.27953053e+00
-7.41050065e-01 -1.40844810e+00 -1.20313287e+00 6.36488711e-03
6.10058367e-01 -6.20576084e-01 -6.69383764e-01 1.78425372e+00
4.01382931e-02 8.59364569e-01 2.91985959e-01 7.89392650e-01
5.89404106e-02 5.40974796e-01 -4.17776763e-01 -4.45072114e-01
8.57028961e-01 -5.20713806e-01 -1.05469942e+00 -3.00793815e-02
9.06754375e-01 -8.24074566e-01 -1.30963296e-01 3.14490318e-01
-6.26074970e-01 -5.64882040e-01 -1.95861053e+00 7.26029873e-01
2.17921197e-01 3.22159022e-01 2.89975226e-01 1.12011445e+00
-5.48772693e-01 5.81221223e-01 -3.84312272e-01 2.80917227e-01
4.98426825e-01 5.08475065e-01 1.30634785e-01 -3.84067255e-03
-1.88241410e+00 6.97937250e-01 5.76437533e-01 5.15633941e-01
-8.94206583e-01 -9.05961633e-01 -7.35142648e-01 2.50284016e-01
5.87138653e-01 -7.02103525e-02 1.20561266e+00 -7.35658884e-01
-1.43745422e+00 -1.34977236e-01 1.97774693e-01 -4.97736484e-01
4.32890654e-01 -2.21902877e-02 -1.10658026e+00 -2.94398405e-02
-1.60467979e-02 -2.84478903e-01 1.13648570e+00 -1.08105838e+00
-7.62088001e-01 -1.16456475e-03 1.64758325e-01 -2.28429154e-01
-3.55558768e-02 -3.13186288e-01 -3.27074111e-01 -1.00182796e+00
1.27659380e-01 -8.91215861e-01 -3.55320334e-01 -2.68796742e-01
-5.53019822e-01 3.98778170e-01 7.77114689e-01 -3.99340570e-01
1.80855000e+00 -2.40412068e+00 -2.24991992e-01 5.76266050e-01
2.92805284e-01 3.18035066e-01 1.23646252e-01 4.10280436e-01
-1.57554641e-01 1.76022992e-01 -2.19008308e-02 6.10463601e-03
-5.29648829e-03 2.41620988e-02 -3.47362161e-01 7.02397823e-01
3.76153171e-01 3.90815079e-01 -1.06222498e+00 -2.55817711e-01
1.34122118e-01 -9.68157649e-02 -1.81446299e-01 -1.49346799e-01
-7.50677586e-02 7.81978488e-01 -7.94460654e-01 6.15345657e-01
1.16860616e+00 -7.38682738e-03 3.12859237e-01 -1.87347978e-01
-2.99187750e-01 -7.47414529e-02 -1.39247608e+00 1.70874417e+00
-6.46363497e-01 6.97382569e-01 -2.11179927e-01 -9.88084078e-01
9.76038635e-01 6.83792293e-01 5.87600350e-01 -1.10521996e+00
4.97102082e-01 5.28667510e-01 3.93188745e-01 2.65993755e-02
3.66644621e-01 3.25881764e-02 -3.63533765e-01 4.99992728e-01
8.96668583e-02 3.59699309e-01 -2.14069009e-01 1.28478542e-01
9.08496737e-01 2.49420991e-03 4.89686042e-01 -4.85669315e-01
6.60548925e-01 -6.46078765e-01 1.18632841e+00 7.65436709e-01
-3.62287313e-01 1.76950395e-01 9.03853953e-01 -2.76329488e-01
-9.48107541e-01 -8.84911776e-01 -4.20012861e-01 7.17161456e-03
5.55096567e-01 -6.21008128e-02 -2.31072918e-01 -3.42000514e-01
1.20412841e-01 5.60328841e-01 -1.72323748e-01 -5.76065302e-01
-5.31573236e-01 -4.69698578e-01 1.07745385e+00 2.36842349e-01
7.37797856e-01 -1.90242901e-01 -7.26594627e-02 6.77291095e-01
2.68706024e-01 -1.35255527e+00 -6.46804571e-01 3.31998616e-01
-3.85655433e-01 -7.98883498e-01 -6.18122756e-01 -4.83098656e-01
2.47146800e-01 4.50655878e-01 4.53260511e-01 4.87563387e-02
-4.40229401e-02 -2.39196882e-01 -3.11576635e-01 -2.94806570e-01
-2.42013633e-01 2.70812623e-02 4.90369827e-01 1.92824394e-01
-3.77426110e-02 -7.16401875e-01 -3.66423786e-01 6.19331241e-01
-7.63784468e-01 -4.54466492e-01 6.56390071e-01 6.92517161e-01
5.66942319e-02 6.54111505e-01 9.50964212e-01 -3.72746587e-01
3.84221435e-01 -8.98009598e-01 -9.64964509e-01 1.43996462e-01
-6.27011001e-01 2.50717551e-01 8.05592477e-01 -2.58381873e-01
-8.20025027e-01 -2.77696073e-01 2.55869955e-01 -4.35828626e-01
4.22812283e-01 7.08321691e-01 -5.47264338e-01 -4.93543357e-01
5.64087629e-01 -5.49793094e-02 -3.61482233e-01 1.29655018e-01
1.76733375e-01 8.76673460e-01 2.96845049e-01 -7.96389878e-01
1.32954943e+00 3.43228430e-01 5.27886868e-01 -8.58516157e-01
-7.12101460e-01 -1.44388929e-01 -4.40143406e-01 -2.27761641e-01
6.35805205e-02 -1.05420232e+00 -6.49202526e-01 4.76555437e-01
-1.19437170e+00 7.77481198e-02 2.94026494e-01 8.99720788e-01
-4.28605974e-01 6.78832591e-01 -2.56895453e-01 -8.15002859e-01
-2.40385868e-02 -1.08685780e+00 5.68275034e-01 3.01538467e-01
2.29476258e-01 -8.67649615e-01 -1.76826924e-01 -4.52065766e-01
3.59185249e-01 5.78319371e-01 1.04307735e+00 -4.04798627e-01
-8.82348835e-01 -1.73830152e-01 -3.74266893e-01 2.18145534e-01
1.62764817e-01 -1.30567014e-01 -8.64399493e-01 -4.95408952e-01
1.10856056e-01 3.18787485e-01 3.96649033e-01 5.54734290e-01
1.05735624e+00 5.94216883e-02 -6.26887202e-01 8.68477523e-01
1.37274384e+00 6.68321013e-01 8.13555360e-01 4.22877014e-01
4.59889233e-01 -1.12706842e-02 9.62369263e-01 1.00386500e+00
4.52017747e-02 7.48465955e-01 4.11988676e-01 4.86738943e-02
3.25158656e-01 1.53005034e-01 2.11600691e-01 5.29630184e-01
3.60896409e-01 -6.36465132e-01 -7.04129457e-01 1.31438166e-01
-1.73853576e+00 -8.07603002e-01 -1.28571168e-01 2.44251180e+00
5.27321279e-01 4.82698679e-01 -4.34295774e-01 4.59887445e-01
8.30308318e-01 3.30183953e-01 -4.68669295e-01 -3.01728666e-01
-1.01979934e-02 -1.29351497e-01 1.08645964e+00 2.78407186e-01
-1.17399323e+00 4.42282647e-01 5.30289221e+00 1.23642385e+00
-1.30476665e+00 -3.05065662e-01 4.69821662e-01 3.84311229e-01
-7.39468634e-02 1.90173775e-01 -7.90429950e-01 7.04748631e-01
1.05895722e+00 -5.11348963e-01 1.76049009e-01 3.94182861e-01
4.43034410e-01 3.91142853e-02 -1.25489914e+00 9.38160539e-01
-3.60035926e-01 -1.19545794e+00 -3.66056979e-01 3.52881849e-01
7.27788746e-01 -5.07030725e-01 2.39494339e-01 4.61385310e-01
-1.59587011e-01 -8.70298982e-01 6.25773013e-01 5.20125985e-01
1.11968017e+00 -1.31336141e+00 8.50438237e-01 4.17043306e-02
-1.36192930e+00 -3.71828377e-01 -1.64668813e-01 5.26325554e-02
2.39476502e-01 1.20068288e+00 -5.75926363e-01 1.41791844e+00
1.34280324e-01 6.52614474e-01 1.43044114e-01 1.43167603e+00
-5.39904125e-02 4.03528750e-01 -2.67197102e-01 3.09042186e-01
5.27281106e-01 8.31836164e-02 7.83696592e-01 9.68356669e-01
7.87589312e-01 -1.03695564e-01 3.16976339e-01 5.32979250e-01
-1.73963815e-01 -4.84858096e-01 -5.53383291e-01 -1.27502859e-01
1.21297526e+00 7.79795051e-01 -1.93998590e-01 6.16317540e-02
-4.02882755e-01 2.28335381e-01 -3.46899003e-01 5.50803244e-01
-9.28369462e-01 -8.84591043e-01 7.40414977e-01 -4.35551941e-01
1.97402090e-01 -8.64952862e-01 -5.29145360e-01 -6.23338699e-01
1.77467540e-01 -8.00354898e-01 4.63278256e-02 -4.90484238e-02
-9.72280920e-01 3.27551812e-01 -2.23491102e-01 -2.11473012e+00
-1.17295012e-01 -2.87403852e-01 -6.47896588e-01 9.77267325e-01
-2.07745600e+00 -8.35104704e-01 6.12258278e-02 3.86581451e-01
1.73504427e-02 -3.74844581e-01 6.05867863e-01 7.51078486e-01
-7.36830771e-01 1.07633698e+00 6.30087912e-01 1.83999866e-01
8.71079147e-01 -8.94208670e-01 2.92372733e-01 9.82145965e-01
-5.45957565e-01 5.97383320e-01 6.14221752e-01 -5.13549685e-01
-1.45145595e+00 -1.08353770e+00 5.66749990e-01 3.05246085e-01
8.53293478e-01 -1.49393380e-01 -3.81237745e-01 4.39266473e-01
-8.36953372e-02 1.13495052e-01 6.47410691e-01 -1.13662228e-01
-2.41045192e-01 -2.95163870e-01 -8.84951055e-01 7.21500278e-01
7.76897550e-01 -4.32482302e-01 -1.66816711e-01 1.15757711e-01
7.85870910e-01 -6.23696566e-01 -9.29402590e-01 3.90618891e-01
6.25043750e-01 -6.29769683e-01 8.75314832e-01 -2.85010308e-01
1.49947464e-01 -6.42137468e-01 -2.82325983e-01 -1.31072342e+00
-2.72749156e-01 -1.25042999e+00 -3.33273262e-01 1.01800191e+00
2.25180686e-01 -5.08423746e-01 4.85355884e-01 -1.92924470e-01
-6.55051112e-01 -3.49964738e-01 -1.44993925e+00 -1.36053491e+00
7.18412101e-02 -4.73993629e-01 8.85128021e-01 1.05520642e+00
1.09559290e-01 -9.12873000e-02 -8.04444015e-01 9.98497367e-01
8.46052527e-01 -2.16683641e-01 5.68544209e-01 -1.04315686e+00
-2.46685445e-01 -3.86903703e-01 -5.18576860e-01 -1.28762603e+00
5.77866100e-02 -6.76356375e-01 -2.51487885e-02 -7.72557497e-01
-8.25716496e-01 -1.05933022e+00 -8.20751429e-01 -2.24291965e-01
1.53426319e-01 -4.89350677e-01 -1.08301856e-01 1.48117334e-01
-3.64626527e-01 6.61601186e-01 1.48411143e+00 -1.13658644e-01
-1.54958501e-01 5.19773722e-01 -6.14379644e-01 1.76120192e-01
9.50246751e-01 -5.46622753e-01 -4.85908508e-01 -2.76792616e-01
2.11002544e-01 7.99050927e-01 -2.90629324e-02 -1.62043524e+00
2.62851566e-01 -1.58970550e-01 2.73471713e-01 -7.11488903e-01
8.16313643e-03 -1.08122730e+00 -1.39098972e-01 7.04191267e-01
1.11223280e-01 -5.66362321e-01 3.79178554e-01 9.47577059e-01
-1.89693257e-01 5.79384714e-02 5.74091554e-01 4.28440422e-01
-8.82715821e-01 6.84862494e-01 -5.24111688e-01 -6.50313646e-02
1.15044093e+00 -2.80829430e-01 -1.97770730e-01 -2.87505031e-01
-3.30606341e-01 7.53280401e-01 -2.61253566e-01 4.67906833e-01
5.00131011e-01 -1.66715920e+00 -4.95492160e-01 1.27001956e-01
2.72727877e-01 -1.85851067e-01 4.55729991e-01 1.07011187e+00
-3.95370424e-01 9.28819060e-01 -3.00791204e-01 -6.40471995e-01
-4.92969483e-01 4.30066347e-01 6.22063935e-01 -2.12437123e-01
-3.46173853e-01 5.46949089e-01 1.10248299e-02 5.40703256e-03
3.68846387e-01 -2.85437375e-01 -5.08550033e-02 -2.57626057e-01
4.77836728e-01 3.07953209e-01 1.69287875e-01 -9.37953666e-02
-4.49811816e-01 4.57071751e-01 -1.60133809e-01 2.95042316e-03
6.33248210e-01 -3.71410429e-01 1.33967221e-01 1.40887201e-01
1.25371969e+00 -1.60417575e-02 -1.25399041e+00 -7.16991305e-01
2.05627561e-01 -5.67113519e-01 3.05123776e-01 -3.76700431e-01
-9.55906332e-01 8.17370296e-01 6.96705103e-01 -3.29565108e-02
9.35915053e-01 -9.02354956e-01 9.23566163e-01 5.07219553e-01
7.99743414e-01 -1.14221859e+00 5.52430227e-02 8.03296149e-01
3.78954679e-01 -9.12304580e-01 1.57931894e-01 -2.02507690e-01
-1.26297340e-01 1.50707889e+00 3.58985633e-01 3.61483768e-02
1.05239213e+00 4.07293290e-01 3.94120105e-02 1.00552969e-01
-1.87869936e-01 8.47476423e-02 5.22087812e-02 6.31499469e-01
1.61359459e-01 1.85867801e-01 -1.74107417e-01 5.51684797e-01
1.23891048e-01 -9.95896161e-02 7.43193269e-01 9.33584988e-01
-4.44009334e-01 -1.23445737e+00 -4.48110819e-01 2.95168787e-01
-2.09505096e-01 -8.47245473e-03 3.75913620e-01 8.31769824e-01
1.08476140e-01 1.12151814e+00 -1.21916719e-01 -7.46534765e-01
7.44330138e-02 -5.89791000e-01 2.21822679e-01 -3.02213281e-01
-1.55901788e-02 -3.16270366e-02 1.54759392e-01 -4.69560742e-01
1.84430629e-02 -4.73928213e-01 -1.42920661e+00 -4.45009828e-01
-7.22138524e-01 3.91197614e-02 5.93416035e-01 1.41097867e+00
3.17255855e-01 1.14429772e+00 1.59287465e+00 -4.85010564e-01
-9.20044839e-01 -5.63175678e-01 -9.29302633e-01 -6.58326983e-01
8.88137996e-01 -7.47733951e-01 -3.37297410e-01 -5.82439661e-01] | [6.327794551849365, 1.4934821128845215] |
309b0b68-88b2-48dc-a0c0-d1c9c9f0ca63 | mhsnet-multi-head-and-spatial-attention | 2201.13392 | null | https://arxiv.org/abs/2201.13392v6 | https://arxiv.org/pdf/2201.13392v6.pdf | MHSnet: Multi-head and Spatial Attention Network with False-Positive Reduction for Pulmonary Nodules Detection | The mortality of lung cancer has ranked high among cancers for many years. Early detection of lung cancer is critical for disease prevention, cure, and mortality rate reduction. However, existing detection methods on pulmonary nodules introduce an excessive number of false positive proposals in order to achieve high sensitivity, which is not practical in clinical situations. In this paper, we propose the multi-head detection and spatial squeeze-and-attention network, MHSnet, to detect pulmonary nodules, in order to aid doctors in the early diagnosis of lung cancers. Specifically, we first introduce multi-head detectors and skip connections to customize for the variety of nodules in sizes, shapes and types and capture multi-scale features. Then, we implement a spatial attention module to enable the network to focus on different regions differently inspired by how experienced clinicians screen CT images, which results in fewer false positive proposals. Lastly, we present a lightweight but effective false positive reduction module with the Linear Regression model to cut down the number of false positive proposals, without any constraints on the front network. Extensive experimental results compared with the state-of-the-art models have shown the superiority of the MHSnet in terms of the average FROC, sensitivity and especially false discovery rate (2.98% and 2.18% improvement in terms of average FROC and sensitivity, 5.62% and 28.33% decrease in terms of false discovery rate and average candidates per scan). The false positive reduction module significantly decreases the average number of candidates generated per scan by 68.11% and the false discovery rate by 13.48%, which is promising to reduce distracted proposals for the downstream tasks based on the detection results. | ['Jinsheng Tao', 'Hua Ji', 'Bo wang', 'Xiangcheng Qiu', 'Yang Yang', 'Airu Yin', 'Jian You', 'Jianyu Xiao', 'Yulong Chen', 'Xinliang Fu', 'Zhaoqi Diao', 'Yanbo Shao', 'Jiayin Zheng', 'Minghao Wang', 'Juanyun Mai'] | 2022-01-31 | null | null | null | null | ['head-detection'] | ['computer-vision'] | [-1.76796631e-03 2.36735627e-01 -1.67810649e-01 -3.69112915e-03
-6.84698164e-01 -2.01538399e-01 2.97995031e-01 -1.33373961e-01
-6.08923495e-01 3.80802065e-01 1.98835842e-02 -3.70471448e-01
-1.25306502e-01 -9.19658542e-01 -3.63685668e-01 -8.32604825e-01
1.28070757e-01 4.20754611e-01 8.51386011e-01 2.18772173e-01
-2.07806662e-01 5.10090172e-01 -1.04856014e+00 3.33235860e-01
8.29009831e-01 7.56874442e-01 7.16052473e-01 3.81138265e-01
4.52911966e-02 5.35045922e-01 -2.47997284e-01 -1.10037521e-01
9.75300297e-02 -4.91705507e-01 -5.19536078e-01 -1.48466051e-01
-5.22545241e-02 -4.89424407e-01 -4.15816665e-01 9.77729976e-01
7.23901451e-01 -1.69608906e-01 6.73950076e-01 -7.11271405e-01
-3.81898046e-01 5.89008451e-01 -8.24638486e-01 5.04720271e-01
-4.47240591e-01 3.82836193e-01 7.66243994e-01 -1.07971418e+00
1.00429572e-01 9.83935714e-01 5.91295958e-01 8.71949613e-01
-6.53114080e-01 -7.84093797e-01 1.58909697e-03 -7.20988913e-03
-1.46186137e+00 -8.90321955e-02 8.60477984e-02 -3.19008261e-01
5.99504471e-01 6.82179630e-01 6.63724542e-01 7.27499008e-01
1.90156445e-01 5.03272593e-01 4.71453846e-01 -2.15978533e-01
-1.48968861e-01 3.69098097e-01 -1.87146485e-01 1.02028883e+00
7.50573575e-01 5.45186251e-02 2.14748740e-01 9.80827361e-02
1.24259257e+00 6.49335980e-01 -3.38452965e-01 -1.88280083e-02
-1.40219247e+00 9.18641210e-01 1.08452916e+00 5.11720300e-01
-3.45228523e-01 1.01901345e-01 2.49447703e-01 -4.73051906e-01
1.67823032e-01 5.31689167e-01 -3.05821389e-01 4.17383939e-01
-6.63053513e-01 -2.01408789e-01 3.70104551e-01 7.15357244e-01
1.85066879e-01 -1.59805208e-01 -9.98557389e-01 8.44392478e-01
4.37741816e-01 5.62702656e-01 7.39905834e-01 -4.71170574e-01
2.65192270e-01 9.38974917e-01 -6.90922439e-02 -8.13246846e-01
-7.48683274e-01 -9.80176151e-01 -1.28160059e+00 -3.65152925e-01
1.82678342e-01 -9.78899673e-02 -1.18368781e+00 1.46569300e+00
2.17319980e-01 2.20828891e-01 -4.66835022e-01 1.02676487e+00
9.29066896e-01 4.46581185e-01 1.96104407e-01 -3.79704118e-01
1.66175485e+00 -1.22495258e+00 -3.28358859e-01 -3.17855597e-01
9.37411070e-01 -4.93318975e-01 1.03311145e+00 -1.68585986e-01
-7.99854577e-01 -4.79770422e-01 -5.79270661e-01 2.74304330e-01
6.67714849e-02 7.30520070e-01 3.16013753e-01 4.65295583e-01
-8.02807271e-01 2.27955222e-01 -9.80978549e-01 -4.38407212e-01
7.70916045e-01 5.80816507e-01 9.47525054e-02 -1.57756716e-01
-9.55247402e-01 7.02356756e-01 4.25288558e-01 9.45429280e-02
-1.04018307e+00 -8.53005111e-01 -3.44516069e-01 5.40729582e-01
6.67825401e-01 -1.05580783e+00 1.25617445e+00 -5.42794347e-01
-1.01778615e+00 5.65725982e-01 -3.93594466e-02 -2.40766913e-01
6.15969837e-01 8.95252675e-02 -1.29432619e-01 6.40275180e-02
2.97792584e-01 8.17461610e-01 4.03408229e-01 -5.74626446e-01
-9.64583814e-01 -1.75142303e-01 -1.86241940e-01 3.55841964e-01
-4.71886128e-01 -1.32053554e-01 -8.86322916e-01 -5.87354302e-01
1.40675977e-01 -1.03043139e+00 -6.98015392e-01 2.28273168e-01
-3.67268741e-01 -3.22366387e-01 7.96924472e-01 -4.18647438e-01
1.25175166e+00 -2.10102940e+00 -2.42760152e-01 1.20500982e-01
5.93976736e-01 5.10242999e-01 5.16933501e-02 -3.73487443e-01
2.65580453e-02 4.62075353e-01 -1.48587242e-01 1.34483382e-01
-5.68915069e-01 3.64468358e-02 3.49201560e-01 2.42044881e-01
4.31425750e-01 1.05263674e+00 -9.37827229e-01 -7.30191827e-01
2.87145346e-01 4.71928269e-01 -7.32513785e-01 2.35224694e-01
2.29853928e-01 5.21288931e-01 -8.03820074e-01 4.94887084e-01
5.82796812e-01 -7.49698758e-01 1.36885449e-01 -1.75129369e-01
-1.98551208e-01 2.65927225e-01 -8.33750844e-01 1.04340243e+00
-4.05867457e-01 1.56931236e-01 -8.70071948e-02 -5.31956911e-01
5.10957778e-01 5.69210649e-01 4.95807976e-01 -4.90880132e-01
1.65774226e-01 2.88667887e-01 4.22965556e-01 -5.54629982e-01
-2.58819491e-01 -2.44210124e-01 1.65907010e-01 -6.96197152e-02
-4.26594049e-01 2.98914999e-01 1.29944071e-01 1.41581014e-01
1.31102860e+00 -5.84841013e-01 5.04207671e-01 -3.34024727e-01
7.29956210e-01 -1.19550288e-01 7.20628798e-01 9.05234456e-01
-1.26326799e-01 5.56158960e-01 3.33476126e-01 -2.74344057e-01
-7.09129810e-01 -9.44211781e-01 -1.22810632e-01 9.10433590e-01
4.74208593e-02 -9.56635922e-03 -5.58881044e-01 -1.03491461e+00
-2.05176905e-01 3.78929079e-01 -5.38555741e-01 -4.09411252e-01
-7.04133153e-01 -1.01555765e+00 4.11714792e-01 7.97498107e-01
6.30798995e-01 -1.14032507e+00 -7.13614047e-01 4.02752571e-02
-7.86407664e-02 -6.96491480e-01 -5.01897097e-01 1.08448491e-01
-9.56532180e-01 -1.09646070e+00 -1.04792523e+00 -9.14304316e-01
9.97768283e-01 5.97219229e-01 8.00829172e-01 4.25017238e-01
-6.85782909e-01 -1.60972461e-01 -2.07385793e-01 -4.73313361e-01
-1.52987644e-01 3.49017084e-01 -2.06655428e-01 -2.73681849e-01
1.46908268e-01 -1.59386825e-02 -8.84452105e-01 5.13461769e-01
-7.99156845e-01 7.21189231e-02 1.36618912e+00 1.01997018e+00
7.18309820e-01 8.27125013e-02 4.20231283e-01 -1.02833045e+00
1.10862471e-01 -6.26501024e-01 -4.25038069e-01 8.27929974e-02
-4.72591847e-01 -1.10151708e-01 5.65282524e-01 -4.61562932e-01
-1.04464507e+00 2.32836068e-01 -7.11276755e-02 -4.05875713e-01
8.90771002e-02 2.21809030e-01 9.99673158e-02 -2.55543627e-02
6.49275720e-01 1.28918052e-01 -2.39925355e-01 -3.62577498e-01
-1.39293671e-01 4.19539958e-01 8.37334320e-02 3.18491638e-01
7.59042382e-01 3.54509741e-01 2.23258480e-01 -5.92533886e-01
-1.02725327e+00 -7.59608746e-01 -4.36139196e-01 -7.43797123e-02
8.72285664e-01 -1.04063785e+00 -6.19503260e-01 3.62711884e-02
-9.30548787e-01 3.13541410e-03 -2.79238492e-01 7.40610301e-01
8.54061395e-02 1.96635216e-01 -5.76329470e-01 -4.87167716e-01
-6.82129443e-01 -1.21307480e+00 8.56142402e-01 4.87777352e-01
-8.61602928e-03 -6.09893501e-01 -4.69455361e-01 1.59320682e-02
6.61698639e-01 -2.48748213e-01 8.79153609e-01 -6.89324200e-01
-8.20675135e-01 -2.26970583e-01 -7.10442603e-01 1.90538764e-02
2.60006279e-01 -2.22545981e-01 -6.31563187e-01 -2.86771715e-01
-2.91850008e-02 1.56545520e-01 1.09838426e+00 8.17458570e-01
1.56938851e+00 -2.99219191e-01 -1.05925560e+00 5.63473046e-01
1.24497283e+00 2.79577225e-01 4.19267893e-01 6.87056803e-04
8.16089928e-01 1.77185580e-01 5.02951205e-01 4.13433403e-01
-1.42353147e-01 3.76655668e-01 7.44871259e-01 -5.82861722e-01
-3.55015337e-01 7.61913601e-03 -6.50896206e-02 6.07304752e-01
-1.02111012e-01 -4.31377232e-01 -9.76943433e-01 6.41451836e-01
-1.66208458e+00 -7.60797024e-01 -5.18908024e-01 2.10353279e+00
5.45321465e-01 2.25781962e-01 1.59660839e-02 -2.95889199e-01
1.06541014e+00 -2.11921334e-01 -4.08350050e-01 2.29051322e-01
5.09644926e-01 4.94360290e-02 6.76026940e-01 4.32496630e-02
-1.07644820e+00 5.78945220e-01 5.45793819e+00 8.91884208e-01
-1.17665696e+00 2.97451407e-01 8.56728971e-01 -2.74494588e-01
4.44193743e-02 -2.55100250e-01 -1.04315007e+00 4.48139876e-01
5.97925246e-01 5.91305904e-02 -1.40279373e-02 8.77985179e-01
3.68282467e-01 6.84828237e-02 -9.57524478e-01 7.99380839e-01
-3.07200938e-01 -1.25605178e+00 1.33197248e-01 1.01398379e-01
4.90117610e-01 2.06163839e-01 -3.53689715e-02 3.56923610e-01
6.25799522e-02 -8.73848498e-01 2.00962856e-01 4.21536475e-01
8.99361312e-01 -5.98119020e-01 1.18230927e+00 6.93646967e-01
-1.24892640e+00 -2.97450930e-01 -5.28998256e-01 3.19547296e-01
-3.49177746e-03 8.42507780e-01 -1.35684693e+00 2.32786611e-01
9.09378409e-01 4.21570987e-01 -6.81647003e-01 1.35022020e+00
-1.98860928e-01 7.33579278e-01 -5.42432785e-01 -4.40119267e-01
2.43097529e-01 2.22775042e-01 4.64598447e-01 1.30089414e+00
6.53802872e-01 2.87019938e-01 2.57285744e-01 9.59224582e-01
-2.65232950e-01 5.54503612e-02 -2.04163983e-01 2.41396382e-01
6.08257115e-01 1.62664795e+00 -8.96945655e-01 -4.26501542e-01
-3.53144377e-01 7.12657511e-01 1.12657594e-02 -4.78207543e-02
-1.32192683e+00 -2.15177312e-01 -1.12577349e-01 5.83546162e-01
3.60560715e-01 4.93956268e-01 -6.03465922e-02 -7.64411628e-01
-1.26510412e-01 -4.47775543e-01 4.83420789e-01 -3.83043528e-01
-1.14925289e+00 5.55496871e-01 -2.65162677e-01 -1.13477743e+00
2.38076240e-01 -5.41030169e-01 -9.85092700e-01 8.10730696e-01
-1.28623104e+00 -8.88822675e-01 -6.47285402e-01 2.72498667e-01
6.86635792e-01 1.19775213e-01 4.34153199e-01 4.93373930e-01
-9.97849703e-01 7.12484360e-01 -8.06220323e-02 1.97502136e-01
6.47524059e-01 -9.35314834e-01 6.00848487e-03 6.59013629e-01
-5.08738518e-01 4.32743758e-01 3.94616760e-02 -6.09163880e-01
-8.48579764e-01 -1.74025524e+00 7.82481670e-01 -1.58754751e-01
1.81082919e-01 1.04013115e-01 -8.01228225e-01 4.37243104e-01
-4.54842567e-01 4.91868913e-01 1.97461426e-01 -3.77651036e-01
2.55486697e-01 -1.48331970e-01 -1.10921860e+00 7.35289931e-01
9.03874755e-01 1.53059527e-01 -1.27398878e-01 5.35237551e-01
8.12067211e-01 -1.96693778e-01 -5.29567599e-01 7.95413494e-01
3.60739499e-01 -8.71271193e-01 1.02934539e+00 -1.39529064e-01
2.52049983e-01 -3.16308379e-01 4.30695564e-01 -6.97029591e-01
-1.34966540e+00 1.28617346e-01 1.41670898e-01 9.47894812e-01
7.73855329e-01 -5.62314272e-01 1.01163971e+00 2.30358124e-01
-4.19875592e-01 -1.09551525e+00 -8.85669768e-01 -4.59996760e-01
-2.12108791e-01 1.60205558e-01 8.63762796e-02 6.14747643e-01
-4.51949835e-01 4.14671242e-01 -1.26826428e-02 2.45818868e-01
1.24505684e-01 -2.87182897e-01 2.34585568e-01 -1.15806460e+00
-4.32363898e-01 -6.08946443e-01 -8.25065002e-02 -1.01136684e+00
-7.70384848e-01 -9.11396146e-01 1.71757400e-01 -1.71277034e+00
8.33352506e-01 -3.28607231e-01 -5.45520186e-01 6.17088377e-01
-5.76274633e-01 1.74500450e-01 -3.75631042e-02 3.83866340e-01
-5.14913559e-01 3.05097997e-01 1.45078015e+00 -6.43000975e-02
-1.44304052e-01 4.56143051e-01 -7.74384081e-01 9.35113549e-01
8.19439530e-01 -8.65428984e-01 -2.89974451e-01 -3.04752111e-01
-1.40936255e-01 9.00750607e-03 4.60856259e-01 -1.11250663e+00
1.57485992e-01 -2.16059655e-01 7.57068872e-01 -7.06236422e-01
1.08243182e-01 -7.19008029e-01 2.03297231e-02 1.23197138e+00
-6.41009435e-02 -2.51147270e-01 1.86101556e-01 8.07196379e-01
1.17801957e-01 -3.25704336e-01 1.11120737e+00 -4.76623267e-01
-3.96658599e-01 4.23537403e-01 -3.52988273e-01 -1.52809948e-01
1.39974701e+00 -1.90119803e-01 -1.42642349e-01 1.52515024e-01
-5.15069425e-01 3.65296900e-01 -7.95024559e-02 8.02061111e-02
5.22153437e-01 -1.21143723e+00 -7.85251021e-01 2.11036608e-01
-1.48222800e-02 4.13310230e-01 4.93590415e-01 1.33299744e+00
-4.94099528e-01 7.21443057e-01 1.79148868e-01 -6.25624299e-01
-1.29915607e+00 4.99593765e-01 6.70738876e-01 -8.26827109e-01
-5.00075996e-01 1.28064573e+00 9.20016289e-01 -1.48554146e-01
7.98882395e-02 -5.67835033e-01 -2.77690500e-01 -2.65001595e-01
2.77448535e-01 5.56002319e-01 1.02128603e-01 -3.36329751e-02
-6.35348082e-01 3.52770358e-01 -4.73516226e-01 5.38055301e-01
1.07276285e+00 2.05321968e-01 -1.86653305e-02 -1.59203410e-01
8.52049947e-01 -3.71754989e-02 -1.05076885e+00 -8.61050412e-02
-1.61414564e-01 -2.18255818e-01 2.98141927e-01 -9.82983768e-01
-1.29364800e+00 8.17049086e-01 8.78181756e-01 1.16436273e-01
1.15653265e+00 3.57173115e-01 6.67061865e-01 2.32224673e-01
-2.14114442e-01 -5.22579789e-01 2.18376458e-01 3.28191340e-01
7.40755856e-01 -1.20048332e+00 1.03234574e-01 -8.19229960e-01
-5.17000914e-01 8.73628616e-01 1.09681129e+00 7.05290679e-03
5.64036667e-01 1.28844783e-01 -3.38424534e-01 -2.74140805e-01
-7.56480932e-01 -6.84769154e-02 4.38553125e-01 2.06635460e-01
6.46708965e-01 2.08889097e-01 -4.97712880e-01 8.98044348e-01
4.56950009e-01 -1.21946849e-01 2.35163167e-01 5.80137432e-01
-8.83569002e-01 -6.10097110e-01 -3.68787885e-01 9.92418647e-01
-6.52081549e-01 -2.22363442e-01 -2.64195800e-01 9.51186776e-01
4.51000810e-01 5.61364293e-01 1.23166025e-01 -1.38101131e-01
4.33278084e-01 -3.77871871e-01 9.46219042e-02 -1.09499061e+00
-5.93467951e-01 4.07460481e-01 -1.59149036e-01 -2.17975691e-01
-6.36721179e-02 -2.82230884e-01 -1.47423589e+00 1.39044911e-01
-9.32474673e-01 -1.15435366e-02 1.04257077e-01 6.70313060e-01
4.10939753e-01 1.09312928e+00 5.42274952e-01 -5.07400811e-01
-7.20143616e-01 -1.13154984e+00 -3.49465072e-01 -9.23910737e-02
1.57799758e-02 -5.35905838e-01 -3.19506973e-01 -2.33290464e-01] | [15.395317077636719, -2.1502885818481445] |
82c3c1de-0445-4fea-9850-a9ca1b7a7842 | gun-source-and-muzzle-head-detection | 2001.11120 | null | https://arxiv.org/abs/2001.11120v1 | https://arxiv.org/pdf/2001.11120v1.pdf | Gun Source and Muzzle Head Detection | There is a surging need across the world for protection against gun violence. There are three main areas that we have identified as challenging in research that tries to curb gun violence: temporal location of gunshots, gun type prediction and gun source (shooter) detection. Our task is gun source detection and muzzle head detection, where the muzzle head is the round opening of the firing end of the gun. We would like to locate the muzzle head of the gun in the video visually, and identify who has fired the shot. In our formulation, we turn the problem of muzzle head detection into two sub-problems of human object detection and gun smoke detection. Our assumption is that the muzzle head typically lies between the gun smoke caused by the shot and the shooter. We have interesting results both in bounding the shooter as well as detecting the gun smoke. In our experiments, we are successful in detecting the muzzle head by detecting the gun smoke and the shooter. | ['Florian Metze', 'Isak Czeresnia Etinger', 'Zhong Zhou', 'Alexander Hauptmann', 'Alexander Waibel'] | 2020-01-29 | null | null | null | null | ['type-prediction', 'head-detection'] | ['computer-code', 'computer-vision'] | [ 2.85104692e-01 -3.08279723e-01 -1.23011604e-01 4.43933159e-01
-5.28868794e-01 -7.78806746e-01 5.52432597e-01 2.71269888e-01
-2.19715983e-01 2.95048356e-01 2.83083409e-01 -3.66836846e-01
9.96009558e-02 -7.44711936e-01 -4.78958458e-01 -5.98400891e-01
3.04818377e-02 1.78951561e-01 6.23915374e-01 -9.58214924e-02
7.17394590e-01 9.47586417e-01 -8.86847258e-01 3.71044815e-01
5.94292209e-02 7.51514554e-01 4.74106431e-01 1.15786350e+00
2.79503465e-01 1.46116805e+00 -9.67955470e-01 -2.00524583e-01
1.08180523e-01 -8.15459907e-01 -8.66499245e-01 -1.48292720e-01
5.04487455e-01 -6.20568216e-01 -4.56183702e-01 8.27186346e-01
1.24563858e-01 3.52506280e-01 8.25577378e-01 -1.36108088e+00
-1.92338794e-01 2.64861703e-01 -1.15903330e+00 1.10786843e+00
6.03421688e-01 -1.04148826e-02 2.90685177e-01 -7.60915756e-01
7.11133063e-01 1.37437510e+00 7.78616846e-01 6.91569269e-01
-5.42763472e-01 -8.41089129e-01 -3.28003436e-01 3.61645728e-01
-1.63043272e+00 -4.24516529e-01 6.44814610e-01 -1.09100294e+00
6.77590370e-01 4.08360451e-01 2.86829710e-01 9.87038672e-01
3.50587189e-01 3.87547970e-01 6.92728639e-01 -6.79329336e-01
1.53810352e-01 -1.46043032e-01 2.31182307e-01 9.78982508e-01
1.83784783e-01 4.31675941e-01 -2.99362093e-01 -4.27839249e-01
8.86015356e-01 5.85571155e-02 -7.37019330e-02 5.36860168e-01
-5.81700683e-01 1.13242638e+00 4.98044550e-01 2.65639901e-01
-4.28086132e-01 4.26634997e-01 3.83869648e-01 -6.57320738e-01
3.45576584e-01 1.46278322e-01 3.91900748e-01 -1.63064644e-01
-1.19255090e+00 6.22106791e-01 5.18878639e-01 4.84967887e-01
1.04191579e-01 -2.51348794e-01 -2.26449341e-01 4.05493885e-01
-6.18677959e-02 3.50451618e-01 -4.04634476e-01 -6.34867132e-01
7.26699829e-01 3.07747483e-01 9.72927809e-02 -1.35647786e+00
-3.40208471e-01 4.61371064e-01 -3.26056927e-01 5.98134100e-01
8.39399993e-01 -6.41460359e-01 -7.43894577e-01 1.28928947e+00
5.87153435e-01 6.23878896e-01 -5.67407191e-01 8.71384859e-01
9.49949265e-01 9.07284796e-01 1.83025435e-01 -1.38695091e-01
1.56049907e+00 -7.92004943e-01 -5.55715919e-01 -3.51003706e-01
5.08656383e-01 -6.27913415e-01 1.47391081e-01 2.88537610e-02
-9.07479048e-01 -1.55511707e-01 -5.85137665e-01 1.37950152e-01
-1.44749954e-01 -1.92278087e-01 3.34750533e-01 5.84697008e-01
-2.71714449e-01 6.36259496e-01 -7.66979635e-01 -5.55465281e-01
4.16353852e-01 -3.27178895e-01 -2.88809091e-01 -1.04515865e-01
-6.26681685e-01 1.31488883e+00 3.33269775e-01 1.16212629e-02
-1.21584642e+00 -4.65519845e-01 -6.95196390e-01 -2.16006070e-01
7.54067302e-01 -1.22925237e-01 1.16405189e+00 -4.54116523e-01
-4.89599556e-01 9.46419060e-01 -3.73230129e-01 -2.27147728e-01
3.54878068e-01 8.60724300e-02 -4.34189856e-01 3.36662889e-01
4.12281454e-01 -2.23604217e-01 8.17210913e-01 -1.41479862e+00
-1.04000533e+00 -5.90660453e-01 -3.95262800e-02 -1.33000612e-01
3.35470498e-01 1.19268250e+00 2.22236127e-01 -6.28898263e-01
-3.45396191e-01 -6.62437022e-01 -4.10037413e-02 -4.71057594e-01
-7.85410941e-01 -2.29743004e-01 1.28119731e+00 -1.04653370e+00
1.45393264e+00 -1.96311593e+00 -4.11727905e-01 1.22568183e-01
3.93431306e-01 5.14451206e-01 3.49309385e-01 6.37491226e-01
-1.83669582e-01 4.30094935e-02 -8.94256830e-02 -5.82950711e-02
-4.35932845e-01 -1.09285049e-01 -6.65636420e-01 7.64682472e-01
-1.59995556e-01 4.37832534e-01 -1.20206058e+00 -4.34858143e-01
2.62911506e-02 1.82155505e-01 -3.48929204e-02 3.38496029e-01
4.26423490e-01 3.62616837e-01 -4.68046129e-01 8.55175912e-01
5.38153648e-01 5.64700425e-01 -7.16072202e-01 1.97221532e-01
-6.21945262e-01 -9.15655047e-02 -1.20385325e+00 4.81543332e-01
1.35066852e-01 6.39930010e-01 5.22557259e-01 -3.88853014e-01
6.92581177e-01 5.10871053e-01 1.93736464e-01 -2.08554849e-01
3.30310374e-01 -4.07453142e-02 -3.32189113e-01 -9.77084279e-01
5.11318386e-01 -7.19091117e-01 1.70995250e-01 5.91727853e-01
-3.78041327e-01 3.86368632e-02 5.00937700e-01 1.08001910e-01
1.47643530e+00 -3.29836786e-01 2.96142817e-01 -1.32499054e-01
3.28269564e-02 3.54303092e-01 3.63264680e-01 9.00400758e-01
-3.37887108e-01 5.99584818e-01 4.48478639e-01 -9.51763764e-02
-6.31372035e-01 -8.99654806e-01 2.53019124e-01 1.34380579e+00
3.12940955e-01 -3.16408455e-01 -1.20604026e+00 -5.83337903e-01
-2.84205317e-01 1.12464714e+00 -9.20712411e-01 -1.24631919e-01
-8.14490855e-01 -2.59567082e-01 8.19186747e-01 8.73454690e-01
6.45496622e-02 -1.24631560e+00 -1.10490966e+00 1.03347972e-01
-3.72164428e-01 -9.37358320e-01 -6.57194018e-01 -4.56528217e-02
-2.70302176e-01 -1.46294403e+00 -5.20416319e-01 -3.59344184e-01
2.15585887e-01 5.98861754e-01 6.58630669e-01 4.13197458e-01
-9.73506510e-01 2.83891886e-01 -4.61438775e-01 -1.07776082e+00
-4.89884228e-01 -6.96885943e-01 -1.20697886e-01 -1.12674400e-01
6.55914471e-02 -3.96810770e-02 -2.41730154e-01 1.37988418e-01
-5.76775014e-01 -2.92303950e-01 -4.12987262e-01 -1.54379055e-01
7.89895803e-02 5.43951750e-01 1.58340201e-01 -5.47672868e-01
7.13928342e-01 -8.95627856e-01 -3.04521233e-01 -9.03944543e-04
4.12149221e-01 -7.95744538e-01 3.74685287e-01 -2.14999452e-01
-9.35743928e-01 -1.73798010e-01 -1.47510037e-01 -5.48033416e-01
-6.98997617e-01 -1.31629437e-01 1.78893954e-01 2.29705915e-01
9.69407141e-01 -3.54576617e-01 -3.98908675e-01 -5.74894965e-01
1.28324524e-01 4.75114286e-01 9.84941244e-01 -2.81153411e-01
8.97087812e-01 7.52000213e-01 2.18647957e-01 -1.02936578e+00
-1.08273017e+00 -8.58966231e-01 -6.31116569e-01 -7.42224038e-01
1.47886539e+00 -1.92707434e-01 -8.50816667e-01 3.22426587e-01
-1.69361413e+00 -7.69003630e-02 1.34330362e-01 2.14754090e-01
-2.29172155e-01 3.66024703e-01 -7.83612788e-01 -1.43313277e+00
-3.74047279e-01 -8.50855172e-01 9.98419523e-01 4.32045728e-01
-4.27468508e-01 -8.74878824e-01 3.55796009e-01 4.88870323e-01
1.30132819e-02 9.03972089e-01 8.21428716e-01 -3.56354713e-01
-1.52221188e-01 -6.44345939e-01 -2.00637951e-01 -5.23762777e-02
1.65156424e-01 2.24317521e-01 -8.57137263e-01 8.16071108e-02
1.15981951e-01 2.42323697e-01 9.10475194e-01 5.80845833e-01
6.06630921e-01 -9.33220759e-02 -7.93603539e-01 3.17844838e-01
1.34233177e+00 7.74137139e-01 3.91833991e-01 3.01199228e-01
9.54023302e-01 4.87194866e-01 7.50774682e-01 2.51328111e-01
-1.88277751e-01 7.26214528e-01 6.88078225e-01 -1.19502530e-01
-2.68196315e-01 -3.58228028e-01 3.89460742e-01 -3.51796895e-01
-7.94499040e-01 -4.84954178e-01 -1.12828803e+00 4.51017827e-01
-1.55488205e+00 -1.86771846e+00 -7.76837945e-01 2.00562596e+00
1.37406111e-01 -1.89612716e-01 7.39641488e-01 1.98554769e-01
1.12749565e+00 1.32212996e-01 1.02796830e-01 -5.74041307e-01
6.36812866e-01 -5.01569360e-02 6.29757047e-01 7.75994837e-01
-1.33181739e+00 7.84167945e-01 6.80960655e+00 9.10668135e-01
-7.38190711e-01 2.47277364e-01 1.31480068e-01 -9.43678319e-02
6.51724458e-01 -1.31124062e-02 -1.07396805e+00 4.64569867e-01
5.44200838e-01 2.41396148e-02 4.09417897e-01 5.24819195e-01
5.38780212e-01 -7.15500414e-01 -9.68572497e-01 6.85992599e-01
3.94453198e-01 -1.35220993e+00 -6.08845651e-01 -1.32042259e-01
2.43719652e-01 -4.35301572e-01 -6.13062739e-01 -1.55634642e-01
1.32468909e-01 -1.16991937e+00 1.21288550e+00 5.03356099e-01
3.55332315e-01 -9.40688372e-01 1.59173757e-01 9.17336702e-01
-1.44658935e+00 -2.23187193e-01 -2.38891229e-01 -4.16793019e-01
7.27822185e-01 4.63353306e-01 -1.08327913e+00 1.09718308e-01
4.84107316e-01 1.86844580e-02 -3.25333804e-01 1.26741910e+00
-4.46270645e-01 6.25202060e-01 -7.06706569e-02 2.40424186e-01
2.89231449e-01 -1.14399754e-01 9.44727838e-01 1.60083127e+00
1.84989721e-01 3.75872076e-01 2.99113065e-01 9.37623084e-01
5.20189643e-01 -2.94008225e-01 -8.18781257e-01 1.08895414e-01
4.29209560e-01 1.11778963e+00 -1.14351392e+00 -2.43459553e-01
8.69763345e-02 6.22334421e-01 5.98208643e-02 -7.15939477e-02
-1.04844022e+00 -4.22805190e-01 3.96810740e-01 6.55933976e-01
3.95467095e-02 9.71701071e-02 -2.30407417e-01 -4.19582576e-01
-1.58352405e-01 -3.47894460e-01 6.75069153e-01 -9.49200869e-01
-9.14395690e-01 5.20349264e-01 4.16329741e-01 -5.70443034e-01
-3.80605698e-01 -4.39929157e-01 -1.45718503e+00 9.12621200e-01
-5.22392869e-01 -1.01379633e+00 -3.20956588e-01 6.18713677e-01
5.24058461e-01 1.81133583e-01 5.48057914e-01 8.38509500e-02
-7.73218453e-01 1.89694166e-01 -4.06230897e-01 5.83844781e-01
-1.00267977e-01 -7.95524359e-01 2.01906711e-01 1.29920137e+00
-1.24983341e-01 2.49049589e-01 9.38675880e-01 -1.20829463e+00
-9.23743963e-01 -1.04891074e+00 1.00407743e+00 -1.01097846e+00
8.34290564e-01 -4.11081672e-01 -6.24546468e-01 8.17846894e-01
-2.50786453e-01 -3.00194263e-01 2.03434408e-01 -2.83315450e-01
-2.01525927e-01 7.73450851e-01 -1.38035512e+00 4.49589312e-01
7.87724853e-01 -3.54582340e-01 -7.95490384e-01 8.11669409e-01
-1.11586340e-01 -4.24045354e-01 -9.82354358e-02 -4.48201261e-02
5.15886992e-02 -9.71630692e-01 1.10638034e+00 -6.13183022e-01
6.41082466e-01 -1.37652874e-01 2.92613775e-01 -9.05499697e-01
-5.41270494e-01 -1.57799423e-01 -2.24562034e-01 1.44140673e+00
-1.93236127e-01 1.81358919e-01 5.28802991e-01 5.52816451e-01
-3.62126231e-02 -4.63889480e-01 -1.02433312e+00 -8.99163604e-01
-2.62841713e-02 -4.37691152e-01 2.20768638e-02 6.91624582e-01
4.17575240e-01 4.15011197e-01 -5.12719274e-01 3.55052829e-01
7.67266810e-01 -1.61322623e-01 4.62882936e-01 -8.50911140e-01
-2.51113743e-01 -5.64856768e-01 -2.26040676e-01 -2.78238833e-01
-2.54611522e-01 -7.71778405e-01 4.15095270e-01 -1.81093514e+00
7.95634329e-01 4.01536040e-02 2.29355916e-01 6.14905119e-01
-3.42380852e-01 1.90873519e-01 5.21947503e-01 -9.84362066e-02
-2.99325176e-02 -6.21546805e-01 1.08259344e+00 1.54100075e-01
1.08569294e-01 2.64079154e-01 -2.90626884e-01 1.52500534e+00
6.40339315e-01 -9.64229047e-01 -2.93158162e-02 -2.65965223e-01
-3.33757222e-01 6.26021504e-01 8.79647195e-01 -7.78788686e-01
1.91372856e-01 -8.17653656e-01 1.43641129e-01 -6.08578622e-01
4.46574688e-01 -6.18600130e-01 3.48282978e-02 5.78610361e-01
3.56320012e-03 8.58053863e-02 9.12880823e-02 3.87380064e-01
4.91413057e-01 -1.16133881e+00 1.16378963e+00 -3.70157093e-01
-5.63504815e-01 2.54890621e-01 -9.68584359e-01 5.91955446e-02
1.54614115e+00 -1.93068400e-01 -3.39555383e-01 -1.33625850e-01
-7.02536821e-01 -2.44128913e-01 7.24643618e-02 4.55391765e-01
5.82859278e-01 -9.99149323e-01 -6.82256401e-01 -1.39245585e-01
-3.16112071e-01 -5.35863340e-01 1.34400681e-01 8.79939139e-01
-6.40580893e-01 1.24518052e-01 1.95021808e-01 2.61060074e-02
-1.94113684e+00 1.00266588e+00 4.81462449e-01 1.12955654e-02
-9.65325773e-01 1.31373715e+00 1.35325760e-01 3.76043767e-01
2.71862745e-01 1.15265094e-01 -3.43450397e-01 1.04143746e-01
1.37923813e+00 1.30728614e+00 -2.37697706e-01 -1.22254825e+00
-6.43099725e-01 2.58745015e-01 1.74321622e-01 1.24981590e-01
1.03016901e+00 2.05371842e-01 -2.31691509e-01 4.37433720e-01
8.93798411e-01 1.87874734e-01 -8.09174180e-01 3.62790048e-01
-1.01894274e-01 -8.72402251e-01 5.02008721e-02 -5.03917038e-01
-8.32971513e-01 8.14999640e-01 3.69421899e-01 3.55867386e-01
6.52752161e-01 5.23920774e-01 1.07347167e+00 -1.79986611e-01
5.35232008e-01 -1.04651392e+00 -1.64179504e-02 5.17155230e-01
9.42494690e-01 -6.81157947e-01 -1.43199727e-01 -6.90300882e-01
-5.87129056e-01 1.04815650e+00 4.47314322e-01 -4.09427673e-01
4.47362751e-01 7.08081603e-01 -3.83044332e-02 -7.07875907e-01
-3.90440188e-02 -2.35647902e-01 7.88427889e-02 6.47955537e-01
9.07097608e-02 2.75739014e-01 -7.55785406e-02 3.15331757e-01
-2.08865091e-01 -1.08563974e-01 6.06568635e-01 1.28662181e+00
-1.15632653e+00 -2.24643707e-01 -1.27719235e+00 4.62063521e-01
-7.38030672e-01 4.02807370e-02 -7.63868690e-01 5.37524581e-01
6.17925942e-01 1.55992651e+00 -2.99245212e-02 -2.51464546e-01
4.90514427e-01 -1.33609146e-01 6.70161903e-01 -8.25700343e-01
-7.60173380e-01 -3.05271223e-02 7.52246380e-02 -3.79699171e-01
1.22274198e-01 -6.05457067e-01 -1.55174577e+00 -5.45239210e-01
-4.95850563e-01 6.42753989e-02 4.46714133e-01 1.39762390e+00
-7.00265765e-01 2.59334326e-01 3.82348120e-01 -1.10635865e+00
-1.29369527e-01 -8.27467561e-01 -1.02518988e+00 4.24688071e-01
4.76940453e-01 -8.07925880e-01 -4.36376125e-01 1.39338106e-01] | [8.346089363098145, 0.18464429676532745] |
23848ae5-0be9-4540-92af-eb4bc6354ea0 | piveted-granite-computational-phenotypes | 1808.02602 | null | http://arxiv.org/abs/1808.02602v1 | http://arxiv.org/pdf/1808.02602v1.pdf | PIVETed-Granite: Computational Phenotypes through Constrained Tensor Factorization | It has been recently shown that sparse, nonnegative tensor factorization of
multi-modal electronic health record data is a promising approach to
high-throughput computational phenotyping. However, such approaches typically
do not leverage available domain knowledge while extracting the phenotypes;
hence, some of the suggested phenotypes may not map well to clinical concepts
or may be very similar to other suggested phenotypes. To address these issues,
we present a novel, automatic approach called PIVETed-Granite that mines
existing biomedical literature (PubMed) to obtain cannot-link constraints that
are then used as side-information during a tensor-factorization based
computational phenotyping process. The resulting improvements are clearly
observed in experiments using a large dataset from VUMC to identify phenotypes
for hypertensive patients. | ['Joydeep Ghosh', 'Bradley A. Malin', 'Jette Henderson', 'Joyce C. Ho'] | 2018-08-08 | null | null | null | null | ['computational-phenotyping'] | ['medical'] | [ 1.14571996e-01 -2.24576950e-01 -2.81099170e-01 -4.02055174e-01
-5.62106967e-01 -6.27128303e-01 -2.54299194e-01 6.98783517e-01
-3.03227380e-02 9.96968627e-01 4.25488651e-01 -3.72025102e-01
-6.48024082e-01 -4.86648858e-01 -3.58362854e-01 -4.54530358e-01
-1.77547514e-01 9.42471266e-01 -3.64198208e-01 1.96585447e-01
7.32585639e-02 3.40714395e-01 -1.14434469e+00 6.97743714e-01
1.17704713e+00 4.51849490e-01 1.82649903e-02 5.33847809e-01
-1.13474265e-01 5.12214124e-01 -3.91721517e-01 -5.68578124e-01
-1.19422100e-01 -1.70147195e-01 -7.09671974e-01 -4.88347560e-02
1.57163218e-01 1.94750965e-01 1.58981636e-01 9.19958770e-01
3.55476528e-01 -2.97965288e-01 4.03005093e-01 -1.06438756e+00
-5.75426996e-01 5.10838270e-01 -5.44368207e-01 1.87467128e-01
5.57597935e-01 -2.46465892e-01 1.05810177e+00 -8.07206929e-01
8.85856628e-01 1.04967093e+00 7.71790862e-01 4.28849667e-01
-1.61270595e+00 -2.61007726e-01 -4.90917712e-02 1.80833921e-01
-1.32294977e+00 -3.15986574e-02 5.43162346e-01 -7.39660680e-01
9.19981182e-01 7.50359595e-01 7.98573315e-01 8.15769911e-01
1.40798748e-01 5.15090108e-01 9.22608674e-01 -1.69686303e-01
7.36010596e-02 -8.84421021e-02 1.58657283e-01 7.48968482e-01
6.20185673e-01 -2.13938922e-01 -6.27305031e-01 -1.35702777e+00
1.07079737e-01 2.75663316e-01 -3.53685379e-01 -2.34711155e-01
-1.44167507e+00 5.79006791e-01 -1.85761169e-01 2.46844262e-01
-8.41947496e-01 -3.62167418e-01 4.80601430e-01 2.75355637e-01
4.89969105e-01 6.58466101e-01 -7.94108570e-01 -1.39977694e-01
-9.10259783e-01 2.77442455e-01 6.98815346e-01 6.81375623e-01
6.91230357e-01 -4.71426934e-01 -1.15392633e-01 9.74487841e-01
2.58927017e-01 8.20937157e-02 4.95076150e-01 -5.57226181e-01
3.63192111e-01 1.17451680e+00 7.70310685e-02 -9.76414025e-01
-6.94928706e-01 -6.33585602e-02 -8.58741820e-01 -5.14322937e-01
2.89937168e-01 -2.79486150e-01 -8.31446528e-01 1.46087742e+00
7.34475911e-01 5.20253241e-01 2.29692757e-02 7.37937331e-01
7.57543206e-01 -4.11598533e-02 3.40392053e-01 -2.41263568e-01
1.64762580e+00 -2.59676456e-01 -6.43576145e-01 3.17607939e-01
9.21080530e-01 -8.44087124e-01 5.97598612e-01 7.50222445e-01
-7.65016675e-01 -5.65540604e-02 -5.90348840e-01 1.00390121e-01
-2.66712070e-01 5.76791286e-01 1.28245950e+00 6.48794055e-01
-7.73884773e-01 6.66469872e-01 -8.79200578e-01 -4.55095142e-01
5.45309305e-01 4.53488946e-01 -5.83581686e-01 -3.79272789e-01
-8.45468342e-01 5.55908024e-01 3.67854089e-01 1.20065764e-01
-4.09298450e-01 -1.31429434e+00 -5.35055161e-01 7.88806845e-03
2.93069959e-01 -1.11557329e+00 4.49645489e-01 -1.99945584e-01
-9.19712484e-01 5.11651218e-01 -6.06331885e-01 -1.33250400e-01
3.72126028e-02 -1.02345787e-01 -7.58624434e-01 2.73852587e-01
3.12729180e-01 2.02414617e-01 2.67107487e-01 -7.09319472e-01
-4.42068189e-01 -7.51710415e-01 -2.36974537e-01 -1.35608539e-01
-3.70923221e-01 1.07623130e-01 -2.71718711e-01 -6.43230557e-01
3.49904627e-01 -1.01542151e+00 -5.64956963e-01 -2.72135615e-01
-8.18333626e-01 -2.13708833e-01 5.89910686e-01 -7.63742387e-01
1.23951805e+00 -1.81355703e+00 4.11837757e-01 3.49476337e-01
7.20486104e-01 1.27328411e-01 2.12004930e-02 5.18503964e-01
-3.85932416e-01 3.80317837e-01 -7.68622160e-02 8.56904238e-02
-5.90270817e-01 2.42534012e-01 -2.74067894e-02 5.04736423e-01
4.82821286e-01 8.13898146e-01 -1.04184186e+00 -3.45382690e-01
4.08564731e-02 4.40310001e-01 -8.47653389e-01 1.54190790e-02
-2.71759629e-01 7.37547278e-01 -8.84376943e-01 9.91823792e-01
7.08886325e-01 -5.92676342e-01 8.19473386e-01 -3.80712777e-01
8.91247541e-02 1.26108965e-02 -8.40320170e-01 1.76009345e+00
7.64280781e-02 -4.40067314e-02 -1.40612796e-01 -9.67062175e-01
4.23869997e-01 6.11840248e-01 1.24658322e+00 9.35164168e-02
-1.79333195e-01 8.87637660e-02 2.74811149e-01 -7.61458158e-01
2.26631984e-01 -1.58272773e-01 3.23293120e-01 3.56016219e-01
-1.13574818e-01 5.32068431e-01 3.73956025e-01 2.98779547e-01
1.48530269e+00 -4.12188657e-03 3.49919409e-01 -2.34917670e-01
5.53508639e-01 4.26316530e-01 1.10157692e+00 2.79885471e-01
2.91774273e-01 4.82254118e-01 7.45917737e-01 -5.08433282e-01
-7.99459994e-01 -6.71800137e-01 -3.19043964e-01 4.70940262e-01
-4.93587285e-01 -1.11842692e+00 -8.16036910e-02 -7.85651445e-01
4.43537325e-01 3.82400937e-02 -7.43965268e-01 1.99114054e-01
-1.38314784e-01 -1.33945262e+00 7.34356582e-01 3.06236923e-01
-4.53178674e-01 -3.92147988e-01 -2.17123106e-01 4.54773784e-01
-2.07668349e-01 -1.10134256e+00 -1.51604548e-01 -7.28054047e-02
-1.16912186e+00 -1.68475556e+00 -6.17734671e-01 -1.25297219e-01
1.08827245e+00 -7.63297752e-02 1.04786646e+00 3.07422012e-01
-6.15831196e-01 1.98027678e-02 -4.47447807e-01 -2.44956553e-01
-3.08901876e-01 -3.09544653e-01 3.39288622e-01 1.17776662e-01
5.64964414e-01 -6.71741664e-01 -6.14948034e-01 1.80719048e-01
-7.55382538e-01 -1.40529081e-01 6.40666962e-01 1.13735116e+00
8.69016886e-01 5.17497510e-02 7.06508398e-01 -1.53227496e+00
6.58780217e-01 -8.86084199e-01 -4.46518779e-01 4.42246199e-01
-7.67109871e-01 7.93689042e-02 7.37878442e-01 -5.30555427e-01
-7.96734154e-01 3.96933585e-01 -8.01830590e-02 -4.51173931e-01
-2.55755514e-01 1.24937022e+00 -1.13137171e-01 -9.07661095e-02
4.30846125e-01 -8.03754199e-03 -1.81852698e-01 -8.61636162e-01
1.11851282e-01 5.52676558e-01 -1.03198938e-01 -7.32474029e-01
3.45874697e-01 6.80492282e-01 2.38380745e-01 -6.89883232e-01
-5.48771203e-01 -8.15770507e-01 -5.18277645e-01 4.39567626e-01
7.50409186e-01 -8.14910173e-01 -1.06158662e+00 -1.33436531e-01
-7.76570439e-01 2.91616201e-01 1.95789814e-01 7.31804371e-01
-1.91455141e-01 6.86398804e-01 -5.17917693e-01 -5.69172204e-01
-3.39139014e-01 -8.67164433e-01 8.82968366e-01 -2.90127069e-01
-5.84626257e-01 -8.50430071e-01 2.96981454e-01 5.40829122e-01
4.83637825e-02 3.58923018e-01 1.74529231e+00 -7.03771949e-01
-1.36922300e-01 -1.72624916e-01 -1.09264746e-01 -1.85769364e-01
4.89356279e-01 1.97078228e-01 -6.27654016e-01 -1.68824896e-01
-4.18491840e-01 -4.98172417e-02 6.70885861e-01 3.84629577e-01
1.19590223e+00 -1.40206233e-01 -7.60707259e-01 6.47759676e-01
1.31876636e+00 8.96785185e-02 4.33935255e-01 -2.59039581e-01
9.21858907e-01 5.45839787e-01 6.73759520e-01 7.37309635e-01
4.50922370e-01 7.52072275e-01 -1.00953102e-01 -1.70254901e-01
2.25172400e-01 2.82509900e-05 -7.87857175e-02 7.41864026e-01
-2.89513111e-01 -1.34392887e-01 -9.78161156e-01 6.88637316e-01
-1.92216182e+00 -6.08627379e-01 -6.91123188e-01 1.99032390e+00
9.09608901e-01 -4.81775999e-01 2.04601571e-01 -1.47555918e-02
7.13562548e-01 -5.24079144e-01 -4.90358204e-01 -1.74165174e-01
-1.68533251e-01 2.17036575e-01 2.94425815e-01 -1.45401239e-01
-7.54587770e-01 5.69606602e-01 7.07295036e+00 2.37622514e-01
-9.48721170e-01 1.75282329e-01 4.02870268e-01 -6.47347271e-02
-5.34466863e-01 2.60008693e-01 -4.03561205e-01 3.80001545e-01
8.87921512e-01 -1.51983961e-01 2.03673989e-01 5.97405434e-01
4.65384930e-01 -1.37062356e-01 -1.24825692e+00 8.64333928e-01
-2.43854672e-01 -1.60838246e+00 4.09612544e-02 3.92742962e-01
5.70969582e-01 1.97378904e-01 -1.43337131e-01 -2.87227571e-01
4.16386515e-01 -8.02138865e-01 -7.18950182e-02 6.07861519e-01
9.43620026e-01 -6.01393998e-01 5.96404612e-01 2.62949169e-02
-9.47703779e-01 -4.92406711e-02 -5.78649163e-01 1.04547188e-01
7.96817094e-02 1.29539359e+00 -1.38627732e+00 1.00786567e+00
5.42676508e-01 8.33406270e-01 -5.32219529e-01 1.22463346e+00
-5.21688201e-02 9.15186286e-01 -2.08568573e-01 2.92422861e-01
-3.27637017e-01 -1.64190814e-01 5.20434439e-01 9.16029871e-01
3.81419927e-01 2.69132078e-01 2.66332120e-01 1.00865734e+00
4.62081060e-02 4.09657031e-01 -5.30746162e-01 -6.61235631e-01
2.15977237e-01 1.42500174e+00 -6.67958379e-01 -4.22885090e-01
-6.51871800e-01 7.44805932e-01 2.42331967e-01 3.97373438e-01
-3.67379159e-01 5.49942367e-02 9.44970310e-01 3.67757753e-02
6.67762011e-02 -9.97322612e-03 -1.88461870e-01 -1.68472576e+00
-4.31363136e-02 -9.53448713e-01 8.50836992e-01 -4.11746442e-01
-1.60304832e+00 2.58484721e-01 -4.22086418e-01 -1.21351826e+00
-9.09881741e-02 -3.69944662e-01 -1.43214226e-01 1.11042869e+00
-1.28679502e+00 -9.96368945e-01 1.20300986e-01 5.87550163e-01
-1.49355933e-01 -2.40837052e-01 1.43193519e+00 5.69565117e-01
-8.51133823e-01 3.56683165e-01 9.27911028e-02 1.01568364e-02
7.08633602e-01 -1.36951983e+00 1.85327128e-01 7.08936870e-01
7.22295940e-02 1.29268134e+00 5.35800338e-01 -1.23213196e+00
-1.95592463e+00 -1.30880725e+00 9.19359982e-01 -4.96289641e-01
6.63933814e-01 -6.41495660e-02 -8.78693044e-01 5.75527787e-01
-4.30245489e-01 1.87588975e-01 1.59422839e+00 8.43604267e-01
-3.43159288e-01 1.42904073e-01 -1.40000570e+00 2.87682414e-01
8.52427602e-01 -2.85052776e-01 -3.30515116e-01 5.43541551e-01
2.07404837e-01 -1.78585991e-01 -1.57712555e+00 3.44540119e-01
5.65165639e-01 -4.71718758e-01 8.11272800e-01 -1.24894321e+00
3.67473602e-01 -5.62331021e-01 7.74953142e-02 -1.30832028e+00
-3.92671585e-01 -5.13267875e-01 -2.00013801e-01 9.79815900e-01
6.39630675e-01 -5.40250301e-01 1.10605252e+00 7.63322175e-01
1.32674882e-02 -7.45561421e-01 -9.18029964e-01 -3.67444128e-01
-3.62211853e-01 -3.10549825e-01 7.87079811e-01 1.44553804e+00
4.33678806e-01 1.93793535e-01 -4.24814194e-01 4.31518406e-01
4.80249047e-01 4.17910725e-01 5.79828858e-01 -1.71027958e+00
-4.65143472e-01 1.86582044e-01 -7.24321604e-01 -2.52354771e-01
-2.89769303e-02 -1.10954392e+00 -5.33647776e-01 -1.52332652e+00
5.66197932e-01 -8.00274849e-01 -5.71666837e-01 8.85540187e-01
-4.37173486e-01 2.22987309e-01 -3.88235718e-01 1.13729917e-01
-2.60765851e-01 4.16292213e-02 9.59310174e-01 -1.69771761e-01
-2.27391198e-01 -5.91980964e-02 -1.12741733e+00 4.39536214e-01
5.02261639e-01 -7.09784210e-01 -4.50275421e-01 -1.86798036e-01
7.12788463e-01 3.88935715e-01 1.99003026e-01 -4.82451141e-01
-2.44000927e-02 -3.91732246e-01 4.90746677e-01 -7.06024110e-01
2.90934622e-01 -7.23608732e-01 5.92671096e-01 5.02692699e-01
-1.96014449e-01 1.77356765e-01 6.45562485e-02 8.31504345e-01
-1.62228435e-01 1.60831273e-01 4.59774807e-02 -7.41063505e-02
-4.27739650e-01 2.87938982e-01 -3.87021542e-01 -2.08155423e-01
9.03507113e-01 4.01044451e-02 -3.26726377e-01 1.04230039e-01
-1.03032970e+00 2.40120083e-01 3.92340869e-01 3.10867965e-01
7.86311626e-01 -1.05945396e+00 -8.73539984e-01 1.23008907e-01
3.93273324e-01 -3.30987662e-01 4.74361837e-01 1.31003928e+00
-3.50251108e-01 5.95606506e-01 -7.90746957e-02 -5.38496137e-01
-1.45182204e+00 7.85765111e-01 -1.31217048e-01 -2.29550108e-01
-8.13168585e-01 7.70831883e-01 -1.09423010e-03 -3.75657171e-01
-3.86099398e-01 -3.65656763e-01 -3.57869059e-01 1.56923056e-01
5.87553978e-01 2.41673961e-01 5.13012946e-01 -5.82489371e-01
-7.83035159e-01 2.19397977e-01 -1.65636629e-01 2.81674206e-01
1.78168380e+00 2.02975273e-01 -6.85734153e-01 2.12266177e-01
8.33637893e-01 4.38993156e-01 -3.03929925e-01 -1.82044692e-02
4.88439709e-01 -6.06539845e-01 -2.70342350e-01 -9.73761916e-01
-1.22447145e+00 3.94465417e-01 2.61655897e-01 -5.61443865e-02
9.94181454e-01 -1.81587040e-01 6.27312422e-01 3.41862917e-01
4.21202451e-01 -7.01969683e-01 -6.84196472e-01 1.26272246e-01
5.33101678e-01 -8.14406216e-01 2.86604106e-01 -9.35907006e-01
-7.40505815e-01 1.13457429e+00 2.16627330e-01 2.46856645e-01
5.98315120e-01 2.52641499e-01 4.66762148e-02 -6.77973688e-01
-1.17776346e+00 -2.36770492e-02 4.26087499e-01 6.53736532e-01
8.14010084e-01 3.60070080e-01 -7.83064485e-01 7.01342881e-01
6.79542264e-03 3.84956986e-01 5.47672629e-01 8.19095731e-01
2.78106898e-01 -1.67563283e+00 -2.87117362e-01 1.08633006e+00
-9.18094218e-01 -4.37793732e-01 -5.54540813e-01 1.24259636e-01
3.75494510e-01 1.03589153e+00 -5.76517284e-01 -6.00897789e-01
1.12641722e-01 2.29671836e-01 3.22296143e-01 -1.05507827e+00
-6.16367519e-01 2.30022252e-01 3.32330108e-01 -6.58406258e-01
-5.31518340e-01 -8.48896861e-01 -1.10779858e+00 3.72952521e-02
-4.15242314e-01 5.24434626e-01 4.14777040e-01 9.27917659e-01
1.00282550e+00 5.76542556e-01 1.36723027e-01 -3.77810858e-02
2.79742572e-02 -6.64313197e-01 -6.35011315e-01 3.85588855e-01
-7.29678292e-03 -7.64411032e-01 2.63310760e-01 5.21968976e-02] | [6.30036735534668, 5.876377105712891] |
d419837e-a0f6-4ce2-b93a-16ac8bc5b28a | ppg-signals-for-hypertension-diagnosis-a | 2304.06952 | null | https://arxiv.org/abs/2304.06952v1 | https://arxiv.org/pdf/2304.06952v1.pdf | PPG Signals for Hypertension Diagnosis: A Novel Method using Deep Learning Models | Hypertension is a medical condition characterized by high blood pressure, and classifying it into its various stages is crucial to managing the disease. In this project, a novel method is proposed for classifying stages of hypertension using Photoplethysmography (PPG) signals and deep learning models, namely AvgPool_VGG-16. The PPG signal is a non-invasive method of measuring blood pressure through the use of light sensors that measure the changes in blood volume in the microvasculature of tissues. PPG images from the publicly available blood pressure classification dataset were used to train the model. Multiclass classification for various PPG stages were done. The results show the proposed method achieves high accuracy in classifying hypertension stages, demonstrating the potential of PPG signals and deep learning models in hypertension diagnosis and management. | ['Brintha Therese A', 'Yaswant T', 'Graham Frederick'] | 2023-04-14 | null | null | null | null | ['photoplethysmography-ppg'] | ['medical'] | [-1.81365654e-01 -1.53430656e-01 -8.85808393e-02 -6.44365132e-01
-3.94792072e-02 -8.20538774e-02 8.16535670e-03 -1.41072035e-01
-1.67417347e-01 9.22015071e-01 4.18734819e-01 -3.76870215e-01
1.65188447e-01 -9.64648128e-01 9.37034041e-02 -6.71972156e-01
-5.52731037e-01 3.26288909e-01 -2.63973653e-01 2.34045401e-01
2.44076177e-01 6.95599020e-01 -8.29490244e-01 2.76690841e-01
9.77534294e-01 1.31817460e+00 -5.44669032e-01 1.18141687e+00
-2.00760484e-01 4.25781906e-01 -7.17356324e-01 1.52991846e-01
6.56558394e-01 -8.52319837e-01 -2.32280567e-01 8.50337669e-02
9.94464755e-01 -7.14302123e-01 -3.98760974e-01 7.50619829e-01
1.33342612e+00 -3.01565200e-01 3.42185795e-01 -6.18161380e-01
-5.89853823e-01 3.48184139e-01 -3.69371176e-01 7.23150849e-01
6.81161433e-02 3.69546026e-01 3.98118019e-01 -3.82833123e-01
-9.78143662e-02 9.91914749e-01 7.63031900e-01 7.85188556e-01
-1.67885387e+00 -4.89027977e-01 -6.24091029e-01 4.10438210e-01
-8.77064764e-01 -4.54694003e-01 4.11900491e-01 -8.05649579e-01
6.45337403e-01 7.51353681e-01 1.16849518e+00 5.81727386e-01
7.68737912e-01 4.42299545e-01 1.81131744e+00 -2.63097882e-01
3.69779795e-01 3.05452466e-01 4.42109108e-01 4.69712526e-01
3.22792798e-01 3.07735950e-01 -2.08916157e-01 -2.52398700e-01
9.88481104e-01 1.33888265e-02 -5.21527588e-01 -8.86476189e-02
-4.72462773e-01 6.26565576e-01 2.85464227e-01 3.41926754e-01
-5.06943047e-01 -9.22006741e-02 4.25346881e-01 5.77554941e-01
4.37980175e-01 7.35856533e-01 -4.05327022e-01 -2.54142344e-01
-1.02553165e+00 2.46882260e-01 1.00061226e+00 1.06255688e-01
2.40314111e-01 -3.00366096e-02 -7.83197165e-01 7.40144610e-01
4.26068783e-01 5.01419425e-01 5.36205292e-01 -1.13831556e+00
-5.51436804e-02 8.37835848e-01 3.61209586e-02 -1.00152779e+00
-7.70600379e-01 -2.68181115e-01 -9.83162344e-01 4.91189599e-01
6.81226730e-01 -2.06467718e-01 -1.04827988e+00 1.05883074e+00
9.54795182e-02 1.34979069e-01 -2.95602441e-01 9.61701572e-01
1.22854722e+00 1.75384521e-01 4.29986000e-01 -1.42112702e-01
1.31766284e+00 -3.69413495e-01 -7.01582253e-01 8.79416764e-02
2.54942864e-01 -3.22141498e-02 5.95946670e-01 2.79180110e-01
-1.00490272e+00 -6.76718056e-01 -7.25193679e-01 -9.11638513e-02
-4.18443114e-01 2.50311848e-02 6.35491431e-01 1.15017104e+00
-9.46977675e-01 9.69527245e-01 -6.87632203e-01 -3.04603100e-01
6.62684739e-01 9.26749930e-02 3.84123586e-02 1.86077148e-01
-1.46052861e+00 1.10913944e+00 -2.49278381e-01 4.98026401e-01
-2.08237782e-01 -1.40672350e+00 -6.61492407e-01 1.65703267e-01
-7.55172908e-01 -6.79474056e-01 7.19474196e-01 1.39654636e-01
-1.60861409e+00 1.18121588e+00 -5.15341898e-03 -6.39304101e-01
7.84647405e-01 -5.00671789e-02 -5.16320705e-01 4.29067373e-01
-2.30149806e-01 2.30923132e-03 6.26012981e-01 -4.58527654e-01
-3.38104963e-01 -9.88839090e-01 -6.86936438e-01 -2.04415575e-01
1.77702084e-01 -3.85876633e-02 3.83981228e-01 2.24149123e-01
2.68822581e-01 -4.64786738e-01 -5.32451689e-01 3.13044786e-01
-3.09572339e-01 1.49338013e-02 4.88038480e-01 -1.19897032e+00
9.97013450e-01 -1.63881993e+00 -3.62486660e-01 3.88651490e-01
8.35726619e-01 7.26728141e-01 4.30470735e-01 -7.79665634e-02
-1.18598556e-02 3.28107476e-01 -1.28588919e-03 2.37813562e-01
2.34362304e-01 9.52200815e-02 7.71275833e-02 5.42607486e-01
-2.95624197e-01 1.22414649e+00 -4.67429012e-01 -2.27187157e-01
1.02538288e+00 4.41459596e-01 1.73710063e-02 2.36092582e-01
2.99329430e-01 7.91653216e-01 -7.52429813e-02 5.70928991e-01
1.03351247e+00 1.08991437e-01 1.32998303e-01 -3.71768534e-01
-2.97069907e-01 7.05956668e-02 -5.19467533e-01 9.68493164e-01
-1.36512406e-02 6.08523607e-01 -6.50656223e-02 -9.70145345e-01
1.39188361e+00 5.38701892e-01 4.71980512e-01 -1.03559005e+00
3.19771945e-01 2.35665768e-01 4.40792769e-01 -1.09793234e+00
-6.29689157e-01 -4.10772532e-01 3.77942264e-01 -5.60743595e-03
-8.66507217e-02 -3.09801400e-02 4.23626363e-01 -3.78060132e-01
9.85114694e-01 -1.93697184e-01 3.45869482e-01 -5.74831009e-01
7.23302424e-01 -5.74813664e-01 4.81470466e-01 9.42120671e-01
-1.17267883e+00 4.50748086e-01 7.26839900e-01 -1.22078335e+00
-1.07121634e+00 -1.01754832e+00 -9.07975495e-01 2.05617398e-01
-4.09290761e-01 1.96814582e-01 -2.67775774e-01 7.79521689e-02
8.81924808e-01 2.16409534e-01 -8.79733443e-01 -2.81878766e-02
-2.98329264e-01 -1.18572044e+00 3.58236521e-01 5.67623734e-01
8.83521438e-01 -9.68834758e-01 -8.09324861e-01 4.33819979e-01
2.22286716e-01 -8.61158490e-01 3.33947778e-01 -2.49336794e-01
-1.10203159e+00 -1.24703300e+00 -1.09281278e+00 -3.55391800e-01
2.57863253e-01 -6.88426614e-01 9.92464423e-01 -4.89082962e-01
-1.21057665e+00 1.53806612e-01 2.76431859e-01 -4.14355785e-01
-3.77925307e-01 -4.01189685e-01 1.03689814e-02 1.18504539e-01
1.39114475e+00 -9.52410042e-01 -1.19885790e+00 -2.04313979e-01
2.17742354e-01 -1.86628029e-01 5.16711414e-01 -2.54784692e-02
4.03967440e-01 -5.32431722e-01 5.53688228e-01 -7.44571507e-01
8.57201636e-01 -1.48162013e-02 -5.77367485e-01 -5.54893017e-02
-7.03813314e-01 -6.50771737e-01 4.44387794e-01 9.97900292e-02
-4.66154486e-01 -6.00271583e-01 -1.50873750e-01 -1.73683967e-02
-5.41514814e-01 2.50623018e-01 2.24817097e-01 -4.42990422e-01
9.83841598e-01 4.70314026e-01 5.17359734e-01 -6.00289226e-01
-9.06866193e-02 8.42108667e-01 5.45885742e-01 -1.24608435e-01
1.28410250e-01 1.04691677e-01 5.14150083e-01 -1.16701698e+00
-7.48514414e-01 -3.45882624e-01 -7.01677859e-01 -4.23459530e-01
9.26123381e-01 -6.61828101e-01 -1.07675338e+00 7.87029505e-01
-2.53008693e-01 -3.05839539e-01 -3.02012742e-01 6.42197013e-01
-3.41277897e-01 4.82826173e-01 -1.12019849e+00 -8.27346325e-01
-9.58517909e-01 -4.96474564e-01 2.69215852e-01 5.97493589e-01
-2.28086576e-01 -1.15983665e+00 4.57558870e-01 4.01359886e-01
1.23959780e+00 9.11310792e-01 8.79254639e-01 -3.59737039e-01
-4.97150049e-02 -6.50808275e-01 -4.49081987e-01 7.27655947e-01
3.54896128e-01 -1.88670278e-01 -8.77926588e-01 -3.58698964e-02
3.21276605e-01 -2.65643988e-02 8.10463488e-01 1.16338241e+00
1.19233263e+00 -8.71897787e-02 7.99762085e-02 8.19446623e-01
1.44067442e+00 5.07813752e-01 1.22656190e+00 2.34334692e-01
4.64988470e-01 1.51335433e-01 -4.29936856e-01 6.27728462e-01
4.07373831e-02 4.22294401e-02 -2.03723475e-01 -5.38576484e-01
-2.87280500e-01 4.88315046e-01 -3.63181740e-01 2.16210306e-01
-4.08900857e-01 8.53208005e-01 -7.29106188e-01 1.97276426e-03
-1.36169648e+00 -1.03071880e+00 -5.02148986e-01 2.13025880e+00
1.13521004e+00 -2.07775086e-01 1.36762127e-01 3.11754625e-02
6.21879518e-01 -1.01957016e-01 -5.94973564e-01 -7.86352932e-01
1.44843176e-01 7.73127079e-01 3.70818257e-01 4.89844054e-01
-1.15100873e+00 7.33861178e-02 7.47345543e+00 -6.91104949e-01
-1.45675194e+00 -4.30321366e-01 8.39555264e-01 1.64205611e-01
6.45682037e-01 -4.30372924e-01 -7.74774671e-01 1.03422618e+00
1.16187406e+00 -4.74716097e-01 1.47067562e-01 5.66899538e-01
7.42624044e-01 -2.91416228e-01 -9.58022773e-01 9.48252618e-01
-1.24230616e-01 -1.20233285e+00 -5.49256921e-01 5.13465963e-02
2.20187992e-01 1.02935517e-02 -2.21484452e-01 3.67068172e-01
-3.97129238e-01 -8.05869222e-01 -6.61626697e-01 1.03975987e+00
6.53379619e-01 -4.68695909e-02 7.70932853e-01 -2.53717303e-01
-5.53806961e-01 -1.05036460e-01 -6.62731469e-01 -5.36421716e-01
5.74796461e-02 1.09246707e+00 -7.41305172e-01 -7.86587596e-02
6.16687775e-01 7.37849891e-01 -6.37122273e-01 1.78348136e+00
-1.69416770e-01 7.94152498e-01 -8.34476277e-02 1.38353378e-01
-3.50501031e-01 -4.47383046e-01 2.63912320e-01 1.17859876e+00
1.13073386e-01 1.38833538e-01 2.06829801e-01 1.51286852e+00
3.63522589e-01 2.50248648e-02 -2.61484832e-01 1.09678313e-01
-8.85229856e-02 1.28563106e+00 -5.75936846e-02 -6.48476481e-01
-3.35815847e-01 2.96937793e-01 -4.31813508e-01 3.45466107e-01
-3.00318688e-01 -6.14813387e-01 5.21638691e-01 3.35402101e-01
-3.42974514e-01 1.55210704e-01 -9.46545720e-01 -1.22408187e+00
-1.86447322e-01 -3.85058939e-01 4.25285667e-01 -3.91308397e-01
-1.28704023e+00 -6.67042807e-02 -3.75676364e-01 -6.66885495e-01
1.92117587e-01 -7.52982557e-01 -8.18218052e-01 1.67875147e+00
-1.80642307e+00 -5.63221097e-01 -8.77976894e-01 1.91248283e-01
-1.01064146e-01 -9.82686877e-02 1.18654227e+00 4.65544194e-01
-6.50570512e-01 1.07581615e-01 -3.44152033e-01 4.28772599e-01
5.62667370e-01 -1.97568977e+00 2.55447298e-01 6.10824585e-01
-1.09794784e+00 4.82081354e-01 7.01972902e-01 -5.68386197e-01
-1.03082848e+00 -9.50517714e-01 1.28107369e+00 -1.13116935e-01
1.96937099e-01 -1.73244596e-01 -7.56480992e-01 3.67372274e-01
-1.64859984e-02 3.67782474e-01 1.18046999e+00 -1.67660445e-01
3.48751135e-02 -4.64240968e-01 -1.67057467e+00 2.23989800e-01
1.93995029e-01 -2.20385820e-01 -8.93377066e-01 4.13948864e-01
-2.86249578e-01 -5.84611058e-01 -1.57618165e+00 6.06839120e-01
9.41944182e-01 -9.82229114e-01 8.12218726e-01 -7.36021042e-01
5.63015878e-01 9.79894549e-02 3.66351366e-01 -1.26178038e+00
-6.65084958e-01 -5.24902463e-01 -5.96345484e-01 6.79621339e-01
1.46033272e-01 -1.00935256e+00 1.07327819e+00 1.65172124e+00
-1.11341951e-02 -5.83639145e-01 -7.32931316e-01 -3.82182151e-01
1.85029805e-01 3.60322386e-01 1.96869567e-01 7.38472402e-01
4.88722533e-01 9.54334885e-02 -4.27419811e-01 -1.48625746e-01
1.10499179e+00 3.66920561e-01 6.43429518e-01 -1.60330009e+00
-2.12908741e-02 -4.20811146e-01 -1.05416286e+00 -4.11408037e-01
-5.56037426e-01 -7.27227628e-01 -4.14646685e-01 -1.96533048e+00
6.41748384e-02 -9.36905369e-02 -5.45994818e-01 5.13940752e-01
-4.58580442e-02 4.26104814e-01 -1.27349392e-01 -1.72083572e-01
3.41739833e-01 -1.47873545e-02 1.39163816e+00 4.59418911e-03
-7.46320724e-01 4.34257120e-01 -5.92997015e-01 5.06339744e-02
1.24644649e+00 9.14369449e-02 9.02642608e-02 4.62830335e-01
-3.77256751e-01 6.85152486e-02 2.82061756e-01 -1.26837349e+00
-1.14223458e-01 5.61427958e-02 1.39997649e+00 -4.16600943e-01
1.82123467e-01 -5.16253531e-01 -2.11175427e-01 1.15525615e+00
-5.06367683e-01 -8.76017869e-01 1.91435754e-01 -1.08468190e-01
6.61823526e-02 1.74657166e-01 1.24798298e+00 -9.33335602e-01
-5.73270619e-01 4.76217508e-01 -6.24459326e-01 -8.49736705e-02
6.82812333e-01 -1.79986328e-01 -5.22497416e-01 -9.96703357e-02
-1.57358122e+00 3.04199278e-01 -3.03462625e-01 -1.79745909e-02
5.12232780e-01 -1.12768722e+00 -1.02443337e+00 6.41891956e-01
-1.10834539e-01 -6.49494350e-01 4.90293026e-01 1.32124686e+00
-9.13898706e-01 5.68349838e-01 -6.99754536e-01 -8.11174273e-01
-1.22672641e+00 8.82235840e-02 1.34215224e+00 9.99466181e-02
-1.21295965e+00 5.64300776e-01 -3.65198493e-01 -2.03160822e-01
3.31190675e-01 -3.90054494e-01 -5.11204422e-01 -3.01165611e-01
1.17463493e+00 6.42854333e-01 -5.26300147e-02 1.94505423e-01
3.27342972e-02 4.77331847e-01 3.00377309e-02 5.00464380e-01
1.24809301e+00 -2.15963110e-01 -5.53452969e-01 4.36815053e-01
1.04018927e+00 -6.88800395e-01 -5.92191815e-01 -1.61587521e-01
-3.01125973e-01 -8.28842998e-01 1.81367874e-01 -1.31968248e+00
-1.18089890e+00 1.06214273e+00 1.48360789e+00 5.40892959e-01
9.22354817e-01 -4.92955178e-01 6.70379758e-01 -6.36425661e-03
-9.92163420e-02 -9.81748998e-01 -5.87810993e-01 -6.23671599e-02
8.06003153e-01 -8.95295322e-01 2.48382643e-01 -2.95279771e-01
-2.61509866e-01 1.45241380e+00 7.37879395e-01 -5.40564358e-01
9.23823476e-01 8.84019956e-02 7.21989870e-01 -3.72030884e-01
-2.79235542e-01 8.11344311e-02 1.14370264e-01 7.55865335e-01
7.21260369e-01 3.73787552e-01 -1.04594493e+00 1.21977016e-01
2.37777047e-02 7.81758428e-01 5.81276774e-01 6.15099132e-01
-8.79052758e-01 -6.70766771e-01 1.54382095e-01 1.15918732e+00
-7.68921375e-01 5.32707088e-02 -2.22295374e-01 4.14274782e-01
-1.81853920e-01 6.55891955e-01 -2.72209272e-02 7.27161989e-02
6.31218970e-01 6.62723780e-01 5.87823629e-01 -4.26040828e-01
-4.81668591e-01 2.61360016e-02 -1.66646928e-01 -6.52108967e-01
-2.14656398e-01 -2.83251047e-01 -6.77605569e-01 -1.38629034e-01
2.70380884e-01 -1.04176417e-01 6.17140412e-01 7.01080680e-01
4.86463845e-01 2.70818621e-01 8.07848632e-01 -6.26543999e-01
-8.46679688e-01 -1.31601954e+00 -1.40625656e+00 5.71066260e-01
5.01934469e-01 -2.49712497e-01 -7.07961738e-01 -2.02154189e-01] | [14.078546524047852, 2.988685131072998] |
73052bb8-a9d4-4bbe-b4b8-35379e662d4f | scimrc-multi-perspective-scientific-machine | 2306.14149 | null | https://arxiv.org/abs/2306.14149v1 | https://arxiv.org/pdf/2306.14149v1.pdf | SciMRC: Multi-perspective Scientific Machine Reading Comprehension | Scientific machine reading comprehension (SMRC) aims to understand scientific texts through interactions with humans by given questions. As far as we know, there is only one dataset focused on exploring full-text scientific machine reading comprehension. However, the dataset has ignored the fact that different readers may have different levels of understanding of the text, and only includes single-perspective question-answer pairs, leading to a lack of consideration of different perspectives. To tackle the above problem, we propose a novel multi-perspective SMRC dataset, called SciMRC, which includes perspectives from beginners, students and experts. Our proposed SciMRC is constructed from 741 scientific papers and 6,057 question-answer pairs. Each perspective of beginners, students and experts contains 3,306, 1,800 and 951 QA pairs, respectively. The extensive experiments on SciMRC by utilizing pre-trained models suggest the importance of considering perspectives of SMRC, and demonstrate its challenging nature for machine comprehension. | ['Xian-Ling Mao', 'Heyan Huang', 'Yuxiang Nie', 'Heqi Zheng', 'Xiao Zhang'] | 2023-06-25 | null | null | null | null | ['reading-comprehension', 'machine-reading-comprehension'] | ['natural-language-processing', 'natural-language-processing'] | [ 2.97981650e-01 2.23887384e-01 1.60036072e-01 -4.20098931e-01
-9.85171080e-01 -8.44804108e-01 6.68264747e-01 5.80833793e-01
-2.10700601e-01 5.64399660e-01 4.87779081e-01 -8.00309360e-01
-3.58940005e-01 -8.19837868e-01 -7.98437357e-01 -2.35752970e-01
6.14768386e-01 4.03473139e-01 1.56102851e-01 -3.44406962e-01
1.00223482e+00 -2.12078437e-01 -1.40188646e+00 4.86488461e-01
1.76518416e+00 3.85380507e-01 5.47646940e-01 8.39485586e-01
-5.02990246e-01 1.07364666e+00 -1.01454842e+00 -6.19156599e-01
-4.56590712e-01 -8.30374002e-01 -1.32918143e+00 -5.93575418e-01
7.14883804e-01 -1.11145861e-01 -6.87957779e-02 1.06439126e+00
4.02287006e-01 1.21646918e-01 6.52767181e-01 -1.05373394e+00
-1.07684469e+00 6.84658349e-01 -6.27307177e-01 3.21746111e-01
1.05121648e+00 -1.27697363e-01 9.30249155e-01 -7.74353087e-01
4.71673787e-01 1.50824845e+00 1.86928391e-01 7.28458405e-01
-6.03404999e-01 -5.53822219e-01 1.64143771e-01 7.39686131e-01
-7.71301568e-01 -1.03139944e-01 5.78282416e-01 -3.66978645e-01
3.24071616e-01 3.94904166e-01 3.03906709e-01 1.19176698e+00
3.04534137e-01 7.08650470e-01 1.64307249e+00 -4.55497921e-01
2.39348382e-01 5.69066666e-02 9.61853802e-01 1.50579050e-01
3.57987523e-01 -5.98269820e-01 -6.55874729e-01 -9.86400172e-02
4.57287937e-01 -3.94902498e-01 -8.20764720e-01 1.11884497e-01
-1.65960789e+00 4.94010508e-01 -2.31710263e-02 2.38676563e-01
-3.59353237e-02 -5.28119504e-01 2.29181394e-01 4.55153346e-01
4.06524427e-02 7.99782753e-01 -5.41782022e-01 -3.63009721e-01
-4.03333426e-01 2.47312650e-01 1.18798923e+00 1.25159812e+00
3.41707557e-01 -7.03463435e-01 -3.08457911e-01 5.95108390e-01
2.17517987e-01 8.32873166e-01 2.40053803e-01 -1.18875754e+00
7.70666659e-01 8.78306329e-01 1.86823934e-01 -1.21130586e+00
-2.03164414e-01 -6.00724638e-01 -7.33555496e-01 -4.30013597e-01
5.70563972e-01 -6.16316982e-02 -4.60559785e-01 1.52621877e+00
7.61684403e-02 -6.71845451e-02 4.38000679e-01 8.96035075e-01
1.64329362e+00 8.90775919e-01 2.51928806e-01 -1.26334265e-01
1.92661309e+00 -1.17812884e+00 -9.22725499e-01 -4.60997745e-02
3.74453276e-01 -9.09805357e-01 1.27818418e+00 7.85562932e-01
-1.27197051e+00 -6.95790350e-01 -8.52976024e-01 -6.51597619e-01
-3.39496851e-01 1.12350984e-02 1.65321305e-01 3.22403878e-01
-7.95278728e-01 5.24180941e-02 -1.21498339e-01 -2.75607437e-01
1.76526114e-01 -2.74728388e-01 -1.76122159e-01 -5.35445333e-01
-1.37296176e+00 1.02385986e+00 1.73717245e-01 -2.24277601e-02
-9.31478202e-01 -1.12071061e+00 -4.43346381e-01 2.68444210e-01
5.29301882e-01 -9.31599855e-01 1.30960155e+00 -5.26302338e-01
-1.45994043e+00 7.73208976e-01 -5.21456718e-01 2.37400383e-01
4.41269010e-01 -4.65059489e-01 -4.56841469e-01 5.74502289e-01
1.85277313e-01 4.22183156e-01 2.26664707e-01 -1.37690318e+00
-3.59295726e-01 -6.03897572e-01 5.72213292e-01 4.74344462e-01
4.08016294e-02 -4.71830228e-03 -1.93528906e-01 -3.79002482e-01
2.02559128e-01 -2.70367175e-01 1.40148088e-01 -3.58078510e-01
-2.36770138e-01 -6.82027996e-01 5.46000838e-01 -9.69205141e-01
1.04407990e+00 -1.54042888e+00 3.58178645e-01 -3.12321246e-01
5.91980338e-01 1.48301825e-01 -3.51352185e-01 8.67091358e-01
1.57727804e-02 3.23945075e-01 -2.17359841e-01 4.47711721e-02
2.48018350e-03 -1.39213979e-01 -4.77333993e-01 -5.86579591e-02
-1.52573913e-01 1.00802064e+00 -1.34059525e+00 -5.21921039e-01
-1.69791952e-01 1.72338393e-02 5.53967850e-03 5.96342444e-01
-5.67820728e-01 7.40705013e-01 -9.12637234e-01 6.38683558e-01
1.07349384e+00 -5.57182789e-01 -2.60173351e-01 1.71922535e-01
-8.76928568e-02 4.21217412e-01 -5.57132185e-01 1.97238207e+00
-3.11248869e-01 6.48209929e-01 -5.42379059e-02 -8.87323141e-01
9.07744765e-01 3.13733071e-01 -1.95681334e-01 -8.85011017e-01
5.81542887e-02 1.84506401e-01 1.08659029e-01 -1.02665770e+00
3.76264691e-01 8.54704753e-02 1.09333463e-01 8.24574709e-01
-4.61362824e-02 3.99736278e-02 2.56460518e-01 7.23734796e-01
1.01433945e+00 -4.62030470e-02 -1.20107234e-02 -6.08550370e-01
1.02087593e+00 1.13060996e-01 2.18310058e-01 1.03464580e+00
-3.05278063e-01 5.33066690e-01 8.44408810e-01 -4.06061336e-02
-4.59168404e-01 -1.05643725e+00 -6.81797862e-02 1.18225455e+00
5.69168925e-01 -4.02461708e-01 -1.00008702e+00 -5.84674954e-01
-4.08024013e-01 1.23439753e+00 -5.10223985e-01 7.69634321e-02
-5.70939481e-01 -2.52862424e-01 4.09203231e-01 2.22302914e-01
9.20001328e-01 -9.78616357e-01 -6.01628959e-01 -4.45053838e-02
-8.48540783e-01 -1.14809418e+00 -2.45564565e-01 -5.10827065e-01
-9.98172164e-01 -1.50081384e+00 -7.20025003e-01 -9.16428864e-01
7.41246223e-01 7.47634292e-01 1.56245184e+00 2.94463068e-01
2.51015455e-01 6.08694732e-01 -7.03421891e-01 -6.59235537e-01
-5.27589262e-01 1.31159410e-01 -5.16094863e-01 -6.28024459e-01
7.32332528e-01 -3.51280332e-01 -7.88700104e-01 3.89697552e-01
-7.24021614e-01 4.84257638e-01 5.93355119e-01 4.18073028e-01
3.62667665e-02 -5.47757208e-01 1.10457659e+00 -9.63825464e-01
8.77433419e-01 -7.43965507e-01 -2.13385776e-01 1.00583160e+00
-3.67460668e-01 -2.45427251e-01 8.61950874e-01 -1.73161447e-01
-1.46375358e+00 -1.02453363e+00 -4.32820134e-02 2.48360947e-01
-4.62983936e-01 7.21881330e-01 -3.36748093e-01 2.05980942e-01
5.61324716e-01 3.55307490e-01 -2.33171675e-02 -6.45787895e-01
4.28214252e-01 8.69795978e-01 6.91471279e-01 -9.36342716e-01
5.47960043e-01 -7.69764185e-02 -2.44504049e-01 -9.18837786e-01
-1.23081875e+00 -5.38707614e-01 -4.62246180e-01 -3.04059356e-01
8.69451344e-01 -9.10328627e-01 -1.12563372e+00 6.33872151e-01
-1.54976332e+00 2.03565583e-01 2.67475188e-01 3.08515400e-01
-2.57966608e-01 6.34389639e-01 -4.96049672e-01 -5.50727367e-01
-5.33701062e-01 -9.40033853e-01 7.97105968e-01 7.88037837e-01
-1.00702256e-01 -1.09037447e+00 -1.60563216e-01 1.32093978e+00
4.89471942e-01 4.43285033e-02 1.60275304e+00 -8.02882671e-01
-6.47356987e-01 2.63985574e-01 -2.86986262e-01 5.51706925e-02
-7.14772791e-02 -2.52042621e-01 -8.19766402e-01 -1.97483242e-01
5.10548711e-01 -5.40145099e-01 5.49967408e-01 -2.01945286e-02
1.47936523e+00 -5.65994121e-02 -6.68082014e-02 -3.39460373e-02
1.12074268e+00 2.42755413e-01 8.52863312e-01 1.85558379e-01
5.85003555e-01 1.16989839e+00 6.20840311e-01 -7.68259317e-02
1.02085674e+00 -1.15521818e-01 2.50721157e-01 2.42368385e-01
2.02636570e-01 -4.33924764e-01 1.11521542e-01 1.64089251e+00
-4.92939651e-02 -5.21427810e-01 -1.04154909e+00 5.19595742e-01
-1.48511386e+00 -8.13254654e-01 -9.06371891e-01 1.90150356e+00
7.82495081e-01 -2.68984854e-01 -5.12320220e-01 -2.09476054e-01
6.99322641e-01 1.83250308e-01 -5.58097661e-01 -4.25458580e-01
-2.07125336e-01 1.60381585e-01 -2.95732886e-01 4.08338457e-01
-4.20368344e-01 5.72536170e-01 6.06553030e+00 8.38136971e-01
-3.48087877e-01 -7.15381205e-02 6.46776795e-01 5.84664762e-01
-8.69516730e-01 3.11349422e-01 -4.64964062e-01 3.30059111e-01
7.75128245e-01 -4.82320279e-01 6.83456808e-02 2.60847330e-01
2.79273570e-01 -5.93066812e-01 -1.28555608e+00 7.79618442e-01
5.25184155e-01 -9.61381912e-01 5.12158811e-01 -4.30296540e-01
6.73562229e-01 -6.24295175e-01 -2.42156252e-01 4.48715925e-01
9.38412622e-02 -1.38137650e+00 1.55161887e-01 8.60296369e-01
1.78578049e-01 -5.75996518e-01 7.77549267e-01 1.00519884e+00
-4.82842594e-01 1.89678729e-01 -3.80344093e-01 -2.57110715e-01
-5.62417787e-03 4.05988723e-01 -2.12717563e-01 1.02687597e+00
6.56417549e-01 6.01784945e-01 -8.86268497e-01 7.35321760e-01
-3.91924560e-01 8.53023827e-01 2.81342030e-01 -5.22779942e-01
-1.92313701e-01 -4.41433251e-01 3.97921383e-01 8.20747793e-01
2.59374470e-01 8.37349594e-01 -3.08463216e-01 1.13902056e+00
-1.43422216e-01 2.92996317e-01 -5.22164740e-02 -7.86173567e-02
6.17397368e-01 1.10471976e+00 -3.44781846e-01 -4.77961808e-01
-4.73430991e-01 7.27450490e-01 3.03346485e-01 5.28862059e-01
-4.41710740e-01 -6.01241052e-01 1.04783103e-01 -2.22992390e-01
-4.42163259e-01 -1.04600936e-01 -6.60937846e-01 -1.51014161e+00
1.00595310e-01 -1.45220482e+00 2.84819752e-01 -1.23477495e+00
-1.47744679e+00 1.16282783e-01 1.65515050e-01 -7.50278413e-01
3.59036207e-01 -5.78102291e-01 -9.98872280e-01 1.18590033e+00
-1.54678178e+00 -9.44664657e-01 -8.60500991e-01 1.64292768e-01
9.75471854e-01 8.88001323e-02 6.51033938e-01 -1.29377663e-01
-4.41265464e-01 2.36233354e-01 8.78229830e-03 -1.70280904e-01
9.66549814e-01 -1.43353546e+00 3.70930046e-01 4.47227776e-01
-3.62437427e-01 1.26514697e+00 8.61237824e-01 -5.35076201e-01
-1.88671815e+00 -3.81677777e-01 1.37956703e+00 -9.83531535e-01
4.41331476e-01 -2.43743226e-01 -1.64072096e+00 4.16205317e-01
9.85847056e-01 -9.64278996e-01 1.01988161e+00 1.08574383e-01
-2.53452539e-01 2.84242779e-01 -8.46457958e-01 7.69714177e-01
1.03861618e+00 -3.91723931e-01 -1.49677646e+00 3.89955103e-01
8.43265951e-01 -5.07940233e-01 -1.12173772e+00 4.52507824e-01
3.40151697e-01 -9.02763486e-01 6.98097587e-01 -7.56179273e-01
1.24445021e+00 -2.34048590e-01 2.56299168e-01 -1.21006656e+00
7.62225874e-03 -1.25156879e-01 -9.96954441e-02 1.28629088e+00
2.62732506e-01 -5.73379278e-01 4.15488929e-01 6.13154173e-01
-1.77589461e-01 -7.99928665e-01 -6.33047640e-01 -1.74530610e-01
8.03758800e-01 -5.88646345e-03 6.48442507e-01 1.13851547e+00
3.12059015e-01 6.28988266e-01 2.18075439e-02 1.22474380e-01
5.06477714e-01 6.41045332e-01 7.12651551e-01 -1.38702488e+00
-1.24069482e-01 -3.39484453e-01 2.93630958e-01 -1.76955175e+00
3.59322950e-02 -6.59582675e-01 -1.85168013e-01 -2.07950497e+00
6.50553048e-01 -2.85280701e-02 2.86031421e-02 -1.77202255e-01
-8.74514461e-01 -5.42290509e-01 -8.22857618e-02 2.42166877e-01
-9.54201221e-01 6.71220243e-01 2.09212971e+00 9.40918922e-02
4.47793037e-01 -2.78532445e-01 -1.27192938e+00 7.24370003e-01
6.92456186e-01 -4.73087728e-02 -7.37215579e-01 -8.95336807e-01
4.36364323e-01 4.86706287e-01 3.59339654e-01 -5.33544779e-01
6.75296783e-01 -4.10581887e-01 2.33986989e-01 -8.49955440e-01
-1.78132862e-01 -2.24064827e-01 -3.21145624e-01 2.16165125e-01
-7.94295013e-01 3.26644361e-01 -2.00208977e-01 6.42418087e-01
-3.58967304e-01 -6.20811105e-01 2.28349164e-01 -5.12490034e-01
-4.63055342e-01 -2.42189914e-01 -4.31052357e-01 7.75121689e-01
8.43582451e-01 3.10471207e-02 -1.20204365e+00 -5.50277352e-01
-1.03564523e-01 1.00043368e+00 2.25843385e-01 7.23892927e-01
6.86478972e-01 -7.95170307e-01 -1.09599924e+00 -3.46550703e-01
2.54159898e-01 4.53467757e-01 7.29112864e-01 5.84702134e-01
-8.19361448e-01 6.54085994e-01 -1.27709493e-01 -4.96819496e-01
-1.17491615e+00 4.15913731e-01 5.64676970e-02 -4.50439379e-03
-3.09917152e-01 7.32943177e-01 4.05605763e-01 -6.76660001e-01
1.37044013e-01 -1.80829555e-01 -9.75075424e-01 8.23542252e-02
7.36717880e-01 8.09602499e-01 -1.65429965e-01 -3.60331774e-01
-1.01838283e-04 9.20182109e-01 -1.98283702e-01 1.81413051e-02
7.55748570e-01 -4.06491160e-01 -6.23189926e-01 6.11743212e-01
9.06677783e-01 1.23206966e-01 -4.39706802e-01 -1.80587083e-01
1.42473489e-01 -4.26403940e-01 -5.41057110e-01 -1.27508557e+00
-2.97257572e-01 1.32292318e+00 6.01845153e-04 1.46144452e-02
9.87127244e-01 -2.16699634e-02 9.44865167e-01 5.80361426e-01
4.36719924e-01 -8.98257434e-01 2.84452230e-01 8.88627887e-01
1.13863611e+00 -1.39753699e+00 -1.37289584e-01 -7.15029180e-01
-4.62889999e-01 1.34172142e+00 1.12826514e+00 3.17216128e-01
8.24369714e-02 -5.50914466e-01 4.12479877e-01 -2.86742240e-01
-1.05326903e+00 4.57961351e-01 2.51030564e-01 4.06454265e-01
7.21152425e-01 -1.42214909e-01 -9.13854003e-01 8.74840438e-01
-5.50957561e-01 -1.26748011e-01 9.79745150e-01 9.79633868e-01
-6.49793088e-01 -1.05630791e+00 -5.15410364e-01 3.09572160e-01
-5.06729424e-01 -4.63799313e-02 -5.15396893e-01 4.46984977e-01
-1.43141538e-01 1.61066651e+00 -2.69724429e-01 1.80597976e-02
3.86170119e-01 -2.25000409e-03 4.66565013e-01 -4.50674802e-01
-7.88843334e-01 -6.39045119e-01 -1.11153126e-01 2.96520535e-02
-3.92121315e-01 -2.66530871e-01 -1.11182952e+00 -4.74269569e-01
-2.29054436e-01 7.41639256e-01 4.87464339e-01 1.12517548e+00
4.48409349e-01 5.05186975e-01 3.03351730e-01 1.58217192e-01
-8.21412504e-01 -1.17858291e+00 -1.32037029e-01 4.53733832e-01
2.67779112e-01 -2.45458558e-01 -4.74808276e-01 -1.36946207e-02] | [11.33234691619873, 8.158875465393066] |
849bf5ca-1fbd-47c4-979b-45e43983b7bf | domain-adaptation-of-thai-word-segmentation | null | null | https://aclanthology.org/2020.emnlp-main.315 | https://aclanthology.org/2020.emnlp-main.315.pdf | Domain Adaptation of Thai Word Segmentation Models using Stacked Ensemble | Like many Natural Language Processing tasks, Thai word segmentation is domain-dependent. Researchers have been relying on transfer learning to adapt an existing model to a new domain. However, this approach is inapplicable to cases where we can interact with only input and output layers of the models, also known as {``}black boxes{''}. We propose a filter-and-refine solution based on the stacked-ensemble learning paradigm to address this black-box limitation. We conducted extensive experimental studies comparing our method against state-of-the-art models and transfer learning. Experimental results show that our proposed solution is an effective domain adaptation method and has a similar performance as the transfer learning method. | ['Sarana Nutanong', 'Ekapol Chuangsuwanich', 'Raheem Sarwar', 'Wannaphong Phatthiyaphaibun', 'Peerat Limkonchotiwat'] | null | null | null | null | emnlp-2020-11 | ['thai-word-tokenization'] | ['natural-language-processing'] | [ 2.68880218e-01 -2.10858118e-02 -3.61362472e-02 -5.54863155e-01
-7.88883686e-01 -5.58305860e-01 3.25719118e-01 -1.20935798e-01
-8.21070492e-01 9.79668915e-01 -1.14503391e-01 -6.17350817e-01
1.01094030e-01 -8.03994060e-01 -8.31695676e-01 -5.11376679e-01
3.36024940e-01 5.97522318e-01 7.41838455e-01 -1.66600883e-01
2.77211756e-01 5.40139265e-02 -1.01449370e+00 4.66152072e-01
1.32085967e+00 5.21132350e-01 4.84418213e-01 3.22234988e-01
-6.48477435e-01 3.29371899e-01 -5.32963276e-01 -5.19420385e-01
9.89581272e-02 -4.30780858e-01 -8.97650242e-01 -8.69950950e-02
1.10345975e-01 -1.24928430e-01 6.05692491e-02 9.99291778e-01
4.28651989e-01 1.78606361e-01 6.98059082e-01 -7.25074828e-01
-9.10428226e-01 5.78319073e-01 -6.25992239e-01 1.84300721e-01
1.95011765e-01 -1.86039671e-01 5.44642389e-01 -9.86270070e-01
5.36716819e-01 1.12892985e+00 6.38849437e-01 8.61664474e-01
-1.23255670e+00 -9.32384551e-01 6.86869681e-01 3.01495254e-01
-9.13933158e-01 -1.45509645e-01 7.63608575e-01 -3.16973776e-01
1.07296133e+00 -1.95624813e-01 3.80041391e-01 1.14345777e+00
4.01732586e-02 1.15948105e+00 1.37283301e+00 -7.49950945e-01
2.49370173e-01 3.36689651e-01 2.76239932e-01 4.22295600e-01
2.15070382e-01 5.40533960e-02 -4.11296129e-01 6.48598671e-02
8.18869829e-01 -2.65360564e-01 6.08223677e-02 -3.41466665e-01
-1.08076668e+00 8.37494493e-01 4.49817359e-01 4.48731631e-01
-2.85992444e-01 -2.92792141e-01 2.95804441e-01 2.79065073e-01
7.29443789e-01 3.75316054e-01 -9.01941895e-01 -1.77304726e-02
-7.62210131e-01 5.98386824e-02 7.25240171e-01 8.53367090e-01
9.08620775e-01 5.12847640e-02 1.48081556e-01 9.99866843e-01
2.27276042e-01 1.64068744e-01 6.32326543e-01 -4.32194293e-01
7.25786328e-01 6.80123627e-01 1.60566077e-03 -2.82280535e-01
-2.67064929e-01 -2.45708466e-01 -4.73028064e-01 2.10748628e-01
6.10593617e-01 -6.24434948e-01 -1.33610797e+00 1.52128553e+00
3.16659421e-01 2.68456250e-01 2.30945155e-01 5.20552695e-01
6.94836259e-01 8.56898725e-01 5.09496391e-01 -1.86591387e-01
1.11600459e+00 -1.16724122e+00 -7.71852553e-01 -6.58699930e-01
4.48772609e-01 -7.86936700e-01 1.17831719e+00 5.10556281e-01
-8.72614801e-01 -6.66339338e-01 -9.43247676e-01 1.32894024e-01
-5.62287867e-01 -1.34759277e-01 5.31110585e-01 9.14478242e-01
-7.16996610e-01 5.48519969e-01 -9.81839180e-01 -6.01658285e-01
4.26459402e-01 5.59254706e-01 -3.59029442e-01 -1.30442772e-02
-1.09287691e+00 1.21711063e+00 8.85712922e-01 1.17688375e-02
-6.77157521e-01 -5.23512065e-01 -7.96205640e-01 -1.34835750e-01
4.51777637e-01 -5.60136497e-01 1.51222646e+00 -1.35791731e+00
-1.89539683e+00 5.16037583e-01 -4.80179906e-01 -3.67668837e-01
2.74941832e-01 -5.88187575e-01 -5.01626194e-01 -1.62301093e-01
-1.38598591e-01 5.66287041e-01 6.40912950e-01 -1.53248835e+00
-8.45671773e-01 -3.57664436e-01 -1.51090696e-01 2.43417144e-01
-5.81721425e-01 2.46623516e-01 -4.86293703e-01 -7.36005187e-01
2.20663875e-01 -7.42407799e-01 -4.08673465e-01 -4.15480405e-01
2.47250567e-03 -3.74753326e-01 1.06152427e+00 -8.63496840e-01
1.47224844e+00 -1.91355264e+00 5.84882572e-02 5.45968823e-02
-3.18683654e-01 8.12329173e-01 -1.14830270e-01 5.36507308e-01
9.51833725e-02 2.00852528e-01 -6.24518096e-01 -1.65365413e-01
-2.52954125e-01 3.74415100e-01 -5.01618795e-02 2.49393806e-02
3.18242967e-01 7.67562211e-01 -9.22158837e-01 -6.03262067e-01
1.74700186e-01 2.39787817e-01 -5.29217780e-01 3.08927089e-01
-3.54869664e-01 6.15372777e-01 -5.41009843e-01 4.00602877e-01
6.88290477e-01 5.02281450e-02 3.83294791e-01 1.63761675e-01
-1.61160186e-01 3.90175402e-01 -1.13910341e+00 1.84336782e+00
-4.63975489e-01 2.96362936e-01 -9.53028202e-02 -1.22837031e+00
1.20875812e+00 3.72927994e-01 1.06783751e-02 -6.19045615e-01
1.08534798e-01 3.67385805e-01 3.83482188e-01 -6.20043993e-01
2.40633562e-01 -3.51886958e-01 -9.31840539e-02 3.81604522e-01
2.78807342e-01 5.51517978e-02 8.35885555e-02 -1.48942664e-01
7.66509712e-01 7.99667656e-01 3.54819059e-01 -2.52740175e-01
4.28702950e-01 3.94874252e-02 9.26379025e-01 5.73872149e-01
-3.19585383e-01 4.17067051e-01 1.79443359e-01 -2.93955147e-01
-8.64828169e-01 -1.12653780e+00 2.43327674e-02 1.51510465e+00
7.12949485e-02 -1.35094509e-01 -1.00949848e+00 -1.00250995e+00
-2.41021961e-01 9.12745416e-01 -4.85570461e-01 -5.61337098e-02
-8.93019676e-01 -7.12115288e-01 3.63120914e-01 9.42373395e-01
8.64726901e-01 -1.35356128e+00 -4.16001409e-01 4.02865022e-01
-1.09069213e-01 -9.78983223e-01 -2.58295923e-01 3.75721097e-01
-1.22843540e+00 -6.50280535e-01 -7.59085238e-01 -1.36859477e+00
6.70241892e-01 -3.77208851e-02 1.00293338e+00 -2.63188452e-01
5.20465136e-01 4.38078865e-02 -6.49464190e-01 -7.03290284e-01
-3.37895989e-01 4.34856296e-01 -1.57329157e-01 1.34517744e-01
8.20064187e-01 -3.90094876e-01 -4.32207853e-01 3.15678149e-01
-8.15660059e-01 2.69887839e-02 7.73882806e-01 8.57709944e-01
4.46812928e-01 -3.48229893e-02 8.97932172e-01 -1.43588912e+00
8.33783865e-01 -4.05568957e-01 -4.69364733e-01 4.69428629e-01
-6.52880669e-01 9.13025215e-02 5.01696169e-01 -5.80405951e-01
-1.70560312e+00 4.84676987e-01 -1.98787123e-01 5.18175997e-02
-5.10542274e-01 6.52122796e-01 -5.41505873e-01 7.23045170e-02
5.16700149e-01 1.29560128e-01 -3.18234980e-01 -7.98556924e-01
2.76200473e-01 9.63332832e-01 4.47553396e-01 -7.62937725e-01
7.14923799e-01 1.03973120e-01 -7.41415918e-01 -6.24711871e-01
-5.34705341e-01 -4.21315134e-01 -1.14703131e+00 -4.16877605e-02
9.67075348e-01 -7.98940063e-01 -9.84252840e-02 6.50171101e-01
-1.30782700e+00 -5.04265428e-01 1.45798638e-01 4.91090685e-01
-3.50652963e-01 3.84984046e-01 -5.42654574e-01 -7.44076431e-01
-2.02263892e-01 -9.96171653e-01 6.54315412e-01 5.68083525e-01
-2.18573228e-01 -1.28710067e+00 1.98107138e-01 3.05437207e-01
3.67484629e-01 -1.09193951e-01 9.86033082e-01 -1.09811151e+00
-9.85751152e-02 -6.91010132e-02 -5.01804948e-02 4.89017040e-01
2.07549915e-01 -1.49690881e-01 -1.11161602e+00 -1.56601906e-01
-4.43962924e-02 -1.79229975e-01 9.15035844e-01 2.67954141e-01
1.09107721e+00 5.29147545e-03 -5.97607732e-01 2.58546829e-01
1.29945993e+00 7.26982296e-01 7.94359922e-01 5.95062554e-01
5.13664603e-01 5.73690891e-01 6.69403374e-01 -1.10097736e-01
5.06989777e-01 4.66743946e-01 -3.46238539e-02 -2.56011367e-01
1.07834153e-01 -4.43989724e-01 4.08170730e-01 9.36074853e-01
-1.31737053e-01 -2.35022053e-01 -1.30820060e+00 8.15168858e-01
-2.07988477e+00 -6.48830175e-01 1.38233826e-01 2.12260461e+00
9.99428511e-01 3.48142713e-01 1.31843716e-01 -1.66203212e-02
9.07553315e-01 -2.84941137e-01 -4.35467392e-01 -7.91372895e-01
1.32277161e-01 6.97458625e-01 1.83573768e-01 4.20415193e-01
-1.30970240e+00 1.46189058e+00 6.73958826e+00 6.74221039e-01
-1.06341720e+00 2.49347717e-01 2.98686087e-01 4.00128603e-01
-2.32954443e-01 9.37381089e-02 -7.22565472e-01 3.97638708e-01
1.03035474e+00 -1.96183007e-02 1.79998502e-01 5.92121363e-01
-1.01876132e-01 -1.53550401e-01 -1.04152596e+00 4.33461666e-01
4.70415130e-02 -9.38705862e-01 8.47336799e-02 -1.42976940e-01
8.85031223e-01 -6.13823086e-02 -9.20576453e-02 5.99246979e-01
4.77026016e-01 -8.36625755e-01 2.74899721e-01 1.58349037e-01
4.95576233e-01 -7.25577593e-01 6.86730623e-01 5.75496137e-01
-9.42415893e-01 -2.12155804e-02 -2.56451637e-01 -1.46437958e-01
1.94491819e-01 1.09877937e-01 -9.47479248e-01 5.71283340e-01
8.27289820e-01 4.44883019e-01 -3.91399354e-01 1.16878915e+00
-5.17763913e-01 9.67090070e-01 -3.91866080e-02 -1.47616893e-01
2.95843154e-01 -2.83845484e-01 1.65537387e-01 1.45613968e+00
2.96250045e-01 2.59934098e-01 2.44196162e-01 5.52117646e-01
6.04174435e-02 2.34917268e-01 -6.74937010e-01 -1.11734672e-02
3.03039461e-01 9.01631832e-01 -8.13959002e-01 -4.30483788e-01
-7.78801739e-01 1.00249279e+00 4.56991673e-01 5.82895398e-01
-5.99629641e-01 -5.20972848e-01 2.90854990e-01 -4.79347864e-03
7.02422440e-01 -3.78022879e-01 -6.03517830e-01 -1.08273506e+00
-9.33471918e-02 -8.85318458e-01 3.91390353e-01 -5.08154154e-01
-1.45589447e+00 6.19241297e-01 1.85114145e-01 -1.17024601e+00
-1.21996313e-01 -9.03406501e-01 -7.50539899e-01 8.75759661e-01
-1.49446869e+00 -1.33375800e+00 1.52955353e-01 7.85290360e-01
7.76773512e-01 -3.78738970e-01 1.04722583e+00 2.29691356e-01
-5.22098243e-01 5.95574737e-01 1.86976448e-01 4.28720951e-01
8.41521680e-01 -1.33710206e+00 4.70245212e-01 1.00821960e+00
2.23954201e-01 6.58533216e-01 4.66220021e-01 -7.43159711e-01
-9.91211176e-01 -1.00550091e+00 9.70735431e-01 -3.98664147e-01
5.06674469e-01 -4.92424607e-01 -1.34870541e+00 8.95581841e-01
7.40841627e-01 -3.22087616e-01 1.03334022e+00 5.06255031e-01
-2.37792462e-01 -2.29059920e-01 -1.06035256e+00 4.93371695e-01
8.36070895e-01 -2.74222106e-01 -1.30366194e+00 3.91694717e-03
5.85337162e-01 -3.85732800e-01 -5.53313494e-01 5.51268458e-01
5.11814594e-01 -7.47696579e-01 5.92375100e-01 -8.29998016e-01
2.12133914e-01 -2.16246441e-01 4.83037494e-02 -1.72773385e+00
-4.28592980e-01 -3.17230195e-01 1.78552270e-01 1.42496789e+00
8.53856385e-01 -7.58610249e-01 7.51210749e-01 5.40234327e-01
-2.91485310e-01 -4.16390538e-01 -9.38059568e-01 -7.95859396e-01
6.80751383e-01 -4.56708461e-01 5.35417140e-01 9.57249582e-01
1.88526601e-01 6.33426309e-01 -1.57012492e-01 2.13756278e-01
2.53324479e-01 1.68803453e-01 5.90509415e-01 -1.13699698e+00
-3.80975991e-01 -3.65329862e-01 1.63384393e-01 -9.61224318e-01
2.71365345e-01 -8.13623011e-01 2.70343274e-01 -1.75165248e+00
1.08804800e-01 -4.31947827e-01 -7.21573591e-01 6.65531993e-01
-4.90004748e-01 2.85243914e-02 3.03039283e-01 -9.27829444e-02
-5.08794427e-01 3.40397239e-01 1.32748377e+00 -5.64387739e-02
-4.34857398e-01 2.14943871e-01 -7.07345486e-01 9.01748538e-01
1.00064087e+00 -4.94916260e-01 -3.58926386e-01 -9.10623074e-01
-1.12808011e-01 -2.47382641e-01 -1.12635113e-01 -9.63654399e-01
2.82507718e-01 -3.12378049e-01 3.54672641e-01 -4.08814967e-01
1.09815992e-01 -9.37254250e-01 -2.03535914e-01 4.01215494e-01
-3.10732163e-02 -8.50901753e-03 6.47778392e-01 3.69580477e-01
-2.97629863e-01 -4.15244460e-01 7.33215690e-01 -3.08512390e-01
-1.18329620e+00 3.16515155e-02 -4.74751562e-01 -1.07411765e-01
1.10705841e+00 -3.31100821e-01 -2.68566161e-01 -1.10023148e-01
-6.87938571e-01 2.47166261e-01 9.00968835e-02 5.51048517e-01
6.29277289e-01 -1.05846918e+00 -7.41604865e-01 2.93975621e-01
2.57551819e-02 1.39924690e-01 -7.95257464e-02 3.66280437e-01
-4.03870702e-01 2.84547061e-01 -4.15453911e-01 -4.56655979e-01
-9.67662275e-01 6.37254834e-01 1.68339729e-01 -3.85929108e-01
-2.75125235e-01 8.72641087e-01 2.73947090e-01 -9.04106081e-01
2.27259502e-01 -1.67307869e-01 -3.46123129e-01 -3.47984545e-02
2.20613688e-01 1.52890146e-01 2.87303794e-02 -2.75531054e-01
-4.47480708e-01 5.47516167e-01 -3.63581389e-01 -2.71432370e-01
1.23712337e+00 4.05393355e-02 1.15183435e-01 7.25672901e-01
6.98687673e-01 -1.92899004e-01 -1.31626678e+00 -4.78946507e-01
3.71146321e-01 -1.94424629e-01 -2.22140819e-01 -1.22566891e+00
-7.32102513e-01 1.26043546e+00 6.69604301e-01 -5.32118045e-02
1.29801488e+00 -2.11488575e-01 8.00161779e-01 3.52983773e-01
4.50578123e-01 -1.41416323e+00 -1.35931130e-02 8.06542337e-01
5.95998943e-01 -1.42753839e+00 -1.20906934e-01 -4.17274356e-01
-8.41597319e-01 9.80096340e-01 1.13413036e+00 -1.95376784e-01
6.76324606e-01 2.21800104e-01 4.12481040e-01 4.18038100e-01
-6.71381414e-01 -2.76814669e-01 2.92467415e-01 8.97958159e-01
7.22995639e-01 3.01636476e-02 -5.51491559e-01 8.63792837e-01
8.23599249e-02 2.71427751e-01 1.41218886e-01 1.11831677e+00
-6.09556258e-01 -1.70498085e+00 -4.53606337e-01 1.84884816e-01
-3.33734661e-01 -1.54244289e-01 -4.51010227e-01 1.01651061e+00
3.68271351e-01 9.75081444e-01 6.19556569e-02 -4.57288623e-01
4.79876965e-01 5.76102316e-01 5.53746879e-01 -1.05070126e+00
-7.76948810e-01 3.83412540e-01 -3.86489108e-02 -9.04580131e-02
-6.42938495e-01 -6.78695798e-01 -1.41038311e+00 1.25293791e-01
-2.69627124e-01 1.86917812e-01 4.53883022e-01 1.21907926e+00
1.97135985e-01 4.30967152e-01 2.58311898e-01 -6.30151749e-01
-3.73054177e-01 -1.25103223e+00 -2.58108288e-01 2.45694935e-01
6.17679395e-02 -5.11248887e-01 2.54468113e-01 3.19747776e-01] | [9.700654983520508, 1.676377296447754] |
d13ec9a3-f58f-42e0-83eb-65ee063cd709 | continual-learning-for-out-of-distribution | 2306.15117 | null | https://arxiv.org/abs/2306.15117v1 | https://arxiv.org/pdf/2306.15117v1.pdf | Continual Learning for Out-of-Distribution Pedestrian Detection | A continual learning solution is proposed to address the out-of-distribution generalization problem for pedestrian detection. While recent pedestrian detection models have achieved impressive performance on various datasets, they remain sensitive to shifts in the distribution of the inference data. Our method adopts and modifies Elastic Weight Consolidation to a backbone object detection network, in order to penalize the changes in the model weights based on their importance towards the initially learned task. We show that when trained with one dataset and fine-tuned on another, our solution learns the new distribution and maintains its performance on the previous one, avoiding catastrophic forgetting. We use two popular datasets, CrowdHuman and CityPersons for our cross-dataset experiments, and show considerable improvements over standard fine-tuning, with a 9% and 18% miss rate percent reduction improvement in the CrowdHuman and CityPersons datasets, respectively. | ['Michael Greenspan', 'Ali Etemad', 'Mahdiyar Molahasani'] | 2023-06-26 | null | null | null | null | ['pedestrian-detection'] | ['computer-vision'] | [-9.70599428e-02 -2.65799344e-01 -8.12057182e-02 -3.56843024e-01
-2.82675803e-01 -2.37619132e-01 5.50683022e-01 2.73606420e-01
-1.03029704e+00 9.45617437e-01 1.56535745e-01 -2.55670715e-02
2.90687650e-01 -7.41048634e-01 -6.63925588e-01 -7.68544376e-01
-3.64008956e-02 6.48468196e-01 1.20697379e+00 -8.56740307e-03
-3.42947058e-02 3.52883220e-01 -1.54497373e+00 2.66379774e-01
5.81428528e-01 5.35967290e-01 9.34769735e-02 1.03801012e+00
2.56859928e-01 4.95426685e-01 -7.42187142e-01 -8.62468541e-01
2.60701507e-01 2.14163437e-01 -4.31975394e-01 1.81389391e-01
8.08499813e-01 -3.97676677e-01 -8.34953427e-01 9.66598570e-01
7.09445834e-01 3.29809427e-01 6.72149599e-01 -1.27434611e+00
-6.56058729e-01 4.33565080e-01 -1.01807594e+00 8.44363689e-01
-7.69155174e-02 8.38503361e-01 5.71412981e-01 -8.65384340e-01
2.93459415e-01 1.43136883e+00 9.83733773e-01 8.12109113e-01
-1.39554071e+00 -7.25216746e-01 5.18911183e-01 3.26297045e-01
-1.33892298e+00 -4.13233936e-01 2.85940081e-01 -6.00641310e-01
9.71228957e-01 -1.19694062e-01 8.79090607e-01 1.40426385e+00
8.24471340e-02 6.85139000e-01 6.08042240e-01 -4.34486687e-01
2.95540035e-01 2.47561902e-01 4.01863664e-01 6.12088501e-01
8.58421385e-01 3.22574079e-01 -4.59016383e-01 -1.26645640e-01
5.60934663e-01 -1.27705574e-01 8.80879611e-02 -5.73093474e-01
-7.33226955e-01 7.82961547e-01 3.77136230e-01 -1.25752091e-01
-2.41245031e-01 3.17516536e-01 5.12222886e-01 3.48069891e-02
3.38245898e-01 1.04487290e-05 -5.28867543e-01 -6.38154522e-02
-8.62719655e-01 5.46925545e-01 5.00516772e-01 8.36306155e-01
5.87180614e-01 -9.27710161e-03 -4.46972400e-01 7.59206772e-01
1.27610341e-01 6.10811055e-01 4.28543478e-01 -6.82913482e-01
4.01226848e-01 3.04191619e-01 3.87499422e-01 -8.49212110e-01
-3.62440199e-01 -5.93125820e-01 -7.19641864e-01 3.33375514e-01
7.83930719e-01 -4.54589009e-01 -1.14302516e+00 1.89653468e+00
4.75039363e-01 3.39083880e-01 -2.69253582e-01 7.07134128e-01
4.52912688e-01 3.18398833e-01 8.57989728e-01 1.41952604e-01
1.22081137e+00 -8.87608767e-01 7.51222903e-03 -5.83240151e-01
6.68702647e-02 -5.64611614e-01 1.10155416e+00 3.17526311e-01
-9.10339177e-01 -9.23672259e-01 -9.00487304e-01 2.30061173e-01
-1.64202541e-01 2.60090139e-02 4.00637060e-01 6.50606632e-01
-1.02284741e+00 4.68963891e-01 -7.93681562e-01 -7.35626876e-01
7.68385589e-01 1.73624218e-01 4.98212466e-04 2.48860437e-02
-8.22970688e-01 9.30784166e-01 6.59997344e-01 -4.46257919e-01
-9.28220809e-01 -9.46498513e-01 -2.65225828e-01 1.86038643e-01
2.38988444e-01 -1.06211829e+00 1.11750841e+00 -5.97871542e-01
-1.07144833e+00 9.37745810e-01 4.90683801e-02 -1.00265789e+00
8.98864746e-01 -4.22897965e-01 -4.34203655e-01 -1.16696537e-01
9.68617424e-02 1.11888397e+00 1.19264901e+00 -1.11525202e+00
-1.25210035e+00 -1.93475321e-01 1.15672536e-02 -1.68158308e-01
-5.21729469e-01 -2.19155133e-01 -5.90534627e-01 -7.77372599e-01
-4.69145238e-01 -9.21868324e-01 -2.92885303e-01 2.61934519e-01
-1.45347416e-01 -1.98437199e-01 6.72631919e-01 -3.45105171e-01
1.22887456e+00 -2.08155870e+00 -2.22510677e-02 -2.18946636e-02
4.13017482e-01 6.60502970e-01 -2.47040033e-01 -2.52516598e-01
-6.42772904e-03 -3.19448531e-01 -8.55790079e-03 -6.01064444e-01
-1.33247554e-01 3.06243181e-01 -2.16006488e-01 3.36424708e-01
2.93140441e-01 6.67537987e-01 -9.03642237e-01 -4.40550357e-01
1.28455505e-01 3.95753056e-01 -7.76905298e-01 6.92145303e-02
-5.15486933e-02 8.84390920e-02 -2.33288743e-02 2.78530806e-01
7.41941214e-01 -2.42248312e-01 -1.81083113e-01 -2.29614958e-01
1.45118413e-02 -4.45657372e-01 -1.07262254e+00 1.08637166e+00
9.98152494e-02 5.85106909e-01 -3.30755740e-01 -6.71349645e-01
6.90941870e-01 -1.36767522e-01 1.39019370e-01 -4.84741569e-01
-8.04512873e-02 -1.95597231e-01 1.32776499e-01 -4.53205258e-01
8.22079539e-01 -1.53630421e-01 8.79951343e-02 9.79367271e-02
1.71178922e-01 4.86091167e-01 5.37713349e-01 3.34386230e-01
1.00839126e+00 -1.67924091e-01 1.14566453e-01 -2.43386358e-01
3.44892830e-01 -5.11650033e-02 5.50731003e-01 1.24728024e+00
-6.91959858e-01 5.67870736e-01 3.34774941e-01 -8.64859343e-01
-1.32453275e+00 -1.41403115e+00 -1.34250581e-01 1.54368842e+00
-1.22117074e-02 -7.61622414e-02 -7.74864256e-01 -7.47323871e-01
4.07151639e-01 7.12659717e-01 -7.84465551e-01 -4.21856821e-01
-7.77855515e-01 -1.43694842e+00 5.77728927e-01 8.59404624e-01
5.34223974e-01 -7.40682244e-01 -8.54141295e-01 2.98234969e-01
6.65308312e-02 -1.00023758e+00 -4.13180321e-01 5.59613109e-02
-4.74441916e-01 -1.20697379e+00 -8.88722479e-01 -5.93635678e-01
5.51961184e-01 3.49352300e-01 1.34990358e+00 1.80807054e-01
-5.32531977e-01 5.54688573e-01 3.70397344e-02 -4.08037394e-01
-2.76445568e-01 3.48721176e-01 3.75699580e-01 -4.86452058e-02
6.12220109e-01 -3.99559587e-01 -6.78048253e-01 2.31251583e-01
-5.95495343e-01 -3.25335026e-01 2.74786174e-01 8.30733836e-01
2.65934974e-01 -3.20822187e-02 5.08686900e-01 -7.04088330e-01
5.64279497e-01 -3.88900667e-01 -6.59910321e-01 1.32134140e-01
-5.15486419e-01 3.76442522e-02 2.18283609e-01 -8.74125063e-01
-1.15310299e+00 -9.31629445e-03 -5.32891089e-03 -4.92225856e-01
-1.10810682e-01 -2.49918193e-01 1.30787447e-01 -2.82005710e-03
1.02244079e+00 -9.33939070e-02 -3.02045554e-01 -3.63724113e-01
3.42989981e-01 1.74452011e-02 9.70756829e-01 -7.70948827e-01
1.02929616e+00 4.70936686e-01 -2.86038756e-01 -6.25880837e-01
-8.60369384e-01 -4.42754984e-01 -7.64160693e-01 -2.28912488e-01
8.17275584e-01 -9.92225528e-01 -6.63718164e-01 6.64750457e-01
-1.10929155e+00 -3.92854184e-01 -7.23704517e-01 1.86953068e-01
-1.85557961e-01 3.75061125e-01 -5.18020689e-01 -7.20933259e-01
-1.75789252e-01 -7.49150455e-01 7.01402903e-01 4.65727836e-01
-2.39417806e-01 -9.57319677e-01 2.25779027e-01 -2.03758225e-01
5.59817433e-01 1.22132860e-01 7.68449903e-01 -3.25223118e-01
-3.16082209e-01 -2.00500533e-01 -4.59187686e-01 2.00727031e-01
-3.75050306e-02 2.38123089e-01 -9.11003172e-01 -8.00026357e-01
-5.93494833e-01 -1.53958276e-01 1.43534470e+00 5.65507233e-01
8.14969122e-01 -2.12698415e-01 -7.08813965e-01 2.80859023e-01
1.20258331e+00 -1.84174448e-01 5.65165520e-01 5.45216203e-01
5.21032751e-01 2.55852848e-01 1.28966406e-01 6.75972044e-01
4.05600578e-01 5.42217255e-01 2.24084839e-01 -6.40340336e-03
-6.72726989e-01 -3.18655014e-01 5.95410205e-02 8.21060911e-02
-1.16168119e-01 -2.76577950e-01 -7.85545945e-01 6.38674140e-01
-2.04489923e+00 -1.24099720e+00 2.14364335e-01 2.35483217e+00
6.88663006e-01 8.61151278e-01 9.15473521e-01 -7.98989981e-02
9.16687012e-01 -7.67916664e-02 -7.45734692e-01 1.08397543e-01
-2.80134529e-01 -7.45677948e-02 4.20755982e-01 5.82476974e-01
-1.34689116e+00 1.05340528e+00 7.08799028e+00 6.09781146e-01
-9.24473405e-01 2.34525755e-01 7.34768867e-01 -4.07860875e-01
3.29865396e-01 -3.00456524e-01 -1.20435107e+00 7.79441416e-01
7.54945457e-01 -1.45335376e-01 1.64110940e-02 9.15758073e-01
-8.30576196e-02 -2.47196004e-01 -9.36035573e-01 8.64720047e-01
-1.50371969e-01 -1.26493371e+00 3.69694680e-01 -2.89494038e-01
7.18605816e-01 7.24680722e-02 2.99265176e-01 6.59946501e-01
7.34483659e-01 -5.50455809e-01 8.46492410e-01 7.68185675e-01
3.76782358e-01 -8.85832965e-01 5.46502233e-01 2.84882575e-01
-1.16841364e+00 -3.26481491e-01 -7.45845020e-01 2.44876463e-02
8.98857638e-02 4.64768738e-01 -8.70111465e-01 -2.24113777e-01
1.13039160e+00 2.48573482e-01 -1.09144747e+00 1.51815641e+00
-1.03913359e-01 6.13098323e-01 -3.55372995e-01 -2.95225102e-02
4.55176570e-02 4.31851417e-01 8.20258856e-01 1.64282620e+00
-1.26780376e-01 -2.20342606e-01 4.60690051e-01 4.27612573e-01
-1.89196449e-02 -5.35624802e-01 -1.33863449e-01 5.94062209e-01
5.30098021e-01 9.91485059e-01 -6.77937090e-01 -7.41612434e-01
-2.35125244e-01 7.77448773e-01 6.59955740e-01 5.54208815e-01
-1.17003393e+00 -1.49644867e-01 9.07791734e-01 5.20628273e-01
8.56084287e-01 -2.01059222e-01 -1.05228797e-01 -1.04221618e+00
-3.69147919e-02 -4.35478508e-01 6.53605580e-01 -2.92244107e-01
-1.65944517e+00 4.42199439e-01 2.61941582e-01 -7.93614805e-01
-1.00346871e-01 -4.97687578e-01 -5.26800036e-01 5.72380245e-01
-1.39256644e+00 -1.03913152e+00 -3.22338253e-01 5.63883722e-01
6.62581325e-01 -4.10533369e-01 3.68675143e-01 5.56671739e-01
-6.87252879e-01 1.06568551e+00 -1.21834807e-01 1.64013743e-01
7.88948357e-01 -1.11847115e+00 7.26540685e-01 1.10359240e+00
-1.89733237e-01 4.50889140e-01 1.00238729e+00 -7.16324389e-01
-6.67990208e-01 -1.47339249e+00 6.33726418e-01 -7.43887424e-01
7.06010759e-01 -1.22373074e-01 -9.58608210e-01 6.82025194e-01
2.24414784e-02 2.73266166e-01 3.65665287e-01 3.11894864e-01
-6.59237027e-01 -1.65143803e-01 -1.38916385e+00 6.64465785e-01
1.28536654e+00 5.83047047e-02 -6.21208966e-01 2.00907603e-01
5.64243555e-01 -2.96193391e-01 -3.09021890e-01 1.80323243e-01
6.91441596e-01 -9.74631250e-01 1.28370619e+00 -1.15952647e+00
-5.75777749e-03 -4.61445570e-01 -1.31591082e-01 -1.27895439e+00
-1.10630643e+00 -3.99730742e-01 -4.35956568e-01 1.07902980e+00
2.55483121e-01 -4.58180904e-01 1.02463603e+00 6.28788173e-01
1.01131439e-01 -4.34163779e-01 -1.12768662e+00 -8.22531044e-01
1.91188663e-01 -2.62136817e-01 4.91021663e-01 2.44029850e-01
-4.99571294e-01 5.70088804e-01 -5.26349366e-01 3.70258927e-01
1.12883222e+00 -3.26467633e-01 1.00904346e+00 -1.34374881e+00
-4.86430794e-01 -5.52101195e-01 -6.26576006e-01 -1.03719032e+00
-1.12942070e-01 -3.84863526e-01 2.75307260e-02 -9.92366910e-01
6.49786472e-01 -3.47711504e-01 -3.81454021e-01 4.67750072e-01
-6.86751425e-01 4.45205688e-01 4.36314255e-01 1.52731284e-01
-1.12496078e+00 4.30100560e-01 6.94520295e-01 -4.61565942e-01
-1.86486349e-01 1.72451511e-01 -6.32118464e-01 9.61460531e-01
5.58145046e-01 -5.80462933e-01 -1.82256028e-01 -6.02634013e-01
-1.99376807e-01 -7.01004744e-01 6.82907462e-01 -1.67011940e+00
4.46787387e-01 3.07203270e-02 1.04861236e+00 -5.53865433e-01
2.53124446e-01 -4.92268205e-01 -7.51735643e-02 7.46922672e-01
-2.47699007e-01 2.26901010e-01 5.89837551e-01 9.58278418e-01
4.81005192e-01 1.86129976e-02 1.37872803e+00 -5.68468384e-02
-9.29228306e-01 4.81597096e-01 -5.25307119e-01 3.64010483e-01
9.90467072e-01 -2.81588346e-01 -3.46682459e-01 3.36267240e-02
-1.03825331e+00 2.72712827e-01 3.71618807e-01 5.04915118e-01
5.67677915e-01 -1.29117560e+00 -8.84339094e-01 7.14524090e-02
1.06281444e-01 -2.77759731e-01 3.02610397e-01 3.60169262e-01
-6.30435720e-02 1.39104858e-01 -4.50762451e-01 -7.14978516e-01
-1.28861368e+00 7.13105023e-01 3.41764718e-01 -3.47600698e-01
-6.17795646e-01 1.11671543e+00 1.88559324e-01 -3.22192535e-02
5.88248193e-01 -2.51436740e-01 2.58666743e-02 2.05869116e-02
7.77582526e-01 8.03111851e-01 -1.53277233e-01 -3.50270182e-01
-3.43802273e-01 2.04141051e-01 -5.95120370e-01 1.85071513e-01
1.14273429e+00 -2.40910172e-01 5.29792368e-01 2.17632577e-01
7.46190071e-01 -2.22024515e-01 -2.02707767e+00 -4.80087548e-01
1.20548364e-02 -6.11183822e-01 -4.97589648e-01 -7.47694790e-01
-9.94462192e-01 5.62408328e-01 1.03479528e+00 1.50800291e-02
6.98550820e-01 -1.83997359e-02 7.96846628e-01 5.76390505e-01
4.57930624e-01 -1.14664876e+00 4.26645786e-01 6.04366601e-01
5.26234448e-01 -1.27707136e+00 1.30554482e-01 7.51619935e-02
-4.21506315e-01 1.00817382e+00 9.80066001e-01 -2.74946690e-01
6.69165134e-01 3.50781560e-01 -3.97898734e-01 1.19737260e-01
-6.73861027e-01 -1.65056497e-01 -7.71529824e-02 1.00124502e+00
-3.23879160e-02 1.16219528e-01 2.55952537e-01 3.42396140e-01
-1.21173833e-03 5.25126830e-02 1.81170553e-01 6.45622313e-01
-9.06647801e-01 -7.00197935e-01 -4.45044816e-01 4.11576867e-01
-5.05230688e-02 4.65791970e-02 -4.85631041e-02 8.21559846e-01
5.07132173e-01 7.02537060e-01 3.25185567e-01 -2.72778332e-01
6.73993051e-01 -3.08326911e-02 6.91235781e-01 -3.88506353e-01
-4.39548731e-01 -3.34931582e-01 -3.27427499e-02 -2.33422205e-01
-2.51022339e-01 -9.13796484e-01 -9.32571769e-01 -7.01619029e-01
-8.45713392e-02 -1.82862207e-01 -9.42984819e-02 7.31192529e-01
3.14350575e-01 5.95084488e-01 1.07172310e-01 -8.89290690e-01
-7.01982081e-01 -8.41738284e-01 -1.99047685e-01 5.52866161e-01
4.32177961e-01 -1.04707646e+00 -7.53667876e-02 1.66290790e-01] | [8.214146614074707, -0.2475028932094574] |
1b66a5d6-43e5-4b02-bdb8-0687acd4c722 | unsupervised-domain-adaptation-for-cross-1 | 2109.05664 | null | https://arxiv.org/abs/2109.05664v3 | https://arxiv.org/pdf/2109.05664v3.pdf | Unsupervised domain adaptation for cross-modality liver segmentation via joint adversarial learning and self-learning | Liver segmentation on images acquired using computed tomography (CT) and magnetic resonance imaging (MRI) plays an important role in clinical management of liver diseases. Compared to MRI, CT images of liver are more abundant and readily available. However, MRI can provide richer quantitative information of the liver compared to CT. Thus, it is desirable to achieve unsupervised domain adaptation for transferring the learned knowledge from the source domain containing labeled CT images to the target domain containing unlabeled MR images. In this work, we report a novel unsupervised domain adaptation framework for cross-modality liver segmentation via joint adversarial learning and self-learning. We propose joint semantic-aware and shape-entropy-aware adversarial learning with post-situ identification manner to implicitly align the distribution of task-related features extracted from the target domain with those from the source domain. In proposed framework, a network is trained with the above two adversarial losses in an unsupervised manner, and then a mean completer of pseudo-label generation is employed to produce pseudo-labels to train the next network (desired model). Additionally, semantic-aware adversarial learning and two self-learning methods, including pixel-adaptive mask refinement and student-to-partner learning, are proposed to train the desired model. To improve the robustness of the desired model, a low-signal augmentation function is proposed to transform MRI images as the input of the desired model to handle hard samples. Using the public data sets, our experiments demonstrated the proposed unsupervised domain adaptation framework reached four supervised learning methods with a Dice score 0.912 plus or minus 0.037 (mean plus or minus standard deviation). | ['Simon Chun-Ho Yu', 'Weitian Chen', 'Jin Hong'] | 2021-09-13 | null | null | null | null | ['liver-segmentation'] | ['medical'] | [ 5.19376516e-01 3.25835258e-01 -2.93088797e-02 -5.41751921e-01
-9.09145713e-01 -6.15107298e-01 3.69024485e-01 -3.32672112e-02
-5.22739768e-01 8.33640456e-01 1.92675784e-01 -2.56154358e-01
5.20080179e-02 -6.73472345e-01 -6.12654567e-01 -1.01847410e+00
-2.30917141e-01 6.83545709e-01 3.53724398e-02 2.67413139e-01
-9.84986126e-02 3.96273166e-01 -5.19527078e-01 1.62274674e-01
1.29838610e+00 1.01716352e+00 2.14036778e-01 5.45501173e-01
5.69947511e-02 6.49608433e-01 -3.69186819e-01 -7.81871825e-02
6.69047952e-01 -8.02590072e-01 -8.61244500e-01 3.94905895e-01
-6.18709065e-02 -2.81536520e-01 -3.21745574e-01 1.30376339e+00
6.55509949e-01 2.23967835e-01 9.96829093e-01 -8.95676851e-01
-6.84026241e-01 5.51329315e-01 -6.31950140e-01 2.31631920e-01
-2.16893367e-02 5.19063890e-01 2.74658144e-01 -5.95640063e-01
5.28212726e-01 7.73962200e-01 4.51756567e-01 6.30423963e-01
-1.26531076e+00 -8.06768000e-01 -1.98356539e-01 -2.25336984e-01
-1.21360946e+00 -1.45928279e-01 9.96274352e-01 -4.58137363e-01
2.43561745e-01 3.41624096e-02 6.48883998e-01 7.91526794e-01
2.20545784e-01 6.17858171e-01 1.63162112e+00 -3.35903347e-01
1.84748828e-01 4.26113546e-01 -4.68202740e-01 8.14419150e-01
-1.67177767e-01 3.38404715e-01 2.44921789e-01 -2.32041180e-01
1.01766932e+00 6.77970797e-02 -2.89822042e-01 -6.78216755e-01
-1.26694477e+00 8.30453277e-01 8.26608479e-01 3.44746590e-01
-6.58329606e-01 -2.49823168e-01 4.87328202e-01 2.09137246e-01
4.21845496e-01 3.95991534e-01 -4.46890354e-01 4.81707245e-01
-8.64508867e-01 -2.06741795e-01 7.18197107e-01 7.19095290e-01
4.27005500e-01 2.75881380e-01 -2.84289986e-01 7.05283880e-01
3.83643270e-01 5.22364557e-01 1.02303779e+00 -7.53129363e-01
1.88209578e-01 4.96355951e-01 -1.68740347e-01 -5.97946942e-01
-3.13961059e-01 -7.17593729e-01 -1.22505534e+00 2.39807874e-01
4.28546876e-01 -3.36075664e-01 -1.28408206e+00 1.63753402e+00
5.89444041e-01 4.95251209e-01 1.67770311e-01 1.17949820e+00
6.75319493e-01 3.86743695e-01 5.37863433e-01 -3.86445045e-01
1.25046849e+00 -9.26197290e-01 -3.94576132e-01 -1.21963218e-01
2.60052055e-01 -7.25498855e-01 8.65974903e-01 1.05520800e-01
-9.84159410e-01 -4.55494642e-01 -9.30848956e-01 4.84268248e-01
5.78960106e-02 -1.35213271e-01 2.90547699e-01 5.66235304e-01
-7.41253078e-01 4.25428480e-01 -1.03282344e+00 -1.01147808e-01
6.87534809e-01 5.54899454e-01 -4.48522151e-01 -3.88554223e-02
-1.13210821e+00 8.26974034e-01 6.55225873e-01 -2.63312757e-01
-1.26799023e+00 -8.49589467e-01 -8.85484695e-01 -1.48435354e-01
2.51917511e-01 -7.52991259e-01 9.68814492e-01 -1.54869699e+00
-1.58073771e+00 1.07823896e+00 3.90880167e-01 -6.05179250e-01
7.35306382e-01 4.44286019e-01 -1.69997349e-01 4.26222205e-01
4.88233902e-02 7.19448566e-01 1.02825534e+00 -1.36381614e+00
-2.14547202e-01 -3.29189748e-01 -3.95479858e-01 4.14128572e-01
-8.84820893e-03 -7.26385936e-02 -1.01631377e-02 -1.08992255e+00
2.34965518e-01 -9.24466848e-01 -3.93230766e-01 5.45020960e-02
-3.44035268e-01 1.90280497e-01 6.56818330e-01 -1.13041019e+00
6.19110286e-01 -2.00920153e+00 8.07647333e-02 4.36556846e-01
1.32603914e-01 2.06731200e-01 -9.37976539e-02 -4.10849005e-01
-4.23777312e-01 -5.36221303e-02 -9.43048716e-01 4.58744653e-02
-3.00777882e-01 2.55501568e-01 2.81270117e-01 8.07890236e-01
1.61545843e-01 8.86625588e-01 -1.05006504e+00 -7.74578333e-01
3.67343396e-01 3.83357465e-01 -4.28285003e-01 6.07095122e-01
1.02802413e-02 1.45904899e+00 -5.46303570e-01 6.01966918e-01
7.66149879e-01 -1.92359015e-01 1.89591721e-01 -2.76600927e-01
4.45484698e-01 -2.92101651e-01 -9.15616453e-01 1.84013379e+00
-3.66025358e-01 -6.00241572e-02 2.45749071e-01 -1.45419455e+00
9.04229701e-01 6.26311421e-01 8.51840973e-01 -5.89376748e-01
3.57728153e-01 1.67992234e-01 2.22434014e-01 -5.70512950e-01
-4.45237845e-01 -7.30169892e-01 5.59090376e-02 5.08417368e-01
2.43784755e-01 -4.44842607e-01 -3.66989553e-01 -1.73091844e-01
8.03623378e-01 2.77353227e-02 4.08999741e-01 -2.12415695e-01
8.62389803e-01 6.58169985e-02 7.58929193e-01 3.60857189e-01
-7.22047865e-01 6.18824422e-01 1.17092654e-02 -3.63305688e-01
-1.28519201e+00 -1.19303238e+00 -2.36397818e-01 6.03611350e-01
7.65492320e-02 6.92926526e-01 -8.51127625e-01 -1.32967556e+00
-4.69736792e-02 5.25836647e-01 -6.05901897e-01 -2.30416387e-01
-6.29256666e-01 -8.50497365e-01 5.21941125e-01 4.37022746e-01
6.56812191e-01 -1.13056242e+00 -2.89915502e-01 1.33266568e-01
-1.32911891e-01 -8.30136478e-01 -8.18520844e-01 2.33053058e-01
-1.06886053e+00 -1.17228568e+00 -1.15868247e+00 -1.07201982e+00
1.13380754e+00 -3.33184421e-01 9.35055315e-01 -1.16418786e-01
-3.00648481e-01 2.17591614e-01 -9.49859023e-02 -2.10290961e-02
-8.01129997e-01 -1.72593534e-01 4.99602966e-02 1.59232602e-01
-3.00774444e-03 -6.28816009e-01 -1.05228996e+00 2.40178496e-01
-1.08094597e+00 -8.89592222e-04 8.84570181e-01 1.24485779e+00
7.11753547e-01 -2.18613297e-02 7.13058710e-01 -9.86101270e-01
4.29185510e-01 -6.33912086e-01 -4.79684442e-01 1.58916011e-01
-6.80904448e-01 1.44129798e-01 8.86377156e-01 -7.10861742e-01
-1.25485849e+00 4.06825453e-01 -5.39643057e-02 -6.37662232e-01
-3.46485943e-01 4.38762367e-01 -4.88530882e-02 -3.06456298e-01
6.53221488e-01 5.64613581e-01 3.11775863e-01 -1.49793625e-01
2.19167471e-01 4.98963296e-01 7.58173645e-01 -5.37184477e-01
9.67050910e-01 2.94117421e-01 -5.39990477e-02 -1.44170731e-01
-3.98638487e-01 -3.19118619e-01 -6.90557301e-01 4.89955805e-02
9.75875258e-01 -8.75995219e-01 -1.93550035e-01 5.44049025e-01
-6.38588250e-01 -4.34849590e-01 -4.86560374e-01 7.80547142e-01
-7.32493758e-01 4.38181639e-01 -6.21709526e-01 -4.72791195e-01
-6.82398081e-01 -1.43644154e+00 4.32865292e-01 3.75340253e-01
1.45395711e-01 -1.32998180e+00 -9.28157642e-02 5.02484739e-01
5.37761450e-01 6.00228190e-01 8.77170742e-01 -1.20135331e+00
-2.92176187e-01 -1.13848552e-01 -2.92963654e-01 8.05249691e-01
3.37677121e-01 -7.44073451e-01 -5.67549288e-01 -5.92029274e-01
4.68947470e-01 -3.94763261e-01 4.40719187e-01 5.89683592e-01
1.23011756e+00 -3.82183552e-01 -1.31333545e-01 8.52666855e-01
1.51659894e+00 4.75402564e-01 3.02642494e-01 1.94114760e-01
4.98703152e-01 3.46552998e-01 4.32499141e-01 3.76649201e-01
2.08031267e-01 2.23609760e-01 4.49643791e-01 -4.47996497e-01
-3.10495973e-01 -1.78789005e-01 1.89250391e-02 6.26462996e-01
2.28537396e-01 2.88083375e-01 -9.68365431e-01 6.27938390e-01
-1.23106217e+00 -6.07824326e-01 3.58767599e-01 2.19659400e+00
1.23972702e+00 -1.63884833e-01 7.25630810e-03 -1.69056773e-01
7.96610773e-01 -2.48022124e-01 -8.33829463e-01 -1.18606547e-02
2.01314896e-01 4.19350207e-01 6.68189406e-01 4.94825214e-01
-1.32094252e+00 5.53704977e-01 5.42748737e+00 6.52470529e-01
-1.35038137e+00 4.80541110e-01 9.55489993e-01 4.53436106e-01
-2.17028037e-01 -1.47429883e-01 1.58517540e-01 6.49064481e-01
7.12843895e-01 -2.08966926e-01 5.37889421e-01 7.21943498e-01
2.62163401e-01 -5.45102581e-02 -1.00164700e+00 7.13111460e-01
9.02744010e-02 -7.84900904e-01 -1.50673747e-01 -1.49296850e-01
1.09316409e+00 -7.16406703e-02 1.74032792e-01 2.63996422e-01
3.94603997e-01 -1.09681654e+00 2.46329352e-01 4.89118248e-01
1.12136483e+00 -5.84115386e-01 9.00002360e-01 5.16282678e-01
-8.48217249e-01 7.13079646e-02 1.40069388e-02 4.86289024e-01
-2.60917190e-02 4.73854989e-01 -1.25658560e+00 6.18547797e-01
3.52712810e-01 3.53479475e-01 -2.41272762e-01 9.95467722e-01
-1.22110069e-01 5.67138791e-01 -1.93284005e-01 5.16232491e-01
1.39555618e-01 -3.77178550e-01 5.80253303e-01 1.10623062e+00
2.55075306e-01 3.43976796e-01 6.28190637e-01 8.69627535e-01
-1.88341007e-01 1.96054354e-01 -4.44748938e-01 2.49581218e-01
3.06357652e-01 1.23437107e+00 -8.37052464e-01 -5.22781909e-01
-2.29780421e-01 1.14514792e+00 -1.64228901e-01 2.02503935e-01
-9.03363407e-01 -1.60021514e-01 -6.51082546e-02 3.94455083e-02
-1.17461681e-01 2.02846989e-01 -4.07004684e-01 -1.10617554e+00
-4.64206010e-01 -9.31432486e-01 5.31464517e-01 -4.49403256e-01
-1.54746866e+00 6.81250155e-01 -9.03755352e-02 -1.31310332e+00
-2.77942628e-01 -2.21033633e-01 -6.68002605e-01 1.25698650e+00
-1.65282738e+00 -1.23073590e+00 -4.45258230e-01 8.49192977e-01
4.30000603e-01 -3.22723389e-01 7.28712797e-01 3.71420026e-01
-1.25481635e-01 6.53486311e-01 1.66485175e-01 5.88023603e-01
8.28541219e-01 -1.39732182e+00 -3.42597008e-01 6.90995157e-01
-4.27541196e-01 1.92023322e-01 5.66912055e-01 -6.60592198e-01
-9.45924997e-01 -1.34988642e+00 2.32712939e-01 2.46506780e-02
3.98811042e-01 2.65565366e-01 -8.70953500e-01 7.65658677e-01
2.63313293e-01 6.55358732e-01 7.09970713e-01 -8.26068640e-01
8.52446705e-02 -9.84811783e-02 -2.03943181e+00 1.71327144e-01
3.48861128e-01 -3.18144917e-01 -7.91642308e-01 4.29200053e-01
5.15556216e-01 -7.59104133e-01 -1.31525946e+00 6.74729049e-01
3.00168246e-01 -5.52679241e-01 1.07594824e+00 -4.86845791e-01
3.08784038e-01 -2.90435404e-01 4.41233963e-02 -1.67265749e+00
-1.32649049e-01 -2.94622660e-01 1.55348793e-01 8.82025957e-01
1.29994661e-01 -6.17330551e-01 9.67774808e-01 4.86336231e-01
-2.12605774e-01 -6.47387862e-01 -8.73580515e-01 -4.96081144e-01
5.78735054e-01 1.50383502e-01 4.15961027e-01 1.32995927e+00
-2.90072471e-01 -1.61897987e-01 -3.42213213e-02 2.47153997e-01
1.01538920e+00 -9.46339667e-02 2.80459762e-01 -8.52975547e-01
-3.11080545e-01 -2.34883070e-01 -2.31895044e-01 -6.82793617e-01
2.63768047e-01 -1.36805248e+00 1.66391030e-01 -1.24064445e+00
5.37085652e-01 -7.97494054e-01 -7.29714096e-01 4.69195247e-01
-2.12802619e-01 2.93725491e-01 3.85194011e-02 2.23634869e-01
-6.90259337e-02 4.73184377e-01 1.79291475e+00 -4.77745205e-01
-9.63626206e-02 2.06767604e-01 -4.73554313e-01 5.88349938e-01
7.82094836e-01 -5.81563532e-01 -4.17212397e-01 -9.23394486e-02
-6.77247405e-01 5.61072588e-01 3.80223602e-01 -8.30592334e-01
1.91071928e-01 -1.05634205e-01 7.45988905e-01 -1.16504841e-01
-1.89687997e-01 -1.21388113e+00 2.42060944e-01 8.60412300e-01
-4.47198510e-01 -3.07597727e-01 -4.78288792e-02 3.68974358e-01
-3.10162544e-01 -3.39616865e-01 1.38192809e+00 -5.40580451e-01
-2.76029766e-01 5.49656272e-01 1.97503064e-02 2.60481060e-01
1.29801345e+00 -3.81991118e-02 2.13888273e-01 -2.69697964e-01
-1.23853159e+00 2.49745309e-01 3.62058610e-01 -8.59889314e-02
7.18626261e-01 -1.36559534e+00 -9.53442872e-01 3.77692848e-01
-1.58664197e-01 3.29629600e-01 2.05781609e-01 1.10189688e+00
-5.49962521e-01 3.35221142e-02 -4.21966970e-01 -6.68133199e-01
-9.12711322e-01 7.17475235e-01 6.88929498e-01 -7.25023270e-01
-4.50012594e-01 6.86294019e-01 2.14351222e-01 -8.71790588e-01
8.93991664e-02 -1.17804386e-01 9.96237025e-02 -4.94004548e-01
1.24774665e-01 7.34166801e-02 -2.41239481e-02 -5.73741317e-01
-1.79641768e-01 3.41952562e-01 3.58348414e-02 -4.23490442e-02
1.22177315e+00 -1.27837136e-01 -1.30069956e-01 -1.70061544e-01
1.21281087e+00 -2.96033502e-01 -1.46995485e+00 -4.07448947e-01
-1.61860183e-01 -2.56379813e-01 1.53307423e-01 -1.28901267e+00
-1.46922159e+00 7.10482121e-01 1.08887863e+00 -2.63863266e-01
1.45193052e+00 -1.35966986e-01 8.08532178e-01 -3.31982732e-01
1.34850278e-01 -7.11389363e-01 1.14780381e-01 1.29962444e-01
6.23052001e-01 -1.55354178e+00 -2.03874826e-01 -3.09133917e-01
-8.98896277e-01 8.33126903e-01 6.61118448e-01 -2.97532141e-01
5.94653010e-01 2.09113300e-01 2.16598213e-01 1.48717076e-01
-1.64669588e-01 1.22318394e-01 3.46437961e-01 8.37695003e-01
3.13584208e-01 2.52843112e-01 -2.72856206e-01 6.69037402e-01
2.67439604e-01 1.69118449e-01 8.61833692e-02 7.32408047e-01
-2.15735525e-01 -1.06186068e+00 -6.00529969e-01 4.93583381e-01
-6.97387516e-01 -3.84650454e-02 2.88886815e-01 5.58497131e-01
1.82742104e-01 5.70496798e-01 -5.22317216e-02 3.35777202e-03
1.21976566e-02 2.47856885e-01 5.59623897e-01 -5.36323965e-01
-7.69167662e-01 1.70831159e-01 -5.32670677e-01 -1.82325959e-01
-4.46897268e-01 -4.89991635e-01 -1.45318246e+00 6.55516610e-02
-1.74756750e-01 3.41358304e-01 5.50329745e-01 8.45693231e-01
-1.32409967e-02 5.52413702e-01 1.12673199e+00 -6.38296545e-01
-9.77364063e-01 -1.07904351e+00 -5.40317714e-01 8.20936859e-01
2.46922925e-01 -3.61609071e-01 -3.57223809e-01 4.23978984e-01] | [14.543741226196289, -2.096034288406372] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.