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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
8e6acb4b-82a1-4ab5-8467-36753c280a40 | conditional-density-estimation-tools-in | 1908.11523 | null | https://arxiv.org/abs/1908.11523v2 | https://arxiv.org/pdf/1908.11523v2.pdf | Conditional Density Estimation Tools in Python and R with Applications to Photometric Redshifts and Likelihood-Free Cosmological Inference | It is well known in astronomy that propagating non-Gaussian prediction uncertainty in photometric redshift estimates is key to reducing bias in downstream cosmological analyses. Similarly, likelihood-free inference approaches, which are beginning to emerge as a tool for cosmological analysis, require a characterization of the full uncertainty landscape of the parameters of interest given observed data. However, most machine learning (ML) or training-based methods with open-source software target point prediction or classification, and hence fall short in quantifying uncertainty in complex regression and parameter inference settings. As an alternative to methods that focus on predicting the response (or parameters) $\mathbf{y}$ from features $\mathbf{x}$, we provide nonparametric conditional density estimation (CDE) tools for approximating and validating the entire probability density function (PDF) $\mathrm{p}(\mathbf{y}|\mathbf{x})$ of $\mathbf{y}$ given (i.e., conditional on) $\mathbf{x}$. As there is no one-size-fits-all CDE method, the goal of this work is to provide a comprehensive range of statistical tools and open-source software for nonparametric CDE and method assessment which can accommodate different types of settings and be easily fit to the problem at hand. Specifically, we introduce four CDE software packages in $\texttt{Python}$ and $\texttt{R}$ based on ML prediction methods adapted and optimized for CDE: $\texttt{NNKCDE}$, $\texttt{RFCDE}$, $\texttt{FlexCode}$, and $\texttt{DeepCDE}$. Furthermore, we present the $\texttt{cdetools}$ package, which includes functions for computing a CDE loss function for tuning and assessing the quality of individual PDFs, along with diagnostic functions. We provide sample code in $\texttt{Python}$ and $\texttt{R}$ as well as examples of applications to photometric redshift estimation and likelihood-free cosmological inference via CDE. | ['Alex I. Malz', 'Rafael Izbicki', 'Niccolò Dalmasso', 'Taylor Pospisil', 'Ann B. Lee', 'Peter E. Freeman'] | 2019-08-30 | null | null | null | null | ['photometric-redshift-estimation'] | ['miscellaneous'] | [-3.58895034e-01 -7.25516304e-02 6.45282269e-01 -7.14109123e-01
-1.03684807e+00 -6.32727206e-01 4.92915213e-01 -5.49569093e-02
-2.91056156e-01 8.51931214e-01 -3.80590260e-01 -8.14457655e-01
-5.45202971e-01 -1.09708846e+00 -7.17676222e-01 -9.71122861e-01
-1.16067313e-01 5.94548047e-01 4.47346359e-01 2.29054436e-01
5.02151132e-01 2.84711808e-01 -1.78951526e+00 -5.83057523e-01
1.02410829e+00 1.20381892e+00 1.57868251e-01 8.00687671e-01
-6.45624399e-02 8.17491040e-02 -5.81395388e-01 -6.76509500e-01
-1.27966297e-04 -3.40348750e-01 -3.73284072e-01 -7.45227635e-01
2.94925541e-01 -3.53318334e-01 -2.61937618e-01 1.15098608e+00
6.16689980e-01 4.60199594e-01 1.28934515e+00 -7.87695944e-01
-5.59345305e-01 1.92342848e-02 -3.56609553e-01 3.01957697e-01
-7.47498274e-02 4.77006465e-01 6.65883064e-01 -8.46350908e-01
1.94519371e-01 8.60121906e-01 6.81346595e-01 1.71158716e-01
-1.38612509e+00 -9.30581391e-01 -3.11799318e-01 -2.87989885e-01
-1.55772126e+00 -2.46685386e-01 1.87631205e-01 -8.79564524e-01
1.02141643e+00 5.05730689e-01 3.96642685e-01 6.73746705e-01
2.80147027e-02 -2.13809758e-02 1.16987836e+00 -4.29023713e-01
4.00070667e-01 1.55721456e-01 1.60422042e-01 5.77578962e-01
4.88628745e-02 4.07278568e-01 -2.60481298e-01 -4.33816463e-01
1.37927020e+00 -5.61813414e-01 -1.22805508e-02 7.78077543e-02
-7.37576008e-01 9.95845377e-01 5.67837432e-03 -2.04813614e-01
1.64300770e-01 4.58925664e-01 -9.21791494e-02 -1.88767657e-01
7.87350953e-01 2.84675211e-01 -6.44709587e-01 -5.65594196e-01
-9.04160261e-01 6.45385861e-01 6.67637646e-01 1.01783442e+00
1.18355906e+00 8.47065151e-02 -7.78751373e-02 1.21068215e+00
5.22417367e-01 7.64691412e-01 2.06343699e-02 -1.47833812e+00
8.06377754e-02 2.44925931e-01 3.30460161e-01 -3.93906057e-01
-4.30573910e-01 -2.92765766e-01 -3.61341089e-01 6.02029443e-01
7.65053511e-01 -2.59398609e-01 -6.16224170e-01 1.76021099e+00
3.32075626e-01 -1.67455841e-02 -4.84214783e-01 5.66520751e-01
9.23686504e-01 5.90027988e-01 -2.82892212e-02 -1.87180400e-01
9.91920114e-01 -2.12080702e-01 7.42398873e-02 -9.57833752e-02
4.58038568e-01 -9.24730480e-01 1.23673975e+00 2.40457088e-01
-1.33800948e+00 -2.92886347e-01 -6.71854496e-01 -1.70053795e-01
-2.75585532e-01 -2.56259382e-01 7.72950470e-01 1.09406924e+00
-1.06217802e+00 6.53936565e-01 -9.71747935e-01 -1.45877168e-01
1.80336073e-01 4.24971431e-01 2.87121266e-01 2.75763601e-01
-7.44944692e-01 6.61798835e-01 -7.89199471e-02 -1.49927855e-01
-7.47221768e-01 -1.10136521e+00 -5.80005527e-01 2.77469873e-01
1.38714194e-01 -4.12889719e-01 1.36350155e+00 -2.00572968e-01
-1.50561512e+00 8.79740655e-01 -4.71155457e-02 3.75220142e-02
2.69676089e-01 2.23773822e-01 -3.07881862e-01 2.30237078e-02
6.42810315e-02 5.17712355e-01 4.23661530e-01 -8.36794496e-01
-2.40548685e-01 -4.29647475e-01 -1.46101102e-01 -4.04862672e-01
-3.81113924e-02 3.99680316e-01 -4.32341218e-01 -4.65128720e-01
3.76154512e-01 -5.81266165e-01 7.85014182e-02 4.78175841e-02
-1.30876452e-01 -2.65147209e-01 2.00266793e-01 -7.62909353e-01
1.07537723e+00 -1.95515275e+00 -2.19658569e-01 4.02294248e-01
1.80462345e-01 -1.98351797e-02 5.71806669e-01 1.73637152e-01
2.93983035e-02 2.18089834e-01 -5.23237526e-01 -2.66056776e-01
4.69462842e-01 -1.44917712e-01 6.03986755e-02 5.58896244e-01
-3.30146365e-02 4.65414286e-01 -5.35030305e-01 -2.52698481e-01
5.65814793e-01 5.20628154e-01 -6.92909300e-01 2.61918783e-01
-5.81180513e-01 7.45352149e-01 -4.00740534e-01 5.04133463e-01
8.43204975e-01 -3.09241265e-01 -2.43988052e-01 2.90338159e-01
-7.68889248e-01 4.51737136e-01 -1.33927929e+00 1.12805116e+00
-2.21755773e-01 4.06786054e-01 2.51807064e-01 -8.67035091e-01
1.28087497e+00 -6.08613119e-02 3.28974009e-01 -4.82514888e-01
1.47023546e-02 3.79350603e-01 -2.92997807e-02 -1.96252584e-01
2.68778354e-01 -3.33325952e-01 1.79961547e-01 6.06598735e-01
2.86013216e-01 -7.75355697e-01 2.23101769e-02 -1.59563590e-02
1.07053733e+00 6.72223091e-01 -1.20578595e-01 -5.89592099e-01
2.74279833e-01 -3.96121293e-01 3.55463624e-01 8.30607116e-01
-1.13886639e-01 6.61346078e-01 7.01391995e-01 -7.80156534e-03
-1.17016757e+00 -1.53548551e+00 -1.02970469e+00 9.83020961e-01
-3.00306797e-01 -2.42270872e-01 -7.36888826e-01 1.31957889e-01
5.87207004e-02 1.13472605e+00 -2.14599043e-01 9.15446207e-02
-2.28878617e-01 -1.31182432e+00 5.88710427e-01 1.92033857e-01
3.07397336e-01 -7.72547841e-01 -3.26039612e-01 -1.88093886e-01
9.58188996e-02 -5.52392304e-01 -7.50595927e-02 1.08833984e-01
-6.17599785e-01 -7.58801281e-01 -5.94217360e-01 -2.27581680e-01
2.33631477e-01 -3.16982359e-01 1.20552480e+00 1.20815285e-01
-4.53924417e-01 7.32531548e-01 -1.26730055e-01 -5.59513092e-01
-2.04863667e-01 -3.29797834e-01 -2.15169005e-02 -4.65862215e-01
5.10345459e-01 -9.62570906e-01 -9.27962124e-01 4.44900692e-01
-7.23060608e-01 -4.05568480e-01 -1.05146132e-01 6.14763737e-01
6.01926029e-01 7.87135959e-02 4.85367686e-01 -3.18208873e-01
8.33031684e-02 -4.49435979e-01 -1.39629257e+00 8.80612805e-02
-7.62297273e-01 2.95032794e-03 3.44803005e-01 -5.19735739e-02
-1.19884253e+00 -5.92668414e-01 -6.34548008e-01 -2.13050738e-01
-2.75230110e-01 4.51696157e-01 6.79655448e-02 -2.09497046e-02
7.21208751e-01 2.02499405e-01 -2.57063806e-02 -7.95357764e-01
8.09765309e-02 6.69926524e-01 7.14778185e-01 -1.17203200e+00
4.85083818e-01 2.15842351e-01 3.23859930e-01 -7.94551492e-01
-4.12109524e-01 -1.97510973e-01 -4.17689860e-01 -4.11839873e-01
9.65810180e-01 -5.91770589e-01 -1.15409124e+00 3.63692045e-01
-4.57480103e-01 -4.93373185e-01 -9.58472192e-02 7.86256075e-01
-6.50398135e-01 5.43116510e-01 -2.44568795e-01 -1.42464793e+00
-7.41213514e-03 -1.36905050e+00 9.35675561e-01 6.94561779e-01
8.91389325e-02 -1.21486866e+00 1.49872983e-02 3.22335184e-01
4.55019236e-01 3.24706398e-02 1.14540684e+00 -2.82185078e-01
-7.98528373e-01 -2.80803263e-01 -4.75279957e-01 3.29898357e-01
-3.62077802e-01 5.30549109e-01 -1.29225743e+00 -1.48535743e-01
1.22850172e-01 -2.73253918e-01 6.34470344e-01 1.05679655e+00
1.44892085e+00 7.30652288e-02 -3.46165039e-02 8.06520402e-01
1.29162157e+00 -7.10202754e-02 8.93093705e-01 1.84897378e-01
4.98480983e-02 7.60440230e-01 1.67327076e-01 7.04578757e-01
2.53938884e-01 6.24264538e-01 3.45454603e-01 3.94137144e-01
2.83780545e-01 1.26161501e-01 2.20566496e-01 2.86966681e-01
-2.27082402e-01 9.41897482e-02 -8.39685619e-01 5.54242581e-02
-1.38532245e+00 -9.88315225e-01 -5.41919231e-01 2.85386705e+00
1.02276397e+00 2.44446158e-01 1.93044811e-01 -4.33453918e-01
5.87839663e-01 -1.89467892e-01 -6.26774132e-01 -2.09174931e-01
6.90948218e-03 8.68339658e-01 3.84521395e-01 7.92341769e-01
-6.29773915e-01 5.49102843e-01 5.35419989e+00 9.40142035e-01
-8.68477762e-01 9.80074704e-02 8.08773518e-01 -1.80017605e-01
-4.20376450e-01 3.36377084e-01 -9.59489107e-01 6.63278401e-01
1.27602649e+00 9.25012049e-04 5.75722933e-01 8.41635346e-01
4.00475174e-01 -7.48168647e-01 -9.09512699e-01 7.96952963e-01
-5.36721408e-01 -1.38786674e+00 -8.62046480e-01 2.17878133e-01
3.65342170e-01 1.89881280e-01 1.88542411e-01 5.95694959e-01
5.85117161e-01 -1.05381918e+00 6.50969088e-01 7.66022861e-01
1.21418321e+00 -5.59167445e-01 2.78032035e-01 3.43940288e-01
-9.43398774e-01 3.52448940e-01 -7.57131338e-01 -1.99936047e-01
7.40170404e-02 1.07966077e+00 -4.98091936e-01 6.47306681e-01
1.32560372e+00 8.84347185e-02 -6.13911390e-01 1.17770743e+00
6.42382447e-03 9.40101445e-01 -5.83177388e-01 -4.67370898e-02
-2.49852166e-01 -6.58182383e-01 4.84877497e-01 1.16279101e+00
7.86921918e-01 2.62915730e-01 -3.60872775e-01 1.63630319e+00
4.17464852e-01 -1.06592529e-01 6.38035461e-02 2.80092303e-02
5.55882752e-01 1.10847068e+00 -8.00646842e-01 -4.79425341e-02
-5.28100908e-01 7.46671706e-02 3.19837391e-01 4.76521105e-01
-8.40673923e-01 -5.53492248e-01 7.88737416e-01 2.63675928e-01
3.57657492e-01 -5.25527298e-01 -5.30660927e-01 -9.82221842e-01
-1.33014470e-01 -2.68620849e-01 3.28946769e-01 -1.18464041e+00
-1.50269032e+00 -5.47168143e-02 5.27146935e-01 -6.44701481e-01
-3.31239402e-01 -1.04685068e+00 -6.36659741e-01 1.62598681e+00
-9.63801026e-01 -5.92018247e-01 -1.59458280e-01 2.42210984e-01
-4.75612143e-03 -2.49999121e-01 8.76470804e-01 1.96129546e-01
-5.66022396e-01 3.89746398e-01 8.48610401e-01 -1.59468889e-01
6.96606576e-01 -1.35454810e+00 1.88416019e-01 6.34101689e-01
-4.58885908e-01 8.71548951e-01 9.98257458e-01 -8.07335854e-01
-1.08080757e+00 -6.20248854e-01 4.72565591e-01 -6.04165494e-01
8.43671739e-01 -5.67298889e-01 -9.15307105e-01 7.95419753e-01
-2.30632752e-01 -7.12609738e-02 6.36625409e-01 3.58635366e-01
-2.10352257e-01 1.03646867e-01 -1.37617600e+00 4.87848014e-01
6.94112659e-01 -3.95411193e-01 -2.76254565e-02 3.87726009e-01
6.44850314e-01 -4.39423710e-01 -1.44658983e+00 2.36353025e-01
6.66830719e-01 -1.47853315e+00 1.11524522e+00 -4.20254320e-02
1.66553736e-01 -1.71778828e-01 -4.15559858e-01 -7.48711884e-01
-2.12657854e-01 -6.77157223e-01 1.66883543e-01 1.44403470e+00
4.34436411e-01 -8.47443104e-01 7.07058966e-01 1.21152556e+00
-3.67649317e-01 -3.57667238e-01 -1.36593544e+00 -7.01214969e-01
8.12494218e-01 -8.60222042e-01 5.33802152e-01 3.66187453e-01
-3.95786673e-01 -3.66452992e-01 1.97533101e-01 3.95469278e-01
8.49678814e-01 -2.43910566e-01 6.87382936e-01 -1.29402745e+00
-6.80589318e-01 -6.94040596e-01 -1.18256509e-01 -7.41614938e-01
-1.86702028e-01 -6.71168268e-01 -8.09557885e-02 -1.30581832e+00
1.54071629e-01 -8.88460696e-01 1.47221491e-01 -6.69015646e-02
-2.92345956e-02 -5.38610891e-02 -4.62968588e-01 1.10319845e-01
3.08771525e-02 4.85331982e-01 7.77587831e-01 3.50743443e-01
7.85495415e-02 2.33738333e-01 -3.06669801e-01 6.88245654e-01
5.14223993e-01 -2.90785044e-01 -2.02965438e-01 -1.50969550e-01
4.55361694e-01 2.05222279e-01 9.42827284e-01 -8.72311115e-01
1.21620698e-02 -4.09251869e-01 5.13698936e-01 -6.59762859e-01
5.81824958e-01 -2.72783749e-02 1.84826180e-01 -1.58152804e-01
1.62567005e-01 -2.35275477e-01 2.62922734e-01 6.67717531e-02
3.84069830e-01 -8.87103140e-01 9.68749404e-01 -3.20589513e-01
2.36431509e-02 1.54071882e-01 -3.73836547e-01 1.93295464e-01
5.64863980e-01 8.34356472e-02 -6.69230103e-01 -6.05413258e-01
-7.75327802e-01 9.18089375e-02 5.99776804e-01 -4.15935457e-01
1.81139246e-01 -8.81499887e-01 -6.13599300e-01 1.15705803e-01
-5.35962656e-02 3.09966147e-01 5.26620448e-01 8.48498940e-01
-6.14314854e-01 4.29924518e-01 3.26088727e-01 -5.74334204e-01
-6.38445079e-01 2.59678066e-01 6.67811036e-01 2.31120214e-01
-1.98426574e-01 1.37415993e+00 2.72648960e-01 -8.06524098e-01
1.74845234e-01 -3.70461702e-01 3.21209371e-01 -5.42316794e-01
4.83155698e-01 8.32234800e-01 1.01575837e-01 -1.81638092e-01
-1.17728904e-01 3.19310278e-01 4.23523575e-01 -5.38589835e-01
1.35159874e+00 -3.39247763e-01 -5.48426807e-01 5.86929679e-01
9.54864383e-01 3.95379588e-02 -1.54995084e+00 -1.14710527e-02
-1.78156644e-01 -5.38776934e-01 -1.03950515e-01 -1.03517318e+00
-4.86545920e-01 9.20005798e-01 6.76845729e-01 4.23128366e-01
7.57785380e-01 3.61364603e-01 -3.66524644e-02 -6.47247629e-03
3.52580130e-01 -1.17141271e+00 -1.72537193e-01 4.33873534e-01
7.20279872e-01 -1.10244429e+00 2.78522611e-01 -2.10109115e-01
4.82184328e-02 1.06397188e+00 6.60166681e-01 1.44481093e-01
1.04788780e+00 1.03223026e-01 -2.88783789e-01 -2.60122389e-01
-3.10698032e-01 5.68587594e-02 2.20503166e-01 7.03734338e-01
6.63725913e-01 -1.46879792e-01 -2.76033849e-01 8.82731616e-01
-7.34058201e-01 -3.17702413e-01 4.21747267e-01 7.10391164e-01
-6.95784271e-01 -1.22368503e+00 -6.82123423e-01 8.15376461e-01
-3.69148463e-01 -2.40310699e-01 5.32576740e-02 5.87138236e-01
1.31569862e-01 1.03286564e+00 2.33669519e-01 9.31910425e-02
-2.15693876e-01 4.20261115e-01 6.54146373e-01 -5.85202038e-01
-3.04233491e-01 3.99837226e-01 4.66352366e-02 -1.90091118e-01
-5.42447791e-02 -1.16440308e+00 -1.39269829e+00 -4.56046194e-01
-1.92653984e-01 1.62547559e-01 8.47691178e-01 9.70927835e-01
-9.27207246e-02 3.67278844e-01 3.52961093e-01 -1.03254437e+00
-5.17653465e-01 -1.13434112e+00 -9.37716544e-01 -2.11673215e-01
-2.30532289e-01 -9.45671082e-01 -7.82626987e-01 -3.36175203e-01] | [7.199306964874268, 3.542888879776001] |
d0bc3eed-44d2-4d70-9545-96bbbbee67e8 | literature-on-hand-gesture-recognition-using | 2207.00329 | null | https://arxiv.org/abs/2207.00329v1 | https://arxiv.org/pdf/2207.00329v1.pdf | Literature on Hand GESTURE Recognition using Graph based methods | Skeleton based recognition systems are gaining popularity and machine learning models focusing on points or joints in a skeleton have proved to be computationally effective and application in many areas like Robotics. It is easy to track points and thereby preserving spatial and temporal information, which plays an important role in abstracting the required information, classification becomes an easy task. In this paper, we aim to study these points but using a cloud mechanism, where we define a cloud as collection of points. However, when we add temporal information, it may not be possible to retrieve the coordinates of a point in each frame and hence instead of focusing on a single point, we can use k-neighbors to retrieve the state of the point under discussion. Our focus is to gather such information using weight sharing but making sure that when we try to retrieve the information from neighbors, we do not carry noise with it. LSTM which has capability of long-term modelling and can carry both temporal and spatial information. In this article we tried to summarise graph based gesture recognition method. | ['Varun Sharma', 'Neha Baranwal'] | 2022-07-01 | null | null | null | null | ['hand-gesture-recognition', 'hand-gesture-recognition-1', 'gesture-recognition'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 6.43972382e-02 -2.02485561e-01 -1.17107652e-01 -2.71371827e-02
-1.99046999e-01 -5.68269268e-02 5.92925906e-01 3.76071960e-01
-7.67133772e-01 6.47467434e-01 -1.80932749e-02 1.26970261e-01
-4.93767619e-01 -9.34673607e-01 -4.98269528e-01 -9.04546559e-01
-3.32946956e-01 5.38947940e-01 7.03099191e-01 -1.16704278e-01
3.91609311e-01 1.03095675e+00 -1.64493430e+00 -1.04363530e-03
2.28153303e-01 7.60602474e-01 2.72021830e-01 5.54576039e-01
-2.76048899e-01 6.35732532e-01 -6.14696324e-01 -1.48716122e-01
7.73843601e-02 -2.63916552e-01 -7.51646101e-01 -4.71117869e-02
1.34891480e-01 -1.56713322e-01 -3.24914366e-01 8.15997601e-01
5.34819901e-01 5.50909221e-01 6.02183700e-01 -9.12801266e-01
1.29770949e-01 2.36654848e-01 -5.18266380e-01 8.47526416e-02
2.92154878e-01 -3.20052594e-01 6.94970489e-01 -4.56470937e-01
8.15288305e-01 7.48426199e-01 5.18045962e-01 5.70630550e-01
-8.18277836e-01 -2.83509880e-01 3.14142346e-01 4.56774950e-01
-1.42357469e+00 -3.17990839e-01 9.94211435e-01 -2.13028938e-01
6.77895963e-01 2.51493275e-01 1.06318963e+00 9.11735356e-01
2.15501100e-01 9.23223019e-01 6.19249523e-01 -5.22943020e-01
3.97716016e-01 -4.76606697e-01 1.37852639e-01 4.61733460e-01
-1.77940056e-02 -9.94627029e-02 -8.21462274e-01 1.50230065e-01
9.84195709e-01 4.16162759e-01 -2.75136799e-01 -5.16768336e-01
-1.24320412e+00 4.73008573e-01 7.81400084e-01 7.29940653e-01
-7.06632614e-01 3.91023248e-01 1.33145303e-01 2.35670000e-01
2.03453094e-01 3.63984145e-02 -7.94366375e-02 -4.51353431e-01
-1.10161102e+00 6.30537653e-03 6.42351389e-01 5.84367335e-01
4.91098315e-01 -1.69117197e-01 1.73452109e-01 6.07355595e-01
8.48288909e-02 2.72192687e-01 6.73597574e-01 -7.61101961e-01
2.63440847e-01 6.37870550e-01 -3.61429490e-02 -1.15610945e+00
-5.53688288e-01 -1.54541627e-01 -1.00626147e+00 5.67779601e-01
6.14645720e-01 1.60247773e-01 -1.02823877e+00 1.52384615e+00
2.10353687e-01 4.30974424e-01 -2.62660712e-01 1.08261752e+00
3.24833572e-01 4.87054586e-01 6.64263070e-02 -1.92435339e-01
1.07070243e+00 -5.39203763e-01 -5.78898132e-01 1.63702771e-01
3.11165899e-01 -5.37408769e-01 4.62872207e-01 5.32174051e-01
-8.63331378e-01 -4.89159763e-01 -1.01896834e+00 4.83006798e-03
-5.71337581e-01 -8.99611562e-02 5.15784502e-01 1.35091752e-01
-1.09969354e+00 8.54588151e-01 -1.39644527e+00 -5.79091787e-01
1.07009433e-01 5.96150815e-01 -5.65613806e-01 7.07776994e-02
-8.95270109e-01 1.03946507e+00 2.63821721e-01 3.82703841e-01
-3.36258441e-01 9.42937285e-03 -5.90990067e-01 -6.94441497e-02
3.68386537e-01 -4.98569399e-01 6.63917601e-01 -7.35150158e-01
-1.48419189e+00 4.04698163e-01 -3.31499547e-01 -4.01723921e-01
5.90359807e-01 -9.47259367e-02 -2.47316703e-01 2.29948521e-01
-1.42489299e-01 6.39908195e-01 1.00913036e+00 -1.04546535e+00
-6.42771244e-01 -6.21295571e-01 6.91999272e-02 3.02254409e-01
-2.94237047e-01 -3.35622877e-01 -6.28823519e-01 -4.62689787e-01
7.16098905e-01 -9.65814710e-01 -1.05057664e-01 3.58836383e-01
-5.38410433e-02 -2.77056605e-01 8.94607306e-01 -7.06254601e-01
1.00045609e+00 -2.11915469e+00 3.89655858e-01 4.66680527e-01
1.39558569e-01 5.75862885e-01 2.30174601e-01 4.66577441e-01
1.29789039e-01 -1.60347432e-01 -2.60525882e-01 -4.31970984e-01
-2.87029505e-01 5.88955700e-01 8.65310282e-02 5.42885303e-01
7.84869567e-02 7.24886775e-01 -7.73554385e-01 -6.08554006e-01
6.56029403e-01 8.74815643e-01 -2.31094539e-01 -3.94563049e-01
-7.00081736e-02 4.80319917e-01 -7.00720906e-01 4.17919278e-01
3.23805511e-01 8.32278505e-02 -6.45710006e-02 -4.89630364e-02
-1.04882188e-01 -2.21275371e-02 -1.45450759e+00 2.07265019e+00
-3.17149282e-01 5.59052229e-01 -1.94484293e-01 -1.20167446e+00
8.95552993e-01 4.23180759e-01 6.95829630e-01 -8.53498459e-01
1.44266725e-01 -2.97103822e-02 -1.07200861e-01 -2.60848731e-01
4.07812268e-01 -6.59681186e-02 2.13237822e-01 4.14419949e-01
-2.25283667e-01 1.56426072e-01 1.30619422e-01 -8.69858414e-02
1.16258168e+00 3.56649280e-01 6.51449040e-02 4.06150758e-01
4.69219714e-01 -6.92858025e-02 2.08172560e-01 4.18789417e-01
-7.37332627e-02 6.96015298e-01 1.05859198e-01 -4.74956483e-01
-7.89054155e-01 -8.39182138e-01 9.93102565e-02 6.38499618e-01
1.24542572e-01 -1.69757783e-01 -4.15625840e-01 -3.41089875e-01
-2.67721862e-01 3.49068105e-01 -6.20348513e-01 -4.00856957e-02
-8.84213567e-01 -3.98845077e-01 2.57864565e-01 4.43022311e-01
6.18844092e-01 -1.18341935e+00 -1.12360287e+00 4.60085481e-01
-8.06115568e-02 -5.86392224e-01 -1.50473431e-01 2.90054709e-01
-1.23988760e+00 -1.06377411e+00 -1.03173685e+00 -6.35398686e-01
7.36002624e-01 2.43448555e-01 6.14467502e-01 2.17575327e-01
-2.56131589e-01 2.81786710e-01 -5.31338513e-01 -2.57806838e-01
1.17607392e-01 1.11561611e-01 -1.44791245e-01 2.28675276e-01
2.77800471e-01 -7.98872054e-01 -5.21718681e-01 1.14581347e-01
-9.62610900e-01 -1.74978793e-01 4.87759471e-01 5.49613535e-01
6.85548663e-01 1.29662335e-01 9.54800546e-02 -2.28136852e-01
4.46770549e-01 -2.54428051e-02 -3.02312106e-01 2.19812289e-01
-2.27729127e-01 2.10746169e-01 2.24981949e-01 -5.15814662e-01
-7.02159286e-01 2.94291079e-01 -6.62081167e-02 -5.78225553e-01
-1.12605400e-01 6.45335793e-01 2.00945344e-02 -1.66060537e-01
2.61836469e-01 3.29540610e-01 1.97085947e-01 -7.03839898e-01
2.86105663e-01 4.73442554e-01 3.23820412e-01 -3.74660701e-01
4.16558981e-01 6.21220231e-01 5.38501799e-01 -1.07976663e+00
-1.85603335e-01 -7.15880334e-01 -1.23113036e+00 -4.84364986e-01
7.62354434e-01 -2.63495743e-01 -7.96642721e-01 4.03047949e-01
-1.14649808e+00 7.32342899e-02 -3.61245006e-01 5.72468340e-01
-5.49668491e-01 4.98837888e-01 -3.63259941e-01 -9.61818874e-01
-5.64747788e-02 -9.06134486e-01 1.01313925e+00 5.21634296e-02
-3.66016746e-01 -1.05856204e+00 1.47545978e-01 -1.20887741e-01
1.86740696e-01 3.82965147e-01 6.39231086e-01 -4.90476310e-01
-6.14679337e-01 -5.60195327e-01 1.87337756e-01 4.62190546e-02
1.53271481e-01 -1.10157125e-01 -6.49643660e-01 -2.18300596e-01
2.45818406e-01 2.89742529e-01 8.60427141e-01 3.80541295e-01
8.38745534e-01 -4.75128368e-03 -5.66781282e-01 2.69399315e-01
1.24957693e+00 2.85342723e-01 8.21460903e-01 4.23646539e-01
6.48962140e-01 7.65352905e-01 5.15351176e-01 6.08342849e-02
1.78816348e-01 9.59981859e-01 5.97864270e-01 6.62597045e-02
-3.66746128e-01 -1.71136603e-01 -8.75264406e-03 9.36893702e-01
-6.61573172e-01 -1.07895121e-01 -8.54626477e-01 5.71258068e-01
-2.14563990e+00 -1.08204079e+00 -1.31792516e-01 2.49386930e+00
4.91883516e-01 8.67204443e-02 -2.11663079e-02 5.89375019e-01
7.08380818e-01 8.55149329e-02 -4.07476008e-01 -1.37905806e-01
7.15824589e-02 2.70122260e-01 4.53287393e-01 4.94264156e-01
-8.67230475e-01 7.62888789e-01 5.52663898e+00 6.86631501e-01
-1.47631693e+00 8.18395019e-02 -3.19822766e-02 -2.90237069e-01
1.27303019e-01 -1.94530785e-02 -4.22244698e-01 5.34508526e-01
6.53099120e-01 3.33714664e-01 2.50135928e-01 4.42006201e-01
2.50524193e-01 -5.44740558e-01 -9.53354001e-01 1.06169343e+00
8.78356174e-02 -9.23621476e-01 1.43945605e-01 1.56563476e-01
2.02788606e-01 -4.47791666e-02 -2.22633898e-01 -1.81416675e-01
-1.65668622e-01 -9.17710006e-01 6.09102011e-01 1.07553279e+00
3.07278544e-01 -6.72932923e-01 6.93224549e-01 7.79528499e-01
-1.09096277e+00 1.96563795e-01 -3.62201929e-01 -2.78669655e-01
2.56265461e-01 3.53874952e-01 -5.55366874e-01 7.72354484e-01
5.68910360e-01 6.56928897e-01 -3.56279373e-01 1.46683788e+00
-2.30139986e-01 1.18884042e-01 -6.48893774e-01 -2.74862051e-01
2.12067500e-01 -2.70617276e-01 5.38367569e-01 6.35112584e-01
4.72773820e-01 2.47262195e-01 5.15336357e-03 3.62153083e-01
4.39948916e-01 -7.34399781e-02 -5.90310276e-01 3.74677241e-01
5.25204837e-02 7.72447526e-01 -1.15329909e+00 -1.99494168e-01
-2.65930206e-01 1.17276704e+00 2.39378557e-01 2.49616355e-01
-4.15016502e-01 -3.56799632e-01 3.64984989e-01 9.77996513e-02
3.88950676e-01 -8.07290494e-01 -1.01872526e-01 -1.01731050e+00
1.66782930e-01 -2.87866086e-01 2.44246453e-01 -7.38297462e-01
-6.91159248e-01 3.25806320e-01 4.29365933e-02 -1.46838391e+00
-4.62147951e-01 -5.19463420e-01 -3.80507469e-01 7.87684619e-01
-1.26853085e+00 -1.14177680e+00 -1.01168759e-01 8.22773814e-01
4.30897415e-01 3.18398595e-01 7.64239609e-01 2.10906044e-01
-1.81722596e-01 7.90260807e-02 1.42103866e-01 3.02306086e-01
5.07483423e-01 -9.47666168e-01 -1.25686906e-03 7.35998929e-01
7.00127065e-01 8.77425551e-01 4.88473713e-01 -7.98238575e-01
-1.42629325e+00 -5.53041101e-01 1.11797345e+00 -3.16302538e-01
3.69964004e-01 4.08504829e-02 -9.79669154e-01 5.22019029e-01
-3.31458867e-01 1.07867429e-02 7.60040581e-02 2.81372666e-02
1.58549249e-01 -1.53901339e-01 -1.00892711e+00 6.88492477e-01
1.17234039e+00 -4.65432972e-01 -6.77053690e-01 1.01053618e-01
9.11051780e-02 -3.74527127e-01 -6.08381748e-01 2.66101182e-01
6.12807453e-01 -8.81356180e-01 8.87330294e-01 -3.31823438e-01
-2.05156893e-01 -4.82730657e-01 1.29895300e-01 -1.26889467e+00
-1.14066720e-01 -9.86677259e-02 -3.62371504e-02 1.00890410e+00
7.70520493e-02 -5.45575917e-01 1.15197194e+00 6.23557568e-01
1.66868657e-01 -4.78493333e-01 -1.47657955e+00 -7.46196628e-01
-2.01981217e-01 -5.99274039e-01 5.33908069e-01 6.37542725e-01
9.45508182e-02 -3.57646495e-02 -3.45922202e-01 1.61108196e-01
5.20987988e-01 7.22147375e-02 5.37278771e-01 -1.45221710e+00
8.65519121e-02 -5.08692145e-01 -8.92837584e-01 -1.09376013e+00
-1.45395566e-02 -6.17093027e-01 -2.73276083e-02 -1.83517826e+00
-1.73130319e-01 -5.39379776e-01 -5.17735422e-01 6.38868392e-01
4.23571467e-01 3.51379693e-01 2.36028880e-01 5.52813113e-01
-4.79203165e-01 3.58130425e-01 1.26229775e+00 -1.84204623e-01
-1.72482923e-01 2.73578405e-01 1.54371053e-01 5.93284667e-01
6.76843166e-01 -3.52282465e-01 -3.41617316e-01 -5.62581599e-01
8.56878702e-03 2.32293278e-01 5.63148379e-01 -1.27818406e+00
9.15095508e-01 6.14340603e-02 5.55379927e-01 -6.34738207e-01
9.41875339e-01 -1.26323140e+00 5.27233601e-01 5.91861844e-01
-8.79357085e-02 -9.42328796e-02 -2.11384058e-01 6.59735262e-01
-4.51973945e-01 -5.32782495e-01 2.70419061e-01 -2.76008189e-01
-9.62975681e-01 4.40166891e-01 -3.25173795e-01 -7.67251253e-01
1.03775764e+00 -6.74504697e-01 2.48908520e-01 -5.68650126e-01
-1.21247649e+00 -4.96856943e-02 6.26229227e-01 5.22765815e-01
7.26426542e-01 -1.17123055e+00 -1.28681853e-01 1.46700382e-01
-2.98582260e-02 7.78229386e-02 1.11572698e-01 8.87475491e-01
-6.39289320e-01 5.92405438e-01 -3.10442299e-01 -6.61648691e-01
-1.34668851e+00 4.48034734e-01 1.31402835e-01 5.65040596e-02
-8.85322928e-01 5.66958785e-01 -4.11418796e-01 9.68813896e-02
4.13353741e-01 -3.54341745e-01 -5.83031774e-01 3.48742932e-01
3.39295983e-01 4.05912280e-01 3.38400066e-01 -8.66806567e-01
-4.73580480e-01 1.16633213e+00 2.58881420e-01 -3.27952355e-01
1.33413839e+00 -8.46039578e-02 -1.50913075e-01 7.75658667e-01
1.06558347e+00 -1.77026346e-01 -1.16836905e+00 -2.34811902e-01
2.16446102e-01 -4.19291556e-01 1.35847375e-01 -4.05462503e-01
-1.04866135e+00 1.06318486e+00 6.73734367e-01 2.95853525e-01
1.03046417e+00 -1.85946330e-01 6.65889621e-01 3.39116573e-01
9.39039886e-01 -9.26166236e-01 -1.84450909e-01 5.10633588e-01
6.89471126e-01 -9.88516092e-01 2.32091732e-02 -1.44608989e-01
-3.06086779e-01 1.32627296e+00 1.23885518e-03 -2.05515355e-01
8.14251125e-01 1.22849958e-03 4.53153849e-02 -1.57560289e-01
-2.46305525e-01 -3.98666829e-01 2.42802769e-01 7.09012926e-01
1.11073792e-01 -4.69490774e-02 -4.20752555e-01 -9.63183418e-02
-2.26956829e-01 1.56546295e-01 7.83855766e-02 1.29961920e+00
-5.73098421e-01 -1.49620092e+00 -4.43682402e-01 2.99463034e-01
-3.45943272e-01 2.89615214e-01 -1.40454993e-01 9.12209213e-01
-3.60388272e-02 7.15774894e-01 2.15671703e-01 -2.61721045e-01
3.30145717e-01 1.77668184e-01 6.57123685e-01 -3.94821525e-01
-1.83389649e-01 -7.43241087e-02 -1.09305061e-01 -5.35145700e-01
-8.48648012e-01 -7.39113390e-01 -1.41383517e+00 -2.75169998e-01
-2.81527877e-01 2.04867437e-01 1.05594134e+00 1.03919280e+00
1.37862757e-01 4.95714277e-01 5.57644852e-02 -1.08774507e+00
-8.30791518e-02 -6.86776400e-01 -6.61432683e-01 2.78034687e-01
3.83568138e-01 -8.87940288e-01 -6.17754273e-02 3.18443514e-02] | [7.871546745300293, 0.32571202516555786] |
29607222-5992-4de6-8129-e6952dd49072 | emergence-of-hierarchical-reference-systems | 2203.13176 | null | https://arxiv.org/abs/2203.13176v2 | https://arxiv.org/pdf/2203.13176v2.pdf | Emergence of hierarchical reference systems in multi-agent communication | In natural language, referencing objects at different levels of specificity is a fundamental pragmatic mechanism for efficient communication in context. We develop a novel communication game, the hierarchical reference game, to study the emergence of such reference systems in artificial agents. We consider a simplified world, in which concepts are abstractions over a set of primitive attributes (e.g., color, style, shape). Depending on how many attributes are combined, concepts are more general ("circle") or more specific ("red dotted circle"). Based on the context, the agents have to communicate at different levels of this hierarchy. Our results show that the agents learn to play the game successfully and can even generalize to novel concepts. To achieve abstraction, they use implicit (omitting irrelevant information) and explicit (indicating that attributes are irrelevant) strategies. In addition, the compositional structure underlying the concept hierarchy is reflected in the emergent protocols, indicating that the need to develop hierarchical reference systems supports the emergence of compositionality. | ['Elia Bruni', 'Marko Duda', 'Xenia Ohmer'] | 2022-03-24 | null | https://aclanthology.org/2022.coling-1.501 | https://aclanthology.org/2022.coling-1.501.pdf | coling-2022-10 | ['novel-concepts'] | ['reasoning'] | [ 1.27923295e-01 3.67246866e-01 3.38950008e-01 -1.15957133e-01
1.73019290e-01 -9.24174011e-01 1.16043746e+00 2.96383739e-01
-4.46698695e-01 7.09095895e-01 2.36490473e-01 1.44045427e-01
-1.83624208e-01 -9.97270167e-01 -4.11959291e-01 -6.44778073e-01
-2.71081805e-01 4.71763581e-01 4.30776387e-01 -8.46945524e-01
2.12403983e-01 2.97179133e-01 -2.06197643e+00 2.33383477e-01
7.00631082e-01 3.17635298e-01 4.64871228e-01 6.04182005e-01
-3.86763036e-01 6.10132337e-01 -9.50707257e-01 -1.06429562e-01
1.52430847e-01 -7.03366816e-01 -9.70770001e-01 -2.31599379e-02
-3.14668834e-01 -1.17513441e-01 3.15944642e-01 1.20656919e+00
4.00556885e-02 1.11023374e-01 6.00867093e-01 -1.26625204e+00
-6.89616024e-01 1.11908031e+00 1.96498096e-01 -3.42561215e-01
7.34140754e-01 1.83019415e-01 1.16633236e+00 -3.05986494e-01
8.10819209e-01 1.24956596e+00 1.28000870e-01 1.19873559e+00
-1.47931850e+00 -3.63808125e-01 5.64067185e-01 -1.01287782e-01
-1.48134267e+00 -2.54000485e-01 7.31318712e-01 -7.04353452e-01
7.39556909e-01 6.10322177e-01 1.13962209e+00 1.06205320e+00
1.37930214e-01 3.44212711e-01 1.35162842e+00 -6.24325871e-01
7.95100093e-01 5.44929922e-01 7.88485184e-02 4.54632193e-01
5.95507503e-01 4.20720041e-01 -6.90506339e-01 -4.55101669e-01
5.65092325e-01 -2.09516793e-01 -3.36399794e-01 -7.41511703e-01
-1.14625907e+00 7.67328203e-01 2.35828593e-01 7.43146896e-01
-3.38972986e-01 7.75660202e-02 9.66537092e-03 6.99069798e-01
-2.25708857e-01 1.16929650e+00 -5.12029946e-01 -1.99567214e-01
7.50602335e-02 2.55492657e-01 1.10388064e+00 1.21953332e+00
8.67963016e-01 -2.93910921e-01 3.41882050e-01 6.52654707e-01
2.74716198e-01 3.67040604e-01 4.80241954e-01 -1.51374531e+00
-2.72126496e-01 8.16891253e-01 3.15348119e-01 -7.08965838e-01
-5.18244028e-01 -3.14773381e-01 -5.44186413e-01 4.27415311e-01
1.82877958e-01 -6.91981241e-02 -2.49645710e-01 2.47305369e+00
3.05295736e-01 -4.98227596e-01 5.62005103e-01 6.70035541e-01
5.06584942e-01 4.44860697e-01 2.76156753e-01 -4.30443972e-01
1.48865783e+00 -3.75632972e-01 -5.04850268e-01 1.19013526e-01
1.03117323e+00 -3.25646192e-01 1.27773821e+00 2.77236491e-01
-1.03520918e+00 -1.95202544e-01 -9.56753552e-01 2.72342116e-01
-6.97828770e-01 -7.04662800e-01 7.49303401e-01 7.96735346e-01
-1.38934696e+00 4.73880827e-01 -3.64506632e-01 -9.05738473e-01
-2.48448402e-01 5.27047813e-01 -2.62754709e-01 4.86638904e-01
-1.12796128e+00 7.57433534e-01 5.68388283e-01 -2.99059242e-01
-8.49654913e-01 -1.68339655e-01 -5.88303387e-01 9.94576812e-02
3.74226272e-01 -1.15763164e+00 1.29082036e+00 -1.51034546e+00
-2.12526202e+00 9.35952783e-01 4.10510629e-01 -1.68727726e-01
2.50469685e-01 1.60618722e-01 -2.43660226e-01 1.75604984e-01
7.09702354e-03 9.10472572e-01 5.69888055e-01 -1.79275358e+00
-1.02484870e+00 -3.33689988e-01 1.04855621e+00 3.94560486e-01
-1.49728552e-01 1.49111569e-01 2.39263728e-01 -3.34401637e-01
1.93777025e-01 -1.12018669e+00 -1.90270003e-02 -2.09045693e-01
-1.43259972e-01 -3.74086440e-01 5.16491413e-01 3.16432625e-01
1.00573325e+00 -2.32909513e+00 7.13141203e-01 1.18278094e-01
5.51651359e-01 -3.27400893e-01 -2.08214015e-01 6.36656880e-01
3.67880911e-01 4.42087382e-01 -2.72128195e-01 1.49392769e-01
1.55667111e-01 6.14836454e-01 1.75281450e-01 8.10402930e-02
-1.05908491e-01 4.66383308e-01 -1.00163591e+00 -3.38447541e-01
4.75246385e-02 3.66946012e-01 -1.00271904e+00 1.76409498e-01
-4.97255862e-01 7.98137844e-01 -8.07308853e-01 1.15733914e-01
1.79355010e-01 -2.70762891e-01 5.06348431e-01 4.28831786e-01
-5.39045870e-01 2.99253285e-01 -1.07337594e+00 1.39454699e+00
-6.50256991e-01 3.42826933e-01 3.47405642e-01 -5.73888958e-01
6.87099099e-01 3.61964136e-01 8.04367438e-02 -3.28278333e-01
3.15842390e-01 3.10918689e-01 7.53233969e-01 -3.94276410e-01
2.81138927e-01 -7.01209009e-02 -3.47393483e-01 6.69376194e-01
-3.98415864e-01 -7.85305023e-01 2.03781456e-01 3.14589798e-01
7.55859554e-01 1.17437944e-01 5.49131811e-01 -9.24987018e-01
8.32351744e-01 -1.84982762e-01 5.62118351e-01 1.10473192e+00
-2.40550444e-01 2.36388613e-02 4.69307035e-01 -5.90006828e-01
-7.88234413e-01 -1.07629323e+00 1.79767773e-01 1.45213699e+00
6.34833872e-01 -6.12798572e-01 -1.10486889e+00 -1.39133722e-01
-4.23984617e-01 9.30595219e-01 -7.03508615e-01 -2.16998145e-01
-6.15302145e-01 -2.36925229e-01 5.66532537e-02 9.39300284e-02
5.00738442e-01 -1.37580621e+00 -1.51913738e+00 2.54846215e-01
-1.30939052e-01 -8.37562382e-01 -1.30177215e-01 9.96885896e-02
-5.93612909e-01 -7.85483599e-01 -8.89888778e-02 -5.81493676e-01
6.43979609e-01 8.24888274e-02 1.25117230e+00 8.13955903e-01
2.08117187e-01 8.73563826e-01 -6.00439429e-01 -3.39314252e-01
-7.95889318e-01 5.74632622e-02 2.17159182e-01 -7.82307088e-02
1.24578260e-01 -7.83246338e-01 -5.38539827e-01 1.32927731e-01
-9.76543367e-01 3.41574103e-01 1.32607460e-01 6.56854212e-01
-1.26665071e-01 -7.87712410e-02 1.85761824e-01 -6.88511014e-01
9.73264337e-01 -1.49411708e-01 -5.70795715e-01 2.16296718e-01
-1.69370562e-01 3.32480311e-01 8.01588356e-01 -6.25755727e-01
-9.21000302e-01 -2.55534589e-01 3.46464574e-01 7.29066610e-01
-4.02154565e-01 2.24525735e-01 -3.09945792e-01 6.23050611e-04
6.44032478e-01 3.99624735e-01 -1.22082114e-01 -1.55310899e-01
5.49951911e-01 6.10462427e-01 -6.14266023e-02 -1.40522635e+00
4.60641801e-01 5.01717925e-01 3.62579897e-02 -1.06966913e+00
-1.90893829e-01 2.35919595e-01 -6.54607058e-01 -1.18149780e-01
8.93835664e-01 -4.53808457e-01 -1.08324242e+00 4.56327438e-01
-1.16067910e+00 -4.98257607e-01 -5.32222092e-01 3.55916411e-01
-9.61692393e-01 2.43678898e-01 -5.18548071e-01 -1.01373231e+00
4.03581783e-02 -1.33200276e+00 7.42794037e-01 2.41245359e-01
-7.01009274e-01 -6.99278057e-01 -2.37862274e-01 -3.59519392e-01
7.10389853e-01 -1.43307120e-01 1.30326760e+00 -7.72598326e-01
-8.42367768e-01 4.42194015e-01 2.61259973e-01 -3.16196263e-01
4.24763858e-01 2.74992526e-01 -5.30954182e-01 -2.53175974e-01
1.89283460e-01 3.13232206e-02 8.89798030e-02 -9.51485336e-02
6.41292512e-01 -5.17991543e-01 -3.46035868e-01 2.76729673e-01
1.07757843e+00 8.37745607e-01 3.40475976e-01 4.99778003e-01
1.15671113e-01 1.03460360e+00 -9.64662507e-02 3.59474361e-01
6.52826309e-01 8.82276952e-01 1.01482019e-01 4.47431624e-01
5.07646166e-02 9.11456943e-02 2.47997910e-01 8.94578636e-01
-3.53288800e-01 5.33451810e-02 -7.24402905e-01 2.53262311e-01
-1.59892786e+00 -9.10877585e-01 1.60428703e-01 2.21532178e+00
1.13645983e+00 1.30290344e-01 8.08670521e-02 -1.97345480e-01
8.03863883e-01 -3.03889245e-01 -1.41098484e-01 -7.16707349e-01
-1.58359662e-01 -2.04062700e-01 -3.33120733e-01 9.13026273e-01
-3.60536575e-01 1.15946841e+00 6.81741858e+00 2.64596254e-01
-9.85620141e-01 1.37258759e-02 9.44255367e-02 3.96839470e-01
-6.46990955e-01 2.33838528e-01 -4.44054723e-01 5.24337590e-01
5.54447293e-01 -2.76632786e-01 1.00380933e+00 3.85955215e-01
-1.92554191e-01 -1.80190578e-01 -1.53372037e+00 5.19680977e-01
-2.72616029e-01 -1.13600910e+00 5.35196304e-01 -4.65761758e-02
3.74762923e-01 -3.70318651e-01 -1.12333342e-01 2.76647568e-01
9.47583854e-01 -6.11720622e-01 1.01926839e+00 2.51272738e-01
4.06421214e-01 -2.56134897e-01 3.12542766e-01 6.12783015e-01
-1.15805447e+00 -3.38990480e-01 5.08152181e-03 -5.93599617e-01
-1.54162794e-01 -4.51862693e-01 -9.71948132e-02 2.61010289e-01
6.69418097e-01 -1.06483936e-01 -2.31985629e-01 3.92941147e-01
-3.85682821e-01 -1.24922156e-01 -5.93578339e-01 -6.32595778e-01
2.49514286e-03 -5.33756554e-01 6.85037255e-01 7.69457757e-01
2.49465823e-01 7.31827140e-01 1.46014959e-01 1.07283711e+00
2.97675282e-01 -6.80087192e-04 -6.95396900e-01 2.33341411e-01
7.31164098e-01 7.23587096e-01 -7.80415297e-01 -5.99101484e-01
-4.82088387e-01 6.13165081e-01 1.68418318e-01 4.12282616e-01
-4.32953060e-01 -1.08369112e-01 8.63270342e-01 -3.87712731e-03
-1.26756713e-01 -4.07207787e-01 -1.64813250e-01 -1.10857141e+00
-5.64137697e-01 -1.00093198e+00 2.48715296e-01 -8.02137017e-01
-1.10215735e+00 8.40283275e-01 4.16546583e-01 -8.55974615e-01
-1.56967223e-01 -3.24966460e-01 -3.07187408e-01 4.98749375e-01
-1.10008562e+00 -9.75857496e-01 -2.25978687e-01 7.35474825e-01
3.88048083e-01 -2.26494595e-01 1.24446881e+00 -3.38334560e-01
-1.66136712e-01 1.82123378e-01 -3.72593999e-01 -8.93938020e-02
-2.05792226e-02 -1.18767464e+00 8.77402350e-02 4.10876304e-01
7.73859993e-02 1.27953506e+00 1.00395775e+00 -3.54809821e-01
-1.15758061e+00 -3.41906995e-01 8.82532954e-01 -5.59566140e-01
6.61031723e-01 -5.59801102e-01 -7.25393414e-01 6.40486658e-01
4.94803399e-01 -6.15283966e-01 8.09877157e-01 1.23423770e-01
-6.20117426e-01 4.97516012e-03 -1.31558132e+00 1.38610995e+00
1.51841652e+00 -5.14888883e-01 -1.00895488e+00 1.86961200e-02
1.15708840e+00 2.42192358e-01 -4.95994955e-01 2.67319113e-01
6.23255372e-01 -1.27865195e+00 5.32011807e-01 -6.47622347e-01
-1.28986454e-02 -5.64894557e-01 -4.83479351e-01 -1.40485823e+00
-5.69482505e-01 -1.00547755e+00 3.14118087e-01 9.36016917e-01
4.63891178e-01 -1.41578937e+00 2.03386635e-01 7.06263125e-01
1.65154710e-02 -7.68298805e-02 -9.63790953e-01 -6.45985186e-01
3.50881100e-01 3.79187129e-02 1.10229552e+00 9.21711385e-01
6.25087023e-01 5.08355618e-01 1.45028159e-01 8.54857266e-02
5.64565301e-01 3.48820448e-01 7.10845411e-01 -1.81732416e+00
-4.23683584e-01 -8.92062962e-01 -3.60394686e-01 -9.88570631e-01
1.56773433e-01 -5.30803978e-01 -9.86417569e-03 -1.16684079e+00
1.90583125e-01 -6.46989286e-01 -2.28889063e-02 2.23578438e-01
2.03725427e-01 -2.71023750e-01 4.85099554e-01 4.93275493e-01
-5.06188989e-01 4.33618456e-01 1.10538840e+00 -4.23242114e-02
-5.10282934e-01 -2.77059227e-01 -8.95548642e-01 9.19298232e-01
1.01989615e+00 -3.94962132e-01 -4.92952526e-01 -4.47243422e-01
6.37843013e-01 -7.24722743e-02 6.95874691e-02 -9.13467765e-01
2.70250171e-01 -6.02600932e-01 -3.89216661e-01 2.82674074e-01
5.21716356e-01 -1.25170887e+00 3.81224364e-01 9.43804026e-01
-7.79420972e-01 1.12281442e-01 -1.93948105e-01 2.66872913e-01
2.03981727e-01 -1.95970625e-01 6.95549369e-01 -4.72674102e-01
-6.08821929e-01 -2.82185584e-01 -8.87727499e-01 6.69147670e-02
1.19847393e+00 -4.78952408e-01 -4.67990249e-01 -3.40567052e-01
-9.48486149e-01 4.75180037e-02 9.26031590e-01 4.06927675e-01
2.88822621e-01 -9.84143436e-01 -4.92531627e-01 3.42841595e-01
5.20413578e-01 -1.21490434e-01 -2.66177416e-01 3.49539548e-01
-5.20661533e-01 9.98545289e-02 -5.36937714e-01 -7.12276936e-01
-9.39208686e-01 7.15454578e-01 5.76846361e-01 4.59206283e-01
-6.29936993e-01 7.67878115e-01 9.71368730e-01 -3.36742878e-01
-3.18878926e-02 -4.93086636e-01 -3.16370517e-01 -2.40008458e-01
3.06453168e-01 -2.58663148e-01 -5.86766601e-01 -7.96836257e-01
-3.66940707e-01 9.49455619e-01 1.27735496e-01 -3.73217165e-01
1.04837441e+00 -5.11545002e-01 -6.20075345e-01 6.01202011e-01
4.42903429e-01 2.05212310e-01 -8.94379735e-01 -1.46463722e-01
1.58134043e-01 -4.10814345e-01 -6.18322670e-01 -5.53239822e-01
-3.90762836e-01 3.89861226e-01 1.85538530e-01 1.02310169e+00
9.54042494e-01 4.13257807e-01 -9.12178457e-02 6.10189557e-01
1.14234054e+00 -1.19190025e+00 1.51956573e-01 7.01103866e-01
9.84512627e-01 -5.79659462e-01 -2.52719104e-01 -7.20000327e-01
-4.38769817e-01 9.68885303e-01 5.28062224e-01 8.58157873e-02
5.46237648e-01 4.00881022e-01 8.87609422e-02 -2.65203118e-01
-1.05566239e+00 -3.80277187e-01 -5.07600188e-01 1.13925827e+00
2.73548365e-01 4.37409431e-01 -7.03679323e-01 3.89325500e-01
-6.71864092e-01 -4.37138736e-01 9.93826628e-01 8.89837325e-01
-7.10764706e-01 -1.23115563e+00 -2.05985487e-01 -3.45539540e-01
-5.25798127e-02 -2.16228161e-02 -7.85112679e-01 9.65389550e-01
3.91111702e-01 1.25677276e+00 2.20411569e-01 -1.73259690e-01
2.98368931e-01 -7.35410303e-02 6.77613795e-01 -9.19751346e-01
-7.31649578e-01 -3.35855126e-01 3.52190388e-03 -3.11232120e-01
-7.65626967e-01 -5.00248909e-01 -1.50941622e+00 -3.56323123e-01
-1.95435047e-01 6.49242043e-01 3.74349236e-01 8.25514019e-01
3.15693200e-01 2.95991689e-01 4.60204214e-01 -4.19584423e-01
-3.30718249e-01 -6.52863562e-01 -4.08833325e-01 6.29418492e-01
4.53048170e-01 -7.50006855e-01 -6.25010967e-01 9.34367105e-02] | [4.331182956695557, 1.4431917667388916] |
14cb8801-07a7-4ed6-a996-4b0302a75169 | ssp-self-supervised-post-training-for | 2307.00569 | null | https://arxiv.org/abs/2307.00569v1 | https://arxiv.org/pdf/2307.00569v1.pdf | SSP: Self-Supervised Post-training for Conversational Search | Conversational search has been regarded as the next-generation search paradigm. Constrained by data scarcity, most existing methods distill the well-trained ad-hoc retriever to the conversational retriever. However, these methods, which usually initialize parameters by query reformulation to discover contextualized dependency, have trouble in understanding the dialogue structure information and struggle with contextual semantic vanishing. In this paper, we propose \fullmodel (\model) which is a new post-training paradigm with three self-supervised tasks to efficiently initialize the conversational search model to enhance the dialogue structure and contextual semantic understanding. Furthermore, the \model can be plugged into most of the existing conversational models to boost their performance. To verify the effectiveness of our proposed method, we apply the conversational encoder post-trained by \model on the conversational search task using two benchmark datasets: CAsT-19 and CAsT-20. Extensive experiments that our \model can boost the performance of several existing conversational search methods. Our source code is available at \url{https://github.com/morecry/SSP}. | ['Rui Yan', 'Ji-Rong Wen', 'Zhao Cao', 'Xiaolong Wu', 'Shen Gao', 'Quan Tu'] | 2023-07-02 | null | null | null | null | ['conversational-search'] | ['natural-language-processing'] | [-6.90948665e-02 3.25692385e-01 -3.41475397e-01 -4.94836897e-01
-8.75585020e-01 -5.63477993e-01 9.90836918e-01 -2.47911006e-01
-3.56518984e-01 8.41042399e-01 6.88668787e-01 -3.62308919e-01
7.25375414e-02 -5.42614818e-01 -3.05176854e-01 -3.42708081e-01
4.73014832e-01 8.75663400e-01 2.70519674e-01 -8.38607728e-01
4.02060211e-01 -4.79053199e-01 -1.41073000e+00 5.75510681e-01
1.37382042e+00 8.40958595e-01 6.45193398e-01 5.43151259e-01
-4.32618469e-01 5.98454833e-01 -6.50626183e-01 -5.28094292e-01
-1.61969960e-01 -7.17265368e-01 -1.43458128e+00 -3.90921861e-01
-1.91814408e-01 -4.20545489e-01 -4.07610297e-01 8.32857490e-01
7.52942741e-01 4.17246401e-01 2.57353008e-01 -1.06049883e+00
-8.28902721e-01 9.68591034e-01 4.39407118e-02 1.79315448e-01
7.97813535e-01 6.20637788e-03 1.24507117e+00 -9.75019217e-01
7.17713535e-01 1.42558908e+00 2.07412109e-01 8.36072624e-01
-7.34821200e-01 -7.42588282e-01 2.20306635e-01 5.82200885e-01
-1.20596528e+00 -5.80703914e-01 9.41553175e-01 1.01162270e-02
1.10825849e+00 4.08199281e-01 4.54394072e-01 1.38534582e+00
-4.09175724e-01 1.09555662e+00 9.01902497e-01 -5.06155491e-01
-1.29536316e-01 2.35563025e-01 3.20642292e-01 5.82539558e-01
-4.45518672e-01 7.76830465e-02 -8.01444173e-01 -3.13776761e-01
3.09316754e-01 -1.33410215e-01 -4.67186421e-01 5.20998091e-02
-9.14956450e-01 8.21063340e-01 4.28124249e-01 2.70098150e-01
-5.80584072e-02 -6.32952303e-02 4.88180846e-01 4.82141912e-01
5.58169544e-01 6.94346666e-01 -3.42777014e-01 -4.30221766e-01
-4.19919401e-01 1.67513236e-01 1.21644974e+00 1.13651228e+00
7.94607401e-01 -6.65994108e-01 -2.96737909e-01 1.49540889e+00
2.84167260e-01 3.60846847e-01 7.65490711e-01 -9.52517629e-01
5.73222399e-01 8.15316737e-01 7.56080970e-02 -7.09929943e-01
-1.54776379e-01 -3.43162626e-01 -6.19893909e-01 -8.14267576e-01
-9.72726345e-02 -2.10446820e-01 -4.60636050e-01 1.71881819e+00
2.88603932e-01 1.38417616e-01 3.48076969e-01 9.60426509e-01
1.38413751e+00 7.90964782e-01 -6.59288615e-02 -2.21474782e-01
1.33896041e+00 -1.74166977e+00 -8.52087736e-01 -2.87010014e-01
6.66463792e-01 -9.15799141e-01 1.36961758e+00 -1.15564344e-02
-9.60381448e-01 -5.51369011e-01 -8.18483174e-01 -2.78925419e-01
-4.43343341e-01 3.40754017e-02 6.90786064e-01 2.89620429e-01
-1.04299545e+00 2.09044904e-01 -4.73837167e-01 -6.08804166e-01
-1.53566942e-01 5.79686426e-02 1.30809397e-01 -9.21118930e-02
-1.87632668e+00 9.17567432e-01 4.07040775e-01 2.89158195e-01
-8.32362771e-01 -2.27566689e-01 -5.66597760e-01 1.32046565e-01
6.60634220e-01 -7.51287639e-01 1.62205386e+00 -6.06884122e-01
-1.72720873e+00 7.63084233e-01 -5.18936753e-01 -6.69493750e-02
1.74346626e-01 -3.49851519e-01 -3.01045299e-01 1.72446612e-02
2.00461730e-01 6.30826831e-01 3.22544396e-01 -1.24084067e+00
-5.15261412e-01 -1.17624745e-01 5.12851238e-01 8.14051211e-01
-3.28974575e-01 4.80950214e-02 -9.92512286e-01 -3.57105553e-01
-1.63726509e-01 -1.08832741e+00 -8.73547718e-02 -6.69656098e-01
-5.51837385e-01 -8.77091885e-01 6.22310936e-01 -6.00765288e-01
1.58176589e+00 -1.90065408e+00 2.81150043e-01 -1.98240250e-01
-1.10774577e-01 4.11543071e-01 -1.85304463e-01 1.09892964e+00
3.07740211e-01 1.61222935e-01 -1.01719439e-01 -4.26963091e-01
1.75133094e-01 2.61079550e-01 -3.53890359e-01 -2.35710278e-01
-2.47494638e-01 1.18971276e+00 -1.15017831e+00 -6.45308912e-01
-3.26934922e-03 2.48341504e-02 -5.36046565e-01 7.73082376e-01
-5.57286263e-01 5.52362025e-01 -7.48436987e-01 4.96331155e-01
3.06097209e-01 -5.17855346e-01 3.16429853e-01 7.50712082e-02
2.68686026e-01 7.46820807e-01 -6.38904631e-01 2.25021076e+00
-6.77616298e-01 3.32484365e-01 1.65038362e-01 -9.45425749e-01
8.98928821e-01 3.72532874e-01 1.95703447e-01 -1.09781229e+00
-1.07529853e-02 2.49830037e-01 -2.98205733e-01 -8.21954668e-01
8.12120318e-01 2.68388748e-01 -3.00564677e-01 8.85147989e-01
-4.41707745e-02 -2.54295290e-01 1.06205903e-01 3.61873209e-01
9.75552738e-01 9.54868942e-02 -6.07343502e-02 -1.19043082e-01
8.46087635e-01 9.81848910e-02 4.05112118e-01 8.02466333e-01
-2.33780220e-02 2.25330353e-01 1.92076296e-01 -6.41582236e-02
-5.96494973e-01 -7.22549498e-01 7.44106099e-02 1.65022254e+00
6.31576002e-01 -6.86935246e-01 -7.14636087e-01 -8.60302091e-01
-2.78431684e-01 6.88993216e-01 -2.25540206e-01 -3.68202448e-01
-6.11715555e-01 -6.44268274e-01 7.23684192e-01 2.50146478e-01
9.52696264e-01 -1.34173417e+00 -9.22436491e-02 1.73559517e-01
-9.48300183e-01 -1.04149747e+00 -6.48062885e-01 -5.67153171e-02
-4.35594797e-01 -1.08585036e+00 -5.43788075e-01 -1.04197979e+00
3.89106721e-01 3.43342304e-01 1.33789623e+00 5.90005636e-01
9.60806981e-02 5.12251556e-01 -8.85011673e-01 -9.67538804e-02
-3.96624237e-01 6.65468514e-01 -4.22692597e-01 -4.49411064e-01
4.83751059e-01 -5.63516319e-01 -1.01660359e+00 7.18565702e-01
-7.13291764e-01 2.18798295e-01 3.97612303e-01 1.20391178e+00
3.55122201e-02 -3.31157655e-01 1.03464663e+00 -9.85438943e-01
1.53633797e+00 -6.29747510e-01 -8.08016807e-02 5.68033338e-01
-1.01405013e+00 1.29195601e-01 3.47218722e-01 -1.58974260e-01
-1.49538875e+00 -4.13080394e-01 -3.17968637e-01 4.05893512e-02
6.77819699e-02 9.04874980e-01 2.92991009e-03 4.20512140e-01
5.34478843e-01 4.65160072e-01 -5.73677979e-02 -7.14072466e-01
6.66364610e-01 1.15015697e+00 4.86663818e-01 -9.91087615e-01
2.93679774e-01 9.17927176e-02 -8.66467059e-01 -4.11370814e-01
-9.61296558e-01 -7.72943199e-01 -3.21858406e-01 -2.13510558e-01
6.43895864e-01 -7.97191620e-01 -8.23698461e-01 3.15909922e-01
-1.34603262e+00 -4.66589779e-01 2.41188869e-01 2.07733527e-01
-3.81564170e-01 5.17182946e-01 -6.53869689e-01 -7.53401279e-01
-7.44150221e-01 -1.08364940e+00 1.09758890e+00 4.11526293e-01
-1.59306332e-01 -1.07439268e+00 2.24194780e-01 9.81490374e-01
6.21575058e-01 -5.12154818e-01 9.41902757e-01 -1.04330659e+00
-4.05970603e-01 3.35946493e-02 -2.22245872e-01 2.30963632e-01
4.67386954e-02 -5.72000444e-01 -9.82677221e-01 -2.16472283e-01
-1.23239405e-01 -7.50625670e-01 8.31187367e-01 -3.12256873e-01
1.01886213e+00 -4.17435378e-01 -5.76856971e-01 1.93902105e-01
9.54452038e-01 3.97866547e-01 4.27061230e-01 3.44827771e-01
3.07956517e-01 6.35854185e-01 7.70612478e-01 2.05111653e-01
8.60242784e-01 9.18816924e-01 2.12961346e-01 2.29421049e-01
-1.98206156e-01 -5.50628543e-01 2.11373925e-01 1.31556344e+00
3.21275704e-02 -4.27978307e-01 -8.51883948e-01 4.93994802e-01
-2.31732154e+00 -7.99289048e-01 4.08611596e-01 1.65134895e+00
1.45108271e+00 -1.13510378e-01 -2.09425479e-01 -6.03911281e-01
7.17499733e-01 3.78060162e-01 -6.81047618e-01 -3.74185383e-01
5.21443412e-02 1.05982637e-02 -2.88852036e-01 8.80628407e-01
-7.15983093e-01 1.31414199e+00 5.73811245e+00 8.59052896e-01
-7.76364386e-01 1.86915860e-01 3.61763537e-01 7.45024160e-02
-4.91832256e-01 3.65967661e-01 -9.58539009e-01 5.71142375e-01
7.64371812e-01 -6.23434842e-01 6.71210051e-01 9.35060680e-01
1.82179548e-02 -5.94674349e-02 -1.14525795e+00 9.31446016e-01
3.30149949e-01 -1.23434663e+00 8.47028643e-02 -3.33765715e-01
5.33768415e-01 -8.20208415e-02 -2.24234566e-01 1.01830339e+00
5.91466069e-01 -1.02712452e+00 -7.33945370e-02 4.83506858e-01
4.97181356e-01 -3.71215403e-01 9.06238437e-01 6.77648246e-01
-1.08221841e+00 -2.73430292e-02 -2.49138221e-01 5.24682924e-02
3.22668552e-01 1.61767006e-01 -1.12172556e+00 7.07068145e-01
6.32796347e-01 5.97475588e-01 -5.83610296e-01 6.40582860e-01
-3.41138631e-01 4.09954250e-01 -2.09042951e-01 -5.20352721e-01
3.66279632e-01 -2.54558980e-01 5.01001477e-01 1.25439858e+00
-3.77828218e-02 2.83251494e-01 3.65737915e-01 7.94366241e-01
-2.25393996e-01 2.34574139e-01 -3.68603736e-01 -3.69021762e-03
8.79231751e-01 9.59853530e-01 -1.22874074e-01 -4.02216762e-01
-2.41690591e-01 1.03736413e+00 4.59883600e-01 5.75984955e-01
-6.59470797e-01 -4.88218725e-01 3.03326041e-01 -2.98366696e-01
-1.15615755e-01 -3.14171538e-02 2.62997150e-01 -1.27708912e+00
2.20098242e-01 -1.17032683e+00 6.21585548e-01 -9.52525377e-01
-1.25192010e+00 8.08503211e-01 1.74008638e-01 -7.35886753e-01
-6.87996566e-01 -2.73778047e-02 -6.38289034e-01 8.16360533e-01
-1.55410671e+00 -1.20292222e+00 -4.53938007e-01 6.54648244e-01
1.15135682e+00 -1.54118106e-01 9.37694550e-01 3.57092232e-01
-4.91274506e-01 6.01514518e-01 -1.90252699e-02 1.84009507e-01
1.03880775e+00 -1.05944514e+00 1.51191190e-01 2.95764506e-01
-1.45935133e-01 9.14886653e-01 4.91851330e-01 -5.98874271e-01
-1.38148832e+00 -5.78748047e-01 1.02182102e+00 -4.73801076e-01
6.31009161e-01 -4.32618350e-01 -8.88747692e-01 5.97689092e-01
8.83756757e-01 -7.59453297e-01 7.34483600e-01 5.54556966e-01
-9.06381980e-02 2.19354336e-03 -5.18713176e-01 7.08783925e-01
1.32510924e+00 -8.32940519e-01 -1.06176877e+00 6.18909538e-01
1.15491283e+00 -5.39118767e-01 -8.15967083e-01 4.72882301e-01
3.68866116e-01 -8.97232413e-01 9.38790619e-01 -5.62565982e-01
3.70088816e-01 1.67572856e-01 1.02885952e-02 -1.47848928e+00
2.59060353e-01 -9.62558985e-01 -5.44493161e-02 1.31306636e+00
6.65281832e-01 -7.48366058e-01 6.66718066e-01 6.86828136e-01
-2.99243003e-01 -8.88565421e-01 -8.89617026e-01 -4.92842287e-01
2.69917324e-02 -1.60209656e-01 7.97750175e-01 9.27873433e-01
6.72747791e-01 1.03027415e+00 -2.84959376e-01 -2.65932590e-01
1.01560146e-01 5.86385131e-01 9.34563994e-01 -9.83829379e-01
-2.48327494e-01 -5.49229681e-01 5.04651248e-01 -1.92524135e+00
4.20449644e-01 -1.04007578e+00 1.83505192e-01 -1.68103552e+00
3.54186535e-01 -6.12786531e-01 -1.65348679e-01 3.26350808e-01
-5.00313461e-01 -4.29850936e-01 -1.06525071e-01 4.49229509e-01
-1.11113334e+00 1.10136950e+00 1.53671384e+00 -1.40797988e-01
-3.32087785e-01 1.20798284e-02 -8.96643221e-01 3.48870397e-01
7.69939899e-01 -2.06091151e-01 -8.65572631e-01 -5.14469206e-01
3.96769315e-01 4.21150655e-01 1.47857219e-01 -2.92401731e-01
8.49270344e-01 -1.20854497e-01 -4.19906139e-01 -6.70275033e-01
6.10972881e-01 -4.27406818e-01 -2.84991413e-01 1.57068849e-01
-8.14753771e-01 1.18619055e-01 -5.60399704e-02 5.57585061e-01
-6.03049636e-01 -3.83360714e-01 1.40145019e-01 -3.30097377e-01
-8.39892983e-01 7.37920925e-02 -8.84790868e-02 3.83271128e-01
4.87280369e-01 3.16874087e-01 -8.53087425e-01 -9.08526421e-01
-7.03985453e-01 9.58696723e-01 3.39906029e-02 8.98415565e-01
4.91004497e-01 -1.17031479e+00 -6.44583285e-01 -1.75257698e-01
3.05402189e-01 2.35828027e-01 2.00502902e-01 5.87119699e-01
-1.23188801e-01 8.45094681e-01 2.71842510e-01 -4.72639591e-01
-1.19491494e+00 4.54936236e-01 3.13325018e-01 -4.54887152e-01
-4.04038429e-01 8.61118376e-01 5.24988919e-02 -9.10066247e-01
4.47220057e-01 9.28867981e-02 -4.02957678e-01 -7.79936928e-03
1.93131760e-01 6.56351894e-02 -2.50183195e-01 -1.91094592e-01
-1.74427211e-01 2.66993672e-01 -3.31423104e-01 -2.88363904e-01
1.05204332e+00 -6.26678407e-01 -3.91301572e-01 1.56166643e-01
1.23454869e+00 -1.85988203e-01 -6.36059463e-01 -7.10886300e-01
1.77845314e-01 -1.82039067e-01 -2.12118641e-01 -1.03063321e+00
-7.29313493e-01 6.50058508e-01 1.79687470e-01 2.94247270e-01
9.65407670e-01 2.88366050e-01 1.04917574e+00 1.02912593e+00
3.01405847e-01 -1.33479297e+00 3.68806750e-01 7.79498756e-01
1.18084097e+00 -1.31910038e+00 -1.82621688e-01 -4.62961346e-01
-7.84904599e-01 7.55061328e-01 1.02032292e+00 2.46231198e-01
7.22330391e-01 -2.30262920e-01 3.03308189e-01 -4.33680475e-01
-1.15029311e+00 -3.71774197e-01 7.90456459e-02 1.76667392e-01
3.76207054e-01 -2.57330090e-01 -6.95513964e-01 7.64521420e-01
-4.93846089e-01 -7.44542032e-02 6.65044188e-02 9.08594012e-01
-3.63072127e-01 -1.45521092e+00 2.48525247e-01 1.10650234e-01
-2.69058626e-02 -4.88848835e-01 -8.17444503e-01 5.10794759e-01
-4.08339322e-01 1.47449481e+00 -2.62480199e-01 -5.28286755e-01
2.86670178e-01 4.47994351e-01 -1.27887502e-01 -7.14828968e-01
-7.29553223e-01 8.24558549e-03 7.63172686e-01 -5.88422060e-01
-4.94258642e-01 -2.09846884e-01 -1.01674235e+00 -1.50773466e-01
-6.41196787e-01 9.98800993e-01 5.13576031e-01 9.13420081e-01
6.66568398e-01 2.52204537e-01 7.50223279e-01 -1.59564182e-01
-5.92126250e-01 -1.38582313e+00 8.27366486e-02 4.36909348e-01
-1.41320884e-01 -6.19683504e-01 -4.32404310e-01 -1.23175010e-01] | [12.17314338684082, 7.847255706787109] |
55b2cbc5-d201-4da6-8340-4281cb151978 | quality-question-answering-with-long-input | 2112.08608 | null | https://arxiv.org/abs/2112.08608v2 | https://arxiv.org/pdf/2112.08608v2.pdf | QuALITY: Question Answering with Long Input Texts, Yes! | To enable building and testing models on long-document comprehension, we introduce QuALITY, a multiple-choice QA dataset with context passages in English that have an average length of about 5,000 tokens, much longer than typical current models can process. Unlike in prior work with passages, our questions are written and validated by contributors who have read the entire passage, rather than relying on summaries or excerpts. In addition, only half of the questions are answerable by annotators working under tight time constraints, indicating that skimming and simple search are not enough to consistently perform well. Our baseline models perform poorly on this task (55.4%) and significantly lag behind human performance (93.5%). | ['Samuel R. Bowman', 'He He', 'Jana Thompson', 'Johnny Ma', 'Vishakh Padmakumar', 'Angelica Chen', 'Jason Phang', 'Nikita Nangia', 'Nitish Joshi', 'Alicia Parrish', 'Richard Yuanzhe Pang'] | 2021-12-16 | null | https://aclanthology.org/2022.naacl-main.391 | https://aclanthology.org/2022.naacl-main.391.pdf | naacl-2022-7 | ['multiple-choice-qa'] | ['natural-language-processing'] | [-3.97959277e-02 3.45293522e-01 3.35544348e-02 -2.55306005e-01
-1.83811164e+00 -1.20727432e+00 4.93379176e-01 6.07706547e-01
-8.39256704e-01 1.23346162e+00 7.83747673e-01 -6.37463689e-01
2.64559165e-02 -4.59618449e-01 -6.72148705e-01 2.36538038e-01
3.02228987e-01 6.00150347e-01 3.54749829e-01 -4.64486957e-01
4.36021775e-01 -4.79783207e-01 -9.45198715e-01 6.53043687e-01
1.67405188e+00 3.65279138e-01 1.70414373e-01 1.19295955e+00
-3.46918702e-01 1.03100789e+00 -1.16815805e+00 -8.19306612e-01
-7.42661431e-02 -5.24475396e-01 -1.63435745e+00 -4.21811998e-01
9.91685212e-01 -4.87516761e-01 -1.88879013e-01 7.23656893e-01
4.51864183e-01 3.93905163e-01 7.19032884e-01 -5.53979635e-01
-1.21185958e+00 8.16484690e-01 1.46279037e-01 4.49654371e-01
1.27990758e+00 2.34340161e-01 1.38778853e+00 -9.24291611e-01
7.21357763e-01 1.05601835e+00 6.75418735e-01 4.86178339e-01
-9.91904259e-01 -2.09668845e-01 6.43806979e-02 3.28802258e-01
-8.81876111e-01 -6.56708121e-01 -1.27682136e-02 -3.85579437e-01
1.33985972e+00 4.84751910e-01 2.94961184e-01 1.27874136e+00
3.92980650e-02 7.38016963e-01 1.09723365e+00 -6.35555685e-01
-7.29694664e-02 -2.15840098e-02 4.14688051e-01 2.97079146e-01
9.51023772e-02 -5.56335211e-01 -5.95698059e-01 -3.14465314e-01
8.96291435e-02 -7.19589889e-01 -6.25447989e-01 5.55965543e-01
-1.34398198e+00 7.18691409e-01 1.13671012e-01 -5.73578570e-03
-3.48902434e-01 -7.49865100e-02 3.33916575e-01 7.16219664e-01
1.37581855e-01 1.22304845e+00 -7.57251620e-01 -7.77405620e-01
-1.08577466e+00 6.65762126e-01 1.37488937e+00 1.05770624e+00
2.26689339e-01 -5.27280569e-01 -7.55875826e-01 9.49276686e-01
-1.59961060e-02 7.51328051e-01 3.79063934e-01 -1.45044208e+00
9.05683935e-01 3.40849191e-01 8.19410920e-01 -7.85171092e-01
-1.28607944e-01 -5.24220169e-01 -2.21048802e-01 -3.86352330e-01
1.03463221e+00 -2.90204167e-01 -6.65934980e-01 1.50778103e+00
-2.32973680e-01 -7.45871365e-01 -1.16731510e-01 7.26371467e-01
1.10303140e+00 6.40224397e-01 2.85159260e-01 -1.48124322e-01
1.58027458e+00 -1.31671035e+00 -9.61693883e-01 -4.37117934e-01
7.24891663e-01 -1.14982224e+00 1.81420386e+00 6.07583284e-01
-1.55808663e+00 -4.53195304e-01 -8.68645430e-01 -5.57798624e-01
-1.41881958e-01 -1.18629187e-01 5.69973923e-02 5.25127351e-01
-1.06416714e+00 5.24744749e-01 -2.50088364e-01 -1.85976297e-01
8.19195509e-02 -4.27386194e-01 -2.57182330e-01 -3.92581612e-01
-1.64591789e+00 1.32242727e+00 1.62536595e-02 -2.62461931e-01
-7.78277516e-01 -6.80887341e-01 -8.04450274e-01 6.25764132e-02
4.23209518e-01 -7.04565048e-01 2.07420349e+00 -4.06383723e-01
-1.31182337e+00 7.14787841e-01 -6.11210108e-01 -3.35701942e-01
7.07052410e-01 -8.74059916e-01 -3.47349286e-01 4.38368052e-01
5.80235898e-01 4.64121282e-01 1.97066486e-01 -6.84846103e-01
-5.29313087e-01 2.74719357e-01 5.40886879e-01 4.97592390e-01
-2.09509552e-01 2.17048943e-01 -4.56759244e-01 -4.72891062e-01
-3.92178260e-02 -4.01311010e-01 -3.48406434e-02 -3.34212035e-01
-4.02492285e-01 -4.94438022e-01 2.65578311e-02 -1.42570186e+00
1.62713718e+00 -1.49614000e+00 -6.93572760e-02 -2.55710393e-01
9.60793793e-02 4.46850230e-04 -3.82619798e-01 9.44944918e-01
5.42893827e-01 6.27292216e-01 -2.00199068e-01 -2.64872283e-01
2.31867045e-01 -1.14803039e-01 -3.52025241e-01 -3.59107405e-02
1.23126402e-01 1.14968598e+00 -1.20831621e+00 -7.00880587e-01
-3.89396161e-01 -2.49340802e-01 -3.90944123e-01 1.79698601e-01
-5.61132729e-01 3.14402223e-01 -2.08103403e-01 6.22866631e-01
2.18556419e-01 -4.91470873e-01 -2.88606167e-01 3.52713972e-01
4.12478596e-02 1.31069243e+00 -7.15254307e-01 2.08024764e+00
-5.63419759e-01 7.06704259e-01 -1.27527907e-01 -2.07871214e-01
3.79586816e-01 5.33844590e-01 -3.28734249e-01 -1.12146461e+00
-3.58603030e-01 2.28458464e-01 4.86343987e-02 -8.92093003e-01
8.77989233e-01 1.91496372e-01 -2.01064482e-01 8.47169161e-01
-3.62709425e-02 -4.19283241e-01 7.91586339e-01 7.48951375e-01
1.36331415e+00 -1.36579379e-01 1.21431574e-01 -1.61099538e-01
4.05687809e-01 5.48939466e-01 3.98222864e-01 1.39431870e+00
-1.83521718e-01 6.24116898e-01 4.24254775e-01 2.13990659e-01
-1.11804128e+00 -1.05351841e+00 -1.60041273e-01 1.21655798e+00
-1.38959140e-01 -8.12150538e-01 -9.96711314e-01 -7.90297151e-01
-2.61193812e-01 1.10456324e+00 -3.79894882e-01 2.78684288e-01
-5.20973444e-01 -5.61905578e-02 8.57861757e-01 5.27362108e-01
4.99998510e-01 -9.94261622e-01 -5.75677037e-01 6.13214195e-01
-1.07246065e+00 -1.06902170e+00 -6.10480607e-01 -1.64031550e-01
-6.14681363e-01 -1.09708714e+00 -7.77108788e-01 -6.29277706e-01
8.55292007e-02 4.10800986e-02 1.96410644e+00 4.83285844e-01
2.29996517e-01 5.09639025e-01 -8.12049866e-01 -3.23023468e-01
-4.46964890e-01 4.08110917e-01 -4.94332999e-01 -1.16806686e+00
4.34899658e-01 -1.14180483e-01 -6.17429495e-01 1.38889253e-01
-5.75701714e-01 -1.08521745e-01 5.98572433e-01 1.13152158e+00
9.73105952e-02 -5.19128561e-01 9.62033153e-01 -8.76376271e-01
1.23228097e+00 -6.67557180e-01 -5.95313788e-04 6.66318655e-01
-5.33923388e-01 -2.91619211e-01 7.10015535e-01 -1.73081547e-01
-1.00906420e+00 -7.90667117e-01 -5.03428161e-01 5.73736846e-01
-2.73597121e-01 8.03666413e-01 1.80632651e-01 3.88250560e-01
1.06816828e+00 -2.41493545e-02 -2.65013814e-01 -5.82612157e-01
5.44289887e-01 7.44042695e-01 7.32989371e-01 -7.28506327e-01
6.27367496e-01 -1.23719878e-01 -7.50655770e-01 -4.45029378e-01
-1.24391937e+00 -4.92124766e-01 -3.54402035e-01 -5.01836985e-02
7.70533442e-01 -9.53376055e-01 -6.18092120e-01 1.42147481e-01
-1.29389238e+00 -6.08981431e-01 -1.63046166e-01 4.74133283e-01
-2.64709115e-01 5.15855491e-01 -1.17682028e+00 -6.66962087e-01
-5.22562981e-01 -6.44626200e-01 5.99772334e-01 3.46834391e-01
-9.97179031e-01 -1.00551200e+00 5.08324290e-03 9.38045621e-01
5.46566963e-01 -8.65303054e-02 9.23983455e-01 -7.81421900e-01
-2.64478832e-01 -2.78846741e-01 2.33912300e-02 2.51795262e-01
-8.42522681e-02 -3.12698297e-02 -7.32707500e-01 -1.41572922e-01
-5.98769113e-02 -9.97014046e-01 6.32633209e-01 3.31001617e-02
9.30530548e-01 -4.38735843e-01 -3.88797671e-02 -1.53754339e-01
9.18920755e-01 2.84560979e-03 4.97278094e-01 5.92728138e-01
1.83205768e-01 5.83781004e-01 8.51162612e-01 2.06386626e-01
8.91213953e-01 1.79967001e-01 6.10771589e-02 3.34300488e-01
-7.13250488e-02 -6.89100564e-01 2.89119422e-01 1.15940559e+00
3.56118858e-01 -6.40969217e-01 -1.13423097e+00 9.65137482e-01
-1.52170992e+00 -9.85872686e-01 -4.37925547e-01 1.92536485e+00
1.22466993e+00 4.87074912e-01 -1.19980760e-01 2.74490239e-03
1.45676851e-01 4.75985967e-02 -3.43696564e-01 -6.21684074e-01
-4.19070333e-01 5.73528588e-01 1.61750853e-01 9.46443856e-01
-6.72810793e-01 7.29865730e-01 7.78443575e+00 6.48608565e-01
-3.10632944e-01 1.77933499e-01 4.16616380e-01 -8.10972154e-02
-9.01148260e-01 1.71968043e-01 -5.39238930e-01 5.97694933e-01
1.18528247e+00 -3.50148320e-01 3.29750657e-01 6.06292725e-01
2.52516270e-01 -5.24774313e-01 -9.44932938e-01 3.17676783e-01
1.83451757e-01 -1.14304769e+00 -1.04630098e-01 -5.42165756e-01
8.73002350e-01 -6.96587339e-02 -1.18134417e-01 6.58908963e-01
5.97449005e-01 -1.57865822e+00 8.26733291e-01 5.54473221e-01
8.16526711e-01 -4.24537092e-01 8.22900951e-01 8.30368936e-01
-5.51612318e-01 4.43305597e-02 -3.94309133e-01 -4.20724511e-01
5.76191545e-01 4.47149396e-01 -6.22882366e-01 5.26460886e-01
7.98894048e-01 2.81977922e-01 -9.14395869e-01 1.17522597e+00
-7.60143042e-01 1.05826926e+00 -1.85259014e-01 -4.13165003e-01
3.60723943e-01 1.76884964e-01 3.74769449e-01 1.45986497e+00
1.49766609e-01 3.53007317e-01 -9.78348628e-02 5.95925093e-01
-3.12721878e-01 2.12289497e-01 -1.80281758e-01 -1.50830880e-01
9.84196544e-01 8.97466958e-01 -7.79811293e-02 -6.11781240e-01
-6.99072421e-01 1.02336776e+00 6.25577390e-01 4.54082221e-01
-5.15168130e-01 -7.87430525e-01 2.06463158e-01 -1.67790234e-01
3.44711654e-02 -3.27242583e-01 -4.63369161e-01 -1.17739451e+00
5.15322566e-01 -1.42308497e+00 5.44750452e-01 -9.75672305e-01
-1.49563527e+00 5.66235125e-01 -3.32486778e-01 -1.06248581e+00
-3.21651340e-01 -4.30147767e-01 -6.49491608e-01 1.32513261e+00
-1.68068957e+00 -4.79175568e-01 -3.38945240e-01 3.63968283e-01
6.95419729e-01 4.16532159e-01 1.01000738e+00 2.72683501e-01
-3.80627485e-03 8.01194072e-01 5.94374016e-02 2.04847381e-01
1.26247764e+00 -1.64792991e+00 7.30193138e-01 8.74773502e-01
1.39950588e-01 1.18264782e+00 9.24542427e-01 -6.94230914e-01
-9.58660841e-01 -4.54685420e-01 1.58449697e+00 -1.14559245e+00
5.52418828e-01 -1.03900224e-01 -1.37199342e+00 4.85770434e-01
9.41474497e-01 -7.44639695e-01 9.37679708e-01 4.29947555e-01
-4.75504130e-01 4.08920288e-01 -9.45124507e-01 6.74530983e-01
9.50215757e-01 -8.95023346e-01 -1.44233835e+00 6.82888865e-01
9.23080683e-01 -9.25950348e-01 -9.53825414e-01 1.78277060e-01
3.38054985e-01 -6.64274931e-01 6.23508811e-01 -9.92472351e-01
8.41341317e-01 -2.55101055e-01 6.91616014e-02 -1.52671766e+00
-3.79693657e-01 -5.82910895e-01 -2.14193672e-01 1.07679939e+00
1.04925168e+00 -2.06942916e-01 4.27195817e-01 9.64443624e-01
-4.24343705e-01 -6.49565458e-01 -7.41057098e-01 -6.42963886e-01
7.49967873e-01 -3.82430196e-01 5.42363584e-01 9.55171883e-01
4.91940081e-01 4.59739536e-01 -5.40049337e-02 -1.80624500e-01
1.78404838e-01 -1.62403315e-01 5.30927002e-01 -7.51600504e-01
-3.98803502e-01 -4.55417305e-01 5.66730917e-01 -1.66215181e+00
-8.66540521e-02 -6.43562317e-01 2.75978506e-01 -2.18708754e+00
1.39528275e-01 -2.90415853e-01 2.67356765e-02 1.71872571e-01
-9.04242814e-01 -1.10436394e-03 2.16898602e-02 6.64631650e-02
-9.70728040e-01 3.73354197e-01 1.39183104e+00 -4.21865433e-02
1.99221388e-01 -3.03662658e-01 -1.04108739e+00 3.78516495e-01
6.20284200e-01 -2.37363368e-01 -3.33473146e-01 -1.06107056e+00
5.82040370e-01 3.61527652e-01 1.21453367e-01 -8.22748125e-01
5.39855719e-01 -2.30093852e-01 4.72777367e-01 -5.42604148e-01
1.70071185e-01 -1.91484347e-01 -3.08158755e-01 1.90666541e-01
-9.22496796e-01 5.72103798e-01 1.92098677e-01 2.32581884e-01
-4.81148928e-01 -7.49164701e-01 2.05463648e-01 -5.34101427e-01
-3.63496482e-01 -2.48564541e-01 -7.16830969e-01 9.62565303e-01
4.27853733e-01 -3.96998934e-02 -9.37983513e-01 -9.54517603e-01
-4.79660571e-01 7.86426544e-01 4.49188590e-01 3.17988515e-01
3.96783471e-01 -1.04125702e+00 -1.06116128e+00 -6.60284698e-01
2.46481255e-01 2.14540362e-02 3.27010512e-01 5.86152196e-01
-6.21871233e-01 7.28069007e-01 1.32524772e-02 -1.81903094e-01
-9.35237169e-01 3.28955092e-02 1.40923023e-01 -3.72163445e-01
-4.24811244e-01 1.08322299e+00 -5.95923066e-01 -3.36053252e-01
2.97460377e-01 -2.68182725e-01 -4.31106269e-01 9.36620235e-02
8.39358568e-01 3.74083668e-01 2.33978346e-01 -6.31084219e-02
-3.95641243e-03 3.66357744e-01 -2.23155677e-01 -6.28381670e-01
7.20601439e-01 -4.15355712e-01 7.14459503e-03 4.23378408e-01
7.45235443e-01 4.91028577e-01 -7.95018733e-01 -3.19126010e-01
3.25946122e-01 -4.18150276e-01 -5.04700661e-01 -1.47549284e+00
2.52780672e-02 8.20562005e-01 -2.73467332e-01 3.39155376e-01
7.37829924e-01 -1.31584093e-01 1.10640752e+00 8.33659708e-01
2.26934671e-01 -1.38741660e+00 2.85009295e-01 1.03213215e+00
1.03485441e+00 -1.36809683e+00 -1.19685411e-01 -7.06669465e-02
-6.95339501e-01 1.01366675e+00 8.57818842e-01 2.65574187e-01
-6.46622032e-02 -1.09275244e-01 3.95603895e-01 5.41757559e-03
-1.24632645e+00 2.64383461e-02 5.02017200e-01 4.07163411e-01
9.26773489e-01 -5.42905740e-02 -7.56433725e-01 7.79043376e-01
-7.67121136e-01 -2.52814025e-01 7.89232433e-01 9.16837454e-01
-6.73096716e-01 -8.98213148e-01 -1.79599121e-01 6.94333196e-01
-7.59555876e-01 -4.54060435e-01 -5.11378527e-01 5.56760073e-01
-3.86844188e-01 1.69009888e+00 -7.57503137e-02 1.55725414e-02
5.22823095e-01 3.77405822e-01 2.89116621e-01 -8.05539250e-01
-9.11452174e-01 -5.05249500e-01 7.84098029e-01 -4.11765903e-01
-1.24124354e-02 -5.96813977e-01 -9.51801717e-01 -4.03490752e-01
-2.38848552e-01 7.83092558e-01 2.00433969e-01 1.00043428e+00
1.20447241e-01 2.53341347e-01 1.06832281e-01 2.90352762e-01
-8.68518472e-01 -1.60504019e+00 5.12495227e-02 6.32375002e-01
4.14881825e-01 -1.50073066e-01 -4.50688422e-01 6.55681118e-02] | [11.380522727966309, 8.101594924926758] |
1d4e69b7-aa13-4668-a0b3-dec1a5e06537 | big-earth-data-and-machine-learning-for | 2211.12584 | null | https://arxiv.org/abs/2211.12584v1 | https://arxiv.org/pdf/2211.12584v1.pdf | Big Earth Data and Machine Learning for Sustainable and Resilient Agriculture | Big streams of Earth images from satellites or other platforms (e.g., drones and mobile phones) are becoming increasingly available at low or no cost and with enhanced spatial and temporal resolution. This thesis recognizes the unprecedented opportunities offered by the high quality and open access Earth observation data of our times and introduces novel machine learning and big data methods to properly exploit them towards developing applications for sustainable and resilient agriculture. The thesis addresses three distinct thematic areas, i.e., the monitoring of the Common Agricultural Policy (CAP), the monitoring of food security and applications for smart and resilient agriculture. The methodological innovations of the developments related to the three thematic areas address the following issues: i) the processing of big Earth Observation (EO) data, ii) the scarcity of annotated data for machine learning model training and iii) the gap between machine learning outputs and actionable advice. This thesis demonstrated how big data technologies such as data cubes, distributed learning, linked open data and semantic enrichment can be used to exploit the data deluge and extract knowledge to address real user needs. Furthermore, this thesis argues for the importance of semi-supervised and unsupervised machine learning models that circumvent the ever-present challenge of scarce annotations and thus allow for model generalization in space and time. Specifically, it is shown how merely few ground truth data are needed to generate high quality crop type maps and crop phenology estimations. Finally, this thesis argues there is considerable distance in value between model inferences and decision making in real-world scenarios and thereby showcases the power of causal and interpretable machine learning in bridging this gap. | ['Vasileios Sitokonstantinou'] | 2022-11-22 | null | null | null | null | ['interpretable-machine-learning'] | ['methodology'] | [ 2.02918917e-01 3.64066124e-01 -3.69981527e-01 -1.35388579e-02
2.52747741e-02 -7.06994832e-01 5.24743319e-01 7.45033324e-01
-1.23852084e-03 8.15225363e-01 1.28471702e-01 -6.24333799e-01
-5.89401722e-01 -1.36883020e+00 -7.44085670e-01 -6.00953758e-01
-5.28906286e-01 2.60288537e-01 -4.60140891e-02 -7.76516497e-01
-1.74259007e-01 6.11595213e-01 -2.11256027e+00 2.50711858e-01
1.01920545e+00 9.62481320e-01 5.37097454e-01 6.60580456e-01
-1.03806891e-01 5.09107709e-01 -1.98515639e-01 4.91355509e-02
1.41954631e-01 -1.86308756e-01 -4.89448786e-01 -1.22705743e-01
-1.15448043e-01 -2.19933033e-01 3.23201776e-01 1.05491805e+00
3.57746661e-01 -2.51547575e-01 2.07821038e-04 -1.34325945e+00
-3.96985084e-01 5.06727815e-01 -2.16540694e-01 1.34126395e-01
2.35052362e-01 9.68365371e-03 4.75168824e-01 -3.61665249e-01
6.91180766e-01 9.78164017e-01 7.94791996e-01 -2.58266747e-01
-8.37355554e-01 -5.41974962e-01 1.12667628e-01 3.77118997e-02
-1.17235100e+00 -3.01370174e-01 -2.04186831e-02 -6.73556030e-01
1.00863862e+00 5.44081151e-01 1.09534049e+00 6.32845402e-01
2.85770983e-01 2.46137470e-01 1.05405521e+00 -5.70717931e-01
3.20930719e-01 6.72118664e-02 -2.58206278e-01 2.82817721e-01
6.85006440e-01 5.21356285e-01 -5.18269658e-01 5.52890487e-02
5.69067717e-01 2.63027489e-01 -7.77418986e-02 -1.09090962e-01
-1.30987620e+00 6.11221433e-01 4.58985418e-01 5.29949248e-01
-9.02274966e-01 -2.27833167e-01 2.96657979e-01 3.85384232e-01
7.70340562e-01 4.42470819e-01 -1.08879948e+00 4.13790978e-02
-1.00447762e+00 1.39599919e-01 8.33333015e-01 8.79954636e-01
8.69705319e-01 2.68324584e-01 5.76075613e-01 3.90220404e-01
4.88368243e-01 8.83439541e-01 1.21986628e-01 -8.85940731e-01
1.79461285e-01 8.24782073e-01 3.08782637e-01 -1.37506068e+00
-6.52654350e-01 -3.12435985e-01 -6.42642617e-01 9.60898846e-02
1.62314951e-01 -4.48551476e-01 -7.55667150e-01 1.31336081e+00
6.64949477e-01 1.28634557e-01 4.02241766e-01 7.05338299e-01
6.96283340e-01 6.93787932e-01 3.69791061e-01 -2.60777265e-01
1.42107117e+00 -1.03757821e-01 -9.56184745e-01 -2.61543244e-01
9.21719551e-01 -3.61754507e-01 7.42274582e-01 2.61557609e-01
-7.09230900e-01 -2.74855256e-01 -9.68851030e-01 5.63578069e-01
-1.22974837e+00 -2.09838435e-01 1.26939344e+00 5.30851066e-01
-7.52740622e-01 7.85186648e-01 -1.02011585e+00 -8.41778517e-01
5.19609869e-01 2.15949401e-01 -4.87708330e-01 -5.71877249e-02
-1.39722133e+00 9.69902277e-01 8.41882765e-01 3.81760150e-01
-7.35125899e-01 -9.40865457e-01 -8.34864557e-01 -4.82463138e-03
3.24538618e-01 -2.72830844e-01 5.27878046e-01 -1.01828766e+00
-1.01033735e+00 8.62197399e-01 4.49728221e-01 -5.74463546e-01
1.90731853e-01 -2.98502117e-01 -9.34434950e-01 3.44264209e-02
3.60297769e-01 5.93355417e-01 2.40694463e-01 -1.07094800e+00
-1.31236684e+00 -7.98969150e-01 -2.14647911e-02 -1.03063002e-01
-4.55496877e-01 1.08097672e-01 4.67500448e-01 -4.46905851e-01
2.86939979e-01 -9.48598921e-01 -2.60116130e-01 -2.81268537e-01
2.49924779e-01 1.53852776e-01 1.02838445e+00 -8.07096303e-01
1.03339374e+00 -1.92849982e+00 -2.07488924e-01 1.04133286e-01
-1.10337190e-01 3.53342980e-01 6.79293498e-02 8.16777587e-01
-4.90435921e-02 3.22166800e-01 -2.74915457e-01 8.70043993e-01
-3.02895457e-01 4.87339765e-01 -2.63730675e-01 2.75993913e-01
1.59579307e-01 7.59292305e-01 -1.36247551e+00 -1.59529105e-01
5.05964696e-01 3.35607886e-01 -2.11468991e-02 -7.01927841e-02
-4.42670584e-01 5.93643486e-01 -3.55870515e-01 9.14284945e-01
6.07827842e-01 -1.59154478e-02 4.26448852e-01 -1.37278304e-01
-7.14354038e-01 -1.86816216e-01 -1.26571953e+00 1.43500292e+00
-4.70576495e-01 5.02617657e-01 2.86775589e-01 -9.43080187e-01
9.59932983e-01 5.57796955e-01 6.56241357e-01 -7.02302277e-01
-1.50123939e-01 5.82784235e-01 -2.35855341e-01 -9.52319801e-01
5.55058479e-01 -1.46420658e-01 1.23084143e-01 1.01884358e-01
1.10603273e-01 -2.16849610e-01 -4.68804389e-02 -3.46298397e-01
6.93338811e-01 6.56590044e-01 4.72918391e-01 -6.87813997e-01
6.79936185e-02 6.30038023e-01 5.43723643e-01 1.00658514e-01
-3.32383923e-02 4.39020768e-02 2.43320301e-01 -9.12575841e-01
-9.18976307e-01 -3.54172856e-01 -3.61848384e-01 1.34676278e+00
-8.07434097e-02 -2.81142890e-01 -3.17760646e-01 -1.51234314e-01
3.63849968e-01 8.14334273e-01 -4.63206500e-01 2.71849483e-01
-6.16028644e-02 -1.15717614e+00 5.01541317e-01 1.73946545e-01
5.02591729e-01 -8.96847188e-01 -1.13794136e+00 1.91915646e-01
-1.81031108e-01 -1.11824155e+00 1.05619669e+00 3.83347213e-01
-1.11718202e+00 -1.00658727e+00 -2.36517206e-01 -2.49852583e-01
4.17947710e-01 3.68538648e-01 1.02078438e+00 -1.87615797e-01
-2.99939990e-01 1.08933114e-01 -6.19458973e-01 -1.00594008e+00
-2.56562710e-01 7.88416043e-02 1.06858462e-03 -4.13308948e-01
6.82189345e-01 -6.88169599e-01 -3.84102464e-01 3.70049983e-01
-1.13244092e+00 -9.32045355e-02 5.61476290e-01 2.91967034e-01
6.69640660e-01 3.25209469e-01 9.09376323e-01 -7.41837859e-01
1.62886813e-01 -1.26048088e+00 -8.57035875e-01 2.44854018e-01
-8.61255407e-01 -3.01768959e-01 1.76142696e-02 1.31999448e-01
-8.76430750e-01 2.22532660e-01 7.87251294e-02 2.91871373e-02
-7.38078356e-01 1.04597974e+00 -2.71129251e-01 1.33108437e-01
8.24051917e-01 -3.72851640e-01 6.22758158e-02 -3.72740477e-01
4.19748068e-01 9.25421178e-01 4.50835437e-01 -1.53064683e-01
4.07129318e-01 7.41532803e-01 9.07875896e-02 -1.35798812e+00
-5.98493755e-01 -3.62808555e-01 -8.71348619e-01 -3.07629496e-01
7.98845172e-01 -1.34854448e+00 -4.15629029e-01 3.10574323e-01
-7.49458432e-01 -4.81784612e-01 -6.37837410e-01 7.05504298e-01
-3.54240268e-01 -5.90687022e-02 1.09494135e-01 -8.79426181e-01
-2.40201056e-01 -4.76394564e-01 9.48454082e-01 2.70837039e-01
-2.36735418e-01 -1.08885217e+00 4.44329232e-02 2.16596752e-01
5.23108840e-01 1.09383774e+00 6.24265790e-01 -3.45574409e-01
-4.96932805e-01 -4.88558650e-01 -9.34779570e-02 1.93934999e-02
4.84540939e-01 1.83426782e-01 -1.08963656e+00 -1.61894828e-01
-1.81956857e-01 -1.14951521e-01 4.28483754e-01 4.19286042e-01
6.23111606e-01 -3.31305176e-01 -4.27255183e-01 5.63126266e-01
1.55070162e+00 3.22277881e-02 8.16855490e-01 6.47196174e-01
5.29241800e-01 1.19423246e+00 1.19243717e+00 4.27892923e-01
5.94063044e-01 2.33091086e-01 9.12490964e-01 -1.87304005e-01
2.21115783e-01 -2.01357469e-01 9.02311280e-02 3.63946706e-01
-4.33779508e-01 -3.39252986e-02 -1.36290419e+00 1.01794147e+00
-2.04083514e+00 -1.11275887e+00 -6.25676453e-01 2.45995259e+00
3.89709115e-01 -2.21599728e-01 -9.87378955e-02 2.63283491e-01
6.00286067e-01 1.60930797e-01 -1.88302279e-01 -3.48803878e-01
-4.96711642e-01 3.00669484e-02 1.17586768e+00 2.17741400e-01
-1.11967361e+00 1.17423475e+00 6.26380539e+00 3.19848001e-01
-1.13600230e+00 2.57177413e-01 1.11208379e-01 8.05251300e-02
-2.82146007e-01 3.04083675e-01 -7.22134769e-01 2.32038110e-01
1.51906419e+00 1.53981090e-01 2.30541065e-01 6.69398665e-01
6.15812421e-01 -3.05148751e-01 -5.18808126e-01 3.18677872e-01
-4.46466625e-01 -1.47953224e+00 -3.34838420e-01 2.70127118e-01
9.20418322e-01 5.61698318e-01 -5.79351783e-01 -2.33580306e-01
7.07687736e-01 -7.71296620e-01 5.60849786e-01 6.10944867e-01
6.59422696e-01 -4.61425960e-01 8.90228927e-01 3.39237243e-01
-1.13832343e+00 -4.53409672e-01 -4.22093451e-01 -4.27136540e-01
-8.43099654e-02 7.57107854e-01 -4.35379595e-01 1.02474785e+00
1.10186493e+00 8.52347791e-01 -2.56587356e-01 7.80337036e-01
-3.17784101e-01 5.14388382e-01 -6.37555301e-01 3.51311713e-01
3.19022030e-01 -1.75525486e-01 5.36002398e-01 9.56530273e-01
6.65687323e-01 2.10660383e-01 -1.30625725e-01 4.39054012e-01
5.92119098e-01 3.35057899e-02 -1.31116152e+00 -3.75024855e-01
5.28010786e-01 1.40067148e+00 -8.28476310e-01 4.98267636e-02
-3.06856334e-01 4.02719498e-01 -4.72883999e-01 2.02538624e-01
-2.08109841e-01 -3.04088950e-01 6.79497421e-01 3.90981168e-01
2.38162875e-01 -1.72309190e-01 -2.81761110e-01 -8.17990243e-01
-3.59322608e-01 -8.97215486e-01 5.25662243e-01 -9.03261542e-01
-7.07894266e-01 1.27958402e-01 1.92755923e-01 -1.20079494e+00
-3.73995483e-01 -3.72000158e-01 -4.15948987e-01 7.89065599e-01
-1.85548162e+00 -1.72453129e+00 -6.92429185e-01 3.36259395e-01
2.56785657e-02 -5.86857647e-02 1.28754711e+00 1.97227538e-01
-3.09207767e-01 -4.55697805e-01 3.89572769e-01 -3.97995651e-01
3.44959438e-01 -1.02637255e+00 2.35968977e-01 7.70990551e-01
-3.16918939e-01 1.49274439e-01 6.24944329e-01 -1.07041764e+00
-1.43475759e+00 -1.31291449e+00 1.08024740e+00 -2.90420234e-01
8.34055007e-01 -3.41432146e-03 -7.60854602e-01 6.91646814e-01
-1.41626269e-01 2.69578367e-01 1.05906105e+00 1.44031942e-01
1.15192696e-01 -4.69779730e-01 -1.32881927e+00 1.11110508e-01
8.41608644e-01 -2.69280374e-01 -2.47872770e-01 6.73184097e-01
6.63961351e-01 -1.96404904e-01 -1.41630626e+00 5.30961514e-01
7.61343420e-01 -8.24180245e-01 7.35007048e-01 -9.49533224e-01
4.36375201e-01 -4.18496072e-01 -6.19234145e-01 -1.36811805e+00
-2.51191586e-01 -4.97434020e-01 1.46347992e-02 1.23060811e+00
2.63365805e-01 -7.26621926e-01 3.59376699e-01 4.84493464e-01
-2.65944451e-01 -2.44610384e-01 -6.74171805e-01 -6.56118572e-01
6.67005181e-02 -4.97426093e-01 1.28670871e+00 1.43337154e+00
7.25489780e-02 -2.78833181e-01 -6.21792153e-02 7.78371036e-01
5.31373680e-01 2.36665219e-01 7.13456452e-01 -1.77084637e+00
3.98958296e-01 1.62577540e-01 -7.85544753e-01 2.61692498e-02
-3.07026416e-01 -5.04558980e-01 -4.26865369e-01 -1.71345890e+00
-5.38220584e-01 -7.47937620e-01 -6.92806989e-02 7.30240345e-01
2.17164457e-01 8.33015889e-02 -5.53686768e-02 3.50947291e-01
-2.72620898e-02 1.32176489e-01 8.15731645e-01 1.22098885e-01
-4.92343456e-01 -7.00249970e-02 -8.17075074e-01 8.23383570e-01
8.23238850e-01 -5.03912985e-01 -2.06665128e-01 -4.87585843e-01
1.15808177e+00 2.49235611e-02 6.68096066e-01 -8.77699792e-01
-5.31503148e-02 -5.64652979e-01 2.82530755e-01 -6.51313722e-01
-4.33982372e-01 -1.21896088e+00 6.68869376e-01 4.63407487e-01
1.72859997e-01 -1.49906471e-01 6.57639682e-01 3.94359440e-01
-2.01888889e-01 -6.65895408e-03 3.25120211e-01 -3.23840171e-01
-9.42822933e-01 -1.54686626e-02 -3.18799615e-01 -2.99007595e-01
1.34854221e+00 -3.45679730e-01 -3.87254477e-01 -1.04327351e-02
-8.20353627e-01 5.95784605e-01 5.10811865e-01 7.01493442e-01
5.34376465e-02 -1.17811608e+00 -9.61556375e-01 3.61728966e-01
2.08975136e-01 2.47378945e-01 2.72374690e-01 8.07604671e-01
-6.37448311e-01 6.49663329e-01 -6.50803745e-01 -6.25759125e-01
-9.29956079e-01 3.87933820e-01 2.61153907e-01 5.50383888e-02
-5.16099513e-01 2.54210800e-01 -4.37915385e-01 -6.33190930e-01
-2.26860717e-01 -3.53412658e-01 -3.44999790e-01 7.19450891e-01
5.02635539e-01 5.25623918e-01 3.53958547e-01 -6.70285404e-01
-3.77103359e-01 3.39549959e-01 6.77146316e-01 1.75766721e-01
2.10920620e+00 -4.48946178e-01 -3.21164727e-01 6.82240725e-01
3.66882831e-01 -2.45505646e-01 -1.10199702e+00 1.99266180e-01
3.67502153e-01 -6.44755483e-01 3.91794533e-01 -9.47984397e-01
-9.21290219e-01 8.48165691e-01 1.00738811e+00 7.73279727e-01
1.13197684e+00 -1.02592088e-01 3.19504708e-01 4.13646579e-01
3.82616788e-01 -1.34992957e+00 -7.72428393e-01 1.34539440e-01
9.67481732e-01 -1.39922094e+00 2.57901579e-01 -2.06083119e-01
-3.63438517e-01 9.88182664e-01 9.53923166e-02 3.59412104e-01
8.39401364e-01 3.99888307e-01 -2.56597617e-04 -6.42444849e-01
-3.90254617e-01 -6.76780879e-01 -4.16731983e-01 1.15782189e+00
3.34545672e-01 5.45266032e-01 9.51981992e-02 4.04242337e-01
-9.14524421e-02 6.21443570e-01 2.61911720e-01 1.08808672e+00
-7.93127954e-01 -8.47510219e-01 -7.21958995e-01 5.09422183e-01
-4.46943253e-01 -5.87540865e-02 -1.92151025e-01 7.88639426e-01
6.49793029e-01 1.04908371e+00 1.69186052e-02 -1.99148387e-01
2.76843429e-01 6.54737726e-02 4.52468283e-02 -5.76430142e-01
-5.42544305e-01 -3.18708904e-02 1.12459742e-01 -5.77269256e-01
-7.78162956e-01 -8.04349899e-01 -1.12632096e+00 -3.91250432e-01
-5.43769062e-01 -2.39408743e-02 1.30570698e+00 8.88178825e-01
7.25976050e-01 5.42670071e-01 4.45820451e-01 -8.89686823e-01
-1.50137112e-01 -9.05852675e-01 -8.10456812e-01 -1.52261555e-01
2.09356606e-01 -4.64659601e-01 -1.63466036e-01 2.23596752e-01] | [9.487431526184082, -1.4811115264892578] |
bd6a8f19-fb63-4952-8f88-0f260291217f | evaluating-the-timing-and-magnitude-of | null | null | https://openreview.net/forum?id=4Gew0VrWfkx | https://openreview.net/pdf?id=4Gew0VrWfkx | Evaluating the timing and magnitude of semantic change in diachronic word embedding models | Recent studies have suggested that diachronic word embedding models are able to track the direction of changes in public perception. Building on these works, we evaluate the ability of diachronic word embedding models to accurately capture such changes both qualitatively and quantitatively, such as their timing and magnitudes. Using a longitudinal dataset on public perception of brands, we found that evolution of word meaning as captured by diachronic word embedding models, trained on New York Times articles, reflected the timing and magnitudes of general consumer awareness of companies. In contrast, this was not the case for other readily available characteristics, such as stock market prices. This comparison is enabled by a new feature extraction method which summarizes the semantic changes encoded in diachronic word embeddings. | ['Anonymous'] | 2022-01-20 | null | null | null | acl-arr-january-2022-1 | ['diachronic-word-embeddings'] | ['natural-language-processing'] | [-4.79052067e-01 9.71257538e-02 -4.26473826e-01 -1.29371464e-01
-3.03552955e-01 -8.04852724e-01 1.15182042e+00 9.30506885e-01
-7.14392960e-01 -3.73675935e-02 1.18964434e+00 -2.78707802e-01
-2.06227899e-01 -1.09654999e+00 -5.61153829e-01 -1.55520692e-01
6.79619312e-02 -1.24686420e-01 -1.63761407e-01 -5.75425625e-01
3.08806509e-01 1.57475770e-01 -1.35473943e+00 9.43581834e-02
2.59030104e-01 8.49741936e-01 -2.41379440e-01 1.91135854e-01
-3.06766808e-01 3.29546124e-01 -1.78522155e-01 -9.17124450e-01
1.49908990e-01 1.79446906e-01 -3.75766993e-01 -3.22498269e-02
3.33218217e-01 -1.18304975e-01 -7.45252013e-01 1.05553925e+00
1.21030279e-01 -1.81214616e-01 6.46122456e-01 -7.40252852e-01
-1.38922298e+00 8.77791941e-01 -9.11722258e-02 5.84317446e-01
4.47107017e-01 2.59556681e-01 1.64788055e+00 -7.61443079e-01
9.70872343e-01 1.35545444e+00 9.91337776e-01 -2.20865399e-01
-1.38719726e+00 -5.52078128e-01 5.01083553e-01 1.44321844e-01
-1.03105986e+00 -7.04950988e-02 8.01263690e-01 -1.04485548e+00
6.60977244e-01 -5.81378862e-03 1.24310052e+00 1.27435446e+00
5.20729184e-01 5.86885773e-02 1.33829427e+00 -1.94741547e-01
-1.15538977e-01 4.44295824e-01 7.20388532e-01 1.42426208e-01
7.63595402e-01 3.15219581e-01 -4.40973133e-01 -2.31013417e-01
3.63674700e-01 2.18254656e-01 6.08948693e-02 5.21811061e-02
-1.25210571e+00 1.17163920e+00 3.97781998e-01 7.81484365e-01
-6.90514684e-01 1.97291970e-01 5.14274001e-01 4.25589919e-01
8.79661500e-01 6.42736733e-01 -6.09242082e-01 -5.24499476e-01
-4.52167630e-01 2.81299055e-01 4.87033606e-01 8.30922127e-02
7.34187484e-01 -4.40648794e-02 -5.38849011e-02 5.80241382e-01
4.35486764e-01 4.07993197e-01 8.84814739e-01 -4.25971478e-01
1.13628104e-01 8.09547186e-01 2.69134372e-01 -1.71379638e+00
-2.89135873e-01 -2.73087859e-01 1.41503349e-01 -9.35599953e-02
2.78588772e-01 2.84294654e-02 -6.27565145e-01 1.81876898e+00
1.18406536e-02 -2.08384067e-01 -8.13729465e-02 3.96559179e-01
5.89124739e-01 4.71545428e-01 3.10729772e-01 5.76204322e-02
1.73624980e+00 -1.13297544e-01 -7.29205608e-01 -4.10107523e-01
6.61936820e-01 -7.82484531e-01 1.18588865e+00 -1.54096663e-01
-8.27448905e-01 -6.64869785e-01 -8.28732789e-01 6.03974983e-02
-9.94057178e-01 -7.62005329e-01 7.88300276e-01 7.30454862e-01
-6.46045685e-01 5.13794243e-01 -4.84627604e-01 -5.41900098e-01
1.76212534e-01 -2.84471095e-01 -2.28550792e-01 1.25445366e-01
-1.44056737e+00 1.09360611e+00 1.86958432e-01 -2.55902708e-01
-2.33466119e-01 -8.51785958e-01 -9.80552018e-01 2.40841266e-02
-1.45675808e-01 -3.34675163e-01 1.12077153e+00 -1.07607007e+00
-1.04192722e+00 8.10147643e-01 2.61254489e-01 -3.06197017e-01
-1.01338089e-01 -1.10503711e-01 -1.08269453e+00 -4.82536018e-01
1.39477804e-01 2.63135284e-01 4.45600241e-01 -8.43785405e-01
-2.74429739e-01 -5.43528318e-01 2.47707754e-01 -2.66011655e-01
-9.01403725e-01 1.12099145e-02 3.10353905e-01 -8.75061870e-01
-8.37064758e-02 -6.60020351e-01 -5.43560348e-02 -1.54853463e-01
3.37054610e-01 -6.25219524e-01 2.54331261e-01 -7.45372355e-01
1.41413534e+00 -2.73783565e+00 -3.08878005e-01 9.04603601e-02
3.39250118e-01 -1.59801885e-01 -2.04540834e-01 9.78188872e-01
-1.69483975e-01 3.74603778e-01 3.37815374e-01 4.60658446e-02
6.33584976e-01 2.64544189e-01 -4.05665696e-01 8.03975821e-01
8.61519277e-02 1.20850849e+00 -8.75858009e-01 1.12897106e-01
1.49054155e-01 4.57755148e-01 -6.71721399e-01 -5.40457487e-01
-5.02391756e-02 -4.09672529e-01 -3.58960748e-01 5.17487645e-01
3.07230830e-01 -1.92983016e-01 1.56129807e-01 -2.39442885e-01
-5.97758174e-01 6.48416221e-01 -6.11623049e-01 1.14639032e+00
-5.32899499e-01 1.06730044e+00 -4.15589303e-01 -6.20867014e-01
7.88243353e-01 2.80492872e-01 3.65596324e-01 -1.21472323e+00
3.20332348e-01 9.81368497e-02 3.86607409e-01 -3.26555759e-01
8.69217694e-01 -5.84182918e-01 -4.47166085e-01 4.41449314e-01
-2.81830579e-01 6.58404268e-03 -4.92723994e-02 1.06110573e-02
7.45927513e-01 -5.29555857e-01 3.51424575e-01 -3.64287913e-01
-9.79451388e-02 1.84949845e-01 4.46225911e-01 2.28938207e-01
-1.58273175e-01 1.44376174e-01 4.28959906e-01 -3.52041841e-01
-1.00499344e+00 -1.26512325e+00 -6.04430079e-01 9.87444222e-01
-3.32962582e-03 -6.27179444e-01 -2.16634110e-01 -1.73100278e-01
6.65894747e-01 1.08324242e+00 -1.14041591e+00 -3.34133476e-01
-1.22532293e-01 -6.28555834e-01 5.17999642e-02 5.56210339e-01
-1.80438802e-01 -6.42113924e-01 -4.18755144e-01 4.77289528e-01
3.38522106e-01 -9.43248749e-01 -2.90893137e-01 -1.59161180e-01
-7.07789540e-01 -9.85623717e-01 -2.96148628e-01 -3.42106313e-01
1.83851227e-01 7.30766132e-02 1.13156629e+00 -3.24429274e-01
-1.30325973e-01 8.19238961e-01 -5.12901485e-01 -8.03629339e-01
-3.72435749e-01 -6.29610345e-02 6.33864626e-02 9.58481878e-02
7.97635257e-01 -5.08577645e-01 -6.45983338e-01 -2.41799310e-01
-1.08424282e+00 -6.31225407e-01 5.44561565e-01 4.46855098e-01
2.20461428e-01 -5.54792322e-02 6.52042329e-01 -6.59379840e-01
1.07985532e+00 -9.42423165e-01 -2.53073335e-01 -4.12529290e-01
-1.36405087e+00 -7.90447816e-02 -1.31296623e-03 -6.55900896e-01
-6.55712187e-01 -9.11650896e-01 -3.42516840e-01 3.19086254e-01
1.19589590e-01 9.32623804e-01 6.65411592e-01 3.10221612e-01
3.28944534e-01 -3.71294618e-02 1.65781781e-01 -5.79709589e-01
7.68607855e-01 7.18908906e-01 1.92029476e-01 -5.16697206e-02
9.56962347e-01 7.13045418e-01 -7.61374593e-01 -6.50497556e-01
-8.52292657e-01 -5.55081129e-01 -4.21651751e-01 -4.02185246e-02
1.20881426e+00 -1.08120036e+00 -4.35222596e-01 1.47590011e-01
-9.59894061e-01 3.91753204e-02 -7.31391132e-01 8.26855898e-01
-9.92913097e-02 2.56649014e-02 -5.93179822e-01 -4.59211171e-01
1.84008271e-01 -8.52665842e-01 5.38806736e-01 -2.51886278e-01
-8.11734915e-01 -1.64491177e+00 6.19208694e-01 1.17320381e-01
6.11010611e-01 4.95927215e-01 1.30968738e+00 -7.63437688e-01
-8.62337500e-02 -5.54923773e-01 -1.20693885e-01 3.30984116e-01
4.41976488e-01 -1.20405875e-01 -7.60707796e-01 -1.23866372e-01
-7.41561577e-02 3.37765887e-02 9.19219375e-01 2.45176747e-01
4.37779278e-01 -4.37204480e-01 -2.64184862e-01 -6.19463362e-02
1.53839231e+00 1.47149786e-01 5.61406016e-01 8.23709369e-01
3.35424393e-01 5.40071666e-01 2.39909738e-01 4.36038941e-01
6.69089198e-01 6.56444192e-01 2.60161966e-01 4.76014987e-02
-4.41739336e-02 -5.48978925e-01 7.16212153e-01 1.13411081e+00
1.80630088e-01 1.53042197e-01 -8.69094849e-01 9.70173538e-01
-1.22711253e+00 -9.41468537e-01 -1.08715810e-01 1.78369594e+00
5.63883841e-01 4.45445210e-01 2.83021212e-01 1.61918718e-02
2.19410852e-01 8.10757995e-01 -1.79060981e-01 -8.99093211e-01
-1.25875458e-01 2.34762564e-01 7.42928982e-01 2.41412833e-01
-4.06557143e-01 5.73151767e-01 7.56679153e+00 1.35038421e-01
-9.92052376e-01 4.31707114e-01 3.08891773e-01 1.11114115e-01
-1.16613436e+00 -2.85248142e-02 -2.89965093e-01 5.16923428e-01
1.22432184e+00 -4.39899266e-01 -2.92550586e-02 5.36337197e-01
3.27408999e-01 3.08733165e-01 -9.50081885e-01 6.02320015e-01
1.62934229e-01 -1.36310673e+00 5.94503595e-04 4.49428648e-01
7.42871940e-01 2.32898500e-02 6.44619465e-01 2.91230708e-01
4.29010153e-01 -9.38591361e-01 1.08317459e+00 4.44456965e-01
4.95588094e-01 -5.77381909e-01 7.18764663e-01 -3.15740824e-01
-7.67325461e-01 -4.91816312e-01 -1.75119773e-01 -7.15366900e-01
3.24079514e-01 6.04202390e-01 -6.51781261e-01 1.59062833e-01
6.59018815e-01 9.87590015e-01 -7.21625149e-01 2.78275877e-01
9.00751054e-02 7.81022310e-01 3.48334581e-01 -8.81182903e-04
5.33940613e-01 -3.70654851e-01 3.41932803e-01 1.04882574e+00
4.04686242e-01 -2.18367442e-01 -5.05769968e-01 7.80009329e-01
-8.65647569e-02 1.49849921e-01 -8.01510453e-01 -1.05466056e+00
3.26330274e-01 9.06961322e-01 -6.08273566e-01 -1.58744678e-01
-1.26380575e+00 5.83489776e-01 1.75321788e-01 2.32612252e-01
-8.43075693e-01 1.35286001e-03 1.44635618e+00 7.25861669e-01
4.66842651e-01 -5.32129705e-01 -2.21102908e-01 -8.80458832e-01
-2.23722652e-01 -6.66621506e-01 2.30516806e-01 -6.08045876e-01
-1.63057983e+00 1.19218402e-01 -1.11873694e-01 -6.18622005e-01
1.64186224e-01 -7.59760022e-01 -5.63737392e-01 5.71846247e-01
-1.43323100e+00 -8.73855889e-01 9.14309919e-02 1.68836787e-01
3.01435858e-01 2.17753172e-01 7.32520819e-01 2.02360511e-01
-3.21440756e-01 2.12543771e-01 2.91583151e-01 -3.50475800e-03
6.87948287e-01 -1.08883417e+00 7.90176034e-01 3.51149917e-01
4.81465757e-01 7.50690997e-01 8.76438975e-01 -6.82369113e-01
-1.20103514e+00 -6.85598075e-01 1.18598425e+00 -8.69585454e-01
1.56984460e+00 -3.36949557e-01 -5.34468710e-01 1.08599365e+00
4.91283804e-01 -5.48288643e-01 1.05346370e+00 6.60346627e-01
-7.34008789e-01 1.63917206e-02 -7.60499716e-01 6.25187695e-01
1.05841064e+00 -9.37476337e-01 -1.04770994e+00 2.75839865e-01
1.26158464e+00 3.77259403e-01 -1.35675156e+00 -1.37505746e-02
8.14687014e-01 -7.06616879e-01 1.12648153e+00 -7.32742608e-01
5.14802575e-01 3.77636969e-01 -5.33305645e-01 -1.47728300e+00
-8.51284921e-01 -2.20544830e-01 3.59559089e-01 1.23423195e+00
4.91523683e-01 -1.22766542e+00 1.69268057e-01 5.19405365e-01
-2.30234206e-01 -6.24592185e-01 -9.62575257e-01 -7.98536658e-01
5.38860559e-01 -5.62731564e-01 8.55615079e-01 1.13793254e+00
-9.29075927e-02 -1.48432562e-02 2.82380849e-01 -1.09353043e-01
-3.68272848e-02 1.78996697e-01 2.62891501e-01 -1.59050953e+00
-5.03342189e-02 -7.15527356e-01 -1.01543915e+00 -5.08252263e-01
3.46392781e-01 -1.05540287e+00 -7.60706484e-01 -1.53715158e+00
2.11242363e-01 -1.35496080e-01 -6.36181295e-01 -1.19242236e-01
9.09992754e-02 1.93299979e-01 3.51564199e-01 -1.14550829e-01
-2.83619855e-02 5.04012465e-01 1.01874363e+00 -8.47436860e-02
1.53557751e-02 -4.86824185e-01 -1.24697983e+00 7.91404486e-01
6.81348860e-01 -4.04042393e-01 -2.03919545e-01 -4.88189787e-01
1.01992989e+00 -4.86736178e-01 5.08417845e-01 -5.37899613e-01
-5.55093765e-01 -2.68850058e-01 1.97339654e-01 -1.84933454e-01
3.68965536e-01 -9.04901028e-01 2.02082440e-01 6.20258987e-01
-4.18489307e-01 9.61601734e-01 3.88223797e-01 8.22470009e-01
-2.93562323e-01 3.74710187e-02 2.36608803e-01 1.09395966e-01
-9.08344209e-01 2.28332311e-01 -7.64081240e-01 1.12497225e-01
1.00029850e+00 -5.22060990e-01 -2.57280767e-01 -3.38964552e-01
-5.40314853e-01 -4.68825668e-01 6.30128443e-01 8.72297227e-01
2.66388923e-01 -1.68979585e+00 -5.91764390e-01 -1.29219508e-02
3.53032768e-01 -1.18591475e+00 1.05743766e-01 8.72067392e-01
4.45022248e-02 5.40796220e-01 -1.06899496e-02 1.38216317e-01
-8.12435746e-01 7.98814297e-01 4.02311161e-02 4.84555103e-02
-4.51436162e-01 4.74713683e-01 1.60474494e-01 -5.59458323e-02
-4.60607141e-01 -9.05019939e-01 -3.72422546e-01 9.85290349e-01
1.28628120e-01 3.12158078e-01 -2.87654698e-01 -8.91726017e-01
-4.00893271e-01 6.01571083e-01 1.56641439e-01 -2.41535321e-01
1.75425124e+00 -4.00627583e-01 -6.85201352e-03 1.24850392e+00
1.49237180e+00 4.33559746e-01 -8.63526285e-01 -1.67827696e-01
1.23874106e-01 -6.02314413e-01 2.16667727e-01 -5.74136019e-01
-7.40415096e-01 4.20081198e-01 8.33660841e-01 6.13949537e-01
4.43072259e-01 3.81165981e-01 1.15209293e+00 -2.79924005e-01
8.45138133e-02 -1.10247374e+00 1.42776102e-01 2.22906753e-01
6.77071750e-01 -1.06780028e+00 -1.50328830e-01 -5.72908558e-02
-4.00517315e-01 8.73783469e-01 -1.00486949e-01 -1.95368260e-01
1.08516169e+00 -2.09522173e-01 -2.63379943e-02 -7.96198308e-01
-6.19813561e-01 -3.24201345e-01 3.42767119e-01 3.10174257e-01
7.61053205e-01 5.87207556e-01 -8.16903293e-01 7.62766361e-01
-5.96118093e-01 -4.45160627e-01 7.33813047e-01 6.05011404e-01
-2.82111168e-01 -1.08836102e+00 -1.51526794e-01 5.09907007e-01
-6.87926829e-01 -1.38232186e-01 -5.07412672e-01 1.15868056e+00
2.86626548e-01 7.93293953e-01 7.60058939e-01 -6.36805594e-01
4.87503618e-01 2.93243200e-01 2.80758351e-01 -7.34993398e-01
-6.83830082e-01 -3.48336637e-01 2.74110496e-01 -4.45000678e-01
-4.42778707e-01 -9.96336341e-01 -8.08351696e-01 -3.88325840e-01
-7.81874657e-02 -7.22945854e-02 6.75069869e-01 7.69413471e-01
4.48665082e-01 5.96991897e-01 5.68494201e-01 -1.39753819e-01
-7.07912683e-01 -1.00186586e+00 -6.28949702e-01 9.00676489e-01
2.89585143e-01 -8.29504311e-01 -5.29731393e-01 -1.34374186e-01] | [10.133910179138184, 8.936760902404785] |
faa613ca-9497-4ea9-b35f-f0fc7179ed35 | efficient-ensemble-architecture-for | 2302.13376 | null | https://arxiv.org/abs/2302.13376v1 | https://arxiv.org/pdf/2302.13376v1.pdf | Efficient Ensemble Architecture for Multimodal Acoustic and Textual Embeddings in Punctuation Restoration using Time-Delay Neural Networks | Punctuation restoration plays an essential role in the post-processing procedure of automatic speech recognition, but model efficiency is a key requirement for this task. To that end, we present EfficientPunct, an ensemble method with a multimodal time-delay neural network that outperforms the current best model by 1.0 F1 points, using less than a tenth of its parameters to process embeddings. We streamline a speech recognizer to efficiently output hidden layer latent vectors as audio embeddings for punctuation restoration, as well as BERT to extract meaningful text embeddings. By using forced alignment and temporal convolutions, we eliminate the need for multi-head attention-based fusion, greatly increasing computational efficiency but also raising performance. EfficientPunct sets a new state of the art, in terms of both performance and efficiency, with an ensemble that weights BERT's purely language-based predictions slightly more than the multimodal network's predictions. | ['Homayoon Beigi', 'Xing Yi Liu'] | 2023-02-26 | null | null | null | null | ['punctuation-restoration'] | ['natural-language-processing'] | [ 2.15240315e-01 7.62238950e-02 2.10363299e-01 -3.29781026e-01
-1.09169257e+00 -5.86646378e-01 6.42372489e-01 5.99322431e-02
-8.63811016e-01 2.60316432e-01 6.86288357e-01 -6.00472093e-01
1.38202652e-01 -5.63310497e-02 -5.87298572e-01 -6.95866644e-01
9.79976580e-02 5.23290813e-01 7.22653493e-02 -1.15602493e-01
2.23239940e-02 3.61403435e-01 -1.40533435e+00 4.12501901e-01
6.53208256e-01 7.38082111e-01 2.64146298e-01 1.10576284e+00
-2.09571421e-01 6.27242029e-01 -6.73471034e-01 -4.96225178e-01
-2.54291505e-01 -1.10617183e-01 -6.91554546e-01 -1.16921820e-01
2.79016167e-01 -3.36293489e-01 -5.03671587e-01 6.79991007e-01
8.20442021e-01 5.09188592e-01 6.24452889e-01 -8.46959710e-01
-7.13044941e-01 9.32550669e-01 6.43688962e-02 4.16995585e-01
-6.66986629e-02 1.39418647e-01 1.19638312e+00 -1.39464855e+00
1.26141161e-01 1.32426882e+00 5.48430324e-01 6.86791062e-01
-1.24639249e+00 -3.27843606e-01 2.91662872e-01 2.76752681e-01
-1.25813603e+00 -1.24661577e+00 4.38380748e-01 -1.16712630e-01
1.81737947e+00 4.61095989e-01 1.98072270e-01 1.06427562e+00
1.47128537e-01 9.46200192e-01 4.48624581e-01 -5.29015422e-01
2.43436322e-01 9.35476124e-02 1.56219587e-01 4.76255715e-01
-5.88641822e-01 8.99590179e-02 -7.06405997e-01 -1.06912516e-01
2.96927154e-01 -1.39932156e-01 -1.49839997e-01 3.21229428e-01
-1.06357121e+00 6.96584940e-01 -1.45598082e-03 3.51657033e-01
-4.10588056e-01 2.42935270e-01 4.82967407e-01 3.23356807e-01
7.56319821e-01 4.11693603e-01 -5.81386983e-01 -4.89378631e-01
-1.26950347e+00 -1.96588397e-01 5.81133246e-01 6.45795584e-01
3.20453465e-01 6.14291787e-01 -2.09650904e-01 1.12652683e+00
1.99533835e-01 3.69979352e-01 7.41916001e-01 -8.33789766e-01
5.14661372e-01 1.38004005e-01 -4.30194810e-02 -3.41234028e-01
-2.48426586e-01 -5.42575181e-01 -7.29221284e-01 2.94579472e-02
2.85449950e-03 -3.24156940e-01 -1.22574198e+00 1.67961752e+00
-5.20058498e-02 1.49032265e-01 9.39288363e-02 5.63166499e-01
5.33574998e-01 9.89522099e-01 1.96761325e-01 -9.74725038e-02
1.28043747e+00 -1.26695371e+00 -8.42026651e-01 -3.41014087e-01
6.24033749e-01 -1.12846291e+00 9.58815217e-01 4.10936922e-01
-1.36253595e+00 -5.08399785e-01 -1.03816271e+00 -3.04210126e-01
-4.19732183e-01 2.54826188e-01 2.63961494e-01 5.76507092e-01
-1.48871446e+00 7.37346172e-01 -9.35613871e-01 -2.08188251e-01
-4.29208912e-02 6.59462154e-01 -5.07750988e-01 2.12644577e-01
-1.08224833e+00 1.14149153e+00 4.57823843e-01 1.03369363e-01
-6.20083332e-01 -5.87699413e-01 -8.90026450e-01 5.31215191e-01
1.42687932e-02 -4.87282783e-01 1.65078449e+00 -7.23062098e-01
-1.80007041e+00 2.26273119e-01 -6.39286637e-01 -6.84599996e-01
1.04141787e-01 -2.86840647e-01 -6.27492726e-01 1.48783118e-01
-4.50302243e-01 8.89252305e-01 1.05948544e+00 -6.32546484e-01
-5.20569026e-01 -2.58844215e-02 -5.46275020e-01 3.16562980e-01
-7.48179615e-01 3.96651179e-01 -5.28418064e-01 -6.67753816e-01
-6.03246838e-02 -6.17568731e-01 -3.25813919e-01 -4.43003446e-01
-2.25083068e-01 -5.32125831e-01 8.12578440e-01 -1.12203741e+00
1.36661279e+00 -2.27028203e+00 2.40980938e-01 -2.79818531e-02
1.83846727e-01 6.38112307e-01 -4.74541336e-01 5.45998514e-01
-2.51256168e-01 3.32018465e-01 -2.29022041e-01 -1.38301456e+00
3.34252506e-01 2.25690961e-01 -5.28588474e-01 1.17795154e-01
6.22919023e-01 8.83169830e-01 -4.42461014e-01 -2.68071830e-01
4.91900861e-01 7.85576463e-01 -6.21997416e-01 1.38276324e-01
-1.14010237e-01 -1.62876353e-01 3.03551883e-01 2.87360758e-01
3.61625284e-01 1.72757849e-01 5.29098623e-02 1.86760481e-02
-3.68405581e-01 9.03545856e-01 -8.02332401e-01 1.53664768e+00
-6.46967471e-01 8.99615109e-01 2.26878226e-01 -7.13969290e-01
7.24032223e-01 7.89666057e-01 3.26787710e-01 -5.40780246e-01
2.11979032e-01 3.17096800e-01 -1.13741934e-01 -1.92129418e-01
8.99675012e-01 -1.03862546e-01 1.51558772e-01 5.54458618e-01
3.88905972e-01 1.39507324e-01 -1.48764387e-01 1.64304987e-01
1.19365132e+00 -2.44999260e-01 -1.74088046e-01 5.94971590e-02
1.74878076e-01 -3.78019035e-01 3.20442110e-01 5.34131587e-01
-1.44183561e-01 7.79872417e-01 2.06694245e-01 -1.03724107e-01
-1.31676471e+00 -1.15505350e+00 1.41249090e-01 1.33099484e+00
-6.47094131e-01 -5.34828126e-01 -9.36949849e-01 -4.29134756e-01
-2.39426494e-01 9.98509943e-01 -2.05389008e-01 -2.81540036e-01
-7.35871375e-01 -5.77472031e-01 8.77515554e-01 8.27372611e-01
-1.33230641e-01 -1.11197376e+00 -8.29925463e-02 6.82064533e-01
-3.50364774e-01 -1.11540449e+00 -8.69355142e-01 7.79568851e-01
-6.78988934e-01 -2.55384535e-01 -7.74235249e-01 -7.89835870e-01
4.73983526e-01 1.45705968e-01 8.51037621e-01 -1.45281032e-01
6.96087182e-02 1.60896495e-01 -3.09871793e-01 -3.59183729e-01
-5.35540164e-01 2.81033605e-01 3.41807604e-01 -9.89397094e-02
4.24453199e-01 -6.74409509e-01 -2.76017785e-01 2.82834619e-02
-9.37411368e-01 -3.83310169e-02 6.83751225e-01 9.99283850e-01
2.32413664e-01 -6.46859482e-02 4.49296117e-01 -2.06918776e-01
5.83408892e-01 -1.53281495e-01 -3.33406150e-01 4.02690433e-02
-3.55597466e-01 3.94966871e-01 7.38901556e-01 -5.72910428e-01
-8.99402380e-01 4.49999794e-02 -6.39003754e-01 -6.80906653e-01
-8.31426084e-02 4.35805619e-01 -7.19379559e-02 1.95484951e-01
4.06785041e-01 2.99526155e-01 6.28186017e-02 -9.14069772e-01
5.88668406e-01 1.11008322e+00 6.64274812e-01 -1.64893344e-01
5.64919651e-01 2.56075449e-02 -7.25465715e-01 -1.18359268e+00
-5.53696036e-01 -6.54000461e-01 -6.81520224e-01 7.14056194e-02
7.67154694e-01 -6.73077643e-01 -5.98166645e-01 3.98519337e-01
-1.56239247e+00 -2.48207718e-01 -2.53209561e-01 6.33107066e-01
-3.50353479e-01 4.22613025e-01 -8.98717225e-01 -9.78769243e-01
-7.49859691e-01 -1.05532408e+00 1.00002515e+00 -8.69887173e-02
-3.23175430e-01 -8.64085138e-01 -6.24102056e-02 3.27514410e-01
6.97321951e-01 -6.72796249e-01 8.57476175e-01 -1.04041767e+00
-3.38420928e-01 -1.96036864e-02 -1.39811352e-01 8.43612492e-01
-1.75269768e-01 9.18346047e-02 -1.39970624e+00 -2.02823550e-01
6.69735810e-03 1.27783343e-01 1.17127812e+00 3.43403906e-01
9.08467829e-01 -3.72335643e-01 -9.91971791e-02 3.75361741e-01
8.58210504e-01 2.72430331e-01 5.63166440e-01 -5.81857376e-02
5.55424094e-01 5.70410430e-01 1.64374355e-02 3.02953422e-01
2.09281430e-01 5.29889703e-01 1.14432171e-01 -1.24498636e-01
-4.28565651e-01 -1.95718542e-01 9.83293295e-01 1.71837687e+00
3.50540459e-01 -5.12757778e-01 -8.50392401e-01 8.20940495e-01
-1.76892889e+00 -9.85869348e-01 1.58843577e-01 2.13178706e+00
7.27308333e-01 2.48364940e-01 -4.73275892e-02 2.58464903e-01
6.78266704e-01 9.44678485e-02 -2.90007710e-01 -8.41066122e-01
-7.28468820e-02 5.48646152e-01 3.56165677e-01 9.02401328e-01
-1.13202226e+00 1.24587274e+00 6.77190113e+00 1.04946482e+00
-1.16706288e+00 3.55663300e-01 4.06558871e-01 -3.56717557e-01
-3.88987720e-01 -1.14789836e-01 -9.38275158e-01 3.77334774e-01
1.73091197e+00 2.63575166e-02 6.88547552e-01 6.38157070e-01
3.10154259e-01 2.32180700e-01 -1.07028854e+00 9.96555328e-01
3.58147740e-01 -1.16956973e+00 6.47210935e-03 -4.68181111e-02
3.09038818e-01 4.08994377e-01 2.55833954e-01 3.85837644e-01
2.49737412e-01 -1.18492615e+00 8.47997069e-01 4.02019739e-01
6.85442567e-01 -9.21867013e-01 6.92376494e-01 3.19570631e-01
-1.08645391e+00 1.01879023e-01 -2.52082199e-01 8.96367431e-02
3.89939189e-01 4.07977968e-01 -1.30067289e+00 1.98025584e-01
3.50429654e-01 2.17086583e-01 -3.30144107e-01 1.10485744e+00
-1.33079346e-02 8.91251624e-01 -6.69876456e-01 -1.14950597e-01
4.11624908e-01 2.59228677e-01 5.82126498e-01 1.57649493e+00
4.76324379e-01 -1.85027376e-01 -1.45855278e-01 4.64020908e-01
-9.58774313e-02 -1.85944159e-02 -3.47158521e-01 -3.29881430e-01
6.34733558e-01 9.54439223e-01 -3.94001186e-01 -3.69434446e-01
-8.13514143e-02 1.27901769e+00 3.84491801e-01 3.76961678e-01
-7.34774530e-01 -6.33707702e-01 9.22484756e-01 -2.85193890e-01
5.34235179e-01 -6.97696209e-01 -2.26608947e-01 -1.01387179e+00
-2.64258143e-02 -6.37999475e-01 -1.02317870e-01 -7.86147833e-01
-1.08018804e+00 1.00439703e+00 -4.06228691e-01 -8.77952635e-01
-5.60031354e-01 -7.81013668e-01 -6.10881031e-01 1.09454322e+00
-1.50765026e+00 -1.16348290e+00 4.14318502e-01 3.62603515e-01
8.72942388e-01 -1.77268401e-01 1.10503840e+00 5.37640274e-01
-7.10681558e-01 9.00791526e-01 3.52507979e-01 1.28583744e-01
7.29501128e-01 -1.24225903e+00 9.53070819e-01 1.13158357e+00
4.79618788e-01 6.74657881e-01 6.58205032e-01 -2.64271110e-01
-1.09214485e+00 -1.00982463e+00 1.64300871e+00 -5.23209512e-01
7.01847255e-01 -5.39786398e-01 -8.70816171e-01 5.99166632e-01
5.88995874e-01 -3.91636312e-01 7.84600973e-01 2.39814326e-01
-2.95818299e-01 1.75323598e-02 -5.94436169e-01 1.00357449e+00
7.97406316e-01 -9.59416330e-01 -7.70251453e-01 1.17172323e-01
1.14121366e+00 -2.42089897e-01 -7.06326842e-01 5.35383448e-02
3.94118428e-01 -5.73494613e-01 8.62142503e-01 -5.56721270e-01
1.52627584e-02 -1.79251730e-01 -2.97530711e-01 -1.50591111e+00
-4.75158602e-01 -9.81868923e-01 -8.30097497e-02 1.42226803e+00
7.33626008e-01 -6.10423028e-01 5.71988225e-01 5.50723851e-01
-8.20258021e-01 -5.96915245e-01 -1.32643628e+00 -7.42230952e-01
5.65784387e-02 -9.62955475e-01 5.45862257e-01 5.11900127e-01
1.17959216e-01 6.32342935e-01 -4.27394986e-01 1.84218153e-01
1.89728379e-01 -6.12262070e-01 2.28110939e-01 -8.88448656e-01
-2.02295363e-01 -6.09426677e-01 -2.31766284e-01 -1.24622548e+00
3.04394692e-01 -8.23111057e-01 2.22355962e-01 -1.32145464e+00
-2.97285408e-01 -1.36200860e-01 -5.31897008e-01 7.86333263e-01
3.31784375e-02 1.72457606e-01 2.85715401e-01 -1.23032756e-01
-5.08867383e-01 8.48440111e-01 5.34871697e-01 -1.69768125e-01
-2.78376162e-01 -2.01277152e-01 -5.00672996e-01 6.07842207e-01
9.47824717e-01 -3.34411830e-01 -2.15160489e-01 -8.61104906e-01
-2.17730805e-01 -5.93532287e-02 2.83536583e-01 -9.87925589e-01
5.46001077e-01 3.06389421e-01 2.94175267e-01 -7.67408967e-01
9.93797064e-01 -4.55494493e-01 -1.88010544e-01 7.57461116e-02
-3.97288442e-01 3.47846001e-01 4.27837610e-01 1.53582782e-01
-3.34621042e-01 -2.94245750e-01 6.07402742e-01 2.58748621e-01
-5.78253686e-01 1.88132748e-01 -8.54024649e-01 -2.50340194e-01
3.51043910e-01 -1.01993583e-01 -3.32270771e-01 -3.25675786e-01
-1.01911747e+00 2.44139247e-02 1.56639516e-01 6.28471732e-01
7.07036793e-01 -1.17922342e+00 -6.64265990e-01 4.97706115e-01
-2.94223487e-01 -3.81102353e-01 3.76005083e-01 8.67408156e-01
-7.09668696e-02 6.08149588e-01 3.20865422e-01 -3.02273273e-01
-1.46700823e+00 2.35510245e-01 2.67808259e-01 -1.17740318e-01
-5.84775746e-01 1.07652807e+00 -1.25119254e-01 -4.79333043e-01
5.11846185e-01 -4.11758840e-01 -9.60852131e-02 1.33859843e-01
7.61053681e-01 1.69282526e-01 6.33204639e-01 -5.76157093e-01
-3.90855998e-01 -2.10388284e-02 -4.56801653e-01 -7.31116772e-01
1.30765641e+00 -1.40069246e-01 6.88720718e-02 3.55415165e-01
1.29530072e+00 -2.18113605e-02 -1.17693889e+00 -2.38735080e-01
1.05110042e-01 1.94813330e-02 3.76016796e-01 -9.47253168e-01
-5.53084731e-01 1.28527296e+00 4.46688652e-01 1.07770965e-01
9.41041768e-01 -9.23178419e-02 1.18793619e+00 3.39580566e-01
-1.41058594e-01 -1.28911293e+00 5.61606102e-02 1.13027453e+00
9.11955774e-01 -8.66144955e-01 -6.12888992e-01 1.73335522e-02
-6.13211274e-01 1.06429493e+00 1.92515463e-01 1.37629658e-01
5.75982571e-01 5.84675252e-01 8.32816958e-02 3.25172514e-01
-1.24655116e+00 -2.02447623e-01 3.56519461e-01 5.42155683e-01
4.69889194e-01 1.68843627e-01 2.33243152e-01 7.14320481e-01
-4.30094510e-01 -5.62592030e-01 1.59729898e-01 5.93394876e-01
-5.90053320e-01 -1.27144694e+00 -3.53486449e-01 3.99007767e-01
-4.44429368e-01 -7.34232605e-01 -2.56774247e-01 2.37378746e-01
-1.96017414e-01 1.20529854e+00 3.65726590e-01 -6.17063463e-01
2.58333862e-01 6.05986297e-01 3.99108455e-02 -6.22675002e-01
-8.62528741e-01 4.28690702e-01 3.30555469e-01 -2.59545058e-01
1.59468457e-01 -6.43210351e-01 -1.17310429e+00 -4.41512108e-01
-4.05837506e-01 2.35199258e-01 9.23576951e-01 1.13310361e+00
6.57997608e-01 6.24541581e-01 5.53134024e-01 -1.10936153e+00
-7.50371516e-01 -1.23355412e+00 -1.97988495e-01 -2.41047293e-01
5.83298743e-01 -2.66961426e-01 -4.91202116e-01 6.27228394e-02] | [14.429940223693848, 6.652266979217529] |
bd8bdb1e-a5ce-44e5-8942-1bef812ac9d6 | a-comparative-study-of-pre-trained-encoders-1 | 2204.04980 | null | https://arxiv.org/abs/2204.04980v1 | https://arxiv.org/pdf/2204.04980v1.pdf | A Comparative Study of Pre-trained Encoders for Low-Resource Named Entity Recognition | Pre-trained language models (PLM) are effective components of few-shot named entity recognition (NER) approaches when augmented with continued pre-training on task-specific out-of-domain data or fine-tuning on in-domain data. However, their performance in low-resource scenarios, where such data is not available, remains an open question. We introduce an encoder evaluation framework, and use it to systematically compare the performance of state-of-the-art pre-trained representations on the task of low-resource NER. We analyze a wide range of encoders pre-trained with different strategies, model architectures, intermediate-task fine-tuning, and contrastive learning. Our experimental results across ten benchmark NER datasets in English and German show that encoder performance varies significantly, suggesting that the choice of encoder for a specific low-resource scenario needs to be carefully evaluated. | ['Leonhard Hennig', 'Christoph Alt', 'Arne Binder', 'Jonas Mikkelsen', 'Yuxuan Chen'] | 2022-04-11 | null | https://aclanthology.org/2022.repl4nlp-1.6 | https://aclanthology.org/2022.repl4nlp-1.6.pdf | repl4nlp-acl-2022-5 | ['low-resource-named-entity-recognition'] | ['natural-language-processing'] | [ 7.45608360e-02 1.70370508e-02 -1.34229332e-01 -4.40484554e-01
-9.91578698e-01 -5.98664463e-01 8.53226483e-01 1.27932221e-01
-1.22566617e+00 8.64494383e-01 6.90788686e-01 -1.34816647e-01
3.73186618e-02 -7.49445379e-01 -4.14861739e-01 -8.30613598e-02
-3.42740417e-02 6.04831219e-01 2.92932183e-01 -4.57931936e-01
-7.96855241e-02 5.53485632e-01 -1.29704165e+00 1.65054098e-01
6.66757822e-01 5.31463206e-01 2.40080923e-01 7.73392677e-01
-5.61573386e-01 8.91445875e-01 -7.16537893e-01 -7.94026077e-01
5.47428876e-02 -3.42057735e-01 -1.05844462e+00 -4.03314233e-01
1.33628950e-01 1.60358809e-02 -4.01929915e-01 7.57690191e-01
1.06098795e+00 5.05243480e-01 5.90235591e-01 -4.86688167e-01
-1.02217638e+00 7.48544753e-01 2.85380810e-01 7.38654673e-01
1.76098160e-02 1.18612908e-01 9.92413104e-01 -9.78482127e-01
9.74642038e-01 8.39937389e-01 1.07110059e+00 9.58691001e-01
-1.05337703e+00 -4.69365448e-01 -2.52337962e-01 9.39742625e-02
-1.30315351e+00 -8.11409771e-01 2.42541060e-01 -1.90358192e-01
1.68845594e+00 -1.79146096e-01 -9.58486646e-02 1.50726128e+00
2.16379669e-02 4.36237067e-01 7.53404677e-01 -5.58464050e-01
3.20366085e-01 2.15078015e-02 2.28894353e-01 4.17556077e-01
4.34785813e-01 1.99391395e-01 -2.89407313e-01 -1.52972430e-01
4.27643865e-01 -3.11030298e-01 -2.83838540e-01 9.55028366e-03
-1.07975733e+00 8.91205847e-01 3.04564267e-01 9.52674448e-01
-6.71280444e-01 -1.08072802e-01 6.67276025e-01 2.78675973e-01
5.84798038e-01 9.99271750e-01 -1.01068974e+00 -5.08141816e-01
-1.26671720e+00 -2.27811664e-01 1.19239068e+00 1.04487717e+00
7.44069874e-01 2.31687889e-01 -5.48724890e-01 1.07088852e+00
-3.42379987e-01 -6.10017069e-02 7.99829543e-01 -5.64001977e-01
6.82240784e-01 2.59322941e-01 1.90975383e-01 -7.90744051e-02
-5.20801187e-01 -2.11325243e-01 -8.72182608e-01 -1.31095499e-01
2.62880743e-01 -8.02727580e-01 -1.23523486e+00 1.69881785e+00
9.88683291e-03 3.59335482e-01 5.78687310e-01 5.03733754e-01
1.08440256e+00 6.53393149e-01 7.24213660e-01 4.11111005e-02
1.48012614e+00 -1.04445398e+00 -7.80810058e-01 -5.33137739e-01
7.80251980e-01 -5.29175818e-01 8.76966000e-01 -3.39595616e-01
-8.61940384e-01 -7.35834599e-01 -8.24799418e-01 -3.75708073e-01
-1.03399384e+00 3.45171124e-01 2.93833107e-01 6.97685778e-01
-8.57658327e-01 9.06865299e-01 -7.53910899e-01 -8.17188144e-01
2.68825114e-01 1.12991422e-01 -5.36373556e-01 -2.53499061e-01
-1.76560831e+00 1.47734356e+00 8.79404306e-01 -1.34293988e-01
-9.25171912e-01 -1.07252574e+00 -1.16764319e+00 4.83731985e-01
2.16895506e-01 -5.67630827e-01 1.33614397e+00 -3.48062873e-01
-1.38386095e+00 9.59982514e-01 1.35331556e-01 -5.57205677e-01
3.11562866e-01 -3.91957313e-01 -9.03920054e-01 -1.49143547e-01
2.72071123e-01 6.51812494e-01 3.43091458e-01 -8.34098876e-01
-4.37031686e-01 2.31424138e-01 2.84594715e-01 4.26264182e-02
-4.43282127e-01 5.13042748e-01 -1.81876317e-01 -5.73463023e-01
-1.05792642e+00 -4.93410677e-01 -5.98841131e-01 -5.62262774e-01
-1.61929220e-01 -2.94829488e-01 3.73164654e-01 -7.07194149e-01
1.36914325e+00 -2.18479943e+00 -2.49779671e-01 -4.62864637e-01
-3.63318741e-01 7.39301682e-01 -5.57647705e-01 8.44656289e-01
-2.18439594e-01 5.12473702e-01 -2.46653050e-01 -4.68345076e-01
1.25684947e-01 2.06959635e-01 2.39478331e-03 -1.57734659e-02
7.40206838e-01 9.64627206e-01 -1.14832926e+00 -5.34226179e-01
1.91555426e-01 7.10201621e-01 -2.33837321e-01 6.64390445e-01
1.17375173e-01 1.46814823e-01 -2.86267787e-01 3.45598340e-01
2.23537356e-01 -2.80673265e-01 1.08523853e-03 -1.71994686e-01
-3.63805056e-01 6.18854284e-01 -1.06775832e+00 1.95089209e+00
-7.20074773e-01 5.10252297e-01 -1.94504514e-01 -6.86795413e-01
8.43802989e-01 6.51682556e-01 6.21009842e-02 -6.90903902e-01
1.18070692e-01 1.78340718e-01 -4.91523109e-02 -4.36624408e-01
7.05376208e-01 -4.12357897e-01 -2.98843205e-01 2.26623744e-01
1.09745407e+00 3.46599668e-01 8.17553163e-01 -1.87176734e-01
1.59656751e+00 6.53756736e-03 8.06201279e-01 -1.69841528e-01
2.33725294e-01 2.05062658e-01 5.83628774e-01 9.45053816e-01
-3.90494198e-01 7.30503559e-01 4.57745306e-02 -4.25956309e-01
-1.21547103e+00 -7.42069960e-01 -3.78513813e-01 1.45902109e+00
-2.96315700e-01 -4.92268860e-01 -7.43046939e-01 -9.80138779e-01
-5.60435593e-01 1.25536668e+00 -7.68007696e-01 -8.22766200e-02
-6.13876760e-01 -9.71738160e-01 1.03228128e+00 7.12234735e-01
4.42451566e-01 -1.55321205e+00 -9.27593291e-01 6.08569264e-01
7.15842173e-02 -1.36207902e+00 -3.23053330e-01 8.70113611e-01
-6.33208871e-01 -9.14973140e-01 -1.05649531e+00 -6.73400581e-01
2.22843170e-01 -5.71838431e-02 1.62852430e+00 -2.54108161e-01
-2.18211055e-01 5.15607238e-01 -7.09005177e-01 -2.94284105e-01
-4.56161916e-01 6.53281510e-01 -1.88504279e-01 -5.50606549e-01
7.03565359e-01 -3.32380861e-01 -2.46242315e-01 3.57689261e-02
-8.55119169e-01 -4.95864689e-01 8.82801712e-01 1.22138035e+00
2.58433878e-01 3.09121003e-03 5.37869930e-01 -1.12034404e+00
7.60419011e-01 -7.13839054e-01 -1.60445616e-01 6.41690671e-01
-1.93208888e-01 3.82736623e-01 7.42078304e-01 -4.72077608e-01
-1.52273703e+00 -1.42015100e-01 -4.87271637e-01 -3.80330384e-01
-5.97003102e-01 6.52062774e-01 -1.72980636e-01 3.59997958e-01
1.10814404e+00 -6.69770613e-02 -8.94950449e-01 -9.62652802e-01
5.83366454e-01 6.22765064e-01 5.16803801e-01 -5.11302233e-01
5.30015647e-01 2.99976896e-02 -5.96793473e-01 -9.29362953e-01
-1.08214974e+00 -6.44645035e-01 -9.68859732e-01 3.16303611e-01
1.12890220e+00 -1.09545898e+00 3.48246276e-01 -1.08037800e-01
-1.43625855e+00 -5.05647004e-01 -6.88680768e-01 5.03553092e-01
-2.15830445e-01 3.09414063e-02 -7.96608388e-01 -7.05371141e-01
-6.09763503e-01 -8.46379817e-01 9.87537086e-01 3.55144024e-01
-2.76066571e-01 -1.22422945e+00 6.14166260e-01 -1.47714898e-01
5.95214784e-01 -1.48139089e-01 6.94843233e-01 -1.51731718e+00
2.29466274e-01 -2.17898831e-01 -1.36094823e-01 3.06720078e-01
-5.17173149e-02 -4.15414631e-01 -1.20211387e+00 -1.85098663e-01
-3.46646816e-01 -5.68764091e-01 8.87426138e-01 -9.61770676e-03
7.21405566e-01 -2.17332896e-02 -2.45975405e-01 4.37877268e-01
1.54816997e+00 -7.64023662e-02 7.03970671e-01 5.08171439e-01
3.78707290e-01 4.48177218e-01 4.32406396e-01 3.58955324e-01
2.10574716e-01 3.41005474e-01 -9.82779488e-02 -4.58292253e-02
-3.81355643e-01 -2.78302789e-01 3.07357490e-01 8.17099214e-01
-3.04305911e-01 -4.50334698e-01 -1.11881638e+00 8.83246422e-01
-1.41741014e+00 -1.17247605e+00 3.59414101e-01 1.81010020e+00
9.84616816e-01 6.97293580e-02 -1.06287926e-01 -3.84021163e-01
9.78623033e-01 2.98059434e-01 -4.84344780e-01 -5.38962960e-01
-1.44071057e-02 7.89009333e-01 6.58700764e-01 5.52288955e-03
-1.34900653e+00 1.15380073e+00 6.85504770e+00 8.76667142e-01
-7.89771199e-01 7.40392268e-01 3.24217886e-01 1.94555223e-01
3.44269201e-02 1.49466619e-02 -1.14745128e+00 3.01431298e-01
1.78402078e+00 -2.05700368e-01 1.79820463e-01 1.05999959e+00
-3.20737869e-01 4.59032983e-01 -1.04713583e+00 6.18370414e-01
6.93970174e-02 -1.34063041e+00 -1.93180040e-01 -1.94788679e-01
7.59924114e-01 8.31660211e-01 -6.77908659e-01 1.23375762e+00
7.59407997e-01 -9.69174683e-01 2.63347447e-01 4.59325969e-01
1.06279469e+00 -5.95805883e-01 1.12060893e+00 3.42395216e-01
-1.21081197e+00 1.28334031e-01 -7.61600435e-01 2.65347540e-01
5.44408381e-01 3.87484938e-01 -6.71787739e-01 6.61537409e-01
7.84983158e-01 4.14184839e-01 -6.29463077e-01 1.08343625e+00
-2.91214675e-01 6.92702830e-01 -1.46085978e-01 -3.41530442e-02
4.16726410e-01 4.60339576e-01 2.32775122e-01 2.02660561e+00
3.52885306e-01 9.93343517e-02 -1.26149267e-01 6.95544124e-01
-4.91168082e-01 2.19253346e-01 -5.96497953e-01 -4.63483393e-01
5.56453764e-01 1.42659259e+00 -4.60072935e-01 -5.15578985e-01
-7.96839356e-01 1.13553405e+00 8.87456000e-01 2.62682021e-01
-6.88274026e-01 -7.83139288e-01 6.98644757e-01 -9.47955027e-02
6.77691936e-01 -1.31922126e-01 4.88600582e-02 -1.53060627e+00
-5.65068841e-01 -5.20178735e-01 7.88032353e-01 -6.34403944e-01
-1.79396069e+00 1.06362569e+00 -1.13841735e-01 -9.68943119e-01
-3.71827841e-01 -7.76003957e-01 -8.45351338e-01 7.69551456e-01
-1.84475636e+00 -1.04209244e+00 -5.59066646e-02 4.33510005e-01
7.09671378e-01 -2.72382408e-01 1.37048292e+00 5.24042547e-01
-5.59881687e-01 5.97080588e-01 1.74572274e-01 6.51083589e-01
8.01792204e-01 -1.23923469e+00 8.99186552e-01 9.50640976e-01
3.67671460e-01 5.97531140e-01 3.80032063e-01 -5.38090110e-01
-8.46498132e-01 -1.33237040e+00 1.45345247e+00 -6.63569331e-01
6.84332252e-01 -4.50903505e-01 -1.05115461e+00 6.87602758e-01
5.82448959e-01 3.14703882e-01 1.09648752e+00 4.82522517e-01
-3.91166270e-01 3.64267081e-01 -1.03376842e+00 4.36299413e-01
1.15974987e+00 -6.37827396e-01 -1.28897321e+00 -1.63363852e-02
7.42571533e-01 -2.10497633e-01 -1.23781407e+00 4.45962042e-01
1.29260495e-01 -6.89374626e-01 9.66850817e-01 -1.05488300e+00
3.95809591e-01 1.36826649e-01 -6.75849617e-02 -1.54768455e+00
-6.19132936e-01 -3.84307772e-01 -1.06708057e-01 1.71917343e+00
6.07429564e-01 -1.06353886e-01 2.66306728e-01 6.44065619e-01
-2.92390168e-01 -3.64803612e-01 -9.00901735e-01 -9.74130929e-01
2.57324040e-01 -3.65775287e-01 4.41701859e-01 1.21841705e+00
1.46112712e-02 8.06997895e-01 -4.06405509e-01 -4.81444411e-03
1.54949099e-01 -4.87990648e-01 4.04806703e-01 -1.18358910e+00
-8.11305493e-02 -2.24503115e-01 -2.63019353e-01 -4.61792171e-01
5.08260429e-01 -8.01721632e-01 2.09151864e-01 -1.75796747e+00
1.82205245e-01 -2.30533883e-01 -7.70919085e-01 8.26867759e-01
-3.29740196e-01 7.03146076e-03 2.39208087e-01 1.22280428e-02
-8.21090162e-01 6.19273067e-01 6.05382323e-01 6.01354465e-02
-1.39997572e-01 -5.39868057e-01 -6.15884781e-01 5.33657134e-01
6.41115010e-01 -8.54717195e-01 -8.76846388e-02 -5.13687551e-01
-3.65065485e-02 -1.78262025e-01 -8.71336982e-02 -1.18412662e+00
2.50035465e-01 -1.20376227e-02 4.71549720e-01 -1.64304301e-01
2.24381700e-01 -5.50129533e-01 -1.07472986e-01 2.07743704e-01
-5.10279894e-01 1.84807509e-01 2.70315766e-01 3.89121592e-01
-3.20190340e-01 -8.13705087e-01 8.97121727e-01 -4.72701430e-01
-1.40102363e+00 2.57386297e-01 -3.06777447e-01 8.57239544e-01
7.00010657e-01 7.39888102e-02 -1.86501667e-01 4.82445583e-03
-6.61323130e-01 -1.45258546e-01 1.38764456e-01 6.63664639e-01
1.64224640e-01 -1.30950272e+00 -8.83393943e-01 -1.95548967e-01
4.80592459e-01 -4.36270684e-01 3.49257380e-01 3.07551265e-01
-9.27221105e-02 4.35580134e-01 -2.78276682e-01 1.57591924e-01
-6.23513401e-01 6.22380018e-01 2.87314266e-01 -9.04648781e-01
-6.32700562e-01 6.78408682e-01 -1.17767230e-01 -7.44012237e-01
-8.45193863e-02 -5.44575509e-04 -5.03056824e-01 2.80742675e-01
7.51584947e-01 3.74008089e-01 2.77273327e-01 -6.89591885e-01
-3.75937730e-01 4.47525717e-02 -3.27295139e-02 -8.72087106e-02
1.58812344e+00 6.64718524e-02 4.18454677e-01 4.86489892e-01
1.24395800e+00 -3.20780545e-01 -9.55604613e-01 -2.86033034e-01
5.19756377e-01 1.00829862e-01 1.03949457e-02 -8.89791965e-01
-6.06631160e-01 1.06090498e+00 3.98676455e-01 -1.41848579e-01
8.87064815e-01 -1.09008722e-01 7.87848592e-01 7.38785505e-01
4.51347947e-01 -1.32008922e+00 -1.59789443e-01 1.10443258e+00
4.82213825e-01 -1.19098067e+00 -3.01858783e-01 1.15013145e-01
-8.93279970e-01 1.11705971e+00 6.63729429e-01 -2.63430387e-01
7.96228647e-01 3.79932135e-01 -1.73917692e-02 -2.84030735e-02
-7.69898295e-01 -9.25104022e-01 3.18861932e-01 5.99475384e-01
8.11350942e-01 -2.56675422e-01 -3.36972326e-01 1.08879924e+00
-4.03060652e-02 2.49574751e-01 3.76064330e-01 1.06254375e+00
-3.21378917e-01 -1.05226183e+00 5.89313842e-02 3.55691403e-01
-9.02916551e-01 -6.92018390e-01 -1.74954712e-01 9.55613375e-01
2.95447350e-01 8.49126399e-01 -1.40924882e-02 -6.27416298e-02
6.12395406e-01 7.04613686e-01 1.83377981e-01 -1.04655409e+00
-9.54871535e-01 -4.60011393e-01 7.94210374e-01 -3.47789049e-01
-5.34454346e-01 -3.93433899e-01 -9.83357191e-01 1.28614530e-01
-3.28916192e-01 3.87277931e-01 4.80685800e-01 1.07201064e+00
6.53213263e-01 5.84192753e-01 -1.56486575e-02 -9.18991923e-01
-5.75471163e-01 -1.24224305e+00 -3.77100825e-01 5.66879630e-01
-8.68695155e-02 -7.02565014e-01 -2.74428457e-01 -3.08268936e-03] | [9.68366527557373, 9.358780860900879] |
8cc662ad-f626-48b6-8857-5b1c9207cc31 | implicit-models-latent-compression-intrinsic | 2210.09186 | null | https://arxiv.org/abs/2210.09186v6 | https://arxiv.org/pdf/2210.09186v6.pdf | Implicit models, latent compression, intrinsic biases, and cheap lunches in community detection | The task of community detection, which aims to partition a network into clusters of nodes to summarize its large-scale structure, has spawned the development of many competing algorithms with varying objectives. Some community detection methods are inferential, explicitly deriving the clustering objective through a probabilistic generative model, while other methods are descriptive, dividing a network according to an objective motivated by a particular application, making it challenging to compare these methods on the same scale. Here we present a solution to this problem that associates any community detection objective, inferential or descriptive, with its corresponding implicit network generative model. This allows us to compute the description length of a network and its partition under arbitrary objectives, providing a principled measure to compare the performance of different algorithms without the need for "ground truth" labels. Our approach also gives access to instances of the community detection problem that are optimal to any given algorithm, and in this way reveals intrinsic biases in popular descriptive methods, explaining their tendency to overfit. Using our framework, we compare a number of community detection methods on artificial networks, and on a corpus of over 500 structurally diverse empirical networks. We find that more expressive community detection methods exhibit consistently superior compression performance on structured data instances, without having degraded performance on a minority of situations where more specialized algorithms perform optimally. Our results undermine the implications of the "no free lunch" theorem for community detection, both conceptually and in practice, since it is confined to unstructured data instances, unlike relevant community detection problems which are structured by requirement. | ['Alec Kirkley', 'Tiago P. Peixoto'] | 2022-10-17 | null | null | null | null | ['community-detection'] | ['graphs'] | [ 4.71483886e-01 3.54357362e-01 -1.76146746e-01 -1.11317396e-01
-4.20012087e-01 -9.65956688e-01 8.56741607e-01 4.69608009e-01
-2.70221025e-01 6.78390563e-01 3.28779578e-01 -5.59462726e-01
-5.89877367e-01 -9.64601219e-01 -1.75522193e-01 -9.34204280e-01
-5.13133228e-01 1.03330278e+00 3.42296898e-01 1.33664757e-01
4.94811714e-01 3.87749463e-01 -1.26877534e+00 1.93034708e-01
6.67594373e-01 4.40660983e-01 1.63600184e-02 7.28135228e-01
-1.34072676e-01 6.79884076e-01 -4.63536024e-01 -5.94636500e-01
1.30736334e-02 -4.33114976e-01 -1.19261456e+00 3.39939743e-01
9.30852294e-02 3.55657756e-01 -2.56831795e-01 1.16674578e+00
3.30610693e-01 -2.25908309e-01 1.22663176e+00 -1.42887831e+00
-3.06137174e-01 1.05005240e+00 -6.80630803e-01 1.92570329e-01
4.01211321e-01 1.61666498e-02 1.63695467e+00 -4.80425328e-01
8.12117279e-01 1.35721767e+00 8.99192512e-01 1.77893505e-01
-2.01137972e+00 -3.95224601e-01 7.34389275e-02 -1.58144236e-01
-1.62100613e+00 -5.12979150e-01 6.33935988e-01 -9.15966988e-01
6.15449131e-01 3.85631025e-01 5.96608877e-01 9.35865521e-01
-1.63394332e-01 4.13107038e-01 1.00819802e+00 -2.38667861e-01
3.32461357e-01 4.72091958e-02 1.75920814e-01 5.81591070e-01
1.00400519e+00 -2.32798919e-01 -2.69650035e-02 -6.44932866e-01
4.59944427e-01 -8.52173418e-02 -3.73706430e-01 -7.67618895e-01
-1.14040065e+00 1.16719949e+00 2.09872186e-01 3.78284216e-01
-6.33407533e-02 5.58556803e-02 5.01802444e-01 1.95588037e-01
3.54027778e-01 5.78417897e-01 -1.04473364e-02 2.64250666e-01
-1.17659211e+00 1.83146581e-01 1.34143531e+00 8.36928785e-01
8.24125588e-01 -3.31956267e-01 1.53281286e-01 3.94860327e-01
4.33573455e-01 -1.54169258e-02 9.71880555e-02 -1.10248685e+00
2.84285933e-01 5.41450620e-01 -2.70573407e-01 -1.44671535e+00
-3.24571073e-01 -6.92082703e-01 -1.23805690e+00 2.91360337e-02
6.57926500e-01 2.66263392e-02 -3.77870023e-01 1.75049031e+00
2.18264852e-02 -3.71894747e-01 -1.13890849e-01 5.18324018e-01
4.87491429e-01 3.01137120e-01 -9.35368240e-02 -5.85535944e-01
1.21663368e+00 -4.37674850e-01 -3.15176725e-01 6.57639205e-02
6.64335966e-01 -6.12682045e-01 6.30821586e-01 3.66558224e-01
-8.68785918e-01 -7.18393538e-04 -9.69090402e-01 3.85626644e-01
-1.78910404e-01 -5.52612185e-01 8.00152242e-01 7.61659443e-01
-1.25501537e+00 7.42169797e-01 -5.20324111e-01 -6.19749665e-01
4.55352575e-01 3.15306842e-01 -3.06685358e-01 7.95753002e-02
-7.76179969e-01 6.91877186e-01 7.16405034e-01 -2.00942948e-01
-7.33786821e-01 -2.19918430e-01 -5.34181595e-01 2.38350391e-01
7.50095785e-01 -7.04067647e-01 7.26124704e-01 -1.10397387e+00
-6.74423993e-01 1.06613743e+00 -1.48337577e-02 -5.20223260e-01
5.26632190e-01 7.11148024e-01 -1.79877102e-01 3.08229744e-01
3.10043514e-01 4.38620299e-01 6.36505544e-01 -1.53744185e+00
-2.88463891e-01 -2.01898500e-01 -8.09110776e-02 2.28461158e-02
-2.40531445e-01 -1.31016225e-02 -2.63933957e-01 -4.72854316e-01
4.44033772e-01 -6.90191090e-01 -5.43992579e-01 -9.89911556e-02
-8.19647074e-01 -3.65496784e-01 5.55694997e-01 -1.36496946e-01
1.35862005e+00 -1.87426269e+00 3.19158643e-01 8.69688511e-01
1.03662014e+00 -2.36446992e-01 1.72543507e-02 8.26452315e-01
-2.72527575e-01 7.60436594e-01 -6.22044921e-01 -2.69156307e-01
1.10845447e-01 2.90657729e-01 -1.80112436e-01 7.70849168e-01
1.44761071e-01 4.99283195e-01 -1.06169522e+00 -9.87522423e-01
-3.05488199e-01 1.32946879e-01 -7.15862691e-01 -2.44233638e-01
8.12262893e-02 -1.56232476e-01 -2.87733376e-01 5.35503149e-01
2.87436306e-01 -7.96488166e-01 9.19629574e-01 2.44357232e-02
5.96130788e-02 3.24756414e-01 -1.48689997e+00 1.14456308e+00
5.12764812e-01 8.18958461e-01 3.49083573e-01 -1.34026372e+00
7.16691554e-01 2.97070414e-01 5.66134036e-01 1.52204111e-01
4.72910292e-02 1.85586646e-01 4.80413049e-01 -1.82521224e-01
3.86328459e-01 -2.08219171e-01 6.38301298e-02 9.23924804e-01
-3.22556496e-02 -2.41590310e-02 6.46405816e-01 7.76002407e-01
1.59955311e+00 -5.05437613e-01 5.23455977e-01 -7.03934193e-01
1.94015205e-01 1.26737714e-01 4.58631456e-01 1.05677617e+00
-2.58809030e-01 6.64783835e-01 1.05882454e+00 -1.45486906e-01
-1.25410914e+00 -1.23604119e+00 -2.28744715e-01 8.28773141e-01
-2.42946930e-02 -6.70967817e-01 -7.09654272e-01 -5.49609900e-01
8.40784684e-02 4.50830758e-02 -6.50833309e-01 2.72632211e-01
-1.82612702e-01 -1.06888497e+00 6.07894719e-01 4.79992926e-02
-5.11568040e-02 -7.81477451e-01 -3.18633467e-01 3.19998801e-01
-3.76395017e-01 -9.05707538e-01 -2.30387136e-01 3.89595032e-01
-1.03316104e+00 -1.53811479e+00 -3.43128502e-01 -6.57779038e-01
8.49204719e-01 1.73364520e-01 1.52728248e+00 4.96928692e-01
-3.04201186e-01 2.08919898e-01 -1.75863922e-01 -2.64062323e-02
-7.98302531e-01 4.40545678e-01 2.15257496e-01 -7.85611495e-02
3.74776810e-01 -9.83258009e-01 -3.55296761e-01 2.04198375e-01
-8.80561471e-01 -1.45350471e-01 6.86353266e-01 6.28086090e-01
1.87886730e-01 5.49950302e-01 3.54170531e-01 -1.03266191e+00
8.95463288e-01 -7.41926730e-01 -3.27991933e-01 8.88315588e-02
-8.61911774e-01 2.85169095e-01 3.57494742e-01 -3.08903545e-01
-3.50212246e-01 -1.00498334e-01 4.08518642e-01 -4.37003486e-02
8.54377635e-03 8.60822141e-01 -9.02016610e-02 1.80978730e-01
9.30170536e-01 1.69336721e-01 3.48743796e-01 -4.27117586e-01
2.31155798e-01 5.28961182e-01 5.52391648e-01 -8.15646112e-01
9.65752125e-01 5.83272398e-01 9.09535885e-02 -9.87346470e-01
-6.44814551e-01 -7.73701847e-01 -7.34507620e-01 -5.59695028e-02
4.62256998e-01 -6.76738977e-01 -8.04845095e-01 -1.47538632e-01
-1.04530561e+00 8.69470239e-02 -1.48407385e-01 1.94370389e-01
-4.70412493e-01 8.42107654e-01 -5.21236122e-01 -9.72637594e-01
-9.73741040e-02 -9.06713128e-01 5.94014585e-01 -4.05679792e-01
-6.63618445e-01 -1.22622573e+00 2.16028899e-01 7.96296969e-02
8.01759213e-02 5.48417509e-01 1.00713289e+00 -9.54448462e-01
-5.33313215e-01 -4.01795767e-02 -5.05325854e-01 -2.00594222e-04
2.11072173e-02 5.54871738e-01 -7.52801955e-01 -5.79001725e-01
-2.26971388e-01 5.47878668e-02 9.73967612e-01 3.29709798e-01
9.05612886e-01 -7.00846374e-01 -7.14909017e-01 2.52394557e-01
1.60842299e+00 -3.02044094e-01 6.01543844e-01 3.82510349e-02
5.20086288e-01 9.71393764e-01 -1.97058827e-01 2.98697203e-01
9.72376913e-02 2.99745053e-01 4.35301512e-01 1.22445533e-02
1.39611617e-01 -1.23670600e-01 1.43502161e-01 8.79748642e-01
-3.78249407e-01 -3.02886099e-01 -1.24711716e+00 6.98704183e-01
-1.83302820e+00 -1.50457895e+00 -5.43750048e-01 2.17469144e+00
1.09253049e+00 6.04350388e-01 5.55500925e-01 3.63898158e-01
1.09080601e+00 3.89773287e-02 -1.57154411e-01 1.40875885e-02
-2.80771613e-01 -1.80946721e-03 3.04057598e-01 4.47969437e-01
-9.86535907e-01 3.59680831e-01 7.37937832e+00 9.84605849e-01
-5.53037047e-01 -1.17645845e-01 5.86480081e-01 2.25509286e-01
-4.77545112e-01 4.82567400e-01 -4.07660186e-01 3.83665860e-01
9.27922428e-01 -4.66213226e-01 3.22005510e-01 6.46875262e-01
1.58621177e-01 -9.33687538e-02 -1.46332598e+00 5.65482199e-01
-2.05938414e-01 -1.49265194e+00 -7.86527712e-03 6.54177606e-01
5.76073289e-01 6.41125906e-03 -2.68892586e-01 -1.21372767e-01
7.31499434e-01 -1.35542119e+00 5.94230950e-01 3.44683468e-01
6.08884633e-01 -5.70949614e-01 4.74514812e-01 4.45674688e-01
-1.19155312e+00 -3.20315585e-02 -2.66909659e-01 -3.06319356e-01
-1.66663915e-01 7.88588524e-01 -9.83989239e-01 4.08928812e-01
5.31343877e-01 6.76306367e-01 -6.49231672e-01 1.14409363e+00
2.05638975e-01 7.07605302e-01 -5.07194221e-01 -1.56045824e-01
2.31305867e-01 -2.92171210e-01 8.53634298e-01 1.45397782e+00
-6.13479735e-03 -1.75546199e-01 2.83992559e-01 1.15879154e+00
2.12639514e-02 7.43680373e-02 -8.84352505e-01 -3.87700289e-01
6.31780803e-01 1.26393235e+00 -1.41037989e+00 -2.97876984e-01
-7.70580769e-02 2.78621465e-01 2.28819743e-01 2.75953978e-01
-3.60599846e-01 -3.59663367e-01 3.91079992e-01 5.62879324e-01
1.90468445e-01 -3.00282270e-01 -2.99257040e-01 -1.04515851e+00
-2.65491009e-01 -9.29475784e-01 5.68227351e-01 -4.15568680e-01
-1.62111855e+00 3.16943228e-01 8.38082433e-02 -1.01891673e+00
-2.49705747e-01 -4.67974216e-01 -5.68766952e-01 6.51809752e-01
-8.88973653e-01 -6.34751499e-01 -5.62163303e-03 4.03634310e-01
1.55827487e-02 7.63258897e-03 6.74999952e-01 1.23042844e-01
-4.16853189e-01 2.70728022e-01 3.01875204e-01 4.16314811e-01
5.04156351e-01 -1.50194967e+00 1.08323745e-01 8.84299815e-01
2.43477151e-01 9.20034885e-01 1.04224443e+00 -6.00561559e-01
-8.56181204e-01 -8.36690605e-01 8.53591084e-01 -7.41678476e-01
1.02424324e+00 -5.08055449e-01 -9.21542645e-01 4.40058500e-01
5.82483299e-02 -2.39980519e-01 6.72086537e-01 5.43104470e-01
-6.48879230e-01 2.47284442e-01 -9.99698043e-01 5.69268048e-01
1.38563550e+00 -5.99695027e-01 -5.24834991e-01 3.26764554e-01
4.56088841e-01 5.39912581e-01 -8.94114077e-01 1.92329079e-01
4.30298716e-01 -1.15754890e+00 9.14016068e-01 -6.24833524e-01
5.94181657e-01 -3.61594379e-01 -1.42668858e-01 -9.43602085e-01
-6.45987093e-01 -6.71040893e-01 -8.77080634e-02 1.15531421e+00
4.09569830e-01 -6.90404713e-01 8.82326543e-01 1.07261144e-01
1.83286563e-01 -5.12720227e-01 -7.51901865e-01 -6.99788928e-01
3.76469418e-02 -1.26817659e-01 2.85520524e-01 1.39723802e+00
8.63002464e-02 7.68317759e-01 2.44378090e-01 -5.40844118e-03
1.20675755e+00 2.42942914e-01 6.25499129e-01 -2.05300546e+00
-4.09835607e-01 -1.15211296e+00 -5.98848224e-01 -5.60217559e-01
7.19587654e-02 -1.15559173e+00 -7.72300437e-02 -1.35453439e+00
8.87253761e-01 -5.97388685e-01 1.69141456e-01 3.49113047e-01
9.59131718e-02 3.35151106e-01 -1.02125786e-01 6.86241925e-01
-8.07602108e-01 1.54701825e-02 8.52380514e-01 -3.31063598e-01
6.03882074e-02 -6.30854070e-02 -1.09786892e+00 7.75820374e-01
4.43750978e-01 -7.13135898e-01 -3.97158682e-01 2.09732398e-01
5.77801287e-01 -1.94903761e-01 5.09885788e-01 -8.37437272e-01
4.59131271e-01 -2.29456462e-02 1.82255611e-01 -4.03229356e-01
-2.07361221e-01 -6.84136689e-01 6.23932719e-01 7.50003219e-01
-3.27400535e-01 -1.72060281e-02 -4.52024460e-01 8.58243823e-01
-5.35149649e-02 -4.51496601e-01 6.81051433e-01 -2.38057375e-01
-3.04972738e-01 2.70377368e-01 -4.92732078e-01 4.33231980e-01
7.08427191e-01 -6.25390708e-01 -2.97425359e-01 -3.87954473e-01
-8.80567312e-01 1.79455876e-01 7.78914928e-01 -1.18600801e-01
2.86038399e-01 -1.04908943e+00 -1.03792810e+00 -1.38579309e-01
1.26808673e-01 8.85501802e-02 -3.67266566e-01 9.48875129e-01
-5.72321475e-01 8.74474645e-02 1.82071269e-01 -8.96826804e-01
-1.26694298e+00 6.76103115e-01 3.40797335e-01 -4.50692922e-01
-4.81954634e-01 5.27473092e-01 2.78975695e-01 -2.49420688e-01
1.32541910e-01 8.40977579e-03 -7.29917958e-02 3.91008645e-01
1.73672125e-01 2.78508574e-01 -1.94376901e-01 -5.01343608e-01
-3.16913366e-01 8.60476419e-02 -2.07156371e-02 -1.42308176e-02
1.29867220e+00 -1.67595848e-01 -5.92382669e-01 3.81515741e-01
1.09945512e+00 3.98694947e-02 -7.66996861e-01 -3.05986613e-01
4.73208249e-01 -2.05391079e-01 -2.30289981e-01 -3.07909846e-01
-7.86595106e-01 5.83805978e-01 -1.25814915e-01 1.07772839e+00
7.54699111e-01 2.29814932e-01 -1.37565523e-01 5.42492986e-01
2.76959300e-01 -8.17949176e-01 6.52532801e-02 3.61354828e-01
5.53947508e-01 -1.12925994e+00 5.20165563e-01 -5.36887288e-01
-1.71359807e-01 1.03221583e+00 -8.60627070e-02 -7.05511495e-02
5.79456568e-01 1.77993253e-01 -6.39378130e-01 -7.12933064e-01
-8.80830348e-01 -1.56028166e-01 1.04980931e-01 7.09825635e-01
3.30544651e-01 1.78391635e-01 -3.89811993e-01 3.21244389e-01
-2.06846073e-01 -5.80044031e-01 8.25313449e-01 6.82754993e-01
-6.38927698e-01 -9.81269479e-01 -3.89030814e-01 7.55497515e-01
-4.17535335e-01 -1.55093953e-01 -7.75594592e-01 1.05475867e+00
-1.37090638e-01 9.41069543e-01 2.20483333e-01 -2.45052800e-01
-1.84403300e-01 1.88472886e-02 3.38823467e-01 -8.93393338e-01
-3.98335904e-01 1.42527670e-01 3.91146123e-01 -1.75484464e-01
-7.27355361e-01 -7.69668043e-01 -9.14793730e-01 -9.35787380e-01
-5.49430311e-01 6.33307695e-01 1.80599660e-01 8.79968643e-01
1.01706073e-01 2.00316772e-01 5.07272363e-01 -7.21794784e-01
-6.39318526e-01 -7.62372732e-01 -8.65584075e-01 3.30323428e-01
1.57403350e-01 -5.55976808e-01 -8.17975044e-01 1.30344674e-01] | [6.964204788208008, 5.3193888664245605] |
2f964756-d8be-46bb-987a-11cd0f72d2fc | safety-aware-task-composition-for-discrete | 2306.17033 | null | https://arxiv.org/abs/2306.17033v1 | https://arxiv.org/pdf/2306.17033v1.pdf | Safety-Aware Task Composition for Discrete and Continuous Reinforcement Learning | Compositionality is a critical aspect of scalable system design. Reinforcement learning (RL) has recently shown substantial success in task learning, but has only recently begun to truly leverage composition. In this paper, we focus on Boolean composition of learned tasks as opposed to functional or sequential composition. Existing Boolean composition for RL focuses on reaching a satisfying absorbing state in environments with discrete action spaces, but does not support composable safety (i.e., avoidance) constraints. We advance the state of the art in Boolean composition of learned tasks with three contributions: i) introduce two distinct notions of safety in this framework; ii) show how to enforce either safety semantics, prove correctness (under some assumptions), and analyze the trade-offs between the two safety notions; and iii) extend Boolean composition from discrete action spaces to continuous action spaces. We demonstrate these techniques using modified versions of value iteration in a grid world, Deep Q-Network (DQN) in a grid world with image observations, and Twin Delayed DDPG (TD3) in a continuous-observation and continuous-action Bullet physics environment. We believe that these contributions advance the theory of safe reinforcement learning by allowing zero-shot composition of policies satisfying safety properties. | ['Zachary Serlin', 'Makai Mann', 'Kevin Leahy'] | 2023-06-29 | null | null | null | null | ['reinforcement-learning-1'] | ['methodology'] | [ 2.69236416e-01 2.48556405e-01 -1.37142971e-01 -1.22040763e-01
-7.12147593e-01 -7.61504591e-01 1.06448770e+00 5.25046512e-02
-5.85994661e-01 1.11204433e+00 1.65015891e-01 -5.66919744e-01
-3.38733822e-01 -8.64562511e-01 -9.77892816e-01 -8.87246668e-01
-7.20021904e-01 3.82086456e-01 3.02684993e-01 -5.23022115e-01
5.58902808e-02 3.59328836e-01 -1.71679759e+00 2.01529741e-01
4.73003417e-01 8.25571537e-01 -4.15077984e-01 8.04539382e-01
2.52664357e-01 1.44632339e+00 -5.07680178e-01 6.51023444e-03
5.44742942e-01 -4.95972991e-01 -1.04329002e+00 -3.78315389e-01
4.20253336e-01 -5.71377635e-01 -3.35518032e-01 1.05836201e+00
4.30457622e-01 4.03108478e-01 2.55654842e-01 -1.95677912e+00
-3.39950383e-01 9.85863745e-01 -1.39291743e-02 -1.74110353e-01
2.72292498e-04 8.99304926e-01 1.14501369e+00 2.52436846e-01
3.66096348e-01 1.33100796e+00 4.58919615e-01 1.11429620e+00
-1.64461768e+00 -6.44860864e-01 4.06078070e-01 7.93878511e-02
-8.79472971e-01 -4.35305476e-01 1.93079874e-01 -3.83334666e-01
1.61979246e+00 1.01439215e-01 9.20206249e-01 1.35035598e+00
8.41782093e-01 7.02755272e-01 1.59340417e+00 -2.13527888e-01
7.69581974e-01 -2.90276766e-01 5.47815040e-02 6.85212493e-01
-3.34899384e-03 9.53922093e-01 -8.99264097e-01 -5.58911026e-01
6.36875451e-01 -2.33604088e-01 -5.80121344e-03 -6.01002812e-01
-1.16885841e+00 7.76213706e-01 3.35941873e-02 -1.26861975e-01
-2.13377446e-01 1.21230042e+00 8.78299773e-01 9.66396749e-01
5.67410700e-02 6.34330451e-01 -5.33712983e-01 -2.12008640e-01
-5.40307879e-01 8.51093113e-01 1.10108101e+00 7.59452343e-01
3.55430901e-01 4.00152683e-01 -3.97176445e-01 -1.23366892e-01
1.84058860e-01 6.48399651e-01 1.00336567e-01 -1.40203011e+00
-5.20161539e-02 -2.30754435e-01 3.09503794e-01 -2.90890243e-02
-4.02103662e-01 3.88921035e-04 -5.79295635e-01 8.68930340e-01
2.59029597e-01 -4.47171539e-01 -6.43564582e-01 2.31143475e+00
2.06814766e-01 1.80670351e-01 4.32115942e-01 6.91206634e-01
4.80287755e-03 5.68370044e-01 1.33216739e-01 -5.19020259e-01
9.68684971e-01 -7.56874442e-01 -6.72439158e-01 4.35431674e-02
3.17333132e-01 9.67414528e-02 1.03504479e+00 7.98593223e-01
-1.23368859e+00 -1.74617112e-01 -1.21633029e+00 2.31839761e-01
1.24169342e-01 -1.01858270e+00 9.09216404e-01 5.34736335e-01
-1.55483031e+00 8.88051212e-01 -1.16680145e+00 2.42149513e-02
1.88879371e-01 7.08981335e-01 -9.48945656e-02 3.68387222e-01
-1.53780639e+00 1.02402222e+00 3.10294628e-01 -4.09531206e-01
-2.25124669e+00 -6.73230290e-01 -7.57046819e-01 9.43968352e-03
9.21952128e-01 -8.84106457e-01 1.98887146e+00 -1.20050311e+00
-1.88801289e+00 2.57169664e-01 4.92751628e-01 -1.02079856e+00
6.25295162e-01 -2.87486434e-01 -4.94200513e-02 5.78972958e-02
-2.07271025e-01 8.04888010e-01 1.00024080e+00 -1.09590244e+00
-8.09321642e-01 -1.85631111e-01 7.04841673e-01 3.23159963e-01
-6.55540153e-02 -1.86688080e-01 6.29636407e-01 -4.74667959e-02
-6.41701341e-01 -1.03694856e+00 -3.77685845e-01 1.62356272e-02
2.36366875e-02 -4.32408601e-01 4.10246670e-01 6.56311586e-02
6.18330359e-01 -2.10183096e+00 3.01887572e-01 1.19914010e-01
4.09937263e-01 -2.43367225e-01 -2.21045449e-01 7.02542245e-01
2.68137902e-01 -2.46275030e-02 -2.29714125e-01 -1.02206372e-01
6.11133814e-01 5.10017991e-01 -8.59222651e-01 8.13268304e-01
1.48143470e-01 8.81684840e-01 -1.31508124e+00 -3.41761112e-01
-7.06261471e-02 1.25443012e-01 -7.68038511e-01 2.64027357e-01
-9.99779761e-01 3.53910178e-01 -2.96757996e-01 2.20173925e-01
1.23026691e-01 2.48139873e-01 5.42386889e-01 4.34730411e-01
-2.75790125e-01 5.51715612e-01 -1.23083925e+00 1.50282907e+00
-2.60890514e-01 1.57995835e-01 3.46346468e-01 -7.24322379e-01
3.73307347e-01 7.05251038e-01 7.02869296e-01 -8.07518840e-01
-1.67258624e-02 4.07477021e-02 2.61726946e-01 -3.91510993e-01
4.44423258e-01 -7.72354066e-01 -3.97704303e-01 7.87698567e-01
1.39670417e-01 -6.72147334e-01 -9.36948508e-02 1.95127070e-01
1.62877524e+00 5.25555134e-01 2.09847867e-01 -6.10881567e-01
2.01027110e-01 8.65579247e-02 5.89648485e-01 1.39001644e+00
-5.20235717e-01 -2.53774345e-01 1.05809355e+00 -4.21700627e-01
-1.07177639e+00 -1.17293131e+00 1.84346572e-01 1.27251697e+00
9.43830088e-02 -5.08140206e-01 -6.47759557e-01 -8.44137192e-01
3.29992503e-01 8.34620237e-01 -7.17246830e-01 -4.94200021e-01
-6.14828408e-01 -7.43425116e-02 1.07711577e+00 5.44771671e-01
4.54047322e-01 -1.38929999e+00 -1.27303231e+00 1.63828194e-01
2.44004890e-01 -4.32783872e-01 -3.19715232e-01 7.88552046e-01
-7.59375274e-01 -9.73791778e-01 2.04132274e-01 -3.34799439e-01
5.92674799e-02 -2.93070301e-02 1.10798597e+00 -4.67297547e-02
1.95716232e-01 9.85664904e-01 6.36859685e-02 -4.24115926e-01
-5.80573320e-01 -2.55061865e-01 4.89393681e-01 -5.15455425e-01
-4.07511331e-02 -5.35134554e-01 -2.31955484e-01 -6.97079077e-02
-1.19969475e+00 2.81107123e-03 1.57982975e-01 1.12707293e+00
4.45993483e-01 1.67737067e-01 3.52575928e-01 -5.96000075e-01
9.50094342e-01 -2.40150109e-01 -1.07564223e+00 2.14083821e-01
-7.20133424e-01 5.22100687e-01 7.85489082e-01 -5.07237613e-01
-9.56997693e-01 -6.83281720e-02 7.54434243e-02 -3.36811304e-01
4.72900048e-02 1.96218088e-01 -4.60740738e-02 4.66556847e-03
8.66432548e-01 2.14754641e-01 4.51498598e-01 2.75025159e-01
5.51626265e-01 7.43997172e-02 3.49225372e-01 -1.53603089e+00
5.04376888e-01 4.71101850e-01 2.77633518e-01 -1.84256896e-01
-2.40197450e-01 1.97081178e-01 -1.44621342e-01 -1.26028150e-01
6.70675933e-01 -7.08711863e-01 -1.68164015e+00 4.86674458e-01
-6.92423403e-01 -1.29376864e+00 -8.32521379e-01 1.31277740e-01
-1.32331467e+00 1.54731691e-01 -8.37236822e-01 -1.29810226e+00
-3.40761334e-01 -1.28652728e+00 9.26418424e-01 -1.11261591e-01
-4.84825596e-02 -6.15794241e-01 4.10489768e-01 -4.30421799e-01
4.40289110e-01 1.29263192e-01 1.00926077e+00 -4.96166378e-01
-6.87823832e-01 5.64512670e-01 4.74506140e-01 4.71542984e-01
-4.63257223e-01 -1.10905707e-01 -8.62070084e-01 -6.54572427e-01
3.07962030e-01 -1.06670046e+00 7.63489127e-01 2.60790735e-01
1.02374482e+00 -5.70555449e-01 9.31157619e-02 2.98429459e-01
1.34676337e+00 2.59175479e-01 5.11554599e-01 4.38309163e-01
5.21324724e-02 5.31100929e-01 6.82693779e-01 6.16766632e-01
2.40232095e-01 4.16689932e-01 7.70232797e-01 4.62786674e-01
2.51787603e-01 -2.75883466e-01 1.05528510e+00 1.92104787e-01
8.49022716e-02 5.70745952e-02 -8.59608650e-01 3.41005534e-01
-2.29589128e+00 -1.19737768e+00 4.97334331e-01 2.24291396e+00
1.39636040e+00 3.56172442e-01 2.96369165e-01 -7.85192698e-02
1.17213391e-01 6.14625588e-02 -8.67339849e-01 -7.79605389e-01
1.08923808e-01 6.43316507e-01 7.10221171e-01 8.22358787e-01
-9.87298965e-01 9.46839392e-01 6.77142715e+00 5.94112337e-01
-1.08768749e+00 3.47316146e-01 3.21501702e-01 -4.89564210e-01
-5.09628415e-01 1.66817486e-01 -5.98236740e-01 8.72707441e-02
1.16690469e+00 -2.57466495e-01 1.00914645e+00 8.06124449e-01
-6.13554418e-02 -1.00140020e-01 -1.61435270e+00 4.40728813e-01
-2.14409217e-01 -1.07888997e+00 -4.97236967e-01 3.22577395e-02
7.55192935e-01 9.09593105e-02 1.91881552e-01 8.47321033e-01
1.29057169e+00 -1.25161493e+00 1.17609978e+00 3.52564484e-01
8.10749233e-01 -9.16687012e-01 3.09149861e-01 5.13526142e-01
-8.35485935e-01 -4.52411085e-01 2.03372929e-02 -4.87620890e-01
-2.00660210e-02 -3.37509178e-02 -5.05852461e-01 2.39969686e-01
8.18329096e-01 2.85173178e-01 3.50841694e-02 5.09971440e-01
-5.02323449e-01 6.22910500e-01 -2.16983184e-01 -1.89662501e-01
5.24383724e-01 5.92979640e-02 5.42249262e-01 7.52364933e-01
-3.27474654e-01 1.55831307e-01 4.89974290e-01 8.26982737e-01
1.35776758e-01 -6.83126748e-01 -7.16664016e-01 -5.70609905e-02
2.83906013e-01 9.35623407e-01 -3.30778211e-01 -3.82940680e-01
-1.78045958e-01 4.99362916e-01 4.41616267e-01 2.62642533e-01
-1.08709121e+00 1.12727419e-01 1.21145391e+00 -2.02947378e-01
3.25735100e-03 -5.12952507e-01 -1.79510161e-01 -9.91170347e-01
-3.51311684e-01 -1.67160439e+00 4.05016840e-01 -3.17804128e-01
-1.33338892e+00 1.91732734e-01 4.80955213e-01 -6.62378967e-01
-2.45298013e-01 -4.23836708e-01 -1.47174671e-01 5.52807570e-01
-1.19572866e+00 -8.87675047e-01 2.46362835e-01 7.87433624e-01
2.04936028e-01 -4.15766193e-03 1.02216673e+00 -2.41623417e-01
-1.59926966e-01 4.00436789e-01 2.85070445e-02 -4.51644510e-01
6.55567050e-01 -1.58396816e+00 4.06558722e-01 7.51603723e-01
-1.76633015e-01 7.21858263e-01 7.99880922e-01 -7.44407296e-01
-1.97853804e+00 -9.56304848e-01 2.49962002e-01 -3.85482371e-01
8.38961005e-01 -5.98917544e-01 -5.69143772e-01 8.62679243e-01
5.01498044e-01 -7.20485002e-02 2.39296034e-01 9.53002945e-02
-6.38412714e-01 -2.64838845e-01 -1.23678637e+00 9.26170766e-01
1.32378602e+00 -7.35221744e-01 -4.88153100e-01 3.96679372e-01
1.29123211e+00 -5.62752783e-01 -7.47725785e-01 4.26588595e-01
6.83848739e-01 -8.68868351e-01 6.40784442e-01 -1.02990437e+00
3.78948808e-01 -4.76673305e-01 -3.29521388e-01 -1.25097656e+00
-2.61428088e-01 -1.09196877e+00 -5.32586396e-01 4.81447190e-01
-2.98772636e-03 -8.96677554e-01 5.27686477e-01 7.18051076e-01
-4.11403954e-01 -8.00979972e-01 -1.07103944e+00 -9.87727046e-01
4.96734709e-01 -5.43147922e-01 6.64657414e-01 6.04008615e-01
3.29197586e-01 5.65753877e-02 -4.21960443e-01 4.80342358e-02
6.07879162e-01 -1.30463660e-01 6.59987211e-01 -6.18845046e-01
-1.03878105e+00 -5.92390060e-01 -6.64802715e-02 -6.85087621e-01
4.71334994e-01 -7.97594011e-01 6.15268528e-01 -1.03795731e+00
2.02568397e-01 -4.63108361e-01 -5.35873950e-01 1.01045489e+00
2.56910592e-01 -2.52195925e-01 2.93396920e-01 4.08067442e-02
-9.55449760e-01 6.64591908e-01 1.09609234e+00 -3.97796512e-01
-5.78064732e-02 -2.81988025e-01 -5.88674188e-01 2.13347360e-01
8.94112349e-01 -3.30713868e-01 -7.97549725e-01 -2.50305653e-01
5.55621624e-01 2.58160472e-01 6.66654348e-01 -9.16944206e-01
2.56786853e-01 -8.45742583e-01 -1.67871222e-01 1.12133631e-02
6.33965880e-02 -7.43475139e-01 1.66682690e-01 1.10766768e+00
-1.03656447e+00 2.73233771e-01 3.87013406e-01 6.10435367e-01
3.58770996e-01 -4.07544896e-02 6.86707675e-01 -2.19208822e-01
-7.18093455e-01 1.42925695e-01 -5.88084877e-01 1.70997351e-01
1.31662214e+00 3.71729523e-01 -6.13904834e-01 -3.09211880e-01
-6.12473965e-01 5.94823599e-01 4.85194355e-01 1.98876545e-01
5.12470067e-01 -9.81011212e-01 -5.06262064e-01 5.83414622e-02
-1.53939947e-01 -5.60057499e-02 1.23221353e-01 9.11710382e-01
-2.16925398e-01 1.90722182e-01 -4.36306387e-01 -4.08955544e-01
-1.13591230e+00 7.61569977e-01 6.75620675e-01 -3.44444662e-01
-6.78031683e-01 7.95433223e-01 7.84611478e-02 -3.60616982e-01
6.09755039e-01 -2.79781938e-01 3.25793058e-01 -3.75423193e-01
4.85899925e-01 1.40418530e-01 -5.64547256e-02 2.63744652e-01
-3.88649523e-01 -2.42828712e-01 -8.04476906e-03 -9.70321953e-01
1.25867629e+00 4.87167448e-01 -3.85563135e-01 5.70385993e-01
4.11786318e-01 -3.97144347e-01 -1.71798968e+00 -1.17621474e-01
-1.86119646e-01 -1.03879392e-01 -4.68356088e-02 -8.83165240e-01
-5.37695169e-01 7.41347671e-01 5.17660797e-01 4.10623610e-01
9.77130890e-01 -2.31584847e-01 5.61787367e-01 7.59156823e-01
8.91600430e-01 -1.22363997e+00 3.33843559e-01 8.94212008e-01
7.42586732e-01 -6.53109670e-01 -9.28046107e-02 3.98215234e-01
-8.14681351e-01 8.41758847e-01 8.28225791e-01 -3.14895391e-01
3.82665336e-01 8.71512353e-01 -4.08198148e-01 -9.75967273e-02
-1.59945571e+00 -3.33609670e-01 -3.72164488e-01 8.53370011e-01
1.41945153e-01 2.20469072e-01 -2.30369657e-01 2.89678842e-01
-5.35434596e-02 2.93391719e-02 5.34868121e-01 1.44675219e+00
-5.35168588e-01 -1.31735480e+00 -2.40823984e-01 2.81292260e-01
-3.81819695e-01 -1.73884466e-01 -2.56927073e-01 7.39166021e-01
1.50614366e-01 8.60311031e-01 2.04101484e-02 -4.36839849e-01
5.04490547e-02 1.78242967e-01 1.08249068e+00 -6.61461115e-01
-1.14620972e+00 -1.96460396e-01 1.52770683e-01 -8.97331297e-01
-2.27686808e-01 -7.35019505e-01 -1.51070845e+00 -6.94640875e-01
4.98183221e-02 1.89957544e-01 4.27627116e-01 7.94877768e-01
9.23176482e-02 7.26103306e-01 3.77722293e-01 -6.15929782e-01
-1.67476475e+00 -4.25747544e-01 -5.68586588e-01 1.09909900e-01
6.49950504e-01 -4.67608422e-01 -2.29552075e-01 -1.81703061e-01] | [4.412973880767822, 2.025777578353882] |
e6822731-e084-4094-8366-60549584ef73 | automatic-exposure-compensation-for-multi | 1805.11211 | null | http://arxiv.org/abs/1805.11211v1 | http://arxiv.org/pdf/1805.11211v1.pdf | Automatic Exposure Compensation for Multi-Exposure Image Fusion | This paper proposes a novel luminance adjustment method based on automatic
exposure compensation for multi-exposure image fusion. Multi-exposure image
fusion is a method to produce images without saturation regions, by using
photos with different exposures. In conventional works, it has been pointed out
that the quality of those multi-exposure images can be improved by adjusting
the luminance of them. However, how to determine the degree of adjustment has
never been discussed. This paper therefore proposes a way to automatically
determines the degree on the basis of the luminance distribution of input
multi-exposure images. Moreover, new weights, called "simple weights", for
image fusion are also considered for the proposed luminance adjustment method.
Experimental results show that the multi-exposure images adjusted by the
proposed method have better quality than the input multi-exposure ones in terms
of the well-exposedness. It is also confirmed that the proposed simple weights
provide the highest score of statistical naturalness and discrete entropy in
all fusion methods. | ['Sayaka Shiota', 'Hitoshi Kiya', 'Yuma Kinoshita'] | 2018-05-29 | null | null | null | null | ['multi-exposure-image-fusion'] | ['computer-vision'] | [ 5.72718024e-01 -4.58742797e-01 3.26260567e-01 -2.76656985e-01
-2.62965590e-01 -2.26085484e-01 4.42951769e-01 1.10431783e-01
-5.98560572e-01 6.45786405e-01 1.36879802e-01 3.30774188e-02
-3.35005999e-01 -8.01781952e-01 -2.37685785e-01 -7.71649480e-01
4.32527691e-01 -4.34234381e-01 1.12939782e-01 -3.46396446e-01
4.11309093e-01 3.45079601e-01 -1.88503349e+00 1.91578344e-02
1.15080535e+00 7.52522528e-01 4.19346064e-01 6.52015150e-01
8.97578225e-02 1.68782368e-01 -7.91528046e-01 -3.11298221e-01
3.88209879e-01 -9.29992437e-01 -3.27071756e-01 4.27214146e-01
4.46562737e-01 -1.95808604e-01 -5.65394200e-02 1.47481203e+00
4.46085811e-01 3.10570449e-01 8.34696293e-01 -9.38697278e-01
-8.56496394e-01 3.99681926e-01 -7.81583011e-01 3.36554229e-01
4.82752979e-01 5.89602888e-02 2.02814996e-01 -5.50979853e-01
1.77064449e-01 1.23870361e+00 3.67548466e-01 1.49556294e-01
-1.27525496e+00 -5.15025020e-01 -2.56115228e-01 3.69338304e-01
-1.44998121e+00 -2.73234189e-01 8.49091351e-01 -2.63416260e-01
3.46449226e-01 5.75557411e-01 6.15194976e-01 1.71944171e-01
8.35342646e-01 -8.83302912e-02 1.87292707e+00 -1.00809097e+00
-1.00079834e-01 5.06561339e-01 1.69156581e-01 5.49388409e-01
3.30926180e-01 2.42422130e-02 -5.45761324e-02 1.58392876e-01
4.17672664e-01 9.24489126e-02 -6.53313220e-01 1.68859378e-01
-1.02076614e+00 4.63673502e-01 2.53235072e-01 9.64021146e-01
-3.20801765e-01 -2.40415379e-01 2.73273885e-02 3.64951342e-01
3.48886520e-01 2.53022164e-01 4.96841595e-02 3.21222037e-01
-9.53845799e-01 -1.28415018e-01 4.21039611e-01 5.73792160e-01
1.00146258e+00 -2.34748125e-02 -2.52966821e-01 5.82902491e-01
1.84829026e-01 7.45982766e-01 5.18429875e-01 -7.52981842e-01
1.38844386e-01 4.85058099e-01 -1.63436588e-02 -1.23243487e+00
-1.66031450e-01 1.62562430e-02 -1.04789948e+00 7.24322379e-01
2.18341291e-01 -1.86970606e-01 -8.14512908e-01 1.63215959e+00
9.84470025e-02 -9.13942158e-02 4.35971677e-01 8.35034251e-01
8.44387054e-01 9.49060023e-01 1.12081327e-01 -8.76602590e-01
1.45446241e+00 -7.04973578e-01 -1.44035614e+00 3.42617452e-01
-3.12764555e-01 -1.23003566e+00 9.23576713e-01 5.83770990e-01
-1.30530238e+00 -1.14748704e+00 -1.41446066e+00 1.71709836e-01
-5.06594479e-01 -6.71742018e-03 2.85017155e-02 9.36816037e-01
-1.03849721e+00 4.88506526e-01 1.50306115e-03 -3.19601744e-01
-2.13753760e-01 2.18990311e-01 -3.03319544e-01 1.55821726e-01
-1.30578029e+00 1.33983147e+00 4.49199766e-01 -8.64661038e-02
-1.46150082e-01 -4.10685837e-01 -7.50201225e-01 2.65460629e-02
2.27528140e-02 -3.78151506e-01 6.64578140e-01 -1.22411609e+00
-1.47148454e+00 5.14905214e-01 -2.55878836e-01 -7.91159943e-02
4.05284226e-01 1.66194260e-01 -9.66478884e-01 5.70042670e-01
-2.53540784e-01 4.91476834e-01 1.05891824e+00 -1.55184507e+00
-6.51070178e-01 -7.87221417e-02 2.65261531e-02 3.58126074e-01
-3.14936161e-01 -7.09674060e-02 -1.85166657e-01 -4.64080721e-01
8.28454364e-03 -5.41870773e-01 -1.07002258e-01 -2.12613538e-01
-9.76927578e-02 1.60849482e-01 1.16629672e+00 -8.55127990e-01
1.32196820e+00 -2.24644899e+00 -1.46086305e-01 4.95259196e-01
-5.66378273e-02 3.08451474e-01 1.60237044e-01 1.94597468e-01
-9.38058496e-02 8.41320083e-02 -3.46094310e-01 3.31142060e-02
-1.64434940e-01 -7.00808391e-02 1.07212938e-01 4.37907159e-01
-1.42917410e-01 2.88005620e-01 -5.95109940e-01 -9.24560010e-01
9.85850275e-01 8.36159348e-01 1.28050163e-01 8.03223103e-02
4.33939874e-01 4.81812000e-01 1.65251985e-01 2.38677964e-01
1.17735124e+00 2.70426691e-01 5.69066443e-02 -7.69858479e-01
-3.53818983e-01 -7.70636857e-01 -1.31609213e+00 1.03254080e+00
-5.01107037e-01 7.32314408e-01 -3.26184094e-01 -3.37918758e-01
1.17441249e+00 5.70494950e-01 5.37219048e-01 -7.84523904e-01
4.89968181e-01 5.78300804e-02 -4.72567976e-02 -6.27777100e-01
9.21539307e-01 -3.31582159e-01 1.34034052e-01 1.43776432e-01
-2.67300963e-01 -4.50927943e-01 6.14689350e-01 -7.68238306e-02
1.66995227e-01 -1.33087590e-01 5.75249732e-01 -2.39208877e-01
1.21699977e+00 -6.18415654e-01 3.11886519e-01 3.07819515e-01
-1.96587235e-01 5.81869900e-01 -4.85780761e-02 3.11423540e-02
-1.22631705e+00 -8.23143601e-01 -4.06631917e-01 2.76072204e-01
8.10726881e-01 5.11045493e-02 -1.07781839e+00 -1.84840299e-02
-3.10143769e-01 5.20516813e-01 -3.77570629e-01 -6.68509603e-02
-2.94024974e-01 -8.42723966e-01 1.19766928e-01 -1.92153439e-01
1.12986565e+00 -1.01078975e+00 -7.81492114e-01 4.03803103e-02
-3.65064085e-01 -8.86604607e-01 -5.00912547e-01 -2.51164049e-01
-6.37679279e-01 -8.95641208e-01 -8.59156311e-01 -7.10332692e-01
8.40960085e-01 5.15843868e-01 6.18171751e-01 4.29155171e-01
-1.74532324e-01 2.60192484e-01 -3.96985471e-01 -3.50387841e-01
-7.97215939e-01 -4.17105317e-01 -3.29139046e-02 3.44372123e-01
2.38056406e-01 -4.61096615e-01 -7.07299411e-01 2.93477327e-01
-1.20430243e+00 -5.36026724e-04 6.21429384e-01 2.66465932e-01
7.29662955e-01 5.22072971e-01 6.20725155e-01 -5.01702487e-01
8.83609831e-01 1.43190026e-02 -6.11484766e-01 6.60731435e-01
-9.44780648e-01 -3.27162305e-03 6.43983722e-01 -3.53499562e-01
-1.53544998e+00 -2.69893140e-01 4.66536991e-02 -3.83458808e-02
-3.37586761e-01 -2.02250704e-02 -2.16232225e-01 -6.25668347e-01
4.27698880e-01 3.08350086e-01 1.12904921e-01 -2.48263344e-01
4.44791794e-01 7.44020462e-01 7.29463696e-01 -2.52832100e-02
7.61820018e-01 3.22296888e-01 2.35005677e-01 -9.64350224e-01
-2.04554945e-01 -1.74338102e-01 -5.85176587e-01 -6.88059986e-01
1.32943487e+00 -5.92422605e-01 -7.89409757e-01 6.65465891e-01
-9.32258546e-01 3.14588517e-01 -7.74162933e-02 6.30771399e-01
-4.85876918e-01 9.42939222e-01 -2.32733369e-01 -8.68241608e-01
-5.01516461e-01 -1.29031777e+00 4.48578238e-01 9.13182378e-01
4.76127453e-02 -1.09187043e+00 -1.12200782e-01 3.98770086e-02
5.59449852e-01 3.49828690e-01 7.38230824e-01 2.57054180e-01
-3.70351851e-01 8.41405317e-02 -1.42549604e-01 5.80875158e-01
5.95360100e-01 3.21698040e-01 -8.91071439e-01 -2.52397984e-01
3.02712381e-01 2.24019185e-01 8.05369258e-01 6.08027577e-01
1.12530863e+00 -9.86070260e-02 4.13871789e-03 6.73308849e-01
2.11050153e+00 6.08798623e-01 1.21861374e+00 3.80222172e-01
3.07065342e-02 2.72902310e-01 7.21370876e-01 2.07330659e-01
2.59388592e-02 4.73710179e-01 1.74415335e-01 -5.85489035e-01
-3.69556516e-01 1.76913559e-01 2.23913163e-01 7.96488881e-01
-1.61272272e-01 -5.83066583e-01 -5.12419641e-02 2.18578592e-01
-1.40316033e+00 -1.32328320e+00 -4.77059335e-01 2.13265586e+00
7.03661084e-01 2.23015975e-02 -2.47387394e-01 8.26943994e-01
1.11609161e+00 1.80347905e-01 -5.87092042e-02 -9.36844528e-01
-5.78198671e-01 2.03078747e-01 5.44731915e-01 9.33559120e-01
-1.01963389e+00 1.94455042e-01 6.59542608e+00 7.98747659e-01
-1.14669943e+00 8.18536133e-02 4.39622283e-01 4.90633875e-01
-4.83938813e-01 1.69333350e-02 -6.28880382e-01 9.25988972e-01
6.07911408e-01 -4.97513831e-01 2.01465368e-01 2.14765053e-02
3.82746786e-01 -8.60973120e-01 -3.67384553e-01 9.98163104e-01
5.11334896e-01 -7.33516097e-01 1.44386932e-01 -5.27697355e-02
1.02173638e+00 -9.74561334e-01 3.75115544e-01 -4.03741151e-01
-3.92509967e-01 -4.79893386e-01 5.48706293e-01 1.11437762e+00
7.09664941e-01 -9.47948873e-01 9.47931767e-01 -1.83439683e-02
-1.27070272e+00 4.10197414e-02 -2.13296965e-01 1.15860268e-01
5.12811065e-01 5.86107314e-01 -1.96937442e-01 9.55487192e-01
3.63800973e-01 1.83657929e-01 -9.51046586e-01 1.16859019e+00
-1.84648052e-01 2.07534656e-01 -1.57907039e-01 8.31121504e-02
-1.08406581e-01 -7.07471907e-01 3.63686532e-01 1.11690009e+00
6.76232874e-01 3.01731974e-01 -2.30409220e-01 7.01428354e-01
3.32413316e-01 3.71700972e-01 -5.38283169e-01 4.66685832e-01
2.82849729e-01 1.30498040e+00 -7.92849064e-01 -6.21980786e-01
-2.98704386e-01 7.94155121e-01 -6.29345059e-01 3.44177514e-01
-9.00328457e-01 -7.67319143e-01 -2.83736121e-02 -2.34705478e-01
6.07123487e-02 -9.91955623e-02 -4.17720169e-01 -7.16137350e-01
-1.72326058e-01 -6.19331777e-01 2.29981095e-01 -1.10505974e+00
-9.20562267e-01 7.14255691e-01 3.43152374e-01 -1.26894236e+00
1.17561214e-01 -1.78008288e-01 -5.60411692e-01 1.22233903e+00
-1.41016781e+00 -7.06629932e-01 -5.94742477e-01 5.91191053e-01
4.54075724e-01 -8.27554166e-02 5.28576255e-01 3.38759899e-01
-3.69267374e-01 5.33083439e-01 2.10869148e-01 -5.58744967e-01
7.59269118e-01 -1.12619245e+00 -4.44931388e-01 1.27942729e+00
-1.64817587e-01 2.18955919e-01 1.01922750e+00 -6.02181017e-01
-5.92926323e-01 -5.09674072e-01 8.65399957e-01 1.59496441e-01
-1.32410511e-01 3.96913677e-01 -1.09224522e+00 4.84906510e-02
1.07469785e+00 -6.56904399e-01 5.91047287e-01 -7.09189534e-01
3.16145271e-01 -4.55717027e-01 -1.65076995e+00 4.88008767e-01
2.07138315e-01 -1.81591049e-01 -7.46224821e-01 -8.65009502e-02
6.65845275e-01 -1.01059151e-03 -1.19697630e+00 3.93151134e-01
4.39114511e-01 -1.28999889e+00 9.10761833e-01 6.07123673e-01
2.85933584e-01 -7.92408705e-01 -5.29888533e-02 -1.16747069e+00
-3.50644767e-01 -3.06423932e-01 4.22358036e-01 1.43539488e+00
2.25710616e-01 -7.20655859e-01 -4.69559096e-02 3.88221771e-01
2.05876321e-01 -2.88382888e-01 -4.97966319e-01 -4.60722387e-01
-4.86458272e-01 2.60632515e-01 7.92806029e-01 6.90719068e-01
-1.89292550e-01 1.11664034e-01 -4.63198423e-01 3.01548868e-01
9.15561974e-01 -1.92391127e-02 4.76205111e-01 -9.75865424e-01
1.06674016e-01 -3.99239898e-01 -4.38497126e-01 -2.76362926e-01
-2.31088042e-01 -2.99423099e-01 -7.78463483e-02 -1.67037642e+00
2.53727883e-01 2.31339671e-02 -5.12337506e-01 -3.37902755e-02
-4.75747764e-01 7.50531912e-01 4.28189129e-01 1.24232158e-01
-8.65083486e-02 3.04188818e-01 1.51535606e+00 -5.87403886e-02
-1.69337615e-01 -1.71423301e-01 -5.40104568e-01 5.97694755e-01
9.09093440e-01 -1.17334463e-01 -5.04959762e-01 -1.75325349e-01
-1.53957874e-01 -3.06160785e-02 2.93702960e-01 -1.44675493e+00
6.25274256e-02 -3.07750016e-01 6.07952952e-01 -8.38685036e-01
3.76957804e-01 -1.15342796e+00 5.91831803e-01 5.68527162e-01
-2.31246546e-01 3.37367684e-01 1.28055260e-01 3.53559375e-01
-2.70202965e-01 -5.63879073e-01 1.21103811e+00 -4.05659713e-02
-8.82102668e-01 -1.89311996e-01 -5.43655455e-01 -6.44136786e-01
1.25673795e+00 -7.50029385e-01 -2.60681540e-01 -4.49651420e-01
-6.11256659e-01 -2.98446894e-01 7.59260178e-01 1.00163221e-01
7.20677435e-01 -1.30068052e+00 -4.24595416e-01 2.46011898e-01
-2.18178526e-01 -7.31821060e-01 6.58354104e-01 8.65969777e-01
-7.20489800e-01 -1.04141533e-01 -7.33684480e-01 -2.82526046e-01
-1.77412891e+00 8.83161783e-01 1.77623495e-01 1.14270203e-01
-4.17314440e-01 1.71461388e-01 -9.05509889e-02 5.08064449e-01
-1.34628251e-01 -2.72117466e-01 -7.10140407e-01 -4.75611873e-02
6.44600928e-01 5.45543551e-01 -1.51026547e-01 -9.89884913e-01
2.05289945e-03 1.14174581e+00 2.86068380e-01 -3.64927888e-01
8.61100018e-01 -7.60063529e-01 -3.34123760e-01 4.55086857e-01
1.05207956e+00 1.45019218e-01 -9.66943681e-01 1.10129915e-01
-5.16210079e-01 -9.49130833e-01 1.37174898e-03 -8.14087212e-01
-8.77456725e-01 8.58291566e-01 1.26457882e+00 4.79505301e-01
1.80589521e+00 -6.91963315e-01 6.10865772e-01 -6.54077306e-02
1.60097048e-01 -1.30950141e+00 -5.19182719e-02 -1.20714404e-01
7.36048579e-01 -1.17413795e+00 2.88153410e-01 -5.02837479e-01
-3.66347641e-01 1.25163543e+00 6.05139554e-01 4.37165536e-02
5.14778793e-01 6.59695342e-02 1.38688311e-01 9.99633521e-02
-1.97655767e-01 -3.68780941e-01 3.43793124e-01 6.22043729e-01
3.25130492e-01 -6.74078688e-02 -1.13357556e+00 -1.12893879e-01
-2.84614395e-02 1.19750928e-02 8.08346748e-01 5.62278569e-01
-9.49843228e-01 -1.03499842e+00 -1.04297101e+00 1.63121000e-01
-4.80644822e-01 2.19004661e-01 6.97425455e-02 7.44998991e-01
6.96966946e-01 1.42224216e+00 -1.02467621e-02 -3.83171916e-01
3.56183618e-01 -2.47517347e-01 7.69042790e-01 1.91435605e-01
-4.66150105e-01 -4.18481380e-02 -4.12217677e-01 1.41311914e-01
-1.10088289e+00 -2.30204344e-01 -9.75086927e-01 -5.38190901e-01
-6.08153880e-01 4.54927444e-01 7.93565035e-01 7.82055557e-01
-2.18638569e-01 6.87699854e-01 1.03711557e+00 -6.09998763e-01
-9.36878100e-02 -9.51665998e-01 -6.14303112e-01 6.26223087e-01
3.55893612e-01 -4.35736150e-01 -5.29376268e-01 4.33818847e-01] | [10.9227933883667, -2.459397792816162] |
694c8dc5-607d-4f84-ad17-b8551de5b79d | interoperability-between-service-composition | null | null | https://aclanthology.org/I13-1145 | https://aclanthology.org/I13-1145.pdf | Interoperability between Service Composition and Processing Pipeline: Case Study on the Language Grid and UIMA | null | ['Toru Ishida', 'Trang Mai Xuan', 'Donghui Lin', 'Yohei Murakami'] | 2013-10-01 | interoperability-between-service-composition-1 | https://aclanthology.org/I13-1145 | https://aclanthology.org/I13-1145.pdf | ijcnlp-2013-10 | ['service-composition'] | ['miscellaneous'] | [-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] | [-7.214504718780518, 3.786614418029785] |
1552f3b0-9775-41c9-9714-9dd85c2e6ebd | highway-long-short-term-memory-rnns-for | 1510.08983 | null | http://arxiv.org/abs/1510.08983v2 | http://arxiv.org/pdf/1510.08983v2.pdf | Highway Long Short-Term Memory RNNs for Distant Speech Recognition | In this paper, we extend the deep long short-term memory (DLSTM) recurrent
neural networks by introducing gated direct connections between memory cells in
adjacent layers. These direct links, called highway connections, enable
unimpeded information flow across different layers and thus alleviate the
gradient vanishing problem when building deeper LSTMs. We further introduce the
latency-controlled bidirectional LSTMs (BLSTMs) which can exploit the whole
history while keeping the latency under control. Efficient algorithms are
proposed to train these novel networks using both frame and sequence
discriminative criteria. Experiments on the AMI distant speech recognition
(DSR) task indicate that we can train deeper LSTMs and achieve better
improvement from sequence training with highway LSTMs (HLSTMs). Our novel model
obtains $43.9/47.7\%$ WER on AMI (SDM) dev and eval sets, outperforming all
previous works. It beats the strong DNN and DLSTM baselines with $15.7\%$ and
$5.3\%$ relative improvement respectively. | ['Dong Yu', 'Sanjeev Khudanpur', 'James Glass', 'Guoguo Chen', 'Yu Zhang', 'Kaisheng Yao'] | 2015-10-30 | null | null | null | null | ['distant-speech-recognition'] | ['speech'] | [ 1.14127181e-01 -6.40480667e-02 -1.39110208e-01 -3.14100325e-01
-7.88374782e-01 -1.60891622e-01 5.78808427e-01 -2.30372459e-01
-7.95365572e-01 9.19337988e-01 2.70375550e-01 -8.05887580e-01
5.69666147e-01 -6.63890183e-01 -8.53154242e-01 -6.78117096e-01
-1.20280758e-01 -1.60192028e-02 6.25654995e-01 -2.20344797e-01
2.13448599e-01 3.21653754e-01 -9.83951151e-01 6.19656324e-01
4.97855902e-01 1.07954586e+00 5.71055055e-01 8.95929515e-01
-2.76674986e-01 1.41761494e+00 -5.54455280e-01 -1.73032492e-01
-2.23003402e-01 -3.02842706e-01 -9.18349981e-01 -3.30620259e-01
1.74695149e-01 -4.36662078e-01 -7.09791839e-01 4.23749268e-01
7.70528138e-01 4.23823416e-01 1.33997589e-01 -5.59635341e-01
-7.68889964e-01 7.53886938e-01 -5.79371870e-01 7.84242988e-01
5.09462282e-02 1.08533874e-01 9.72675979e-01 -1.07287586e+00
3.71024191e-01 1.12970400e+00 5.05393445e-01 8.06100488e-01
-8.04752350e-01 -7.78162539e-01 5.46628654e-01 5.17708123e-01
-1.22815514e+00 -8.89084101e-01 3.88643354e-01 -2.96394657e-02
1.97011125e+00 1.60873994e-01 2.06173211e-01 1.37807333e+00
1.90848261e-01 1.11119640e+00 7.03281283e-01 -4.49577153e-01
-9.23361480e-02 -2.78110921e-01 2.88656980e-01 8.98618639e-01
-4.25099880e-01 1.33432761e-01 -8.58743727e-01 2.73767859e-01
7.43952990e-01 1.83433041e-01 -2.72684395e-01 5.01280487e-01
-1.48162961e+00 5.99431038e-01 6.66053593e-01 5.29726863e-01
-4.34444755e-01 3.98643523e-01 7.34322488e-01 5.96342504e-01
6.95771098e-01 -1.96076483e-01 -6.34640157e-01 -4.74430889e-01
-1.02535534e+00 -4.54816699e-01 4.45852518e-01 8.80369723e-01
5.64467132e-01 7.11071014e-01 -4.07324016e-01 1.16653430e+00
2.25522116e-01 3.39708835e-01 8.92403662e-01 -6.44406974e-01
8.01617503e-01 1.94687381e-01 -3.28672767e-01 -6.00801528e-01
-1.49332985e-01 -6.54379547e-01 -1.08419418e+00 -1.55062303e-01
-1.70094296e-01 -9.57971960e-02 -1.28912425e+00 1.68328381e+00
-3.23919922e-01 5.94887853e-01 2.01901808e-01 4.50397938e-01
6.71478689e-01 1.17598784e+00 1.78923473e-01 -2.59678125e-01
7.97399044e-01 -1.46713388e+00 -8.08409750e-01 -4.47163582e-01
1.22743320e+00 -8.75566363e-01 1.11471641e+00 -1.15627330e-02
-1.28798902e+00 -6.95518434e-01 -1.13247406e+00 -1.74773708e-01
-3.18250626e-01 -3.95957977e-02 1.52224168e-01 2.45181873e-01
-1.50990450e+00 5.74355900e-01 -1.12342620e+00 -1.71652287e-01
3.20329249e-01 5.04215658e-01 -3.83852087e-02 1.58425093e-01
-1.33368492e+00 8.76539886e-01 2.34515786e-01 3.86364102e-01
-1.18175459e+00 -5.36352277e-01 -8.33031237e-01 -7.57256374e-02
-4.86705266e-03 -6.20119333e-01 1.43590200e+00 -9.08688366e-01
-1.67785764e+00 7.67098904e-01 -5.90647221e-01 -1.13650239e+00
2.68781573e-01 -3.42645556e-01 -4.65148717e-01 -1.10701129e-01
-3.24210942e-01 1.04964221e+00 5.57454824e-01 -5.27623653e-01
-7.50997305e-01 1.73712205e-02 -2.38837019e-01 3.80107090e-02
-3.37177664e-01 1.67698398e-01 -1.17786139e-01 -9.94558334e-01
9.86509174e-02 -9.78704393e-01 -2.04217792e-01 -4.41469401e-01
-3.47840846e-01 -3.47284824e-01 1.27665305e+00 -8.99474382e-01
1.54232550e+00 -2.05072904e+00 -1.09035689e-02 -2.85920233e-01
9.04852226e-02 8.03715885e-01 -2.89522707e-01 3.16076100e-01
1.02117717e-01 9.83475298e-02 -1.24404900e-01 -6.13054335e-01
-2.23474547e-01 5.64855397e-01 -6.31583393e-01 1.84252679e-01
9.35880765e-02 1.18653321e+00 -4.83697712e-01 -2.89351851e-01
2.29946539e-01 7.41524935e-01 -1.32626072e-01 2.68994838e-01
-1.73938736e-01 3.70972842e-01 -5.67761101e-02 4.03659344e-01
2.16022551e-01 -1.25899792e-01 -1.81998193e-01 2.72995889e-01
-2.67921835e-01 8.87616813e-01 -3.86068970e-01 2.03577828e+00
-7.94328511e-01 9.94158566e-01 -4.16010112e-01 -1.07000935e+00
1.21396017e+00 6.74069285e-01 -2.53409624e-01 -1.22822726e+00
-1.05442621e-01 4.30964470e-01 3.52833271e-02 -1.28031880e-01
3.58534217e-01 -8.95881057e-02 2.43032977e-01 3.30033541e-01
4.79858927e-02 7.21922815e-01 -2.91374236e-01 9.75311175e-02
1.16794121e+00 3.67734097e-02 -1.38926446e-01 -1.68561220e-01
6.20165706e-01 -5.94919384e-01 6.78635538e-01 6.16355598e-01
-1.90404028e-01 4.73352820e-01 8.17810446e-02 -6.06303394e-01
-1.06064951e+00 -1.01869023e+00 3.47098559e-01 1.41428065e+00
-4.25923973e-01 -1.41210601e-01 -5.01758575e-01 -3.84407997e-01
-6.42855644e-01 7.04979479e-01 -4.27225292e-01 -2.13815585e-01
-1.23623478e+00 -4.00264829e-01 9.52109516e-01 9.97936845e-01
9.32427406e-01 -1.31161880e+00 -5.09352982e-01 5.35133779e-01
-3.53845209e-01 -1.23089385e+00 -6.06181324e-01 1.51428506e-01
-1.16467869e+00 -2.16301784e-01 -1.16087186e+00 -1.11357355e+00
1.65537283e-01 2.55402774e-01 1.11907434e+00 2.42664851e-02
7.57177696e-02 -3.10551852e-01 -4.21205789e-01 4.89651449e-02
-1.91416264e-01 5.65172613e-01 -1.13753565e-01 -1.98033214e-01
3.63695443e-01 -8.38240683e-01 -7.48072982e-01 3.36901098e-01
-4.03662741e-01 3.91513288e-01 6.66796029e-01 1.03409505e+00
5.89450955e-01 -6.57555044e-01 8.40906620e-01 -5.51543593e-01
2.41206720e-01 -2.63944209e-01 -2.31858820e-01 2.13480189e-01
-5.53120375e-01 3.57567519e-01 5.06960928e-01 -2.49124333e-01
-1.16435421e+00 -4.68762964e-01 -5.80748796e-01 -4.96881247e-01
3.57277840e-02 3.63964736e-01 1.86649770e-01 2.24170417e-01
2.27633059e-01 6.06428087e-01 -4.32424277e-01 -6.65865004e-01
8.59494731e-02 5.68521559e-01 3.68551165e-01 -2.40200579e-01
1.78501397e-01 2.90402591e-01 -5.59456050e-01 -9.18568671e-01
-6.91506088e-01 -2.25203067e-01 -5.06213784e-01 8.51615518e-02
9.08463657e-01 -1.03104186e+00 -5.46556354e-01 5.87609589e-01
-1.45132208e+00 -7.91568995e-01 1.16826165e-02 5.20028591e-01
-4.28646982e-01 7.69355074e-02 -1.29925525e+00 -7.79018104e-01
-7.90483952e-01 -1.14559340e+00 7.05215514e-01 -1.40853018e-01
-2.15868294e-01 -1.14301312e+00 4.82450500e-02 8.34746882e-02
7.91593552e-01 -4.27735820e-02 7.04811752e-01 -6.20813370e-01
-7.45730221e-01 1.34925351e-01 -3.79453212e-01 7.21901774e-01
-4.03580964e-02 -2.29160219e-01 -1.14939940e+00 -4.26679611e-01
-1.92149743e-01 -5.46171702e-02 1.53002477e+00 4.44694102e-01
9.34745610e-01 -4.33285147e-01 -3.90471905e-01 5.45091271e-01
1.04777288e+00 6.24035716e-01 9.88078654e-01 4.69875306e-01
8.93664837e-01 2.60519058e-01 1.22318640e-01 8.07351694e-02
3.13997000e-01 5.36471367e-01 5.54932617e-02 -2.70916700e-01
-4.15293783e-01 -3.64946157e-01 9.32241738e-01 1.60573542e+00
-2.35490669e-02 -3.14128429e-01 -9.92756248e-01 6.26879156e-01
-1.96124017e+00 -8.79671514e-01 1.12545885e-01 1.91166067e+00
8.05605233e-01 5.76367199e-01 -8.46438035e-02 2.73658425e-01
7.94905782e-01 5.79703093e-01 -4.11791772e-01 -9.23809946e-01
-2.94037282e-01 6.08865857e-01 5.68292320e-01 6.97010577e-01
-7.69191861e-01 1.26302040e+00 5.85485363e+00 1.05764079e+00
-1.34270191e+00 4.27329868e-01 9.67021286e-01 -1.92239195e-01
-1.61207452e-01 -5.38749248e-02 -1.00827575e+00 3.89841050e-01
1.81291878e+00 2.20125049e-01 5.32591715e-02 4.94004011e-01
5.42254627e-01 1.85447231e-01 -8.74008119e-01 9.49254751e-01
-2.22239181e-01 -1.63036442e+00 1.81269869e-01 -5.49661778e-02
6.56764686e-01 6.22970343e-01 4.14331228e-01 4.49438900e-01
3.37218523e-01 -1.15726602e+00 5.63004315e-01 4.66333926e-01
6.92897379e-01 -1.03568840e+00 7.95721412e-01 3.21425617e-01
-1.35763538e+00 4.23006937e-02 -3.29149812e-01 -3.17835599e-01
4.35872912e-01 4.38806236e-01 -7.69035280e-01 2.59631962e-01
7.22947240e-01 9.58424687e-01 -2.78497487e-01 4.92654413e-01
-1.80752024e-01 8.54095459e-01 -6.02318086e-02 2.65451148e-02
8.82683098e-01 1.54739186e-01 3.00300568e-01 1.64610851e+00
4.03440684e-01 -2.40093425e-01 -1.47359639e-01 4.36828077e-01
-3.72194111e-01 -1.82359800e-01 -5.89119434e-01 -7.51481159e-03
3.95540804e-01 6.45019054e-01 -4.33061182e-01 -5.48275709e-01
-5.00150919e-01 1.29286909e+00 3.27063322e-01 5.87763309e-01
-1.00247836e+00 -4.90587384e-01 6.46229506e-01 -6.32859394e-02
5.52045643e-01 -5.32073200e-01 -1.79346591e-01 -1.06500077e+00
8.46798047e-02 -5.57138503e-01 4.48843062e-01 -6.36679590e-01
-7.86867797e-01 1.06682801e+00 -6.41256511e-01 -1.00873721e+00
-4.29041058e-01 -3.39256167e-01 -6.48013949e-01 8.99214506e-01
-2.05481887e+00 -1.25224245e+00 3.28225791e-01 7.04510093e-01
1.13436425e+00 -2.00743958e-01 7.99365938e-01 6.57979548e-01
-7.99913406e-01 8.18504095e-01 2.53492314e-02 2.61545479e-01
4.04840946e-01 -8.06925356e-01 9.92580533e-01 9.57206905e-01
1.78974271e-01 6.02363169e-01 3.39684844e-01 -2.53717542e-01
-8.75476480e-01 -1.27643645e+00 1.28884542e+00 -1.49753150e-02
6.36448205e-01 -3.43264639e-01 -1.08469474e+00 1.13327336e+00
5.86562335e-01 -4.91313376e-02 4.43010896e-01 -7.14929998e-02
-6.14051938e-01 -1.13334231e-01 -6.00628078e-01 4.31785941e-01
1.21532428e+00 -8.32124352e-01 -4.99969631e-01 -9.78877768e-02
1.37627256e+00 -9.68262777e-02 -5.72636127e-01 4.73892570e-01
4.37661469e-01 -9.89978254e-01 9.35412824e-01 -3.98760051e-01
1.33956537e-01 1.64869241e-02 -4.06290621e-01 -1.02265680e+00
5.79599142e-02 -8.34098279e-01 -4.25502986e-01 1.22591031e+00
8.68726313e-01 -8.45578194e-01 5.37527084e-01 1.57127783e-01
-7.03959346e-01 -8.50285888e-01 -1.13250566e+00 -9.19814408e-01
2.47788250e-01 -5.31453192e-01 2.91795313e-01 7.02201188e-01
-1.63568735e-01 4.49910730e-01 -6.53939724e-01 -1.25691026e-01
1.16078079e-01 -2.48125479e-01 1.47467852e-01 -4.72694755e-01
-1.69032425e-01 -4.07152057e-01 -2.63978899e-01 -1.70775092e+00
2.80036300e-01 -7.04449296e-01 -8.06217119e-02 -1.45666683e+00
-7.84239992e-02 -2.62722284e-01 -8.53537083e-01 5.74495137e-01
2.65236676e-01 1.67757541e-01 3.63059156e-03 1.10019885e-01
-6.94611490e-01 6.10039771e-01 9.30600166e-01 -3.96255106e-02
-1.76801890e-01 -1.13914505e-01 -1.17151171e-01 6.42178118e-01
1.06427872e+00 -2.52750516e-01 -2.64894068e-01 -1.05229342e+00
-6.43471479e-02 1.59808487e-01 3.29558700e-01 -9.61385608e-01
4.70304281e-01 3.68923813e-01 1.61884412e-01 -8.96260560e-01
4.65806335e-01 -1.34890288e-01 -2.73885429e-01 6.86979949e-01
-7.22520828e-01 5.96553266e-01 3.52974623e-01 4.09968674e-01
-5.43335915e-01 4.70662951e-01 9.25124526e-01 -2.42750794e-01
-9.63446200e-01 3.46460462e-01 -5.84138811e-01 -1.40838146e-01
6.96438789e-01 -1.57970965e-01 -2.80535251e-01 -3.35364163e-01
-9.21982408e-01 3.50388288e-02 -1.96859345e-01 5.36351204e-01
1.04966676e+00 -1.37120521e+00 -6.84666991e-01 2.60723621e-01
-2.63133317e-01 -2.15218529e-01 3.87281835e-01 9.27190423e-01
-4.14406925e-01 8.63623738e-01 -1.59731582e-02 -5.34308314e-01
-1.26343727e+00 3.00536245e-01 3.72737467e-01 -3.72021198e-01
-7.34088123e-01 1.37408674e+00 2.60863543e-01 -8.58614743e-02
5.39738297e-01 -5.10759175e-01 1.35442726e-02 -9.63188186e-02
6.42831326e-01 6.07021749e-01 2.60303408e-01 -5.69628119e-01
-4.16748375e-01 2.22306401e-01 -4.28234935e-01 -2.65438020e-01
1.28602803e+00 -5.02329350e-01 -1.88268557e-01 7.90213764e-01
1.63508952e+00 -6.50592804e-01 -1.08184791e+00 -5.47081649e-01
3.41999590e-01 -2.26969030e-02 1.53015316e-01 -5.87564826e-01
-1.25285423e+00 1.51467943e+00 7.17077494e-01 -8.66647586e-02
1.01555181e+00 -7.50177801e-02 1.81738901e+00 2.88944721e-01
6.55696169e-02 -9.43835080e-01 2.39165783e-01 9.59850490e-01
5.78321397e-01 -8.44351113e-01 -9.11508262e-01 3.81782562e-01
-4.46421713e-01 1.15362155e+00 4.60835159e-01 -2.02989236e-01
5.64374328e-01 3.16728055e-01 1.23287469e-01 1.19059987e-01
-1.47093654e+00 -1.69296339e-01 1.60923786e-02 1.13570176e-01
6.98964894e-01 -7.25048734e-03 -7.93960020e-02 -2.00244058e-02
2.64323615e-02 -5.99227101e-02 1.48845583e-01 8.58721554e-01
-7.40396321e-01 -1.13710916e+00 1.39968649e-01 1.42272174e-01
-6.48873746e-01 -6.32910848e-01 1.46971121e-01 3.87528002e-01
-2.68548340e-01 9.65213656e-01 2.18585283e-01 -5.22598386e-01
2.48751029e-01 2.60852218e-01 1.22213252e-01 -3.59075993e-01
-5.19066691e-01 2.01880947e-01 2.48460785e-01 -5.14919877e-01
-3.39073360e-01 -2.68554986e-01 -1.39395964e+00 -6.44649923e-01
-2.33185545e-01 -2.66426709e-02 4.85169619e-01 1.07060957e+00
6.06385112e-01 8.71911883e-01 5.04029155e-01 -5.10895908e-01
-1.30710229e-01 -1.02509296e+00 -1.42787740e-01 -1.76257029e-01
7.60673046e-01 -1.25137284e-01 -2.70088732e-01 1.13030769e-01] | [10.918557167053223, 6.326347351074219] |
62887cea-4c71-4275-8389-e4a3effec481 | monodvps-a-self-supervised-monocular-depth | 2210.07577 | null | https://arxiv.org/abs/2210.07577v1 | https://arxiv.org/pdf/2210.07577v1.pdf | MonoDVPS: A Self-Supervised Monocular Depth Estimation Approach to Depth-aware Video Panoptic Segmentation | Depth-aware video panoptic segmentation tackles the inverse projection problem of restoring panoptic 3D point clouds from video sequences, where the 3D points are augmented with semantic classes and temporally consistent instance identifiers. We propose a novel solution with a multi-task network that performs monocular depth estimation and video panoptic segmentation. Since acquiring ground truth labels for both depth and image segmentation has a relatively large cost, we leverage the power of unlabeled video sequences with self-supervised monocular depth estimation and semi-supervised learning from pseudo-labels for video panoptic segmentation. To further improve the depth prediction, we introduce panoptic-guided depth losses and a novel panoptic masking scheme for moving objects to avoid corrupting the training signal. Extensive experiments on the Cityscapes-DVPS and SemKITTI-DVPS datasets demonstrate that our model with the proposed improvements achieves competitive results and fast inference speed. | ['Sergiu Nedevschi', 'Andra Petrovai'] | 2022-10-14 | null | null | null | null | ['panoptic-segmentation'] | ['computer-vision'] | [ 4.44399744e-01 -1.96559504e-01 -5.50898910e-01 -5.39084256e-01
-9.24334049e-01 -8.28768253e-01 5.02876520e-01 -6.69406712e-01
-2.83495516e-01 5.09851933e-01 -3.09054609e-02 -3.05434048e-01
8.17893073e-02 -9.32952881e-01 -9.69320893e-01 -9.24711287e-01
-1.02554962e-01 6.98071837e-01 3.77757370e-01 5.29433131e-01
1.53064385e-01 3.56435716e-01 -1.47448432e+00 3.16518903e-01
7.46099651e-01 1.21948004e+00 4.76619393e-01 8.67420912e-01
-5.95270023e-02 8.52452755e-01 -7.80597106e-02 -4.50250842e-02
1.18055212e+00 -9.99299586e-02 -7.55982637e-01 7.87311018e-01
1.22851014e+00 -1.00421298e+00 -5.39751232e-01 1.19997156e+00
-1.27355412e-01 2.65381664e-01 4.73020196e-01 -1.20829129e+00
-2.17802897e-01 1.83126539e-01 -9.64175403e-01 3.73548061e-01
2.89437100e-02 1.01323448e-01 1.08940542e+00 -5.64216495e-01
7.11429000e-01 1.34487987e+00 4.96834010e-01 3.63402218e-01
-9.30787981e-01 -8.53562772e-01 5.27429044e-01 4.42241952e-02
-1.29747164e+00 -2.49550492e-01 6.27891600e-01 -5.24018586e-01
7.60153830e-01 1.77464157e-01 5.98217905e-01 5.64843357e-01
-2.65509933e-01 8.31658185e-01 1.02802968e+00 6.25384077e-02
1.01061933e-01 -7.98499733e-02 -2.50692874e-01 8.08349967e-01
5.44854887e-02 1.95525706e-01 -4.37532157e-01 6.49063140e-02
1.16017151e+00 3.32782686e-01 -3.09292197e-01 -3.77651155e-01
-1.18800569e+00 6.30753875e-01 3.42069477e-01 -3.00695360e-01
-2.31717616e-01 3.69867653e-01 3.19912732e-02 2.38716528e-01
1.05295587e+00 8.33416507e-02 -8.21606398e-01 3.09386641e-01
-1.37741601e+00 3.06564808e-01 4.59869862e-01 1.17089713e+00
1.31934035e+00 1.32388368e-01 3.71494234e-01 4.91033405e-01
4.10230011e-01 1.13433826e+00 -6.45336136e-02 -1.78375256e+00
8.85003924e-01 2.93934762e-01 2.64198780e-01 -9.88513947e-01
-2.52019733e-01 9.10925716e-02 -6.17620647e-01 1.03811592e-01
5.25141537e-01 -1.59191728e-01 -1.28478217e+00 1.59459388e+00
7.41482794e-01 8.73595357e-01 3.07558328e-02 1.28861904e+00
4.89960521e-01 1.09366381e+00 -2.34340563e-01 -2.11721435e-01
1.02363932e+00 -1.03898418e+00 -3.07381928e-01 -4.94957507e-01
4.18332189e-01 -3.97239000e-01 5.13450503e-01 2.58630663e-01
-7.95895219e-01 -4.96124744e-01 -7.72243083e-01 -3.33489567e-01
-5.35675697e-02 -6.11820519e-02 7.56089687e-01 4.68814790e-01
-1.13404000e+00 4.25533980e-01 -9.37558651e-01 -1.96647584e-01
3.86999816e-01 2.97443926e-01 -1.47466928e-01 -2.27187350e-01
-9.66667891e-01 2.39486739e-01 3.37272525e-01 -4.20136889e-03
-1.41115201e+00 -8.91235411e-01 -9.62337673e-01 -2.46955022e-01
5.59279740e-01 -6.07353449e-01 1.13730526e+00 -9.22065496e-01
-1.24924910e+00 1.19935012e+00 -1.33080289e-01 -6.53900981e-01
4.91797179e-01 -2.61421710e-01 -8.56931359e-02 9.78204072e-01
3.47738206e-01 1.54897904e+00 1.10523820e+00 -1.17322063e+00
-1.38389242e+00 -4.70807701e-01 1.10489547e-01 6.47808075e-01
6.71804994e-02 -3.59221816e-01 -7.79335558e-01 -4.13267195e-01
6.95246816e-01 -1.13711262e+00 -4.33484077e-01 2.78230458e-01
-2.63489276e-01 3.21618617e-01 8.86705458e-01 -9.38404918e-01
4.14621681e-01 -1.93522489e+00 4.53039184e-02 -2.39096005e-02
1.91789299e-01 -2.47818809e-02 -1.44635230e-01 -4.49864388e-01
1.87737614e-01 -1.95567220e-01 -6.21634841e-01 -5.07036924e-01
-4.19217139e-01 5.00547409e-01 -6.54739320e-01 6.14364684e-01
-1.54897690e-01 6.60941303e-01 -9.07128453e-01 -5.63221216e-01
7.30621994e-01 1.22175239e-01 -6.95059240e-01 1.49723858e-01
-8.51711392e-01 7.21802711e-01 -6.14308536e-01 9.26230252e-01
1.27734673e+00 -3.58438343e-01 -9.33078006e-02 1.31243631e-01
-2.99474239e-01 3.93985301e-01 -1.05870736e+00 1.87001801e+00
-2.40572870e-01 5.97489357e-01 3.81432205e-01 -8.57115328e-01
6.25795603e-01 2.06650034e-01 7.94949532e-01 -5.60744166e-01
-9.98453200e-02 2.04516891e-02 -7.54708827e-01 -5.63037634e-01
8.43783021e-01 -6.18824810e-02 1.67219967e-01 4.86666441e-01
-2.18658850e-01 -8.29130411e-01 -1.70576721e-01 2.05717400e-01
6.21115685e-01 4.47020471e-01 -1.90942422e-01 -8.50568041e-02
4.25794750e-01 3.56917411e-01 9.54380035e-01 6.13402903e-01
-1.80953637e-01 9.65404391e-01 1.34818569e-01 -5.38713038e-01
-1.09894967e+00 -1.09549546e+00 -8.08337480e-02 8.27928841e-01
3.94468576e-01 2.10941583e-01 -5.66946924e-01 -5.05931973e-01
4.95293457e-03 4.05975103e-01 -3.99161458e-01 6.50554597e-01
-7.02843189e-01 -9.87780690e-01 3.61041337e-01 2.92514980e-01
7.65241146e-01 -5.97455025e-01 -6.26820326e-01 2.52208244e-02
-5.80890298e-01 -1.83689404e+00 -5.18311024e-01 1.24176659e-01
-1.29313242e+00 -1.05966055e+00 -7.59360373e-01 -7.92902291e-01
4.98891264e-01 1.00497484e+00 8.30802977e-01 -3.90984893e-01
1.04543693e-01 3.69414270e-01 1.56112341e-03 1.40557572e-01
-1.58932611e-01 8.85416493e-02 -1.53711706e-01 9.49235708e-02
3.51333350e-01 -7.57370412e-01 -6.68573618e-01 3.03475618e-01
-8.60054791e-01 4.45247680e-01 2.03199759e-01 1.92703858e-01
9.03657615e-01 -1.19096346e-01 -1.04202569e-01 -8.16833019e-01
-7.55688071e-01 -4.76129740e-01 -1.38738763e+00 2.31117243e-03
-4.78593349e-01 -3.43956113e-01 1.63575649e-01 -1.20055132e-01
-1.34060049e+00 5.15409172e-01 1.04807965e-01 -9.93200719e-01
-1.95573360e-01 -7.02104643e-02 -6.45620525e-02 -1.74247950e-01
2.40335912e-01 2.60217398e-01 -2.45794266e-01 -4.76367503e-01
5.77903450e-01 5.73965847e-01 9.02484536e-01 -3.93500239e-01
9.45605695e-01 1.28237629e+00 -2.23548319e-02 -9.06967103e-01
-1.18767869e+00 -9.13324594e-01 -8.19748998e-01 -8.37109238e-02
1.22230375e+00 -1.78789842e+00 -3.69147927e-01 7.71134198e-01
-1.17554474e+00 -5.99735141e-01 -9.05825719e-02 5.57430148e-01
-6.84916615e-01 7.71819711e-01 -8.52519333e-01 -6.33525014e-01
-2.94294596e-01 -8.98956060e-01 1.40671194e+00 6.60475343e-02
4.52665269e-01 -1.02286303e+00 7.37589644e-03 7.51144886e-01
-3.31905305e-01 1.48946315e-01 5.53050399e-01 -1.30120233e-01
-1.77319372e+00 3.59814018e-01 -6.54941559e-01 4.17647392e-01
1.95242926e-01 -1.34144410e-01 -1.11213255e+00 -1.59735568e-02
3.53491932e-01 -9.55383331e-02 1.19660676e+00 8.89700115e-01
1.10032344e+00 -2.50646949e-01 -4.10725176e-01 1.39129603e+00
1.44474375e+00 2.47022465e-01 3.70342106e-01 4.06998277e-01
1.23076916e+00 8.16326499e-01 8.70110273e-01 4.13563251e-01
7.52389312e-01 2.86140263e-01 6.94526315e-01 9.28225666e-02
-4.70802449e-02 -2.94424027e-01 2.05470979e-01 6.48419738e-01
3.16311955e-01 -4.90139544e-01 -7.73781776e-01 8.65590453e-01
-1.78474581e+00 -8.13467264e-01 -2.93760985e-01 2.00917077e+00
5.90405524e-01 -1.87031537e-01 -3.38692427e-01 -1.96712688e-01
6.55102074e-01 8.00209701e-01 -6.64232314e-01 4.54318523e-01
-4.36998814e-01 -2.22497091e-01 1.17830420e+00 8.32994938e-01
-1.52032197e+00 1.30898595e+00 6.18743849e+00 5.28676927e-01
-1.06966794e+00 3.99195284e-01 8.74306679e-01 -3.20392221e-01
-3.67434978e-01 7.31348991e-02 -1.06418216e+00 3.37972909e-01
6.24819756e-01 3.21262032e-01 5.65599501e-01 6.80394888e-01
3.24635446e-01 -3.52422416e-01 -8.86179209e-01 1.20368695e+00
7.37570375e-02 -1.57100463e+00 1.94882095e-01 9.66655165e-02
1.39902925e+00 8.71505499e-01 1.08939365e-01 -3.71337891e-01
4.88732606e-01 -2.59922683e-01 6.08085871e-01 1.71314150e-01
6.74519539e-01 -4.23529863e-01 5.07697538e-02 1.82400852e-01
-1.15386343e+00 -1.88766226e-01 -5.20512104e-01 -1.45633817e-01
5.97525358e-01 6.71255350e-01 -6.21632636e-01 5.70794642e-01
9.35235202e-01 1.31247222e+00 -1.49028465e-01 9.45164740e-01
-3.83177072e-01 4.25952554e-01 -7.23310411e-01 7.62972176e-01
7.48380542e-01 -7.56969154e-01 6.30049229e-01 7.68557727e-01
4.43925977e-01 4.29829627e-01 3.46003056e-01 8.12428772e-01
-9.64214653e-02 -3.31064761e-01 -6.50849223e-01 2.40541264e-01
3.29405099e-01 1.18536615e+00 -8.53772402e-01 -6.47740602e-01
-5.31750083e-01 8.54801416e-01 -3.38694900e-01 4.89240199e-01
-6.60385191e-01 4.90957111e-01 1.10130858e+00 -1.66190285e-02
5.57484925e-01 -3.43583047e-01 -4.83704120e-01 -1.43563855e+00
-2.42224410e-01 -5.04138112e-01 5.41710377e-01 -1.11268485e+00
-7.07323134e-01 1.70046881e-01 2.88103282e-01 -1.23442638e+00
-2.96196133e-01 -4.57259029e-01 -2.09214821e-01 3.93904418e-01
-1.97394502e+00 -1.10372317e+00 -4.99949753e-01 5.92532277e-01
8.06932628e-01 2.45676830e-01 4.34774011e-02 5.45153558e-01
-1.88384205e-01 -2.43501544e-01 2.55262494e-01 1.77508146e-02
4.57274705e-01 -9.99865890e-01 7.63890922e-01 1.22070611e+00
1.70315966e-01 -2.34025404e-01 4.57652986e-01 -7.87359774e-01
-8.51971209e-01 -1.67195868e+00 7.29045212e-01 -5.10380447e-01
6.24367058e-01 -2.65948564e-01 -8.49538743e-01 1.06336701e+00
-2.74513721e-01 -1.32697284e-01 2.97983661e-02 -5.83292603e-01
-2.44436696e-01 -2.33145654e-01 -1.01482594e+00 3.32123131e-01
1.23599899e+00 -7.30695665e-01 -4.49089766e-01 8.33990097e-01
1.18642056e+00 -6.18954003e-01 -3.40015382e-01 4.19410616e-01
2.02144727e-01 -7.46367931e-01 1.26275539e+00 1.42880142e-01
4.25357431e-01 -4.74503875e-01 -4.77809429e-01 -8.08479607e-01
2.51694113e-01 -5.85102797e-01 1.97419897e-01 8.74661088e-01
1.21598177e-01 -3.54683131e-01 1.24621594e+00 4.70261723e-01
-2.09835321e-01 1.83767438e-01 -1.07188845e+00 -5.48956156e-01
-1.20222226e-01 -6.65830314e-01 5.60433090e-01 1.09543884e+00
-7.75768399e-01 -1.61964148e-02 -8.07041883e-01 7.83325851e-01
9.84370291e-01 4.97909963e-01 6.86704695e-01 -1.15599334e+00
-1.50881514e-01 1.44145519e-01 -8.98595154e-03 -1.83476400e+00
3.56789678e-01 -7.12600112e-01 2.85388172e-01 -1.14863455e+00
1.16730578e-01 -4.92608160e-01 5.21189012e-02 2.35068217e-01
1.66735753e-01 4.51060385e-01 -7.23763481e-02 4.69181895e-01
-6.96455240e-01 4.54871237e-01 1.39518476e+00 -2.19533995e-01
-3.13888639e-01 1.33604750e-01 -5.93021996e-02 8.34756076e-01
5.07710993e-01 -8.31375897e-01 -7.58901596e-01 -1.12448418e+00
1.60920322e-01 4.70623404e-01 5.12201965e-01 -8.77365887e-01
2.63483465e-01 -4.23694164e-01 9.71759334e-02 -1.31035483e+00
7.56174147e-01 -8.10431480e-01 8.12644362e-02 2.03971252e-01
-3.46842930e-02 -1.66576505e-01 -1.14049539e-02 8.39482963e-01
-1.19152129e-01 -5.65983960e-03 8.09914947e-01 -3.97842884e-01
-1.10257518e+00 8.95313084e-01 -5.00889838e-01 -4.17927131e-02
8.59305799e-01 -3.03031176e-01 -2.03938991e-01 -3.24537665e-01
-4.23866391e-01 6.22268796e-01 6.16487801e-01 4.82277095e-01
4.92054254e-01 -7.09272921e-01 -4.85740125e-01 3.45505655e-01
-1.82516575e-01 7.15996325e-01 4.53515112e-01 6.24803841e-01
-1.01523042e+00 6.47303760e-01 -8.11525658e-02 -1.14155018e+00
-1.03815591e+00 5.14512777e-01 5.00267804e-01 2.09395260e-01
-9.21464801e-01 9.80174363e-01 9.19572711e-01 -7.14390278e-01
1.74568757e-01 -9.56491888e-01 6.17300421e-02 1.25363410e-01
4.14142132e-01 3.04958194e-01 -2.49699488e-01 -5.39846957e-01
-1.35187060e-01 7.31583893e-01 3.32992263e-02 -3.85403156e-01
1.36816657e+00 -8.42499554e-01 -1.43608257e-01 3.99302781e-01
1.17495537e+00 -5.47564089e-01 -2.10574770e+00 -6.06661141e-01
-2.11618140e-01 -7.89638817e-01 2.31683716e-01 -2.81380594e-01
-1.50467885e+00 9.59346116e-01 4.18770134e-01 -4.49851453e-01
1.11684585e+00 -1.29881710e-01 1.13135374e+00 6.48305357e-01
4.39351112e-01 -1.01689494e+00 -1.53580263e-01 4.94617194e-01
1.83277763e-02 -1.47263765e+00 9.18738991e-02 -7.02965081e-01
-4.23676133e-01 6.95315540e-01 5.46120048e-01 -1.24703757e-01
6.31234825e-01 -2.78724115e-02 3.50167811e-01 -1.72582954e-01
-7.39527762e-01 -2.31363311e-01 -5.91543168e-02 5.23682177e-01
-5.04525602e-01 -5.90768047e-02 5.29617369e-01 -2.90503412e-01
6.58127591e-02 -1.77421758e-03 5.08895576e-01 4.46030170e-01
-6.11973524e-01 -5.11889994e-01 -4.91311759e-01 2.37852439e-01
-3.49353105e-01 -3.35880429e-01 4.84040715e-02 7.01122463e-01
1.30969256e-01 8.20201457e-01 6.83168471e-01 -1.64410740e-01
-4.82437462e-01 -1.85648501e-01 2.87773162e-01 -5.30853212e-01
-1.23874441e-01 4.44923788e-01 -9.03400183e-02 -4.70269382e-01
-9.92781579e-01 -9.39117610e-01 -1.17284799e+00 -4.47369307e-01
1.26231490e-02 3.07427663e-02 5.04582524e-01 1.04611063e+00
9.45588723e-02 -1.04711782e-02 9.20335889e-01 -1.13330603e+00
-5.63468263e-02 -6.51921988e-01 -6.95590317e-01 1.23906001e-01
7.49794662e-01 -3.70407790e-01 -6.71562493e-01 3.78247082e-01] | [8.5147123336792, -2.0112040042877197] |
c3dfc4a6-c769-4a34-99f9-d823bd19b492 | a-multi-task-multi-stage-transitional | 2301.11749 | null | https://arxiv.org/abs/2301.11749v1 | https://arxiv.org/pdf/2301.11749v1.pdf | A Multi-task Multi-stage Transitional Training Framework for Neural Chat Translation | Neural chat translation (NCT) aims to translate a cross-lingual chat between speakers of different languages. Existing context-aware NMT models cannot achieve satisfactory performances due to the following inherent problems: 1) limited resources of annotated bilingual dialogues; 2) the neglect of modelling conversational properties; 3) training discrepancy between different stages. To address these issues, in this paper, we propose a multi-task multi-stage transitional (MMT) training framework, where an NCT model is trained using the bilingual chat translation dataset and additional monolingual dialogues. We elaborately design two auxiliary tasks, namely utterance discrimination and speaker discrimination, to introduce the modelling of dialogue coherence and speaker characteristic into the NCT model. The training process consists of three stages: 1) sentence-level pre-training on large-scale parallel corpus; 2) intermediate training with auxiliary tasks using additional monolingual dialogues; 3) context-aware fine-tuning with gradual transition. Particularly, the second stage serves as an intermediate phase that alleviates the training discrepancy between the pre-training and fine-tuning stages. Moreover, to make the stage transition smoother, we train the NCT model using a gradual transition strategy, i.e., gradually transiting from using monolingual to bilingual dialogues. Extensive experiments on two language pairs demonstrate the effectiveness and superiority of our proposed training framework. | ['Jinsong Su', 'Min Zhang', 'Hongji Wang', 'Jinan Xu', 'Jie zhou', 'Fandong Meng', 'Yunlong Liang', 'Chulun Zhou'] | 2023-01-27 | null | null | null | null | ['nmt'] | ['computer-code'] | [ 2.23722249e-01 -9.99685898e-02 -9.45780426e-02 -4.39408422e-01
-1.05204618e+00 -4.29868579e-01 7.54319549e-01 -3.88677061e-01
-4.58717167e-01 7.91719794e-01 3.92104864e-01 -7.45688260e-01
4.29729760e-01 -3.04573685e-01 -3.38299870e-01 -5.03787875e-01
4.61365640e-01 8.01258266e-01 2.14376599e-01 -5.13868451e-01
1.84197314e-02 -8.96713361e-02 -8.32717359e-01 4.58498597e-01
1.46155322e+00 5.52741289e-01 4.73297507e-01 4.39566255e-01
-6.01203263e-01 6.04926407e-01 -6.51478767e-01 -4.21964735e-01
9.60868523e-02 -1.18356550e+00 -1.11099935e+00 2.40483224e-01
-2.16734916e-01 -2.27833554e-01 5.60350493e-02 9.57506478e-01
5.66489577e-01 1.65480793e-01 4.01264876e-01 -9.80319083e-01
-6.18838787e-01 9.45077181e-01 -4.95279402e-01 2.32539684e-01
2.98813522e-01 3.71315181e-01 8.68981421e-01 -1.01260686e+00
3.25325429e-01 1.50313485e+00 5.06442189e-01 8.99506032e-01
-1.14246631e+00 -6.24828339e-01 1.34890273e-01 2.49671772e-01
-1.13684380e+00 -7.16148674e-01 8.60357523e-01 -4.13334250e-01
8.02031636e-01 1.96526796e-01 4.81474727e-01 1.22275472e+00
-7.99772441e-02 7.13508606e-01 1.55329537e+00 -5.61242461e-01
-7.97921047e-02 4.23379302e-01 -6.60781264e-02 4.43990380e-01
-6.25238955e-01 -8.31914395e-02 -3.09115201e-01 -3.92359607e-02
6.56414628e-01 -4.81820494e-01 -2.07936555e-01 2.30384752e-01
-1.36798584e+00 8.01901817e-01 2.98458152e-02 5.64672291e-01
-1.52193055e-01 -6.06299162e-01 7.68834114e-01 7.61934280e-01
5.21869659e-01 1.90985784e-01 -4.36731905e-01 -3.38292569e-01
-6.71626508e-01 -2.27545738e-01 8.55465770e-01 1.14743686e+00
7.11087525e-01 -2.60452181e-02 -3.15261990e-01 1.41045046e+00
1.64406151e-01 2.67068297e-01 9.30556238e-01 -4.50271517e-01
1.14403784e+00 5.26095510e-01 -1.45751595e-01 -2.96129376e-01
-1.82296991e-01 -2.04504281e-01 -8.52694988e-01 -4.59631383e-01
4.21447545e-01 -5.32435060e-01 -4.33157593e-01 1.80469620e+00
3.61680597e-01 -5.32149002e-02 4.28792298e-01 1.01012254e+00
9.79961634e-01 9.41938162e-01 1.17472529e-01 -7.71857798e-01
1.41039443e+00 -1.49925327e+00 -9.77781653e-01 -8.20572972e-02
6.62773550e-01 -1.18166840e+00 1.51198602e+00 -2.28734706e-02
-1.15798199e+00 -7.78298020e-01 -7.71294117e-01 -1.11971259e-01
-1.00788958e-02 3.66727173e-01 3.13116252e-01 5.36176026e-01
-8.32007587e-01 7.12074414e-02 -5.53019285e-01 -3.36876422e-01
-1.92204803e-01 1.93701059e-01 -1.13803066e-01 1.35234833e-01
-1.56760228e+00 1.06609619e+00 2.87881494e-01 3.17237824e-01
-6.90042794e-01 -2.58491039e-01 -7.09450185e-01 -2.00023111e-02
3.68019819e-01 -4.50541854e-01 1.54298103e+00 -1.27698052e+00
-2.26562405e+00 7.01284349e-01 -2.30946496e-01 6.02762215e-03
8.82676840e-01 1.08516552e-02 -3.17145288e-01 -3.20597231e-01
1.04517303e-01 3.85932863e-01 4.98832256e-01 -1.09058785e+00
-7.50252187e-01 -1.99599639e-01 8.62795264e-02 7.50795066e-01
-4.04169798e-01 2.42369771e-01 -5.29094875e-01 -5.16175270e-01
6.28295448e-03 -9.91620123e-01 -1.09399848e-01 -7.63705850e-01
-3.99725378e-01 -4.37268376e-01 6.51342750e-01 -8.72595787e-01
1.21033633e+00 -1.90688038e+00 2.17674062e-01 -2.49480754e-01
-2.06100732e-01 3.74719054e-01 -3.37934166e-01 6.67904079e-01
2.08252236e-01 -1.81375995e-01 -2.42022097e-01 -5.52678108e-01
-9.11661983e-02 1.64790437e-01 -2.50956360e-02 2.73763910e-02
2.52564818e-01 7.71830440e-01 -8.97322774e-01 -8.31712961e-01
1.84426177e-02 8.19379985e-02 -3.66808206e-01 5.94376802e-01
-1.76065043e-01 1.23954058e+00 -4.84240234e-01 4.17396724e-01
5.64254105e-01 9.73988101e-02 4.91940022e-01 -6.37167767e-02
-4.72755879e-01 9.67962086e-01 -7.26645172e-01 1.74641049e+00
-8.77415955e-01 3.17239553e-01 2.82221049e-01 -9.13971364e-01
1.06124914e+00 7.49565303e-01 2.15489000e-01 -7.93410778e-01
2.81243443e-01 4.00500715e-01 4.77779478e-01 -7.39033937e-01
5.39714575e-01 -4.58520144e-01 -1.55325398e-01 6.99946404e-01
1.26550764e-01 -4.70017828e-02 1.15541764e-01 -2.11066619e-01
5.92926383e-01 1.54267997e-01 1.62508458e-01 -2.77060717e-01
1.07459962e+00 1.04544058e-01 9.28956151e-01 1.81510061e-01
-4.37154055e-01 2.21863478e-01 4.67884928e-01 -1.46515265e-01
-8.67168605e-01 -7.05471158e-01 2.74387207e-02 1.33819222e+00
8.06234851e-02 -2.44525045e-01 -9.64144766e-01 -9.55121338e-01
-5.88213027e-01 5.40441811e-01 -2.47845054e-01 -3.64286117e-02
-9.60553229e-01 -9.14634764e-01 6.73175156e-01 2.89228231e-01
8.69796336e-01 -1.23909211e+00 9.56442058e-02 2.59809494e-01
-9.11566317e-01 -1.21057892e+00 -9.64214265e-01 4.68013845e-02
-8.55138183e-01 -7.13448822e-01 -6.49162710e-01 -1.26654565e+00
5.90047717e-01 3.31836045e-01 8.65654707e-01 4.41967510e-02
5.59828460e-01 -2.93218583e-01 -4.55306500e-01 1.39306009e-01
-1.11018050e+00 4.44985956e-01 1.07147865e-01 -1.61436573e-03
3.40634972e-01 -5.58669388e-01 -2.49566883e-01 6.55357838e-01
-4.28921193e-01 5.62771857e-01 9.58613396e-01 1.28192413e+00
2.07799092e-01 -3.06812495e-01 9.48039949e-01 -1.03673208e+00
1.13612592e+00 -2.91731119e-01 -3.92960668e-01 5.06478310e-01
-4.37028706e-01 -7.40237907e-02 9.80810881e-01 -7.31331944e-01
-1.53960371e+00 -3.67119499e-02 -5.54491282e-01 -8.38249922e-02
-3.36303711e-02 6.18387461e-01 -4.82794493e-01 1.89862624e-01
4.31215048e-01 5.97990155e-01 6.60429373e-02 -4.61073846e-01
2.47475445e-01 1.26707315e+00 3.88355434e-01 -8.97464275e-01
6.71253681e-01 -3.55441123e-01 -6.81217730e-01 -5.88758945e-01
-5.16736090e-01 -2.94762343e-01 -9.07584131e-01 -1.91641554e-01
8.13935161e-01 -8.60012412e-01 -7.02333570e-01 3.52628171e-01
-1.27977586e+00 -6.78098559e-01 1.93321288e-01 7.47225821e-01
-4.48301315e-01 4.05698180e-01 -1.08650684e+00 -6.71004057e-01
-5.50258100e-01 -1.39667511e+00 6.02758169e-01 1.61593959e-01
-1.88801177e-02 -9.85497594e-01 1.89538583e-01 8.05216193e-01
4.60374117e-01 -3.17279130e-01 1.04527378e+00 -8.12879145e-01
-1.42115399e-01 2.92054623e-01 -1.20951407e-01 5.17599821e-01
4.51585472e-01 -1.49170116e-01 -8.40743005e-01 -3.47596705e-01
3.45633507e-01 -5.84898174e-01 2.00928926e-01 -1.89373642e-01
4.85354960e-01 -5.61537445e-01 -5.92891648e-02 2.59877324e-01
8.30482066e-01 3.89830053e-01 3.26766461e-01 2.10810751e-01
7.08500981e-01 8.83426666e-01 7.53122687e-01 -1.23665303e-01
8.01879227e-01 7.66470492e-01 -1.32469296e-01 -1.75331503e-01
-5.20073213e-02 -2.96377599e-01 6.93222284e-01 2.03129554e+00
-9.45290998e-02 1.79175720e-01 -8.38729858e-01 5.14731944e-01
-1.82345378e+00 -7.35647142e-01 -1.43537283e-01 2.17481947e+00
1.35853994e+00 2.69845963e-01 3.92300636e-01 -2.84478724e-01
9.93854940e-01 5.41931950e-02 -4.32165265e-01 -6.20864391e-01
2.01483071e-01 -3.22681278e-01 -1.22703150e-01 6.61014974e-01
-8.09233665e-01 1.25602663e+00 4.97080469e+00 8.63285422e-01
-1.38610947e+00 6.20092034e-01 5.45458972e-01 4.52425271e-01
-2.48437285e-01 1.93624511e-01 -7.43743002e-01 5.84209919e-01
9.46337163e-01 -1.12168364e-01 5.80638885e-01 5.02681315e-01
4.49689299e-01 2.97270685e-01 -1.03890431e+00 5.70024788e-01
-5.03148958e-02 -8.13703537e-01 9.35121402e-02 -1.83279827e-01
6.84474885e-01 2.73162080e-03 -3.75797242e-01 8.38185430e-01
3.24110448e-01 -5.60865462e-01 5.62231719e-01 -7.75771365e-02
8.69404018e-01 -6.29247665e-01 7.89222896e-01 8.90658200e-01
-1.19664884e+00 1.11282490e-01 -3.30529541e-01 3.17898430e-02
1.79992765e-01 1.51178837e-02 -9.67281580e-01 8.33230317e-01
2.63020158e-01 3.57595950e-01 -2.24889487e-01 4.58879858e-01
-3.55063558e-01 7.37547338e-01 2.52842139e-02 -1.49208039e-01
3.78844380e-01 -6.77627981e-01 4.33721274e-01 1.44489121e+00
4.78466861e-02 3.27909589e-02 4.60125476e-01 7.36087203e-01
-8.84733051e-02 5.49256980e-01 -2.42415190e-01 7.81092495e-02
7.55316079e-01 1.27596033e+00 -3.66833538e-01 -3.01476091e-01
-6.97210550e-01 1.03587914e+00 4.85913247e-01 3.14995438e-01
-7.08807349e-01 -2.48374835e-01 2.08454445e-01 -2.49320254e-01
-6.99587613e-02 -1.68269217e-01 -2.63864785e-01 -1.26361728e+00
1.33643616e-02 -1.29291487e+00 1.16576724e-01 -1.85523957e-01
-1.35127509e+00 1.13728631e+00 -1.64216742e-01 -1.44540918e+00
-5.06684482e-01 -4.02378589e-02 -1.00555062e+00 1.25902140e+00
-1.43229592e+00 -1.47068715e+00 -9.23829749e-02 6.53085172e-01
1.09414208e+00 -2.01563433e-01 8.86449635e-01 5.88423908e-01
-8.13938498e-01 9.23061192e-01 2.26851124e-02 3.13594192e-01
9.97082770e-01 -1.07810044e+00 3.66397530e-01 7.66862929e-01
-3.03436071e-01 7.33278692e-01 4.06012326e-01 -4.62713927e-01
-1.03478050e+00 -9.93998706e-01 1.35764527e+00 -4.09595966e-02
6.58504426e-01 -5.99418163e-01 -1.01473129e+00 7.88826823e-01
4.78409588e-01 -6.04278982e-01 7.68825889e-01 4.09717143e-01
-5.62214255e-02 -1.66949287e-01 -8.35488617e-01 7.35082686e-01
7.63364494e-01 -6.89990103e-01 -8.64937127e-01 2.95767903e-01
6.96459711e-01 -6.44646108e-01 -9.68537927e-01 3.24484855e-01
4.51605022e-01 -6.62364066e-01 4.39057678e-01 -3.49038035e-01
3.63396525e-01 -1.52263492e-01 1.88401565e-01 -1.59048939e+00
-9.54918861e-02 -9.68989491e-01 4.58801448e-01 1.71111882e+00
5.09400189e-01 -7.08276212e-01 2.36367807e-01 1.90001771e-01
-5.25302172e-01 -7.37389505e-01 -1.06068254e+00 -7.08560705e-01
3.55022371e-01 7.37964064e-02 6.12852573e-01 1.35448575e+00
5.70346713e-01 1.15057802e+00 -4.98161703e-01 -1.37920290e-01
8.85669589e-02 3.31818342e-01 9.51122642e-01 -8.67649257e-01
-3.71318072e-01 -4.66832578e-01 5.23725271e-01 -1.53008926e+00
1.66555285e-01 -7.68197119e-01 5.15389800e-01 -1.22810400e+00
3.06341439e-01 -7.23427534e-01 6.61894530e-02 3.73207182e-01
-5.64577758e-01 -8.59829858e-02 8.61739144e-02 6.26025677e-01
-4.38033789e-01 9.32846427e-01 1.70489645e+00 4.28894311e-02
-4.90532637e-01 4.07687515e-01 -5.01311004e-01 4.59140569e-01
8.53228688e-01 -3.95786554e-01 -5.45205116e-01 -5.12555957e-01
-4.33648467e-01 6.43812120e-01 -2.94061601e-01 -6.18767560e-01
1.65861830e-01 -2.44685590e-01 -1.78471804e-01 -3.98658633e-01
2.24142537e-01 -4.65717375e-01 -2.62685448e-01 4.95516688e-01
-5.35847247e-01 2.32180774e-01 6.58706948e-02 8.06033313e-02
-3.74393255e-01 -2.50395268e-01 7.25275636e-01 -1.60011604e-01
-2.31136873e-01 4.14028317e-02 -6.06959283e-01 1.27495259e-01
5.71180284e-01 4.72013094e-02 -3.91981095e-01 -3.35567445e-01
-6.17793083e-01 5.55938303e-01 2.20793560e-01 6.75807059e-01
5.94134443e-02 -1.36745369e+00 -9.63989913e-01 1.58013061e-01
-5.70469163e-03 -9.17268544e-02 2.99220622e-01 1.23958921e+00
-3.09552550e-01 4.23231751e-01 -3.16579938e-01 -6.70614541e-01
-1.30269003e+00 3.32425237e-01 3.09654862e-01 -6.50538564e-01
-3.98791134e-01 5.14418900e-01 2.69778579e-01 -1.10818839e+00
1.33294061e-01 -1.91627473e-01 -3.29206854e-01 -8.19819793e-02
2.11609751e-01 1.19926758e-01 1.50819309e-02 -9.68723714e-01
-2.33680859e-01 2.87048727e-01 -3.01036865e-01 -4.41731185e-01
8.27466726e-01 -6.22243106e-01 -2.17239246e-01 6.02793038e-01
9.38924730e-01 1.47679046e-01 -1.18902063e+00 -6.54830277e-01
-4.71251011e-02 -1.48548648e-01 -4.60671544e-01 -8.57346892e-01
-6.63978517e-01 1.02972281e+00 2.01571882e-01 1.68293238e-01
1.00915515e+00 -2.34458447e-01 1.18658793e+00 2.54047424e-01
1.71215981e-01 -1.20876694e+00 -2.69073807e-03 9.50263679e-01
8.46830070e-01 -1.28232276e+00 -4.77589190e-01 -4.70069826e-01
-1.00623763e+00 1.02864647e+00 1.01665437e+00 3.66727680e-01
2.34555244e-01 -7.66462237e-02 7.11874962e-01 2.43809998e-01
-8.99768591e-01 -1.13433503e-01 2.19933227e-01 3.01535279e-01
8.24205399e-01 2.33405456e-02 -9.14847136e-01 7.41110027e-01
-3.95406783e-01 -1.91152051e-01 2.43433207e-01 6.46830738e-01
-1.99527547e-01 -1.44966376e+00 -2.56523073e-01 -7.37474253e-03
-3.68107617e-01 -2.31017902e-01 -4.84083623e-01 6.80201173e-01
7.73618668e-02 1.30036259e+00 -1.51854917e-01 -6.58607662e-01
4.09399807e-01 1.81827813e-01 1.50529504e-01 -7.73756742e-01
-1.12738204e+00 4.43727374e-01 3.57744098e-01 -2.60747392e-02
-4.40446585e-01 -6.24775231e-01 -1.04868698e+00 -2.80084014e-01
-5.74416161e-01 8.03790629e-01 5.04310966e-01 1.34680915e+00
1.52281476e-02 5.64899504e-01 1.06766164e+00 -6.76372111e-01
-7.71730661e-01 -1.71599066e+00 -1.13899177e-02 2.16816336e-01
1.40185252e-01 -2.66940266e-01 -9.81618017e-02 7.15184957e-02] | [12.838099479675293, 8.272337913513184] |
83bba57c-3615-4646-8f06-58184dc91915 | inverse-reinforcement-learning-without | 2303.14623 | null | https://arxiv.org/abs/2303.14623v2 | https://arxiv.org/pdf/2303.14623v2.pdf | Inverse Reinforcement Learning without Reinforcement Learning | Inverse Reinforcement Learning (IRL) is a powerful set of techniques for imitation learning that aims to learn a reward function that rationalizes expert demonstrations. Unfortunately, traditional IRL methods suffer from a computational weakness: they require repeatedly solving a hard reinforcement learning (RL) problem as a subroutine. This is counter-intuitive from the viewpoint of reductions: we have reduced the easier problem of imitation learning to repeatedly solving the harder problem of RL. Another thread of work has proved that access to the side-information of the distribution of states where a strong policy spends time can dramatically reduce the sample and computational complexities of solving an RL problem. In this work, we demonstrate for the first time a more informed imitation learning reduction where we utilize the state distribution of the expert to alleviate the global exploration component of the RL subroutine, providing an exponential speedup in theory. In practice, we find that we are able to significantly speed up the prior art on continuous control tasks. | ['Zhiwei Steven Wu', 'J. Andrew Bagnell', 'Sanjiban Choudhury', 'Gokul Swamy'] | 2023-03-26 | null | null | null | null | ['continuous-control'] | ['playing-games'] | [ 1.02086999e-01 4.63238806e-01 -3.04387450e-01 2.42459789e-01
-8.85822296e-01 -7.50668764e-01 4.85063583e-01 -8.34891647e-02
-7.61113763e-01 1.05338097e+00 -1.36200055e-01 -7.38177419e-01
-2.64521062e-01 -5.08907735e-01 -9.72190619e-01 -8.01769793e-01
-1.87955007e-01 4.69956547e-01 -4.27659824e-02 -4.27597433e-01
6.36811376e-01 4.97145116e-01 -1.29121864e+00 -4.29125249e-01
6.57964289e-01 7.63009071e-01 2.78044939e-01 8.66087139e-01
2.69442677e-01 1.01498699e+00 -3.63904685e-01 3.69303524e-01
4.78131890e-01 -7.31566906e-01 -8.72718334e-01 4.15393338e-02
-4.32770289e-02 -4.95097816e-01 -4.06683236e-01 1.10692179e+00
4.03458565e-01 3.25130761e-01 3.97192657e-01 -1.44506800e+00
-6.74522072e-02 5.62164307e-01 -5.78022301e-01 -5.59382327e-02
2.25054920e-01 5.79748631e-01 9.22647297e-01 -1.98481351e-01
6.12167478e-01 1.17734468e+00 3.52711260e-01 6.62042201e-01
-1.19148695e+00 -5.16726375e-01 1.59797236e-01 2.36360818e-01
-8.55039775e-01 -8.98713768e-02 4.87054020e-01 -2.47283161e-01
8.21976006e-01 -2.13815738e-03 7.54513383e-01 8.10574770e-01
6.26230910e-02 9.17009175e-01 1.41947472e+00 -4.80858862e-01
5.63526571e-01 -6.35426790e-02 -2.23502472e-01 9.36381996e-01
7.48083070e-02 6.21888816e-01 -1.73942447e-01 -1.56033918e-01
9.37822282e-01 -2.49265693e-02 -3.22895408e-01 -8.21659148e-01
-1.04692161e+00 1.09307957e+00 2.90816635e-01 -1.21535305e-02
-2.68377900e-01 5.64544201e-01 4.17978168e-01 1.00757849e+00
-2.09051028e-01 7.72671282e-01 -6.40619338e-01 -5.20213664e-01
-3.48961264e-01 4.52339709e-01 1.22440195e+00 6.93234682e-01
9.88008916e-01 6.34409040e-02 2.70940065e-01 2.97200620e-01
5.15208244e-02 4.71368581e-01 3.43821317e-01 -1.59422410e+00
4.37400818e-01 2.89308161e-01 4.88315433e-01 -2.74810374e-01
-1.51115745e-01 -1.61147222e-01 -1.93537563e-01 1.04467094e+00
7.46957123e-01 -5.47255874e-01 -4.82618123e-01 1.94600475e+00
4.30921793e-01 5.88749461e-02 1.92197606e-01 7.89153874e-01
-4.18780655e-01 5.15427470e-01 -3.83265972e-01 -3.61116558e-01
8.08920085e-01 -1.07724881e+00 -1.49273500e-01 -2.12434962e-01
7.73676634e-01 -3.24652910e-01 1.24187350e+00 6.33071423e-01
-1.21734512e+00 -2.05556661e-01 -1.04555416e+00 2.37662077e-01
2.48362646e-02 -5.94777949e-02 6.69722021e-01 3.24736506e-01
-9.10434008e-01 1.15993762e+00 -8.82429004e-01 -4.61736210e-02
3.30828249e-01 5.88010967e-01 -2.47934297e-01 4.90711965e-02
-8.54686439e-01 1.12619460e+00 1.54343173e-01 -9.34535041e-02
-1.42748868e+00 -6.19883776e-01 -7.14851916e-01 1.21425726e-01
1.22948623e+00 -2.94616699e-01 1.82669938e+00 -1.03683162e+00
-2.04582739e+00 2.87083179e-01 1.32478401e-01 -4.73754168e-01
7.22672939e-01 -2.87481278e-01 6.57538354e-01 6.51080310e-02
3.72927785e-02 2.51584917e-01 1.08622909e+00 -8.71381640e-01
-5.67447901e-01 -4.75337684e-01 4.95518386e-01 3.56640726e-01
1.22201540e-01 -5.71526647e-01 2.00970359e-02 -1.66924030e-01
-4.53127652e-01 -1.38514614e+00 -5.88901401e-01 5.89707047e-02
7.90508464e-02 -4.89410937e-01 5.04342973e-01 -2.23246247e-01
5.95129728e-01 -2.08883190e+00 5.78557789e-01 1.56530961e-01
1.99039921e-01 3.13576669e-01 -2.58364767e-01 6.64556861e-01
2.48703593e-03 -1.34930223e-01 -7.69081563e-02 4.63128537e-02
4.30660665e-01 3.35481256e-01 -5.66918910e-01 7.15215266e-01
-2.01268524e-01 1.06707644e+00 -1.21803844e+00 -2.11211786e-01
1.95394084e-01 2.00808365e-02 -7.31479704e-01 4.11836237e-01
-6.25685990e-01 5.84730089e-01 -6.59144580e-01 -1.87137797e-02
1.24445409e-01 -2.67582595e-01 4.20838505e-01 5.00670016e-01
-2.60013849e-01 4.24881339e-01 -1.29221201e+00 1.67300510e+00
-7.37608135e-01 5.05196929e-01 4.20367658e-01 -1.19737589e+00
3.95371497e-01 1.80656686e-01 3.77234489e-01 -6.07048452e-01
8.28532577e-02 3.13525945e-01 1.43470287e-01 -3.35827887e-01
1.13119960e-01 -1.73995480e-01 -1.42253503e-01 8.90247285e-01
-1.03761427e-01 -4.66577411e-01 5.36650047e-02 1.16046006e-02
1.34417510e+00 6.52339160e-01 3.86993706e-01 -5.60561977e-02
4.56640720e-01 2.68783748e-01 3.31236511e-01 1.11054850e+00
-1.46393791e-01 -2.89074749e-01 1.06760311e+00 -2.68784940e-01
-1.10998082e+00 -9.75126922e-01 5.62331140e-01 1.06779671e+00
-1.95392817e-02 -2.08597884e-01 -7.26440430e-01 -7.66896904e-01
1.67025715e-01 5.35791695e-01 -7.21524119e-01 -1.09371498e-01
-7.69366980e-01 -7.12176263e-02 2.32108250e-01 4.71073389e-01
2.36421123e-01 -1.25765336e+00 -1.10228467e+00 2.56260365e-01
4.13815737e-01 -7.36855507e-01 -5.54773867e-01 4.41530466e-01
-9.74180102e-01 -1.29487014e+00 -5.85427761e-01 -5.89001000e-01
6.08879745e-01 1.07630320e-01 8.34680080e-01 1.58689722e-01
-4.72250581e-01 7.11980343e-01 5.35832047e-02 -3.40945810e-01
-4.96661007e-01 5.97773260e-03 -8.30515325e-02 -6.88576758e-01
-1.51734203e-01 -7.08337724e-01 -6.11536920e-01 8.40052962e-02
-5.30868351e-01 -6.23577721e-02 6.63757622e-01 1.07724917e+00
3.53960127e-01 6.10632915e-03 4.34760749e-01 -7.23356426e-01
8.58284235e-01 -2.29592279e-01 -1.32814026e+00 1.08444743e-01
-6.78524077e-01 7.29712903e-01 9.74549830e-01 -6.70041144e-01
-7.58098066e-01 2.49653429e-01 2.20493555e-01 -4.28286463e-01
2.45824814e-01 2.96196193e-01 2.82840312e-01 -2.37374485e-01
4.50783879e-01 4.42955941e-01 7.31412768e-01 -2.27165774e-01
4.07923162e-01 2.60349929e-01 3.37624520e-01 -9.91025209e-01
7.40959942e-01 2.45047554e-01 3.95507276e-01 -5.88886976e-01
-7.71089673e-01 -2.92135745e-01 -1.37460381e-01 -5.36491536e-02
4.23236847e-01 -7.25562334e-01 -1.69806266e+00 5.38699999e-02
-6.50515497e-01 -1.08809876e+00 -6.34189725e-01 3.77035558e-01
-1.29614675e+00 4.78657633e-01 -4.58075613e-01 -1.01095903e+00
1.17434701e-02 -1.27824092e+00 7.39856541e-01 9.98974070e-02
3.32283298e-03 -7.33857691e-01 3.59053463e-01 -1.95876770e-02
3.21371347e-01 1.07987747e-01 7.48655677e-01 -3.44546020e-01
-7.54663765e-01 3.47849056e-02 6.03824928e-02 1.77025780e-01
-1.38165146e-01 -5.30445933e-01 -4.63331431e-01 -6.09645128e-01
2.97549605e-01 -7.36850023e-01 5.60171604e-01 1.71641216e-01
1.11237478e+00 -4.49810177e-01 1.30944448e-02 2.47713402e-01
1.48105311e+00 1.35143012e-01 4.41119999e-01 4.71302837e-01
3.13420296e-01 4.67868835e-01 8.83786798e-01 6.18893266e-01
1.44628167e-01 5.77104628e-01 4.02298838e-01 2.12657467e-01
2.53795803e-01 -5.42576551e-01 6.27722859e-01 1.82846114e-01
-2.08888546e-01 4.37050343e-01 -4.87470508e-01 3.80035162e-01
-2.25364089e+00 -9.55722392e-01 5.23586333e-01 2.31810880e+00
1.02026427e+00 3.83507200e-02 5.02620399e-01 -1.00505471e-01
3.46869826e-01 -1.76682651e-01 -1.07380617e+00 -4.76994574e-01
7.13985384e-01 4.37700391e-01 7.62898088e-01 8.85680676e-01
-7.35872149e-01 9.20305192e-01 6.23359966e+00 7.10332036e-01
-9.14871633e-01 -1.54261991e-01 6.17444664e-02 -1.78041130e-01
7.18412474e-02 3.66160721e-01 -5.89931965e-01 2.88395643e-01
8.98945570e-01 -2.89122671e-01 1.35528576e+00 1.21740091e+00
2.07947597e-01 -4.75505233e-01 -1.40230668e+00 8.48169267e-01
-3.91805232e-01 -1.05881417e+00 -5.85638404e-01 4.40092683e-01
6.96384490e-01 7.04275444e-02 1.09598488e-01 6.66283607e-01
7.02892840e-01 -1.00309002e+00 3.67995828e-01 1.25005260e-01
4.56962496e-01 -1.01965344e+00 3.18971992e-01 7.54433990e-01
-8.36331010e-01 -4.87983793e-01 -3.56101722e-01 -5.55605829e-01
-1.44423530e-01 -1.73691854e-01 -8.55016410e-01 2.66127102e-03
2.23695576e-01 3.20839792e-01 5.32199517e-02 6.41464055e-01
-5.31552434e-01 3.69852275e-01 -4.74170148e-01 -4.40052122e-01
7.22579479e-01 -5.11492372e-01 4.25648898e-01 6.24010146e-01
1.67602226e-01 -2.37309746e-02 3.03349793e-01 9.62209821e-01
7.29500353e-02 -2.19694987e-01 -7.71993220e-01 -3.94737035e-01
1.15062080e-01 1.05783939e+00 -4.61990505e-01 -2.17484921e-01
-2.05331609e-01 8.41088772e-01 7.44609773e-01 2.30020329e-01
-7.77053297e-01 -3.55821878e-01 5.69637716e-01 -2.50574619e-01
6.84093654e-01 -5.28633595e-01 1.32085651e-01 -9.89332676e-01
-1.58600025e-02 -1.18084586e+00 1.09242693e-01 -3.95563751e-01
-7.07033277e-01 -6.94552734e-02 -1.35800585e-01 -8.67794633e-01
-8.60954762e-01 -6.77796721e-01 -4.12682831e-01 7.07944274e-01
-1.62401223e+00 -3.25036645e-01 1.34920314e-01 5.77823520e-01
6.43120825e-01 -6.42760396e-02 6.71530902e-01 -2.57098496e-01
-2.55105048e-01 2.89327979e-01 2.13687658e-01 -2.04547435e-01
2.75349021e-01 -1.61904287e+00 1.11545190e-01 4.98405665e-01
-1.34777352e-01 4.27155674e-01 8.49890053e-01 -2.79105276e-01
-2.11879969e+00 -3.83226663e-01 2.82440394e-01 -1.85132325e-01
1.08064425e+00 -4.13221717e-02 -4.44059849e-01 7.40482926e-01
-7.45420251e-03 -1.84167355e-01 -1.83602311e-02 6.58072829e-02
-1.37166798e-01 1.09431013e-01 -7.62563050e-01 8.38561773e-01
7.68210649e-01 -6.63832903e-01 -6.56834483e-01 3.03639978e-01
5.00407338e-01 -4.52123225e-01 -7.49644458e-01 -1.00432999e-01
5.40615797e-01 -6.71434522e-01 9.49426770e-01 -5.98231673e-01
3.78930390e-01 -3.28702718e-01 2.09708661e-01 -1.40764856e+00
2.73612112e-01 -1.31349146e+00 -4.16494608e-01 3.84864509e-01
1.65720567e-01 -8.36349189e-01 7.65283465e-01 4.30353910e-01
2.63427645e-01 -9.62119341e-01 -8.24091613e-01 -1.01942551e+00
2.93649435e-01 -3.30355130e-02 2.08081380e-01 4.01587576e-01
3.75906110e-01 3.77213418e-01 -3.66289169e-01 -8.08242857e-02
7.29075551e-01 3.31045866e-01 1.04362833e+00 -8.43410552e-01
-1.07239366e+00 -4.50193703e-01 1.48048028e-01 -1.37549496e+00
4.20017391e-01 -5.91423929e-01 3.89089286e-01 -1.09334731e+00
6.09825887e-02 -4.89724070e-01 -1.51349381e-01 4.51566905e-01
1.68991107e-02 -3.92084241e-01 4.57914978e-01 3.13853808e-02
-7.94841826e-01 4.43313509e-01 1.62673235e+00 1.36449277e-01
-3.50429684e-01 1.88810617e-01 -7.06595123e-01 7.21289158e-01
1.03164661e+00 -5.92128217e-01 -7.04782784e-01 -6.36406988e-02
4.10800844e-01 5.63323677e-01 4.52953100e-01 -6.76305175e-01
2.30680317e-01 -3.52395594e-01 -1.16540246e-01 -1.45491332e-01
2.17308491e-01 -6.49143457e-01 -2.59423167e-01 1.14193869e+00
-6.40470862e-01 7.68219158e-02 -3.94979008e-02 7.41270423e-01
2.39217401e-01 -5.87895453e-01 7.64515221e-01 -3.64363581e-01
-4.61636215e-01 2.47338593e-01 -5.57531714e-01 3.90381485e-01
1.01211333e+00 1.18481100e-01 -2.28356533e-02 -6.02551639e-01
-6.20870948e-01 3.82355839e-01 6.63220286e-01 4.53667641e-02
3.59431118e-01 -7.52913356e-01 -2.46957630e-01 -4.51242030e-02
-4.21893299e-01 -2.47643963e-01 -1.46180868e-01 1.01348913e+00
-3.15478504e-01 3.31431061e-01 -3.33609402e-01 -2.06213653e-01
-1.07199430e+00 7.63223231e-01 4.95991230e-01 -6.03399456e-01
-9.70652580e-01 4.34461087e-01 4.18635868e-02 -3.38059098e-01
4.39681768e-01 -1.97052896e-01 2.51760095e-01 -3.64192426e-01
4.83801186e-01 6.63692653e-01 -4.52316135e-01 2.94420511e-01
-2.74501182e-02 4.12761569e-01 -1.08186506e-01 -5.36848545e-01
1.52285886e+00 -1.74869802e-02 -8.34800769e-03 3.21768224e-01
1.11141658e+00 -3.45822096e-01 -1.87843537e+00 -8.77413824e-02
1.16275132e-01 -2.33369678e-01 -5.64056337e-02 -6.70525849e-01
-5.92636287e-01 8.41054022e-01 3.98095161e-01 1.05526514e-01
8.61895680e-01 -1.13850966e-01 6.93095684e-01 1.18541431e+00
7.13465333e-01 -1.53302217e+00 2.84655720e-01 6.28021657e-01
7.44405627e-01 -1.18470240e+00 9.17652473e-02 2.13658154e-01
-6.74710155e-01 1.24630642e+00 3.88419569e-01 -7.10733712e-01
1.74554586e-01 3.09867471e-01 -3.53102624e-01 -1.03377447e-01
-8.00519645e-01 -2.45609730e-01 -3.62716287e-01 3.50784451e-01
-1.12261869e-01 -2.37885844e-02 -2.54733682e-01 -2.95989010e-02
-3.40173692e-02 2.30259493e-01 5.71262181e-01 1.16523647e+00
-7.32160926e-01 -1.40387428e+00 -2.02641353e-01 1.54077828e-01
-3.20009112e-01 2.27941364e-01 -1.45258352e-01 1.02793598e+00
-4.05300677e-01 6.47362709e-01 -3.14700097e-01 2.13561624e-01
7.07381666e-02 -1.49546355e-01 1.18119824e+00 -5.47764719e-01
-3.74933958e-01 -6.57400712e-02 -1.76901922e-01 -1.08251226e+00
-3.52519713e-02 -5.18330812e-01 -1.52548981e+00 -2.30871245e-01
-5.90112992e-02 2.75670856e-01 7.00071990e-01 1.12769318e+00
1.70808688e-01 2.38130018e-01 8.46987844e-01 -9.96400714e-01
-1.59329593e+00 -4.64823753e-01 -5.18127143e-01 3.57034691e-02
7.23228633e-01 -7.42935777e-01 -5.69099367e-01 -4.81898665e-01] | [4.156332015991211, 1.8093974590301514] |
05329ac9-1d0c-427f-9ba5-1b08b13689b8 | efficient-and-scalable-high-order-portfolios | 2206.02412 | null | https://arxiv.org/abs/2206.02412v1 | https://arxiv.org/pdf/2206.02412v1.pdf | Efficient and Scalable High-Order Portfolios Design via Parametric Skew-t Distribution | Since Markowitz's mean-variance framework, optimizing a portfolio that maximizes the profit and minimizes the risk has been ubiquitous in the financial industry. Initially, profit and risk were measured by the first two moments of the portfolio's return, a.k.a. the mean and variance, which are sufficient to characterize a Gaussian distribution. However, it is broadly believed that the first two moments are not enough to capture the characteristics of the returns' behavior, which have been recognized to be asymmetric and heavy-tailed. Although there is ample evidence that portfolio designs involving the third and fourth moments, i.e., skewness and kurtosis, will outperform the conventional mean-variance framework, they are non-trivial. Specifically, in the classical framework, the memory and computational cost of computing the skewness and kurtosis grow sharply with the number of assets. To alleviate the difficulty in high-dimensional problems, we consider an alternative expression for high-order moments based on parametric representations via a generalized hyperbolic skew-t distribution. Then, we reformulate the high-order portfolio optimization problem as a fixed-point problem and propose a robust fixed-point acceleration algorithm that solves the problem in an efficient and scalable manner. Empirical experiments also demonstrate that our proposed high-order portfolio optimization framework is of low complexity and significantly outperforms the state-of-the-art methods by 2 to 4 orders of magnitude. | ['Daniel P. Palomar', 'Jiaxi Ying', 'Rui Zhou', 'Xiwen Wang'] | 2022-06-06 | null | null | null | null | ['portfolio-optimization'] | ['time-series'] | [-4.53856915e-01 -2.92401642e-01 -1.60840124e-01 -1.48630828e-01
-5.93275130e-01 -9.06465352e-01 2.12672696e-01 4.77011576e-02
-2.60051310e-01 7.79752016e-01 -2.25117698e-01 -6.54223919e-01
-7.28649616e-01 -8.49657595e-01 -4.69729930e-01 -7.74304867e-01
-2.04936951e-01 5.02774954e-01 -1.12997822e-01 1.72305107e-02
7.33763397e-01 6.00099683e-01 -1.40692699e+00 -2.67676324e-01
8.92654121e-01 1.69565904e+00 -1.61613256e-01 3.39215755e-01
-1.44799381e-01 2.58338779e-01 -4.38591391e-01 -7.32379317e-01
7.03939557e-01 -2.18206540e-01 -1.66570976e-01 7.90839344e-02
-2.58878946e-01 -4.98247743e-01 1.21855333e-01 1.20084703e+00
2.85359204e-01 1.09983675e-01 7.77853549e-01 -1.30285561e+00
-5.87063432e-01 3.13780308e-01 -9.09356534e-01 2.97614068e-01
-1.88169386e-02 -1.43213093e-01 1.22233701e+00 -8.36839795e-01
-4.27753441e-02 9.26508904e-01 6.78309679e-01 -6.52835593e-02
-9.30305123e-01 -3.83472443e-01 -1.59090292e-02 -1.73513457e-01
-1.18520510e+00 7.62579590e-02 7.14278221e-01 -5.14607012e-01
6.01549506e-01 4.51508284e-01 6.89220250e-01 4.39144582e-01
4.74219829e-01 6.05829120e-01 1.03527141e+00 -1.26895383e-01
3.55461657e-01 3.00349593e-01 -4.92776558e-02 1.04565032e-01
9.06089723e-01 1.26669943e-01 -2.25092098e-01 -6.50062680e-01
8.09337676e-01 2.88862288e-01 -1.58450186e-01 -4.35543031e-01
-8.57098281e-01 1.18288124e+00 -2.89234996e-01 6.65766671e-02
-7.77573586e-01 1.28581068e-02 1.80115372e-01 2.48252928e-01
4.84872162e-01 3.97501349e-01 -4.46448267e-01 -4.22274381e-01
-1.02322519e+00 5.84844708e-01 1.15726268e+00 8.86453688e-01
4.44240421e-01 5.30255772e-03 -1.15981214e-01 3.35239798e-01
4.74041075e-01 7.25525260e-01 2.87969321e-01 -9.62259173e-01
5.21659136e-01 5.48099399e-01 5.37675440e-01 -1.02909780e+00
-2.80114770e-01 -8.19988906e-01 -7.08587229e-01 1.83783650e-01
8.27052951e-01 -2.47876480e-01 -3.42405066e-02 1.43508041e+00
2.45254815e-01 -8.39567855e-02 5.88793382e-02 7.99234927e-01
-1.74144968e-01 5.13093531e-01 -5.93401134e-01 -5.99427640e-01
1.22155869e+00 -6.61405087e-01 -5.98225474e-01 5.74786887e-02
2.15682283e-01 -9.74659145e-01 9.61605430e-01 5.11889458e-01
-1.29393327e+00 1.72243327e-01 -9.39377487e-01 6.18306339e-01
3.97284403e-02 -1.44732237e-01 7.53532290e-01 1.08177257e+00
-7.23244607e-01 7.53488421e-01 -5.01545787e-01 4.02459443e-01
2.30663732e-01 1.39923587e-01 2.07890213e-01 5.00511110e-01
-8.55478942e-01 4.86841947e-01 2.76444227e-01 6.92776665e-02
-3.66453022e-01 -9.68650460e-01 -2.16545194e-01 4.75088030e-01
5.07118523e-01 -4.98820513e-01 1.20330751e+00 -6.92370117e-01
-1.59927952e+00 3.53909284e-01 1.69483691e-01 -4.65977937e-01
8.17605317e-01 -2.15448260e-01 -5.90639599e-02 9.04270336e-02
-7.40581825e-02 -5.04469514e-01 7.89562821e-01 -7.58364439e-01
-6.95195854e-01 -3.34157199e-01 -4.69936728e-02 -4.97955717e-02
-5.28168023e-01 1.22638434e-01 1.17575236e-01 -8.82526934e-01
2.42099524e-01 -8.15420926e-01 -2.07782641e-01 -2.48695344e-01
-2.61066496e-01 1.38681471e-01 2.53484368e-01 -5.74966073e-01
1.50343359e+00 -1.98443794e+00 -5.14885843e-01 6.34529293e-01
-7.11292773e-02 -8.81768689e-02 3.34078878e-01 5.60420632e-01
-9.72076058e-02 1.17918596e-01 -1.96371302e-01 5.54669313e-02
4.18332934e-01 -1.07040003e-01 -7.01459765e-01 7.03571916e-01
-5.85581250e-02 6.60210013e-01 -6.77793443e-01 -3.04869022e-02
-2.18863547e-01 9.70270783e-02 -7.19684660e-01 1.74448967e-01
1.90690607e-01 3.84230390e-02 -5.44649839e-01 6.37894034e-01
8.49899948e-01 -5.19602001e-01 1.38920574e-02 1.47779435e-01
-3.39840382e-01 7.92201906e-02 -1.58843040e+00 6.91743910e-01
-5.23930080e-02 2.15642288e-01 -1.51124567e-01 -1.10814130e+00
1.04105377e+00 2.33479440e-01 7.82964289e-01 -3.80498916e-01
6.39883950e-02 6.36575341e-01 -1.29108801e-01 -8.81847888e-02
6.63574398e-01 -6.19884849e-01 -1.48032114e-01 7.90338159e-01
-5.69613814e-01 1.05595112e-01 3.26044947e-01 -2.69833475e-01
8.02865386e-01 -2.74271369e-01 5.16369462e-01 -4.76956785e-01
3.88415933e-01 -3.71974349e-01 6.75224960e-01 5.48544765e-01
2.43363157e-02 4.82083052e-01 1.11247802e+00 -8.14978704e-02
-1.01742578e+00 -1.26581216e+00 -2.11722672e-01 6.85631812e-01
-1.76500767e-01 8.68376810e-03 -6.78746521e-01 -3.24421704e-01
5.71452677e-01 8.77871156e-01 -3.30886424e-01 1.65145949e-01
-1.64940044e-01 -1.06173623e+00 1.33189455e-01 3.87056500e-01
3.21730673e-01 -4.40637022e-01 -9.65783060e-01 3.92114818e-01
3.94061536e-01 -7.29130208e-01 -7.00130403e-01 1.51273170e-02
-1.01732898e+00 -1.13498151e+00 -1.16148889e+00 1.07953884e-01
5.33028722e-01 1.84922218e-01 9.92559552e-01 -3.04612130e-01
-7.93926120e-02 5.40552855e-01 -2.26822793e-01 -6.84277654e-01
1.99466974e-01 -3.96295339e-01 9.46029276e-02 3.27847540e-01
1.38785690e-01 -5.43209553e-01 -7.71723747e-01 3.69072974e-01
-9.45659876e-01 -6.34189606e-01 6.94048047e-01 8.13707173e-01
7.12740541e-01 3.74294192e-01 8.63174140e-01 -5.68855584e-01
1.04944980e+00 -5.88917196e-01 -1.15227878e+00 4.39023077e-01
-9.39006925e-01 9.35139209e-02 5.75741410e-01 -5.05829811e-01
-9.25770760e-01 -4.75279957e-01 5.23984253e-01 -4.26673621e-01
6.08154058e-01 7.16776550e-01 6.54972941e-02 -4.21046801e-02
-4.99073789e-02 3.13462973e-01 5.35482876e-02 -4.12664324e-01
7.96476901e-02 5.50919533e-01 2.86702514e-01 -8.30462992e-01
7.38674819e-01 3.65613282e-01 4.93206650e-01 -5.01531720e-01
-9.65848982e-01 -5.71840107e-01 -1.48025319e-01 -1.27779394e-02
4.15492982e-01 -3.68081093e-01 -1.16685629e+00 4.26195681e-01
-9.50103521e-01 2.55157739e-01 -5.35270691e-01 7.09259689e-01
-9.48140442e-01 4.48692977e-01 -2.92728186e-01 -1.41273332e+00
-4.96798843e-01 -9.40387905e-01 6.48345470e-01 2.75061190e-01
9.25412849e-02 -1.22645581e+00 1.16348930e-01 -1.28954668e-02
5.74577808e-01 3.90813440e-01 9.03613806e-01 -9.90052760e-01
-5.21307588e-01 -5.13373435e-01 -2.23829389e-01 4.99245614e-01
2.05721129e-02 -1.15037672e-01 -3.41178775e-01 -3.22139919e-01
6.80411339e-01 2.58587748e-01 1.70789793e-01 5.72864473e-01
1.01837945e+00 -6.79307222e-01 1.68397814e-01 7.36452699e-01
1.43361270e+00 3.38672757e-01 3.59744340e-01 7.26859152e-01
-9.39255804e-02 7.18378603e-01 8.81203949e-01 1.39178872e+00
1.27489269e-01 2.43114173e-01 3.62183094e-01 4.45590407e-01
9.92808700e-01 -3.44367959e-02 3.26709569e-01 7.52794325e-01
-1.16256297e-01 -1.68929994e-02 -8.48871768e-01 3.27152789e-01
-1.87224042e+00 -1.17230523e+00 1.22633250e-02 2.70603991e+00
6.19116068e-01 2.70882547e-01 3.99322897e-01 8.13077167e-02
6.18351996e-01 -3.55516337e-02 -6.90615833e-01 -3.62029314e-01
-2.30721116e-01 -1.02244485e-02 9.94266391e-01 9.07447562e-02
-8.52034330e-01 1.05625637e-01 6.81059504e+00 8.32154274e-01
-9.34477568e-01 -2.81857222e-01 8.19165409e-01 -2.29772821e-01
-6.25140369e-01 1.04730971e-01 -1.03868210e+00 8.48720074e-01
8.75718296e-01 -1.03927267e+00 4.14503574e-01 1.21318436e+00
2.26610214e-01 -3.32096107e-02 -9.43420410e-01 9.90443468e-01
-2.06319734e-01 -1.08795905e+00 -2.16629565e-01 6.09502614e-01
8.51986408e-01 -5.63928485e-01 5.78216553e-01 1.10886462e-01
-2.05690593e-01 -8.47195685e-01 8.06587279e-01 9.68881369e-01
2.77527988e-01 -1.19030368e+00 8.98244023e-01 4.80096310e-01
-1.04160714e+00 -3.02695036e-01 -6.48698390e-01 -4.92735505e-02
4.31760281e-01 1.17874670e+00 -3.76643211e-01 4.98802036e-01
5.29955506e-01 2.00130790e-01 -1.43697292e-01 1.41431952e+00
1.54888049e-01 3.34335059e-01 -6.05130374e-01 -2.28139505e-01
3.93392205e-01 -9.48941708e-01 6.38006866e-01 7.08178937e-01
1.08093333e+00 1.90824255e-01 -1.51221901e-01 9.15751696e-01
2.01283574e-01 2.60084629e-01 -2.67595828e-01 -4.23640162e-01
3.38095516e-01 1.02093840e+00 -6.52690411e-01 -1.57461792e-01
-5.61110139e-01 4.24587399e-01 -2.76539952e-01 3.36927265e-01
-8.56632113e-01 -6.20357692e-01 7.18153358e-01 2.65882075e-01
4.78162080e-01 -3.12642038e-01 -6.38849258e-01 -9.84724104e-01
5.19068241e-01 -7.86787570e-01 3.24533612e-01 -6.50519654e-02
-1.49343240e+00 2.04594955e-01 1.51059926e-01 -1.37076867e+00
-6.02281868e-01 -8.17860842e-01 -6.25413179e-01 9.46284711e-01
-1.55974150e+00 -4.03818607e-01 2.75029540e-01 4.13365155e-01
-9.41690132e-02 -5.46736777e-01 4.21676964e-01 2.81235188e-01
-4.45392102e-01 7.14290440e-01 8.73475552e-01 -3.31517547e-01
3.11669976e-01 -1.38958001e+00 2.50099171e-02 5.97450733e-01
-2.43775710e-01 7.65996516e-01 8.63339126e-01 -6.40766144e-01
-1.47859311e+00 -5.67529321e-01 6.36481762e-01 2.23285612e-02
1.25527072e+00 1.80051699e-01 -7.99690962e-01 4.35235173e-01
-3.09903145e-01 -8.03768635e-02 8.65647852e-01 -1.59428433e-01
7.56731778e-02 -3.03585798e-01 -1.07560444e+00 6.01595402e-01
4.45869982e-01 -1.38794169e-01 -5.39274096e-01 3.50219756e-01
2.12378889e-01 3.14207338e-02 -1.06857705e+00 3.06219339e-01
6.81174278e-01 -1.29975736e+00 9.53328133e-01 -3.10758978e-01
-5.05327173e-02 2.53998227e-02 -4.08504128e-01 -9.34085906e-01
-1.53218582e-01 -1.16048431e+00 -2.38613158e-01 1.16269505e+00
2.97954351e-01 -1.28481662e+00 9.18738186e-01 7.80279517e-01
2.09234446e-01 -1.22953260e+00 -1.13460779e+00 -1.34702444e+00
1.70232758e-01 -3.58547807e-01 8.47594440e-01 4.97879893e-01
-7.89435804e-02 -3.18763673e-01 -3.34188074e-01 2.15728655e-02
1.03576565e+00 5.70646644e-01 5.90607524e-01 -1.34530163e+00
-6.31357014e-01 -8.07355464e-01 -2.92051166e-01 -1.13058996e+00
5.94334826e-02 -5.68731904e-01 -3.67081165e-01 -9.93788362e-01
1.75436527e-01 -3.91165525e-01 -4.35191661e-01 -1.95066869e-01
-1.78305786e-02 -3.87096226e-01 9.86604914e-02 3.49568039e-01
-2.12705240e-01 5.94772816e-01 9.88613069e-01 3.02585304e-01
-2.15169191e-01 5.60282111e-01 -8.13577771e-01 9.29949164e-01
7.95793891e-01 -4.69533503e-01 -4.51692313e-01 1.91583917e-01
6.31370246e-01 3.41135979e-01 2.50838511e-02 -5.03571391e-01
2.61710078e-01 -6.12763643e-01 9.82835367e-02 -9.33815837e-01
2.97261387e-01 -7.16472507e-01 2.43724719e-01 3.88884515e-01
1.06552675e-01 4.89492863e-01 -2.15148062e-01 5.76602578e-01
-3.18604916e-01 -6.91148162e-01 5.90969920e-01 2.75115725e-02
9.46420059e-02 3.56379360e-01 -1.36773154e-01 2.13309422e-01
1.29835105e+00 -2.07056478e-01 -2.56374449e-01 -5.53265214e-01
-2.12611958e-01 1.04361899e-01 4.57256883e-01 -7.08517209e-02
7.12028444e-01 -1.38399386e+00 -6.07288420e-01 8.32322836e-02
-4.09014404e-01 -2.96096742e-01 7.53689781e-02 1.21150517e+00
-5.61668873e-01 6.80249453e-01 -5.96825546e-03 -2.68504351e-01
-7.24322081e-01 6.08226359e-01 2.26625085e-01 -6.67142808e-01
-2.50550687e-01 4.93945926e-01 2.87358105e-01 2.93361861e-02
7.28291646e-02 -2.81449229e-01 1.26305252e-01 3.51658553e-01
6.50437593e-01 8.85133326e-01 -2.19620213e-01 -2.58853137e-01
-2.12656528e-01 7.20498025e-01 1.12506717e-01 -3.36217582e-01
1.50529337e+00 -1.47376042e-02 -2.60642290e-01 4.96212572e-01
1.07844210e+00 1.24642760e-01 -1.21392691e+00 6.42381608e-02
5.94018519e-01 -8.16071749e-01 -1.77539259e-01 -1.79400563e-01
-1.01230609e+00 7.63168275e-01 2.15700299e-01 6.10088825e-01
1.08082616e+00 -4.22441483e-01 8.15179706e-01 3.41981173e-01
4.26698208e-01 -1.30532336e+00 2.28258707e-02 5.17968595e-01
9.19945240e-01 -8.13701868e-01 2.75279820e-01 -2.04280764e-01
-7.80644655e-01 1.25463283e+00 -4.91869450e-02 -2.86540300e-01
9.83483255e-01 3.14976305e-01 -3.21062833e-01 1.74071223e-01
-5.96660554e-01 1.87123597e-01 5.11932671e-01 1.60244122e-01
3.14773947e-01 1.26218557e-01 -8.01640749e-01 1.25304210e+00
-5.08310258e-01 -2.66988337e-01 5.14627516e-01 8.91973615e-01
-6.74248993e-01 -1.07687914e+00 -6.97380841e-01 6.15778089e-01
-1.03769052e+00 2.72340383e-02 -1.62035152e-02 7.15002120e-01
-5.86710691e-01 8.27607095e-01 1.48199514e-01 1.10788062e-01
4.64761436e-01 -1.15131050e-01 1.87998250e-01 -2.44127646e-01
-3.62518162e-01 3.30320418e-01 -2.93657988e-01 -4.21027124e-01
1.01933457e-01 -9.46106493e-01 -9.41096723e-01 -5.97992480e-01
-3.02377701e-01 3.06161135e-01 7.59796441e-01 7.44004548e-01
7.02141002e-02 1.21745713e-01 1.00817955e+00 -5.98706484e-01
-1.89261734e+00 -5.44232428e-01 -1.29541779e+00 1.02586865e-01
2.54704151e-02 -7.82755136e-01 -8.19022119e-01 -4.47826058e-01] | [5.0334014892578125, 3.9650909900665283] |
2664add4-4530-44fa-ae88-09541f17e82e | activation-regression-for-continuous-domain | 2204.07030 | null | https://arxiv.org/abs/2204.07030v1 | https://arxiv.org/pdf/2204.07030v1.pdf | Activation Regression for Continuous Domain Generalization with Applications to Crop Classification | Geographic variance in satellite imagery impacts the ability of machine learning models to generalise to new regions. In this paper, we model geographic generalisation in medium resolution Landsat-8 satellite imagery as a continuous domain adaptation problem, demonstrating how models generalise better with appropriate domain knowledge. We develop a dataset spatially distributed across the entire continental United States, providing macroscopic insight into the effects of geography on crop classification in multi-spectral and temporally distributed satellite imagery. Our method demonstrates improved generalisability from 1) passing geographically correlated climate variables along with the satellite data to a Transformer model and 2) regressing on the model features to reconstruct these domain variables. Combined, we provide a novel perspective on geographic generalisation in satellite imagery and a simple-yet-effective approach to leverage domain knowledge. Code is available at: \url{https://github.com/samar-khanna/cropmap} | ['Bharath Hariharan', 'Kavita Bala', 'Bram Wallace', 'Samar Khanna'] | 2022-04-14 | null | null | null | null | ['crop-classification'] | ['miscellaneous'] | [ 1.62088171e-01 -2.05044121e-01 -2.79214889e-01 -4.93083388e-01
-7.26614654e-01 -1.03258014e+00 6.03302896e-01 -6.67079836e-02
-2.73404807e-01 9.15440202e-01 2.44950011e-01 -7.74946153e-01
-2.35138685e-01 -1.09827268e+00 -8.12907338e-01 -7.23893702e-01
-6.36990070e-01 6.74157068e-02 -1.42666057e-01 -3.79036367e-01
-1.16861232e-01 5.93739152e-01 -1.07497323e+00 -1.64606303e-01
1.16926038e+00 4.40499872e-01 6.98331058e-01 7.15853810e-01
2.42091641e-01 1.25311106e-01 -1.20232537e-01 2.47280806e-01
6.03385866e-01 -3.93896669e-01 -8.16665173e-01 1.62266299e-01
4.07546431e-01 -4.13829356e-01 -2.87216783e-01 1.08231008e+00
1.32334739e-01 1.41960710e-01 8.23357522e-01 -1.06459463e+00
-9.74093080e-01 7.58036137e-01 -8.18766534e-01 4.63893980e-01
-1.46759123e-01 2.78275371e-01 8.92204046e-01 -2.77543515e-01
6.35039449e-01 9.22263086e-01 1.12518334e+00 -2.08528429e-01
-1.64631093e+00 -7.89150298e-01 3.72216344e-01 -1.22637644e-01
-1.81920457e+00 -1.68753773e-01 2.42029071e-01 -7.77324080e-01
6.50266409e-01 4.26158249e-01 5.44341147e-01 6.08245492e-01
2.97841609e-01 2.59683102e-01 1.41249943e+00 -2.85801172e-01
1.81887597e-02 1.41787287e-02 -7.94345215e-02 8.13512281e-02
2.87596464e-01 4.90227818e-01 4.83162142e-02 -3.19600344e-01
9.53675210e-01 2.37712950e-01 -2.40748689e-01 -1.96484923e-01
-1.06922269e+00 8.55339289e-01 1.09773552e+00 2.58517474e-01
-8.29132259e-01 -1.87805481e-02 -4.94724810e-02 2.62971580e-01
9.03193295e-01 4.80137825e-01 -7.73340583e-01 3.06979269e-01
-1.24876714e+00 3.52736264e-01 3.04897636e-01 8.93851399e-01
1.20849049e+00 1.50354758e-01 7.67049372e-01 8.51619244e-01
8.57742578e-02 1.07362270e+00 3.54231268e-01 -8.91093850e-01
1.75405592e-01 3.30311835e-01 1.88489184e-01 -1.08151639e+00
-5.70416093e-01 -5.68574846e-01 -7.54393518e-01 1.18953027e-02
3.60040426e-01 -5.44877172e-01 -1.00280142e+00 1.70751584e+00
1.68508038e-01 2.55854636e-01 1.41794709e-02 8.85691345e-01
2.15349182e-01 9.23807919e-01 5.57403445e-01 2.31222317e-01
1.07287788e+00 -2.30717659e-01 -1.29560933e-01 -5.39751112e-01
7.54382908e-01 -2.41199657e-01 8.40005279e-01 -1.13989376e-02
-4.21118826e-01 -4.65219826e-01 -8.81033301e-01 3.77961606e-01
-8.30125868e-01 3.57376821e-02 9.25357342e-01 3.92722160e-01
-1.41420543e+00 6.11182034e-01 -9.50716794e-01 -1.05567861e+00
4.37134117e-01 6.32954836e-02 -4.38522667e-01 -6.14454448e-02
-1.21425533e+00 9.07616079e-01 7.99579561e-01 1.36008039e-01
-5.04176080e-01 -8.41419578e-01 -8.37792039e-01 3.03250756e-02
4.48672427e-03 -3.19488913e-01 8.64012897e-01 -1.36221862e+00
-8.70932400e-01 8.00453782e-01 -5.24304807e-02 -4.84977841e-01
-9.93618965e-02 6.31588474e-02 -6.79654181e-01 -1.19156808e-01
6.55187786e-01 6.25998139e-01 2.38246962e-01 -1.26328123e+00
-1.09583235e+00 -7.85243571e-01 -1.01106934e-01 2.41579309e-01
-1.32524267e-01 5.13887517e-02 3.44039738e-01 -9.04455781e-01
2.66422153e-01 -1.12698114e+00 -5.01468718e-01 -1.91308856e-01
1.35664687e-01 3.56732756e-01 3.92446548e-01 -1.20902300e+00
1.18207455e+00 -2.24373293e+00 3.71545777e-02 3.19391340e-01
-2.74314046e-01 1.23556219e-02 -4.19387996e-01 5.21476448e-01
-2.73963153e-01 5.42319238e-01 -6.34951532e-01 6.31832182e-01
-1.76460266e-01 4.88002956e-01 -5.20876586e-01 5.16953170e-01
4.53290373e-01 7.78044224e-01 -8.83227468e-01 -1.27534373e-02
2.27436602e-01 2.02113479e-01 -1.07248828e-01 -1.76280811e-01
-7.75878802e-02 4.92419541e-01 -4.44945723e-01 6.45329773e-01
1.21990037e+00 -1.62002176e-01 5.03307760e-01 2.50374585e-01
-5.07112980e-01 1.24177560e-01 -9.30060506e-01 1.46565711e+00
-2.43616730e-01 5.09607673e-01 1.78014293e-01 -9.92719054e-01
9.13656235e-01 -1.51865661e-01 1.98256031e-01 -5.49662948e-01
-3.96745354e-01 1.48341030e-01 -1.09625503e-01 -3.03521812e-01
6.38826549e-01 -3.19065332e-01 -1.64401606e-01 4.06279355e-01
5.24134226e-02 -2.46357337e-01 -2.13245139e-01 -1.10080048e-01
3.08221042e-01 4.82727230e-01 7.23767817e-01 -8.76128495e-01
-2.10288148e-02 8.45707774e-01 5.16493499e-01 6.81597114e-01
8.19996744e-03 2.91299134e-01 1.82213783e-01 -4.65245545e-01
-1.08692145e+00 -1.15684438e+00 -7.41869211e-01 1.35904348e+00
-2.60129750e-01 1.49842529e-02 -2.47248530e-01 -2.55689710e-01
5.76909602e-01 9.17685986e-01 -7.47720122e-01 1.86383292e-01
-9.57257152e-02 -1.35835719e+00 8.28666866e-01 5.61806560e-01
5.49965382e-01 -3.79962921e-01 -4.92567480e-01 -3.22909132e-02
-1.96066186e-01 -5.68872511e-01 -9.57361143e-03 4.57469016e-01
-1.16289079e+00 -5.24813831e-01 -8.57072890e-01 -4.12051171e-01
3.32712293e-01 5.07129431e-01 9.47098672e-01 -2.05641463e-01
-3.67482081e-02 3.10953677e-01 -3.72037053e-01 -3.05844337e-01
-2.51212150e-01 4.97187227e-01 -8.17114338e-02 -3.02790701e-01
5.68523169e-01 -7.43129253e-01 -4.28726107e-01 3.76112014e-01
-1.03517163e+00 -2.29179710e-01 5.20728707e-01 7.01861262e-01
3.12938899e-01 1.61928132e-01 5.39602458e-01 -5.37016392e-01
6.16690181e-02 -1.23474753e+00 -6.63885117e-01 2.54334211e-01
-4.73397881e-01 -2.30524480e-01 2.34284729e-01 -1.15136467e-01
-1.17190552e+00 2.67310113e-01 4.14365858e-01 3.37166786e-01
-5.72906494e-01 1.37926483e+00 1.39833570e-01 -2.92351115e-02
1.10880768e+00 3.16985637e-01 1.65850133e-01 -4.85899001e-01
6.08296216e-01 8.51208091e-01 6.29739702e-01 -5.95953584e-01
9.44189548e-01 6.81326926e-01 -2.24640146e-01 -1.17891765e+00
-2.64098406e-01 -5.35734713e-01 -1.21921468e+00 4.72376831e-02
5.18396497e-01 -1.59782100e+00 3.80326569e-01 6.05837405e-01
-3.38407695e-01 -9.64339137e-01 9.80132818e-03 6.06170595e-01
-2.75622070e-01 2.36280859e-01 -1.00425720e-01 -5.05177379e-01
2.43822381e-01 -5.07339895e-01 7.83110559e-01 2.99097180e-01
-2.37775579e-01 -1.56832767e+00 2.54020810e-01 -2.35200837e-01
7.07449377e-01 5.27447939e-01 6.53272808e-01 -4.10373569e-01
-4.44630623e-01 1.04355715e-01 -4.76424426e-01 2.41841897e-01
7.34858692e-01 2.08486810e-01 -8.53297293e-01 -3.61801982e-01
-2.80783802e-01 3.27091664e-02 1.11742818e+00 8.51277232e-01
6.50320232e-01 -2.47869775e-01 -5.37642002e-01 1.08436632e+00
1.81683886e+00 3.63197215e-02 5.38767278e-01 8.88559401e-01
3.16577464e-01 6.37487888e-01 6.42306447e-01 3.24123561e-01
5.96109748e-01 3.60301733e-01 2.27321252e-01 -4.58494484e-01
3.16651851e-01 -3.90781797e-02 1.31092325e-01 -6.95068762e-02
-2.79990137e-01 1.45508721e-01 -1.52593446e+00 1.09345794e+00
-1.78472042e+00 -1.20805669e+00 -2.79449642e-01 2.10362267e+00
8.39368582e-01 -6.98088944e-01 2.58171588e-01 -6.49011314e-01
7.19014645e-01 4.32581782e-01 -6.11214101e-01 -6.49848208e-02
-5.84082961e-01 1.62518308e-01 1.41706085e+00 5.46069682e-01
-1.47380137e+00 1.22416210e+00 6.96026611e+00 3.84781063e-01
-1.52773285e+00 -6.47756830e-02 4.53197807e-01 -2.43452322e-02
-4.09797013e-01 1.90142050e-01 -4.66700166e-01 1.59224451e-01
1.19246554e+00 -4.54223961e-01 5.06456196e-01 6.54108822e-01
4.43876594e-01 -3.55841517e-01 -3.00397813e-01 1.33875534e-01
-2.25434989e-01 -1.15458179e+00 -6.56119213e-02 2.00790718e-01
8.87028813e-01 6.95330083e-01 3.09861511e-01 7.97507167e-02
1.03894246e+00 -9.72645819e-01 4.87094134e-01 4.01621908e-01
9.68155503e-01 -4.05031174e-01 4.24634010e-01 9.09066200e-02
-1.19020677e+00 -2.72351593e-01 -4.89608675e-01 -4.57890540e-01
-1.53718501e-01 2.21540406e-01 -8.77451360e-01 8.01089227e-01
1.14073551e+00 1.14804995e+00 -7.83728063e-01 7.74231970e-01
-1.28827229e-01 8.49497080e-01 -6.10474110e-01 6.78131461e-01
5.24556935e-01 -3.71518523e-01 3.01603824e-01 1.24193156e+00
6.54916167e-01 4.44524050e-01 1.08016387e-01 8.82408798e-01
5.79717159e-01 -8.46598074e-02 -9.69441116e-01 -2.54559983e-02
6.17202997e-01 9.36301947e-01 -5.15898407e-01 -1.40761048e-01
-3.92960012e-01 6.96314991e-01 -1.67019181e-02 8.81601095e-01
-5.09368539e-01 -2.10326582e-01 9.95622575e-01 9.98426750e-02
4.34786052e-01 -5.85851014e-01 -4.38251823e-01 -1.08329296e+00
-2.89291143e-01 -8.12437952e-01 5.82846224e-01 -1.11459064e+00
-1.20851231e+00 2.22480252e-01 4.14076179e-01 -1.11094427e+00
-3.47943306e-01 -5.92202067e-01 -4.36807305e-01 1.59620368e+00
-1.94598424e+00 -1.65220177e+00 -2.98876435e-01 4.54959899e-01
-1.03841349e-02 5.48077188e-02 1.04308856e+00 -3.13536644e-01
-2.53180116e-01 7.59818628e-02 9.76443827e-01 -3.45597304e-02
9.09953713e-01 -1.13176012e+00 8.42492640e-01 9.90630329e-01
-3.15704733e-01 8.12143087e-01 4.60419446e-01 -7.72253633e-01
-8.67118180e-01 -1.54283369e+00 7.78075814e-01 -2.78672665e-01
1.14149642e+00 1.74715638e-01 -1.14076734e+00 1.42990291e+00
7.23908991e-02 -3.53123128e-01 8.21010768e-01 4.60081935e-01
-3.41670126e-01 -2.96143860e-01 -1.19141483e+00 4.39360082e-01
7.32145786e-01 -6.62519753e-01 -4.98068422e-01 2.56428957e-01
3.42484474e-01 2.81650312e-02 -1.27560580e+00 1.34842664e-01
6.77292049e-01 -5.64976752e-01 8.47017884e-01 -7.60632694e-01
3.13753366e-01 -4.55244362e-01 -6.67221129e-01 -1.92416847e+00
-1.02398205e+00 -4.37365472e-02 1.03371882e+00 1.09800410e+00
6.19333267e-01 -1.17591429e+00 1.51390821e-01 5.26132345e-01
1.66251928e-01 3.15173090e-01 -6.88761771e-01 -1.07938027e+00
1.02001834e+00 -2.45747454e-02 1.00299585e+00 1.37187684e+00
8.73061568e-02 -4.45210963e-01 -4.08562981e-02 1.01057708e+00
4.59706604e-01 1.48395970e-01 7.31069624e-01 -1.41583252e+00
5.20048141e-02 -3.05175990e-01 -4.52396631e-01 -8.44639838e-01
1.02773853e-01 -8.22206736e-01 -2.53787279e-01 -1.18351817e+00
1.66853651e-01 -7.77180970e-01 -1.01724312e-01 8.44089150e-01
-2.69739330e-01 1.87124893e-01 1.01167947e-01 4.55369294e-01
2.21802756e-01 1.57344460e-01 6.25774562e-01 -1.69235885e-01
-3.01521540e-01 -1.04807161e-01 -9.65174079e-01 3.86638850e-01
1.32312024e+00 -5.00379801e-01 -1.34572357e-01 -8.10586691e-01
9.45922360e-02 -1.22876674e-01 8.51197779e-01 -7.48259723e-01
-9.21611786e-02 -9.76114333e-01 4.34829503e-01 -3.66171181e-01
-2.14247167e-01 -7.98645020e-01 7.28270411e-01 1.72668040e-01
-1.40389651e-01 6.23537675e-02 8.15295517e-01 3.54894817e-01
-3.38172987e-02 3.27329375e-02 5.52346289e-01 -3.62318784e-01
-1.30900395e+00 2.40918621e-01 -7.00796902e-01 -2.25905374e-01
8.07680070e-01 -2.65281230e-01 -4.39503551e-01 -1.81368083e-01
-8.65532279e-01 4.15917426e-01 1.09911215e+00 3.97247314e-01
-1.89606339e-01 -1.04423714e+00 -1.23690999e+00 4.78227586e-01
3.05165172e-01 -2.83223093e-01 4.34320480e-01 5.19494593e-01
-6.42051220e-01 5.50522685e-01 -5.84882915e-01 -8.65635276e-01
-9.15965199e-01 2.31781483e-01 6.64845645e-01 2.78588384e-01
-3.57582837e-01 5.60651958e-01 4.49844860e-02 -8.91153574e-01
-6.23160779e-01 -4.62159812e-01 2.99940426e-02 1.73133522e-01
2.40499109e-01 -6.76036850e-02 -3.40529144e-01 -8.28346252e-01
-3.34550947e-01 6.02260709e-01 3.18659782e-01 -2.13684246e-01
1.74494088e+00 -6.58182323e-01 -1.94238245e-01 6.42521918e-01
8.54763806e-01 -4.37760323e-01 -1.44631577e+00 -4.82984096e-01
1.58318251e-01 -5.91584563e-01 3.60298723e-01 -1.06405008e+00
-8.68457437e-01 5.58502257e-01 6.88446820e-01 2.99153507e-01
1.26325047e+00 -5.12733459e-02 -5.02063408e-02 2.39658713e-01
3.06215405e-01 -8.61072958e-01 -1.11467755e+00 3.73253167e-01
8.82988811e-01 -1.30274785e+00 2.76039034e-01 5.80248572e-02
-7.72851050e-01 9.80502248e-01 1.28593355e-01 -1.72108695e-01
1.02632892e+00 6.92609278e-03 2.47621417e-01 -3.15926448e-02
-2.88918972e-01 -4.65463728e-01 -1.49508268e-01 1.20685387e+00
3.50668132e-01 7.71199763e-01 5.72301820e-02 3.59304339e-01
-3.20567369e-01 3.05181891e-01 5.55133343e-01 6.78035736e-01
-4.12058055e-01 -7.60812521e-01 -7.74493337e-01 5.00735581e-01
-7.96929523e-02 -4.64214981e-01 -2.61861414e-01 1.08749211e+00
-1.22816801e-01 7.16517985e-01 3.29239875e-01 -1.02282874e-01
4.64773588e-02 2.08669767e-01 -3.49941701e-02 -6.23947322e-01
-2.56617457e-01 -1.37692302e-01 -7.85458907e-02 -1.80416971e-01
-7.08071172e-01 -1.30614638e+00 -7.48644650e-01 -7.00195551e-01
-2.17134222e-01 -1.86691303e-02 7.35974789e-01 6.10674322e-01
6.00525737e-01 1.05735257e-01 6.79915488e-01 -1.07724798e+00
-6.51808679e-01 -1.11301148e+00 -1.19685698e+00 -6.39224723e-02
5.70215046e-01 -3.52601081e-01 -5.14012277e-01 2.35251158e-01] | [9.438830375671387, -1.5778048038482666] |
08e7472b-f885-4481-baea-c3c10d6866af | a-frequency-domain-constraint-for-synthetic-x | 2105.06887 | null | https://arxiv.org/abs/2105.06887v2 | https://arxiv.org/pdf/2105.06887v2.pdf | A Frequency Domain Constraint for Synthetic and Real X-ray Image Super Resolution | Synthetic X-ray images are simulated X-ray images projected from CT data. High-quality synthetic X-ray images can facilitate various applications such as surgical image guidance systems and VR training simulations. However, it is difficult to produce high-quality arbitrary view synthetic X-ray images in real-time due to different CT slice thickness, high computational cost, and the complexity of algorithms. Our goal is to generate high-resolution synthetic X-ray images in real-time by upsampling low-resolution images with deep learning-based super-resolution methods. Reference-based Super Resolution (RefSR) has been well studied in recent years and has shown higher performance than traditional Single Image Super-Resolution (SISR). It can produce fine details by utilizing the reference image but still inevitably generates some artifacts and noise. In this paper, we introduce frequency domain loss as a constraint to further improve the quality of the RefSR results with fine details and without obvious artifacts. To the best of our knowledge, this is the first paper utilizing the frequency domain for the loss functions in the field of super-resolution. We achieved good results in evaluating our method on both synthetic and real X-ray image datasets. | ['WonSook Lee', 'Jae Chul Koh', 'Qing Ma'] | 2021-05-14 | null | null | null | null | ['reference-based-super-resolution'] | ['computer-vision'] | [ 5.20811379e-01 1.63004562e-01 -3.98999751e-02 -2.86174953e-01
-1.44837904e+00 1.28341332e-01 3.18606228e-01 -4.77726668e-01
-1.24241106e-01 1.12733495e+00 3.79075766e-01 3.38394158e-02
-3.24607104e-01 -9.47542429e-01 -6.45283878e-01 -6.69070423e-01
1.54080361e-01 2.87508637e-01 3.29313099e-01 -4.54635561e-01
-8.00643954e-03 4.87951547e-01 -1.19786692e+00 6.04130030e-01
9.46799457e-01 4.96799111e-01 8.56526017e-01 4.96107906e-01
2.68920988e-01 7.43812144e-01 -3.83303672e-01 9.45722759e-02
2.20236912e-01 -8.16314220e-01 -8.95332754e-01 -2.48627171e-01
4.24577892e-01 -6.88078165e-01 -6.27659202e-01 1.02675807e+00
7.94393182e-01 1.47133648e-01 2.56289363e-01 -2.00393423e-01
-6.74003601e-01 6.86192393e-01 -8.39042246e-01 5.20543575e-01
4.88482088e-01 -1.37460651e-02 2.41071120e-01 -7.49872565e-01
1.04256213e+00 9.02705550e-01 5.67026794e-01 1.05913627e+00
-1.21480632e+00 -7.64315546e-01 -5.18815815e-01 -1.34536907e-01
-9.77662444e-01 2.20219456e-02 7.79801607e-01 -1.13446400e-01
4.23317522e-01 5.32769024e-01 4.41013813e-01 1.16946721e+00
7.23177433e-01 3.45821828e-01 1.68555534e+00 -2.60626674e-01
7.51646385e-02 -1.06126137e-01 -3.70684594e-01 7.29983389e-01
-8.49066954e-03 6.47019327e-01 -4.50748742e-01 -6.05257526e-02
1.79789257e+00 2.00952634e-01 -8.76282096e-01 1.89929083e-02
-1.33436096e+00 6.11174405e-01 7.28172541e-01 6.13476336e-01
-5.22153199e-01 9.64521915e-02 1.81885973e-01 -6.25343472e-02
3.18791956e-01 8.59765828e-01 4.78255674e-02 1.29404599e-02
-9.56452250e-01 1.49861991e-01 -3.91287543e-02 6.96916580e-01
1.19073398e-01 3.03156465e-01 -5.41523024e-02 9.32964981e-01
-1.87520564e-01 1.50311708e-01 8.80835593e-01 -1.13660443e+00
1.47327185e-01 9.91431903e-03 1.37134641e-01 -3.81140113e-01
-3.77505034e-01 -8.67956936e-01 -1.15447581e+00 6.90119743e-01
1.02574095e-01 1.59071252e-01 -1.17637682e+00 1.49524426e+00
2.12485343e-01 3.18463653e-01 1.54322147e-01 1.54474783e+00
1.36351883e+00 5.73128700e-01 -1.52466521e-01 -5.62153339e-01
1.15253615e+00 -6.60481870e-01 -1.04546285e+00 -3.47220190e-02
1.66872799e-01 -8.98378789e-01 1.30856752e+00 3.61226588e-01
-1.69914985e+00 -4.74647909e-01 -1.22960413e+00 -2.38051545e-02
6.26370251e-01 -5.30031696e-03 6.64985061e-01 2.76371628e-01
-9.33485210e-01 8.81209493e-01 -1.07672799e+00 1.32212877e-01
6.11994684e-01 1.56761989e-01 -3.76710445e-01 -3.26927364e-01
-1.30870867e+00 8.90637875e-01 -1.44248396e-01 -3.51908714e-01
-6.78282678e-01 -1.54638493e+00 -7.52136528e-01 -1.66173667e-01
2.46534258e-01 -8.34931076e-01 1.27357030e+00 -5.25472820e-01
-1.61131752e+00 7.76309848e-01 6.43413141e-02 -1.44519895e-01
6.93655491e-01 3.29398736e-02 -4.88945931e-01 5.14506161e-01
1.45241663e-01 2.50206649e-01 4.59288210e-01 -1.71828377e+00
-3.97397012e-01 -5.63040972e-01 4.09000972e-03 2.95993835e-01
6.43549025e-01 -1.38875976e-01 -1.17060784e-02 -8.73591900e-01
5.19878805e-01 -8.06442916e-01 -5.70074677e-01 4.99236956e-02
-1.75701931e-01 6.80736840e-01 4.03484046e-01 -7.16442823e-01
9.29260910e-01 -1.85601532e+00 -8.98093581e-02 -2.34888449e-01
4.57388759e-01 1.05276540e-01 1.81662157e-01 1.37875909e-02
-3.73269886e-01 -8.52241367e-02 -4.06113923e-01 4.79045138e-02
-8.80634248e-01 4.61819256e-03 -4.21212524e-01 4.84594613e-01
-3.48219216e-01 8.79913092e-01 -1.17046452e+00 -6.29383326e-01
3.93289477e-01 1.12920940e+00 -4.30981129e-01 6.44243658e-02
2.35424742e-01 1.14939344e+00 -5.28978348e-01 3.28507036e-01
8.64769161e-01 -4.05259609e-01 -1.49301514e-01 -5.32209218e-01
-2.46930093e-01 2.09019586e-01 -7.82181025e-01 2.11529398e+00
-9.63395655e-01 5.12390912e-01 1.10087566e-01 -4.66470540e-01
5.21057904e-01 4.08694088e-01 8.02852154e-01 -1.23125112e+00
-2.53727306e-02 3.24532241e-01 -1.69456169e-01 -3.58907402e-01
4.98538435e-01 -1.00610852e+00 2.78017879e-01 4.03214395e-01
-2.92884588e-01 -4.96937215e-01 -2.83610463e-01 1.42233893e-01
7.25722611e-01 7.62217492e-02 2.51400713e-02 -2.28706524e-01
2.13477746e-01 1.60102144e-01 1.19438030e-01 5.57117105e-01
2.33386427e-01 1.31327534e+00 -1.17537983e-01 -3.76624435e-01
-1.41389465e+00 -1.39468992e+00 -6.82840943e-01 2.22325921e-01
6.01067305e-01 1.05730742e-02 -4.81237978e-01 -1.75093725e-01
-5.19220293e-01 7.16796219e-01 -6.56030953e-01 -9.98704657e-02
-1.03834832e+00 -9.22585607e-01 1.00494809e-01 5.86432695e-01
4.54339385e-01 -8.94844294e-01 -8.51178110e-01 2.84259439e-01
-5.20085692e-01 -1.11096704e+00 -4.63276297e-01 -1.55139044e-01
-1.13489401e+00 -8.51850569e-01 -1.16205955e+00 -6.28638387e-01
6.85000718e-01 3.09294701e-01 1.20933783e+00 -6.77484795e-02
-8.13995600e-01 -1.26534745e-01 -2.01294273e-01 -1.00564035e-02
-7.45909393e-01 -4.80869710e-01 -9.86251980e-02 -5.42135179e-01
-5.97667098e-01 -5.03724098e-01 -1.04129016e+00 3.59031618e-01
-1.03203392e+00 4.60280299e-01 7.13604689e-01 1.13727796e+00
1.25203037e+00 2.79370457e-01 3.98074567e-01 -1.12119830e+00
5.05909264e-01 -6.42497316e-02 -4.56377655e-01 -1.36026829e-01
-4.71240759e-01 3.51506770e-02 8.91943753e-01 -2.69141644e-01
-1.56518114e+00 -3.55173200e-01 -3.99105996e-01 -4.47260499e-01
1.02139071e-01 2.80162185e-01 5.02648830e-01 -3.84857565e-01
1.08132768e+00 5.02170861e-01 5.58508225e-02 -3.16606194e-01
-9.03779492e-02 2.54707992e-01 6.45056129e-01 -9.34887007e-02
5.48424959e-01 1.00973165e+00 3.51039648e-01 -7.27753341e-01
-9.62522388e-01 -3.30443941e-02 -1.25827745e-01 -2.02409983e-01
7.93499351e-01 -9.40961778e-01 -4.80571628e-01 9.73342825e-03
-6.18505657e-01 -2.79298097e-01 -3.47216606e-01 1.10249138e+00
-8.32314432e-01 3.32099408e-01 -9.76318359e-01 -2.12965846e-01
-3.81424844e-01 -1.64579523e+00 1.16016269e+00 2.38301396e-01
7.01424479e-02 -6.35460317e-01 -5.83389960e-03 3.46550077e-01
7.33175695e-01 7.17465043e-01 6.79545939e-01 6.34028673e-01
-8.41069460e-01 1.70951024e-01 -3.01654398e-01 5.53694405e-02
3.43492478e-01 -5.47958732e-01 -6.53710604e-01 -5.13422668e-01
5.96526623e-01 -2.39706352e-01 8.16117525e-01 9.99749959e-01
1.39113998e+00 7.19446614e-02 -3.25969070e-01 1.18220806e+00
2.03814077e+00 2.87980199e-01 8.57163727e-01 1.47092640e-01
5.08680165e-01 1.21529572e-01 8.63706589e-01 3.50959837e-01
-2.33142406e-01 7.76529074e-01 3.07360649e-01 -6.21411920e-01
-6.58660710e-01 -2.07295015e-01 -3.99526656e-01 3.09438080e-01
-4.77051914e-01 1.84899479e-01 -7.41782069e-01 2.68790573e-01
-1.23021102e+00 -1.20413935e+00 -1.50962830e-01 2.09645414e+00
9.73850608e-01 -6.28442392e-02 -3.98727536e-01 -1.23208486e-01
5.04947901e-01 6.22110851e-02 -5.46545565e-01 1.61783785e-01
1.00583293e-01 5.83502173e-01 4.98396754e-01 8.05138290e-01
-6.95987225e-01 3.98329288e-01 6.44769096e+00 9.02505040e-01
-1.56871247e+00 3.98446470e-01 6.87283456e-01 -5.25352657e-01
-5.45488119e-01 -4.34055030e-01 -4.36943531e-01 3.59429479e-01
4.26187992e-01 -2.73437738e-01 3.28036815e-01 6.17379904e-01
3.07259858e-01 -2.24927023e-01 -7.83107936e-01 1.31028891e+00
-3.60617712e-02 -1.89855576e+00 1.06939543e-02 -3.18210386e-02
1.02311122e+00 7.11213872e-02 3.68269056e-01 -1.98593065e-01
1.72829568e-01 -1.43459678e+00 1.53355852e-01 4.87131596e-01
1.64992070e+00 -1.01002300e+00 4.43550855e-01 1.38661057e-01
-7.99244344e-01 2.84707010e-01 -4.22265679e-01 5.60180008e-01
5.11505604e-01 6.55444980e-01 -6.31914794e-01 8.03387582e-01
6.80752814e-01 3.68504107e-01 8.21389779e-02 8.60436201e-01
1.85815006e-01 3.01029682e-01 -2.01191679e-02 5.73626816e-01
1.35836646e-01 -9.12590325e-02 6.95652783e-01 6.45592570e-01
3.64123046e-01 8.32387030e-01 -1.31801233e-01 9.78215575e-01
1.28578722e-01 -1.50979340e-01 -3.81446391e-01 6.05448365e-01
9.61619169e-02 9.32806253e-01 -5.69782376e-01 -2.65918136e-01
-3.54900807e-01 7.57307291e-01 -1.46545947e-01 4.22792792e-01
-9.28412557e-01 -1.35598913e-01 2.30689019e-01 9.16539490e-01
-2.49062717e-01 -2.71368548e-02 -5.98808885e-01 -1.20231366e+00
-1.33764176e-02 -6.65053666e-01 1.93855822e-01 -1.01678002e+00
-9.23602283e-01 1.32662940e+00 2.45751794e-02 -1.74036229e+00
-3.65727633e-01 -2.60984874e-03 -2.35897601e-01 1.18640339e+00
-1.81138992e+00 -8.28768969e-01 -6.53710365e-01 5.19096851e-01
7.25552082e-01 1.29030898e-01 7.40562677e-01 2.99388945e-01
2.44542405e-01 3.73004049e-01 3.26104164e-01 -3.09353560e-01
5.90047538e-01 -1.15427864e+00 9.02421325e-02 5.09078145e-01
-4.99003083e-01 4.29645300e-01 9.16173697e-01 -8.27474952e-01
-1.19893277e+00 -9.65982914e-01 9.94790345e-02 -2.84758002e-01
-5.70105854e-03 9.95467305e-02 -1.09129429e+00 3.71630758e-01
-1.85263176e-02 3.45493227e-01 5.83392978e-01 -4.84039515e-01
3.42219293e-01 1.70010731e-01 -1.55612469e+00 5.82898080e-01
9.99541581e-01 -3.76004994e-01 -3.64124715e-01 3.50267440e-01
7.85969317e-01 -1.21088469e+00 -1.22981143e+00 7.42436409e-01
3.91414285e-01 -1.20095456e+00 1.31541169e+00 -2.87351906e-02
1.09865141e+00 -1.33773804e-01 1.54539376e-01 -1.50499964e+00
-4.70715165e-01 -5.11576653e-01 4.08567876e-01 2.92425901e-01
2.36346543e-01 -4.66012627e-01 9.00962353e-01 3.02760720e-01
-4.10801977e-01 -1.03902996e+00 -8.78096640e-01 -5.90484500e-01
-1.09290620e-02 -4.35069911e-02 5.16210437e-01 1.03582954e+00
-6.25438839e-02 -4.54705730e-02 -2.59878546e-01 2.88930744e-01
1.05596948e+00 5.38324535e-01 1.46840975e-01 -7.16514647e-01
-3.68552208e-01 -1.98576704e-01 -6.20669425e-02 -7.31927454e-01
-2.89010584e-01 -8.20951700e-01 -1.04660794e-01 -1.87015378e+00
3.19972634e-01 -7.25218236e-01 -7.78969228e-02 -1.65583402e-01
-2.77782053e-01 6.26426041e-01 -1.21290475e-01 4.73160833e-01
1.39660046e-01 4.28209037e-01 2.43918729e+00 2.73702502e-01
-2.78365165e-01 1.69185862e-01 -6.79717422e-01 5.77759266e-01
6.21515036e-01 -2.81107515e-01 -7.46428072e-01 -5.39130270e-01
-2.01729853e-02 1.08662701e+00 2.85095841e-01 -9.04410720e-01
-5.22451699e-02 -1.62396118e-01 7.19482064e-01 -6.27306163e-01
5.82651556e-01 -5.51553309e-01 4.99378055e-01 6.39222741e-01
-4.58588094e-01 -2.09864020e-01 1.25271887e-01 2.41641507e-01
-3.57370794e-01 2.45991778e-02 1.59209907e+00 -8.03542852e-01
-2.47709200e-01 4.53066289e-01 1.24738820e-01 4.37014066e-02
9.24250662e-01 -3.79305720e-01 -1.25261247e-01 -3.04682612e-01
-9.67777967e-01 -2.55605012e-01 7.49126375e-01 1.76373184e-01
1.32056189e+00 -1.32832754e+00 -1.10694313e+00 3.03626567e-01
-2.59392172e-01 3.39686364e-01 9.40670490e-01 6.74393952e-01
-9.58045661e-01 3.24608326e-01 -4.83597279e-01 -6.16713822e-01
-1.15999937e+00 5.73760033e-01 5.46288967e-01 -5.19273162e-01
-1.35631037e+00 7.42560327e-01 6.63831174e-01 -1.65575184e-02
-2.25111395e-01 -1.77727908e-01 -1.75681576e-01 -7.80794621e-01
9.62650537e-01 4.68368568e-02 9.07458141e-02 -5.12410164e-01
-9.58556533e-02 9.47243333e-01 -2.64846206e-01 -2.13521287e-01
1.51971734e+00 -1.56103849e-01 2.81941175e-01 -3.50573994e-02
1.06347990e+00 8.04775045e-04 -1.31567895e+00 -2.13325873e-01
-8.12428176e-01 -1.05075932e+00 6.79157436e-01 -9.38367426e-01
-1.30871618e+00 6.20140553e-01 7.39243746e-01 -2.91980654e-01
1.35045230e+00 2.78236389e-01 1.10342455e+00 -4.86380517e-01
7.12308288e-01 -7.13045359e-01 1.54542938e-01 -2.21013099e-01
1.17225695e+00 -1.12973762e+00 4.10735488e-01 -8.95507932e-01
-6.38788164e-01 8.40809405e-01 6.63975656e-01 -2.74504304e-01
3.91976148e-01 6.53020918e-01 1.08151086e-01 -3.00242186e-01
-6.56090856e-01 1.55926347e-01 1.14646912e-01 5.60489476e-01
5.52880347e-01 4.21898030e-02 -5.36826253e-01 3.18637341e-01
-4.60593849e-01 3.73356491e-01 9.48813975e-01 7.60184348e-01
-2.01139182e-01 -7.51959801e-01 -4.32808280e-01 4.80117351e-01
-9.27957654e-01 -1.39636829e-01 5.47672391e-01 7.27247655e-01
-3.71761471e-01 3.65786672e-01 -1.58803120e-01 1.00664817e-01
4.64735538e-01 -7.99123943e-01 1.04344630e+00 -5.29646933e-01
-2.50136793e-01 1.73337221e-01 -1.70655385e-01 -8.76384258e-01
-2.70660460e-01 -2.32589141e-01 -1.65996516e+00 -5.65828905e-02
-1.19009651e-01 -9.34277102e-03 7.04782665e-01 3.12261283e-01
6.98861703e-02 1.10385537e+00 7.57267952e-01 -8.06461573e-01
-3.67254525e-01 -6.76110387e-01 -7.62880325e-01 5.06366074e-01
5.53535521e-01 -4.78848487e-01 -2.74551332e-01 -1.68125793e-01] | [13.547468185424805, -2.4770026206970215] |
fd0e06f2-f79e-4baf-b6bf-7394b552e823 | a-fast-and-accurate-unconstrained-face | 1408.1656 | null | https://arxiv.org/abs/1408.1656v3 | https://arxiv.org/pdf/1408.1656v3.pdf | A Fast and Accurate Unconstrained Face Detector | We propose a method to address challenges in unconstrained face detection, such as arbitrary pose variations and occlusions. First, a new image feature called Normalized Pixel Difference (NPD) is proposed. NPD feature is computed as the difference to sum ratio between two pixel values, inspired by the Weber Fraction in experimental psychology. The new feature is scale invariant, bounded, and is able to reconstruct the original image. Second, we propose a deep quadratic tree to learn the optimal subset of NPD features and their combinations, so that complex face manifolds can be partitioned by the learned rules. This way, only a single soft-cascade classifier is needed to handle unconstrained face detection. Furthermore, we show that the NPD features can be efficiently obtained from a look up table, and the detection template can be easily scaled, making the proposed face detector very fast. Experimental results on three public face datasets (FDDB, GENKI, and CMU-MIT) show that the proposed method achieves state-of-the-art performance in detecting unconstrained faces with arbitrary pose variations and occlusions in cluttered scenes. | ['Stan Z. Li', 'Anil K. Jain', 'Shengcai Liao'] | 2014-08-06 | null | null | null | null | ['robust-face-recognition'] | ['computer-vision'] | [-7.87587613e-02 -1.14075094e-01 -1.22745223e-01 -5.52116275e-01
-3.29354048e-01 -5.20563960e-01 3.32455724e-01 -5.95252454e-01
-4.39313024e-01 3.42686921e-01 -3.70899320e-01 6.74787164e-02
1.11731075e-01 -4.16581899e-01 -6.38337135e-01 -7.50993013e-01
-1.72116935e-01 9.74372923e-02 9.35856029e-02 1.82323575e-01
-1.73102249e-03 8.59070122e-01 -1.71016526e+00 1.39590651e-02
5.83895326e-01 1.51681340e+00 6.25494719e-02 2.09427729e-01
3.31398815e-01 1.49552658e-01 -4.08896565e-01 -5.05764008e-01
7.04066098e-01 -1.21358499e-01 -3.40191156e-01 3.70480001e-01
8.66689146e-01 -7.75242329e-01 -5.08005500e-01 1.31821227e+00
7.36332417e-01 2.64507849e-02 6.41935050e-01 -1.30252421e+00
-5.82583666e-01 -6.43962398e-02 -7.87228405e-01 1.53545961e-01
6.11179709e-01 3.21538597e-01 6.36791229e-01 -1.44780028e+00
6.55245841e-01 1.98893189e+00 4.03817087e-01 6.22248590e-01
-1.33059239e+00 -9.20734525e-01 1.99306533e-01 3.05528879e-01
-1.73871922e+00 -7.54966497e-01 6.30930007e-01 -3.10728669e-01
5.91500998e-01 1.64601207e-01 6.17649078e-01 9.58885431e-01
-3.64697091e-02 7.20890999e-01 1.06727803e+00 -3.39227945e-01
1.59599781e-01 7.19872117e-02 -2.48159468e-01 1.16635931e+00
5.20708382e-01 -9.90658328e-02 -3.77806664e-01 -8.60210881e-02
1.08795571e+00 8.93415883e-02 -3.67810637e-01 -4.18807536e-01
-8.24551404e-01 6.56915247e-01 4.79838729e-01 -1.38674468e-01
2.19960231e-03 -3.21334183e-01 6.24168403e-02 1.66145965e-01
2.90625930e-01 -1.51167542e-01 -3.57471138e-01 4.74897802e-01
-5.29686570e-01 9.47975591e-02 6.81088507e-01 9.40563023e-01
8.01188946e-01 -8.57580006e-02 -3.02633882e-01 8.17467511e-01
5.26180089e-01 9.89002228e-01 2.37303719e-01 -9.70558524e-01
2.05249637e-01 4.68217522e-01 2.33964339e-01 -1.27574706e+00
-5.37395835e-01 -1.98455155e-01 -6.96315110e-01 2.99777687e-01
5.63791871e-01 -7.99738094e-02 -7.53306270e-01 1.77885377e+00
7.46735752e-01 7.15770200e-02 -4.24818367e-01 1.12923634e+00
8.69634569e-01 2.67896473e-01 -1.74998626e-01 -6.33867919e-01
1.68879914e+00 -5.23494542e-01 -7.06659913e-01 -3.91516238e-01
1.30516559e-01 -5.35335124e-01 8.71521413e-01 2.79315382e-01
-8.22255075e-01 -6.29112720e-01 -1.00657368e+00 -5.43959774e-02
-2.12199494e-01 5.98192155e-01 6.43037319e-01 8.47950220e-01
-9.60741282e-01 4.57611173e-01 -7.67969728e-01 -3.14681798e-01
7.43104041e-01 5.68806648e-01 -4.22858357e-01 -1.51427045e-01
-7.42907763e-01 6.08097136e-01 1.43677831e-01 4.24317747e-01
-7.63404369e-01 -3.54890168e-01 -9.10780847e-01 8.12882930e-03
5.36322176e-01 -3.62333953e-01 8.71705532e-01 -8.03063452e-01
-1.59067357e+00 1.14470947e+00 -4.63660181e-01 1.20129578e-01
4.14467871e-01 1.77033674e-02 -3.42286259e-01 4.80048060e-01
4.12655771e-02 8.68478119e-01 1.46470892e+00 -9.13924813e-01
-3.51769984e-01 -8.52318406e-01 -5.39432466e-02 1.28492475e-01
-7.65278816e-01 3.01136792e-01 -6.85202181e-01 -3.29523593e-01
3.05033833e-01 -9.04280365e-01 1.65396661e-01 7.56717622e-01
-2.94510126e-01 -5.71317375e-01 8.59923184e-01 -6.69795513e-01
8.54907334e-01 -2.31302786e+00 1.29199520e-01 1.73091307e-01
2.89282352e-01 4.35871542e-01 -3.17037165e-01 -4.09787267e-01
4.06810939e-02 -1.11078948e-01 2.73896568e-02 -3.08353007e-01
1.21765654e-03 -1.15908287e-01 -1.15238003e-01 8.18418324e-01
5.28914094e-01 7.50421286e-01 -7.10735321e-01 -6.45262241e-01
1.86245684e-02 5.46401262e-01 -5.22475064e-01 2.10473597e-01
1.15859997e-03 4.96556073e-01 -4.51956987e-01 9.72372413e-01
1.17355192e+00 -3.72314528e-02 1.94110408e-01 -3.87343466e-01
1.57438099e-01 -1.44689471e-01 -1.41615283e+00 1.22230434e+00
-1.34370551e-01 5.49593985e-01 2.16973901e-01 -7.93371081e-01
8.80692542e-01 9.08726677e-02 1.16540097e-01 -5.37957907e-01
5.08426309e-01 1.74656153e-01 -6.77061304e-02 -4.26929593e-01
-2.14840949e-01 3.39478403e-01 3.31510574e-01 6.99130818e-02
1.85422733e-01 1.70261711e-01 2.57133186e-01 -2.93708086e-01
6.93600595e-01 7.57934973e-02 3.93355846e-01 -2.73895264e-01
7.63268650e-01 -8.19341004e-01 9.44278061e-01 2.90956527e-01
-6.07850373e-01 3.45690131e-01 6.47190154e-01 -4.92906123e-01
-4.93085444e-01 -1.20823014e+00 -5.27454436e-01 1.03980160e+00
3.55157763e-01 -8.64518657e-02 -1.13302445e+00 -6.80467606e-01
1.36487022e-01 -1.62186399e-01 -6.41060948e-01 -2.75480375e-02
-5.48394501e-01 -6.93874419e-01 1.81932330e-01 3.60762030e-01
6.29048944e-01 -8.83277237e-01 -4.46370959e-01 -2.60787040e-01
3.76106910e-02 -1.44483018e+00 -8.94620597e-01 -3.88832003e-01
-7.11345136e-01 -1.17785811e+00 -5.02254725e-01 -1.08855879e+00
1.05962932e+00 5.86867452e-01 4.80555773e-01 8.87296349e-02
-7.89236605e-01 3.72622460e-01 7.68088773e-02 -2.07758099e-01
2.67100006e-01 -4.75675792e-01 5.40415168e-01 5.68540931e-01
4.84005213e-01 -3.55261177e-01 -8.07493329e-01 5.15355229e-01
-4.87225860e-01 -4.62893069e-01 6.08272195e-01 8.26424658e-01
6.74160480e-01 -7.24648386e-02 5.19622803e-01 -3.81311893e-01
1.52028978e-01 8.76275524e-02 -8.95668566e-01 2.50590354e-01
-1.61380440e-01 -2.41855919e-01 5.20720541e-01 -8.17606628e-01
-1.07792377e+00 4.40851957e-01 2.67897218e-01 -6.84603274e-01
-2.80711036e-02 -1.60262302e-01 -6.17759585e-01 -6.61886930e-01
5.22044718e-01 1.28795862e-01 1.22361042e-01 -3.28963876e-01
3.76992166e-01 7.63867319e-01 6.30269051e-01 -2.34963223e-01
1.11073339e+00 7.25606382e-01 1.22534998e-01 -1.12599134e+00
-6.86459899e-01 -2.19374090e-01 -9.61191416e-01 -3.54200751e-01
6.37983918e-01 -1.02554750e+00 -1.27208006e+00 5.80214679e-01
-1.15241659e+00 2.50209123e-01 1.86983675e-01 4.27239746e-01
-2.55576134e-01 5.40738463e-01 -5.81439495e-01 -1.00353110e+00
-4.04889822e-01 -1.13204598e+00 1.32691836e+00 5.47233641e-01
2.17150375e-01 -5.14597952e-01 -6.57449305e-01 2.06577212e-01
9.04087201e-02 3.02538633e-01 5.96378446e-01 -2.60103047e-01
-7.37710953e-01 -3.03616971e-01 -4.60160881e-01 2.70968705e-01
1.77914605e-01 -2.55143400e-02 -1.14576721e+00 -8.18789721e-01
2.29467258e-01 -4.47313637e-01 8.05502355e-01 2.66101331e-01
1.32236278e+00 -4.33513254e-01 -4.45699900e-01 8.56250107e-01
9.69568014e-01 5.01733422e-02 3.10265958e-01 -2.52993196e-01
6.20608628e-01 4.97931182e-01 5.60440302e-01 4.81623232e-01
2.35949174e-01 7.81926692e-01 4.32910115e-01 6.00768141e-02
-2.85527647e-01 -1.48580357e-01 5.49102068e-01 4.17506158e-01
6.15892895e-02 3.42969716e-01 -2.96619773e-01 2.95742042e-02
-1.49895358e+00 -7.31019735e-01 2.21755683e-01 2.36071110e+00
7.19337285e-01 2.96990648e-02 2.07088217e-01 -8.40913057e-02
9.06159699e-01 -1.29357621e-01 -8.72220099e-01 2.11720690e-01
-9.05438066e-02 1.29483551e-01 2.02808022e-01 3.83476973e-01
-1.32421887e+00 1.09492326e+00 6.05793381e+00 8.87311816e-01
-1.19813085e+00 1.10282451e-01 6.57442391e-01 -2.93867499e-01
3.41795832e-01 -4.04821336e-01 -1.13353074e+00 4.56986874e-01
3.01060319e-01 -2.93547995e-02 7.05151021e-01 1.04291236e+00
7.16691241e-02 7.94678479e-02 -1.21924937e+00 1.44215381e+00
3.40896219e-01 -6.32395625e-01 -1.06616639e-01 3.35577950e-02
4.88261580e-01 -5.50714612e-01 5.35308361e-01 2.31045470e-01
-3.72040689e-01 -9.05540347e-01 5.61750591e-01 2.81900913e-01
9.79724407e-01 -5.74851394e-01 3.40749830e-01 1.34532705e-01
-1.34894240e+00 -3.62993479e-01 -8.59420598e-01 9.06436145e-02
-3.91799241e-01 4.68702674e-01 -7.29920685e-01 3.56156453e-02
6.79810464e-01 5.13938189e-01 -6.42322898e-01 9.92116272e-01
-2.50099987e-01 1.47643000e-01 -4.44471002e-01 -8.22140947e-02
-4.00630772e-01 -2.09265873e-01 6.37789369e-01 9.53622699e-01
1.92389265e-01 1.59467772e-01 3.39632928e-01 8.55117917e-01
-3.97223979e-01 1.78485021e-01 -5.18019319e-01 2.93042183e-01
6.52746141e-01 1.70645249e+00 -8.38048756e-01 -5.98524027e-02
-4.22264099e-01 1.11903691e+00 4.80783463e-01 4.47269708e-01
-7.52027273e-01 -3.50512743e-01 7.82166779e-01 -1.47380698e-02
5.38318217e-01 -1.15115672e-01 1.33057803e-01 -1.26472056e+00
5.31396031e-01 -7.62710094e-01 2.97285229e-01 -3.73720676e-01
-1.20876753e+00 5.61489880e-01 -1.30459383e-01 -1.24186218e+00
7.95829669e-02 -1.16071415e+00 -3.83854657e-01 6.44650102e-01
-1.46150279e+00 -9.93079424e-01 -4.95624959e-01 6.29953027e-01
4.51478571e-01 -2.01682940e-01 5.93034029e-01 3.97969365e-01
-1.01902914e+00 1.04500484e+00 -2.14182034e-01 3.15731943e-01
8.32671344e-01 -8.27360630e-01 2.15912968e-01 8.26525390e-01
3.29641663e-02 7.23041892e-01 3.17860574e-01 -4.12360281e-01
-1.88903475e+00 -1.12489235e+00 3.72816265e-01 -3.52854192e-01
2.86311090e-01 -6.90434098e-01 -8.85285676e-01 4.92355764e-01
-3.94589305e-01 7.16262817e-01 3.68089020e-01 -2.62197733e-01
-6.76154971e-01 -5.34690380e-01 -1.45388329e+00 5.35133541e-01
1.41540825e+00 -4.65386450e-01 -3.52062106e-01 6.74734116e-01
5.42622268e-01 -4.62688178e-01 -5.90310633e-01 3.86256933e-01
8.08743179e-01 -8.70070636e-01 1.12183237e+00 -4.64914858e-01
-2.95269430e-01 -4.06318724e-01 -1.81754157e-02 -9.25675154e-01
-4.73695576e-01 -7.94400096e-01 -4.60777700e-01 1.02256763e+00
4.35147323e-02 -8.13712060e-01 5.55857122e-01 4.60453361e-01
3.56201142e-01 -7.99410582e-01 -1.31323016e+00 -1.02077734e+00
-3.81471962e-01 6.69646263e-02 4.40491229e-01 5.96044898e-01
-2.04042420e-01 1.89495042e-01 -2.76821524e-01 5.07979095e-01
1.02029920e+00 9.36971977e-02 6.19466364e-01 -1.52399409e+00
-1.19820751e-01 -2.74956018e-01 -8.14672053e-01 -1.29712749e+00
4.52869147e-01 -8.04948092e-01 -1.17418490e-01 -9.27938104e-01
3.93572271e-01 8.02329034e-02 -2.06272714e-02 5.01085281e-01
-3.73860419e-01 4.82327461e-01 1.17674097e-01 1.43720567e-01
-5.68874121e-01 6.19226515e-01 1.48408663e+00 -1.82809651e-01
-1.01381429e-01 -6.77743703e-02 -6.43675506e-01 8.60579431e-01
3.59555423e-01 -1.49920747e-01 -1.01804890e-01 -2.87264198e-01
-3.90959144e-01 -2.00105399e-01 3.58238041e-01 -9.27435219e-01
2.43524373e-01 -4.62912656e-02 1.09637284e+00 -2.91735202e-01
5.39777994e-01 -6.49706006e-01 -4.71146643e-01 6.20621920e-01
1.68796688e-01 -4.38434035e-01 2.10759088e-01 6.08624339e-01
2.48451889e-01 5.79844378e-02 1.19548702e+00 3.00539255e-01
-5.28645098e-01 6.54239178e-01 1.43631279e-01 -1.13674730e-01
1.22685742e+00 -1.08646683e-01 -2.11112186e-01 -1.10145368e-01
-5.50862491e-01 2.58343726e-01 1.97932214e-01 6.38481617e-01
7.85436153e-01 -1.54713070e+00 -7.10119307e-01 6.88654840e-01
-7.50244856e-02 -2.47428879e-01 1.88817188e-01 7.92006969e-01
-2.27181703e-01 3.50427806e-01 -3.04401875e-01 -7.32398450e-01
-1.39157295e+00 7.84532368e-01 4.20131803e-01 4.61089671e-01
-3.32266927e-01 8.70785832e-01 3.35957170e-01 -2.70323813e-01
3.94829541e-01 -1.15544200e-01 -1.57639861e-01 2.47183368e-01
1.03894842e+00 3.22027028e-01 -5.27523868e-02 -7.59756207e-01
-5.16551673e-01 9.68234122e-01 -8.27471018e-02 2.45577991e-01
9.97724712e-01 -1.02661207e-01 -2.72023469e-01 1.08545208e-02
1.39441991e+00 -1.83261722e-01 -1.56242430e+00 -4.53999728e-01
-3.44896555e-01 -9.13383245e-01 -1.48925379e-01 -2.48635888e-01
-1.27274787e+00 8.06981683e-01 9.74597335e-01 1.65255629e-02
1.27436352e+00 -9.44490370e-04 5.13926506e-01 6.53320014e-01
4.72288430e-01 -8.59367132e-01 2.93972313e-01 2.03486860e-01
1.14502788e+00 -1.45944154e+00 3.16943787e-02 -9.39931214e-01
-1.86608151e-01 1.19342160e+00 9.51165020e-01 6.17958531e-02
6.87002599e-01 -5.44368587e-02 -1.77695394e-01 1.30661532e-01
-3.27400029e-01 -3.42584640e-01 4.67961848e-01 7.22297072e-01
9.82335284e-02 7.80277997e-02 -1.10659741e-01 4.54393148e-01
-3.56541052e-02 -2.45544747e-01 1.29385898e-02 5.38145661e-01
-6.90131843e-01 -6.99387908e-01 -5.68000495e-01 5.07465243e-01
-2.48711750e-01 3.00050169e-01 -3.63575459e-01 5.63129187e-01
5.79220094e-02 9.38305080e-01 1.81900501e-01 -1.99149400e-01
2.09106296e-01 -2.43484452e-02 8.07395279e-01 -5.72268128e-01
2.81018227e-01 1.13976285e-01 -5.27540922e-01 -7.41152525e-01
6.84982399e-03 -5.97074151e-01 -1.05943429e+00 -6.89954460e-02
-3.83674651e-01 -3.56146157e-01 5.14661670e-01 8.12290251e-01
3.41414839e-01 5.18732630e-02 1.06306481e+00 -1.22533429e+00
-8.78659070e-01 -1.03294098e+00 -6.98340356e-01 4.14834827e-01
3.55169237e-01 -1.15754807e+00 -5.28754890e-01 -2.81669572e-02] | [13.361499786376953, 0.5574584603309631] |
f8277761-9077-4994-992a-3fe5e4513d7e | loosely-synchronized-search-for-multi-agent | 2103.04516 | null | https://arxiv.org/abs/2103.04516v2 | https://arxiv.org/pdf/2103.04516v2.pdf | Loosely Synchronized Search for Multi-agent Path Finding with Asynchronous Actions | Multi-agent path finding (MAPF) determines an ensemble of collision-free paths for multiple agents between their respective start and goal locations. Among the available MAPF planners for workspace modeled as a graph, A*-based approaches have been widely investigated due to their guarantees on completeness and solution optimality, and have demonstrated their efficiency in many scenarios. However, almost all of these A*-based methods assume that each agent executes an action concurrently in that all agents start and stop together. This article presents a natural generalization of MAPF with asynchronous actions (MAPF-AA) where agents do not necessarily start and stop concurrently. The main contribution of the work is a proposed approach called Loosely Synchronized Search (LSS) that extends A*-based MAPF planners to handle asynchronous actions. We show LSS is complete and finds an optimal solution if one exists. We also combine LSS with other existing MAPF methods that aims to trade-off optimality for computational efficiency. Numerical results are presented to corroborate the performance of LSS and the applicability of the proposed method is verified in the Robotarium, a remotely accessible swarm robotics research platform. | ['Howie Choset', 'Sivakumar Rathinam', 'Zhongqiang Ren'] | 2021-03-08 | null | null | null | null | ['multi-agent-path-finding'] | ['playing-games'] | [ 1.07524008e-01 2.23204255e-01 -4.09094989e-02 2.41203278e-01
-1.82543397e-01 -8.71405423e-01 7.17503786e-01 4.48129982e-01
-5.35205960e-01 1.14929974e+00 -3.08797240e-01 -1.76653892e-01
-1.05412471e+00 -8.37695062e-01 -5.01457751e-01 -8.73608112e-01
-6.89827502e-01 7.73493528e-01 6.12949252e-01 -5.25592327e-01
5.41158736e-01 7.03290939e-01 -1.42813003e+00 -6.14141941e-01
1.08524275e+00 5.04158080e-01 9.01510060e-01 6.39451623e-01
8.46229196e-02 4.78840262e-01 -5.93135178e-01 3.15546751e-01
5.41747332e-01 -3.23508620e-01 -9.71295893e-01 4.14411761e-02
-4.99694258e-01 -8.53886083e-02 -1.05146751e-01 9.09451723e-01
3.03241730e-01 4.28439051e-01 4.96089011e-01 -2.06637907e+00
1.74021021e-01 4.68475282e-01 -4.82813001e-01 -8.37631226e-02
5.38337708e-01 4.86598760e-02 5.80734968e-01 -3.35985690e-01
9.15397584e-01 1.07016325e+00 4.60018545e-01 1.48270443e-01
-7.85625339e-01 -2.30587259e-01 1.82934985e-01 3.24480563e-01
-1.26085746e+00 -1.01555251e-02 4.18879092e-01 -9.01219398e-02
1.17565739e+00 2.05366552e-01 7.07748771e-01 3.09685141e-01
7.60654330e-01 3.08929384e-01 1.21435416e+00 -4.31333840e-01
6.72503948e-01 -3.96751642e-01 -9.17223934e-03 6.88038766e-01
5.31420410e-01 1.56015456e-01 -2.61451602e-01 -3.66750628e-01
7.53117323e-01 -1.56728417e-01 -1.47949085e-01 -7.09250748e-01
-1.69034350e+00 6.84827924e-01 3.95787686e-01 2.92278051e-01
-8.20568144e-01 9.42641422e-02 1.52588502e-01 3.39132041e-01
-2.38293126e-01 6.74544096e-01 -8.82052556e-02 -2.05262408e-01
-2.52850264e-01 5.91932476e-01 1.09875500e+00 1.19014251e+00
7.26804793e-01 -2.19239995e-01 3.26680750e-01 2.77251035e-01
1.44674048e-01 5.28848588e-01 -2.95970794e-02 -1.34942210e+00
3.24140012e-01 6.90847933e-01 7.51572311e-01 -1.18370569e+00
-9.95645463e-01 -1.68083787e-01 -4.77517694e-01 8.07839870e-01
3.54735374e-01 -1.97117642e-01 -1.83891922e-01 1.53542554e+00
5.76959014e-01 1.46004245e-01 4.83338714e-01 8.49485576e-01
1.35609135e-01 8.60384166e-01 -2.69686431e-01 -5.82226634e-01
1.01610088e+00 -1.19556880e+00 -7.44022369e-01 -2.23714218e-01
6.07241154e-01 -5.79389513e-01 5.55554211e-01 4.88004178e-01
-1.04719865e+00 -1.82381719e-02 -1.18106437e+00 7.61341214e-01
-3.27320963e-01 -2.36632317e-01 4.76186752e-01 3.39956224e-01
-1.38112748e+00 5.31548321e-01 -9.90876317e-01 -8.51111293e-01
-3.72456670e-01 7.02790082e-01 -4.96865511e-01 3.73568684e-02
-7.45090067e-01 1.25291169e+00 6.31320059e-01 1.31994724e-01
-9.71513033e-01 -3.86672989e-02 -5.34843087e-01 -6.65864870e-02
9.52079356e-01 -5.62894344e-01 1.23939276e+00 -5.21736205e-01
-1.82824612e+00 8.11672583e-03 2.04537362e-02 -4.85270530e-01
4.36721593e-01 2.01147571e-01 2.07616054e-02 2.25047380e-01
3.93763661e-01 3.66356015e-01 1.63164914e-01 -1.36179376e+00
-9.46779132e-01 -1.49462372e-01 4.70494777e-01 6.69656992e-01
-3.84523124e-02 -9.24888179e-02 1.04444996e-01 -6.50670603e-02
9.79277268e-02 -1.27988565e+00 -7.60031343e-01 -1.87187791e-01
-2.19570518e-01 -4.33636367e-01 7.27808595e-01 1.55756593e-01
8.74635875e-01 -1.65116465e+00 5.01788199e-01 3.52944136e-01
-7.82509986e-03 -7.35187680e-02 -3.05946410e-01 1.49327445e+00
4.84911233e-01 -2.51935363e-01 -2.33558744e-01 -1.07804611e-01
7.86661282e-02 6.51028216e-01 2.52955444e-02 8.95823359e-01
-2.47806653e-01 3.75697106e-01 -1.27384436e+00 -5.01499951e-01
4.83160883e-01 -9.10401996e-03 -2.46716008e-01 -2.49670744e-02
-2.41612658e-01 3.99911910e-01 -6.35274291e-01 3.39265317e-01
6.74977720e-01 5.94687834e-02 4.74865198e-01 4.22909647e-01
-1.11997759e+00 -4.97309305e-02 -1.64907813e+00 1.79231346e+00
-3.35986704e-01 2.35362485e-01 6.27972364e-01 -9.63209867e-01
1.01712656e+00 2.33486041e-01 9.16354537e-01 -4.52349007e-01
1.11672640e-01 6.22398913e-01 1.42939433e-01 -2.76364684e-01
6.87961996e-01 3.16059709e-01 -2.06910133e-01 6.18651688e-01
-3.30649614e-01 -1.57674193e-01 6.27404153e-01 1.27042711e-01
1.51427412e+00 2.05435425e-01 7.32173204e-01 -7.71166980e-01
7.59814322e-01 6.96627200e-01 5.78367174e-01 9.82608795e-01
-4.43996161e-01 -7.58933201e-02 1.95591450e-01 -3.32915276e-01
-6.66543484e-01 -1.02135646e+00 4.36152279e-01 7.35351980e-01
1.29082799e+00 -4.34640557e-01 -5.37859499e-01 -3.69755983e-01
-2.08555251e-01 6.40111566e-01 -4.07112241e-01 2.69411772e-01
-9.11485255e-01 -4.43493545e-01 1.68020357e-04 -1.11194208e-01
6.18269503e-01 -1.01209211e+00 -1.52314150e+00 6.52287543e-01
-1.51173323e-01 -9.18577015e-01 -1.37900651e-01 2.41296679e-01
-4.66030806e-01 -1.50525594e+00 -4.11833256e-01 -9.67734933e-01
8.04229140e-01 8.25487614e-01 4.77574527e-01 1.38853163e-01
1.08655713e-01 6.32550538e-01 -8.11886430e-01 -3.52043748e-01
-5.46885788e-01 1.19244553e-01 3.06637228e-01 -2.79317826e-01
-4.03013289e-01 -5.54989874e-01 -5.20080984e-01 7.66503036e-01
-6.39711618e-01 2.56284997e-02 6.63948298e-01 4.59158927e-01
5.06522894e-01 5.95080018e-01 7.02576220e-01 -1.89389199e-01
7.82232523e-01 -6.16987884e-01 -8.88633966e-01 2.33966008e-01
-6.22404218e-01 -1.11418113e-01 8.17206681e-01 -1.45359516e-01
-7.99752474e-01 2.50508666e-01 3.48077118e-01 3.79844976e-04
-8.91744345e-02 5.99034250e-01 7.40949810e-02 -4.18675125e-01
2.97559619e-01 2.65620649e-01 4.50579405e-01 4.42042528e-03
1.18914500e-01 3.27186525e-01 5.22510648e-01 -4.35659111e-01
7.05399156e-01 6.04472399e-01 5.17810822e-01 -6.98912561e-01
1.96443260e-01 -5.50362945e-01 -4.29525673e-01 -6.50361836e-01
5.77742517e-01 -2.12472320e-01 -1.22996807e+00 3.29715937e-01
-1.28241086e+00 -4.14561272e-01 5.22536822e-02 4.18408215e-01
-1.11223447e+00 3.16701800e-01 -6.12969883e-02 -1.17746353e+00
-3.84054109e-02 -1.31911969e+00 6.24506652e-01 3.76991689e-01
-8.82628709e-02 -9.06988978e-01 3.09018403e-01 -3.37019026e-01
4.70400572e-01 7.32384324e-01 3.87112886e-01 -5.83555281e-01
-7.00932145e-01 1.52484044e-01 2.28030905e-01 -7.99188554e-01
3.08605582e-01 -2.27709591e-01 1.77440181e-01 -6.45175815e-01
-9.37767997e-02 2.28882030e-01 -1.02225550e-01 4.16285843e-01
-8.33280981e-02 -5.04900575e-01 -8.81605327e-01 -2.34835539e-02
1.70339549e+00 7.97432601e-01 3.22605252e-01 1.26589453e+00
-7.57967010e-02 7.81904399e-01 1.31824517e+00 6.96913123e-01
5.19584239e-01 9.51634049e-01 1.16669774e+00 2.46085539e-01
2.88560987e-01 2.55280882e-01 5.23912013e-01 5.41804254e-01
-2.86060899e-01 -6.24860168e-01 -8.95228207e-01 7.01206982e-01
-2.44274855e+00 -8.19187701e-01 -4.05526996e-01 2.14022374e+00
7.75077343e-02 -1.75747961e-01 4.07415390e-01 2.60323554e-01
9.49708879e-01 -1.32026076e-01 -3.42303395e-01 -3.74038577e-01
-3.26555292e-03 -3.30144614e-01 9.93947208e-01 7.37770915e-01
-9.20629680e-01 6.56088114e-01 5.21579695e+00 3.74844611e-01
-6.33968234e-01 1.67690404e-02 -4.46454257e-01 1.48590043e-01
1.18255459e-01 1.91081971e-01 -3.73840272e-01 1.91031486e-01
7.05735683e-01 -5.62667608e-01 7.67573953e-01 7.34745443e-01
6.14301622e-01 -5.82749009e-01 -5.89901865e-01 6.24794006e-01
-3.05431128e-01 -1.21763825e+00 -5.68930268e-01 2.20066562e-01
7.24142432e-01 -1.50521860e-01 -4.14060235e-01 -1.09621756e-01
6.55574918e-01 -5.68646371e-01 9.24224913e-01 1.98739886e-01
-4.12906297e-02 -1.06913555e+00 7.62225866e-01 7.80007482e-01
-1.54254627e+00 -4.19838607e-01 -2.20016792e-01 -5.32057405e-01
7.72887051e-01 -5.95889054e-02 -1.20419812e+00 1.25072217e+00
5.53071618e-01 3.35272849e-01 -3.23697105e-02 1.45938754e+00
-5.96028343e-02 -1.71499640e-01 -5.81482589e-01 -6.60250723e-01
6.98788941e-01 -5.76432586e-01 1.14710057e+00 6.75051272e-01
5.48495471e-01 6.67696670e-02 7.76889682e-01 4.97635961e-01
9.01611447e-01 1.16297610e-01 -7.33424723e-01 3.08868676e-01
8.91227305e-01 1.05301034e+00 -1.31250656e+00 1.31654799e-01
-1.69625610e-01 6.97376847e-01 8.69659409e-02 1.30264089e-01
-8.75811040e-01 -7.56936789e-01 6.44140065e-01 -8.92115384e-02
-7.25715011e-02 -7.71244168e-01 -1.07107475e-01 -2.19302207e-01
-1.76407218e-01 -4.86170948e-01 2.91935682e-01 -6.51162267e-01
-6.59160018e-01 6.91908658e-01 4.03630942e-01 -1.54844952e+00
-5.17380118e-01 -2.55923063e-01 -4.97655511e-01 4.04997081e-01
-1.46787226e+00 -1.01624131e+00 -4.86637890e-01 5.75014710e-01
7.22411096e-01 -1.70986772e-01 7.75926232e-01 -2.40075305e-01
-2.68385351e-01 -2.06050500e-01 2.31817335e-01 -6.16289973e-01
2.02945769e-01 -1.18743443e+00 5.88222779e-02 1.04013598e+00
-4.68790412e-01 4.74874169e-01 1.09836376e+00 -9.34123099e-01
-2.02010179e+00 -8.08437586e-01 5.75296342e-01 1.14834175e-01
6.72851384e-01 8.08999017e-02 -3.11020166e-01 6.35738969e-01
5.68045378e-01 -4.14603412e-01 -7.46973380e-02 -5.07693887e-01
6.62517369e-01 -2.24543866e-02 -1.15858984e+00 7.83932447e-01
1.07374704e+00 5.98399341e-01 -2.73692727e-01 3.50736350e-01
4.58230704e-01 -4.14732367e-01 -6.29327357e-01 4.29412723e-01
1.99239373e-01 -1.13656867e+00 6.93395615e-01 1.51509345e-01
-2.66336858e-01 -9.45112050e-01 2.67760418e-02 -1.64099205e+00
-2.48058900e-01 -1.14695430e+00 3.61065775e-01 8.71226847e-01
-2.75439443e-03 -1.26016247e+00 5.74410915e-01 1.57571316e-01
-6.74147010e-01 -5.48495352e-01 -1.32681918e+00 -1.34584773e+00
-3.25305551e-01 9.35293958e-02 6.45313799e-01 5.99461138e-01
4.32531357e-01 -1.37616590e-01 -1.18377224e-01 7.26306438e-01
7.59968579e-01 2.93066055e-01 1.05122113e+00 -1.17205048e+00
-7.97670931e-02 -4.38789248e-01 -2.16409355e-01 -7.96491444e-01
1.60190552e-01 -6.29764557e-01 4.28948641e-01 -2.16906500e+00
-3.60400975e-01 -8.84728789e-01 1.95635960e-01 4.37787503e-01
4.48794544e-01 -3.71987313e-01 4.54946518e-01 4.22428966e-01
-7.15849638e-01 4.15632904e-01 1.32930124e+00 1.27756402e-01
-6.28817081e-01 4.91488390e-02 -2.32160501e-02 4.67446417e-01
1.00162256e+00 -4.60894972e-01 -7.59703815e-01 -2.89101779e-01
8.16085488e-02 6.40289068e-01 1.80965856e-01 -1.10730731e+00
7.77829945e-01 -7.49706507e-01 -7.27590203e-01 -4.91880536e-01
3.58706772e-01 -1.18331671e+00 6.88620329e-01 1.11139464e+00
-9.82384011e-03 6.17820740e-01 -7.22501725e-02 8.15264463e-01
-5.43995388e-02 -4.83745784e-01 4.38386887e-01 -2.43077710e-01
-9.80132163e-01 -1.09952763e-01 -8.00457060e-01 -4.97952223e-01
1.93498576e+00 -4.26730454e-01 -7.47122288e-01 -3.54111642e-01
-4.11898494e-01 5.63069940e-01 6.33810163e-01 1.25428244e-01
5.33148825e-01 -9.69353378e-01 -4.82324600e-01 -2.53542453e-01
-1.09920323e-01 -1.25767547e-03 1.75641738e-02 1.27066195e+00
-9.94821906e-01 5.14601171e-01 -5.73251486e-01 -4.02974457e-01
-1.19147658e+00 6.19801044e-01 1.63667247e-01 -2.75788069e-01
-7.45483935e-01 3.17963034e-01 -1.30803734e-01 -3.37946475e-01
1.29195079e-01 -3.81337665e-02 -1.95339903e-01 -3.73845279e-01
2.44310021e-01 9.33273315e-01 -1.59940466e-01 -5.55043817e-01
-7.60028243e-01 4.72967446e-01 5.49500942e-01 -4.41708922e-01
1.41518867e+00 -5.08841634e-01 -4.33597982e-01 -7.62476400e-03
4.36494112e-01 2.36713141e-02 -1.12683892e+00 3.35128337e-01
1.79192930e-01 -3.97235543e-01 -3.51738751e-01 -6.31893039e-01
-6.12622619e-01 -9.80428606e-02 1.46483853e-01 5.79798102e-01
1.13832939e+00 -1.85864419e-01 4.66640890e-01 5.83097398e-01
1.37350190e+00 -9.69746053e-01 4.93069887e-02 6.22029543e-01
9.11524951e-01 -6.52301431e-01 -1.64327845e-02 -8.19908023e-01
-3.79084557e-01 1.26804769e+00 6.47106409e-01 -4.74597871e-01
1.12137385e-01 5.25401354e-01 -1.33240312e-01 -4.33531143e-02
-6.16751313e-01 -3.67607951e-01 -7.96836019e-01 9.73363698e-01
-5.40251434e-01 -6.94493651e-02 -1.03751826e+00 2.00071782e-01
-2.61514902e-01 -9.16980878e-02 1.12483644e+00 1.63290882e+00
-8.20118606e-01 -1.20429635e+00 -7.28626668e-01 -1.12726860e-01
3.23048025e-01 5.99676192e-01 -2.35316291e-01 1.21676981e+00
-6.11230358e-02 1.41212738e+00 -4.81729507e-02 -1.81929782e-01
4.06299919e-01 -5.01489878e-01 5.79239309e-01 -2.03622386e-01
-7.07683325e-01 -1.18143775e-01 3.63315314e-01 -5.70842385e-01
-7.13823259e-01 -8.23466599e-01 -1.79391241e+00 -2.83433527e-01
-2.67191648e-01 5.82174897e-01 8.48282516e-01 9.03436601e-01
3.68816257e-01 4.00467962e-01 7.97493458e-01 -1.10224128e+00
-4.17039663e-01 -5.68293750e-01 -4.77008373e-01 -2.80704916e-01
1.39404491e-01 -1.17814457e+00 -2.75744230e-01 -4.24243867e-01] | [4.942138195037842, 1.7074941396713257] |
c43cce28-ec23-48b4-872b-365b6461a5f6 | a-matter-of-annotation-an-empirical-study-on | 2305.08752 | null | https://arxiv.org/abs/2305.08752v1 | https://arxiv.org/pdf/2305.08752v1.pdf | A Matter of Annotation: An Empirical Study on In Situ and Self-Recall Activity Annotations from Wearable Sensors | Research into the detection of human activities from wearable sensors is a highly active field, benefiting numerous applications, from ambulatory monitoring of healthcare patients via fitness coaching to streamlining manual work processes. We present an empirical study that compares 4 different commonly used annotation methods utilized in user studies that focus on in-the-wild data. These methods can be grouped in user-driven, in situ annotations - which are performed before or during the activity is recorded - and recall methods - where participants annotate their data in hindsight at the end of the day. Our study illustrates that different labeling methodologies directly impact the annotations' quality, as well as the capabilities of a deep learning classifier trained with the data respectively. We noticed that in situ methods produce less but more precise labels than recall methods. Furthermore, we combined an activity diary with a visualization tool that enables the participant to inspect and label their activity data. Due to the introduction of such a tool were able to decrease missing annotations and increase the annotation consistency, and therefore the F1-score of the deep learning model by up to 8% (ranging between 82.1 and 90.4% F1-score). Furthermore, we discuss the advantages and disadvantages of the methods compared in our study, the biases they may could introduce and the consequences of their usage on human activity recognition studies and as well as possible solutions. | ['Kristof Van Laerhoven', 'Alexander Hoelzemann'] | 2023-05-15 | null | null | null | null | ['human-activity-recognition', 'human-activity-recognition'] | ['computer-vision', 'time-series'] | [ 3.39219958e-01 4.18986082e-01 -2.14999616e-01 -3.64092946e-01
-3.65811110e-01 -5.20414531e-01 3.32804680e-01 6.19035423e-01
-6.97210312e-01 7.41750419e-01 3.73993963e-01 1.78444311e-01
-2.01082230e-01 -5.03542900e-01 -3.64296764e-01 -5.97722471e-01
-4.70754318e-02 2.92510390e-01 7.47345909e-02 2.86324561e-01
-9.50499997e-02 2.75376081e-01 -1.68590760e+00 2.29197428e-01
3.07247221e-01 9.79030132e-01 -9.58391204e-02 4.27862972e-01
8.95654410e-03 5.59343219e-01 -1.04828060e+00 -2.65677590e-02
-2.75146798e-03 -3.24200690e-01 -4.73962724e-01 1.11473866e-01
3.38938743e-01 -9.23929214e-02 2.76668102e-01 5.30536354e-01
6.95541382e-01 6.75795823e-02 3.23094130e-01 -9.79115605e-01
-7.92837590e-02 2.76281357e-01 -1.45296201e-01 1.14060596e-01
8.25202525e-01 1.77809566e-01 3.35523456e-01 -3.12591553e-01
4.92625535e-01 5.63204646e-01 1.13571334e+00 5.12354493e-01
-1.24533737e+00 -5.75566173e-01 -1.97704554e-01 -2.01565892e-01
-1.33686686e+00 -4.40890968e-01 6.06385529e-01 -9.60465610e-01
8.28154385e-01 4.40597057e-01 8.94057512e-01 1.40796268e+00
-2.29120604e-03 3.39404345e-01 1.16998732e+00 -5.86997926e-01
4.52034235e-01 5.98502815e-01 3.70742679e-01 5.58921099e-01
6.80981815e-01 -9.47177783e-02 -6.82611346e-01 -1.80477321e-01
6.20984674e-01 2.20583558e-01 -3.22903752e-01 -2.90519744e-01
-1.20749199e+00 3.81882399e-01 3.01457010e-02 9.00973976e-01
-5.80696821e-01 9.88293812e-02 5.56383789e-01 -1.30946323e-01
4.64222044e-01 6.13052249e-01 -4.68608111e-01 -5.50757825e-01
-9.93678868e-01 1.36265010e-01 8.83678913e-01 5.94096899e-01
3.95584375e-01 -4.92353976e-01 -6.66143894e-01 5.43758452e-01
1.52578637e-01 4.13982719e-02 7.71124601e-01 -8.15292656e-01
1.39383003e-01 8.55722129e-01 5.32377899e-01 -9.31576490e-01
-9.05035496e-01 -4.97184128e-01 -5.37772059e-01 7.41362274e-02
7.51424968e-01 -4.02174145e-01 -4.73493993e-01 1.58578944e+00
3.23191136e-01 -2.15132117e-01 -5.27398109e-01 8.38866532e-01
6.00732386e-01 -9.80361104e-02 5.01218259e-01 -4.35707986e-01
1.53397810e+00 -7.67377198e-01 -1.24551296e+00 -1.44205257e-01
9.59049463e-01 -4.44588065e-01 1.38490355e+00 5.89659452e-01
-8.31499994e-01 -8.87703359e-01 -1.19444370e+00 1.51060775e-01
-5.99084735e-01 4.02075231e-01 3.88049185e-01 1.22334552e+00
-7.74608135e-01 1.01370668e+00 -1.13582790e+00 -6.99445546e-01
4.49765593e-01 4.56361443e-01 -3.18721950e-01 5.28477371e-01
-8.04340243e-01 9.30859745e-01 3.71199727e-01 4.00614040e-03
-4.63037580e-01 -4.71990913e-01 -5.92821598e-01 -5.89305572e-02
3.12597603e-01 -3.22397053e-01 1.35083318e+00 -9.44108129e-01
-1.34064531e+00 1.09750473e+00 4.23550978e-02 -3.57253671e-01
9.37967360e-01 -6.35971665e-01 -5.10392189e-01 -2.73420274e-01
-2.54935697e-02 2.38950402e-01 2.11715817e-01 -8.30476642e-01
-4.56961334e-01 -5.24005771e-01 -1.71966791e-01 -7.50530511e-02
-5.07266164e-01 -4.23017442e-01 -1.83187842e-01 -4.04637635e-01
-2.63027310e-01 -8.67403984e-01 1.86388880e-01 1.11265421e-01
-1.21547081e-01 -2.07109943e-01 7.59180367e-01 -7.72394478e-01
1.43745625e+00 -2.18016553e+00 -3.02565426e-01 1.59011096e-01
2.88433194e-01 4.99807417e-01 4.53100234e-01 4.94320601e-01
3.02162934e-02 4.11705524e-02 -1.03838123e-01 -5.74849069e-01
4.26255092e-02 3.03515404e-01 3.28574717e-01 5.47995687e-01
-3.38719279e-01 6.18960083e-01 -8.67666423e-01 -2.95339078e-01
4.86985415e-01 7.17928171e-01 6.08413368e-02 3.77631962e-01
8.03139359e-02 6.95305824e-01 1.87130198e-02 4.00340974e-01
5.67800887e-02 -4.08377610e-02 4.47772920e-01 -3.15155476e-01
-2.11627707e-01 3.93881410e-01 -1.15369976e+00 1.91866457e+00
-5.51189899e-01 6.93617404e-01 -3.57245564e-01 -9.52358663e-01
9.64848816e-01 5.33588588e-01 7.60710478e-01 -7.88221836e-01
1.64097458e-01 9.60019901e-02 -2.21850842e-01 -9.88897562e-01
2.73253918e-01 2.36004502e-01 8.68732631e-02 5.37170172e-01
-3.02413087e-02 7.11350560e-01 1.83741570e-01 -4.53108519e-01
1.17174649e+00 5.21699131e-01 6.72956705e-01 -3.05064887e-01
2.85747707e-01 -2.48964116e-01 2.55376369e-01 8.07952702e-01
-4.56148356e-01 3.87678057e-01 2.47551382e-01 -7.39615798e-01
-6.82360053e-01 -6.67020261e-01 6.50652684e-03 1.09946179e+00
-1.31870985e-01 -5.41654348e-01 -9.62480545e-01 -8.48025024e-01
-1.99967846e-01 4.95355994e-01 -8.88998091e-01 -1.52435824e-01
-4.07239467e-01 -5.21154702e-01 6.88259661e-01 7.48989046e-01
5.37404716e-01 -1.21926868e+00 -1.53243804e+00 1.80309594e-01
-3.00201952e-01 -8.27056944e-01 -1.02012455e-01 4.29428339e-01
-1.03554070e+00 -1.11934531e+00 -6.37131095e-01 -3.91529709e-01
3.85299951e-01 -2.56467521e-01 1.15769672e+00 -3.27266045e-02
-2.58679122e-01 5.63683510e-01 -3.99787664e-01 -6.69090092e-01
-1.43452004e-01 2.55560070e-01 3.76488492e-02 -2.61886208e-03
8.56587410e-01 -6.23389781e-01 -7.01331258e-01 2.20354453e-01
-5.40790737e-01 -2.65997738e-01 3.14486444e-01 2.54898906e-01
5.29800713e-01 -4.64652777e-01 2.62317330e-01 -1.12082422e+00
8.26172948e-01 -2.52577901e-01 -1.93264887e-01 1.28290847e-01
-9.82411623e-01 -2.26789080e-02 6.27485439e-02 -6.41523898e-01
-6.03506804e-01 3.41857105e-01 -2.92797368e-02 -2.08405145e-02
-6.08597755e-01 1.32062390e-01 -4.74340357e-02 1.54671535e-01
1.15540266e+00 -2.03514308e-01 -1.36290744e-01 -6.33238614e-01
-1.85370788e-01 8.15531373e-01 4.79720891e-01 -3.38073075e-01
8.62310156e-02 4.13112998e-01 -1.29776418e-01 -7.93140292e-01
-8.70842457e-01 -6.49967074e-01 -7.87731647e-01 -5.98827004e-01
1.00846088e+00 -6.25765324e-01 -8.96743417e-01 2.94411600e-01
-9.75095868e-01 -5.25820851e-01 -6.48902893e-01 4.60788965e-01
-4.19286013e-01 1.18613288e-01 -2.12095410e-01 -1.24905539e+00
-3.88094723e-01 -6.72477722e-01 1.15401244e+00 2.73202896e-01
-1.12338924e+00 -9.30424988e-01 2.57714301e-01 4.80367839e-01
4.46887076e-01 9.27223206e-01 3.14274728e-01 -7.69309282e-01
2.45550826e-01 -4.06584710e-01 2.41088733e-01 1.99788615e-01
4.16536540e-01 -6.56176448e-01 -1.45427978e+00 -1.16348214e-01
9.95458364e-02 -9.83956158e-02 1.79273039e-01 2.02440530e-01
1.14924181e+00 -2.55617648e-01 -5.84098697e-01 -1.90807681e-03
1.18324697e+00 3.19851935e-01 7.93270648e-01 3.84584844e-01
6.07575774e-01 7.01925278e-01 5.87154388e-01 3.84836644e-01
-1.82939574e-01 9.84084606e-01 2.08931610e-01 -2.14512765e-01
-1.11574799e-01 -1.39061525e-01 3.38785261e-01 3.78265798e-01
-4.76783842e-01 -3.11479904e-02 -7.30522692e-01 4.47461039e-01
-1.98139000e+00 -6.98167622e-01 -3.41943353e-01 2.59680438e+00
7.13386118e-01 1.95884615e-01 5.81900120e-01 7.19145298e-01
5.35987437e-01 -2.67408118e-02 -1.89224333e-01 -4.21055406e-01
5.40243983e-01 2.73654222e-01 5.94833970e-01 1.89326823e-01
-1.07952905e+00 4.97302003e-02 6.17992735e+00 1.13952294e-01
-9.80327308e-01 5.59506714e-01 3.18264574e-01 -3.26112211e-01
4.25203264e-01 -4.32231843e-01 -4.39375609e-01 7.69187391e-01
1.17426383e+00 5.76171577e-01 3.87106203e-02 9.18865740e-01
5.97360969e-01 -5.53604722e-01 -1.36349583e+00 1.19075096e+00
1.17252462e-01 -8.25969398e-01 -6.45252347e-01 1.58577979e-01
3.64656687e-01 -2.96618521e-01 -5.24440169e-01 1.72416583e-01
-5.83271563e-01 -7.39715695e-01 6.65393114e-01 9.86450851e-01
6.84751630e-01 -4.16866422e-01 9.23439443e-01 3.83017093e-01
-8.19307148e-01 1.03818066e-02 4.42028433e-01 -6.03139818e-01
1.32486001e-01 8.18420291e-01 -9.15824115e-01 1.04008615e-01
9.51016307e-01 3.92235428e-01 -6.19459271e-01 1.00397241e+00
-1.89874500e-01 5.91553092e-01 -2.64519691e-01 -1.93613932e-01
-3.41424793e-01 5.50376959e-02 8.01453367e-02 1.46357703e+00
3.14487606e-01 -3.23692083e-01 1.21677823e-01 7.23354518e-01
2.17366740e-01 -3.28508019e-02 -6.99137270e-01 -1.04123659e-01
2.66000807e-01 1.12684071e+00 -7.96009958e-01 -2.60057747e-01
-2.24210382e-01 1.01363444e+00 3.06141227e-02 3.00225206e-02
-8.23253572e-01 -4.48687613e-01 3.88928086e-01 9.03266907e-01
-1.51012868e-01 -1.71761796e-01 -4.75832731e-01 -6.14360631e-01
4.01885867e-01 -5.93566358e-01 3.15899283e-01 -7.14984715e-01
-8.21761131e-01 2.03936741e-01 8.76792818e-02 -1.01669335e+00
-1.75766870e-01 -4.10722435e-01 -2.25327283e-01 6.10702872e-01
-7.33553648e-01 -7.77635694e-01 -9.59033728e-01 2.92156190e-01
3.45011860e-01 3.00006419e-01 1.22411668e+00 7.72461951e-01
-4.76898521e-01 4.56209034e-01 -3.61621708e-01 8.87564123e-02
7.19007492e-01 -1.09188318e+00 -8.11301395e-02 4.19481009e-01
2.78627634e-01 5.46946049e-01 6.06825650e-01 -5.53031206e-01
-6.54846966e-01 -7.97717571e-01 1.17839932e+00 -8.40521872e-01
6.08883239e-02 -4.08344179e-01 -7.63474345e-01 6.59607291e-01
4.66462448e-02 -1.64038375e-01 1.01816666e+00 3.50063235e-01
7.57901818e-02 -2.36149371e-01 -1.28234935e+00 2.08653584e-01
1.22973156e+00 -5.33950627e-01 -4.45135057e-01 4.05456215e-01
1.57678738e-01 -5.46943665e-01 -1.00455701e+00 3.01731229e-01
1.08786607e+00 -1.15044463e+00 4.77152139e-01 -2.92212456e-01
-8.01742524e-02 -2.77587652e-01 2.64952064e-01 -9.47995186e-01
-2.14181826e-01 -3.61356169e-01 -3.79700452e-01 1.27357078e+00
3.06732744e-01 -4.72837031e-01 7.04106510e-01 7.79184580e-01
-3.66241448e-02 -4.79375839e-01 -5.58299243e-01 -5.83444536e-01
-8.59884381e-01 -5.12318313e-01 2.66086400e-01 1.06742275e+00
1.82792649e-01 1.86215043e-01 -3.74915451e-01 -2.45375380e-01
9.12694260e-02 -5.77402174e-01 7.43987441e-01 -1.64452708e+00
-3.90526623e-01 -1.00166142e-01 -4.55940098e-01 -3.18854183e-01
-5.33467948e-01 -4.59953189e-01 -7.31418002e-03 -1.46362710e+00
-1.18562020e-01 -1.68820664e-01 -3.98004740e-01 6.86698914e-01
2.55754143e-01 3.80506724e-01 -1.05144516e-01 2.40266979e-01
-5.96959770e-01 -2.32494563e-01 6.51424527e-01 8.56086537e-02
-6.12913489e-01 1.86458305e-01 -4.74930406e-01 7.43975163e-01
8.42396975e-01 -7.35466599e-01 -3.75392139e-01 -1.61929518e-01
3.39497685e-01 -3.89228404e-01 4.21919942e-01 -1.58466387e+00
-7.85167590e-02 3.14639866e-01 5.85278213e-01 -7.45632052e-02
3.18911940e-01 -1.14275324e+00 5.43166697e-01 7.37603009e-01
-4.70205486e-01 -2.95647327e-02 1.86159313e-01 5.16925991e-01
1.12224743e-01 -3.23669225e-01 3.77827346e-01 -3.27244788e-01
-3.72563064e-01 -3.77580941e-01 -4.18339908e-01 -2.88676858e-01
9.80746627e-01 -5.56547940e-01 -1.02759421e-01 -2.31987268e-01
-1.17628694e+00 -2.85099506e-01 1.62222579e-01 4.88442212e-01
-2.17392579e-01 -1.13527024e+00 1.39101148e-01 1.11240216e-01
3.86445075e-01 -1.69814199e-01 3.74845602e-02 1.19166064e+00
-4.24649328e-01 2.93595999e-01 -3.54373813e-01 -7.36542046e-01
-1.16618502e+00 3.00017864e-01 4.75199282e-01 -2.95559049e-01
-5.07705331e-01 2.93423235e-01 -5.20798206e-01 -1.41708449e-01
9.05414760e-01 -6.72906756e-01 -4.19755071e-01 3.77869308e-01
4.71173108e-01 8.62631440e-01 5.91174722e-01 -1.71944976e-01
-5.58457911e-01 3.50659907e-01 4.22136158e-01 -6.93534538e-02
9.41606998e-01 -3.50349098e-02 3.15962642e-01 1.04870141e+00
8.02644968e-01 3.80074754e-02 -9.92474854e-01 1.82674736e-01
3.84377122e-01 -1.49020419e-01 -1.36942089e-01 -1.23720205e+00
-7.18583405e-01 7.94775546e-01 1.44540167e+00 5.52956402e-01
1.00786757e+00 -1.63661420e-01 5.31972885e-01 2.62862980e-01
2.70544648e-01 -1.45174944e+00 -2.61151837e-03 -1.96409643e-01
5.12368381e-01 -1.11084068e+00 3.00070215e-02 -1.46379218e-01
-4.49562013e-01 8.91450584e-01 4.36729550e-01 1.78224027e-01
4.75505948e-01 1.48323700e-01 1.32296488e-01 -4.08996642e-01
-8.06661397e-02 -2.63470620e-01 2.54000574e-01 9.47662473e-01
9.93320942e-01 2.98031181e-01 -9.13770378e-01 8.14616442e-01
-6.91332594e-02 7.41453052e-01 2.49972299e-01 1.08789217e+00
-2.06578851e-01 -1.01897824e+00 -3.50114942e-01 4.60709363e-01
-5.39146423e-01 7.73652911e-01 -6.12566769e-01 8.63754809e-01
8.32137525e-01 1.05188930e+00 6.09848015e-02 -4.21422094e-01
8.06561530e-01 4.92867470e-01 4.65168357e-01 -7.37865448e-01
-1.03600907e+00 -1.65384591e-01 4.08642530e-01 -7.54657865e-01
-9.12525475e-01 -6.78279638e-01 -8.86643946e-01 2.17447624e-01
-2.32255623e-01 4.56282198e-02 9.34326410e-01 1.20748067e+00
4.43369001e-01 8.37415040e-01 -6.13946691e-02 -8.03385437e-01
-1.20049432e-01 -1.46888006e+00 -4.45997864e-01 5.62437534e-01
1.73369035e-01 -7.08838880e-01 -1.54609248e-01 3.71723086e-01] | [7.486680507659912, 0.8135167956352234] |
90a0a433-5d01-4337-9a39-fc47d043676e | topic-switch-adapted-japanese-dialogue-system | 2302.11280 | null | https://arxiv.org/abs/2302.11280v1 | https://arxiv.org/pdf/2302.11280v1.pdf | Topic-switch adapted Japanese Dialogue System based on PLATO-2 | Large-scale open-domain dialogue systems such as PLATO-2 have achieved state-of-the-art scores in both English and Chinese. However, little work explores whether such dialogue systems also work well in the Japanese language. In this work, we create a large-scale Japanese dialogue dataset, Dialogue-Graph, which contains 1.656 million dialogue data in a tree structure from News, TV subtitles, and Wikipedia corpus. Then, we train PLATO-2 using Dialogue-Graph to build a large-scale Japanese dialogue system, PLATO-JDS. In addition, to improve the PLATO-JDS in the topic switch issue, we introduce a topic-switch algorithm composed of a topic discriminator to switch to a new topic when user input differs from the previous topic. We evaluate the user experience by using our model with respect to four metrics, namely, coherence, informativeness, engagingness, and humanness. As a result, our proposed PLATO-JDS achieves an average score of 1.500 for the human evaluation with human-bot chat strategy, which is close to the maximum score of 2.000 and suggests the high-quality dialogue generation capability of PLATO-2 in Japanese. Furthermore, our proposed topic-switch algorithm achieves an average score of 1.767 and outperforms PLATO-JDS by 0.267, indicating its effectiveness in improving the user experience of our system. | ['Kazushi Ikeda', 'Gen Hattori', 'Kazunori Matsumoto', 'Yanan Wang', 'Jianming Wu', 'Donghuo Zeng'] | 2023-02-22 | null | null | null | null | ['dialogue-generation', 'dialogue-generation'] | ['natural-language-processing', 'speech'] | [-4.09413993e-01 4.01831388e-01 1.86109856e-01 -2.55587906e-01
-7.71652818e-01 -7.13409841e-01 7.31853783e-01 -1.61735684e-01
-3.28689665e-01 9.94009674e-01 5.22880375e-01 -2.14073405e-01
4.23950702e-01 -6.54496670e-01 2.56635398e-01 -4.13831055e-01
1.17744938e-01 6.91488087e-01 5.73010147e-01 -8.85594070e-01
2.50417471e-01 -4.09552991e-01 -7.74748325e-01 2.86143184e-01
1.45366538e+00 6.47115529e-01 4.29410428e-01 8.52654338e-01
-1.76695168e-01 6.97454393e-01 -1.28099358e+00 -4.67333943e-01
-2.53838301e-01 -8.32797408e-01 -1.27578998e+00 2.33474895e-02
-1.81132182e-01 -6.41859531e-01 -3.95389587e-01 6.70416236e-01
8.82575035e-01 4.18484658e-01 4.88672525e-01 -1.21564913e+00
-4.62446362e-01 8.07434678e-01 -8.60774219e-02 -1.36868238e-01
8.44757915e-01 4.28784400e-01 1.24638116e+00 -6.49313271e-01
7.68891394e-01 1.37454033e+00 4.14949685e-01 8.66208971e-01
-1.04583049e+00 -3.88981849e-01 -3.71733867e-02 -2.54556388e-01
-8.68064821e-01 -2.16878474e-01 5.38456500e-01 -3.64422679e-01
8.16631198e-01 4.12604779e-01 5.78301311e-01 1.30400336e+00
1.14258409e-01 1.02006662e+00 1.14929914e+00 -1.63300633e-01
1.10840768e-01 4.04804677e-01 2.98589677e-01 4.03388888e-01
-4.88625914e-01 -6.27444804e-01 -5.91129839e-01 -3.80520761e-01
5.93349099e-01 -5.48368573e-01 -3.92106444e-01 4.53535795e-01
-1.26715696e+00 1.10280061e+00 -1.97896850e-03 2.96319187e-01
-1.64404035e-01 -5.74669242e-01 7.26851702e-01 6.22338176e-01
5.84614813e-01 9.07321155e-01 -4.50558126e-01 -8.58651400e-01
-3.11871737e-01 4.97727484e-01 1.56707644e+00 1.10065961e+00
2.29016453e-01 -2.59726763e-01 -7.80766129e-01 1.39998424e+00
5.04772775e-02 3.95108044e-01 6.12308800e-01 -1.10779083e+00
5.52351475e-01 6.88085675e-01 2.12266922e-01 -7.15531886e-01
-4.86800253e-01 1.90579906e-01 -7.46692240e-01 -4.09745991e-01
5.93449414e-01 -8.86529207e-01 -9.95866656e-02 1.60786438e+00
2.66363144e-01 -7.73347795e-01 4.94473308e-01 7.56276011e-01
1.31856024e+00 1.13723052e+00 -2.25615695e-01 -2.28749812e-01
1.52309692e+00 -1.45769322e+00 -1.03775084e+00 -2.38436516e-02
8.10844243e-01 -8.05144787e-01 1.60759211e+00 4.80785072e-01
-8.65050256e-01 -3.39018643e-01 -6.43679559e-01 -3.81542146e-02
3.50856446e-02 1.66147739e-01 3.92379880e-01 6.78334475e-01
-8.89216900e-01 1.84628755e-01 -2.77968198e-01 -6.90712988e-01
-4.01485562e-01 -1.26959115e-01 -6.23345934e-02 4.00347769e-01
-1.73344111e+00 8.60976160e-01 3.47273856e-01 -2.37123385e-01
-6.85722053e-01 -3.03405702e-01 -5.18843949e-01 -2.22333483e-02
6.66690886e-01 -3.59792233e-01 1.90187299e+00 -3.28449368e-01
-2.36456037e+00 4.75497186e-01 2.14994207e-01 -1.77410871e-01
4.22777563e-01 -2.75444269e-01 -3.99579674e-01 1.58327684e-01
2.33586490e-01 6.05098307e-01 1.75630540e-01 -9.24121559e-01
-7.72371233e-01 4.63212840e-02 5.74227095e-01 7.61195421e-01
-3.98850530e-01 5.29670641e-02 -6.39372349e-01 -4.63805884e-01
-4.34207588e-01 -1.01971841e+00 -1.67877629e-01 -4.27809149e-01
-8.94414067e-01 -7.58251429e-01 6.89313650e-01 -7.30868042e-01
1.75170958e+00 -2.08471036e+00 5.42594157e-02 -2.43708685e-01
2.55677938e-01 2.43244916e-01 -1.47581361e-02 9.99753475e-01
7.88857639e-01 1.78184226e-01 -3.34751233e-03 -1.82724625e-01
9.46826711e-02 -1.25168666e-01 -1.37627069e-02 -2.71066308e-01
-4.04652171e-02 6.46089673e-01 -9.47336376e-01 -4.87460077e-01
-6.43636435e-02 -2.21100733e-01 -6.02331042e-01 7.75330901e-01
-5.13978004e-01 6.13348007e-01 -5.22916019e-01 2.18501821e-01
1.92246392e-01 -2.32957959e-01 3.30084413e-01 3.10264409e-01
-1.79724783e-01 6.87375903e-01 -6.40925884e-01 1.73447025e+00
-6.16436720e-01 7.77293563e-01 7.54024759e-02 -2.47996859e-02
1.18613100e+00 6.44465625e-01 2.36147180e-01 -5.75246572e-01
3.95106450e-02 9.95736197e-02 2.44286418e-01 -5.00094473e-01
1.11932683e+00 2.21128806e-01 -5.69937289e-01 7.06013441e-01
2.62637347e-01 -3.60031068e-01 3.85850817e-01 6.06898963e-01
1.09001446e+00 -4.37344283e-01 4.17896569e-01 -3.75998199e-01
4.55950052e-01 -3.01871765e-02 2.08000705e-01 7.84094393e-01
-3.39490861e-01 4.22554582e-01 1.15414882e+00 2.04798803e-01
-7.87328959e-01 -7.14520216e-01 8.60853121e-02 1.41511273e+00
1.29349470e-01 -6.91114962e-01 -1.20257103e+00 -8.38082254e-01
-4.12547857e-01 9.25563991e-01 -3.02334160e-01 2.40779035e-02
-3.48101288e-01 -5.64212620e-01 8.49686146e-01 -3.07874288e-02
1.03017342e+00 -1.28048825e+00 -3.36301357e-01 3.58440787e-01
-9.15221989e-01 -1.14438605e+00 -9.47802126e-01 -2.50488847e-01
-4.42736566e-01 -7.81525910e-01 -9.74162221e-01 -7.78229952e-01
1.78487211e-01 2.43206039e-01 1.05763280e+00 -9.13182944e-02
4.47440565e-01 3.43266241e-02 -7.38039196e-01 2.50081029e-02
-7.90882647e-01 6.29199266e-01 -1.76222593e-01 -3.94890964e-01
1.10251963e-01 -9.86879468e-02 -5.49695969e-01 8.46217990e-01
-4.83961552e-01 2.66647309e-01 2.20017657e-01 1.23111928e+00
-4.54747736e-01 -1.40887246e-01 8.55300784e-01 -1.05884099e+00
1.67803681e+00 -5.27368426e-01 -2.02260062e-01 2.46650517e-01
-4.73587811e-01 -1.69023350e-01 7.64196336e-01 -5.04164338e-01
-1.44191456e+00 -5.94059050e-01 -4.73467320e-01 5.73172569e-01
5.26978895e-02 6.55702174e-01 -2.12240964e-01 4.47890610e-01
7.14669645e-01 2.22260848e-01 1.47655740e-01 -4.63733554e-01
3.79196376e-01 1.35212457e+00 1.96815342e-01 -5.99105895e-01
2.24706128e-01 -3.08972180e-01 -9.77586329e-01 -1.14373004e+00
-6.28681242e-01 -5.34715652e-01 -1.80493891e-01 -3.63685161e-01
8.81203711e-01 -8.74560773e-01 -1.15747023e+00 8.70405138e-01
-1.43832266e+00 -7.77829349e-01 6.75051659e-02 9.59783643e-02
-3.45236242e-01 4.06596065e-01 -1.00306606e+00 -9.01409686e-01
-6.19865954e-01 -1.14130843e+00 8.17369521e-01 5.04214525e-01
-5.89897096e-01 -1.04872358e+00 2.53549516e-01 5.06749809e-01
3.81630272e-01 -2.75896937e-01 6.64875627e-01 -1.27279139e+00
1.37373274e-02 6.17683865e-02 -1.14874102e-01 2.68152207e-01
1.03735358e-01 -7.60784447e-02 -7.81815946e-01 -2.51828134e-01
3.82318944e-02 -6.98810697e-01 2.44981527e-01 -1.04348071e-01
4.78622079e-01 -5.87212980e-01 -2.52934489e-02 -1.93352580e-01
4.57241476e-01 6.56183779e-01 5.24266481e-01 1.53461188e-01
3.06594223e-01 7.84031272e-01 9.72700179e-01 7.94628799e-01
8.40656877e-01 8.96028578e-01 -2.00951353e-01 7.53272921e-02
2.16799214e-01 -3.43123049e-01 6.54356837e-01 1.40616930e+00
1.75573826e-01 -6.83450222e-01 -7.97959507e-01 3.00963312e-01
-1.85817599e+00 -6.43173397e-01 -9.71297473e-02 1.96048319e+00
1.48210251e+00 3.27203274e-01 5.77906668e-01 -4.36730444e-01
6.52927518e-01 3.57364863e-01 -2.57555872e-01 -6.05559230e-01
-1.42183888e-03 -4.11260039e-01 -2.01361641e-01 7.12799430e-01
-8.40166926e-01 1.19053352e+00 5.97098494e+00 1.05239308e+00
-8.41607630e-01 1.57684848e-01 7.32406437e-01 2.26383567e-01
-9.49303284e-02 -5.09262979e-02 -8.67277205e-01 7.51480818e-01
9.61520374e-01 -6.32179141e-01 3.05289477e-01 9.03559327e-01
2.25844190e-01 -3.61928821e-01 -8.56174171e-01 5.24271905e-01
-6.09588176e-02 -8.86160493e-01 -7.42891952e-02 1.20684132e-01
6.21749461e-01 -4.18950379e-01 -3.21287602e-01 8.38374555e-01
6.00158632e-01 -6.49930060e-01 1.93307132e-01 7.92551786e-02
6.36491001e-01 -4.12052095e-01 8.80026639e-01 6.42377794e-01
-8.01796854e-01 3.06711853e-01 -1.61438450e-01 -2.48604920e-03
5.07896185e-01 2.67238438e-01 -1.24765778e+00 4.17690039e-01
3.09016317e-01 2.50904709e-01 -2.58700699e-01 6.14990950e-01
-2.50762314e-01 7.50711322e-01 -2.37114966e-01 -7.30961859e-01
3.94616783e-01 -2.11335689e-01 7.01919734e-01 1.23163295e+00
1.08633228e-02 3.88709009e-01 4.52265054e-01 5.51416934e-01
-1.98041201e-01 3.23421687e-01 -4.82860953e-01 -8.53058025e-02
8.74281645e-01 1.25718689e+00 -2.67263621e-01 -4.10182983e-01
-2.56170064e-01 1.16876733e+00 1.35702014e-01 1.77257106e-01
-7.04743862e-01 -8.71665239e-01 4.10713106e-01 -4.15305406e-01
-1.64339140e-01 -2.01221034e-01 -6.81514442e-02 -9.68576789e-01
-1.25522345e-01 -1.27727127e+00 2.12000713e-01 -5.34534454e-01
-1.34634936e+00 1.16659343e+00 -1.05948159e-02 -1.06731927e+00
-5.20133674e-01 -2.42120162e-01 -8.86781633e-01 8.52704942e-01
-5.81258655e-01 -7.66357720e-01 -2.57176787e-01 3.70081395e-01
1.14108038e+00 -3.57514173e-01 9.65277016e-01 3.25599201e-02
-6.37653053e-01 7.14958131e-01 1.95898697e-01 2.68376827e-01
1.11587024e+00 -1.50518131e+00 7.09721565e-01 2.52453625e-01
-4.39686030e-01 6.52396679e-01 7.26109207e-01 -7.24356055e-01
-9.26841855e-01 -6.60172164e-01 9.85394180e-01 -3.18286479e-01
7.69620657e-01 -5.39391041e-01 -8.58030856e-01 2.54453272e-01
9.42660570e-01 -9.18375909e-01 7.92664230e-01 4.78899747e-01
7.21908137e-02 2.69972563e-01 -1.04025888e+00 9.20270383e-01
6.89530730e-01 -4.59554821e-01 -5.97196281e-01 4.79748815e-01
1.18827558e+00 -6.66254222e-01 -1.22788668e+00 -1.14031389e-01
4.71879035e-01 -8.63792777e-01 3.91025037e-01 -4.45664227e-01
5.53929985e-01 3.19285065e-01 1.13967299e-01 -1.73351276e+00
-1.64225008e-02 -1.34232485e+00 1.80808783e-01 1.50047719e+00
7.40599990e-01 -7.74884462e-01 4.33072656e-01 7.43799090e-01
-1.63469911e-01 -8.33563447e-01 -7.58991539e-01 -5.64408422e-01
1.21692203e-01 2.17975870e-01 4.06347305e-01 9.35501456e-01
7.81338513e-01 1.14009643e+00 -7.12997079e-01 -5.98427653e-01
-2.34692153e-02 -3.14894430e-02 1.13482130e+00 -1.07815289e+00
-3.19259733e-01 -4.04366046e-01 3.58946383e-01 -1.89519536e+00
-1.25311553e-01 -3.19341719e-01 3.62445444e-01 -1.44413912e+00
5.93757108e-02 -4.26822782e-01 5.09814918e-01 2.60939151e-01
-4.92391437e-01 -2.23285452e-01 3.58433753e-01 2.45467365e-01
-8.73678267e-01 1.03113866e+00 1.68005991e+00 4.17509563e-02
-7.29081094e-01 2.94407457e-01 -7.88175642e-01 5.18222332e-01
9.57218707e-01 -9.39005762e-02 -3.31129581e-01 -5.30714355e-02
-1.58505127e-01 6.73382342e-01 -4.37971920e-01 -5.35911500e-01
2.51433283e-01 -1.35457724e-01 -5.14710307e-01 -4.40004796e-01
4.44527119e-01 -2.73591578e-01 -3.66050214e-01 2.77075469e-01
-6.98523462e-01 4.17906269e-02 -7.76273608e-02 3.73314410e-01
-3.32584053e-01 -4.17272270e-01 4.90566432e-01 -1.64325386e-01
-5.85840821e-01 -1.14179112e-01 -1.01056862e+00 6.52374089e-01
7.84157097e-01 7.52709582e-02 -7.67149091e-01 -9.38490093e-01
-4.68110263e-01 8.06645632e-01 1.68067411e-01 5.52534103e-01
3.21637690e-01 -1.16645396e+00 -7.30745971e-01 -2.49682456e-01
2.45641559e-01 -1.58109799e-01 2.67726690e-01 6.88987434e-01
-5.09232581e-01 6.08784735e-01 -1.62225813e-01 -5.20049453e-01
-1.35216618e+00 -2.55550981e-01 4.43954282e-02 -6.45984948e-01
-4.50949907e-01 6.36631131e-01 1.61812946e-01 -9.66597855e-01
3.74518335e-01 -9.74873155e-02 -4.06260461e-01 9.77999195e-02
2.75965810e-01 4.79534984e-01 -3.56965005e-01 -3.28878611e-01
9.45552532e-03 -2.47871682e-01 -2.68972844e-01 -6.36386812e-01
6.85958207e-01 -3.55351061e-01 -2.00976893e-01 6.65172279e-01
8.86134326e-01 1.05699457e-01 -9.23611462e-01 -3.30670476e-01
-4.85217050e-02 -2.79326826e-01 -5.16246319e-01 -1.20376182e+00
-4.63230997e-01 6.66081786e-01 8.93190652e-02 7.57821381e-01
7.86878467e-01 -9.54599231e-02 1.05078626e+00 8.62873495e-01
3.20444286e-01 -1.32414329e+00 5.72713375e-01 1.23300958e+00
1.13487923e+00 -1.36850667e+00 -3.82445782e-01 -4.23439980e-01
-1.55121183e+00 8.58177006e-01 1.24063790e+00 3.79350007e-01
3.51077765e-01 -1.83231235e-01 3.63357902e-01 -7.38078281e-02
-1.12415528e+00 1.02368154e-01 1.13967970e-01 3.70619714e-01
8.18092227e-01 2.39582047e-01 -6.44062281e-01 8.13204110e-01
-4.96845067e-01 -3.45921129e-01 7.86938965e-01 6.62465453e-01
-5.63639402e-01 -1.08343649e+00 2.83003822e-02 3.88962269e-01
-2.09728807e-01 -1.76983729e-01 -8.60672951e-01 7.09434569e-01
-7.83698201e-01 1.45684063e+00 -2.43703127e-01 -6.85442686e-01
5.03204286e-01 1.07605085e-02 -1.66592121e-01 -9.89864409e-01
-1.16615665e+00 1.48640499e-01 8.78506899e-01 -2.32523635e-01
4.62860987e-02 -3.23282689e-01 -1.07945526e+00 -4.70732927e-01
-7.07585812e-01 7.62649894e-01 2.14678019e-01 6.42700076e-01
3.15926135e-01 3.89719754e-01 9.98047471e-01 -4.16419595e-01
-8.37628722e-01 -1.78450572e+00 -5.31963527e-01 1.82381243e-01
-1.45296842e-01 -3.78725082e-01 -9.69252884e-02 -3.88798475e-01] | [12.817649841308594, 8.081599235534668] |
fe0f4b3d-aadc-4edc-9367-3a299436e0a4 | all-in-one-exploring-unified-video-language | 2203.07303 | null | https://arxiv.org/abs/2203.07303v1 | https://arxiv.org/pdf/2203.07303v1.pdf | All in One: Exploring Unified Video-Language Pre-training | Mainstream Video-Language Pre-training models \cite{actbert,clipbert,violet} consist of three parts, a video encoder, a text encoder, and a video-text fusion Transformer. They pursue better performance via utilizing heavier unimodal encoders or multimodal fusion Transformers, resulting in increased parameters with lower efficiency in downstream tasks. In this work, we for the first time introduce an end-to-end video-language model, namely \textit{all-in-one Transformer}, that embeds raw video and textual signals into joint representations using a unified backbone architecture. We argue that the unique temporal information of video data turns out to be a key barrier hindering the design of a modality-agnostic Transformer. To overcome the challenge, we introduce a novel and effective token rolling operation to encode temporal representations from video clips in a non-parametric manner. The careful design enables the representation learning of both video-text multimodal inputs and unimodal inputs using a unified backbone model. Our pre-trained all-in-one Transformer is transferred to various downstream video-text tasks after fine-tuning, including text-video retrieval, video-question answering, multiple choice and visual commonsense reasoning. State-of-the-art performances with the minimal model FLOPs on nine datasets demonstrate the superiority of our method compared to the competitive counterparts. The code and pretrained model have been released in https://github.com/showlab/all-in-one. | ['Mike Zheng Shou', 'XiaoHu Qie', 'Ying Shan', 'Jianping Wu', 'Guanyu Cai', 'Xudong Lin', 'Yuying Ge', 'Rui Yan', 'Yixiao Ge', 'Alex Jinpeng Wang'] | 2022-03-14 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Wang_All_in_One_Exploring_Unified_Video-Language_Pre-Training_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Wang_All_in_One_Exploring_Unified_Video-Language_Pre-Training_CVPR_2023_paper.pdf | cvpr-2023-1 | ['visual-commonsense-reasoning'] | ['reasoning'] | [ 2.53802180e-01 -8.06209445e-02 -3.18481863e-01 -2.93981224e-01
-1.11820018e+00 -6.77962244e-01 8.40941191e-01 -3.31121624e-01
-4.59063292e-01 3.94065917e-01 3.98328871e-01 -3.76427323e-01
5.89921251e-02 -3.14466476e-01 -9.89183247e-01 -7.13348210e-01
2.06499055e-01 1.33273318e-01 1.28990903e-01 -1.52326792e-01
3.26665230e-02 -1.29200399e-01 -1.43671608e+00 8.50615621e-01
5.74311733e-01 1.12517536e+00 2.77774334e-01 7.67521560e-01
-1.95333421e-01 1.38709247e+00 -2.49123678e-01 -7.61008263e-01
8.23645294e-02 -4.52831298e-01 -7.98474967e-01 -4.99178506e-02
3.58569175e-01 -6.17252111e-01 -9.84815300e-01 6.83874249e-01
4.97546375e-01 2.52454996e-01 5.52829385e-01 -1.38662195e+00
-9.80784059e-01 7.59008706e-01 -5.33864498e-01 2.83104062e-01
4.05400336e-01 4.23514307e-01 1.13948691e+00 -1.03390217e+00
6.52797878e-01 1.27250302e+00 4.41181093e-01 5.29539704e-01
-9.17381108e-01 -6.19686306e-01 3.26226383e-01 5.54881215e-01
-1.35855663e+00 -6.19852662e-01 6.86629772e-01 -4.58381087e-01
1.11270630e+00 1.86043948e-01 3.93329710e-01 1.68817341e+00
1.01564765e-01 1.23574591e+00 7.01716304e-01 -1.70383975e-01
-1.30077407e-01 -9.92912147e-03 -3.50939661e-01 8.36659729e-01
-3.10157716e-01 -1.49413913e-01 -7.42016196e-01 1.38223246e-01
7.18514323e-01 3.23435932e-01 -3.22741926e-01 -1.61565498e-01
-1.37924922e+00 8.27595413e-01 3.02340418e-01 3.53284508e-01
-3.11077982e-01 5.36971033e-01 8.24424922e-01 3.97748739e-01
4.25923526e-01 -2.55226269e-02 -3.67954373e-01 -4.00496036e-01
-1.10016632e+00 6.37791008e-02 4.95122343e-01 1.07803822e+00
4.71308917e-01 6.46307468e-02 -5.67770422e-01 6.99324548e-01
2.76364267e-01 5.29899538e-01 5.84903240e-01 -1.01605606e+00
8.34672689e-01 3.28208476e-01 -3.87636960e-01 -6.93527400e-01
-7.71820620e-02 -6.33305460e-02 -7.00380147e-01 -3.60526770e-01
2.19133899e-01 -9.23073441e-02 -1.13244128e+00 1.81411099e+00
3.55344191e-02 3.56615812e-01 -8.49267526e-04 1.01450050e+00
8.32597196e-01 8.04816127e-01 2.07806319e-01 2.43650768e-02
1.46105480e+00 -1.18872166e+00 -7.31168628e-01 -1.48116136e-02
8.09614897e-01 -7.42664754e-01 1.15065455e+00 2.70499259e-01
-1.11928725e+00 -4.47142243e-01 -8.72971296e-01 -6.37084782e-01
-3.60160232e-01 2.77628392e-01 3.97689670e-01 1.61467135e-01
-9.24357831e-01 3.12031358e-01 -7.85433412e-01 -3.24054182e-01
4.07866508e-01 1.19374208e-01 -4.81623769e-01 -3.41241121e-01
-1.26469409e+00 7.54526615e-01 4.16650444e-01 1.57635584e-01
-1.11096919e+00 -6.22391939e-01 -1.08565891e+00 -3.34950760e-02
7.05360353e-01 -9.64293420e-01 1.25387740e+00 -1.17139184e+00
-1.61242080e+00 8.13714743e-01 -2.57030398e-01 -4.84037399e-01
6.20121539e-01 -3.09589058e-01 -2.76398689e-01 6.23741865e-01
1.10414267e-01 8.06967139e-01 1.22936714e+00 -9.49188232e-01
-4.93947417e-01 -8.43703449e-02 3.94647419e-01 1.99295402e-01
-4.84757483e-01 6.92582801e-02 -9.27311361e-01 -9.38813090e-01
-4.04936999e-01 -8.53555441e-01 2.60190785e-01 -5.17298356e-02
-3.97290498e-01 -2.96880871e-01 8.73192608e-01 -7.81817913e-01
1.32749259e+00 -2.35998130e+00 5.22993088e-01 -2.87157089e-01
2.30624467e-01 7.33554512e-02 -4.47604150e-01 6.77127838e-01
-1.60385385e-01 1.15482152e-01 -4.21867482e-02 -6.92287087e-01
2.33321995e-01 2.26408958e-01 -3.98889691e-01 4.86700594e-01
4.28421527e-01 1.13709629e+00 -8.63657475e-01 -7.17884243e-01
3.06044787e-01 7.90234983e-01 -6.42634213e-01 2.10461855e-01
-3.44600320e-01 2.06698164e-01 -4.83910203e-01 8.73541117e-01
2.66777128e-01 -4.21437383e-01 4.82332334e-02 -4.61078674e-01
4.55097221e-02 2.49457195e-01 -7.11759746e-01 2.18470454e+00
-2.94831187e-01 7.48599648e-01 -2.10322603e-03 -1.17238760e+00
3.03426862e-01 6.73254550e-01 5.81700325e-01 -9.45445180e-01
2.57825881e-01 6.27983287e-02 -4.31720465e-01 -9.51498687e-01
5.44995368e-01 -1.67086273e-01 -2.21814334e-01 3.22445184e-01
4.87682253e-01 1.70091331e-01 3.34669054e-01 5.51272392e-01
9.83672976e-01 5.82219183e-01 -1.14085205e-01 2.47806504e-01
4.33582038e-01 -1.24290451e-01 1.70233116e-01 5.96437395e-01
-9.79656801e-02 8.12485039e-01 6.39744759e-01 -1.08108632e-01
-9.65967238e-01 -9.18991566e-01 1.77273363e-01 1.57917762e+00
-7.69348629e-03 -7.40096033e-01 -4.19939697e-01 -6.17852628e-01
-4.94479276e-02 6.18427277e-01 -7.33450592e-01 -2.61325747e-01
-5.15471816e-01 -2.80763447e-01 9.41508412e-01 6.29274368e-01
2.79516399e-01 -7.78916001e-01 -4.63149965e-01 7.19614998e-02
-5.82481563e-01 -1.60229862e+00 -6.80552602e-01 6.06121421e-02
-6.27646029e-01 -8.83612037e-01 -8.63635540e-01 -5.47953904e-01
3.48792523e-01 4.31801081e-01 7.20642805e-01 -6.60597160e-02
-1.40240556e-02 7.34298944e-01 -7.03200877e-01 -2.94323377e-02
-5.70158251e-02 -1.35545164e-01 -1.89576358e-01 2.63266444e-01
1.54561728e-01 -4.96341228e-01 -6.57541454e-01 1.44513562e-01
-1.17174399e+00 3.01487446e-01 6.22248173e-01 8.94427776e-01
4.75826442e-01 -2.80628979e-01 1.78280860e-01 -4.74997371e-01
3.68952632e-01 -7.85635054e-01 -1.22115090e-01 3.49654913e-01
-6.49466962e-02 2.08177149e-01 7.13164628e-01 -6.13991857e-01
-9.31021094e-01 -8.16629156e-02 2.92732641e-02 -1.25492525e+00
1.65924504e-01 7.40534425e-01 -6.67718202e-02 2.90639907e-01
3.17808747e-01 3.94509256e-01 1.27791180e-04 -4.09178168e-01
7.31528580e-01 3.69603872e-01 6.73156857e-01 -7.58378685e-01
7.50197232e-01 5.36342621e-01 -2.30350122e-01 -7.02593565e-01
-7.45591044e-01 -5.30981064e-01 -4.03115928e-01 -2.85455048e-01
1.02129495e+00 -1.41078126e+00 -7.11370230e-01 3.26416641e-01
-1.16806698e+00 -4.47589844e-01 -1.15204968e-01 5.21163404e-01
-6.12642407e-01 5.51912069e-01 -8.00559938e-01 -5.87382257e-01
-3.59497964e-01 -1.17845821e+00 1.29370749e+00 -4.86670397e-02
1.16160534e-01 -7.56218612e-01 -3.21388245e-01 8.12415302e-01
3.95518482e-01 1.06430106e-01 7.04333246e-01 -6.26663804e-01
-7.10914910e-01 -1.07200600e-01 -4.29057032e-01 3.29577208e-01
-2.07506746e-01 1.15096189e-01 -1.07001197e+00 -2.86954701e-01
-1.27759233e-01 -6.53611302e-01 1.27697456e+00 2.86094338e-01
1.27473962e+00 -3.40529829e-01 -3.60124186e-02 8.47050071e-01
1.33896804e+00 1.16260315e-03 7.28980184e-01 3.39406967e-01
9.58539844e-01 3.20192903e-01 3.87426645e-01 5.38166404e-01
7.21315265e-01 6.02629244e-01 3.98488015e-01 7.48263150e-02
-5.77648543e-02 -2.55334139e-01 9.22815382e-01 7.55443037e-01
-1.95297092e-01 -4.64115471e-01 -7.31169760e-01 5.46924412e-01
-2.09381270e+00 -1.26405239e+00 1.17241338e-01 1.73054349e+00
7.89591193e-01 -7.57391378e-02 2.09247306e-01 -2.12085955e-02
4.36507970e-01 3.33721936e-01 -4.65822995e-01 -5.52813858e-02
-1.17906809e-01 -1.46967927e-02 2.58217931e-01 2.91973799e-01
-1.06883168e+00 1.00145841e+00 4.93159342e+00 1.02662706e+00
-1.36489367e+00 3.61896843e-01 3.86292905e-01 -4.95868951e-01
-4.69051033e-01 6.24752603e-02 -5.01394033e-01 5.41624367e-01
1.16166890e+00 -2.13257037e-02 4.65392917e-01 4.99960601e-01
2.05415711e-01 6.20686114e-02 -1.15686274e+00 1.15717173e+00
3.12433153e-01 -1.46693063e+00 3.14190656e-01 -1.26641810e-01
2.78700709e-01 2.18299106e-01 1.09020159e-01 5.86280882e-01
-1.82236061e-01 -9.19188261e-01 1.18669176e+00 5.04914463e-01
1.06675458e+00 -4.03049231e-01 4.62051183e-01 4.06042226e-02
-1.33002579e+00 -1.88780025e-01 1.41086802e-02 1.28239259e-01
3.29425484e-01 3.76240164e-01 -3.56648773e-01 8.61140192e-01
7.11752117e-01 9.17075455e-01 -4.04057026e-01 6.41997814e-01
-3.93810533e-02 5.21220565e-01 -2.33470231e-01 3.51291269e-01
5.20690501e-01 8.00421685e-02 4.68797445e-01 1.60856295e+00
3.44474077e-01 1.01328805e-01 1.17253400e-01 8.05308938e-01
-2.62145102e-01 -9.95671377e-02 -6.22724295e-01 -5.38175762e-01
2.73420811e-01 1.12776637e+00 -4.07198071e-01 -3.80970508e-01
-6.84890568e-01 1.19219816e+00 1.21673964e-01 5.74721277e-01
-1.33134091e+00 -1.98351309e-01 3.79820317e-01 -9.89020243e-02
6.07227325e-01 -3.06043833e-01 1.44278944e-01 -1.61733770e+00
1.93646789e-01 -9.22867298e-01 5.53469419e-01 -8.52287948e-01
-1.18909967e+00 5.32872975e-01 2.49218926e-01 -1.26515925e+00
-3.81259143e-01 -7.43464410e-01 -4.39136237e-01 5.17878711e-01
-1.66281939e+00 -1.59648108e+00 -2.84293979e-01 1.18847775e+00
6.95968926e-01 -9.33424681e-02 4.81376350e-01 5.53659499e-01
-7.50860989e-01 6.66398406e-01 -1.89499706e-02 2.47914895e-01
8.35876584e-01 -9.44248021e-01 1.29811577e-02 8.65650892e-01
-7.72738531e-02 5.08006394e-01 3.88547957e-01 -3.27460498e-01
-1.94969940e+00 -8.57515931e-01 6.73063993e-01 -4.21534091e-01
1.02470767e+00 -4.53646839e-01 -7.95478225e-01 9.39370453e-01
4.74354178e-01 8.82274192e-03 6.32164598e-01 -2.06666276e-01
-6.78624690e-01 4.86207977e-02 -6.18969381e-01 5.70044935e-01
9.35598016e-01 -1.13148248e+00 -6.11214817e-01 2.81366676e-01
9.97271180e-01 -3.92185211e-01 -8.69389415e-01 2.47569531e-01
6.22405946e-01 -7.74018466e-01 1.05937696e+00 -6.30528748e-01
9.66967046e-01 -1.67011321e-01 -4.09958929e-01 -7.24213481e-01
-1.68431982e-01 -8.88492227e-01 -4.43549693e-01 1.28617728e+00
3.23194832e-01 -2.50961870e-01 3.28228682e-01 3.83323431e-01
-2.98760891e-01 -7.56744444e-01 -9.58959937e-01 -4.82852757e-01
9.42723267e-03 -6.51383936e-01 5.55731170e-02 1.01626515e+00
1.34580821e-01 5.44148386e-01 -7.30354786e-01 -9.83902514e-02
3.67667407e-01 -8.74926448e-02 5.52408159e-01 -6.05304062e-01
-3.75944525e-01 -5.25609612e-01 -2.00322896e-01 -1.27287710e+00
2.69435406e-01 -9.81211066e-01 -5.08624874e-02 -1.40113270e+00
3.34358573e-01 1.20945886e-01 -4.15380418e-01 7.72553384e-01
-7.16107264e-02 1.19329795e-01 4.88578588e-01 2.31181875e-01
-8.65918398e-01 7.53294051e-01 1.28010082e+00 -2.73750603e-01
1.33182168e-01 -5.48175752e-01 -6.34993613e-01 4.01814252e-01
4.38760817e-01 -2.63591588e-01 -5.64546645e-01 -8.41318786e-01
2.10318163e-01 3.20869476e-01 6.18855834e-01 -6.08555734e-01
3.28232288e-01 -1.00060426e-01 2.32749179e-01 -4.71591473e-01
7.35573053e-01 -7.19065070e-01 -1.23711340e-01 2.95075178e-02
-4.17100817e-01 1.87811568e-01 2.77496815e-01 6.01014614e-01
-3.61741334e-01 6.84865564e-02 4.21424270e-01 -6.15899675e-02
-8.98140132e-01 3.42008531e-01 -2.49783695e-01 9.64867026e-02
8.95429194e-01 -8.73198211e-02 -5.96172750e-01 -6.07223332e-01
-6.02072537e-01 3.71096671e-01 1.30420297e-01 7.75878310e-01
7.33945370e-01 -1.26093781e+00 -7.83604801e-01 -3.56707051e-02
1.47758111e-01 -2.41512597e-01 5.73768139e-01 1.43101954e+00
-2.83536285e-01 6.34171367e-01 -5.74732125e-02 -6.32950485e-01
-9.89149868e-01 6.71625495e-01 1.97726622e-01 -1.68569401e-01
-8.15165758e-01 8.32552016e-01 3.85995775e-01 1.85231958e-02
3.89405608e-01 -4.40471828e-01 1.38120100e-01 2.00796440e-01
5.59501112e-01 8.17400664e-02 -1.65010542e-01 -6.44362867e-01
-3.66608113e-01 4.37658757e-01 -1.92772374e-01 -1.87092453e-01
1.26503050e+00 -2.59026229e-01 5.54513000e-02 4.78218496e-01
1.56839132e+00 -4.25513059e-01 -1.27965605e+00 -3.16428721e-01
-3.21928293e-01 -3.20506871e-01 1.50095984e-01 -6.20386600e-01
-1.19869959e+00 9.41215873e-01 2.29159698e-01 -3.48061211e-02
1.21931899e+00 1.41508639e-01 1.05653596e+00 3.17486286e-01
3.23815644e-02 -1.01543176e+00 3.53063852e-01 5.32401204e-01
9.68736589e-01 -1.22576606e+00 -2.36505195e-01 2.39298039e-04
-9.88375962e-01 1.16730952e+00 4.19326544e-01 1.10318974e-01
4.63475108e-01 1.35875046e-01 -2.49607012e-01 -1.94447756e-01
-1.20686090e+00 -2.77901798e-01 4.55989420e-01 2.04917982e-01
4.50629324e-01 -1.61115527e-01 -5.60416989e-02 6.59961700e-01
1.77735239e-01 3.78595114e-01 1.73428208e-01 1.04605591e+00
-1.35796620e-02 -8.75012100e-01 -1.62653670e-01 2.02914387e-01
-5.33752859e-01 -2.58567393e-01 -7.56476745e-02 9.21869338e-01
1.93204563e-02 8.20350587e-01 -1.07252868e-02 -6.25080645e-01
1.78937823e-01 2.72932321e-01 6.37058616e-01 -1.38601974e-01
-5.58616281e-01 3.46171767e-01 1.43982932e-01 -8.21342289e-01
-5.89592695e-01 -5.17775178e-01 -1.31290925e+00 -4.51741010e-01
6.03769347e-03 -1.58790872e-01 4.83866453e-01 1.13475263e+00
4.79436815e-01 5.76810956e-01 3.71424258e-01 -1.07956231e+00
-3.95797491e-01 -8.13475847e-01 -1.73159406e-01 5.20891786e-01
5.32128870e-01 -6.07618332e-01 -4.02799338e-01 3.41775447e-01] | [10.24433708190918, 0.9402483701705933] |
bf718834-7639-4f47-9abf-719a72d00483 | hybrid-life-integrating-biological-artificial | 2212.00285 | null | https://arxiv.org/abs/2212.00285v1 | https://arxiv.org/pdf/2212.00285v1.pdf | Hybrid Life: Integrating Biological, Artificial, and Cognitive Systems | Artificial life is a research field studying what processes and properties define life, based on a multidisciplinary approach spanning the physical, natural and computational sciences. Artificial life aims to foster a comprehensive study of life beyond "life as we know it" and towards "life as it could be", with theoretical, synthetic and empirical models of the fundamental properties of living systems. While still a relatively young field, artificial life has flourished as an environment for researchers with different backgrounds, welcoming ideas and contributions from a wide range of subjects. Hybrid Life is an attempt to bring attention to some of the most recent developments within the artificial life community, rooted in more traditional artificial life studies but looking at new challenges emerging from interactions with other fields. In particular, Hybrid Life focuses on three complementary themes: 1) theories of systems and agents, 2) hybrid augmentation, with augmented architectures combining living and artificial systems, and 3) hybrid interactions among artificial and biological systems. After discussing some of the major sources of inspiration for these themes, we will focus on an overview of the works that appeared in Hybrid Life special sessions, hosted by the annual Artificial Life Conference between 2018 and 2022. | ['Keisuke Suzuki', 'Lana Sinapayen', 'Olaf Witkowski', 'Hiroyuki Iizuka', 'Manuel Baltieri'] | 2022-12-01 | null | null | null | null | ['artificial-life'] | ['miscellaneous'] | [-3.74325621e-03 2.48415500e-01 2.92060729e-02 5.99175453e-01
4.96038586e-01 -7.58694589e-01 1.06500375e+00 8.09238702e-02
1.75668932e-02 9.06406522e-01 9.28752944e-02 -4.50543202e-02
5.22528216e-02 -9.79165375e-01 -3.47388625e-01 -1.00527728e+00
-3.37740779e-01 2.48266816e-01 -1.66437790e-01 -8.63889456e-01
1.81700930e-01 4.61090535e-01 -2.26381445e+00 -3.11244994e-01
9.77914333e-01 4.69587296e-01 1.05128460e-01 7.90346980e-01
3.63265723e-02 7.23369777e-01 -6.30193472e-01 -1.62026614e-01
3.19265085e-03 -1.01602840e+00 -6.16188228e-01 -1.99973062e-01
-5.12610435e-01 5.01882195e-01 -3.01114649e-01 6.81784451e-01
4.34323043e-01 -1.48258552e-01 6.37820899e-01 -1.44592476e+00
-9.80252802e-01 3.75205785e-01 1.82115287e-01 -1.13474175e-01
5.04785240e-01 7.38695204e-01 5.48151135e-01 -7.55173922e-01
6.33594871e-01 1.20826530e+00 8.25931132e-01 9.93091702e-01
-1.15877354e+00 -1.36595726e-01 -7.56828487e-02 4.76608388e-02
-1.06221652e+00 -6.17383361e-01 1.78693116e-01 -7.53964007e-01
1.25405049e+00 3.63129795e-01 1.44731164e+00 1.11372316e+00
7.79528081e-01 5.80539048e-01 1.12487233e+00 -4.91834849e-01
6.48080766e-01 3.34287703e-01 2.09975198e-01 4.18884337e-01
7.75442123e-01 5.79075694e-01 -5.03961563e-01 -1.03943326e-01
6.07908726e-01 -3.62287670e-01 -2.86807895e-01 -8.65945220e-02
-1.61388242e+00 3.46385598e-01 -1.91321708e-02 8.48386109e-01
-5.64402699e-01 4.37848985e-01 1.30583048e-01 2.32311949e-01
2.55744994e-01 7.09070981e-01 -5.40244937e-01 -1.02768637e-01
-1.01347171e-01 1.53007925e-01 1.43484092e+00 8.88874292e-01
6.46234035e-01 2.05823973e-01 2.17521936e-01 5.54225981e-01
5.78158319e-01 6.02544904e-01 4.93650317e-01 -9.72658336e-01
-8.16590965e-01 7.39761114e-01 1.34278769e-02 -4.22847599e-01
-4.54124659e-01 -4.20603633e-01 -1.01834953e+00 3.65550637e-01
2.07258105e-01 -1.04307212e-01 -7.40956485e-01 1.71788943e+00
4.86119717e-01 -4.91787344e-02 5.50148845e-01 3.67143542e-01
1.03572369e+00 9.21357453e-01 -1.37192162e-03 -4.53981996e-01
1.47831702e+00 -1.06756246e+00 -7.26335287e-01 -1.20857053e-01
4.27097470e-01 -3.91889751e-01 6.93986177e-01 4.05148834e-01
-1.22828114e+00 -1.64031461e-01 -1.24799657e+00 1.18694015e-01
-1.11171114e+00 -6.09140396e-01 8.60679805e-01 8.97815824e-01
-1.51623845e+00 4.71730232e-01 -8.17426741e-01 -1.34604359e+00
-1.04112029e-01 5.53266481e-02 -9.10742506e-02 3.10312420e-01
-1.23862696e+00 1.43740666e+00 2.93664217e-01 -2.30126157e-01
-1.03996301e+00 -6.93521976e-01 -5.78335702e-01 -3.77668858e-01
1.57753676e-01 -1.44861090e+00 7.21978605e-01 -7.39132226e-01
-1.62999153e+00 1.19060636e+00 -5.15360292e-03 -6.01610363e-01
2.60061532e-01 -3.30322534e-02 -4.32353675e-01 -1.89747930e-01
-2.00872645e-01 4.28548992e-01 -6.30724728e-02 -1.56401908e+00
-3.34300280e-01 -3.13009679e-01 8.16176459e-02 -1.43906653e-01
-2.72620678e-01 -1.50977403e-01 2.24385713e-03 -6.63537025e-01
-2.38602594e-01 -1.05031824e+00 -4.38575864e-01 2.54265964e-01
4.41745035e-02 -3.44889909e-01 7.21973300e-01 -1.44533023e-01
1.02640426e+00 -1.67500007e+00 5.11025190e-01 -4.77392107e-01
4.44922835e-01 3.24358821e-01 -2.14698212e-03 1.31940556e+00
2.92291224e-01 3.67156148e-01 -4.98226404e-01 6.86266348e-02
-8.64046961e-02 3.13070238e-01 4.89658639e-02 4.28369492e-01
-5.72598986e-02 1.11104453e+00 -1.22263718e+00 -2.05984354e-01
4.61382687e-01 8.13375831e-01 -2.80818734e-02 2.02112138e-01
-3.96705389e-01 6.42498970e-01 1.03505896e-02 9.16409492e-01
2.08262727e-01 -2.93892443e-01 -6.96668029e-02 1.83300182e-01
-4.16830033e-01 -1.96823835e-01 -6.10483825e-01 1.55206871e+00
-2.97952056e-01 6.03912771e-01 2.99148828e-01 -8.39969218e-01
8.72383177e-01 6.71401858e-01 7.26841629e-01 -4.86186296e-01
2.21960410e-01 2.89015979e-01 9.69316512e-02 -6.96451068e-01
2.66223773e-02 -2.18856424e-01 1.93766318e-02 4.99435276e-01
1.44729450e-01 -5.40716112e-01 5.29131472e-01 -2.81074252e-02
1.35120094e+00 7.13953912e-01 6.61486268e-01 -8.53351355e-01
9.70371306e-01 -4.44219932e-02 5.96150935e-01 5.15199721e-01
-7.07075536e-01 3.06918435e-02 -7.28147924e-02 -7.95343101e-01
-1.03012753e+00 -1.12942290e+00 -1.82843432e-01 8.79034579e-01
6.22892737e-01 -5.52770257e-01 -1.15418875e+00 5.85744455e-02
-1.01855189e-01 6.72702909e-01 -9.66078937e-01 -3.23135436e-01
-5.38918197e-01 -1.07065201e+00 5.28217196e-01 -1.97944138e-02
6.11534297e-01 -1.60756648e+00 -8.80197883e-01 2.71071225e-01
-1.30133018e-01 -7.66080678e-01 3.70508105e-01 2.07412854e-01
-6.60790622e-01 -8.56652796e-01 -6.38709426e-01 -7.83536613e-01
2.03074098e-01 1.95068911e-01 1.49080873e+00 6.60254180e-01
-6.76789701e-01 6.76314771e-01 -4.41813886e-01 -7.27150559e-01
-9.60866511e-01 -5.36363535e-02 2.93845087e-01 -4.66702789e-01
1.61521211e-01 -8.02748859e-01 -6.26452208e-01 1.91460446e-01
-1.00084043e+00 2.68331379e-01 3.78063560e-01 7.48702943e-01
6.05324507e-02 -1.86922252e-01 1.08275521e+00 -2.89635032e-01
4.22030687e-01 -8.11798513e-01 -4.29073982e-02 5.71893394e-01
-7.49703944e-01 -1.42637178e-01 6.92429364e-01 -1.61545902e-01
-7.07670808e-01 -2.73672730e-01 -1.19999051e-01 4.96363997e-01
-2.33007938e-01 1.96871832e-01 -4.41167027e-01 7.76242791e-03
7.27334023e-01 5.82823634e-01 4.04770166e-01 -8.78123268e-02
4.42441821e-01 5.16953170e-01 3.95689040e-01 -6.39926493e-01
5.58976412e-01 7.24201143e-01 8.06473047e-02 -1.29294920e+00
-3.03767741e-01 8.10582191e-02 -6.15777194e-01 -6.68155313e-01
8.36739302e-01 -3.37295145e-01 -7.42675126e-01 8.47795784e-01
-1.07387114e+00 -3.82673591e-01 -8.54043424e-01 1.74670383e-01
-8.86888623e-01 1.28442273e-01 -7.37188458e-01 -1.22252035e+00
-6.11942828e-01 -7.82986701e-01 8.79747570e-01 5.61622798e-01
-5.10538042e-01 -1.54304624e+00 6.75016105e-01 1.58827171e-01
7.25133002e-01 6.88422561e-01 6.48918867e-01 -2.58600980e-01
-3.81894559e-01 1.83634385e-02 4.06748295e-01 4.01678775e-03
4.06765670e-01 5.53995609e-01 -9.97967005e-01 -3.30170691e-01
1.63451642e-01 -4.83550541e-02 5.89266241e-01 3.15282531e-02
1.90086827e-01 -1.50855675e-01 -7.51264930e-01 2.79055357e-01
1.40634155e+00 5.61397314e-01 8.94239485e-01 5.41512489e-01
1.73742309e-01 1.15101898e+00 3.85365158e-01 5.83825350e-01
4.66254115e-01 6.85451180e-02 5.98232687e-01 4.36414219e-03
-2.50737339e-01 1.34332463e-01 4.31225777e-01 1.17787445e+00
-1.53553143e-01 -5.69456577e-01 -1.17818069e+00 7.07719564e-01
-1.76012790e+00 -9.56224620e-01 -2.76602954e-01 2.27104449e+00
6.28451467e-01 -2.86633939e-01 3.16381961e-01 1.27606556e-01
8.74094367e-01 -3.74033928e-01 -9.41440821e-01 -2.02763349e-01
-7.05406785e-01 3.69502716e-02 -2.67919172e-02 2.48178527e-01
-7.58402526e-01 1.02380466e+00 8.22314262e+00 4.43702936e-01
-9.59030271e-01 -7.76553825e-02 4.07426119e-01 2.77420133e-01
-4.52048153e-01 4.53904957e-01 -4.49106067e-01 6.06363773e-01
1.21835268e+00 -6.64140403e-01 6.59798503e-01 2.42619455e-01
2.21359357e-01 -2.77138591e-01 -1.17643416e+00 4.62827444e-01
-3.62810940e-02 -1.35618460e+00 3.32359523e-02 3.69846702e-01
6.55109048e-01 1.80306390e-01 -3.80466938e-01 1.51427954e-01
5.42968929e-01 -1.26330435e+00 8.83068681e-01 9.90007818e-01
3.08475018e-01 -2.78877556e-01 4.36906874e-01 5.93148410e-01
-1.19628954e+00 -1.51240215e-01 -7.13446587e-02 -4.67594892e-01
3.54360759e-01 6.40160441e-01 1.08566023e-01 4.12849039e-01
1.00689948e+00 8.04561794e-01 -3.53092998e-01 1.11114252e+00
7.66822770e-02 3.10091734e-01 -1.57171100e-01 -6.09039545e-01
-4.53513443e-01 -4.30131882e-01 7.13629007e-01 1.07277274e+00
2.24951908e-01 3.68674099e-01 -2.86105722e-01 1.08236134e+00
5.94055116e-01 -1.93068311e-01 -1.02306294e+00 -7.01709092e-01
6.01831675e-01 1.47399771e+00 -1.00987267e+00 -5.80064237e-01
-1.84337318e-01 4.70489651e-01 -3.40447098e-01 9.18468162e-02
-7.13464141e-01 -3.89798254e-01 8.24564338e-01 6.53992891e-02
-3.06132585e-01 -5.42960644e-01 -5.64372957e-01 -1.13690197e+00
-6.00007772e-01 -9.22972560e-01 -1.32008493e-02 -8.04223239e-01
-1.36616838e+00 4.68077004e-01 -3.28785062e-01 -1.09149110e+00
-1.43918633e-01 -4.18562770e-01 -5.28517067e-01 3.07383537e-01
-8.64343762e-01 -1.26421952e+00 -5.54168642e-01 -2.58333981e-01
3.22984755e-01 -8.76463875e-02 1.14873207e+00 -1.11267842e-01
-6.74481750e-01 4.31254655e-02 4.50477958e-01 -6.42663002e-01
3.38591605e-01 -1.02544069e+00 8.21884036e-01 4.27764237e-01
-4.43662435e-01 7.58632720e-01 1.01242018e+00 -5.61699748e-01
-1.98133910e+00 -4.90391225e-01 7.16329098e-01 -7.29410827e-01
5.65413117e-01 -5.19521832e-01 -6.48404360e-01 3.89786780e-01
8.05624366e-01 -5.03449857e-01 8.97803545e-01 -5.20197093e-01
2.42168196e-02 1.37892485e-01 -1.49092996e+00 8.74673247e-01
1.40677130e+00 -3.36941853e-02 -4.10375893e-01 2.94882119e-01
1.04663074e+00 3.71817887e-01 -1.29006863e+00 4.53227460e-01
1.06749249e+00 -1.10437047e+00 1.00043881e+00 -1.06416449e-01
2.36924052e-01 -6.69149816e-01 5.46211712e-02 -1.01143909e+00
-4.34577256e-01 -9.90569413e-01 -3.47898215e-01 1.40131497e+00
6.50715977e-02 -1.26622343e+00 4.00629133e-01 3.84928912e-01
-2.38319263e-01 -9.27781045e-01 -6.10460460e-01 -1.03927791e+00
6.41804755e-01 1.84525102e-01 7.19074368e-01 9.20179844e-01
3.97805721e-01 5.14144123e-01 -8.34807940e-03 -5.17800152e-01
7.34455585e-01 -2.12252632e-01 9.21930969e-01 -1.52559757e+00
-1.68105721e-01 -6.24774873e-01 -5.58954239e-01 -6.48763359e-01
-2.26270989e-01 -6.40542030e-01 1.86905622e-01 -1.88331854e+00
2.72978991e-01 -9.89505053e-02 -1.11330405e-01 1.25095518e-02
9.75507274e-02 3.20130050e-01 -1.31050870e-02 4.11592215e-01
-4.68311876e-01 6.05512977e-01 1.20099747e+00 7.37252980e-02
-3.99812199e-02 -5.60252368e-01 -8.85100663e-01 7.61312604e-01
7.74361253e-01 1.13432586e-01 -2.24765331e-01 -2.19692774e-02
4.97873813e-01 -3.66267413e-02 3.01342726e-01 -1.44660628e+00
3.09549630e-01 -4.74871784e-01 -2.41224334e-01 -3.92548323e-01
3.91952306e-01 -7.44320512e-01 8.52479458e-01 9.84176457e-01
1.94633588e-01 -3.10332272e-02 2.68352367e-02 3.18741828e-01
2.68922180e-01 -1.45695731e-01 9.59056795e-01 -5.67542315e-01
-4.19002116e-01 -3.85912880e-02 -1.15401411e+00 -1.58002019e-01
1.82116485e+00 -7.83541620e-01 -7.62626588e-01 -3.18523236e-02
-8.28847885e-01 3.91741246e-01 1.25453329e+00 4.84105080e-01
1.70977280e-01 -9.32001352e-01 -6.97974741e-01 -6.56142309e-02
2.46683359e-01 -3.34092766e-01 -1.33237571e-01 6.40623868e-01
-8.88755322e-01 5.06164432e-01 -3.66564512e-01 -6.06936336e-01
-7.45469570e-01 9.68989074e-01 5.91962814e-01 2.14851648e-01
-3.52041185e-01 4.97646242e-01 3.91086161e-01 -6.51576161e-01
-1.65701602e-02 1.48656085e-01 -2.95205206e-01 -1.24565855e-01
4.19501722e-01 8.30991447e-01 -3.62391233e-01 -6.77624345e-01
-4.89207268e-01 8.54650855e-01 6.61257386e-01 -1.43092811e-01
1.47276342e+00 -5.54543674e-01 -9.90209997e-01 1.12591934e+00
5.37560403e-01 -4.87288058e-01 -6.75557435e-01 2.20202357e-01
1.23955138e-01 2.38054499e-01 -6.05717838e-01 -1.00011361e+00
-5.31396449e-01 6.75388813e-01 5.48723698e-01 9.12579000e-01
1.16021621e+00 1.82092711e-01 8.69889677e-01 1.78312346e-01
8.85744154e-01 -9.09601212e-01 2.34912604e-01 5.59817016e-01
1.08031404e+00 -7.69466698e-01 -2.55451556e-02 -4.20756221e-01
1.37641773e-01 9.81685400e-01 7.29251504e-01 -1.66703895e-01
9.85192120e-01 6.74805343e-01 -8.81394297e-02 -1.64555073e-01
-1.27255297e+00 -3.63728344e-01 -3.73041540e-01 1.04048574e+00
9.98202026e-01 -1.61551073e-01 -9.79493499e-01 2.96802223e-01
-2.35407874e-01 9.61656198e-02 6.27283752e-01 1.21430326e+00
-7.87297368e-01 -1.34373140e+00 -4.42513078e-01 -2.70927008e-02
-1.81873590e-01 8.35866556e-02 -9.66643810e-01 5.23555934e-01
4.73723948e-01 1.29695451e+00 -2.12393045e-01 -4.32576835e-01
-8.65615532e-03 -1.43375874e-01 5.92155755e-01 -3.10793996e-01
-8.97430539e-01 -5.10713041e-01 1.66811019e-01 -2.75037825e-01
-6.34953320e-01 -5.51512539e-01 -1.34006333e+00 -7.64753997e-01
-3.47607642e-01 3.65398496e-01 9.09465551e-01 8.05867374e-01
5.91052949e-01 4.94531423e-01 3.93807232e-01 -1.02540755e+00
1.79828122e-01 -7.27431357e-01 -3.73659223e-01 -7.15292916e-02
1.70886859e-01 -4.73912895e-01 -5.01109958e-01 4.46125269e-01] | [5.605409145355225, 4.158647060394287] |
8bd70d5d-d335-40d1-80cc-27108c6fe5ba | iflyea-a-chinese-essay-assessment-system-with | null | null | https://aclanthology.org/2021.acl-demo.29 | https://aclanthology.org/2021.acl-demo.29.pdf | IFlyEA: A Chinese Essay Assessment System with Automated Rating, Review Generation, and Recommendation | Automated Essay Assessment (AEA) aims to judge students{'} writing proficiency in an automatic way. This paper presents a Chinese AEA system IFlyEssayAssess (IFlyEA), targeting on evaluating essays written by native Chinese students from primary and junior schools. IFlyEA provides multi-level and multi-dimension analytical modules for essay assessment. It has state-of-the-art grammar level analysis techniques, and also integrates components for rhetoric and discourse level analysis, which are important for evaluating native speakers{'} writing ability, but still challenging and less studied in previous work. Based on the comprehensive analysis, IFlyEA provides application services for essay scoring, review generation, recommendation, and explainable analytical visualization. These services can benefit both teachers and students during the process of writing teaching and learning. | ['Ting Liu', 'Shijin Wang', 'Bo Zhu', 'Zhichao Sheng', 'Ruiji Fu', 'Wei Song', 'Xiao Hu', 'Jiefu Gong'] | 2021-08-01 | null | null | null | acl-2021-5 | ['review-generation'] | ['natural-language-processing'] | [-4.26151305e-01 8.89313295e-02 -1.95004120e-01 -1.07797449e-02
-8.21611226e-01 -9.92089093e-01 4.40939903e-01 5.50469279e-01
2.36496851e-02 5.17436206e-01 3.29683036e-01 -1.09874380e+00
-6.91682518e-01 -8.99651945e-01 1.00620359e-01 -2.53324687e-01
8.28678310e-01 4.14798796e-01 5.16551323e-02 -4.51784313e-01
1.21730411e+00 3.99994344e-01 -1.64300537e+00 1.59839466e-01
1.64796913e+00 2.55093694e-01 1.72572881e-01 1.12944543e+00
-8.19882751e-01 1.41003478e+00 -1.01467824e+00 -7.98602104e-01
-6.74569666e-01 -8.70306432e-01 -1.01068783e+00 -1.85874864e-01
6.63977921e-01 -2.16452330e-01 3.72588895e-02 1.11469376e+00
1.29676938e-01 -7.18677696e-03 8.03091109e-01 -7.50428677e-01
-1.12212181e+00 8.42458427e-01 -2.53862292e-01 1.44564524e-01
7.34067678e-01 -2.68425524e-01 1.00480068e+00 -1.12146282e+00
4.13709372e-01 1.09220707e+00 5.62561393e-01 4.59330618e-01
-4.21344697e-01 -5.85094333e-01 3.48654799e-02 4.86770034e-01
-6.26604915e-01 -1.10468373e-01 8.51632535e-01 -8.93069506e-01
4.18602675e-01 3.26506883e-01 9.34145033e-01 6.01795435e-01
2.30370656e-01 8.66804659e-01 1.59736657e+00 -8.35002124e-01
7.14472383e-02 1.94922820e-01 1.16152453e+00 9.58422780e-01
3.23745281e-01 -8.83481741e-01 -8.06881785e-01 3.03667963e-01
1.61771193e-01 -3.90212834e-01 4.32654060e-02 5.85198045e-01
-8.97932589e-01 9.47791815e-01 -6.21347487e-01 4.61710811e-01
-2.34834589e-02 -4.74674165e-01 2.42189571e-01 5.82030058e-01
1.82244971e-01 5.55927694e-01 -8.74501169e-02 -9.08032477e-01
-9.62203324e-01 3.72296810e-01 1.13642240e+00 8.14745843e-01
1.28084272e-02 1.60018802e-01 -6.17472291e-01 8.25746894e-01
6.42822325e-01 6.34633958e-01 7.46236324e-01 -8.69582832e-01
5.55482090e-01 1.31506419e+00 -4.47946727e-01 -1.00139689e+00
-1.37161806e-01 -3.49009812e-01 -2.83755153e-01 3.97158235e-01
4.22081321e-01 -1.27700856e-02 -3.74859095e-01 1.07590044e+00
6.06842376e-02 -5.74199498e-01 1.56206056e-01 5.67543924e-01
1.91038954e+00 5.48179388e-01 3.04649681e-01 -2.82677680e-01
1.69429529e+00 -1.24250877e+00 -1.35947001e+00 2.47338131e-01
1.00128889e+00 -1.22745216e+00 1.65461111e+00 8.09527636e-01
-1.72531152e+00 -5.92375875e-01 -1.21574497e+00 -5.61073065e-01
-4.09778893e-01 7.24236906e-01 2.25807324e-01 1.35586643e+00
-8.64365458e-01 2.50108570e-01 -3.69440705e-01 1.26734376e-01
3.95712674e-01 -6.38391599e-02 1.38174117e-01 2.26002961e-01
-1.01949239e+00 7.98418045e-01 -3.26933056e-01 -4.97855902e-01
-1.96675360e-01 -1.01916540e+00 -6.94950819e-01 4.71577197e-01
6.38602674e-02 -2.34110072e-01 1.57294536e+00 -1.46474257e-01
-2.27325225e+00 1.11264181e+00 -2.54601650e-02 1.96249217e-01
3.76455069e-01 -3.22849452e-01 -2.60900587e-01 2.81885684e-01
3.27103078e-01 -2.04709187e-01 1.16173424e-01 -4.16558266e-01
-4.17588681e-01 -4.73250091e-01 -2.26874687e-02 5.54403067e-01
-7.84134924e-01 3.79966557e-01 -1.65052384e-01 -8.42673957e-01
2.85748065e-01 -3.59493375e-01 5.75762153e-01 -2.97084481e-01
-2.09675267e-01 -1.19666266e+00 9.26000297e-01 -1.16022313e+00
2.13355422e+00 -1.52375913e+00 -2.22570166e-01 3.57204676e-01
5.45071006e-01 6.18984580e-01 1.76586851e-01 6.29081964e-01
2.45157182e-01 4.38729644e-01 2.35467359e-01 -1.90157741e-01
6.06702983e-01 -3.76096487e-01 -8.15592706e-03 -1.53788507e-01
-2.80729294e-01 8.93884420e-01 -1.07269490e+00 -8.27295959e-01
1.25313893e-01 -2.02491537e-01 -5.32980412e-02 2.12069184e-01
-1.37663975e-01 6.93252459e-02 -6.42525196e-01 9.16118205e-01
2.66451359e-01 -2.27829009e-01 1.09780274e-01 8.98824811e-01
-6.16172850e-01 9.68581438e-01 -8.91852379e-01 1.26875257e+00
-5.44448197e-01 1.01963449e+00 -1.41003892e-01 -6.03615820e-01
1.37090826e+00 2.10451886e-01 -7.41761327e-02 -8.12754631e-01
1.19997330e-01 4.64407325e-01 1.17933258e-01 -8.16965163e-01
7.34034121e-01 4.26215887e-01 -9.17357281e-02 1.31693947e+00
-1.79270916e-02 -5.70221245e-01 8.15680146e-01 7.86994159e-01
7.82648504e-01 2.13828921e-01 4.91812706e-01 -6.68279529e-01
1.05150032e+00 -1.01988032e-01 -1.35829896e-01 6.36676788e-01
-9.50580835e-02 2.44877338e-01 7.15861678e-01 -8.15300345e-02
-7.66364634e-01 -8.37760091e-01 -3.76504287e-02 1.38227224e+00
-4.48856950e-01 -8.05380940e-01 -1.18047714e+00 -5.53734303e-01
-3.92706782e-01 7.61171162e-01 -3.01775098e-01 3.37783426e-01
-4.45433348e-01 -1.29948482e-01 4.33089942e-01 2.96473861e-01
5.98551214e-01 -1.24623394e+00 -4.94643003e-01 2.96529561e-01
-1.52604416e-01 -7.02865303e-01 -3.83091956e-01 -2.98145354e-01
-4.00055259e-01 -1.17320013e+00 -5.20632446e-01 -1.06596971e+00
4.94704038e-01 1.85546935e-01 1.24876010e+00 7.50117302e-01
-5.16945869e-02 6.66264176e-01 -4.46385503e-01 -8.87152135e-01
-7.71830916e-01 3.88260424e-01 -2.34641984e-01 -8.40223968e-01
7.90384889e-01 -4.01739955e-01 -1.55073941e-01 -6.33389428e-02
-3.59847218e-01 2.52278477e-01 2.28109598e-01 4.09558356e-01
1.35586694e-01 -1.94206581e-01 9.95049655e-01 -1.07279599e+00
1.71061015e+00 -1.33652955e-01 -6.21795177e-01 6.51543379e-01
-1.02045274e+00 -2.54865348e-01 6.52649462e-01 -1.29853308e-01
-1.24604750e+00 -1.00411415e+00 -3.12748104e-01 3.91827375e-01
-1.15919895e-01 8.50272655e-01 -1.07478062e-02 -3.47061038e-01
6.84679091e-01 9.76096839e-02 6.87148049e-02 -2.37988397e-01
-1.43340245e-01 9.60224748e-01 6.39432013e-01 -9.36345756e-01
2.76123822e-01 -6.42823458e-01 1.96031287e-01 -8.04135084e-01
-1.25988448e+00 -5.10068476e-01 -7.22753823e-01 -9.88773227e-01
4.17803645e-01 -6.72283709e-01 -1.55571747e+00 4.93260235e-01
-8.77933383e-01 -3.89027148e-01 -1.34078756e-01 5.22416174e-01
-1.42364115e-01 4.20346320e-01 -7.89514840e-01 -8.20360363e-01
-5.40439844e-01 -1.02798557e+00 3.72671813e-01 1.03595150e+00
-5.23576856e-01 -1.29424942e+00 2.55997390e-01 1.08253336e+00
7.11480007e-02 -2.56653726e-01 1.25771105e+00 -7.26538360e-01
-6.27898127e-02 1.18476383e-01 -6.54148906e-02 1.54337049e-01
-4.80189353e-01 5.93318701e-01 -7.02263236e-01 3.93312812e-01
2.92749126e-02 -6.43334389e-01 3.92693788e-01 2.18714654e-01
1.30852604e+00 -5.39429486e-01 4.11207110e-01 -4.55956580e-03
9.27357256e-01 1.70364305e-01 4.71940160e-01 6.01291001e-01
6.15616858e-01 8.48692656e-01 6.56903148e-01 3.79720658e-01
9.07922685e-01 2.26756752e-01 -1.64071977e-01 6.88641250e-01
-6.49969995e-01 -1.49653524e-01 6.54367387e-01 2.03308058e+00
-1.34254575e-01 -5.38426824e-03 -1.28759861e+00 7.67670572e-01
-1.63058090e+00 -1.01840031e+00 -1.11874485e+00 1.51579511e+00
1.11024928e+00 -3.94203141e-02 1.77530125e-01 6.26459002e-01
2.97436923e-01 -1.27860621e-01 8.09552222e-02 -1.31132054e+00
6.08062185e-02 8.34930956e-01 -3.60628188e-01 9.12353516e-01
-4.12551254e-01 8.13555479e-01 5.88615513e+00 1.07823098e+00
-7.44797766e-01 -1.48117453e-01 4.92659181e-01 2.98358530e-01
-6.04516745e-01 -3.89430493e-01 -1.04341137e+00 3.68362874e-01
1.04510653e+00 -5.10903418e-01 -9.22530591e-02 6.81600869e-01
8.43353122e-02 -2.40519643e-01 -4.43502337e-01 4.09054190e-01
2.56199121e-01 -1.58022404e+00 -1.51328966e-01 -1.52239755e-01
1.09922624e+00 -8.54490280e-01 2.73720592e-01 6.38462424e-01
4.10578161e-01 -1.05076206e+00 7.75940299e-01 6.33948505e-01
5.12280047e-01 -7.39079177e-01 6.43450975e-01 3.61956447e-01
-7.56943762e-01 4.48726714e-02 -6.17561005e-02 -6.97607577e-01
-5.74229479e-01 2.97091365e-01 -5.73762953e-01 3.43268424e-01
3.65957856e-01 4.43483621e-01 -9.03280020e-01 7.63107419e-01
-9.91255879e-01 1.27624345e+00 2.67924160e-01 -9.94851112e-01
2.29578555e-01 -6.21586204e-01 3.44575107e-01 1.23406005e+00
5.74683487e-01 6.93657994e-01 3.56830545e-02 7.55436897e-01
-1.11453377e-01 7.01362133e-01 8.51106718e-02 -1.74338564e-01
8.29952061e-01 1.33167553e+00 -7.19148099e-01 -4.33652043e-01
-2.86609560e-01 2.26799220e-01 2.89278388e-01 9.73130986e-02
-1.20286651e-01 -6.50834024e-01 2.69794688e-02 4.10816930e-02
-2.57378221e-01 -2.16519058e-01 -1.31989217e+00 -7.72194564e-01
-1.23494898e-03 -1.12790751e+00 4.30677980e-01 -6.52950108e-01
-9.98119473e-01 -7.92684630e-02 -3.26262623e-01 -9.20712233e-01
-2.23856136e-01 -9.67710257e-01 -1.37017047e+00 1.05414093e+00
-1.24025583e+00 -1.04555428e+00 -5.64421535e-01 3.23784709e-01
8.21288705e-01 -6.84632957e-01 9.25680459e-01 -4.08726744e-02
-6.57912076e-01 7.98853517e-01 -1.73819866e-02 4.74359989e-02
5.56748271e-01 -1.79160607e+00 -5.63690998e-02 8.08399916e-01
1.07603244e-01 4.76158381e-01 3.44581217e-01 -5.31568766e-01
-1.08731389e+00 -3.79975915e-01 1.67711699e+00 -4.89913702e-01
9.38120127e-01 3.98754925e-02 -1.03476143e+00 2.67296340e-02
5.41114569e-01 -8.62774730e-01 1.18659115e+00 3.92507434e-01
-8.17629416e-03 3.44409257e-01 -6.24181092e-01 7.26210356e-01
5.00675201e-01 -7.28300810e-01 -9.78706062e-01 1.99698046e-01
1.68641493e-01 -6.07774675e-01 -1.16982162e+00 -5.10027222e-02
7.21125782e-01 -9.37505424e-01 5.78405142e-01 -4.13094521e-01
1.19942081e+00 -1.39195542e-03 6.72069013e-01 -9.28358078e-01
-1.98438108e-01 -6.10253572e-01 -3.75828356e-01 1.48134005e+00
2.63746917e-01 -1.37958348e-01 8.91373158e-01 5.03129959e-01
-6.62912726e-01 -9.47394550e-01 -3.09773177e-01 -2.85270274e-01
5.42204976e-01 -2.53576785e-01 5.30859888e-01 1.21054351e+00
6.78591669e-01 3.63896668e-01 4.24272180e-01 -3.73906285e-01
4.05676931e-01 4.12642390e-01 6.89322472e-01 -1.53208387e+00
2.88465202e-01 -1.50549257e+00 -6.44061863e-02 -6.23106003e-01
3.30598652e-01 -7.54766047e-01 -4.18090373e-01 -1.84742093e+00
5.81869557e-02 -7.21888766e-02 4.08113748e-01 4.02584404e-01
-5.39350808e-01 1.41259953e-01 6.15246780e-02 1.12716042e-01
-6.40241265e-01 2.52481639e-01 1.71987391e+00 1.33199066e-01
-2.19640613e-01 1.90641195e-01 -8.68200421e-01 8.19662631e-01
1.01366341e+00 -2.26601228e-01 -6.02875769e-01 -6.01841062e-02
5.50923705e-01 4.23815958e-02 4.01863419e-02 -7.06264675e-01
6.07467353e-01 -5.44425607e-01 3.20475727e-01 -9.42135811e-01
-5.65307140e-01 1.19589925e-01 -5.84839404e-01 2.16047347e-01
-5.32540858e-01 3.11076611e-01 1.05009675e-01 -6.10764802e-01
-4.83852416e-01 -1.04282236e+00 5.59917450e-01 -3.16248275e-02
-3.25081378e-01 -1.99989408e-01 -8.92355561e-01 9.38584358e-02
8.64750087e-01 -4.55166876e-01 -8.77829194e-01 -5.22183597e-01
-4.45369482e-01 4.08583015e-01 3.66799012e-02 1.96951941e-01
6.25749707e-01 -1.26373208e+00 -1.00896049e+00 6.06440976e-02
-8.67085978e-02 -9.15470645e-02 2.73135215e-01 7.76455402e-01
-1.19227552e+00 7.50234663e-01 -3.64992797e-01 8.79990030e-03
-1.71287620e+00 -2.96965152e-01 -1.99563041e-01 -7.82760978e-01
-1.91003874e-01 7.93017030e-01 -2.56661177e-01 -6.32678032e-01
2.49297380e-01 -7.36670243e-03 -1.10737813e+00 3.58449191e-01
9.02002275e-01 1.04461181e+00 3.11538368e-01 -3.60672086e-01
3.11014742e-01 4.96204615e-01 -8.32114718e-04 -2.37736985e-01
9.42372322e-01 -2.24772498e-01 -3.03675711e-01 6.49164915e-01
2.57302374e-01 8.08731556e-01 -3.80301893e-01 1.23296916e-01
3.05633038e-01 -1.41962498e-01 3.70045565e-02 -1.34283912e+00
-2.64371306e-01 1.34231818e+00 -1.62580028e-01 4.35246319e-01
7.74789512e-01 -5.06481171e-01 2.48922676e-01 5.68416357e-01
-4.86546040e-01 -1.63711226e+00 4.05354649e-01 1.01313388e+00
1.00544238e+00 -9.05356705e-01 3.15852731e-01 -3.67414832e-01
-4.31539178e-01 1.84000683e+00 9.59332466e-01 2.89009273e-01
4.27060902e-01 1.20397307e-01 2.08902761e-01 -3.35310489e-01
-6.40840173e-01 2.33764157e-01 9.46527064e-01 2.44742945e-01
1.13851047e+00 2.84701139e-01 -1.17514312e+00 1.22667837e+00
-8.84510577e-01 -4.04499114e-01 1.06271756e+00 7.45817184e-01
-7.93670118e-01 -1.32310545e+00 -7.67968178e-01 4.75753874e-01
-7.42559016e-01 -2.56352842e-01 -7.06937611e-01 7.50736833e-01
-2.24733248e-01 1.00157785e+00 -3.52018088e-01 8.55800733e-02
-7.66811892e-02 4.18573678e-01 6.05817199e-01 -6.80662274e-01
-1.16214967e+00 -2.04150990e-01 2.59215236e-01 3.05936873e-01
-3.57820213e-01 -8.69005799e-01 -1.37468314e+00 -7.34550953e-01
-3.45185876e-01 5.52955091e-01 8.37065935e-01 1.14982283e+00
-2.09849954e-01 8.55805099e-01 3.05293977e-01 1.23222061e-01
-3.81666929e-01 -1.32406557e+00 -5.58028996e-01 -3.20465773e-01
-5.34890033e-02 -2.12927118e-01 -4.24281396e-02 8.29881728e-02] | [11.281232833862305, 9.248703002929688] |
2c3a46a4-4275-491d-a4c0-2ab3e6578bdc | information-screening-whilst-exploiting | 2305.11719 | null | https://arxiv.org/abs/2305.11719v2 | https://arxiv.org/pdf/2305.11719v2.pdf | Information Screening whilst Exploiting! Multimodal Relation Extraction with Feature Denoising and Multimodal Topic Modeling | Existing research on multimodal relation extraction (MRE) faces two co-existing challenges, internal-information over-utilization and external-information under-exploitation. To combat that, we propose a novel framework that simultaneously implements the idea of internal-information screening and external-information exploiting. First, we represent the fine-grained semantic structures of the input image and text with the visual and textual scene graphs, which are further fused into a unified cross-modal graph (CMG). Based on CMG, we perform structure refinement with the guidance of the graph information bottleneck principle, actively denoising the less-informative features. Next, we perform topic modeling over the input image and text, incorporating latent multimodal topic features to enrich the contexts. On the benchmark MRE dataset, our system outperforms the current best model significantly. With further in-depth analyses, we reveal the great potential of our method for the MRE task. Our codes are open at https://github.com/ChocoWu/MRE-ISE. | ['Tat-Seng Chua', 'Lidong Bing', 'Yixin Cao', 'Hao Fei', 'Shengqiong Wu'] | 2023-05-19 | null | null | null | null | ['relation-extraction'] | ['natural-language-processing'] | [ 3.58270764e-01 3.55654985e-01 -2.33139783e-01 -1.44481465e-01
-8.40959549e-01 -5.60881019e-01 6.48265302e-01 2.09757164e-01
-2.40910910e-02 3.05979997e-01 6.44149184e-01 -1.55809641e-01
-1.12473980e-01 -7.04741180e-01 -4.90556121e-01 -5.49562275e-01
1.36575893e-01 3.36461663e-01 1.99994147e-02 3.01373415e-02
1.17450058e-01 3.57714146e-02 -1.46719944e+00 6.59119964e-01
8.32375646e-01 9.12182629e-01 2.12003931e-01 4.63940293e-01
-2.64635533e-01 7.29941070e-01 -7.23454058e-02 -7.40172625e-01
-1.80062220e-01 -1.84691653e-01 -1.04195225e+00 5.83287597e-01
1.41897053e-01 -7.30915368e-02 -5.67111075e-01 1.11928046e+00
2.57165879e-01 -3.90224010e-02 6.35011673e-01 -1.38568819e+00
-5.48605621e-01 6.89621210e-01 -6.88094020e-01 -2.56162971e-01
4.32615370e-01 -1.09542221e-01 1.33643782e+00 -1.28358710e+00
9.40565765e-01 1.41716886e+00 3.39665264e-01 2.51007378e-01
-1.23540175e+00 -4.34599370e-01 4.43931133e-01 2.57213891e-01
-1.42515719e+00 -4.14347172e-01 1.04376674e+00 -4.01928782e-01
7.73259461e-01 3.15741688e-01 5.35281539e-01 1.15230870e+00
-4.04866003e-02 1.07673299e+00 1.01786053e+00 -3.94977033e-01
-7.75536001e-02 9.62201953e-02 2.60734707e-01 9.64701056e-01
7.38087595e-02 -2.75512993e-01 -8.85544538e-01 -2.76315570e-01
3.86678725e-01 -5.99959083e-02 -3.01752120e-01 -5.80940545e-01
-1.09880972e+00 7.71087289e-01 2.25634187e-01 1.68835118e-01
-3.56517404e-01 -1.50481582e-01 2.12551370e-01 5.99983111e-02
5.85970879e-01 2.30785802e-01 -2.65671253e-01 5.23742512e-02
-7.33053565e-01 -3.89180593e-02 6.89394355e-01 1.04321206e+00
8.97074938e-01 -5.95388234e-01 -3.14456165e-01 9.26331520e-01
5.79124570e-01 1.37966767e-01 -3.71608138e-02 -9.21622574e-01
7.16987252e-01 9.53270853e-01 -2.54968196e-01 -1.24872148e+00
-3.57087195e-01 -3.29615116e-01 -8.17982256e-01 -4.49223191e-01
1.99199781e-01 2.88108643e-02 -8.23082089e-01 1.61177802e+00
4.77704465e-01 -1.34514317e-01 2.10819021e-02 6.89790249e-01
9.79699910e-01 6.09266579e-01 3.36293697e-01 -8.86527896e-02
1.81083262e+00 -1.13812447e+00 -9.92503285e-01 -3.20534050e-01
5.68591833e-01 -8.20186079e-01 9.92405891e-01 3.05557400e-01
-9.07876313e-01 -1.97798163e-01 -7.94639230e-01 -3.51434648e-01
-4.45784301e-01 4.05240953e-01 6.37490094e-01 2.50920326e-01
-9.45401788e-01 1.94699794e-01 -8.06082726e-01 -4.06669259e-01
6.29086494e-01 1.75960973e-01 -5.27539790e-01 -3.71364444e-01
-1.05609059e+00 4.84199166e-01 5.31453550e-01 1.45971850e-01
-7.12704182e-01 -5.38500249e-01 -1.09441030e+00 1.75549854e-02
1.06935740e+00 -8.00382197e-01 8.11111689e-01 -5.57570457e-01
-1.16839683e+00 8.50615382e-01 -4.50662374e-01 4.28592712e-02
1.54084086e-01 -3.32647055e-01 -9.90318581e-02 4.11201686e-01
-1.39133297e-02 7.32484937e-01 7.99907148e-01 -1.64252520e+00
-3.57483536e-01 -5.28159380e-01 1.86197460e-02 4.11774814e-01
-5.28871179e-01 4.40059118e-02 -1.13939857e+00 -7.62707651e-01
3.87171388e-01 -8.03229868e-01 -1.58291355e-01 -3.62897158e-01
-8.01154315e-01 -2.97702670e-01 7.85737634e-01 -8.21126461e-01
1.61041582e+00 -2.26896811e+00 5.22225976e-01 4.23835397e-01
6.98206782e-01 -1.50187522e-01 -3.61054063e-01 6.30342126e-01
-9.70557109e-02 1.86579749e-01 -3.25182855e-01 -6.70587182e-01
1.00305006e-01 1.19493484e-01 -1.99995860e-01 1.35005340e-01
3.83568794e-01 1.25359523e+00 -8.35572243e-01 -8.06489289e-01
1.02990316e-02 4.99504596e-01 -3.72403085e-01 2.19610661e-01
-3.32350969e-01 4.42589074e-01 -6.08394921e-01 8.96611989e-01
6.27887964e-01 -6.51744425e-01 4.71243829e-01 -6.32734478e-01
1.90490380e-01 2.89929003e-01 -1.13248765e+00 1.70804358e+00
-9.06389579e-02 4.88541692e-01 3.03355187e-01 -8.12188566e-01
7.21676350e-01 1.10238947e-01 5.54607511e-01 -5.55831969e-01
4.17207815e-02 -1.37123451e-01 -5.10516047e-01 -4.84507293e-01
7.08214164e-01 1.68079510e-01 -8.89572799e-02 4.52755362e-01
2.93955952e-01 1.47778377e-01 5.80469668e-02 9.20603216e-01
9.37583864e-01 2.20061570e-01 2.64778405e-01 8.11967067e-03
2.83216417e-01 -1.07013337e-01 1.88912377e-01 5.54426193e-01
4.13763896e-02 5.63312352e-01 9.19359744e-01 9.00211185e-02
-7.40391552e-01 -8.36808145e-01 9.35223624e-02 1.21734095e+00
1.80482849e-01 -1.17409992e+00 -7.71276474e-01 -8.95712614e-01
-2.10967049e-01 5.29089332e-01 -6.05824888e-01 -2.65645850e-02
-1.43592626e-01 -8.41937840e-01 2.81547815e-01 3.04627091e-01
1.98772326e-01 -7.02305675e-01 -1.15527406e-01 -1.08593535e-02
-7.30592430e-01 -1.39384353e+00 -4.03304428e-01 -9.90072638e-03
-7.06486225e-01 -9.35586929e-01 -4.92791861e-01 -4.94886041e-01
7.01877356e-01 4.17483002e-01 1.19112372e+00 9.27149728e-02
-2.23114192e-01 8.27032030e-01 -3.84769976e-01 7.52875656e-02
-1.76924229e-01 -7.60431914e-03 -5.35587907e-01 2.49699190e-01
2.57939130e-01 -5.11158407e-01 -5.31560361e-01 2.29943261e-01
-9.01884377e-01 4.93114501e-01 6.30014896e-01 7.87299037e-01
7.34416425e-01 6.65801093e-02 1.45160764e-01 -1.12371492e+00
5.30666411e-01 -7.03563511e-01 -4.04228538e-01 4.25415993e-01
-5.80058753e-01 4.49523926e-02 -3.52473333e-02 -2.54883945e-01
-1.33478141e+00 5.61748333e-02 7.51487464e-02 -3.80703866e-01
-8.09618831e-02 8.78120065e-01 -6.69282615e-01 2.74405062e-01
7.20941648e-02 6.05958067e-02 -3.52613896e-01 -5.28919101e-01
8.30367029e-01 5.56425989e-01 5.15690804e-01 -7.14779794e-01
5.07838130e-01 5.45396984e-01 -1.18964911e-02 -9.10117567e-01
-1.04003668e+00 -6.46058023e-01 -6.29302621e-01 -2.73108155e-01
8.61173987e-01 -1.12713742e+00 -4.24757063e-01 4.13030058e-01
-1.15558577e+00 -1.73983112e-01 -3.45831066e-02 1.71096116e-01
-2.91627914e-01 7.79791236e-01 -6.73048615e-01 -8.09377134e-01
-3.24542038e-02 -1.02816939e+00 1.51084983e+00 -3.59296948e-02
-1.10158414e-01 -9.91702139e-01 -5.63957952e-02 8.72732043e-01
-1.25387743e-01 5.54672107e-02 1.09766579e+00 -5.30077159e-01
-8.53726387e-01 8.51270556e-02 -6.30401313e-01 -3.21073867e-02
-1.12335585e-01 -5.91990426e-02 -1.01836872e+00 -8.07003751e-02
-1.91841528e-01 -2.97448248e-01 1.26235008e+00 1.60238743e-01
1.15835929e+00 -2.07641914e-01 -4.40010071e-01 5.93084455e-01
1.26110530e+00 -3.16569299e-01 5.31282723e-01 1.72969088e-01
1.08648956e+00 1.00587964e+00 6.03692532e-01 5.99055052e-01
1.00122118e+00 6.31375790e-01 5.13350546e-01 -2.39486113e-01
-2.28200838e-01 -4.42681700e-01 3.16829145e-01 9.51340139e-01
-7.61570185e-02 -4.92929846e-01 -9.96928751e-01 6.22123182e-01
-2.10196042e+00 -6.45648360e-01 -1.81248143e-01 1.76525605e+00
7.22305000e-01 -1.68443039e-01 7.72213563e-02 -1.94173753e-01
7.09551156e-01 3.57146740e-01 -1.27919331e-01 1.64568424e-01
-2.03496769e-01 -2.12326929e-01 1.23350419e-01 4.69961911e-01
-1.30148232e+00 1.09933233e+00 5.24941826e+00 9.65458453e-01
-5.24465382e-01 2.14348450e-01 6.63080096e-01 1.90209180e-01
-6.99406624e-01 3.21191251e-01 -7.18866706e-01 1.12693772e-01
7.61401236e-01 1.92443117e-01 4.02396351e-01 4.69367772e-01
-5.29582947e-02 -2.59939730e-01 -9.79435384e-01 9.76479530e-01
1.13423519e-01 -1.20589495e+00 2.86323607e-01 2.52393782e-01
5.65504611e-01 -1.85376093e-01 1.12859361e-01 1.19295284e-01
1.35716677e-01 -7.70841062e-01 6.31157577e-01 5.83749413e-01
6.89582407e-01 -5.58311462e-01 5.67940176e-01 1.37582630e-01
-1.36377287e+00 9.32860076e-02 -1.04664139e-01 3.20691794e-01
1.01338163e-01 7.00793922e-01 -4.58203673e-01 1.10380912e+00
5.54258943e-01 9.31024551e-01 -8.58983338e-01 5.24140596e-01
-2.79196590e-01 5.41032374e-01 -2.29326919e-01 3.68670732e-01
7.26184696e-02 -3.60878974e-01 7.37934530e-01 1.40233123e+00
1.85174197e-01 1.69737533e-01 2.16403961e-01 9.90015864e-01
-1.90496922e-01 1.79201201e-01 -5.74059308e-01 -3.84508282e-01
2.49647379e-01 1.54411972e+00 -7.83094585e-01 -3.96768630e-01
-6.72810018e-01 1.01061893e+00 4.95471627e-01 6.33051753e-01
-5.43134570e-01 4.50668186e-02 2.37154186e-01 -2.42230386e-01
3.61421943e-01 -2.34991565e-01 -2.93927789e-01 -1.49499774e+00
1.18356444e-01 -9.13784206e-01 6.37793243e-01 -8.59752178e-01
-1.21178532e+00 4.29638028e-01 1.79442450e-01 -8.32935154e-01
6.72111362e-02 -5.79949617e-01 -2.45333314e-01 5.79487026e-01
-1.47089553e+00 -1.59914684e+00 -2.91493148e-01 7.34756589e-01
4.19294298e-01 1.18280627e-01 6.40302181e-01 2.72896647e-01
-7.24638164e-01 4.73087817e-01 -2.38689601e-01 1.22200899e-01
6.13670111e-01 -1.13249695e+00 9.24281776e-02 7.95671761e-01
4.07766253e-01 6.00833595e-01 3.87481987e-01 -9.09059823e-01
-1.56302953e+00 -1.13147414e+00 9.90560770e-01 -5.32930195e-01
1.04305482e+00 -7.15770125e-01 -9.86269176e-01 7.71902621e-01
3.31140429e-01 -3.08157712e-01 7.58730352e-01 4.11436707e-01
-6.10473335e-01 3.29343021e-01 -5.67344904e-01 6.91768110e-01
1.09932721e+00 -8.32071364e-01 -4.88869816e-01 3.02094728e-01
9.02928114e-01 -2.27054179e-01 -1.03490400e+00 3.74821812e-01
3.91313404e-01 -7.88166106e-01 1.01295221e+00 -3.96729320e-01
6.50965035e-01 -8.17504525e-02 -3.38869542e-01 -9.45664227e-01
-1.61904261e-01 -7.85066485e-01 -3.89164865e-01 1.58538532e+00
5.15911520e-01 -3.03579748e-01 5.96952140e-01 4.90943372e-01
-1.86734367e-02 -8.10579360e-01 -6.80153966e-01 -2.34506518e-01
-2.91426927e-01 -7.19055593e-01 4.17553842e-01 9.15704787e-01
2.52808094e-01 7.03063071e-01 -4.85475779e-01 2.64842629e-01
7.18695343e-01 1.70738041e-01 6.55732810e-01 -1.03055632e+00
-2.68434316e-01 -4.11236852e-01 -1.34067833e-02 -9.86255348e-01
1.85391888e-01 -9.07479227e-01 -1.46006867e-01 -1.57787383e+00
8.98550570e-01 5.74523322e-02 -2.85284162e-01 6.46123528e-01
-4.74074721e-01 1.50926262e-01 3.41772765e-01 2.84263492e-01
-1.19396436e+00 7.98882902e-01 1.37436700e+00 -1.05753385e-01
-6.59543425e-02 -2.63583690e-01 -8.65190625e-01 7.17597604e-01
5.12932420e-01 -3.41478109e-01 -3.65122050e-01 -2.26101935e-01
4.81369078e-01 8.29149485e-02 6.05620801e-01 -3.53238046e-01
2.57198513e-01 5.14521338e-02 1.58216029e-01 -8.14418316e-01
4.40987498e-01 -7.75839806e-01 7.58051127e-02 -1.55386358e-01
-2.83050418e-01 -1.60371587e-01 1.80648446e-01 7.49808908e-01
-3.85424674e-01 -1.64346248e-02 3.02348405e-01 5.81457503e-02
-7.00963914e-01 2.91030020e-01 -2.84668356e-01 2.09617149e-03
6.43297315e-01 1.10436857e-01 -6.77957177e-01 -6.22534752e-01
-1.00977099e+00 3.75905126e-01 4.10653323e-01 2.74762481e-01
8.17593694e-01 -1.29750538e+00 -4.92384046e-01 1.77395001e-01
5.23491800e-01 -7.69953802e-02 4.37724233e-01 1.12728477e+00
2.31710374e-01 3.71592999e-01 3.12448055e-01 -5.71644366e-01
-1.44436932e+00 6.99092448e-01 -1.24817677e-01 -5.70833027e-01
-6.07602894e-01 6.61447644e-01 5.87431788e-01 -3.71585697e-01
2.78870821e-01 -1.09296963e-01 -3.53161991e-01 3.53219569e-01
2.01743245e-01 2.33174250e-01 -1.89784750e-01 -8.17964673e-01
-2.73673296e-01 5.74783087e-01 -7.77807310e-02 -2.48917207e-01
1.19637370e+00 -6.81276739e-01 -4.99065548e-01 2.67273724e-01
1.09968483e+00 1.18085936e-01 -1.14754176e+00 -5.61337829e-01
4.21669632e-01 -2.97183514e-01 1.57294407e-01 -7.88639128e-01
-9.02187526e-01 8.40198636e-01 1.47087947e-01 1.77234039e-01
1.38071394e+00 6.15420878e-01 5.17495275e-01 2.53794104e-01
-1.67607516e-01 -9.58248019e-01 2.90433913e-01 4.53677148e-01
8.99039149e-01 -1.34108949e+00 1.57654643e-01 -9.12327647e-01
-9.02705371e-01 8.68554235e-01 3.30569267e-01 4.39681560e-01
6.95209563e-01 1.33789673e-01 -1.18627705e-01 -7.73802102e-01
-9.29697573e-01 -4.46317047e-01 8.68635416e-01 3.31948042e-01
3.41871530e-01 1.42905086e-01 -4.69256528e-02 9.29386973e-01
1.57711834e-01 -2.60682851e-01 2.09060237e-01 6.46670997e-01
-1.23317704e-01 -1.21967196e+00 -3.14695209e-01 1.65729448e-01
-3.30456793e-01 -3.75064522e-01 -7.47619987e-01 8.05788457e-01
-8.23033527e-02 1.06300628e+00 -1.93431273e-01 -4.65786010e-01
1.76542521e-01 2.10117653e-01 3.05157602e-01 -5.24967432e-01
-1.88637197e-01 7.51515567e-01 3.50666523e-01 -8.92349541e-01
-6.90409839e-01 -7.42295027e-01 -9.70389664e-01 9.53151882e-02
-2.74894744e-01 -1.55729884e-02 6.05836868e-01 9.40286517e-01
5.64682722e-01 5.14820933e-01 3.28817636e-01 -7.42695153e-01
3.41691375e-02 -7.75627375e-01 -3.83414090e-01 4.24688369e-01
2.11572886e-01 -6.42613530e-01 -3.19268703e-01 1.72334731e-01] | [10.523110389709473, 1.3403679132461548] |
cc9a00b6-6de9-484b-b2b2-9ae0b552f52f | learning-synthetic-environments-for | 2101.09721 | null | https://arxiv.org/abs/2101.09721v3 | https://arxiv.org/pdf/2101.09721v3.pdf | Learning Synthetic Environments for Reinforcement Learning with Evolution Strategies | This work explores learning agent-agnostic synthetic environments (SEs) for Reinforcement Learning. SEs act as a proxy for target environments and allow agents to be trained more efficiently than when directly trained on the target environment. We formulate this as a bi-level optimization problem and represent an SE as a neural network. By using Natural Evolution Strategies and a population of SE parameter vectors, we train agents in the inner loop on evolving SEs while in the outer loop we use the performance on the target task as a score for meta-updating the SE population. We show empirically that our method is capable of learning SEs for two discrete-action-space tasks (CartPole-v0 and Acrobot-v1) that allow us to train agents more robustly and with up to 60% fewer steps. Not only do we show in experiments with 4000 evaluations that the SEs are robust against hyperparameter changes such as the learning rate, batch sizes and network sizes, we also show that SEs trained with DDQN agents transfer in limited ways to a discrete-action-space version of TD3 and very well to Dueling DDQN. | ['Frank Hutter', 'Thomas Nierhoff', 'Fabio Ferreira'] | 2021-01-24 | null | null | null | null | ['acrobot'] | ['playing-games'] | [ 8.39480013e-02 7.92896561e-03 3.10046375e-01 3.78581695e-02
-5.17391026e-01 -5.76322317e-01 8.23927343e-01 -9.25497040e-02
-1.24456501e+00 1.26933837e+00 -3.83884847e-01 3.53670679e-02
-1.08140029e-01 -7.65961349e-01 -1.03695464e+00 -9.06833887e-01
-4.88822877e-01 9.32151139e-01 5.42938948e-01 -7.06191063e-01
1.40464887e-01 4.09668446e-01 -1.78088057e+00 -2.40938097e-01
1.04303896e+00 5.91602266e-01 1.19219080e-01 1.05371356e+00
5.32569885e-01 7.68544316e-01 -1.32166564e+00 -5.62487617e-02
4.94580984e-01 -6.48953736e-01 -5.48989654e-01 1.16594873e-01
-6.01882525e-02 -2.51955874e-02 2.24027913e-02 9.53127325e-01
7.03647852e-01 5.25508463e-01 5.80405295e-01 -1.36988425e+00
-2.81346589e-01 6.23090208e-01 -1.76032990e-01 1.31373584e-01
3.35525535e-02 7.54891157e-01 4.60227638e-01 -2.84446448e-01
7.10197806e-01 1.35464156e+00 6.76251054e-01 8.60104442e-01
-1.34653544e+00 -4.56113219e-01 4.00264040e-02 -1.37375310e-01
-9.37231660e-01 -2.70091295e-01 2.29737133e-01 -3.19679058e-03
1.19943523e+00 1.22992814e-01 9.76352811e-01 1.26112986e+00
3.56650829e-01 7.13782430e-01 1.07818627e+00 -3.60027641e-01
8.49233389e-01 2.44000584e-01 -5.28979123e-01 6.46366239e-01
2.62137562e-01 7.17031240e-01 -2.22231060e-01 -1.42991841e-01
8.99891138e-01 -5.35847723e-01 -1.21104486e-01 -5.78702867e-01
-1.34421480e+00 1.12734807e+00 4.72491980e-01 8.07818472e-02
-7.45661080e-01 5.96446753e-01 5.36857426e-01 7.43380129e-01
9.49601009e-02 1.23490274e+00 -5.89851439e-01 -5.08613944e-01
-4.41620678e-01 7.01070726e-01 1.04490793e+00 4.71959144e-01
4.95437264e-01 6.91240847e-01 -9.77745429e-02 8.66028905e-01
-5.93845174e-02 5.62040865e-01 9.19317842e-01 -1.27552307e+00
2.64352977e-01 3.03105742e-01 4.35124636e-01 -3.41931880e-01
-4.95432228e-01 -5.38701653e-01 -2.94551194e-01 9.93630409e-01
2.53633708e-01 -9.45939958e-01 -9.91447449e-01 1.99094737e+00
4.94891644e-01 2.08638132e-01 6.73528612e-01 6.48978710e-01
2.74440736e-01 6.78320289e-01 1.61765087e-02 -8.77236873e-02
8.51966500e-01 -1.22561419e+00 -1.89192116e-01 -3.63454521e-01
6.83856130e-01 -1.95932716e-01 1.06000030e+00 3.66179943e-01
-1.36016572e+00 -4.45282072e-01 -1.27445209e+00 8.74666631e-01
-4.22367126e-01 -2.30091736e-01 4.36943859e-01 6.44801021e-01
-1.42141342e+00 9.86185670e-01 -1.06446290e+00 -3.33589584e-01
1.20669238e-01 6.49999559e-01 -1.08936159e-02 4.43688363e-01
-1.35595429e+00 1.21672320e+00 8.03883791e-01 -3.02406639e-01
-1.39186704e+00 -4.95354712e-01 -7.33231187e-01 -3.22086439e-02
4.28635478e-01 -7.48770177e-01 1.46745420e+00 -1.29489541e+00
-2.13159752e+00 3.02094579e-01 6.09228849e-01 -7.67053783e-01
7.28519022e-01 2.56603688e-01 -1.29013255e-01 -8.12462494e-02
-2.52348065e-01 1.06547105e+00 1.00825417e+00 -1.37574017e+00
-9.53141809e-01 3.64574380e-02 3.40881675e-01 7.05260277e-01
-2.60887206e-01 -6.59822822e-02 1.03503995e-01 -5.45283854e-01
-6.45103931e-01 -1.33401620e+00 -4.25138026e-01 -3.77503514e-01
-4.97614667e-02 -1.34103119e-01 6.04562938e-01 -8.78443420e-02
5.73772848e-01 -1.76036692e+00 7.13029861e-01 1.34045377e-01
-3.94442827e-02 5.61333418e-01 -6.98334157e-01 2.69923329e-01
4.23571393e-02 -4.42289487e-02 -3.73312652e-01 -2.37802923e-01
2.24759638e-01 4.89541650e-01 1.82697028e-01 2.16728255e-01
2.65985548e-01 9.88353789e-01 -1.02973270e+00 -1.22967467e-01
-1.69200554e-01 4.15948063e-01 -5.66954374e-01 2.95206726e-01
-6.00611210e-01 4.19062644e-01 -4.44087952e-01 4.10524867e-02
1.02253795e-01 -2.49409638e-02 -4.87848967e-02 6.03392124e-01
-6.30669948e-03 -1.27379522e-01 -1.25497520e+00 1.28037333e+00
-5.27671635e-01 5.49593151e-01 1.65717661e-01 -7.54921019e-01
7.93145359e-01 1.23787463e-01 2.44883686e-01 -9.28365409e-01
1.66100562e-01 1.20136961e-01 5.34357905e-01 -2.93437630e-01
5.01127958e-01 2.03114282e-03 -1.59002766e-02 7.44674563e-01
1.43813685e-01 -4.99317080e-01 7.08114326e-01 -2.32478574e-01
1.26265538e+00 2.41497561e-01 -1.72571599e-01 -1.73820347e-01
3.77463490e-01 1.91796929e-01 4.45418775e-01 1.01711547e+00
-4.21247818e-02 1.09226428e-01 5.58395982e-01 -3.89825016e-01
-1.39525115e+00 -9.85346377e-01 6.71100691e-02 1.29388428e+00
-8.23811740e-02 3.49368788e-02 -9.57680583e-01 -7.29306757e-01
9.26338658e-02 1.06094742e+00 -9.83624816e-01 -4.69013184e-01
-7.91886330e-01 -1.06268597e+00 7.88973629e-01 3.59269738e-01
6.48250878e-01 -1.68488801e+00 -1.29182088e+00 4.41022903e-01
5.42929232e-01 -4.96963471e-01 -3.19037050e-01 6.67043746e-01
-7.01401770e-01 -8.93447042e-01 -9.84390676e-01 -7.09506392e-01
4.99733180e-01 -4.09199804e-01 1.18253362e+00 2.14408845e-01
-2.20532075e-01 2.81228125e-01 -2.30635837e-01 -5.14117897e-01
-1.09096944e+00 1.44677758e-01 3.16620141e-01 -4.26163077e-01
-1.93057716e-01 -5.43803275e-01 -4.20379817e-01 4.62284923e-01
-9.65416312e-01 -1.02055356e-01 4.39052999e-01 1.19279110e+00
1.30532786e-01 1.13975607e-01 7.02363789e-01 -7.46905982e-01
1.28871882e+00 -2.97371596e-01 -1.02260804e+00 3.39284927e-01
-6.00029945e-01 4.86178577e-01 7.91336477e-01 -9.66499865e-01
-9.47764397e-01 -2.43773267e-01 -2.78559178e-02 -2.37151250e-01
2.87258849e-02 2.27761760e-01 3.58803064e-01 -3.79688293e-01
1.12364924e+00 3.95509928e-01 3.13952297e-01 1.50148347e-01
1.90117881e-01 1.77385241e-01 1.72567368e-01 -8.84725273e-01
8.83352757e-01 -1.90758601e-01 -1.23243280e-01 -4.45761204e-01
-1.66609198e-01 5.50231099e-01 -1.79700255e-01 -6.35291561e-02
6.98066711e-01 -5.48500121e-01 -1.01123583e+00 6.64935112e-01
-6.66985452e-01 -1.35593975e+00 -8.72649670e-01 3.26807857e-01
-8.32742751e-01 -3.16780061e-01 -5.72872043e-01 -6.53614402e-01
-9.69313830e-02 -1.37050819e+00 7.18314409e-01 5.28726041e-01
7.02133402e-02 -1.28068912e+00 6.53871000e-01 -1.60797030e-01
7.12723196e-01 4.01265651e-01 7.57481039e-01 -7.14772940e-01
-1.49496302e-01 1.86920375e-01 3.67216587e-01 4.01040196e-01
-5.48947155e-02 6.64655045e-02 -6.05181813e-01 -8.12727928e-01
-1.58639073e-01 -9.17199850e-01 5.10496318e-01 1.92863360e-01
5.78987479e-01 -3.44967484e-01 -5.38796745e-02 5.85792899e-01
1.17059124e+00 6.94154978e-01 5.31651318e-01 1.04268682e+00
2.70636082e-01 3.99273634e-01 5.86655736e-01 3.66384983e-01
2.11698607e-01 6.77818120e-01 5.29335380e-01 -4.21375819e-02
5.98257631e-02 5.07211015e-02 5.47769964e-01 4.95485365e-01
-8.13536048e-02 -4.52287048e-01 -8.41640890e-01 2.22216442e-01
-1.92486060e+00 -8.77488256e-01 5.72968662e-01 2.23537898e+00
9.90179956e-01 2.94703007e-01 5.74977577e-01 -2.99075633e-01
5.37132680e-01 -1.49589613e-01 -1.09520710e+00 -5.04950881e-01
-2.54333317e-01 2.97516733e-01 4.86765563e-01 4.40052003e-01
-7.97458708e-01 1.03728473e+00 6.59597158e+00 4.38988566e-01
-1.22187686e+00 -1.12113528e-01 6.44796848e-01 -3.71696770e-01
1.77177154e-02 -4.58622485e-01 -4.49215889e-01 5.88736057e-01
1.18737113e+00 -2.58250147e-01 9.97971237e-01 6.61599278e-01
1.20090470e-02 -1.48545057e-01 -1.14095962e+00 6.35236740e-01
-1.22145586e-01 -1.18049550e+00 -3.58916074e-01 2.81294156e-02
1.19359255e+00 4.14242089e-01 3.06136310e-01 9.13174629e-01
1.09414697e+00 -1.17127717e+00 6.94891214e-01 1.06592603e-01
4.19736445e-01 -1.10728705e+00 7.78150976e-01 4.71275181e-01
-6.62980020e-01 -3.12728733e-01 -4.43882436e-01 1.64334223e-01
-1.02647118e-01 -5.34244895e-01 -1.00935638e+00 1.61328271e-01
5.76295018e-01 1.67564407e-01 -6.59482360e-01 1.24563384e+00
-1.92088798e-01 3.69691312e-01 -4.80095237e-01 -5.05656719e-01
6.52262330e-01 -2.42866620e-01 6.57524645e-01 8.16841185e-01
2.89102733e-01 -1.37050852e-01 5.28368428e-02 6.71890795e-01
-8.62691260e-04 -2.63017058e-01 -4.34133291e-01 -1.79722473e-01
5.14055610e-01 9.50380981e-01 -6.55300796e-01 -2.40856022e-01
2.56115317e-01 8.80850136e-01 6.36586189e-01 6.16856813e-01
-9.55766737e-01 -4.84471202e-01 7.06755221e-01 -3.84504110e-01
5.53675652e-01 -1.76769480e-01 3.21567804e-01 -1.04500759e+00
-4.87481445e-01 -1.51760852e+00 7.84340650e-02 -6.89331472e-01
-1.00020182e+00 1.05668724e+00 -1.08831905e-01 -1.01596797e+00
-8.56480420e-01 -4.39324021e-01 -6.74402833e-01 6.33899748e-01
-1.20289600e+00 -4.07117009e-01 -3.39040346e-02 4.94665563e-01
5.04001558e-01 -7.53445745e-01 9.10298586e-01 -1.99129596e-01
-7.85589933e-01 4.96487588e-01 5.47843277e-01 -2.02960342e-01
4.54716057e-01 -1.55598271e+00 6.11326694e-01 3.83282304e-01
-5.24331406e-02 2.67791867e-01 1.11525345e+00 -4.92835462e-01
-1.17978764e+00 -9.61856008e-01 -1.54377624e-01 -3.30017358e-01
6.31183982e-01 -2.79677093e-01 -7.24569261e-01 4.83978599e-01
4.68786716e-01 -1.28840163e-01 7.41922855e-02 -1.60837963e-01
1.18224263e-01 2.11366341e-02 -1.37551963e+00 9.18202817e-01
9.19143856e-01 1.46234155e-01 -5.19150376e-01 5.21708727e-01
6.43000782e-01 -7.79039621e-01 -8.48064184e-01 1.21589132e-01
1.21630646e-01 -8.16871762e-01 8.96557927e-01 -8.89463723e-01
2.26956204e-01 -2.22370252e-01 1.52556941e-01 -2.25696325e+00
-1.14979327e-01 -5.98636448e-01 8.69941339e-02 6.71301305e-01
6.64549589e-01 -1.10792041e+00 9.03706491e-01 3.90907109e-01
1.20825559e-01 -8.02676558e-01 -9.64489043e-01 -1.07243526e+00
4.18835998e-01 1.75216034e-01 6.92368686e-01 6.69795811e-01
-4.31503445e-01 3.33825886e-01 -1.90676227e-01 -1.11300863e-01
6.14644587e-01 -3.68517160e-01 8.36421192e-01 -1.04492044e+00
-8.75389993e-01 -6.23843729e-01 -1.49799734e-01 -6.39136672e-01
3.24176759e-01 -4.29672658e-01 2.97010511e-01 -1.12602627e+00
-1.06861725e-01 -7.61479855e-01 -1.89515099e-01 4.22298521e-01
-1.75399914e-01 8.16192478e-02 4.01648462e-01 5.32827973e-02
-7.25262821e-01 9.03151512e-01 1.37767875e+00 -9.65152681e-02
-4.58862722e-01 -5.09839691e-02 -2.50955135e-01 4.12963420e-01
9.02065456e-01 -5.74759901e-01 -7.95358717e-01 -4.93531376e-01
3.76169443e-01 5.46071306e-02 1.51938707e-01 -1.25556600e+00
1.83672950e-01 -3.27035964e-01 5.03727436e-01 3.10058743e-01
5.51787853e-01 -5.80478549e-01 1.64799333e-01 8.63949895e-01
-5.47297478e-01 5.71418047e-01 5.04984438e-01 3.81541014e-01
1.04865670e-01 -6.02683008e-01 9.03438926e-01 -1.55008540e-01
-4.54885304e-01 7.52234533e-02 -6.23463571e-01 3.43080550e-01
1.32116508e+00 -3.23217422e-01 -4.64381337e-01 -4.33299601e-01
-6.44133449e-01 7.97427237e-01 8.04288566e-01 4.60988045e-01
4.39213783e-01 -1.06129146e+00 -8.92314196e-01 1.54570177e-01
-2.00605750e-01 -1.05897650e-01 -1.45641804e-01 3.86744082e-01
-7.73092508e-01 6.34113029e-02 -7.34438121e-01 -5.15251637e-01
-1.10250843e+00 3.56079847e-01 1.00050914e+00 -3.40843618e-01
-3.33521962e-01 1.06028426e+00 -5.42887636e-02 -8.94491494e-01
4.47934031e-01 3.59087400e-02 -2.68357575e-01 -2.61950195e-01
3.76208246e-01 3.10214490e-01 -1.43188566e-01 -2.50202775e-01
-2.33129144e-01 1.98094964e-01 -1.52138084e-01 -6.25766695e-01
1.64870596e+00 4.44652051e-01 3.67514968e-01 2.60693789e-01
7.60667801e-01 -4.78168726e-01 -1.98580146e+00 1.53344691e-01
-1.13808423e-01 -1.42558798e-01 -1.40127599e-01 -1.04975295e+00
-1.00389850e+00 2.91191459e-01 8.41371477e-01 3.71798813e-01
8.90385032e-01 -4.26535726e-01 3.54858607e-01 8.54484260e-01
4.88883048e-01 -1.35410762e+00 5.54006338e-01 7.14080393e-01
9.78672504e-01 -1.09460282e+00 -2.22334951e-01 5.70130944e-01
-9.29034054e-01 1.09691465e+00 8.80461454e-01 -4.02148217e-01
-6.00373670e-02 4.17345494e-01 1.95741951e-01 -5.56864701e-02
-1.23932350e+00 -4.12809104e-02 -2.34935358e-01 9.50214982e-01
-7.56996647e-02 -1.89816251e-01 1.99703053e-01 -3.30577612e-01
-2.72690892e-01 -2.74525642e-01 5.99774539e-01 1.11479938e+00
-5.35894632e-01 -1.31248546e+00 -3.52489889e-01 3.06917101e-01
-9.50986668e-02 2.73763895e-01 -2.92572320e-01 1.15273511e+00
-2.04801932e-02 5.89118361e-01 2.39241302e-01 -2.28097051e-01
4.92754668e-01 -1.90593883e-01 7.19829857e-01 -5.05228460e-01
-1.14965892e+00 -3.71490896e-01 3.52923512e-01 -4.66395050e-01
-1.31227404e-01 -7.54328191e-01 -1.31639802e+00 -2.40572616e-01
-1.35857821e-01 5.49657345e-01 7.12219357e-01 6.83614492e-01
1.92360133e-01 8.84375691e-01 6.28978908e-01 -1.13858414e+00
-1.12851954e+00 -8.03816140e-01 -3.43973756e-01 2.83752471e-01
3.60281438e-01 -7.99636364e-01 -4.17650759e-01 -3.90118748e-01] | [4.120968818664551, 1.8131026029586792] |
992b41d9-ba1c-448a-b3c4-19796ceb46fd | gvdoc-graph-based-visual-document | 2305.17219 | null | https://arxiv.org/abs/2305.17219v1 | https://arxiv.org/pdf/2305.17219v1.pdf | GVdoc: Graph-based Visual Document Classification | The robustness of a model for real-world deployment is decided by how well it performs on unseen data and distinguishes between in-domain and out-of-domain samples. Visual document classifiers have shown impressive performance on in-distribution test sets. However, they tend to have a hard time correctly classifying and differentiating out-of-distribution examples. Image-based classifiers lack the text component, whereas multi-modality transformer-based models face the token serialization problem in visual documents due to their diverse layouts. They also require a lot of computing power during inference, making them impractical for many real-world applications. We propose, GVdoc, a graph-based document classification model that addresses both of these challenges. Our approach generates a document graph based on its layout, and then trains a graph neural network to learn node and graph embeddings. Through experiments, we show that our model, even with fewer parameters, outperforms state-of-the-art models on out-of-distribution data while retaining comparable performance on the in-distribution test set. | ['Ashish Verma', 'Catherine Finegan-Dollak', 'Mohammed J. Zaki', 'Fnu Mohbat'] | 2023-05-26 | null | null | null | null | ['graph-attention', 'document-classification'] | ['graphs', 'natural-language-processing'] | [ 2.96727661e-03 -2.46426746e-01 -3.32475275e-01 -3.34658831e-01
-7.01514781e-01 -1.11994004e+00 9.16240454e-01 3.73994827e-01
-2.66063539e-03 2.12674126e-01 -1.78057179e-01 -6.77575171e-01
-2.51748767e-02 -7.78970420e-01 -5.89577198e-01 -6.44729912e-01
-2.78886128e-02 1.08654225e+00 3.85832369e-01 1.34509085e-02
1.44785255e-01 4.89118516e-01 -1.34704041e+00 5.23462057e-01
5.98364949e-01 9.52792466e-01 1.35545611e-01 8.27623665e-01
-5.81626058e-01 8.14032078e-01 -9.99881148e-01 -4.78552848e-01
-6.78425729e-02 -3.52065295e-01 -5.50945640e-01 3.67839098e-01
1.01839769e+00 -2.78159082e-01 -5.47778070e-01 1.06139815e+00
5.32910705e-01 -1.62192494e-01 1.15275991e+00 -1.69231606e+00
-1.41284609e+00 3.95974427e-01 -5.64201653e-01 2.10366413e-01
3.62371296e-01 -1.62780970e-01 1.15130985e+00 -8.61879051e-01
7.45393515e-01 1.37221408e+00 5.97589791e-01 3.38327944e-01
-1.35773146e+00 -3.82437497e-01 3.61988485e-01 2.17497766e-01
-1.26113832e+00 -7.64415637e-02 8.32548320e-01 -5.55302143e-01
1.12483060e+00 7.68459812e-02 5.81842840e-01 1.53904355e+00
2.23810837e-01 1.09326708e+00 9.67462599e-01 -4.57740933e-01
2.67259687e-01 1.45432875e-01 9.29767266e-02 8.63520801e-01
3.75774056e-01 -3.23945880e-01 -3.48306328e-01 -9.45936069e-02
4.07711834e-01 -1.00924904e-02 -4.54091340e-01 -8.76855791e-01
-8.15381527e-01 8.29152465e-01 4.50143635e-01 3.52096587e-01
2.39494383e-01 2.11304933e-01 4.35446024e-01 3.44792277e-01
5.50632179e-01 1.52912885e-01 -8.91553164e-02 9.79274884e-03
-1.07837617e+00 8.99709612e-02 7.68889129e-01 1.25880682e+00
4.64272916e-01 2.00027470e-02 -2.92857796e-01 9.93219852e-01
1.72574922e-01 6.18884087e-01 5.68216980e-01 -3.01747061e-02
5.67884028e-01 7.27326035e-01 -5.34931958e-01 -1.01234329e+00
-3.17562759e-01 -5.58461964e-01 -8.86707425e-01 8.34588856e-02
6.51992440e-01 4.25294548e-01 -1.47821820e+00 1.39850485e+00
-7.25641251e-02 -2.99497724e-01 -3.21690403e-02 6.95460021e-01
8.24485481e-01 5.02502441e-01 -8.82530212e-02 4.20607269e-01
1.35574079e+00 -9.40997481e-01 -4.03170317e-01 -5.72153449e-01
6.59773290e-01 -6.67834163e-01 1.32835364e+00 4.30146486e-01
-7.34781206e-01 -4.37397748e-01 -1.06077981e+00 -1.42083719e-01
-7.89689720e-01 2.32978493e-01 5.22224426e-01 7.05986083e-01
-1.15263140e+00 2.86033541e-01 -4.82446641e-01 -5.88022470e-01
6.54735923e-01 2.35536285e-02 -5.81258476e-01 -5.69055140e-01
-7.18257427e-01 7.81001031e-01 3.33655268e-01 -2.15611100e-01
-1.03770721e+00 -6.12328708e-01 -8.72142553e-01 2.65798658e-01
2.29911774e-01 -5.52526951e-01 1.09433234e+00 -8.23787808e-01
-7.93219030e-01 1.00028777e+00 1.38352364e-01 -2.40212716e-02
6.02495134e-01 3.35482985e-01 -5.95337749e-01 2.21885145e-01
9.77626815e-02 5.15234590e-01 1.16063845e+00 -1.38717020e+00
-4.62168276e-01 -4.73866791e-01 1.07788600e-01 6.35662153e-02
-6.88363552e-01 -3.86798054e-01 -8.18978369e-01 -6.10814512e-01
-7.98407756e-03 -8.79246831e-01 3.23383659e-01 5.62506504e-02
-7.01635897e-01 -2.87600130e-01 1.37999165e+00 -4.40503359e-01
9.65777040e-01 -2.13299775e+00 -7.27779940e-02 3.16149950e-01
3.86434346e-01 7.22211227e-02 -4.35540795e-01 5.87123632e-01
-1.20364269e-03 2.70610958e-01 1.82083637e-01 -4.25946444e-01
1.67332545e-01 1.78491533e-01 -3.79113853e-01 5.16195595e-01
2.93691486e-01 7.48748720e-01 -8.38971257e-01 -5.73311090e-01
1.36956185e-01 4.73025233e-01 -2.35437855e-01 1.93466648e-01
-4.52384561e-01 -5.95581830e-02 -2.58965105e-01 8.98411632e-01
7.65278101e-01 -7.31345177e-01 6.20449543e-01 -2.43372291e-01
6.07534587e-01 -1.49841622e-01 -7.76290894e-01 1.56138778e+00
-3.75259042e-01 1.08192539e+00 -7.25926757e-02 -1.25543678e+00
9.07448530e-01 -8.74342322e-02 -1.60649061e-01 -1.03539062e+00
9.47872102e-02 6.48061410e-02 8.79791752e-02 -4.34487939e-01
6.96836770e-01 3.94208282e-02 -7.34839737e-02 3.89264733e-01
2.84799606e-01 -1.84710130e-01 5.18303573e-01 6.28445387e-01
1.31130362e+00 -1.62618637e-01 -1.00266710e-01 -1.52309015e-01
2.13118210e-01 -8.49132687e-02 -1.25844643e-01 8.69080544e-01
1.77039690e-02 8.63413692e-01 9.21262443e-01 -3.20649058e-01
-1.07533622e+00 -1.25363278e+00 -3.16158943e-02 1.07649493e+00
1.27511650e-01 -5.18767416e-01 -5.28861701e-01 -1.18122160e+00
3.93455297e-01 6.60841227e-01 -6.20949090e-01 -7.89533183e-02
-1.54989272e-01 -5.97132981e-01 6.80998147e-01 7.48243988e-01
1.19874820e-01 -6.16890550e-01 -2.91478764e-02 -5.62153161e-02
9.56995413e-02 -1.34174252e+00 -3.21859330e-01 2.21387342e-01
-4.50768024e-01 -1.30978191e+00 -8.75837922e-01 -1.03564262e+00
9.82669234e-01 3.98988366e-01 1.52803648e+00 1.78180188e-01
-4.83275086e-01 7.55810380e-01 -3.44863176e-01 -2.65323341e-01
-5.21464586e-01 1.79704040e-01 -4.12037939e-01 -1.77684426e-01
4.51474220e-01 -2.91926444e-01 -3.17186236e-01 3.25209379e-01
-1.01964211e+00 -2.81923056e-01 5.60055077e-01 1.04445350e+00
3.98349196e-01 3.38193655e-01 2.81377256e-01 -9.88338232e-01
1.03291416e+00 -4.77380216e-01 -5.19760549e-01 6.55385673e-01
-7.19695807e-01 1.58091113e-01 9.72373605e-01 -6.38976753e-01
-5.00660717e-01 -4.66406405e-01 3.40232909e-01 -7.61991084e-01
-5.20656630e-02 5.42889118e-01 -1.45832658e-01 1.17086515e-01
6.95388019e-01 2.43315697e-01 -9.28033143e-02 -4.29580122e-01
3.75420809e-01 7.99508870e-01 5.16972542e-01 -6.88322604e-01
8.87339056e-01 3.85350436e-01 6.26174659e-02 -1.01610243e+00
-7.91496158e-01 -4.28519845e-01 -4.93274540e-01 -2.18248293e-01
6.17055118e-01 -8.42388034e-01 -3.96882027e-01 5.42004049e-01
-9.61219788e-01 -4.03252363e-01 6.84667602e-02 6.77641630e-02
-1.93534389e-01 5.43318748e-01 -3.59982342e-01 -7.70551741e-01
-1.80102494e-02 -9.95039165e-01 1.25522876e+00 -3.02496534e-02
-2.40999684e-02 -1.47086251e+00 -1.74755931e-01 1.61731601e-01
3.16551715e-01 2.14845493e-01 1.40188766e+00 -8.54630649e-01
-5.71055114e-01 -3.90889555e-01 -5.83142579e-01 3.99354458e-01
1.15235366e-01 2.36280158e-01 -1.11365104e+00 -7.60752678e-01
-8.22326958e-01 -5.87865770e-01 1.01965642e+00 2.60431528e-01
1.55035520e+00 2.52689198e-02 -5.91187119e-01 5.71306884e-01
1.49844241e+00 -6.69242516e-02 5.49199224e-01 3.57316226e-01
1.03074861e+00 3.29375386e-01 3.41148734e-01 1.56384364e-01
4.34215903e-01 5.32852173e-01 5.60893059e-01 -2.13718668e-01
-5.59112191e-01 -4.45286453e-01 1.99057788e-01 4.45157617e-01
4.47801292e-01 -1.19348562e+00 -1.13881075e+00 7.77442515e-01
-1.64580464e+00 -8.07076335e-01 -4.21563275e-02 1.93572628e+00
3.60646307e-01 3.43071967e-01 2.01693892e-01 2.25140229e-01
7.79512167e-01 3.25517535e-01 -4.49482262e-01 -5.40608168e-01
-3.85800332e-01 1.47179857e-01 4.63629127e-01 4.83362377e-02
-1.06819797e+00 7.68108010e-01 6.43059158e+00 1.10248029e+00
-1.27877605e+00 -2.31281921e-01 6.26847148e-01 9.18286368e-02
-5.08772850e-01 -1.91601664e-01 -7.13877618e-01 5.09611011e-01
7.78860390e-01 7.10373521e-02 3.71168435e-01 1.11265266e+00
-5.63725531e-01 3.18651572e-02 -1.40909421e+00 1.22531474e+00
7.48088598e-01 -1.36063397e+00 3.26256871e-01 9.57002118e-02
3.93651098e-01 1.13300316e-01 2.00481996e-01 3.25841963e-01
3.91619742e-01 -1.25973427e+00 1.01678264e+00 -5.62190413e-02
1.22552490e+00 -6.48044527e-01 5.20391941e-01 2.16012776e-01
-1.06575823e+00 2.88197976e-02 -5.19833207e-01 3.44528586e-01
-3.04067671e-01 6.89949870e-01 -1.17076159e+00 4.53291088e-01
6.82982087e-01 7.41518259e-01 -1.17813742e+00 8.93901169e-01
-3.51905197e-01 4.83992457e-01 -1.17912866e-01 -3.00771236e-01
2.84231037e-01 8.37913901e-02 1.63193122e-01 1.43038237e+00
4.82211202e-01 -6.87863469e-01 4.34365720e-01 7.01467335e-01
-3.27524304e-01 -1.79975659e-01 -8.90314043e-01 -6.12669468e-01
4.08052862e-01 1.32185364e+00 -9.59124804e-01 -3.20564151e-01
-4.36009616e-01 1.03146374e+00 6.01308286e-01 5.41377068e-01
-7.10258842e-01 -5.98526239e-01 3.13863426e-01 2.36803591e-01
6.52692258e-01 -2.17136994e-01 1.90638974e-01 -1.23436320e+00
2.65856117e-01 -9.18260753e-01 6.67346537e-01 -9.50057566e-01
-1.69195533e+00 6.90883160e-01 -1.26393870e-01 -1.29745710e+00
-3.27572286e-01 -1.28436995e+00 -3.96290243e-01 6.38314724e-01
-1.56200659e+00 -1.60023034e+00 -5.32313466e-01 5.74660242e-01
2.39139974e-01 -3.93980861e-01 8.52763772e-01 2.97370672e-01
-4.02817965e-01 8.92476797e-01 3.94023478e-01 5.17857254e-01
7.46555030e-01 -1.63138247e+00 5.34673989e-01 6.54591382e-01
5.28099954e-01 2.60812730e-01 5.05344450e-01 -4.39038038e-01
-1.64109051e+00 -1.09334826e+00 5.95224500e-01 -6.30535185e-01
7.63174832e-01 -8.95238578e-01 -8.61714184e-01 8.73591959e-01
1.39686644e-01 3.70469749e-01 6.49178684e-01 3.53880316e-01
-1.03522575e+00 8.88786465e-02 -1.08001041e+00 4.94424880e-01
1.06293118e+00 -8.26911330e-01 -4.11442161e-01 5.89001834e-01
1.03728496e-01 -5.02260268e-01 -6.50214016e-01 -1.22748271e-01
5.57143927e-01 -9.72702742e-01 9.82223034e-01 -6.72384560e-01
5.05597472e-01 -2.58654565e-01 -1.63327321e-01 -1.61364055e+00
-4.77653801e-01 -1.73000693e-01 -1.66021287e-01 1.42226851e+00
4.38147217e-01 -5.84064364e-01 8.70543182e-01 1.61134470e-02
-1.42803369e-02 -5.16865909e-01 -5.59512019e-01 -1.24891675e+00
5.93641475e-02 -4.54931825e-01 4.37473178e-01 1.07731688e+00
-1.47140399e-01 5.87144852e-01 -1.07615646e-02 9.05934647e-02
6.81503296e-01 2.39385605e-01 8.60776782e-01 -1.24762094e+00
-3.12149614e-01 -4.90441382e-01 -8.87233496e-01 -9.61817920e-01
3.68421525e-01 -1.28982949e+00 -2.78282702e-01 -1.90932858e+00
2.78916061e-01 -4.66903508e-01 -7.24200755e-02 5.21930575e-01
1.28015995e-01 3.38179141e-01 2.32891589e-01 8.91379267e-02
-6.20087683e-01 3.04892629e-01 1.16551340e+00 -8.60072851e-01
3.57981116e-01 -4.98871654e-01 -6.33403301e-01 4.50728983e-01
4.86542583e-01 -3.12138587e-01 -7.84849226e-01 -6.94009602e-01
1.38351768e-01 -3.31429332e-01 4.71060425e-01 -9.55902934e-01
1.10503687e-02 7.98954517e-02 8.84541810e-01 -8.49955380e-01
2.89487988e-01 -8.08691084e-01 -2.40054920e-01 1.22105561e-01
-1.81402132e-01 3.56641054e-01 2.74760723e-01 7.52286494e-01
-2.34869719e-01 -2.24316865e-01 5.62572360e-01 6.93170652e-02
-7.86333203e-01 2.93669194e-01 -2.70779133e-01 5.02926171e-01
8.76186609e-01 -4.15554762e-01 -1.09772885e+00 -4.89928782e-01
-3.34202677e-01 9.80640352e-02 7.70201921e-01 6.83584750e-01
5.05652726e-01 -1.23360419e+00 -5.08930206e-01 4.60223228e-01
6.28960073e-01 -1.84674412e-01 1.76989496e-01 4.00406867e-01
-6.08794391e-01 2.16596812e-01 9.54719633e-03 -1.05346322e+00
-1.33175397e+00 7.18150079e-01 2.23753825e-01 -4.12585735e-01
-6.00677133e-01 8.44310403e-01 2.81267792e-01 -3.87906998e-01
2.74870574e-01 -3.21591228e-01 1.45504534e-01 2.57541448e-01
4.13597256e-01 6.95618941e-03 3.61527741e-01 -3.69328260e-01
-5.28354287e-01 5.23956299e-01 -3.12572479e-01 1.28036961e-01
1.00034690e+00 3.08294773e-01 1.36556700e-01 2.23764852e-01
1.51289046e+00 -6.48302510e-02 -8.82847905e-01 -2.76068635e-02
-5.14793769e-02 -6.94239497e-01 7.35462308e-02 -8.67748559e-01
-1.05464494e+00 1.10815334e+00 3.56108397e-01 8.04787993e-01
8.88064325e-01 1.24400109e-01 4.37270880e-01 3.80660295e-01
2.21756622e-02 -1.22661746e+00 4.64124262e-01 4.24895585e-01
8.21973681e-01 -1.15052855e+00 1.19432546e-01 -2.87120074e-01
-5.52583754e-01 1.31786823e+00 5.71607769e-01 -2.60009896e-02
5.35518169e-01 4.21865582e-01 8.32637325e-02 -4.03535813e-01
-7.48969555e-01 -7.51677677e-02 4.22170132e-01 1.16136909e+00
3.79135758e-01 -5.58684766e-02 4.15923417e-01 -2.71425694e-02
-1.92618102e-01 -4.14324284e-01 4.61282700e-01 8.87069046e-01
-9.20210704e-02 -1.12021482e+00 -2.87968844e-01 6.77018344e-01
-1.56171128e-01 -5.50723970e-02 -6.92965090e-01 1.20129240e+00
-2.59804934e-01 9.07093048e-01 4.47632641e-01 -3.87026668e-01
3.06659132e-01 4.98617813e-02 8.57236445e-01 -5.12026310e-01
-1.69281334e-01 -6.07510135e-02 3.10020864e-01 -2.68739164e-01
-3.18715945e-02 -3.91596913e-01 -8.98973286e-01 -4.19524223e-01
-3.37149709e-01 -3.34647633e-02 7.73862243e-01 6.43626809e-01
4.22951758e-01 7.44924784e-01 3.84080768e-01 -6.31746531e-01
-7.12296367e-01 -7.60211706e-01 -9.11277115e-01 7.59534419e-01
3.51085991e-01 -6.65969551e-01 -4.72355604e-01 -1.60997137e-01] | [11.50561237335205, 2.4737002849578857] |
c54a914f-b6bf-4a37-ae4e-72058e90fc06 | prompt-free-diffusion-taking-text-out-of-text | 2305.16223 | null | https://arxiv.org/abs/2305.16223v2 | https://arxiv.org/pdf/2305.16223v2.pdf | Prompt-Free Diffusion: Taking "Text" out of Text-to-Image Diffusion Models | Text-to-image (T2I) research has grown explosively in the past year, owing to the large-scale pre-trained diffusion models and many emerging personalization and editing approaches. Yet, one pain point persists: the text prompt engineering, and searching high-quality text prompts for customized results is more art than science. Moreover, as commonly argued: "an image is worth a thousand words" - the attempt to describe a desired image with texts often ends up being ambiguous and cannot comprehensively cover delicate visual details, hence necessitating more additional controls from the visual domain. In this paper, we take a bold step forward: taking "Text" out of a pre-trained T2I diffusion model, to reduce the burdensome prompt engineering efforts for users. Our proposed framework, Prompt-Free Diffusion, relies on only visual inputs to generate new images: it takes a reference image as "context", an optional image structural conditioning, and an initial noise, with absolutely no text prompt. The core architecture behind the scene is Semantic Context Encoder (SeeCoder), substituting the commonly used CLIP-based or LLM-based text encoder. The reusability of SeeCoder also makes it a convenient drop-in component: one can also pre-train a SeeCoder in one T2I model and reuse it for another. Through extensive experiments, Prompt-Free Diffusion is experimentally found to (i) outperform prior exemplar-based image synthesis approaches; (ii) perform on par with state-of-the-art T2I models using prompts following the best practice; and (iii) be naturally extensible to other downstream applications such as anime figure generation and virtual try-on, with promising quality. Our code and models are open-sourced at https://github.com/SHI-Labs/Prompt-Free-Diffusion. | ['Humphrey Shi', 'Irfan Essa', 'Gao Huang', 'Zhangyang Wang', 'Jiayi Guo', 'Xingqian Xu'] | 2023-05-25 | null | null | null | null | ['image-variation', 'conditional-text-to-image-synthesis', 'virtual-try-on', 'prompt-engineering'] | ['computer-vision', 'computer-vision', 'computer-vision', 'natural-language-processing'] | [ 5.71526349e-01 -9.28391470e-04 1.58230305e-01 -2.64769644e-01
-5.82511246e-01 -6.32264137e-01 8.93818438e-01 -2.18121901e-01
-3.14653426e-01 5.26394129e-01 2.73023069e-01 -4.57448065e-01
1.82566062e-01 -5.71420789e-01 -8.22924376e-01 -5.25585175e-01
4.13988978e-01 3.62907290e-01 2.51995862e-01 -2.46439293e-01
3.55456114e-01 3.25597644e-01 -1.48746502e+00 3.75910848e-01
7.18031824e-01 7.94579864e-01 7.10254431e-01 7.42448926e-01
-3.39314938e-01 8.44846487e-01 -6.65376663e-01 -6.04752958e-01
1.44125074e-01 -7.39945948e-01 -7.81199694e-01 2.71322787e-01
5.44678271e-01 -3.96265775e-01 -3.59618038e-01 8.57064128e-01
5.08572519e-01 9.56657063e-03 4.68645483e-01 -1.07685602e+00
-1.09834361e+00 3.58197570e-01 -6.07555270e-01 -8.86508971e-02
4.50253338e-01 4.94366378e-01 7.86816537e-01 -1.11043036e+00
1.09759581e+00 1.09193528e+00 4.11754698e-01 7.39287198e-01
-1.29850900e+00 -5.19578040e-01 2.24557281e-01 1.41986862e-01
-1.28921258e+00 -3.90614569e-01 8.13891172e-01 -4.14753973e-01
9.66313660e-01 2.86442846e-01 8.71653438e-01 1.53809595e+00
1.90727308e-01 8.16144943e-01 1.13070452e+00 -5.02325654e-01
1.91179261e-01 2.52688795e-01 -4.44493264e-01 6.87635064e-01
-1.03594311e-01 1.59850270e-01 -5.05919993e-01 3.23180407e-01
9.08787787e-01 2.65172012e-02 -3.31490070e-01 -1.00083008e-01
-1.29840660e+00 6.57784998e-01 2.27014467e-01 3.72397542e-01
-3.01659852e-01 3.26913297e-01 3.08784962e-01 4.37880993e-01
4.38356698e-01 3.59962076e-01 -1.81920737e-01 -2.45324746e-01
-1.23948705e+00 4.60637212e-01 5.64521372e-01 1.20314336e+00
7.87182927e-01 1.89755619e-01 -2.97824115e-01 8.68583083e-01
2.44096182e-02 6.04809284e-01 3.47494423e-01 -1.02641928e+00
3.24196905e-01 3.86357397e-01 3.86830792e-02 -8.42524052e-01
-4.71235886e-02 -3.69259834e-01 -8.04615736e-01 5.60168266e-01
3.08153540e-01 -1.62898555e-01 -1.16311300e+00 1.63157880e+00
2.01226547e-01 -3.76030691e-02 -2.56140292e-01 8.89423251e-01
6.89018548e-01 9.17935848e-01 1.29380599e-01 1.32550836e-01
1.25820065e+00 -1.05692184e+00 -5.42329609e-01 -3.83836538e-01
3.95490289e-01 -1.08196020e+00 1.54829872e+00 4.35641736e-01
-1.18660820e+00 -5.94527185e-01 -1.03517616e+00 -3.55553240e-01
-4.53539759e-01 2.00644180e-01 5.18748879e-01 4.76824284e-01
-1.33920467e+00 4.85904604e-01 -3.31634849e-01 -5.13224721e-01
3.97472322e-01 -2.08375957e-02 -3.59715819e-01 -3.11267167e-01
-8.50445211e-01 7.91427851e-01 1.35544613e-01 -3.15692239e-02
-1.02162254e+00 -8.13705206e-01 -7.21332610e-01 -8.81152377e-02
7.09054291e-01 -1.03517044e+00 1.23087859e+00 -1.42376697e+00
-1.63696539e+00 9.02777970e-01 -2.53004909e-01 -1.00007348e-01
6.82000220e-01 -1.98584646e-01 -2.03277826e-01 2.17824548e-01
1.32770076e-01 9.82062221e-01 1.30307925e+00 -1.58457720e+00
-4.51841354e-01 4.44915667e-02 2.10335732e-01 2.38801822e-01
-1.36280507e-01 1.16183743e-01 -8.09189618e-01 -9.89352345e-01
-3.91353071e-01 -9.52668488e-01 -1.76335663e-01 3.42153609e-01
-3.69549245e-01 9.21193436e-02 1.08925521e+00 -6.55005753e-01
1.19146574e+00 -2.16989684e+00 7.03238174e-02 -4.97988649e-02
1.74870193e-01 3.22372913e-01 -4.52826619e-01 6.43848479e-01
-1.66674823e-01 1.94007590e-01 -3.27187538e-01 -4.57712084e-01
-1.73289832e-02 1.89217523e-01 -3.70338649e-01 8.36984161e-03
2.91522175e-01 1.12300646e+00 -9.18909729e-01 -5.48686743e-01
4.73003656e-01 6.27496302e-01 -7.06906497e-01 1.08234406e-01
-5.24242818e-01 4.93387967e-01 -9.39315110e-02 4.11413550e-01
4.14917231e-01 -4.09352422e-01 8.14408213e-02 -2.77346492e-01
-2.37153411e-01 5.03297634e-02 -1.11145556e+00 2.03395176e+00
-6.50619328e-01 8.71537983e-01 1.85870513e-01 -8.68150473e-01
7.77437210e-01 2.81302184e-01 2.17258438e-01 -8.49975169e-01
2.76480224e-02 2.23321870e-01 -2.40798786e-01 -5.94497859e-01
6.23675346e-01 5.63015230e-03 9.02559832e-02 3.98439288e-01
1.65381104e-01 -5.28073192e-01 3.56533706e-01 5.07318437e-01
8.37032318e-01 5.99916577e-01 -3.00216079e-02 -1.34404063e-01
3.50481808e-01 -1.15521904e-02 2.28795126e-01 7.73520827e-01
1.02381602e-01 9.28226292e-01 4.16225612e-01 -2.26054966e-01
-1.37714982e+00 -9.75315571e-01 2.34880134e-01 9.44603324e-01
1.67697445e-01 -4.56962228e-01 -8.33197236e-01 -5.14327645e-01
-3.40433002e-01 9.39829290e-01 -6.52782619e-01 2.16143191e-01
-4.79588985e-01 -1.81160033e-01 3.12817991e-01 4.81578082e-01
4.65225965e-01 -1.22443295e+00 -6.20711625e-01 2.97648430e-01
-1.39308572e-01 -9.30953562e-01 -8.19080889e-01 1.43362209e-01
-6.52470291e-01 -6.55662000e-01 -1.11517406e+00 -7.98574448e-01
7.51401365e-01 3.86222214e-01 1.07422173e+00 2.63303220e-01
-5.15264690e-01 5.31841397e-01 -4.43993986e-01 -3.53527248e-01
-5.76594055e-01 -1.82671458e-01 -4.99488324e-01 -1.85142942e-02
-1.04441442e-01 -6.19078040e-01 -9.20135736e-01 1.88671976e-01
-1.30666518e+00 6.13821924e-01 6.53667390e-01 9.13939178e-01
4.87265706e-01 -1.11328192e-01 4.06114221e-01 -9.40438330e-01
5.58136106e-01 -9.44627523e-02 -4.47298348e-01 2.52906412e-01
-7.00203300e-01 -6.51423112e-02 7.20541179e-01 -4.83691990e-01
-1.31963527e+00 -1.30952090e-01 -2.18266308e-01 -4.22950357e-01
-1.48740366e-01 3.58206779e-01 -6.83880523e-02 1.45039633e-01
6.61641657e-01 3.34255487e-01 -1.47479381e-02 -2.87402898e-01
8.12298238e-01 4.65926439e-01 5.38760006e-01 -6.81021273e-01
8.63271475e-01 4.97992784e-01 -2.15973973e-01 -9.25247610e-01
-5.65988958e-01 -5.12347110e-02 -4.02171940e-01 -3.53163958e-01
9.43877637e-01 -7.37613022e-01 -5.15629649e-01 4.66899693e-01
-1.28883970e+00 -8.80284786e-01 -3.71481299e-01 1.80351902e-02
-5.43443382e-01 4.54710782e-01 -5.49558938e-01 -4.66395885e-01
-3.75523031e-01 -1.34386885e+00 1.08510518e+00 3.28410238e-01
-3.57698679e-01 -9.76820529e-01 -2.67438948e-01 3.07613403e-01
7.10708559e-01 2.73630116e-02 9.70308661e-01 -1.05805315e-01
-7.60035992e-01 4.23267484e-02 -4.80120808e-01 4.26938713e-01
-6.42415136e-02 2.54348129e-01 -9.68575776e-01 -1.06370844e-01
-2.19311491e-01 -3.26483279e-01 6.59938455e-01 2.94796437e-01
1.12819767e+00 -2.92222321e-01 -1.74995437e-01 5.89996696e-01
1.53817451e+00 3.48070502e-01 8.55814099e-01 3.16938072e-01
7.08300173e-01 5.96692979e-01 4.13006246e-01 3.70971411e-01
3.87025535e-01 6.52726352e-01 2.52202988e-01 -3.66079897e-01
-6.97789311e-01 -5.28743863e-01 2.77682245e-01 8.22904348e-01
6.10874826e-03 -3.21933895e-01 -7.22647727e-01 6.93526328e-01
-1.67056644e+00 -1.13900435e+00 -1.53452996e-02 1.96430695e+00
9.37659323e-01 2.98947357e-02 -1.06636465e-01 -1.01365492e-01
4.38174784e-01 2.36255929e-01 -5.43836355e-01 -4.78253841e-01
-1.43345624e-01 3.89109254e-01 3.20704609e-01 6.29880607e-01
-7.82219112e-01 1.17784524e+00 5.75137377e+00 1.00910914e+00
-1.41920209e+00 1.21565469e-01 7.26898432e-01 -1.60347387e-01
-6.24464810e-01 3.10940713e-01 -5.30349672e-01 4.63474900e-01
6.09659135e-01 -2.18077302e-01 6.91706359e-01 7.70914078e-01
4.32420611e-01 -2.51479745e-01 -1.06213248e+00 1.14823961e+00
9.44797993e-02 -1.52774668e+00 2.69958466e-01 -1.06126614e-01
7.43187487e-01 -1.53589934e-01 2.59469986e-01 2.46931061e-01
2.67310619e-01 -9.86413836e-01 1.08914554e+00 4.04948652e-01
1.25226784e+00 -3.88614476e-01 1.82257220e-01 1.05548717e-01
-1.00573444e+00 9.06963870e-02 -2.61609852e-01 1.29750043e-01
3.11118990e-01 5.31876326e-01 -5.71287632e-01 5.60317278e-01
7.68542528e-01 5.64279377e-01 -6.10124886e-01 7.35865355e-01
-3.98305476e-01 5.81122398e-01 -1.64565101e-01 -1.14351556e-01
2.97207236e-01 -1.63979843e-01 3.59739959e-01 1.53875363e+00
3.96783352e-01 1.14284143e-01 7.30296373e-02 1.05117476e+00
-5.02112880e-02 1.23764552e-01 -7.01983988e-01 -1.53382182e-01
3.81442368e-01 1.26026416e+00 -7.88207233e-01 -6.00950241e-01
-6.38369322e-01 1.61038578e+00 -8.01694521e-04 6.36520207e-01
-9.28581834e-01 -5.85307360e-01 3.59807223e-01 1.43867299e-01
4.71594930e-01 -2.35537857e-01 -4.34870720e-01 -1.15072608e+00
2.23444658e-03 -1.14441991e+00 -1.70527220e-01 -1.38368261e+00
-1.13136816e+00 6.84323192e-01 -2.14672238e-02 -1.08578086e+00
-1.95067853e-01 -5.45812905e-01 -5.77137947e-01 9.60468054e-01
-1.53294528e+00 -1.29654121e+00 -4.83184963e-01 7.00606287e-01
9.25028086e-01 2.33371228e-01 5.75023651e-01 4.62048531e-01
-2.78508157e-01 5.76954007e-01 -5.08318581e-02 -7.68823028e-02
8.70299637e-01 -1.10262084e+00 5.83392859e-01 9.03436363e-01
1.68689191e-01 6.24436438e-01 7.47981489e-01 -7.25790381e-01
-1.50043046e+00 -9.81763601e-01 8.29980969e-01 -4.06648904e-01
7.00400710e-01 -5.03156066e-01 -7.97707677e-01 5.99208236e-01
6.43692911e-01 -3.98189247e-01 3.03350687e-01 -3.94383550e-01
-3.16362500e-01 -9.89829823e-02 -7.19466984e-01 1.11509311e+00
1.22054851e+00 -5.59303045e-01 -2.73322374e-01 2.63547719e-01
7.45960951e-01 -3.99514228e-01 -5.37244081e-01 -6.49006516e-02
5.35987735e-01 -1.12252247e+00 9.07116890e-01 -2.42938213e-02
8.00501347e-01 -3.28558028e-01 -2.08682846e-02 -1.26120031e+00
-2.51741052e-01 -1.03236401e+00 3.60227674e-01 1.40290511e+00
4.91563082e-01 -4.26741093e-01 6.51733696e-01 6.64164662e-01
-2.00410262e-01 -6.71175599e-01 -5.93757749e-01 -6.80546105e-01
-2.93113869e-02 -6.61046743e-01 3.28473061e-01 8.54195714e-01
-1.78209946e-01 5.11187851e-01 -5.38757503e-01 -2.57684827e-01
4.26776618e-01 -3.38263623e-02 1.00072312e+00 -8.06054950e-01
-5.59201837e-01 -6.54956102e-01 6.18353337e-02 -1.45227933e+00
-3.10419321e-01 -8.67276609e-01 2.36264393e-02 -1.66815782e+00
1.29599363e-01 -3.72563958e-01 3.16721201e-01 6.29636049e-01
-3.13628644e-01 2.68546313e-01 5.04795074e-01 1.61970913e-01
-2.72479147e-01 5.58955789e-01 1.57495308e+00 -2.15243876e-01
-1.51702881e-01 -5.37032485e-01 -7.88781345e-01 4.98879701e-01
5.74285328e-01 -2.70703048e-01 -7.91099608e-01 -5.96367776e-01
3.32646936e-01 2.93370280e-02 5.45079827e-01 -8.31441164e-01
1.39712512e-01 -1.81979209e-01 3.68635982e-01 -4.13370878e-01
3.25770587e-01 -8.42058182e-01 3.27688575e-01 1.46587923e-01
-3.23607773e-01 1.38819411e-01 2.12857023e-01 4.19419557e-01
-4.31894697e-02 -3.50926608e-01 6.85811758e-01 -2.64561146e-01
-8.75193298e-01 2.42662817e-01 -4.22011256e-01 2.92841196e-02
7.78110623e-01 -4.60786253e-01 -3.14384848e-01 -6.59591377e-01
-4.77016270e-01 -1.00569427e-01 7.89826393e-01 4.44037706e-01
7.12662160e-01 -1.02963173e+00 -6.32938683e-01 1.22262433e-01
9.55245346e-02 -7.98230246e-02 4.32774156e-01 7.37605929e-01
-5.43928206e-01 1.56136006e-01 1.01291895e-01 -5.40900171e-01
-1.02625954e+00 7.96196342e-01 -6.79363832e-02 -1.46741435e-01
-9.88876522e-01 8.20446730e-01 6.34440362e-01 -5.44825047e-02
1.42420530e-01 -2.86563784e-01 3.55456799e-01 -2.80600011e-01
6.06680930e-01 1.26285730e-02 -1.88923314e-01 -2.47972205e-01
2.64164526e-02 5.70271730e-01 -2.23238394e-01 -5.72455466e-01
1.20027328e+00 -3.55419427e-01 1.78491727e-01 2.97006130e-01
1.11369896e+00 -1.39347002e-01 -1.51041281e+00 -1.28626987e-01
-2.28496656e-01 -4.94985819e-01 -4.00150716e-02 -1.13867438e+00
-9.53644931e-01 1.02291393e+00 3.36100519e-01 7.79729038e-02
1.29119372e+00 -9.18506533e-02 8.58544767e-01 8.38424414e-02
2.04791367e-01 -1.06393659e+00 4.28491265e-01 3.02642882e-01
1.21445608e+00 -9.86199260e-01 -1.66188762e-01 -2.50864387e-01
-1.05545342e+00 1.06887615e+00 3.93304050e-01 -6.59624115e-02
3.92486334e-01 4.71087992e-01 3.19587708e-01 -2.23329335e-01
-9.00156617e-01 -8.12149942e-02 4.89275157e-02 8.47534001e-01
6.13240480e-01 -2.56176859e-01 -2.34153032e-01 2.19118401e-01
-1.06123937e-02 2.74049401e-01 5.63171744e-01 7.84145832e-01
-2.34323099e-01 -1.22554314e+00 -3.20334584e-01 2.43799254e-01
-2.69742787e-01 -3.87917697e-01 -2.78140068e-01 8.67669761e-01
-4.50736918e-02 9.33536649e-01 -1.52158827e-01 -2.00948402e-01
3.30323786e-01 1.05531611e-01 6.07480288e-01 -4.60833549e-01
-5.90623856e-01 2.09653214e-01 3.42533253e-02 -6.27404273e-01
-2.33502910e-01 -4.93354052e-01 -1.10465002e+00 -4.13492948e-01
-6.90399855e-02 -3.33132803e-01 9.84402299e-01 7.72473276e-01
6.13553822e-01 5.17472088e-01 4.04893667e-01 -1.10085666e+00
-1.51388898e-01 -8.22283387e-01 -2.53187537e-01 6.16069257e-01
6.52359948e-02 -4.42405820e-01 -1.56560808e-01 5.97123921e-01] | [11.306215286254883, -0.20923607051372528] |
0f803a30-6db0-49e6-a22e-299a25d128e4 | modar-using-motion-forecasting-for-3d-object-1 | 2306.03206 | null | https://arxiv.org/abs/2306.03206v1 | https://arxiv.org/pdf/2306.03206v1.pdf | MoDAR: Using Motion Forecasting for 3D Object Detection in Point Cloud Sequences | Occluded and long-range objects are ubiquitous and challenging for 3D object detection. Point cloud sequence data provide unique opportunities to improve such cases, as an occluded or distant object can be observed from different viewpoints or gets better visibility over time. However, the efficiency and effectiveness in encoding long-term sequence data can still be improved. In this work, we propose MoDAR, using motion forecasting outputs as a type of virtual modality, to augment LiDAR point clouds. The MoDAR modality propagates object information from temporal contexts to a target frame, represented as a set of virtual points, one for each object from a waypoint on a forecasted trajectory. A fused point cloud of both raw sensor points and the virtual points can then be fed to any off-the-shelf point-cloud based 3D object detector. Evaluated on the Waymo Open Dataset, our method significantly improves prior art detectors by using motion forecasting from extra-long sequences (e.g. 18 seconds), achieving new state of the arts, while not adding much computation overhead. | ['Dragomir Anguelov', 'Chenxi Liu', 'Yin Zhou', 'Charles R. Qi', 'Yingwei Li'] | 2023-06-05 | modar-using-motion-forecasting-for-3d-object | http://openaccess.thecvf.com//content/CVPR2023/html/Li_MoDAR_Using_Motion_Forecasting_for_3D_Object_Detection_in_Point_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Li_MoDAR_Using_Motion_Forecasting_for_3D_Object_Detection_in_Point_CVPR_2023_paper.pdf | cvpr-2023-1 | ['3d-object-detection', 'motion-forecasting'] | ['computer-vision', 'computer-vision'] | [ 3.42990495e-02 -3.53397906e-01 -1.87143058e-01 -2.49898389e-01
-8.13010812e-01 -8.15395832e-01 7.64482021e-01 1.44529700e-01
-3.13795477e-01 2.17352346e-01 -2.21085161e-01 -1.35189891e-01
2.02537134e-01 -8.83232653e-01 -8.15658569e-01 -4.45170462e-01
-1.49832338e-01 8.51114392e-01 9.80090499e-01 -1.59585193e-01
5.05027957e-02 1.08045185e+00 -1.96819413e+00 2.40918860e-01
4.71756488e-01 1.25438976e+00 6.94336057e-01 7.75749624e-01
-2.43804947e-01 3.81309897e-01 -3.12214494e-01 -7.14423181e-03
4.00179386e-01 4.87447828e-01 -3.66002768e-02 7.78046995e-02
9.09188986e-01 -6.27058208e-01 -5.08477330e-01 6.60018802e-01
1.97952524e-01 2.03592539e-01 3.34447145e-01 -1.44427049e+00
-1.39165133e-01 -2.36235838e-02 -6.22395098e-01 3.26444149e-01
6.09211922e-01 4.30104077e-01 7.00174510e-01 -1.29438591e+00
8.47487688e-01 1.38024759e+00 5.82512081e-01 3.09309095e-01
-8.57152641e-01 -7.28870690e-01 4.21364456e-01 3.93238693e-01
-1.46158624e+00 -3.62305552e-01 6.73534095e-01 -6.08705699e-01
1.21874821e+00 2.77804077e-01 7.76750386e-01 1.00845909e+00
-5.21727167e-02 8.01245570e-01 4.11766917e-01 2.04219967e-01
1.69976085e-01 -2.31156901e-01 -1.16045279e-02 5.08046746e-01
2.44006187e-01 4.80008155e-01 -8.60376358e-01 -2.11197749e-01
4.50398088e-01 5.22697091e-01 -2.38120452e-01 -6.47602975e-01
-1.40497899e+00 5.42711437e-01 7.16938257e-01 -2.42556170e-01
-6.16474211e-01 3.59806001e-01 -2.20841030e-03 8.75922963e-02
6.50167763e-01 -3.35524052e-01 -5.01329422e-01 -2.17050493e-01
-9.85889256e-01 5.25011837e-01 1.80699557e-01 1.24901927e+00
8.72950554e-01 -8.49542301e-03 -5.38218953e-02 1.10093869e-01
6.53479934e-01 1.31789637e+00 -2.35505059e-01 -1.04739511e+00
7.59424090e-01 6.94433153e-01 4.23151821e-01 -8.28898251e-01
-3.11000437e-01 -2.89800853e-01 -3.25527132e-01 5.13564706e-01
8.67119208e-02 3.72492760e-01 -1.00378060e+00 1.26440418e+00
9.76555228e-01 7.24712789e-01 -2.47847706e-01 1.33431065e+00
7.50042379e-01 9.20853257e-01 -1.56932220e-01 1.36403427e-01
1.33874488e+00 -6.08868718e-01 -3.54684144e-01 -4.31002021e-01
5.80492318e-01 -7.69689679e-01 5.57036638e-01 3.73864323e-01
-8.64503205e-01 -6.62089169e-01 -9.57986712e-01 -1.01816416e-01
-3.24146867e-01 -2.18892485e-01 3.86678338e-01 2.84459263e-01
-8.81059766e-01 3.48766148e-01 -1.17141449e+00 -2.46873468e-01
5.78044832e-01 2.22706571e-01 -3.11560273e-01 -5.01318812e-01
-6.29280984e-01 8.02197576e-01 4.80156168e-02 -1.06850304e-01
-1.00632262e+00 -1.10429275e+00 -8.00042450e-01 -4.33865599e-02
4.94535029e-01 -9.65982199e-01 1.12675703e+00 -1.37148425e-01
-9.05628920e-01 6.86691940e-01 -4.71311808e-01 -5.95213473e-01
5.73914886e-01 -4.76762027e-01 -4.19926316e-01 1.42466038e-01
2.58467525e-01 9.53590930e-01 8.33167732e-01 -1.37781310e+00
-1.14025235e+00 -8.36234987e-01 -6.82941452e-02 2.06447408e-01
3.91404659e-01 -1.87051088e-01 -5.42362928e-01 -1.99778959e-01
5.74289620e-01 -1.18074477e+00 -2.13280737e-01 6.28748000e-01
4.14829291e-02 -2.08484501e-01 1.53935170e+00 -2.00478435e-01
3.25300306e-01 -2.15920663e+00 -6.90544676e-03 -7.02264160e-02
1.72554120e-01 2.00601667e-01 -9.93743166e-02 3.07128102e-01
2.54888624e-01 -1.67636678e-01 -1.14274332e-02 -5.87858438e-01
-2.43191607e-02 3.58100027e-01 -6.54577196e-01 5.98361373e-01
9.83328521e-02 9.36715841e-01 -1.08543122e+00 -2.68114090e-01
8.48349154e-01 8.29806030e-01 -3.73748749e-01 2.68774107e-02
-5.86551368e-01 5.64606607e-01 -3.78061414e-01 9.42418277e-01
1.02135003e+00 -1.04174860e-01 -5.23745060e-01 3.36545110e-02
-4.08231795e-01 2.59095371e-01 -1.15876305e+00 2.19090199e+00
-2.59618312e-01 8.51583421e-01 2.16425750e-02 -3.76016825e-01
8.44451249e-01 1.17267206e-01 6.87359154e-01 -5.73014855e-01
-9.99411866e-02 1.00317925e-01 -5.57357490e-01 -3.03334862e-01
8.77510190e-01 2.49603659e-01 5.66856824e-02 3.57005857e-02
-2.84166873e-01 -2.55870938e-01 -2.43389085e-02 2.13359177e-01
1.33982337e+00 2.60065913e-01 -1.07809283e-01 4.13438201e-01
3.12235564e-01 4.88298297e-01 4.69450027e-01 6.63597405e-01
-6.28399178e-02 6.02156997e-01 -3.58141333e-01 -5.34575820e-01
-1.04088509e+00 -1.46521461e+00 -1.06397010e-01 7.26301789e-01
5.46554267e-01 -2.48993605e-01 1.44320205e-01 -5.52558482e-01
4.85303402e-01 6.10876441e-01 -6.69531822e-02 1.35314301e-01
-6.89626932e-01 9.37427953e-02 8.18887427e-02 5.27909994e-01
1.27647325e-01 -5.11487365e-01 -1.23637962e+00 2.73292363e-01
-2.17953891e-01 -1.55058277e+00 -2.37254575e-01 -2.05958441e-01
-1.10614324e+00 -7.87470698e-01 -4.12399083e-01 -1.41678929e-01
1.95597693e-01 1.27047992e+00 1.10208893e+00 -5.33900484e-02
-7.75915682e-02 5.88114083e-01 -5.50518095e-01 -6.45471334e-01
-1.26962736e-01 -2.97473371e-01 2.98245907e-01 -1.02382876e-01
4.52301234e-01 -6.41956627e-01 -7.15761185e-01 3.33379179e-01
-5.70665479e-01 6.12073354e-02 2.85197526e-01 2.41645679e-01
8.63921404e-01 -4.64726627e-01 -6.33712485e-02 -8.25936273e-02
-4.14869189e-01 -6.36175692e-01 -1.08313322e+00 -3.45667928e-01
-3.19575876e-01 -2.43936971e-01 4.68255989e-02 -4.03548568e-01
-7.17774630e-01 5.48711956e-01 1.00707993e-01 -1.21303189e+00
-1.94677845e-01 2.73005098e-01 2.67739948e-02 -5.18337041e-02
5.87502658e-01 3.75755057e-02 -1.75274670e-01 -6.18462145e-01
6.47178769e-01 4.05979991e-01 7.75903106e-01 -1.90012068e-01
1.16677547e+00 1.19279039e+00 2.45521903e-01 -8.24904084e-01
-5.35148084e-01 -1.05955875e+00 -8.19619179e-01 -5.13310015e-01
7.75474787e-01 -1.36440921e+00 -6.69522166e-01 -1.35275070e-02
-1.55479956e+00 -3.78974271e-03 -2.23169342e-01 5.61407387e-01
-3.69155824e-01 2.45295048e-01 -1.30123079e-01 -9.71673846e-01
-8.52428675e-02 -9.79084432e-01 1.77394319e+00 -3.02028842e-02
7.01098591e-02 -3.61343652e-01 7.35156387e-02 3.33597898e-01
-1.15428038e-01 3.92413586e-01 9.52666849e-02 -2.82691047e-02
-1.56312346e+00 -5.81171751e-01 -2.29717031e-01 -4.12382364e-01
-3.00661206e-01 -4.00481634e-02 -1.05852342e+00 -2.91831583e-01
-1.94125548e-01 1.24894969e-01 1.01568496e+00 3.36944222e-01
7.28817880e-01 1.96846619e-01 -8.86371136e-01 6.64184272e-01
1.34054565e+00 1.45405918e-01 3.14858913e-01 1.85095981e-01
8.21408093e-01 2.84903586e-01 1.23865414e+00 6.28470361e-01
5.67112684e-01 1.06568050e+00 1.09579086e+00 8.81622732e-02
-2.79821664e-01 -2.65986383e-01 1.99427843e-01 3.81692350e-01
-8.09070691e-02 -2.83560663e-01 -1.17288947e+00 6.45185769e-01
-2.01988268e+00 -1.01751590e+00 -6.00824356e-01 2.28778028e+00
-1.59807980e-01 1.25708938e-01 3.96643244e-02 2.10204581e-03
4.59382057e-01 3.01665276e-01 -7.22535789e-01 3.42505276e-01
-1.15192756e-02 -1.71443507e-01 7.27863133e-01 3.75439733e-01
-8.90479326e-01 8.90311658e-01 4.80309153e+00 3.52973968e-01
-1.23987198e+00 4.48404431e-01 -3.70508671e-01 -5.44192135e-01
-1.98988557e-01 2.30472073e-01 -1.08721161e+00 4.17690128e-01
9.88804579e-01 1.49520651e-01 8.29753205e-02 9.68165815e-01
3.19282383e-01 -1.87799454e-01 -1.08617032e+00 1.31741536e+00
-9.85497758e-02 -1.71517825e+00 -1.09771103e-01 4.48956698e-01
4.36767638e-01 9.08774734e-01 4.55438904e-02 1.72001719e-01
8.77629891e-02 -4.58486497e-01 1.14429688e+00 4.25155967e-01
4.82439965e-01 -7.07144678e-01 3.92569482e-01 8.67712140e-01
-1.60160840e+00 1.08722821e-02 -4.62825596e-01 -2.00773314e-01
7.14012921e-01 6.74776912e-01 -1.01730835e+00 7.22128391e-01
9.06866908e-01 6.98854446e-01 -4.00668979e-01 1.32630765e+00
3.80576216e-03 1.97938234e-01 -8.56811404e-01 1.70215458e-01
3.78692269e-01 -1.52972993e-02 1.13412547e+00 8.39079559e-01
8.13495219e-01 3.90503824e-01 5.57715952e-01 8.12131763e-01
2.35672712e-01 -4.51634884e-01 -1.13810718e+00 4.67859387e-01
7.30178177e-01 1.14965975e+00 -5.40094376e-01 -4.48799849e-01
-4.79675978e-01 7.59263456e-01 1.06996536e-01 2.49338597e-01
-9.16549802e-01 3.27663302e-01 1.15709472e+00 4.10545856e-01
7.40004778e-01 -8.96737516e-01 -1.32227927e-01 -1.20963645e+00
2.09827676e-01 -1.57721207e-01 1.20828755e-01 -1.16415012e+00
-7.53508985e-01 3.82732064e-01 7.99001940e-03 -1.94307101e+00
-2.79311717e-01 -4.85247076e-01 -1.40433252e-01 7.46252596e-01
-1.61437249e+00 -1.26506984e+00 -7.63967693e-01 5.20401955e-01
7.34246731e-01 1.47272721e-01 3.81574661e-01 4.40451473e-01
9.06739309e-02 -2.39848912e-01 -2.47657388e-01 -4.28518891e-01
2.90872425e-01 -8.16261172e-01 7.57625818e-01 9.79742050e-01
7.35148489e-01 -4.41461690e-02 7.57963181e-01 -9.31757927e-01
-1.84817505e+00 -1.28813124e+00 6.71381950e-01 -1.02395165e+00
3.49987298e-01 -5.69389522e-01 -9.49186802e-01 6.54955566e-01
-3.73028576e-01 4.72567618e-01 1.43212333e-01 -2.08725199e-01
-3.48463178e-01 -6.84572533e-02 -9.54704821e-01 6.19975746e-01
1.48679769e+00 -3.87443274e-01 -4.21916485e-01 3.95773679e-01
1.12275672e+00 -9.59438920e-01 -6.04902387e-01 6.32910430e-01
4.74618286e-01 -8.55791390e-01 1.54576385e+00 -2.89364249e-01
2.43743951e-03 -9.18103456e-01 -5.54818869e-01 -1.06294656e+00
-2.65349358e-01 -4.09651846e-01 -7.71584928e-01 7.38667428e-01
1.51964482e-02 -3.54973823e-01 1.04389715e+00 3.35537225e-01
-5.11214316e-01 -3.26218367e-01 -1.40257370e+00 -1.03048849e+00
-5.64289331e-01 -1.24981689e+00 9.85264242e-01 5.21530509e-01
-4.82454568e-01 3.27326030e-01 -3.15370500e-01 8.73991966e-01
8.95422518e-01 3.63130689e-01 1.31756735e+00 -1.45714140e+00
1.29360959e-01 -1.77326471e-01 -9.76728678e-01 -1.66062307e+00
-9.03358161e-02 -9.50125992e-01 1.22654922e-01 -1.41209745e+00
-4.05541182e-01 -5.52524030e-01 9.12412480e-02 1.18775159e-01
8.59344436e-04 2.43178666e-01 6.93267226e-01 4.49347436e-01
-5.65135121e-01 5.89806080e-01 8.99189711e-01 -2.15173647e-01
-9.30852517e-02 1.04961611e-01 2.14373738e-01 7.17710197e-01
3.90270919e-01 -6.59358442e-01 -4.02268171e-01 -8.24881792e-01
6.84664100e-02 4.85287249e-01 7.99088836e-01 -1.31060326e+00
4.04468864e-01 -9.90410596e-02 4.19351190e-01 -1.93601274e+00
1.20372832e+00 -1.37761390e+00 5.33457220e-01 4.34907049e-01
3.17334384e-01 2.36341834e-01 2.32227802e-01 9.56808031e-01
1.11653522e-01 1.90446883e-01 3.33039701e-01 5.84357139e-03
-1.04785681e+00 8.48585486e-01 -1.07215337e-01 -5.42662859e-01
1.22447133e+00 -5.10045946e-01 -3.37267846e-01 -1.50893658e-01
-4.35157299e-01 4.42199707e-01 7.36985981e-01 7.93411136e-01
1.06904495e+00 -1.48705065e+00 -6.18676901e-01 1.65335536e-01
4.68462259e-01 5.57802737e-01 4.10969794e-01 8.72452676e-01
-3.74358833e-01 5.75992584e-01 1.08319238e-01 -1.68640161e+00
-1.43419337e+00 7.91166902e-01 -4.79187481e-02 2.38161281e-01
-1.16179967e+00 7.06201375e-01 1.99255258e-01 -3.34036976e-01
-3.50264646e-02 -6.56670809e-01 1.57760009e-01 1.40623637e-02
7.49926329e-01 4.35557365e-01 2.18356356e-01 -8.62990320e-01
-6.68035746e-01 8.21465194e-01 3.27530354e-01 -3.05832565e-01
1.18074143e+00 -3.28054786e-01 4.58864659e-01 6.92942381e-01
1.03856170e+00 -9.10932124e-02 -1.69507968e+00 -4.30044562e-01
-6.64813742e-02 -1.05113852e+00 2.15628415e-01 -3.83679748e-01
-9.37252462e-01 9.26690757e-01 9.17511702e-01 5.10844178e-02
5.53992331e-01 3.91811073e-01 1.00959122e+00 3.56580466e-01
1.03892827e+00 -4.34495628e-01 -2.09341690e-01 5.66034079e-01
7.83690393e-01 -1.40051162e+00 6.24877736e-02 -6.57409251e-01
-4.35568482e-01 8.98622036e-01 3.02017212e-01 -9.56005901e-02
3.83501977e-01 2.28735074e-01 7.24345371e-02 -4.06308234e-01
-9.58654404e-01 -3.97351086e-01 2.97873646e-01 9.52703774e-01
-3.00626069e-01 1.30850524e-01 4.60170090e-01 -3.85662429e-02
-1.52116239e-01 -6.41494989e-02 1.93242580e-01 9.92732227e-01
-6.63848162e-01 -7.15243518e-01 -7.23750830e-01 2.14757651e-01
2.11847365e-01 1.93230405e-01 1.21850297e-02 7.39692628e-01
2.13579923e-01 9.20890749e-01 5.52775264e-01 -5.10596693e-01
5.24787366e-01 -1.67519495e-01 4.85207826e-01 -5.74020088e-01
-8.53678733e-02 4.35352251e-02 5.07293753e-02 -1.00295866e+00
-3.45105827e-01 -1.11119473e+00 -1.42872548e+00 -3.19919914e-01
-2.65203148e-01 -3.37362587e-01 1.10920799e+00 8.29409063e-01
7.05142677e-01 2.68073887e-01 4.16846454e-01 -1.67805922e+00
-2.09066302e-01 -5.36769450e-01 -1.13677993e-01 1.46796435e-01
7.43170738e-01 -9.90127683e-01 -6.29764646e-02 -2.43508792e-03] | [7.1104736328125, -2.3084030151367188] |
61ff0c7c-956e-41d3-9780-28ef313f32c8 | batman-bilateral-attention-transformer-in | 2208.01159 | null | https://arxiv.org/abs/2208.01159v4 | https://arxiv.org/pdf/2208.01159v4.pdf | BATMAN: Bilateral Attention Transformer in Motion-Appearance Neighboring Space for Video Object Segmentation | Video Object Segmentation (VOS) is fundamental to video understanding. Transformer-based methods show significant performance improvement on semi-supervised VOS. However, existing work faces challenges segmenting visually similar objects in close proximity of each other. In this paper, we propose a novel Bilateral Attention Transformer in Motion-Appearance Neighboring space (BATMAN) for semi-supervised VOS. It captures object motion in the video via a novel optical flow calibration module that fuses the segmentation mask with optical flow estimation to improve within-object optical flow smoothness and reduce noise at object boundaries. This calibrated optical flow is then employed in our novel bilateral attention, which computes the correspondence between the query and reference frames in the neighboring bilateral space considering both motion and appearance. Extensive experiments validate the effectiveness of BATMAN architecture by outperforming all existing state-of-the-art on all four popular VOS benchmarks: Youtube-VOS 2019 (85.0%), Youtube-VOS 2018 (85.3%), DAVIS 2017Val/Testdev (86.2%/82.2%), and DAVIS 2016 (92.5%). | ['Mei Chen', 'Li Fuxin', 'Gaurav Mittal', 'Jialin Yuan', 'Ye Yu'] | 2022-08-01 | null | null | null | null | ['semi-supervised-video-object-segmentation', 'visual-object-tracking'] | ['computer-vision', 'computer-vision'] | [-3.68303806e-02 -3.64294410e-01 -3.56024384e-01 -2.99196303e-01
-5.85746825e-01 -5.69069684e-01 1.42455176e-01 -4.11881864e-01
-3.05099428e-01 5.91080964e-01 1.11931518e-01 -3.24461609e-02
3.22766870e-01 -3.80984098e-01 -9.12084877e-01 -4.01843995e-01
1.97933704e-01 1.50677204e-01 7.33699143e-01 1.09091140e-01
1.82565063e-01 1.83090270e-01 -1.25192952e+00 2.79923826e-01
1.03790307e+00 1.26687431e+00 2.25073367e-01 8.35608542e-01
-1.96490243e-01 8.74501228e-01 -3.81140202e-01 -2.39581466e-01
6.21127307e-01 -4.20524746e-01 -1.08867717e+00 3.34205478e-01
1.25696349e+00 -6.47356391e-01 -4.91175771e-01 1.10455477e+00
2.38455996e-01 2.66474575e-01 5.73477745e-01 -1.64569712e+00
-8.62499535e-01 2.25457028e-01 -9.38807189e-01 7.06906617e-01
1.49706945e-01 6.35688901e-01 8.29988658e-01 -8.04133415e-01
9.97204185e-01 1.40386224e+00 4.46996510e-01 6.66792333e-01
-1.18955350e+00 -7.34791100e-01 3.21576089e-01 4.02847946e-01
-1.25642776e+00 -4.13060844e-01 7.11398959e-01 -7.10621417e-01
7.11624742e-01 4.44148369e-02 8.50363910e-01 8.96021783e-01
-9.08818096e-02 1.33887482e+00 8.02684844e-01 1.12413235e-01
-1.26531288e-01 3.96904759e-02 5.29864803e-02 7.72293329e-01
1.34293303e-01 -6.39574528e-02 -5.73189378e-01 4.46740925e-01
1.00543582e+00 9.63188335e-03 -7.04837203e-01 -4.63012397e-01
-1.28184330e+00 5.07281482e-01 5.40614665e-01 -9.02987737e-03
-2.17895404e-01 2.93076515e-01 4.41821843e-01 -3.78271267e-02
4.46210057e-01 -5.40589392e-02 -3.91043067e-01 -1.17819116e-01
-1.38147473e+00 1.84725240e-01 4.79158461e-01 1.24184263e+00
7.14252234e-01 4.59484786e-01 -5.85983574e-01 5.56428075e-01
5.31450272e-01 4.17876959e-01 2.10404575e-01 -1.40751421e+00
6.82335436e-01 4.39746052e-01 1.75314486e-01 -9.06840324e-01
3.36221457e-02 -2.26892918e-01 -7.45945156e-01 6.05992451e-02
5.81661820e-01 -6.18080646e-02 -1.13459754e+00 1.52063107e+00
5.20919681e-01 8.42711449e-01 -4.60340790e-02 1.36986983e+00
1.16087353e+00 8.41497898e-01 2.23734021e-01 -1.45001844e-01
9.22197044e-01 -1.63476408e+00 -8.47657442e-01 -1.79321259e-01
2.09075913e-01 -7.15359390e-01 1.07750452e+00 1.39975533e-01
-1.34874880e+00 -8.73253763e-01 -9.82821465e-01 -4.72147405e-01
1.41120881e-01 -4.39006202e-02 2.95360327e-01 4.77513850e-01
-1.06082845e+00 6.19578421e-01 -8.44002724e-01 -9.21422094e-02
9.03021932e-01 1.35696545e-01 -1.10839590e-01 -2.30429068e-01
-8.75507951e-01 2.27574408e-01 2.24147096e-01 1.71246618e-01
-1.36277521e+00 -1.22676444e+00 -9.68106091e-01 -1.86775811e-02
4.79529232e-01 -9.37383950e-01 9.81823206e-01 -1.13365781e+00
-1.34190953e+00 7.88671672e-01 -4.35033321e-01 -4.64998990e-01
1.08548939e+00 -4.18701828e-01 -8.27617422e-02 4.71896172e-01
4.68224049e-01 1.23459506e+00 1.01553369e+00 -1.23935878e+00
-8.16532016e-01 -9.83421952e-02 -2.71140099e-01 1.63907021e-01
3.89875397e-02 -1.51690081e-01 -1.08969641e+00 -8.23723674e-01
-5.99760450e-02 -6.50673151e-01 4.82564326e-03 3.87454897e-01
-5.38027406e-01 -3.60360116e-01 1.31687629e+00 -8.39775681e-01
1.09448147e+00 -1.92839992e+00 1.75715014e-01 -2.21602365e-01
3.38462144e-01 6.99002862e-01 -2.69238919e-01 -4.25061315e-01
5.77309094e-02 1.74216300e-01 -3.73885691e-01 -5.00965357e-01
-3.21222037e-01 2.39481628e-02 -1.45483404e-01 5.39616346e-01
3.97716671e-01 1.44194281e+00 -1.07021558e+00 -8.62818122e-01
5.09523988e-01 4.74252582e-01 -7.18294084e-01 3.92541379e-01
-2.42406160e-01 6.14243031e-01 -2.52221912e-01 1.03563237e+00
7.59026110e-01 -2.60813832e-01 -3.22465569e-01 -4.96202618e-01
-3.20996828e-02 -9.61422101e-02 -1.15292525e+00 1.77275336e+00
-8.09491426e-02 1.11999011e+00 1.14191949e-01 -7.83470333e-01
6.61281824e-01 6.96791485e-02 8.85141194e-01 -6.43183053e-01
2.01455280e-01 -1.11557856e-01 -2.74270773e-01 -7.91354597e-01
4.51614112e-01 4.84469116e-01 5.88014662e-01 1.19029894e-01
2.98869580e-01 -1.42300799e-01 5.73046386e-01 4.24284458e-01
6.02688909e-01 4.08793479e-01 -1.26485839e-01 -2.68493503e-01
6.97353840e-01 -1.47495314e-01 1.02393460e+00 4.86533433e-01
-1.02486646e+00 9.87305343e-01 4.44558412e-01 -3.87996584e-01
-1.05408084e+00 -1.24224734e+00 7.90317208e-02 6.35708332e-01
6.20534003e-01 -4.25194412e-01 -8.19188178e-01 -1.07722044e+00
1.15540206e-01 3.62789512e-01 -6.64000034e-01 1.00957863e-01
-6.60006285e-01 -1.02681264e-01 3.44650000e-01 7.40481734e-01
9.77499187e-01 -1.14349723e+00 -4.56656456e-01 1.19957484e-01
-6.88017428e-01 -1.67525721e+00 -1.23510528e+00 -6.31432354e-01
-9.39191937e-01 -1.26213896e+00 -1.21439767e+00 -8.87329221e-01
4.36732024e-01 7.21101105e-01 1.10619617e+00 -8.69591832e-02
-2.37499446e-01 5.05395889e-01 -1.91289261e-01 7.02950405e-03
-1.63648531e-01 -2.29100153e-01 -2.60127783e-01 4.58838284e-01
-2.52466127e-02 -2.30539575e-01 -1.03172517e+00 5.58403194e-01
-8.42653394e-01 4.24967669e-02 3.37122649e-01 6.68712199e-01
8.51285875e-01 -4.99159575e-01 2.15287030e-01 -6.22939229e-01
-5.42041026e-02 -5.45753956e-01 -5.79508483e-01 3.26493770e-01
-4.22315538e-01 -1.76729634e-01 2.27566153e-01 -4.84182298e-01
-9.33330476e-01 1.87469080e-01 1.51710078e-01 -1.08409464e+00
-1.26437858e-01 -1.11589998e-01 -2.37637922e-01 -2.84810662e-01
1.49465531e-01 2.26647675e-01 -1.32164896e-01 -3.00941199e-01
3.27171296e-01 5.47743618e-01 8.21774125e-01 -2.35109091e-01
7.70963550e-01 8.71913135e-01 -2.58190691e-01 -7.56298542e-01
-9.29929614e-01 -8.37107718e-01 -6.97129250e-01 -5.15813112e-01
1.32521999e+00 -9.16847587e-01 -8.67522001e-01 5.08991599e-01
-1.25828195e+00 -6.48049951e-01 -2.83882439e-01 2.88367897e-01
-5.79620123e-01 6.25401318e-01 -5.81866324e-01 -5.74940503e-01
-5.26996195e-01 -1.42791748e+00 1.13773620e+00 5.54477155e-01
4.79476415e-02 -9.34399784e-01 -2.09705815e-01 8.19317877e-01
2.41127506e-01 2.84125000e-01 1.58583760e-01 -9.39599238e-03
-9.85639513e-01 4.47956234e-01 -8.42699945e-01 6.79491282e-01
1.17703661e-01 2.96291262e-01 -8.31453502e-01 -2.93904990e-01
-3.72436911e-01 -1.04849659e-01 1.20834565e+00 8.19830239e-01
1.11978114e+00 -1.03104031e-02 -1.56944558e-01 1.20414627e+00
1.26424372e+00 2.46813297e-01 6.43964767e-01 2.14175180e-01
1.41316998e+00 4.92953956e-01 8.27428222e-01 6.30963147e-02
4.92195129e-01 7.37294197e-01 5.78890800e-01 -1.99766472e-01
-7.29176164e-01 -2.10624218e-01 5.29793799e-01 4.51740563e-01
2.03092955e-02 -2.79155016e-01 -6.84864759e-01 1.03014350e+00
-1.85059667e+00 -8.50041449e-01 -3.84475380e-01 1.89583266e+00
6.52471185e-01 1.02315336e-01 2.11219817e-01 -1.19632021e-01
8.13564062e-01 3.61774087e-01 -8.32077861e-01 -3.94255891e-02
-2.89042115e-01 -6.60597682e-02 6.18065298e-01 6.08647287e-01
-1.41734242e+00 1.25996506e+00 5.06361914e+00 6.65262103e-01
-1.08885133e+00 2.85022318e-01 8.72810662e-01 -1.46954909e-01
-4.65378948e-02 -1.99625298e-01 -8.14685464e-01 6.66866779e-01
3.62433702e-01 3.24372049e-05 3.15131575e-01 6.40694439e-01
4.64517891e-01 -2.48591423e-01 -9.15580034e-01 1.28692520e+00
2.54440755e-01 -1.65288031e+00 2.29080752e-01 -2.79208779e-01
1.41392696e+00 1.20788641e-01 1.27480492e-01 -4.24909703e-02
-2.51540113e-02 -8.01909506e-01 1.07665431e+00 4.09296006e-01
8.31911922e-01 -4.73066866e-01 5.65982461e-01 -2.00877383e-01
-1.64503646e+00 1.27684876e-01 -8.98221955e-02 5.18958271e-01
6.47951245e-01 2.02721730e-01 -4.95065570e-01 5.15997648e-01
1.23961639e+00 1.50035179e+00 -5.74886978e-01 1.30739355e+00
-1.08379669e-01 7.09552705e-01 -1.84178889e-01 4.43763494e-01
6.07854545e-01 -3.16641659e-01 9.04578269e-01 1.33834016e+00
-6.60634860e-02 -1.69563815e-01 2.94152677e-01 1.09385788e+00
-1.96302816e-01 7.10166618e-02 -1.63262039e-01 8.47531483e-02
5.18280789e-02 1.01190126e+00 -8.09118092e-01 -6.27109170e-01
-3.56314003e-01 1.15950477e+00 1.77731570e-02 7.74444282e-01
-1.01462495e+00 -1.52137592e-01 9.34886634e-01 1.34059101e-01
7.04595208e-01 -1.47947863e-01 -9.98051614e-02 -1.33310366e+00
7.04799592e-02 -4.60711509e-01 5.34921765e-01 -8.68868530e-01
-1.06702816e+00 3.52130413e-01 -7.66786486e-02 -1.26321864e+00
1.48780346e-01 -3.26256871e-01 -6.97444022e-01 5.74783802e-01
-1.82798088e+00 -9.95463490e-01 -7.17294455e-01 8.10503781e-01
1.06593680e+00 6.88937530e-02 -2.18487740e-01 5.61022580e-01
-7.50920773e-01 5.42740405e-01 -1.15437143e-01 4.67473894e-01
8.54796112e-01 -1.20721877e+00 5.54362297e-01 1.11527967e+00
2.15543523e-01 1.32331222e-01 3.07149202e-01 -8.05008769e-01
-1.15392137e+00 -1.61503506e+00 4.07299459e-01 -3.01275611e-01
4.23603296e-01 -1.33116812e-01 -1.08592105e+00 7.19797492e-01
2.99489766e-01 7.76804388e-01 -2.63698455e-02 -9.09968793e-01
-2.67617494e-01 -2.18340382e-01 -1.06538272e+00 5.98490059e-01
1.21371877e+00 -3.30395490e-01 -1.98894441e-01 1.45097286e-01
8.79472852e-01 -6.86471701e-01 -7.39400268e-01 4.22906190e-01
4.20393556e-01 -9.82817411e-01 1.18381917e+00 -7.71407545e-01
4.23274100e-01 -6.62502229e-01 5.10351844e-02 -9.10790980e-01
2.73101665e-02 -1.09237230e+00 -3.61611873e-01 1.37268031e+00
-6.70493916e-02 -2.58968264e-01 1.07319534e+00 4.77883995e-01
-2.12959900e-01 -6.23707175e-01 -7.68371820e-01 -7.35747516e-01
-2.31527984e-01 -5.90348005e-01 1.09776624e-01 9.34847713e-01
-6.67744577e-01 -7.47262267e-03 -3.89011443e-01 2.43823230e-02
9.83374178e-01 6.56374767e-02 9.26207840e-01 -8.00361931e-01
-1.02950968e-01 -5.01048386e-01 -5.83532870e-01 -1.54825616e+00
1.38202548e-01 -8.71947706e-01 -2.04747282e-02 -1.52652228e+00
1.70538083e-01 -8.05256963e-02 -2.69930661e-01 1.69625521e-01
-4.99485582e-01 6.52552366e-01 6.15321159e-01 1.84473947e-01
-9.91811752e-01 7.40109444e-01 1.74078369e+00 -4.23209131e-01
-5.08881867e-01 -1.34844705e-01 -2.40377009e-01 7.63121068e-01
6.69232070e-01 -3.78337324e-01 -3.24662775e-01 -5.36836624e-01
-5.43750644e-01 -7.41396751e-03 5.82077205e-01 -9.41279829e-01
2.96042979e-01 -1.02379322e-01 3.71859044e-01 -7.66853869e-01
6.08526319e-02 -5.86131394e-01 -1.36481494e-01 4.84616548e-01
-2.25168109e-01 -4.75471318e-02 2.02779576e-01 7.49449849e-01
-3.55028123e-01 9.12185013e-02 1.01270318e+00 1.81496799e-01
-1.24774253e+00 8.42774272e-01 1.11264531e-02 7.70636499e-01
1.30997193e+00 -4.62515444e-01 -8.34616721e-02 -4.06190932e-01
-6.72369838e-01 6.37713790e-01 1.46961331e-01 6.92727566e-01
1.02624834e+00 -1.12914276e+00 -8.01975846e-01 1.59319028e-01
-1.35763034e-01 3.78150523e-01 5.79438746e-01 1.11241579e+00
-7.71711648e-01 3.38680953e-01 -3.97275478e-01 -1.04745924e+00
-1.36216748e+00 4.59419101e-01 4.45738375e-01 2.90826350e-01
-6.48544371e-01 1.07095909e+00 4.63297695e-01 2.65462577e-01
4.55633193e-01 -5.47046304e-01 -9.21558067e-02 -8.11562911e-02
5.14181137e-01 6.73465192e-01 -4.65502262e-01 -9.67669725e-01
-3.19031030e-01 9.93139148e-01 -3.35740447e-02 8.28133225e-02
9.75674212e-01 -5.01907527e-01 1.62886202e-01 2.41530821e-01
1.47812450e+00 -2.62994260e-01 -2.12068582e+00 -3.47322583e-01
-2.26443037e-01 -9.25158381e-01 1.08179972e-01 -3.94762814e-01
-1.86827767e+00 1.07363439e+00 5.95679939e-01 -2.24749044e-01
9.35496509e-01 -6.00495562e-02 1.26357365e+00 -2.86245197e-01
-6.84609637e-02 -8.31813276e-01 2.91683644e-01 2.96157837e-01
7.31547117e-01 -1.59172022e+00 -1.12651296e-01 -6.35631382e-01
-9.76856828e-01 1.00644755e+00 9.28417087e-01 -2.77864724e-01
5.42013466e-01 -5.66942915e-02 1.23261847e-01 1.93223089e-01
-4.69135135e-01 -4.35852766e-01 8.06994438e-01 6.63064718e-01
1.81655705e-01 -4.03127968e-01 1.99447460e-02 1.65768489e-01
3.16358417e-01 1.98415760e-02 4.32543069e-01 4.89209801e-01
-9.59699079e-02 -4.54170316e-01 -2.00513154e-01 1.49231881e-01
-3.70596766e-01 -6.77404702e-02 -2.63732493e-01 7.69541204e-01
2.32100666e-01 9.01242912e-01 2.70294338e-01 -6.11217469e-02
1.99144915e-01 -2.77458221e-01 3.23729008e-01 -2.22724140e-01
-4.79367107e-01 2.59419888e-01 -2.86662757e-01 -1.15479171e+00
-6.29663110e-01 -8.01973999e-01 -1.34405375e+00 -2.44450778e-01
-8.92770756e-03 -6.36365265e-02 3.97257924e-01 8.19834709e-01
3.74475747e-01 6.80806756e-01 4.03460711e-01 -1.11047494e+00
-1.41471541e-02 -5.65499127e-01 -2.26576224e-01 7.40811706e-01
6.54351115e-01 -7.16478229e-01 -3.13390285e-01 4.94731665e-01] | [9.166528701782227, -0.17317698895931244] |
60bdd847-89b3-4add-899c-007c8ce845cc | robust-fair-clustering-a-novel-fairness | 2210.01953 | null | https://arxiv.org/abs/2210.01953v3 | https://arxiv.org/pdf/2210.01953v3.pdf | Robust Fair Clustering: A Novel Fairness Attack and Defense Framework | Clustering algorithms are widely used in many societal resource allocation applications, such as loan approvals and candidate recruitment, among others, and hence, biased or unfair model outputs can adversely impact individuals that rely on these applications. To this end, many fair clustering approaches have been recently proposed to counteract this issue. Due to the potential for significant harm, it is essential to ensure that fair clustering algorithms provide consistently fair outputs even under adversarial influence. However, fair clustering algorithms have not been studied from an adversarial attack perspective. In contrast to previous research, we seek to bridge this gap and conduct a robustness analysis against fair clustering by proposing a novel black-box fairness attack. Through comprehensive experiments, we find that state-of-the-art models are highly susceptible to our attack as it can reduce their fairness performance significantly. Finally, we propose Consensus Fair Clustering (CFC), the first robust fair clustering approach that transforms consensus clustering into a fair graph partitioning problem, and iteratively learns to generate fair cluster outputs. Experimentally, we observe that CFC is highly robust to the proposed attack and is thus a truly robust fair clustering alternative. | ['Hongfu Liu', 'Prasant Mohapatra', 'Peizhao Li', 'Anshuman Chhabra'] | 2022-10-04 | null | null | null | null | ['graph-partitioning'] | ['graphs'] | [ 4.03207447e-03 6.93708062e-02 -1.59050018e-01 -3.77401441e-01
-4.34872538e-01 -9.57017720e-01 6.43024266e-01 1.42315865e-01
-1.47071674e-01 8.63800585e-01 -2.43686140e-01 -4.59088147e-01
-1.99278846e-01 -9.53757882e-01 -4.40855205e-01 -5.62299550e-01
-5.12738302e-02 3.94899011e-01 -9.28386226e-02 -4.66266498e-02
4.25723106e-01 3.79891574e-01 -1.23185754e+00 -3.22922826e-01
1.46011806e+00 4.44632202e-01 -8.75544846e-01 4.19401467e-01
3.52895498e-01 6.84486926e-01 -7.86493897e-01 -1.38345122e+00
6.03079796e-01 -5.08875251e-01 -9.10106122e-01 -3.81643921e-01
2.44203433e-01 -3.85633379e-01 -3.29143330e-02 1.36384928e+00
7.07313716e-01 7.92694092e-02 7.37848938e-01 -2.04123878e+00
-8.20086718e-01 1.16829491e+00 -6.80408895e-01 -2.59662926e-01
2.50465006e-01 3.19828957e-01 1.01210594e+00 -3.97225529e-01
3.11116546e-01 1.44411433e+00 5.53803682e-01 1.00755215e+00
-1.36950123e+00 -1.32776952e+00 1.25718461e-02 -1.36395618e-01
-1.75709534e+00 -5.68470657e-01 6.09791338e-01 -2.10531339e-01
1.97599709e-01 6.82405412e-01 4.55544353e-01 9.83171284e-01
-7.75983706e-02 4.49804008e-01 1.31156874e+00 -3.33209932e-01
6.46697223e-01 -9.27285030e-02 -5.97001938e-03 1.78447694e-01
7.80639648e-01 2.92051822e-01 -1.90291658e-01 -7.13592529e-01
2.21236050e-01 -1.94807693e-01 -8.92852843e-02 -3.36436987e-01
-5.63221872e-01 9.74484682e-01 7.01425135e-01 -4.12465073e-02
1.75549630e-02 6.66138232e-01 2.32938096e-01 2.82000899e-01
3.02972794e-01 3.57766271e-01 2.97592491e-01 9.45329815e-02
-1.27216768e+00 5.56039512e-01 8.67397487e-01 6.90614641e-01
4.13833499e-01 -3.72184888e-02 -9.25203860e-02 2.68831819e-01
3.65822375e-01 6.52155936e-01 -2.18661264e-01 -1.52853823e+00
6.08572476e-02 4.34296727e-01 7.83064067e-02 -1.48422050e+00
2.50556227e-02 -3.38504612e-01 -1.14713216e+00 1.98352039e-01
4.29020256e-01 -2.13919148e-01 -2.92625785e-01 2.11923575e+00
4.64729905e-01 4.12220389e-01 -1.64201379e-01 9.63745058e-01
2.68153757e-01 2.41346601e-02 4.69025314e-01 -1.84777111e-01
7.57777572e-01 -3.97026867e-01 -5.91276944e-01 2.30475545e-01
9.99582186e-02 -7.35484362e-01 8.42665136e-01 2.77551502e-01
-1.22822618e+00 -4.10542227e-02 -1.03225338e+00 3.26877207e-01
-2.29316577e-01 -7.08408952e-01 7.98325896e-01 1.67037320e+00
-1.13521266e+00 5.85119128e-01 -5.57071567e-01 -4.04706240e-01
7.83988476e-01 5.58713555e-01 -2.67348420e-02 -1.55953184e-01
-1.46787107e+00 6.91636682e-01 -1.43594682e-01 -5.46639524e-02
-8.19821119e-01 -7.34898508e-01 -4.14615303e-01 2.64257938e-01
3.77206028e-01 -7.71481752e-01 7.87645638e-01 -1.15743566e+00
-1.35409069e+00 8.83455336e-01 2.64815897e-01 -6.37079537e-01
8.79453361e-01 1.80290252e-01 -2.73232669e-01 9.04912949e-02
2.22408831e-01 6.87069952e-01 8.66146147e-01 -1.58442390e+00
-2.49918684e-01 -3.86891007e-01 2.60456473e-01 2.60739513e-02
-5.93312442e-01 3.23626876e-01 5.00140861e-02 -6.23156250e-01
-5.01782119e-01 -1.27536666e+00 -5.29578567e-01 -9.63587165e-02
-8.97687793e-01 -1.42539263e-01 5.19169509e-01 1.54717604e-03
1.41117847e+00 -1.87177122e+00 -1.26982331e-01 9.91973460e-01
5.12633681e-01 2.60790765e-01 8.52712765e-02 1.70517474e-01
2.19956651e-01 9.59367633e-01 -3.48507315e-01 -2.58768022e-01
5.92183888e-01 -1.31781951e-01 -2.97227144e-01 8.14727247e-01
-1.43224716e-01 8.17399621e-01 -7.93883264e-01 -5.60717463e-01
1.56079524e-03 1.38796568e-01 -8.22220087e-01 1.66360259e-01
2.77770460e-01 2.52718151e-01 -2.79687136e-01 7.02800035e-01
1.15801001e+00 -6.08702973e-02 6.30059242e-01 2.91238487e-01
3.50989223e-01 -4.19924587e-01 -1.09164524e+00 1.07193553e+00
1.28272355e-01 1.49955779e-01 4.28634971e-01 -9.95960414e-01
6.54684246e-01 2.23322779e-01 4.25214231e-01 -2.88073033e-01
4.42907929e-01 2.32527986e-01 2.80666262e-01 3.57106596e-01
6.94349527e-01 -3.07549864e-01 -6.24145329e-01 1.02934587e+00
-4.97949362e-01 -1.59902140e-01 -1.30810425e-01 8.20237637e-01
1.02669179e+00 -5.60646117e-01 1.72636569e-01 -6.30633771e-01
4.42921281e-01 -2.17719898e-01 6.24106407e-01 1.03103113e+00
-1.06505632e+00 4.44294572e-01 4.87632275e-01 -8.62003118e-02
-6.43821001e-01 -1.36543810e+00 1.09996229e-01 9.44949508e-01
4.83885974e-01 -2.96107322e-01 -1.29581738e+00 -8.25886905e-01
2.94952303e-01 5.56975961e-01 -6.40347302e-01 -6.12865090e-01
-1.79517061e-01 -6.65297925e-01 1.40654576e+00 2.31465712e-01
5.18797159e-01 -6.41019464e-01 -3.41189742e-01 -1.37305319e-01
-2.30395809e-01 -7.43077338e-01 -7.93859005e-01 -4.55687284e-01
-4.66857284e-01 -1.41159189e+00 -4.89432096e-01 -3.73260260e-01
8.15324724e-01 4.05454069e-01 1.30482590e+00 8.54335308e-01
-2.88721859e-01 5.20599484e-01 -2.58588701e-01 -2.54112363e-01
-6.38472199e-01 7.95703679e-02 4.99574780e-01 6.88783750e-02
4.65021759e-01 -7.67988801e-01 -9.25322831e-01 4.57900494e-01
-9.90779281e-01 -5.01110792e-01 5.44522591e-02 3.95288497e-01
4.99779060e-02 2.69578844e-01 1.12574410e+00 -1.26654530e+00
1.20398378e+00 -5.63563406e-01 -4.62274164e-01 3.54997873e-01
-8.28808308e-01 -3.72580945e-01 9.17453110e-01 -2.42935687e-01
-8.28609824e-01 -2.60902911e-01 2.72222191e-01 -3.57611477e-01
-2.31033918e-02 1.15913473e-01 -3.57309788e-01 -3.40609372e-01
7.79090166e-01 -4.41783637e-01 -3.63049731e-02 2.12768137e-01
8.86645913e-01 7.01605737e-01 6.03903115e-01 -9.79978204e-01
1.34971583e+00 4.40341443e-01 2.19750553e-01 -1.13757104e-01
-2.87323326e-01 -1.04559097e-03 -4.46027309e-01 -4.10162419e-01
6.42135799e-01 -6.51498139e-01 -1.37992978e+00 3.67226839e-01
-7.19985902e-01 -2.60880172e-01 1.01715796e-01 -6.60145879e-02
-2.40394622e-01 6.27050996e-01 -5.34382641e-01 -9.58663702e-01
-6.89897358e-01 -9.89758015e-01 2.28948101e-01 4.33922470e-01
-4.72157985e-01 -8.50242734e-01 -2.11740300e-01 4.34931904e-01
6.78450048e-01 5.70797026e-01 6.86788976e-01 -7.32597291e-01
-5.25566757e-01 -1.58197671e-01 -1.86521903e-01 3.99196446e-02
-1.04896024e-01 5.88315368e-01 -8.95347118e-01 -6.70683622e-01
-5.11673927e-01 -5.09594619e-01 5.16823232e-01 1.90562412e-01
1.11490643e+00 -5.67752182e-01 -1.33447319e-01 3.35161507e-01
1.16975820e+00 -2.36373052e-01 5.89721203e-01 -1.16493084e-01
4.75081921e-01 7.77430058e-01 5.01340389e-01 5.90785921e-01
5.23114145e-01 5.30148149e-01 6.04126513e-01 8.03383440e-03
1.90473393e-01 -2.80031264e-01 2.69062161e-01 3.82402390e-01
-1.40263855e-01 -2.37560138e-01 -8.49168599e-01 3.37809592e-01
-1.88562250e+00 -1.20816922e+00 -2.53217630e-02 2.37482929e+00
9.14476693e-01 1.39792025e-01 3.76309037e-01 3.08564305e-01
1.11935449e+00 -8.36272389e-02 -5.91605723e-01 -5.92839360e-01
7.97191486e-02 4.54708099e-01 5.01345456e-01 3.90814185e-01
-1.08010387e+00 1.05154073e+00 6.61347342e+00 1.05939460e+00
-7.56450832e-01 -1.70706045e-02 9.94020045e-01 -2.64802754e-01
-5.14266729e-01 2.00698003e-01 1.08381778e-01 6.58195078e-01
8.41727734e-01 -8.16012621e-01 7.76471138e-01 7.74255335e-01
1.12210862e-01 2.34823391e-01 -9.36203003e-01 6.73481345e-01
-1.42007202e-01 -1.00533628e+00 3.85330431e-02 1.72363549e-01
9.95503783e-01 -7.25719273e-01 4.26325113e-01 3.37619126e-01
1.24975491e+00 -1.48227870e+00 6.71547532e-01 1.37178507e-02
7.87268400e-01 -1.38914168e+00 4.37588871e-01 1.15585580e-01
-1.03011465e+00 3.87544371e-02 -5.33198059e-01 -1.72886312e-01
-4.77674119e-02 6.94362998e-01 -3.03353012e-01 6.12418711e-01
6.15495801e-01 2.53717184e-01 -6.78450406e-01 8.73313189e-01
-2.80874342e-01 8.70040357e-01 -1.86829686e-01 7.40917772e-02
-6.98452815e-02 2.92691048e-02 3.25159431e-01 8.40306222e-01
9.05286297e-02 2.29161948e-01 2.66687125e-01 1.31261671e+00
-8.35596919e-01 -1.21611981e-02 -4.88775641e-01 3.04031104e-01
1.23641050e+00 1.38180542e+00 -9.84804213e-01 -1.37482315e-01
1.43900335e-01 7.74583399e-01 1.36281848e-01 2.69580543e-01
-1.05524611e+00 -3.82816434e-01 8.28899145e-01 -8.21437512e-04
-4.10327137e-01 1.83580831e-01 -6.26314402e-01 -1.03029978e+00
-4.86849606e-01 -1.16721749e+00 5.43961883e-01 -9.05298516e-02
-1.68705606e+00 1.32842213e-01 -3.46462041e-01 -1.06188130e+00
-2.12681945e-02 2.24859744e-01 -1.04524112e+00 6.81154370e-01
-1.30046415e+00 -1.26984739e+00 -3.89241762e-02 9.80015278e-01
-3.81516308e-01 -1.20593861e-01 7.69112051e-01 3.81800681e-01
-6.41499341e-01 1.41229546e+00 -1.25484064e-01 1.25487626e-01
9.92987514e-01 -1.26812649e+00 3.16987336e-01 1.45277703e+00
-1.47808760e-01 1.16974866e+00 6.75104439e-01 -7.41502583e-01
-1.09860361e+00 -1.19157207e+00 5.00497997e-01 -5.56591570e-01
3.85951877e-01 -4.39816445e-01 -6.14388049e-01 3.45706195e-01
1.87924773e-01 9.13259238e-02 1.07677913e+00 1.83029864e-02
-6.03741586e-01 -9.66621041e-02 -1.67613828e+00 9.22455549e-01
1.06824303e+00 -6.04637563e-01 -8.17082226e-02 -1.62781090e-01
6.29173934e-01 5.24561517e-02 -7.88479984e-01 2.40972087e-01
6.14814103e-01 -1.09926248e+00 1.05596769e+00 -4.87238377e-01
3.07856351e-01 -3.55268925e-01 -4.83708046e-02 -1.27620292e+00
-2.42873818e-01 -1.00082684e+00 1.68190375e-01 1.68523526e+00
1.15196139e-01 -5.66471994e-01 9.79027808e-01 1.28033602e+00
4.11735088e-01 -1.06931500e-01 -1.06688380e+00 -6.57260478e-01
8.07719171e-01 -2.68636793e-01 1.17393279e+00 1.40687311e+00
2.85809577e-01 -9.28779766e-02 -6.99777424e-01 1.42244712e-01
1.25012505e+00 2.35565484e-01 1.01630747e+00 -1.30694366e+00
5.64989299e-02 -8.58549535e-01 -1.61179423e-01 -1.04614228e-01
5.47116637e-01 -1.09294701e+00 -6.13029227e-02 -9.37022567e-01
4.58895117e-01 -8.73811364e-01 -1.71206385e-01 3.02417278e-01
-6.76218688e-01 6.26541257e-01 6.63396895e-01 4.24488962e-01
-8.79625022e-01 3.85981709e-01 8.01987767e-01 -2.52607197e-01
1.12153575e-01 2.95746684e-01 -1.36730528e+00 4.37903702e-01
8.84852469e-01 -6.07327640e-01 -4.93810683e-01 1.91924527e-01
1.97788671e-01 -1.52335659e-01 3.98585945e-01 -6.80432260e-01
3.69347006e-01 -5.78287065e-01 1.11954719e-01 6.05310909e-02
-3.23307842e-01 -9.66096222e-01 2.72674531e-01 7.19284177e-01
-5.01872659e-01 -1.46862984e-01 -2.90089726e-01 6.02066278e-01
1.04641750e-01 2.31276050e-01 9.35415924e-01 3.93370679e-03
-1.19541800e-02 4.48846906e-01 -3.90516788e-01 3.72760743e-01
1.36749935e+00 -1.06351497e-02 -7.68873036e-01 -7.78423190e-01
-3.73183280e-01 6.26271069e-01 1.04456401e+00 1.71140864e-01
3.61280471e-01 -1.45385420e+00 -8.19677413e-01 -1.40162870e-01
-1.71414837e-01 -3.69483292e-01 1.30723894e-01 5.04224420e-01
-3.70786220e-01 -6.13150112e-02 -3.34763467e-01 -1.35450259e-01
-1.25328350e+00 7.18653440e-01 3.25435698e-01 -2.32867196e-01
2.15403527e-01 6.74302399e-01 -6.28561229e-02 -7.27617919e-01
3.30248594e-01 4.06076193e-01 1.21353418e-01 -2.51300901e-01
3.23942065e-01 7.71914184e-01 -1.95597127e-01 -7.30915129e-01
-6.72595084e-01 3.28284025e-01 3.31667989e-01 -5.21516129e-02
7.87705243e-01 -3.25852871e-01 -4.06734258e-01 -4.12363291e-01
6.41273737e-01 3.19966227e-01 -8.37032855e-01 2.51943231e-01
-7.27374852e-02 -8.82813156e-01 -3.65015835e-01 -8.00687551e-01
-1.39148629e+00 9.33196962e-01 3.77315700e-01 6.87285960e-01
1.10017312e+00 -5.51513672e-01 7.02744722e-01 -1.02841459e-01
5.32030344e-01 -1.08349800e+00 2.38664988e-02 -1.64063260e-01
2.12113768e-01 -1.10638952e+00 2.35030428e-01 -5.20620465e-01
-6.29735291e-01 5.85373700e-01 8.19488585e-01 -1.78156301e-01
6.41366541e-01 1.83572382e-01 1.14799067e-01 9.76769626e-02
-4.97716993e-01 1.36932656e-01 2.27733459e-02 9.59443331e-01
1.70268640e-01 7.40857720e-01 -4.91427034e-01 1.02322352e+00
-2.39352643e-01 -1.09014459e-01 9.18492138e-01 6.70369864e-01
-2.16220692e-01 -1.41900873e+00 -5.75409889e-01 3.28912646e-01
-9.07177210e-01 -3.55287665e-03 -9.59220529e-01 5.20761430e-01
-9.54179168e-02 1.50792646e+00 -2.65857697e-01 -5.41111708e-01
-7.00450130e-03 -2.78794944e-01 -7.49068707e-02 -1.39189690e-01
-1.18261099e+00 -4.51629102e-01 -2.10503355e-01 -6.39910161e-01
-6.28529847e-01 -3.53984743e-01 -1.13015699e+00 -1.55215037e+00
-4.12967622e-01 4.84199584e-01 2.53447257e-02 5.16898513e-01
2.90789098e-01 -4.66745459e-02 1.10932636e+00 -6.35045648e-01
-7.06396878e-01 -3.38101357e-01 -7.77335227e-01 8.90880227e-01
-2.74431825e-01 -4.84958380e-01 -5.58055222e-01 -2.65050024e-01] | [7.3006720542907715, 5.2504496574401855] |
ebda8f31-9758-436f-b52d-899224ec62c4 | very-fast-approximate-counterfactual | 2303.02883 | null | https://arxiv.org/abs/2303.02883v1 | https://arxiv.org/pdf/2303.02883v1.pdf | Very fast, approximate counterfactual explanations for decision forests | We consider finding a counterfactual explanation for a classification or regression forest, such as a random forest. This requires solving an optimization problem to find the closest input instance to a given instance for which the forest outputs a desired value. Finding an exact solution has a cost that is exponential on the number of leaves in the forest. We propose a simple but very effective approach: we constrain the optimization to only those input space regions defined by the forest that are populated by actual data points. The problem reduces to a form of nearest-neighbor search using a certain distance on a certain dataset. This has two advantages: first, the solution can be found very quickly, scaling to large forests and high-dimensional data, and enabling interactive use. Second, the solution found is more likely to be realistic in that it is guided towards high-density areas of input space. | ['Suryabhan Singh Hada', 'Miguel Á. Carreira-Perpiñán'] | 2023-03-06 | null | null | null | null | ['counterfactual-explanation'] | ['miscellaneous'] | [ 4.92726952e-01 4.08331096e-01 -6.83973253e-01 -3.14394325e-01
-5.66380918e-01 -8.03201497e-01 5.74040949e-01 9.70282704e-02
-3.73399854e-01 1.13435912e+00 1.20750725e-01 -7.96337843e-01
-4.60590124e-01 -1.28049958e+00 -6.71723127e-01 -4.98179972e-01
-1.91685766e-01 9.06109571e-01 2.26236746e-01 1.88337252e-01
4.37428027e-01 7.16481268e-01 -1.78658164e+00 8.58015195e-02
8.34185541e-01 7.98471630e-01 3.32765311e-01 7.03567445e-01
-2.24791184e-01 1.10836424e-01 -4.97593462e-01 -3.92871380e-01
5.61903954e-01 -3.75109583e-01 -1.00784087e+00 1.48530692e-01
2.59206414e-01 -1.04360655e-01 2.43536249e-01 8.15271735e-01
-1.21669201e-02 1.76192731e-01 7.47004449e-01 -1.51901281e+00
-3.52786809e-01 3.75139177e-01 -6.41534209e-01 7.55811781e-02
3.23761523e-01 2.20565423e-01 1.02761185e+00 -4.98867571e-01
8.35562468e-01 1.21708739e+00 4.69629705e-01 3.89233887e-01
-1.93664765e+00 -4.88309652e-01 3.66629630e-01 -5.62505499e-02
-1.24585176e+00 -1.77259564e-01 3.79010409e-01 -3.13227117e-01
8.23367953e-01 6.82794929e-01 9.97052491e-01 4.52689826e-01
3.34845036e-02 4.36073333e-01 1.22671080e+00 -4.06384319e-01
6.11372709e-01 1.41056687e-01 -1.97668925e-01 4.24457699e-01
4.36352253e-01 3.47681820e-01 -4.21741098e-01 -5.72009802e-01
7.88700879e-01 1.81149870e-01 -2.65539765e-01 -8.03273976e-01
-9.99872983e-01 1.04908359e+00 7.57388413e-01 -2.26081032e-02
-4.34200764e-01 1.81964725e-01 -5.39623722e-02 3.02944154e-01
3.53427708e-01 8.34779263e-01 -5.80211580e-01 1.04911298e-01
-1.18890512e+00 7.14082241e-01 9.82248127e-01 6.13790095e-01
9.13337111e-01 -4.38216031e-01 3.39557081e-01 4.14343506e-01
1.55665755e-01 3.89396757e-01 9.23146904e-02 -1.23091304e+00
5.77627778e-01 6.06395304e-01 4.52928007e-01 -7.09992707e-01
-1.52002901e-01 -4.42024976e-01 -5.71884930e-01 5.41027546e-01
9.38094854e-01 -2.21757982e-02 -8.51896942e-01 1.56247878e+00
5.80152214e-01 -1.25656500e-01 1.45524461e-03 9.84568179e-01
-1.64532721e-01 7.26134002e-01 -4.31316495e-02 -5.79900861e-01
1.07971370e+00 -5.37502766e-01 -3.64704341e-01 -6.72489285e-01
4.42472637e-01 -4.38792408e-01 1.25087762e+00 2.45119080e-01
-7.05767691e-01 7.14625269e-02 -9.62104321e-01 1.31822929e-01
-4.83802676e-01 -4.54928815e-01 9.22874689e-01 4.94284451e-01
-7.35503554e-01 7.76277065e-01 -3.98287028e-01 -3.55817288e-01
3.66634339e-01 3.89201254e-01 -2.96010017e-01 -3.12485546e-01
-8.72144938e-01 8.21467817e-01 3.94241363e-01 -4.84088100e-02
-5.71028233e-01 -6.67459309e-01 -5.60074568e-01 1.37140155e-01
6.91202939e-01 -6.90484524e-01 1.17246890e+00 -9.96172428e-01
-7.71203697e-01 7.68629372e-01 -6.42490029e-01 -5.49570620e-01
6.51880860e-01 6.87163323e-02 -1.40875736e-02 -4.19914454e-01
4.74243522e-01 6.04740560e-01 5.44412673e-01 -1.36804390e+00
-6.96424901e-01 -8.12974453e-01 1.30019680e-01 4.04815555e-01
2.08455130e-01 -1.81002825e-01 3.69713083e-02 -2.94328213e-01
5.64455926e-01 -9.62193072e-01 -9.38228190e-01 1.78866774e-01
-5.01420140e-01 1.03835808e-02 6.13808453e-01 -3.08373332e-01
1.23168480e+00 -1.75434566e+00 -1.83445945e-01 5.55611372e-01
2.88528260e-02 -3.15674186e-01 1.53196275e-01 4.57056314e-01
1.86314657e-02 6.71627283e-01 -5.29936075e-01 4.02478755e-01
-2.28731096e-01 4.35373783e-01 -5.48191011e-01 4.37061995e-01
8.61607939e-02 6.48777843e-01 -9.32020068e-01 -5.15264511e-01
4.85851094e-02 -3.05982679e-01 -5.73868215e-01 1.16431862e-01
-4.97261167e-01 3.61921340e-02 -3.93691540e-01 3.83043677e-01
7.62018085e-01 -1.09566405e-01 4.76455599e-01 5.69705606e-01
-3.49054784e-01 4.42528963e-01 -1.64072692e+00 1.03367496e+00
-5.32483160e-01 4.21972007e-01 -6.34882003e-02 -7.28926539e-01
9.61941779e-01 -5.59430197e-02 3.46440226e-01 -1.43063188e-01
-4.86310929e-01 3.20892572e-01 -1.16383083e-01 -1.92403737e-02
2.61553168e-01 -6.99054956e-01 2.65218448e-02 8.73452246e-01
-4.46397096e-01 -4.27922159e-01 7.63022676e-02 -7.80096203e-02
1.04111922e+00 -4.99785990e-02 8.23947787e-01 -4.73841757e-01
-2.18529448e-01 4.85327393e-01 5.61646461e-01 9.74867105e-01
1.64158195e-01 4.20402169e-01 7.68884540e-01 -7.18769491e-01
-9.33721840e-01 -1.09825945e+00 -2.98497647e-01 7.10341394e-01
7.86957815e-02 -1.64478511e-01 -5.07903457e-01 -8.19241405e-01
3.53334516e-01 9.93080616e-01 -7.15517759e-01 2.35155538e-01
-3.66831303e-01 -5.29886127e-01 -2.92584568e-01 3.53259712e-01
3.66308779e-01 -9.55939770e-01 -9.26682413e-01 2.07526624e-01
-2.11904094e-01 -4.77761239e-01 -2.77320832e-01 6.62045360e-01
-1.61636806e+00 -1.09111071e+00 -4.66215849e-01 -4.12703663e-01
8.62049758e-01 4.43109602e-01 1.26113808e+00 -5.21332286e-02
-1.39245763e-01 -4.13614094e-01 1.55749530e-01 -5.70094645e-01
3.81376245e-03 2.29065359e-01 -1.76223606e-01 -5.09376466e-01
5.28284788e-01 -7.05475092e-01 -3.99721175e-01 4.38019514e-01
-4.97806430e-01 -3.34345060e-03 5.71425438e-01 8.02083254e-01
8.25044870e-01 4.29963470e-01 4.29368883e-01 -1.18240857e+00
5.64772308e-01 -6.89525425e-01 -7.31553614e-01 2.65924454e-01
-1.00667572e+00 3.28102201e-01 6.13026738e-01 -2.85766900e-01
-6.57828271e-01 6.39542341e-01 4.87836897e-01 -9.93020833e-02
-4.06098157e-01 5.12702346e-01 -2.52881587e-01 4.67018932e-01
1.03279805e+00 3.57891582e-02 -1.33290485e-01 -5.62515795e-01
5.93313873e-01 6.09200180e-01 3.77159089e-01 -5.02819002e-01
8.74235570e-01 3.03117633e-01 3.23559791e-01 -5.87472022e-01
-7.38082290e-01 -2.10636824e-01 -7.57704914e-01 -1.69335708e-01
4.05919909e-01 -3.32692623e-01 -6.62938833e-01 -2.70316869e-01
-9.32624519e-01 -3.78992081e-01 -7.04959333e-01 2.47353479e-01
-7.16369271e-01 -2.51839072e-01 4.80588615e-01 -1.19575250e+00
1.58405900e-01 -7.15682685e-01 7.78438866e-01 1.42868906e-01
-5.08474827e-01 -7.35626519e-01 7.51174316e-02 2.43026599e-01
1.18427388e-01 4.04503703e-01 1.11991215e+00 -4.43202078e-01
-1.07274997e+00 -2.67761588e-01 -1.57279864e-01 -4.44202185e-01
1.20888062e-01 2.66228169e-02 -7.83897996e-01 -1.84426263e-01
-2.12922558e-01 -6.39535785e-02 7.92432308e-01 6.08204126e-01
1.25826621e+00 -9.92734849e-01 -7.25690126e-01 4.00933683e-01
1.64043701e+00 1.40383944e-01 3.12223047e-01 5.12078106e-01
-1.33726345e-02 9.19048131e-01 1.06460118e+00 2.31747761e-01
2.04735994e-01 6.78404570e-01 6.71576023e-01 -2.45082662e-01
4.69709128e-01 -6.74928904e-01 -1.80120602e-01 -4.48481560e-01
-2.15643719e-02 -1.08611286e-01 -1.08285666e+00 9.76154864e-01
-2.08207488e+00 -1.25870681e+00 -1.48665786e-01 2.71431470e+00
5.50882399e-01 1.66077346e-01 4.81174797e-01 2.82633096e-01
5.45271754e-01 1.10847846e-01 -8.77930582e-01 -7.20917821e-01
2.16088131e-01 7.68909752e-02 7.51532257e-01 6.83874726e-01
-7.70642936e-01 6.67215407e-01 7.34751225e+00 4.09273922e-01
-7.16323435e-01 -1.96405888e-01 8.30440521e-01 -4.74066883e-01
-5.67339182e-01 4.40278083e-01 -5.68344831e-01 3.18650126e-01
8.31234455e-01 -4.64871615e-01 8.59088004e-01 1.08690774e+00
3.63949835e-01 -6.63590133e-01 -1.14396656e+00 4.58229363e-01
-6.45961583e-01 -1.52527189e+00 -1.40081599e-01 6.94095254e-01
6.52888954e-01 -1.73482165e-01 -1.14952803e-01 -1.87289029e-01
8.76419306e-01 -1.35510373e+00 5.67575872e-01 2.68338799e-01
9.01040912e-01 -1.03387415e+00 4.07966405e-01 8.42340708e-01
-1.00950873e+00 -2.89664537e-01 -3.68292242e-01 -4.34315324e-01
-8.29428434e-03 8.72586668e-01 -1.23861325e+00 1.78189084e-01
8.02447557e-01 8.34886581e-02 -1.45618170e-01 1.08735120e+00
-2.47325480e-01 5.36995530e-01 -7.31974602e-01 -2.79882461e-01
9.91770923e-02 -2.82629043e-01 3.90294611e-01 9.54325855e-01
2.71324039e-01 1.64282937e-02 2.48543665e-01 8.79137933e-01
2.28091046e-01 1.18462868e-01 -1.40409720e+00 1.78439632e-01
6.95810497e-01 9.07707274e-01 -8.70311379e-01 4.80463393e-02
-4.82501388e-02 6.30814373e-01 4.48553711e-01 2.52005607e-01
-4.28893954e-01 -3.04089457e-01 8.02987337e-01 5.19824147e-01
1.44520417e-01 -3.56994793e-02 -7.90971875e-01 -8.07761133e-01
1.74499482e-01 -8.52677166e-01 5.95700324e-01 -6.32116973e-01
-9.85496938e-01 3.85211051e-01 1.10302173e-01 -9.36114371e-01
-6.09494925e-01 -3.96097183e-01 -5.90461493e-01 1.21326435e+00
-1.11617160e+00 -6.32984579e-01 7.11078849e-03 3.64532262e-01
2.76665866e-01 2.87067890e-01 1.01246393e+00 -4.69336808e-01
-5.32950833e-03 9.91434753e-02 -4.50075902e-02 -4.53061223e-01
7.97857414e-04 -1.63621914e+00 4.68393594e-01 9.00974333e-01
3.52069736e-01 6.67407453e-01 9.31708694e-01 -6.86253250e-01
-9.39281642e-01 -8.94621313e-01 1.36009169e+00 -3.24063927e-01
4.84452367e-01 -2.83916771e-01 -6.49277747e-01 7.82018304e-01
-3.18683386e-01 -5.03089987e-02 6.02934182e-01 7.22947896e-01
-1.60506546e-01 -1.52440161e-01 -1.64659762e+00 8.19449186e-01
1.29153407e+00 -3.24859053e-01 -2.96830386e-01 4.79677677e-01
3.32762599e-01 -1.27141133e-01 -6.54844820e-01 1.48267016e-01
6.29709601e-01 -8.60883594e-01 7.63251245e-01 -9.53029394e-01
6.19318187e-01 -4.65481967e-01 -4.09122318e-01 -1.52276218e+00
-3.31080824e-01 -6.74770534e-01 -1.66975695e-03 9.53595281e-01
9.14586067e-01 -6.87072992e-01 1.19989109e+00 9.06073749e-01
6.85240090e-01 -1.21496296e+00 -1.24680090e+00 -8.22147012e-01
8.87110531e-02 -3.28072160e-01 1.15711880e+00 7.39289761e-01
-8.09438340e-03 4.08231348e-01 -6.50225282e-02 1.48634151e-01
8.05502832e-01 9.29926872e-01 9.22402859e-01 -1.53514218e+00
-1.48263693e-01 -1.11364484e-01 -1.79461241e-01 -1.12624645e+00
-1.64619178e-01 -8.87449384e-01 -4.85010371e-02 -1.76450253e+00
2.53139883e-01 -9.68192220e-01 2.24882886e-01 5.10096908e-01
-1.60072938e-01 -2.58063257e-01 1.81108952e-01 3.26890141e-01
1.54275432e-01 1.17534474e-01 1.15612435e+00 2.11006910e-01
-2.96523571e-01 6.96436524e-01 -1.05451727e+00 6.23170853e-01
8.51133466e-01 -9.20697868e-01 -4.45654362e-01 -2.25017853e-02
-4.46591452e-02 3.97166401e-01 4.77516830e-01 -5.82728982e-01
-1.17082573e-01 -1.05210149e+00 5.50928950e-01 -5.81525326e-01
1.98922291e-01 -1.18647110e+00 6.30207837e-01 5.61041534e-01
-5.68987370e-01 1.37632892e-01 -1.49583548e-01 6.92720711e-01
2.13590592e-01 -6.65594578e-01 7.86949217e-01 -4.36094970e-01
-2.86793232e-01 6.89198524e-02 -3.46472204e-01 -8.97887424e-02
1.17777514e+00 -5.76106369e-01 -1.54617921e-01 -5.88244975e-01
-5.64226031e-01 3.18575084e-01 8.29957485e-01 3.96297388e-02
3.12481046e-01 -1.08869636e+00 -5.11492848e-01 5.60111441e-02
3.72815989e-02 3.11137348e-01 -3.90688956e-01 1.35391191e-01
-1.81601703e-01 4.66737062e-01 -2.84354426e-02 -4.84951228e-01
-1.20889366e+00 7.13782787e-01 4.15491104e-01 -5.35910189e-01
-6.42965138e-01 5.86002409e-01 -3.94583633e-03 -6.41799748e-01
-1.00917593e-01 -1.92597494e-01 2.72841394e-01 -3.97282802e-02
3.78279001e-01 3.79741281e-01 -2.49940559e-01 -1.10197276e-01
-4.83023942e-01 1.56280339e-01 2.98729539e-01 -5.94791532e-01
1.48867035e+00 2.65754580e-01 -6.14437275e-02 4.30226058e-01
7.81597495e-01 -4.35864702e-02 -1.36720896e+00 -1.48350909e-01
2.73015648e-01 -1.13747501e+00 7.38166943e-02 -1.04603553e+00
-8.48647475e-01 6.10721648e-01 3.42867404e-01 6.13270104e-01
1.02249479e+00 1.16885811e-01 9.16268304e-02 3.90042931e-01
6.53607786e-01 -7.84743667e-01 -4.76422608e-01 8.24326202e-02
9.99377131e-01 -1.02796543e+00 3.10362488e-01 -4.13827300e-01
-6.17150962e-01 1.02664006e+00 6.60437763e-01 -6.13836497e-02
6.52744949e-01 3.77862304e-01 2.70826835e-02 -6.47141114e-02
-1.00826740e+00 -2.53653944e-01 -2.91492511e-02 8.52455258e-01
1.50161952e-01 4.51147795e-01 -2.03655437e-01 2.21206442e-01
-7.22554803e-01 1.39685437e-01 3.76412183e-01 7.50904262e-01
-6.85653150e-01 -9.21525359e-01 -4.72183406e-01 1.01059866e+00
-1.47495717e-01 2.41505224e-02 -7.24511981e-01 8.42792690e-01
6.12041503e-02 1.06058073e+00 2.64096886e-01 1.97844822e-02
4.87964958e-01 1.72801957e-01 2.50863403e-01 -7.45139599e-01
-3.64985764e-01 -2.36184075e-01 9.65230167e-02 -5.71717381e-01
-6.80584460e-02 -9.37981725e-01 -1.05794120e+00 -5.68823040e-01
-6.13413334e-01 3.96984577e-01 6.66752636e-01 7.14963198e-01
2.42236078e-01 -5.41516021e-02 8.03547859e-01 -5.97752810e-01
-6.80906117e-01 -5.70353091e-01 -7.10678279e-01 -1.39053300e-01
3.72767717e-01 -5.78267753e-01 -3.32313806e-01 -3.68494540e-01] | [8.616141319274902, 5.459455490112305] |
50aa13a2-c2fa-4de6-b74f-cf4c6ea77fd1 | clip-guided-prototype-modulating-for-few-shot | 2303.02982 | null | https://arxiv.org/abs/2303.02982v1 | https://arxiv.org/pdf/2303.02982v1.pdf | CLIP-guided Prototype Modulating for Few-shot Action Recognition | Learning from large-scale contrastive language-image pre-training like CLIP has shown remarkable success in a wide range of downstream tasks recently, but it is still under-explored on the challenging few-shot action recognition (FSAR) task. In this work, we aim to transfer the powerful multimodal knowledge of CLIP to alleviate the inaccurate prototype estimation issue due to data scarcity, which is a critical problem in low-shot regimes. To this end, we present a CLIP-guided prototype modulating framework called CLIP-FSAR, which consists of two key components: a video-text contrastive objective and a prototype modulation. Specifically, the former bridges the task discrepancy between CLIP and the few-shot video task by contrasting videos and corresponding class text descriptions. The latter leverages the transferable textual concepts from CLIP to adaptively refine visual prototypes with a temporal Transformer. By this means, CLIP-FSAR can take full advantage of the rich semantic priors in CLIP to obtain reliable prototypes and achieve accurate few-shot classification. Extensive experiments on five commonly used benchmarks demonstrate the effectiveness of our proposed method, and CLIP-FSAR significantly outperforms existing state-of-the-art methods under various settings. The source code and models will be publicly available at https://github.com/alibaba-mmai-research/CLIP-FSAR. | ['Nong Sang', 'Deli Zhao', 'Yingya Zhang', 'Changxin Gao', 'Jun Cen', 'Shiwei Zhang', 'Xiang Wang'] | 2023-03-06 | null | null | null | null | ['few-shot-action-recognition'] | ['computer-vision'] | [ 3.51049900e-01 -4.96875972e-01 -4.49567705e-01 -2.44492859e-01
-1.13799739e+00 -3.27498496e-01 6.76650107e-01 -2.92524129e-01
-1.87602982e-01 4.42698151e-01 4.92144972e-01 2.45862827e-01
2.74465904e-02 -1.85749456e-01 -7.87429690e-01 -8.36094975e-01
2.59209871e-01 1.20343693e-01 2.50989527e-01 -2.08185732e-01
1.77938491e-01 -1.04300022e-01 -1.63111556e+00 7.05469489e-01
7.18892574e-01 1.17334378e+00 5.28260767e-01 3.91071618e-01
1.44902006e-01 9.50534046e-01 -1.46071553e-01 -1.84166491e-01
2.45198071e-01 -6.55052841e-01 -3.68972331e-01 3.96145433e-01
4.43270743e-01 -3.87642682e-01 -6.08789861e-01 1.05812478e+00
4.58384246e-01 4.87863272e-01 5.69951594e-01 -1.29825985e+00
-6.81268930e-01 5.72321713e-01 -7.08033323e-01 4.03204769e-01
3.31647873e-01 4.99106824e-01 1.08309782e+00 -1.45147479e+00
6.12948775e-01 1.30935907e+00 4.49805975e-01 4.73445296e-01
-1.03178763e+00 -7.83875883e-01 3.99363011e-01 7.35870004e-01
-1.74109709e+00 -8.53749096e-01 8.40942621e-01 -5.40174127e-01
7.35375345e-01 4.41530533e-02 7.92626321e-01 1.44754708e+00
-1.73882842e-01 1.18232310e+00 9.50873852e-01 -2.45441765e-01
2.63435274e-01 -2.14935597e-02 -9.70293507e-02 6.74435258e-01
-2.07771748e-01 2.30878256e-02 -7.62305975e-01 1.75432667e-01
5.72783530e-01 3.71343374e-01 -6.29307687e-01 -3.25197637e-01
-1.14071262e+00 7.91385949e-01 2.26425439e-01 4.47697252e-01
-3.74736339e-01 1.18383914e-01 5.79199195e-01 1.57258511e-01
5.40541232e-01 2.04041600e-01 -2.23910227e-01 -2.50248939e-01
-1.07559311e+00 1.17038406e-01 2.44311437e-01 1.03944695e+00
4.54600453e-01 1.59842357e-01 -6.68116450e-01 1.06951070e+00
7.16699734e-02 5.32812476e-01 5.60282648e-01 -9.00943696e-01
5.39146006e-01 2.19575509e-01 -4.02287468e-02 -8.21854472e-01
4.89433855e-02 -3.74041557e-01 -7.54626095e-01 -2.48456374e-01
1.44039556e-01 1.88270688e-01 -8.83572876e-01 1.64074123e+00
2.78662056e-01 6.91549420e-01 -9.50408354e-02 1.16038167e+00
7.35216081e-01 8.72718513e-01 2.04179719e-01 -3.08742315e-01
1.39409220e+00 -1.22125733e+00 -6.44769669e-01 -2.74548918e-01
3.55926335e-01 -6.22923672e-01 1.25907946e+00 2.61977881e-01
-1.05512750e+00 -7.36102343e-01 -9.96887684e-01 -9.12533998e-02
-9.14819539e-02 1.63110510e-01 3.85976821e-01 1.33288786e-01
-7.61764765e-01 3.07956159e-01 -8.31427038e-01 -3.84309739e-01
9.57146406e-01 -1.46777883e-01 -2.72705704e-01 -6.44836426e-01
-1.20964193e+00 4.00326073e-01 5.52441478e-01 7.57098645e-02
-1.13258886e+00 -9.14733171e-01 -1.03710854e+00 1.18834428e-01
9.30645049e-01 -5.86248577e-01 1.31578290e+00 -1.18497300e+00
-1.52254045e+00 6.31965697e-01 -2.21196502e-01 -3.55728149e-01
5.73013604e-01 -2.54391611e-01 -3.35286230e-01 5.30203104e-01
1.28776923e-01 8.40513825e-01 1.12380385e+00 -1.06761134e+00
-7.93878138e-01 -1.79608136e-01 7.97910523e-03 3.85012060e-01
-5.82188845e-01 1.25237573e-02 -8.89797628e-01 -9.09985542e-01
-3.54103804e-01 -7.69958198e-01 1.16389230e-01 2.00947434e-01
-1.55755773e-01 -2.13226661e-01 7.65043020e-01 -5.72452366e-01
1.12731838e+00 -2.31731987e+00 3.41205716e-01 -3.69117677e-01
-6.85457364e-02 5.17364264e-01 -3.92945737e-01 5.72276294e-01
3.87358181e-02 -2.65850365e-01 -3.24684858e-01 -4.77625370e-01
-2.65686330e-03 -6.09728508e-02 -5.10584652e-01 5.64797223e-01
3.02278131e-01 1.17648911e+00 -9.86236691e-01 -6.91629946e-01
3.33256036e-01 6.17309630e-01 -4.28184360e-01 2.59940892e-01
-4.16746765e-01 4.04482454e-01 -5.74421942e-01 9.21059191e-01
3.36237967e-01 -3.52089435e-01 -2.00352237e-01 -4.19212699e-01
-1.54692635e-01 -3.22609723e-01 -8.27112079e-01 2.08573771e+00
-1.60978377e-01 7.28483737e-01 -7.48096928e-02 -1.33951735e+00
5.49885035e-01 2.86355317e-01 5.32849014e-01 -9.24726665e-01
1.97779804e-01 1.26567595e-02 -2.71737576e-01 -7.94253707e-01
3.77199113e-01 -3.07759285e-01 6.22539930e-02 1.97629049e-01
2.28953436e-01 1.17464565e-01 4.23195630e-01 2.91307569e-01
9.42722559e-01 3.49426508e-01 3.51062655e-01 2.41202321e-02
6.15113199e-01 -9.63989049e-02 6.83403015e-01 6.43885553e-01
-4.50021952e-01 7.13268995e-01 2.64306724e-01 1.31263837e-01
-8.48997891e-01 -9.38937783e-01 -1.53880017e-02 1.41548038e+00
1.82774439e-01 -4.49344784e-01 -4.17170107e-01 -5.37151456e-01
-7.96798170e-02 6.57422125e-01 -6.35165036e-01 -2.37305477e-01
-4.97602642e-01 -2.91696072e-01 2.61796057e-01 6.27341330e-01
4.05374557e-01 -9.84486878e-01 -3.82599920e-01 2.97086616e-03
-3.92131746e-01 -1.42201698e+00 -8.87417436e-01 -2.91528851e-01
-5.24989188e-01 -9.95007277e-01 -1.13931215e+00 -7.97769725e-01
4.43335831e-01 8.23191762e-01 7.97433615e-01 -1.43538281e-01
-4.07906771e-01 6.85760736e-01 -8.25597465e-01 -1.81779742e-01
-3.94398272e-02 -2.49666318e-01 -1.58120960e-01 3.58014345e-01
2.82686442e-01 -5.68997264e-01 -7.74681926e-01 2.63929605e-01
-1.07767892e+00 3.76154840e-01 7.74210632e-01 9.65946198e-01
8.29787374e-01 -1.97962672e-01 6.72783494e-01 -6.28462553e-01
2.79973418e-01 -7.09334850e-01 -3.67306054e-01 3.18064928e-01
-8.98266286e-02 -2.79018968e-01 7.59441972e-01 -6.89192653e-01
-1.11814320e+00 1.44643709e-01 4.49666530e-02 -1.19841003e+00
2.65412219e-02 6.27223909e-01 -2.78324157e-01 1.84019685e-01
2.80547738e-01 5.79882145e-01 -1.56873599e-01 -3.41294110e-01
5.21082163e-01 7.58696496e-01 6.38967454e-01 -5.53947449e-01
7.40265846e-01 7.03450263e-01 -2.65369922e-01 -9.40579236e-01
-1.20107114e+00 -8.70343029e-01 -4.64675784e-01 -4.96596724e-01
8.15221190e-01 -1.31887281e+00 -2.38227487e-01 5.07505357e-01
-8.02215219e-01 -5.89778364e-01 -3.08892071e-01 4.48197871e-01
-7.65204549e-01 5.09390414e-01 -5.48480153e-01 -6.84065402e-01
-3.68021995e-01 -9.46582854e-01 1.19410861e+00 2.44148836e-01
2.54722625e-01 -6.81285799e-01 -2.52875667e-02 5.71171761e-01
2.13258967e-01 3.44517082e-02 4.84249234e-01 -5.69141626e-01
-5.81710815e-01 -1.04386292e-01 -3.38792235e-01 4.28046227e-01
-2.04363447e-02 -2.62690276e-01 -9.54781234e-01 -4.69566703e-01
3.01021151e-02 -7.39424467e-01 1.23129427e+00 3.48848313e-01
1.29968953e+00 -1.23844385e-01 -1.85090542e-01 5.91144979e-01
1.38058197e+00 4.11077775e-02 4.94875103e-01 -1.28686223e-02
7.68469155e-01 4.33073670e-01 1.17527056e+00 7.81412542e-01
2.77804255e-01 8.85251701e-01 1.05252445e-01 2.44026273e-01
-4.29212570e-01 -3.79056603e-01 5.75081766e-01 9.08240557e-01
7.97652230e-02 -2.78649956e-01 -7.92674541e-01 5.28044999e-01
-2.16563416e+00 -1.26899493e+00 3.66577148e-01 1.91965854e+00
8.02825451e-01 -6.65140674e-02 1.83352485e-01 -1.31838113e-01
8.97143006e-01 5.76184392e-01 -7.62509465e-01 4.27054614e-01
-1.40485436e-01 -7.58410618e-02 -3.64115648e-02 8.84165391e-02
-1.26948476e+00 1.03727436e+00 4.26770544e+00 1.21804476e+00
-9.73146081e-01 3.30600172e-01 5.86024284e-01 -4.42481339e-01
1.45221040e-01 -9.92875025e-02 -7.74317026e-01 6.73160553e-01
6.78241551e-01 -3.05314451e-01 3.95211965e-01 7.91732848e-01
4.33331907e-01 -2.14434750e-02 -1.14630187e+00 1.26772118e+00
6.72822416e-01 -1.42181110e+00 1.00434355e-01 -3.44681025e-01
7.97230005e-01 8.12472403e-02 2.32382953e-01 6.25860393e-01
-2.74004906e-01 -7.23051667e-01 9.41136301e-01 6.82970405e-01
9.35673773e-01 -5.71603298e-01 3.83353829e-01 1.47886127e-01
-1.53399038e+00 -4.39725399e-01 -4.20904487e-01 6.65143579e-02
3.63983333e-01 4.15028989e-01 -1.88020751e-01 4.97661263e-01
5.56307793e-01 1.35199022e+00 -4.35805261e-01 1.17924500e+00
-1.22115470e-01 5.76241314e-01 1.18908502e-01 1.61163986e-01
3.48605245e-01 -1.08082995e-01 6.02222383e-01 1.26916957e+00
2.55261511e-01 4.43073481e-01 4.38026071e-01 8.74364972e-01
-1.36379510e-01 2.21364543e-01 -2.26862118e-01 -3.99775922e-01
5.30087411e-01 1.15439928e+00 -5.54105878e-01 -5.02860188e-01
-7.19093144e-01 1.05211401e+00 2.91755468e-01 4.70102400e-01
-1.03382313e+00 -2.03340843e-01 5.06731391e-01 -1.07240766e-01
8.13336253e-01 -7.31595233e-02 4.00472522e-01 -1.46647918e+00
1.74544543e-01 -1.00177073e+00 5.06740630e-01 -9.51528907e-01
-1.41865647e+00 2.48351917e-01 1.75561041e-01 -1.49163342e+00
-8.03418830e-02 -4.37054873e-01 -6.44194603e-01 3.50273848e-01
-1.61498153e+00 -1.51359093e+00 -4.35934722e-01 8.04859400e-01
1.25017226e+00 -1.10299818e-01 3.59021544e-01 4.61442620e-01
-8.11054409e-01 5.54651678e-01 2.45688632e-01 7.50275031e-02
9.44883883e-01 -6.44876540e-01 -5.36799617e-03 9.04842615e-01
2.47278661e-01 1.67501733e-01 6.37125492e-01 -5.24164021e-01
-1.65887094e+00 -1.41786146e+00 1.47596315e-01 -8.45965520e-02
8.58598411e-01 -4.51290071e-01 -1.06647909e+00 5.53306162e-01
4.00212072e-02 4.20251608e-01 6.13481462e-01 -3.55769545e-01
-5.74854553e-01 -1.22604415e-01 -5.30460656e-01 6.07322395e-01
1.24822509e+00 -6.44889355e-01 -5.16890407e-01 4.93729919e-01
7.38179207e-01 -1.35030642e-01 -7.33903348e-01 2.76128322e-01
6.13462389e-01 -8.85143638e-01 1.00658369e+00 -4.74398822e-01
6.39134228e-01 -1.77921757e-01 -4.51509148e-01 -1.17737663e+00
-2.43049741e-01 -4.18883264e-01 -4.60173875e-01 1.32137990e+00
-8.29422176e-02 -3.15798312e-01 4.25978601e-01 1.09727241e-01
-4.25763965e-01 -8.91440094e-01 -8.07191074e-01 -9.42768693e-01
-7.97979161e-02 -4.85735983e-01 1.47442579e-01 8.68855119e-01
5.69690578e-02 3.02011698e-01 -6.85110927e-01 -2.01077193e-01
7.21873283e-01 3.07316303e-01 7.01787293e-01 -7.69560516e-01
-6.04936421e-01 -4.84242022e-01 -4.12122369e-01 -1.17210686e+00
4.47933018e-01 -8.11847150e-01 3.81201893e-01 -1.25437748e+00
6.77147746e-01 3.27814883e-03 -4.62088406e-01 4.20921028e-01
-4.09494281e-01 3.84803236e-01 5.36532879e-01 4.24731970e-01
-1.31160021e+00 1.08817387e+00 1.20745885e+00 -3.51615161e-01
-1.37963891e-01 -2.58113682e-01 -6.39646173e-01 5.90540051e-01
6.77946329e-01 -3.82710487e-01 -5.65538645e-01 -1.46938935e-01
-1.23849608e-01 1.45073742e-01 4.14570749e-01 -1.03691411e+00
2.30253071e-01 -2.78994143e-01 2.54431486e-01 -4.47716117e-01
6.46939814e-01 -4.95288998e-01 -9.41125825e-02 2.77728528e-01
-4.85518277e-01 -4.58750129e-01 4.94480990e-02 1.02123380e+00
-3.25284183e-01 -5.74310422e-02 8.74858916e-01 -4.07787077e-02
-1.17982996e+00 5.96459985e-01 -5.60730472e-02 2.82371312e-01
1.18844235e+00 -8.35951343e-02 -5.00979185e-01 -3.88892204e-01
-3.36299688e-01 3.79972935e-01 5.01591086e-01 7.26936102e-01
8.63333464e-01 -1.32973313e+00 -8.74865234e-01 5.05179316e-02
6.27174377e-01 -3.34665954e-01 8.35469544e-01 1.36864424e+00
4.09451388e-02 4.20521557e-01 -1.61591217e-01 -6.56244874e-01
-1.25865531e+00 8.62327695e-01 7.50947073e-02 1.23152457e-01
-1.03136396e+00 7.37530947e-01 6.20971322e-01 3.68355721e-01
4.11033332e-01 4.19232659e-02 -1.34004429e-01 8.35403875e-02
9.72451329e-01 2.87839890e-01 -3.67028594e-01 -9.39989150e-01
-1.82029039e-01 5.88514447e-01 -2.43046179e-01 -4.56233323e-02
1.36264944e+00 -3.42337251e-01 3.79540831e-01 7.04896808e-01
1.28313935e+00 -5.04745066e-01 -1.72282588e+00 -4.85955775e-01
-3.29620749e-01 -7.52159834e-01 9.35684294e-02 -5.23434460e-01
-1.11236703e+00 9.69146609e-01 5.70270181e-01 -2.98950434e-01
1.24059415e+00 2.42251381e-01 9.38637555e-01 3.43454212e-01
2.21189216e-01 -1.11829102e+00 6.31796062e-01 3.57647389e-01
1.05882847e+00 -1.43995750e+00 7.40635097e-02 -2.92984396e-01
-9.49479878e-01 8.37630749e-01 4.72899437e-01 -6.40481859e-02
4.27922130e-01 -2.20815167e-01 -1.98439136e-01 -1.16245352e-01
-9.26523983e-01 -4.58805710e-01 4.61318672e-01 4.53090400e-01
8.05906579e-02 -1.46735117e-01 -1.50279149e-01 8.75103414e-01
4.59694147e-01 2.10603401e-01 2.52180248e-01 1.00588667e+00
-4.81248468e-01 -6.58768773e-01 -1.72679916e-01 4.36112612e-01
-2.33555257e-01 -1.34343818e-01 -1.82743251e-01 6.68495357e-01
-5.78129403e-02 8.79824996e-01 -6.39261156e-02 -3.75650525e-01
1.64432287e-01 8.39631855e-02 5.23315907e-01 -5.71529508e-01
-1.38144493e-01 3.50082397e-01 -1.31971985e-01 -8.43345642e-01
-6.51416361e-01 -8.73966396e-01 -9.82379377e-01 8.86139497e-02
-2.29436785e-01 -1.02559380e-01 2.84079671e-01 9.91217434e-01
4.88224030e-01 4.17987019e-01 5.89607656e-01 -1.19553947e+00
-5.42830050e-01 -9.29348111e-01 -6.31397963e-01 5.82800090e-01
1.61315382e-01 -1.01721084e+00 -4.17304575e-01 2.79507220e-01] | [9.398029327392578, 0.8979353308677673] |
3472eedb-e3d6-49c1-8c8e-dfb14f32b4a7 | sampling-theorems-for-unsupervised-learning | 2203.12513 | null | https://arxiv.org/abs/2203.12513v2 | https://arxiv.org/pdf/2203.12513v2.pdf | Sensing Theorems for Unsupervised Learning in Linear Inverse Problems | Solving an ill-posed linear inverse problem requires knowledge about the underlying signal model. In many applications, this model is a priori unknown and has to be learned from data. However, it is impossible to learn the model using observations obtained via a single incomplete measurement operator, as there is no information about the signal model in the nullspace of the operator, resulting in a chicken-and-egg problem: to learn the model we need reconstructed signals, but to reconstruct the signals we need to know the model. Two ways to overcome this limitation are using multiple measurement operators or assuming that the signal model is invariant to a certain group action. In this paper, we present necessary and sufficient sensing conditions for learning the signal model from measurement data alone which only depend on the dimension of the model and the number of operators or properties of the group action that the model is invariant to. As our results are agnostic of the learning algorithm, they shed light into the fundamental limitations of learning from incomplete data and have implications in a wide range set of practical algorithms, such as dictionary learning, matrix completion and deep neural networks. | ['Mike Davies', 'Dongdong Chen', 'Julián Tachella'] | 2022-03-23 | null | null | null | null | ['matrix-completion'] | ['methodology'] | [ 7.65922189e-01 2.84389496e-01 -8.64795595e-03 -1.89440906e-01
-4.07658935e-01 -6.08989120e-01 2.05411807e-01 -1.86360911e-01
-2.82019556e-01 5.88041127e-01 2.60749876e-01 -2.36709550e-01
-6.18291378e-01 -4.88741964e-01 -8.10150623e-01 -9.70484853e-01
9.91070569e-02 5.14644325e-01 -4.85235125e-01 -3.81300360e-01
1.04879942e-02 3.63831997e-01 -1.21522057e+00 -1.04375355e-01
4.03261572e-01 9.86024857e-01 3.29310864e-01 6.10230684e-01
1.94743574e-01 7.84039617e-01 -3.93158138e-01 3.70395362e-01
7.15978682e-01 -8.12047899e-01 -5.38186789e-01 2.54665256e-01
8.19493607e-02 -2.48142958e-01 -5.94856858e-01 1.33901179e+00
2.70689130e-01 5.36808036e-02 6.21576965e-01 -8.17213058e-01
-3.73573810e-01 5.40387392e-01 -7.42141679e-02 5.63871972e-02
2.63631403e-01 -3.18295121e-01 1.08398700e+00 -7.34572172e-01
5.33274293e-01 6.86053455e-01 8.58539999e-01 4.03023750e-01
-1.58898425e+00 -3.85813922e-01 -3.21163476e-01 1.38970554e-01
-1.41911530e+00 -8.91896427e-01 1.04654181e+00 -4.88231570e-01
3.02602142e-01 3.30007941e-01 5.27763128e-01 1.06418753e+00
2.85143945e-02 3.33000660e-01 1.04363763e+00 -8.68144155e-01
4.59374547e-01 2.18119651e-01 3.36602107e-02 6.17533088e-01
2.08598509e-01 3.41187954e-01 -6.24341488e-01 -1.50485024e-01
1.01409638e+00 -1.74689189e-01 -8.23402226e-01 -9.31167483e-01
-1.35904574e+00 1.10735035e+00 1.92948163e-01 4.69349951e-01
-5.60353816e-01 -3.84433311e-03 4.40488197e-02 6.65733397e-01
-9.53946188e-02 7.07603872e-01 -3.60297233e-01 1.99407041e-01
-7.67451942e-01 -8.38799030e-02 1.25801563e+00 6.40460432e-01
1.10657239e+00 2.62398601e-01 5.17949343e-01 4.10720885e-01
2.03960001e-01 8.19490612e-01 3.27808559e-01 -1.15512419e+00
2.73406982e-01 -3.19719277e-02 2.29092658e-01 -1.04499543e+00
-5.18797994e-01 -7.86944389e-01 -9.94016349e-01 7.91793168e-02
8.77388597e-01 -4.48494911e-01 -6.23194098e-01 1.97902822e+00
-1.31568061e-02 3.45696181e-01 9.20601040e-02 1.19291162e+00
2.77421743e-01 4.41957086e-01 -9.22982335e-01 -4.44970459e-01
7.64189839e-01 -2.25436717e-01 -8.86572480e-01 -6.13330662e-01
6.69789970e-01 -7.98852801e-01 7.07390606e-01 6.15056753e-01
-7.44139791e-01 -4.28320408e-01 -1.36871946e+00 9.86427963e-02
2.53232628e-01 6.77777007e-02 5.74815631e-01 5.00867665e-01
-7.23566294e-01 5.94995797e-01 -7.83523440e-01 -2.53669053e-01
-2.31587902e-01 4.91672426e-01 -5.13752937e-01 -3.10775489e-01
-1.11961329e+00 1.06444228e+00 2.57011443e-01 3.83177370e-01
-7.22274363e-01 -5.43311954e-01 -6.98413193e-01 9.88164842e-02
4.50396150e-01 -5.43803871e-01 1.04233050e+00 -1.00115359e+00
-1.15273869e+00 4.84189570e-01 -6.93326741e-02 -4.49770123e-01
2.35640123e-01 6.88683987e-02 -3.50548655e-01 3.28327835e-01
-1.24605283e-01 -1.85614273e-01 1.27479970e+00 -1.23013830e+00
3.11244186e-02 -4.03789967e-01 1.19089626e-01 1.83725096e-02
-3.32518667e-02 -6.66026115e-01 1.34739250e-01 -4.41657841e-01
1.01255417e+00 -1.18685067e+00 -3.61492664e-01 1.50297478e-01
-1.37709543e-01 7.31656373e-01 7.29846835e-01 -8.38978529e-01
8.02458525e-01 -2.26530409e+00 4.83429104e-01 6.14160061e-01
1.09731331e-01 -1.30450979e-01 -2.73348629e-01 8.48748505e-01
-3.42120826e-01 -4.13938642e-01 -2.78463244e-01 -1.14740409e-01
-9.98831540e-02 4.64427531e-01 -6.59525812e-01 9.57655907e-01
-2.66446233e-01 4.60611522e-01 -6.18708432e-01 1.78741112e-01
3.17966878e-01 5.44584215e-01 -5.69936097e-01 9.92444381e-02
-9.24037099e-02 1.09206736e+00 -5.60014725e-01 1.84555650e-02
3.62766057e-01 -2.23415047e-01 2.67847151e-01 -4.48539734e-01
7.54477084e-02 2.51628280e-01 -1.81407785e+00 1.85446012e+00
-6.38363004e-01 6.11566365e-01 6.64292276e-01 -1.81724584e+00
8.50989401e-01 5.68657756e-01 8.51458788e-01 -6.53130531e-01
1.89647213e-01 3.34224939e-01 2.38905251e-01 -6.54615104e-01
-6.57700375e-02 -7.00882137e-01 1.81405693e-01 5.72380900e-01
3.05953979e-01 -8.19426849e-02 -1.20629303e-01 1.17831603e-01
1.29898047e+00 -2.65314311e-01 3.02342713e-01 -3.64523053e-01
2.34831944e-01 -1.92564592e-01 6.95862412e-01 9.99178588e-01
3.19632441e-01 6.13600194e-01 2.47825250e-01 -3.69685739e-01
-1.03034091e+00 -7.57646322e-01 -3.62226397e-01 2.20368370e-01
-7.46843815e-02 -1.65408939e-01 -3.51890564e-01 -2.35656481e-02
-3.28744620e-01 5.79101741e-01 -5.53550541e-01 -5.01228034e-01
-6.08437061e-01 -5.96078873e-01 1.63887348e-02 -3.13382074e-02
3.36921006e-01 -4.71326411e-01 -4.60384816e-01 3.11112285e-01
-3.41130584e-01 -1.32856476e+00 -2.39401057e-01 4.40085083e-01
-9.01960194e-01 -1.27211010e+00 -2.84882575e-01 -4.26006764e-01
7.72901952e-01 2.69861817e-01 8.67092550e-01 -1.23419255e-01
-3.10798407e-01 8.46999049e-01 -1.40257239e-01 -4.09796983e-01
-4.29298818e-01 -5.05156517e-01 1.72509447e-01 5.02269864e-01
-6.91306517e-02 -1.01477897e+00 -1.65215895e-01 1.75198123e-01
-8.47145975e-01 -1.42999351e-01 7.27244258e-01 9.95540202e-01
4.17558551e-01 4.95817780e-01 3.35530728e-01 -7.72001326e-01
4.35081631e-01 -2.42148891e-01 -5.96060932e-01 8.30162168e-02
-2.64682204e-01 5.57876170e-01 7.13607490e-01 -5.87950408e-01
-7.87558734e-01 5.95284700e-01 -1.31938830e-01 -3.63391787e-01
1.64037794e-01 9.18005824e-01 -3.06174159e-01 -2.95203745e-01
8.95105064e-01 3.85691017e-01 2.78232813e-01 -7.24539936e-01
1.77056953e-01 1.70086682e-01 4.16904181e-01 -5.21745741e-01
8.60229611e-01 6.39966488e-01 5.65262735e-01 -1.24029374e+00
-1.10974467e+00 -4.93873656e-01 -8.38960230e-01 1.21720634e-01
5.55381715e-01 -9.30042326e-01 -4.32085007e-01 7.76974559e-02
-8.83504391e-01 -2.22997993e-01 -6.91026568e-01 9.58007693e-01
-6.41946197e-01 3.64580661e-01 -2.17423692e-01 -7.74937809e-01
1.81391463e-01 -8.68130028e-01 6.81778431e-01 -4.76444334e-01
-2.03354746e-01 -1.15618169e+00 -3.40272719e-03 1.27468392e-01
4.30290908e-01 7.59817213e-02 9.47809398e-01 -2.99809247e-01
-6.96749330e-01 -6.15010262e-01 3.32681298e-01 6.20695770e-01
1.08494014e-01 -8.53349984e-01 -7.15185285e-01 -5.58349907e-01
9.39278066e-01 -8.91192481e-02 7.69663751e-01 4.93535668e-01
6.99339330e-01 -4.18094665e-01 -3.43701281e-02 7.23682761e-01
1.49345076e+00 -1.81614179e-02 3.12688798e-01 -2.41942599e-01
5.80001354e-01 5.68031013e-01 -3.92829180e-02 4.40293252e-01
-9.12791789e-02 6.80311143e-01 3.75363052e-01 1.65922716e-01
6.01821765e-02 -2.70413250e-01 1.69658557e-01 1.00938261e+00
1.23630390e-01 1.68527797e-01 -6.97095692e-01 2.64666975e-01
-1.67119563e+00 -9.61943209e-01 -2.08112458e-03 2.52333593e+00
4.03003514e-01 -2.18097568e-02 -1.99287146e-01 7.50345707e-01
3.25993776e-01 -3.63923833e-02 -6.99154258e-01 1.20741345e-01
-2.49840051e-01 2.98303425e-01 5.36017299e-01 8.31552446e-01
-7.22287834e-01 1.38064370e-01 6.65306950e+00 3.42019022e-01
-1.39565027e+00 8.58811066e-02 -3.20293158e-01 2.57378221e-01
-3.59132230e-01 5.17038584e-01 -3.66052449e-01 8.44642967e-02
7.16204584e-01 -1.00114278e-01 1.00445402e+00 2.93159187e-01
2.10257992e-01 -5.96183948e-02 -1.45333588e+00 1.07665956e+00
1.88920096e-01 -1.06037998e+00 -2.30893910e-01 4.14670438e-01
5.30883372e-01 -1.44680440e-01 -1.71354756e-01 6.13629483e-02
-1.02933660e-01 -8.49694490e-01 6.09947801e-01 6.62581146e-01
3.97111624e-01 -2.32699424e-01 4.75609750e-01 1.02431905e+00
-6.29024684e-01 -2.28985190e-01 -4.90048736e-01 -6.68646753e-01
1.94650307e-01 1.04259706e+00 -7.44778454e-01 3.40152144e-01
6.00682497e-02 7.64752984e-01 -9.82376188e-02 7.86131322e-01
-1.81704506e-01 7.51580656e-01 -5.61146080e-01 4.60157871e-01
-1.50122344e-01 -6.29840136e-01 7.64314950e-01 5.11115968e-01
7.00923860e-01 6.09688699e-01 3.00982356e-01 8.11364532e-01
3.02475512e-01 -2.63473779e-01 -7.31994331e-01 -3.28961074e-01
2.88098514e-01 9.24664199e-01 -4.40206885e-01 1.53268993e-01
-5.67977309e-01 6.59305811e-01 -1.14732929e-01 6.54131234e-01
-1.75326064e-01 -1.86194628e-02 3.31207097e-01 1.38046160e-01
2.96189517e-01 -7.57190585e-01 -2.74775565e-01 -1.39135325e+00
2.28390038e-01 -1.05065656e+00 2.72804424e-02 -5.14451802e-01
-9.89289701e-01 4.75935675e-02 -1.24681532e-01 -1.29973638e+00
-5.22921324e-01 -5.64005971e-01 -3.37878525e-01 8.36656153e-01
-9.29727256e-01 -8.09852898e-01 -2.86328676e-03 7.77927995e-01
-8.59655067e-02 -9.87019241e-02 1.09362757e+00 1.88448220e-01
3.94462310e-02 -4.68575768e-02 3.79186898e-01 9.42225382e-02
3.31232488e-01 -8.94576967e-01 -3.42908114e-01 7.43638337e-01
6.74196064e-01 8.41545165e-01 1.20231390e+00 -1.57061219e-01
-2.17319775e+00 -4.72987950e-01 5.42587698e-01 -3.01070064e-01
6.84611201e-01 -5.48314214e-01 -8.15489054e-01 1.01712191e+00
-2.54063785e-01 2.97380269e-01 5.97839296e-01 2.79997498e-01
-2.17898637e-01 -1.44803673e-01 -8.64161611e-01 1.12335943e-01
7.95543492e-01 -7.79084206e-01 -5.04198134e-01 3.90573651e-01
2.48708129e-01 -3.62823129e-01 -6.05371058e-01 3.92131388e-01
5.48025429e-01 -9.73223567e-01 1.12360668e+00 -5.04385173e-01
-8.18614364e-02 -2.47025222e-01 -7.05605090e-01 -1.53996634e+00
-3.06336761e-01 -5.13621867e-01 -1.95260212e-01 5.65096200e-01
2.81240612e-01 -6.79880619e-01 6.77733839e-01 5.85321248e-01
-1.25925243e-01 -5.00582814e-01 -1.10983407e+00 -8.27943087e-01
-3.20501290e-02 -5.09061873e-01 3.02318096e-01 1.26996148e+00
-1.20000914e-01 7.42129505e-01 -7.00880706e-01 6.64955199e-01
7.81894505e-01 2.03985676e-01 6.05067611e-01 -1.35868716e+00
-9.32949722e-01 2.06918836e-01 -3.04782242e-01 -1.40960526e+00
2.35272080e-01 -8.63923073e-01 -3.36491056e-02 -1.31028843e+00
-1.78798884e-01 -5.75316608e-01 -1.08610138e-01 6.28848970e-02
5.20859420e-01 8.60038097e-04 2.78071523e-01 3.17145616e-01
-4.16464023e-02 4.12491381e-01 1.03391814e+00 -2.05000907e-01
-5.05558141e-02 2.93587089e-01 -6.60663664e-01 7.99309552e-01
4.67627376e-01 -3.90410334e-01 -6.39121175e-01 -6.78936541e-01
6.67011976e-01 6.03407979e-01 5.07781148e-01 -1.22760725e+00
4.21208411e-01 -6.51073456e-03 2.87543595e-01 7.03771636e-02
8.44568908e-01 -1.13909721e+00 5.28953433e-01 4.23431665e-01
-3.53502125e-01 -3.66818637e-01 -2.05391273e-01 7.41486549e-01
-1.46285459e-01 -7.06601322e-01 5.87867796e-01 -2.29526520e-01
-4.96539354e-01 2.43547764e-02 -2.62877584e-01 2.90676374e-02
3.79639328e-01 -1.34726390e-01 2.81785011e-01 -9.97348130e-01
-1.18516302e+00 -2.09895596e-01 2.14340448e-01 8.94732922e-02
5.12661636e-01 -1.18755662e+00 -6.27616882e-01 8.12077820e-01
-1.05610110e-01 -2.50975490e-01 1.24040432e-01 7.95033216e-01
-7.26191550e-02 5.25482059e-01 -1.09201614e-02 -5.64684331e-01
-8.11695695e-01 6.04899645e-01 4.07129377e-01 1.76077709e-01
-6.84924304e-01 4.83284682e-01 1.85019389e-01 -5.67864656e-01
9.51783285e-02 -1.13673933e-01 3.32858503e-01 -2.06015825e-01
4.07831758e-01 1.64833143e-01 1.79751128e-01 -8.85283947e-01
1.44090885e-02 7.06493616e-01 4.10204560e-01 -3.92154247e-01
1.25485122e+00 -2.42851034e-01 -2.63788849e-01 7.21831977e-01
1.41218674e+00 2.95969427e-01 -9.93673742e-01 -7.02471912e-01
-1.60923034e-01 -4.05960083e-01 4.17278469e-01 -4.73658144e-01
-9.90852833e-01 9.36361492e-01 4.31160182e-01 2.83391863e-01
1.16296196e+00 -9.85819176e-02 4.09262210e-01 9.57168877e-01
6.21567070e-01 -8.31172764e-01 -4.84104827e-02 5.26552677e-01
1.10741270e+00 -1.12306583e+00 1.21218279e-01 -2.94827372e-01
-3.51756886e-02 1.11856246e+00 -9.00863409e-02 -3.68334323e-01
1.02853763e+00 3.96624692e-02 6.46425635e-02 -1.77228838e-01
-2.58878887e-01 -7.09235445e-02 3.82252425e-01 6.13684833e-01
8.98971185e-02 -8.99806321e-02 -5.59001043e-02 2.49242216e-01
-4.52962577e-01 -6.90147132e-02 7.37444162e-01 7.76255846e-01
-4.95768636e-01 -1.23514211e+00 -6.33800447e-01 4.70463663e-01
-1.32919967e-01 2.13130474e-01 -2.68396765e-01 6.00127637e-01
5.42580858e-02 1.01993072e+00 -4.28425819e-01 -1.05230793e-01
3.01921338e-01 -4.56486605e-02 8.36823821e-01 -7.44638383e-01
5.64526856e-01 1.36482060e-01 7.90167414e-03 -3.17782372e-01
-3.22847933e-01 -7.51652062e-01 -1.07109225e+00 -5.24990410e-02
-5.15910685e-01 3.59046072e-01 8.55838895e-01 1.35367107e+00
1.20956935e-01 2.53310382e-01 5.13561606e-01 -8.08987319e-01
-1.07108033e+00 -7.88013220e-01 -1.06155229e+00 2.56310225e-01
9.49359775e-01 -5.92355132e-01 -5.40373266e-01 8.78715422e-03] | [7.0635480880737305, 4.446740627288818] |
07063ac9-e330-48d4-a0ad-95a6d040f74e | 3dfill-reference-guided-image-inpainting-by | 2211.04831 | null | https://arxiv.org/abs/2211.04831v1 | https://arxiv.org/pdf/2211.04831v1.pdf | 3DFill:Reference-guided Image Inpainting by Self-supervised 3D Image Alignment | Most existing image inpainting algorithms are based on a single view, struggling with large holes or the holes containing complicated scenes. Some reference-guided algorithms fill the hole by referring to another viewpoint image and use 2D image alignment. Due to the camera imaging process, simple 2D transformation is difficult to achieve a satisfactory result. In this paper, we propose 3DFill, a simple and efficient method for reference-guided image inpainting. Given a target image with arbitrary hole regions and a reference image from another viewpoint, the 3DFill first aligns the two images by a two-stage method: 3D projection + 2D transformation, which has better results than 2D image alignment. The 3D projection is an overall alignment between images and the 2D transformation is a local alignment focused on the hole region. The entire process of image alignment is self-supervised. We then fill the hole in the target image with the contents of the aligned image. Finally, we use a conditional generation network to refine the filled image to obtain the inpainting result. 3DFill achieves state-of-the-art performance on image inpainting across a variety of wide view shifts and has a faster inference speed than other inpainting models. | ['Long Zeng', 'Xinyu Zhang', 'Hailong Ma', 'Xinyuan Zhao', 'Liang Zhao'] | 2022-11-09 | null | null | null | null | ['image-inpainting'] | ['computer-vision'] | [ 3.99798214e-01 -2.70618312e-02 -2.50543088e-01 -2.20153153e-01
-7.38623738e-01 -2.01355278e-01 4.75871533e-01 -5.35291910e-01
-1.17266357e-01 5.35150290e-01 2.08684683e-01 -2.10033972e-02
3.47065806e-01 -6.75033271e-01 -9.39807057e-01 -6.10245347e-01
7.71765769e-01 3.64298016e-01 3.86054456e-01 -7.75780901e-02
4.17855889e-01 4.20744300e-01 -1.13969290e+00 2.79960424e-01
6.77243590e-01 8.18935812e-01 7.56503642e-01 6.18630171e-01
-3.19567293e-01 6.85563743e-01 -4.15530443e-01 -3.73436511e-01
6.85540259e-01 -7.75920808e-01 -7.80224562e-01 7.70511866e-01
6.68727219e-01 -7.24357426e-01 -2.72284985e-01 1.31827986e+00
3.23683858e-01 -2.99565140e-02 1.96203411e-01 -1.15905774e+00
-6.01024330e-01 2.50168353e-01 -1.13058746e+00 -5.53531237e-02
6.73828781e-01 -1.85044743e-02 4.33332354e-01 -1.16056883e+00
1.02065790e+00 1.22980940e+00 3.84669423e-01 4.99988109e-01
-1.10929859e+00 -4.00921792e-01 2.21798196e-01 -2.23678183e-02
-1.32412052e+00 -2.60619432e-01 1.18709791e+00 -3.00062895e-01
6.86262190e-01 1.33969069e-01 7.26674676e-01 5.67852974e-01
4.22319680e-01 6.12168789e-01 1.14273846e+00 -6.90301836e-01
-9.80850309e-02 -8.93719271e-02 -4.43704545e-01 6.97606266e-01
-2.58489847e-01 1.67835295e-01 -5.31569958e-01 1.68736726e-01
1.27797663e+00 4.98051792e-01 -4.98086751e-01 -3.83270413e-01
-1.32616270e+00 6.77363992e-01 4.80500787e-01 1.05729833e-01
-3.59001428e-01 -7.31857792e-02 1.00122876e-02 3.43979239e-01
6.14219844e-01 2.75751859e-01 -1.88767985e-01 2.91035205e-01
-1.28302574e+00 2.06610292e-01 4.10167187e-01 1.05484104e+00
9.75300729e-01 -1.15112290e-01 1.06177397e-01 9.21142578e-01
3.10630888e-01 4.02464688e-01 3.62972885e-01 -1.39061058e+00
7.08989799e-01 4.09490764e-01 1.80730745e-01 -1.21537030e+00
1.77263394e-02 -4.26879227e-02 -9.55765367e-01 5.02400637e-01
1.32515252e-01 1.86572939e-01 -8.19240212e-01 1.48055005e+00
4.45023119e-01 -3.06168851e-02 -2.99244449e-02 9.08115625e-01
7.24965394e-01 9.83806729e-01 -6.45040572e-01 -4.85375375e-01
1.21135116e+00 -1.74970496e+00 -8.75788510e-01 -4.82554644e-01
1.71350911e-01 -1.30041659e+00 8.55508804e-01 5.21810889e-01
-1.53378654e+00 -8.33067656e-01 -9.72481668e-01 -5.89760423e-01
1.82430059e-01 -8.31665099e-03 1.01335101e-01 -4.23743352e-02
-1.10002732e+00 3.05526793e-01 -5.13218939e-01 -3.12256515e-01
1.75216883e-01 -6.08400665e-02 -7.45285511e-01 -4.62908626e-01
-7.31457829e-01 1.01020706e+00 3.28580052e-01 -2.92662643e-02
-6.76458895e-01 -4.66302395e-01 -1.05805254e+00 -2.80301720e-01
3.26774925e-01 -9.46592987e-01 1.06164920e+00 -8.68262112e-01
-1.51388526e+00 1.20473909e+00 -4.88322943e-01 -2.90468484e-01
7.42170095e-01 -4.85463560e-01 -1.38362572e-01 2.90929109e-01
5.65617263e-01 1.11805069e+00 1.14808822e+00 -1.41612279e+00
-5.23680985e-01 -2.78975248e-01 -1.27893105e-01 7.06895411e-01
3.38444382e-01 -4.34862413e-02 -9.97257650e-01 -7.63611555e-01
7.36991823e-01 -6.28704131e-01 -3.49041969e-01 2.96001583e-01
-4.51648802e-01 2.69806623e-01 1.21649384e+00 -9.45327759e-01
1.08479953e+00 -2.21025324e+00 2.95799345e-01 -2.70038337e-01
1.32518262e-01 1.54468957e-02 -1.57532126e-01 5.24571359e-01
-2.62316138e-01 -2.30474994e-01 -3.20127904e-01 -6.83260918e-01
-6.35007620e-01 2.47681588e-01 -4.82363552e-01 2.87319750e-01
5.66586386e-03 8.55681181e-01 -8.68787766e-01 -6.81792855e-01
6.54835522e-01 2.90419072e-01 -6.16629660e-01 6.46980882e-01
-1.03083901e-01 6.92565322e-01 -1.08755216e-01 4.87192482e-01
1.01748276e+00 -7.59659857e-02 -1.83970302e-01 -5.17197013e-01
-2.78406233e-01 5.53621501e-02 -1.00658894e+00 2.34168172e+00
-4.47943658e-01 5.28811693e-01 3.95670459e-02 -7.19119370e-01
1.21200132e+00 5.32718562e-02 4.95222360e-01 -5.34821868e-01
6.75451127e-04 1.58200532e-01 -3.24725658e-01 -6.27436757e-01
4.30454880e-01 -2.08199248e-01 3.02208513e-01 6.75158024e-01
-1.10723644e-01 -9.29353118e-01 8.58949050e-02 1.89589739e-01
5.51166058e-01 6.37447476e-01 4.86435473e-01 1.68066159e-01
5.11832416e-01 -2.04481319e-01 4.11994994e-01 4.53602701e-01
4.44904529e-02 1.47657359e+00 1.91394016e-01 -7.51435995e-01
-1.28325701e+00 -1.01264429e+00 1.01025015e-01 2.44283259e-01
6.48560524e-01 -4.07179058e-01 -8.59943807e-01 -5.10026813e-01
-5.13347208e-01 5.30294955e-01 -4.87350255e-01 -2.84764823e-02
-6.42855704e-01 -2.82187879e-01 -1.54590309e-01 2.95914888e-01
1.17370129e+00 -1.12577975e+00 -5.26337981e-01 1.82783738e-01
-7.50849187e-01 -1.22009706e+00 -1.16506743e+00 -1.24516211e-01
-1.14140713e+00 -1.00523794e+00 -8.13024580e-01 -1.16061270e+00
1.13862228e+00 8.63407433e-01 1.14673042e+00 1.26120344e-01
1.09250560e-01 -2.56714467e-02 -1.90643013e-01 -2.88453624e-02
-4.82098699e-01 -4.04663712e-01 -2.07449034e-01 1.17339194e-02
-5.71148880e-02 -7.12151229e-01 -6.66416585e-01 4.00353789e-01
-1.06369185e+00 7.05679893e-01 6.57151222e-01 1.05827188e+00
9.09321666e-01 1.82474956e-01 -1.42693698e-01 -7.16609120e-01
2.62960345e-01 -5.78619726e-02 -4.42756772e-01 1.57657281e-01
-3.21571529e-01 -1.27205342e-01 5.07399261e-01 -3.54540497e-01
-1.07150602e+00 2.12132603e-01 -2.46418223e-01 -8.42515945e-01
-5.62022552e-02 1.80988938e-01 -3.08821619e-01 1.39425382e-01
4.21428919e-01 5.49607694e-01 4.50147808e-01 -3.95622671e-01
4.19701517e-01 4.43567038e-01 7.86812603e-01 -2.66283125e-01
8.65045667e-01 5.77898383e-01 -2.16627121e-01 -4.90433693e-01
-1.19471002e+00 -2.88825452e-01 -9.82143104e-01 -1.12157367e-01
9.97256875e-01 -9.29954946e-01 -3.92538048e-02 6.51056528e-01
-1.48108971e+00 -2.89482892e-01 -4.35331672e-01 4.75584477e-01
-7.13617504e-01 7.70269573e-01 -6.38636529e-01 -2.18079954e-01
-4.17205393e-01 -1.38691056e+00 1.40095651e+00 2.16104612e-01
-5.06406948e-02 -8.93810391e-01 1.08977281e-01 4.76448506e-01
5.56665771e-02 1.18299790e-01 5.85005760e-01 2.16235220e-01
-7.66011178e-01 2.36259103e-02 -2.81386614e-01 6.05985582e-01
4.31661725e-01 -1.08690068e-01 -5.27326226e-01 -1.72197804e-01
6.22284651e-01 -2.07397595e-01 5.72859943e-01 5.64148307e-01
1.02078092e+00 -2.40382791e-01 -2.08239943e-01 9.12981689e-01
1.53028023e+00 4.04631793e-01 1.08180857e+00 4.93138939e-01
7.43211865e-01 4.67068374e-01 8.72876167e-01 8.42344612e-02
2.53669620e-01 8.92745972e-01 5.01324236e-01 -4.30560082e-01
-3.56242508e-01 -7.34456480e-01 2.85067827e-01 9.61361289e-01
1.55160695e-01 2.81686522e-02 -3.51651222e-01 3.75136733e-01
-1.79841316e+00 -1.07425940e+00 -1.12977609e-01 2.16679907e+00
9.82715428e-01 3.48629244e-02 -2.94588059e-01 -3.98689555e-03
7.91502416e-01 4.13540691e-01 -4.30196583e-01 -2.28380814e-01
-1.15833625e-01 -1.54615238e-01 1.99546814e-01 9.46507156e-01
-8.29982579e-01 1.03007007e+00 6.48082161e+00 7.08076477e-01
-1.07030833e+00 1.63475424e-01 8.20769310e-01 1.67063370e-01
-2.65201598e-01 3.20776463e-01 -7.40261436e-01 5.41574001e-01
-4.85874489e-02 2.36508295e-01 4.93905306e-01 6.66638911e-01
3.01069975e-01 -5.78440487e-01 -1.01784062e+00 1.40771973e+00
5.59382081e-01 -1.45290446e+00 3.59293193e-01 -6.35260046e-02
1.10705674e+00 -2.16962010e-01 -1.00431897e-01 -3.03088099e-01
-1.18647806e-01 -8.01952004e-01 7.70185828e-01 5.36888301e-01
8.28566849e-01 -4.98115063e-01 6.10738635e-01 6.04809046e-01
-9.04429376e-01 2.54546762e-01 -5.35702348e-01 -1.31415218e-01
6.60581350e-01 6.86518252e-01 -3.86386186e-01 6.26487672e-01
7.81162679e-01 9.04856265e-01 -4.16559547e-01 8.94861877e-01
-4.41673487e-01 -1.32834658e-01 -1.97996363e-01 7.93549478e-01
1.85532197e-01 -7.76105881e-01 4.43985403e-01 5.90565383e-01
6.36897981e-01 9.58648995e-02 2.69573718e-01 7.82542765e-01
-2.87890173e-02 -4.03074175e-02 -1.05084813e+00 5.60302138e-01
3.67780477e-01 1.23164582e+00 -7.27067471e-01 -4.94449586e-01
-5.53161085e-01 1.51616073e+00 1.52521014e-01 2.01254025e-01
-8.29563737e-01 -2.10445374e-01 5.39701581e-02 1.37963235e-01
3.48558813e-01 -1.64948210e-01 -1.76114276e-01 -1.28401601e+00
1.72249809e-01 -8.77481401e-01 7.68184513e-02 -1.51576328e+00
-9.86593306e-01 7.23102152e-01 2.15350352e-02 -1.72372973e+00
-1.22692451e-01 -2.24238664e-01 -7.54536569e-01 9.45744574e-01
-1.40382123e+00 -1.05286610e+00 -5.64138710e-01 6.92783058e-01
1.01193571e+00 -5.92269339e-02 5.33036947e-01 1.01237111e-01
-1.80493802e-01 1.31569907e-01 -1.40620351e-01 -3.27849668e-03
9.67357397e-01 -9.26677287e-01 3.84303123e-01 1.17796278e+00
2.11289525e-01 4.44126308e-01 5.33157110e-01 -7.50958562e-01
-1.07932866e+00 -1.04846406e+00 9.30213153e-01 -4.50385571e-01
8.80311653e-02 -6.24866895e-02 -7.76569784e-01 7.30911732e-01
6.98129952e-01 2.34778449e-01 1.44274622e-01 -7.80645788e-01
-2.16599107e-01 -1.76662624e-01 -1.11264408e+00 7.05082953e-01
1.13494241e+00 -4.57659990e-01 -7.11321354e-01 4.02625650e-01
8.07782292e-01 -9.18267667e-01 -6.71709418e-01 1.43560931e-01
3.30694914e-01 -1.34206629e+00 9.99781430e-01 -1.31045645e-02
9.13734734e-01 -7.73500204e-01 -1.47998720e-01 -1.27044988e+00
-1.39553994e-01 -7.96971142e-01 1.47653356e-01 1.04855645e+00
-1.00261278e-01 -3.94627362e-01 6.61623240e-01 2.23240495e-01
-1.80784225e-01 -8.08157504e-01 -4.87086743e-01 -3.90727431e-01
-2.64301330e-01 -1.77155569e-01 4.15090501e-01 7.38070488e-01
-4.35966909e-01 4.29893315e-01 -6.87115192e-01 8.25910717e-02
6.63183093e-01 5.69976330e-01 1.18474209e+00 -6.82797730e-01
-2.88773119e-01 -1.19520366e-01 -1.22884572e-01 -1.65082729e+00
-1.75244585e-01 -5.67555130e-01 1.37218207e-01 -1.56880152e+00
4.12924111e-01 -1.06084719e-01 4.31269914e-01 2.72209883e-01
-2.46064812e-01 5.75180709e-01 3.02390575e-01 5.43147624e-01
-3.51215571e-01 5.46590149e-01 1.89895558e+00 -1.34969756e-01
-2.82946736e-01 -7.44540095e-02 -6.57127321e-01 8.65737677e-01
6.03769004e-01 -4.01100397e-01 -6.49402142e-01 -7.22427070e-01
-7.84914270e-02 4.86267865e-01 2.86264300e-01 -7.86064506e-01
3.08528036e-01 -2.14129135e-01 6.16644025e-01 -1.14691150e+00
4.86588061e-01 -8.62948537e-01 5.11981547e-01 3.91767800e-01
-1.39122948e-01 4.96911108e-01 -2.31610909e-01 3.53048056e-01
-6.68182909e-01 -5.02314150e-01 9.26967025e-01 -5.59772909e-01
-5.28638422e-01 4.93518472e-01 3.99326868e-02 -8.67818110e-03
1.06602561e+00 -5.67512333e-01 -1.55987274e-02 -6.64088845e-01
-6.78069711e-01 4.86327223e-02 9.40416873e-01 3.89721930e-01
1.08412170e+00 -1.50866604e+00 -5.78334212e-01 6.59611166e-01
-2.90991843e-01 6.55912042e-01 3.37580085e-01 8.34360301e-01
-9.52763021e-01 4.34107892e-02 -4.31523472e-01 -9.28270161e-01
-1.27810168e+00 6.39761984e-01 4.29832965e-01 -3.86864841e-01
-9.22087550e-01 7.05176413e-01 7.14763582e-01 -5.09830236e-01
2.03367919e-02 -6.37083426e-02 7.78700113e-02 -3.17269385e-01
6.79356515e-01 -7.93878436e-02 -1.14637814e-01 -6.44608855e-01
2.16615442e-02 1.12706125e+00 -3.31028789e-01 -3.33736926e-01
1.17599058e+00 -5.16662061e-01 -4.94539291e-01 2.09634393e-01
1.30371845e+00 1.99077830e-01 -1.67529011e+00 -3.27845961e-01
-7.44176924e-01 -9.83124495e-01 -1.90970749e-01 -3.88902873e-01
-1.34283853e+00 8.45606387e-01 2.66409695e-01 -7.39605799e-02
1.41406298e+00 -1.14719182e-01 9.44259167e-01 -5.16477562e-02
4.83901352e-01 -8.64268780e-01 4.93946105e-01 3.88355345e-01
1.21114647e+00 -1.17378390e+00 3.80116671e-01 -6.88306332e-01
-5.52302539e-01 1.23527324e+00 8.03809881e-01 -1.93722770e-01
5.62752306e-01 1.13437817e-01 2.91272134e-01 -6.07667603e-02
-5.05404234e-01 2.21108049e-01 4.71706167e-02 7.42056549e-01
1.20303161e-01 -4.69390363e-01 -3.27621371e-01 -3.41359563e-02
-2.02326700e-01 -5.46508329e-03 6.00498259e-01 7.15852618e-01
-2.29526013e-01 -1.31585002e+00 -6.40716970e-01 -1.05118677e-02
-1.64571136e-01 -1.15184158e-01 -2.43055463e-01 8.15383792e-01
2.01292917e-01 7.00180590e-01 5.93358129e-02 -2.32185990e-01
1.89826190e-01 -2.41253331e-01 7.77289808e-01 -7.43778527e-01
1.61118433e-02 4.88507003e-01 -4.30813849e-01 -7.08127499e-01
-7.48605907e-01 -5.27398169e-01 -9.96316314e-01 -7.41215423e-02
-3.03593546e-01 -4.55994122e-02 3.29013616e-01 7.96860099e-01
3.99853766e-01 1.12803541e-01 8.82403076e-01 -1.22486055e+00
-1.68279871e-01 -8.02610278e-01 -4.68972355e-01 4.75963801e-01
2.27668196e-01 -4.44193751e-01 -3.05204362e-01 4.29673791e-01] | [10.735343933105469, -1.4451358318328857] |
0c23448f-21bc-4e82-9ffc-f00d8c77a1ec | decoding-visemes-improving-machine-lipreading-1 | 1710.01169 | null | http://arxiv.org/abs/1710.01169v1 | http://arxiv.org/pdf/1710.01169v1.pdf | Decoding visemes: improving machine lipreading | To undertake machine lip-reading, we try to recognise speech from a visual
signal. Current work often uses viseme classification supported by language
models with varying degrees of success. A few recent works suggest phoneme
classification, in the right circumstances, can outperform viseme
classification. In this work we present a novel two-pass method of training
phoneme classifiers which uses previously trained visemes in the first pass.
With our new training algorithm, we show classification performance which
significantly improves on previous lip-reading results. | ['Richard Harvey', 'Helen L. Bear'] | 2017-10-03 | null | null | null | null | ['lipreading'] | ['computer-vision'] | [ 6.14897490e-01 8.89739245e-02 -4.03464675e-01 -1.48622721e-01
-8.09453547e-01 -2.27710575e-01 9.25329864e-01 -2.14341611e-01
-4.14787650e-01 9.16619539e-01 5.27099371e-01 -5.54753840e-01
4.14161980e-01 -3.23888987e-01 -3.42828780e-01 -5.32450199e-01
4.57006812e-01 3.67315002e-02 2.65105575e-01 -2.31198043e-01
7.97171712e-01 3.51532310e-01 -2.29228258e+00 6.52354360e-01
5.77425539e-01 9.69861507e-01 3.37117016e-01 1.27421546e+00
-5.58267415e-01 5.84316432e-01 -4.17675763e-01 -3.38375121e-01
2.51652468e-02 -7.54228175e-01 -9.00370121e-01 1.28477156e-01
5.01597643e-01 6.37629777e-02 1.30174503e-01 8.23579729e-01
8.57667208e-01 -5.41770346e-02 1.06476605e+00 -1.14758050e+00
-5.75269520e-01 2.69809902e-01 -2.87234992e-01 1.26689255e-01
8.57446492e-01 1.19848140e-02 5.83991706e-01 -1.01892734e+00
3.35346788e-01 1.12144089e+00 7.27111280e-01 1.11079729e+00
-1.11174560e+00 -4.35911089e-01 -3.30871135e-01 3.83303493e-01
-1.25336254e+00 -1.27686667e+00 8.74382079e-01 -2.90736049e-01
1.29581606e+00 5.61052442e-01 7.55296588e-01 9.77883816e-01
-3.24301398e-03 1.00347495e+00 1.71403968e+00 -1.16972601e+00
-7.82486871e-02 6.27742290e-01 -2.57555723e-01 6.36567473e-01
-3.16772103e-01 2.78585017e-01 -7.16534019e-01 2.76618302e-01
2.68602580e-01 -7.67176509e-01 -4.34820533e-01 -1.89495489e-01
-1.08361530e+00 6.49848580e-01 -2.12514371e-01 4.13550258e-01
-1.59784988e-01 -1.85828984e-01 4.99214321e-01 2.98330337e-01
5.73783100e-01 1.68151017e-02 -4.23913717e-01 -5.41558206e-01
-1.27662611e+00 -4.82246757e-01 1.02648914e+00 6.66372359e-01
4.10697192e-01 7.15473220e-02 6.20741062e-02 1.32306874e+00
6.15170240e-01 5.31856358e-01 9.79780316e-01 -2.63060778e-01
1.14340663e-01 -5.44127226e-02 -3.69927347e-01 -3.94101560e-01
-1.46613628e-01 4.50996369e-01 -3.43059659e-01 6.41182303e-01
5.53774297e-01 1.80262595e-01 -1.19648540e+00 1.38763571e+00
7.86135867e-02 1.28195539e-01 4.81829941e-01 2.77246535e-01
1.14123487e+00 6.69244111e-01 2.74141073e-01 -5.92657447e-01
1.02577388e+00 -8.85641992e-01 -1.02175188e+00 1.45230010e-01
5.16816974e-01 -1.29165411e+00 1.10867369e+00 7.77593970e-01
-1.21099055e+00 -5.75979888e-01 -8.42982471e-01 -1.46662802e-01
-8.24346423e-01 4.10037071e-01 5.07259429e-01 1.29388511e+00
-1.57688272e+00 1.98485926e-01 -6.24588244e-02 -6.67492449e-01
1.56688228e-01 6.85582280e-01 -4.64876741e-01 3.66076618e-01
-8.01111877e-01 1.11428082e+00 4.75054651e-01 -6.80785477e-02
-2.24553928e-01 -5.69688447e-04 -1.05376244e+00 -3.61411750e-01
1.34282351e-01 -1.70031101e-01 1.28994334e+00 -1.05973279e+00
-2.25327373e+00 1.34609509e+00 -6.67986810e-01 -2.14792117e-01
5.07832944e-01 2.24057227e-01 -7.14260578e-01 2.09594727e-01
-5.76461911e-01 1.17508280e+00 1.36623311e+00 -1.48815835e+00
-6.93904102e-01 1.90474033e-01 -4.32365417e-01 2.55981803e-01
8.36210772e-02 1.74897388e-01 -2.03412548e-01 -5.60919106e-01
-1.04041353e-01 -4.98896986e-01 7.90108740e-01 -4.74663563e-02
-2.66162664e-01 -7.71621406e-01 8.13297689e-01 -8.18012297e-01
1.07469475e+00 -2.11468244e+00 -1.41369760e-01 8.08157995e-02
-2.57546324e-02 6.73968136e-01 -9.57914367e-02 1.99229911e-01
-3.23510766e-01 1.80069938e-01 -1.21849962e-01 -7.87354171e-01
1.39468834e-01 1.66052535e-01 -1.35769710e-01 3.92819077e-01
8.22733250e-03 1.06131029e+00 -5.03160298e-01 -1.01706195e+00
7.08444595e-01 6.14139736e-01 -2.26709306e-01 6.97948039e-02
8.37103799e-02 -2.09214926e-01 4.35557961e-01 6.80920899e-01
6.70177579e-01 3.88849944e-01 -7.92439356e-02 -9.47653688e-03
-2.93675721e-01 5.79564810e-01 -9.15690422e-01 1.42452025e+00
-7.18500435e-01 1.48502827e+00 1.76491633e-01 -1.01398718e+00
8.88946235e-01 6.35232449e-01 1.38191491e-01 -6.22732460e-01
2.90971905e-01 3.66554260e-01 -1.48003325e-01 -9.91318464e-01
3.65378678e-01 -5.78374743e-01 6.21432126e-01 2.54617929e-01
1.73388168e-01 -4.53238577e-01 -1.39044315e-01 -3.54386151e-01
9.68167186e-02 1.28211230e-01 6.98257029e-01 -2.30957046e-01
1.15068233e+00 -3.68176669e-01 -2.93149650e-01 4.64326948e-01
-4.49359238e-01 5.42158425e-01 5.77737391e-02 7.74823800e-02
-7.67490208e-01 -9.38813746e-01 -4.46538836e-01 1.24709821e+00
-8.06579441e-02 -2.32225403e-01 -1.02320695e+00 -6.41009569e-01
-3.89403224e-01 7.34873950e-01 -5.43989480e-01 2.23796129e-01
-3.42044264e-01 -2.35481232e-01 6.15008533e-01 2.53162891e-01
2.86638349e-01 -1.72602034e+00 -3.35241288e-01 2.87364312e-02
-2.74885749e-03 -7.83390582e-01 -3.97898972e-01 1.26493156e-01
-5.26784778e-01 -7.80327976e-01 -1.19213986e+00 -1.43483841e+00
3.07641238e-01 6.77771121e-02 8.04660082e-01 3.35582912e-01
-1.84423372e-01 6.67499423e-01 -4.51182991e-01 -7.68268645e-01
-8.80838871e-01 4.74782400e-02 1.52830109e-01 -7.94559568e-02
7.86970317e-01 -2.09426478e-01 5.51043376e-02 -3.02082766e-02
-7.08811462e-01 1.57583222e-01 6.71785057e-01 7.59342670e-01
2.14731187e-01 -1.77648067e-01 6.67030990e-01 -3.15964073e-01
9.25938129e-01 -1.83040395e-01 -1.51356399e-01 4.40342873e-01
-6.55728698e-01 4.94154580e-02 1.54910311e-01 -7.39342093e-01
-8.79613638e-01 1.22760110e-01 -7.32883334e-01 -1.09444894e-02
-8.18871677e-01 1.66741945e-02 -2.18479246e-01 -5.10060310e-01
6.01297259e-01 6.46612167e-01 5.03987670e-01 -6.79656982e-01
3.06757271e-01 1.49837708e+00 6.14794970e-01 1.71152223e-02
2.41896436e-01 8.40262100e-02 -2.77453929e-01 -1.49100912e+00
3.59898247e-02 -6.43583894e-01 -8.42098057e-01 -6.06934428e-01
8.52966726e-01 -4.31710154e-01 -9.00305450e-01 9.32126999e-01
-1.12309277e+00 -3.80034894e-01 -3.67916435e-01 6.85914099e-01
-9.31530476e-01 5.39859831e-01 -6.00668527e-02 -1.32537186e+00
-9.24781486e-02 -1.00919569e+00 1.09811699e+00 1.97209105e-01
-1.32654294e-01 -1.29067481e+00 2.19972685e-01 3.67277861e-01
3.82371694e-01 -4.92372245e-01 5.73967874e-01 -5.81638753e-01
-3.21240500e-02 9.37280804e-02 -1.55048102e-01 5.93715787e-01
4.32115495e-01 4.25007455e-02 -1.39835453e+00 2.32199896e-02
-8.37432519e-02 -4.67740268e-01 1.08243155e+00 5.47001779e-01
1.05684316e+00 -1.83948770e-01 -3.99634749e-01 2.98395872e-01
1.20516169e+00 2.53467500e-01 9.12567198e-01 2.88116857e-02
3.66984248e-01 7.37677395e-01 1.32382333e-01 -2.11871430e-01
2.76041836e-01 6.02814376e-01 -5.46995513e-02 -2.84785181e-01
-7.77915418e-01 -3.91741127e-01 5.06295502e-01 9.35490787e-01
-7.98837692e-02 -6.87672734e-01 -8.75059843e-01 5.76980412e-01
-1.16211712e+00 -1.01493537e+00 1.27497554e-01 1.94421411e+00
9.33562696e-01 1.67783633e-01 5.08998811e-01 9.23427820e-01
8.37536097e-01 1.38386905e-01 9.37112197e-02 -9.62786138e-01
-7.58914649e-02 4.80373353e-01 3.49778324e-01 9.66106832e-01
-8.79114270e-01 1.17163253e+00 8.21628284e+00 1.12167871e+00
-1.52772033e+00 5.70780970e-02 5.51035255e-02 1.84722602e-01
-1.52047098e-01 -4.02493805e-01 -6.27307475e-01 3.67528677e-01
1.10148811e+00 2.04720393e-01 5.11550188e-01 3.49446535e-01
4.11835499e-02 -6.09347284e-01 -9.63978052e-01 1.50957358e+00
8.86489153e-01 -1.17612863e+00 3.15439105e-02 1.08681642e-01
2.01368973e-01 -2.43547678e-01 2.32166618e-01 1.13317415e-01
-4.26930040e-01 -1.37422907e+00 6.16907597e-01 6.04039729e-01
1.31491029e+00 -5.34172416e-01 4.25588816e-01 4.56115991e-01
-1.33000588e+00 2.08742857e-01 -1.31197199e-02 4.68351692e-02
1.47891641e-01 -2.56607503e-01 -1.13903391e+00 1.09014250e-01
1.75000325e-01 5.34463882e-01 -4.59814280e-01 1.43845105e+00
-1.47627085e-01 6.79047287e-01 -4.21436727e-01 -6.18952096e-01
-5.43185100e-02 2.39928275e-01 6.06347382e-01 1.75442564e+00
1.83792040e-01 7.81491324e-02 -3.34135324e-01 2.62000769e-01
1.78481042e-01 7.69750476e-01 -9.28800702e-01 -7.20622344e-03
-4.98833321e-02 8.46296608e-01 -5.94507039e-01 -3.79021913e-01
-3.73943955e-01 9.34436977e-01 -1.04627199e-01 1.46341547e-01
-2.26502985e-01 -4.61087048e-01 3.40189010e-01 -4.00417969e-02
1.75029010e-01 -9.42926705e-02 -2.45351180e-01 -8.96341503e-01
-1.81413472e-01 -8.68057132e-01 -1.12749608e-02 -1.04889143e+00
-9.36128855e-01 6.11113012e-01 -2.65960470e-02 -9.35534120e-01
-5.87000668e-01 -1.12568200e+00 -6.75358057e-01 8.05922449e-01
-1.68958569e+00 -1.11649752e+00 9.68555361e-02 6.29517198e-01
1.10059559e+00 -5.43464124e-01 8.94361198e-01 -2.47677900e-02
2.91559398e-02 8.80740106e-01 -2.09880576e-01 -5.05794119e-03
7.15265930e-01 -1.21499407e+00 -9.30705387e-03 4.26784545e-01
6.24547720e-01 9.82064307e-02 7.23069966e-01 -4.57017660e-01
-8.03951204e-01 -1.34893119e-01 1.53188396e+00 -5.18098712e-01
2.93119162e-01 -4.50183690e-01 -7.92005658e-01 7.94589520e-02
7.63533890e-01 -4.21934932e-01 7.52716780e-01 -2.11603925e-01
-5.48731051e-02 2.47747570e-01 -1.38333130e+00 2.77274877e-01
8.66680503e-01 -1.08944213e+00 -1.09075034e+00 1.78573713e-01
1.45452365e-01 -2.37253979e-01 -2.48886481e-01 3.33785504e-01
8.96860540e-01 -8.40268195e-01 8.30675900e-01 -4.15163070e-01
-1.18054986e-01 -6.16222881e-02 2.63082748e-03 -1.28325236e+00
4.09242958e-01 -8.43031943e-01 3.97064015e-02 1.21619451e+00
5.59015572e-01 -6.72220170e-01 6.84089601e-01 2.07379982e-02
1.87995732e-01 -4.99048173e-01 -1.13128352e+00 -7.41258025e-01
1.57352701e-01 -7.27724373e-01 2.93227255e-01 4.93717760e-01
4.54635620e-01 2.96286196e-01 -4.93838996e-01 -4.36775357e-01
3.90245318e-01 -5.13872862e-01 4.95571762e-01 -1.23719895e+00
6.10772558e-02 -8.77420425e-01 -7.00829506e-01 -1.22480917e+00
5.92328370e-01 -9.25971150e-01 4.58473474e-01 -1.55179071e+00
-1.42211750e-01 -6.60483092e-02 -2.01093614e-01 5.36468208e-01
1.51406750e-02 6.03519857e-01 2.50826716e-01 -8.38897824e-02
-7.68136457e-02 1.83627725e-01 9.18257177e-01 -2.29299411e-01
-3.34091067e-01 2.18961209e-01 -3.17330658e-01 6.25006497e-01
1.02898145e+00 -8.34625363e-02 -4.68356818e-01 -3.24811414e-02
-1.53760046e-01 -1.80716842e-01 3.19004476e-01 -1.09588301e+00
2.97810853e-01 1.75996795e-01 3.05544764e-01 -8.56738687e-01
6.97377980e-01 -6.54303253e-01 -3.48879993e-01 1.93147227e-01
-5.08896589e-01 -1.59481522e-02 4.89420831e-01 3.60906988e-01
-2.75205493e-01 -6.28740847e-01 9.44141328e-01 7.31217340e-02
-9.32489276e-01 -2.39507511e-01 -8.99733067e-01 -1.68851644e-01
9.49656785e-01 -9.53381598e-01 -6.20102417e-03 -5.98366141e-01
-9.59663928e-01 -3.78594637e-01 4.49313670e-01 4.90809470e-01
8.77192676e-01 -1.02925599e+00 -5.86801708e-01 5.99952817e-01
2.23478209e-02 -9.18638349e-01 -3.23507518e-01 8.93421829e-01
-4.30101633e-01 5.58224976e-01 -3.37938935e-01 -5.73402464e-01
-1.91627097e+00 6.35658503e-01 4.37200338e-01 6.84438407e-01
-5.06442547e-01 1.09170508e+00 -2.53281534e-01 -4.80625825e-03
5.67992628e-01 -2.88699210e-01 -7.98834264e-01 3.27662349e-01
5.36015868e-01 2.77509153e-01 3.51895206e-02 -1.17250347e+00
-2.68233210e-01 9.09652352e-01 2.14196950e-01 -7.45050073e-01
7.38097548e-01 -3.31523776e-01 2.18549117e-01 7.30288327e-01
1.41572583e+00 5.69535315e-01 -9.02116418e-01 1.52103052e-01
-1.59142062e-01 -4.14032042e-01 2.60199606e-01 -9.23053861e-01
-6.19575858e-01 1.29784870e+00 9.10175085e-01 5.24651587e-01
1.36228931e+00 1.50882736e-01 4.99840438e-01 1.08141109e-01
-2.28974987e-02 -1.39679420e+00 -2.64737517e-01 2.89359212e-01
7.20342696e-01 -1.49550235e+00 -3.33636791e-01 -4.82800752e-01
-6.39598370e-01 1.35064459e+00 1.79082930e-01 3.91797811e-01
8.17827523e-01 3.16606402e-01 3.99743527e-01 1.77290946e-01
-6.12066150e-01 -8.08184803e-01 7.37879336e-01 1.12390769e+00
5.19024134e-01 2.38803402e-03 -5.42944551e-01 -2.60292947e-01
-5.30292988e-01 9.26880538e-02 2.68752664e-01 7.06255794e-01
-8.03879499e-01 -1.32779229e+00 -3.88581365e-01 3.75680089e-01
-3.79692525e-01 -4.58508968e-01 -8.43544841e-01 8.94915700e-01
3.34926128e-01 1.22060418e+00 2.73042805e-02 -4.27061439e-01
-1.85104012e-02 6.74359620e-01 7.70189524e-01 -3.71796578e-01
-3.52025598e-01 2.27277145e-01 2.32297733e-01 -2.09318250e-01
-8.87756348e-01 -8.42327058e-01 -1.01481855e+00 -2.45295703e-01
-6.13113344e-01 1.00293703e-01 1.13496625e+00 1.07171082e+00
-2.63854057e-01 -2.64270026e-02 3.85954529e-01 -9.81675506e-01
-8.49145129e-02 -1.06691766e+00 -5.29824018e-01 9.48585793e-02
8.86433661e-01 -6.57116771e-01 -7.22644985e-01 3.91659647e-01] | [14.301273345947266, 4.990958213806152] |
5fcc1363-c909-4200-a653-8e220fc1a5ee | clickbait-classification-and-spoiling-using | 2306.14907 | null | https://arxiv.org/abs/2306.14907v1 | https://arxiv.org/pdf/2306.14907v1.pdf | Clickbait Classification and Spoiling Using Natural Language Processing | Clickbait is the practice of engineering titles to incentivize readers to click through to articles. Such titles with sensationalized language reveal as little information as possible. Occasionally, clickbait will be intentionally misleading, so natural language processing (NLP) can scan the article and answer the question posed by the clickbait title, or spoil it. We tackle two tasks: classifying the clickbait into one of 3 types (Task 1), and spoiling the clickbait (Task 2). For Task 1, we propose two binary classifiers to determine the final spoiler type. For Task 2, we experiment with two approaches: using a question-answering model to identify the span of text of the spoiler, and using a large language model (LLM) to generate the spoiler. Because the spoiler is contained in the article, we frame the second task as a question-answering approach for identifying the starting and ending positions of the spoiler. We created models for Task 1 that were better than the baselines proposed by the dataset authors and engineered prompts for Task 2 that did not perform as well as the baselines proposed by the dataset authors due to the evaluation metric performing worse when the output text is from a generative model as opposed to an extractive model. | ['Elisa Ferracane', 'Adhitya Thirumala'] | 2023-06-16 | null | null | null | null | ['classification-1', 'question-answering'] | ['methodology', 'natural-language-processing'] | [ 2.90592194e-01 4.54388291e-01 -8.69120136e-02 -1.54326141e-01
-1.00675631e+00 -9.91151750e-01 7.86747694e-01 3.38909686e-01
-3.19158792e-01 4.92910802e-01 2.63491422e-01 -7.89186895e-01
4.85700630e-02 -4.70862180e-01 -9.55855370e-01 -1.03239812e-01
5.46546519e-01 5.89967012e-01 4.24702644e-01 -2.32470855e-01
7.24069297e-01 -7.18001574e-02 -1.52031648e+00 6.53722286e-01
1.01094460e+00 8.37585807e-01 5.17843187e-01 1.11265099e+00
-6.21265054e-01 1.00484502e+00 -1.13979983e+00 -5.47806919e-01
-4.49599652e-03 -6.23431206e-01 -9.75013733e-01 -6.68020546e-02
9.06753659e-01 -1.04675218e-01 1.25778154e-01 1.01637352e+00
1.07862651e-01 -3.08868915e-01 4.74475741e-01 -1.16344452e+00
-6.12562537e-01 9.72645640e-01 -3.27301949e-01 2.31452510e-02
7.90151656e-01 5.65590709e-02 1.39666271e+00 -8.06804657e-01
7.94782162e-01 1.20333242e+00 3.86141002e-01 5.86065590e-01
-1.26545393e+00 -4.54824746e-01 -2.16741897e-02 2.03653365e-01
-9.99749601e-01 -3.72738063e-01 8.20238531e-01 -7.77419865e-01
5.31938255e-01 7.72876620e-01 2.63287395e-01 1.59002674e+00
1.02202713e-01 1.16127086e+00 1.32003438e+00 -5.90353191e-01
2.38540545e-01 6.03385448e-01 4.54057932e-01 3.27057183e-01
1.48546584e-02 1.13913938e-01 -6.23664737e-01 -2.63370335e-01
2.38597877e-02 -5.55311739e-01 -1.89662784e-01 -3.94779108e-02
-1.27035511e+00 1.10917413e+00 2.10862026e-01 2.55265266e-01
-4.12545204e-01 9.50210541e-02 3.43045384e-01 2.56465793e-01
2.94288337e-01 1.16354120e+00 -3.71400774e-01 -2.53558427e-01
-1.38875461e+00 9.37410414e-01 1.44235790e+00 7.96369553e-01
2.90015459e-01 -6.29229307e-01 -6.89817488e-01 7.07474232e-01
3.31234783e-01 4.24199432e-01 4.65706229e-01 -5.39894402e-01
6.54673934e-01 3.90019745e-01 6.17579222e-01 -1.03825986e+00
-1.59063682e-01 -4.42736447e-01 -9.75754708e-02 2.02366002e-02
7.29778886e-01 -1.45054877e-01 -9.67407405e-01 1.43594742e+00
-1.22630261e-01 -6.24018669e-01 -3.00156027e-01 1.01415110e+00
9.26267922e-01 8.51250887e-01 -2.12655286e-03 1.55882835e-01
1.78068554e+00 -1.12141120e+00 -1.08969438e+00 -5.88426590e-01
4.61197078e-01 -1.26361549e+00 1.58499980e+00 5.15994787e-01
-9.76480901e-01 -5.80814004e-01 -1.20347321e+00 -2.03930512e-01
-5.73380888e-01 3.00371528e-01 7.76544213e-02 4.56963539e-01
-6.54820681e-01 3.60370129e-01 -1.59682736e-01 -1.88144743e-01
6.84816688e-02 -4.41893369e-01 4.10772890e-01 3.70082349e-01
-1.61593771e+00 9.39544559e-01 2.43371472e-01 -2.78173387e-01
-8.15131009e-01 -8.43212366e-01 -5.38662553e-01 8.08378682e-02
6.09179497e-01 -5.03957629e-01 1.74871755e+00 -9.02174175e-01
-1.13198078e+00 9.41786945e-01 -1.55916959e-01 -5.22059500e-01
8.81248355e-01 -5.51120460e-01 -2.60918289e-01 4.74765152e-02
4.17344093e-01 3.79931331e-01 1.21631420e+00 -1.35579705e+00
-6.31334960e-01 -2.21476406e-01 1.83768515e-02 -1.07278220e-01
2.22166300e-01 1.43919706e-01 -1.14109293e-01 -7.29967296e-01
-1.28819704e-01 -7.82164752e-01 1.68445066e-01 -3.77702266e-01
-1.28215468e+00 -4.92510766e-01 8.06829512e-01 -1.11681080e+00
1.69594157e+00 -2.01321650e+00 -3.77865970e-01 -3.92581932e-02
3.95396471e-01 3.04970086e-01 7.84014463e-02 5.69792509e-01
-6.54326677e-02 7.06790328e-01 2.30641156e-01 -1.14737615e-01
4.22784984e-01 -3.16425443e-01 -9.34368670e-01 -1.28865823e-01
3.04664597e-02 8.60733151e-01 -9.48088050e-01 -3.50058526e-01
-3.16533059e-01 -1.46315068e-01 -3.14780563e-01 5.70368767e-01
-8.74026775e-01 -7.73835033e-02 -4.21972692e-01 2.79793829e-01
4.83873546e-01 -3.64715725e-01 -3.12828869e-01 -1.06854640e-01
-3.15225959e-01 9.01922166e-01 -7.33087838e-01 1.12942505e+00
-4.68697399e-01 9.74132299e-01 9.13703367e-02 -3.34789395e-01
1.01978743e+00 1.22788139e-01 -2.22960413e-01 -6.89316809e-01
-1.67942166e-01 3.62606585e-01 -1.09774753e-01 -1.05486715e+00
6.76764727e-01 -3.96440141e-02 -4.55364794e-01 6.08717084e-01
-2.14254469e-01 -3.22192460e-01 3.48178238e-01 5.19719541e-01
1.17196214e+00 1.13039032e-01 4.94403057e-02 -9.84965712e-02
3.10570329e-01 2.76337981e-01 -1.26291022e-01 1.45180786e+00
-7.27987587e-02 4.21345681e-01 8.20490539e-01 -1.95114121e-01
-1.25196636e+00 -1.01142311e+00 2.19318300e-01 1.37736440e+00
6.89236820e-02 -6.76309884e-01 -8.92366111e-01 -9.04350400e-01
-1.46485150e-01 1.59190881e+00 -5.58680713e-01 -8.16751942e-02
-2.71060884e-01 -1.24382684e-02 5.96630514e-01 3.04628294e-02
4.22814578e-01 -9.41796303e-01 -7.70272493e-01 1.44452929e-01
-7.74177551e-01 -8.98848593e-01 -7.51762927e-01 3.18219632e-01
-2.42949232e-01 -9.47629929e-01 -6.75015330e-01 -5.54529071e-01
3.50533038e-01 -2.93641001e-01 1.16531777e+00 -1.12838045e-01
-8.56821314e-02 -6.54822141e-02 -4.42942590e-01 -7.41918921e-01
-8.84316564e-01 1.88490167e-01 -5.60639679e-01 -4.09939140e-02
5.33565283e-01 1.85348913e-01 -3.88715386e-01 3.10539484e-01
-1.07704163e+00 2.64611691e-01 5.87614655e-01 6.12226307e-01
-2.02544197e-01 -2.36249581e-01 3.77367646e-01 -1.17864108e+00
1.32940006e+00 -4.62829560e-01 -2.83567846e-01 3.79950494e-01
-5.39841652e-01 4.70888495e-01 7.27344096e-01 -5.16036630e-01
-8.44031453e-01 -2.42988870e-01 -3.42751056e-01 -1.56924710e-01
-3.84496480e-01 5.00061214e-01 6.79627433e-02 5.10514677e-01
1.02301538e+00 2.50573814e-01 -1.69757411e-01 -7.63494432e-01
3.36738944e-01 8.84021878e-01 4.52677727e-01 -3.52983832e-01
9.91180062e-01 -1.48742422e-01 -6.09919906e-01 -7.26012707e-01
-1.23968971e+00 -5.07370710e-01 -1.39605209e-01 -3.18195939e-01
8.08555424e-01 -2.29983494e-01 -7.27810383e-01 3.15697789e-01
-1.69501972e+00 3.37179489e-02 -1.91755638e-01 7.28230551e-03
-4.08709973e-01 2.27085352e-01 -3.16257447e-01 -8.36725175e-01
-2.26153433e-01 -1.03401709e+00 9.72900867e-01 1.67649299e-01
-8.03997159e-01 -7.45838284e-01 9.58601236e-02 7.88592279e-01
5.76445043e-01 1.54375464e-01 1.31492102e+00 -1.49664235e+00
-2.81775802e-01 -6.03939772e-01 -1.49490938e-01 1.65001899e-01
-2.80895323e-01 1.53753147e-01 -9.42141473e-01 3.58719170e-01
1.74233839e-01 -5.06751299e-01 6.63024306e-01 -3.78097384e-03
1.03666043e+00 -1.00408101e+00 -1.52522266e-01 2.40268316e-02
9.40511048e-01 6.91296384e-02 5.99819124e-01 4.66388702e-01
3.23197097e-01 1.03120494e+00 6.24166667e-01 1.76665977e-01
2.93585300e-01 7.98967540e-01 4.26928252e-01 1.98040888e-01
-1.44803569e-01 -1.11740506e+00 4.87098277e-01 4.18956816e-01
7.26716101e-01 -3.27689439e-01 -8.54645789e-01 4.02941346e-01
-1.63529396e+00 -7.50976801e-01 -5.03490269e-01 2.16055989e+00
9.81650829e-01 4.80846405e-01 2.10670590e-01 1.40826017e-01
7.29330480e-01 3.55707735e-01 -2.43028298e-01 -7.50139654e-01
1.88048378e-01 -7.96626434e-02 4.10237730e-01 5.73743522e-01
-1.05477965e+00 7.07177579e-01 6.27609205e+00 9.21235502e-01
-1.03046417e+00 -3.84433866e-02 5.00249207e-01 2.33998328e-01
-5.04151881e-01 2.43274629e-01 -1.15384674e+00 9.73483324e-01
1.06223714e+00 -1.26496062e-01 3.63385975e-01 9.39863145e-01
5.82131684e-01 -2.09535465e-01 -1.22265351e+00 7.01957822e-01
2.66495645e-01 -1.14386451e+00 3.33190225e-02 -2.12805465e-01
1.43596336e-01 -4.32799816e-01 -2.29701605e-02 5.03393710e-01
2.05637619e-01 -1.02658772e+00 1.31967998e+00 5.97072840e-01
2.63401955e-01 -6.22773692e-02 6.09108329e-01 7.60676742e-01
-4.01065141e-01 2.30681114e-02 2.31208965e-01 1.40520046e-02
2.01083586e-01 5.74930966e-01 -1.38707268e+00 2.70671910e-03
4.17268127e-01 7.18305120e-03 -1.12351894e+00 1.18204129e+00
-7.08533406e-01 9.61169541e-01 -7.75172636e-02 -8.97893667e-01
2.74764866e-01 2.51154423e-01 1.10023844e+00 1.11982870e+00
-1.19985845e-02 -6.93374515e-01 5.59356995e-02 1.53783143e+00
-4.41125892e-02 1.47956476e-01 -3.68863463e-01 -6.65690243e-01
4.46560234e-01 1.20165944e+00 -4.65054214e-01 -3.73110294e-01
1.03169322e-01 9.48003232e-01 1.26794381e-02 3.79720747e-01
-8.54927242e-01 -1.01729727e+00 -2.05742970e-01 6.47785187e-01
1.59975797e-01 1.14238769e-01 -5.59634566e-01 -9.91592765e-01
1.50686041e-01 -1.07114208e+00 1.34148449e-01 -1.18962216e+00
-1.23591173e+00 4.99096155e-01 -1.62765980e-01 -7.26845086e-01
-2.32330412e-01 -4.79660541e-01 -7.29640186e-01 9.39416766e-01
-1.22962153e+00 -9.00858223e-01 -2.48062313e-01 -1.40219823e-01
7.20328927e-01 4.30273473e-01 4.01312560e-01 1.01486459e-01
-1.95752040e-01 4.39645827e-01 -3.79574358e-01 2.02699021e-01
8.25760782e-01 -1.61759353e+00 3.08289438e-01 6.48165584e-01
2.20930338e-01 9.03005064e-01 1.46732748e+00 -7.19819605e-01
-1.13204956e+00 -7.43825555e-01 1.65793836e+00 -6.44564092e-01
1.13390183e+00 -7.56186724e-01 -8.82382393e-01 2.94952273e-01
3.46708864e-01 -7.85140276e-01 5.96580863e-01 5.35923429e-02
-4.06724691e-01 3.02457720e-01 -7.81305850e-01 6.07707441e-01
2.16740683e-01 -5.32234192e-01 -1.26930630e+00 5.83123744e-01
9.12246823e-01 -3.01300228e-01 -1.54781222e-01 -2.01735273e-01
4.61511433e-01 -6.74474299e-01 5.54780960e-01 -9.17381942e-01
9.85696971e-01 -1.12321697e-01 1.34804606e-01 -1.32065392e+00
2.32991241e-02 -8.39899421e-01 -1.63174585e-01 1.31478810e+00
8.35057914e-01 -3.11642349e-01 5.04439294e-01 6.13452435e-01
9.52693149e-02 -5.63134491e-01 -5.21281362e-01 -6.58688307e-01
1.36171803e-02 -3.78092617e-01 3.21982145e-01 4.47715610e-01
7.30439946e-02 9.64888692e-01 -3.80221099e-01 -2.72787035e-01
5.18346786e-01 1.88099056e-01 6.80411339e-01 -1.01849401e+00
-2.77228326e-01 -5.98231852e-01 3.56098473e-01 -1.58984792e+00
-2.34855831e-01 -8.62370074e-01 5.91413319e-01 -1.31049597e+00
8.48711506e-02 -1.97179541e-01 2.44776353e-01 3.05412233e-01
-2.88330793e-01 -2.16050401e-01 4.05992329e-01 2.48134673e-01
-4.88674909e-01 1.13508642e-01 1.13582659e+00 -2.11915776e-01
-1.92022845e-01 4.51388389e-01 -1.04021168e+00 6.92716539e-01
4.82954770e-01 -4.71077949e-01 8.74136761e-03 -1.23031316e-02
9.07780707e-01 1.64857119e-01 7.00117826e-01 -6.02369666e-01
4.03162122e-01 -7.61172548e-02 6.42335638e-02 -8.91627669e-01
-6.39497042e-02 -6.46247625e-01 -3.73031527e-01 4.39270973e-01
-1.20011926e+00 -2.81888455e-01 -1.38516054e-01 3.48037928e-01
3.38479243e-02 -9.17031586e-01 5.53307056e-01 -1.46087125e-01
-2.12043896e-01 -3.66073847e-01 -7.43301094e-01 3.22374225e-01
7.56634831e-01 5.28904423e-02 -6.83552802e-01 -6.53606057e-01
-6.87110424e-01 1.56241074e-01 1.82173595e-01 7.80502200e-01
4.62659925e-01 -8.45081329e-01 -6.59489512e-01 2.32123453e-02
1.43751770e-01 -3.92329782e-01 -2.15326831e-01 6.50010169e-01
-6.16739452e-01 8.78477633e-01 3.09170932e-01 -3.17828745e-01
-9.84180212e-01 5.67183077e-01 4.89315614e-02 -5.85589528e-01
-1.59100041e-01 7.74290442e-01 -6.44562468e-02 -3.85376900e-01
3.92284453e-01 -3.84956777e-01 -3.19089323e-01 3.76680315e-01
6.41765058e-01 2.93728113e-01 1.38534173e-01 -1.97980046e-01
-1.78697467e-01 -8.81983042e-02 -3.97261709e-01 -4.27050948e-01
8.64479661e-01 -1.08273640e-01 1.21294603e-01 7.12740123e-01
1.08521497e+00 2.47605383e-01 -7.64358282e-01 -3.68550420e-02
5.89211643e-01 -3.96581143e-01 -1.37414900e-04 -1.50098073e+00
-1.05332762e-01 9.15719807e-01 8.89081694e-03 9.81096447e-01
4.53761578e-01 3.06700110e-01 9.13225472e-01 1.87498301e-01
-1.42385915e-01 -1.22682178e+00 3.80592763e-01 6.62603140e-01
1.24508262e+00 -1.04130828e+00 -2.83002436e-01 -2.71632910e-01
-8.03298414e-01 1.25572073e+00 6.00350022e-01 9.76737887e-02
3.27744126e-01 1.37980700e-01 1.59353748e-01 -5.03716886e-01
-8.35925221e-01 1.48064986e-01 6.00255013e-01 1.32079497e-01
3.76386255e-01 -4.30989489e-02 -5.70778370e-01 9.30328727e-01
-7.52664745e-01 -2.52705775e-02 5.01816869e-01 8.51742804e-01
-7.26571500e-01 -6.49722040e-01 -5.40781975e-01 7.40802467e-01
-7.47159660e-01 -2.19277591e-01 -1.00130975e+00 4.67052728e-01
-1.84914917e-01 1.24493527e+00 -5.85940480e-02 -5.96882463e-01
2.84011602e-01 3.51006031e-01 -1.21069700e-01 -6.56231582e-01
-7.87378728e-01 6.63639382e-02 4.02740449e-01 -3.93596828e-01
1.84896037e-01 -3.15898538e-01 -6.87284768e-01 1.25433505e-01
-4.51313585e-01 5.39303303e-01 1.18919766e+00 9.90853608e-01
1.95350155e-01 3.86053205e-01 7.33331859e-01 -2.02674165e-01
-1.15904081e+00 -1.18342173e+00 -2.44359881e-01 6.15515590e-01
4.90289986e-01 -3.05755675e-01 -7.87826777e-01 1.46484226e-01] | [11.989368438720703, 8.991891860961914] |
72a6ab0e-35af-45c0-9b37-8bfacf04b594 | physics-informed-machine-learning-method-for-1 | 2108.00037 | null | https://arxiv.org/abs/2108.00037v1 | https://arxiv.org/pdf/2108.00037v1.pdf | Physics-Informed Machine Learning Method for Large-Scale Data Assimilation Problems | We develop a physics-informed machine learning approach for large-scale data assimilation and parameter estimation and apply it for estimating transmissivity and hydraulic head in the two-dimensional steady-state subsurface flow model of the Hanford Site given synthetic measurements of said variables. In our approach, we extend the physics-informed conditional Karhunen-Lo\'{e}ve expansion (PICKLE) method for modeling subsurface flow with unknown flux (Neumann) and varying head (Dirichlet) boundary conditions. We demonstrate that the PICKLE method is comparable in accuracy with the standard maximum a posteriori (MAP) method, but is significantly faster than MAP for large-scale problems. Both methods use a mesh to discretize the computational domain. In MAP, the parameters and states are discretized on the mesh; therefore, the size of the MAP parameter estimation problem directly depends on the mesh size. In PICKLE, the mesh is used to evaluate the residuals of the governing equation, while the parameters and states are approximated by the truncated conditional Karhunen-Lo\'{e}ve expansions with the number of parameters controlled by the smoothness of the parameter and state fields, and not by the mesh size. For a considered example, we demonstrate that the computational cost of PICKLE increases near linearly (as $N_{FV}^{1.15}$) with the number of grid points $N_{FV}$, while that of MAP increases much faster as $N_{FV}^{3.28}$. We demonstrated that once trained for one set of Dirichlet boundary conditions (i.e., one river stage), the PICKLE method provides accurate estimates of the hydraulic head for any value of the Dirichlet boundary conditions (i.e., for any river stage). | ['Alexandre M. Tartakovsky', 'David A. Barajas-Solano', 'Yu-Hong Yeung'] | 2021-07-30 | null | null | null | null | ['physics-informed-machine-learning'] | ['graphs'] | [-1.41969666e-01 -2.65690003e-04 3.67198735e-01 5.35098426e-02
-7.31775284e-01 -3.78839314e-01 3.34166110e-01 2.91945159e-01
-3.88795882e-01 1.00668550e+00 -2.05563217e-01 -7.26943970e-01
-2.25071847e-01 -1.12214613e+00 -7.43258357e-01 -1.02122486e+00
-4.85686809e-01 5.47509432e-01 4.09199186e-02 -3.24305773e-01
1.01630598e-01 5.46512246e-01 -1.48297584e+00 -4.48080122e-01
1.03835034e+00 1.01904881e+00 -5.76548949e-02 8.47340107e-01
2.57460680e-02 4.09123600e-01 -2.83591926e-01 5.45301259e-01
1.97386608e-01 -5.72216749e-01 -7.57997036e-01 -3.27111036e-02
6.88016042e-02 -4.11619037e-01 1.61779858e-02 9.36408341e-01
3.74801964e-01 7.00576425e-01 1.09644854e+00 -8.22035551e-01
1.59528688e-01 -1.50438678e-02 -6.93924844e-01 -9.27304663e-03
-2.44849652e-01 1.65844649e-01 6.83809817e-01 -8.34755778e-01
2.83614278e-01 8.67651463e-01 9.99901235e-01 2.15913713e-01
-1.45145023e+00 -3.74241561e-01 4.19369753e-04 -5.32579184e-01
-1.55975187e+00 -4.00588453e-01 3.45438719e-01 -8.88369560e-01
8.24922979e-01 2.77532011e-01 8.67832303e-01 -2.92231068e-02
2.50129789e-01 3.92820127e-02 1.05904293e+00 -5.90019166e-01
8.02408338e-01 -6.46457374e-02 -1.95157796e-03 7.23026812e-01
2.35252470e-01 2.44208157e-01 1.27308398e-01 -5.67141414e-01
1.12276804e+00 -4.33350325e-01 -4.13238525e-01 -1.18517213e-01
-6.56043649e-01 1.08742070e+00 1.00061990e-01 4.84449379e-02
-4.78753656e-01 3.23646784e-01 1.73683316e-01 4.33574095e-02
6.75040126e-01 5.29475749e-01 -8.37308109e-01 -4.79237065e-02
-1.16564023e+00 4.32496130e-01 1.08679235e+00 6.07588351e-01
1.10329056e+00 4.24497247e-01 2.14153185e-01 8.76899779e-01
4.48441058e-01 1.18862307e+00 -5.98446913e-02 -1.54086804e+00
-5.10055199e-02 8.72270856e-03 7.26906896e-01 -6.22419775e-01
-4.60842162e-01 -1.81541398e-01 -9.51208234e-01 3.96680444e-01
8.17954957e-01 -8.62598538e-01 -9.72851574e-01 1.66112614e+00
5.00218213e-01 2.02626571e-01 2.01985955e-01 6.16128027e-01
2.78107822e-01 1.25413764e+00 1.62669659e-01 -5.77558637e-01
1.10872221e+00 -4.28829223e-01 -3.66460949e-01 -2.60327667e-01
9.91737127e-01 -4.36993480e-01 9.23833311e-01 -1.10192904e-02
-1.14190495e+00 -9.04645398e-02 -8.05316806e-01 4.42739546e-01
-1.68908641e-01 -2.89459527e-01 4.85127389e-01 3.48582745e-01
-1.10727382e+00 8.14143240e-01 -1.12877917e+00 -1.40517816e-01
-3.58585417e-02 9.81320366e-02 3.18525638e-03 2.46242791e-01
-1.22715998e+00 8.02640378e-01 -3.26600671e-02 4.47704852e-01
-8.99352014e-01 -9.99615192e-01 -9.89310384e-01 2.92463034e-01
-2.60235578e-01 -4.28257823e-01 1.11177778e+00 -4.15928066e-01
-1.51025152e+00 4.22907472e-01 -4.73912209e-01 9.77625418e-03
5.24990261e-01 1.80305943e-01 -9.11429897e-02 8.90460685e-02
1.67262942e-01 3.09643507e-01 3.97540003e-01 -1.51529205e+00
-2.77069211e-01 -4.98127565e-02 -1.51342496e-01 1.82509363e-01
1.23170927e-01 -2.36517519e-01 -5.76595701e-02 -1.35081261e-01
3.15859050e-01 -1.01625633e+00 -6.07598007e-01 1.44505009e-01
-1.80739146e-02 2.50130445e-01 4.38689262e-01 -9.33409572e-01
9.81296301e-01 -2.08671737e+00 -8.65057632e-02 5.60155571e-01
-1.44162685e-01 4.46358882e-02 3.12525064e-01 4.70912308e-01
1.23633161e-01 1.84690878e-01 -1.07537639e+00 -1.45941004e-01
-4.22525853e-01 4.02860075e-01 -1.82875872e-01 7.77704656e-01
-8.37738514e-02 2.68694848e-01 -7.88094819e-01 -2.36782700e-01
4.09118205e-01 5.14836252e-01 -7.16187537e-01 3.22824381e-02
-1.80061951e-01 8.58298838e-01 -4.77034360e-01 1.23958282e-01
1.02597570e+00 -3.99882883e-01 1.42129019e-01 2.12444425e-01
-5.78528166e-01 -1.23869196e-01 -1.54195178e+00 1.02904069e+00
-9.58814919e-01 3.15776616e-01 7.65498340e-01 -9.80059862e-01
9.08681870e-01 3.37357700e-01 5.97488761e-01 -4.04584855e-01
-1.47237200e-02 5.66138566e-01 -2.89006859e-01 -3.34722906e-01
2.33797133e-01 -7.34742641e-01 7.20005855e-02 3.93549860e-01
-3.33198935e-01 -6.63904786e-01 -1.18528157e-01 -9.66284499e-02
8.69445801e-01 3.51312011e-02 1.76244423e-01 -9.77438629e-01
4.46224421e-01 3.89947772e-01 5.46528339e-01 6.62506521e-01
2.48186328e-02 3.23175728e-01 6.63336337e-01 -2.21990108e-01
-1.21218514e+00 -9.06490982e-01 -8.61595809e-01 6.51173592e-01
2.48897642e-01 7.01375455e-02 -7.38441527e-01 2.33204499e-01
2.18458861e-01 7.94476211e-01 -5.55033863e-01 1.71387553e-01
-7.79558897e-01 -1.20422161e+00 4.03480023e-01 4.97666627e-01
6.70706987e-01 -7.80216873e-01 -6.21533811e-01 3.31191897e-01
-1.57971326e-02 -7.58638263e-01 1.87891066e-01 2.51077116e-01
-1.09846818e+00 -9.00566876e-01 -7.21720278e-01 -4.44082230e-01
7.79113591e-01 -4.59904522e-01 8.33356857e-01 1.93290040e-01
-4.46239077e-02 1.85802668e-01 -3.26928347e-02 -2.90981591e-01
-5.43766320e-01 -3.38017792e-01 -2.48537719e-01 -3.19879413e-01
-2.30655253e-01 -5.02237737e-01 -6.53494120e-01 2.19566330e-01
-7.47225225e-01 -1.46226794e-01 -1.53737634e-01 9.33113873e-01
5.64966619e-01 -1.84007175e-02 4.72725332e-01 -6.65075421e-01
2.54830033e-01 -6.22677386e-01 -1.09868896e+00 -1.52367115e-01
-4.49489772e-01 -2.75130700e-02 5.39651871e-01 -1.74228609e-01
-1.06176066e+00 -1.07338391e-01 -2.45862737e-01 -1.75611898e-01
-1.33765504e-01 9.02843416e-01 3.08204681e-01 -9.62106138e-02
6.45092785e-01 1.23686299e-01 3.77737656e-02 -5.53859055e-01
7.06519783e-02 5.90191126e-01 3.57173234e-01 -9.54418123e-01
5.99335134e-01 5.62552631e-01 4.13494945e-01 -1.30835497e+00
-3.46684217e-01 -1.66474536e-01 -5.62887192e-01 -7.82076269e-02
7.65946329e-01 -9.43038940e-01 -7.42336929e-01 6.33309543e-01
-7.75721967e-01 -8.90018165e-01 -5.57478607e-01 6.87982976e-01
-6.04984522e-01 1.36210456e-01 -7.70510852e-01 -1.25277388e+00
-1.79217279e-01 -9.81378138e-01 9.41935122e-01 -3.60458158e-02
-1.66986093e-01 -1.41662228e+00 9.58015919e-02 -1.51554585e-01
7.62556314e-01 5.44729471e-01 1.22552168e+00 6.77164970e-03
-2.11302638e-01 -4.94508632e-02 -1.50818154e-01 2.54243404e-01
-6.99915960e-02 1.21760525e-01 -8.97639275e-01 -3.86461854e-01
3.11611593e-01 1.70558095e-01 6.81414425e-01 9.08046842e-01
7.29066610e-01 -4.30055052e-01 -3.58182162e-01 7.56645441e-01
1.70731091e+00 3.00051689e-01 4.11962897e-01 9.11239982e-02
4.20258164e-01 3.75465393e-01 2.34773993e-01 7.98291326e-01
1.89805001e-01 9.29051265e-02 1.26987115e-01 -3.83247375e-01
3.12339544e-01 -6.48234338e-02 -6.66849986e-02 5.09310782e-01
-1.72583073e-01 -4.79097739e-02 -1.36219013e+00 8.12351465e-01
-1.46933031e+00 -5.68704188e-01 -2.75631607e-01 2.47031713e+00
8.52473140e-01 -2.15157896e-01 -3.53184760e-01 2.82313637e-02
7.14957595e-01 6.41608611e-02 -5.18063486e-01 -4.99851823e-01
2.02178329e-01 4.61459577e-01 8.54242086e-01 1.18928766e+00
-7.98972309e-01 5.18012166e-01 6.11142492e+00 2.22673148e-01
-1.26259720e+00 3.24992016e-02 5.46236873e-01 2.89550096e-01
-2.33491600e-01 2.76786655e-01 -6.94067836e-01 4.13145036e-01
1.24613178e+00 3.04078683e-02 3.33194286e-01 5.36329985e-01
6.69075668e-01 -5.99338949e-01 -7.34838545e-01 3.01816404e-01
-5.51480532e-01 -1.17028880e+00 -4.79261249e-01 1.34682804e-01
8.94946158e-01 1.49563417e-01 -4.64717031e-01 1.58338875e-01
6.25011146e-01 -9.52267349e-01 4.53140259e-01 4.37437385e-01
1.06096518e+00 -4.91561294e-01 7.24339247e-01 6.01341665e-01
-1.28284049e+00 -2.15721526e-03 -1.50853947e-01 -3.67861986e-01
6.24778807e-01 8.27667713e-01 -3.87325168e-01 1.74106881e-01
7.84582794e-01 1.03812166e-01 3.44778299e-01 9.23887074e-01
4.24135663e-02 9.05604243e-01 -9.53347445e-01 2.72228211e-01
2.51951903e-01 -5.41621387e-01 4.86585766e-01 9.47469175e-01
6.04847312e-01 7.16495752e-01 2.35577300e-01 9.83560145e-01
2.20309123e-01 3.10858525e-02 -2.61045605e-01 4.47515368e-01
5.37894011e-01 6.82078421e-01 -5.35677135e-01 -3.41666996e-01
-1.31624207e-01 1.87440306e-01 -2.07800955e-01 7.68283904e-01
-6.04744434e-01 -4.60159063e-01 8.02815020e-01 5.92451334e-01
4.07273263e-01 -3.64352882e-01 -3.99496764e-01 -8.33499253e-01
-1.90985486e-01 -2.12901235e-01 1.39206305e-01 -7.17753291e-01
-9.92135346e-01 2.40400836e-01 5.47259748e-01 -8.59096825e-01
-3.36055160e-01 -4.70827788e-01 -7.63798475e-01 1.56370533e+00
-1.66150761e+00 -5.82201123e-01 -2.45236769e-01 2.61062503e-01
-1.38110086e-01 4.94279593e-01 1.01800382e+00 1.36016637e-01
-4.49079752e-01 -5.10237459e-03 8.93719912e-01 2.48327866e-01
1.22595564e-01 -8.97682786e-01 2.13229761e-01 6.97739124e-01
-9.78356004e-01 4.75639343e-01 1.04564273e+00 -7.58337319e-01
-1.22682679e+00 -1.04037857e+00 7.16466784e-01 2.01628804e-01
6.51299417e-01 -1.65855158e-02 -1.30135489e+00 6.91303909e-01
-1.16807312e-01 5.16972780e-01 2.36137912e-01 -1.31722182e-01
3.43487293e-01 -2.70315371e-02 -1.46685958e+00 2.49559969e-01
2.87306368e-01 -2.88052708e-01 -3.93882655e-02 3.45175356e-01
3.76668245e-01 -5.33828497e-01 -1.25228822e+00 7.48038352e-01
3.07844192e-01 -5.39086998e-01 5.18198371e-01 -2.52103359e-01
2.11808741e-01 -4.11478072e-01 -3.40192556e-01 -1.33382618e+00
-1.56258032e-01 -4.91343856e-01 1.03465304e-01 8.29234719e-01
6.48283064e-01 -1.08475685e+00 4.93857533e-01 9.16914523e-01
-8.79242793e-02 -7.91582048e-01 -1.23807907e+00 -4.10671026e-01
9.01805341e-01 -2.68387109e-01 3.39886218e-01 1.03003001e+00
-1.79675192e-01 -1.63550720e-01 -2.89730746e-02 6.21641636e-01
6.69092298e-01 8.69960114e-02 3.53359848e-01 -1.44363475e+00
-3.35947156e-01 -5.18262200e-02 2.40638360e-01 -9.95200694e-01
-2.65815072e-02 -3.81038338e-01 4.59729701e-01 -1.49720383e+00
-3.25709373e-01 -1.09398103e+00 9.42763239e-02 4.89132643e-01
3.94337177e-02 -1.83835536e-01 -2.08076477e-01 2.58332491e-01
5.71162105e-01 4.55995530e-01 1.12226200e+00 1.28201991e-01
-6.10706270e-01 -1.15334868e-01 -7.38475397e-02 7.38403320e-01
7.63533771e-01 -3.72704476e-01 -1.93705425e-01 -4.38942790e-01
1.32099554e-01 7.63459921e-01 3.39591295e-01 -8.95299554e-01
6.23352565e-02 -4.27897602e-01 3.49095583e-01 -2.74181813e-01
3.70530605e-01 -4.46816593e-01 2.67502487e-01 3.99570882e-01
-6.16167523e-02 -1.44246951e-01 6.94280326e-01 3.82717341e-01
-7.00030401e-02 -2.23302320e-01 1.29756832e+00 -3.87709469e-01
-2.69483238e-01 1.08731531e-01 -6.74120486e-01 1.53182790e-01
8.81944776e-01 -9.02090594e-02 -1.57884687e-01 -4.06194001e-01
-7.26712465e-01 3.66216213e-01 5.79245150e-01 -5.86812973e-01
1.96407199e-01 -7.78441310e-01 -7.11277544e-01 6.02870762e-01
-3.68245393e-01 6.35208189e-01 3.19895118e-01 8.08915257e-01
-9.65830445e-01 -2.68814117e-02 1.64856568e-01 -6.58611536e-01
-4.32393849e-01 -3.07680797e-02 9.73833203e-01 -1.34712309e-01
-5.12153983e-01 8.73501897e-01 1.00069076e-01 -6.01514161e-01
-3.34647894e-01 -3.06446671e-01 3.35160941e-02 3.46899568e-03
2.20045537e-01 6.04496062e-01 -1.30946320e-02 -6.49739563e-01
-1.08303972e-01 8.25399458e-01 6.79793894e-01 -3.87855351e-01
1.34161067e+00 -1.57424480e-01 -2.72975564e-01 5.44650793e-01
1.13548350e+00 -1.80984452e-01 -1.54443121e+00 3.56660672e-02
-6.35105133e-01 -8.40070546e-02 4.00816530e-01 -6.07973635e-01
-9.86878514e-01 8.62410009e-01 4.67944980e-01 1.50986731e-01
8.30446422e-01 -2.30241761e-01 5.83471596e-01 1.98932022e-01
1.75834402e-01 -8.75640869e-01 -8.51111352e-01 8.28193724e-01
4.90190268e-01 -9.19611692e-01 1.25471771e-01 -3.36685151e-01
-2.66029239e-01 7.56178379e-01 3.52573782e-01 -3.14584002e-02
1.25769973e+00 6.79717839e-01 3.01990032e-01 -1.93449944e-01
-2.81260431e-01 3.12685817e-01 -2.64977872e-01 9.40872356e-02
3.23051363e-01 6.80476204e-02 -1.54446766e-01 -7.59013891e-02
-3.27115543e-02 -9.03877988e-03 5.14300525e-01 1.11980796e+00
-7.32570112e-01 -5.58901429e-01 -5.91401994e-01 6.35254741e-01
-7.25295395e-02 -2.21402049e-01 6.01824820e-01 7.25241780e-01
2.56103631e-02 8.60726297e-01 4.03275996e-01 3.46320987e-01
1.60077482e-01 3.25838268e-01 1.50568802e-02 -4.01518941e-01
-2.15061307e-01 1.92990363e-01 2.13916954e-02 -5.69021627e-02
-3.03634137e-01 -7.45085001e-01 -1.83586192e+00 -5.16023397e-01
-3.83940548e-01 7.69460082e-01 6.66521728e-01 1.05799282e+00
1.07011624e-01 2.26590037e-01 5.20480394e-01 -1.21503413e+00
-5.79356432e-01 -1.10410774e+00 -1.11085236e+00 -8.12980160e-03
5.89674771e-01 -7.43179500e-01 -8.94042790e-01 5.98029373e-03] | [6.494843482971191, 3.3833484649658203] |
7b8256d8-7f71-45f8-b662-ba3ba2191684 | distilled-mid-fusion-transformer-networks-for | 2305.03810 | null | https://arxiv.org/abs/2305.03810v1 | https://arxiv.org/pdf/2305.03810v1.pdf | Distilled Mid-Fusion Transformer Networks for Multi-Modal Human Activity Recognition | Human Activity Recognition is an important task in many human-computer collaborative scenarios, whilst having various practical applications. Although uni-modal approaches have been extensively studied, they suffer from data quality and require modality-specific feature engineering, thus not being robust and effective enough for real-world deployment. By utilizing various sensors, Multi-modal Human Activity Recognition could utilize the complementary information to build models that can generalize well. While deep learning methods have shown promising results, their potential in extracting salient multi-modal spatial-temporal features and better fusing complementary information has not been fully explored. Also, reducing the complexity of the multi-modal approach for edge deployment is another problem yet to resolve. To resolve the issues, a knowledge distillation-based Multi-modal Mid-Fusion approach, DMFT, is proposed to conduct informative feature extraction and fusion to resolve the Multi-modal Human Activity Recognition task efficiently. DMFT first encodes the multi-modal input data into a unified representation. Then the DMFT teacher model applies an attentive multi-modal spatial-temporal transformer module that extracts the salient spatial-temporal features. A temporal mid-fusion module is also proposed to further fuse the temporal features. Then the knowledge distillation method is applied to transfer the learned representation from the teacher model to a simpler DMFT student model, which consists of a lite version of the multi-modal spatial-temporal transformer module, to produce the results. Evaluation of DMFT was conducted on two public multi-modal human activity recognition datasets with various state-of-the-art approaches. The experimental results demonstrate that the model achieves competitive performance in terms of effectiveness, scalability, and robustness. | ['Claude Sammut', 'Binghao Li', 'Lina Yao', 'Jingcheng Li'] | 2023-05-05 | null | null | null | null | ['human-activity-recognition', 'feature-engineering', 'human-activity-recognition'] | ['computer-vision', 'methodology', 'time-series'] | [ 2.69627780e-01 -2.81778514e-01 -3.22666943e-01 -9.66576114e-02
-1.14334750e+00 -1.18669719e-01 6.09986544e-01 -1.00043342e-01
-3.25963318e-01 5.96441031e-01 4.14608449e-01 1.99840620e-01
-3.73286903e-01 -6.71987295e-01 -4.16646689e-01 -8.81928682e-01
4.29039933e-02 2.62760729e-01 3.47566187e-01 -1.56430751e-02
-2.96124928e-02 2.88550526e-01 -1.83260477e+00 5.45295119e-01
9.29887176e-01 1.28282237e+00 6.72757030e-02 5.08779049e-01
1.27179734e-03 9.57396567e-01 -4.61743861e-01 1.25017747e-01
2.76659746e-02 -4.41815972e-01 -7.48735368e-01 2.05314413e-01
8.08153227e-02 -6.63237870e-02 -3.84602368e-01 5.45401216e-01
6.96021914e-01 5.70789754e-01 6.61215365e-01 -1.41156662e+00
-4.90477592e-01 5.99421486e-02 -6.68042898e-01 3.45784098e-01
9.09819007e-01 9.72965062e-02 6.21906936e-01 -8.53681803e-01
2.93810397e-01 1.00851262e+00 5.86196184e-01 3.50912780e-01
-8.41908395e-01 -5.29008210e-01 2.26368308e-01 5.06405890e-01
-1.43247485e+00 -2.87036210e-01 9.06015515e-01 -2.82102853e-01
1.06919932e+00 2.11548313e-01 9.28582132e-01 1.22261310e+00
5.35986722e-02 1.28672385e+00 1.16413724e+00 -3.49922478e-01
5.01959957e-02 1.36861414e-01 -2.46064607e-02 8.54890823e-01
-1.65019467e-01 -8.85068700e-02 -1.02170050e+00 5.20175435e-02
5.74043095e-01 4.98838872e-01 -4.20805186e-01 -3.74423206e-01
-1.32488990e+00 4.89952803e-01 6.49997652e-01 7.61148334e-01
-7.73624957e-01 -1.60865083e-01 3.13076824e-01 -2.33435910e-02
2.29594529e-01 9.05789435e-02 -3.61074746e-01 -4.17477906e-01
-1.17160392e+00 1.39137721e-02 4.53154892e-01 5.71156204e-01
7.24037111e-01 -1.06619813e-01 -4.69025284e-01 8.79507661e-01
3.91262412e-01 3.91441673e-01 8.51048827e-01 -7.87055254e-01
4.88786280e-01 1.10116017e+00 -2.81061918e-01 -9.01927888e-01
-4.61627603e-01 -3.29351395e-01 -9.09536421e-01 -1.70043200e-01
2.12612554e-01 4.94322553e-02 -8.82465422e-01 1.57587981e+00
5.25546432e-01 5.60695887e-01 1.57694712e-01 1.02108955e+00
8.66169751e-01 7.14002490e-01 2.78937966e-01 -1.87672749e-01
1.58773744e+00 -1.18184590e+00 -5.93911767e-01 -3.69568110e-01
5.41346371e-01 -4.17626053e-01 8.92684698e-01 3.36972773e-01
-8.50354433e-01 -5.79294562e-01 -1.09263277e+00 -2.19131075e-02
-6.18605793e-01 9.36225522e-03 7.08990932e-01 4.73215669e-01
-6.31215274e-01 1.89548865e-01 -8.61008763e-01 -5.70350945e-01
6.66670680e-01 1.31418496e-01 -7.37269163e-01 -3.53470206e-01
-1.25086594e+00 9.22854722e-01 5.84364057e-01 2.01874096e-02
-8.69535923e-01 -6.96387589e-01 -1.00955868e+00 -2.37544551e-02
4.95656252e-01 -8.52204382e-01 1.12482345e+00 -8.12022448e-01
-1.38612878e+00 4.37875956e-01 -3.34285229e-01 -2.00718626e-01
3.37018698e-01 1.02502219e-02 -6.92226946e-01 3.82602841e-01
1.96657434e-01 6.00277662e-01 6.51586771e-01 -1.03782940e+00
-1.02042067e+00 -5.27644038e-01 -6.57983124e-02 5.97524226e-01
-7.80373096e-01 -3.55770856e-01 -6.47170722e-01 -8.25848639e-01
7.26386607e-02 -7.90181518e-01 -3.20116356e-02 -1.20819345e-01
-1.03084721e-01 -3.05525690e-01 1.21588790e+00 -6.04839861e-01
1.33619332e+00 -2.03823471e+00 1.29224151e-01 1.91233158e-01
-1.81951169e-02 2.95083076e-01 -8.78203660e-02 3.47722948e-01
3.62714455e-02 -4.33169961e-01 -2.39099532e-01 -3.33284169e-01
2.73666717e-02 2.62333572e-01 1.13898568e-01 3.82817507e-01
1.09543771e-01 1.10005260e+00 -8.77269328e-01 -5.52378714e-01
5.54258704e-01 8.50413322e-01 -3.35733980e-01 2.29825243e-01
4.97794487e-02 4.90778685e-01 -5.98762870e-01 1.04086912e+00
2.13961706e-01 -3.77287328e-01 -1.25452191e-01 -3.74170333e-01
1.54384732e-01 -7.11350590e-02 -1.23891890e+00 2.09793687e+00
-4.67677772e-01 3.65938514e-01 -2.91366518e-01 -1.26055622e+00
5.22609174e-01 5.23767710e-01 9.60132182e-01 -1.01859927e+00
1.79354295e-01 -2.90669687e-02 -3.45826149e-01 -7.54670382e-01
4.30961221e-01 -1.52986914e-01 -3.86374444e-01 4.08650219e-01
2.79167831e-01 3.24852139e-01 2.02562883e-01 1.55323401e-01
1.21942043e+00 3.62402946e-01 2.43809447e-01 2.05829352e-01
7.02080488e-01 2.58521009e-02 6.66994572e-01 2.94379443e-01
-4.28101212e-01 3.75335395e-01 -1.90092757e-01 -2.38230467e-01
-2.50381678e-01 -1.10448253e+00 2.83673108e-01 1.32950747e+00
2.28712231e-01 -4.44748849e-01 -7.21336424e-01 -8.02640021e-01
-1.36128098e-01 3.94124687e-01 -7.17828751e-01 -4.61614937e-01
-3.31453681e-01 -7.41257131e-01 6.39336407e-01 8.11861694e-01
1.10502434e+00 -1.05436730e+00 -8.98674726e-01 1.61848485e-01
-6.99776173e-01 -1.00726914e+00 -4.80485797e-01 1.94718108e-01
-6.77640021e-01 -1.15929461e+00 -7.97908425e-01 -6.78360641e-01
4.51948941e-01 6.02709651e-01 6.07515395e-01 -2.58794039e-01
-4.18498665e-01 7.64922798e-01 -4.28721815e-01 -3.33680212e-01
2.24455103e-01 2.92976517e-02 7.29546025e-02 2.54473925e-01
6.66929245e-01 -6.01711690e-01 -6.80408239e-01 3.45936269e-01
-1.08382082e+00 1.52605856e-02 9.15680230e-01 9.01406884e-01
5.62030554e-01 3.54179800e-01 6.75047755e-01 -3.09970528e-01
5.09716332e-01 -5.02834558e-01 2.18062714e-01 3.77527684e-01
-4.50475246e-01 9.29509178e-02 3.88162106e-01 -8.12779367e-01
-1.35403705e+00 1.39299810e-01 5.89386784e-02 -5.87049603e-01
-3.19102794e-01 8.11047256e-01 -5.45807540e-01 4.55324389e-02
6.10060513e-01 6.08472526e-01 -1.69956982e-01 -2.82976478e-01
2.81187832e-01 6.95071816e-01 7.64051616e-01 -6.55988216e-01
7.33313262e-01 4.24756676e-01 -2.51559764e-01 -7.90669620e-01
-8.87525022e-01 -6.48724973e-01 -4.84174490e-01 -4.44952041e-01
9.90873516e-01 -1.08081675e+00 -5.70377469e-01 6.65588558e-01
-7.26136148e-01 -9.15946141e-02 -2.79701680e-01 5.64864337e-01
-5.16065657e-01 3.48383516e-01 -3.61652404e-01 -7.16399670e-01
-2.64656454e-01 -1.05540264e+00 1.13449872e+00 7.22920954e-01
-2.86612630e-01 -9.89915550e-01 1.42825708e-01 1.05642772e+00
2.85332620e-01 2.40106747e-01 5.57482541e-01 -5.78339458e-01
-5.52304506e-01 -4.12831187e-01 -3.04589570e-02 -2.18566675e-02
2.24431783e-01 -5.84268928e-01 -1.12673855e+00 -2.61150748e-01
-1.90867320e-01 -3.81738365e-01 7.95701027e-01 2.69168824e-01
1.06523275e+00 -6.17566593e-02 -6.13179147e-01 4.44295496e-01
9.37090218e-01 1.98295325e-01 6.61955237e-01 4.45304304e-01
7.27810442e-01 4.19098049e-01 9.04206574e-01 4.71425027e-01
8.23975682e-01 6.23792052e-01 2.29564697e-01 -1.07026972e-01
-5.77426106e-02 -2.92340547e-01 6.43008292e-01 5.97022057e-01
-1.15179308e-01 -1.52791124e-02 -7.40732789e-01 6.46042049e-01
-2.11314726e+00 -1.38411880e+00 3.47848386e-01 1.75378203e+00
6.00280404e-01 -1.26966968e-01 3.10291290e-01 5.53785563e-01
3.89663458e-01 2.03042150e-01 -5.79465687e-01 1.03759877e-01
-5.34844361e-02 -4.87628859e-03 5.73088452e-02 -6.79281205e-02
-1.17853153e+00 5.74939728e-01 5.44018602e+00 9.75883365e-01
-9.99104857e-01 2.91380942e-01 1.68739915e-01 -1.91272333e-01
-2.11368836e-02 -2.68824458e-01 -5.55526614e-01 4.43384051e-01
8.04979026e-01 1.81048270e-02 4.02786016e-01 6.41684115e-01
1.51470572e-01 -4.97068614e-01 -1.01920271e+00 1.39597821e+00
3.26681763e-01 -1.03023112e+00 -8.86137132e-03 1.39880285e-01
5.85891247e-01 -2.86393821e-01 -1.17221184e-01 6.81713879e-01
-4.91176881e-02 -9.17840481e-01 6.65799022e-01 8.84679079e-01
3.48682463e-01 -7.79839635e-01 6.20041847e-01 7.65473962e-01
-1.79485500e+00 -3.48726273e-01 2.83233881e-01 1.50331911e-02
2.31016159e-01 4.83308971e-01 -5.39702833e-01 9.19354498e-01
9.07061577e-01 8.14493179e-01 -6.18037105e-01 9.54838932e-01
-4.18019108e-02 2.72693932e-01 -2.56354600e-01 2.04402089e-01
7.13208988e-02 2.38769073e-02 3.23457122e-01 1.19145977e+00
4.99285787e-01 1.95381075e-01 4.60281700e-01 4.73893523e-01
3.14181149e-02 -9.16923806e-02 -5.45313895e-01 -2.25296929e-01
3.90638977e-01 1.38207150e+00 -5.33955157e-01 -4.15386975e-01
-6.39101803e-01 1.13168669e+00 1.76423609e-01 3.32935035e-01
-1.01340044e+00 -2.81506807e-01 5.05793333e-01 2.69055516e-02
1.93522081e-01 4.90679853e-02 -7.58542567e-02 -1.28858900e+00
2.31645271e-01 -9.28203523e-01 9.53946590e-01 -8.60400558e-01
-1.18534005e+00 4.30746764e-01 1.58890843e-01 -1.36343670e+00
-3.53829116e-01 -4.06502396e-01 -5.96225262e-01 7.52186298e-01
-1.39572525e+00 -1.68158793e+00 -6.17844403e-01 1.11201346e+00
6.37095332e-01 -3.00614715e-01 9.26807940e-01 5.45150816e-01
-8.61492038e-01 6.22249126e-01 -4.69792157e-01 1.33184060e-01
6.04169190e-01 -9.81780827e-01 -4.88531679e-01 1.02374125e+00
-2.73969420e-03 6.25317335e-01 2.43752658e-01 -4.43366766e-01
-1.65339541e+00 -1.14566219e+00 6.40095830e-01 -3.89044702e-01
3.31775278e-01 3.54858511e-03 -9.09493446e-01 5.58601975e-01
2.22839817e-01 2.85631716e-01 1.10123396e+00 -8.60866234e-02
-2.69426525e-01 -1.47495896e-01 -1.20716631e+00 4.09771651e-01
1.07939589e+00 -6.92904592e-01 -8.45571041e-01 -1.52287915e-01
2.13993266e-01 -2.06024587e-01 -1.18123293e+00 6.03133202e-01
6.32558167e-01 -8.73725533e-01 1.04387391e+00 -2.78022826e-01
9.41889510e-02 -6.46528125e-01 -2.95495480e-01 -1.35411346e+00
-2.82522619e-01 -2.23355576e-01 -7.25920439e-01 1.30623329e+00
1.03155471e-01 -5.30442178e-01 7.98262239e-01 4.34230477e-01
-2.25010246e-01 -8.21603715e-01 -1.05668151e+00 -6.78323567e-01
-4.63379741e-01 -5.83326519e-01 4.60379809e-01 9.75609958e-01
3.89951795e-01 4.56660122e-01 -3.48961890e-01 1.29894152e-01
4.07768309e-01 1.69555783e-01 7.75573552e-01 -1.23347807e+00
-2.67921090e-01 -4.36086774e-01 -4.30694997e-01 -1.10226738e+00
6.55325204e-02 -8.76826346e-01 5.41009866e-02 -1.86997330e+00
4.39675212e-01 5.58592603e-02 -5.17716467e-01 7.89880216e-01
-1.41185597e-01 3.02628785e-01 8.07143673e-02 8.70154873e-02
-8.66293848e-01 7.83653677e-01 1.14433706e+00 -5.31661510e-01
-3.34457040e-01 -1.50468782e-01 -7.32560277e-01 6.38724923e-01
5.15265226e-01 -1.47821069e-01 -6.72735393e-01 -1.37049693e-03
-2.71037281e-01 1.49176791e-01 4.63189691e-01 -1.45719576e+00
5.91641307e-01 -2.31428146e-01 7.79183805e-01 -7.56977677e-01
7.39491403e-01 -9.25974727e-01 2.65030742e-01 2.74229258e-01
2.87420931e-03 -2.08768651e-01 2.14885339e-01 6.96479499e-01
-4.42789912e-01 2.85460621e-01 6.21989965e-01 5.17559191e-03
-1.08406341e+00 4.45211798e-01 -5.29667497e-01 -1.13853522e-01
1.37558794e+00 -7.08718300e-01 -1.96335882e-01 -2.70529240e-01
-7.34333396e-01 3.40786159e-01 1.60245493e-01 7.35296726e-01
7.20632851e-01 -1.72297347e+00 -2.51019418e-01 2.52311438e-01
5.44942141e-01 -1.52434595e-02 6.49668157e-01 1.24131310e+00
1.77252382e-01 3.20855349e-01 -2.47800097e-01 -7.55771875e-01
-1.12048674e+00 5.23588121e-01 4.56739753e-01 -4.52006787e-01
-4.82387066e-01 4.59026366e-01 -3.32398079e-02 -3.32026482e-01
4.11358595e-01 -1.82077900e-01 -2.57757574e-01 3.41154933e-01
7.25286007e-01 4.77948636e-01 1.57600105e-01 -9.66616333e-01
-6.46868289e-01 6.50346339e-01 2.40496218e-01 -1.61238030e-01
1.24907768e+00 -1.27085492e-01 2.51621723e-01 3.79482836e-01
1.05982566e+00 -4.92166102e-01 -1.15404522e+00 -4.95988339e-01
-3.42716463e-02 -2.91446924e-01 1.73869848e-01 -8.48918855e-01
-1.18413305e+00 8.88584077e-01 8.09322536e-01 -8.04411992e-02
1.60486245e+00 -7.44863153e-02 1.04660165e+00 2.39134967e-01
4.24537629e-01 -1.32894361e+00 5.55168390e-01 4.52862829e-01
5.47957718e-01 -1.21297681e+00 -2.05375507e-01 -1.50572494e-01
-7.19988704e-01 7.63444245e-01 8.23904276e-01 3.51567537e-01
6.55068636e-01 1.83508158e-01 -4.17630896e-02 -2.32239202e-01
-5.69447458e-01 -4.63905454e-01 5.31748235e-01 8.27133834e-01
2.79503196e-01 -1.54302046e-01 6.37306198e-02 1.04006743e+00
2.60089487e-01 3.29427421e-01 -1.92022026e-01 1.32718706e+00
-4.73498344e-01 -8.95583034e-01 -5.88196874e-01 3.47088724e-01
-1.20763861e-01 4.22758996e-01 -1.35710359e-01 5.58512568e-01
4.25002098e-01 1.05026484e+00 -2.53680646e-01 -6.71525896e-01
6.14947081e-01 3.44571203e-01 4.77472275e-01 -3.37005228e-01
-6.03352189e-01 5.64844050e-02 -1.25085741e-01 -6.54824317e-01
-8.09246004e-01 -6.43217742e-01 -1.48438215e+00 -1.89344347e-01
-9.35838148e-02 6.68211803e-02 3.49146485e-01 1.28048790e+00
4.48272347e-01 8.90324771e-01 3.45656067e-01 -1.14490926e+00
-6.62258267e-02 -1.01597738e+00 -3.61309171e-01 5.14785707e-01
6.11267611e-02 -1.05600607e+00 -2.20289384e-03 7.14525059e-02] | [7.949157238006592, 0.754128098487854] |
8a42a8e4-ae82-4bf9-b0a2-1d2bdd3e3474 | keep-it-consistent-topic-aware-storytelling | 1911.04192 | null | https://arxiv.org/abs/1911.04192v2 | https://arxiv.org/pdf/1911.04192v2.pdf | Keep it Consistent: Topic-Aware Storytelling from an Image Stream via Iterative Multi-agent Communication | Visual storytelling aims to generate a narrative paragraph from a sequence of images automatically. Existing approaches construct text description independently for each image and roughly concatenate them as a story, which leads to the problem of generating semantically incoherent content. In this paper, we propose a new way for visual storytelling by introducing a topic description task to detect the global semantic context of an image stream. A story is then constructed with the guidance of the topic description. In order to combine the two generation tasks, we propose a multi-agent communication framework that regards the topic description generator and the story generator as two agents and learn them simultaneously via iterative updating mechanism. We validate our approach on VIST dataset, where quantitative results, ablations, and human evaluation demonstrate our method's good ability in generating stories with higher quality compared to state-of-the-art methods. | ['Ying Cheng', 'Ruize Wang', 'Piji Li', 'Haijun Shan', 'Xuanjing Huang', 'Zhongyu Wei', 'Qi Zhang', 'Ji Zhang'] | 2019-11-11 | null | https://aclanthology.org/2020.coling-main.204 | https://aclanthology.org/2020.coling-main.204.pdf | coling-2020-8 | ['visual-storytelling'] | ['natural-language-processing'] | [ 4.10405695e-01 4.93487418e-01 1.42108083e-01 -5.72030731e-02
-8.74493420e-01 -7.51858413e-01 1.29137254e+00 -3.34179923e-02
1.17298216e-01 8.00270498e-01 7.85850286e-01 2.13993877e-01
4.84740436e-01 -8.36914957e-01 -6.51133537e-01 -5.40226281e-01
3.48651886e-01 4.79218751e-01 2.59307981e-01 -2.26046532e-01
2.26770371e-01 -3.17129284e-01 -1.37125635e+00 8.66725743e-01
6.45736158e-01 4.10106212e-01 5.76270878e-01 4.91218954e-01
-4.73586261e-01 1.16250348e+00 -8.58759940e-01 -4.64499593e-01
-7.59475976e-02 -1.11296248e+00 -9.33355212e-01 6.76462233e-01
9.23507884e-02 -2.97564268e-01 -1.57914951e-01 8.54938865e-01
4.07205373e-01 6.41293153e-02 8.65514278e-01 -1.47650445e+00
-6.11414373e-01 1.02991498e+00 -7.64867902e-01 -3.36163729e-01
6.84289813e-01 2.54929125e-01 1.01363873e+00 -6.93290114e-01
1.23277271e+00 1.28037894e+00 2.99577951e-01 7.27846980e-01
-1.27891719e+00 -4.93631512e-01 4.07970220e-01 1.10559277e-01
-1.21320808e+00 -3.55204850e-01 1.07845950e+00 -5.87807417e-01
6.49416506e-01 1.85211927e-01 8.73340189e-01 1.46394205e+00
-9.82060134e-02 1.09637034e+00 1.04010081e+00 -2.85075396e-01
3.74552339e-01 2.31410936e-01 -3.02088469e-01 6.02873504e-01
-3.78634185e-02 3.87039334e-02 -8.36203933e-01 -4.92561832e-02
7.37484634e-01 -4.70060468e-01 -1.70903444e-01 -4.41052258e-01
-1.62740409e+00 9.42317009e-01 2.01778263e-01 2.07569644e-01
-5.51888704e-01 4.20450687e-01 4.26305979e-01 -1.53188452e-01
4.42767769e-01 3.11010897e-01 6.56520844e-01 1.45255700e-01
-9.38723922e-01 7.98854470e-01 7.77881384e-01 1.20435607e+00
2.97288299e-01 1.03130430e-01 -7.88652778e-01 5.51178336e-01
2.31105521e-01 2.26063162e-01 3.78800780e-01 -7.26369262e-01
4.81099904e-01 5.14623225e-01 2.53348857e-01 -9.50099409e-01
-1.03052676e-01 -2.43570477e-01 -8.36015701e-01 3.47681940e-01
1.76484883e-01 -2.54818082e-01 -8.40147018e-01 1.70053434e+00
4.44780916e-01 3.90616536e-01 3.97917837e-01 9.87690628e-01
1.10473096e+00 1.04227293e+00 2.64203489e-01 -4.04370517e-01
1.56549633e+00 -1.24281657e+00 -9.50096607e-01 -2.91328400e-01
7.82587081e-02 -7.55145252e-01 9.41481054e-01 2.73850083e-01
-1.40360427e+00 -3.48083466e-01 -1.11105049e+00 5.60722463e-02
-6.00665845e-02 6.29909635e-02 3.85661364e-01 1.36544973e-01
-8.13763142e-01 6.80160150e-02 -3.86197209e-01 -2.73615241e-01
4.64588702e-01 -3.93658936e-01 -1.10091642e-01 1.77394807e-01
-8.94224942e-01 6.47571564e-01 7.39874363e-01 -4.13301468e-01
-1.43565726e+00 -4.00127292e-01 -7.97310889e-01 1.48010245e-02
4.84119058e-01 -1.15616858e+00 1.38500452e+00 -9.70851004e-01
-1.46304584e+00 8.10417116e-01 -2.01575190e-01 -4.63259786e-01
7.88567781e-01 8.50302428e-02 3.39866206e-02 2.28848293e-01
4.21642751e-01 1.02645326e+00 1.01027954e+00 -2.02264690e+00
-7.52360642e-01 9.84599069e-02 3.41541767e-02 7.20279574e-01
-9.61143151e-02 -6.54973462e-02 -4.97670203e-01 -1.01645660e+00
-6.60668164e-02 -6.76968038e-01 -1.54069304e-01 -4.32560563e-01
-7.12434232e-01 -2.55975038e-01 9.03367817e-01 -5.93272507e-01
1.06176054e+00 -1.95859218e+00 4.88439232e-01 -1.64711908e-01
3.65752220e-01 -1.63981557e-01 -1.98730245e-01 7.03215957e-01
1.23212054e-01 7.44757652e-02 -5.17365575e-01 -7.23114431e-01
9.11740512e-02 -4.26541306e-02 -7.05974042e-01 1.25015870e-01
1.48289889e-01 9.97266829e-01 -1.20272446e+00 -7.74369061e-01
1.15039378e-01 4.25353825e-01 -2.19498962e-01 4.25633281e-01
-8.09662044e-01 4.99191880e-01 -4.34234530e-01 2.05620185e-01
3.32875669e-01 -2.55027860e-01 2.86420763e-01 -9.88369063e-02
-2.68618036e-02 -2.22337209e-02 -1.13996816e+00 2.09947157e+00
-2.28672311e-01 7.07607567e-01 -3.51525009e-01 -7.46327877e-01
7.77986348e-01 7.03779638e-01 7.66315609e-02 -5.07195890e-01
1.74128469e-02 -3.18986386e-01 -4.86704201e-01 -6.47173822e-01
6.52566254e-01 -4.50509220e-01 -3.67463678e-01 9.86492217e-01
9.80610959e-03 -4.59240586e-01 3.42945039e-01 5.76202810e-01
8.76002729e-01 4.62297678e-01 3.86566758e-01 1.83682606e-01
2.64770120e-01 3.67727399e-01 1.18076131e-01 9.40190494e-01
1.73450008e-01 1.09137666e+00 7.72490859e-01 -3.47701907e-01
-1.34163606e+00 -1.30958486e+00 5.40335596e-01 6.63471758e-01
4.66366023e-01 -6.31584823e-01 -9.79139566e-01 -7.81924963e-01
-4.34955746e-01 9.68413234e-01 -7.04506934e-01 1.21073291e-01
-5.37263811e-01 -5.77387154e-01 3.86634350e-01 2.19852537e-01
7.16769576e-01 -1.47721612e+00 -9.32845652e-01 3.52510214e-01
-7.06309974e-01 -1.22137892e+00 -4.08169121e-01 -5.54091096e-01
-3.60271543e-01 -7.58186638e-01 -8.46917927e-01 -8.50067854e-01
6.69050932e-01 3.61251473e-01 1.17415321e+00 -1.12923108e-01
-1.50136679e-01 2.45750085e-01 -4.45258349e-01 -4.22105789e-01
-9.09952939e-01 -1.10020205e-01 -5.59988022e-01 2.19047129e-01
-4.26961094e-01 -5.77074170e-01 -5.74975908e-01 -1.02048881e-01
-1.12959731e+00 1.11496985e+00 5.13529301e-01 6.91851556e-01
4.25908059e-01 5.56581542e-02 7.28112876e-01 -9.73845720e-01
9.88418937e-01 -6.42790675e-01 -2.46274158e-01 2.88684160e-01
-2.62566239e-01 1.48117900e-01 5.49328685e-01 -3.80499780e-01
-1.61580014e+00 2.45962605e-01 4.10447091e-01 -1.89543039e-01
-2.60788918e-01 3.41825306e-01 -2.62912456e-02 6.82448924e-01
5.55552363e-01 5.50492704e-01 -2.58623153e-01 -2.04672683e-02
8.15939307e-01 3.92640740e-01 7.81566143e-01 -4.86517161e-01
9.18901980e-01 7.67396390e-01 -2.28395820e-01 -5.28862119e-01
-7.37317741e-01 -1.15148216e-01 -1.63989842e-01 -6.43239319e-01
1.05523014e+00 -1.11361074e+00 -3.72358322e-01 2.76692450e-01
-1.71442556e+00 -1.82602078e-01 -3.36958915e-01 8.18875507e-02
-9.86813962e-01 1.15935192e-01 -2.87352502e-01 -7.41522908e-01
-3.80175442e-01 -9.42199290e-01 1.25570893e+00 7.27890134e-02
-5.66625655e-01 -7.48694479e-01 2.30588168e-01 3.32986176e-01
1.29522830e-01 9.18928921e-01 7.55886078e-01 -2.08439916e-01
-7.76411235e-01 1.07058614e-01 -3.41192245e-01 -3.07231516e-01
-1.34222144e-02 -2.25939244e-01 -9.35890138e-01 -2.22759396e-02
-7.93290213e-02 -2.95629293e-01 7.85991192e-01 1.54624388e-01
6.92386806e-01 -6.63339615e-01 -4.39377367e-01 2.12504715e-01
1.55958903e+00 3.00258756e-01 7.57306457e-01 2.95354247e-01
7.54033267e-01 8.16959918e-01 5.50072014e-01 6.63102269e-01
5.14210701e-01 7.44019687e-01 4.21428174e-01 -1.06085531e-01
-5.22357941e-01 -6.75999224e-01 2.58193523e-01 3.46317649e-01
6.14434034e-02 -6.78470969e-01 -7.72749305e-01 8.46997261e-01
-2.36155701e+00 -1.38382041e+00 -8.73527303e-02 1.68450069e+00
7.71431625e-01 -4.82587423e-03 3.04662049e-01 -1.44984677e-01
8.10009778e-01 5.31327069e-01 -1.66776583e-01 -1.97427243e-01
-1.98299557e-01 -3.66239458e-01 -2.16928229e-01 4.19521928e-01
-8.63217950e-01 1.20219302e+00 6.08701658e+00 9.42099094e-01
-8.10215175e-01 3.11877042e-01 6.82556272e-01 -2.03280210e-01
-6.90703094e-01 1.62429899e-01 -4.08219427e-01 4.47615415e-01
2.60415941e-01 -5.52258611e-01 2.22844079e-01 5.57155073e-01
4.37566012e-01 -2.47440323e-01 -8.99043322e-01 9.07815874e-01
4.56578106e-01 -1.87183106e+00 5.56509733e-01 -2.47515455e-01
1.00053918e+00 -6.47234976e-01 -1.68137342e-01 -1.79883409e-02
5.83571851e-01 -9.93968308e-01 1.40637529e+00 4.58060920e-01
4.96299446e-01 -7.66411483e-01 3.82948011e-01 3.88433456e-01
-1.19876468e+00 2.70555496e-01 5.69348484e-02 -7.64867589e-02
7.13258684e-01 2.85160005e-01 -1.06379235e+00 6.15458965e-01
3.25467080e-01 6.57471597e-01 -3.58216345e-01 8.45601439e-01
-5.61776638e-01 3.84678394e-01 2.18271971e-01 -1.40720308e-01
2.33215675e-01 -5.61154895e-02 9.52228904e-01 1.29150045e+00
3.18353891e-01 2.88157433e-01 2.07117006e-01 1.36486769e+00
-8.54499564e-02 6.51449114e-02 -8.93064857e-01 -6.03564142e-04
3.36169511e-01 1.30855739e+00 -1.03916228e+00 -6.65829778e-01
-4.39314395e-02 1.32849979e+00 3.22178245e-01 4.12723571e-01
-1.01965427e+00 -1.16355337e-01 1.47865647e-02 -1.02935340e-02
1.26855522e-01 -1.30646443e-02 -5.13440549e-01 -1.09876180e+00
1.54276729e-01 -7.96011090e-01 1.59972802e-01 -1.34401572e+00
-8.52013469e-01 8.63072157e-01 3.30610514e-01 -1.40694022e+00
-5.69788396e-01 2.89323598e-01 -1.00765145e+00 4.76874083e-01
-1.12239897e+00 -1.56820178e+00 -6.11842930e-01 2.99828321e-01
1.01254427e+00 -2.07460672e-01 6.64518237e-01 -3.30535918e-01
-1.87302336e-01 5.54882921e-02 -3.02957296e-01 -2.58593131e-02
4.36319888e-01 -1.25657058e+00 6.02598727e-01 1.01509917e+00
3.47416699e-01 -1.82298981e-02 1.11094129e+00 -8.47454071e-01
-9.00876582e-01 -1.10552216e+00 7.46339619e-01 -2.90504575e-01
4.27534610e-01 -6.56852782e-01 -6.49507582e-01 5.34408629e-01
1.04185057e+00 -7.45472431e-01 3.06855410e-01 -5.19849360e-01
-3.08979541e-01 3.58211428e-01 -9.43985403e-01 1.06385446e+00
1.16309392e+00 -7.94730186e-02 -7.59821773e-01 3.38217527e-01
9.60795224e-01 -3.78753275e-01 -2.64098853e-01 -6.87937438e-02
3.66379648e-01 -9.47016239e-01 8.19533348e-01 -2.34641075e-01
1.11409771e+00 -5.98368824e-01 1.89927176e-01 -1.44133210e+00
-8.17912072e-02 -9.70864236e-01 5.41560538e-02 1.60037839e+00
2.78412610e-01 1.26952892e-02 7.29204476e-01 2.31801495e-02
-6.88736290e-02 -2.23088726e-01 -6.96495473e-01 -4.90037233e-01
-2.72802085e-01 -1.82437599e-01 6.09995425e-01 9.27750945e-01
3.08593035e-01 8.94662380e-01 -6.68343842e-01 -7.11873593e-03
8.17800641e-01 3.82241249e-01 8.99792373e-01 -7.27610648e-01
-2.90811986e-01 -3.99514765e-01 -7.21345767e-02 -5.58060288e-01
1.80100098e-01 -7.96322346e-01 4.16551441e-01 -1.97865474e+00
6.50296330e-01 -6.16778135e-02 4.31862235e-01 4.23600733e-01
-2.17474490e-01 2.11991444e-01 5.61583519e-01 1.65936574e-01
-7.93593168e-01 7.04332173e-01 1.36540055e+00 -3.00057948e-01
-2.72951275e-01 -5.46625376e-01 -8.28383982e-01 5.75551748e-01
6.46411717e-01 -5.75533271e-01 -7.46649325e-01 -4.65907544e-01
1.68912917e-01 3.90118659e-01 6.08814001e-01 -1.09365511e+00
2.04965293e-01 -2.62498856e-01 2.53481269e-01 -6.58643007e-01
3.45690459e-01 -3.90260607e-01 4.82819647e-01 2.20264778e-01
-6.02224827e-01 2.67079426e-03 4.75701466e-02 6.75189555e-01
-3.10311586e-01 -1.21269390e-01 5.29678464e-01 -5.17004728e-01
-6.72012806e-01 -5.85247166e-02 -5.42299271e-01 1.31649047e-01
1.38503385e+00 -9.68542695e-02 -5.71388960e-01 -7.45330811e-01
-7.01952517e-01 2.84303337e-01 4.00890589e-01 6.56169415e-01
9.44555640e-01 -1.60655427e+00 -1.15936291e+00 -1.74037144e-01
2.74435073e-01 1.55974865e-01 2.68282384e-01 9.00150165e-02
-4.62476373e-01 1.04475841e-01 -2.13002756e-01 -5.57378650e-01
-1.27366233e+00 6.04207039e-01 -1.26946671e-02 -4.25446689e-01
-8.85594308e-01 5.11776626e-01 8.04385841e-01 2.25863367e-01
1.44718274e-01 1.80221692e-01 -3.96307111e-01 2.07147375e-01
7.43697882e-01 5.64238951e-02 -5.84514856e-01 -5.98251760e-01
1.48381189e-01 3.99989456e-01 -6.79758340e-02 -8.58515441e-01
1.21478832e+00 -3.49199384e-01 9.10862610e-02 4.30797279e-01
8.47864449e-01 -2.76207745e-01 -1.28153586e+00 2.95884274e-02
-1.77993059e-01 -2.69657761e-01 -2.66606718e-01 -8.96654546e-01
-7.17389107e-01 4.91322190e-01 2.77861327e-01 3.70575488e-01
9.35516357e-01 3.34964365e-01 7.70445585e-01 -1.68109506e-01
1.43651709e-01 -9.13447201e-01 5.85538447e-01 5.64276390e-02
1.41052175e+00 -1.15745854e+00 -2.39035860e-02 -5.58433115e-01
-1.15235472e+00 8.41512382e-01 6.05668783e-01 -1.05342366e-01
-8.65745023e-02 1.64994061e-01 5.08941486e-02 -3.98320585e-01
-1.04500628e+00 -2.79636741e-01 2.04172969e-01 6.67403817e-01
1.53070554e-01 5.51918894e-02 -4.49754477e-01 5.88171482e-01
-2.47717828e-01 8.82712305e-02 9.17486906e-01 8.41289759e-01
-3.20359379e-01 -9.53301311e-01 -4.22313392e-01 -5.60857430e-02
-1.18225031e-01 1.47317001e-03 -6.62075460e-01 6.80901349e-01
4.09045303e-03 9.67463017e-01 1.00361183e-01 -3.02423805e-01
1.79201812e-01 5.34696952e-02 3.73020798e-01 -7.18414903e-01
-5.90912819e-01 2.82199055e-01 2.64889866e-01 -3.52183431e-01
-5.36466300e-01 -6.65989876e-01 -1.31864500e+00 9.12477355e-03
1.13233492e-01 7.54119530e-02 6.28373921e-01 8.99503708e-01
2.32883066e-01 8.11500609e-01 5.52873611e-01 -7.45743990e-01
8.91133845e-02 -7.51448631e-01 -3.01176906e-01 8.52510273e-01
2.24217787e-01 -2.67731160e-01 -5.44341728e-02 6.79840088e-01] | [11.182755470275879, 0.7346868515014648] |
0839b64d-6b12-46e6-8b9a-7d6d2b4a6543 | writing-by-memorizing-hierarchical-retrieval | 2106.06471 | null | https://arxiv.org/abs/2106.06471v1 | https://arxiv.org/pdf/2106.06471v1.pdf | Writing by Memorizing: Hierarchical Retrieval-based Medical Report Generation | Medical report generation is one of the most challenging tasks in medical image analysis. Although existing approaches have achieved promising results, they either require a predefined template database in order to retrieve sentences or ignore the hierarchical nature of medical report generation. To address these issues, we propose MedWriter that incorporates a novel hierarchical retrieval mechanism to automatically extract both report and sentence-level templates for clinically accurate report generation. MedWriter first employs the Visual-Language Retrieval~(VLR) module to retrieve the most relevant reports for the given images. To guarantee the logical coherence between sentences, the Language-Language Retrieval~(LLR) module is introduced to retrieve relevant sentences based on the previous generated description. At last, a language decoder fuses image features and features from retrieved reports and sentences to generate meaningful medical reports. We verified the effectiveness of our model by automatic evaluation and human evaluation on two datasets, i.e., Open-I and MIMIC-CXR. | ['Fenglong Ma', 'Quanzeng You', 'Muchao Ye', 'Xingyi Yang'] | 2021-05-25 | null | https://aclanthology.org/2021.acl-long.387 | https://aclanthology.org/2021.acl-long.387.pdf | acl-2021-5 | ['medical-report-generation'] | ['medical'] | [ 4.59333271e-01 -1.17787912e-01 -8.63147825e-02 -4.08965796e-01
-1.49955404e+00 -4.73950118e-01 4.88328844e-01 3.69852483e-01
-2.49573067e-01 6.24421537e-01 4.79322612e-01 -3.00700516e-01
2.05020830e-01 -5.33282042e-01 -2.39559740e-01 -3.49878341e-01
3.00187439e-01 1.86978295e-01 2.60836542e-01 1.85776740e-01
4.47651297e-01 2.69252211e-01 -1.31359816e+00 7.14614451e-01
8.64212692e-01 7.60779917e-01 7.41581440e-01 1.03658724e+00
-3.53231490e-01 1.07132649e+00 -8.24199855e-01 -3.18075031e-01
-2.54188180e-01 -9.70898390e-01 -7.93827355e-01 3.18233281e-01
4.90534008e-02 -3.59907150e-01 -4.87853199e-01 1.06893384e+00
8.43485355e-01 -3.63805741e-01 9.25567389e-01 -5.90934098e-01
-6.86268628e-01 4.58223283e-01 -6.62191033e-01 4.11574930e-01
9.19212759e-01 2.02400878e-01 5.84466338e-01 -9.15199935e-01
1.09885263e+00 9.83091414e-01 1.08666517e-01 5.31668723e-01
-1.03422427e+00 -6.35930181e-01 -1.11764282e-01 -1.23775266e-01
-1.78087211e+00 -5.82660437e-01 7.42845297e-01 -4.11445975e-01
9.25426602e-01 6.14694834e-01 3.77609640e-01 9.58194256e-01
7.95032740e-01 7.67729580e-01 1.06807446e+00 -4.82294589e-01
3.23170088e-02 4.27261919e-01 1.90798156e-02 8.77980947e-01
8.11094269e-02 -2.98076242e-01 -5.79060197e-01 -1.16045639e-01
6.31568491e-01 -1.81900814e-01 -5.48047900e-01 3.68499368e-01
-1.40662098e+00 7.51885235e-01 2.87012994e-01 4.41301018e-01
-4.12036717e-01 -1.50994539e-01 4.71726418e-01 1.46124199e-01
4.36455458e-01 3.19542378e-01 3.53229225e-01 3.24592382e-01
-1.18066907e+00 1.86373577e-01 5.98094642e-01 1.15782952e+00
1.63383514e-01 -3.47855240e-01 -1.00331616e+00 8.53622198e-01
4.51808870e-01 8.08645308e-01 7.54000187e-01 -5.67033231e-01
8.09378684e-01 5.16407847e-01 1.77580509e-02 -1.38826692e+00
-1.19954266e-01 -3.34363073e-01 -9.67961311e-01 -3.53377908e-01
-4.11014616e-01 1.82636604e-01 -1.00917673e+00 1.22305119e+00
6.33067265e-02 -3.16988170e-01 3.49050969e-01 9.49080229e-01
1.62324119e+00 9.65246320e-01 1.65386423e-01 -6.84604943e-01
1.44376159e+00 -7.49471068e-01 -1.15639424e+00 -6.40896941e-03
4.75010067e-01 -1.06062770e+00 9.00573552e-01 1.83090493e-02
-1.31522226e+00 -4.70402211e-01 -1.09540749e+00 -2.09176421e-01
1.91420298e-02 5.75496554e-01 -1.22050093e-02 8.86085853e-02
-1.21295512e+00 -1.32101685e-01 -4.08524424e-01 -3.45692486e-01
2.18603536e-01 1.37303904e-01 -4.07683641e-01 -1.48452193e-01
-1.01318312e+00 6.96675420e-01 3.46621960e-01 5.15192281e-04
-8.79231930e-01 -2.81212240e-01 -8.99575830e-01 -1.79807007e-01
1.00557327e-01 -9.68928933e-01 1.38114178e+00 -4.82220501e-01
-1.06037235e+00 1.24560511e+00 -3.53907108e-01 -1.60667300e-01
4.88956481e-01 2.73714542e-01 -5.08116901e-01 9.19174254e-01
4.13596272e-01 9.09093678e-01 9.49616849e-01 -1.34911740e+00
-4.28384185e-01 -1.53695434e-01 -3.28139454e-01 3.38176697e-01
-1.23423204e-01 3.01858246e-01 -9.72012758e-01 -1.03715050e+00
9.97031629e-02 -6.27946079e-01 -3.11586797e-01 -1.54175878e-01
-7.30676651e-01 -7.62780905e-02 3.30458701e-01 -9.29849625e-01
1.90102565e+00 -2.30656266e+00 -5.84033728e-02 2.36039430e-01
5.09663463e-01 5.32330237e-02 -3.04690927e-01 3.70900273e-01
3.51181068e-02 1.63836375e-01 -2.74274349e-01 -4.09871310e-01
-4.48520690e-01 -1.03763722e-01 -5.62061906e-01 1.68381989e-01
2.55627185e-01 9.52637970e-01 -7.05847919e-01 -1.46907270e+00
-5.89107163e-02 2.54738450e-01 -3.19193184e-01 5.17273009e-01
-1.40440494e-01 4.59258407e-01 -7.95484304e-01 5.45261621e-01
2.73980349e-01 -5.23622692e-01 1.52239650e-01 -2.54992902e-01
7.94760510e-02 2.84980088e-01 -6.89831495e-01 1.73184991e+00
-4.05792385e-01 3.39960605e-01 -2.56677330e-01 -3.80863339e-01
9.32861328e-01 7.69209921e-01 3.58291537e-01 -9.46594417e-01
6.99558258e-02 2.30253860e-01 -4.39562708e-01 -1.06477094e+00
5.49134493e-01 -1.77109838e-01 -3.05983424e-01 5.92094779e-01
-2.45952249e-01 -3.88825893e-01 1.34826526e-01 6.64353848e-01
1.24067700e+00 -2.57074475e-01 4.92542535e-01 1.94882765e-01
6.95361972e-01 2.45449319e-01 1.79442212e-01 9.61908698e-01
-1.27072344e-02 1.16194046e+00 3.17019552e-01 -1.48489073e-01
-7.89064884e-01 -1.12700236e+00 6.37347028e-02 4.91899490e-01
2.74880201e-01 -6.26685917e-01 -6.63519800e-01 -7.96755791e-01
-5.28712511e-01 7.87272871e-01 -4.26411688e-01 -1.82978928e-01
-3.76356840e-01 -5.42490482e-01 5.32743633e-01 3.06289762e-01
2.65011996e-01 -1.33129263e+00 -6.43637538e-01 4.18779969e-01
-6.08274102e-01 -1.14400816e+00 -9.19714391e-01 -3.98688674e-01
-6.48378193e-01 -8.05247545e-01 -9.19554710e-01 -8.75404298e-01
1.13326848e+00 1.18405156e-01 1.14854085e+00 2.46687263e-01
-7.05312014e-01 3.34572583e-01 -5.30735791e-01 -1.52793512e-01
-9.88646924e-01 -1.06619317e-02 -2.96570182e-01 -1.64182156e-01
-1.21520124e-01 -9.06800255e-02 -6.87778950e-01 -3.05289596e-01
-1.20478821e+00 4.91029203e-01 1.01877725e+00 6.53577626e-01
8.27997804e-01 -1.91104040e-01 4.75177497e-01 -6.53756917e-01
1.19193101e+00 -1.33148283e-01 -3.75205070e-01 6.13067091e-01
-4.35203910e-01 2.81406939e-01 4.92427438e-01 -3.16009611e-01
-1.03725708e+00 2.12127268e-02 -2.64130563e-01 -3.60762596e-01
1.07421853e-01 6.71453297e-01 3.54487687e-01 4.11067188e-01
4.99956995e-01 8.60606909e-01 -9.11801159e-02 -2.44208977e-01
1.46301389e-01 1.01409006e+00 5.84407806e-01 -9.47477147e-02
5.61858237e-01 2.83530205e-01 -4.11201119e-01 -5.15471578e-01
-8.82411480e-01 -4.43127185e-01 -2.53282011e-01 -4.91730124e-01
1.15669179e+00 -9.26459253e-01 -4.93021429e-01 -6.02849983e-02
-1.58216894e+00 4.92438287e-01 1.21925038e-03 5.04499614e-01
-4.45140779e-01 3.21979076e-01 -8.98463428e-01 -6.92296326e-01
-9.76807594e-01 -1.47291338e+00 1.33695340e+00 1.88691556e-01
-3.95765632e-01 -5.08068085e-01 8.09198394e-02 3.11710685e-01
1.45208865e-01 1.05814248e-01 1.00609469e+00 -3.04559648e-01
-6.21594071e-01 -3.17291945e-01 -4.88108069e-01 -1.01356134e-02
2.35023856e-01 -6.26109019e-02 -7.35822856e-01 -1.70740634e-01
8.75305608e-02 -3.26441646e-01 8.83078158e-01 1.76848054e-01
1.18390870e+00 -4.80797917e-01 -6.18259788e-01 1.31251574e-01
1.31579673e+00 4.51842397e-01 6.64268076e-01 -2.50430882e-01
3.10194105e-01 3.25716496e-01 6.50029480e-01 3.76689702e-01
4.61434871e-01 5.00066400e-01 -2.74430245e-01 -1.10421099e-01
-3.65226239e-01 -3.68290573e-01 3.76970589e-01 1.55138338e+00
3.68341833e-01 -2.74908155e-01 -8.73222053e-01 4.71813858e-01
-1.51308751e+00 -7.88720489e-01 1.11688703e-01 1.91350055e+00
1.31537700e+00 1.33221000e-01 -2.77303308e-01 -2.14102000e-01
5.01018286e-01 1.69942111e-01 -1.44415364e-01 -2.63993770e-01
2.24815115e-01 1.03270851e-01 1.65115427e-02 5.51193953e-01
-7.95549333e-01 6.97024107e-01 6.86449242e+00 9.33687687e-01
-1.18792439e+00 1.70455933e-01 7.11956501e-01 2.16469839e-02
-5.52542627e-01 -4.33469892e-01 -7.94307947e-01 3.81387293e-01
6.15306020e-01 -4.15434927e-01 -2.47467935e-01 4.96687502e-01
4.66573596e-01 -3.43695641e-01 -8.65040541e-01 1.10993302e+00
6.04461849e-01 -1.60446906e+00 6.94433510e-01 -2.21489504e-01
4.38874781e-01 -4.20978487e-01 4.96435128e-02 1.06752068e-01
-1.21714674e-01 -9.25676286e-01 6.30891860e-01 1.05480349e+00
1.27399158e+00 -5.02744079e-01 7.36239970e-01 3.15729141e-01
-1.13871968e+00 3.42422724e-01 -2.26667985e-01 4.51452374e-01
9.89591479e-02 5.70028543e-01 -1.28395009e+00 8.36112082e-01
3.16826671e-01 5.07364213e-01 -9.15582240e-01 9.13244903e-01
-1.11309983e-01 3.74244422e-01 1.20917842e-01 -2.33829364e-01
9.68464464e-03 1.41472265e-01 5.82108736e-01 1.38758469e+00
3.07484806e-01 2.66984195e-01 2.85730481e-01 9.57043588e-01
-1.20078586e-01 4.82704967e-01 -8.58399093e-01 -2.84989327e-01
1.79909155e-01 1.20644557e+00 -9.44081426e-01 -5.79323888e-01
-3.19434106e-01 1.12627649e+00 6.29671514e-02 2.43076727e-01
-8.15023184e-01 -5.35450697e-01 -2.14913994e-01 1.44747362e-01
1.16572082e-02 -4.50202264e-02 -1.30211830e-01 -1.14198375e+00
3.43395054e-01 -8.67905080e-01 4.26548392e-01 -1.12237442e+00
-1.19453108e+00 9.54110742e-01 -5.71532510e-02 -1.66180360e+00
-4.82050449e-01 -6.35046735e-02 -2.13437691e-01 8.18724096e-01
-1.31550956e+00 -9.92477715e-01 -1.07768022e-01 4.69749719e-01
7.86980093e-01 -1.25717938e-01 9.82073307e-01 2.73594171e-01
-4.69386846e-01 4.41464275e-01 -4.40798283e-01 2.51072377e-01
7.46885538e-01 -8.81214201e-01 -2.43238974e-02 7.14149594e-01
5.02895042e-02 7.60074735e-01 5.08941472e-01 -9.41568851e-01
-1.20139658e+00 -1.11895180e+00 1.35807049e+00 -2.55443066e-01
2.87578225e-01 -2.99205512e-01 -7.19859779e-01 2.99870998e-01
3.94277513e-01 -2.32566953e-01 6.87831819e-01 -8.52530658e-01
-1.27390534e-01 -3.08594525e-01 -9.49519753e-01 9.10919011e-01
7.43664503e-01 -6.56017959e-01 -7.39277422e-01 4.58136439e-01
1.05356896e+00 -4.79348779e-01 -7.46086001e-01 4.09152180e-01
2.78062761e-01 -5.37552953e-01 7.58958757e-01 -1.62899241e-01
8.74682724e-01 -5.57643652e-01 -1.12620452e-02 -7.42464185e-01
8.54305476e-02 -3.13546240e-01 1.97153002e-01 1.17815769e+00
6.63595557e-01 -1.60901859e-01 6.99500367e-03 2.42497176e-01
6.87373504e-02 -7.97653913e-01 -7.20437646e-01 -1.84980720e-01
-4.91215527e-01 -5.47434270e-01 2.04651773e-01 5.70886314e-01
1.60264477e-01 4.83339489e-01 -2.24747539e-01 1.05401419e-01
3.28301668e-01 4.43348229e-01 3.71827990e-01 -5.17022669e-01
-2.01969996e-01 -2.47375295e-01 -1.58201963e-01 -8.57587457e-01
-4.91966084e-02 -1.11667836e+00 4.42664981e-01 -1.98215365e+00
8.59569192e-01 -1.48176089e-01 -7.07231909e-02 3.46678436e-01
-4.35377121e-01 4.93012905e-01 2.34846681e-01 4.69506860e-01
-1.05460453e+00 3.21185112e-01 1.62900591e+00 -4.18241233e-01
-1.79787427e-01 -1.19911216e-01 -7.81957150e-01 4.41324770e-01
4.13368851e-01 -6.32179379e-01 -4.10432637e-01 -2.58776128e-01
1.68459088e-01 7.46080637e-01 2.28625327e-01 -9.28499341e-01
3.53120625e-01 -6.56678528e-02 5.10619640e-01 -9.96778429e-01
6.03127340e-03 -4.40571219e-01 2.92200577e-02 5.45593739e-01
-7.55961359e-01 3.54209989e-01 1.14043213e-01 3.52978408e-01
-4.99769211e-01 -1.94516763e-01 5.29914081e-01 -3.90341163e-01
-2.99796671e-01 1.75126389e-01 -6.02734745e-01 -1.09051228e-01
1.00363111e+00 4.61380072e-02 -1.92937031e-01 -4.86424029e-01
-6.99666381e-01 2.53229320e-01 1.09780863e-01 5.40960193e-01
1.43463922e+00 -1.21109414e+00 -9.51722801e-01 1.51257813e-01
5.26500225e-01 -8.37837607e-02 2.71028876e-01 8.76278400e-01
-7.74944484e-01 5.72042644e-01 3.43637735e-01 -6.34795249e-01
-1.50487709e+00 5.81217766e-01 7.33898431e-02 -7.21466362e-01
-8.37960780e-01 5.49175918e-01 2.98541576e-01 -5.44808954e-02
6.38578311e-02 -3.21245134e-01 -3.36501569e-01 -1.44696906e-01
9.22460675e-01 -5.06169081e-01 3.19767930e-02 -6.03540778e-01
-3.92411470e-01 4.78110433e-01 -4.80720550e-01 -4.57057983e-01
9.82173443e-01 -3.43861789e-01 -2.72211969e-01 4.48183060e-01
1.15576315e+00 1.08061068e-01 -4.34894472e-01 -4.52848412e-02
1.58735388e-03 -1.63278505e-01 -1.39886960e-01 -8.73228073e-01
-7.77898788e-01 5.58688223e-01 4.40884441e-01 -1.62926152e-01
1.32269239e+00 3.17964345e-01 7.22818494e-01 2.15161845e-01
3.26730132e-01 -8.71405184e-01 4.15710926e-01 5.30722812e-02
1.27163541e+00 -9.91041481e-01 1.80479318e-01 -4.52085376e-01
-8.43136311e-01 9.42103624e-01 3.66012484e-01 3.44018579e-01
4.10818309e-01 3.62360269e-01 3.83102298e-01 -3.37715358e-01
-8.63840520e-01 -9.35189128e-02 5.05488276e-01 1.79899290e-01
7.42776573e-01 -1.17873810e-01 -7.17501462e-01 4.87005830e-01
-9.21456441e-02 3.55993569e-01 4.83402461e-01 1.08745241e+00
-2.24162653e-01 -7.64469385e-01 -5.25406599e-01 5.75106919e-01
-6.31587148e-01 -1.86427474e-01 -2.70288855e-01 3.00317109e-01
-1.32012814e-01 1.11542094e+00 -1.06630258e-01 -3.35774481e-01
2.01434076e-01 -5.84789254e-02 3.91260773e-01 -9.36728776e-01
-5.35338163e-01 2.44586453e-01 -9.71328244e-02 -4.54840779e-01
-4.71092492e-01 -3.83641422e-01 -1.33683562e+00 2.53413409e-01
-9.30131972e-02 1.62399948e-01 3.96550238e-01 8.77979517e-01
3.64349693e-01 9.19275761e-01 5.98260522e-01 -2.03028947e-01
-3.01410198e-01 -8.76187682e-01 -4.14658904e-01 4.80514526e-01
2.69363254e-01 -2.04470102e-02 1.18414715e-01 5.95455050e-01] | [15.050836563110352, -1.3898720741271973] |
e68ec98a-1b49-4dd9-9f75-6b232d8f56f9 | heterogeneous-knowledge-transfer-in-video | 1511.04798 | null | http://arxiv.org/abs/1511.04798v2 | http://arxiv.org/pdf/1511.04798v2.pdf | Heterogeneous Knowledge Transfer in Video Emotion Recognition, Attribution and Summarization | Emotion is a key element in user-generated videos. However, it is difficult
to understand emotions conveyed in such videos due to the complex and
unstructured nature of user-generated content and the sparsity of video frames
expressing emotion. In this paper, for the first time, we study the problem of
transferring knowledge from heterogeneous external sources, including image and
textual data, to facilitate three related tasks in understanding video emotion:
emotion recognition, emotion attribution and emotion-oriented summarization.
Specifically, our framework (1) learns a video encoding from an auxiliary
emotional image dataset in order to improve supervised video emotion
recognition, and (2) transfers knowledge from an auxiliary textual corpora for
zero-shot recognition of emotion classes unseen during training. The proposed
technique for knowledge transfer facilitates novel applications of emotion
attribution and emotion-oriented summarization. A comprehensive set of
experiments on multiple datasets demonstrate the effectiveness of our
framework. | ['Yu-Gang Jiang', 'Yanwei Fu', 'Leonid Sigal', 'Boyang Li', 'Baohan Xu'] | 2015-11-16 | null | null | null | null | ['video-emotion-recognition'] | ['computer-vision'] | [ 6.24573708e-01 -5.96533865e-02 -4.11761224e-01 -4.42329824e-01
-6.53708756e-01 -4.74556804e-01 1.55882016e-01 -2.72109713e-02
-1.70681044e-01 7.39049375e-01 7.26689458e-01 4.15293515e-01
4.27705884e-01 -2.78831214e-01 -8.42793763e-01 -6.45611763e-01
-1.14081465e-01 -1.68779463e-01 -3.42364907e-01 1.79358870e-01
2.98415810e-01 1.09895788e-01 -1.81984115e+00 7.23844647e-01
7.80766368e-01 1.47544146e+00 1.27176598e-01 7.53870487e-01
-1.27056435e-01 1.59493387e+00 -6.64142072e-01 -2.53427923e-01
-2.52686918e-01 -7.86776304e-01 -8.36110234e-01 7.92798936e-01
1.12336256e-01 -3.16564560e-01 -2.18721747e-01 1.04693937e+00
4.35768634e-01 3.80780548e-01 7.13672698e-01 -1.50284457e+00
-7.62636900e-01 3.53817850e-01 -4.89842385e-01 1.67540997e-01
6.08153522e-01 -1.00105964e-01 7.56805062e-01 -9.17328537e-01
1.05754125e+00 9.98623133e-01 2.10878462e-01 6.99795008e-01
-6.39772952e-01 -5.87328851e-01 2.16392055e-01 3.77664357e-01
-1.13465321e+00 -6.61439121e-01 1.13424337e+00 -5.25316656e-01
6.68888748e-01 2.37713218e-01 7.15907097e-01 1.49236178e+00
-1.60061002e-01 1.33665276e+00 8.25886607e-01 -2.12068155e-01
3.52608323e-01 3.42912227e-01 -4.55296524e-02 6.09006941e-01
-3.61161590e-01 -4.65890855e-01 -9.56621230e-01 -9.97025594e-02
5.70098460e-01 6.23206496e-02 -6.82932794e-01 -1.48587599e-01
-1.18554950e+00 8.06985378e-01 -3.20530720e-02 1.27260596e-01
-7.52421916e-01 6.90907091e-02 9.84236300e-01 4.15835410e-01
9.31929886e-01 2.80344993e-01 -4.78893876e-01 -5.42583823e-01
-7.93561101e-01 -3.12853783e-01 8.41512561e-01 1.11342001e+00
7.64205277e-01 2.79168427e-01 -3.02587133e-02 8.38193953e-01
-6.07614145e-02 4.95437026e-01 7.22609103e-01 -1.08435512e+00
4.99420434e-01 3.98088247e-01 -3.09791300e-03 -1.47637212e+00
2.38813497e-02 6.55855387e-02 -9.80338395e-01 -4.61833417e-01
-3.33311796e-01 -5.47622502e-01 -5.95147550e-01 1.70885313e+00
1.55354738e-01 5.46270072e-01 4.68535244e-01 9.74679410e-01
1.20052421e+00 1.19525993e+00 1.77647606e-01 -6.99010670e-01
1.20294642e+00 -1.16500843e+00 -9.78488386e-01 -6.75233006e-02
5.96911430e-01 -5.44704020e-01 8.35497797e-01 3.24068397e-01
-9.51501131e-01 -4.89435613e-01 -6.86002791e-01 3.54670957e-02
-1.01836890e-01 3.87996078e-01 4.58148122e-01 1.55301630e-01
-6.67898059e-01 5.01665100e-02 -5.54449975e-01 -4.69034910e-01
7.10439086e-01 1.80613577e-01 -6.88372672e-01 -4.54801023e-02
-1.13910556e+00 3.51356179e-01 3.89823377e-01 3.90847623e-02
-9.97631848e-01 -5.88903368e-01 -1.18381989e+00 3.78195271e-02
5.65794468e-01 -4.46733505e-01 1.12113988e+00 -2.33789039e+00
-1.55954111e+00 7.48898149e-01 -4.11002338e-01 -2.48420864e-01
1.58332288e-02 -3.83340150e-01 -3.99976701e-01 8.99253428e-01
-4.30656876e-03 6.66543484e-01 1.22898412e+00 -1.28893268e+00
-5.62431633e-01 -3.47251385e-01 -6.49484098e-02 3.99416983e-01
-7.99389124e-01 1.33253172e-01 -5.68485081e-01 -8.46898735e-01
-4.79440719e-01 -6.61741793e-01 -7.43704662e-02 -3.46231580e-01
-1.90321162e-01 3.75654846e-02 1.05727351e+00 -7.71626472e-01
1.12409592e+00 -2.25316834e+00 5.02191901e-01 -5.03658392e-02
4.57148775e-02 -1.24227174e-01 -3.94778371e-01 3.37599069e-01
-2.58635759e-01 2.84956079e-02 -3.30664329e-02 -1.73281118e-01
-2.24873856e-01 2.54786044e-01 -5.74962437e-01 1.63949519e-01
4.87991244e-01 8.51997912e-01 -1.08959007e+00 -6.95227861e-01
-4.92389537e-02 5.67956030e-01 -5.64448476e-01 6.12781882e-01
-1.82125196e-01 6.21967256e-01 -8.21465731e-01 7.85017133e-01
1.16688937e-01 -3.27563822e-01 1.25526637e-02 -6.19850516e-01
7.58430064e-02 -3.87996197e-01 -8.69106650e-01 1.75458658e+00
-4.98512447e-01 7.88301766e-01 6.64360076e-02 -1.33686638e+00
6.37092352e-01 8.25758517e-01 8.26236308e-01 -4.19935614e-01
4.02359515e-01 -3.61736625e-01 -7.25262761e-01 -1.12284362e+00
7.28850126e-01 -2.16749132e-01 -3.62165481e-01 6.23650551e-01
5.19061565e-01 6.77658319e-02 1.44366607e-01 4.39038903e-01
8.44734848e-01 4.11834046e-02 4.72168803e-01 1.81267634e-01
5.88734448e-01 -7.00940751e-03 6.55395448e-01 3.92836750e-01
-1.83510616e-01 5.39432704e-01 7.09414184e-01 -4.48624879e-01
-6.64783835e-01 -4.12497997e-01 3.09724450e-01 1.32632136e+00
1.90228522e-01 -4.75919455e-01 -6.29968882e-01 -7.82149255e-01
-4.53432620e-01 3.88690174e-01 -7.72966325e-01 -3.20010036e-01
-6.78802207e-02 -3.95351291e-01 2.99642146e-01 5.61785102e-01
3.49952698e-01 -1.43211412e+00 -7.16799617e-01 1.46784382e-02
-7.97910750e-01 -1.47162426e+00 -4.31984544e-01 -1.15812987e-01
-6.39836788e-01 -1.03353715e+00 -5.97149014e-01 -8.44506145e-01
9.86888945e-01 3.27983767e-01 1.15181911e+00 9.08271670e-02
-2.63219953e-01 1.10291982e+00 -9.27122355e-01 -4.06108677e-01
-2.67719567e-01 -2.21180573e-01 -3.63890529e-02 6.52573884e-01
1.92714140e-01 -3.19354802e-01 -4.00163352e-01 3.77671719e-02
-1.24493957e+00 1.02810286e-01 3.77569079e-01 8.53979945e-01
5.61903417e-01 2.90308923e-01 7.78178096e-01 -8.96939337e-01
6.33999109e-01 -7.70391583e-01 7.61469379e-02 2.11678445e-01
2.61995643e-01 -3.64504635e-01 6.43141270e-01 -5.78048348e-01
-1.41929352e+00 2.15580881e-01 1.58266932e-01 -7.36919999e-01
-2.46534854e-01 6.70414925e-01 -3.48552823e-01 9.82844532e-02
2.02755183e-01 3.67695689e-01 -8.94603953e-02 7.80730322e-02
4.90691811e-01 8.32946777e-01 4.84573960e-01 -6.60036266e-01
2.08132684e-01 5.52745283e-01 -5.04101455e-01 -1.18533599e+00
-1.00309646e+00 -4.88953859e-01 -4.19285089e-01 -6.97766006e-01
1.13640559e+00 -1.26084280e+00 -5.95557153e-01 2.81164765e-01
-1.13968706e+00 -5.51506877e-02 -3.12911898e-01 4.38229799e-01
-8.89292419e-01 4.98379827e-01 -7.22554266e-01 -7.57561684e-01
-4.41728979e-01 -1.03622508e+00 1.05379820e+00 2.96865582e-01
-2.00680181e-01 -9.58562374e-01 -1.70920029e-01 6.09498739e-01
7.34962225e-02 3.81178975e-01 6.33213103e-01 -5.44887424e-01
-2.44831786e-01 -2.96571374e-01 -2.01409101e-01 6.51736259e-01
2.03844830e-01 1.58627960e-03 -8.40479195e-01 -1.20880818e-02
1.80123881e-01 -1.24971867e+00 7.82368064e-01 3.13905746e-01
1.42139959e+00 -5.04941046e-01 9.43048596e-02 1.92530036e-01
1.28932881e+00 1.76000133e-01 6.25564098e-01 -4.82077263e-02
6.94845319e-01 7.80984938e-01 9.06751513e-01 9.31906819e-01
4.23367977e-01 1.46193445e-01 3.89577985e-01 -1.14830613e-01
3.88353467e-01 -9.55298766e-02 6.67727053e-01 1.24463868e+00
-3.12574685e-01 -5.07133067e-01 -4.16299701e-01 5.44121563e-01
-2.07596564e+00 -1.44430578e+00 3.80143493e-01 1.64804995e+00
7.68591285e-01 -4.50346172e-01 -1.02250166e-01 -1.84707001e-01
7.61510134e-01 4.39117610e-01 -6.10612094e-01 -3.99832398e-01
-1.37519300e-01 -1.62686706e-01 -9.81702581e-02 8.80032778e-03
-1.24517202e+00 9.44144666e-01 5.63395739e+00 5.60753286e-01
-1.24874473e+00 1.89752489e-01 9.06432748e-01 -2.48941407e-01
-7.25564286e-02 -2.36731544e-01 -1.42098412e-01 3.64554077e-01
8.39882910e-01 -4.20895904e-01 2.40161389e-01 9.70022082e-01
1.79871723e-01 -4.40104790e-02 -1.02582514e+00 1.20901740e+00
9.15333569e-01 -1.23601007e+00 4.22901422e-01 -5.16381145e-01
9.87166345e-01 -3.60278398e-01 -1.09856583e-01 4.41307783e-01
-3.20397735e-01 -6.77572608e-01 5.99079370e-01 6.23998404e-01
6.36876762e-01 -9.00427699e-01 7.46856868e-01 1.83350831e-01
-1.13134515e+00 -5.74098602e-02 -2.49187082e-01 -5.50047718e-02
3.91953617e-01 4.63727057e-01 -3.10436159e-01 4.73043919e-01
6.76990688e-01 1.34056818e+00 -1.32539064e-01 4.58287716e-01
-1.57870784e-01 6.50528908e-01 1.45684928e-01 1.70702770e-01
2.01770276e-01 -1.82877585e-01 3.93353611e-01 1.38283491e+00
2.91812718e-01 5.85219264e-01 2.84147292e-01 3.42396170e-01
-5.41001260e-01 4.07655418e-01 -7.01342523e-01 -5.80664158e-01
9.49915349e-02 1.40152228e+00 -5.64095676e-01 -6.78377688e-01
-6.07708216e-01 1.32528925e+00 2.14861125e-01 6.18982673e-01
-1.01726937e+00 -2.41198525e-01 4.52186197e-01 -5.75259447e-01
3.97795856e-01 1.97614774e-01 5.18687248e-01 -1.75303936e+00
-3.56254689e-02 -1.03488290e+00 4.13879395e-01 -1.14313376e+00
-1.24494898e+00 7.04213619e-01 -3.78750354e-01 -1.30742085e+00
-2.85342485e-01 -3.33986402e-01 -4.85063076e-01 3.26634869e-02
-1.38832259e+00 -9.95075822e-01 -5.53095102e-01 9.98456001e-01
1.12711382e+00 -1.76851049e-01 8.77074957e-01 3.01042438e-01
-6.24427557e-01 2.03458995e-01 -8.78425464e-02 1.83046982e-01
1.00569952e+00 -8.48594308e-01 -7.13571191e-01 6.58940613e-01
2.93025859e-02 1.03131950e-01 6.28236711e-01 -6.48866117e-01
-1.84994781e+00 -1.36965632e+00 3.93571556e-01 -3.38355973e-02
7.51183093e-01 -1.02666371e-01 -7.21683741e-01 8.40243936e-01
4.59483504e-01 8.84128436e-02 1.28882456e+00 -1.52366042e-01
-2.30204970e-01 1.06505260e-01 -8.37799489e-01 4.08030838e-01
7.99294949e-01 -6.64090753e-01 -6.39057398e-01 3.39012265e-01
7.40768254e-01 -5.53969815e-02 -1.06122911e+00 4.09192502e-01
5.58157742e-01 -6.98062301e-01 7.52777040e-01 -1.03669429e+00
1.16158462e+00 9.93314832e-02 -3.38803411e-01 -1.44846833e+00
1.09448813e-01 -4.68282908e-01 -2.80375361e-01 1.49382341e+00
6.80102929e-02 -3.36686596e-02 6.47678375e-01 5.65917552e-01
-5.34638502e-02 -5.18228710e-01 -3.69103163e-01 -1.92026392e-01
-6.69709206e-01 -4.80681360e-01 7.99592584e-02 1.28545237e+00
2.44110614e-01 5.42709529e-01 -1.03564596e+00 5.40704317e-02
4.08102006e-01 4.02870893e-01 8.21958244e-01 -8.19686234e-01
-5.32549545e-02 3.67151177e-03 -6.33425117e-01 -9.28785384e-01
7.80378640e-01 -5.70584774e-01 1.99829414e-01 -1.27661920e+00
6.36722028e-01 2.70424694e-01 -3.46884757e-01 3.47706825e-01
-7.26775154e-02 4.86257464e-01 2.42306471e-01 5.68037890e-02
-1.39811456e+00 9.61215019e-01 1.22122598e+00 -1.73914403e-01
-4.54689339e-02 -4.25570011e-01 -7.80124843e-01 9.84436870e-01
5.62488198e-01 -4.12749767e-01 -6.74727023e-01 -3.36051852e-01
2.99670070e-01 5.37365675e-01 2.63858795e-01 -5.79219282e-01
7.04244748e-02 -3.42422098e-01 5.02620161e-01 -5.63062727e-01
3.79548907e-01 -9.85391557e-01 -1.11024901e-01 -5.33093289e-02
-4.99008179e-01 -9.75211263e-02 1.52519539e-01 8.13899934e-01
-8.35453808e-01 -1.26524925e-01 3.61720711e-01 -1.99767455e-01
-1.15349507e+00 4.38575149e-01 -7.41458297e-01 2.90441126e-01
1.15376568e+00 -1.61693811e-01 -1.89788952e-01 -8.89030516e-01
-9.39279318e-01 2.41671816e-01 3.74638110e-01 6.06436312e-01
9.83597875e-01 -1.44331598e+00 -5.82989872e-01 1.02933086e-02
4.13502753e-01 -3.20685923e-01 7.63646364e-01 8.05154860e-01
-1.06351301e-01 9.87205654e-03 -4.63237554e-01 -3.98110271e-01
-1.51806653e+00 5.90915978e-01 -2.30580270e-01 1.17952771e-01
-3.79696697e-01 6.38943553e-01 5.56214452e-01 1.06586210e-01
3.44035000e-01 1.32648706e-01 -4.16536719e-01 4.29979235e-01
9.02269065e-01 2.54942626e-02 -4.36323702e-01 -1.07627738e+00
-3.35027911e-02 5.41904807e-01 3.62927318e-02 1.97270647e-01
1.52510548e+00 -5.13407052e-01 -2.48647824e-01 6.45739853e-01
1.54009032e+00 -2.56951004e-01 -1.10247660e+00 -1.35379061e-01
-1.97088957e-01 -4.33948189e-01 -1.69488210e-02 -3.78239959e-01
-1.36599660e+00 7.61092305e-01 1.42923325e-01 -2.83004623e-02
1.68578422e+00 4.51535694e-02 9.70416069e-01 5.76450109e-01
8.04365352e-02 -1.30501282e+00 6.77179039e-01 3.82903427e-01
9.79969621e-01 -1.37966907e+00 -1.71574280e-01 -2.93130249e-01
-1.42587030e+00 1.10647619e+00 6.11922443e-01 1.29100755e-01
4.98781830e-01 -4.03461084e-02 1.22640103e-01 -2.15098143e-01
-1.05489433e+00 -1.25997029e-02 2.67026544e-01 3.50784034e-01
3.92279655e-01 -1.44721776e-01 -7.91640859e-03 8.70663047e-01
2.59140998e-01 4.06707972e-01 6.11362219e-01 1.07183003e+00
-3.77526939e-01 -6.57190263e-01 -1.97427690e-01 4.29738075e-01
-6.89412594e-01 8.54973793e-02 -4.40238267e-01 3.76962572e-01
-6.53759241e-02 9.01314557e-01 -1.13185951e-02 -2.92354882e-01
1.19732179e-01 1.48521617e-01 3.04814667e-01 -5.20016849e-01
-3.97765398e-01 1.63011000e-01 2.08656311e-01 -7.95882165e-01
-1.04916000e+00 -5.37031054e-01 -1.25962543e+00 3.55839312e-01
-5.88961132e-02 4.51352865e-01 5.33335090e-01 9.27489579e-01
6.07932866e-01 4.54565018e-01 8.72983694e-01 -1.07642019e+00
6.35714233e-02 -6.92151666e-01 -6.28243625e-01 9.92122591e-01
2.30153903e-01 -4.49134558e-01 -3.84623468e-01 9.23340499e-01] | [13.151470184326172, 4.844359874725342] |
1ddc8e17-8278-4a6c-8d55-a9bc95211a88 | adapting-language-audio-models-as-few-shot | 2305.17719 | null | https://arxiv.org/abs/2305.17719v1 | https://arxiv.org/pdf/2305.17719v1.pdf | Adapting Language-Audio Models as Few-Shot Audio Learners | We presented the Treff adapter, a training-efficient adapter for CLAP, to boost zero-shot classification performance by making use of a small set of labelled data. Specifically, we designed CALM to retrieve the probability distribution of text-audio clips over classes using a set of audio-label pairs and combined it with CLAP's zero-shot classification results. Furthermore, we designed a training-free version of the Treff adapter by using CALM as a cosine similarity measure. Experiments showed that the proposed Treff adapter is comparable and even better than fully-supervised methods and adaptation methods in low-shot and data-abundant scenarios. While the Treff adapter shows that combining large-scale pretraining and rapid learning of domain-specific knowledge is non-trivial for obtaining generic representations for few-shot learning, it is still limited to audio classification tasks. In the future, we will explore how to use audio-language models in diverse audio domains. | ['Wenwu Wang', 'Mark D. Plumbley', 'Emmanouil Benetos', 'Huy Phan', 'Haohe Liu', 'Xubo Liu', 'Jinhua Liang'] | 2023-05-28 | null | null | null | null | ['audio-classification'] | ['audio'] | [ 2.78482974e-01 -2.55112827e-01 -1.84564367e-01 -5.31691611e-01
-1.45160890e+00 -4.70786721e-01 3.69165152e-01 3.36890399e-01
-3.78684670e-01 4.22766626e-01 4.02916700e-01 2.92905211e-01
-2.47959375e-01 -5.58091938e-01 -4.62071061e-01 -4.51772511e-01
-2.68377513e-01 6.07635081e-01 6.87645912e-01 -2.00707540e-01
9.68387425e-02 -1.01544097e-01 -2.20982552e+00 6.60725713e-01
1.69962913e-01 1.32700181e+00 1.96473151e-01 8.94274712e-01
-5.96660301e-02 7.88055420e-01 -3.86406273e-01 -1.42286673e-01
-5.25677344e-03 -5.91970921e-01 -8.52782130e-01 -1.71849489e-01
4.70871687e-01 4.17895149e-03 -2.11353786e-02 5.96898735e-01
8.80717874e-01 7.35855043e-01 8.21229458e-01 -1.23629701e+00
-2.38054499e-01 6.39288187e-01 -2.47482732e-01 3.99268478e-01
6.62313938e-01 -2.50142187e-01 1.26050377e+00 -1.16532743e+00
5.42395055e-01 1.05639517e+00 9.49531138e-01 4.91650701e-01
-9.65331733e-01 -8.76549661e-01 -2.80541241e-01 6.30892038e-01
-1.66232109e+00 -7.86409855e-01 5.75222194e-01 -3.41541082e-01
1.06881571e+00 2.07189128e-01 4.75109756e-01 1.19433117e+00
-1.61144137e-01 7.45271206e-01 7.18098879e-01 -8.64611983e-01
6.72966719e-01 1.00551240e-01 2.00981513e-01 3.65832269e-01
-6.48902714e-01 4.00739983e-02 -9.66708958e-01 -4.33885306e-01
1.07573420e-01 -7.34225512e-02 -1.08634770e-01 -3.85578513e-01
-9.70667362e-01 8.72039318e-01 2.64850352e-02 4.36017722e-01
-7.24096596e-02 -3.09048630e-02 8.25915337e-01 6.56475186e-01
6.10012591e-01 5.03864110e-01 -3.78582835e-01 -6.38453841e-01
-1.34624588e+00 -1.22616219e-03 8.83271337e-01 1.08720028e+00
7.11421311e-01 -1.48777990e-02 -3.56033415e-01 1.27848208e+00
-1.06972113e-01 3.77908736e-01 7.65089691e-01 -1.10531902e+00
2.60220557e-01 -1.48957577e-02 -2.13696748e-01 -6.23741448e-01
-2.22136095e-01 -2.04961643e-01 -4.74665582e-01 -2.02216431e-01
5.97779192e-02 -4.21460941e-02 -8.67440164e-01 1.73323357e+00
1.52595654e-01 7.79618084e-01 -3.17340940e-02 7.80566216e-01
6.91618562e-01 7.71412432e-01 6.52253255e-02 -2.80518085e-01
1.37413812e+00 -1.02846813e+00 -5.23987234e-01 -4.11221199e-02
7.82157362e-01 -8.02754104e-01 1.37978375e+00 3.64323109e-01
-9.79083359e-01 -6.14073277e-01 -1.07897663e+00 1.11565128e-01
-4.85184431e-01 -3.01702291e-01 4.39164311e-01 6.11918271e-01
-7.34515011e-01 7.42936969e-01 -5.01759171e-01 -7.24605143e-01
3.60994250e-01 -6.23247139e-02 -3.50573838e-01 -3.44830543e-01
-1.54102063e+00 5.36620021e-01 4.69360739e-01 -7.45094776e-01
-1.42844462e+00 -8.25937629e-01 -8.56493831e-01 4.23710495e-01
4.29763645e-01 -2.67100364e-01 1.58470166e+00 -8.10024321e-01
-1.60475290e+00 6.07768953e-01 6.41773175e-03 -5.60407996e-01
-2.17260420e-01 -1.51157185e-01 -6.67263806e-01 4.87082958e-01
-3.75339296e-03 6.86683416e-01 1.07731664e+00 -5.06476820e-01
-6.45566225e-01 -1.12597123e-01 -1.12593845e-01 1.12643294e-01
-9.00381804e-01 2.86103070e-01 -4.13043350e-01 -6.05924487e-01
-3.27764958e-01 -6.60092473e-01 5.23125269e-02 -3.03401519e-03
2.93569416e-01 -2.08145693e-01 6.65879726e-01 -2.31369451e-01
1.36598253e+00 -2.50753284e+00 -1.22063294e-01 1.01610683e-01
-2.54706413e-01 4.26971108e-01 -6.68343484e-01 8.14819813e-01
-1.05727360e-01 -2.88880080e-01 -7.08978251e-02 -3.86781216e-01
1.92692056e-02 1.82028502e-01 -6.16930604e-01 1.01916201e-01
1.61320865e-02 5.10105550e-01 -1.25424206e+00 -7.42458224e-01
1.85172170e-01 3.84850323e-01 -7.03783751e-01 2.64618188e-01
-1.45222738e-01 -6.36250228e-02 5.58225401e-02 5.45322537e-01
2.08091259e-01 5.82016865e-03 -1.96172819e-01 1.11504391e-01
1.98510379e-01 2.45824158e-01 -1.18910408e+00 2.30004954e+00
-7.05625474e-01 6.33434832e-01 -2.88998663e-01 -1.05893481e+00
9.18867111e-01 7.80093610e-01 3.68171811e-01 -6.08779371e-01
2.26570085e-01 8.55069086e-02 -3.59798789e-01 -5.57018995e-01
4.44508433e-01 -4.66770530e-01 -3.73055398e-01 6.23348773e-01
9.66692746e-01 -2.13844657e-01 3.39408100e-01 3.36257190e-01
1.30979347e+00 -1.41658545e-01 4.09531385e-01 7.75380656e-02
4.07315522e-01 -2.15504676e-01 3.50293100e-01 8.72373760e-01
-2.33410060e-01 1.06634235e+00 -7.56781921e-02 -9.60811749e-02
-8.67614746e-01 -1.00273097e+00 -2.58694142e-01 1.95775080e+00
-3.22751492e-01 -1.06466806e+00 -5.85028172e-01 -4.79883790e-01
-2.38082141e-01 9.31151569e-01 -4.55102116e-01 -5.30043602e-01
4.80411574e-02 -3.23927701e-01 8.54347944e-01 5.45404315e-01
-3.27554420e-02 -9.62149382e-01 -5.72218955e-01 3.53525996e-01
-2.72027254e-01 -9.53673661e-01 -6.82274401e-01 5.94109297e-01
-6.13538802e-01 -8.61231089e-01 -9.01332438e-01 -7.06171095e-01
-1.04323879e-01 5.06213188e-01 9.89474297e-01 -5.80100834e-01
-4.15560901e-01 6.28267050e-01 -9.32979643e-01 -5.73175311e-01
-2.33891770e-01 1.81067482e-01 2.35805035e-01 4.57816906e-02
6.11470222e-01 -9.69559669e-01 -2.96865463e-01 4.61443603e-01
-6.36578798e-01 -3.93450677e-01 1.18623674e-01 8.07508349e-01
4.70751613e-01 -7.24342987e-02 7.69734383e-01 -5.12633383e-01
6.74803436e-01 -5.18232226e-01 1.00353457e-01 2.73449719e-01
-3.45947385e-01 -5.86251318e-02 4.79580194e-01 -9.49118078e-01
-9.15648937e-01 2.29019579e-02 -2.73451090e-01 -8.67727876e-01
-2.35842600e-01 3.89796764e-01 1.98749706e-01 1.80095121e-01
1.12000728e+00 1.44059971e-01 -2.27086544e-01 -7.38352776e-01
5.16697705e-01 1.08399701e+00 5.15032649e-01 -6.20970368e-01
3.87452513e-01 3.12405467e-01 -4.67730701e-01 -1.01381075e+00
-1.32588112e+00 -1.17101836e+00 -5.71579754e-01 -2.48933181e-01
6.10969067e-01 -1.07845223e+00 -1.53583929e-01 9.21639726e-02
-8.10785651e-01 -2.07356438e-01 -9.44551110e-01 7.84033298e-01
-8.38940859e-01 2.73217469e-01 -4.97482687e-01 -9.16142702e-01
-3.52221996e-01 -4.50078160e-01 1.17593646e+00 -9.75060984e-02
-5.22720873e-01 -7.06411242e-01 7.94068038e-01 3.97445410e-02
5.42548239e-01 -4.29550052e-01 6.35143459e-01 -1.08835983e+00
2.05960378e-01 -4.61570293e-01 6.88276216e-02 4.16874379e-01
-8.93309191e-02 -3.55101526e-01 -1.57354331e+00 -3.08311015e-01
-7.11254328e-02 -1.05590725e+00 1.02953303e+00 2.44011134e-02
1.24192369e+00 1.43669844e-01 -9.46079791e-02 4.20581162e-01
1.07560563e+00 8.58983472e-02 4.61922169e-01 -3.84541340e-02
1.47360340e-01 3.79926205e-01 9.73159969e-01 7.42141068e-01
2.56560016e-02 9.17361915e-01 -1.78041998e-02 2.50379086e-01
-2.81271935e-01 -4.50156778e-01 5.15646398e-01 1.26770353e+00
-4.50054333e-02 7.95260966e-02 -8.80999565e-01 7.40911245e-01
-1.77746940e+00 -1.30979419e+00 4.21549499e-01 2.24220896e+00
1.05619633e+00 1.60236597e-01 2.88086861e-01 5.35309792e-01
6.43274605e-01 1.26469180e-01 -1.67806134e-01 -3.46180081e-01
2.38200292e-01 6.46962404e-01 6.68257521e-03 2.17376515e-01
-1.15420377e+00 7.98074543e-01 6.81328964e+00 1.44708204e+00
-7.96889186e-01 8.03808928e-01 -1.41568407e-01 -5.14834285e-01
1.41260773e-01 3.65119912e-02 -8.31870973e-01 4.06916708e-01
1.68160188e+00 -3.89837205e-01 2.23121375e-01 1.04226840e+00
-1.99510381e-01 1.54694825e-01 -1.23457706e+00 1.14172971e+00
4.84134316e-01 -1.08454049e+00 -7.11313263e-02 -5.62643528e-01
5.18254101e-01 1.24232359e-01 -2.39782065e-01 9.24381614e-01
1.24275617e-01 -6.34721577e-01 5.82844377e-01 3.74936044e-01
9.92159486e-01 -7.24342763e-01 6.90066457e-01 3.65037739e-01
-1.43598855e+00 -2.60984242e-01 -6.61798298e-01 -2.28822395e-01
2.11451307e-01 4.64711368e-01 -9.86481786e-01 4.16605949e-01
8.68206084e-01 8.09436500e-01 -4.85333502e-01 1.21657753e+00
1.12039812e-01 8.76679361e-01 -3.48563522e-01 8.88549984e-02
2.22114369e-01 4.75986332e-01 3.94298285e-01 1.57634544e+00
5.98886967e-01 -1.78241190e-02 2.24806637e-01 2.51190573e-01
4.67480123e-02 3.29890192e-01 -6.85226262e-01 -5.77654541e-02
6.26696825e-01 1.23922086e+00 -3.78546774e-01 -5.61208963e-01
-3.46845806e-01 8.64474654e-01 2.95502305e-01 1.38062522e-01
-7.09393084e-01 -7.98114240e-01 4.55354601e-01 6.68216273e-02
5.57834208e-01 1.89379171e-01 3.29563230e-01 -1.20927656e+00
-4.06800032e-01 -6.73360527e-01 7.80507505e-01 -1.01292086e+00
-1.48592448e+00 5.45474052e-01 2.87607402e-01 -1.73185241e+00
-5.36949039e-01 -1.73719659e-01 -6.85551465e-01 3.08309466e-01
-1.34199178e+00 -9.54782546e-01 -1.35133296e-01 9.43499565e-01
9.02205348e-01 -4.86445129e-01 1.49123263e+00 5.51290631e-01
-6.25331551e-02 7.33789504e-01 2.40375012e-01 -8.24239776e-02
1.24333584e+00 -9.74886239e-01 4.18290496e-02 2.70435601e-01
6.87447131e-01 2.72068560e-01 6.64459407e-01 -4.55747284e-02
-8.79708648e-01 -1.08260441e+00 8.26398253e-01 -2.35424116e-01
9.37454879e-01 -5.17776906e-01 -1.09345663e+00 4.40124482e-01
2.35728309e-01 1.23442091e-01 1.35815299e+00 5.27900636e-01
-8.12202036e-01 -4.50901300e-01 -7.84746766e-01 8.70483667e-02
1.02330148e+00 -1.10619938e+00 -1.06758142e+00 4.01328653e-01
7.73092151e-01 3.62697989e-02 -9.13446307e-01 2.14190453e-01
5.17994285e-01 -6.08216703e-01 9.93264437e-01 -7.46577144e-01
2.16100946e-01 -1.00292442e-02 -5.89475453e-01 -1.39680612e+00
-2.93348640e-01 -4.46248472e-01 -1.65002197e-01 1.45178366e+00
2.66679943e-01 -4.24006358e-02 4.34261113e-01 -8.39983448e-02
-1.37445226e-01 -3.23841929e-01 -1.14137304e+00 -1.21236742e+00
-2.22921550e-01 -9.63485301e-01 3.59733224e-01 9.28177893e-01
7.05794573e-01 6.57832563e-01 -7.39393950e-01 -2.22976387e-01
5.07239223e-01 7.34951347e-02 6.05460823e-01 -1.41395366e+00
-5.69952369e-01 8.99066851e-02 -6.46223187e-01 -5.74854493e-01
2.42412463e-01 -1.11512029e+00 2.34397814e-01 -9.16811109e-01
2.92686343e-01 -1.83406070e-01 -8.91699255e-01 5.94294310e-01
3.19749802e-01 5.68515837e-01 2.09073454e-01 1.57681018e-01
-1.19580364e+00 6.58991218e-01 7.04369664e-01 -3.31094801e-01
-1.86829552e-01 8.43114313e-03 -3.52853924e-01 6.65778220e-01
6.62382007e-01 -8.51533532e-01 -7.36950994e-01 2.12573409e-02
6.17149919e-02 1.15882322e-01 6.50403053e-02 -1.48493278e+00
4.11432087e-01 8.67935270e-02 -1.98390260e-02 -4.86954212e-01
7.71999717e-01 -7.17724144e-01 -1.05994880e-01 -7.00169289e-03
-7.28231907e-01 -5.30054867e-01 1.12021819e-01 7.52194405e-01
-5.41697621e-01 -4.99321133e-01 8.63277614e-01 -1.48298502e-01
-9.16165233e-01 1.76136211e-01 -4.49637413e-01 4.50055778e-01
9.78527367e-01 5.64350486e-02 6.73981905e-02 -5.63568771e-01
-9.29575443e-01 -3.36347893e-03 1.93040043e-01 6.07730329e-01
7.17118084e-01 -1.55823600e+00 -5.08019567e-01 2.77871788e-01
7.45899975e-01 -5.10857940e-01 3.04150075e-01 6.34371638e-01
1.71260580e-01 3.55017006e-01 -3.34893405e-01 -6.79934323e-01
-1.46852946e+00 7.99166858e-01 -1.96520649e-02 -1.26000822e-01
-5.65392494e-01 9.15986896e-01 -1.45066887e-01 -3.46959382e-01
6.05659306e-01 -1.84307005e-02 -1.79124638e-01 4.53722060e-01
1.11220288e+00 2.23743811e-01 2.50076622e-01 -2.36103311e-01
-3.78087401e-01 6.14249885e-01 -1.45080671e-01 -5.29059291e-01
1.45263088e+00 -9.57200155e-02 4.39898252e-01 1.15299702e+00
1.27102435e+00 -3.75180483e-01 -9.74373043e-01 -6.01553500e-01
2.43569352e-02 -4.94102389e-01 1.19670659e-01 -5.52962363e-01
-6.44545853e-01 1.40785563e+00 8.92832756e-01 5.13724238e-02
1.17802191e+00 1.90046787e-01 8.09112966e-01 6.93583369e-01
5.86216927e-01 -1.43971395e+00 4.63427663e-01 6.95926487e-01
6.82953537e-01 -1.05253124e+00 -2.27463931e-01 2.40534525e-02
-8.48234355e-01 1.02355158e+00 2.45702326e-01 9.86984372e-02
9.66708720e-01 2.90501952e-01 -9.28320140e-02 -1.62127577e-02
-1.29402018e+00 -5.56221902e-01 3.82729799e-01 7.39738524e-01
3.53987873e-01 -1.46154940e-01 4.62372787e-02 7.10102677e-01
-1.27199829e-01 4.14725989e-01 1.88014671e-01 1.05622125e+00
-7.91877031e-01 -1.04474068e+00 -1.29232600e-01 3.44223291e-01
-2.57787406e-01 -1.76911652e-01 -9.90874991e-02 2.95089632e-01
4.25858349e-02 9.93221343e-01 1.61281019e-01 -6.76738381e-01
3.00592899e-01 8.01855206e-01 3.64902705e-01 -9.98569369e-01
-3.92777443e-01 7.69819021e-02 1.83415368e-01 -5.96883893e-01
-5.17505407e-01 -4.75261956e-01 -7.79940903e-01 7.00112432e-02
-4.57397938e-01 4.32686359e-01 4.16015953e-01 8.83893192e-01
4.23885375e-01 4.51371968e-01 6.73878014e-01 -1.01463139e+00
-6.58268332e-01 -1.19902289e+00 -9.32864010e-01 5.12847543e-01
8.14498812e-02 -8.12793314e-01 -6.05681062e-01 1.08054288e-01] | [15.203974723815918, 5.088539123535156] |
5784cbad-9793-423d-b432-d2f235680a9e | multi-scale-efficient-graph-transformer-for | 2305.15773 | null | https://arxiv.org/abs/2305.15773v1 | https://arxiv.org/pdf/2305.15773v1.pdf | Multi-scale Efficient Graph-Transformer for Whole Slide Image Classification | The multi-scale information among the whole slide images (WSIs) is essential for cancer diagnosis. Although the existing multi-scale vision Transformer has shown its effectiveness for learning multi-scale image representation, it still cannot work well on the gigapixel WSIs due to their extremely large image sizes. To this end, we propose a novel Multi-scale Efficient Graph-Transformer (MEGT) framework for WSI classification. The key idea of MEGT is to adopt two independent Efficient Graph-based Transformer (EGT) branches to process the low-resolution and high-resolution patch embeddings (i.e., tokens in a Transformer) of WSIs, respectively, and then fuse these tokens via a multi-scale feature fusion module (MFFM). Specifically, we design an EGT to efficiently learn the local-global information of patch tokens, which integrates the graph representation into Transformer to capture spatial-related information of WSIs. Meanwhile, we propose a novel MFFM to alleviate the semantic gap among different resolution patches during feature fusion, which creates a non-patch token for each branch as an agent to exchange information with another branch by cross-attention. In addition, to expedite network training, a novel token pruning module is developed in EGT to reduce the redundant tokens. Extensive experiments on TCGA-RCC and CAMELYON16 datasets demonstrate the effectiveness of the proposed MEGT. | ['Jun Shi', 'Shihui Ying', 'Jun Wang', 'Juncheng Li', 'Saisai Ding'] | 2023-05-25 | null | null | null | null | ['whole-slide-images'] | ['computer-vision'] | [ 1.35968313e-01 -2.08479837e-01 -4.99385372e-02 -8.38392228e-02
-9.75573242e-01 -1.25795320e-01 2.38400415e-01 3.22803378e-01
-1.86553076e-01 3.61850053e-01 1.01372227e-01 -9.12728757e-02
-3.90066743e-01 -9.91447568e-01 -5.56231499e-01 -1.10431361e+00
3.76807839e-01 1.10538332e-02 5.48081815e-01 -2.40542591e-01
8.72006789e-02 5.75503707e-01 -1.20322955e+00 3.79703313e-01
1.01228392e+00 1.17590380e+00 5.13308764e-01 3.77815664e-01
-5.15593171e-01 5.35736084e-01 -6.17836356e-01 -1.13420673e-01
-1.44093543e-01 -1.47537842e-01 -4.85731423e-01 -7.06538279e-03
3.83427262e-01 4.98971678e-02 -4.68815863e-01 1.03413713e+00
5.49787521e-01 -1.02848254e-01 4.26121444e-01 -1.16216874e+00
-5.35484076e-01 2.90834278e-01 -9.15744603e-01 3.16014349e-01
-6.87625781e-02 -2.55086627e-02 8.78534317e-01 -8.32691729e-01
5.43231130e-01 1.09229517e+00 6.23656034e-01 1.06395572e-01
-6.51808560e-01 -6.88464880e-01 1.89103931e-01 3.12018812e-01
-1.72875714e+00 -3.52680162e-02 9.52979743e-01 1.74291074e-01
9.56934512e-01 3.68254036e-01 9.42798138e-01 3.17171752e-01
6.52591884e-01 9.61535811e-01 9.51085269e-01 -2.40708798e-01
-1.76302925e-01 -7.84282386e-02 6.91064149e-02 1.06755126e+00
2.15883002e-01 -4.26767766e-01 -6.18754208e-01 -3.05445325e-02
8.75847399e-01 6.07093751e-01 -3.07405889e-01 -8.26665983e-02
-1.21380138e+00 6.86428845e-01 9.79509711e-01 6.60824001e-01
-2.12632328e-01 2.12180018e-02 4.22334433e-01 1.68329403e-01
5.11425257e-01 -6.69882223e-02 -2.17328846e-01 3.88992637e-01
-6.48415029e-01 -1.84665263e-01 1.09822460e-01 6.57719731e-01
9.62013781e-01 -3.38911623e-01 -2.93870628e-01 9.04022634e-01
3.67382526e-01 3.39169949e-01 8.85939360e-01 -2.66393363e-01
3.85826290e-01 1.27850044e+00 -6.45029187e-01 -1.16327715e+00
-5.42816222e-01 -6.52233422e-01 -1.15516508e+00 -2.05832332e-01
-1.03309013e-01 4.01226372e-01 -1.07867455e+00 1.29422057e+00
6.36069179e-01 6.81238651e-01 5.08492216e-02 8.79570544e-01
1.22202897e+00 5.29944420e-01 -9.19030532e-02 -9.49532837e-02
1.60637665e+00 -1.04711187e+00 -5.42114496e-01 -5.61731681e-02
8.13516080e-01 -7.28048921e-01 8.43089938e-01 -2.42502242e-03
-9.25202072e-01 -4.99890178e-01 -1.14506567e+00 -2.81325281e-01
-6.08850658e-01 3.10330600e-01 5.98565578e-01 1.16578624e-01
-9.13978279e-01 2.26986438e-01 -7.61381924e-01 -2.88020372e-01
7.24033415e-01 4.09099638e-01 -5.71206033e-01 -3.59643102e-01
-9.81991827e-01 5.83611488e-01 1.93662435e-01 1.90918162e-01
-4.89139020e-01 -8.20719659e-01 -1.01713729e+00 2.11939454e-01
3.55627775e-01 -4.84958857e-01 6.26249194e-01 -5.99313855e-01
-1.19797444e+00 5.84056556e-01 -3.74204844e-01 8.19729418e-02
1.44635923e-02 5.31128526e-01 -3.37514430e-01 4.80433941e-01
2.15924546e-01 5.87253869e-01 7.59831488e-01 -9.61039603e-01
-8.48329902e-01 -6.92632675e-01 -6.79231435e-02 5.32339513e-01
-7.87681699e-01 -2.55857229e-01 -7.98510134e-01 -7.41991162e-01
4.58479613e-01 -6.54642105e-01 -1.62348002e-01 -5.27101848e-03
-4.72092867e-01 -3.96267474e-01 1.18305743e+00 -5.33753276e-01
1.06300318e+00 -2.29672527e+00 2.61391729e-01 3.38013440e-01
6.10648513e-01 4.12948221e-01 -4.77222502e-01 1.19321896e-02
-2.00870875e-02 2.81158499e-02 1.80229843e-02 -3.51325333e-01
-4.70086873e-01 3.85130227e-01 2.73284912e-01 3.76381636e-01
2.88529485e-01 1.19802296e+00 -7.59951413e-01 -9.30497169e-01
4.56833571e-01 5.38424075e-01 -3.37917328e-01 7.97218233e-02
3.22958291e-01 5.34096882e-02 -7.71293938e-01 9.09874082e-01
9.05917764e-01 -4.58185107e-01 -5.65713905e-02 -7.64680445e-01
-1.24541685e-01 -1.06565408e-01 -9.47044253e-01 1.95152986e+00
-3.29035044e-01 1.48789123e-01 1.95688412e-01 -1.20528138e+00
8.07026148e-01 1.54295936e-01 7.06121325e-01 -7.88007557e-01
1.24838024e-01 1.19322814e-01 -2.54151255e-01 -3.72064143e-01
2.04743102e-01 -1.31669015e-01 9.23377648e-03 2.33708527e-02
1.72107533e-01 -6.43530022e-03 8.01219046e-02 3.61504704e-01
1.25470448e+00 -4.07815754e-01 3.06987613e-01 -1.26302361e-01
8.21042776e-01 -1.00761980e-01 9.52411473e-01 2.37129048e-01
-1.18415751e-01 5.58847129e-01 3.58141452e-01 -3.42724532e-01
-6.14965022e-01 -1.04185760e+00 6.24255277e-02 6.73055291e-01
5.02822578e-01 -4.88746226e-01 -4.11204129e-01 -1.00595582e+00
7.21028522e-02 -2.25246504e-01 -7.24746287e-01 -3.97546709e-01
-6.14580154e-01 -9.54501271e-01 4.04017985e-01 6.57052040e-01
6.43970013e-01 -6.91908181e-01 -3.09620649e-01 2.87221074e-01
-2.26207197e-01 -8.78799438e-01 -7.06102133e-01 1.13359302e-01
-6.11273348e-01 -1.21115136e+00 -7.62273610e-01 -1.00844014e+00
9.31502461e-01 9.67581928e-01 6.15481555e-01 3.58274639e-01
-6.86332285e-01 9.62759107e-02 -4.70632046e-01 -3.61062318e-01
2.37340163e-02 6.18707612e-02 -4.12165940e-01 1.19073115e-01
2.66988754e-01 -3.61308634e-01 -7.95895457e-01 3.24093461e-01
-1.04650605e+00 2.45983928e-01 9.64267373e-01 1.22710371e+00
1.05934119e+00 3.28769743e-01 4.55798090e-01 -5.84654152e-01
4.90367025e-01 -2.86467552e-01 -2.65363723e-01 5.80050588e-01
-2.28118464e-01 -2.47146308e-01 7.99407303e-01 -3.41678470e-01
-9.40038562e-01 -1.20655023e-01 -1.41756907e-01 -6.41039491e-01
3.16801846e-01 7.78558195e-01 -3.14983696e-01 -6.62887335e-01
9.58249718e-02 4.43984270e-01 2.33590201e-01 -1.90457135e-01
-1.18048526e-02 6.00976169e-01 2.00567856e-01 -1.26224414e-01
7.30513453e-01 5.53436399e-01 1.52311787e-01 -7.12100744e-01
-7.23837376e-01 -5.59810579e-01 -3.22096378e-01 7.56153688e-02
8.02977562e-01 -8.92755747e-01 -8.26688766e-01 6.96891308e-01
-7.97460318e-01 2.30847932e-02 -2.61072010e-01 3.67407620e-01
-1.03801526e-01 3.52206141e-01 -7.83666790e-01 -2.07637832e-01
-6.00914180e-01 -1.31681502e+00 1.46622956e+00 8.25590372e-01
4.87129867e-01 -8.39263260e-01 -1.49319455e-01 3.91800970e-01
4.60903436e-01 1.19128063e-01 9.09303665e-01 -3.98957103e-01
-6.87606335e-01 -3.10548604e-01 -7.33776391e-01 1.71529695e-01
4.81711388e-01 6.38543665e-02 -6.42486811e-01 -3.98239523e-01
-1.52619973e-01 -2.07060933e-01 1.11141515e+00 5.01154721e-01
1.29276025e+00 -3.36277648e-04 -8.97968769e-01 7.88433194e-01
1.70703256e+00 1.71656817e-01 4.34861720e-01 1.64451808e-01
1.16357601e+00 1.82273656e-01 7.09792614e-01 2.39077494e-01
6.94321871e-01 5.28972924e-01 2.70305425e-01 -5.65765023e-01
-3.30390513e-01 -1.04205891e-01 7.60544371e-03 1.08904195e+00
1.76542014e-01 -1.54090032e-01 -6.31859601e-01 4.49235559e-01
-1.83390510e+00 -5.20508826e-01 2.85095543e-01 1.75765967e+00
5.06107569e-01 -5.51077798e-02 -3.30970436e-01 2.39161208e-01
6.64520800e-01 2.47608930e-01 -5.90363264e-01 7.00798184e-02
-2.00087130e-01 2.77903765e-01 3.45962912e-01 3.15228552e-02
-9.75561440e-01 7.68950760e-01 4.63383627e+00 1.50362587e+00
-1.34648263e+00 2.01608390e-01 6.03916287e-01 -2.94509567e-02
-3.98134083e-01 -2.61121988e-01 -8.24348569e-01 3.94584864e-01
3.56148958e-01 -2.10579798e-01 1.17822751e-01 4.89140958e-01
7.11548626e-02 -3.58150378e-02 -3.49885225e-01 9.75411415e-01
2.67249733e-01 -1.71122801e+00 3.10184956e-01 1.02502450e-01
3.41523111e-01 5.21414019e-02 -4.44373377e-02 2.30399549e-01
-5.01855202e-02 -7.71847367e-01 1.72231033e-01 4.80493695e-01
9.12987709e-01 -8.92034948e-01 9.37660992e-01 1.29001513e-01
-1.94488764e+00 -2.53697664e-01 -6.53430641e-01 6.44634843e-01
-2.65358508e-01 7.48211384e-01 -6.22281909e-01 1.15344083e+00
5.86435020e-01 1.03410709e+00 -8.40779483e-01 9.89816666e-01
1.55992627e-01 1.23558193e-01 -4.96021062e-01 -6.17026119e-03
3.38529676e-01 -7.42295012e-02 3.05799067e-01 1.01489377e+00
6.10352874e-01 1.64557502e-01 3.32663089e-01 3.85240048e-01
3.34424414e-02 1.70184344e-01 -3.34379345e-01 8.55728090e-02
6.35950923e-01 1.83461261e+00 -8.34196568e-01 -3.63620579e-01
-7.01481581e-01 8.61522973e-01 6.23221695e-01 1.66922703e-01
-7.17153013e-01 -6.65054142e-01 4.88911361e-01 -4.73764911e-02
3.81138533e-01 1.45245016e-01 -2.20388267e-02 -1.21084917e+00
2.06631258e-01 -7.04620004e-01 7.67590404e-01 -5.51981211e-01
-1.24666524e+00 4.49995667e-01 -3.25585753e-01 -1.37001383e+00
5.44298112e-01 -4.86983716e-01 -1.01238334e+00 6.93277299e-01
-2.00641465e+00 -1.81716359e+00 -7.19196975e-01 9.54151750e-01
3.66718680e-01 -8.32460821e-02 6.45832360e-01 3.75950605e-01
-9.44050848e-01 8.25035572e-01 -8.22247379e-03 1.93546996e-01
5.34379900e-01 -1.12847531e+00 -2.26967074e-02 5.15995562e-01
-5.87835573e-02 5.35063863e-01 -8.84960741e-02 -5.27175963e-01
-1.67633426e+00 -1.38352382e+00 3.63255411e-01 3.68442595e-01
4.79910940e-01 -7.01856986e-02 -1.08617091e+00 4.45429027e-01
-1.27521783e-01 5.75778306e-01 8.06779861e-01 -2.13972762e-01
-1.77834794e-01 -5.23118436e-01 -1.16023517e+00 3.58193606e-01
8.14719260e-01 -3.41330081e-01 -2.98586249e-01 3.60030442e-01
7.85267711e-01 -5.67908287e-01 -1.19880009e+00 6.95921779e-01
3.45776230e-01 -7.76865005e-01 1.05393255e+00 7.61089176e-02
1.61619186e-01 -6.04354560e-01 1.08272687e-01 -1.48920131e+00
-5.31425774e-01 -5.18157445e-02 1.46504417e-01 1.26011932e+00
5.43765090e-02 -8.80896628e-01 9.31269169e-01 -1.06565878e-01
-5.93213916e-01 -1.29322386e+00 -1.34878516e+00 -4.95419621e-01
-2.38792449e-01 7.28224367e-02 8.93389165e-01 8.46741796e-01
5.72000071e-02 2.12089866e-01 1.17053032e-01 1.99272096e-01
5.42897522e-01 2.83461779e-01 4.99011070e-01 -1.04495323e+00
-5.27543426e-02 -5.59738159e-01 -8.41576457e-01 -7.11160243e-01
-4.20687571e-02 -1.06056774e+00 -2.19026655e-01 -1.73239410e+00
4.03583378e-01 -6.37547135e-01 -7.88531959e-01 6.78206384e-01
-4.86963958e-01 4.80380744e-01 -1.10937692e-02 6.13945648e-02
-7.15108514e-01 7.00376272e-01 1.70142293e+00 -4.36892748e-01
2.71767695e-02 -4.49830800e-01 -8.10536504e-01 5.70373297e-01
5.52993298e-01 -2.11808011e-01 -3.79644066e-01 -2.83384472e-01
-1.58810601e-01 8.75504017e-02 3.89679492e-01 -1.21294904e+00
4.87314850e-01 -1.27957270e-01 8.03787470e-01 -9.23218846e-01
3.94445002e-01 -6.89854264e-01 -5.62529266e-02 4.04472172e-01
7.87085444e-02 -1.70549184e-01 2.15907544e-01 5.45097053e-01
-5.46145380e-01 1.18386887e-01 7.12814867e-01 -8.07064250e-02
-5.02857149e-01 8.36237490e-01 3.68891507e-02 -3.97688806e-01
1.24482787e+00 -3.04555982e-01 -6.44161701e-01 4.42769557e-01
-3.99920583e-01 5.68555832e-01 3.47499043e-01 3.10959548e-01
1.06368423e+00 -1.43519318e+00 -5.92166662e-01 5.79178035e-01
3.99082333e-01 4.68804628e-01 9.12537694e-01 9.50092614e-01
-2.71359950e-01 2.41360053e-01 -2.45198160e-01 -6.51021838e-01
-1.56333542e+00 3.96122605e-01 3.38877946e-01 -8.17595601e-01
-6.90142930e-01 1.15147865e+00 4.70837325e-01 -2.14749455e-01
-2.23319963e-01 -3.49550784e-01 -3.58819544e-01 1.84452794e-02
6.16828144e-01 -2.46942174e-02 2.99422622e-01 -5.67047358e-01
-4.32717592e-01 9.46778953e-01 -6.77691460e-01 3.80728245e-01
1.26418090e+00 -1.14109412e-01 -3.92773598e-01 6.30834177e-02
1.44370401e+00 -1.20676205e-01 -9.96898174e-01 -5.42173147e-01
-5.29191494e-01 -4.47606653e-01 2.68773675e-01 -3.65478188e-01
-1.51661706e+00 8.12592387e-01 6.20376766e-01 -8.94772410e-02
1.50604892e+00 2.13687252e-02 1.26291978e+00 7.83711001e-02
1.57735258e-01 -9.62756634e-01 2.60588944e-01 1.47150934e-01
5.69780707e-01 -1.00191855e+00 2.84320675e-02 -6.08212352e-01
-3.96291107e-01 1.10713112e+00 7.06524849e-01 -1.42052397e-01
6.95576370e-01 3.98882240e-01 -1.54688563e-02 -4.57829386e-01
-6.64578915e-01 -1.17459685e-01 2.87329555e-01 4.53955472e-01
6.93408623e-02 8.41813013e-02 -1.57905847e-01 5.32458127e-01
1.99954227e-01 1.33552209e-01 1.64115742e-01 1.05160499e+00
-4.80238914e-01 -1.00154710e+00 -1.92412123e-01 9.09460127e-01
-2.26374134e-01 6.56449469e-03 -3.24672498e-02 6.28743112e-01
3.29973310e-01 8.05514038e-01 -4.52469615e-03 -7.13358700e-01
3.68294537e-01 -4.65523362e-01 3.25004995e-01 -5.35650074e-01
-6.37169838e-01 3.25654358e-01 -4.88117278e-01 -5.22548079e-01
-3.15690964e-01 -3.79297227e-01 -1.51312041e+00 -1.49165774e-02
-4.67519939e-01 2.57146746e-01 4.27948117e-01 9.64837074e-01
3.88247043e-01 9.88921285e-01 8.85651886e-01 -6.23401821e-01
-5.93294352e-02 -8.05498362e-01 -6.36670589e-01 1.98383167e-01
2.35890076e-01 -6.86048031e-01 -1.88771859e-01 -4.18174565e-01] | [15.031338691711426, -2.7585055828094482] |
2022340d-37bf-4e17-9298-44820bcda990 | boosting-detection-in-crowd-analysis-via | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Wu_Boosting_Detection_in_Crowd_Analysis_via_Underutilized_Output_Features_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Wu_Boosting_Detection_in_Crowd_Analysis_via_Underutilized_Output_Features_CVPR_2023_paper.pdf | Boosting Detection in Crowd Analysis via Underutilized Output Features | Detection-based methods have been viewed unfavorably in crowd analysis due to their poor performance in dense crowds. However, we argue that the potential of these methods has been underestimated, as they offer crucial information for crowd analysis that is often ignored. Specifically, the area size and confidence score of output proposals and bounding boxes provide insight into the scale and density of the crowd. To leverage these underutilized features, we propose Crowd Hat, a plug-and-play module that can be easily integrated with existing detection models. This module uses a mixed 2D-1D compression technique to refine the output features and obtain the spatial and numerical distribution of crowd-specific information. Based on these features, we further propose region-adaptive NMS thresholds and a decouple-then-align paradigm that address the major limitations of detection-based methods. Our extensive evaluations on various crowd analysis tasks, including crowd counting, localization, and detection, demonstrate the effectiveness of utilizing output features and the potential of detection-based methods in crowd analysis. Our code is available at https://github.com/wskingdom/Crowd-Hat. | ['Fengyu Yang', 'Shaokai Wu'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['crowd-counting'] | ['computer-vision'] | [-2.80666292e-01 -2.79958904e-01 2.52539128e-01 -1.86701298e-01
-5.83424747e-01 -6.63688838e-01 6.37260258e-01 4.13204104e-01
-6.53977513e-01 6.73233807e-01 4.52577412e-01 -1.91667497e-01
4.58677620e-01 -6.78049386e-01 -1.89750358e-01 -5.68993747e-01
4.00977507e-02 3.26503098e-01 8.30164552e-01 -2.56986290e-01
3.80849600e-01 5.58267236e-01 -1.82833874e+00 1.32340580e-01
7.02465415e-01 7.35032558e-01 7.31275007e-02 1.13352811e+00
-1.10299036e-01 6.76216543e-01 -1.01232517e+00 -4.96591121e-01
3.67157817e-01 -3.34517169e-03 -2.08258912e-01 -1.72955826e-01
4.55210268e-01 -7.77896345e-01 -3.56909513e-01 6.71674252e-01
8.94126534e-01 1.71063095e-01 5.94091117e-01 -1.14082599e+00
-1.65059060e-01 1.03291616e-01 -9.98845518e-01 8.46832514e-01
7.61140645e-01 6.01856112e-01 6.35333717e-01 -1.07508516e+00
3.77689660e-01 1.36590123e+00 7.61302233e-01 2.67264754e-01
-7.38989592e-01 -6.20047212e-01 6.79472685e-02 -2.36528873e-01
-1.62755775e+00 -7.34259546e-01 3.48986149e-01 -8.38340998e-01
9.83768165e-01 2.47358128e-01 7.41878211e-01 5.41062236e-01
5.59275821e-02 8.98866475e-01 7.56211638e-01 -3.68647575e-01
1.61395490e-01 6.45061508e-02 -1.13298252e-01 7.93131769e-01
6.94495380e-01 -1.97467238e-01 -8.21673632e-01 -5.04822075e-01
5.43985188e-01 1.49670944e-01 -1.09620474e-01 -1.71798803e-02
-1.05590987e+00 7.70581305e-01 2.86039144e-01 -8.65395069e-02
-1.98088825e-01 2.53187895e-01 3.55687141e-01 -3.73241454e-01
7.79721439e-01 3.24823767e-01 1.48911208e-01 -5.72424114e-01
-1.08444190e+00 7.70251215e-01 5.26637912e-01 8.83373380e-01
8.12918067e-01 -2.64719576e-01 -5.46442330e-01 3.39150399e-01
1.84257746e-01 8.44062448e-01 -8.70132372e-02 -1.12912118e+00
6.77889764e-01 8.92055333e-01 6.56849742e-01 -1.36042750e+00
-4.66915101e-01 7.71009997e-02 -1.97466627e-01 1.79850250e-01
6.92926168e-01 -4.11838114e-01 -5.79974115e-01 1.27768672e+00
7.07574844e-01 7.85125494e-02 -3.78012329e-01 9.83159840e-01
8.35777879e-01 2.99779534e-01 -7.59503916e-02 5.69240563e-02
1.28880394e+00 -7.28305042e-01 -5.42084575e-01 -2.36555636e-01
6.79138303e-01 -6.62951410e-01 9.03337419e-01 -8.30751061e-02
-1.13978946e+00 -3.97699118e-01 -8.21293652e-01 -1.27688246e-02
-4.04300332e-01 1.28234848e-01 5.53649962e-01 7.77922153e-01
-8.78430486e-01 1.12613887e-01 -8.22446883e-01 -3.40978205e-01
7.37558186e-01 2.65743583e-01 -6.21970035e-02 7.60239810e-02
-6.57304466e-01 7.50777245e-01 -8.01240504e-02 1.38852326e-02
-3.92053872e-01 -4.85010713e-01 -8.64480972e-01 -2.87729353e-01
4.70111221e-01 -5.64115822e-01 1.30702817e+00 2.97748372e-02
-1.03837872e+00 6.32592380e-01 -7.84508109e-01 -3.64610583e-01
9.69092488e-01 -3.93914938e-01 4.95621450e-02 4.53001261e-01
6.30122185e-01 5.79254866e-01 5.69703341e-01 -1.23647559e+00
-1.12175369e+00 -3.40777785e-01 8.73541087e-03 2.54240662e-01
-1.61120400e-01 2.58186877e-01 -5.33643723e-01 -3.50373536e-01
-1.48242995e-01 -6.37004972e-01 -3.29236031e-01 1.03819840e-01
-2.77997285e-01 -5.67893032e-03 8.41208041e-01 -5.99453270e-01
1.48536825e+00 -2.08712125e+00 -5.99634767e-01 2.02232689e-01
7.21744597e-01 4.07821119e-01 2.08091646e-01 3.24435413e-01
9.21313345e-01 1.14052929e-01 9.69476253e-02 -6.13534272e-01
-1.65843934e-01 -1.91782832e-01 -2.08826080e-01 6.69614196e-01
3.65286529e-01 1.02659464e+00 -1.21709096e+00 -8.41325462e-01
4.06228781e-01 4.22152102e-01 -4.88270164e-01 2.61694670e-01
9.05938819e-02 4.66355950e-01 -4.59722251e-01 1.05317974e+00
9.22125697e-01 -1.79813474e-01 -2.05355674e-01 1.73776165e-01
-4.12009329e-01 -4.23302539e-02 -1.17490590e+00 1.06085300e+00
1.32996172e-01 7.45034635e-01 4.63457853e-02 -4.96691801e-02
7.16950595e-01 -2.38697790e-02 2.94113755e-01 -3.72427046e-01
1.78412765e-01 2.03934550e-01 -2.73100287e-01 -5.33025444e-01
1.13665020e+00 3.34276766e-01 3.47032957e-03 5.30307531e-01
-5.38639486e-01 -4.85627502e-02 5.05956590e-01 3.74267489e-01
1.35269701e+00 3.86801511e-02 5.13864458e-01 3.38870622e-02
1.82367116e-01 1.51088998e-01 3.46282631e-01 1.04798722e+00
-6.09665036e-01 8.37903738e-01 4.24763650e-01 -5.42470157e-01
-9.10658717e-01 -8.82761002e-01 1.58146873e-01 1.32435167e+00
2.99520671e-01 -4.54522163e-01 -8.64174247e-01 -5.68323791e-01
3.75691146e-01 2.07424000e-01 -7.27401197e-01 3.98142517e-01
-8.38054776e-01 -6.53806031e-01 8.33795249e-01 7.43430734e-01
3.11199009e-01 -6.27377510e-01 -1.36623991e+00 1.35648865e-02
-2.50613123e-01 -1.10945952e+00 -5.08812189e-01 -2.72874206e-01
-4.34904605e-01 -1.20401716e+00 -8.34285498e-01 -1.48509875e-01
7.19796240e-01 9.81422007e-01 8.54583859e-01 4.61942464e-01
-3.76173407e-01 5.03895104e-01 -4.51156676e-01 -8.22830677e-01
-9.00573656e-02 -6.27775937e-02 2.24279895e-01 -1.62413061e-01
7.27008343e-01 -2.11139217e-01 -8.33254457e-01 5.06366909e-01
-5.35014212e-01 -3.33923399e-01 1.76530093e-01 3.53761882e-01
2.75949538e-01 -4.37101871e-01 3.02030206e-01 -4.59733188e-01
8.02884758e-01 -5.42497694e-01 -7.10618794e-01 1.66747672e-03
-1.17089964e-01 -3.60321075e-01 2.47233603e-02 -3.65583867e-01
-8.73951674e-01 2.72291414e-02 3.11743766e-01 -3.49277586e-01
-1.30368039e-01 4.73080426e-02 8.56642053e-02 -1.67270556e-01
7.13207185e-01 -7.11311470e-04 1.16274081e-01 -1.98149145e-01
1.58770844e-01 8.30666423e-01 5.59790969e-01 -5.02617836e-01
6.04757965e-01 1.18465221e+00 -2.06489265e-01 -8.92582595e-01
-8.15618634e-01 -1.02189660e+00 -6.99317575e-01 -4.40681934e-01
6.80750966e-01 -1.15861976e+00 -9.31724191e-01 4.13138568e-01
-1.38447094e+00 -1.70872897e-01 -3.01494241e-01 2.78853446e-01
-7.29390159e-02 6.27355278e-01 -3.36350471e-01 -1.44264698e+00
-1.66556627e-01 -1.01862872e+00 1.47347474e+00 5.19490361e-01
-3.74495000e-01 -6.44406021e-01 9.24186409e-02 3.68742794e-01
4.85981494e-01 4.25690025e-01 -1.58734359e-02 -5.95584810e-01
-5.62840879e-01 -5.24134457e-01 -4.11416322e-01 -2.05933303e-01
-1.64815217e-01 1.22294649e-01 -1.16638267e+00 -3.52156669e-01
-5.10732651e-01 -7.70786256e-02 9.69077051e-01 6.12947404e-01
7.31687605e-01 -9.37922746e-02 -6.27560675e-01 4.46061373e-01
9.72923338e-01 -3.10903370e-01 4.27957714e-01 3.78796101e-01
5.80386639e-01 6.16434693e-01 6.60126388e-01 1.11634922e+00
8.09806824e-01 5.85295975e-01 3.53942752e-01 -1.60292555e-02
-1.20504774e-01 -3.93985182e-01 1.85545981e-01 3.08371663e-01
-4.65136886e-01 -4.13342893e-01 -1.04258466e+00 6.43614650e-01
-1.98006105e+00 -1.04472148e+00 -2.27398083e-01 2.06790352e+00
3.44312727e-01 5.61250150e-02 7.88949907e-01 -5.65717928e-02
1.00605333e+00 2.58482635e-01 -3.48450184e-01 7.02380110e-03
-2.09334433e-01 -2.40921706e-01 6.07555449e-01 5.15368104e-01
-1.08918321e+00 9.75825846e-01 6.86606646e+00 6.03890002e-01
-7.19520509e-01 1.84091926e-01 3.74498785e-01 -3.94983470e-01
4.70863990e-02 -7.44866282e-02 -1.33530807e+00 6.59516513e-01
3.89788091e-01 -6.64556101e-02 1.02757081e-01 7.84940302e-01
4.80241209e-01 -8.56827736e-01 -8.39789808e-01 1.00577402e+00
9.87804234e-02 -1.39939368e+00 -1.53996542e-01 4.08279657e-01
5.36554098e-01 5.98472022e-02 -7.18248188e-02 9.67109874e-02
2.66111970e-01 -8.34322214e-01 1.10605466e+00 3.89890194e-01
6.58861458e-01 -6.06310487e-01 1.01320791e+00 2.90153295e-01
-1.54470778e+00 -1.31889477e-01 -5.64748228e-01 -5.87960660e-01
5.40823638e-01 6.51903629e-01 -1.21603131e+00 -6.93022683e-02
7.27694988e-01 2.76711494e-01 -8.16188097e-01 1.19132531e+00
-8.50813910e-02 4.15327907e-01 -6.23700857e-01 -3.93241316e-01
4.87818196e-02 2.46389657e-01 8.03852558e-01 1.71359062e+00
3.64905596e-01 1.47930935e-01 3.38959664e-01 8.44026506e-01
1.44780234e-01 2.07562018e-02 -7.15365469e-01 3.54532868e-01
9.84181881e-01 1.07488275e+00 -9.70054567e-01 -3.68389338e-01
-3.05359155e-01 4.86409932e-01 4.02429461e-01 3.87053639e-01
-8.92994463e-01 -3.35717171e-01 6.42506301e-01 6.83305442e-01
5.11669040e-01 -4.05788392e-01 -4.33428913e-01 -1.01494205e+00
4.27010268e-01 -4.96008605e-01 2.96464503e-01 -3.88173997e-01
-9.74828959e-01 3.17913890e-01 2.80302435e-01 -1.16719985e+00
-1.55138895e-01 -4.05367374e-01 -7.27510691e-01 6.70969963e-01
-1.29324317e+00 -1.12874103e+00 -7.53578663e-01 4.32170570e-01
3.18443656e-01 -2.12636039e-01 2.09054962e-01 4.15478870e-02
-4.62391526e-01 6.15624726e-01 -1.71029478e-01 3.59195054e-01
5.55942059e-01 -1.00227320e+00 6.52958095e-01 1.02835655e+00
-3.38047892e-01 6.24766529e-01 6.71588004e-01 -9.63023186e-01
-1.14854765e+00 -9.14664745e-01 6.72194004e-01 -9.87001479e-01
4.74329084e-01 -4.73582178e-01 -5.87159216e-01 2.38057271e-01
-2.29750365e-01 2.65665293e-01 6.05494380e-01 -1.22532286e-01
-1.05676256e-01 2.37684458e-01 -1.18745410e+00 5.80671549e-01
1.03956664e+00 -1.46503255e-01 -2.86712080e-01 8.19674805e-02
5.48045039e-01 -5.82667351e-01 -4.36462283e-01 1.49350300e-01
7.13591814e-01 -1.37888014e+00 8.94494772e-01 -1.31866351e-01
4.20939326e-02 -6.66561604e-01 -2.58527100e-01 -8.67427051e-01
-1.24044284e-01 -7.36347437e-01 -3.66117954e-01 1.10528326e+00
3.19352448e-01 -6.92857981e-01 7.77680278e-01 7.06892788e-01
1.58587784e-01 -6.40312910e-01 -1.05078852e+00 -7.14211464e-01
-3.82871360e-01 -3.60224456e-01 8.12501013e-01 4.69145060e-01
6.30047545e-02 1.89900547e-02 -2.68718988e-01 2.02634797e-01
5.28174400e-01 -1.76425621e-01 1.49123824e+00 -9.96352196e-01
-1.17741758e-02 -3.51964772e-01 -5.22897065e-01 -1.20122015e+00
-2.72943258e-01 -1.78205207e-01 1.73967659e-01 -1.47025859e+00
4.83079761e-01 -4.01255280e-01 3.94209951e-01 2.02509224e-01
-5.38839400e-01 4.05894160e-01 3.80812049e-01 5.25534451e-01
-1.09686351e+00 2.43400112e-01 9.46064949e-01 1.74176082e-01
-3.39890927e-01 2.20832862e-02 -5.68145096e-01 9.46324885e-01
8.77935886e-01 -4.14074928e-01 1.60442337e-01 -6.59639180e-01
2.70178288e-01 -4.17379707e-01 4.62733269e-01 -1.32418740e+00
5.16635299e-01 -2.50132769e-01 4.64203537e-01 -7.90107906e-01
4.75362748e-01 -1.88540950e-01 -4.34528589e-01 3.51117402e-01
3.61114182e-02 3.32118899e-01 1.87346548e-01 7.31164753e-01
6.12396412e-02 1.50226597e-02 4.63548422e-01 -2.55879253e-01
-5.04642963e-01 1.00943662e-01 -3.72836024e-01 3.35509449e-01
1.12981486e+00 -7.58461714e-01 -6.17158234e-01 -5.43932021e-01
-1.05263256e-01 2.71132439e-01 8.34154069e-01 5.48286662e-02
7.35074162e-01 -8.59263659e-01 -9.19989169e-01 9.57524553e-02
2.68817484e-01 2.31321603e-01 -4.05131206e-02 1.01234865e+00
-7.90241897e-01 1.83119163e-01 8.44966024e-02 -8.77413869e-01
-1.33159423e+00 3.47358882e-01 6.10476993e-02 -1.35994479e-01
-4.63627785e-01 9.71608579e-01 8.11489746e-02 -2.27848947e-01
2.71261275e-01 -2.78216094e-01 3.14959548e-02 5.60025545e-03
1.15263402e+00 1.09057021e+00 -3.34147573e-01 -9.32778776e-01
-7.00929880e-01 3.68656725e-01 1.97368965e-01 -2.54023433e-01
1.07587266e+00 -3.61959100e-01 4.92805131e-02 1.07050806e-01
6.17014349e-01 6.32222772e-01 -1.50450015e+00 -1.43768400e-01
-1.65845826e-01 -1.00535619e+00 -4.22915250e-01 -2.78776795e-01
-6.14417493e-01 8.39706779e-01 3.45562279e-01 2.08847567e-01
5.51929355e-01 1.52274668e-01 7.39179254e-01 2.05662102e-01
5.29157639e-01 -1.22153485e+00 1.29876778e-01 5.02925932e-01
5.29387951e-01 -1.50028849e+00 3.67864192e-01 -4.15627062e-01
-5.98760128e-01 8.50418627e-01 7.17289388e-01 9.99384075e-02
3.14037293e-01 7.11990118e-01 -1.81110818e-02 -3.77112538e-01
-3.97486955e-01 -5.87614119e-01 -1.02961235e-01 8.76503348e-01
2.67694116e-01 1.83478847e-01 -8.14143643e-02 3.95734757e-01
-1.80049792e-01 -1.02663398e-01 5.13354719e-01 1.23112798e+00
-9.36453521e-01 -3.84927839e-01 -9.86683905e-01 5.07669747e-01
-4.33861345e-01 3.66342105e-02 -4.03775424e-01 5.93507051e-01
2.95668423e-01 1.32982314e+00 3.01231183e-02 -5.06476760e-01
3.42865914e-01 -4.68226135e-01 2.43060574e-01 -5.90399146e-01
-4.87142295e-01 -9.27514806e-02 1.29678220e-01 -6.20453894e-01
-4.15254205e-01 -7.13994563e-01 -1.06577432e+00 -5.72483122e-01
-5.00749648e-01 4.57117073e-02 5.12836516e-01 8.33747923e-01
6.13256693e-01 -2.79202666e-02 2.37314835e-01 -1.45179427e+00
-3.69336337e-01 -9.61510837e-01 -3.19731742e-01 2.79732078e-01
3.94150615e-01 -9.88794029e-01 -3.65045309e-01 -1.46922007e-01] | [8.273000717163086, -0.40276414155960083] |
a21ded50-ed9f-47e7-b23c-5d8243f2d8cc | bldnet-a-semi-supervised-change-detection | 2201.10389 | null | https://arxiv.org/abs/2201.10389v1 | https://arxiv.org/pdf/2201.10389v1.pdf | BLDNet: A Semi-supervised Change Detection Building Damage Framework using Graph Convolutional Networks and Urban Domain Knowledge | Change detection is instrumental to localize damage and understand destruction in disaster informatics. While convolutional neural networks are at the core of recent change detection solutions, we present in this work, BLDNet, a novel graph formulation for building damage change detection and enable learning relationships and representations from both local patterns and non-stationary neighborhoods. More specifically, we use graph convolutional networks to efficiently learn these features in a semi-supervised framework with few annotated data. Additionally, BLDNet formulation allows for the injection of additional contextual building meta-features. We train and benchmark on the xBD dataset to validate the effectiveness of our approach. We also demonstrate on urban data from the 2020 Beirut Port Explosion that performance is improved by incorporating domain knowledge building meta-features. | ['Mariette Awad', 'Ali Ismail'] | 2022-01-25 | null | null | null | null | ['semi-supervised-change-detection'] | ['computer-vision'] | [ 2.56880552e-01 1.62300691e-02 7.65538663e-02 -2.34808400e-01
-4.94161218e-01 -4.21020418e-01 7.74141431e-01 1.05918097e+00
-3.59455854e-01 5.87914288e-01 7.68977821e-01 -2.71214008e-01
-3.15197736e-01 -1.47800863e+00 -8.53116870e-01 -2.34852821e-01
-6.41985834e-01 1.74776554e-01 2.03339085e-01 -6.92422271e-01
-1.48259863e-01 8.70787084e-01 -1.24773729e+00 2.54423440e-01
6.05253637e-01 8.13263953e-01 6.95202947e-02 6.00061595e-01
2.01437116e-01 7.47085392e-01 -2.55443960e-01 3.09741259e-01
3.76457632e-01 6.78920969e-02 -1.12982285e+00 -9.67252553e-02
4.65022951e-01 -3.76218200e-01 -7.79220402e-01 5.64751208e-01
7.25182652e-01 2.14880615e-01 6.21187627e-01 -7.78269172e-01
-5.55571735e-01 4.40488964e-01 -3.25256497e-01 9.63031590e-01
6.40286386e-01 1.40787527e-01 1.05125666e+00 -7.63055086e-01
7.59081066e-01 1.13516366e+00 1.25262213e+00 -3.02272588e-01
-1.14789510e+00 -1.25557214e-01 1.94276839e-01 4.68883604e-01
-1.26426995e+00 -1.72855720e-01 9.13998067e-01 -4.96358722e-01
1.50194371e+00 6.92885295e-02 7.10124671e-01 9.23992276e-01
1.06123775e-01 5.57303667e-01 7.80674636e-01 -5.28464496e-01
2.26719424e-01 -6.80747926e-01 2.12849542e-01 1.03778577e+00
3.87869924e-01 2.13733122e-01 -4.68932956e-01 4.41118181e-02
5.68370819e-01 3.19178194e-01 -4.00887877e-01 -1.20523654e-01
-1.04557991e+00 7.24384367e-01 1.30107594e+00 5.24205565e-01
-5.84549725e-01 4.54882264e-01 4.93048519e-01 2.05253318e-01
7.49838352e-01 2.52032071e-01 -2.56667286e-01 2.14219138e-01
-8.59048724e-01 -2.42500640e-02 3.74096811e-01 5.35332620e-01
1.11462772e+00 -1.42784072e-02 -3.18164915e-01 4.39942598e-01
-6.05882108e-02 4.35288578e-01 3.55464406e-02 -3.86568934e-01
6.23249173e-01 1.06859386e+00 -1.86974648e-02 -1.54494965e+00
-1.13124025e+00 -5.63323557e-01 -1.16616356e+00 7.06955194e-02
-8.01263452e-02 3.14278863e-02 -1.05572426e+00 1.35853255e+00
4.39368337e-01 3.30072463e-01 -3.69698882e-01 6.03467762e-01
7.22658992e-01 3.82147998e-01 -8.12803209e-02 1.83792159e-01
9.22823071e-01 -6.64238572e-01 -5.05070865e-01 -2.99865782e-01
7.73704708e-01 6.29233345e-02 1.04610860e+00 -7.55389705e-02
-4.95659739e-01 -4.95747060e-01 -1.03219855e+00 -1.06103756e-01
-7.17113018e-01 -2.95841694e-01 3.07304204e-01 3.92991304e-01
-1.17163527e+00 6.95800781e-01 -1.04988074e+00 -8.52874577e-01
8.74748707e-01 9.60791111e-02 -5.91349900e-01 -3.45589489e-01
-1.18474674e+00 9.76964772e-01 5.46387196e-01 2.83348322e-01
-1.14762855e+00 -7.42137253e-01 -1.08579171e+00 1.84477001e-01
1.43998768e-02 -6.68077767e-01 5.49142420e-01 -3.66599530e-01
-6.42402589e-01 7.08816648e-01 2.12263301e-01 -8.79940629e-01
3.49216908e-01 -3.85749608e-01 -3.99803400e-01 3.65826994e-01
1.79015592e-01 3.88484657e-01 4.87169713e-01 -1.18128443e+00
-4.68072504e-01 -3.21040004e-01 3.06827635e-01 6.66947514e-02
-4.27561611e-01 -3.40278417e-01 1.22238785e-01 -7.75183201e-01
1.66996375e-01 -6.16865695e-01 -2.61260033e-01 -2.27695510e-01
-4.47089761e-01 6.55946210e-02 9.97031987e-01 -1.03357363e+00
1.34516931e+00 -1.88919032e+00 -1.88827291e-02 3.78218204e-01
2.16692269e-01 4.95315231e-02 -2.46376887e-01 7.81301022e-01
-1.34602502e-01 3.05467516e-01 -8.09881091e-01 -1.23975128e-01
-4.39770855e-02 3.59154135e-01 -2.68468916e-01 6.14452839e-01
4.32799488e-01 1.03467214e+00 -9.20074642e-01 -1.04371570e-01
4.06920940e-01 5.16232312e-01 -5.66112518e-01 -4.39347327e-02
-3.82478796e-02 4.10491467e-01 -2.96756297e-01 7.91533113e-01
5.57494342e-01 8.92032385e-02 -9.80730057e-02 -2.38556668e-01
-5.44038275e-03 2.44312793e-01 -9.56558049e-01 1.74356163e+00
-4.51828510e-01 7.02777803e-01 -2.56475300e-01 -1.43449628e+00
9.12840784e-01 6.85148388e-02 6.13071799e-01 -9.46746707e-01
-2.28990257e-01 -2.29528457e-01 -7.14471877e-01 -5.94668329e-01
5.77291965e-01 4.38492186e-02 -2.39158228e-01 4.07022059e-01
-1.12925403e-01 -1.53045475e-01 -4.08542529e-02 1.33422911e-01
1.84799838e+00 -1.57384753e-01 3.48170191e-01 -2.66767085e-01
2.00407699e-01 2.51454294e-01 2.75017291e-01 8.07365537e-01
-1.39175519e-01 7.41605639e-01 1.80208251e-01 -9.09823954e-01
-6.19330645e-01 -1.11122000e+00 1.05811745e-01 8.08313131e-01
-6.22341633e-02 -4.82755274e-01 -3.86692137e-01 -1.08323550e+00
9.63590741e-02 7.30643690e-01 -7.25908458e-01 -1.96901575e-01
-8.76316786e-01 -8.19097936e-01 7.95271397e-01 7.08341777e-01
6.82561576e-01 -1.02558923e+00 -7.48059034e-01 3.41773331e-01
-2.58460432e-01 -9.00027931e-01 -7.82714486e-02 4.07603651e-01
-7.16732919e-01 -1.23744118e+00 9.55517069e-02 -7.92250991e-01
5.33179998e-01 3.95996571e-01 1.17387307e+00 2.79635966e-01
-8.31014931e-01 8.77808213e-01 -5.77186286e-01 -1.36832938e-01
2.58644260e-02 3.03777158e-01 -1.25178561e-01 -3.07529598e-01
-2.21371129e-02 -9.06078935e-01 -7.05566585e-01 -1.27791911e-01
-9.91294861e-01 -2.56939977e-01 2.26921454e-01 6.06924295e-01
5.00879884e-01 4.55548733e-01 6.04343593e-01 -5.55803418e-01
5.54624975e-01 -7.65039444e-01 -5.94636202e-02 2.18364611e-01
-6.04079723e-01 -6.80748327e-03 2.02378273e-01 2.32630283e-01
-8.05883288e-01 9.74069610e-02 -4.41797636e-02 -6.19913393e-04
-5.01669228e-01 1.08532333e+00 -7.94259608e-02 8.18262249e-03
1.07205081e+00 -7.69891217e-02 -5.65654755e-01 -5.89147091e-01
5.32357275e-01 4.75646764e-01 7.87667215e-01 -4.36298817e-01
1.10418761e+00 7.70050824e-01 1.02942117e-01 -8.42615604e-01
-6.95280373e-01 -6.80737674e-01 -1.17022550e+00 -2.18319863e-01
8.26758146e-01 -1.04345107e+00 9.56146866e-02 4.75284576e-01
-1.00996137e+00 -6.69178903e-01 -5.89045286e-01 8.05133507e-02
-4.25391644e-01 2.74771303e-01 -3.70158911e-01 -3.97580862e-01
-4.90111917e-01 -3.98101717e-01 1.15708458e+00 -1.27229840e-01
1.07871071e-02 -1.31149626e+00 5.16454458e-01 3.41077968e-02
4.13723081e-01 1.23423004e+00 1.02767646e+00 -4.04032648e-01
-5.45977175e-01 -3.03130269e-01 -2.32979447e-01 1.03891313e-01
6.30252063e-01 -4.77966309e-01 -7.51825809e-01 -4.59878206e-01
-4.39312220e-01 -5.69587341e-03 1.25663841e+00 2.83638805e-01
8.69516253e-01 -4.37030196e-01 -4.75876868e-01 6.06173337e-01
1.68879902e+00 -3.68177176e-01 6.90897286e-01 8.11353028e-01
8.59701872e-01 3.95065516e-01 2.83528507e-01 6.99663222e-01
7.26099610e-01 4.95092332e-01 8.77474964e-01 -3.81921709e-01
-5.14148295e-01 -3.03495854e-01 3.01412672e-01 4.22087520e-01
4.12096493e-02 -2.73659319e-01 -1.70244753e+00 1.00804365e+00
-1.85956335e+00 -1.15539670e+00 -3.45855594e-01 1.73827565e+00
5.60474813e-01 5.49833141e-02 -4.45059091e-02 4.75071758e-01
6.42375052e-01 3.85625064e-01 -3.68657947e-01 3.95065993e-02
-4.12572443e-01 5.87364972e-01 5.30964017e-01 5.40746272e-01
-1.55347562e+00 9.36866581e-01 6.30265808e+00 2.67613679e-01
-7.37016439e-01 3.19171906e-01 2.91299164e-01 -1.18649350e-02
-3.97461116e-01 9.47726890e-03 -2.75631517e-01 -3.89783084e-02
6.48000896e-01 2.31228322e-01 3.74781042e-01 5.45043230e-01
4.00020033e-01 -7.78718591e-02 -8.37373435e-01 5.82790852e-01
8.20489153e-02 -1.62202716e+00 -9.36200842e-02 -1.72563910e-01
8.29176724e-01 4.39405054e-01 -2.42348313e-01 2.40115568e-01
3.02320272e-01 -1.04161024e+00 6.85493946e-01 6.20126009e-01
6.66234374e-01 -7.99220860e-01 6.09528065e-01 1.76307529e-01
-1.46705544e+00 -2.84985125e-01 -7.25800171e-02 -3.25327665e-01
2.83936769e-01 8.52008641e-01 -1.43032801e+00 1.05358851e+00
9.94787931e-01 1.25156546e+00 -1.05967176e+00 1.05022275e+00
-4.68487978e-01 8.60828221e-01 -4.59715128e-01 6.54361010e-01
2.61895210e-01 2.09293649e-01 6.32365167e-01 1.54406536e+00
2.66207129e-01 -5.14304899e-02 4.77695465e-01 5.74465215e-01
-4.27530967e-02 -2.12067232e-01 -9.38941777e-01 2.88574606e-01
2.12349057e-01 1.07394326e+00 -7.68344402e-01 1.11762561e-01
-1.65521368e-01 9.16987181e-01 6.14328027e-01 3.22021574e-01
-6.29878521e-01 -4.46196914e-01 5.87699473e-01 4.32175696e-01
4.48973298e-01 -7.05160618e-01 -3.13195616e-01 -8.78521144e-01
-2.32252717e-01 -4.31083292e-01 6.50369048e-01 -6.16611540e-01
-1.18450224e+00 3.99240494e-01 1.37357578e-01 -9.49755847e-01
1.16740517e-01 -2.44591013e-01 -9.63844121e-01 4.20376837e-01
-1.81827343e+00 -1.55462444e+00 -7.73953736e-01 7.75506079e-01
2.47297987e-01 2.14614511e-01 7.11594701e-01 2.28876233e-01
-7.01396883e-01 1.40797719e-01 9.81428996e-02 3.44937265e-01
3.53042662e-01 -1.36117136e+00 8.30406070e-01 1.25765789e+00
2.96517611e-01 8.10642391e-02 4.85992134e-01 -9.97410297e-01
-9.26872432e-01 -1.70830858e+00 5.96346855e-01 -2.78449148e-01
5.08389115e-01 -2.50808626e-01 -9.15778816e-01 8.09612453e-01
6.29023537e-02 2.00361699e-01 3.29515487e-01 1.96005896e-01
-3.41399699e-01 -9.64007154e-02 -1.23853970e+00 4.20431010e-02
1.59022880e+00 -7.53442168e-01 -6.91470921e-01 5.30639291e-01
7.49687374e-01 -1.90986201e-01 -8.44507396e-01 6.02177441e-01
-1.16459720e-01 -9.02105331e-01 1.17454028e+00 -6.09924197e-01
7.48252869e-02 -4.54861552e-01 -4.90715772e-01 -1.49733770e+00
-5.72993636e-01 -2.35677466e-01 -2.06397966e-01 1.12839770e+00
-1.97170451e-02 -3.88772398e-01 4.74981904e-01 6.17594644e-02
-4.68014151e-01 -3.41484904e-01 -1.23527837e+00 -7.58097470e-01
-7.85681158e-02 -6.18492007e-01 8.15151930e-01 9.82935667e-01
-1.44513547e-01 1.70647264e-01 -1.24216706e-01 7.35852063e-01
3.27827007e-01 -4.93609607e-02 5.13474226e-01 -1.25076520e+00
2.20292792e-01 -2.17926621e-01 -8.15913975e-01 -4.28099185e-01
1.94874614e-01 -1.22575891e+00 -7.75808692e-02 -2.22271776e+00
-7.95638114e-02 -6.40029311e-01 -5.87211788e-01 1.06682026e+00
-8.57966691e-02 1.01650894e-01 -2.66274154e-01 2.90746447e-02
-3.14485878e-01 8.70544255e-01 5.26069641e-01 -6.79464459e-01
-1.62480682e-01 -3.03962111e-01 -1.94793627e-01 3.64336193e-01
1.01865268e+00 -6.62710905e-01 -2.87296385e-01 -7.79431581e-01
2.60519981e-01 -5.30422032e-01 9.37251627e-01 -1.42857814e+00
1.29276320e-01 -2.58159335e-03 3.51080090e-01 -7.97804177e-01
-1.27517179e-01 -7.46957898e-01 2.70540953e-01 6.48893952e-01
-7.15376586e-02 2.85504162e-01 4.07799631e-01 1.05247736e+00
-1.04232095e-01 7.58081824e-02 4.00843978e-01 1.56774577e-02
-8.96177411e-01 4.36885536e-01 -1.96363658e-01 4.62964699e-02
8.34856808e-01 -8.25172439e-02 -5.14068902e-01 -1.54584974e-01
-6.04310274e-01 1.39893480e-02 5.49258769e-01 4.07708108e-01
7.34162807e-01 -1.45739651e+00 -6.91272318e-01 1.12219557e-01
1.31587982e-01 2.58369535e-01 2.54524261e-01 6.75611973e-01
-9.09240901e-01 8.02076608e-02 -2.57551163e-01 -5.45645356e-01
-8.03327084e-01 5.95735252e-01 4.17720735e-01 -4.89128709e-01
-1.00070667e+00 5.55920839e-01 -2.91040331e-01 -6.98120654e-01
-1.11595176e-01 -6.65893257e-01 -2.11016208e-01 2.82102346e-01
2.61488706e-01 5.92509508e-01 5.65921366e-01 -6.51819110e-01
-6.78336740e-01 2.63953507e-01 4.47533250e-01 1.46933928e-01
2.02602673e+00 -2.12785721e-01 -1.03026062e-01 1.75989419e-01
1.12106299e+00 -2.54994035e-01 -1.18599904e+00 -3.93750489e-01
4.61711884e-01 -2.37488925e-01 4.53655154e-01 -8.54992926e-01
-9.12832558e-01 6.21926367e-01 1.10157669e+00 2.30510354e-01
1.24003947e+00 -2.79431604e-03 7.75033355e-01 7.50485599e-01
4.40431893e-01 -9.66802478e-01 3.67103666e-01 5.38467228e-01
1.21989727e+00 -1.19208848e+00 1.88616887e-01 9.17832274e-03
-2.27492690e-01 1.08320117e+00 3.57284158e-01 -2.58856267e-01
9.85272646e-01 8.61508325e-02 -3.72838885e-01 -6.58734322e-01
-2.49639466e-01 -6.87059700e-01 2.00440526e-01 9.87292469e-01
-1.46281391e-01 1.47000879e-01 3.17302823e-01 2.74624169e-01
6.35358989e-02 -2.30638221e-01 3.44174534e-01 1.34565544e+00
-8.05971563e-01 -8.30589652e-01 -2.50812978e-01 4.26760972e-01
1.28806025e-01 -1.49869949e-01 -2.56292850e-01 8.20080936e-01
3.48587334e-01 1.01125491e+00 -1.58676524e-02 -5.00694215e-01
6.29417419e-01 -1.62474722e-01 2.63291597e-01 -7.64147937e-01
-8.60837519e-01 -5.95835984e-01 1.35570377e-01 -5.85887909e-01
-5.42859316e-01 -8.43156099e-01 -1.24918377e+00 -1.36967376e-01
-4.86706197e-02 -3.07583719e-01 3.91813755e-01 9.32296574e-01
5.95076203e-01 6.76397145e-01 6.95638955e-01 -9.06307042e-01
1.36899007e-02 -1.08538616e+00 -3.05632502e-01 4.61469740e-01
5.85898340e-01 -6.90864563e-01 -1.51259720e-01 -1.43256621e-04] | [9.639846801757812, -1.2887578010559082] |
b051af18-7601-4592-9999-85f714154b44 | fourier-mixed-window-attention-accelerating | 2307.00493 | null | https://arxiv.org/abs/2307.00493v1 | https://arxiv.org/pdf/2307.00493v1.pdf | Fourier-Mixed Window Attention: Accelerating Informer for Long Sequence Time-Series Forecasting | We study a fast local-global window-based attention method to accelerate Informer for long sequence time-series forecasting. While window attention is local and a considerable computational saving, it lacks the ability to capture global token information which is compensated by a subsequent Fourier transform block. Our method, named FWin, does not rely on query sparsity hypothesis and an empirical approximation underlying the ProbSparse attention of Informer. Through experiments on univariate and multivariate datasets, we show that FWin transformers improve the overall prediction accuracies of Informer while accelerating its inference speeds by 40 to 50 %. We also show in a nonlinear regression model that a learned FWin type attention approaches or even outperforms softmax full attention based on key vectors extracted from an Informer model's full attention layer acting on time series data. | ['Jack Xin', 'Nhat Thanh Tran'] | 2023-07-02 | null | null | null | null | ['time-series-forecasting'] | ['time-series'] | [ 2.16528565e-01 -1.29240662e-01 -4.28558290e-01 -4.35256571e-01
-1.13258100e+00 -3.43910486e-01 5.35005629e-01 1.81549028e-01
-5.52421868e-01 6.16421044e-01 3.92983884e-01 -4.54326510e-01
-4.76017706e-02 -5.56654811e-01 -9.93995070e-01 -4.53459084e-01
-5.07180810e-01 4.22467440e-01 6.56755362e-03 2.44308226e-02
2.99113363e-01 1.79029316e-01 -1.09708476e+00 4.51848984e-01
5.56266308e-01 1.55109501e+00 2.68760830e-01 1.06848693e+00
-5.89459687e-02 1.29779279e+00 -8.83852184e-01 -1.48438171e-01
1.97630703e-01 -1.61100239e-01 -7.52603114e-01 -7.02068150e-01
1.97252020e-01 -6.05224431e-01 -4.29619253e-01 4.88645375e-01
5.03450334e-01 4.25904363e-01 3.52733999e-01 -1.16413248e+00
-9.23989415e-01 1.19842374e+00 -3.18798155e-01 7.96892941e-01
3.15297157e-01 2.78906673e-01 1.44181478e+00 -8.20411325e-01
3.24302912e-01 9.62631345e-01 1.21846604e+00 -4.62503657e-02
-1.14575684e+00 -6.00267768e-01 3.60382944e-01 5.66841662e-01
-1.26540112e+00 -3.51401299e-01 6.81556106e-01 -4.94315177e-02
1.85196888e+00 5.03415287e-01 5.20511627e-01 8.63738418e-01
2.91548163e-01 1.20643103e+00 3.84357989e-01 -4.36540209e-02
1.81383759e-01 -2.91611135e-01 5.84586859e-01 2.78900564e-01
-2.85171986e-01 1.89157259e-02 -7.62998044e-01 -4.43345785e-01
4.81042296e-01 3.23335707e-01 -3.24276060e-01 7.01823354e-01
-1.31069744e+00 8.08095872e-01 3.89626831e-01 3.75157624e-01
-9.77274835e-01 7.11012423e-01 4.12087381e-01 5.65654516e-01
8.31196010e-01 7.12101161e-01 -1.07631421e+00 -5.78451395e-01
-1.33363569e+00 1.53064936e-01 9.01790261e-01 8.17480981e-01
6.63417399e-01 3.85277510e-01 -4.46507335e-01 3.82873595e-01
-2.28377208e-01 5.99673986e-01 8.60136032e-01 -7.04308093e-01
4.57826465e-01 2.01318398e-01 2.06861123e-01 -8.53324056e-01
-4.93505538e-01 -6.64236724e-01 -7.70054221e-01 -5.55840909e-01
6.68425858e-02 -3.74816805e-01 -7.91839421e-01 1.52468002e+00
-1.03369355e-01 6.50458872e-01 -3.41953963e-01 5.90652168e-01
4.21937376e-01 1.21577370e+00 9.71346945e-02 -3.79622072e-01
1.07469058e+00 -9.96584117e-01 -7.46590912e-01 -1.10498190e-01
6.47077203e-01 -6.90177321e-01 8.74216616e-01 4.23757315e-01
-1.36111069e+00 -7.04404771e-01 -7.77322233e-01 -2.97359586e-01
-3.65587652e-01 4.54437062e-02 7.66056418e-01 1.08883448e-01
-1.00379729e+00 7.69337296e-01 -1.03591347e+00 -2.15518922e-02
3.72712135e-01 3.84719044e-01 4.32108045e-02 2.11422697e-01
-1.34038031e+00 7.80749321e-01 3.21980834e-01 3.90878599e-03
-6.88077509e-01 -1.42495787e+00 -6.85780227e-01 6.68139696e-01
1.56073540e-01 -4.88787681e-01 1.41224968e+00 -9.14979756e-01
-1.51028502e+00 -6.88591152e-02 -6.59757793e-01 -1.39463127e+00
2.32029125e-01 -7.52456009e-01 -2.86505520e-01 7.61714354e-02
-2.33455777e-01 2.96776116e-01 9.20002818e-01 -3.24732095e-01
-4.67169434e-01 1.06182620e-01 -1.91700816e-01 -3.80352527e-01
-3.03737104e-01 2.52673268e-01 -1.91764027e-01 -9.67474520e-01
-4.51403320e-01 -4.96151328e-01 -4.91578907e-01 -3.75062227e-01
-3.43821049e-01 -5.57572365e-01 8.79289329e-01 -8.84553730e-01
1.53146160e+00 -1.79263663e+00 -3.89723808e-01 3.34608406e-01
9.57436785e-02 4.24073309e-01 -3.84118468e-01 8.47446442e-01
-2.46931896e-01 9.14820582e-02 -1.13835424e-01 -4.32132304e-01
7.27365613e-02 8.82799402e-02 -9.98038411e-01 5.45088291e-01
5.43242514e-01 1.31819165e+00 -7.82991111e-01 -3.95179763e-02
1.18284831e-02 5.90892136e-01 -6.62403464e-01 4.09492344e-01
-4.47756171e-01 -1.86984882e-01 -3.20200324e-01 4.76114959e-01
4.44584250e-01 -8.42205048e-01 -8.33179578e-02 -1.87408358e-01
-9.41387936e-02 6.31588578e-01 -7.23070800e-01 1.42566848e+00
-6.97681844e-01 8.41207266e-01 -2.67185599e-01 -1.28579223e+00
7.20482409e-01 4.99488890e-01 7.75054395e-01 -6.14194930e-01
-6.29675388e-02 -1.97820455e-01 -1.96760029e-01 -5.58648884e-01
7.22903073e-01 1.26331806e-01 1.23712696e-01 8.29939365e-01
6.57795519e-02 5.68992019e-01 -1.06610648e-01 8.97244513e-02
1.39810061e+00 -1.32122740e-01 2.08699465e-01 -1.39666274e-01
4.03156102e-01 -1.33311182e-01 5.09359300e-01 1.17475319e+00
1.71539307e-01 3.81722599e-01 3.13223094e-01 -7.79903650e-01
-9.82619107e-01 -7.25756049e-01 1.76521894e-02 1.67067623e+00
-5.44263184e-01 -6.21803880e-01 -2.22435236e-01 -4.25699830e-01
1.97342589e-01 9.56542671e-01 -7.53949165e-01 -8.06690939e-03
-7.86568463e-01 -8.00486505e-01 6.46987557e-01 9.37913120e-01
6.29615337e-02 -9.71868873e-01 -6.66694760e-01 6.53729558e-01
-2.98378736e-01 -1.02518570e+00 -8.71677577e-01 2.95513511e-01
-7.90306926e-01 -8.26156616e-01 -8.91319335e-01 -1.42494902e-01
2.59280682e-01 4.38454375e-02 1.18338215e+00 -1.94838524e-01
-1.45158350e-01 2.57782370e-01 -3.18357587e-01 -5.53742468e-01
1.94888804e-02 4.96068925e-01 -1.38240978e-01 5.31937256e-02
5.47130823e-01 -7.40821481e-01 -6.62655413e-01 -4.43395525e-02
-6.99595034e-01 -2.34667331e-01 5.70791781e-01 9.60233033e-01
3.18264693e-01 -3.28493983e-01 8.76917005e-01 -7.81321466e-01
7.55107343e-01 -8.36728871e-01 -7.23145783e-01 7.90845230e-02
-8.34004462e-01 2.57736027e-01 8.49840641e-01 -8.71066093e-01
-6.69937491e-01 -2.61083096e-01 -4.53965366e-01 -6.91334009e-01
3.00150812e-01 9.71737504e-01 4.84824806e-01 3.03097934e-01
4.34151441e-01 5.20606220e-01 -2.39740506e-01 -6.06798947e-01
3.39799583e-01 4.27085578e-01 4.61347699e-01 -2.42397189e-01
5.23664415e-01 3.81484777e-01 -4.53395784e-01 -9.07860994e-01
-7.28166759e-01 -6.48207128e-01 -9.72350165e-02 6.15304969e-02
6.84923470e-01 -8.49427402e-01 -1.15239346e+00 2.21268058e-01
-1.36251771e+00 -5.67632258e-01 -3.90994519e-01 6.78472877e-01
-4.53002363e-01 1.21626861e-01 -8.91621828e-01 -1.19778645e+00
-8.64299119e-01 -4.91366714e-01 1.09721899e+00 -3.77513289e-01
-5.15037119e-01 -1.07421315e+00 -4.14573140e-02 -1.43301979e-01
1.01748145e+00 -3.61720324e-02 6.17267191e-01 -1.16725707e+00
-8.24959993e-01 -5.23332953e-01 -2.79066145e-01 2.35978976e-01
-2.56759822e-01 -1.77522868e-01 -1.07852173e+00 -1.70425013e-01
-6.69815466e-02 -1.22546487e-01 1.34585953e+00 7.11960137e-01
1.39514101e+00 -9.75024462e-01 -2.99427688e-01 7.23499715e-01
1.37821817e+00 1.19588487e-01 3.94853383e-01 -1.57080397e-01
5.26862502e-01 -1.33387558e-02 3.14239204e-01 8.56060922e-01
2.74270713e-01 2.36148521e-01 2.19869345e-01 -7.92792216e-02
3.20358336e-01 -1.77266255e-01 6.64445460e-01 1.03977013e+00
-3.07054847e-01 -4.30811912e-01 -8.02703381e-01 8.21322799e-01
-1.83070552e+00 -1.26491642e+00 1.01321854e-01 2.09897351e+00
8.57896328e-01 2.69983131e-02 1.16413236e-01 9.82682630e-02
3.07797253e-01 5.47353804e-01 -6.58538938e-01 -3.63727272e-01
-2.57648472e-02 5.31098545e-01 7.84511328e-01 7.01671243e-01
-8.78004193e-01 7.55169570e-01 7.44846487e+00 7.67038643e-01
-1.17943871e+00 2.57385343e-01 6.30358458e-01 -3.56462270e-01
-3.29875022e-01 -2.74930954e-01 -7.40047514e-01 4.58521217e-01
1.71767294e+00 -4.64002728e-01 5.95667243e-01 8.88636827e-01
2.73458213e-01 2.59206682e-01 -1.19713461e+00 7.03656316e-01
-5.93059957e-02 -1.74690044e+00 -2.16601357e-01 -1.62055805e-01
6.32864833e-01 7.26663291e-01 -3.50874513e-02 3.67392719e-01
5.68320632e-01 -1.10124445e+00 5.43005824e-01 7.90416241e-01
4.86033171e-01 -7.42379785e-01 8.35895240e-01 4.58360672e-01
-1.30290890e+00 -2.41130188e-01 -3.85565847e-01 -3.67399931e-01
2.88193911e-01 7.44961441e-01 -1.03902662e+00 2.83677518e-01
5.45434892e-01 8.11401963e-01 -8.28332976e-02 8.22647929e-01
2.90519536e-01 1.42485487e+00 -6.49526000e-01 -2.10893005e-01
3.74842018e-01 3.72283220e-01 5.29421985e-01 1.51164687e+00
3.96501720e-01 5.35561107e-02 -2.63398066e-02 8.48188043e-01
-2.88759489e-02 -3.71535197e-02 -1.00874715e-01 -3.22008461e-01
3.87800366e-01 6.88971102e-01 -8.53065252e-02 -7.73552835e-01
-5.22231758e-01 8.78036618e-01 2.35350668e-01 8.77838790e-01
-1.02900445e+00 -5.28078675e-01 6.89805567e-01 -7.16487970e-03
1.02065694e+00 -2.52583295e-01 -1.65488616e-01 -1.12915516e+00
-9.86342728e-02 -5.75586557e-01 5.11238337e-01 -6.83052659e-01
-1.42469954e+00 5.60712039e-01 -2.58021951e-01 -8.62318695e-01
-6.01513326e-01 -3.27418774e-01 -9.42430198e-01 1.22597623e+00
-1.63582289e+00 -9.62005258e-01 2.00487316e-01 6.73492432e-01
5.87447286e-01 1.39745727e-01 8.58978868e-01 2.27963701e-01
-3.27212065e-01 7.24765539e-01 2.02626601e-01 2.70187110e-01
3.95411909e-01 -1.20320046e+00 8.60892177e-01 7.00994551e-01
8.90970379e-02 8.08981359e-01 8.79019976e-01 -4.67483282e-01
-1.69539785e+00 -1.24425912e+00 1.43741977e+00 -2.14871809e-01
9.10206676e-01 -2.02843949e-01 -1.17629147e+00 1.03078806e+00
4.53991801e-01 9.02753100e-02 6.66354895e-01 4.26830918e-01
-4.23073530e-01 -2.11732835e-01 -6.34796798e-01 -3.53302271e-03
5.66047966e-01 -7.50022173e-01 -6.91588521e-01 4.58612382e-01
1.10378671e+00 -2.48624936e-01 -1.21273232e+00 5.87784834e-02
6.33624911e-01 -5.17587662e-01 1.12142742e+00 -9.49672937e-01
4.33003247e-01 2.26832792e-01 -8.54226202e-02 -9.32560325e-01
-3.35649014e-01 -1.17587852e+00 -8.22896421e-01 9.53682482e-01
5.56661367e-01 -9.11526859e-01 6.65191352e-01 3.96431535e-01
8.32974166e-03 -8.24843347e-01 -8.52919936e-01 -7.99613237e-01
-5.08451983e-02 -1.17562711e+00 9.29242074e-01 8.45696747e-01
8.24456736e-02 2.53991961e-01 -6.66553915e-01 3.10066998e-01
1.94515020e-01 3.19694191e-01 5.66495955e-01 -9.10278141e-01
-6.36373222e-01 -4.95664716e-01 -4.62596044e-02 -1.51653540e+00
1.53845185e-02 -6.42889678e-01 2.80550551e-02 -1.27627015e+00
-2.56690681e-01 -6.54801279e-02 -8.43726873e-01 5.15687764e-01
-3.98137540e-01 9.04568098e-03 6.56412318e-02 2.17862248e-01
-7.00917482e-01 6.95038199e-01 5.46927989e-01 -1.96004376e-01
-2.02101901e-01 3.81388992e-01 -4.27454412e-01 3.70608777e-01
7.48886406e-01 -4.92406756e-01 -2.77608722e-01 -3.72927070e-01
4.63604003e-01 2.91060597e-01 3.45852673e-01 -7.60693669e-01
4.90309805e-01 -5.86900348e-03 4.19470370e-01 -1.01450551e+00
3.49918574e-01 -6.32681727e-01 -2.40381613e-01 4.48752075e-01
-7.37563670e-01 4.45184410e-01 1.39307514e-01 6.88578546e-01
-2.86622763e-01 2.98513889e-01 2.32624084e-01 8.63509029e-02
-3.47219139e-01 5.54115474e-01 -7.20586181e-01 1.03605829e-01
3.59666139e-01 3.64375651e-01 -7.26685002e-02 -8.23358297e-01
-4.48077708e-01 3.67190719e-01 -3.48871350e-01 2.17435002e-01
5.15093327e-01 -1.22034657e+00 -9.25709605e-01 3.42674673e-01
-1.55885592e-01 -3.05823058e-01 1.83518901e-01 1.04068613e+00
-8.87706503e-02 8.75257313e-01 4.70095992e-01 -4.94844526e-01
-8.82770240e-01 8.67414176e-01 9.85730141e-02 -6.96119905e-01
-5.68949342e-01 1.11553502e+00 -1.17491089e-01 1.09744091e-02
4.93167102e-01 -1.08890510e+00 1.64033905e-01 2.83217500e-03
9.84671712e-01 3.81137937e-01 -4.23156237e-03 -1.10629834e-02
-2.45882809e-01 2.04191744e-01 -9.80858132e-02 -3.17695662e-02
1.56781769e+00 8.81947950e-02 1.44902557e-01 6.45704150e-01
1.36614323e+00 -5.28863929e-02 -1.19249821e+00 -6.00858152e-01
1.84399113e-02 -2.18093783e-01 1.88176349e-01 -6.79107487e-01
-9.18345094e-01 9.23755348e-01 7.35375807e-02 6.41149700e-01
1.29015768e+00 -3.99549268e-02 1.42837811e+00 7.26337969e-01
-1.88997209e-01 -1.00112009e+00 -4.56200272e-01 7.21650481e-01
9.70519841e-01 -1.17821622e+00 -8.73966217e-02 2.95933276e-01
-3.12454224e-01 1.08363044e+00 -4.79251631e-02 -2.70150840e-01
1.08606160e+00 5.14722645e-01 -1.42062396e-01 1.15521960e-01
-1.56185961e+00 -5.62391281e-02 5.22603095e-01 3.33859116e-01
5.42394578e-01 -1.62222251e-01 1.92939267e-02 9.23917949e-01
-2.15775341e-01 2.91150004e-01 -1.09271102e-01 5.41295052e-01
-3.80078912e-01 -5.93995869e-01 -1.32941112e-01 5.93283892e-01
-8.42019320e-01 -6.33715451e-01 1.60086021e-01 6.83058858e-01
-1.99409291e-01 7.78095782e-01 4.84176159e-01 -4.53373224e-01
-2.14825407e-01 3.00344110e-01 -1.85660109e-01 6.70469776e-02
-9.68380213e-01 5.13605848e-02 8.96424130e-02 -9.42235231e-01
-2.40104869e-01 -6.22072220e-01 -1.03083110e+00 -5.31854391e-01
-3.07686388e-01 1.99324191e-01 4.17049766e-01 1.07000804e+00
6.17711902e-01 7.26575911e-01 5.32441378e-01 -7.06237435e-01
-6.38622701e-01 -1.35360515e+00 -1.72410756e-01 -3.39201614e-02
9.06147838e-01 1.14420734e-01 -2.39356965e-01 2.22209767e-01] | [7.171378135681152, 3.0531742572784424] |
1ad3fa79-cb98-49de-8bf1-e3473c799413 | safe-exploration-for-optimizing-contextual | 2002.00467 | null | https://arxiv.org/abs/2002.00467v1 | https://arxiv.org/pdf/2002.00467v1.pdf | Safe Exploration for Optimizing Contextual Bandits | Contextual bandit problems are a natural fit for many information retrieval tasks, such as learning to rank, text classification, recommendation, etc. However, existing learning methods for contextual bandit problems have one of two drawbacks: they either do not explore the space of all possible document rankings (i.e., actions) and, thus, may miss the optimal ranking, or they present suboptimal rankings to a user and, thus, may harm the user experience. We introduce a new learning method for contextual bandit problems, Safe Exploration Algorithm (SEA), which overcomes the above drawbacks. SEA starts by using a baseline (or production) ranking system (i.e., policy), which does not harm the user experience and, thus, is safe to execute, but has suboptimal performance and, thus, needs to be improved. Then SEA uses counterfactual learning to learn a new policy based on the behavior of the baseline policy. SEA also uses high-confidence off-policy evaluation to estimate the performance of the newly learned policy. Once the performance of the newly learned policy is at least as good as the performance of the baseline policy, SEA starts using the new policy to execute new actions, allowing it to actively explore favorable regions of the action space. This way, SEA never performs worse than the baseline policy and, thus, does not harm the user experience, while still exploring the action space and, thus, being able to find an optimal policy. Our experiments using text classification and document retrieval confirm the above by comparing SEA (and a boundless variant called BSEA) to online and offline learning methods for contextual bandit problems. | ['Rolf Jagerman', 'Ilya Markov', 'Maarten de Rijke'] | 2020-02-02 | null | null | null | null | ['safe-exploration'] | ['robots'] | [-2.17986125e-02 1.00077532e-01 -7.48758137e-01 7.42756799e-02
-1.04629016e+00 -8.04280221e-01 6.01111650e-01 2.17482179e-01
-6.34637237e-01 1.22684586e+00 2.45178014e-01 -6.70787811e-01
-5.52781224e-01 -6.45164788e-01 -8.46528471e-01 -1.00468171e+00
5.23386709e-02 7.59358287e-01 2.66522437e-01 2.10660234e-01
3.48753989e-01 2.26851881e-01 -1.44936204e+00 3.28773081e-01
9.66876149e-01 1.02212179e+00 1.38256043e-01 5.40861666e-01
-1.18416332e-01 6.63813710e-01 -5.54697931e-01 -2.08863094e-01
3.11365575e-01 -4.63410288e-01 -9.53330278e-01 -4.39313442e-01
5.05063795e-02 -6.59586549e-01 -9.38814308e-04 1.15291667e+00
2.01334029e-01 5.95671535e-01 2.52599001e-01 -8.42587352e-01
-2.12910414e-01 8.06996822e-01 -2.78261811e-01 -1.00093253e-01
3.65483880e-01 -2.38096982e-01 1.35696387e+00 -2.73815095e-01
4.36567366e-01 1.27304339e+00 2.05560654e-01 5.27978897e-01
-1.48247600e+00 -6.48053825e-01 7.54379392e-01 -6.70040250e-02
-8.24702024e-01 -9.93756503e-02 6.58445120e-01 -4.28587757e-02
3.43710840e-01 7.92807221e-01 1.06034565e+00 1.09856045e+00
-1.64055228e-01 1.14394045e+00 1.45618236e+00 -4.60264713e-01
7.85402536e-01 3.08021456e-01 8.64534304e-02 4.50621873e-01
3.47170234e-01 5.26836395e-01 -5.46211660e-01 -6.82336748e-01
3.47521931e-01 8.30943286e-02 -3.62424552e-01 -8.12710643e-01
-1.03050900e+00 8.91679347e-01 3.45103085e-01 1.82781339e-01
-7.23311067e-01 2.52514958e-01 2.17057586e-01 3.05286378e-01
4.69244510e-01 7.00641096e-01 -4.72648114e-01 -3.85469943e-01
-1.03020239e+00 5.80935001e-01 6.52522922e-01 1.98699862e-01
6.81119800e-01 -3.99915934e-01 -4.37227368e-01 8.83636117e-01
1.24462523e-01 3.87319028e-01 6.33619845e-01 -8.93325686e-01
4.53309387e-01 4.34066176e-01 8.46514344e-01 -5.90243042e-01
-1.86776340e-01 -3.77850890e-01 -2.24240527e-01 3.62187386e-01
5.65241337e-01 -2.93735594e-01 -9.85298395e-01 1.72398722e+00
3.38965386e-01 -2.08317563e-02 1.62800122e-02 1.08360195e+00
1.19557448e-01 1.00551975e+00 -2.23449208e-02 -6.66236997e-01
9.08775866e-01 -8.11745107e-01 -6.88756883e-01 -4.86794025e-01
6.15541518e-01 -2.91257948e-01 1.17905676e+00 6.18278801e-01
-1.10591388e+00 -1.25820220e-01 -9.37341511e-01 7.02229381e-01
-1.89669341e-01 -2.20197737e-01 7.77709484e-01 7.03745425e-01
-8.28664422e-01 7.20757782e-01 -8.19685519e-01 -4.72246036e-02
2.60368824e-01 3.25232476e-01 1.61540404e-01 -6.53857812e-02
-1.27330554e+00 8.53096724e-01 6.06373131e-01 -1.07782423e-01
-9.78122771e-01 -6.32770002e-01 -3.57379466e-01 3.32264483e-01
8.95295203e-01 -2.47652188e-01 1.53018820e+00 -1.32693827e+00
-1.83086419e+00 9.47829932e-02 9.85704586e-02 -5.12188554e-01
7.97075570e-01 -6.02079213e-01 -6.51335865e-02 -2.98434675e-01
-8.67092013e-02 4.22502816e-01 7.07249999e-01 -1.34471762e+00
-8.88956547e-01 -2.87247688e-01 4.31838810e-01 5.25963068e-01
-4.15078908e-01 -3.11002046e-01 -4.17339712e-01 -2.86772847e-01
2.75824089e-02 -1.33590305e+00 -3.24378699e-01 -5.44572771e-01
-4.23787355e-01 -1.03331491e-01 4.54162598e-01 -4.49962467e-01
1.37770593e+00 -1.97114575e+00 -3.75170857e-02 6.71852410e-01
-5.00696838e-01 3.33991081e-01 -3.89476642e-02 2.77709484e-01
4.12141569e-02 2.98840046e-01 -9.69292670e-02 1.05293028e-01
-1.53538391e-01 2.05078661e-01 -5.81705451e-01 3.36965650e-01
-6.16787136e-01 5.75356185e-01 -1.17904329e+00 -1.83086500e-01
9.86501202e-02 -1.97769538e-01 -6.90587878e-01 1.21496640e-01
-6.71673238e-01 3.37196082e-01 -7.21666813e-01 2.56836027e-01
3.41312468e-01 6.20794818e-02 5.45257390e-01 3.60356659e-01
-1.97560027e-01 3.90002161e-01 -1.34467053e+00 1.07552266e+00
-6.03401363e-01 3.36941570e-01 -3.35556030e-01 -1.12899280e+00
5.90986431e-01 3.67056519e-01 5.57835758e-01 -7.89020538e-01
-1.85921639e-01 3.41732800e-01 -1.14312351e-01 -1.17034204e-01
1.62551612e-01 6.22533448e-03 1.12948887e-01 7.95169413e-01
-5.16845226e-01 2.21032202e-01 -7.15570971e-02 -2.63050422e-02
9.18168485e-01 2.98956811e-01 2.24921480e-01 -1.88240096e-01
2.90711880e-01 -4.67915013e-02 6.74361050e-01 1.38692605e+00
1.31873503e-01 -3.05650979e-02 9.13675129e-01 -5.57932198e-01
-5.88332832e-01 -9.11139667e-01 9.97043177e-02 1.36524808e+00
3.24626029e-01 -5.91443963e-02 -5.14617682e-01 -1.10373902e+00
1.00809261e-01 9.57337737e-01 -6.06933773e-01 -2.24724323e-01
-4.89705622e-01 -6.91179693e-01 -1.69276893e-01 1.98979005e-01
4.54971403e-01 -1.18904459e+00 -1.01227558e+00 2.94427633e-01
-2.88816452e-01 -4.78800923e-01 -3.96438420e-01 2.16724738e-01
-1.11172855e+00 -9.13562179e-01 -6.82586372e-01 -1.10896163e-01
5.72221041e-01 1.60155788e-01 7.75512576e-01 4.81696092e-02
4.35033172e-01 1.37167230e-01 -2.67622918e-01 -3.91078144e-01
-2.37726524e-01 -1.69430912e-01 -6.92596063e-02 -2.17434485e-02
-7.58449286e-02 -9.60741565e-02 -7.60807753e-01 2.47612834e-01
-7.14833558e-01 3.87401134e-02 5.23175299e-01 1.17443883e+00
5.37831604e-01 2.01221392e-01 6.58858716e-01 -1.13611567e+00
8.51931095e-01 -1.80746809e-01 -1.21746898e+00 4.73058999e-01
-9.18842912e-01 5.11895478e-01 6.62132144e-01 -7.09892154e-01
-1.25162673e+00 -1.81665659e-01 2.86938339e-01 -1.74748451e-01
2.88181633e-01 7.64048100e-01 -6.77646026e-02 4.61819887e-01
6.79466367e-01 -4.33232673e-02 -2.33668953e-01 -5.45318127e-01
3.65509182e-01 6.03040040e-01 2.21370071e-01 -7.89691806e-01
3.75946790e-01 3.65515471e-01 -2.03490257e-01 -2.07571477e-01
-1.08384204e+00 -4.04760748e-01 1.59180745e-01 -2.59016514e-01
3.51137996e-01 -5.05269766e-01 -6.87958717e-01 -3.07733715e-02
-8.31934452e-01 -5.45130432e-01 -2.02969864e-01 6.78181708e-01
-6.02342486e-01 -5.98864770e-03 -8.82541686e-02 -1.39440656e+00
-1.86859131e-01 -1.06490386e+00 5.42197645e-01 2.31773213e-01
-2.35846967e-01 -7.62214363e-01 6.20379485e-02 2.55267948e-01
1.46948919e-01 1.57422200e-01 1.16695964e+00 -7.03523040e-01
-5.66153049e-01 -2.71464735e-01 9.01620910e-02 1.08196445e-01
-1.65553503e-02 -2.06327111e-01 -8.10190916e-01 -5.53616405e-01
-2.29535833e-01 -2.44970471e-01 8.22179794e-01 6.22638404e-01
1.37473500e+00 -8.76100481e-01 -5.29846489e-01 1.96021706e-01
1.13112700e+00 8.81100178e-01 4.26872939e-01 8.75074327e-01
-1.15317143e-02 5.89624047e-01 1.06463885e+00 3.49546224e-01
-2.55071133e-01 8.97618234e-01 4.68570411e-01 1.33608403e-02
5.55384219e-01 -3.81953299e-01 5.04098296e-01 -1.79117501e-01
-4.60256226e-02 -1.28541917e-01 -6.66780531e-01 5.60017526e-01
-2.39291954e+00 -8.38696301e-01 5.73356211e-01 2.89535165e+00
9.31299567e-01 4.35808480e-01 3.36148173e-01 -6.13090061e-02
5.54770410e-01 8.02566856e-02 -9.48125839e-01 -7.79212356e-01
4.33581531e-01 1.87662710e-02 5.65036595e-01 5.20755529e-01
-9.61794555e-01 8.92575443e-01 5.42421293e+00 6.10843003e-01
-1.19575560e+00 -2.22396981e-02 7.60683775e-01 -5.82362235e-01
-3.98427010e-01 3.69464159e-01 -5.26884794e-01 6.58905566e-01
9.48771238e-01 -2.47562900e-01 8.94877136e-01 9.39672232e-01
2.81690717e-01 -4.48862016e-01 -1.15101731e+00 6.87084019e-01
-4.50788468e-01 -1.34501040e+00 -3.58453803e-02 1.33719876e-01
9.05022204e-01 -1.79591104e-01 6.21737838e-02 5.17685235e-01
8.06385040e-01 -6.27099395e-01 9.02928770e-01 3.13831121e-01
4.78651851e-01 -1.08283877e+00 9.36598897e-01 7.27059186e-01
-3.48349959e-01 -7.28031158e-01 -1.76659599e-01 7.76509121e-02
-1.37023225e-01 5.40406585e-01 -6.89834297e-01 2.31910884e-01
7.28509128e-01 -3.37164663e-02 -7.87315250e-04 1.17453349e+00
-3.71878505e-01 6.58135891e-01 -2.92475671e-01 -5.06402671e-01
6.88309908e-01 -2.36856058e-01 6.17150664e-01 7.93964565e-01
3.21490765e-01 -1.37567334e-02 4.25269842e-01 3.70397061e-01
-6.10344969e-02 2.90598124e-01 -4.39482629e-01 9.78900790e-02
5.95325291e-01 6.73500597e-01 -5.19533277e-01 -4.22076732e-01
-6.67817295e-02 4.90082800e-01 3.38086486e-01 6.32897198e-01
-5.54689527e-01 6.22011647e-02 5.63416183e-01 -1.09775357e-01
1.85431734e-01 5.11838794e-01 -1.10150687e-01 -7.85374761e-01
9.43001825e-04 -1.03930295e+00 7.19841778e-01 -4.84194010e-01
-5.84966838e-01 1.41257644e-01 4.70567375e-01 -1.03234494e+00
-5.24923503e-01 -2.64037311e-01 -3.72559816e-01 6.03374243e-01
-1.35760725e+00 -4.73028660e-01 2.86628455e-01 2.32182294e-01
3.67729992e-01 4.84992750e-02 6.46307766e-01 -3.06884050e-01
-3.59085917e-01 3.99276733e-01 6.59950852e-01 -3.89957070e-01
5.19748449e-01 -1.31944931e+00 -2.76761740e-01 5.22006989e-01
2.41418704e-01 5.37623942e-01 7.43491650e-01 -6.12307668e-01
-1.19657791e+00 -8.63126099e-01 4.07406718e-01 -1.09569458e-02
4.41045225e-01 -2.93066949e-01 -7.72594690e-01 4.61520195e-01
-3.30493152e-01 -4.06887323e-01 3.77649397e-01 6.31202579e-01
-8.98951218e-02 -5.36061525e-01 -1.13129103e+00 1.09846985e+00
6.10723972e-01 -2.05236107e-01 -4.97958839e-01 4.51927960e-01
5.18580556e-01 -2.84479767e-01 -2.45215818e-01 3.85239333e-01
1.03404760e+00 -1.02682316e+00 8.00734460e-01 -7.99859226e-01
7.44231567e-02 7.85588250e-02 5.99101745e-02 -1.73243332e+00
-1.61637515e-01 -6.92384660e-01 -4.00833815e-01 9.15874124e-01
5.98907351e-01 -8.69603038e-01 8.87127578e-01 7.47376025e-01
3.38269502e-01 -1.06602001e+00 -1.18328667e+00 -9.96639907e-01
1.47155762e-01 -2.99376845e-01 9.61918712e-01 4.98902023e-01
1.21584564e-01 2.06964947e-02 -5.80713153e-01 -2.84900051e-02
4.07949746e-01 6.46915615e-01 6.35336101e-01 -1.10650134e+00
-3.90062332e-01 -5.74462414e-01 4.65901494e-01 -1.21048903e+00
8.12240690e-02 -3.73022318e-01 3.99891883e-01 -1.41966939e+00
3.26881051e-01 -6.85966969e-01 -9.00488019e-01 5.70386410e-01
-2.94669867e-01 -4.00530368e-01 2.40745291e-01 3.36611807e-01
-5.55859208e-01 2.09955528e-01 1.12408686e+00 -1.17236279e-01
-7.53428698e-01 4.61121172e-01 -9.77042258e-01 5.59284747e-01
7.46323645e-01 -6.12996876e-01 -4.84298617e-01 8.11535865e-02
5.24952292e-01 5.84638834e-01 1.50735199e-01 -4.65656012e-01
-1.17683597e-01 -5.83283782e-01 2.98133254e-01 -4.55182910e-01
1.52309999e-01 -8.00423265e-01 1.50548965e-01 8.71469975e-01
-6.98047221e-01 -2.65228003e-01 9.20420811e-02 7.46839464e-01
9.87348184e-02 -4.68477517e-01 7.32373834e-01 -1.56642362e-01
-2.31200591e-01 6.78312108e-02 -4.79096949e-01 -3.00368905e-01
7.88938284e-01 -1.42993838e-01 -2.01131001e-01 -6.94813371e-01
-4.78201956e-01 5.33780813e-01 3.37553442e-01 3.44872952e-01
2.76216865e-01 -1.12926733e+00 -3.95451784e-01 -1.48434192e-01
3.58024538e-02 -1.94445565e-01 -1.52588263e-01 4.64615643e-01
-9.17880423e-03 7.72831440e-01 3.30808312e-01 -3.50604147e-01
-1.07940769e+00 6.60210669e-01 3.05699229e-01 -8.52300406e-01
-4.76094097e-01 4.80131388e-01 2.63549775e-01 -2.38643184e-01
3.79800886e-01 -1.63886815e-01 -2.17713445e-01 1.75435066e-01
4.27269489e-01 2.09026158e-01 1.71726659e-01 2.73348033e-01
-2.27002248e-01 7.69661292e-02 -2.22356662e-01 -5.55594862e-01
1.31601739e+00 1.23157561e-01 1.26978159e-01 4.47963089e-01
7.92573392e-01 1.12559438e-01 -1.33045971e+00 -2.88517743e-01
4.31945443e-01 -8.54960382e-01 4.25025076e-01 -1.22456419e+00
-6.84596300e-01 3.30694616e-01 7.07599521e-01 5.13422251e-01
1.05831742e+00 -2.27782041e-01 3.25508803e-01 7.10399628e-01
3.25970680e-01 -1.39602363e+00 6.46668896e-02 2.98654824e-01
7.44871557e-01 -1.03070641e+00 1.11318797e-01 3.89526010e-01
-7.65540302e-01 6.77013576e-01 4.79843616e-01 1.82978824e-01
3.97434592e-01 -2.21822247e-01 4.17003185e-02 9.07768831e-02
-9.34450209e-01 -9.36701670e-02 2.67867804e-01 1.69707239e-01
1.50999159e-01 3.29150289e-01 -4.78227466e-01 1.69829279e-01
-1.86616600e-01 -7.19723180e-02 1.64199114e-01 7.76962578e-01
-4.92184252e-01 -1.14617682e+00 -6.76485181e-01 8.28031957e-01
-5.71378648e-01 1.15573213e-01 -4.03331608e-01 6.17186010e-01
-2.08321676e-01 9.30667460e-01 -3.86420786e-02 1.15981348e-01
3.31132114e-01 1.60530046e-01 1.61107838e-01 -4.34462667e-01
-3.33101422e-01 3.26288193e-01 3.08246106e-01 -6.11359298e-01
-1.40635803e-01 -8.43904793e-01 -7.91548729e-01 -9.24611092e-03
-6.95235431e-01 7.23867238e-01 6.85327709e-01 9.75298882e-01
-4.28190501e-03 2.70558357e-01 9.21442747e-01 -4.17171121e-01
-9.00967479e-01 -7.31597781e-01 -3.35477293e-01 3.16342354e-01
5.65310538e-01 -8.62214863e-01 -2.18955442e-01 -7.08637059e-01] | [4.562575817108154, 3.2390694618225098] |
03c397e9-c73c-4dcc-a172-1261a03e792e | influencerrank-discovering-effective | 2304.01897 | null | https://arxiv.org/abs/2304.01897v2 | https://arxiv.org/pdf/2304.01897v2.pdf | InfluencerRank: Discovering Effective Influencers via Graph Convolutional Attentive Recurrent Neural Networks | As influencers play considerable roles in social media marketing, companies increase the budget for influencer marketing. Hiring effective influencers is crucial in social influencer marketing, but it is challenging to find the right influencers among hundreds of millions of social media users. In this paper, we propose InfluencerRank that ranks influencers by their effectiveness based on their posting behaviors and social relations over time. To represent the posting behaviors and social relations, the graph convolutional neural networks are applied to model influencers with heterogeneous networks during different historical periods. By learning the network structure with the embedded node features, InfluencerRank can derive informative representations for influencers at each period. An attentive recurrent neural network finally distinguishes highly effective influencers from other influencers by capturing the knowledge of the dynamics of influencer representations over time. Extensive experiments have been conducted on an Instagram dataset that consists of 18,397 influencers with their 2,952,075 posts published within 12 months. The experimental results demonstrate that InfluencerRank outperforms existing baseline methods. An in-depth analysis further reveals that all of our proposed features and model components are beneficial to discover effective influencers. | ['Wei Wang', 'Jinyoung Han', 'Jyun-Yu Jiang', 'Seungbae Kim'] | 2023-04-04 | null | null | null | null | ['marketing'] | ['miscellaneous'] | [ 1.23117767e-01 2.23326251e-01 -6.90166891e-01 -5.41865528e-01
-1.36380866e-01 -4.01060283e-01 9.28985476e-01 1.24194935e-01
-1.72066599e-01 4.85792726e-01 6.86461627e-01 -3.58943433e-01
-5.68406165e-01 -1.30189395e+00 -6.85734272e-01 -3.51199329e-01
-4.87056613e-01 5.81384599e-01 -3.87813188e-02 -7.34882116e-01
3.62247318e-01 2.51471609e-01 -1.29010499e+00 2.32996821e-01
6.78868651e-01 6.62691116e-01 -1.55595556e-01 6.79520130e-01
-6.96692988e-02 1.02684999e+00 -6.63867712e-01 -5.26996374e-01
-1.03836786e-02 -3.04768741e-01 -4.81093228e-01 -1.06382683e-01
-1.77702725e-01 -2.58182794e-01 -9.73515391e-01 7.13785529e-01
1.04460254e-01 2.34623742e-03 5.21193147e-01 -1.06420565e+00
-7.85249591e-01 1.92532468e+00 -6.62263870e-01 6.14032745e-01
3.50173563e-01 -2.45664492e-01 1.73475254e+00 -2.01468393e-01
7.90837646e-01 1.61415040e+00 4.90533471e-01 -9.66910198e-02
-9.99011576e-01 -8.02519858e-01 7.17303336e-01 -6.61229789e-02
-7.28623271e-01 5.84128723e-02 1.00213766e+00 -1.77518144e-01
6.79528475e-01 8.43386948e-02 1.40325618e+00 1.43257093e+00
3.64006400e-01 1.37744713e+00 8.14386964e-01 4.32191670e-01
-4.38876927e-01 -2.04852089e-01 6.94750965e-01 3.45003724e-01
2.55603075e-01 2.01734453e-01 -4.43673402e-01 -3.80613625e-01
6.15409255e-01 7.09408998e-01 -4.34735641e-02 3.47618848e-01
-1.10163450e+00 1.34660208e+00 1.12890494e+00 6.42407358e-01
-4.24137235e-01 4.06108648e-01 1.01219788e-01 8.41184735e-01
1.13862264e+00 7.40976214e-01 -3.40408146e-01 -7.08483458e-02
-5.67227483e-01 9.78141204e-02 8.91790152e-01 4.92956936e-01
9.48594570e-01 -1.63554832e-01 6.17075153e-02 6.10637128e-01
6.25358462e-01 6.06373250e-01 2.43786335e-01 -2.77032405e-01
2.05186576e-01 1.28682375e+00 -4.20211852e-01 -1.62936091e+00
-4.96492267e-01 -8.98428440e-01 -7.39203274e-01 -7.43285477e-01
-1.60195008e-01 -3.08868915e-01 -7.62650907e-01 1.14789104e+00
2.42291361e-01 3.36332947e-01 -4.63451862e-01 6.96966827e-01
9.68350053e-01 9.68055308e-01 -1.00186273e-01 -1.14701632e-02
7.33196139e-01 -7.45713770e-01 -7.03056097e-01 -1.38585016e-01
5.24890780e-01 -4.58099753e-01 6.31919205e-01 -9.72666070e-02
-5.34094036e-01 -1.78980216e-01 -1.14173460e+00 3.20081234e-01
-3.79259139e-01 -4.19416279e-01 1.53066707e+00 -4.28388976e-02
-6.67291403e-01 1.06418979e+00 -3.99234295e-01 9.16057527e-02
7.51362383e-01 5.23503721e-01 1.49953157e-01 -1.19108818e-01
-2.06614947e+00 2.12572321e-01 -3.85666072e-01 8.73647109e-02
-1.13167346e+00 -8.19879353e-01 -3.51463169e-01 -1.27180457e-01
4.53468949e-01 -3.55550945e-01 9.66008067e-01 -1.18025351e+00
-9.53105927e-01 4.32120562e-01 2.56572038e-01 -5.05627751e-01
3.05756062e-01 -7.15861917e-02 -7.69508898e-01 -8.72073025e-02
4.11080830e-02 -5.21241203e-02 9.72754002e-01 -8.65748405e-01
-7.85856724e-01 -4.87565756e-01 4.66351151e-01 2.17227399e-01
-3.04408371e-01 2.62725037e-02 -3.39158446e-01 -2.89458007e-01
1.69866662e-02 -9.34174597e-01 -5.59078217e-01 -1.17292738e+00
-5.62182069e-01 -6.59459889e-01 4.24913436e-01 -1.48269031e-02
1.52454865e+00 -1.79094994e+00 1.97366968e-01 6.54609978e-01
9.71390009e-01 -1.05377980e-01 -2.48490512e-01 2.26544604e-01
3.30768168e-01 5.58045864e-01 3.24994415e-01 3.71071011e-01
-1.49027288e-01 -1.65999755e-01 -3.33844990e-01 4.36804682e-01
-1.27182364e-01 1.28213608e+00 -1.09987855e+00 -3.59890275e-02
-1.58456475e-01 9.50604677e-01 -1.48560405e-01 -9.11921859e-02
-3.26247245e-01 2.48308331e-01 -1.64212155e+00 6.92686915e-01
7.15398043e-02 -6.05598390e-01 1.95672318e-01 3.42160583e-01
2.70414799e-01 7.16510773e-01 -5.19882083e-01 9.15617466e-01
-2.94889420e-01 7.63587296e-01 1.76004663e-01 -6.15407586e-01
1.10244787e+00 -1.59705266e-01 8.20054889e-01 -6.33450091e-01
6.15382314e-01 1.92652658e-01 3.19181085e-01 6.49005175e-02
6.79728389e-01 2.78082997e-01 -5.05981982e-01 1.21789849e+00
-4.19456899e-01 1.69248879e-01 4.46415901e-01 9.82075989e-01
1.62950444e+00 -6.03405535e-01 -2.73976922e-02 -8.42400491e-02
4.77725230e-02 -1.52207851e-01 2.99621433e-01 8.13093662e-01
-5.70402667e-02 1.14567563e-01 8.36747289e-01 -7.43885279e-01
-6.50160670e-01 -5.41605294e-01 -1.00783706e-01 1.53687012e+00
6.90349638e-02 -3.95898193e-01 -2.54468799e-01 -9.00822580e-01
6.26135886e-01 2.06587657e-01 -1.20455515e+00 -2.56501496e-01
-7.45925963e-01 -1.09943748e+00 1.41405523e-01 1.06847160e-01
2.25953847e-01 -1.35631239e+00 4.62500840e-01 4.59932804e-01
7.12479055e-02 -4.07419264e-01 -4.51059699e-01 -1.06769949e-01
-8.34493637e-01 -1.40083075e+00 -4.50368255e-01 -1.90511644e-01
7.07665563e-01 6.69653893e-01 1.64265168e+00 6.09366119e-01
2.43657120e-02 3.16605330e-01 -5.85588515e-01 -3.74687612e-01
-2.84343660e-01 8.46788406e-01 -1.05212480e-01 3.86348218e-01
8.86046350e-01 -8.76108766e-01 -7.52918601e-01 6.03482485e-01
-7.12903321e-01 -3.33277136e-01 8.61135900e-01 1.09506346e-01
3.42860103e-01 -1.95082761e-02 8.07329357e-01 -1.75960076e+00
1.25071359e+00 -1.02929044e+00 -2.01777697e-01 -1.05681375e-01
-1.14105070e+00 -1.81759730e-01 2.28208661e-01 -6.82109952e-01
-6.29834652e-01 -5.08089066e-01 2.93157339e-01 2.18574390e-01
5.91839254e-01 8.76099586e-01 2.36743554e-01 4.09985572e-01
6.10129774e-01 -1.80628359e-01 -2.05859900e-01 -2.95998544e-01
4.99106228e-01 8.36611211e-01 -5.14180481e-01 1.99705511e-02
7.31500447e-01 3.37365091e-01 -4.80699420e-01 -5.89113951e-01
-1.34107304e+00 -5.36088526e-01 -2.12737769e-01 -5.60389817e-01
1.28597304e-01 -9.96706247e-01 -8.42156947e-01 4.68716651e-01
-6.57978714e-01 1.65902004e-01 -8.55944157e-02 4.97840166e-01
3.13891590e-01 -4.48347628e-01 -9.67599392e-01 -4.25834090e-01
-6.04901135e-01 -8.21155131e-01 5.93276501e-01 2.59086162e-01
-2.52649754e-01 -1.08520436e+00 1.39736459e-01 4.51398432e-01
8.27677310e-01 4.34509993e-01 4.33593750e-01 -9.15140390e-01
-1.09500575e+00 -6.76800311e-01 -9.45444182e-02 -3.11751932e-01
1.24739252e-01 2.85303622e-01 -3.84776175e-01 -2.13193282e-01
-5.82911253e-01 9.95459408e-03 1.64753389e+00 9.36340213e-01
6.16171539e-01 -3.37755322e-01 -8.61463845e-01 2.17481494e-01
6.06288791e-01 4.47350405e-02 3.60675216e-01 2.14631855e-01
1.13528001e+00 4.95779097e-01 4.76012796e-01 4.37041581e-01
4.22235966e-01 -3.01598245e-03 6.05910003e-01 -1.57971606e-01
3.43428999e-01 -6.26187325e-01 6.11454248e-01 1.24325180e+00
-4.50037181e-01 -1.94185987e-01 -4.86815512e-01 5.41952670e-01
-1.77137721e+00 -9.24351275e-01 -4.56255227e-01 1.58509636e+00
6.53333962e-01 4.97860670e-01 1.83207214e-01 -9.11571458e-02
1.13174164e+00 9.64478910e-01 -1.02644336e+00 -1.60211906e-01
-2.44646758e-01 -1.63738698e-01 5.20888150e-01 6.73579872e-01
-1.03458750e+00 8.46871912e-01 5.81636095e+00 5.73823571e-01
-7.70387053e-01 -9.85927433e-02 9.92603958e-01 -5.63628316e-01
-9.24126923e-01 -6.39286116e-02 -1.15596235e+00 3.20228606e-01
1.12539041e+00 -6.09069586e-01 6.29390955e-01 8.63158166e-01
1.05151199e-01 4.70619529e-01 -7.18468666e-01 2.47982472e-01
1.98573872e-01 -1.41731334e+00 2.47945320e-02 4.33593214e-01
1.16442919e+00 7.33710587e-01 1.52358174e-01 5.73863208e-01
9.79932010e-01 -1.15784144e+00 -3.99734974e-02 5.76075852e-01
5.69249317e-02 -1.16059041e+00 6.52747154e-01 -2.13795919e-02
-1.15242124e+00 -2.16759130e-01 -5.53331852e-01 -3.53224754e-01
2.05781162e-01 1.47068322e+00 -8.17582428e-01 2.57238120e-01
3.66215348e-01 1.98217618e+00 -4.98627663e-01 1.82720810e-01
-6.58677816e-01 9.85780120e-01 -1.38919592e-01 -9.48318422e-01
4.32438135e-01 -4.18600470e-01 6.82114244e-01 6.56366348e-01
8.54925662e-02 -1.87669262e-01 1.33993044e-01 6.47402465e-01
-7.61683166e-01 2.67535418e-01 -8.62262964e-01 -9.96345758e-01
3.04959118e-01 1.60396576e+00 -6.86986744e-01 -2.35988885e-01
-1.02615459e-02 1.75611511e-01 2.00412482e-01 2.94835418e-01
-7.48450935e-01 -4.02607381e-01 7.61372924e-01 8.31390858e-01
6.35906383e-02 2.22074777e-01 2.90814757e-01 -8.69466245e-01
-5.17687798e-01 -7.10700452e-01 3.95523459e-01 -3.14603120e-01
-1.57321036e+00 1.01570332e+00 -4.77610528e-01 -7.13483870e-01
-1.64198875e-01 -2.28230670e-01 -4.93454695e-01 3.77533466e-01
-1.43332005e+00 -1.11933410e+00 6.36145845e-02 3.15940589e-01
2.53027111e-01 -2.81429321e-01 1.15074657e-01 1.29383564e-01
-4.36374009e-01 -1.75343808e-02 -1.79997161e-01 3.48078549e-01
2.02018186e-01 -1.08912396e+00 6.07137144e-01 6.48583695e-02
4.37580764e-01 8.51037383e-01 3.39899689e-01 -8.98135781e-01
-1.49748731e+00 -1.15090919e+00 1.06010389e+00 -5.73924363e-01
1.22149599e+00 -2.07227677e-01 -4.55894113e-01 7.96669900e-01
2.09130228e-01 -3.99816930e-01 8.77343774e-01 1.07946503e+00
-2.46085793e-01 -2.87443459e-01 -5.44142306e-01 4.69325960e-01
1.35831881e+00 -3.59967321e-01 -4.11864877e-01 5.66125333e-01
1.33490932e+00 1.14986457e-01 -9.63784277e-01 3.24458897e-01
5.44936776e-01 -6.92345321e-01 6.63757980e-01 -7.79344618e-01
1.04971337e+00 8.45709722e-03 3.51888418e-01 -1.64815867e+00
-4.98767227e-01 -7.65065253e-01 -3.21529329e-01 1.08253253e+00
1.08774948e+00 -7.40693569e-01 6.52551830e-01 2.44852424e-01
3.68153453e-01 -9.76942301e-01 -2.65050352e-01 3.33067238e-01
-2.26742715e-01 -4.06504154e-01 1.02301502e+00 1.11212802e+00
-2.34522313e-01 1.09690273e+00 -4.25862879e-01 -5.27907491e-01
2.95052946e-01 6.26083076e-01 6.20132148e-01 -2.01121378e+00
-3.29243451e-01 -5.75256169e-01 -2.77612388e-01 -1.13606048e+00
3.36680353e-01 -1.20213699e+00 -5.16106188e-01 -1.71827066e+00
6.10484600e-01 -6.61532402e-01 -8.75244379e-01 2.14709252e-01
-1.35514334e-01 9.22069624e-02 -3.02576452e-01 4.94338602e-01
-9.88125384e-01 5.01156926e-01 2.01272440e+00 -7.15359092e-01
-3.88442546e-01 6.49750292e-01 -1.40569592e+00 6.25334740e-01
5.98903537e-01 -7.60003209e-01 -5.95378160e-01 -8.84454250e-02
1.42587805e+00 -2.10409343e-01 -4.82463151e-01 -9.93475765e-02
-1.34515375e-01 1.95232183e-02 1.36302754e-01 -6.40666306e-01
-6.63774237e-02 -3.90716523e-01 8.50763917e-02 2.60431886e-01
-9.80968595e-01 -6.01848140e-02 -6.36554420e-01 1.23611653e+00
-1.92105547e-01 4.03650671e-01 1.45710662e-01 -1.69000283e-01
-4.57139648e-02 9.60197687e-01 -4.42095190e-01 1.54956449e-02
6.70007765e-01 3.85412037e-01 -4.26428050e-01 -9.07480180e-01
-7.35617578e-01 4.45373088e-01 6.00346588e-02 9.35702920e-01
5.21972299e-01 -1.34115899e+00 -1.17415082e+00 2.55386345e-02
-1.26377000e-02 -1.22214839e-01 9.84372199e-02 7.87682533e-01
1.86351657e-01 1.02690510e-01 3.84631574e-01 -3.76291752e-01
-1.23088849e+00 1.26642123e-01 1.04516270e-02 -8.33489001e-01
-5.25971115e-01 8.99027228e-01 -1.42478913e-01 -7.31531978e-01
-1.29215971e-01 -2.05332726e-01 -8.95765901e-01 6.98821247e-01
7.34157383e-01 3.12238544e-01 -3.21494579e-01 -6.67680085e-01
-2.26293020e-02 1.46768421e-01 -6.67600453e-01 3.81885111e-01
1.99111414e+00 2.02679727e-02 -4.33313340e-01 5.36528349e-01
1.39091504e+00 -9.41060483e-02 -9.31553721e-01 -3.56554300e-01
-1.65182233e-01 -3.74436855e-01 4.50267494e-01 -6.33340418e-01
-1.74335837e+00 4.77632552e-01 -1.96931884e-01 8.10830355e-01
5.30082405e-01 3.98078799e-01 9.10211861e-01 4.58381116e-01
2.29285851e-01 -1.04248643e+00 7.62431771e-02 8.24778736e-01
7.78165400e-01 -1.13518620e+00 4.02127028e-01 -2.58323371e-01
-5.23543537e-01 6.12413168e-01 1.25987560e-01 -5.40849328e-01
1.35641801e+00 6.65038303e-02 -9.69595611e-02 -8.55885208e-01
-1.04772723e+00 -1.93338707e-01 2.80375421e-01 -1.21994078e-01
7.23321080e-01 6.55150235e-01 -4.73493606e-01 7.04166710e-01
-3.39460999e-01 -5.02956331e-01 5.08634508e-01 4.97035086e-01
-6.24389112e-01 -9.32098687e-01 1.96492121e-01 1.15888596e+00
-6.85794234e-01 -1.09667800e-01 -1.12616014e+00 6.37383044e-01
-5.04245281e-01 1.10088885e+00 7.77090341e-03 -9.84744966e-01
8.58861133e-02 -5.26425123e-01 -8.09055418e-02 -3.91502619e-01
-1.01714730e+00 2.37730388e-02 3.60330552e-01 -5.59523582e-01
-6.39507592e-01 -6.24118090e-01 -1.14537430e+00 -9.64658678e-01
-5.28441072e-01 2.44777933e-01 7.10753262e-01 8.62496793e-01
5.74728906e-01 1.08359575e+00 1.37341440e+00 -3.96629691e-01
-1.97842285e-01 -1.41912198e+00 -1.06822217e+00 5.95315814e-01
3.27473879e-02 -6.27054870e-01 -9.10740376e-01 -7.88733959e-01] | [7.306843280792236, 6.067253112792969] |
ec01a7e8-0017-4713-ba12-ff59cec94ecb | tech-report-a-fast-multiscale-spatial | 1712.01770 | null | http://arxiv.org/abs/1712.01770v3 | http://arxiv.org/pdf/1712.01770v3.pdf | Tech Report: A Fast Multiscale Spatial Regularization for Sparse Hyperspectral Unmixing | Sparse hyperspectral unmixing from large spectral libraries has been
considered to circumvent limitations of endmember extraction algorithms in many
applications. This strategy often leads to ill-posed inverse problems, which
can benefit from spatial regularization strategies. While existing spatial
regularization methods improve the problem conditioning and promote piecewise
smooth solutions, they lead to large nonsmooth optimization problems. Thus,
efficiently introducing spatial context in the unmixing problem remains a
challenge, and a necessity for many real world applications. In this paper, a
novel multiscale spatial regularization approach for sparse unmixing is
proposed. The method uses a signal-adaptive spatial multiscale decomposition
based on superpixels to decompose the unmixing problem into two simpler
problems, one in the approximation domain and another in the original domain.
Simulation results using both synthetic and real data indicate that the
proposed method can outperform state-of-the-art Total Variation-based
algorithms with a computation time comparable to that of their unregularized
counterparts. | ['Cédric Richard', 'José Carlos Moreira Bermudez', 'Tales Imbiriba', 'Ricardo Augusto Borsoi'] | 2017-12-05 | null | null | null | null | ['hyperspectral-unmixing'] | ['computer-vision'] | [ 6.07074738e-01 -7.16355979e-01 1.22577228e-01 2.15600263e-02
-1.06108081e+00 -4.98119354e-01 2.95505643e-01 -2.22555816e-01
-7.64832869e-02 1.06455481e+00 5.96006699e-02 -1.25474766e-01
-2.95696139e-01 -5.46644628e-01 -4.61698890e-01 -1.47949719e+00
4.31834072e-01 9.98218283e-02 -6.37651086e-02 -6.43964261e-02
-1.47777963e-02 4.11194712e-01 -1.41914535e+00 1.51956528e-01
1.55703855e+00 8.13656449e-01 4.68202084e-01 1.41423464e-01
-9.51802582e-02 5.74528873e-01 -2.80104041e-01 5.54032087e-01
4.33929324e-01 -5.98151863e-01 -2.06129462e-01 5.81991792e-01
4.76081461e-01 1.18395880e-01 1.29719004e-02 1.47366047e+00
4.53849792e-01 5.85021079e-01 6.77367449e-01 -7.88316369e-01
-5.09292543e-01 1.75201997e-01 -1.30074203e+00 2.63495110e-02
-2.53913462e-01 -7.10793212e-02 5.39518833e-01 -1.11416411e+00
2.35880867e-01 9.48081791e-01 8.61710548e-01 -2.01795310e-01
-1.70578444e+00 -6.59509718e-01 8.96498784e-02 -3.01000662e-02
-1.67759585e+00 -4.53668326e-01 1.04649901e+00 -6.65764093e-01
4.36883718e-01 5.63359261e-01 4.86670703e-01 3.60352427e-01
-2.61885673e-01 4.44314182e-01 1.76694250e+00 -4.64357972e-01
2.99923539e-01 -1.20995447e-01 3.19639236e-01 2.80542105e-01
6.53151155e-01 -1.24332542e-02 -2.42399395e-01 -3.99030030e-01
7.58180141e-01 9.81238410e-02 -8.22123766e-01 -3.86459827e-01
-1.04038393e+00 8.03220093e-01 4.83218431e-01 4.14136380e-01
-8.01423609e-01 -2.14094371e-01 -2.45551437e-01 -2.93628667e-02
9.17142212e-01 1.79432422e-01 -7.69862086e-02 5.73005438e-01
-1.62695980e+00 3.39925110e-01 4.30906683e-01 4.89620298e-01
9.56943452e-01 4.54935282e-01 -1.05815612e-01 1.21576869e+00
6.15162373e-01 9.49990630e-01 1.29311040e-01 -8.70031476e-01
2.82764137e-01 5.05978644e-01 3.78677994e-01 -1.14308107e+00
-1.84212431e-01 -7.08999038e-01 -1.24417782e+00 2.85439074e-01
5.12108922e-01 -8.48999247e-02 -9.56471324e-01 1.38715446e+00
5.12161613e-01 5.46571374e-01 -4.77572065e-03 1.22119296e+00
5.54919362e-01 1.11866093e+00 4.67241928e-02 -8.35303485e-01
9.19610679e-01 -9.62669969e-01 -1.08447742e+00 -4.55563456e-01
1.89769551e-01 -1.01814067e+00 5.86874247e-01 4.67808545e-01
-8.20953548e-01 -3.58370185e-01 -9.60289478e-01 2.82116920e-01
-3.19894180e-02 3.28506172e-01 3.02678436e-01 5.66529810e-01
-6.01140916e-01 4.88463342e-01 -7.81039178e-01 -2.28877828e-01
2.79249728e-01 1.62416827e-02 -1.20314404e-01 -3.83588284e-01
-8.06883514e-01 6.10009074e-01 2.19840989e-01 5.53372502e-01
-6.55122697e-01 -6.00632906e-01 -6.76695764e-01 -1.20254397e-01
3.65884364e-01 -3.81536216e-01 5.15709758e-01 -1.24436867e+00
-1.26154840e+00 5.65722167e-01 -4.90043640e-01 5.07635549e-02
2.52035588e-01 -1.47526816e-01 -5.78046858e-01 1.36235729e-01
2.58174658e-01 -1.50610670e-01 1.29555321e+00 -1.65448844e+00
-2.98303008e-01 -4.03948754e-01 -6.49939418e-01 2.96971411e-01
-2.37656265e-01 -2.30128810e-01 -1.60595745e-01 -1.14958465e+00
6.96517766e-01 -8.80252600e-01 -6.32297099e-01 2.84356494e-02
-3.26577216e-01 5.97441673e-01 9.71284986e-01 -1.17463768e+00
1.19116640e+00 -2.27698135e+00 4.66878921e-01 3.35868984e-01
-1.03720099e-01 3.58258545e-01 -1.82376713e-01 2.50760525e-01
-3.52956742e-01 -2.96832919e-01 -1.13246346e+00 1.82469357e-02
-4.92990762e-01 7.73355067e-02 -1.56417236e-01 1.15228033e+00
-1.22970358e-01 3.67102534e-01 -8.00224066e-01 -3.01275939e-01
4.27997679e-01 7.48460114e-01 -2.83340544e-01 2.49563918e-01
-1.61300097e-02 8.51462901e-01 -3.01371425e-01 6.73150301e-01
1.33212316e+00 -2.96881676e-01 3.01589608e-01 -4.32686597e-01
-5.74419022e-01 -5.29019713e-01 -1.64013171e+00 1.62842524e+00
-3.46743345e-01 4.14524227e-01 1.17309475e+00 -1.42714071e+00
6.95277631e-01 3.49080771e-01 5.90076089e-01 -1.62289336e-01
-1.02843314e-01 5.79327524e-01 -2.69119710e-01 -5.43077290e-01
2.58025020e-01 -8.66367579e-01 6.89968050e-01 2.32618690e-01
-3.52211356e-01 -2.81452239e-01 1.11982763e-01 -2.19444469e-01
5.09176552e-01 2.34932095e-01 3.92639786e-01 -8.90861213e-01
8.28326583e-01 2.75693685e-01 6.79817259e-01 4.02620286e-01
7.40076823e-04 6.83135509e-01 -1.66389331e-01 -6.43490627e-02
-6.07236028e-01 -8.42775643e-01 -4.26135540e-01 6.00532711e-01
3.54223371e-01 1.88036472e-01 -7.08438933e-01 -4.32082377e-02
-7.10376650e-02 6.69059455e-01 -1.50315911e-01 3.11022967e-01
-5.03655314e-01 -1.59087884e+00 -7.54557922e-03 8.27421844e-02
7.64560640e-01 -4.09798145e-01 -2.06355199e-01 4.45539504e-01
-6.73856139e-01 -9.27509725e-01 -2.77173966e-01 5.78226857e-02
-1.23621738e+00 -1.00619900e+00 -1.16973114e+00 -5.94994426e-01
8.21985543e-01 9.19055521e-01 8.40186119e-01 -2.71816790e-01
-2.69009620e-01 1.24065340e-01 -2.85658419e-01 -8.51774588e-02
-1.56792283e-01 -3.41661572e-01 -1.03120148e-01 6.00245237e-01
1.11105636e-01 -7.48184800e-01 -5.95519960e-01 4.44015294e-01
-1.27661729e+00 -3.46545801e-02 5.86329103e-01 1.01866961e+00
9.59415138e-01 3.39113355e-01 5.15908301e-01 -7.70157814e-01
4.72091913e-01 -5.37739396e-01 -8.61622930e-01 3.09127778e-01
-5.57377279e-01 -1.57038108e-01 4.84632403e-01 -3.60144943e-01
-1.42592883e+00 3.68613690e-01 2.51610845e-01 -4.21128631e-01
-9.96368304e-02 8.46190453e-01 -3.39050531e-01 -4.75855470e-01
7.84482062e-01 5.93207002e-01 1.92946628e-01 -7.77532160e-01
3.68260145e-01 6.32755995e-01 3.58859867e-01 -5.62575102e-01
1.00200999e+00 8.54855180e-01 2.00105593e-01 -1.67682326e+00
-7.86249280e-01 -9.89936233e-01 -4.84882206e-01 -1.75433025e-01
7.37180650e-01 -1.18405855e+00 8.41430128e-02 6.07231855e-01
-9.05232668e-01 -2.44916961e-01 -2.51139820e-01 4.69431520e-01
-2.11617842e-01 7.07669914e-01 -2.83668607e-01 -1.00357842e+00
-1.32654399e-01 -1.08765674e+00 7.72636414e-01 9.13137868e-02
1.23082586e-01 -9.60886121e-01 1.61781684e-01 5.73205233e-01
6.50790334e-01 5.20639300e-01 6.86163127e-01 1.05615005e-01
-5.37440419e-01 -1.02844335e-01 -4.31575000e-01 5.40534139e-01
3.62020135e-01 -2.96040922e-01 -1.07758164e+00 -4.44451541e-01
6.14300728e-01 6.99876100e-02 1.01158822e+00 1.00428140e+00
8.22410524e-01 -2.25562423e-01 -3.81352186e-01 7.67874122e-01
1.88916826e+00 -3.53259593e-02 6.09052062e-01 1.66595817e-01
7.84523427e-01 6.63902223e-01 5.54252684e-01 3.91615033e-01
-3.58469605e-01 4.39111084e-01 4.02916819e-01 -5.40556371e-01
-1.06186479e-01 3.04675817e-01 1.34870857e-01 8.84555399e-01
-3.29400361e-01 -1.14503793e-01 -7.13540912e-01 7.08873034e-01
-2.02966070e+00 -9.92343485e-01 -9.35844362e-01 2.27049685e+00
7.10737824e-01 -8.45713317e-01 -2.99053907e-01 2.51676947e-01
1.02377760e+00 5.97034276e-01 -3.78366858e-01 3.46979171e-01
-5.59794605e-01 9.98523980e-02 6.46670341e-01 8.17703247e-01
-1.28700101e+00 6.42980516e-01 6.07068872e+00 1.02383280e+00
-1.16578484e+00 5.43610811e-01 4.56506371e-01 5.69588616e-02
-2.05780178e-01 6.10214435e-02 -2.47583911e-01 3.40082288e-01
4.52800155e-01 1.91786617e-01 7.43808448e-01 3.26104075e-01
6.90420628e-01 -3.81113857e-01 -3.33247513e-01 1.16752839e+00
9.11636744e-03 -1.01233363e+00 -6.39020503e-02 -1.03562303e-01
1.29476345e+00 -1.36836246e-01 -1.25307530e-01 -2.47129560e-01
-1.31327137e-01 -1.09551919e+00 5.02380908e-01 5.19733608e-01
7.14438915e-01 -4.66343313e-01 5.27697504e-01 4.56889689e-01
-1.28848791e+00 7.19957575e-02 -3.47234190e-01 -2.68114865e-01
4.07352835e-01 1.37476242e+00 6.64126873e-02 7.59492815e-01
4.51670110e-01 7.99496710e-01 -2.60810629e-02 1.11342645e+00
-3.03933889e-01 8.65980804e-01 -4.52290833e-01 5.19775391e-01
1.47964284e-01 -1.11677468e+00 1.01567125e+00 1.12349916e+00
7.22609639e-01 5.49911439e-01 1.82987779e-01 1.14758492e+00
1.87508091e-01 3.75428528e-01 -5.82859874e-01 -8.54961425e-02
1.74563006e-01 1.33428705e+00 -6.81137264e-01 -2.58639783e-01
-5.99353254e-01 9.76501167e-01 -2.97165781e-01 1.03111708e+00
-6.45244181e-01 -6.44115582e-02 5.84501982e-01 3.02071333e-01
2.58540809e-01 -3.31794888e-01 -4.97204214e-01 -1.36678171e+00
-1.12193532e-01 -1.07080781e+00 3.29069942e-01 -6.13797724e-01
-1.22743368e+00 3.03336591e-01 -2.28182867e-01 -1.38501191e+00
4.91739661e-01 -6.41556203e-01 -4.96434748e-01 1.32903171e+00
-1.73109889e+00 -1.23603606e+00 -5.48269451e-01 5.39854467e-01
4.87210840e-01 1.32638127e-01 6.54416740e-01 4.37863261e-01
-7.39834249e-01 -3.68474603e-01 8.71972203e-01 -4.47264701e-01
6.42523468e-01 -9.04329896e-01 -6.63815737e-01 1.47119093e+00
-1.70693219e-01 4.84438300e-01 9.09560859e-01 -7.18400121e-01
-1.26124942e+00 -1.33513021e+00 3.87975961e-01 2.32567608e-01
5.35565376e-01 2.99471885e-01 -1.17042637e+00 4.34206307e-01
3.33546162e-01 3.57784808e-01 8.93379271e-01 -2.95641333e-01
5.54113872e-02 -2.69819617e-01 -1.27694762e+00 3.30563754e-01
4.72666085e-01 -2.30660692e-01 -3.27279925e-01 4.49948370e-01
6.79235533e-02 -9.79282632e-02 -6.49149597e-01 5.30755579e-01
2.55622596e-01 -9.13392365e-01 1.15876961e+00 -8.21572319e-02
2.76357651e-01 -8.35265636e-01 -4.60640430e-01 -1.50833988e+00
-6.01613343e-01 -5.93004704e-01 -5.31840846e-02 1.11885846e+00
3.35304916e-01 -7.11237252e-01 4.31708813e-01 3.87188882e-01
-1.22768581e-01 -2.13189065e-01 -7.96351850e-01 -8.43907118e-01
4.46499400e-02 -7.34560117e-02 7.58905262e-02 1.35812724e+00
-3.40054065e-01 1.39537185e-01 -5.29281914e-01 6.57411277e-01
1.31989741e+00 5.95805347e-01 2.91052401e-01 -1.27841222e+00
-2.19264552e-01 -3.21573853e-01 2.59385079e-01 -9.13544476e-01
1.51681438e-01 -8.66566122e-01 4.55325454e-01 -1.67585599e+00
1.12800151e-01 -3.50764304e-01 -3.57656121e-01 1.17108405e-01
-4.46008563e-01 4.22540873e-01 -2.05912858e-01 4.79675055e-01
1.43620610e-01 5.73103011e-01 1.20116317e+00 -5.67149937e-01
-4.43604171e-01 -2.34563306e-01 -5.28892696e-01 8.47316265e-01
6.17755294e-01 -4.60120231e-01 -3.59688580e-01 -5.03734648e-01
-4.54958491e-02 1.41017273e-01 3.65458548e-01 -1.05303264e+00
9.48430412e-03 -5.86035609e-01 1.37302294e-01 -4.57952529e-01
3.52187812e-01 -1.12330914e+00 6.70142055e-01 2.45722964e-01
1.97071880e-01 -6.85974121e-01 3.25298905e-01 7.22002208e-01
-5.32792866e-01 -3.48075956e-01 1.29626203e+00 -1.52915806e-01
-5.36743283e-01 9.95888934e-02 -4.73742247e-01 -2.36641377e-01
9.39706624e-01 -1.45069405e-01 -3.71232256e-02 -1.98853865e-01
-7.24511445e-01 -9.91621315e-02 5.11057615e-01 -4.13217038e-01
5.01851082e-01 -1.31041765e+00 -8.93803179e-01 1.20440185e-01
-1.75715774e-01 1.17491238e-01 4.95560467e-01 1.31616473e+00
-7.57093787e-01 9.70067978e-02 3.14467587e-02 -6.46759868e-01
-1.08975291e+00 5.19691229e-01 4.35113102e-01 -1.80396080e-01
-3.88338983e-01 7.32567132e-01 3.74931067e-01 -3.22708458e-01
-2.50259310e-01 -1.52577832e-01 -8.66735354e-02 2.36346573e-01
4.97884482e-01 8.04833710e-01 -4.32310104e-02 -9.78868127e-01
-1.86104015e-01 8.05188715e-01 7.32098043e-01 -6.49230927e-02
1.56010187e+00 -2.99174905e-01 -8.60275030e-01 3.51290375e-01
9.82039332e-01 2.63706326e-01 -1.23928487e+00 -2.25227058e-01
-2.02876195e-01 -7.08097637e-01 6.43869579e-01 -5.77858210e-01
-1.17460549e+00 9.40093815e-01 7.35530794e-01 2.88103521e-01
1.41764224e+00 -5.68149507e-01 5.17718315e-01 8.26041847e-02
9.03769284e-02 -1.05494952e+00 -2.99955159e-01 1.83587179e-01
1.06287801e+00 -1.31164491e+00 4.99586076e-01 -9.49804008e-01
-1.01552747e-01 8.93626153e-01 1.82998225e-01 -1.01556934e-01
7.77141511e-01 7.00785369e-02 1.97456852e-01 -8.31125602e-02
3.00678611e-01 -2.26682708e-01 3.71593744e-01 4.37714338e-01
5.87496758e-01 1.86696887e-01 -5.38762331e-01 5.42379618e-01
7.81454504e-01 -5.63099049e-02 1.17151901e-01 8.30113292e-01
-6.47284865e-01 -9.17482197e-01 -1.35327744e+00 4.33606088e-01
-3.56066734e-01 -3.01912010e-01 4.72340211e-02 4.19145942e-01
1.10901408e-01 1.19442832e+00 -4.36616391e-01 2.70635456e-01
1.01166658e-01 9.63945314e-02 3.05742770e-01 -4.27648395e-01
-1.95371099e-02 7.65355647e-01 -1.22119583e-01 -2.17726141e-01
-1.06487048e+00 -6.48444772e-01 -9.27555501e-01 8.31265226e-02
-7.18457043e-01 2.70319402e-01 4.54947114e-01 8.28311503e-01
4.64188941e-02 3.77805501e-01 4.79409367e-01 -1.25152361e+00
-4.63422567e-01 -1.14818311e+00 -1.23497868e+00 3.53237569e-01
4.33690757e-01 -6.46830857e-01 -6.05841160e-01 3.69118840e-01] | [10.144158363342285, -2.062161445617676] |
e278d6bd-d62c-45c6-9536-d30920915bd7 | using-answer-set-programming-for-pattern | 1409.7777 | null | http://arxiv.org/abs/1409.7777v1 | http://arxiv.org/pdf/1409.7777v1.pdf | Using Answer Set Programming for pattern mining | Serial pattern mining consists in extracting the frequent sequential patterns
from a unique sequence of itemsets. This paper explores the ability of a
declarative language, such as Answer Set Programming (ASP), to solve this issue
efficiently. We propose several ASP implementations of the frequent sequential
pattern mining task: a non-incremental and an incremental resolution. The
results show that the incremental resolution is more efficient than the
non-incremental one, but both ASP programs are less efficient than dedicated
algorithms. Nonetheless, this approach can be seen as a first step toward a
generic framework for sequential pattern mining with constraints. | ['René Quiniou', 'Yves Moinard', 'Thomas Guyet'] | 2014-09-27 | null | null | null | null | ['sequential-pattern-mining'] | ['natural-language-processing'] | [ 2.66732126e-01 3.75574142e-01 -4.55579698e-01 -6.25333846e-01
-5.87901706e-03 -4.81339186e-01 1.33460850e-01 3.99727851e-01
-1.96346909e-01 9.24753129e-01 -1.91121355e-01 -5.55969536e-01
-5.30772388e-01 -1.44943416e+00 -5.82614601e-01 -5.98592274e-02
-5.41211486e-01 1.08529639e+00 9.81120110e-01 -1.40779227e-01
2.51618952e-01 4.14975613e-01 -2.10567403e+00 9.69035745e-01
6.93997622e-01 1.11601484e+00 -5.17923385e-03 3.71713489e-01
-6.97740614e-01 9.80489075e-01 -1.68161631e-01 -2.11307034e-01
4.60618764e-01 -3.09057325e-01 -1.16430306e+00 -6.44507036e-02
-3.02865654e-01 -8.65415558e-02 3.69502991e-01 8.87582242e-01
-3.57721031e-01 -1.27658561e-01 -1.56747758e-01 -1.53885305e+00
1.31303281e-01 9.19187605e-01 -5.97188294e-01 2.70719588e-01
1.16721582e+00 -1.87849805e-01 1.16976666e+00 -8.43118429e-01
9.84693885e-01 8.10792208e-01 7.55110502e-01 8.70515630e-02
-8.70773554e-01 -3.64509910e-01 2.66052514e-01 4.85083997e-01
-1.27660859e+00 -2.11556360e-01 2.86890209e-01 -6.71076030e-02
1.68639898e+00 9.59393859e-01 1.25237429e+00 1.57083541e-01
-7.36802965e-02 8.02645802e-01 1.08276975e+00 -8.44034910e-01
4.21614498e-01 3.10839266e-01 8.87663364e-01 5.70065618e-01
1.13301790e+00 2.88527161e-01 -6.82504654e-01 -7.76127219e-01
2.14177683e-01 -4.07919735e-02 -6.68649301e-02 -4.37139809e-01
-6.80851460e-01 8.27916324e-01 -5.33755660e-01 3.67805034e-01
-5.49404323e-01 -3.45094502e-01 6.18899643e-01 9.91099656e-01
-6.15127198e-02 4.02965635e-01 -6.22331500e-01 -1.17873743e-01
-8.57578754e-01 8.84891748e-01 1.70310533e+00 1.29011750e+00
7.02278197e-01 -1.84487984e-01 2.52160877e-01 1.94981858e-01
2.57734895e-01 4.11601923e-02 8.42287242e-02 -6.65747225e-01
1.18064947e-01 1.13141918e+00 4.68715876e-01 -8.63186717e-01
-2.28957415e-01 1.29053608e-01 7.58370683e-02 4.22105342e-02
3.24950904e-01 -8.23785067e-02 -4.08871323e-01 1.14143825e+00
5.67752421e-01 -1.46400273e-01 1.24755159e-01 6.15560174e-01
4.75040406e-01 7.50561118e-01 -3.93305123e-02 -1.18723917e+00
1.22705054e+00 -7.16271281e-01 -7.67258763e-01 7.60490075e-02
5.10356605e-01 -4.63438958e-01 4.87082511e-01 9.39304829e-01
-1.47626615e+00 -3.47134061e-02 -1.28319788e+00 5.83320439e-01
-2.97545493e-01 -6.94588006e-01 1.14115572e+00 8.55897963e-01
-7.57223010e-01 3.21389705e-01 -7.10397065e-01 -3.79202127e-01
1.31918406e-02 6.65282607e-01 -2.46570364e-01 -1.12572804e-01
-9.08191383e-01 7.81087935e-01 1.05866146e+00 -3.33810180e-01
1.11132905e-01 -7.21595109e-01 -5.87443709e-01 1.19759716e-01
8.84076118e-01 -6.39907241e-01 1.34447670e+00 -1.10282493e+00
-1.02101970e+00 9.85705912e-01 -3.49943906e-01 -1.00886714e+00
3.96829724e-01 -6.23957440e-03 -6.93695247e-01 2.19196483e-01
-5.64992912e-02 -1.74190983e-01 1.38617650e-01 -7.87132680e-01
-1.40999198e+00 -4.16810840e-01 3.47358510e-02 -6.55306354e-02
-1.08701818e-01 7.06196129e-01 -4.46108356e-02 -1.86849877e-01
2.09689215e-02 -5.19022048e-01 -4.96789753e-01 -3.49504203e-01
-2.49588583e-02 -5.42771637e-01 8.02107871e-01 -1.18901439e-01
1.72567070e+00 -1.66869259e+00 -2.40534917e-01 7.80345321e-01
-2.07676589e-01 3.62677849e-03 4.18446273e-01 6.89344227e-01
-1.48345396e-01 8.11526552e-02 -3.10325623e-01 6.08647346e-01
-2.50386745e-01 6.90392494e-01 -4.23804790e-01 -4.13489528e-03
7.64120892e-02 5.25604129e-01 -6.69817448e-01 -6.18892729e-01
-1.99900851e-01 -6.76206827e-01 -7.75467038e-01 5.68307899e-02
-8.69413912e-01 -5.57978332e-01 -3.41042221e-01 7.94202626e-01
7.79550970e-01 3.51279825e-02 1.05033016e+00 5.36103368e-01
-7.79144347e-01 3.36292833e-01 -1.89939106e+00 9.30837870e-01
2.71646529e-01 -2.50723988e-01 1.66181117e-01 -1.17445147e+00
1.06872547e+00 3.42288554e-01 8.88294399e-01 -7.98051476e-01
-4.44764704e-01 4.85658228e-01 1.50546001e-03 -8.34512115e-01
4.37401980e-01 -2.62730449e-01 -1.93710290e-02 9.43718433e-01
-4.49369729e-01 4.31619048e-01 7.50623941e-01 -1.46607265e-01
1.38675630e+00 -1.65506527e-01 1.00148165e+00 -6.13240540e-01
6.68264329e-01 1.01511002e+00 1.14199388e+00 7.71355152e-01
4.13948745e-01 2.12009102e-01 3.66659433e-01 -1.48718572e+00
-7.27173746e-01 -9.01095212e-01 -1.05166838e-01 6.32117629e-01
1.07889026e-01 -9.08799112e-01 -1.69287369e-01 -8.10139298e-01
2.76904821e-01 7.67066896e-01 -1.40674382e-01 2.92579114e-01
-1.02323103e+00 -7.98019350e-01 1.27807438e-01 5.82233012e-01
1.20180994e-01 -1.29627395e+00 -1.42204678e+00 7.29900241e-01
-5.03741615e-02 -7.68286884e-01 2.70760447e-01 6.05813265e-01
-9.33579028e-01 -1.50862086e+00 5.72114825e-01 -6.59763455e-01
2.48686731e-01 2.52757102e-01 1.25999510e+00 5.49644470e-01
-3.52693349e-01 2.20869645e-01 -7.88485229e-01 -1.19884896e+00
-2.59823382e-01 -3.40775251e-01 -2.67487392e-02 -7.60847628e-02
1.12699342e+00 -6.23615861e-01 -1.79264337e-01 3.09231013e-01
-8.93308222e-01 -2.56125629e-01 1.89056799e-01 4.00165796e-01
1.00180876e+00 5.69887459e-01 5.62370002e-01 -1.37599444e+00
5.80266774e-01 -7.39822268e-01 -6.16219997e-01 5.62960088e-01
-9.84901130e-01 -2.22017825e-01 5.98011494e-01 -1.57394707e-02
-1.04178178e+00 4.42982525e-01 -1.83143631e-01 1.46599606e-01
-1.69914410e-01 9.45799291e-01 -4.26442884e-02 2.06640974e-01
5.40960431e-01 2.63128787e-01 9.33704227e-02 -3.09563100e-01
-3.78695548e-01 3.14583451e-01 1.75995395e-01 -6.49363518e-01
4.40885812e-01 6.05043828e-01 -1.04001656e-01 -9.25664961e-01
-3.25273693e-01 -8.80294502e-01 -2.54447252e-01 3.46777029e-02
2.13604689e-01 -3.05614889e-01 -2.46224552e-01 -1.21155061e-01
-1.05961633e+00 -1.01482213e-01 -9.64043617e-01 1.04684480e-01
-7.45513082e-01 4.33391154e-01 -1.79402664e-01 -8.44226420e-01
-4.84878272e-01 -4.11972672e-01 2.23004401e-01 -1.53207749e-01
-8.36124361e-01 -3.21619958e-01 5.04518747e-01 -1.16549991e-01
-2.73572486e-02 5.08086085e-01 1.11123919e+00 -1.27364147e+00
-4.79928106e-01 -3.05025399e-01 2.42767170e-01 -4.31124240e-01
-1.38042957e-01 3.14008832e-01 -2.42421657e-01 3.14501487e-02
3.57079357e-01 -6.87739998e-02 1.92568526e-01 2.27700710e-01
1.04439640e+00 -7.43151009e-01 -5.95468402e-01 2.51571864e-01
1.74016583e+00 7.69840539e-01 8.47045243e-01 6.82930589e-01
-2.94043660e-01 7.79784441e-01 1.24525416e+00 8.02317619e-01
2.87737310e-01 6.41968191e-01 2.04231858e-01 4.40133780e-01
4.11399037e-01 8.88038427e-02 -3.96942534e-02 3.34887981e-01
-3.51651132e-01 1.52628630e-01 -1.06702852e+00 5.64290047e-01
-2.28544474e+00 -1.69559395e+00 -7.93622553e-01 1.97327292e+00
7.61036873e-01 3.12657416e-01 7.51865149e-01 8.33774030e-01
4.83939469e-01 -4.60881680e-01 -7.17901736e-02 -1.30092883e+00
-2.22959593e-01 5.59785008e-01 9.53011401e-03 3.98929834e-01
-7.40034580e-01 5.10992050e-01 7.50234509e+00 2.24239320e-01
-5.96108556e-01 -1.66801170e-01 -6.29017428e-02 -1.41062707e-01
-6.03890240e-01 2.69471616e-01 -9.01683629e-01 3.02947551e-01
1.18205035e+00 -5.95259070e-01 -2.81054694e-02 1.25274241e+00
1.52006252e-02 -1.98086336e-01 -1.26067102e+00 5.20013034e-01
-2.56584704e-01 -1.96708846e+00 3.03034902e-01 7.46216923e-02
3.92132610e-01 -2.39089027e-01 -7.58144319e-01 7.04919547e-02
2.19145700e-01 -7.27872014e-01 9.20978487e-01 3.19026172e-01
2.14725912e-01 -8.69604647e-01 6.93303704e-01 5.74202597e-01
-1.23543954e+00 -5.70921361e-01 -3.37987393e-01 -6.53028429e-01
2.49907300e-01 7.92182982e-01 -7.12099493e-01 8.39697659e-01
1.17116261e+00 2.54231811e-01 1.28928926e-02 1.34480739e+00
2.27942899e-01 4.13358212e-01 -7.98071325e-01 -9.82285887e-02
-2.42995797e-03 -2.17186704e-01 7.04462945e-01 1.44693244e+00
1.19834505e-01 5.14267862e-01 5.12949646e-01 5.20074368e-01
6.31202221e-01 2.16856644e-01 -8.53761613e-01 4.23312843e-01
6.61912143e-01 6.06990576e-01 -8.25205863e-01 -2.29884952e-01
-8.69440973e-01 2.72319466e-01 1.87256902e-01 -1.30456880e-01
-5.56591690e-01 -3.00637364e-01 3.69917393e-01 5.85552394e-01
7.53586292e-01 6.90547451e-02 -7.35227585e-01 -7.18680382e-01
5.31819344e-01 -1.32648408e+00 1.30647242e+00 -8.47752020e-02
-8.57984245e-01 3.87779027e-01 3.73972863e-01 -8.52584600e-01
-3.68616670e-01 -3.29133183e-01 -8.74014556e-01 3.14661086e-01
-1.20418262e+00 -6.44404590e-01 -6.86817393e-02 7.17798948e-01
5.03959656e-01 -1.63435545e-02 1.09851849e+00 -3.25346887e-02
-3.03962111e-01 3.67002875e-01 -6.15065336e-01 -6.39231563e-01
-1.20332457e-01 -1.21282411e+00 -3.39263752e-02 1.02633214e+00
5.94025254e-02 6.75706506e-01 8.67280364e-01 -8.04000854e-01
-1.78807163e+00 -7.57713735e-01 1.33449686e+00 -1.60681173e-01
4.73501235e-01 7.70885274e-02 -9.33641553e-01 8.13744426e-01
2.02878252e-01 -1.32432252e-01 1.07533085e+00 2.48919502e-01
-1.85243890e-01 -2.80730098e-01 -1.45195043e+00 2.83697724e-01
1.37635517e+00 2.47635573e-01 -1.11998677e+00 1.15713216e-01
5.50653636e-01 1.38301894e-01 -8.27739954e-01 7.23137558e-01
8.95523071e-01 -1.32720983e+00 7.69948006e-01 -9.35004532e-01
3.05449069e-01 -3.95668805e-01 -2.13044390e-01 -3.94766241e-01
-4.10686255e-01 -8.54958355e-01 -6.09574556e-01 6.46217525e-01
7.98591018e-01 -6.73430204e-01 1.08996320e+00 7.09552288e-01
-1.44257009e-01 -1.09761035e+00 -9.44257498e-01 -9.64270234e-01
-4.50109422e-01 -5.91192782e-01 8.58680129e-01 1.01263404e+00
7.96557307e-01 -4.31343429e-02 -1.30547751e-02 3.93837644e-03
4.99123156e-01 1.09759474e+00 5.29215276e-01 -1.63551009e+00
-6.09347165e-01 -2.30616093e-01 -2.09422767e-01 -4.23357248e-01
-2.06280440e-01 -7.79357374e-01 -4.05681193e-01 -1.28254318e+00
1.79276586e-01 -7.97877610e-01 -1.30360186e-01 6.43967211e-01
4.54457462e-01 -3.36664915e-01 -3.07716101e-01 1.34004220e-01
-7.24771023e-01 -3.87709498e-01 4.76758212e-01 3.46581377e-02
-5.50429881e-01 2.46147394e-01 -7.91406214e-01 8.42089415e-01
9.06118035e-01 -7.55335927e-01 -4.35742527e-01 -2.51599364e-02
7.05657840e-01 4.63613719e-01 -3.64005774e-01 -6.85810566e-01
7.05822051e-01 -8.34932745e-01 -1.41921893e-01 -8.29960346e-01
-3.11171770e-01 -1.11276996e+00 7.35482812e-01 1.08250451e+00
3.59564498e-02 4.05294746e-01 1.75214887e-01 6.01869941e-01
-3.95722449e-01 -7.29931056e-01 3.59021127e-01 -4.57622588e-01
-9.90839005e-01 1.79490689e-02 -7.31255174e-01 -1.11456119e-01
1.72524869e+00 -1.05922222e+00 -9.86943301e-03 2.99399734e-01
-9.33962107e-01 3.33889842e-01 3.82034421e-01 -2.27737837e-02
8.63286972e-01 -8.46016884e-01 -6.02347255e-01 4.20527399e-01
2.21919268e-01 1.09148584e-01 -2.91028023e-01 8.23047459e-01
-8.85044158e-01 5.82446754e-01 -4.95261252e-01 -3.67232919e-01
-1.52865052e+00 1.03910887e+00 7.88162425e-02 -6.49213135e-01
-9.55044091e-01 5.58971465e-01 -5.15917182e-01 -1.29802406e-01
1.87061757e-01 -2.68309861e-01 -2.30058908e-01 1.42542110e-03
9.71143961e-01 5.70300221e-01 3.53000313e-01 2.90729105e-01
-6.85037971e-01 1.66425526e-01 -2.99188405e-01 2.06334785e-01
1.67653251e+00 2.52883285e-01 -7.25255072e-01 2.62285918e-01
2.26049453e-01 7.47146383e-02 -3.31577092e-01 3.72376330e-02
8.79002213e-01 -4.48110163e-01 -5.47259748e-01 -7.02076435e-01
-5.38122535e-01 -3.12073797e-01 -2.47058868e-01 8.02567303e-01
1.23560321e+00 -6.01053908e-02 6.08077228e-01 7.09677935e-01
9.32613254e-01 -1.26186180e+00 -5.54700196e-01 3.70734096e-01
9.88235593e-01 -6.90834761e-01 2.85313904e-01 -1.20037031e+00
-3.85264546e-01 1.34608972e+00 6.57103598e-01 -3.44528228e-01
8.25137734e-01 1.20057130e+00 -6.18925154e-01 -5.46257913e-01
-1.35439408e+00 -5.26257940e-02 -5.08531988e-01 4.84216213e-01
3.32629532e-02 2.16728762e-01 -1.13625395e+00 8.47900152e-01
-2.65819013e-01 4.16101724e-01 5.84437311e-01 1.70315099e+00
-6.53949261e-01 -1.40928292e+00 -4.04749453e-01 7.30964363e-01
-6.62462473e-01 3.05764496e-01 -8.38106990e-01 1.11383498e+00
2.60501832e-01 9.49246645e-01 2.01166034e-01 -3.77851993e-01
7.19029248e-01 2.62510091e-01 7.24561632e-01 -8.92156720e-01
-9.18057382e-01 -4.45616663e-01 8.78000617e-01 -7.05589354e-01
-4.53954190e-01 -1.05001485e+00 -1.46044636e+00 -3.79975080e-01
-9.69949588e-02 6.89684868e-01 1.02022156e-01 6.58239603e-01
1.08354175e-02 -2.28097633e-01 5.08904755e-01 3.38816196e-01
-6.39119744e-01 -4.90025133e-02 -5.43237925e-01 1.53595597e-01
-3.43252987e-01 -3.27122808e-01 1.03960216e-01 -1.41700566e-01] | [8.323698043823242, 6.323720932006836] |
0f275a47-99d3-492e-aef6-6526b910e1a5 | umad-universal-model-adaptation-under-domain | 2112.08553 | null | https://arxiv.org/abs/2112.08553v1 | https://arxiv.org/pdf/2112.08553v1.pdf | UMAD: Universal Model Adaptation under Domain and Category Shift | Learning to reject unknown samples (not present in the source classes) in the target domain is fairly important for unsupervised domain adaptation (UDA). There exist two typical UDA scenarios, i.e., open-set, and open-partial-set, and the latter assumes that not all source classes appear in the target domain. However, most prior methods are designed for one UDA scenario and always perform badly on the other UDA scenario. Moreover, they also require the labeled source data during adaptation, limiting their usability in data privacy-sensitive applications. To address these issues, this paper proposes a Universal Model ADaptation (UMAD) framework which handles both UDA scenarios without access to the source data nor prior knowledge about the category shift between domains. Specifically, we aim to learn a source model with an elegantly designed two-head classifier and provide it to the target domain. During adaptation, we develop an informative consistency score to help distinguish unknown samples from known samples. To achieve bilateral adaptation in the target domain, we further maximize localized mutual information to align known samples with the source classifier and employ an entropic loss to push unknown samples far away from the source classification boundary, respectively. Experiments on open-set and open-partial-set UDA scenarios demonstrate that UMAD, as a unified approach without access to source data, exhibits comparable, if not superior, performance to state-of-the-art data-dependent methods. | ['Ran He', 'Jiashi Feng', 'Dapeng Hu', 'Jian Liang'] | 2021-12-16 | null | null | null | null | ['universal-domain-adaptation'] | ['computer-vision'] | [ 4.34754044e-01 2.13679045e-01 -4.86870855e-01 -7.16758013e-01
-9.03083682e-01 -6.94443464e-01 4.81914252e-01 4.38168757e-02
-4.30654407e-01 9.93243575e-01 -1.45812511e-01 -7.88747892e-03
-8.87036771e-02 -6.06844008e-01 -6.88599885e-01 -8.75364959e-01
2.82719463e-01 5.75725734e-01 1.79189339e-01 1.40797850e-02
-1.03627801e-01 2.98903823e-01 -1.33011496e+00 1.42181098e-01
1.19189608e+00 1.06263912e+00 -1.54665234e-02 -1.15796244e-02
-1.45419970e-01 6.83165640e-02 -6.28795505e-01 -6.50725603e-01
6.48280025e-01 -5.78661323e-01 -6.51032984e-01 1.34716853e-01
3.21032017e-01 -1.20109727e-03 -1.48120329e-01 1.30977190e+00
5.44246137e-01 1.60266429e-01 9.94060576e-01 -1.47621989e+00
-8.18372726e-01 3.19603086e-01 -5.76318026e-01 4.10204306e-02
2.40431100e-01 -5.33778928e-02 7.82857835e-01 -9.30525184e-01
5.49221516e-01 1.00883746e+00 5.98058105e-01 9.51566696e-01
-1.45252120e+00 -1.03272295e+00 4.75288510e-01 8.54443535e-02
-1.51908398e+00 -6.10230327e-01 8.94047678e-01 -3.79465014e-01
1.18939005e-01 3.40963334e-01 1.39846593e-01 1.47948050e+00
-1.79051131e-01 9.08345342e-01 1.16484046e+00 -2.76390225e-01
6.07131124e-01 8.00566554e-01 1.91822633e-01 4.02278155e-02
3.87159467e-01 9.97387022e-02 -4.18602794e-01 -4.68271434e-01
3.82116616e-01 -3.54354605e-02 -4.64965880e-01 -9.86486852e-01
-1.02165449e+00 6.73666000e-01 2.75843203e-01 -2.54811831e-02
-1.41055793e-01 -8.28514874e-01 2.89256334e-01 4.68055606e-01
3.36996406e-01 2.67435521e-01 -6.52253330e-01 3.77147108e-01
-6.64455831e-01 1.94790527e-01 7.97511637e-01 1.42290151e+00
8.26439321e-01 -2.69998878e-01 -1.40749708e-01 9.42485034e-01
2.52156287e-01 6.40781403e-01 7.69227564e-01 -4.73406017e-01
6.04891300e-01 5.37883759e-01 2.25940123e-01 -7.54429996e-01
-1.79496691e-01 -5.38791478e-01 -1.04227459e+00 2.46767011e-02
8.89695287e-01 -9.47179645e-02 -8.15758169e-01 2.12374282e+00
6.22559428e-01 1.71637028e-01 5.40994585e-01 8.48060489e-01
5.71748316e-01 3.08596253e-01 -6.22597411e-02 -4.80260730e-01
1.09393704e+00 -4.97085840e-01 -6.45069361e-01 -3.51411909e-01
4.06571001e-01 -5.56568801e-01 1.12970173e+00 3.05249184e-01
-6.70927048e-01 -5.14649153e-01 -1.18074667e+00 1.57579631e-01
-6.70661151e-01 1.57888964e-01 9.63086188e-02 8.04183125e-01
-3.56609911e-01 3.53612006e-01 -4.30864006e-01 -4.98362094e-01
6.92988694e-01 3.52093041e-01 -5.50098956e-01 -2.81171560e-01
-1.35253716e+00 7.62871742e-01 4.53763723e-01 -1.83274880e-01
-8.20644915e-01 -6.31258428e-01 -8.37464929e-01 -1.23984717e-01
4.75690514e-01 -6.02611780e-01 1.07187462e+00 -1.15542614e+00
-1.42428863e+00 1.03799951e+00 -1.99738041e-01 -5.04565597e-01
9.13517177e-01 -3.03129237e-02 -6.91253841e-01 -3.67366791e-01
1.65159956e-01 4.54081714e-01 1.03472209e+00 -1.42953777e+00
-7.11110234e-01 -6.47252679e-01 -3.89825106e-01 3.52747172e-01
-6.08515918e-01 -3.87848198e-01 -3.56857628e-01 -7.43598700e-01
1.39283746e-01 -7.99880505e-01 -1.14237266e-02 1.93857521e-01
-6.06103897e-01 -5.09687625e-02 1.01890290e+00 -4.13547754e-01
1.18843782e+00 -2.33182096e+00 4.95987348e-02 4.88808662e-01
-3.07463948e-02 4.37434614e-01 -1.28072008e-01 -5.88682368e-02
-2.78081596e-01 -1.51985586e-01 -7.24370658e-01 -2.76162684e-01
2.90646721e-02 4.12107050e-01 -3.37053418e-01 5.98409057e-01
3.49751972e-02 2.84877986e-01 -8.61934781e-01 -2.82255054e-01
-7.24751642e-03 1.69872969e-01 -5.37802279e-01 3.74890387e-01
5.68137765e-02 6.96840823e-01 -4.24986452e-01 5.67280650e-01
1.16243398e+00 -9.43415537e-02 1.69858515e-01 -4.75987981e-05
3.62569481e-01 -1.03681132e-01 -1.52681732e+00 1.51823819e+00
-2.55257338e-01 1.99963167e-01 1.67219788e-01 -1.10452807e+00
1.07866538e+00 2.53465265e-01 4.64712799e-01 -4.35177386e-01
3.70571725e-02 3.68133217e-01 1.20112700e-02 -2.50570059e-01
7.98258483e-02 -2.50234604e-01 -1.61788642e-01 5.24122491e-02
1.56084120e-01 2.20658183e-01 -1.44610777e-01 -1.69456646e-01
6.30749404e-01 -4.55146581e-02 6.99846268e-01 -3.01654667e-01
7.64276922e-01 -3.17008764e-01 9.92229640e-01 7.67346919e-01
-6.07018232e-01 9.37240303e-01 2.81098366e-01 -1.10039048e-01
-6.98990703e-01 -1.30640447e+00 -5.42612135e-01 8.88126016e-01
5.82565546e-01 4.88821268e-02 -7.34795868e-01 -1.26623476e+00
2.76789993e-01 9.16987598e-01 -6.57550395e-01 -5.54452479e-01
-1.97746515e-01 -8.75533640e-01 3.00395280e-01 2.40804598e-01
6.76034093e-01 -4.76440370e-01 -1.00787573e-01 1.02032229e-01
-2.17377305e-01 -8.87918532e-01 -7.57215261e-01 3.90489697e-01
-5.99790514e-01 -1.22943449e+00 -1.00722885e+00 -7.55966067e-01
8.68006766e-01 1.15731135e-01 7.43465662e-01 -7.70237625e-01
2.65015811e-01 1.80833608e-01 -2.29875222e-01 -5.30984342e-01
-5.00201404e-01 9.68882069e-02 4.30139422e-01 6.92353904e-01
8.34764361e-01 -5.38785577e-01 -3.43383700e-01 8.40387106e-01
-8.21917295e-01 -3.91734153e-01 5.46764135e-01 9.05497015e-01
8.14102590e-01 9.15480331e-02 9.72605646e-01 -1.15720022e+00
6.02369249e-01 -9.07272995e-01 -4.73963320e-01 4.11394566e-01
-8.22056949e-01 -6.68970570e-02 9.60454762e-01 -9.08740103e-01
-1.12123346e+00 3.65032494e-01 1.52919874e-01 -6.86908245e-01
-5.06520391e-01 1.55751169e-01 -1.02972519e+00 5.17677665e-02
9.49170887e-01 3.11008871e-01 5.75423166e-02 -6.60867870e-01
1.80002615e-01 1.04945624e+00 6.19774938e-01 -5.37725329e-01
9.40093994e-01 4.29947495e-01 -3.25305700e-01 -6.22348249e-01
-7.53569067e-01 -6.80591404e-01 -9.67163622e-01 2.34489337e-01
5.52338302e-01 -8.93295050e-01 -1.16792284e-01 5.15287161e-01
-7.38374412e-01 -3.15857120e-02 -5.66275895e-01 5.25700986e-01
-4.88148510e-01 4.18062717e-01 2.38092661e-01 -6.99295640e-01
-1.14854433e-01 -1.02025223e+00 6.86885238e-01 2.64386594e-01
-1.61524326e-01 -8.75142276e-01 -1.67157099e-01 1.83961451e-01
2.21804261e-01 9.30876136e-02 7.54751861e-01 -1.51156354e+00
-1.15399703e-01 -3.96320790e-01 1.30620664e-02 6.18623734e-01
5.14611363e-01 -6.04660273e-01 -1.22051299e+00 -4.91334260e-01
8.00729319e-02 -9.00372714e-02 6.36551559e-01 2.47438043e-01
1.18723321e+00 -5.70584834e-01 -4.45468575e-01 6.50413990e-01
9.87699211e-01 3.26333255e-01 3.49561870e-01 1.28692895e-01
4.79786396e-01 5.71559906e-01 7.67148137e-01 5.75552046e-01
2.50715405e-01 6.68155789e-01 2.46587679e-01 8.23947564e-02
2.74744369e-02 -4.31087494e-01 4.32374179e-01 4.33451504e-01
5.42410254e-01 -3.74616981e-01 -6.75580740e-01 6.12985611e-01
-1.69633126e+00 -7.07751632e-01 1.69069350e-01 2.85759902e+00
1.11882412e+00 9.51246098e-02 2.43150160e-01 9.89545719e-04
9.38855112e-01 -1.12713121e-01 -1.17382419e+00 3.86574604e-02
-2.94475943e-01 -1.93301052e-01 6.50973201e-01 2.83651322e-01
-1.54668379e+00 5.52608371e-01 5.65562487e+00 8.98356915e-01
-9.94330287e-01 1.23195045e-01 5.16803503e-01 2.83406787e-02
-2.72497833e-01 -1.91806346e-01 -8.34193528e-01 7.23082781e-01
6.43586814e-01 -3.81208241e-01 2.73563057e-01 1.10690260e+00
-2.73597240e-01 2.24274233e-01 -1.47088397e+00 1.03143525e+00
4.39591184e-02 -7.25087941e-01 1.32739544e-01 6.38138354e-02
6.54297173e-01 -2.29352459e-01 2.23110795e-01 4.69348550e-01
2.48517603e-01 -5.19718885e-01 5.61241627e-01 2.83022761e-01
1.02586126e+00 -7.29218245e-01 7.38683999e-01 6.93598628e-01
-9.33298826e-01 -1.54155150e-01 -5.07083237e-01 3.59275997e-01
-2.04184264e-01 5.34082592e-01 -6.76144004e-01 6.36253655e-01
6.83282137e-01 7.02944696e-01 -5.59222579e-01 1.10904598e+00
3.10012489e-03 4.33989286e-01 -5.31325221e-01 3.09760332e-01
-2.60904729e-01 -2.39482939e-01 9.08589840e-01 9.20561075e-01
3.19745302e-01 3.56972055e-03 2.71394491e-01 7.78849483e-01
-2.81573832e-01 2.61619538e-01 -7.73867369e-01 3.57315332e-01
7.63053238e-01 7.18747377e-01 -1.84851274e-01 -2.50169247e-01
-4.47413206e-01 1.23788273e+00 9.57012847e-02 6.24171734e-01
-6.79946661e-01 -5.13220310e-01 9.71310318e-01 2.13235199e-01
2.14069873e-01 3.85336876e-01 -3.24460059e-01 -1.54162848e+00
2.59198189e-01 -9.85344470e-01 9.30110872e-01 -2.49312103e-01
-1.85794365e+00 4.50551063e-01 1.51813999e-01 -1.78074706e+00
-1.25305220e-01 -3.88629228e-01 -3.22091579e-01 9.12437141e-01
-1.55408144e+00 -9.45773602e-01 -1.35244370e-01 1.00605857e+00
2.67423570e-01 -3.61306101e-01 9.35477674e-01 4.72878605e-01
-6.79341435e-01 1.40091407e+00 7.04944789e-01 1.55928671e-01
1.27842081e+00 -1.13461936e+00 -1.05120450e-01 8.77387047e-01
-4.11839075e-02 6.72499537e-01 5.80392182e-01 -5.10706663e-01
-1.03109908e+00 -1.37133324e+00 5.43021500e-01 -4.56149369e-01
1.71538711e-01 -5.66290319e-01 -1.36053312e+00 8.52581918e-01
-1.68261588e-01 3.40933233e-01 9.39994931e-01 1.02538228e-01
-5.99831462e-01 -4.76545244e-01 -1.76126397e+00 3.19109410e-01
8.79295051e-01 -3.91275644e-01 -7.53938675e-01 2.50815898e-01
3.91892523e-01 -4.14075166e-01 -8.56372356e-01 3.74181002e-01
2.88555980e-01 -7.14232087e-01 1.00402749e+00 -6.93610489e-01
-1.53003424e-01 -4.79903072e-01 -2.98984200e-01 -1.41246033e+00
-2.28925258e-01 -5.59663951e-01 -1.89255208e-01 1.56437206e+00
4.73069221e-01 -1.09716082e+00 8.47429633e-01 8.71417999e-01
6.73798844e-02 -3.03091705e-01 -1.26027215e+00 -1.13258135e+00
3.88182819e-01 -1.28259718e-01 7.09269106e-01 1.31745231e+00
-6.25822768e-02 1.60869986e-01 -4.20769155e-01 5.56408226e-01
8.85877311e-01 1.22474097e-01 8.42069983e-01 -1.46234989e+00
-6.96499348e-02 -2.76079208e-01 -2.54659832e-01 -1.10804307e+00
2.73433328e-01 -1.08837175e+00 -3.28434259e-02 -9.23578739e-01
8.74946862e-02 -6.03230178e-01 -6.02629364e-01 4.99483198e-01
-2.77364165e-01 -1.26386926e-01 1.88253392e-02 4.64806139e-01
-3.61426145e-01 6.80708885e-01 1.01144910e+00 -1.86877087e-01
-4.47602838e-01 4.44225669e-01 -9.37071264e-01 6.92254186e-01
7.28866875e-01 -5.78510165e-01 -6.42803490e-01 -7.75237679e-02
-5.72925806e-01 -2.94919014e-01 1.89000905e-01 -1.09330904e+00
1.74050704e-01 -3.87012780e-01 6.21446848e-01 -3.17451626e-01
1.49135903e-01 -1.20657194e+00 8.47198293e-02 1.31793067e-01
-5.97054482e-01 -7.21320391e-01 9.13223475e-02 9.89624441e-01
-2.65293092e-01 -2.25826338e-01 1.29014444e+00 1.59480751e-01
-7.13053226e-01 4.59049642e-01 5.27924486e-02 2.68478870e-01
1.36688316e+00 -3.85078728e-01 -9.61989537e-03 -2.81213522e-01
-9.39642191e-01 3.74623746e-01 5.78775585e-01 4.20469046e-01
3.85308594e-01 -1.42882812e+00 -6.33247137e-01 7.46529877e-01
6.43741488e-01 2.69330561e-01 1.99002475e-01 5.23944259e-01
3.47335130e-01 1.75534710e-01 -1.22483529e-01 -5.52475274e-01
-1.12529051e+00 9.86824334e-01 4.54106361e-01 -2.90295687e-02
-3.00317347e-01 8.92031968e-01 5.74961960e-01 -9.92578685e-01
4.45672333e-01 -1.37004510e-01 -9.49327350e-02 2.16809392e-01
5.54295242e-01 2.75073469e-01 -4.35075760e-02 -6.73078537e-01
-4.51439440e-01 3.95077825e-01 -2.78969377e-01 2.99840838e-01
7.78189957e-01 -4.34262156e-01 4.28663909e-01 3.59759778e-01
1.42493808e+00 -1.38090570e-02 -1.41094828e+00 -8.01199853e-01
2.50108279e-02 -6.86558664e-01 -4.55854118e-01 -9.11858439e-01
-9.68680382e-01 8.50372732e-01 8.62099767e-01 4.53034751e-02
1.28913248e+00 -7.28487372e-02 5.49855471e-01 3.32561582e-01
2.82601446e-01 -1.25589669e+00 -2.88574845e-01 2.05973670e-01
8.28053296e-01 -1.58921945e+00 -8.68549496e-02 -4.06253368e-01
-9.08029854e-01 8.76447439e-01 8.86077583e-01 2.46296614e-01
7.61389852e-01 -1.22713432e-01 9.75778177e-02 4.12985772e-01
-3.14795732e-01 -4.77905907e-02 3.87236685e-01 1.09331584e+00
-1.43397301e-01 4.66200933e-02 -1.32730659e-02 1.06377673e+00
5.36549576e-02 -8.31760019e-02 2.73932666e-01 6.35090947e-01
-1.35760874e-01 -1.16473424e+00 -5.76946259e-01 4.46446985e-01
-3.17983866e-01 2.09637970e-01 -4.99826312e-01 8.14331591e-01
2.90732563e-01 7.60679424e-01 -7.92033449e-02 -3.32574308e-01
6.87235057e-01 3.77877891e-01 4.89144139e-02 -4.57213402e-01
-2.43945211e-01 -6.50720075e-02 -2.30170339e-01 -3.11159313e-01
-1.81618989e-01 -8.68636012e-01 -9.66652870e-01 -6.42169937e-02
-4.03198272e-01 2.37522885e-01 2.46338934e-01 6.42397344e-01
6.27521276e-01 1.21624032e-02 7.44998395e-01 -4.13214266e-01
-1.00594258e+00 -8.61563146e-01 -8.14200222e-01 7.07243502e-01
5.20987809e-01 -7.55964696e-01 -4.96212274e-01 8.86771679e-02] | [10.388480186462402, 3.199429512023926] |
ad636de3-dde0-4b4a-a32a-2b68132613e4 | learning-to-predict-stereo-reliability | null | null | http://openaccess.thecvf.com/content_cvpr_2017/html/Poggi_Learning_to_Predict_CVPR_2017_paper.html | http://openaccess.thecvf.com/content_cvpr_2017/papers/Poggi_Learning_to_Predict_CVPR_2017_paper.pdf | Learning to Predict Stereo Reliability Enforcing Local Consistency of Confidence Maps | Confidence measures estimate unreliable disparity assignments performed by a stereo matching algorithm and, as recently proved, can be used for several purposes. This paper aims at increasing, by means of a deep network, the effectiveness of state-of-the-art confidence measures exploiting the local consistency assumption. We exhaustively evaluated our proposal on 23 confidence measures, including 5 top-performing ones based on random-forests and CNNs, training our networks with two popular stereo algorithms and a small subset (25 out of 194 frames) of the KITTI 2012 dataset. Experimental results show that our approach dramatically increases the effectiveness of all the 23 confidence measures on the remaining frames. Moreover, without re-training, we report a further cross-evaluation on KITTI 2015 and Middlebury 2014 confirming that our proposal provides remarkable improvements for each confidence measure even when dealing with significantly different input data. To the best of our knowledge, this is the first method to move beyond conventional pixel-wise confidence estimation.
| ['Stefano Mattoccia', 'Matteo Poggi'] | 2017-07-01 | null | null | null | cvpr-2017-7 | ['stereo-matching'] | ['computer-vision'] | [-3.10790958e-03 8.78084227e-02 2.07392693e-01 -4.04005677e-01
-9.84959364e-01 -3.27857554e-01 8.84415448e-01 1.80858597e-01
-9.39484060e-01 1.16182530e+00 -6.20153956e-02 -1.59893766e-01
-1.15788974e-01 -6.31140590e-01 -8.02934647e-01 -5.26931703e-01
-2.70122051e-01 4.93297428e-01 7.65399635e-01 3.29922438e-02
6.10520065e-01 3.72839987e-01 -1.98458469e+00 7.01398998e-02
8.39772701e-01 1.47181642e+00 -9.55843180e-02 6.95763707e-01
3.68025482e-01 7.96838939e-01 -4.49672669e-01 -8.67920101e-01
2.81317174e-01 4.85477690e-03 -7.40633011e-01 -1.69220984e-01
1.07358539e+00 -4.84476089e-01 -1.72892828e-02 9.52974141e-01
4.36500669e-01 -9.48645547e-02 6.11153364e-01 -9.22634900e-01
-1.11251496e-01 4.85384464e-01 -3.96375597e-01 2.71747231e-01
2.80505568e-01 2.47805476e-01 7.68612444e-01 -9.03697729e-01
7.04168022e-01 8.59067857e-01 1.25289345e+00 8.49300846e-02
-1.12429869e+00 -6.77351892e-01 -1.61214352e-01 5.64122438e-01
-1.31486320e+00 -4.65039462e-01 3.35425675e-01 -5.04062057e-01
1.11612129e+00 -1.05705157e-01 5.66833615e-01 1.09785819e+00
-1.61414012e-01 3.78300905e-01 1.59818387e+00 -4.71631557e-01
2.56847084e-01 -1.70512959e-01 -1.89205650e-02 6.71486974e-01
3.61229807e-01 5.39582670e-01 -5.13226390e-01 1.82148833e-02
6.42347932e-01 -7.95897484e-01 -3.32902163e-01 -6.36038542e-01
-1.36394155e+00 5.93447566e-01 4.46117371e-01 3.34435970e-01
-1.53427929e-01 3.48241717e-01 4.58946347e-01 1.42258361e-01
6.34408832e-01 2.43156493e-01 -4.93337035e-01 -3.37071568e-01
-1.43968379e+00 4.90698107e-02 7.25749731e-01 7.21950889e-01
9.54681396e-01 -3.00038964e-01 -1.05080485e-01 6.16488755e-01
-1.11334994e-02 4.69379455e-01 1.74925640e-01 -1.36215699e+00
3.79740328e-01 1.35758683e-01 2.52899319e-01 -9.13198948e-01
-3.92772824e-01 -5.16112328e-01 -8.68146300e-01 7.24391818e-01
9.06208217e-01 2.00114362e-02 -9.00113285e-01 1.70579243e+00
7.06434622e-02 5.50102949e-01 -2.63928980e-01 8.38868678e-01
4.89797711e-01 -1.26456870e-02 -4.76975963e-02 -1.52081400e-01
9.63479698e-01 -7.85384119e-01 -2.01617315e-01 1.08010639e-02
2.98211157e-01 -9.09999430e-01 5.52398860e-01 8.40478957e-01
-1.05009556e+00 -8.74245405e-01 -1.33650219e+00 1.20405659e-01
-3.07347804e-01 2.37619057e-01 6.10398889e-01 7.31034994e-01
-1.41891265e+00 1.09403610e+00 -6.82049811e-01 -3.30562413e-01
4.44380999e-01 1.98030949e-01 -6.10781133e-01 -6.64222166e-02
-1.14497709e+00 1.17749035e+00 4.64339584e-01 2.98409641e-01
-5.86926639e-01 -4.20629025e-01 -7.97533095e-01 -1.51324645e-01
3.92120123e-01 -8.75761211e-01 1.19319510e+00 -8.85339499e-01
-1.39072251e+00 1.31607509e+00 1.14941716e-01 -9.53621805e-01
1.30397606e+00 -4.96667206e-01 -1.46238118e-01 2.47248411e-01
2.00253829e-01 9.79349315e-01 7.97633231e-01 -1.08023310e+00
-9.02977586e-01 -8.50797594e-02 8.42737854e-02 -3.41193050e-01
1.39177188e-01 -2.91671723e-01 -5.44044912e-01 -4.28033918e-01
1.06685519e-01 -8.15999806e-01 -2.69153640e-02 4.77128215e-02
-3.48214090e-01 -7.09129423e-02 -4.50262660e-03 -4.24709231e-01
9.79279995e-01 -1.89887035e+00 7.96719268e-02 3.47143799e-01
1.19792625e-01 3.30860794e-01 7.02565908e-02 1.13510169e-01
-4.80681434e-02 -6.61907867e-02 -4.28898275e-01 -6.98400855e-01
-6.46478217e-03 8.65455791e-02 -6.00218326e-02 7.11310446e-01
3.82388234e-01 5.36370277e-01 -8.56311977e-01 -8.04575145e-01
7.84805000e-01 5.86101770e-01 -3.49768609e-01 -1.85456842e-01
1.02418764e-02 4.31313068e-01 7.69924000e-02 2.77394176e-01
9.20823872e-01 3.94687615e-02 -5.30622639e-02 -3.67456764e-01
-4.01038140e-01 3.31739560e-02 -1.36993659e+00 1.82734096e+00
-3.95548731e-01 9.35096443e-01 -4.31368232e-01 -7.55056858e-01
8.61811936e-01 3.79814059e-02 1.72917962e-01 -7.78127968e-01
1.40030488e-01 6.99235976e-01 -5.42771369e-02 -9.15644169e-02
6.03605092e-01 1.51891872e-01 2.36694172e-01 -5.59801050e-02
4.23499405e-01 -1.68514267e-01 6.18609369e-01 7.76077155e-03
1.01184344e+00 3.86473238e-01 2.93569177e-01 -4.52118069e-01
8.80301893e-01 -3.28351378e-01 2.09360734e-01 1.11943388e+00
-5.26768565e-01 1.24020243e+00 4.42587286e-01 -6.08974159e-01
-1.23296750e+00 -1.12917840e+00 -4.93013471e-01 4.88220930e-01
2.55992822e-02 -3.03510308e-01 -7.00981438e-01 -5.34568071e-01
-9.76124685e-03 4.83722538e-01 -9.59010839e-01 1.84594065e-01
-5.00426054e-01 -4.59113479e-01 7.05430329e-01 5.26289880e-01
9.49315965e-01 -9.24385905e-01 -8.28832686e-01 1.46048874e-01
-1.73482805e-01 -1.25641465e+00 1.43671900e-01 1.27969801e-01
-7.87667930e-01 -1.40029490e+00 -7.81001329e-01 -4.04329419e-01
8.60985816e-02 -6.56885356e-02 1.70987582e+00 2.24033996e-01
-6.75013140e-02 1.60858035e-01 -3.14770341e-01 -9.43876803e-03
-2.98536688e-01 1.31557882e-01 -1.55718520e-01 -2.69685805e-01
2.15343907e-01 -6.46835625e-01 -6.52880132e-01 4.33350563e-01
-5.65387607e-01 3.40578966e-02 4.55917478e-01 8.16901565e-01
4.68903303e-01 -2.55266815e-01 4.01143618e-02 -6.07534409e-01
8.44330415e-02 4.29853536e-02 -9.63951230e-01 4.70477901e-02
-6.43102348e-01 2.59268373e-01 2.04527617e-01 5.67738228e-02
-9.03770864e-01 -4.52798270e-02 -3.83698612e-01 -2.01145127e-01
-2.90236086e-01 1.92101255e-01 4.77957457e-01 -2.09714428e-01
8.12985361e-01 -1.07100211e-01 -3.61694336e-01 -1.88130111e-01
3.26890945e-01 2.46874064e-01 1.01962948e+00 -5.54302216e-01
6.55819654e-01 7.30415881e-01 3.70897740e-01 -2.93358892e-01
-8.66411388e-01 -4.83605176e-01 -7.30653524e-01 -3.63262624e-01
7.89755702e-01 -9.35897171e-01 -8.32913756e-01 8.53825033e-01
-1.18595970e+00 -2.86537409e-01 -9.83708799e-02 6.04572892e-01
-8.20352793e-01 7.03080058e-01 -5.44543028e-01 -8.57506692e-01
-1.36624455e-01 -1.05692673e+00 1.28445423e+00 -4.40581515e-03
-1.90805703e-01 -6.98451519e-01 3.70765805e-01 1.17234036e-01
4.71488297e-01 2.89200038e-01 1.46189839e-01 -1.56536683e-01
-6.09506965e-01 -1.40465766e-01 -5.14704287e-01 5.96642733e-01
-3.56660843e-01 4.52465206e-01 -1.51497996e+00 -1.46950334e-02
-1.59182101e-01 -4.87437725e-01 1.55209208e+00 6.40364408e-01
1.06772959e+00 4.65880483e-01 6.99880868e-02 6.52910888e-01
1.58096921e+00 -3.96824360e-01 1.08932579e+00 9.18862820e-01
2.41303995e-01 4.42478091e-01 8.40341210e-01 4.96129513e-01
2.87269056e-01 7.20616102e-01 7.50862837e-01 2.11608279e-02
-2.41079092e-01 6.71913326e-02 -1.13903694e-01 2.50113249e-01
-7.39301920e-01 -2.79006716e-02 -8.65693390e-01 5.48124909e-01
-1.91227901e+00 -1.15107679e+00 -2.79047340e-01 2.53608704e+00
6.95584357e-01 6.77673042e-01 -4.94181737e-02 4.60970253e-01
7.21057773e-01 1.15678132e-01 -1.51178658e-01 -1.12007454e-01
-4.18132454e-01 5.55911243e-01 6.95588708e-01 5.80816031e-01
-1.49595869e+00 6.28979146e-01 6.17722750e+00 9.12303567e-01
-8.27126861e-01 -5.25460616e-02 7.92311609e-01 1.69930026e-01
1.80412799e-01 -2.07406804e-02 -5.50336123e-01 5.08291066e-01
7.52300501e-01 3.91187191e-01 6.17410131e-02 8.03115845e-01
-1.22035727e-01 -8.14627767e-01 -1.14767551e+00 1.07669532e+00
1.47683611e-02 -1.44630361e+00 -6.77045941e-01 -1.43327296e-01
9.12569344e-01 3.73718530e-01 -4.11840975e-02 1.72470510e-01
9.43089053e-02 -1.08659697e+00 1.01079917e+00 7.17640579e-01
5.96993864e-01 -7.42456615e-01 1.29992759e+00 -2.83316486e-02
-9.83320594e-01 3.45112175e-01 -3.52984786e-01 -1.78276330e-01
1.92900121e-01 1.07438481e+00 -4.59994346e-01 7.09905207e-01
9.95764554e-01 7.65975177e-01 -8.23529840e-01 1.36348426e+00
-5.65398455e-01 3.40927541e-01 -5.50291181e-01 1.87914923e-01
3.28925699e-01 2.11154725e-02 3.49766731e-01 1.30495095e+00
4.57767040e-01 -5.24332702e-01 -4.70839590e-01 6.54923201e-01
6.12243749e-02 -1.96072727e-01 -4.59802866e-01 7.39973724e-01
2.99811065e-01 1.17321646e+00 -7.39533126e-01 -5.00796616e-01
-3.28769088e-01 8.38115036e-01 6.06333077e-01 -3.42419855e-02
-1.02589405e+00 -9.67259035e-02 4.40073878e-01 -1.53180748e-01
7.13418663e-01 -3.35831165e-01 -3.21839511e-01 -1.14271879e+00
2.62192458e-01 -5.34092188e-01 2.23838702e-01 -1.05268133e+00
-1.27370572e+00 7.55420387e-01 1.26632109e-01 -1.37178016e+00
-6.70928657e-01 -7.85453141e-01 -2.76255131e-01 8.78197134e-01
-1.78933632e+00 -8.05776298e-01 -5.46286047e-01 4.08417225e-01
2.36418888e-01 1.22886576e-01 7.30353713e-01 3.82057697e-01
-3.70847397e-02 4.44193959e-01 2.89382078e-02 -1.29068419e-01
9.62596416e-01 -1.25940430e+00 5.21613777e-01 1.08306646e+00
1.69483751e-01 3.44184369e-01 9.59673464e-01 -2.65157193e-01
-2.95351714e-01 -5.55645466e-01 9.79052961e-01 -4.78821129e-01
7.30877757e-01 -3.91962901e-02 -7.87577450e-01 2.24639148e-01
2.87160933e-01 1.00691907e-01 1.60444021e-01 1.81729198e-01
-5.71328819e-01 -4.95703630e-02 -1.16839194e+00 3.39826494e-01
1.18472266e+00 -5.52098215e-01 -5.18953025e-01 4.59712893e-02
2.74387807e-01 -6.46703839e-01 -8.19395542e-01 8.69154572e-01
9.13953424e-01 -2.27835870e+00 1.08376610e+00 1.10657208e-01
7.08097458e-01 -4.15697277e-01 -2.70341158e-01 -1.12817383e+00
2.69881207e-02 -1.63419813e-01 -2.15831660e-02 1.14312017e+00
2.99180895e-01 -6.89970493e-01 7.75138795e-01 2.48281002e-01
-2.53417134e-01 -4.72147077e-01 -1.36097455e+00 -8.25925410e-01
7.94110969e-02 -1.02311397e+00 3.40983570e-01 5.49500763e-01
-4.52313632e-01 -1.38514966e-01 -5.53384364e-01 -9.72719397e-03
8.64089429e-01 -6.88699111e-02 7.66831100e-01 -1.22485042e+00
-4.91942763e-01 -6.10939682e-01 -9.65098143e-01 -7.19574153e-01
-5.00923507e-02 -2.44409084e-01 3.08946639e-01 -1.08827019e+00
1.06103979e-01 -3.08482081e-01 -3.22041601e-01 1.83167547e-01
-1.46274313e-01 8.14890623e-01 6.64357468e-02 5.07821068e-02
-6.78360283e-01 2.84042209e-01 8.62884283e-01 -3.41716260e-02
4.64356869e-01 1.55259892e-02 -8.47811475e-02 9.19817567e-01
8.48111689e-01 -3.48023713e-01 3.58740054e-02 -4.35260803e-01
5.53320169e-01 -3.29607248e-01 7.62239575e-01 -1.77378833e+00
1.53442770e-01 4.26140189e-01 4.66793001e-01 -6.95637405e-01
2.50086129e-01 -7.76595354e-01 2.05111966e-01 5.02104998e-01
7.94720661e-04 -1.04399607e-01 3.99466723e-01 4.13605273e-01
-4.09868538e-01 -3.46614987e-01 9.67734158e-01 8.95599052e-02
-1.07964146e+00 -9.71811265e-02 -5.76843247e-02 3.36257629e-02
6.58833802e-01 -4.47231591e-01 -4.50112760e-01 -3.17485631e-01
-6.13714516e-01 -9.98762697e-02 6.98155642e-01 2.72846464e-02
2.54578114e-01 -1.16301799e+00 -6.57291412e-01 -1.35357212e-02
3.40227425e-01 -5.83364330e-02 2.37880647e-01 1.06760705e+00
-8.41338813e-01 4.83849615e-01 -4.80057299e-01 -1.07251227e+00
-1.16651404e+00 2.95424938e-01 3.63544941e-01 -5.61536014e-01
-3.25311542e-01 9.06250477e-01 -1.80751592e-01 -1.57927230e-01
2.80230284e-01 -6.66319072e-01 -9.79390368e-02 -9.02598351e-03
2.97556162e-01 4.86457229e-01 4.00628418e-01 -5.29382110e-01
-4.99169797e-01 8.88271689e-01 3.13432276e-01 -3.46984029e-01
1.11824954e+00 -1.80833712e-01 9.66637135e-02 2.60324806e-01
1.05201447e+00 -8.49595517e-02 -1.52213418e+00 8.87933560e-03
-1.99989807e-02 -7.60600626e-01 8.47472325e-02 -7.50676692e-01
-1.05467439e+00 9.16570961e-01 9.62076068e-01 8.76099616e-02
1.12842679e+00 -1.86613008e-01 2.30875686e-01 3.57360572e-01
8.14447343e-01 -1.04891586e+00 -1.21731654e-01 5.79524636e-01
6.64854228e-01 -1.71293557e+00 1.55875891e-01 -3.30216587e-01
-1.98583335e-01 1.28853297e+00 4.14945543e-01 -3.20502847e-01
4.12307203e-01 1.32417947e-01 -6.26272857e-02 1.36995032e-01
-5.07013142e-01 -6.20606005e-01 2.89128006e-01 6.92980945e-01
4.66076553e-01 -1.45819306e-01 -5.96573353e-01 -1.23750433e-01
-2.03661293e-01 3.65735292e-01 4.17000592e-01 7.06960738e-01
-4.09356266e-01 -9.50357735e-01 -1.81757808e-01 2.26421043e-01
-4.09251750e-01 -3.49948972e-01 -5.00404686e-02 1.26018083e+00
4.05141503e-01 9.19056892e-01 1.14183731e-01 -3.43637794e-01
2.71627754e-01 -2.54906982e-01 7.69414008e-01 -6.97193518e-02
-4.51488078e-01 -2.55700558e-01 4.30823326e-01 -8.59327793e-01
-1.08311629e+00 -7.12462127e-01 -6.40782773e-01 -5.12535512e-01
-3.25556636e-01 -2.94617385e-01 6.48475170e-01 9.73680317e-01
-1.56290442e-01 2.95952737e-01 2.21909776e-01 -1.30740166e+00
-3.32614869e-01 -1.19436646e+00 -3.74931723e-01 4.74232882e-01
2.09895819e-01 -1.02265871e+00 -6.17245734e-01 -1.47346884e-01] | [8.699316024780273, -1.8167709112167358] |
8085febe-aaed-4fc5-9620-6da5bf7d0cd2 | risk-averse-decision-making-under-uncertainty | 2109.04082 | null | https://arxiv.org/abs/2109.04082v1 | https://arxiv.org/pdf/2109.04082v1.pdf | Risk-Averse Decision Making Under Uncertainty | A large class of decision making under uncertainty problems can be described via Markov decision processes (MDPs) or partially observable MDPs (POMDPs), with application to artificial intelligence and operations research, among others. Traditionally, policy synthesis techniques are proposed such that a total expected cost or reward is minimized or maximized. However, optimality in the total expected cost sense is only reasonable if system behavior in the large number of runs is of interest, which has limited the use of such policies in practical mission-critical scenarios, wherein large deviations from the expected behavior may lead to mission failure. In this paper, we consider the problem of designing policies for MDPs and POMDPs with objectives and constraints in terms of dynamic coherent risk measures, which we refer to as the constrained risk-averse problem. For MDPs, we reformulate the problem into a infsup problem via the Lagrangian framework and propose an optimization-based method to synthesize Markovian policies. For MDPs, we demonstrate that the formulated optimization problems are in the form of difference convex programs (DCPs) and can be solved by the disciplined convex-concave programming (DCCP) framework. We show that these results generalize linear programs for constrained MDPs with total discounted expected costs and constraints. For POMDPs, we show that, if the coherent risk measures can be defined as a Markov risk transition mapping, an infinite-dimensional optimization can be used to design Markovian belief-based policies. For stochastic finite-state controllers (FSCs), we show that the latter optimization simplifies to a (finite-dimensional) DCP and can be solved by the DCCP framework. We incorporate these DCPs in a policy iteration algorithm to design risk-averse FSCs for POMDPs. | ['Aaron D. Ames', 'Richard M. Murray', 'Michel D. Ingham', 'Ugo Rosolia', 'Mohamadreza Ahmadi'] | 2021-09-09 | null | null | null | null | ['decision-making-under-uncertainty', 'decision-making-under-uncertainty'] | ['medical', 'reasoning'] | [ 1.81843862e-01 6.38494551e-01 -4.40988481e-01 -3.11359260e-02
-8.43248963e-01 -5.86138606e-01 4.66539025e-01 9.70684886e-02
-6.38712108e-01 1.14367437e+00 -6.10955618e-02 -5.70755482e-01
-6.92380965e-01 -8.20188642e-01 -7.27339327e-01 -9.63566065e-01
-2.96398789e-01 6.48321092e-01 -2.09816117e-02 -1.86606701e-02
3.90208885e-02 4.90999371e-01 -1.15009272e+00 -4.36910599e-01
9.00626242e-01 1.14924324e+00 3.49387258e-01 5.19391000e-01
2.16041967e-01 7.46804893e-01 -5.05266845e-01 -2.74087973e-02
3.81230682e-01 -2.05287486e-01 -7.41383791e-01 8.19614828e-02
-7.62620151e-01 -6.01412773e-01 1.35972112e-01 1.22501528e+00
3.67835402e-01 5.34889221e-01 6.63211882e-01 -1.62805068e+00
1.54767735e-02 3.26753318e-01 -3.88416409e-01 -4.18471515e-01
1.62761509e-01 1.85814545e-01 8.53422761e-01 -1.69655234e-01
3.09506863e-01 1.61875093e+00 -2.87953448e-02 7.91331470e-01
-1.25728166e+00 -6.56044297e-03 6.05899990e-01 -1.36324093e-01
-9.72718596e-01 7.78305605e-02 1.41063437e-01 -4.27138090e-01
7.24218607e-01 3.05094540e-01 6.49892390e-01 7.69886255e-01
6.81075692e-01 1.00629354e+00 1.17666399e+00 -2.20628425e-01
9.82944131e-01 -1.74077451e-01 -1.24559551e-01 3.15755844e-01
6.44613683e-01 4.86544192e-01 3.78785372e-01 -4.31267679e-01
4.03559983e-01 1.81610212e-01 -8.94938931e-02 -4.10867393e-01
-1.04737759e+00 1.02449501e+00 -1.10332586e-01 -2.53471464e-01
-8.30988288e-01 2.78914511e-01 1.38047606e-01 3.29956412e-01
6.80802688e-02 5.12965977e-01 -1.93859816e-01 -4.89372462e-02
-3.64350021e-01 6.88611925e-01 1.15625930e+00 9.06254828e-01
1.59100175e-01 2.44826108e-01 -5.34020483e-01 1.33021578e-01
4.67506319e-01 8.44582438e-01 -3.47372204e-01 -1.39887917e+00
5.48692048e-01 2.23494824e-02 1.13405859e+00 -4.25217658e-01
-3.44880700e-01 -4.57763523e-02 -3.35868716e-01 7.37596571e-01
3.13807040e-01 -5.45461297e-01 -8.07886422e-01 1.92543519e+00
1.59882158e-01 -4.35762823e-01 3.46391976e-01 8.00443172e-01
-5.38007855e-01 9.81034696e-01 -1.20251603e-01 -1.04356802e+00
1.04734659e+00 -3.55985016e-01 -1.06086934e+00 -8.45709890e-02
2.08459646e-01 -1.17279872e-01 8.46736133e-01 7.20593035e-01
-1.33331001e+00 3.77383471e-01 -1.14187264e+00 7.62523472e-01
3.45882207e-01 -4.82035935e-01 6.82595968e-02 6.63522243e-01
-8.62990737e-01 6.59460366e-01 -1.12005401e+00 -1.29330605e-01
1.34653151e-01 4.36157137e-01 3.70770246e-01 1.48795648e-02
-1.04154420e+00 1.20282495e+00 6.22622013e-01 1.08976677e-01
-1.94064975e+00 -9.01957080e-02 -5.88474095e-01 6.21116012e-02
1.26279664e+00 -3.91904116e-01 1.48951864e+00 -4.32813883e-01
-1.85544097e+00 1.69979960e-01 3.24785352e-01 -4.97274578e-01
7.79642999e-01 -9.83781517e-02 -7.40665644e-02 2.21604303e-01
2.83104810e-03 1.82513669e-02 9.60274816e-01 -1.37417674e+00
-6.92035675e-01 -3.64900045e-02 8.02531421e-01 3.83787692e-01
3.18765976e-02 1.82311669e-01 9.70836356e-02 -2.01414868e-01
-2.10507110e-01 -1.37758017e+00 -9.04476523e-01 -3.52387041e-01
-6.10083520e-01 7.18308985e-02 4.99725163e-01 -4.67175663e-01
1.15709066e+00 -1.68851435e+00 5.66762567e-01 2.55079031e-01
-2.04255879e-01 -1.09953538e-01 1.44263193e-01 5.68206608e-01
3.91324908e-01 1.65556565e-01 -6.58054173e-01 -2.55734682e-01
4.20323431e-01 8.03976178e-01 -5.22036731e-01 5.62491298e-01
1.02992579e-01 4.60375071e-01 -9.22507584e-01 -1.91922918e-01
6.70478120e-02 -2.98571706e-01 -5.84832430e-01 2.92596996e-01
-6.77744985e-01 5.61872661e-01 -1.01919258e+00 5.28079689e-01
4.40538704e-01 3.71454865e-01 3.03412616e-01 5.49017549e-01
-3.40405464e-01 -1.54023513e-01 -1.30161786e+00 9.93475795e-01
-7.05278099e-01 -2.29729782e-03 6.97370052e-01 -1.11529958e+00
5.31366885e-01 4.43587273e-01 6.86599791e-01 -5.73932081e-02
2.88333476e-01 1.10927589e-01 -1.55212775e-01 -2.12550104e-01
5.56193471e-01 -7.40513921e-01 -6.26896739e-01 6.74871147e-01
-3.71974021e-01 -5.27695000e-01 -1.14405259e-01 -6.38300329e-02
1.14122272e+00 -1.50027215e-01 5.18132031e-01 -4.74773914e-01
3.36981684e-01 5.91565780e-02 9.24555361e-01 7.70480156e-01
-3.63253921e-01 2.98696887e-02 1.07154524e+00 1.77516609e-01
-7.38492489e-01 -1.06710863e+00 7.04832515e-03 5.83866715e-01
3.17192584e-01 1.02721930e-01 -6.78788722e-01 -5.86925149e-01
1.40616288e-02 1.19386280e+00 -3.24969262e-01 -1.72553867e-01
-3.54943633e-01 -7.58889318e-01 6.22102246e-03 1.49919525e-01
1.49774820e-01 -7.31650472e-01 -9.38183606e-01 6.08207762e-01
1.25933766e-01 -8.86051178e-01 -1.43446341e-01 2.82280624e-01
-7.87054360e-01 -8.38841438e-01 -9.17500556e-01 7.32117472e-03
6.58644855e-01 -1.47944808e-01 5.62555432e-01 -6.06440723e-01
3.29140753e-01 7.59367585e-01 -6.48737922e-02 -8.47879529e-01
-5.26622355e-01 -5.34497499e-01 6.08869135e-01 8.30229819e-02
-3.66473794e-01 -2.18581781e-01 -3.23824674e-01 4.69622642e-01
-1.04928887e+00 -1.07692957e-01 2.52394259e-01 7.42988169e-01
9.20656204e-01 4.77438271e-01 6.16068423e-01 -3.56231511e-01
1.01689136e+00 -3.92938733e-01 -1.21783018e+00 5.01217663e-01
-6.16826117e-01 6.53656960e-01 6.65253699e-01 -4.20300901e-01
-1.31179547e+00 -1.77215934e-02 2.65200287e-01 -3.71375382e-01
2.27728501e-01 5.93534291e-01 -6.41824841e-01 5.92875183e-02
-2.30952892e-02 -4.08701977e-04 1.23340949e-01 -2.07992852e-01
2.00650916e-01 5.71294129e-01 8.35099742e-02 -1.27552426e+00
6.11679852e-01 5.96471429e-01 3.93518060e-01 -3.67307901e-01
-7.98125029e-01 1.48913756e-01 5.24114333e-02 -4.85837609e-01
9.78342712e-01 -6.20074630e-01 -1.24952340e+00 6.98491037e-02
-9.79602039e-01 -3.43087137e-01 -6.63763702e-01 7.07193732e-01
-1.43314826e+00 1.25086293e-01 -2.92277008e-01 -1.73494363e+00
1.76414296e-01 -1.46314955e+00 5.91111183e-01 9.58267376e-02
1.70711562e-01 -8.50990415e-01 8.89791772e-02 -1.62458159e-02
1.64154232e-01 7.74173021e-01 7.35759258e-01 -2.34464288e-01
-6.56017959e-01 -1.93243474e-01 4.55548555e-01 5.45914710e-01
-1.93317816e-01 -2.92565584e-01 -4.62555707e-01 -6.15226865e-01
5.74930370e-01 -3.14384878e-01 3.12805980e-01 6.76864862e-01
8.58172238e-01 -8.37955713e-01 -3.02630961e-01 1.68682396e-01
1.68682063e+00 9.12196636e-01 3.55648369e-01 4.19491410e-01
1.86825946e-01 9.92129385e-01 1.18056738e+00 9.41621423e-01
2.22125426e-01 5.61611533e-01 1.01900697e+00 6.52029693e-01
1.04142308e+00 -1.07470065e-01 8.61277163e-01 2.39156306e-01
-4.36539650e-01 -3.26617330e-01 -7.85280228e-01 4.92603719e-01
-2.18954492e+00 -9.83345568e-01 1.93405479e-01 2.75723529e+00
6.30766511e-01 1.61040679e-01 2.33944744e-01 -1.43579707e-01
8.35535944e-01 -1.38733312e-01 -9.63791251e-01 -8.30825448e-01
7.27685392e-02 -2.43281797e-01 1.03304040e+00 6.18833303e-01
-8.97924840e-01 3.74896586e-01 5.93264961e+00 7.67388821e-01
-6.12594962e-01 2.12286606e-01 4.18615520e-01 -4.46366996e-01
-3.08158904e-01 2.57890254e-01 -7.89850950e-01 4.90774602e-01
1.13349247e+00 -7.26103723e-01 7.18776226e-01 9.15418565e-01
8.96929622e-01 -4.04197216e-01 -1.03300393e+00 5.33150911e-01
-8.31324339e-01 -7.23474681e-01 -5.75507820e-01 4.89945412e-01
9.19951022e-01 -3.48932385e-01 -3.93173322e-02 2.06489310e-01
1.11441278e+00 -8.48626435e-01 1.19880986e+00 4.88300771e-01
5.26005268e-01 -1.24224198e+00 5.21197498e-01 6.44686520e-01
-9.85973537e-01 -6.92734480e-01 -3.27415764e-01 -1.94655731e-01
9.26932633e-01 5.28565764e-01 -2.88783997e-01 6.36672258e-01
2.50467598e-01 1.50455907e-01 6.85169160e-01 8.87830317e-01
-4.06054258e-01 3.12925458e-01 -5.17050982e-01 -2.13880718e-01
4.96420175e-01 -5.70432603e-01 9.13612843e-01 4.91236269e-01
5.11104941e-01 2.15786904e-01 5.41004658e-01 8.87088418e-01
5.73754847e-01 -4.35877472e-01 -4.49995488e-01 -3.37802619e-01
3.44176620e-01 8.69134963e-01 -7.57571280e-01 -1.15119807e-01
-1.60358220e-01 6.17075384e-01 -1.70737877e-01 4.57938731e-01
-9.50054646e-01 -2.63049960e-01 1.06712389e+00 -3.69770467e-01
-9.23232269e-03 -2.38202497e-01 -1.89256057e-01 -9.60520864e-01
8.51280093e-02 -6.94369733e-01 4.13827538e-01 -2.82744259e-01
-1.14652586e+00 2.79097140e-01 5.16105950e-01 -1.39956081e+00
-5.24869919e-01 -5.16749561e-01 -6.56246781e-01 7.77553082e-01
-1.26015222e+00 -4.27718759e-01 5.37200689e-01 5.84068775e-01
1.47763178e-01 2.01839581e-01 5.27491450e-01 -3.12958062e-01
-7.70288348e-01 -1.88068956e-01 6.16728544e-01 -6.38981462e-01
4.70239557e-02 -1.39046800e+00 -2.30869099e-01 9.57086623e-01
-9.43368971e-01 1.41913712e-01 1.13924122e+00 -7.36981928e-01
-1.68735123e+00 -9.90676165e-01 1.96840882e-01 -4.45753857e-02
8.10625076e-01 -1.58874527e-01 -5.23303151e-01 5.89777470e-01
-1.20427065e-01 -2.19211042e-01 1.53892832e-02 -3.97988558e-01
6.14657581e-01 2.26928070e-01 -1.45663571e+00 7.14112520e-01
8.85339797e-01 -1.28105208e-02 -5.80206573e-01 4.35108125e-01
9.84143972e-01 -2.05616444e-01 -8.42341125e-01 3.32866311e-01
3.13690275e-01 -3.07238817e-01 7.14283943e-01 -8.63859296e-01
-2.64294762e-02 -3.44809115e-01 -3.54110777e-01 -1.56559312e+00
3.90419513e-02 -1.14956582e+00 -2.76955903e-01 8.30378294e-01
2.44021922e-01 -7.78089643e-01 3.77968311e-01 1.02746034e+00
-2.38150135e-01 -7.62002349e-01 -1.21833229e+00 -1.54327846e+00
1.85733661e-01 -6.55051351e-01 5.13254642e-01 3.24494034e-01
2.92387486e-01 -4.18697298e-01 -5.77500463e-01 3.34186316e-01
8.93840313e-01 3.24493200e-02 1.26088306e-01 -7.33954489e-01
-6.88509107e-01 -1.74410000e-01 8.63591507e-02 -7.94067919e-01
1.77705020e-01 -3.71047795e-01 7.39993274e-01 -1.71395648e+00
-3.83629873e-02 -6.34619355e-01 -2.10321322e-01 2.54266024e-01
1.09241322e-01 -1.00746202e+00 7.55423844e-01 2.19693050e-01
-5.50748944e-01 9.74623144e-01 1.38059986e+00 -1.32356703e-01
-4.33669567e-01 6.39504254e-01 -5.13876081e-01 7.37996519e-01
7.06498742e-01 -5.92960775e-01 -6.69672310e-01 -4.70452756e-02
2.76132941e-01 1.02264357e+00 -2.94530299e-02 -8.01065207e-01
-1.32474184e-01 -1.22061205e+00 -7.05165327e-01 -5.38694978e-01
4.63751674e-01 -9.53700006e-01 3.79758060e-01 9.81952190e-01
-3.36108863e-01 -3.85235772e-02 -1.16091155e-01 1.05343270e+00
-1.54056875e-02 -6.37334645e-01 8.41261327e-01 -5.75835593e-02
-3.37422162e-01 4.26357597e-01 -1.03059745e+00 2.46893112e-02
1.57746160e+00 2.55022347e-01 2.53041591e-02 -7.75711834e-01
-9.70536649e-01 7.74103045e-01 3.42472583e-01 -5.38420118e-02
5.49949527e-01 -1.08798826e+00 -3.59052956e-01 -5.89398026e-01
-3.74026895e-01 1.35149449e-01 1.66203648e-01 9.27505672e-01
-1.44420221e-01 5.70667267e-01 -1.65406018e-01 -2.51839638e-01
-6.18433654e-01 7.34163880e-01 3.24473590e-01 -6.46432996e-01
-1.53559536e-01 4.03790146e-01 2.01310456e-01 -1.56645223e-01
2.35418841e-01 -5.23865581e-01 3.90739664e-02 9.53616649e-02
5.32251239e-01 5.78871727e-01 -2.52498031e-01 -1.83795407e-01
-1.71910986e-01 6.35982752e-02 3.24770302e-01 -8.84020984e-01
1.34609592e+00 -3.43873113e-01 5.90280853e-02 1.98019147e-01
6.52543426e-01 -4.39069420e-01 -1.87064254e+00 7.26313964e-02
2.83120245e-01 -1.53027698e-01 1.97045445e-01 -4.77694869e-01
-7.19865859e-01 6.77401602e-01 2.20674768e-01 3.96238804e-01
1.08806610e+00 -2.59740591e-01 4.43621367e-01 5.25318623e-01
9.84352589e-01 -1.36715591e+00 -2.35323370e-01 6.89597070e-01
8.89486015e-01 -8.10936511e-01 -2.52489537e-01 -1.21253200e-01
-1.09262455e+00 9.18986320e-01 2.03237206e-01 -1.59545809e-01
6.25408351e-01 6.28134072e-01 -5.40729523e-01 3.27357531e-01
-8.86495829e-01 -4.04226959e-01 -1.76760703e-01 6.00664020e-01
-4.69539851e-01 7.04196155e-01 -6.72759414e-01 9.19308603e-01
2.27933392e-01 2.12897677e-02 8.85835528e-01 1.60786152e+00
-7.26234019e-01 -1.02498066e+00 -7.91997850e-01 3.51120472e-01
-4.17675734e-01 3.64703000e-01 4.89641763e-02 4.95965809e-01
-3.44693691e-01 1.20089102e+00 -5.40866107e-02 -1.77017704e-01
3.37636441e-01 -1.33037820e-01 3.86951417e-01 -7.14249313e-01
-1.88216753e-02 3.66272293e-02 3.60366464e-01 -6.93130434e-01
-3.82353872e-01 -7.99174309e-01 -1.35816061e+00 -3.06333929e-01
-2.63155103e-01 1.24997199e-01 7.67311990e-01 1.24253762e+00
-1.30163789e-01 5.38279414e-01 8.26428711e-01 -7.60338306e-01
-1.45300138e+00 -5.25599182e-01 -9.64127183e-01 -3.04452181e-01
1.42727777e-01 -8.56913269e-01 -3.99392843e-01 -4.86714840e-01] | [4.551140308380127, 2.4308853149414062] |
a9f4da55-a49c-4f9d-b46a-0c873cf21ea8 | simple-diffusion-end-to-end-diffusion-for | 2301.11093 | null | https://arxiv.org/abs/2301.11093v1 | https://arxiv.org/pdf/2301.11093v1.pdf | simple diffusion: End-to-end diffusion for high resolution images | Currently, applying diffusion models in pixel space of high resolution images is difficult. Instead, existing approaches focus on diffusion in lower dimensional spaces (latent diffusion), or have multiple super-resolution levels of generation referred to as cascades. The downside is that these approaches add additional complexity to the diffusion framework. This paper aims to improve denoising diffusion for high resolution images while keeping the model as simple as possible. The paper is centered around the research question: How can one train a standard denoising diffusion models on high resolution images, and still obtain performance comparable to these alternate approaches? The four main findings are: 1) the noise schedule should be adjusted for high resolution images, 2) It is sufficient to scale only a particular part of the architecture, 3) dropout should be added at specific locations in the architecture, and 4) downsampling is an effective strategy to avoid high resolution feature maps. Combining these simple yet effective techniques, we achieve state-of-the-art on image generation among diffusion models without sampling modifiers on ImageNet. | ['Tim Salimans', 'Jonathan Heek', 'Emiel Hoogeboom'] | 2023-01-26 | null | null | null | null | ['conditional-image-generation'] | ['computer-vision'] | [ 2.83036590e-01 1.40366927e-01 1.80110902e-01 -6.87410235e-02
-5.07356107e-01 -1.50026739e-01 6.27681851e-01 -4.23879236e-01
-7.07946479e-01 6.47198260e-01 5.07076383e-01 2.32355148e-01
-2.18290344e-01 -1.01642644e+00 -4.09110665e-01 -7.96349347e-01
2.67376184e-01 1.03946917e-01 6.75845444e-01 -3.77623796e-01
3.82352799e-01 6.15263045e-01 -1.41686440e+00 4.22020137e-01
8.69340360e-01 5.53302884e-01 5.65770507e-01 5.79408884e-01
-1.56848505e-01 9.54346776e-01 -6.69768333e-01 -3.29701364e-01
3.63374859e-01 -7.69416213e-01 -7.78879941e-01 2.25593746e-01
5.21679163e-01 -8.01275730e-01 -4.15595829e-01 1.40870512e+00
5.98887742e-01 -8.85770246e-02 5.50647557e-01 -6.55474126e-01
-1.32041240e+00 8.96672249e-01 -9.45001364e-01 7.39676535e-01
-1.36665657e-01 2.14925170e-01 5.65837502e-01 -6.50514901e-01
8.61686885e-01 1.55085719e+00 5.14867961e-01 7.91419089e-01
-1.66618252e+00 -5.46490490e-01 4.47517008e-01 -9.89567861e-03
-1.23615324e+00 -4.97177809e-01 7.50933945e-01 -3.29375088e-01
1.03224385e+00 -1.54148281e-01 5.72768927e-01 1.39503467e+00
2.55036652e-01 3.69733900e-01 1.47807229e+00 -1.94763154e-01
2.84808427e-01 5.64424023e-02 1.17773332e-01 3.25489253e-01
3.46636832e-01 5.76722808e-02 -4.96110678e-01 1.78465337e-01
1.62573826e+00 -6.61709234e-02 -3.12778682e-01 -8.76911264e-03
-1.15202081e+00 9.23422515e-01 5.81608951e-01 6.32413745e-01
-8.64092648e-01 2.31617227e-01 4.79009934e-02 4.86590385e-01
7.35722125e-01 5.26007593e-01 1.56935491e-03 7.91625530e-02
-1.14950395e+00 1.57624647e-01 2.64988631e-01 5.35253108e-01
7.25892127e-01 4.60985601e-01 -1.79422811e-01 8.33676815e-01
9.46998745e-02 6.40414888e-03 6.75766051e-01 -1.29082561e+00
3.12102795e-01 2.69186199e-01 -1.56976730e-02 -7.96276271e-01
-2.94714093e-01 -4.96898234e-01 -1.15561748e+00 7.65129685e-01
3.47757250e-01 -3.90429705e-01 -1.23386979e+00 1.76072180e+00
-4.98342961e-02 1.60282403e-01 -3.87568679e-03 1.10566783e+00
6.07625782e-01 6.25722289e-01 -2.31906530e-02 -1.68717325e-01
1.07571197e+00 -9.95530486e-01 -9.60998178e-01 -3.86603922e-01
2.93153733e-01 -7.38532245e-01 1.07084441e+00 4.07073170e-01
-1.51537359e+00 -6.26203597e-01 -1.28925872e+00 -3.04809421e-01
-3.95916224e-01 -4.75340448e-02 4.46561396e-01 7.22331703e-01
-1.62019062e+00 8.31898868e-01 -9.00483489e-01 -2.73639381e-01
4.70229387e-01 3.52943957e-01 -1.73160166e-01 -1.83212221e-01
-1.16240335e+00 1.09670424e+00 -7.93268532e-03 7.70172477e-02
-9.95301962e-01 -6.84336245e-01 -4.99818057e-01 -4.24676985e-02
-5.40964445e-03 -9.23353553e-01 7.75002360e-01 -1.40740848e+00
-1.54606402e+00 6.66698396e-01 -1.61383674e-01 -6.27144992e-01
6.79076433e-01 -2.25375473e-01 -1.67560264e-01 3.53531986e-01
-2.02076752e-02 9.67293859e-01 1.27825153e+00 -1.45131528e+00
-5.90712845e-01 -4.31928575e-01 1.50586769e-01 2.31580108e-01
-5.24474561e-01 1.85862795e-01 -2.88237840e-01 -8.47585261e-01
2.32354239e-01 -6.70083225e-01 -6.33585095e-01 -5.27866594e-02
-2.63373014e-02 2.82226264e-01 9.90670681e-01 -7.28521049e-01
1.12343955e+00 -1.97714090e+00 4.66554344e-01 -4.69876826e-02
4.85980928e-01 1.70590937e-01 -3.46892476e-01 -2.03971080e-02
-1.28332406e-01 5.10719657e-01 -2.15413511e-01 -2.53306210e-01
-4.07318741e-01 4.60040569e-02 -1.48152769e-01 3.53954911e-01
4.05453771e-01 8.69755328e-01 -5.16528249e-01 -1.62446961e-01
2.16357693e-01 1.01588655e+00 -5.57899654e-01 -2.35759422e-01
1.22065894e-01 5.29714465e-01 -3.24164271e-01 -6.15629591e-02
9.44666803e-01 -2.27800950e-01 -2.01882616e-01 -2.91023523e-01
-3.45926374e-01 1.02524079e-01 -1.44280601e+00 1.60561502e+00
-2.84024775e-01 5.21510243e-01 3.44529301e-01 -6.75241590e-01
8.67684662e-01 1.83038265e-01 3.15564752e-01 -8.50552917e-01
-1.24863818e-01 -7.13954866e-02 1.54083237e-01 -2.93413937e-01
5.08600175e-01 8.09142273e-03 5.51610529e-01 5.62386811e-01
-6.30022436e-02 -7.77066797e-02 2.70937860e-01 1.27990916e-01
1.10518169e+00 6.69632927e-02 -2.42723033e-01 -3.40352654e-01
2.49291375e-01 -1.41133830e-01 4.65633273e-01 8.15493226e-01
-1.02966309e-01 9.81469929e-01 5.92126846e-01 -1.10950872e-01
-1.38196719e+00 -8.99430394e-01 1.45410389e-01 7.12240696e-01
1.16537668e-01 -1.11693360e-01 -1.15667391e+00 -3.45777273e-01
-5.46827316e-01 5.63460529e-01 -9.22595024e-01 -8.40403792e-03
-7.04030573e-01 -1.08211994e+00 3.95897239e-01 5.97438693e-01
9.87675548e-01 -8.97384107e-01 -6.17908657e-01 1.99866340e-01
4.41279309e-03 -8.65341902e-01 -2.98946321e-01 1.42403796e-01
-1.30516934e+00 -4.54985857e-01 -1.24554241e+00 -7.40354717e-01
8.44582617e-01 5.26815712e-01 8.92731428e-01 -3.43203023e-02
-2.06260875e-01 9.66350064e-02 -2.13084206e-01 1.64654240e-01
-4.00815547e-01 1.71511635e-01 -3.17157447e-01 -3.10935695e-02
2.18270704e-01 -6.98556542e-01 -1.00544512e+00 1.97383821e-01
-1.16274512e+00 3.45589295e-02 8.32883596e-01 7.21844673e-01
7.15767503e-01 4.98515993e-01 6.24113619e-01 -1.00243759e+00
1.19613349e+00 -1.42645210e-01 -4.95221287e-01 -1.36674792e-01
-8.42484176e-01 1.04554653e-01 5.29390037e-01 -7.11153328e-01
-1.55702198e+00 -1.23392373e-01 -2.93824524e-01 -3.41666788e-01
-2.09060341e-01 1.83484018e-01 5.89344800e-02 -1.32965848e-01
1.07689095e+00 7.00423047e-02 7.49607235e-02 -5.98520100e-01
5.14574349e-01 2.36360326e-01 7.62110204e-02 -2.15944499e-01
7.04272091e-01 8.91207874e-01 -2.68585652e-01 -9.60801899e-01
-5.07460535e-01 6.46729469e-02 -9.83808279e-01 -9.28474404e-03
1.11984229e+00 -9.29410696e-01 -4.88672741e-02 6.06554031e-01
-9.27309632e-01 -5.25282800e-01 -4.04413968e-01 3.10577542e-01
-3.65708381e-01 2.98404694e-01 -1.13983321e+00 -6.26643479e-01
-1.72586009e-01 -1.48981082e+00 6.40829980e-01 3.40173423e-01
-2.20773518e-01 -9.79013562e-01 -3.69527549e-01 1.67804450e-01
1.05429208e+00 -7.10004941e-02 8.24006557e-01 5.20653576e-02
-7.36241341e-01 2.42091864e-01 -4.76976424e-01 6.78161025e-01
7.45250508e-02 -4.32409719e-02 -8.60533953e-01 -2.20735595e-01
4.99287277e-01 -1.05543047e-01 1.18192554e+00 7.65212297e-01
9.70551252e-01 -2.42974192e-01 -9.70434994e-02 5.50294161e-01
1.44485140e+00 -7.17833862e-02 1.10957217e+00 5.24963081e-01
6.11844957e-01 6.77830458e-01 1.49468020e-01 2.05725878e-02
1.88333288e-01 3.97185177e-01 1.82763427e-01 -4.87909526e-01
-7.89557934e-01 -7.46645704e-02 3.71256858e-01 6.58682644e-01
-2.70542115e-01 1.02803834e-01 -5.83657861e-01 6.08632147e-01
-1.66464663e+00 -1.13091993e+00 -3.06124151e-01 1.89704883e+00
7.65366673e-01 2.93720961e-01 -4.85176817e-02 6.30300269e-02
7.53229022e-01 3.47882420e-01 -5.96080959e-01 -1.83054805e-01
-4.81717795e-01 5.08259377e-03 6.22049928e-01 6.35900974e-01
-8.41822863e-01 1.04250169e+00 7.06400919e+00 6.33457482e-01
-1.28983688e+00 4.04650956e-01 8.86418998e-01 -3.13348591e-01
-2.83556223e-01 6.71544159e-03 -1.00297141e+00 3.04647088e-01
8.99580359e-01 1.43362712e-02 6.41956806e-01 4.81231242e-01
4.43316579e-01 -9.85842794e-02 -7.15891540e-01 7.81839788e-01
5.43877892e-02 -1.31684983e+00 3.26356292e-01 1.91948920e-01
9.75738227e-01 9.10019055e-02 5.51914454e-01 -1.14215948e-01
4.44125295e-01 -1.13782430e+00 6.11991346e-01 6.08697414e-01
4.15134877e-01 -7.81546772e-01 4.74466950e-01 3.23723912e-01
-6.89989567e-01 -1.82331219e-01 -7.64501810e-01 7.42919594e-02
3.36583346e-01 6.04485631e-01 9.48189944e-02 6.40111193e-02
1.04768765e+00 5.74169695e-01 -7.02907860e-01 8.88831675e-01
-4.30245072e-01 3.88880014e-01 -1.96262196e-01 5.01009941e-01
2.87125111e-01 -3.89749587e-01 2.74934679e-01 1.09727836e+00
3.67675483e-01 1.70468371e-02 -3.58355075e-01 1.13124263e+00
5.37196882e-02 -2.26881549e-01 -5.32411933e-01 1.89087272e-01
1.59522206e-01 1.23768997e+00 -9.36229050e-01 -3.85212034e-01
-6.01867557e-01 1.32521927e+00 3.00019592e-01 8.03004682e-01
-5.75383782e-01 -9.76147428e-02 5.61135590e-01 4.55215394e-01
6.06297016e-01 -2.68028378e-01 -6.76656783e-01 -1.00691533e+00
-3.02604049e-01 -1.05344319e+00 1.10769600e-01 -8.73355925e-01
-1.35641491e+00 8.44114542e-01 -1.17261916e-01 -7.26818621e-01
-1.39465600e-01 -2.93058306e-01 -2.93456435e-01 1.20026922e+00
-1.74962473e+00 -1.00181234e+00 -2.34956190e-01 7.46137559e-01
6.71755552e-01 1.76038489e-01 6.32588804e-01 5.04980624e-01
-5.47488451e-01 2.72322446e-01 3.27924490e-02 -4.74897400e-02
6.10645652e-01 -1.04286039e+00 5.02775431e-01 1.12839592e+00
-3.98757942e-02 7.85316288e-01 6.99316919e-01 -7.71749556e-01
-9.82303619e-01 -9.29562211e-01 5.63100219e-01 -3.64006668e-01
6.27592266e-01 -2.00766236e-01 -1.31959641e+00 5.73196054e-01
5.82066119e-01 -2.09749848e-01 1.98619023e-01 -2.36541182e-01
-1.83286384e-01 -4.75758836e-02 -1.25296485e+00 1.00380647e+00
1.06566405e+00 -3.31238240e-01 -3.88072193e-01 -3.77996112e-05
5.13059676e-01 -1.77663699e-01 -9.84787583e-01 1.06421165e-01
1.43574670e-01 -1.25315952e+00 1.11978066e+00 -2.05147684e-01
6.67549908e-01 -2.94965565e-01 2.31957272e-01 -1.35581338e+00
-7.90125549e-01 -4.47406948e-01 -8.04130360e-02 1.33126843e+00
5.98746717e-01 -6.01881206e-01 8.52079093e-01 7.54622340e-01
2.61094958e-01 -3.69388551e-01 -8.50900531e-01 -4.93183732e-01
4.64486301e-01 8.61066356e-02 3.37912232e-01 9.30757284e-01
-6.50248349e-01 5.50574243e-01 -4.54053968e-01 1.58656031e-01
7.87556767e-01 -3.63343447e-01 4.02320206e-01 -1.05778527e+00
-2.01275006e-01 -7.03826487e-01 -3.70282301e-04 -1.19329512e+00
-1.28182799e-01 -3.92984867e-01 -2.74300873e-01 -1.99125886e+00
6.17549792e-02 -8.66196528e-02 -2.27735087e-01 1.79817930e-01
-5.81182688e-02 3.90879005e-01 8.74307752e-02 4.20206904e-01
2.38080916e-04 3.50170821e-01 1.73159635e+00 -6.39484152e-02
-2.14420795e-01 -2.56525308e-01 -9.44653988e-01 8.31685603e-01
7.60649323e-01 -3.93642008e-01 -8.22718203e-01 -9.92243290e-01
5.02087362e-02 -3.07659116e-02 3.05708557e-01 -7.85269320e-01
4.28907245e-01 1.50982738e-02 7.38830864e-01 -3.36377382e-01
4.69388962e-01 -7.94964075e-01 8.91676694e-02 3.74148279e-01
-4.85193670e-01 2.50888109e-01 1.12642303e-01 4.17845964e-01
-2.04770580e-01 -3.25830549e-01 1.28293049e+00 -5.91101408e-01
-8.02469909e-01 7.67337307e-02 -6.45223200e-01 -1.95165008e-01
8.12810600e-01 -4.34609860e-01 -5.60181379e-01 -4.18411285e-01
-7.06959426e-01 -2.23552242e-01 7.15623200e-01 5.28012097e-01
5.79272211e-01 -9.00502026e-01 -7.49415874e-01 1.01768523e-01
-5.62651694e-01 -1.72796875e-01 5.19277811e-01 7.13399351e-01
-4.31665927e-01 1.03919776e-02 -4.78608251e-01 -2.53339916e-01
-9.76846874e-01 6.27290189e-01 4.82373476e-01 -2.73169786e-01
-1.14811110e+00 9.57684398e-01 3.58883411e-01 1.77874684e-01
1.67340145e-01 -9.77118462e-02 -3.99501503e-01 2.35189646e-01
7.48060942e-01 3.05428147e-01 -8.84318948e-02 -6.35817170e-01
7.72112655e-03 6.86337829e-01 -5.54974556e-01 -5.19007266e-01
1.47420752e+00 -5.11178911e-01 -2.16517672e-01 2.05496550e-01
8.48871708e-01 -2.42182866e-01 -1.80425560e+00 -1.28260836e-01
-1.26197711e-01 -4.74692047e-01 6.54936492e-01 -6.54338658e-01
-1.44610417e+00 1.05430245e+00 9.42267001e-01 3.03429008e-01
1.27739894e+00 -2.82112241e-01 5.90618014e-01 -1.38625681e-01
-1.36162015e-02 -1.31777585e+00 1.92994252e-01 3.59079987e-01
9.48426604e-01 -1.00688136e+00 1.19930558e-01 -2.10034743e-01
-8.84627223e-01 8.93235028e-01 7.83212841e-01 -4.72300470e-01
6.32122278e-01 5.93519390e-01 2.79505998e-01 -2.56496578e-01
-8.56906235e-01 -2.56941736e-01 -1.31303072e-01 7.81426072e-01
6.82634354e-01 -6.35512888e-01 -4.77147877e-01 2.20938325e-01
-4.98246998e-02 5.01847789e-02 7.69439578e-01 7.29034185e-01
-5.07966816e-01 -9.80293036e-01 -3.57752174e-01 4.15456295e-01
-7.95580864e-01 -2.40159646e-01 -2.80580997e-01 6.01565361e-01
-1.36190966e-01 9.97842431e-01 2.22827829e-02 -2.86921151e-02
4.06948805e-01 -1.89979002e-01 4.97535586e-01 -4.66514707e-01
-6.45273447e-01 4.16253984e-01 -2.40752220e-01 -5.41182101e-01
-4.94279534e-01 -3.73992056e-01 -9.37586427e-01 -3.36101592e-01
-2.34882697e-01 -2.28571072e-01 7.10982323e-01 5.86931467e-01
5.64420700e-01 9.01580393e-01 1.61091015e-01 -9.50561941e-01
-5.18287778e-01 -1.00030291e+00 -6.38532579e-01 4.00647104e-01
3.15628886e-01 -4.42066938e-01 -3.92642528e-01 1.59917563e-01] | [11.21985912322998, -1.7896032333374023] |
385cd010-6d8c-4cb4-a379-ab76b8958887 | acoustic-model-adaptation-from-raw-waveforms | 1909.13759 | null | https://arxiv.org/abs/1909.13759v1 | https://arxiv.org/pdf/1909.13759v1.pdf | Acoustic Model Adaptation from Raw Waveforms with SincNet | Raw waveform acoustic modelling has recently gained interest due to neural networks' ability to learn feature extraction, and the potential for finding better representations for a given scenario than hand-crafted features. SincNet has been proposed to reduce the number of parameters required in raw-waveform modelling, by restricting the filter functions, rather than having to learn every tap of each filter. We study the adaptation of the SincNet filter parameters from adults' to children's speech, and show that the parameterisation of the SincNet layer is well suited for adaptation in practice: we can efficiently adapt with a very small number of parameters, producing error rates comparable to techniques using orders of magnitude more parameters. | ['Peter Bell', 'Ondřej Klejch', 'Joachim Fainberg', 'Erfan Loweimi', 'Steve Renals'] | 2019-09-30 | null | null | null | null | ['acoustic-modelling'] | ['speech'] | [ 3.13269377e-01 2.17150941e-01 3.62680972e-01 -5.31815886e-01
-5.39087176e-01 -4.62976754e-01 4.22631443e-01 1.74251482e-01
-1.01080155e+00 4.29271936e-01 2.82656014e-01 -1.67927444e-01
-2.24976555e-01 -3.94423813e-01 -5.05019963e-01 -5.04147768e-01
-3.12415421e-01 3.20661724e-01 6.15819633e-01 -1.80580065e-01
-1.33730248e-01 5.25053740e-01 -1.91322052e+00 1.33549407e-01
4.15992647e-01 9.09816444e-01 3.27936441e-01 1.13129544e+00
1.23085834e-01 4.81982380e-02 -8.50001097e-01 -4.11175013e-01
-9.62065440e-03 -3.97572577e-01 -3.74707818e-01 -4.55774248e-01
4.45693910e-01 -1.22616209e-01 -1.93653837e-01 6.90895081e-01
9.31479812e-01 5.98164022e-01 7.14163303e-01 -6.32036984e-01
-2.90505588e-01 8.89162123e-01 4.24856581e-02 4.37323004e-01
2.57079363e-01 -1.60202786e-01 9.91842270e-01 -7.46581316e-01
6.33746535e-02 1.36512649e+00 9.88400102e-01 9.64525044e-01
-1.40265715e+00 -8.09925497e-01 9.85748991e-02 2.22835183e-01
-1.40005314e+00 -9.22931194e-01 4.44843322e-01 -2.07177117e-01
1.39834547e+00 2.32216716e-01 8.60551536e-01 9.82345879e-01
-1.96101338e-01 5.64493179e-01 4.60342944e-01 -6.77564800e-01
3.20463151e-01 -1.96301169e-03 -2.49325290e-01 1.27248183e-01
9.86290281e-04 3.18584651e-01 -5.70169866e-01 -1.11096069e-01
8.89266491e-01 -5.40643811e-01 -3.88418026e-02 7.26637617e-02
-6.79724336e-01 8.67979109e-01 3.33471335e-02 3.83329719e-01
-2.71994501e-01 4.76503968e-01 5.93314409e-01 5.02004087e-01
3.60196948e-01 4.34765428e-01 -7.32727706e-01 -5.47528148e-01
-8.13352346e-01 3.11215639e-01 8.86769354e-01 8.99864495e-01
2.34419480e-01 6.97349906e-01 4.09095705e-01 1.48087454e+00
1.57869935e-01 4.34131652e-01 7.94243753e-01 -8.78914237e-01
2.77220011e-01 -1.75248653e-01 -1.72965273e-01 -6.55147076e-01
-5.44933677e-01 -3.68512690e-01 -5.23769736e-01 1.01323985e-01
5.30841470e-01 -4.75835800e-01 -1.05785334e+00 1.99577391e+00
2.61248648e-01 3.72788429e-01 -1.81895643e-01 4.19062614e-01
5.56341410e-01 6.06390417e-01 1.62185192e-01 -1.24716282e-01
1.00023985e+00 -3.64152521e-01 -5.49818218e-01 -4.91429478e-01
3.58571857e-01 -8.07271540e-01 8.09038401e-01 7.97673583e-01
-1.41093147e+00 -7.15641022e-01 -1.13195300e+00 2.94275433e-01
-3.97841066e-01 -3.59251499e-01 7.65103996e-01 1.12851989e+00
-1.22195804e+00 7.93485522e-01 -8.26707900e-01 -1.10042833e-01
2.84227580e-01 1.03875554e+00 -2.34361738e-01 3.32634956e-01
-1.26966524e+00 1.01805997e+00 7.77333796e-01 1.43424764e-01
-6.14835143e-01 -7.54289150e-01 -9.60787535e-01 3.20269257e-01
-1.05752803e-01 -3.26971650e-01 1.55065465e+00 -6.70536160e-01
-1.70959115e+00 1.90771610e-01 1.01919524e-01 -5.30344486e-01
-1.25164524e-01 -2.83476382e-01 -7.39311814e-01 1.96072087e-01
-7.15773284e-01 7.47889280e-01 1.14322817e+00 -6.17859662e-01
-6.72361374e-01 1.93557739e-01 -3.02790105e-01 -5.67862801e-02
-6.05800033e-01 2.44403273e-01 -6.62194788e-02 -8.73575985e-01
1.01280540e-01 -6.19357944e-01 -3.62069964e-01 -2.30792210e-01
3.00137609e-01 -3.17506671e-01 3.28142107e-01 -5.75052321e-01
1.40809846e+00 -2.08399916e+00 -1.86463490e-01 3.86146545e-01
-2.39043087e-01 7.07302272e-01 -5.02231419e-01 3.99367899e-01
-2.59435505e-01 -8.78955871e-02 -6.12750985e-02 -2.74772644e-01
5.91378845e-02 3.74135375e-01 -5.08352034e-02 2.67922699e-01
2.25146607e-01 4.43133891e-01 -6.83730841e-01 -2.60393709e-01
1.06220461e-01 8.12704325e-01 -9.14962351e-01 2.05408722e-01
-3.49383876e-02 3.74745764e-02 -2.16552809e-01 -5.10034757e-03
2.96001464e-01 4.47484493e-01 4.46140330e-04 2.08465040e-01
5.56066781e-02 8.26305151e-01 -1.49429822e+00 1.30974257e+00
-8.44311953e-01 7.67252147e-01 1.67441025e-01 -9.10875440e-01
1.02275932e+00 7.91031539e-01 1.91536516e-01 -5.35787523e-01
3.03710788e-01 4.13299859e-01 6.13124847e-01 -4.11807835e-01
-2.03783423e-01 -5.28839171e-01 1.48246095e-01 3.34849447e-01
4.59690124e-01 -4.66145337e-01 6.24283999e-02 -3.55189174e-01
1.03599036e+00 -2.02579141e-01 3.19859028e-01 -1.55420423e-01
4.94945228e-01 -7.20056951e-01 2.93363750e-01 7.03997374e-01
2.66346574e-01 5.59344649e-01 -2.07089171e-01 -4.38749284e-01
-1.02413785e+00 -1.01592469e+00 -2.50893384e-01 1.45421910e+00
-8.44518542e-01 -2.68959433e-01 -9.53830242e-01 -6.22255430e-02
-2.20941491e-02 8.70145679e-01 -5.35945714e-01 -2.61971146e-01
-1.00914443e+00 -5.40578127e-01 8.33727360e-01 7.55377471e-01
-1.96306318e-01 -1.36292648e+00 -7.88429439e-01 6.54243112e-01
5.50983608e-01 -8.12172294e-01 -6.21398032e-01 6.57760203e-01
-9.84799564e-01 -5.04793108e-01 -5.83221197e-01 -8.91003370e-01
3.10154766e-01 -3.62555295e-01 8.82437527e-01 1.39280394e-01
-3.24457109e-01 4.30302143e-01 -3.16017479e-01 -8.11241627e-01
-4.50323105e-01 2.32970029e-01 3.66330981e-01 -2.77138233e-01
3.25112909e-01 -1.24939203e+00 -4.56145793e-01 -4.42212336e-02
-9.03113544e-01 -4.05989736e-01 5.53106189e-01 8.40813160e-01
2.23283675e-02 1.63200483e-01 9.45736587e-01 -7.12613583e-01
6.89148724e-01 -9.06546563e-02 -4.41656053e-01 -2.26806268e-01
-3.93839270e-01 1.20253004e-01 7.13756263e-01 -1.01403368e+00
-8.95495236e-01 1.42520458e-01 -7.71659732e-01 -1.59396589e-01
-4.93684530e-01 1.72012255e-01 -2.66363807e-02 -9.79806408e-02
6.97977781e-01 1.67899597e-02 -1.60808176e-01 -8.30840945e-01
5.03558218e-01 6.17209852e-01 7.88087189e-01 -5.99976361e-01
8.89112651e-01 -2.37268165e-01 -1.67299762e-01 -1.11040521e+00
-4.55287516e-01 -3.22731972e-01 -8.01875949e-01 7.26115033e-02
5.05686998e-01 -5.26087463e-01 -9.43318248e-01 2.56764323e-01
-1.22650957e+00 -3.71926695e-01 -4.80443954e-01 8.25168490e-01
-5.63137293e-01 -2.52293069e-02 -4.67705786e-01 -1.07562304e+00
-3.46125871e-01 -6.25918090e-01 5.28558910e-01 3.55861753e-01
-5.54895937e-01 -1.29371810e+00 2.11396024e-01 -3.86040598e-01
5.73329449e-01 -1.48061529e-01 1.14292943e+00 -1.23045003e+00
2.39020541e-01 -4.47452247e-01 3.54770184e-01 7.21226335e-01
7.67741818e-03 -2.63465077e-01 -1.37275517e+00 -3.96515697e-01
1.06689505e-01 -2.27452219e-01 6.66229069e-01 4.50525969e-01
9.48516428e-01 -4.13365841e-01 -6.88983127e-02 5.80346227e-01
9.80084598e-01 6.39545321e-01 3.03887069e-01 1.00461900e-01
3.75851691e-01 5.48342228e-01 2.22802088e-02 3.88295174e-01
3.93663496e-02 6.16784036e-01 -4.16791532e-03 1.01149619e-01
-3.32929224e-01 -2.75338888e-01 1.59237131e-01 1.27970588e+00
-1.91834792e-01 -2.11692005e-01 -7.18313754e-01 8.00132453e-01
-1.35619795e+00 -7.40408182e-01 2.38771364e-01 2.34782887e+00
1.08401489e+00 5.25453210e-01 3.31095338e-01 5.66044688e-01
5.11197805e-01 -1.13272041e-01 -2.66031474e-01 -1.00273168e+00
2.32776448e-01 1.07003725e+00 1.99861154e-01 6.88900113e-01
-7.49436915e-01 7.67881811e-01 8.28211880e+00 1.03002787e+00
-8.46814394e-01 9.60120261e-02 1.63785815e-02 -2.85068542e-01
-2.15714112e-01 -3.24106514e-01 -8.33108783e-01 3.46724659e-01
1.56363297e+00 1.50796413e-01 7.27141798e-01 4.95784938e-01
8.35920274e-02 7.85926282e-02 -1.47956634e+00 8.55666220e-01
-8.69201422e-02 -7.93221474e-01 -7.27941766e-02 -2.49518782e-01
3.59691232e-01 -1.88847780e-01 2.63243150e-02 3.47500592e-01
1.53326645e-01 -1.25211740e+00 7.37714171e-01 2.21878186e-01
7.01335907e-01 -1.02764785e+00 4.19845730e-01 3.56368005e-01
-1.22294545e+00 -2.10938305e-01 -5.71218908e-01 -3.33365381e-01
-2.66427491e-02 2.35752046e-01 -1.27253640e+00 -2.90965259e-01
5.09296417e-01 -6.06712587e-02 -4.12855625e-01 1.39765632e+00
-1.10499717e-01 1.03382790e+00 -7.90615439e-01 -3.17798615e-01
9.90644321e-02 2.81882614e-01 3.26713085e-01 1.64076245e+00
4.62911516e-01 3.20657402e-01 -2.76725769e-01 3.03633094e-01
4.13815975e-01 4.29202281e-02 -3.39938223e-01 6.00334210e-03
9.05572653e-01 8.47499669e-01 -4.20434445e-01 5.04192598e-02
-3.57726395e-01 5.58967471e-01 2.00600401e-01 2.46343300e-01
-3.67942333e-01 -8.70084763e-01 6.39567554e-01 2.38176391e-01
7.61056125e-01 -3.28404725e-01 -1.04441449e-01 -3.33573729e-01
-4.66050804e-01 -8.55499148e-01 2.30909601e-01 -4.45507020e-01
-9.46691871e-01 7.56853640e-01 3.02749038e-01 -6.56852663e-01
-9.85440075e-01 -8.22891653e-01 -8.56054127e-01 1.01847661e+00
-8.66110086e-01 -7.70199537e-01 5.29765248e-01 3.67020100e-01
5.31411707e-01 -2.87515998e-01 1.12494051e+00 3.72106165e-01
-8.78795609e-02 1.08719873e+00 -1.32814482e-01 -5.37350178e-02
3.47282857e-01 -1.25231242e+00 8.96711648e-01 5.24051964e-01
4.04057562e-01 6.73835695e-01 1.15070462e+00 -2.03918934e-01
-9.59347546e-01 -4.27915215e-01 9.97572005e-01 -1.68742865e-01
6.94387555e-01 -7.48346508e-01 -1.15021229e+00 4.64493454e-01
9.85646248e-02 -1.15559623e-01 6.69740260e-01 3.00739944e-01
-2.72571743e-01 -2.66108036e-01 -9.36215043e-01 5.11755645e-01
1.08699811e+00 -4.61328447e-01 -7.58625031e-01 -1.31916985e-01
7.76664019e-01 -3.33640158e-01 -9.79743242e-01 1.99013621e-01
8.59314740e-01 -7.32458770e-01 1.14547384e+00 -4.20868158e-01
-3.31393898e-01 8.79471153e-02 -1.21041805e-01 -1.71711791e+00
-4.31222677e-01 -8.94862831e-01 -1.61744937e-01 1.41546798e+00
5.44755876e-01 -6.75904751e-01 6.99225605e-01 5.07018328e-01
-3.52593541e-01 -7.36250043e-01 -1.22971344e+00 -8.07874441e-01
1.78152323e-01 -8.42656910e-01 7.41868615e-01 2.91142941e-01
-1.99668165e-02 3.68722230e-01 -2.00895086e-01 1.44212157e-01
2.67317653e-01 -7.11981237e-01 3.15017492e-01 -1.34438765e+00
-7.96098232e-01 -5.73265612e-01 -5.28539658e-01 -9.90522683e-01
-4.02767435e-02 -5.22291064e-01 3.37822556e-01 -9.93060827e-01
-5.35616398e-01 -4.49770004e-01 -4.33413893e-01 5.08890867e-01
-2.77950689e-02 3.17286462e-01 1.71050593e-01 -5.85124373e-01
-1.31847262e-02 3.66222233e-01 9.87868369e-01 1.28689647e-01
-2.57818997e-01 5.07823944e-01 -4.96606261e-01 1.04938447e+00
5.97170472e-01 -6.45218194e-01 -7.70112038e-01 -4.64026093e-01
1.43552274e-01 -5.43694720e-02 6.87826127e-02 -1.33101916e+00
3.30058545e-01 -6.40430748e-02 8.50367129e-01 -2.86068976e-01
7.41250992e-01 -7.50913501e-01 1.77668631e-01 2.80618995e-01
-5.31011462e-01 -5.11045335e-03 5.79283714e-01 2.16273084e-01
1.43224467e-02 -9.45169866e-01 9.13908422e-01 -1.07491883e-02
-4.67076302e-01 -6.28594980e-02 -6.48589194e-01 1.24107145e-01
1.80409878e-01 -3.21266890e-01 2.24459812e-01 -5.43474972e-01
-9.91759062e-01 -1.98715135e-01 -1.91789493e-01 2.65454173e-01
6.59121871e-01 -1.20648015e+00 -4.69204634e-01 6.54628754e-01
-4.09271985e-01 -8.80801529e-02 4.63101603e-02 3.43449384e-01
-2.24054083e-01 2.85840690e-01 -5.61644286e-02 -2.23782107e-01
-1.20971382e+00 2.74877250e-01 3.72749060e-01 2.66064435e-01
-7.08966732e-01 1.46085775e+00 8.64771530e-02 -2.11809084e-01
6.69497848e-01 -4.62105870e-01 -3.71681273e-01 1.09309085e-01
7.94339538e-01 5.04208922e-01 2.41768584e-01 -4.05974358e-01
-3.60277146e-01 5.95434546e-01 -1.18414074e-01 -4.45508927e-01
1.58232474e+00 1.03398576e-01 4.04322714e-01 4.72106606e-01
1.02114975e+00 -2.30027512e-02 -1.27017641e+00 -1.23050325e-01
1.90276042e-01 -2.39656225e-01 4.82724831e-02 -8.65867972e-01
-7.05172598e-01 9.66348648e-01 7.81285048e-01 2.66225547e-01
1.32140434e+00 -1.67298421e-01 8.22206497e-01 5.22043645e-01
1.99833021e-01 -1.13681221e+00 -1.12941012e-01 5.24048209e-01
7.95357049e-01 -5.31612635e-01 -4.88136858e-02 -2.10476875e-01
-2.12596953e-01 1.28510225e+00 3.13698947e-01 -3.75835001e-01
8.74279499e-01 6.39625788e-01 1.20473184e-01 1.54904976e-01
-1.08536327e+00 -1.03449695e-01 6.26722395e-01 1.08513725e+00
6.33730114e-01 -1.83544513e-02 4.68625082e-03 7.59761512e-01
-6.76467657e-01 -5.21334708e-01 3.21848720e-01 5.43210268e-01
-8.64675581e-01 -1.32106960e+00 -4.04065341e-01 5.45214057e-01
-5.73834658e-01 -2.13300258e-01 -1.73776019e-02 8.31298828e-01
3.44051659e-01 1.01507962e+00 8.71868655e-02 -3.83556485e-01
7.07034588e-01 4.38124746e-01 8.17036390e-01 -8.76242578e-01
-7.75388360e-01 5.71021080e-01 2.46423542e-01 -1.67002544e-01
-8.04088712e-02 -9.31385994e-01 -1.32453299e+00 1.71167940e-01
-4.62887585e-01 2.74309486e-01 6.97966516e-01 1.10662067e+00
-2.80909538e-01 5.96068382e-01 4.46593791e-01 -1.11914325e+00
-6.77887857e-01 -1.12659252e+00 -6.07815981e-01 -1.55986100e-01
6.03579342e-01 -6.36961937e-01 -3.51250947e-01 4.51418646e-02] | [15.085411071777344, 5.6948323249816895] |
b51f2ed5-eb4c-4dba-85b2-371ca715952d | pushing-the-limits-of-chatgpt-on-nlp-tasks | 2306.09719 | null | https://arxiv.org/abs/2306.09719v1 | https://arxiv.org/pdf/2306.09719v1.pdf | Pushing the Limits of ChatGPT on NLP Tasks | Despite the success of ChatGPT, its performances on most NLP tasks are still well below the supervised baselines. In this work, we looked into the causes, and discovered that its subpar performance was caused by the following factors: (1) token limit in the prompt does not allow for the full utilization of the supervised datasets; (2) mismatch between the generation nature of ChatGPT and NLP tasks; (3) intrinsic pitfalls of LLMs models, e.g., hallucination, overly focus on certain keywords, etc. In this work, we propose a collection of general modules to address these issues, in an attempt to push the limits of ChatGPT on NLP tasks. Our proposed modules include (1) a one-input-multiple-prompts strategy that employs multiple prompts for one input to accommodate more demonstrations; (2) using fine-tuned models for better demonstration retrieval; (3) transforming tasks to formats that are more tailored to the generation nature; (4) employing reasoning strategies that are tailored to addressing the task-specific complexity; (5) the self-verification strategy to address the hallucination issue of LLMs; (6) the paraphrase strategy to improve the robustness of model predictions. We conduct experiments on 21 datasets of 10 representative NLP tasks, including question answering, commonsense reasoning, natural language inference, sentiment analysis, named entity recognition, entity-relation extraction, event extraction, dependency parsing, semantic role labeling, and part-of-speech tagging. Using the proposed assemble of techniques, we are able to significantly boost the performance of ChatGPT on the selected NLP tasks, achieving performances comparable to or better than supervised baselines, or even existing SOTA performances. | ['Guoyin Wang', 'Fei Wu', 'Lingjuan Lyu', 'Fei Cheng', 'Jiwei Li', 'Tianwei Zhang', 'Shuhe Wang', 'Zhen Wan', 'Xiaoya Li', 'Linfeng Dong', 'Xiaofei Sun'] | 2023-06-16 | null | null | null | null | ['question-answering', 'natural-language-inference', 'sentiment-analysis', 'event-extraction', 'dependency-parsing', 'semantic-role-labeling', 'part-of-speech-tagging', 'relation-extraction'] | ['natural-language-processing', 'natural-language-processing', 'natural-language-processing', 'natural-language-processing', 'natural-language-processing', 'natural-language-processing', 'natural-language-processing', 'natural-language-processing'] | [ 7.61038810e-02 4.16473895e-01 -1.63239077e-01 -2.89220691e-01
-8.88272643e-01 -7.21341968e-01 7.58037508e-01 3.76170278e-01
-4.11342591e-01 8.27473640e-01 5.13679326e-01 -4.69934374e-01
-9.80713367e-02 -5.68683267e-01 -4.98112947e-01 -3.29724610e-01
9.44045708e-02 5.78671753e-01 4.93549943e-01 -3.73322845e-01
4.34224159e-01 3.00900757e-01 -1.41412139e+00 6.94886982e-01
9.78596926e-01 7.72163391e-01 2.78802991e-01 4.71101820e-01
-5.21639347e-01 1.46038151e+00 -8.67735445e-01 -3.57862443e-01
-1.12218261e-01 -5.01498520e-01 -1.03544378e+00 -2.63805687e-01
2.27092877e-02 -6.75092041e-02 -1.37862116e-01 7.10657716e-01
4.61956561e-01 1.48851946e-01 7.54794061e-01 -1.44184017e+00
-6.82296693e-01 8.58148336e-01 -9.03060213e-02 1.45969078e-01
6.95798635e-01 3.73530716e-01 9.26637113e-01 -6.89431727e-01
7.48759806e-01 1.47074199e+00 5.12504697e-01 6.34347975e-01
-9.90027428e-01 -6.35222435e-01 3.00860871e-02 3.70607346e-01
-1.14354217e+00 -4.01874393e-01 4.47569549e-01 -1.84048265e-01
1.57945585e+00 1.32655740e-01 3.00388336e-01 1.46997821e+00
1.70978069e-01 7.55420685e-01 1.36982608e+00 -6.27905607e-01
2.69933462e-01 4.64376748e-01 3.05558264e-01 4.76964086e-01
-2.73608506e-01 -1.29936159e-01 -6.90609634e-01 -2.37347618e-01
4.03436214e-01 -4.59177881e-01 -1.92610890e-01 1.81181446e-01
-1.30347109e+00 8.84790957e-01 1.51658103e-01 6.85676575e-01
-3.14862579e-01 -1.15633219e-01 6.00662887e-01 3.95544529e-01
2.37765148e-01 8.74952316e-01 -8.40878427e-01 -4.04229343e-01
-8.48882437e-01 4.04208899e-01 1.19847655e+00 9.06067431e-01
4.70227063e-01 -1.25753284e-01 -4.29512590e-01 8.27325881e-01
8.55074897e-02 2.08536670e-01 8.45637977e-01 -9.02620375e-01
8.75877976e-01 6.94810510e-01 -3.50335613e-02 -6.11769497e-01
-5.88289499e-01 -1.45014912e-01 -3.12397987e-01 -2.43860617e-01
5.40489018e-01 -1.92859769e-01 -6.70757890e-01 2.11827254e+00
1.25645310e-01 -1.21377088e-01 4.11433488e-01 5.49830019e-01
1.11910570e+00 7.37416744e-01 5.34970522e-01 -2.27104887e-01
1.61206973e+00 -1.03809452e+00 -9.09193993e-01 -3.38786095e-01
9.03810740e-01 -8.16034079e-01 1.40830147e+00 2.12431565e-01
-1.02787757e+00 -5.22868335e-01 -7.87741423e-01 -1.57025829e-01
-6.79187536e-01 1.23425201e-01 5.98803163e-01 3.82926971e-01
-8.43511820e-01 6.20347142e-01 -4.27206159e-01 -7.55921781e-01
6.89046877e-03 2.45223958e-02 -3.53525460e-01 1.18182667e-01
-1.53093886e+00 1.25152576e+00 7.03756332e-01 -2.28034779e-01
-6.33670390e-01 -6.52215958e-01 -9.05550897e-01 1.41900077e-01
5.79984665e-01 -5.86128652e-01 1.32239783e+00 -6.99920297e-01
-1.52961779e+00 8.01072299e-01 -1.11506380e-01 -5.19341886e-01
3.84374201e-01 -2.37302810e-01 -4.05990630e-01 1.98274374e-01
4.02101308e-01 6.82188690e-01 4.75117773e-01 -8.12086523e-01
-3.78779054e-01 -2.00408503e-01 2.14098766e-01 4.00724351e-01
-1.82821110e-01 3.19472790e-01 -3.07014044e-02 -5.81963658e-01
-6.49988726e-02 -9.71975982e-01 6.86935410e-02 -6.01031840e-01
-5.84402800e-01 -6.40095174e-01 8.51462305e-01 -6.40076280e-01
1.18653095e+00 -2.20277643e+00 -2.21550409e-02 -4.56629604e-01
-1.31891742e-01 4.57006127e-01 -2.05004916e-01 9.41160202e-01
-2.50581592e-01 2.48716474e-01 6.18392974e-02 -3.97040397e-01
7.80255944e-02 4.09891248e-01 -5.09085715e-01 -1.81868672e-01
4.40117836e-01 9.18114662e-01 -8.59379768e-01 -7.60933340e-01
9.07283947e-02 9.99884084e-02 -2.32521161e-01 3.30459684e-01
-4.78034705e-01 3.97745848e-01 -4.77889925e-01 5.14461994e-01
2.97571301e-01 -1.93478525e-01 -5.82029894e-02 -2.02312887e-01
-2.11353585e-01 1.07413459e+00 -1.15529108e+00 1.54240561e+00
-5.56526005e-01 3.61189127e-01 -1.45877331e-01 -8.40371370e-01
6.77214026e-01 6.75563514e-01 -3.13823447e-02 -5.40332079e-01
-8.32019150e-02 2.08510026e-01 1.23441003e-01 -9.35633123e-01
3.77621263e-01 -3.51613462e-01 -2.08891749e-01 4.41102535e-01
3.79962146e-01 -2.12940037e-01 3.23830456e-01 3.93587351e-01
1.28663802e+00 1.36273891e-01 4.84425694e-01 -1.64147332e-01
5.10375917e-01 2.93123156e-01 5.00382781e-01 7.47401655e-01
-3.44846755e-01 4.15723264e-01 7.39607811e-01 -6.75896034e-02
-5.66465378e-01 -8.72877896e-01 5.56352697e-02 1.15536332e+00
-2.10036844e-01 -5.06306112e-01 -6.75991476e-01 -8.27663898e-01
-2.23007411e-01 1.36627972e+00 -3.61090064e-01 -1.34535745e-01
-5.43502152e-01 -6.77830398e-01 1.15396893e+00 6.59137547e-01
7.93026567e-01 -1.58453286e+00 -5.96600711e-01 3.69577706e-01
-5.36455274e-01 -1.56375682e+00 -1.82736173e-01 6.19187593e-01
-8.10202539e-01 -1.06083274e+00 -2.14143261e-01 -8.36704969e-01
1.89964443e-01 -1.12258464e-01 1.02239156e+00 -6.46847561e-02
1.92726031e-01 2.90932834e-01 -5.87219000e-01 -4.07679468e-01
-7.03591585e-01 1.62996739e-01 -1.58363923e-01 -5.05781472e-01
3.34897757e-01 -5.76588333e-01 -2.56914616e-01 1.16833106e-01
-9.31234002e-01 4.40423265e-02 7.98543096e-01 7.11487055e-01
7.36281127e-02 1.13128778e-02 9.16816533e-01 -1.03115082e+00
9.55481291e-01 -4.60025996e-01 6.87060654e-02 3.16454470e-01
-5.10299444e-01 5.12439385e-02 7.13517666e-01 -6.76556528e-01
-1.32942951e+00 -2.74291694e-01 -2.60646909e-01 -1.03471749e-01
-5.56004405e-01 5.04737496e-01 -3.39732021e-01 2.90422022e-01
7.52694070e-01 4.37924087e-01 -1.35215670e-01 -3.94059092e-01
3.83983195e-01 6.63856983e-01 3.26359898e-01 -7.89653718e-01
5.91961384e-01 -7.65252337e-02 -3.17621350e-01 -6.63540781e-01
-9.51284707e-01 -2.97392398e-01 -3.53454173e-01 2.08086669e-01
9.34688985e-01 -9.08824265e-01 -5.00036180e-01 3.45678717e-01
-1.31142020e+00 -5.53964376e-01 -1.80007830e-01 3.86953235e-01
-4.55216438e-01 4.91130322e-01 -1.11823916e+00 -8.86082053e-01
-4.79919434e-01 -9.80235219e-01 8.46262872e-01 3.63080442e-01
-6.11817122e-01 -8.71493399e-01 1.10119963e-02 7.41454899e-01
5.11445463e-01 9.20705199e-02 1.38439679e+00 -1.40443385e+00
-2.43467227e-01 5.26577085e-02 -1.92408755e-01 5.22443175e-01
-1.52154401e-01 -1.30407035e-01 -9.38207865e-01 1.75785303e-01
2.31315523e-01 -6.98918164e-01 4.86478239e-01 -1.93305954e-01
7.31882572e-01 -5.55030107e-01 -2.70077199e-01 1.61284164e-01
1.06699049e+00 3.90761763e-01 8.27174187e-01 4.90254492e-01
3.11285079e-01 8.65653753e-01 7.16338694e-01 7.40624517e-02
6.32639945e-01 5.60793281e-01 2.25557625e-01 1.97481602e-01
-2.73534447e-01 -5.76531827e-01 6.11562550e-01 5.77023268e-01
1.77720949e-01 -2.97286481e-01 -8.40996325e-01 7.13677168e-01
-1.93236530e+00 -8.86415780e-01 -2.66947418e-01 1.82670259e+00
8.87686789e-01 2.85533071e-01 -6.38957173e-02 9.60630625e-02
5.28512895e-01 2.74533451e-01 -2.05390275e-01 -7.75516629e-01
-8.35521147e-02 1.98362425e-01 -7.04669282e-02 1.54503569e-01
-7.81228781e-01 1.15123379e+00 5.54898405e+00 8.62982571e-01
-9.38056171e-01 2.71083295e-01 2.40544289e-01 2.41461277e-01
3.74990664e-02 1.94888115e-01 -8.99227202e-01 5.47860622e-01
9.27980900e-01 -1.49768731e-02 2.75808394e-01 7.63880789e-01
-5.96997403e-02 -1.21190391e-01 -1.17173612e+00 5.80174685e-01
1.14279315e-01 -1.02652752e+00 1.23982122e-02 -2.75072724e-01
3.12237382e-01 1.37273848e-01 -2.76516467e-01 9.19092774e-01
3.66154276e-02 -9.27313805e-01 7.41217613e-01 9.84215993e-04
2.48115346e-01 -3.96569639e-01 7.27834523e-01 9.04695332e-01
-8.30263138e-01 -6.70024604e-02 -1.12332873e-01 -3.29416394e-01
1.32984594e-01 4.43824708e-01 -1.17063868e+00 5.18145621e-01
5.87600112e-01 2.62518138e-01 -5.87817013e-01 7.46557772e-01
-7.85295486e-01 7.59953320e-01 -2.89300054e-01 -3.43012929e-01
3.25936526e-01 2.17987865e-01 7.51461089e-01 1.55223298e+00
1.68306708e-01 1.46679834e-01 2.51272563e-02 1.12264729e+00
1.23518147e-01 2.44215310e-01 -4.61508691e-01 -2.42371887e-01
6.85222983e-01 1.32622957e+00 -6.69739068e-01 -4.10782844e-01
-3.15037400e-01 8.44406188e-01 2.96965986e-01 3.12960804e-01
-8.80449653e-01 -4.76915002e-01 2.14433134e-01 1.83806226e-01
2.57487625e-01 -3.58752534e-02 -9.63036343e-02 -1.14109671e+00
8.91476348e-02 -1.09180248e+00 5.90244889e-01 -1.05639017e+00
-1.27504754e+00 7.00867236e-01 2.32507959e-01 -7.44441032e-01
-4.80532020e-01 -4.45793480e-01 -9.45356607e-01 8.25777113e-01
-1.46750498e+00 -1.26314950e+00 3.72515805e-02 7.04581797e-01
5.69292068e-01 1.50943631e-02 1.10682917e+00 2.29843736e-01
-4.80825335e-01 3.24996442e-01 -6.72069430e-01 1.00651674e-01
9.59158778e-01 -1.39438450e+00 1.71893716e-01 7.24429488e-01
1.45179316e-01 7.47890770e-01 7.87889183e-01 -5.96126795e-01
-1.11209500e+00 -8.05939734e-01 1.69557989e+00 -6.12081170e-01
8.19915771e-01 -2.98406065e-01 -1.16774380e+00 8.14851820e-01
2.76947588e-01 -3.95773590e-01 6.96277976e-01 2.62167692e-01
-4.79875356e-01 2.45360211e-01 -1.34229088e+00 5.24362922e-01
7.60331452e-01 -6.05692744e-01 -1.30276155e+00 5.29264033e-01
8.27527225e-01 -4.95534182e-01 -6.48993909e-01 3.11257213e-01
1.53293490e-01 -8.34635913e-01 7.31941998e-01 -6.50146663e-01
6.47544205e-01 -2.26617575e-01 -5.75039983e-02 -1.17826831e+00
-9.95108113e-02 -6.00318432e-01 -5.56037687e-02 1.71339619e+00
5.82811952e-01 -8.05955052e-01 3.02161902e-01 7.78280139e-01
-1.93643868e-01 -7.60210037e-01 -9.53594387e-01 -5.54505527e-01
5.04800379e-02 -3.94855529e-01 3.54503393e-01 1.01423156e+00
4.17759955e-01 9.34742749e-01 -8.35054442e-02 1.00944713e-01
1.70578897e-01 1.32507622e-01 5.54049730e-01 -9.49446559e-01
-5.99367261e-01 -1.78151429e-01 2.77842786e-02 -8.67304802e-01
3.68413001e-01 -8.08591962e-01 -1.63419433e-02 -1.66511226e+00
2.10866764e-01 -3.11915338e-01 -5.04984567e-03 9.30286050e-01
-2.12285534e-01 -1.64675906e-01 4.22481716e-01 1.65426031e-01
-5.73944330e-01 2.88697511e-01 1.06425655e+00 2.69436389e-01
-3.20757508e-01 -7.81284645e-02 -1.00586975e+00 8.09349418e-01
8.20329070e-01 -6.45073712e-01 -4.20306116e-01 -2.98757017e-01
2.73712605e-01 3.77578467e-01 4.41489935e-01 -8.80876660e-01
3.01570773e-01 -1.78690162e-02 1.10922486e-01 -4.11989540e-01
3.35718453e-01 -4.27488178e-01 -2.43143201e-01 4.39530969e-01
-4.43134397e-01 6.34355843e-03 4.28137630e-01 2.74080217e-01
-2.88818389e-01 -5.16218126e-01 5.17857969e-01 -3.14094514e-01
-6.41984642e-01 -4.45850462e-01 -5.83583772e-01 2.54283279e-01
7.97034681e-01 6.44926652e-02 -6.08807743e-01 -3.92938703e-01
-6.27915382e-01 2.81386465e-01 2.20180303e-01 6.37965381e-01
2.02683225e-01 -8.94891798e-01 -5.54211676e-01 -1.14634648e-01
1.11744918e-01 -2.26061821e-01 1.01719707e-01 9.25704896e-01
-1.66736320e-01 6.53565824e-01 5.95678203e-03 -1.36200905e-01
-1.22026598e+00 5.09992838e-01 1.25601202e-01 -8.75261068e-01
-5.26771009e-01 6.85870051e-01 1.55491322e-01 -6.56525850e-01
3.27423871e-01 -5.40956259e-01 -4.07500923e-01 6.86776638e-02
4.68503416e-01 1.79680660e-01 1.46524727e-01 -4.23125267e-01
-5.17062604e-01 1.01952806e-01 -1.75666749e-01 -1.26003176e-01
1.14580786e+00 -1.51830064e-02 -1.09046511e-03 5.07730424e-01
8.57914746e-01 -4.51183468e-02 -5.72655797e-01 -1.05956495e-01
1.14565171e-01 7.70300329e-02 -2.36216187e-01 -1.41540980e+00
-4.02642548e-01 7.32753813e-01 -2.32966654e-02 3.69906425e-01
9.22900915e-01 1.84616789e-01 9.89117444e-01 4.32853162e-01
3.99005294e-01 -1.08554447e+00 1.58022121e-01 7.64427543e-01
8.33811700e-01 -9.83892083e-01 -2.61525035e-01 -4.54619467e-01
-9.24297094e-01 1.14731848e+00 6.66326702e-01 1.28497258e-01
1.67190209e-01 3.10022742e-01 7.37315193e-02 -2.47244447e-01
-1.16409147e+00 -1.53265923e-01 1.31303854e-02 4.80559617e-01
5.68421543e-01 -7.36250803e-02 -5.51495373e-01 1.03308785e+00
-2.19493061e-01 7.34676570e-02 2.87334919e-01 8.40820312e-01
-2.48784646e-01 -1.15845311e+00 -3.34882468e-01 1.78585574e-01
-5.92463791e-01 -3.32517982e-01 -6.12135768e-01 1.05137980e+00
2.79903442e-01 1.18494081e+00 -3.43528360e-01 -7.38571137e-02
6.84917688e-01 6.01060987e-01 3.72614145e-01 -8.52856636e-01
-1.16322732e+00 -2.73587793e-01 5.08632362e-01 -6.26353383e-01
-3.73658597e-01 -6.98450267e-01 -1.42484951e+00 8.00181180e-02
-2.50148267e-01 2.92140633e-01 5.10613680e-01 1.31641603e+00
4.23334211e-01 4.13155854e-01 5.49110882e-02 -4.10879612e-01
-8.93031895e-01 -1.45373130e+00 -3.38532507e-01 4.69098270e-01
-1.40939027e-01 -5.25477469e-01 -6.63840592e-01 -1.02951467e-01] | [10.628684043884277, 8.632710456848145] |
a797516a-d147-4025-858d-735c8c57ecef | clrnet-cross-layer-refinement-network-for | 2203.10350 | null | https://arxiv.org/abs/2203.10350v1 | https://arxiv.org/pdf/2203.10350v1.pdf | CLRNet: Cross Layer Refinement Network for Lane Detection | Lane is critical in the vision navigation system of the intelligent vehicle. Naturally, lane is a traffic sign with high-level semantics, whereas it owns the specific local pattern which needs detailed low-level features to localize accurately. Using different feature levels is of great importance for accurate lane detection, but it is still under-explored. In this work, we present Cross Layer Refinement Network (CLRNet) aiming at fully utilizing both high-level and low-level features in lane detection. In particular, it first detects lanes with high-level semantic features then performs refinement based on low-level features. In this way, we can exploit more contextual information to detect lanes while leveraging local detailed lane features to improve localization accuracy. We present ROIGather to gather global context, which further enhances the feature representation of lanes. In addition to our novel network design, we introduce Line IoU loss which regresses the lane line as a whole unit to improve the localization accuracy. Experiments demonstrate that the proposed method greatly outperforms the state-of-the-art lane detection approaches. | ['Xiaofei He', 'Deng Cai', 'Zheng Yang', 'Wenjian Tang', 'Yang Liu', 'Yifei HUANG', 'Tu Zheng'] | 2022-03-19 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Zheng_CLRNet_Cross_Layer_Refinement_Network_for_Lane_Detection_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Zheng_CLRNet_Cross_Layer_Refinement_Network_for_Lane_Detection_CVPR_2022_paper.pdf | cvpr-2022-1 | ['lane-detection'] | ['computer-vision'] | [-4.12913747e-02 -2.22131476e-01 -3.23893100e-01 -7.15265810e-01
-2.67967671e-01 -2.10817471e-01 5.72317719e-01 -1.12907372e-01
-4.89774436e-01 4.90244418e-01 1.81890145e-01 -4.40698802e-01
9.99212917e-03 -9.76534605e-01 -5.42892158e-01 -7.09265590e-01
-3.10189258e-02 -4.08864260e-01 1.06245720e+00 -3.60593319e-01
5.00354409e-01 8.21864843e-01 -1.71045017e+00 1.42697334e-01
9.22124684e-01 1.05411875e+00 3.35268497e-01 3.52503389e-01
-4.88182306e-01 9.06727374e-01 1.05716400e-02 2.06609681e-01
1.73055962e-01 -1.87214270e-01 -1.37395740e-01 8.96753073e-02
7.03060865e-01 -3.34215075e-01 -4.08522904e-01 1.16495502e+00
8.90383795e-02 -5.72801903e-02 4.72476125e-01 -1.31313229e+00
1.13403447e-01 2.27363601e-01 -8.23583663e-01 1.57842159e-01
3.08301020e-02 4.25236672e-01 8.90208602e-01 -7.75497675e-01
5.37491083e-01 1.33588433e+00 6.95599437e-01 1.27918705e-01
-8.12203586e-01 -8.04267704e-01 4.71219450e-01 5.04774392e-01
-1.23722899e+00 -4.73644406e-01 1.29327250e+00 -2.85796911e-01
3.63549262e-01 -1.26273245e-01 3.63639086e-01 6.19910598e-01
2.69646347e-01 9.30166364e-01 1.15169632e+00 -2.22215399e-01
-7.59843215e-02 -3.22789252e-02 5.10190606e-01 1.14653111e+00
4.43827957e-01 1.33132964e-01 -3.86618555e-01 4.45035577e-01
8.88084233e-01 1.36435226e-01 -2.67923981e-01 -8.42554033e-01
-1.30301666e+00 7.16551960e-01 8.01203907e-01 2.55088866e-01
-3.13767731e-01 2.06358314e-01 3.79699856e-01 -1.00572892e-01
-1.06245346e-01 4.71245758e-02 -1.83041200e-01 7.61075690e-02
-8.01235497e-01 -4.30431888e-02 4.83884364e-01 9.71052587e-01
1.46484292e+00 1.48624852e-01 -6.64967746e-02 7.93410242e-01
4.71238226e-01 4.70019817e-01 5.36683500e-02 -9.46981251e-01
5.58231175e-01 7.64889419e-01 -1.07292712e-01 -1.43729579e+00
-8.93180609e-01 -5.78525305e-01 -8.71192753e-01 4.15083319e-01
4.27000970e-01 1.09252436e-02 -9.39822555e-01 1.61707532e+00
1.52944148e-01 6.28041506e-01 6.81312308e-02 1.01340199e+00
6.97513580e-01 6.32977724e-01 9.61275697e-02 6.87481910e-02
1.48313057e+00 -1.27201009e+00 -6.94678247e-01 -5.30850232e-01
8.13283682e-01 -6.63017988e-01 9.34667408e-01 3.85994539e-02
-5.32169402e-01 -8.07163596e-01 -1.31251264e+00 -6.63458407e-02
-7.48400509e-01 2.66781509e-01 6.56148612e-01 5.87988436e-01
-9.05284643e-01 -5.30224554e-02 -2.16321245e-01 -4.22069639e-01
3.54762346e-01 1.33109227e-01 -5.22020698e-01 -4.40622211e-01
-1.35025120e+00 7.52299368e-01 5.03358006e-01 4.64988172e-01
-2.54095018e-01 -4.76082712e-01 -1.32333779e+00 -8.58533382e-02
7.61885405e-01 -3.48467559e-01 7.95267403e-01 -2.77628034e-01
-1.22795618e+00 3.57812613e-01 -3.70554686e-01 -5.13850808e-01
4.63998318e-01 1.28262416e-01 -6.45499468e-01 1.86850712e-01
2.41687402e-01 8.63800526e-01 7.11038828e-01 -1.52045774e+00
-1.40006828e+00 -7.33102709e-02 -1.63799021e-02 -2.77630817e-02
-3.27085331e-02 -3.88766289e-01 -6.22901201e-01 -2.90266067e-01
5.28697431e-01 -6.46149397e-01 -3.95828784e-01 9.29260701e-02
-3.68522853e-01 -2.34163105e-01 1.34583056e+00 -3.32908750e-01
1.28910565e+00 -2.18527913e+00 -7.20239878e-01 6.64567769e-01
4.29811954e-01 2.76241630e-01 -2.14132711e-01 1.18605249e-01
4.01818514e-01 -1.62381217e-01 -8.11837390e-02 -5.44577204e-02
-7.47148274e-03 2.04877406e-01 -6.62648082e-02 4.30665642e-01
2.81045645e-01 8.72700512e-01 -8.64850819e-01 -6.14089727e-01
7.24750280e-01 2.80857712e-01 -5.08757412e-01 -8.34212154e-02
2.17884883e-01 2.70871729e-01 -7.58112609e-01 5.56427658e-01
9.93387580e-01 7.61537999e-02 -1.68754652e-01 -7.16189265e-01
-7.00893223e-01 1.03414673e-02 -1.21323574e+00 1.37250149e+00
-6.09243393e-01 8.28647554e-01 1.20007090e-01 -8.28315258e-01
1.17586565e+00 -4.39930588e-01 1.45107731e-01 -1.07150292e+00
-3.33801620e-02 2.98568726e-01 5.33647984e-02 -3.73943865e-01
5.85517287e-01 4.68016386e-01 -4.50279474e-01 -1.74420685e-01
-4.09597546e-01 3.88259321e-01 1.74604401e-01 5.27355634e-02
1.04132175e+00 -5.49158901e-02 2.50382185e-01 -2.16448843e-01
1.09799457e+00 -6.39778003e-02 7.53405094e-01 9.66849327e-01
-5.18111289e-01 1.94963798e-01 4.57558602e-01 -2.16048986e-01
-8.75308156e-01 -8.42372060e-01 1.57846622e-02 8.60828578e-01
7.78811336e-01 -1.79106981e-01 -5.99596560e-01 -8.45509529e-01
2.57619351e-01 5.19854367e-01 -4.64469135e-01 -5.20612486e-02
-9.13269162e-01 -1.00130580e-01 4.77581352e-01 5.73314548e-01
1.06699538e+00 -8.53736281e-01 -6.31353676e-01 3.34088713e-01
2.59426951e-01 -1.36527991e+00 -5.90697825e-01 8.23858380e-02
-4.70256239e-01 -1.10213792e+00 -3.83443028e-01 -8.99324477e-01
7.08959877e-01 9.42997873e-01 5.26506603e-01 8.81842226e-02
-1.67683929e-01 2.09695175e-02 -2.53050983e-01 -9.44174677e-02
-8.69003609e-02 2.44145423e-01 -2.83714116e-01 2.15117499e-01
4.42369163e-01 -3.36356461e-01 -8.15881193e-01 4.81351256e-01
-4.98414338e-01 3.42123508e-01 1.18998754e+00 8.44963372e-01
3.31850678e-01 1.25079900e-01 6.57760918e-01 -8.81819665e-01
2.97361732e-01 -3.20409685e-01 -7.35330582e-01 -6.48141578e-02
-4.67072755e-01 1.34863153e-01 6.82482481e-01 8.21399391e-02
-1.08267939e+00 1.48913786e-01 -2.65824050e-01 -3.05711448e-01
-5.76476455e-01 3.56183648e-01 -5.23827314e-01 -3.81099164e-01
1.05585970e-01 2.60286331e-01 1.50569648e-01 -4.30600494e-01
5.34230053e-01 6.20521188e-01 7.71664262e-01 -1.15471147e-01
9.88383055e-01 6.10248625e-01 3.96387964e-01 -8.97661626e-01
-7.61106431e-01 -9.47208464e-01 -8.88852894e-01 -5.44781148e-01
7.71764934e-01 -7.14315116e-01 -9.03103948e-01 4.00189012e-01
-9.92425740e-01 -5.91521822e-02 2.29828313e-01 3.42436433e-01
-5.03458798e-01 4.89561290e-01 -2.50371575e-01 -6.62724257e-01
1.46917865e-01 -1.26073575e+00 1.06880021e+00 3.22930306e-01
4.05569345e-01 -9.69947755e-01 -3.81636798e-01 1.53261170e-01
4.43863899e-01 2.14335307e-01 7.18309581e-01 -3.45733792e-01
-1.03090000e+00 -2.23309457e-01 -9.32104707e-01 1.87074870e-01
1.48912907e-01 -3.06492120e-01 -8.15617025e-01 5.36761917e-02
-4.10015464e-01 3.10882002e-01 1.37329543e+00 2.47884586e-01
1.06702280e+00 -1.19475238e-02 -6.02321923e-01 4.59075809e-01
1.39117169e+00 1.93898007e-01 7.30039954e-01 6.84336007e-01
1.18686008e+00 7.64436066e-01 1.01785481e+00 2.71792300e-02
6.15256429e-01 5.85898459e-01 4.73396331e-01 -5.05093575e-01
-4.57769275e-01 -3.09123516e-01 3.01610649e-01 5.85186064e-01
4.80638415e-01 -1.68809127e-02 -8.31133783e-01 4.03244168e-01
-2.17558432e+00 -1.06056166e+00 -4.25823361e-01 1.87808549e+00
2.51646310e-01 6.31055534e-01 -3.45639735e-02 1.53485417e-01
8.35094213e-01 5.79544067e-01 -3.79351854e-01 -2.48498172e-01
-1.83032587e-01 -5.61274588e-01 1.12806165e+00 5.94173372e-01
-1.37230980e+00 1.26702547e+00 5.19304991e+00 9.71435785e-01
-1.20170629e+00 -2.56507993e-01 2.62370914e-01 8.11792195e-01
-1.87538534e-01 -2.75808983e-02 -1.18314469e+00 4.89439934e-01
4.42852527e-01 2.47032985e-01 -1.28454044e-02 9.62960958e-01
7.03769743e-01 -3.67361248e-01 -7.04791248e-01 8.22640061e-01
1.49934798e-01 -1.35471439e+00 -1.59824733e-02 2.06057295e-01
5.60161829e-01 -1.01409972e-01 -1.27223954e-02 3.68060142e-01
-2.20604464e-01 -5.88703573e-01 6.76740348e-01 7.26017237e-01
4.61630046e-01 -9.62248504e-01 9.98323619e-01 4.80001390e-01
-1.74533856e+00 -1.93949029e-01 -2.55582511e-01 2.54555464e-01
3.13226938e-01 6.10062540e-01 -9.41057146e-01 5.73898792e-01
1.83638856e-01 1.02077985e+00 -8.12103808e-01 1.23871601e+00
-1.39645532e-01 5.27735412e-01 -2.97722131e-01 1.02929376e-01
8.00179183e-01 -1.76720023e-01 4.27348286e-01 1.63394737e+00
2.60179907e-01 -2.04430372e-01 6.62847996e-01 5.56313515e-01
2.06050679e-01 5.62321097e-02 -8.20319593e-01 7.93259740e-01
5.68337142e-01 1.32787287e+00 -6.80457890e-01 -2.28608474e-01
-7.65944421e-01 5.36240041e-01 1.63384184e-01 3.55674624e-01
-7.94086933e-01 -7.99887538e-01 8.32168639e-01 2.91613340e-01
3.56992483e-01 -4.81921732e-01 -2.54884809e-01 -7.29480922e-01
-1.44542664e-01 -1.49860948e-01 1.73787437e-02 -6.95632458e-01
-1.02610254e+00 2.48025239e-01 -1.74294397e-01 -1.53119516e+00
-1.98388055e-01 -7.57279932e-01 -6.21668518e-01 8.22812915e-01
-2.43722272e+00 -1.40145075e+00 -7.35148370e-01 4.48016763e-01
6.38131022e-01 -8.97300094e-02 -6.79882690e-02 4.17034626e-01
-5.41449070e-01 5.89545131e-01 -7.15884119e-02 2.66713887e-01
7.55879462e-01 -8.76169562e-01 2.33622149e-01 1.11335230e+00
-4.14122581e-01 5.50835013e-01 5.28317988e-01 -6.07044816e-01
-1.25472140e+00 -1.48806190e+00 7.07556665e-01 3.46053727e-02
6.64226055e-01 -2.38306284e-01 -9.14882243e-01 2.28962123e-01
-1.52759925e-01 1.40503004e-01 1.69979967e-02 -2.67516941e-01
-3.19334626e-01 -5.72926283e-01 -1.04347634e+00 8.95879626e-01
1.24988210e+00 -4.71753210e-01 -3.14795643e-01 -4.08693165e-01
5.36434472e-01 3.09560969e-02 -4.23469037e-01 6.85066283e-01
5.48262656e-01 -1.14618313e+00 9.24251735e-01 1.82666704e-01
-1.18794329e-01 -1.02289331e+00 4.08963934e-02 -9.31595266e-01
-4.92721736e-01 -2.56039083e-01 1.08750217e-01 1.30213833e+00
2.72801310e-01 -8.84349763e-01 9.29193437e-01 -1.03231840e-01
-5.30154943e-01 -6.12417221e-01 -7.21318901e-01 -7.44279027e-01
-4.83093053e-01 -7.87958086e-01 6.51494324e-01 6.45963192e-01
-3.31773728e-01 2.43195564e-01 -4.30794716e-01 4.87279654e-01
8.54660153e-01 3.97661328e-02 1.00660610e+00 -1.27788353e+00
6.49283230e-01 -7.64422715e-01 -9.01026070e-01 -1.52836192e+00
4.64426279e-01 -7.33283877e-01 3.74213845e-01 -1.64806664e+00
-1.24563694e-01 -6.00049794e-01 -4.63837087e-01 3.83875906e-01
-7.77381435e-02 3.62578213e-01 1.94171354e-01 2.64627039e-02
-9.20990109e-01 3.32420230e-01 1.26939094e+00 -1.46043122e-01
-1.27495900e-01 -1.01933882e-01 -4.32363689e-01 8.62441182e-01
1.05134034e+00 1.10589899e-01 -3.44647884e-01 2.22265739e-02
-4.04361695e-01 -2.28555888e-01 5.28418183e-01 -1.28801775e+00
7.22735286e-01 -1.56175196e-01 2.81279683e-01 -1.29190469e+00
1.58321857e-01 -1.02752888e+00 -4.16073292e-01 3.88811886e-01
-2.02500865e-01 -1.81835890e-01 1.07489854e-01 7.68074989e-01
-4.00581598e-01 3.74145024e-02 7.84245312e-01 1.72179013e-01
-1.94866002e+00 1.18426345e-01 -5.09698153e-01 -4.88554120e-01
1.06918383e+00 -7.66513884e-01 -5.59341729e-01 -2.52349675e-01
-1.79372907e-01 6.50630414e-01 4.93356526e-01 5.93408942e-01
7.40942597e-01 -1.40304470e+00 -3.69651616e-01 5.45159638e-01
5.68099856e-01 -1.82219759e-01 3.08447570e-01 9.90744472e-01
-5.60128331e-01 7.15394914e-01 -2.90807307e-01 -8.36433053e-01
-1.09218347e+00 5.26538908e-01 1.76857516e-01 1.98773406e-02
-6.20825589e-01 1.98009670e-01 5.03948033e-01 -1.30889401e-01
1.81079209e-01 -4.64873910e-01 -6.60660982e-01 4.81950231e-02
4.73995537e-01 4.56118256e-01 -1.58168226e-01 -1.04209280e+00
-3.79791766e-01 1.07337320e+00 -8.18032324e-02 2.89857954e-01
7.41295993e-01 -6.40362084e-01 -6.07371628e-02 3.22101533e-01
1.30601525e+00 2.53938019e-01 -1.47950399e+00 -2.09008440e-01
3.11590582e-01 -4.76995438e-01 3.17648739e-01 -5.06181479e-01
-1.00505412e+00 8.53492439e-01 6.41769886e-01 -3.12751792e-02
9.94000852e-01 -2.71578223e-01 9.02318358e-01 4.01964247e-01
6.52821302e-01 -1.02820909e+00 -1.10856093e-01 7.74369776e-01
3.88676137e-01 -1.39504170e+00 -2.64750421e-01 -8.77069294e-01
-3.53396714e-01 1.25456071e+00 8.70704770e-01 -7.29289055e-02
6.05153561e-01 2.91387975e-01 1.97789446e-01 -5.62460050e-02
-2.69858330e-01 -7.97204554e-01 4.56724644e-01 5.67836583e-01
2.46788990e-02 -4.80623879e-02 -2.45364532e-01 2.11489275e-01
1.94585875e-01 -2.95856416e-01 3.32195371e-01 7.90005326e-01
-1.22342169e+00 -9.26951647e-01 -3.85380834e-01 5.11408448e-01
3.27702135e-01 8.40796381e-02 1.47636667e-01 1.01363790e+00
4.26644683e-01 8.40359867e-01 1.23852447e-01 -4.71393287e-01
5.55878341e-01 -1.96785912e-01 -1.32378832e-01 -2.27934092e-01
1.61647886e-01 -9.34710726e-04 2.56110549e-01 -8.35295856e-01
-2.49046326e-01 -5.32244563e-01 -1.40370858e+00 -2.41848961e-01
-1.18960179e-01 -7.06268325e-02 7.16905177e-01 7.84464419e-01
3.95413965e-01 5.55436254e-01 9.27992761e-01 -1.09671748e+00
2.22641286e-02 -5.26874781e-01 -5.11298060e-01 4.91969921e-02
6.71110630e-01 -9.65259135e-01 -2.96511501e-01 -2.87919044e-01] | [8.054889678955078, -1.492987036705017] |
089c6322-b806-42da-bc12-447147b1d8b3 | genie-nf-ai-identifying-neurofibromatosis | 2304.13429 | null | https://arxiv.org/abs/2304.13429v1 | https://arxiv.org/pdf/2304.13429v1.pdf | GENIE-NF-AI: Identifying Neurofibromatosis Tumors using Liquid Neural Network (LTC) trained on AACR GENIE Datasets | In recent years, the field of medicine has been increasingly adopting artificial intelligence (AI) technologies to provide faster and more accurate disease detection, prediction, and assessment. In this study, we propose an interpretable AI approach to diagnose patients with neurofibromatosis using blood tests and pathogenic variables. We evaluated the proposed method using a dataset from the AACR GENIE project and compared its performance with modern approaches. Our proposed approach outperformed existing models with 99.86% accuracy. We also conducted NF1 and interpretable AI tests to validate our approach. Our work provides an explainable approach model using logistic regression and explanatory stimulus as well as a black-box model. The explainable models help to explain the predictions of black-box models while the glass-box models provide information about the best-fit features. Overall, our study presents an interpretable AI approach for diagnosing patients with neurofibromatosis and demonstrates the potential of AI in the medical field. | ['Hamdan Abdellatef', 'Muhammed Nadir Yalçın', 'Amin Jafari', 'Fırat Sefaoğlu', 'Omid Hamza', 'Ali Davar', 'Elnaz Abedini', 'Ferhat Atasoy', 'Michael Bidollahkhani'] | 2023-04-26 | null | null | null | null | ['explainable-models'] | ['computer-vision'] | [ 3.24669361e-01 6.79173589e-01 -4.87204134e-01 -5.29376268e-01
-9.66361612e-02 -1.52396142e-01 1.47991002e-01 1.73835933e-01
3.79836783e-02 1.01357055e+00 2.64875621e-01 -5.99515915e-01
-7.77702034e-01 -5.01733959e-01 -1.40451957e-02 -4.55575824e-01
-1.50305256e-01 1.31905425e+00 -3.13557059e-01 -8.67673159e-02
1.60517722e-01 4.78040874e-01 -1.13824284e+00 6.29926383e-01
1.46523821e+00 8.68774831e-01 -2.18808934e-01 9.37728465e-01
-8.39412659e-02 1.07707572e+00 -5.52461565e-01 -3.83346558e-01
1.48245633e-01 -4.30369556e-01 -8.86887610e-01 -7.90051557e-03
-2.30303273e-01 -2.61933953e-01 -1.38828438e-02 4.33099866e-01
5.65700582e-04 -4.46289927e-01 1.05402660e+00 -1.31271112e+00
-8.07521343e-01 7.81379998e-01 -4.73332852e-01 -4.17716652e-02
3.30513865e-01 3.78040612e-01 8.93980324e-01 -3.84960741e-01
5.69649041e-01 1.36846197e+00 5.16386747e-01 7.53009677e-01
-1.05511200e+00 -4.53278899e-01 9.56031904e-02 4.44914281e-01
-1.08131611e+00 1.34682730e-01 1.32224977e-01 -5.92727363e-01
9.76341426e-01 7.58844376e-01 1.27013493e+00 4.91540730e-01
5.95780611e-01 6.22836471e-01 1.27024865e+00 -5.19096434e-01
2.88362890e-01 2.74689142e-02 4.14744467e-01 1.14566159e+00
5.07153690e-01 3.01605314e-01 -3.75043571e-01 -3.87842447e-01
6.76118553e-01 2.76454240e-01 1.08376041e-01 2.98989296e-01
-9.83438432e-01 7.76126266e-01 2.57357895e-01 3.18298489e-01
-5.74626327e-01 2.57989883e-01 1.11663423e-01 1.04771992e-02
5.07085323e-01 6.57232285e-01 -5.79556346e-01 1.70634970e-01
-8.33073199e-01 2.53944188e-01 7.84904063e-01 5.17841935e-01
-6.38828576e-02 1.03676967e-01 -2.80057967e-01 4.65156943e-01
4.77985680e-01 3.56742293e-01 6.43700421e-01 -6.82589173e-01
-2.01128140e-01 1.14693260e+00 5.32948785e-02 -6.85555637e-01
-8.21301103e-01 -5.96121192e-01 -8.93602908e-01 1.82253420e-01
2.68693388e-01 -6.61210641e-02 -1.51924050e+00 1.13512158e+00
-9.39860865e-02 1.19636953e-01 9.86636430e-02 8.95779014e-01
5.95549881e-01 9.21471566e-02 3.31385374e-01 -2.71315482e-02
1.39350247e+00 -9.35496092e-01 -7.72051394e-01 1.70663357e-01
8.01523805e-01 -3.52246702e-01 5.97111464e-01 9.99664009e-01
-9.79049444e-01 -6.21108599e-02 -7.75830269e-01 4.52282801e-02
-2.38441154e-01 3.13784510e-01 1.31714308e+00 6.08733237e-01
-1.00661218e+00 2.65568912e-01 -9.51784909e-01 -5.04170060e-01
4.18110341e-01 8.07026744e-01 -2.74368167e-01 9.42995921e-02
-8.64708960e-01 1.09370542e+00 5.19219279e-01 -2.00756378e-02
-5.33791006e-01 -7.35162735e-01 -2.22335160e-01 5.35475016e-02
2.69140244e-01 -1.22209120e+00 9.77766275e-01 -7.57501066e-01
-1.14115167e+00 6.44692838e-01 -4.37569499e-01 -1.02105308e+00
3.18359792e-01 -4.07096781e-02 -6.32626772e-01 2.45794684e-01
-5.04948080e-01 7.34648645e-01 3.60982120e-01 -6.72498941e-01
-7.81442225e-01 -4.14088339e-01 -2.40996361e-01 -2.92055786e-01
1.33354023e-01 -2.13757232e-01 4.25573587e-01 -5.55994689e-01
1.99431658e-01 -8.59916270e-01 -5.44547200e-01 2.06165999e-01
-7.05540955e-01 -1.74641728e-01 2.90282369e-01 -8.76126230e-01
1.34577048e+00 -1.37396085e+00 1.80374265e-01 6.38522983e-01
1.12212682e+00 3.29464525e-01 3.36868256e-01 2.76469231e-01
-3.27883601e-01 6.72859728e-01 -1.43297464e-01 2.96689600e-01
-1.63514391e-01 3.97668481e-01 1.29983068e-01 -2.47545112e-02
6.99274898e-01 1.02128708e+00 -6.54961407e-01 -3.47374469e-01
3.41824859e-01 4.28969741e-01 -8.22386265e-01 6.85253646e-03
-4.20553654e-01 4.66687381e-01 -5.92863381e-01 9.58252907e-01
1.88763127e-01 -3.41592968e-01 2.76505888e-01 6.06435873e-02
1.82092294e-01 -2.61441648e-01 -5.48428893e-01 1.10206389e+00
9.85596627e-02 4.22082663e-01 -6.01668358e-01 -8.31566513e-01
1.19394147e+00 4.52206224e-01 5.87057233e-01 -2.68544406e-01
2.28369385e-01 3.46074790e-01 8.23713362e-01 -7.60755837e-01
-1.99401751e-01 -4.02249515e-01 5.10956049e-01 1.28869146e-01
-3.40306520e-01 -1.01142218e-02 1.23266153e-01 2.10230928e-02
1.44971144e+00 -1.87839568e-01 1.06895065e+00 -1.01555482e-01
7.84339786e-01 4.63755250e-01 5.69284260e-01 6.76623821e-01
-2.42327988e-01 6.94240928e-02 5.00531316e-01 -1.10712337e+00
-1.10377491e+00 -6.91858292e-01 -2.93160677e-01 1.38724267e-01
-3.32141310e-01 -3.61366004e-01 -6.07463241e-01 -5.99437714e-01
2.08748415e-01 8.28512371e-01 -8.06578338e-01 -2.29919240e-01
-1.24606062e-02 -8.30583096e-01 5.03293097e-01 6.39084399e-01
1.69102818e-01 -4.92380619e-01 -4.62793142e-01 1.89343244e-01
1.04238391e-01 -4.25663054e-01 3.20146084e-01 1.09594591e-01
-9.69802201e-01 -1.28691947e+00 -3.13848704e-01 -2.18843907e-01
8.61540258e-01 -3.80376637e-01 9.69242632e-01 3.92127901e-01
-9.43261504e-01 -1.27731729e-02 -4.24194813e-01 -9.64083314e-01
-6.23911321e-01 -1.68241426e-01 1.02399066e-01 -3.19159567e-01
8.24316919e-01 -2.03654587e-01 -7.50688672e-01 2.83190429e-01
-7.32556760e-01 6.53131068e-01 1.00047123e+00 9.38027263e-01
7.29031265e-01 1.86449140e-01 5.97661138e-01 -1.13304007e+00
6.13803625e-01 -7.48144507e-01 -8.02270696e-02 3.71179700e-01
-1.18833554e+00 4.31519985e-01 6.07930660e-01 -2.03217953e-01
-8.28348100e-01 3.94823775e-02 1.83634311e-02 -5.31259850e-02
-3.95536631e-01 6.01497531e-01 3.01818609e-01 2.89945249e-02
7.32798219e-01 -9.77196768e-02 3.42340887e-01 -4.56139594e-01
2.02717081e-01 9.61047530e-01 3.83440048e-01 -2.12918252e-01
6.59579396e-01 3.91444474e-01 5.49349129e-01 -1.76111296e-01
-5.57693601e-01 -2.79046118e-01 -4.84502286e-01 -2.15925664e-01
8.27121258e-01 -3.94974589e-01 -8.82465899e-01 -1.34790555e-01
-7.10744441e-01 2.69524306e-01 -2.01737642e-01 5.71732461e-01
-7.03223348e-01 -1.16616033e-01 -3.37003917e-01 -1.24630356e+00
-6.96375251e-01 -1.12943733e+00 8.26284230e-01 7.80885890e-02
-9.20071244e-01 -9.80006576e-01 5.90313226e-02 6.72659516e-01
3.88254285e-01 4.59606588e-01 1.55237293e+00 -1.14633453e+00
-2.78769165e-01 -4.85469848e-01 -2.34189555e-01 -2.53738135e-01
1.58385009e-01 2.63816148e-01 -7.71818221e-01 2.47877136e-01
-2.78800637e-01 1.21009359e-02 7.52713323e-01 5.75188756e-01
1.30099821e+00 -4.83985156e-01 -5.09810388e-01 4.73575681e-01
1.26197970e+00 7.35826433e-01 6.12491667e-01 1.81391835e-01
2.13168785e-01 4.31502610e-01 6.74529254e-01 4.57756966e-01
3.39854747e-01 3.69280338e-01 3.25869441e-01 -5.10688186e-01
-6.93446323e-02 -1.23442203e-01 -3.21722537e-01 2.15580165e-01
-6.45147741e-01 -4.35279012e-01 -1.26859844e+00 3.41337286e-02
-2.00326991e+00 -8.08127284e-01 -2.68527120e-01 1.68841302e+00
5.82050920e-01 5.72693981e-02 3.96315098e-01 5.08158207e-01
3.48284870e-01 -1.04847491e+00 -7.06304848e-01 -9.58964586e-01
-7.63718784e-02 2.73722976e-01 2.43563473e-01 4.86820132e-01
-5.45505762e-01 6.88244820e-01 7.86927748e+00 5.64428270e-01
-9.30108130e-01 -1.97579682e-01 1.03164685e+00 -1.26953289e-01
-4.27548587e-01 -1.31650358e-01 -3.94316047e-01 2.46786684e-01
1.15568757e+00 -2.13066503e-01 4.68768179e-01 6.55247748e-01
6.36239946e-01 -1.61545515e-01 -1.18972421e+00 5.31291842e-01
-1.49676308e-01 -1.35426927e+00 4.52635616e-01 4.01976675e-01
3.89101326e-01 -4.38242555e-01 -9.56068486e-02 -8.04283842e-02
4.85215038e-01 -1.58051312e+00 1.22669257e-01 1.23893189e+00
6.28766656e-01 -5.80740750e-01 1.13337028e+00 4.81081568e-02
-6.00244343e-01 -4.68625784e-01 -2.20409632e-01 -3.81015599e-01
-2.67319739e-01 6.66420519e-01 -1.79779649e+00 5.01164079e-01
1.07696950e-01 4.90381122e-01 -4.39258873e-01 1.16930616e+00
-8.61349851e-02 8.01586390e-01 -8.81286561e-02 -2.10696682e-01
3.43273431e-02 -1.91222280e-01 4.58210766e-01 8.66246939e-01
1.75548062e-01 4.02125776e-01 -1.98464170e-01 1.03923976e+00
6.43992364e-01 3.95282894e-01 -2.61910826e-01 -4.72383887e-01
1.18175440e-01 9.74229097e-01 -5.79452455e-01 -5.06754518e-01
-2.58082598e-01 4.46553677e-01 4.52638641e-02 9.38541561e-05
-5.76765716e-01 5.07480875e-02 4.00026709e-01 2.93486506e-01
-3.05143565e-01 2.26125240e-01 -8.18549573e-01 -9.36906397e-01
-4.50296104e-01 -1.04733956e+00 6.85617864e-01 -9.72157717e-01
-1.13278329e+00 7.10478604e-01 -1.39713541e-01 -1.00772250e+00
-6.07250392e-01 -9.52212393e-01 -4.08886462e-01 8.16669464e-01
-1.08602679e+00 -1.40291047e+00 -2.59983271e-01 -1.53107720e-03
3.52101892e-01 -6.03533268e-01 1.14859629e+00 -3.25433582e-01
-6.96665287e-01 3.38872164e-01 -2.42714509e-02 -3.41864079e-01
3.20423096e-01 -1.41831696e+00 9.63974968e-02 3.57893884e-01
-3.25522900e-01 7.09463120e-01 8.36610496e-01 -9.17926788e-01
-1.04973745e+00 -7.32045054e-01 7.63833821e-01 -3.19427282e-01
6.70768142e-01 3.20098907e-01 -6.55904889e-01 4.66027409e-01
-1.97876319e-01 -5.05266309e-01 1.39794290e+00 1.79511666e-01
-5.61521053e-02 -1.62277068e-03 -1.68548870e+00 5.75110495e-01
8.25925231e-01 1.13820978e-01 -4.80285376e-01 5.13069153e-01
4.59169567e-01 -2.51044333e-02 -1.14644158e+00 4.38443840e-01
8.90920162e-01 -7.47774601e-01 6.76225185e-01 -1.43376482e+00
6.30460739e-01 -2.92731613e-01 1.84975341e-01 -1.17750132e+00
-6.08465254e-01 -3.86169553e-01 -3.88031721e-01 4.58063722e-01
7.23978281e-01 -6.36552334e-01 7.70586252e-01 1.27228773e+00
6.02478795e-02 -1.41431427e+00 -4.32101160e-01 -3.59149277e-01
-1.39983714e-01 -2.99095601e-01 8.26960683e-01 5.37470996e-01
3.84328961e-01 -2.14042708e-01 -3.14553410e-01 2.26228889e-02
4.41635609e-01 -2.64854699e-01 2.87342340e-01 -1.57238495e+00
-6.20928168e-01 -4.24488634e-01 -1.15830636e+00 -6.25327602e-03
-2.05497116e-01 -9.96552646e-01 -6.92014575e-01 -1.58949268e+00
6.29254699e-01 -3.93346280e-01 -3.94503146e-01 8.56039166e-01
-1.40947312e-01 7.40956292e-02 -9.78657696e-03 2.76356012e-01
1.21668346e-01 -2.97716200e-01 1.18411100e+00 -3.07450294e-01
-1.49487585e-01 -1.26428381e-02 -1.11135602e+00 9.12839592e-01
1.07135451e+00 -2.53532797e-01 -5.41194141e-01 -9.79013070e-02
2.10639998e-01 1.48192674e-01 4.15028661e-01 -8.81346047e-01
-7.75641622e-03 -5.67480087e-01 8.47662985e-01 -4.05698776e-01
2.37324804e-01 -7.97029316e-01 5.17640650e-01 1.13955772e+00
-4.90943819e-01 -2.81390011e-01 4.70884517e-02 4.97567773e-01
1.01050109e-01 -1.21488959e-01 5.03964424e-01 -4.76265736e-02
-2.96405584e-01 6.51091710e-02 -6.42162681e-01 -5.12368500e-01
1.26939011e+00 -4.04688388e-01 -6.11299098e-01 -3.97260278e-01
-1.27606666e+00 1.52557209e-01 2.01455250e-01 -1.82234570e-02
7.95454502e-01 -7.12070107e-01 -7.94601619e-01 1.90621093e-01
1.04387686e-01 -4.13867921e-01 -8.62517655e-02 8.91266584e-01
-1.02818370e+00 1.11049056e+00 -4.56930280e-01 -4.25531566e-01
-1.40641892e+00 4.37611759e-01 5.09368718e-01 -2.62052625e-01
-3.74657631e-01 4.87541080e-01 -4.06881534e-02 1.54724764e-02
9.90509540e-02 -7.31619537e-01 -3.00133705e-01 -6.83791220e-01
6.94094718e-01 5.16307712e-01 -1.89263850e-01 -6.98614791e-02
-2.90764600e-01 3.98346573e-01 -3.43227714e-01 8.16022381e-02
1.29257882e+00 1.57900423e-01 -1.52312219e-01 2.26680890e-01
1.22377761e-01 -3.61888021e-01 -2.82442212e-01 2.30802670e-01
1.33907378e-01 -2.36808181e-01 1.97762430e-01 -1.83765125e+00
-7.52508283e-01 6.79913163e-01 7.25016356e-01 1.54538155e-01
1.25921857e+00 -1.32833794e-01 4.15062964e-01 5.30676425e-01
2.49767706e-01 -6.59451246e-01 -5.69420159e-01 -2.23987445e-01
8.21814716e-01 -1.10867119e+00 2.25662783e-01 -7.11121559e-01
-6.74921036e-01 1.29798949e+00 5.97603023e-01 1.71186596e-01
3.31140280e-01 2.53296912e-01 8.60374942e-02 -2.59692580e-01
-1.33651435e+00 -1.06212817e-01 5.86105049e-01 7.83759177e-01
5.97893775e-01 6.30797982e-01 -8.82685602e-01 1.03854871e+00
-3.63249302e-01 6.12279415e-01 4.10376191e-01 5.24462819e-01
-5.63685536e-01 -1.35295475e+00 -3.11727405e-01 1.22530150e+00
-1.97906688e-01 -1.16981089e-01 -1.26375079e+00 8.46520066e-01
3.05138975e-01 9.22363460e-01 -6.83861524e-02 -6.58221245e-01
-1.84756890e-01 2.97003984e-01 2.78581619e-01 -4.20004189e-01
-5.76264143e-01 -4.66325581e-02 4.24793929e-01 -4.17483479e-01
-2.35695198e-01 -2.89784342e-01 -1.75996864e+00 -4.12746519e-01
-2.92924464e-01 7.92268813e-02 6.43527985e-01 1.17868483e+00
5.50015092e-01 7.07694173e-01 2.05553621e-01 2.54640937e-01
-1.85014099e-01 -8.18291724e-01 -5.90452850e-01 1.64719358e-01
1.37691006e-01 -7.04069674e-01 -1.65397078e-01 4.56068702e-02] | [8.302999496459961, 5.863886833190918] |
9130acca-9e0d-4355-b4d0-2e1562a5a4d0 | confidence-based-ensembles-of-end-to-end | 2306.15824 | null | https://arxiv.org/abs/2306.15824v1 | https://arxiv.org/pdf/2306.15824v1.pdf | Confidence-based Ensembles of End-to-End Speech Recognition Models | The number of end-to-end speech recognition models grows every year. These models are often adapted to new domains or languages resulting in a proliferation of expert systems that achieve great results on target data, while generally showing inferior performance outside of their domain of expertise. We explore combination of such experts via confidence-based ensembles: ensembles of models where only the output of the most-confident model is used. We assume that models' target data is not available except for a small validation set. We demonstrate effectiveness of our approach with two applications. First, we show that a confidence-based ensemble of 5 monolingual models outperforms a system where model selection is performed via a dedicated language identification block. Second, we demonstrate that it is possible to combine base and adapted models to achieve strong results on both original and target data. We validate all our results on multiple datasets and model architectures. | ['Boris Ginsburg', 'Aleksandr Laptev', 'Vitaly Lavrukhin', 'Igor Gitman'] | 2023-06-27 | null | null | null | null | ['language-identification', 'model-selection', 'language-identification', 'speech-recognition'] | ['audio', 'methodology', 'natural-language-processing', 'speech'] | [ 4.19703871e-02 1.78765118e-01 1.06696654e-02 -6.51230633e-01
-1.28090286e+00 -8.30540419e-01 7.09354341e-01 -2.29451135e-01
-6.38264298e-01 7.40804493e-01 1.15569837e-01 -5.69098473e-01
1.85344040e-01 -1.30235806e-01 -4.73671228e-01 -1.86555386e-01
7.23797157e-02 8.67055893e-01 4.27434474e-01 -2.62391269e-01
-1.22815199e-01 3.51476252e-01 -1.34816408e+00 4.79227185e-01
1.17291272e+00 6.95848584e-01 2.45802268e-01 8.43086600e-01
-7.54825771e-02 3.70219409e-01 -7.68967032e-01 -5.74828923e-01
1.94955960e-01 -1.74209073e-01 -5.89931071e-01 -5.54590859e-02
4.56980437e-01 1.66972339e-01 -4.48615626e-02 8.73913050e-01
7.84899473e-01 2.21518297e-02 5.70799470e-01 -1.01819193e+00
-5.65506876e-01 8.75906348e-01 6.85114786e-02 1.05335593e-01
3.12458009e-01 1.06706902e-01 7.45618999e-01 -1.15802383e+00
4.81656253e-01 1.21395135e+00 7.63827562e-01 7.41721511e-01
-1.32195652e+00 -6.37881041e-01 3.51671785e-01 -2.74479743e-02
-1.44928455e+00 -1.14675438e+00 4.27465618e-01 -3.20380300e-01
1.42791259e+00 2.16825709e-01 -3.88623364e-02 1.45463336e+00
-3.48691940e-01 7.50520110e-01 1.20749903e+00 -7.57216096e-01
3.02550107e-01 6.78113461e-01 8.77415463e-02 5.11923790e-01
-1.69789586e-02 3.15774977e-01 -7.33203530e-01 -2.06434578e-01
8.82206038e-02 -7.53720820e-01 -1.75340936e-01 -6.70644194e-02
-1.18283892e+00 6.64280653e-01 -2.08160773e-01 5.46840906e-01
-2.34125391e-01 -3.85468304e-01 1.51398554e-01 5.37700832e-01
5.32352388e-01 4.86368954e-01 -9.08360124e-01 -2.02861398e-01
-1.27792442e+00 -1.14391282e-01 1.13837063e+00 1.00892591e+00
4.38040853e-01 2.02592894e-01 -9.81191844e-02 1.40375912e+00
2.85020918e-01 6.33385599e-01 7.79186845e-01 -7.38205850e-01
5.38316369e-01 3.81413251e-01 7.15122446e-02 -2.37567484e-01
-1.11511469e-01 -7.79631019e-01 -4.66185570e-01 2.98280120e-01
3.96066576e-01 -3.01053345e-01 -1.11699951e+00 1.88904333e+00
-1.97849441e-02 1.26801059e-01 4.07045156e-01 4.29634959e-01
4.17338669e-01 4.30952996e-01 1.98129520e-01 -2.59345084e-01
9.53355253e-01 -1.03741598e+00 -3.46912980e-01 -5.67033052e-01
7.47995317e-01 -7.53192365e-01 8.67450714e-01 6.96105003e-01
-9.85790312e-01 -6.37843668e-01 -1.02033949e+00 3.69114518e-01
-4.19677138e-01 4.98533756e-01 1.97931439e-01 1.00672722e+00
-1.32532954e+00 4.32915002e-01 -5.37207365e-01 -5.22425771e-01
3.19599845e-02 4.79216486e-01 -5.72655857e-01 1.64065361e-01
-1.21497524e+00 1.37593424e+00 5.53820550e-01 -1.56255946e-01
-8.64221394e-01 -2.25219592e-01 -5.02549946e-01 -1.57710403e-01
1.85465798e-01 -6.32205367e-01 1.56759119e+00 -1.30362523e+00
-1.71936500e+00 8.86198997e-01 -1.89457998e-01 -6.79949343e-01
6.13755763e-01 -9.69527438e-02 -9.72961068e-01 -4.14800614e-01
-1.09366648e-01 4.14268315e-01 8.56283188e-01 -1.31152391e+00
-9.87901688e-01 -2.21075237e-01 -2.88491428e-01 1.85402095e-01
-4.24407303e-01 2.90044218e-01 -4.08013821e-01 -4.76168722e-01
-1.01273796e-02 -9.46268976e-01 -1.77312315e-01 -5.45649827e-01
-3.61130327e-01 -1.49697319e-01 6.38237715e-01 -9.36092019e-01
1.58602703e+00 -1.92145956e+00 2.32802257e-01 2.24107370e-01
-2.30783314e-01 7.00807571e-01 -2.10916609e-01 4.10771489e-01
-9.34971869e-02 4.67287064e-01 -3.80481064e-01 -6.21345699e-01
1.95431001e-02 2.42950186e-01 -2.02105001e-01 -2.68655494e-02
1.99259192e-01 4.06880200e-01 -6.62344575e-01 -4.25497979e-01
5.14255874e-02 4.19756770e-01 -4.04263586e-01 1.66006237e-01
-1.69665813e-01 3.72849375e-01 -2.53149718e-01 5.98903596e-01
4.11283582e-01 -5.98085262e-02 4.13345069e-01 4.29783240e-02
-6.57842457e-02 5.29694319e-01 -1.26130593e+00 1.63124287e+00
-7.32000947e-01 3.60644728e-01 2.23066628e-01 -7.70352006e-01
9.34457064e-01 5.74841738e-01 -6.74241632e-02 -2.60285705e-01
-2.32491493e-01 7.33652890e-01 3.76642585e-01 -1.22802213e-01
2.23563895e-01 -1.54060766e-01 -9.89362225e-02 3.44971567e-01
4.03871804e-01 -1.33036241e-01 7.96485692e-02 3.47108096e-02
1.01777601e+00 1.48586892e-02 3.23270887e-01 -2.08956435e-01
7.38517761e-01 -9.27182138e-02 4.81235236e-01 1.00678027e+00
-3.62073094e-01 5.07306039e-01 -5.82209341e-02 -1.35485502e-02
-1.11383140e+00 -1.08476925e+00 -1.28035516e-01 1.29122031e+00
-4.84571010e-01 -2.79646277e-01 -8.07949066e-01 -9.48360264e-01
2.78268699e-02 1.10263813e+00 -3.26335788e-01 4.77468688e-03
-4.89594638e-01 -5.59123456e-01 9.68255818e-01 4.73413467e-01
3.23115677e-01 -9.04777884e-01 -7.22835809e-02 3.25117230e-01
-6.41108304e-02 -1.05650294e+00 -3.69665742e-01 3.99898738e-01
-7.92822778e-01 -6.04003489e-01 -5.13724744e-01 -7.20293224e-01
2.45286077e-01 -1.74383342e-01 1.36976957e+00 -8.52571651e-02
5.93475282e-01 4.63622957e-01 -4.39804524e-01 -4.45572197e-01
-1.14142942e+00 5.88882029e-01 4.70572472e-01 -4.82675768e-02
3.66678923e-01 -4.43794698e-01 3.10328193e-02 4.76657480e-01
-4.04486209e-01 -9.29492787e-02 7.20420301e-01 1.06504834e+00
1.86774984e-01 -3.62292558e-01 9.42307055e-01 -7.42992759e-01
7.92953849e-01 -4.71422791e-01 -3.53226215e-01 7.73662984e-01
-7.45122135e-01 3.15216452e-01 3.93923461e-01 -5.42462826e-01
-1.13118124e+00 1.64180458e-01 -2.46496677e-01 -3.04538965e-01
-3.30530614e-01 6.45098925e-01 -1.72627985e-01 1.30844310e-01
1.01554954e+00 1.62831903e-01 2.84312479e-02 -7.75886178e-01
2.08388627e-01 1.34300780e+00 4.39760655e-01 -6.52394831e-01
5.08258402e-01 -2.02480227e-01 -7.56954551e-01 -7.95246840e-01
-4.17903900e-01 -2.74208665e-01 -6.58638537e-01 -1.52284160e-01
3.77563953e-01 -9.82261062e-01 -1.68412685e-01 4.66359496e-01
-1.26049721e+00 -2.95359015e-01 -4.86845849e-03 6.47629261e-01
-3.89065474e-01 2.45869413e-01 -1.55598432e-01 -1.23266864e+00
-2.87356108e-01 -1.30681598e+00 9.53044713e-01 -5.87283261e-02
-4.59513366e-01 -1.10536718e+00 2.26915032e-01 3.14312965e-01
4.82734472e-01 -5.09964228e-01 6.75271988e-01 -1.51666188e+00
-1.06862716e-01 -4.05162632e-01 3.47482979e-01 7.82877862e-01
7.90966451e-02 1.80606067e-01 -1.48182118e+00 -4.07741249e-01
-2.67136872e-01 -2.82168269e-01 9.03354645e-01 5.43015487e-02
6.56977057e-01 1.28950588e-02 -4.73664016e-01 1.42252669e-01
9.26996946e-01 4.00072187e-01 2.16688573e-01 9.78189781e-02
3.67655307e-01 6.94063187e-01 2.86037296e-01 -1.89781487e-01
2.59844005e-01 1.00843883e+00 -1.03059478e-01 2.40643203e-01
-2.95622081e-01 -2.84819394e-01 6.76400244e-01 1.28250313e+00
1.21501870e-02 -4.48480099e-01 -1.30538452e+00 6.60464227e-01
-1.68590164e+00 -9.17409003e-01 3.03073525e-01 2.58456016e+00
8.92726362e-01 2.82600701e-01 2.07690969e-01 -1.42696917e-01
7.93213308e-01 -2.74161130e-01 -6.12472951e-01 -3.42329681e-01
-3.58726531e-01 3.26272547e-01 3.00343633e-01 9.02554095e-01
-9.74715292e-01 1.15168047e+00 7.61696148e+00 8.94717991e-01
-9.54566240e-01 3.59437346e-01 6.69881880e-01 -1.92628726e-02
-4.27949071e-01 1.32608846e-01 -1.01119757e+00 4.57719713e-01
1.46923518e+00 -2.46279955e-01 3.75707895e-01 9.03594732e-01
-6.53116927e-02 3.46968621e-02 -1.17798746e+00 6.96749449e-01
6.33260682e-02 -8.99468362e-01 -1.34354547e-01 -6.51183426e-02
5.34740329e-01 3.72824848e-01 5.02964146e-02 7.54567564e-01
7.39054322e-01 -1.06326747e+00 7.86809325e-01 4.12790179e-01
8.51426840e-01 -4.80137825e-01 5.36558330e-01 7.51311660e-01
-6.96843982e-01 -1.25594854e-01 2.97711361e-02 3.51143628e-01
1.32535055e-01 2.01821700e-01 -1.37395191e+00 5.54234743e-01
4.31376427e-01 1.99592933e-01 -6.40580833e-01 8.80824685e-01
-4.35935296e-02 8.77090871e-01 -6.25763118e-01 -1.62787158e-02
6.59822151e-02 9.00491178e-02 6.33484244e-01 1.41024649e+00
5.66127837e-01 -3.08868140e-01 9.44510251e-02 3.96096230e-01
1.52370438e-01 2.57554054e-01 -8.08954835e-01 -6.39318228e-02
8.43445241e-01 8.97713482e-01 -1.59942269e-01 -4.36476529e-01
-5.55871725e-01 9.03869033e-01 6.99711800e-01 4.28091496e-01
-4.34828162e-01 7.53985122e-02 6.48037314e-01 -1.79831654e-01
2.44742736e-01 -2.38998592e-01 -1.81349978e-01 -1.25952041e+00
-1.68490306e-01 -1.51612520e+00 3.76725495e-01 -6.85605943e-01
-1.37054086e+00 1.11729372e+00 -1.53574506e-02 -1.05470312e+00
-8.61704230e-01 -7.69314349e-01 -3.02714318e-01 1.37798011e+00
-9.52009022e-01 -1.13609171e+00 3.44647795e-01 2.46178478e-01
7.18382657e-01 -9.57861185e-01 1.19121444e+00 2.89901912e-01
-5.18385172e-01 8.98036540e-01 2.43962765e-01 5.57305962e-02
9.54383850e-01 -1.01944864e+00 5.71867406e-01 9.31790590e-01
6.33237779e-01 6.14420533e-01 7.75089443e-01 -6.58756435e-01
-8.23426008e-01 -6.60858572e-01 1.18304741e+00 -1.03214157e+00
4.70739394e-01 -2.69578815e-01 -9.54528987e-01 7.92455673e-01
2.69537389e-01 -3.42690259e-01 7.20727980e-01 4.94099021e-01
-4.95193660e-01 -1.22071914e-01 -1.12563479e+00 5.42169333e-01
1.03588426e+00 -7.88812220e-01 -6.09139025e-01 1.54041827e-01
5.01592100e-01 -1.80305332e-01 -7.55445480e-01 4.65868473e-01
6.10944033e-01 -1.11908102e+00 6.78256452e-01 -9.51525450e-01
-3.16385835e-01 -2.42414609e-01 -5.60674727e-01 -1.81943059e+00
-9.36742797e-02 -4.89507198e-01 -7.19062015e-02 1.15364194e+00
1.25460470e+00 -7.86601484e-01 3.79993051e-01 7.51202166e-01
-3.48528385e-01 -4.51695234e-01 -1.17706180e+00 -9.10694122e-01
3.05559665e-01 -8.15797031e-01 5.50336480e-01 9.91318464e-01
-7.89916590e-02 4.75113779e-01 -2.28263676e-01 3.56140137e-01
3.18560958e-01 -3.71501148e-01 7.49046385e-01 -1.17888081e+00
-6.49940908e-01 -5.65453470e-01 -2.52317786e-01 -7.39533842e-01
3.75578880e-01 -9.57391024e-01 5.12619652e-02 -1.20607555e+00
-1.55694231e-01 -5.89871228e-01 -5.04749238e-01 5.07434189e-01
-1.65431678e-01 -9.87278670e-02 1.52862713e-01 2.29126677e-01
-3.37829471e-01 1.77025422e-01 5.45361459e-01 -6.67151958e-02
-3.19462687e-01 1.76508665e-01 -6.44048929e-01 6.98481679e-01
8.47569704e-01 -3.75598401e-01 -4.80482936e-01 -5.27440071e-01
-1.48631975e-01 8.51707009e-04 -3.59156840e-02 -1.24401402e+00
2.95762271e-01 1.78765640e-01 1.31940991e-01 -2.48939082e-01
4.56603914e-01 -6.18884444e-01 4.17072654e-01 1.40740633e-01
-2.81590074e-01 5.89096881e-02 3.18400681e-01 3.97547215e-01
-2.95645982e-01 -3.26662868e-01 7.55071819e-01 -9.54042748e-02
-7.14079738e-01 -4.91071716e-02 -2.04443410e-01 -4.08785492e-02
7.52806246e-01 -2.47398838e-01 -1.20746806e-01 -5.59925497e-01
-1.18354058e+00 2.51473654e-02 3.99836183e-01 6.31758571e-01
2.39050880e-01 -1.16912436e+00 -1.14101708e+00 3.79638553e-01
2.38887116e-01 -5.59888780e-01 -9.61199105e-02 6.85736597e-01
-8.17487463e-02 4.73830402e-01 2.09548697e-01 -6.66049957e-01
-1.33958137e+00 2.96415478e-01 7.60892272e-01 -3.36537361e-01
6.67479858e-02 9.40814197e-01 -1.74420223e-01 -1.02117395e+00
2.41012916e-01 -9.91579220e-02 -6.26083687e-02 -3.12025920e-02
2.89320111e-01 1.47301823e-01 3.70289981e-01 -8.92613947e-01
-5.71725905e-01 2.11732194e-01 -3.21397990e-01 -7.51968145e-01
9.65229750e-01 -1.59649272e-02 4.72311020e-01 6.50031805e-01
7.16093659e-01 2.34448731e-01 -1.14944744e+00 -3.02927375e-01
2.54624426e-01 -7.91534707e-02 3.45975645e-02 -1.38091159e+00
-6.42286837e-01 7.88165867e-01 7.78834283e-01 1.78600326e-01
9.58717942e-01 -1.60533294e-01 1.74821630e-01 4.51423347e-01
6.44195378e-01 -1.24368155e+00 -4.38966930e-01 6.90160036e-01
9.06703234e-01 -1.45379150e+00 -2.93639392e-01 -1.49026349e-01
-9.15619135e-01 7.62602389e-01 5.23208082e-01 3.85891408e-01
7.17928648e-01 3.96305203e-01 4.29791808e-01 4.44512725e-01
-1.18160975e+00 -3.09459001e-01 4.55113679e-01 8.65034282e-01
6.96335316e-01 2.00519890e-01 -7.37912357e-02 8.47106159e-01
-2.40585789e-01 -9.85907540e-02 7.99527466e-02 5.91176212e-01
-4.29920077e-01 -1.64826238e+00 -3.19405824e-01 3.50129098e-01
-2.79472649e-01 -4.55730438e-01 -6.46541417e-01 8.02256167e-01
8.86070803e-02 1.27824581e+00 -2.29159594e-01 -6.19470477e-01
5.04458308e-01 8.98432791e-01 3.57956231e-01 -8.23870897e-01
-7.73069739e-01 8.87196139e-03 7.48319447e-01 -8.40722844e-02
-2.14621633e-01 -8.08201969e-01 -7.18970656e-01 -1.73059921e-03
-4.65589017e-01 3.43935460e-01 8.46354306e-01 9.28391397e-01
5.97343206e-01 1.24192625e-01 4.06215101e-01 -4.39483374e-01
-1.07326746e+00 -1.46366704e+00 -3.53838891e-01 1.63092420e-01
1.85550705e-01 -6.09325171e-01 -4.66112524e-01 1.57994375e-01] | [14.403002738952637, 6.668713569641113] |
e9caf8d9-0fd8-4186-8806-b69485b38694 | supervised-relation-classification-as-two-way | null | null | https://openreview.net/forum?id=Drjb0jGXtGe | https://openreview.net/pdf?id=Drjb0jGXtGe | Supervised Relation Classification as Two-way Span-Prediction | Most of the current supervised relation classification (RC) algorithms use a single embedding to represent the relation between a pair of entities. We argue that a better approach is to treat the RC task as a Span-Prediction (SP) problem, similar to Question Answering (QA). We present an SP-based system for RC and evaluate its performance compared to the embedding-based system. We demonstrate that by adding a few improvements, the supervised SP objective works significantly better than the standard classification-based objective. We achieve state-of-the-art results on the TACRED, SemEval task 8, and the CRE datasets. | ['Anonymous'] | 2021-09-17 | null | null | null | acl-arr-september-2021-9 | ['relation-classification'] | ['natural-language-processing'] | [ 7.09700659e-02 7.24071085e-01 -3.90966266e-01 -4.21703905e-01
-8.85123491e-01 -3.64432037e-01 7.42071509e-01 7.81856775e-01
-2.45393574e-01 6.11438274e-01 4.51052815e-01 -5.58240354e-01
-4.22759652e-01 -9.18179035e-01 -4.22683388e-01 -9.97261629e-02
-2.98590779e-01 9.90759611e-01 6.82821989e-01 -5.39861143e-01
-2.54229046e-02 1.68384820e-01 -1.39916992e+00 6.58354580e-01
9.71065164e-01 1.23322153e+00 -5.22803664e-01 6.40255094e-01
-2.99532056e-01 1.24892151e+00 -4.83082294e-01 -1.16234922e+00
-2.92340040e-01 -2.37497345e-01 -1.89503062e+00 -5.20385981e-01
6.67766750e-01 2.96401203e-01 -5.43837130e-01 5.15203655e-01
3.83482069e-01 1.17175102e-01 7.29827166e-01 -1.50768876e+00
-9.48033392e-01 7.11097717e-01 4.39237282e-02 2.27451667e-01
7.87833452e-01 -7.71117330e-01 1.93521035e+00 -7.97194421e-01
9.39717233e-01 1.12786913e+00 7.91766882e-01 1.19840160e-01
-1.44760847e+00 8.61647725e-02 -3.20488691e-01 1.07782173e+00
-1.30499637e+00 -3.94907713e-01 6.37059569e-01 -1.60791859e-01
1.67513168e+00 5.29572368e-01 1.94671333e-01 7.28349686e-01
1.48247153e-01 9.13132608e-01 6.18067563e-01 -8.58756661e-01
-8.39815140e-02 1.36938959e-01 8.72283816e-01 7.33554244e-01
2.06019729e-01 -1.94861591e-01 -4.12951589e-01 -6.39982939e-01
-2.55277693e-01 -5.58158934e-01 -5.52611291e-01 -4.63005275e-01
-1.04182744e+00 9.42654014e-01 5.25839508e-01 3.64375055e-01
-2.07242012e-01 1.44211976e-02 5.96159995e-01 6.47865593e-01
8.14411283e-01 1.01921296e+00 -8.80784869e-01 2.12129019e-03
-4.94067371e-01 4.14760023e-01 1.50928020e+00 8.32108855e-01
4.92038369e-01 -6.08602524e-01 -4.95782524e-01 9.48057294e-01
4.61580269e-02 -2.57574171e-01 1.62963882e-01 -1.02564061e+00
6.01064324e-01 8.78933191e-01 -5.35261445e-03 -1.03196478e+00
-4.29417312e-01 -3.82717550e-01 -4.09649611e-01 -3.68249089e-01
2.35352874e-01 1.40690982e-01 -3.80963922e-01 1.29027414e+00
3.88190985e-01 3.56160961e-02 4.98587251e-01 3.79452258e-01
1.21918976e+00 6.05642796e-01 1.20177642e-01 -7.21133947e-02
1.61473393e+00 -1.26524377e+00 -9.98881817e-01 -8.32605660e-02
1.19249928e+00 -5.21921396e-01 7.14265347e-01 1.27479639e-02
-7.23311305e-01 -3.13872069e-01 -1.10979521e+00 -6.39054239e-01
-8.71067882e-01 1.50398508e-01 9.33040082e-01 4.36570168e-01
-6.84581637e-01 7.56179988e-01 -3.45593035e-01 -5.30280411e-01
1.78434700e-01 6.22419454e-02 -7.87593603e-01 1.98267102e-02
-1.70251632e+00 1.55204427e+00 6.62949741e-01 -1.68329984e-01
-5.30105922e-03 -8.02435517e-01 -1.13276398e+00 2.81683028e-01
5.14391661e-01 -6.30892634e-01 1.02566171e+00 -5.13224937e-02
-9.79461312e-01 1.03289700e+00 -4.23308492e-01 -9.11128879e-01
-6.60413504e-02 -3.47446382e-01 -7.56826997e-01 3.21525276e-01
-1.86038554e-01 3.18667561e-01 2.83262610e-01 -1.27179658e+00
-4.36911166e-01 -2.49903291e-01 3.55036974e-01 -7.40580633e-02
-3.55265737e-01 2.52054870e-01 -4.02541533e-02 -2.31815487e-01
2.10604630e-02 -8.11044097e-01 1.72315121e-01 -8.32305327e-02
-5.08162737e-01 -1.21149647e+00 8.87846470e-01 -8.37125421e-01
1.57511687e+00 -1.70891452e+00 2.90995359e-01 -2.49976680e-01
3.72123331e-01 5.62307596e-01 -1.15062907e-01 9.81476665e-01
-5.91256380e-01 1.12738207e-01 -1.45468578e-01 -4.72139269e-01
1.61644280e-01 4.58505213e-01 -5.21157444e-01 1.89095780e-01
4.50759232e-01 1.29352176e+00 -9.69065905e-01 -8.08911443e-01
-2.86086947e-01 7.93707520e-02 -2.82544017e-01 3.28649282e-01
-3.63881767e-01 -3.77005070e-01 -1.47146687e-01 5.80074608e-01
2.59784341e-01 -3.53460431e-01 5.29286325e-01 -5.18890381e-01
4.21282619e-01 8.88406038e-01 -7.97470272e-01 1.38047457e+00
-4.90954220e-01 9.67781544e-01 -5.53755641e-01 -1.20065546e+00
9.08773363e-01 6.88842356e-01 3.53800237e-01 -3.30250323e-01
-2.67766982e-01 1.17367625e-01 1.69550702e-01 -8.16013098e-01
7.57263184e-01 1.71193108e-01 2.02690288e-02 3.91283512e-01
5.07232785e-01 5.58478869e-02 2.61651903e-01 3.54862899e-01
1.50639200e+00 1.23018593e-01 4.60035890e-01 -2.31009871e-01
6.20880127e-01 3.10116768e-01 2.65858531e-01 2.60624081e-01
-2.50676479e-02 3.16758096e-01 9.75716591e-01 -4.21875805e-01
-8.64990234e-01 -9.17460740e-01 -5.10502338e-01 1.01033890e+00
3.47653925e-02 -1.02703035e+00 -2.94606984e-01 -1.22438526e+00
2.87450522e-01 1.07341814e+00 -7.93684483e-01 -1.80317760e-01
-6.01391375e-01 -4.31540608e-01 7.87765920e-01 6.61804318e-01
3.86002988e-01 -9.05603886e-01 -2.00274736e-01 3.29413801e-01
-4.91810948e-01 -1.34012961e+00 -8.74902755e-02 2.72347629e-01
-6.23495817e-01 -1.23561168e+00 -8.40696245e-02 -1.06677973e+00
1.75153464e-01 -1.00578237e-02 1.79595101e+00 1.41842246e-01
-8.37077945e-02 2.82479465e-01 -7.72543311e-01 -1.89070553e-01
-7.52597451e-02 5.24072349e-01 -2.98781604e-01 -1.83750764e-01
6.40493512e-01 -4.44973946e-01 -1.90235779e-01 2.14314103e-01
-6.79175735e-01 -3.73530805e-01 1.57117903e-01 1.12797582e+00
3.41779947e-01 4.71705869e-02 7.22836494e-01 -1.29922235e+00
9.27311540e-01 -5.74838161e-01 -1.11251548e-02 9.61229920e-01
-1.01130307e+00 4.96506929e-01 5.25146127e-01 -1.73356742e-01
-8.29807699e-01 -4.22306150e-01 -2.79831529e-01 -1.75917104e-01
2.54143402e-02 7.95227170e-01 -8.50550234e-02 1.66425943e-01
9.34952796e-01 -2.28392273e-01 -2.21210212e-01 -5.73881865e-01
6.50992095e-01 8.32605541e-01 3.79921138e-01 -6.18425727e-01
6.07209682e-01 9.65805277e-02 1.39310390e-01 -5.18012643e-01
-1.29500353e+00 -7.66268134e-01 -6.28323317e-01 1.89383671e-01
8.37082803e-01 -5.26716948e-01 -7.35381424e-01 -3.55135232e-01
-1.55947924e+00 3.77556086e-02 -3.87947291e-01 1.49607599e-01
-3.52170080e-01 3.94748569e-01 -6.99601710e-01 -6.76676869e-01
-5.76349497e-01 -5.13014197e-01 9.65385497e-01 -1.93694130e-01
-5.55171371e-01 -1.24143302e+00 4.81753170e-01 6.98691964e-01
2.30555728e-01 2.12633610e-01 1.47409606e+00 -1.10621524e+00
-1.28658533e-01 -4.67123508e-01 -3.36271971e-01 1.98933586e-01
-9.42883492e-02 -1.64290160e-01 -7.99283981e-01 2.12104395e-01
-4.43458438e-01 -5.56775689e-01 1.09323442e+00 -3.81939083e-01
1.11046219e+00 -3.10443431e-01 -5.06648540e-01 1.77146390e-01
1.41814899e+00 -2.44728655e-01 8.46701741e-01 2.64113277e-01
5.14228642e-01 1.01765132e+00 8.38635683e-01 -2.12558582e-01
9.42828357e-01 9.85359371e-01 2.04191834e-01 2.79498816e-01
-2.90917665e-01 -4.23232079e-01 -5.80336452e-02 7.30633378e-01
-2.09346086e-01 -3.16062331e-01 -1.13173938e+00 8.70288372e-01
-2.16687202e+00 -1.16038156e+00 -6.95599675e-01 1.65932441e+00
1.15055847e+00 6.81531802e-03 -1.25871181e-01 3.38321984e-01
2.76475728e-01 3.13548684e-01 -7.45034125e-03 -6.88200533e-01
-1.69884056e-01 6.14155829e-01 2.13392347e-01 5.56162238e-01
-1.23139477e+00 9.27192330e-01 7.30836010e+00 8.00853491e-01
-2.23413095e-01 2.89494038e-01 2.02143684e-01 6.35374427e-01
-4.21987593e-01 2.81707704e-01 -7.67904103e-01 -1.31353326e-02
1.28131509e+00 -1.79653227e-01 7.11934865e-02 8.08861613e-01
-7.37644911e-01 -9.75891277e-02 -1.64316487e+00 7.32227683e-01
2.99715102e-01 -1.64008987e+00 -2.17623174e-01 -1.29889548e-01
4.35983360e-01 -2.07732767e-01 -3.35577071e-01 6.31150365e-01
1.18359745e-01 -1.15558195e+00 2.43765414e-01 5.80088079e-01
5.57808876e-01 -4.96349961e-01 1.18325794e+00 1.57980874e-01
-1.23069787e+00 8.70418400e-02 -4.84697223e-01 5.97627498e-02
1.29358932e-01 5.92418492e-01 -6.94117308e-01 1.15714991e+00
5.62940478e-01 5.58970332e-01 -8.96561384e-01 8.88491452e-01
-6.28404856e-01 7.46562243e-01 -1.49555914e-02 -1.25939354e-01
-1.76342934e-01 1.46277666e-01 4.70285356e-01 1.23732388e+00
-2.82470703e-01 5.07892966e-02 -1.79971874e-01 4.76513028e-01
-3.12062889e-01 1.96030661e-01 -5.91294289e-01 -1.11850873e-01
4.93923008e-01 1.12174714e+00 7.45703876e-02 -4.17665541e-01
-4.05110568e-01 9.51746106e-01 1.03660655e+00 8.02813023e-02
-6.32653832e-01 -7.82354772e-01 5.17467380e-01 -2.78065145e-01
5.05709827e-01 -1.52399793e-01 -2.53757685e-01 -1.29976368e+00
2.37931252e-01 -7.64632404e-01 8.44439030e-01 -6.70032501e-01
-1.49387145e+00 6.99886203e-01 8.72046947e-02 -1.04277301e+00
-3.40500891e-01 -6.16727233e-01 -5.57402313e-01 7.68896818e-01
-1.78808153e+00 -1.24324512e+00 -5.08575328e-02 1.45742357e-01
1.29831672e-01 -6.79882914e-02 1.54896832e+00 2.30867147e-01
-4.34464216e-01 7.28179514e-01 -3.52146775e-02 4.08529699e-01
6.93765581e-01 -1.70018363e+00 1.63358629e-01 3.99459958e-01
6.72281146e-01 5.08042395e-01 7.56406426e-01 -2.38970309e-01
-1.25421524e+00 -9.30926025e-01 1.98984921e+00 -8.75975132e-01
1.02412927e+00 -1.85481548e-01 -1.05109310e+00 9.53939319e-01
6.55184567e-01 2.71699458e-01 1.05028796e+00 9.97097492e-01
-8.75571311e-01 -1.12463169e-01 -9.06158566e-01 4.17665780e-01
1.11283064e+00 -7.58708000e-01 -1.47362876e+00 5.69235861e-01
1.10203171e+00 -2.04541728e-01 -1.73896646e+00 5.39387584e-01
3.90037149e-01 -6.40089273e-01 1.12360644e+00 -1.22667837e+00
7.65273988e-01 -1.76515162e-01 -2.78715789e-01 -1.46264136e+00
-1.47410601e-01 -3.52846712e-01 -1.02017915e+00 1.20122790e+00
9.39926863e-01 -7.95967937e-01 5.72356582e-01 6.60458207e-01
5.86203933e-02 -1.19193804e+00 -1.07531285e+00 -9.76878524e-01
2.61940777e-01 -2.35009402e-01 5.35057008e-01 1.08025074e+00
5.67398071e-01 1.05870891e+00 -9.61785465e-02 3.27156484e-01
3.45373541e-01 7.01105714e-01 4.78310347e-01 -1.32579195e+00
-3.83126289e-01 -1.20669812e-01 -7.30705023e-01 -8.61864805e-01
6.35189056e-01 -1.14761019e+00 -3.21667194e-01 -1.90391409e+00
-6.82898909e-02 -4.19470996e-01 -2.64287591e-01 6.42578900e-01
-4.53042179e-01 -7.82956406e-02 7.25502819e-02 1.82679202e-02
-9.09734488e-01 6.56815410e-01 6.86146021e-01 -3.12808573e-01
2.59427011e-01 -2.09855214e-01 -5.23444176e-01 2.04599917e-01
5.68229020e-01 -5.36813676e-01 -2.28437632e-01 -3.11342031e-01
5.25860667e-01 1.02011792e-01 1.82127446e-01 -6.01579607e-01
4.23564017e-01 2.79637694e-01 -5.39048873e-02 -6.02311492e-01
5.36699653e-01 -6.96985960e-01 -2.33530015e-01 2.73408979e-01
-7.04046905e-01 -1.55510113e-01 -1.77252576e-01 4.69580114e-01
-6.62859917e-01 -4.91958678e-01 8.96073431e-02 3.87346625e-01
-5.04854560e-01 5.57396933e-02 1.45246789e-01 4.06078517e-01
8.55949581e-01 3.59425992e-01 -7.86768496e-01 -3.70371282e-01
-7.29431272e-01 5.91689169e-01 -2.46372104e-01 3.49836349e-01
5.37763894e-01 -1.45250583e+00 -8.97073805e-01 -5.86591244e-01
7.34128237e-01 -2.49002829e-01 -2.14632526e-01 8.41656804e-01
-3.46879244e-01 7.91421533e-01 4.85769778e-01 -4.69154827e-02
-1.42836714e+00 6.20496631e-01 3.54577094e-01 -9.73921895e-01
-4.84534800e-01 8.43192577e-01 -7.14823365e-01 -7.47431576e-01
9.61891115e-02 -1.01370037e-01 -7.07960010e-01 3.40416253e-01
3.55913103e-01 5.44011593e-01 3.71252835e-01 -3.05671066e-01
-3.93226892e-01 1.23748451e-01 -7.62476772e-02 1.16941527e-01
1.43955827e+00 3.05197656e-01 -4.46427584e-01 4.85622466e-01
1.47146451e+00 -1.50577486e-01 -3.11202109e-01 -6.88234806e-01
6.69691026e-01 -3.81745964e-01 -1.72794238e-03 -7.98675060e-01
-4.56675857e-01 7.40698636e-01 -3.26243974e-02 8.85007679e-01
7.59867549e-01 5.50825834e-01 9.33213294e-01 7.26816177e-01
3.61078978e-01 -9.26426768e-01 -2.25759968e-01 7.10704684e-01
1.06216931e+00 -1.18926871e+00 2.27955520e-01 -8.11904132e-01
-7.64838755e-01 1.23963976e+00 3.57381940e-01 -1.63499415e-02
8.85369599e-01 -1.04942724e-01 -3.28348875e-01 -4.93107319e-01
-1.36509836e+00 -6.66780233e-01 7.40317822e-01 5.60748756e-01
6.40200496e-01 9.52065215e-02 -6.76226079e-01 6.25554323e-01
-3.08471322e-01 -3.65306407e-01 1.98021591e-01 9.29997861e-01
-1.81626439e-01 -1.52467680e+00 1.60904586e-01 7.08823681e-01
-4.58752811e-01 -2.59587348e-01 -5.45047820e-01 7.03080654e-01
3.48889604e-02 1.20753598e+00 2.15342268e-03 -5.50861597e-01
5.41567445e-01 6.97979867e-01 5.33732116e-01 -8.04536939e-01
-5.41551948e-01 -9.38805878e-01 1.16402948e+00 -5.92784524e-01
-6.21003628e-01 -6.21642947e-01 -9.64495778e-01 -3.26328814e-01
-8.21799278e-01 4.33395028e-01 2.97108173e-01 1.01816392e+00
5.17765164e-01 4.00458276e-01 4.97312486e-01 1.97737530e-01
-5.66347778e-01 -1.14996660e+00 -3.88206244e-01 4.84797508e-01
-6.63295090e-02 -6.62129998e-01 -2.81640410e-01 -2.44726136e-01] | [9.537646293640137, 8.475790023803711] |
e68a569d-b431-45be-8573-85fd385cd143 | a-comprehensive-survey-on-graph-anomaly | 2106.07178 | null | https://arxiv.org/abs/2106.07178v5 | https://arxiv.org/pdf/2106.07178v5.pdf | A Comprehensive Survey on Graph Anomaly Detection with Deep Learning | Anomalies represent rare observations (e.g., data records or events) that deviate significantly from others. Over several decades, research on anomaly mining has received increasing interests due to the implications of these occurrences in a wide range of disciplines. Anomaly detection, which aims to identify rare observations, is among the most vital tasks in the world, and has shown its power in preventing detrimental events, such as financial fraud, network intrusion, and social spam. The detection task is typically solved by identifying outlying data points in the feature space and inherently overlooks the relational information in real-world data. Graphs have been prevalently used to represent the structural information, which raises the graph anomaly detection problem - identifying anomalous graph objects (i.e., nodes, edges and sub-graphs) in a single graph, or anomalous graphs in a database/set of graphs. However, conventional anomaly detection techniques cannot tackle this problem well because of the complexity of graph data. For the advent of deep learning, graph anomaly detection with deep learning has received a growing attention recently. In this survey, we aim to provide a systematic and comprehensive review of the contemporary deep learning techniques for graph anomaly detection. We compile open-sourced implementations, public datasets, and commonly-used evaluation metrics to provide affluent resources for future studies. More importantly, we highlight twelve extensive future research directions according to our survey results covering unsolved and emerging research problems and real-world applications. With this survey, our goal is to create a "one-stop-shop" that provides a unified understanding of the problem categories and existing approaches, publicly available hands-on resources, and high-impact open challenges for graph anomaly detection using deep learning. | ['Leman Akoglu', 'Hui Xiong', 'Quan Z. Sheng', 'Chuan Zhou', 'Jian Yang', 'Shan Xue', 'Jia Wu', 'Xiaoxiao Ma'] | 2021-06-14 | null | null | null | null | ['graph-anomaly-detection'] | ['graphs'] | [ 1.33787515e-02 2.08811555e-02 3.08846477e-02 -1.86357304e-01
1.14087500e-01 -2.49719232e-01 3.45185697e-01 8.88904870e-01
1.42040655e-01 4.20967191e-01 -2.54590183e-01 -4.17697847e-01
-2.79686362e-01 -1.11754143e+00 -5.85313261e-01 -6.22190595e-01
-8.19721401e-01 3.72166783e-01 4.78325151e-02 -3.02050591e-01
2.74660110e-01 8.40308666e-01 -1.44986510e+00 -1.51655162e-02
7.79984474e-01 1.11289239e+00 -6.32533550e-01 4.18685526e-01
-4.79955673e-01 7.66525328e-01 -6.94293737e-01 -6.50921822e-01
1.13563828e-01 -3.56219739e-01 -6.18305504e-01 2.01926291e-01
5.27292967e-01 -1.13990583e-01 -7.64983714e-01 1.25980520e+00
2.34175742e-01 8.65596980e-02 4.81314093e-01 -1.83990157e+00
-7.12621033e-01 4.19715255e-01 -9.10911322e-01 7.74643719e-01
2.90378243e-01 3.73891555e-02 1.21081662e+00 -5.56845486e-01
3.05481106e-01 9.23065722e-01 6.93047941e-01 3.21815223e-01
-8.96027505e-01 -6.18883431e-01 6.65580332e-01 5.15323520e-01
-1.19901693e+00 2.39644665e-03 1.06539595e+00 -3.03826451e-01
1.09005857e+00 4.11323756e-01 7.33521342e-01 1.10348439e+00
4.09828573e-01 6.78582370e-01 1.91856056e-01 -1.95840225e-01
1.08069137e-01 -6.14816964e-01 3.85490805e-01 8.38695586e-01
8.82078409e-01 -9.74755734e-02 -2.84607321e-01 -4.29729342e-01
3.46390069e-01 5.65198421e-01 -5.36225960e-02 -4.32113826e-01
-8.04636121e-01 8.77007186e-01 6.43378019e-01 4.81184721e-01
-4.16369200e-01 4.38821837e-02 7.90381253e-01 7.29951918e-01
6.61961854e-01 3.99127126e-01 -1.75562322e-01 5.50134219e-02
-2.57251054e-01 2.79600501e-01 8.28564584e-01 8.51465225e-01
5.63623190e-01 6.11214638e-01 1.59255579e-01 7.14210093e-01
1.42967269e-01 1.09335646e-01 3.00587744e-01 -8.19368288e-02
3.10346216e-01 1.30839348e+00 -6.17077053e-01 -1.62937379e+00
-6.49601877e-01 -6.13644600e-01 -1.38869703e+00 -8.95552188e-02
3.79902750e-01 1.95863564e-02 -8.68715048e-01 1.32655132e+00
4.03824240e-01 4.69961345e-01 -2.41595492e-01 6.29488945e-01
9.56543148e-01 2.78429836e-01 -1.73118770e-01 -1.11420065e-01
9.61836815e-01 -6.20417118e-01 -7.44661331e-01 -3.47391456e-01
8.49783361e-01 -3.44975442e-01 9.16573048e-01 3.05190653e-01
-4.96931642e-01 -9.91377793e-03 -9.22862887e-01 1.87691942e-01
-7.67945588e-01 -6.23334289e-01 9.28245664e-01 4.65031445e-01
-8.13782632e-01 6.60743058e-01 -8.22097361e-01 -5.84982097e-01
6.89349234e-01 3.21155190e-01 -5.05601406e-01 -1.67618617e-01
-1.02865636e+00 3.72861952e-01 3.64336371e-01 7.32612088e-02
-5.59634626e-01 -3.60089391e-01 -1.03127623e+00 1.19503155e-01
7.54032791e-01 -2.66362518e-01 7.43885636e-01 -9.28201497e-01
-5.65537333e-01 9.00171220e-01 1.97644815e-01 -6.50280952e-01
2.30389267e-01 -1.45431653e-01 -1.18057692e+00 -2.75157005e-01
3.20723280e-02 -5.34935951e-01 8.67149055e-01 -7.75350809e-01
-5.20157635e-01 -8.21433067e-01 -8.30451772e-02 -1.13171518e-01
-5.77813029e-01 -4.41901237e-02 -1.53670028e-01 -8.31612825e-01
3.04241210e-01 -7.06196129e-01 -3.75988424e-01 -2.88053662e-01
-7.97857881e-01 -4.47619706e-01 1.23126459e+00 -3.12330723e-01
1.63443363e+00 -2.33454323e+00 -1.93088442e-01 5.48125148e-01
9.30106521e-01 2.76280016e-01 1.01125604e-02 6.59241021e-01
-4.86046404e-01 1.93347424e-01 -4.73446399e-01 -1.81707670e-03
-2.94815332e-01 3.43611538e-01 -3.27632576e-01 8.07641447e-01
2.48532861e-01 8.94625008e-01 -9.70056891e-01 -1.06139980e-01
2.10817263e-01 -3.25933658e-02 -2.85911292e-01 2.11416543e-01
-1.60890192e-01 3.44724745e-01 -5.38687468e-01 1.07134223e+00
5.65538168e-01 -4.82234359e-01 1.25950173e-01 2.22688019e-01
3.48866552e-01 -7.51895159e-02 -1.11935794e+00 1.11374223e+00
1.59563676e-01 5.17894566e-01 -2.12249517e-01 -1.62005579e+00
1.06734014e+00 -5.45263775e-02 6.70213461e-01 -7.52494931e-01
7.58046005e-03 3.70896101e-01 4.56263125e-01 -4.16488200e-01
3.19802970e-01 4.15485680e-01 -1.77872062e-01 4.91225570e-01
-1.14339381e-01 5.21524131e-01 3.53218555e-01 4.13158745e-01
1.63049722e+00 -7.44489551e-01 6.05830669e-01 8.59040469e-02
6.60723150e-01 -1.81248203e-01 4.27815765e-01 8.00837576e-01
-3.09006304e-01 3.24600071e-01 9.78249192e-01 -1.27801585e+00
-6.75864398e-01 -9.60196674e-01 1.09964848e-01 9.85762417e-01
-1.00284785e-01 -6.10478520e-01 -2.73551404e-01 -1.23334217e+00
4.76499766e-01 3.89873207e-01 -6.53673470e-01 -6.34065926e-01
-7.32702315e-01 -9.51323211e-01 5.24701416e-01 4.46863174e-01
2.96200931e-01 -1.36860168e+00 1.41440660e-01 1.91636965e-01
1.54873312e-01 -1.09774125e+00 -1.57489002e-01 -1.23660050e-01
-1.04133081e+00 -1.70578194e+00 -2.78273243e-02 -7.27386296e-01
7.41420448e-01 2.60688812e-01 1.44534469e+00 6.61860883e-01
-4.77496266e-01 3.78208578e-01 -4.09461230e-01 -6.01194322e-01
-2.73928881e-01 4.78558466e-02 3.44661385e-01 2.31274068e-01
8.77691090e-01 -7.82889724e-01 -3.96143734e-01 7.91571066e-02
-1.09196484e+00 -7.82731056e-01 4.79198873e-01 6.23278856e-01
6.04124546e-01 2.52921999e-01 8.42739284e-01 -1.17893147e+00
9.81545150e-01 -1.02564013e+00 -5.80614746e-01 4.05801423e-02
-7.97641873e-01 -2.54921079e-01 9.32745755e-01 -7.55123347e-02
-1.69779345e-01 -4.39099550e-01 -1.35949373e-01 -4.96593177e-01
-3.60400468e-01 7.72438288e-01 -1.52097508e-01 -2.27735773e-01
6.18601680e-01 3.21056485e-01 1.76779434e-01 -4.15784061e-01
1.11379921e-01 4.67120647e-01 3.92156780e-01 -3.74027729e-01
6.14829004e-01 3.97463173e-01 2.56714255e-01 -1.05921638e+00
-6.91050410e-01 -6.27298534e-01 -3.78447473e-01 -2.12027788e-01
3.24161142e-01 -4.40021992e-01 -5.64287126e-01 7.38891602e-01
-8.44420135e-01 2.31115758e-01 -4.16155577e-01 -3.42330560e-02
5.95065858e-03 8.30011547e-01 -4.72256750e-01 -7.36409307e-01
-5.70610344e-01 -7.21604645e-01 6.72158122e-01 8.54519382e-02
-1.79895923e-01 -1.25497341e+00 8.53429884e-02 -1.28406674e-01
3.11464399e-01 8.09907615e-01 1.13134193e+00 -1.48640919e+00
-5.35836935e-01 -7.91262090e-01 -2.44996727e-01 3.21490645e-01
4.05006498e-01 1.04157105e-01 -7.01874316e-01 -5.38807690e-01
-2.39206910e-01 1.22000530e-01 7.14850724e-01 2.35970974e-01
1.70987236e+00 -4.01401371e-01 -4.39053178e-01 6.49749041e-01
1.12598085e+00 1.99978828e-01 4.50143099e-01 2.83829957e-01
1.21888506e+00 3.75605434e-01 2.21396327e-01 4.94105190e-01
1.71683401e-01 2.13562608e-01 9.27549243e-01 -1.95242857e-04
3.14460367e-01 -4.80830930e-02 6.21449761e-02 8.04085016e-01
-2.03601122e-01 -5.18988907e-01 -1.16984785e+00 5.09406626e-01
-1.87031031e+00 -8.76875043e-01 -5.27023077e-01 2.17448139e+00
-5.42299934e-02 3.64236206e-01 3.44745070e-01 3.33080947e-01
8.74616683e-01 3.87756526e-01 -9.20379281e-01 -4.45129573e-01
-3.92939039e-02 6.38409853e-02 2.21831158e-01 -9.51136127e-02
-1.24748087e+00 7.32651532e-01 5.82836533e+00 4.58929151e-01
-9.90829885e-01 -4.36524510e-01 5.96944809e-01 1.25419617e-01
-3.22215222e-02 -2.90590554e-01 -3.35323095e-01 5.22634864e-01
7.82489538e-01 -4.23517495e-01 3.02328944e-01 1.05266309e+00
-1.61178783e-01 2.95748174e-01 -1.23665869e+00 1.05220759e+00
1.50189251e-01 -1.17449653e+00 3.94332677e-01 2.99270660e-01
4.63336736e-01 2.40224496e-01 3.42921987e-02 3.76612782e-01
1.21204399e-01 -1.11845374e+00 -1.65831059e-01 3.22117299e-01
5.36358118e-01 -9.63661790e-01 1.01362693e+00 1.55422896e-01
-1.34026980e+00 -3.24118763e-01 -2.39509806e-01 -4.06343609e-01
-2.00368062e-01 1.04148126e+00 -4.72982258e-01 8.17950487e-01
8.72293591e-01 1.24603498e+00 -5.73948622e-01 1.11577451e+00
7.31897578e-02 4.80627984e-01 -2.35800266e-01 9.14082527e-02
2.08631277e-01 -4.30873215e-01 9.12202895e-01 8.67226958e-01
1.80021733e-01 -2.02555135e-01 3.75538528e-01 6.56370640e-01
-3.52452815e-01 2.44458020e-01 -1.29138291e+00 -4.38272119e-01
2.62612224e-01 1.15548790e+00 -7.44220018e-01 -1.05989225e-01
-7.74678648e-01 6.91010654e-01 5.24701774e-01 2.88525045e-01
-5.17370462e-01 -5.90406835e-01 1.00085700e+00 2.29961455e-01
-2.39045307e-01 -9.24213305e-02 -1.32479280e-01 -1.30332136e+00
2.79270142e-01 -1.10179079e+00 1.13520980e+00 7.18389899e-02
-1.87162042e+00 6.36490464e-01 -2.46828496e-01 -1.25987971e+00
-2.52265573e-01 -6.54266775e-01 -1.01207376e+00 5.55764079e-01
-1.14958334e+00 -7.81024754e-01 -6.69307947e-01 7.79387355e-01
3.01431507e-01 -6.33097768e-01 7.70606518e-01 3.89732838e-01
-9.07999694e-01 6.56183183e-01 9.93908420e-02 6.48346186e-01
3.26545089e-01 -1.31458819e+00 1.17408240e+00 1.09751368e+00
3.76662940e-01 4.33564961e-01 4.79472548e-01 -8.68455291e-01
-1.30681729e+00 -1.24258173e+00 5.29338658e-01 -3.84400040e-01
1.03175700e+00 -3.77646655e-01 -1.46956182e+00 9.73303735e-01
-3.54351074e-01 6.95476592e-01 6.72129154e-01 4.32218820e-01
-2.38474846e-01 -2.32465461e-01 -1.08123422e+00 5.67991436e-01
1.35652518e+00 -3.91555250e-01 -2.10120738e-01 4.25024122e-01
4.68627840e-01 -3.90424848e-01 -5.48727632e-01 4.76252943e-01
1.79556444e-01 -1.18898535e+00 7.46767163e-01 -1.13344193e+00
1.66118190e-01 -3.52981240e-02 1.72715366e-01 -1.38534033e+00
-3.29746485e-01 -5.44541597e-01 -9.92771685e-01 8.42828870e-01
9.42393485e-03 -1.10409665e+00 9.74343181e-01 2.82980055e-01
-3.88116837e-01 -9.97020721e-01 -8.41645360e-01 -8.33149612e-01
-2.49104768e-01 -4.91692960e-01 8.89212251e-01 1.29789007e+00
-1.04363248e-01 1.84851155e-01 -1.75731927e-01 2.96841621e-01
6.99725866e-01 1.10158131e-01 9.47670043e-01 -1.74468446e+00
2.98478037e-01 -6.65565431e-01 -1.15898097e+00 -4.40372497e-01
1.35641307e-01 -1.06049275e+00 -6.36899650e-01 -1.41979229e+00
-2.21138090e-01 -2.66679943e-01 -5.52291512e-01 4.58705217e-01
-1.74471006e-01 3.51402722e-02 -4.03611034e-01 2.07130924e-01
-7.28177428e-01 4.51613188e-01 9.99212682e-01 -2.26828322e-01
-1.19570412e-01 2.84095138e-01 -7.91180253e-01 9.73977983e-01
8.79071534e-01 -4.09547180e-01 -3.10983896e-01 -1.44418553e-01
4.74092931e-01 -3.68510574e-01 3.70773435e-01 -1.03296089e+00
9.21580866e-02 7.62832351e-03 2.28335917e-01 -4.19683307e-01
-3.36314201e-01 -9.40715432e-01 -2.50232577e-01 4.47950840e-01
1.67006448e-01 7.22044051e-01 1.47493780e-01 1.18645835e+00
-4.20592636e-01 1.86278120e-01 4.71699089e-01 -1.21133037e-01
-9.58241403e-01 9.89622951e-01 -1.22046486e-01 3.64770323e-01
1.33560133e+00 -1.87929526e-01 -4.27721083e-01 -4.61992145e-01
-7.13083565e-01 3.94139349e-01 2.06356138e-01 8.14708948e-01
8.45383465e-01 -1.35528135e+00 -7.57822514e-01 6.19052351e-01
6.32265389e-01 2.64502585e-01 3.04848641e-01 7.06901848e-01
-6.28333330e-01 1.84105244e-02 -2.65615374e-01 -5.57339013e-01
-1.15047717e+00 7.19635367e-01 4.16423559e-01 -3.60803604e-01
-1.12684178e+00 5.16618311e-01 1.33771980e-02 -2.63696551e-01
3.33350658e-01 -1.91807970e-01 -3.70700240e-01 -1.51381284e-01
4.31253135e-01 5.47906876e-01 4.61401880e-01 -4.87968981e-01
-5.92487395e-01 8.27127397e-02 -5.62012732e-01 1.06020832e+00
1.22974217e+00 1.75421953e-01 -4.64857012e-01 5.28872550e-01
9.55981135e-01 -2.64826268e-02 -4.82031047e-01 -3.38286430e-01
4.28636461e-01 -4.94379669e-01 -2.90341556e-01 -2.21822992e-01
-1.49499357e+00 6.31296456e-01 3.67725670e-01 1.13193750e+00
1.04762125e+00 1.20609753e-01 8.63635778e-01 5.53597629e-01
1.55610338e-01 -8.39034617e-01 2.91759610e-01 6.49167836e-01
9.10287142e-01 -1.43447280e+00 8.92133713e-02 -3.84574324e-01
-3.21710259e-01 1.18040955e+00 1.01042175e+00 -4.10679817e-01
8.14485252e-01 -4.76853065e-02 -1.16602555e-01 -7.77092099e-01
-3.23206306e-01 -9.74601060e-02 4.14342552e-01 7.17955232e-01
4.85129237e-01 3.84378210e-02 -5.06293811e-02 4.24536318e-01
-1.06120028e-01 -6.42915666e-01 4.66753423e-01 8.09239507e-01
-1.81117833e-01 -9.00550365e-01 -5.51350787e-02 1.37478364e+00
-6.73102379e-01 1.17507624e-02 -7.19616234e-01 9.08064663e-01
-1.94516972e-01 7.45829701e-01 4.58548039e-01 -4.08754677e-01
6.02131784e-01 9.29684564e-03 -8.58286917e-02 -5.08371770e-01
-2.75533766e-01 -6.03115678e-01 -3.63493562e-02 -8.01801205e-01
7.73920342e-02 -3.45220506e-01 -1.06429696e+00 -7.59312153e-01
-2.33519033e-01 5.95176667e-02 1.50769353e-01 9.39070523e-01
6.66763127e-01 8.34334910e-01 5.25910020e-01 -1.74018875e-01
-2.81405300e-01 -8.36962044e-01 -9.24267888e-01 7.23741770e-01
6.10771954e-01 -7.20084429e-01 -6.37600541e-01 -7.38397360e-01] | [6.641246795654297, 5.80620813369751] |
92bdc894-85bf-43e8-ab98-26cecc25e76a | betrayed-by-motion-camouflaged-object | 2011.11630 | null | https://arxiv.org/abs/2011.11630v1 | https://arxiv.org/pdf/2011.11630v1.pdf | Betrayed by Motion: Camouflaged Object Discovery via Motion Segmentation | The objective of this paper is to design a computational architecture that discovers camouflaged objects in videos, specifically by exploiting motion information to perform object segmentation. We make the following three contributions: (i) We propose a novel architecture that consists of two essential components for breaking camouflage, namely, a differentiable registration module to align consecutive frames based on the background, which effectively emphasises the object boundary in the difference image, and a motion segmentation module with memory that discovers the moving objects, while maintaining the object permanence even when motion is absent at some point. (ii) We collect the first large-scale Moving Camouflaged Animals (MoCA) video dataset, which consists of over 140 clips across a diverse range of animals (67 categories). (iii) We demonstrate the effectiveness of the proposed model on MoCA, and achieve competitive performance on the unsupervised segmentation protocol on DAVIS2016 by only relying on motion. | ['Andrew Zisserman', 'Weidi Xie', 'Charig Yang', 'Hala Lamdouar'] | 2020-11-23 | null | null | null | null | ['motion-segmentation'] | ['computer-vision'] | [ 3.83753628e-01 -3.34184140e-01 -1.44374207e-01 -1.45335957e-01
-2.57874489e-01 -7.72465050e-01 3.65937382e-01 -4.31958675e-01
-5.28364241e-01 3.12987417e-01 -1.65150747e-01 7.67863914e-02
8.08822289e-02 -3.74929279e-01 -9.68244314e-01 -9.02425766e-01
-2.76484907e-01 1.19405068e-01 7.45292187e-01 -1.59107000e-02
5.28113186e-01 3.90905976e-01 -1.69710541e+00 1.61021456e-01
8.49706352e-01 1.01124334e+00 2.29731724e-01 7.36749947e-01
7.12949485e-02 8.42473030e-01 -7.25228250e-01 -1.06232405e-01
3.81436378e-01 -5.34165800e-01 -1.07264662e+00 4.58905131e-01
9.27730560e-01 -4.36926752e-01 -4.88603085e-01 1.28208363e+00
-7.01662078e-02 3.48912865e-01 6.07540369e-01 -1.26649392e+00
-8.43776047e-01 4.61544067e-01 -5.88623106e-01 6.14896119e-01
1.00254476e-01 4.36126232e-01 6.35818899e-01 -5.64031601e-01
9.29249287e-01 1.09739017e+00 5.42099416e-01 1.04365766e+00
-1.06658900e+00 -6.00196421e-01 4.28030640e-01 4.61274981e-01
-1.20510161e+00 -4.31644827e-01 8.22407722e-01 -6.91012561e-01
5.95715940e-01 1.86395988e-01 9.77892578e-01 9.50944901e-01
2.96173900e-01 1.09005725e+00 8.05725574e-01 1.04687594e-01
3.76799256e-01 -5.83236575e-01 4.25149351e-01 6.75470054e-01
1.84460923e-01 7.64486417e-02 -3.17980528e-01 -1.28011117e-02
6.30358934e-01 1.12347841e-01 -5.06784618e-01 -5.27709603e-01
-1.33034432e+00 5.43270528e-01 3.98618609e-01 3.72683287e-01
-4.72269244e-02 1.74939260e-01 2.45600149e-01 1.82981089e-01
2.38002136e-01 4.58401591e-01 -5.01713574e-01 6.54054387e-03
-1.23164225e+00 3.30395579e-01 6.71829522e-01 1.05398369e+00
5.85248172e-01 1.20682001e-01 -1.13685519e-01 4.75572348e-01
2.57215947e-01 6.12195075e-01 6.06465876e-01 -1.28739488e+00
9.44686532e-02 3.01254541e-01 1.24503613e-01 -1.24160230e+00
-2.16110751e-01 1.55575559e-01 -7.90118039e-01 1.37133867e-01
3.57651949e-01 -7.54918084e-02 -1.30874860e+00 1.91948664e+00
4.74333853e-01 1.03371477e+00 -1.42407892e-02 1.17805588e+00
1.19502938e+00 6.11951172e-01 3.04168403e-01 -1.56824276e-01
1.30216134e+00 -1.41532910e+00 -7.89018393e-01 -1.47161677e-01
4.15888637e-01 -5.16146541e-01 7.19500959e-01 2.86806732e-01
-1.08979738e+00 -6.85100079e-01 -1.31246698e+00 -6.63296804e-02
-3.91015530e-01 -1.86603859e-01 8.13534737e-01 3.83488059e-01
-9.05888617e-01 7.80706584e-01 -1.12578285e+00 -1.26439855e-01
8.21713388e-01 3.31348628e-01 -4.34997231e-01 2.10733756e-01
-7.78762162e-01 4.82790142e-01 4.46630985e-01 3.73251289e-01
-1.38534641e+00 -7.15781510e-01 -9.48383033e-01 -1.05952717e-01
3.09853822e-01 -6.61108971e-01 8.53861809e-01 -1.50334775e+00
-1.41821349e+00 1.07296979e+00 -3.84888165e-02 -6.45698488e-01
6.88282311e-01 -3.39883119e-01 -4.89096552e-01 5.94280779e-01
-2.23364830e-02 1.30067933e+00 1.30515778e+00 -1.12411165e+00
-8.88881266e-01 -2.39182711e-01 -5.61985262e-02 -1.68084264e-01
-6.19357526e-02 -5.71257062e-02 -8.09141397e-01 -1.08928883e+00
5.15094213e-02 -1.13161004e+00 -1.46902934e-01 2.28465851e-02
-2.23995790e-01 2.21405122e-02 1.41813552e+00 -7.57138073e-01
1.18555844e+00 -2.24729347e+00 6.25984371e-01 -3.01213235e-01
4.09369916e-01 6.09826863e-01 -4.17549223e-01 -4.96755332e-01
-5.87984771e-02 -1.91697758e-02 -8.11551571e-01 -3.03294390e-01
-3.51218343e-01 2.31975213e-01 -4.55606192e-01 8.19494545e-01
4.68672365e-01 1.23775029e+00 -8.35922539e-01 -5.85869372e-01
2.22604141e-01 1.36877790e-01 -7.35566437e-01 2.18264326e-01
-3.37456226e-01 8.36270988e-01 -2.92157888e-01 8.51912677e-01
1.08153927e+00 -8.79055485e-02 -1.40193060e-01 -1.02782398e-02
-1.04388334e-02 -1.91565916e-01 -1.04228818e+00 1.92784154e+00
5.34328222e-01 6.99063778e-01 1.84793890e-01 -8.57874691e-01
3.43610644e-01 -7.69814178e-02 8.26762855e-01 -3.37244540e-01
3.73116881e-01 -7.80359507e-02 1.21885501e-02 -6.49509549e-01
3.98524731e-01 5.17954171e-01 6.65462464e-02 1.99206874e-01
3.26578617e-01 -1.28963336e-01 4.66573387e-01 2.08787918e-02
1.19556987e+00 4.04030234e-01 -2.12502643e-01 -5.68620622e-01
6.13125563e-01 2.38296270e-01 8.52434337e-01 6.96609914e-01
-7.78393984e-01 6.78319275e-01 1.57507479e-01 -7.52397537e-01
-8.45995426e-01 -9.72308636e-01 -7.05541074e-02 8.67955387e-01
8.62538636e-01 2.09945440e-02 -1.10274851e+00 -9.62142766e-01
-1.31192237e-01 1.62490666e-01 -8.05498779e-01 -1.84176788e-01
-9.13848400e-01 -7.47694194e-01 4.65028435e-01 4.26463693e-01
8.92015636e-01 -1.23217046e+00 -8.39034021e-01 -1.69452175e-01
-5.04321516e-01 -1.11389434e+00 -8.80628049e-01 -2.48719063e-02
-9.38766301e-01 -1.28848851e+00 -7.35081375e-01 -1.19027102e+00
7.04844177e-01 6.30160630e-01 8.36288393e-01 4.10241216e-01
-2.38618836e-01 3.04710090e-01 -2.80786037e-01 -2.37735156e-02
-1.15698956e-01 -3.74543965e-02 -9.87196565e-02 9.83124319e-03
2.91689992e-01 -4.50395495e-01 -6.81630135e-01 5.78277946e-01
-1.22144783e+00 -4.04923968e-02 2.83758044e-01 6.17494106e-01
6.82761133e-01 -2.92753428e-02 2.55361766e-01 -7.39632607e-01
-1.79030538e-01 -4.50032473e-01 -8.22262585e-01 5.40485494e-02
9.86064896e-02 -2.10690990e-01 2.55432874e-01 -9.51426029e-01
-7.20042288e-01 3.31549525e-01 -5.18287206e-03 -7.41848648e-01
-4.31017637e-01 -2.64307827e-01 -1.90429702e-01 -4.43577766e-01
2.99429268e-01 8.74144137e-02 -1.95541337e-01 -5.45771420e-01
4.94191557e-01 2.90301055e-01 1.14227498e+00 -3.39392275e-01
9.24866021e-01 9.76765335e-01 -1.68454334e-01 -7.65804350e-01
-6.93087995e-01 -5.76064527e-01 -1.10886872e+00 -2.07565710e-01
1.33550131e+00 -8.21041167e-01 -8.00993919e-01 9.32090342e-01
-1.14921463e+00 -5.06360114e-01 -2.01496318e-01 3.82253021e-01
-5.63952684e-01 6.81109428e-01 -8.03026915e-01 -2.61350960e-01
-3.11534047e-01 -1.09820092e+00 9.04318631e-01 5.47507226e-01
-1.07827984e-01 -8.11489105e-01 4.14957367e-02 4.86966133e-01
3.82259153e-02 4.55657214e-01 6.09474242e-01 -4.37369227e-01
-9.79945838e-01 1.61783785e-01 1.62385795e-02 3.47077221e-01
3.26222032e-02 2.37049952e-01 -8.90761852e-01 -4.41195637e-01
1.92447305e-01 -8.52352306e-02 1.25434816e+00 5.70825219e-01
1.29030514e+00 -2.34300241e-01 -3.93362999e-01 1.08458221e+00
1.01859903e+00 4.48937267e-01 8.49474132e-01 4.03560370e-01
9.55636263e-01 3.97265673e-01 7.39043653e-01 -2.41233353e-02
1.84431091e-01 5.34848034e-01 5.59495270e-01 -5.54129966e-02
-3.80030543e-01 -5.82548268e-02 3.90216678e-01 9.55807865e-01
2.83721928e-02 -1.57631144e-01 -4.22587872e-01 8.55710566e-01
-1.92123640e+00 -1.06488216e+00 -1.17233716e-01 1.82390654e+00
5.03529727e-01 1.90163534e-02 8.12177435e-02 -1.79446980e-01
8.17897141e-01 3.02426457e-01 -6.92571819e-01 3.89475487e-02
-3.72859448e-01 5.97181320e-02 5.40068567e-01 2.13330328e-01
-1.85465395e+00 1.33634889e+00 7.04053879e+00 6.85006618e-01
-1.18847620e+00 1.12145841e-01 4.05204743e-01 1.23153783e-01
1.36806652e-01 6.21720916e-03 -5.23053348e-01 7.93567002e-01
3.90172571e-01 3.24630946e-01 5.10721147e-01 8.90059233e-01
-8.49580094e-02 -1.00428432e-01 -1.00932646e+00 8.16772044e-01
3.59446526e-01 -1.39991558e+00 1.69018865e-01 -2.41109416e-01
1.07327330e+00 -3.35550644e-02 2.49667123e-01 -2.93330103e-03
1.18891738e-01 -8.81308496e-01 1.04550314e+00 4.87533778e-01
3.09969753e-01 -4.90651876e-01 5.10500491e-01 2.21346304e-01
-1.29124975e+00 -1.23012282e-01 -3.61936957e-01 1.14770927e-01
3.12635340e-02 6.45516440e-02 -2.10917488e-01 3.08919787e-01
1.02065194e+00 1.07796443e+00 -6.27955139e-01 1.35418928e+00
-1.75692514e-01 6.03113651e-01 -1.27432540e-01 4.19350266e-01
2.48171017e-01 -3.11853230e-01 8.28617752e-01 1.13875759e+00
3.01347282e-02 1.79486275e-01 2.02291787e-01 7.82903671e-01
-2.35357821e-01 -2.99492836e-01 -3.20840210e-01 -6.76371455e-02
1.76112697e-01 1.33528173e+00 -1.16250765e+00 -4.80892986e-01
-3.15806478e-01 1.25093043e+00 2.11710092e-02 3.96047801e-01
-1.19594634e+00 -3.77482265e-01 9.36117649e-01 -1.61706671e-01
8.89707744e-01 -4.10070211e-01 -8.99664536e-02 -1.32452667e+00
-1.77660421e-01 -7.92360783e-01 5.59903085e-01 -7.33814538e-01
-1.01518095e+00 4.75859046e-01 -1.07296884e-01 -1.29096949e+00
1.10971563e-01 -5.72997987e-01 -4.10315663e-01 1.04045920e-01
-1.35610390e+00 -1.11882186e+00 -4.13418531e-01 6.21966064e-01
7.56109774e-01 5.46389967e-02 2.99603492e-01 6.44114435e-01
-6.00710690e-01 3.15849692e-01 -1.14398651e-01 4.48082358e-01
5.53499401e-01 -9.29224908e-01 5.50670087e-01 1.26060617e+00
3.42871040e-01 4.92118090e-01 5.23797035e-01 -8.92151833e-01
-1.44365311e+00 -1.38009167e+00 3.32927197e-01 -7.41345882e-01
5.91391385e-01 -3.03652525e-01 -1.10264933e+00 8.00442517e-01
8.62588547e-03 2.38601565e-01 2.52868623e-01 -7.64176071e-01
-2.31517851e-01 2.97354311e-01 -1.25873458e+00 6.97024524e-01
1.47127247e+00 -1.84389755e-01 -8.90802085e-01 3.35537456e-02
9.96619046e-01 -6.90866649e-01 -5.15314102e-01 4.84442621e-01
4.61818695e-01 -5.20587206e-01 9.98443782e-01 -8.98531497e-01
4.57948416e-01 -8.53145599e-01 1.85719505e-03 -8.14619958e-01
-3.46213609e-01 -8.81703794e-01 -3.91126007e-01 1.13646317e+00
-1.32728174e-01 -2.70755738e-01 7.71123946e-01 3.95943314e-01
-2.31965706e-01 -2.77162045e-01 -1.07734740e+00 -5.60335457e-01
-1.21770106e-01 -3.87158871e-01 5.11735618e-01 1.11785376e+00
-5.78573048e-01 -1.54959798e-01 -5.52473724e-01 3.24132055e-01
6.30113900e-01 4.19468671e-01 7.94193089e-01 -1.05515385e+00
-8.74306187e-02 -3.72714251e-01 -7.91265666e-01 -1.56454086e+00
4.50632900e-01 -7.37825930e-01 4.28256273e-01 -7.98248291e-01
3.22102636e-01 -6.77306205e-02 -3.24205726e-01 4.50835705e-01
-2.11954638e-01 6.31246388e-01 2.13396087e-01 4.04693037e-01
-1.10254681e+00 4.47703362e-01 1.41433775e+00 -5.47427952e-01
-3.49579662e-01 -1.77727044e-01 -4.74923044e-01 1.26089776e+00
3.35697919e-01 -6.63498878e-01 -8.75994936e-02 -7.10490882e-01
-6.79320514e-01 -3.28443855e-01 5.93583822e-01 -1.10836494e+00
2.14220002e-01 -3.71916533e-01 5.23864031e-01 -6.28980219e-01
7.02581704e-02 -8.23858023e-01 2.45458663e-01 6.66891277e-01
-1.55070141e-01 -4.26290631e-02 2.30489492e-01 8.53103638e-01
-1.23741090e-01 -8.33068136e-03 1.18272245e+00 1.22505486e-01
-1.41115701e+00 4.95717317e-01 -4.96262938e-01 2.25754142e-01
1.24155867e+00 -1.18302606e-01 -4.30607587e-01 2.59615839e-01
-5.82744539e-01 2.48425409e-01 6.95315659e-01 7.76370287e-01
4.96432453e-01 -1.22463846e+00 -2.63760954e-01 2.72356689e-01
-1.41513839e-01 -1.73260588e-02 4.63080645e-01 7.95563698e-01
-8.97474468e-01 1.93388656e-01 -5.49208939e-01 -7.73177803e-01
-1.34654844e+00 8.75095785e-01 3.15077990e-01 3.57241273e-01
-8.91740143e-01 9.94373858e-01 4.22185987e-01 7.17712566e-02
2.10652351e-01 -6.15488708e-01 -3.20716202e-01 -1.39549375e-01
7.42558658e-01 4.23278242e-01 -2.43882373e-01 -1.10360777e+00
-5.60461104e-01 9.66069758e-01 2.61576567e-02 2.33771920e-01
1.16297519e+00 -2.68702954e-01 -2.55119592e-01 1.95754588e-01
1.03008878e+00 -1.20904170e-01 -1.71218741e+00 3.61952260e-02
1.39879763e-01 -7.63555050e-01 -1.95567116e-01 -4.00070667e-01
-1.51295149e+00 6.06574833e-01 8.07432473e-01 -1.02692269e-01
1.27141416e+00 1.35734836e-02 1.29119337e+00 3.06357145e-01
3.19977105e-01 -1.15091038e+00 1.60625800e-01 3.84514809e-01
4.26172823e-01 -1.08457935e+00 -1.97999880e-01 -4.36488539e-01
-5.89999735e-01 7.81383991e-01 5.76452792e-01 -4.28661823e-01
6.02160275e-01 1.88093975e-01 2.48613413e-02 -2.19909623e-01
-4.59292173e-01 -3.05254698e-01 3.83315116e-01 8.04474652e-01
-2.77466565e-01 -2.11945042e-01 -2.77494937e-01 6.93566978e-01
9.60532203e-02 -1.05170459e-02 2.68864334e-01 9.68927681e-01
-5.22411048e-01 -5.25631607e-01 -2.18459725e-01 -3.13348584e-02
-6.21527970e-01 1.93315700e-01 -6.02005661e-01 6.76128685e-01
5.98598719e-01 1.04340565e+00 3.14001232e-01 -3.32400531e-01
1.24579944e-01 -1.90428793e-01 3.76421541e-01 -2.46107280e-01
-4.72076893e-01 5.34269102e-02 -4.38084841e-01 -8.76614630e-01
-8.55908215e-01 -5.15311182e-01 -1.33892334e+00 -4.07191753e-01
-1.75827309e-01 5.18374480e-02 3.64284664e-01 9.88535821e-01
3.10490251e-01 5.28495789e-01 3.53641778e-01 -1.07685542e+00
1.37255773e-01 -5.74809134e-01 -5.40815294e-01 7.48861969e-01
6.42654896e-01 -7.48740375e-01 -4.63160068e-01 8.06993902e-01] | [9.213602066040039, -0.17378944158554077] |
fe4505c7-bf70-411b-9069-455f5cad8f50 | progressive-motion-context-refine-network-for | 2211.06024 | null | https://arxiv.org/abs/2211.06024v1 | https://arxiv.org/pdf/2211.06024v1.pdf | Progressive Motion Context Refine Network for Efficient Video Frame Interpolation | Recently, flow-based frame interpolation methods have achieved great success by first modeling optical flow between target and input frames, and then building synthesis network for target frame generation. However, above cascaded architecture can lead to large model size and inference delay, hindering them from mobile and real-time applications. To solve this problem, we propose a novel Progressive Motion Context Refine Network (PMCRNet) to predict motion fields and image context jointly for higher efficiency. Different from others that directly synthesize target frame from deep feature, we explore to simplify frame interpolation task by borrowing existing texture from adjacent input frames, which means that decoder in each pyramid level of our PMCRNet only needs to update easier intermediate optical flow, occlusion merge mask and image residual. Moreover, we introduce a new annealed multi-scale reconstruction loss to better guide the learning process of this efficient PMCRNet. Experiments on multiple benchmarks show that proposed approaches not only achieve favorable quantitative and qualitative results but also reduces current model size and running time significantly. | ['Jie Yang', 'Jinfeng Liu', 'Lingtong Kong'] | 2022-11-11 | null | null | null | null | ['video-frame-interpolation'] | ['computer-vision'] | [ 2.33850151e-01 -2.21072659e-01 -2.10401416e-01 -3.44699144e-01
-3.77923787e-01 -2.23191813e-01 3.35265785e-01 -4.61654812e-01
-3.89376760e-01 7.72753894e-01 2.50020653e-01 -2.28836745e-01
3.97128254e-01 -8.85705411e-01 -7.64377415e-01 -5.25115728e-01
2.26055726e-01 -2.64065564e-01 6.19985342e-01 9.55457613e-02
2.65023500e-01 2.45136559e-01 -1.26516974e+00 2.94228047e-01
1.01592100e+00 1.00602889e+00 6.00180924e-01 6.15369260e-01
-2.27072164e-01 1.49644172e+00 -3.05510014e-01 -2.41436794e-01
3.64365280e-01 -5.61737239e-01 -7.64623940e-01 1.52640402e-01
4.25247580e-01 -9.65758204e-01 -7.16689110e-01 8.82626057e-01
4.86719906e-01 3.25422764e-01 2.70895883e-02 -1.03792274e+00
-4.40238357e-01 4.92366821e-01 -7.91709661e-01 3.91750753e-01
2.74448246e-01 7.08185077e-01 6.59490287e-01 -9.04311061e-01
7.31098115e-01 1.43161225e+00 3.70516926e-01 5.62717557e-01
-9.27424371e-01 -7.82540202e-01 4.47897911e-01 5.40007651e-01
-1.17922926e+00 -5.66493571e-01 7.68054485e-01 -1.88850746e-01
6.53047144e-01 1.77658126e-01 6.92937851e-01 8.59675050e-01
9.45124328e-02 9.97659087e-01 7.80408323e-01 -2.53050588e-02
1.66562460e-02 -3.50276202e-01 -3.20209026e-01 8.29055846e-01
-1.24261782e-01 1.40125528e-01 -4.88577664e-01 2.48434201e-01
1.38001513e+00 -1.17703035e-01 -6.03310943e-01 2.64653116e-02
-1.30071974e+00 3.90980333e-01 6.02946699e-01 -4.07589413e-02
-3.09228063e-01 4.72296298e-01 2.98708171e-01 3.66120711e-02
2.88950861e-01 -6.56442419e-02 -3.53935480e-01 -1.37495622e-01
-1.07614636e+00 1.91852063e-01 3.52244020e-01 9.54969406e-01
1.08213103e+00 3.19201320e-01 -5.07986486e-01 6.47818923e-01
4.68708694e-01 3.30874443e-01 2.93496192e-01 -1.48891866e+00
7.26628065e-01 4.87661839e-01 3.14162731e-01 -1.17620265e+00
-1.01705365e-01 -2.65438795e-01 -1.08764958e+00 8.89659524e-02
4.20084208e-01 -1.53558537e-01 -8.76714110e-01 1.48214447e+00
5.24264395e-01 9.98124957e-01 -2.73842454e-01 1.17265284e+00
6.52001917e-01 1.04726171e+00 1.46387398e-01 -3.17794204e-01
1.09241879e+00 -1.53440285e+00 -5.99187016e-01 -3.38964015e-01
4.83163595e-01 -9.03906345e-01 1.03090799e+00 2.12383792e-01
-1.54220974e+00 -9.72810507e-01 -9.30926621e-01 -6.13395393e-01
4.31559950e-01 1.09206550e-01 7.16042042e-01 2.86877006e-01
-9.93817329e-01 6.10863805e-01 -1.04803407e+00 2.96683133e-01
7.90109158e-01 4.04301941e-01 2.73406487e-02 -4.73448038e-01
-1.09097123e+00 4.98377293e-01 3.32780302e-01 6.20836854e-01
-9.13029313e-01 -7.95809507e-01 -8.74360621e-01 1.40286073e-01
2.18280345e-01 -1.00237966e+00 1.02316868e+00 -1.10290194e+00
-1.83501780e+00 1.41258970e-01 -5.82544744e-01 -4.43753183e-01
6.18766010e-01 -3.35592806e-01 -1.91669822e-01 1.30520299e-01
-1.25989139e-01 1.08635366e+00 9.18021619e-01 -9.98236477e-01
-1.06796026e+00 1.95515588e-01 2.94518501e-01 2.28859171e-01
-1.92835584e-01 -3.88831496e-02 -9.41665649e-01 -7.24124551e-01
5.43418853e-03 -6.86623156e-01 -6.40445530e-01 3.16075236e-01
-5.67978770e-02 -6.72236532e-02 8.84450316e-01 -8.75056267e-01
1.54519832e+00 -1.93460536e+00 1.38268933e-01 -1.25897646e-01
2.97117919e-01 4.90290880e-01 -2.46759281e-01 -3.22245091e-01
2.63603032e-01 -9.63300392e-02 -2.74965286e-01 -4.43274468e-01
-3.75175476e-01 3.89023423e-01 -3.19123536e-01 2.58750945e-01
3.41948062e-01 9.45146084e-01 -1.02320564e+00 -7.07664847e-01
6.13145649e-01 7.02307522e-01 -1.12229788e+00 1.94862813e-01
-3.66158217e-01 9.18184459e-01 -4.85458255e-01 3.79476845e-01
9.43937421e-01 -5.03642499e-01 -1.65853910e-02 -4.41800714e-01
-1.20669447e-01 2.70741493e-01 -1.26506937e+00 1.92043662e+00
-7.71050870e-01 6.04076684e-01 -1.87868159e-02 -8.03849161e-01
6.87212408e-01 1.63663924e-01 4.87058163e-01 -6.77844644e-01
1.72576949e-01 1.35618066e-02 -2.86524557e-02 -3.68280530e-01
4.83112037e-01 3.85560691e-01 5.30513346e-01 1.32096097e-01
-2.09725201e-01 2.56308764e-01 2.15290084e-01 -1.09872110e-02
8.15071762e-01 6.23957455e-01 -1.08921960e-01 -4.68273787e-03
1.11658204e+00 -3.39062154e-01 1.12828696e+00 3.99000823e-01
-1.92549407e-01 7.84931004e-01 3.78503472e-01 -7.47197390e-01
-9.70355511e-01 -8.47006619e-01 2.16177598e-01 8.36648941e-01
5.75779021e-01 -3.69605929e-01 -5.78605056e-01 -4.76146400e-01
-5.37653148e-01 2.91873842e-01 -2.85611451e-01 1.39250234e-01
-1.29586399e+00 -6.18801296e-01 7.99243078e-02 5.10898948e-01
9.87246096e-01 -1.04575419e+00 -6.17049217e-01 6.26198113e-01
-6.69811189e-01 -1.41437185e+00 -8.82332981e-01 -6.21096373e-01
-1.00186872e+00 -8.31771791e-01 -7.80650675e-01 -9.03405845e-01
7.25231946e-01 3.24578732e-01 1.11432207e+00 4.90024537e-01
-2.32374500e-02 -3.35212797e-01 -1.95829630e-01 2.49476969e-01
-2.46822566e-01 1.67641968e-01 -5.07712841e-01 2.66906470e-01
-1.15619801e-01 -6.74133003e-01 -1.36010706e+00 4.75927442e-01
-9.87132668e-01 6.39756083e-01 3.59717667e-01 7.94950604e-01
5.93788147e-01 -1.90222368e-01 4.11304802e-01 -6.06686294e-01
2.02228725e-02 -2.48244449e-01 -8.37611139e-01 2.18126372e-01
-4.05781627e-01 5.40419184e-02 7.76465833e-01 -2.58515269e-01
-1.40781677e+00 2.43945017e-01 -3.17427874e-01 -6.22888148e-01
1.63877994e-01 3.34503353e-02 -2.41753384e-01 -9.39408690e-02
2.24397138e-01 3.41710657e-01 -2.26907670e-01 -1.93848059e-01
2.00325593e-01 3.12615484e-01 6.84700072e-01 -4.50829536e-01
6.84484065e-01 7.00879753e-01 1.14113197e-01 -4.12774354e-01
-7.03016460e-01 -2.14924499e-01 -4.85185981e-01 -2.47752726e-01
9.23115015e-01 -1.24233794e+00 -8.73265207e-01 5.26067197e-01
-1.56398630e+00 -5.43634236e-01 9.74869635e-03 4.88591492e-01
-3.95107239e-01 4.96275723e-01 -8.11928034e-01 -3.83340657e-01
-4.38162237e-01 -1.53876579e+00 1.05165410e+00 3.48017484e-01
1.88507721e-01 -9.40664291e-01 -2.60843605e-01 3.08122307e-01
4.39564168e-01 1.29576296e-01 4.59931403e-01 5.68513691e-01
-1.30343175e+00 3.21303636e-01 -6.43349290e-01 3.63249421e-01
2.44104907e-01 5.29936515e-02 -7.17418849e-01 -2.65130967e-01
3.33784074e-02 1.67279124e-01 9.20814514e-01 4.74218398e-01
1.49886751e+00 -3.58900815e-01 -2.00167552e-01 1.07909358e+00
1.46888733e+00 2.05331013e-01 1.01178420e+00 2.38110721e-01
1.15828860e+00 3.11851799e-01 6.03888452e-01 4.75200236e-01
6.70490503e-01 6.76932693e-01 2.88721263e-01 -1.90466702e-01
-5.63399136e-01 -2.28134573e-01 5.50316393e-01 8.32230568e-01
-2.77077556e-01 -2.91056842e-01 -5.83496571e-01 4.31161195e-01
-1.96100879e+00 -9.07999754e-01 -1.77263960e-01 1.96458328e+00
8.97751749e-01 1.96551442e-01 -1.42876565e-01 -1.48361117e-01
7.22096503e-01 3.65610331e-01 -6.34915352e-01 1.34796381e-01
-5.52951843e-02 3.00017506e-01 4.67382133e-01 9.43374336e-01
-8.78575742e-01 1.18603170e+00 4.92327166e+00 9.50016439e-01
-1.28100705e+00 1.81961432e-01 1.10707366e+00 -1.42344356e-01
-3.97172719e-01 3.39525312e-01 -8.26464534e-01 7.85790324e-01
5.34069121e-01 1.52073801e-01 5.10304868e-01 3.97197336e-01
7.42975354e-01 -1.30287081e-01 -8.91501427e-01 1.11969030e+00
-4.22933877e-01 -1.80530918e+00 2.01796994e-01 -1.19308569e-01
9.45292532e-01 -1.01224616e-01 -1.54085308e-01 4.25504036e-02
2.14317277e-01 -8.67881775e-01 5.78069270e-01 5.23387671e-01
5.72195292e-01 -6.83559358e-01 4.38356608e-01 9.89365503e-02
-1.49366772e+00 -9.00670961e-02 -5.47660887e-01 -1.84260771e-01
5.61468124e-01 6.07807934e-01 -2.27111101e-01 5.67393363e-01
5.76138794e-01 1.08672154e+00 -5.05413532e-01 1.06916535e+00
-2.05161050e-01 5.76362491e-01 -2.74407804e-01 4.19997692e-01
2.76284426e-01 -2.64809549e-01 2.42826834e-01 9.81220126e-01
3.39985609e-01 1.99904710e-01 3.02190542e-01 7.87573099e-01
-2.79896818e-02 -7.35431910e-02 -8.01867098e-02 5.93602061e-01
5.11225045e-01 1.18096852e+00 -7.43006468e-01 -5.05374134e-01
-5.90555549e-01 1.12725484e+00 1.66689575e-01 4.75713491e-01
-1.15067327e+00 -8.95766094e-02 6.76305771e-01 1.92569032e-01
3.63036513e-01 -4.00917053e-01 -2.43145674e-01 -1.46821105e+00
4.42058742e-02 -6.55281365e-01 5.57871610e-02 -7.38926172e-01
-6.40759468e-01 5.74608505e-01 -3.22172612e-01 -1.46512532e+00
-1.89651981e-01 -1.80073157e-01 -5.85784137e-01 7.96120226e-01
-1.91169024e+00 -1.03427565e+00 -5.29964685e-01 8.32083285e-01
8.94751847e-01 2.73173451e-01 7.09213093e-02 7.73645520e-01
-7.86881268e-01 5.83985209e-01 -2.25560069e-01 2.25075215e-01
6.74172759e-01 -7.26936042e-01 7.60144532e-01 1.11406064e+00
-1.33904606e-01 3.75069499e-01 3.23104441e-01 -5.08166671e-01
-1.30214155e+00 -1.45654178e+00 5.80100119e-01 -1.80927649e-01
3.24893147e-01 -1.01259306e-01 -8.20468545e-01 5.14415145e-01
2.77397752e-01 5.55194378e-01 -9.95211303e-02 -6.64819717e-01
3.16118151e-02 -3.76572102e-01 -8.62102687e-01 9.35762167e-01
1.30564535e+00 -1.28372565e-01 1.62433758e-01 2.24388525e-01
8.83952320e-01 -7.45358765e-01 -7.43928730e-01 4.63985860e-01
4.26530600e-01 -1.18190515e+00 1.17162561e+00 -6.93909377e-02
8.83951485e-01 -7.20786691e-01 1.72586650e-01 -7.67360270e-01
-2.95120806e-01 -9.61740255e-01 -4.07781541e-01 1.21721995e+00
-2.05483381e-02 -3.41595024e-01 9.67592835e-01 5.66995740e-01
-1.78889647e-01 -8.66959512e-01 -6.87916577e-01 -2.82412946e-01
-1.47845715e-01 -3.33159566e-01 7.71415055e-01 7.10816205e-01
-5.98928332e-01 3.41017634e-01 -7.36300528e-01 1.81776449e-01
5.59739828e-01 2.04363912e-01 8.62708509e-01 -6.58221722e-01
-4.10584360e-01 -3.73215556e-01 -2.36810535e-01 -1.90311348e+00
1.12149261e-01 -4.80193704e-01 9.34713557e-02 -1.44457793e+00
-1.72469258e-01 -6.82603002e-01 -7.30823874e-02 2.01047197e-01
-5.48795283e-01 4.09813195e-01 3.58738601e-01 2.39184603e-01
-5.86425126e-01 5.28719544e-01 1.93308902e+00 1.40970768e-02
-2.34991059e-01 -1.30225271e-01 -3.48637521e-01 7.69844413e-01
5.88825703e-01 -1.55194372e-01 -7.40466356e-01 -9.21482325e-01
5.15183508e-02 5.10748744e-01 5.44595778e-01 -1.20788431e+00
3.23958516e-01 -4.17090625e-01 6.03313982e-01 -4.72572565e-01
1.61242291e-01 -6.79130912e-01 1.23856820e-01 4.36653554e-01
-7.43405372e-02 1.07831448e-01 -1.38421068e-02 4.84388381e-01
-3.24675828e-01 -4.94869910e-02 8.64048600e-01 -1.26253843e-01
-9.88818347e-01 7.92939901e-01 -1.47572607e-01 -9.44588333e-02
7.47443855e-01 -3.55732232e-01 -1.26497984e-01 -3.58553559e-01
-5.12599468e-01 3.24843884e-01 3.58032525e-01 4.18513685e-01
6.44268930e-01 -1.31937885e+00 -7.29026020e-01 2.77334690e-01
-5.49737632e-01 6.12718105e-01 5.75393438e-01 8.37020934e-01
-9.48367476e-01 1.41077504e-01 -1.54942393e-01 -6.53421998e-01
-8.40678394e-01 4.64364469e-01 3.45543087e-01 -3.72555107e-01
-6.76444173e-01 9.08948004e-01 6.41713440e-01 9.34722051e-02
1.31077379e-01 -6.35586441e-01 -2.01388020e-02 -4.07525837e-01
6.35322809e-01 4.83814538e-01 -2.39364848e-01 -6.04951739e-01
-4.88942489e-02 6.15683913e-01 -1.36972561e-01 1.27638966e-01
9.93800938e-01 -6.12552881e-01 -1.74093679e-01 -2.31985033e-01
1.17779982e+00 -2.92681426e-01 -2.00610542e+00 -1.24544352e-01
-3.58139575e-01 -8.15779507e-01 1.68573141e-01 -2.33304650e-01
-1.55091441e+00 8.52259576e-01 6.09380424e-01 -3.88746023e-01
1.43961477e+00 -6.52381837e-01 1.45204663e+00 8.04982707e-02
2.59669811e-01 -8.04516613e-01 5.15951551e-02 5.09387434e-01
5.15595436e-01 -1.10493195e+00 4.64859344e-02 -6.77819610e-01
-3.78575951e-01 1.11419213e+00 9.61016417e-01 -3.02048117e-01
5.27944982e-01 1.39056280e-01 -3.56913134e-02 4.04937774e-01
-8.02790225e-01 9.99789536e-02 1.43317342e-01 4.34034258e-01
4.79839295e-01 -5.79623044e-01 -3.23023736e-01 6.97417706e-02
1.02131091e-01 5.00063896e-01 5.29761076e-01 6.93973422e-01
-2.45247811e-01 -1.29879296e+00 -1.70038953e-01 2.59957224e-01
-4.31673259e-01 -2.61292964e-01 5.62905967e-01 3.95590156e-01
2.27382511e-01 8.13144207e-01 1.50317982e-01 -1.88946873e-01
4.12038341e-02 -5.16958773e-01 5.60086250e-01 -2.05197796e-01
-4.46492791e-01 3.34970832e-01 -2.27284208e-01 -8.29471529e-01
-7.31294215e-01 -3.89412969e-01 -1.30162334e+00 -4.26253051e-01
-2.10358381e-01 -7.93325752e-02 3.73603851e-02 9.11748827e-01
5.31381130e-01 6.84981465e-01 8.16695392e-01 -1.07110500e+00
1.31099131e-02 -6.74937606e-01 3.87456492e-02 1.70464858e-01
5.23285985e-01 -3.31194133e-01 -6.28034212e-03 5.35250187e-01] | [10.72938346862793, -1.3926918506622314] |
6cd2e104-b1bb-4ab8-8d6d-52b483baf646 | towards-a-one-stop-solution-to-both-aspect | null | null | https://ieeexplore.ieee.org/document/8489042/authors#authors | https://ieeexplore.ieee.org/document/8489042/authors#authors | Towards a One-stop Solution to Both Aspect Extraction and Sentiment Analysis Tasks with Neural Multi-task Learning | Previous studies usually divided aspect-based sentiment analysis into several subtasks in pipeline, i.e., first aspect term and/or opinion term extraction, then aspect-based sentiment prediction, resulting in error propagation and external resources dependency. To overcome the problems mentioned above, in this work we present a novel one-stop solution on aspect-based sentiment analysis. Specifically, we propose a novel multi-task neural learning framework to jointly tackle aspect extraction and sentiment prediction subtasks at the same time, and leverage attention mechanisms to learn the joint representation of aspect-sentiment relationship. We have conducted extensive comparative experiments on two benchmark datasets from SemEval-2014. The experiment results demonstrate the effectiveness of our proposed solution. Especially, our multi-task model outperforms the state-of-the-art systems on aspect extraction subtask. | ['Wenting Wang', 'Man Lan', 'Feixiang Wang'] | 2018-10-14 | null | null | null | ieee-2018-10 | ['term-extraction', 'aspect-extraction', 'aspect-based-sentiment-analysis'] | ['natural-language-processing', 'natural-language-processing', 'natural-language-processing'] | [ 1.39312908e-01 -2.29065511e-02 -3.37180257e-01 -6.33837521e-01
-1.00130343e+00 -3.94976407e-01 5.24467647e-01 1.75166100e-01
-4.86435086e-01 3.75737548e-01 3.93029183e-01 -2.64091581e-01
2.31202766e-01 -7.12116957e-01 -5.24195850e-01 -4.37032193e-01
3.80943775e-01 3.01565379e-01 -5.06816916e-02 -1.50708050e-01
3.80835295e-01 -2.07236782e-01 -1.24235392e+00 4.83967900e-01
5.90574622e-01 1.11231637e+00 -2.12081522e-01 4.59416658e-01
-6.57766700e-01 8.17378819e-01 -7.00902224e-01 -8.09961438e-01
-1.32072970e-01 1.35774314e-02 -6.59389496e-01 2.69697487e-01
-2.48228312e-01 1.00134730e-01 3.54921252e-01 9.09829259e-01
4.75356013e-01 -4.20417301e-02 6.67110264e-01 -1.20376551e+00
-8.87061834e-01 7.65465081e-01 -1.12528455e+00 -5.88588156e-02
2.23494247e-01 -8.16632584e-02 1.43542182e+00 -1.09096527e+00
2.05581293e-01 1.04231524e+00 5.64170718e-01 1.95517212e-01
-7.36281633e-01 -6.64457560e-01 9.63706732e-01 9.77727026e-02
-9.73159313e-01 -2.73776859e-01 1.09053493e+00 -2.90784299e-01
1.54922318e+00 -4.27440591e-02 8.14315915e-01 8.54093254e-01
3.85457665e-01 1.28556895e+00 1.07690370e+00 -1.82758406e-01
5.10470662e-03 3.74158204e-01 5.99193633e-01 5.10351658e-01
4.31652963e-01 -4.27005291e-01 -6.60552680e-01 -1.12778842e-01
3.57565694e-02 2.80542802e-02 6.52751923e-02 1.74289849e-02
-9.56728518e-01 7.56582022e-01 1.33152619e-01 2.00930864e-01
-5.69323242e-01 7.60178315e-03 7.82688439e-01 3.87654960e-01
1.09497905e+00 3.75148743e-01 -1.31960905e+00 3.42210755e-03
-6.34252965e-01 1.19312607e-01 1.02936041e+00 1.04694307e+00
7.98613608e-01 9.18719321e-02 -2.98459440e-01 8.19576621e-01
6.95936203e-01 5.22830844e-01 5.09186983e-01 4.51006219e-02
7.12366164e-01 1.04964757e+00 -2.19279304e-01 -8.56296539e-01
-5.43139577e-01 -6.85502768e-01 -7.03285635e-01 -1.61504641e-01
-4.50624853e-01 -6.10649526e-01 -1.02562857e+00 1.30566895e+00
5.01735389e-01 3.48004047e-03 1.18869871e-01 5.90451062e-01
1.34860051e+00 6.02591574e-01 3.56094688e-01 -2.10262865e-01
1.83751833e+00 -1.66761315e+00 -8.00579607e-01 -5.44752955e-01
6.09414220e-01 -1.14017308e+00 1.06903505e+00 4.26855117e-01
-7.16244698e-01 -2.89875001e-01 -1.14999104e+00 2.32544281e-02
-7.24310935e-01 4.30225402e-01 1.17252862e+00 4.62256283e-01
-7.26050735e-01 4.25481610e-02 -8.48865688e-01 9.31683704e-02
4.85916585e-01 2.76942879e-01 -2.45245427e-01 3.09391856e-01
-1.03217208e+00 4.88327533e-01 2.17018444e-02 3.15615177e-01
-7.09659636e-01 -6.84517264e-01 -1.10837555e+00 1.26764610e-01
5.27529478e-01 -1.01900995e+00 1.41944325e+00 -1.14555717e+00
-1.34431732e+00 8.22001755e-01 -5.25088072e-01 5.92681915e-02
-2.62794435e-01 -6.73044324e-01 -5.41977048e-01 -5.20317197e-01
2.81530291e-01 -7.63248838e-03 8.42722476e-01 -1.26473618e+00
-7.65582263e-01 -5.57243764e-01 4.80190247e-01 3.53285909e-01
-5.92054367e-01 3.78955245e-01 -8.14909041e-01 -5.71749389e-01
-2.14773640e-01 -7.03505099e-01 -6.23833120e-01 -6.53923094e-01
-7.21931219e-01 -4.82640535e-01 4.74509150e-01 -4.66864169e-01
1.33632660e+00 -1.87526858e+00 -1.22789824e-02 -1.90770581e-01
1.53363094e-01 2.79524535e-01 -4.04372245e-01 2.52446115e-01
-2.56569862e-01 2.36078024e-01 -1.04380168e-01 -9.25565422e-01
7.95060694e-02 -3.07673603e-01 -3.62166941e-01 2.73763463e-02
4.46492374e-01 1.13411093e+00 -6.09678090e-01 -4.72353727e-01
-3.29681337e-01 6.60924792e-01 -4.63546813e-01 2.93245733e-01
-4.92762119e-01 1.34831965e-01 -9.98109579e-01 1.05578995e+00
7.20992148e-01 -3.73167098e-01 -2.15508014e-01 -3.18965405e-01
1.12597212e-01 7.83002138e-01 -8.88674617e-01 1.60193634e+00
-1.05275369e+00 3.57035071e-01 5.63681088e-02 -7.00181544e-01
8.14049661e-01 4.01930034e-01 4.76239592e-01 -5.70900083e-01
2.91514426e-01 -4.67495359e-02 -3.20749253e-01 -5.01125157e-01
7.57662773e-01 -3.53675693e-01 -3.49821895e-01 8.01845372e-01
2.70930380e-01 -1.26638576e-01 2.21135125e-01 -3.06066927e-02
7.72250056e-01 3.89061481e-01 5.25279522e-01 -2.81950478e-02
8.58288527e-01 4.43863653e-04 6.46298707e-01 2.22390458e-01
8.89260545e-02 4.76905674e-01 8.45801234e-01 -5.00390768e-01
-4.72710490e-01 -4.06516075e-01 1.71743214e-01 1.32530749e+00
-1.89159869e-03 -9.52846110e-01 -4.51952279e-01 -1.27508307e+00
-3.26118201e-01 5.53728580e-01 -8.16442966e-01 7.29107633e-02
-3.99346024e-01 -1.25212181e+00 8.45073387e-02 5.82063794e-01
3.58464003e-01 -1.38569152e+00 2.76281592e-02 2.03597650e-01
-5.93288802e-03 -1.22874987e+00 -4.30316687e-01 2.39628181e-01
-6.68614209e-01 -9.40630674e-01 -3.97400439e-01 -7.72831142e-01
6.48927987e-01 4.22349840e-01 1.71172655e+00 6.08079471e-02
1.41345739e-01 1.19432338e-01 -5.42832136e-01 -8.55227232e-01
4.15982902e-01 6.70144916e-01 -4.83219028e-01 5.23866192e-02
9.51027155e-01 -6.60930514e-01 -6.90507233e-01 -2.04203315e-02
-8.73654842e-01 -4.98851724e-02 1.04341948e+00 6.99263155e-01
9.60912108e-01 2.43160743e-02 5.61263382e-01 -1.53127408e+00
8.61621320e-01 -6.62449896e-01 -4.88235295e-01 1.75495312e-01
-9.58639443e-01 -1.31613269e-01 7.31235921e-01 -1.18831575e-01
-1.24710608e+00 -9.47696567e-02 -4.25558031e-01 1.05917364e-01
-6.50751516e-02 1.15602911e+00 -4.77431148e-01 3.58290434e-01
6.16314933e-02 4.21085835e-01 -6.69009924e-01 -5.24594367e-01
4.10483986e-01 8.23219359e-01 -2.64702380e-01 -3.74666721e-01
6.71035469e-01 4.37303007e-01 -2.86119908e-01 -4.92324293e-01
-1.60470712e+00 -6.49740458e-01 -3.79618257e-01 1.48688450e-01
8.18188310e-01 -1.46042049e+00 -4.69860822e-01 6.04773402e-01
-1.33654165e+00 2.60872900e-01 -1.37841165e-01 2.14800581e-01
-2.42625162e-01 8.45853090e-02 -6.20013177e-01 -7.48898804e-01
-1.19511938e+00 -1.22088325e+00 1.44663537e+00 3.96884292e-01
-7.23840594e-02 -1.10859942e+00 3.91238272e-01 6.32177770e-01
4.77016658e-01 -1.04741305e-01 7.56257355e-01 -9.99040246e-01
-5.81712842e-01 -4.70422029e-01 -2.55721837e-01 2.50763834e-01
2.90549129e-01 -2.77804583e-02 -1.12199306e+00 -5.01955347e-03
1.37181044e-01 -2.62330741e-01 1.10985136e+00 2.45802879e-01
8.61805260e-01 -2.08175585e-01 -2.23698661e-01 7.08019793e-01
1.43837500e+00 -5.62354065e-02 2.67536044e-01 7.68038452e-01
9.67733622e-01 4.95721281e-01 9.69702065e-01 3.15362483e-01
9.82255220e-01 3.15431416e-01 1.98499829e-01 -1.89148471e-01
1.05776049e-01 -4.25291173e-02 5.41375160e-01 1.40889275e+00
1.33844912e-01 -2.01744214e-01 -4.86107022e-01 8.85481417e-01
-1.84184182e+00 -2.20678777e-01 -1.23339944e-01 1.54128683e+00
7.64338911e-01 4.70570773e-01 -1.57169610e-01 -2.16557071e-01
1.37590379e-01 7.95635760e-01 -5.04965842e-01 -7.33279943e-01
-1.29873246e-01 2.30935201e-01 4.15986218e-02 2.89541483e-01
-1.48407936e+00 1.15336788e+00 5.30264521e+00 7.23454118e-01
-9.31420803e-01 3.52436244e-01 7.04842746e-01 -2.57726192e-01
-8.14265251e-01 1.45500228e-01 -1.01275265e+00 1.23387046e-01
6.54694796e-01 -3.02574486e-01 -2.77757701e-02 1.14994788e+00
-1.27588034e-01 1.16974749e-01 -7.81805098e-01 6.11695647e-01
4.02886659e-01 -8.68368924e-01 2.75471121e-01 -3.22601944e-01
8.97423863e-01 2.44192734e-01 6.45702472e-03 7.44281948e-01
1.91331029e-01 -8.02754641e-01 2.84769624e-01 2.56593108e-01
3.42908412e-01 -9.88099635e-01 1.21920872e+00 -3.83545980e-02
-1.71916735e+00 3.59477043e-01 -1.76952556e-01 4.38817544e-03
3.87028039e-01 1.14026260e+00 -2.27078378e-01 8.46665621e-01
6.58528209e-01 1.19470155e+00 -5.05307615e-01 7.75817871e-01
-7.55766153e-01 5.15933037e-01 1.46919593e-01 -2.73152709e-01
3.24769586e-01 -2.58796066e-01 5.34026861e-01 1.27973843e+00
6.53644130e-02 -2.18344733e-01 1.15164012e-01 5.63000262e-01
-6.88369647e-02 5.72245717e-01 -4.53128368e-01 -3.56589049e-01
-2.15612307e-01 1.91130829e+00 -5.45567214e-01 -4.85772789e-01
-1.00189388e+00 6.60274327e-01 4.14380729e-01 4.29699838e-01
-6.96667314e-01 -5.37619412e-01 8.72127056e-01 -4.74833190e-01
6.84173107e-01 -2.79057585e-02 -7.56706893e-01 -1.63564825e+00
4.74110723e-01 -1.00256288e+00 3.96855891e-01 -5.52690387e-01
-1.23096871e+00 1.14215100e+00 -4.90666747e-01 -1.15677619e+00
-1.88222185e-01 -5.31291962e-01 -9.31739151e-01 9.16687548e-01
-2.17606950e+00 -1.73714018e+00 2.88459957e-02 3.73072088e-01
8.52214158e-01 -2.81259686e-01 8.62121046e-01 3.65088135e-01
-7.51663923e-01 6.01664126e-01 -4.50096965e-01 1.37822151e-01
6.01905584e-01 -1.30027103e+00 8.62215698e-01 7.49526083e-01
1.63087785e-01 8.22242856e-01 3.48736882e-01 -6.33985341e-01
-1.52863240e+00 -1.11430919e+00 1.20197713e+00 -5.25048673e-01
9.44716871e-01 -3.51358414e-01 -5.98748446e-01 8.90264630e-01
5.71083188e-01 -4.80436862e-01 1.06974578e+00 8.02114725e-01
-5.83880424e-01 -2.37156227e-01 -5.48807383e-01 6.02969348e-01
5.98857105e-01 -5.30566394e-01 -5.71926713e-01 2.47821644e-01
1.29954982e+00 -2.00838804e-01 -8.07255089e-01 6.98779941e-01
5.78193426e-01 -6.93030298e-01 7.16311991e-01 -7.67782152e-01
9.62565899e-01 -3.35284710e-01 2.39468645e-02 -1.47218919e+00
-8.52993652e-02 -3.29689264e-01 -3.06396067e-01 1.69239485e+00
1.13551569e+00 -4.92367327e-01 7.68510103e-01 4.21867102e-01
4.94144410e-02 -1.41725659e+00 -4.54244643e-01 7.12326402e-03
-1.21860981e-01 -7.03466654e-01 8.26194525e-01 6.20055735e-01
-1.98786065e-01 1.23445857e+00 -3.64848524e-01 6.94273189e-02
3.11575681e-01 8.33568633e-01 8.45628202e-01 -7.88979828e-01
-6.08170331e-01 -5.29677272e-01 6.08262199e-04 -9.80239391e-01
3.84999663e-01 -6.19410098e-01 -4.20414060e-02 -1.74698949e+00
6.06156051e-01 -6.86405525e-02 -6.73915207e-01 5.16204178e-01
-8.95904005e-01 3.08809765e-02 -1.86148509e-02 -2.26152524e-01
-1.07800424e+00 9.75865006e-01 1.29813516e+00 -4.80463028e-01
-1.02023438e-01 3.79865736e-01 -1.55792129e+00 8.13367188e-01
7.09409952e-01 -7.22389519e-01 -4.26201224e-01 -7.54654884e-01
9.46967006e-01 -4.25141484e-01 -4.50003982e-01 -4.13299292e-01
1.23791553e-01 8.39274451e-02 1.60841152e-01 -8.49830151e-01
3.68781686e-01 -8.69809985e-01 -4.99228597e-01 -8.19471665e-03
-1.08293362e-01 2.51736343e-01 3.86448830e-01 5.09218693e-01
-6.46087170e-01 -9.62038636e-02 7.25713596e-02 -9.62319076e-02
-4.98852611e-01 5.02180994e-01 -2.34054610e-01 1.05465531e-01
5.91188788e-01 5.91860473e-01 -3.96110475e-01 -2.31497392e-01
-4.71305013e-01 4.42860395e-01 1.97095927e-02 7.20335841e-01
5.44591784e-01 -1.18961644e+00 -6.79754376e-01 2.27326497e-01
5.06549835e-01 1.96641088e-01 2.99209446e-01 8.54516864e-01
4.87241298e-02 4.99067694e-01 3.75424087e-01 -2.41494820e-01
-1.27053344e+00 5.31313539e-01 8.08222741e-02 -1.01151979e+00
-1.06328681e-01 1.00692952e+00 3.62309068e-01 -1.01978457e+00
6.54401258e-03 -1.93304673e-01 -8.71880054e-01 3.88164580e-01
6.20242834e-01 -3.36193204e-01 3.16790223e-01 -5.30284107e-01
-5.03356755e-01 1.00986123e+00 -5.94692051e-01 1.32306432e-02
1.58123744e+00 -2.13988155e-01 -5.75770378e-01 5.24483383e-01
1.20781803e+00 1.07278176e-01 -5.87674499e-01 -3.80744100e-01
4.67726476e-02 -1.64084151e-01 1.04607441e-01 -8.74711156e-01
-1.50993013e+00 7.99655735e-01 3.80135141e-02 2.26613373e-01
1.45555949e+00 1.82562210e-02 1.07955205e+00 4.83337998e-01
5.77052794e-02 -9.89710808e-01 -1.51932448e-01 7.68181324e-01
7.01151848e-01 -1.53303576e+00 3.87430161e-01 -4.51001704e-01
-9.65569198e-01 6.50028229e-01 8.99133801e-01 -1.96001679e-01
1.16753566e+00 4.71485019e-01 4.39993382e-01 -5.30476034e-01
-1.41434276e+00 -3.95379156e-01 5.03291607e-01 2.85634607e-01
9.95175958e-01 1.10839970e-01 -5.40087581e-01 1.53810418e+00
-2.19046995e-01 -1.06351957e-01 2.04793319e-01 8.99364889e-01
-6.01178296e-02 -1.19687915e+00 2.21031919e-01 6.16915524e-01
-9.33557212e-01 -6.96171820e-01 -5.26023507e-01 4.29545611e-01
-1.91907004e-01 1.12017787e+00 -3.44938397e-01 -5.53498149e-01
6.40692592e-01 -7.31206313e-02 -1.49278432e-01 -7.21617937e-01
-1.18712354e+00 4.05453801e-01 4.83482629e-01 -6.81653321e-01
-6.06675088e-01 -4.11019415e-01 -8.91000450e-01 2.18491450e-01
-5.30746937e-01 2.76841223e-01 1.03262460e+00 1.23322690e+00
6.43422127e-01 1.06385672e+00 8.64193678e-01 -4.85693961e-01
-1.01358175e-01 -1.23023748e+00 -4.05409992e-01 1.93448365e-01
2.74347186e-01 -2.47718439e-01 -3.43642324e-01 -2.04381496e-01] | [11.474335670471191, 6.646939754486084] |
81a2bd82-aec9-4ed5-a06d-8462367c98dd | benchmarking-robustness-of-3d-object | 2303.11040 | null | https://arxiv.org/abs/2303.11040v1 | https://arxiv.org/pdf/2303.11040v1.pdf | Benchmarking Robustness of 3D Object Detection to Common Corruptions in Autonomous Driving | 3D object detection is an important task in autonomous driving to perceive the surroundings. Despite the excellent performance, the existing 3D detectors lack the robustness to real-world corruptions caused by adverse weathers, sensor noises, etc., provoking concerns about the safety and reliability of autonomous driving systems. To comprehensively and rigorously benchmark the corruption robustness of 3D detectors, in this paper we design 27 types of common corruptions for both LiDAR and camera inputs considering real-world driving scenarios. By synthesizing these corruptions on public datasets, we establish three corruption robustness benchmarks -- KITTI-C, nuScenes-C, and Waymo-C. Then, we conduct large-scale experiments on 24 diverse 3D object detection models to evaluate their corruption robustness. Based on the evaluation results, we draw several important findings, including: 1) motion-level corruptions are the most threatening ones that lead to significant performance drop of all models; 2) LiDAR-camera fusion models demonstrate better robustness; 3) camera-only models are extremely vulnerable to image corruptions, showing the indispensability of LiDAR point clouds. We release the benchmarks and codes at https://github.com/kkkcx/3D_Corruptions_AD. We hope that our benchmarks and findings can provide insights for future research on developing robust 3D object detection models. | ['Jun Zhu', 'Xingxing Wei', 'Hang Su', 'Xiao Yang', 'Yikai Wang', 'Zijian Zhu', 'Jinlai Zhang', 'Caixin Kang', 'Yinpeng Dong'] | 2023-03-20 | null | null | null | null | ['robust-3d-object-detection'] | ['computer-vision'] | [-3.85650992e-01 -6.18913889e-01 -2.26453152e-02 -3.50070179e-01
-6.48956835e-01 -6.70433700e-01 7.38939285e-01 -1.34469151e-01
-3.45956296e-01 2.62529552e-01 -2.46030688e-01 -4.96816337e-01
2.76582986e-01 -7.50225306e-01 -8.51657987e-01 -6.23044550e-01
-8.10501948e-02 -3.69419977e-02 7.28558481e-01 -3.50576371e-01
2.37982303e-01 9.59681571e-01 -2.13530087e+00 -1.82294980e-01
8.62300396e-01 1.03525233e+00 -6.72215447e-02 6.11106396e-01
2.76073545e-01 4.25274789e-01 -6.86051190e-01 -2.11989954e-01
6.69077516e-01 1.76684931e-01 8.52951929e-02 -6.61794692e-02
5.94558656e-01 -5.11711001e-01 -5.13300836e-01 1.12032044e+00
6.04191959e-01 -1.13505960e-01 6.49433136e-01 -1.87887669e+00
-3.56579155e-01 -3.66260231e-01 -6.47517622e-01 3.75499249e-01
2.53805667e-01 9.64153230e-01 3.70536625e-01 -9.44613576e-01
3.74337912e-01 1.48719645e+00 7.70021081e-01 2.54145354e-01
-7.78287172e-01 -1.08844531e+00 -1.45130055e-02 3.46648127e-01
-1.55963087e+00 -5.58726549e-01 6.19506478e-01 -4.46110874e-01
1.17064214e+00 3.78730804e-01 4.26943541e-01 1.21823668e+00
5.98519504e-01 5.89305401e-01 1.19264352e+00 1.29081190e-01
1.56457692e-01 1.99230209e-01 3.88942927e-01 5.25472343e-01
9.23014998e-01 8.82629633e-01 -5.83405316e-01 8.02858323e-02
9.19538960e-02 -2.63743490e-01 1.92790806e-01 -2.90760458e-01
-1.12817395e+00 6.02756023e-01 6.22413695e-01 -3.70496184e-01
5.93734309e-02 2.66250789e-01 4.06483084e-01 2.57149100e-01
3.01510692e-01 -6.63039610e-02 -1.97576266e-02 -8.65247473e-02
-3.39575917e-01 5.05273283e-01 3.99177879e-01 1.42152286e+00
9.42732930e-01 5.01472317e-02 1.66772995e-02 3.23133856e-01
4.43472981e-01 1.32203412e+00 -1.76973745e-01 -1.03360236e+00
6.39787436e-01 5.59543848e-01 3.35190982e-01 -1.12427652e+00
-4.38784212e-01 -1.45680010e-01 -7.39221871e-01 7.08573282e-01
7.05010369e-02 5.34995943e-02 -9.64433014e-01 1.26240170e+00
4.52841133e-01 2.49905169e-01 1.05254792e-01 1.02932787e+00
1.11465156e+00 4.87590551e-01 -8.36725160e-03 1.36047944e-01
1.23162210e+00 -4.12054509e-01 -6.37900174e-01 -4.74533856e-01
6.74761236e-01 -8.09185922e-01 9.15852070e-01 3.59245129e-02
-5.83423018e-01 -8.40800285e-01 -1.30864632e+00 -6.37013242e-02
-6.07750356e-01 -1.95014387e-01 3.06337029e-01 8.64674449e-01
-6.95901394e-01 -2.39850506e-01 -7.83918619e-01 -3.36625993e-01
4.59859967e-01 -5.47674811e-03 -3.18013281e-01 -3.62950653e-01
-1.05047524e+00 1.37652171e+00 1.53768331e-01 1.60121098e-01
-1.16738200e+00 -6.03365064e-01 -8.39777827e-01 -5.47530413e-01
3.58575881e-01 -6.20272160e-01 9.82003033e-01 3.21268737e-02
-8.27885449e-01 9.63407874e-01 -3.67890328e-01 -4.38886583e-01
6.96112752e-01 -3.38453352e-01 -5.75463831e-01 -2.29797587e-01
2.82705665e-01 7.96063542e-01 7.93013334e-01 -1.66714525e+00
-8.41955364e-01 -4.21219051e-01 -3.62706184e-02 1.21523678e-01
2.42792711e-01 -4.28040661e-02 -5.82045019e-01 -1.28942981e-01
1.13308817e-01 -1.04867160e+00 -2.14671284e-01 1.75995335e-01
-5.34940958e-01 -2.06077322e-01 1.19521415e+00 5.34030683e-02
7.48234332e-01 -2.58394313e+00 -6.22762978e-01 5.38547523e-02
2.31638044e-01 4.10089672e-01 -2.24537149e-01 2.01585472e-01
3.01230490e-01 2.74485916e-01 -1.12164393e-01 -3.06295484e-01
1.43423632e-01 4.28220361e-01 -4.66982067e-01 7.41984606e-01
5.72896302e-01 9.67083693e-01 -7.01895058e-01 -4.73099023e-01
7.78022468e-01 3.99947703e-01 -3.25934321e-01 4.43221964e-02
1.78252384e-01 2.00461894e-01 -5.18998325e-01 9.82129633e-01
1.20590591e+00 4.11293685e-01 -6.61852956e-01 -1.67879879e-01
-3.01490426e-01 1.92043483e-01 -1.24770272e+00 9.14370477e-01
-6.98038265e-02 7.73271322e-01 1.73140708e-02 -4.37834620e-01
1.04738760e+00 -1.86578006e-01 1.90477312e-01 -9.87334847e-01
1.73071951e-01 2.76889354e-01 -1.54049695e-01 -5.22768080e-01
7.07748592e-01 1.28698140e-01 -3.38997424e-01 1.97471548e-02
-5.10024905e-01 -8.85304153e-01 7.48081133e-02 2.87152499e-01
1.18082941e+00 -2.80395508e-01 -2.21291166e-02 -1.34002775e-01
2.66237825e-01 3.04658145e-01 5.43615997e-01 9.80617285e-01
-9.66224253e-01 5.34285545e-01 2.72565931e-01 -3.65661174e-01
-8.39432001e-01 -1.39403200e+00 -3.98826778e-01 1.80416226e-01
7.46652961e-01 -2.77864248e-01 -2.70159364e-01 -5.10226369e-01
7.30958581e-01 7.71573722e-01 -3.87931854e-01 -4.79921728e-01
-2.30549097e-01 -7.00787723e-01 8.79203141e-01 4.27567959e-01
7.71722734e-01 -5.48116744e-01 -8.66063356e-01 -2.52778143e-01
-1.77327231e-01 -1.65537786e+00 3.99892330e-02 1.05727755e-01
-6.99269295e-01 -1.16673231e+00 8.17176253e-02 -2.98890263e-01
4.56443340e-01 1.19340003e+00 1.03071332e+00 2.84023792e-01
-3.48954707e-01 4.00819898e-01 -2.85534918e-01 -1.07179821e+00
-5.07550299e-01 -4.19758409e-01 5.54074109e-01 -4.77647692e-01
6.65912747e-01 -3.35800909e-02 -4.07556117e-01 8.47517610e-01
-7.35079050e-01 -3.23946178e-01 4.01095659e-01 1.57877550e-01
5.58107972e-01 2.75013149e-01 5.50774410e-02 -1.75706327e-01
3.87581229e-01 -5.07392168e-01 -1.10563231e+00 -4.69420463e-01
-4.87719506e-01 -3.00616235e-01 8.72932933e-03 -8.65553319e-02
-6.83208942e-01 1.33101240e-01 -3.33502367e-02 -5.99995673e-01
-5.08251905e-01 2.13983469e-02 -2.10830271e-01 -2.55369902e-01
8.96939516e-01 9.47131366e-02 -1.84475724e-02 -1.79426491e-01
2.73260295e-01 7.20908821e-01 5.79640687e-01 -2.94290423e-01
1.58064508e+00 9.53722000e-01 1.25966325e-01 -1.16171193e+00
-4.52165455e-01 -5.87567985e-01 -4.89376664e-01 -5.35165846e-01
6.96509361e-01 -1.36896479e+00 -6.52279317e-01 7.78696597e-01
-1.21794522e+00 -1.16173416e-01 -2.07731221e-02 4.32870448e-01
-3.34686249e-01 4.54026490e-01 -1.27777770e-01 -9.72826123e-01
3.58523846e-01 -1.53531468e+00 1.27773595e+00 5.55541068e-02
6.45961761e-02 -2.86981136e-01 -2.58123297e-02 5.54037094e-01
2.48973966e-01 3.79358053e-01 3.86371255e-01 1.46413445e-01
-9.47546721e-01 -4.64596987e-01 -5.44054031e-01 4.29459900e-01
-1.17311683e-02 4.51086938e-01 -1.26262558e+00 -2.55575359e-01
-1.02693960e-01 -1.92364469e-01 1.00091064e+00 1.77811131e-01
8.86288643e-01 3.17347914e-01 -4.65450555e-01 6.59829617e-01
1.21580899e+00 -7.81150982e-02 7.59573996e-01 3.57485414e-01
5.44958651e-01 3.56077462e-01 9.96827900e-01 2.37586558e-01
4.77930218e-01 5.11590958e-01 1.09503090e+00 -1.23770714e-01
-2.97392607e-01 -1.83706179e-01 6.79101586e-01 4.92453724e-01
-2.61199325e-02 -2.63151437e-01 -1.05347621e+00 5.24725258e-01
-1.55178344e+00 -7.98981071e-01 -8.12407911e-01 2.13484645e+00
2.17360929e-01 5.97422302e-01 -2.68672891e-02 2.67718285e-01
5.27707577e-01 2.95486242e-01 -6.44018292e-01 -1.51785791e-01
-3.43788683e-01 -3.94912243e-01 1.07549310e+00 4.15687829e-01
-1.14370632e+00 8.38474512e-01 6.57305622e+00 4.84554708e-01
-9.49287415e-01 5.55907004e-02 2.87441522e-01 -2.98225105e-01
-5.67321777e-02 -9.29623097e-02 -1.30766952e+00 5.79693675e-01
9.41376925e-01 -1.12727351e-01 -7.30772838e-02 7.22620785e-01
6.17527127e-01 -4.76224959e-01 -7.41214991e-01 8.79009962e-01
-1.12707265e-01 -1.05747414e+00 -3.95580865e-02 2.47178853e-01
5.95062256e-01 7.35042334e-01 1.11288689e-01 2.99904794e-01
5.58754325e-01 -7.61845291e-01 9.33222294e-01 1.63882196e-01
6.42859995e-01 -6.23412907e-01 8.13784778e-01 5.05942464e-01
-1.16674459e+00 -9.18104351e-02 -6.47733867e-01 -1.94525555e-01
1.22573987e-01 8.71194363e-01 -6.04103982e-01 4.85936701e-01
1.19026196e+00 8.08295190e-01 -8.84599328e-01 1.09812891e+00
-3.93998146e-01 5.35843432e-01 -7.17123032e-01 -2.77234744e-02
2.86810964e-01 8.71181339e-02 9.86830652e-01 1.15852427e+00
3.15229416e-01 5.43523543e-02 -8.14686716e-02 9.09643471e-01
1.04801841e-01 -5.56527674e-01 -1.36265838e+00 4.64061171e-01
6.99728668e-01 1.01933742e+00 -3.85559201e-01 -1.65838569e-01
-5.86765766e-01 3.20006609e-01 -1.26618475e-01 4.63053763e-01
-1.24036515e+00 -1.21268108e-01 1.46588099e+00 1.49769604e-01
1.89852402e-01 -7.86423028e-01 -6.16219223e-01 -8.73847663e-01
2.48867214e-01 -6.40428424e-01 -1.90311447e-02 -9.56140220e-01
-1.14451778e+00 1.27215862e-01 2.51731902e-01 -1.66973269e+00
3.41311693e-01 -8.22326362e-01 -4.72898155e-01 5.84367633e-01
-2.08172035e+00 -8.45036924e-01 -8.49189222e-01 5.38669586e-01
2.78722554e-01 2.36380585e-02 3.03251773e-01 3.72758090e-01
-6.77347422e-01 3.04972589e-01 -2.22324595e-01 -3.42181548e-02
7.79431105e-01 -6.71040297e-01 8.24047446e-01 1.15056157e+00
-1.82894170e-01 1.82877794e-01 7.73619533e-01 -7.53203928e-01
-1.81751978e+00 -1.45401335e+00 6.73720717e-01 -1.13707864e+00
6.19822264e-01 -4.72796202e-01 -7.36011147e-01 4.37400877e-01
-1.28437489e-01 3.17140639e-01 1.21371798e-01 -3.06272358e-01
-4.76920068e-01 -3.15879852e-01 -1.08359563e+00 5.63025415e-01
1.21881127e+00 -3.41743886e-01 -4.13804531e-01 2.47427166e-01
8.06561351e-01 -5.66183805e-01 -5.33866882e-01 7.42805481e-01
3.04206163e-01 -1.31316829e+00 1.14179933e+00 -1.66675791e-01
5.23522533e-02 -8.35894942e-01 -5.18630326e-01 -1.02627826e+00
-1.78316250e-01 -1.78731933e-01 3.26782465e-02 9.29724813e-01
2.42197514e-01 -1.03519654e+00 5.08748353e-01 4.51526821e-01
-4.21492606e-01 -4.36714441e-02 -1.34882915e+00 -1.29276121e+00
3.26108448e-02 -1.10333037e+00 7.90939629e-01 6.31179214e-01
-5.99488497e-01 -4.83478829e-02 -7.13011697e-02 8.72149944e-01
8.71001363e-01 -2.84212887e-01 1.43738008e+00 -1.00121582e+00
5.78915179e-01 -4.63122725e-01 -8.64654124e-01 -8.05990815e-01
-8.41248557e-02 -5.39058328e-01 3.60421896e-01 -1.21198559e+00
-4.47528996e-02 -6.67900145e-01 -1.50684103e-01 4.12032425e-01
-2.59785414e-01 5.98768294e-01 2.59200335e-01 3.22173417e-01
-5.14468133e-01 5.94201744e-01 1.05577207e+00 -4.03123349e-01
1.05289310e-01 1.51868016e-01 -4.04125750e-01 5.89918792e-01
7.97446907e-01 -5.70842266e-01 -1.66243076e-01 -6.36548877e-01
1.57759085e-01 -4.79163557e-01 9.79990184e-01 -1.43968594e+00
1.19838238e-01 -3.24142247e-01 1.67900190e-01 -1.16724539e+00
4.83832806e-01 -9.76573229e-01 1.04570051e-03 7.65293777e-01
5.68451345e-01 1.12251297e-01 7.27202773e-01 4.83462304e-01
-1.62722796e-01 2.24619031e-01 8.73912871e-01 1.42848402e-01
-1.08897638e+00 3.09864938e-01 -6.31654918e-01 1.06953504e-02
1.29182112e+00 -6.35053694e-01 -7.29327440e-01 -2.26273954e-01
2.12433130e-01 5.21346688e-01 7.52496064e-01 9.32546735e-01
7.13986754e-01 -1.34354031e+00 -8.71896923e-01 5.56680024e-01
7.89894342e-01 2.67405063e-01 2.00823899e-02 7.32777596e-01
-5.42988777e-01 4.39797223e-01 -1.51891112e-01 -1.28733444e+00
-1.20814002e+00 4.56721634e-01 3.66725266e-01 4.58825976e-01
-4.41438079e-01 5.85579574e-01 7.76766911e-02 -4.79378104e-01
2.28658006e-01 -7.85784483e-01 3.82855147e-01 -2.00917423e-01
2.88663805e-01 5.10950029e-01 3.78522217e-01 -9.98844504e-01
-7.84688532e-01 7.25347459e-01 4.10999298e-01 2.95961469e-01
7.99153745e-01 -1.64897427e-01 3.74760658e-01 2.26228550e-01
8.99337709e-01 -1.06697604e-01 -1.32848334e+00 7.16843084e-02
-3.38117778e-01 -7.28564322e-01 1.85524374e-01 -4.83184278e-01
-9.38044071e-01 9.15229440e-01 8.57833922e-01 1.34852692e-01
8.02155316e-01 -2.13322658e-02 5.70200026e-01 3.91874969e-01
5.18573046e-01 -9.01078224e-01 -1.76889569e-01 7.82907903e-01
8.19059610e-01 -1.63409269e+00 2.18639493e-01 -5.04555047e-01
-4.44211334e-01 4.22595650e-01 7.81583607e-01 -1.69702470e-02
6.74777210e-01 5.15456855e-01 4.86008555e-01 -3.15518469e-01
-6.60464287e-01 -4.98523176e-01 4.25774455e-02 8.67490351e-01
-3.29239041e-01 1.94907457e-01 2.77604908e-01 4.86598387e-02
-2.13221222e-01 -1.76332295e-01 4.25550610e-01 9.60476041e-01
-5.66292226e-01 -6.20977461e-01 -8.51355672e-01 3.94207805e-01
3.18110466e-01 3.17531615e-01 -5.76816797e-01 1.06560159e+00
2.67982870e-01 1.37042713e+00 3.35574299e-02 -6.53585374e-01
9.53295648e-01 -3.22817206e-01 1.71553567e-01 -3.13010991e-01
-1.40268028e-01 -4.68481183e-01 1.17267184e-01 -9.63975370e-01
-2.18363136e-01 -7.72605658e-01 -1.01833987e+00 -7.08475709e-01
-5.05961120e-01 -4.34508771e-01 8.87608230e-01 7.15061247e-01
5.92270434e-01 3.49956781e-01 6.78749084e-01 -1.07139194e+00
-5.32260060e-01 -6.97646081e-01 -3.65514576e-01 3.62988830e-01
5.34468234e-01 -1.16965067e+00 -7.99897373e-01 -2.74912596e-01] | [7.923798084259033, -1.9954513311386108] |
6813b39f-6913-4943-ac51-cee8f630dd46 | traffic-surveillance-using-vehicle-license | 2012.02218 | null | https://arxiv.org/abs/2012.02218v1 | https://arxiv.org/pdf/2012.02218v1.pdf | Traffic Surveillance using Vehicle License Plate Detection and Recognition in Bangladesh | Computer vision coupled with Deep Learning (DL) techniques bring out a substantial prospect in the field of traffic control, monitoring and law enforcing activities. This paper presents a YOLOv4 object detection model in which the Convolutional Neural Network (CNN) is trained and tuned for detecting the license plate of the vehicles of Bangladesh and recognizing characters using tesseract from the detected license plates. Here we also present a Graphical User Interface (GUI) based on Tkinter, a python package. The license plate detection model is trained with mean average precision (mAP) of 90.50% and performed in a single TESLA T4 GPU with an average of 14 frames per second (fps) on real time video footage. | ['Raiyan Ibne Hafiz', 'Mahmudul Haque', 'Muhaiminul Islam Akash', 'Md. Saif Hassan Onim'] | 2020-12-03 | null | null | null | null | ['license-plate-detection'] | ['computer-vision'] | [-2.23544151e-01 -6.70928478e-01 1.08294569e-01 8.65629017e-02
-5.23616195e-01 -7.21086323e-01 5.63363910e-01 -2.88579524e-01
-6.04907393e-01 3.46412539e-01 -5.91041028e-01 -5.79240501e-01
5.56385577e-01 -9.96312976e-01 -9.36359704e-01 -5.59229553e-01
4.54820730e-02 2.95913368e-01 8.07036579e-01 -7.13383332e-02
4.78267550e-01 8.78540695e-01 -1.30368316e+00 3.63891602e-01
4.27748322e-01 9.66019511e-01 4.23722789e-02 1.14141512e+00
2.65343726e-01 9.62292373e-01 -6.29369736e-01 -8.72121513e-01
7.81439722e-01 4.59542155e-01 -6.36256039e-02 1.47108629e-01
8.32203507e-01 -7.76700258e-01 -8.26346278e-01 1.05598474e+00
2.51347363e-01 -1.74592603e-02 6.95405364e-01 -1.13115036e+00
-3.09996933e-01 -6.01546586e-01 -7.67448664e-01 8.30280066e-01
-5.34925796e-02 5.52606940e-01 1.65634081e-01 -1.15373695e+00
4.07987773e-01 9.16280091e-01 1.10107470e+00 1.38445720e-01
-5.36200941e-01 -9.47304606e-01 -8.87156904e-01 4.58781600e-01
-1.71851945e+00 -4.83645111e-01 3.55274707e-01 -8.51605117e-01
1.10265899e+00 6.49355054e-02 5.49517155e-01 4.47789252e-01
5.89949608e-01 6.91909611e-01 8.25553656e-01 -2.98659354e-01
2.50787102e-02 3.65221441e-01 2.44720042e-01 9.56387818e-01
5.20465374e-01 6.94767460e-02 -7.36599267e-02 3.09923023e-01
1.17498767e+00 -4.35875803e-02 5.54529190e-01 3.41631979e-01
-5.91232836e-01 9.82867360e-01 1.21635519e-01 -1.00209489e-01
-3.57747942e-01 2.47285202e-01 6.56759977e-01 -2.90012695e-02
4.00890023e-01 -1.84699818e-01 -1.21920399e-01 -7.32287988e-02
-9.69364583e-01 3.42049718e-01 3.09924722e-01 9.80152786e-01
6.30287766e-01 4.88298953e-01 -1.53439894e-01 5.92826724e-01
2.74690837e-01 7.57759213e-01 -1.95135653e-01 -7.83127904e-01
5.67595065e-01 5.90489149e-01 9.93854180e-02 -1.22102523e+00
-1.87281325e-01 -5.86005785e-02 -4.20305789e-01 7.14317977e-01
4.56199855e-01 -5.40239453e-01 -8.33630204e-01 4.11919892e-01
1.37644947e-01 3.96283865e-01 -2.97160089e-01 8.33598673e-01
7.74411440e-01 1.16185820e+00 6.71221269e-03 3.96288753e-01
1.56913316e+00 -1.04735684e+00 -3.01170886e-01 -3.45772892e-01
5.07917047e-01 -8.71046901e-01 5.27122915e-01 3.44506562e-01
-8.68719101e-01 -7.29536831e-01 -1.17817855e+00 -1.47468776e-01
-6.44937336e-01 8.02914977e-01 3.15045089e-01 1.06707251e+00
-9.00009573e-01 -8.21532831e-02 -5.95468819e-01 -1.88526809e-01
6.71224833e-01 4.44161296e-01 -3.34462106e-01 -1.08230449e-02
-7.20588684e-01 8.75023782e-01 9.88170952e-02 4.59777899e-02
-9.66332555e-01 -5.24525464e-01 -5.96492350e-01 -1.29085273e-01
2.48373136e-01 2.68386547e-02 9.81930435e-01 -9.44967270e-01
-1.35351753e+00 1.31468332e+00 3.42340976e-01 -5.78052819e-01
7.85461664e-01 -1.43465653e-01 -7.08414793e-01 2.26117641e-01
1.66722596e-01 5.11309683e-01 7.88011789e-01 -5.15797317e-01
-1.03066444e+00 -2.65154928e-01 -1.77825131e-02 -1.21825129e-01
1.30077064e-01 7.11587369e-01 -8.75494182e-01 -3.51969302e-01
-6.48211062e-01 -9.47505951e-01 2.54754126e-01 1.14357166e-01
-1.79534093e-01 -3.27393860e-01 1.06828094e+00 -9.15315330e-01
6.66082919e-01 -2.11548018e+00 -9.17755783e-01 2.19504833e-01
4.37494889e-02 1.09464121e+00 3.21826577e-01 1.95589408e-01
2.37996519e-01 -3.61813784e-01 1.07596070e-01 7.70338848e-02
-1.33264914e-01 -1.38763011e-01 -1.36713088e-02 9.59654093e-01
6.75642788e-02 6.59802616e-01 -2.66021937e-01 -4.65858817e-01
7.25616574e-01 4.95310366e-01 -4.47088718e-01 6.02512285e-02
2.67110229e-01 8.69448185e-02 -2.36361161e-01 7.89727807e-01
1.29334319e+00 2.69962013e-01 -2.85767198e-01 4.99041229e-02
-7.67683864e-01 -3.37166637e-02 -1.17371833e+00 6.95451617e-01
1.81435794e-02 1.54686797e+00 1.44895554e-01 -6.79517508e-01
8.43400300e-01 2.32618153e-01 8.79039392e-02 -7.72594452e-01
4.59659636e-01 4.32172678e-02 -3.63862157e-01 -9.84329462e-01
6.13500714e-01 1.89211637e-01 1.09848700e-01 2.89392602e-02
-6.22590035e-02 3.96344393e-01 2.85488546e-01 -8.13062862e-02
9.75850940e-01 -3.61406595e-01 3.15192081e-02 -1.63202569e-01
6.80911183e-01 4.74207073e-01 2.50270545e-01 9.10843611e-01
-5.69980383e-01 3.09185714e-01 4.57295537e-01 -8.81561935e-01
-1.66113019e+00 -8.91898692e-01 -2.11159036e-01 1.16167593e+00
-8.41627494e-02 2.42056206e-01 -8.42191875e-01 -1.44774392e-01
-7.33131543e-02 3.91756862e-01 -3.36419195e-01 2.57246852e-01
-8.49026978e-01 -5.28451920e-01 9.11710083e-01 7.16901362e-01
1.00349772e+00 -1.11042166e+00 -6.19180381e-01 9.86882076e-02
7.27712333e-01 -1.50447738e+00 -2.59002507e-01 -4.45214480e-01
-3.41794908e-01 -1.11540747e+00 -6.62272215e-01 -1.08161426e+00
3.70604336e-01 5.06932080e-01 7.72751927e-01 -1.25645474e-01
-3.42280447e-01 8.08113217e-02 1.50392130e-01 -7.20264554e-01
-3.67086560e-01 -2.94988900e-01 3.71864922e-02 3.72799709e-02
8.74511361e-01 4.76359278e-02 -4.34243292e-01 1.77967861e-01
-7.15629637e-01 1.55439451e-01 5.82580984e-01 -3.45506519e-02
-3.71573642e-02 -5.98549061e-02 9.62335542e-02 -5.28959751e-01
4.56832051e-01 -4.44759369e-01 -1.44783652e+00 -2.64688879e-02
1.03652962e-02 -9.05319631e-01 4.81352657e-01 1.48036145e-02
-8.10322106e-01 3.27639073e-01 -4.27239060e-01 -7.51414001e-01
-4.23882574e-01 -1.39500931e-01 3.05375516e-01 -5.30160964e-01
4.19798940e-01 5.16875684e-01 1.73968226e-02 -1.41806051e-01
-8.72622430e-02 9.61673975e-01 8.60093415e-01 -1.58132136e-01
7.03506470e-01 5.57760477e-01 -1.21425018e-01 -1.31894195e+00
-2.82456875e-01 -6.77298307e-01 -5.95684469e-01 -9.12107468e-01
1.22385657e+00 -1.03015292e+00 -1.28640687e+00 1.02640259e+00
-1.17992759e+00 -1.54001582e-02 6.13342643e-01 4.83358592e-01
-7.74109513e-02 2.49941751e-01 -8.25002789e-01 -8.72502983e-01
-3.47535133e-01 -1.02989244e+00 1.06255865e+00 5.04944742e-01
5.30037761e-01 -6.61617994e-01 6.17636293e-02 8.48812222e-01
4.72610354e-01 1.76040217e-01 2.95724183e-01 -4.92917657e-01
-8.30228686e-01 -9.18619573e-01 -7.90232003e-01 6.43486679e-01
-5.33662558e-01 2.86600918e-01 -1.10623217e+00 7.08412230e-02
-2.56392598e-01 1.48898959e-02 6.66407287e-01 6.82205200e-01
8.80975723e-01 -2.36369342e-01 -1.50825992e-01 5.52799940e-01
1.67466795e+00 5.17338514e-01 8.94381225e-01 7.55267441e-01
8.13063145e-01 1.38397753e-01 3.33458692e-01 4.10156369e-01
1.48714587e-01 6.93383694e-01 2.77609050e-01 -2.37771720e-01
-5.40033318e-02 2.46698275e-01 5.09607196e-01 1.40127957e-01
-3.50575596e-01 -1.41245335e-01 -1.35284853e+00 1.81985557e-01
-1.36849725e+00 -1.18386209e+00 -7.50926733e-01 1.90516663e+00
1.99436545e-01 3.60419631e-01 4.75428581e-01 -3.50151956e-03
1.27616692e+00 -1.84927776e-01 -1.98802337e-01 -7.45261669e-01
1.16290621e-01 5.73015213e-03 1.25727355e+00 5.27188301e-01
-1.34785438e+00 1.17784500e+00 6.41524982e+00 9.72525299e-01
-1.35881197e+00 1.55920565e-01 6.29276276e-01 1.80581748e-01
9.14754868e-01 -5.13493359e-01 -9.72828746e-01 5.38650751e-01
1.06305230e+00 1.03383981e-01 1.53969452e-01 1.12650132e+00
4.52923506e-01 -2.65711099e-01 -3.85538042e-01 9.06065822e-01
1.68441176e-01 -1.95066011e+00 -4.75184210e-02 2.96987802e-01
4.21912014e-01 5.02316713e-01 2.23354504e-01 4.63942319e-01
-1.29023921e-02 -8.98451269e-01 8.47575426e-01 3.61611128e-01
9.95555878e-01 -1.16370249e+00 8.96367729e-01 5.09138823e-01
-9.85202134e-01 -3.22209150e-02 -7.70305276e-01 -1.90070480e-01
-2.52794951e-01 -5.57292029e-02 -1.38657510e+00 -2.64975667e-01
7.77247667e-01 5.29276133e-01 -5.17393470e-01 1.20421898e+00
3.46237719e-01 8.56464863e-01 -1.83794469e-01 -1.04558304e-01
8.01032066e-01 -3.76594484e-01 5.81021726e-01 1.94890213e+00
2.12168038e-01 5.36481105e-02 -2.07862273e-01 5.82544625e-01
-5.68695478e-02 -3.15825671e-01 -5.18584490e-01 3.65871638e-01
3.28928113e-01 1.26645207e+00 -8.08844268e-01 -5.59555888e-01
-6.85307264e-01 6.86065257e-01 -2.08539143e-01 -5.92208952e-02
-1.29590929e+00 -5.56209207e-01 6.79208994e-01 5.36474705e-01
6.42964900e-01 -5.37606716e-01 -3.02434087e-01 -9.47343588e-01
-1.56600878e-01 -4.07391429e-01 6.84869215e-02 -8.32733214e-01
-9.21651304e-01 1.13626398e-01 -2.63452888e-01 -1.29143119e+00
2.82139122e-01 -1.33971930e+00 -9.37773287e-01 5.65395236e-01
-1.26283455e+00 -1.11547124e+00 -3.33329827e-01 7.58822560e-01
9.09366727e-01 -8.21074188e-01 2.47111857e-01 7.95558512e-01
-9.92925227e-01 4.02783662e-01 5.76928377e-01 9.29993093e-01
1.38161585e-01 -8.72429848e-01 6.31728411e-01 1.07473433e+00
-5.14062941e-01 2.15533793e-01 5.78023672e-01 -5.48836946e-01
-1.49070847e+00 -1.45219731e+00 6.35253966e-01 -3.44099134e-01
6.06033027e-01 -5.02270818e-01 -6.19937599e-01 8.21825743e-01
4.00015771e-01 2.36293107e-01 4.55781311e-01 -7.68188477e-01
-7.98831582e-02 -2.13939101e-01 -1.36306071e+00 4.67003286e-01
2.00218529e-01 -3.19356978e-01 -2.48660490e-01 5.85925817e-01
-1.58748880e-01 -5.38033247e-01 -3.51385921e-01 -2.69839883e-01
6.36788428e-01 -1.05113280e+00 1.19686079e+00 -1.26626164e-01
9.08274800e-02 -5.39264560e-01 2.70794169e-03 -2.17897832e-01
-4.22932833e-01 -3.22650820e-01 1.85543329e-01 1.07748759e+00
4.29068655e-02 -3.59298766e-01 9.29916084e-01 4.38851357e-01
-2.72827685e-01 -1.08398385e-01 -1.32537282e+00 -6.94485426e-01
-1.25835449e-01 -7.97633350e-01 -1.65701304e-02 5.93924701e-01
-6.80671692e-01 1.15756087e-01 -6.33408308e-01 4.88226414e-01
6.18829489e-01 -5.69711685e-01 9.06674385e-01 -9.19127166e-01
1.46062553e-01 -3.21344703e-01 -1.16260099e+00 -6.16625369e-01
-1.99682578e-01 -4.43874419e-01 -2.83273011e-01 -9.67524230e-01
2.99299359e-01 -2.42071953e-02 1.01506971e-01 2.79353827e-01
4.54186171e-01 9.37950850e-01 2.68122882e-01 4.10732806e-01
-6.59957647e-01 -2.46466264e-01 6.36240840e-01 -2.43518770e-01
2.47594699e-01 4.39878345e-01 -2.69429758e-02 9.45110321e-01
8.43604624e-01 -6.32651985e-01 3.25042784e-01 -4.67194080e-01
1.70466214e-01 3.37403491e-02 6.21343493e-01 -1.40712082e+00
5.50398588e-01 1.44632995e-01 8.49059999e-01 -9.22591925e-01
3.35232377e-01 -6.32809639e-01 -2.55904049e-01 8.16986501e-01
1.09927393e-01 1.80931151e-01 6.71171844e-01 3.75780731e-01
1.53387800e-01 -2.88913935e-01 1.14957392e+00 -1.47682233e-02
-1.08737278e+00 1.95942700e-01 -1.32645631e+00 -4.66494232e-01
1.42563665e+00 -8.00809443e-01 -4.22015667e-01 3.59019972e-02
-2.12005883e-01 -2.41785258e-01 2.46432126e-01 2.73452222e-01
4.96200979e-01 -1.14372420e+00 -1.01003122e+00 3.83993506e-01
-2.63815839e-02 -6.17805839e-01 2.68901110e-01 4.92462158e-01
-1.68431640e+00 9.17531133e-01 -7.52237439e-01 -6.85568690e-01
-1.33850694e+00 3.01030636e-01 6.19020402e-01 3.74305546e-01
-7.27781892e-01 6.98206782e-01 -9.04155970e-02 -1.68552026e-01
-4.32650112e-02 1.98846310e-01 -5.33555329e-01 -2.32611969e-01
7.75022089e-01 9.83935773e-01 1.81692690e-01 -1.09716094e+00
-5.82651377e-01 6.76634789e-01 -1.31289616e-01 7.66335875e-02
1.41553259e+00 3.51714671e-01 5.59917018e-02 -1.75855190e-01
1.14913833e+00 2.98909079e-02 -1.57704544e+00 5.20837903e-02
-1.18915588e-01 -5.07565558e-01 3.44138980e-01 -4.55431789e-01
-1.20494962e+00 5.79280138e-01 9.72525477e-01 1.88080147e-01
6.27006531e-01 -2.94893682e-01 8.29172671e-01 4.95950282e-01
-7.04008490e-02 -1.29923356e+00 -3.19764763e-01 5.78783035e-01
4.20622677e-01 -1.32847297e+00 -9.28703845e-02 -1.15781575e-01
-7.87546277e-01 1.62387407e+00 6.43830180e-01 -8.05716634e-01
3.75295311e-01 6.32537425e-01 1.73169076e-02 -3.12016219e-01
-1.96695790e-01 -7.74790272e-02 3.09400380e-01 6.46508753e-01
1.67109549e-01 1.85785398e-01 -2.12717522e-02 -1.18401228e-03
1.66162759e-01 6.76542148e-02 7.60048151e-01 8.94191980e-01
-9.66562510e-01 -1.74673423e-01 -8.27482939e-01 3.44536155e-01
-7.74013937e-01 -5.05092219e-02 -1.71421945e-01 9.90086675e-01
4.97859329e-01 8.49023104e-01 6.53867245e-01 -3.00819129e-01
3.26981395e-02 -3.94636095e-01 9.04268175e-02 -3.74247134e-01
-6.59469724e-01 -1.52609795e-01 -5.62777780e-02 -1.79217041e-01
1.08516231e-01 -5.85558474e-01 -8.62862170e-01 -9.44505036e-01
1.16048595e-02 -2.87935406e-01 9.86312032e-01 9.85656977e-01
1.29981831e-01 1.14907965e-01 4.81880784e-01 -1.07405937e+00
2.01864883e-01 -8.35535169e-01 -6.43778563e-01 -3.57719660e-01
3.40509385e-01 -2.81772286e-01 5.95721416e-02 1.41294241e-01] | [9.898487091064453, -4.867773532867432] |
e9e2f0ee-c3c2-4084-b9ec-b93394d1e252 | leaf-segmentation-and-counting-with-deep | 2012.11486 | null | https://arxiv.org/abs/2012.11486v1 | https://arxiv.org/pdf/2012.11486v1.pdf | Leaf Segmentation and Counting with Deep Learning: on Model Certainty, Test-Time Augmentation, Trade-Offs | Plant phenotyping tasks such as leaf segmentation and counting are fundamental to the study of phenotypic traits. Since it is well-suited for these tasks, deep supervised learning has been prevalent in recent works proposing better performing models at segmenting and counting leaves. Despite good efforts from research groups, one of the main challenges for proposing better methods is still the limitation of labelled data availability. The main efforts of the field seem to be augmenting existing limited data sets, and some aspects of the modelling process have been under-discussed. This paper explores such topics and present experiments that led to the development of the best-performing method in the Leaf Segmentation Challenge and in another external data set of Komatsuna plants. The model has competitive performance while been arguably simpler than other recently proposed ones. The experiments also brought insights such as the fact that model cardinality and test-time augmentation may have strong applications in object segmentation of single class and high occlusion, and regarding the data distribution of recently proposed data sets for benchmarking. | ['Lihong Zheng', 'Douglas Pinto Sampaio Gomes'] | 2020-12-21 | null | null | null | null | ['plant-phenotyping'] | ['computer-vision'] | [ 3.41234118e-01 2.30955258e-01 -3.12429339e-01 -4.25432026e-01
-5.05467117e-01 -7.46812284e-01 2.81660348e-01 4.81148034e-01
-2.21102178e-01 5.42209387e-01 -5.53638101e-01 -4.07676399e-01
-3.76146913e-01 -7.97038734e-01 -4.10016567e-01 -8.30191195e-01
-1.57472461e-01 1.12051654e+00 4.77705240e-01 6.10301942e-02
7.14024156e-02 8.58624279e-01 -1.74778867e+00 1.29852548e-01
7.93230295e-01 9.34199214e-01 5.49712956e-01 8.51309180e-01
-3.56039286e-01 3.70332479e-01 -6.47457480e-01 -2.14334279e-01
2.09406585e-01 -3.21393967e-01 -1.09870648e+00 6.32024705e-01
6.32335484e-01 -9.36653689e-02 3.32355350e-01 7.26746142e-01
6.27726436e-01 -3.25879812e-01 5.26577473e-01 -1.35751987e+00
-4.23090935e-01 8.74249279e-01 -6.74073696e-01 3.11940759e-02
-3.11437845e-01 4.42723259e-02 8.97009134e-01 -2.90102392e-01
5.91596484e-01 9.60251927e-01 7.57964492e-01 2.80073613e-01
-1.47543073e+00 1.04521848e-02 2.59896610e-02 3.75318617e-01
-1.07520032e+00 2.52163038e-02 3.22832227e-01 -4.06244189e-01
3.61464053e-01 6.01916134e-01 5.60648322e-01 6.96872234e-01
-4.87908483e-01 1.19493067e+00 1.12196231e+00 -5.75808704e-01
3.37662935e-01 1.86728746e-01 2.16586530e-01 5.24093449e-01
3.17425102e-01 -6.33839965e-02 2.46186450e-01 -4.63950634e-02
5.21082222e-01 -4.76082265e-01 -9.67631713e-02 -9.30909932e-01
-1.00052381e+00 8.54722261e-01 1.77400291e-01 6.12437427e-01
-1.04284599e-01 -3.48880917e-01 6.07649207e-01 -1.59161627e-01
4.61344451e-01 4.95305628e-01 -1.06816876e+00 -4.83783782e-02
-1.39254236e+00 4.29246634e-01 1.11633348e+00 8.82877588e-01
3.17914903e-01 -5.02818860e-02 -4.88845795e-01 8.60609055e-01
-2.86190677e-02 1.95403636e-01 -1.55757546e-01 -1.28096008e+00
-5.18150337e-04 7.42499411e-01 3.86290886e-02 -4.82920080e-01
-7.26260245e-01 -4.90347028e-01 -8.50602508e-01 3.83558720e-01
1.05516911e+00 1.39086425e-01 -9.86780941e-01 1.47410595e+00
2.62885749e-01 -1.20605260e-01 -5.09426594e-01 6.85027897e-01
7.13959873e-01 3.54978710e-01 3.04012865e-01 -1.86449066e-01
1.27886665e+00 -5.53142786e-01 -4.99708444e-01 -9.10700113e-02
8.29867423e-01 -8.77210855e-01 8.27996075e-01 6.74326479e-01
-1.11126328e+00 -7.39316106e-01 -7.62571931e-01 1.01131916e-01
-5.50901055e-01 7.46681392e-01 9.61764693e-01 9.60928142e-01
-8.35318804e-01 1.00402677e+00 -7.47744739e-01 -8.73241603e-01
1.07630587e+00 3.75620186e-01 -1.16957873e-01 2.62048785e-02
-3.30484092e-01 7.80797899e-01 6.59421444e-01 3.42497319e-01
-8.92042875e-01 -6.64837480e-01 -2.92387217e-01 3.48517805e-01
7.60837078e-01 -3.66008162e-01 1.09811711e+00 -4.46631014e-01
-1.39757574e+00 1.33689392e+00 2.81189561e-01 -6.28047585e-01
6.17313027e-01 1.62900507e-01 -2.56708004e-02 -2.46084899e-01
-5.61436638e-02 9.92960215e-01 4.30076838e-01 -1.22384632e+00
-8.22992861e-01 -6.49177909e-01 -1.02219477e-01 -4.84337986e-01
-3.04645479e-01 -1.08699858e-01 -1.96625024e-01 -3.16666186e-01
2.76901454e-01 -1.03966796e+00 -3.73728245e-01 2.00729355e-01
-5.13957262e-01 -2.69467443e-01 1.06192017e+00 -5.52806377e-01
8.35453928e-01 -1.70226181e+00 1.83349073e-01 -2.56032377e-01
-3.90391573e-02 8.04780066e-01 -2.89873958e-01 3.47910076e-01
-2.69160092e-01 2.94672579e-01 -5.21368206e-01 -1.52350292e-01
-7.04198703e-02 4.04337227e-01 -1.27071530e-01 4.30806577e-01
5.62550783e-01 7.83646703e-01 -4.77578044e-01 -5.57685077e-01
5.10893643e-01 1.91851526e-01 -1.45835727e-01 7.06789866e-02
-6.75522447e-01 5.85565805e-01 -2.86727726e-01 1.01380014e+00
1.18266714e+00 -3.81277353e-02 1.07970543e-01 -2.24926516e-01
-4.84269172e-01 -8.31991956e-02 -1.28975892e+00 1.39915133e+00
-4.26747426e-02 5.86274922e-01 1.90185905e-01 -1.54461813e+00
7.69548357e-01 8.62088576e-02 8.04327548e-01 -2.72297412e-01
3.33428293e-01 1.19141027e-01 3.77352983e-01 -4.57708776e-01
1.65759444e-01 4.99018788e-01 5.14730513e-01 9.54659935e-03
2.66783774e-01 -5.89941263e-01 1.13071966e+00 -9.44614783e-02
9.72977281e-01 6.56307220e-01 2.34082907e-01 -5.12451410e-01
5.18324077e-01 4.61463124e-01 4.97231513e-01 6.07790291e-01
-4.92187083e-01 7.10132122e-01 8.84173155e-01 -3.58966291e-01
-1.08194256e+00 -4.13565040e-01 -6.09497428e-01 1.14189863e+00
-2.64526755e-01 -2.76168108e-01 -8.89871836e-01 -7.29013920e-01
7.05436012e-03 5.73900998e-01 -5.73149741e-01 2.94002682e-01
-3.53970885e-01 -1.27667964e+00 5.71270168e-01 6.36858702e-01
4.40196842e-01 -1.32343256e+00 -7.33166635e-01 3.46230596e-01
2.19306573e-02 -1.42234564e+00 6.39382780e-01 6.76161647e-01
-1.35205293e+00 -1.18744504e+00 -8.42021704e-01 -6.20464265e-01
2.46481061e-01 1.08365864e-01 1.47059810e+00 2.98033096e-02
-9.10138071e-01 -7.80807436e-02 -4.93515819e-01 -7.90470600e-01
-4.97465968e-01 6.46768391e-01 -6.37638927e-01 -5.48380435e-01
2.38024160e-01 -4.39335406e-01 -3.41832936e-01 3.36764604e-01
-9.75927413e-01 -2.61373192e-01 5.81432700e-01 8.40258956e-01
5.84195852e-01 -9.19150130e-04 3.40591460e-01 -1.37203801e+00
9.25888214e-03 -1.61920801e-01 -1.14115894e+00 4.04215157e-01
-4.80120152e-01 -1.89220414e-01 3.91204715e-01 -3.44369054e-01
-7.75747299e-01 2.54441708e-01 -2.45432928e-01 3.15814674e-01
-8.82741690e-01 3.35265964e-01 -5.03027976e-01 -2.02653468e-01
6.83950365e-01 -3.16088796e-01 -1.00231111e-01 -7.21726537e-01
3.44247520e-01 1.89586446e-01 2.68815637e-01 -5.75290918e-01
7.90044427e-01 5.86585104e-01 4.76531744e-01 -9.90503669e-01
-1.02719223e+00 -4.63625193e-01 -1.02867031e+00 1.30322918e-01
5.99624276e-01 -3.62435073e-01 -8.91461313e-01 5.63396931e-01
-9.41844642e-01 -4.64474589e-01 -7.87334621e-01 1.90954283e-02
-6.19892180e-01 6.43997669e-01 -4.14745778e-01 -8.41188371e-01
-1.77874770e-02 -1.11113894e+00 1.05449498e+00 4.76706862e-01
9.14522335e-02 -8.98602068e-01 -1.60761088e-01 4.40955013e-01
3.52225155e-01 2.60727882e-01 1.08025742e+00 -5.56429446e-01
-6.33944094e-01 -5.35299242e-01 -5.58089077e-01 5.99832177e-01
-1.37507588e-01 6.28249526e-01 -1.35045505e+00 -2.24269956e-01
-1.93099707e-01 -5.41471303e-01 9.07660723e-01 9.79610920e-01
1.83606482e+00 6.63306177e-01 -4.37944591e-01 5.07270873e-01
1.38691115e+00 -3.24097462e-02 6.24072313e-01 2.05575764e-01
4.87460673e-01 1.00743353e+00 8.20316494e-01 4.54770684e-01
1.07625656e-01 5.89566529e-01 9.70089197e-01 -5.35049558e-01
-2.04002425e-01 4.55239832e-01 -4.84889060e-01 3.14480811e-01
-2.23822519e-01 -6.04029298e-01 -8.95308077e-01 6.03075624e-01
-1.71874881e+00 -7.66233206e-01 -7.88479090e-01 2.06837010e+00
4.30883259e-01 2.71643072e-01 3.62270743e-01 6.55185461e-01
4.22756195e-01 2.13961229e-02 -4.77494121e-01 -3.40128809e-01
-5.10546744e-01 3.93213302e-01 6.49227321e-01 1.24313995e-01
-1.38112402e+00 8.39008868e-01 6.36691713e+00 7.88316667e-01
-7.09314108e-01 -5.94937801e-02 9.56065536e-01 5.52761555e-01
3.55827630e-01 4.94441569e-01 -1.02073061e+00 1.91250384e-01
8.02532017e-01 4.14431840e-01 9.28066671e-02 9.45768833e-01
1.56109631e-01 -5.83446324e-01 -1.08089936e+00 2.60165930e-01
-1.86390072e-01 -1.06379855e+00 -2.72112817e-01 1.71573892e-01
6.16792142e-01 4.62344140e-02 -1.36183143e-01 4.16448683e-01
1.13433316e-01 -8.15776825e-01 2.30635792e-01 6.56985939e-02
4.18972939e-01 -4.15292293e-01 9.22037959e-01 5.84218204e-01
-9.62406695e-01 -1.97053999e-01 -8.96251857e-01 1.04763411e-01
-9.75305066e-02 9.39209878e-01 -9.02760625e-01 6.75392449e-01
7.62320340e-01 3.34816068e-01 -1.28529727e+00 1.56482768e+00
1.34219110e-01 9.62693453e-01 -5.40029585e-01 1.96052954e-01
3.33174914e-02 -3.57886344e-01 2.36545280e-01 1.11141431e+00
1.58002019e-01 -5.06739557e-01 3.16629499e-01 1.03224957e+00
3.60724688e-01 8.13642293e-02 -3.67416739e-01 -3.96974951e-01
1.55469961e-02 1.91150618e+00 -1.32562017e+00 -9.59058031e-02
-1.16914786e-01 7.47922838e-01 1.38699353e-01 -6.45422935e-02
-7.39575148e-01 5.66687919e-02 1.76409762e-02 2.46295840e-01
4.71073985e-01 -7.29838312e-02 -5.92847645e-01 -5.18142462e-01
-2.17215806e-01 -7.30377376e-01 4.72997129e-01 -4.66830671e-01
-1.06390905e+00 3.16693485e-01 4.65222076e-02 -9.36297178e-01
5.57065308e-02 -1.09574068e+00 -3.14888388e-01 7.43924022e-01
-1.41541350e+00 -1.44571471e+00 -5.87746918e-01 -2.44331837e-01
7.18198776e-01 -6.70402274e-02 1.15745771e+00 3.91260505e-01
-5.96122980e-01 1.21023901e-01 2.30479196e-01 -2.50466049e-01
5.61259627e-01 -1.47543919e+00 4.01091129e-01 7.63027132e-01
4.65115726e-01 -1.80118829e-01 6.97145104e-01 -2.23931238e-01
-9.14196014e-01 -7.55518198e-01 8.22640657e-01 -6.16895735e-01
4.20393437e-01 -2.29634851e-01 -9.55939353e-01 3.68828237e-01
9.88118574e-02 1.76467285e-01 4.82601047e-01 3.97049114e-02
2.28447735e-01 -7.64804259e-02 -1.12005424e+00 2.60474712e-01
8.50589275e-01 1.39247522e-01 2.15915471e-01 6.55305505e-01
7.21877441e-02 -3.56641471e-01 -7.70571470e-01 9.11904275e-01
3.29804540e-01 -1.10554504e+00 9.62230384e-01 -5.81894755e-01
3.13759059e-01 -1.81594491e-01 4.35860828e-02 -1.12639713e+00
-4.87219632e-01 -2.97938526e-01 -7.59633305e-03 1.81543458e+00
1.30883783e-01 5.69820739e-02 1.17648125e+00 1.57897651e-01
-2.15622168e-02 -6.01828694e-01 -4.36204761e-01 -8.05994332e-01
2.20496163e-01 -2.61312366e-01 3.12854111e-01 7.14911044e-01
-9.18062747e-01 8.20840821e-02 -3.67325023e-02 8.05722401e-02
5.54736793e-01 3.90568227e-01 7.72821784e-01 -1.90869081e+00
-3.10062058e-02 -5.64039469e-01 -3.34588319e-01 -6.10803306e-01
5.94649091e-02 -9.89474058e-01 -4.93165478e-02 -1.58654118e+00
3.19428951e-01 -4.58909035e-01 1.33392796e-01 3.60864043e-01
-1.54664710e-01 2.06372812e-01 2.25714818e-01 -2.65074581e-01
-2.45365024e-01 8.23779125e-03 1.31535614e+00 -1.63431659e-01
-2.71002413e-03 5.39762795e-01 -4.43743885e-01 7.86098301e-01
9.96598184e-01 -5.54020643e-01 -2.41885841e-01 -4.59799469e-01
1.81996820e-05 -3.35704356e-01 5.43958724e-01 -1.16989720e+00
-2.03351900e-01 -3.38807218e-02 5.30481458e-01 -1.16551650e+00
2.05772102e-01 -1.05428684e+00 -6.68892786e-02 5.34478188e-01
-2.09034801e-01 -2.60264844e-01 3.64499599e-01 1.24499142e-01
5.69321141e-02 -8.51660788e-01 1.12789845e+00 -2.56524950e-01
-7.86730945e-01 3.14364403e-01 -1.80764198e-01 5.12914993e-02
1.07723510e+00 -5.20750523e-01 -1.49751499e-01 9.40855816e-02
-8.96626949e-01 1.52129188e-01 1.02788009e-01 3.07732671e-01
-2.77796298e-01 -6.50535285e-01 -8.96430731e-01 -5.23549330e-04
2.01720282e-01 -2.70218085e-02 1.46850199e-01 8.94502699e-01
-8.27677608e-01 5.81994534e-01 -6.25762582e-01 -1.08609617e+00
-1.32725453e+00 4.27978814e-01 1.38182819e-01 -5.46962082e-01
-3.54902238e-01 8.43688488e-01 -6.82704821e-02 -6.76152110e-01
5.65144718e-01 -5.78454912e-01 -5.25860250e-01 3.82799923e-01
8.83063953e-03 7.49101102e-01 4.35891122e-01 -1.44756168e-01
-1.81130111e-01 3.39801013e-01 1.43902615e-01 4.55454141e-01
1.56563687e+00 2.24411160e-01 -1.93251953e-01 5.58944762e-01
5.08901000e-01 -4.37064737e-01 -1.09802401e+00 6.21713251e-02
7.69630432e-01 -4.19530988e-01 3.14362869e-02 -1.17434299e+00
-1.12971020e+00 1.14879513e+00 1.12320888e+00 8.11140418e-01
1.15257573e+00 -3.63074429e-02 2.77599931e-01 2.89321393e-01
1.92252859e-01 -1.16313827e+00 -5.81845462e-01 4.29318696e-01
5.75830758e-01 -1.69799078e+00 1.30987778e-01 -9.15409267e-01
-1.71164110e-01 1.22335327e+00 6.44319892e-01 1.15470916e-01
4.69421238e-01 7.14489520e-01 -2.77033709e-02 -1.54302076e-01
-4.95037615e-01 -8.39833379e-01 -5.37186163e-03 1.00895464e+00
8.49945962e-01 8.83014500e-02 -6.27022982e-01 3.23756993e-01
2.22373068e-01 2.12717786e-01 4.63887930e-01 8.91474783e-01
-2.72728235e-01 -1.56175983e+00 -3.14426512e-01 5.24906099e-01
-4.42305923e-01 3.03847700e-01 -6.15371883e-01 1.10703540e+00
4.72940356e-01 7.13127911e-01 -2.62403667e-01 3.42099071e-01
5.22643745e-01 9.34071317e-02 7.96498895e-01 -7.68154621e-01
-6.64989948e-01 -5.49840480e-02 -9.63344872e-02 -3.48412484e-01
-4.89016920e-01 -7.47234166e-01 -7.33898222e-01 -1.64825588e-01
-8.06011379e-01 -1.90189585e-01 8.36411893e-01 8.15969706e-01
3.46327946e-02 9.26792920e-01 3.03686410e-01 -7.93652356e-01
-6.48524761e-01 -1.16233277e+00 -7.70251334e-01 3.59037846e-01
-3.24818581e-01 -5.16764522e-01 1.43723667e-03 2.23731384e-01] | [9.126411437988281, -1.5114164352416992] |
aba770f7-d7f7-4baa-9052-86f345b84a10 | unsupervised-domain-adaptation-with-4 | 2009.00520 | null | https://arxiv.org/abs/2009.00520v1 | https://arxiv.org/pdf/2009.00520v1.pdf | Unsupervised Domain Adaptation with Progressive Adaptation of Subspaces | Unsupervised Domain Adaptation (UDA) aims to classify unlabeled target domain by transferring knowledge from labeled source domain with domain shift. Most of the existing UDA methods try to mitigate the adverse impact induced by the shift via reducing domain discrepancy. However, such approaches easily suffer a notorious mode collapse issue due to the lack of labels in target domain. Naturally, one of the effective ways to mitigate this issue is to reliably estimate the pseudo labels for target domain, which itself is hard. To overcome this, we propose a novel UDA method named Progressive Adaptation of Subspaces approach (PAS) in which we utilize such an intuition that appears much reasonable to gradually obtain reliable pseudo labels. Speci fically, we progressively and steadily refine the shared subspaces as bridge of knowledge transfer by adaptively anchoring/selecting and leveraging those target samples with reliable pseudo labels. Subsequently, the refined subspaces can in turn provide more reliable pseudo-labels of the target domain, making the mode collapse highly mitigated. Our thorough evaluation demonstrates that PAS is not only effective for common UDA, but also outperforms the state-of-the arts for more challenging Partial Domain Adaptation (PDA) situation, where the source label set subsumes the target one. | ['Weikai Li', 'Songcan Chen'] | 2020-09-01 | null | null | null | null | ['partial-domain-adaptation'] | ['methodology'] | [ 3.25678766e-01 8.51111785e-02 -4.15417910e-01 -2.75312424e-01
-1.07766604e+00 -8.70729029e-01 5.74833632e-01 -8.29878449e-02
-2.13218346e-01 1.06242394e+00 2.68194944e-01 4.11721431e-02
-2.07028627e-01 -5.27699351e-01 -5.52893579e-01 -9.66037214e-01
3.51650387e-01 6.25029743e-01 2.05008999e-01 -2.15179563e-01
5.75523414e-02 3.30151796e-01 -1.24263620e+00 9.02641658e-03
1.25669491e+00 7.59478152e-01 1.69735298e-01 -7.38238320e-02
-2.69217759e-01 3.58879685e-01 -4.23573107e-01 -1.86466679e-01
5.48934937e-01 -5.67696989e-01 -9.01419580e-01 4.43760216e-01
2.16033190e-01 -1.67951539e-01 -6.97689429e-02 1.21290815e+00
3.43812972e-01 2.66040564e-01 9.91990626e-01 -1.28623915e+00
-6.02084816e-01 4.04298693e-01 -7.80708790e-01 5.66818602e-02
1.08952321e-01 -1.08582154e-01 8.53075802e-01 -9.26140428e-01
6.39471471e-01 1.16158330e+00 5.82675755e-01 7.99704909e-01
-1.34457195e+00 -9.70469892e-01 4.19742525e-01 2.21266765e-02
-1.54237664e+00 -5.04883766e-01 1.07651877e+00 -3.98458779e-01
2.14265689e-01 8.98714513e-02 -2.70180823e-03 1.15901136e+00
-4.83857214e-01 8.21246505e-01 1.26446295e+00 -4.78701651e-01
4.90132123e-01 5.95472515e-01 2.57251710e-01 2.58391201e-01
8.69690478e-02 -1.07409738e-01 -4.61289555e-01 -3.26049387e-01
5.67909479e-01 3.13669071e-02 -3.35058451e-01 -9.27751720e-01
-1.20848751e+00 6.54538691e-01 2.27555975e-01 2.45271362e-02
-4.64472741e-01 -6.75272107e-01 4.82683390e-01 4.83276665e-01
5.34279227e-01 3.97576392e-01 -7.07747519e-01 7.58941397e-02
-7.24626124e-01 1.65009543e-01 4.81849134e-01 1.19081867e+00
9.66289759e-01 -3.64626870e-02 -2.93125864e-02 1.16356075e+00
2.60133862e-01 3.61221254e-01 7.19251394e-01 -8.32518697e-01
5.18234789e-01 8.23238313e-01 4.12459940e-01 -4.81898278e-01
-1.05178729e-01 -4.31034476e-01 -7.92027235e-01 9.03976634e-02
6.88557625e-01 -2.18355268e-01 -1.05816948e+00 2.11475897e+00
7.39224434e-01 2.71210223e-01 4.74812061e-01 8.93086076e-01
2.96605498e-01 4.74716485e-01 3.17697018e-01 -4.98361707e-01
1.15244687e+00 -8.46997857e-01 -5.19573212e-01 -2.37248123e-01
7.17862725e-01 -6.18643820e-01 1.18730283e+00 2.06929430e-01
-5.12282193e-01 -7.25116432e-01 -1.09410369e+00 2.63194323e-01
-2.25701302e-01 2.38552168e-02 2.77332276e-01 5.71478367e-01
-7.35160828e-01 4.79421884e-01 -5.18406391e-01 -5.88070750e-01
4.84955847e-01 3.49023461e-01 -5.39584398e-01 -2.52315074e-01
-1.20191073e+00 5.70549369e-01 6.57323897e-01 -3.32357436e-01
-8.06092739e-01 -7.60200024e-01 -6.92354143e-01 -1.91679940e-01
5.86659431e-01 -3.16337645e-01 1.17194426e+00 -1.22446811e+00
-1.46551311e+00 8.53123128e-01 -1.94918901e-01 -4.02878106e-01
4.61468995e-01 -2.16461256e-01 -5.23445070e-01 -1.95396990e-01
2.61683822e-01 7.21711576e-01 1.07444561e+00 -1.52158368e+00
-7.36086786e-01 -5.98097980e-01 -2.29374468e-02 5.40556848e-01
-9.13903296e-01 -3.64035010e-01 -1.80475011e-01 -6.34918630e-01
2.43076935e-01 -1.10855722e+00 -1.95217714e-01 -1.94354400e-01
-3.12423617e-01 -4.27006513e-01 1.01968586e+00 -4.91232306e-01
1.25928521e+00 -2.40087676e+00 2.37183318e-01 1.49091333e-01
1.52380794e-01 4.26865906e-01 -1.35728583e-01 2.17231259e-01
-2.89719760e-01 -2.38692150e-01 -5.35243571e-01 -1.37740120e-01
-1.44383848e-01 2.50045747e-01 -6.39573514e-01 3.07064354e-01
1.35433629e-01 3.17580730e-01 -9.84534383e-01 -4.98387039e-01
-4.86639924e-02 1.07007779e-01 -4.08733130e-01 3.19271207e-01
-2.30086342e-01 8.67199481e-01 -7.75025964e-01 5.61011791e-01
8.88937294e-01 -2.95069456e-01 3.42942357e-01 -1.89383551e-01
1.24932706e-01 3.37351263e-02 -1.42427278e+00 1.74338484e+00
-1.50730118e-01 1.40683502e-01 6.80774227e-02 -1.14635074e+00
1.17490828e+00 3.19945157e-01 6.57894552e-01 -3.03264678e-01
-9.17166919e-02 3.59873027e-01 -1.49508804e-01 -2.29165986e-01
2.26301268e-01 -4.10973459e-01 -1.00987129e-01 3.07553351e-01
-7.85748009e-03 2.58972675e-01 -1.33488789e-01 1.06894210e-01
8.88241112e-01 3.92809063e-01 5.44139147e-01 -4.04226780e-01
8.86469722e-01 2.78252125e-01 9.03306425e-01 3.66092503e-01
-6.64953649e-01 5.73190987e-01 1.98290020e-01 -1.13469400e-01
-1.00847948e+00 -1.21911347e+00 -1.53144225e-01 1.15469873e+00
3.44131351e-01 -3.85184437e-02 -7.07986295e-01 -1.26207769e+00
-2.20517311e-02 7.25242257e-01 -5.02831995e-01 -3.58162284e-01
-4.92886573e-01 -5.88907063e-01 3.12535584e-01 5.80464244e-01
7.60277152e-01 -8.02085578e-01 1.00110779e-02 1.94150433e-01
-3.19855809e-01 -9.00345683e-01 -5.60402930e-01 2.63087332e-01
-1.06614172e+00 -9.00626242e-01 -1.03817451e+00 -9.44747567e-01
8.33152294e-01 6.52432919e-01 7.19674766e-01 -5.14374495e-01
4.33547050e-01 1.80603907e-01 -4.79593456e-01 -1.84173673e-01
-6.31153882e-01 1.34517565e-01 3.86731684e-01 2.27415904e-01
7.28826046e-01 -6.71795428e-01 -4.82065737e-01 5.96874416e-01
-9.08725739e-01 -1.67932769e-03 5.41651368e-01 9.37848687e-01
7.16877043e-01 3.17815244e-01 1.08429694e+00 -1.20601368e+00
5.65807223e-01 -7.82349527e-01 -3.96622628e-01 3.37693959e-01
-7.93883383e-01 1.44072667e-01 9.90521252e-01 -7.49228179e-01
-1.40991938e+00 3.43091607e-01 1.90262347e-01 -7.02970684e-01
-5.41149437e-01 2.39156187e-01 -4.86938238e-01 1.81104347e-01
1.00917768e+00 2.97968447e-01 -1.07451349e-01 -6.72979116e-01
3.69419336e-01 8.79943013e-01 4.02394503e-01 -7.13369548e-01
1.06590521e+00 5.42920649e-01 -1.59740016e-01 -5.75728059e-01
-8.82464945e-01 -8.28672171e-01 -8.33241880e-01 7.55902529e-02
5.52201688e-01 -1.12364757e+00 6.71081468e-02 3.19901854e-01
-7.50014484e-01 -1.59924462e-01 -3.35217655e-01 4.61603791e-01
-4.51644152e-01 5.56637347e-01 -2.12485835e-01 -6.69629753e-01
-1.15368851e-01 -9.07083213e-01 8.34654808e-01 2.86393344e-01
-2.91669428e-01 -9.34831023e-01 7.73583874e-02 3.36748898e-01
1.63651392e-01 1.82086095e-01 1.01854563e+00 -1.05122745e+00
-1.23120435e-01 6.09306991e-02 -2.52953321e-01 5.67015707e-01
6.70726836e-01 -6.77570283e-01 -1.10838401e+00 -5.72767317e-01
9.93129164e-02 -3.36688578e-01 5.20214677e-01 9.72063318e-02
8.23515713e-01 -6.49779588e-02 -4.12342817e-01 2.68633068e-01
1.28575659e+00 2.84388334e-01 1.86815694e-01 3.23158890e-01
7.89200723e-01 6.84759974e-01 1.17598188e+00 4.28500682e-01
3.16102892e-01 7.26114452e-01 4.05111313e-02 -2.44723819e-02
-1.90468490e-01 -3.51203173e-01 4.45190847e-01 8.75203729e-01
1.95040718e-01 -8.56651179e-03 -9.33438599e-01 6.91753209e-01
-1.86810780e+00 -4.36545521e-01 2.59040117e-01 2.44747519e+00
1.11878288e+00 1.39489874e-01 3.84493411e-01 1.21192932e-01
9.26353812e-01 -1.57767490e-01 -8.59262586e-01 1.49629861e-01
-3.68111804e-02 -2.06470266e-01 4.49229628e-01 2.05497801e-01
-1.21215999e+00 9.46805000e-01 5.44970417e+00 1.01193190e+00
-9.05789614e-01 1.61374360e-01 4.17822719e-01 3.26358885e-01
-1.87066659e-01 4.74636033e-02 -7.69344866e-01 5.02351880e-01
7.66225696e-01 -3.47336620e-01 2.62549341e-01 1.16662526e+00
2.55693365e-02 1.68149173e-01 -1.23127997e+00 8.32374632e-01
-1.08861096e-01 -7.64880121e-01 1.48931444e-01 1.68083813e-02
9.14308131e-01 -2.31335223e-01 1.28345760e-02 6.67449474e-01
4.44543779e-01 -3.99182022e-01 3.54969889e-01 5.60111068e-02
9.26040888e-01 -6.63594127e-01 4.98937815e-01 5.87265372e-01
-1.04105890e+00 -1.67927697e-01 -5.36550999e-01 1.99255377e-01
-1.60822242e-01 3.18627536e-01 -1.10589051e+00 5.83196342e-01
4.98415440e-01 7.10083604e-01 -4.53315139e-01 8.64456236e-01
-4.52915542e-02 7.76522338e-01 -1.75620764e-01 4.90471005e-01
1.24197133e-01 -4.16219771e-01 7.00086296e-01 8.77668917e-01
3.71782511e-01 1.46103591e-01 2.92701840e-01 5.42035878e-01
-2.53644824e-01 2.85903782e-01 -7.84003973e-01 6.44935668e-02
7.87425756e-01 9.07887042e-01 -6.64722383e-01 -4.65229273e-01
-5.11154115e-01 1.19901085e+00 3.88801873e-01 6.80305123e-01
-7.55450606e-01 -1.02512114e-01 6.82817757e-01 9.37846005e-02
1.31385565e-01 9.78977978e-03 -3.11273605e-01 -1.21056879e+00
7.50728548e-02 -1.01415014e+00 7.42250741e-01 -4.40476298e-01
-1.53146279e+00 4.22565758e-01 7.92280585e-02 -1.85055602e+00
-1.23199135e-01 -1.76610410e-01 -2.65929878e-01 8.37491393e-01
-1.71534491e+00 -1.05205441e+00 -2.76575178e-01 8.46384823e-01
9.44578648e-01 -3.36274743e-01 9.15678978e-01 4.25374150e-01
-5.83970726e-01 7.61558414e-01 4.93636280e-01 -1.47414356e-01
1.40549111e+00 -1.28967440e+00 4.12564650e-02 7.46798277e-01
-7.02578351e-02 6.96768045e-01 7.03574538e-01 -6.51876569e-01
-1.03625321e+00 -1.27245998e+00 5.19841850e-01 -4.26298857e-01
5.70633173e-01 -2.87890702e-01 -1.33574688e+00 6.75365865e-01
-1.83673799e-01 -1.37044594e-01 6.65374279e-01 1.77200943e-01
-5.83481431e-01 -3.10837418e-01 -1.37319887e+00 5.87116182e-01
9.19610918e-01 -4.26793277e-01 -7.80632079e-01 3.28599423e-01
8.29738438e-01 -1.57051623e-01 -8.36668849e-01 5.02542615e-01
1.59995839e-01 -7.21997976e-01 9.66221750e-01 -5.78972042e-01
1.65161595e-01 -5.81654072e-01 -1.39033616e-01 -1.46939504e+00
-4.21912581e-01 -4.31730002e-01 -8.14530626e-02 1.60673571e+00
2.80398756e-01 -6.10102654e-01 1.02365947e+00 6.36258364e-01
-5.29144295e-02 -3.04765999e-01 -9.30494070e-01 -1.02283931e+00
2.33484194e-01 -7.27247971e-04 6.48780942e-01 1.48025942e+00
7.83720538e-02 3.93375784e-01 -3.76361400e-01 3.21313292e-01
7.37675011e-01 1.42291293e-01 7.76270092e-01 -1.29150808e+00
-1.85242653e-01 -1.30366787e-01 -1.23451620e-01 -1.26141655e+00
1.16017893e-01 -8.09791565e-01 1.41102746e-01 -1.02297795e+00
2.30932698e-01 -8.33314240e-01 -7.64264643e-01 4.97094631e-01
-3.84174138e-01 1.46198226e-02 5.14413044e-02 7.17098653e-01
-6.30681515e-01 5.09009302e-01 1.23573661e+00 7.96444830e-04
-4.81763124e-01 1.33037165e-01 -9.24045205e-01 6.78531468e-01
7.50508785e-01 -5.12603641e-01 -8.54763985e-01 -1.83995634e-01
-4.17604864e-01 -9.30941924e-02 2.66168192e-02 -9.05653715e-01
7.27303773e-02 -3.87071490e-01 2.79609859e-01 -3.84235233e-01
9.92562175e-02 -1.08630908e+00 -4.27782014e-02 1.19383834e-01
-3.37599695e-01 -4.47224408e-01 2.00538293e-01 9.39323783e-01
-3.55594605e-01 -9.98056307e-02 1.10187113e+00 1.16815478e-01
-1.03747392e+00 1.73374459e-01 -3.04622315e-02 1.48389846e-01
1.25897312e+00 -3.74641806e-01 3.78488423e-03 -2.51084387e-01
-6.70839369e-01 4.05894935e-01 4.93212461e-01 4.82327849e-01
3.10911655e-01 -1.51524627e+00 -6.42293215e-01 2.89443165e-01
4.26197767e-01 2.11112902e-01 2.82345414e-01 6.29746974e-01
1.76971838e-01 3.25088084e-01 -2.17662349e-01 -5.74866652e-01
-1.08977497e+00 8.91067863e-01 5.72396675e-03 -2.60705262e-01
-5.11942685e-01 7.72070587e-01 7.37687528e-01 -4.99014169e-01
2.94516176e-01 3.32916901e-02 -3.45260918e-01 1.17427483e-01
3.37540478e-01 3.98329377e-01 -7.17252940e-02 -6.18356884e-01
-3.97201300e-01 4.63186920e-01 -4.41640615e-01 1.80761755e-01
1.12948442e+00 -5.93224466e-01 4.01738733e-01 3.04337800e-01
1.02992618e+00 1.30107347e-02 -1.71489775e+00 -8.18486810e-01
1.86622739e-01 -3.63332570e-01 -3.41137469e-01 -8.48687291e-01
-7.01106846e-01 7.42780328e-01 7.23264217e-01 3.65872532e-02
1.45471370e+00 -9.96551216e-02 8.33348393e-01 2.14211076e-01
4.86746252e-01 -1.27256405e+00 1.63044840e-01 2.79539526e-01
6.21377885e-01 -1.48270297e+00 -8.98066759e-02 -6.32486463e-01
-9.85141456e-01 8.65343571e-01 9.19484735e-01 2.36862991e-03
5.23942590e-01 -2.12444559e-01 2.01532006e-01 1.56552553e-01
-3.20181131e-01 -9.87430811e-02 1.56593069e-01 7.35006511e-01
7.42058083e-02 7.00984746e-02 -2.58900255e-01 8.15347075e-01
2.18135893e-01 -2.24894851e-01 2.69949675e-01 8.54788780e-01
-4.62675303e-01 -1.36074913e+00 -5.98134518e-01 1.67093188e-01
-1.37167543e-01 1.45337746e-01 -3.88472289e-01 7.36156583e-01
1.22583322e-01 7.85753429e-01 -3.88666838e-01 -3.10072601e-01
4.55173790e-01 4.21516240e-01 1.33269861e-01 -7.10726380e-01
6.87361835e-03 3.17336351e-01 -1.44541010e-01 -1.67886510e-01
-3.91847581e-01 -7.64678776e-01 -1.26088548e+00 5.10241240e-02
-4.39747781e-01 2.34600559e-01 1.93450406e-01 1.00904405e+00
4.15331483e-01 2.45121285e-01 7.77079701e-01 -4.12316114e-01
-1.01483583e+00 -9.94542658e-01 -7.44447351e-01 8.48391533e-01
3.15716207e-01 -1.02416408e+00 -3.01655442e-01 2.67226756e-01] | [10.329292297363281, 3.1257221698760986] |
80c56a7d-cbe1-4282-b478-d5f3462ff41b | low-light-image-enhancement-by-learning | 2303.13412 | null | https://arxiv.org/abs/2303.13412v1 | https://arxiv.org/pdf/2303.13412v1.pdf | Low-Light Image Enhancement by Learning Contrastive Representations in Spatial and Frequency Domains | Images taken under low-light conditions tend to suffer from poor visibility, which can decrease image quality and even reduce the performance of the downstream tasks. It is hard for a CNN-based method to learn generalized features that can recover normal images from the ones under various unknow low-light conditions. In this paper, we propose to incorporate the contrastive learning into an illumination correction network to learn abstract representations to distinguish various low-light conditions in the representation space, with the purpose of enhancing the generalizability of the network. Considering that light conditions can change the frequency components of the images, the representations are learned and compared in both spatial and frequency domains to make full advantage of the contrastive learning. The proposed method is evaluated on LOL and LOL-V2 datasets, the results show that the proposed method achieves better qualitative and quantitative results compared with other state-of-the-arts. | ['Ziliang Feng', 'Ming Yang', 'Bokai Liu', 'Tingting Liu', 'Gui Fu', 'Xiaoguang Tu', 'Yi Huang'] | 2023-03-23 | null | null | null | null | ['image-enhancement', 'low-light-image-enhancement'] | ['computer-vision', 'computer-vision'] | [ 1.68705955e-01 -5.07677138e-01 -2.35018618e-02 -3.47563416e-01
-2.39002571e-01 -3.42076421e-01 4.64266300e-01 -4.02043074e-01
-1.60465926e-01 7.35715568e-01 1.51199088e-01 1.54113054e-01
-4.91112061e-02 -6.85423851e-01 -5.91345489e-01 -1.11167037e+00
1.42574906e-01 -7.26278543e-01 9.56153423e-02 -2.76077807e-01
2.69703031e-01 6.32002592e-01 -1.95074975e+00 4.24207598e-01
8.62347066e-01 1.02967918e+00 3.82413954e-01 3.70961219e-01
1.75830089e-02 6.98870063e-01 -5.54356396e-01 2.06270352e-01
5.67242503e-01 -2.44157553e-01 -2.41874784e-01 4.54836749e-02
7.79572368e-01 -3.98012072e-01 -7.72620976e-01 1.19160497e+00
5.75536609e-01 2.97983468e-01 6.17660463e-01 -1.02598560e+00
-9.99812663e-01 -2.36668691e-01 -7.08810091e-01 4.64201510e-01
2.32409284e-01 5.46200573e-01 6.82054877e-01 -8.85813236e-01
1.91437066e-01 1.24705875e+00 5.33630967e-01 3.70630562e-01
-1.03686559e+00 -9.14311647e-01 1.13104090e-01 5.74964404e-01
-1.21530473e+00 -5.69231331e-01 8.29977930e-01 -2.77718186e-01
6.87762618e-01 -7.37439543e-02 4.80801255e-01 9.17705178e-01
5.51807344e-01 3.39092731e-01 1.56392789e+00 -2.60611087e-01
-2.38984346e-01 -1.55304745e-01 1.42225195e-02 6.92905486e-01
5.60747623e-01 5.54917753e-01 -6.30770206e-01 4.52186882e-01
6.12253249e-01 2.34726250e-01 -8.73508990e-01 1.07512683e-01
-9.74588156e-01 3.71374249e-01 1.02304125e+00 2.13726133e-01
-2.70892471e-01 1.70951843e-01 1.39098406e-01 2.50120431e-01
4.57137734e-01 2.41553873e-01 -1.55566469e-01 4.91643935e-01
-6.74702227e-01 -8.80029704e-03 1.16069816e-01 6.42202735e-01
1.18075430e+00 3.20036203e-01 -4.71494704e-01 7.78534949e-01
2.94532597e-01 6.88312948e-01 3.79132688e-01 -7.92900920e-01
2.65476048e-01 4.28140849e-01 1.24777146e-01 -1.20590758e+00
-4.82293785e-01 -7.00832725e-01 -1.07965565e+00 5.13375938e-01
1.87801495e-01 8.17753822e-02 -1.02982533e+00 1.54157686e+00
1.05268462e-02 4.73307639e-01 1.67489424e-01 1.26203501e+00
1.21353638e+00 9.73158956e-01 -1.13991484e-01 -4.30043280e-01
1.07055831e+00 -9.62530196e-01 -9.33734834e-01 -3.55612338e-01
9.92838144e-02 -8.35313559e-01 1.16939414e+00 5.12965679e-01
-1.00825322e+00 -1.10388517e+00 -1.35800421e+00 -2.55504370e-01
-2.79053748e-01 2.90144891e-01 6.57603562e-01 4.80154753e-01
-9.21644688e-01 5.89802623e-01 -4.29902554e-01 -4.84133735e-02
5.91012478e-01 -8.52362141e-02 -2.32715264e-01 -6.78758979e-01
-1.24958968e+00 7.74638176e-01 2.29651690e-01 5.44381142e-01
-8.90182853e-01 -5.55048823e-01 -7.00581968e-01 1.00143440e-01
-3.71306911e-02 -3.70116174e-01 5.54956079e-01 -9.20863330e-01
-1.34337044e+00 6.50801957e-01 -1.25989527e-01 -9.71915647e-02
1.03618577e-01 6.46283701e-02 -5.37216842e-01 3.19110781e-01
6.51449710e-03 4.53424245e-01 9.97349560e-01 -1.45528138e+00
-5.27824879e-01 -2.33117595e-01 3.50457340e-01 3.16022217e-01
-1.70856476e-01 -3.94741595e-01 -5.42974830e-01 -5.14960706e-01
4.12582643e-02 -8.01389754e-01 1.96536809e-01 3.26398671e-01
-1.31473780e-01 -9.18010063e-03 7.44576037e-01 -6.04128957e-01
7.43924260e-01 -2.32317162e+00 -2.92269260e-01 -2.84195721e-01
5.58479503e-02 5.74325919e-01 -3.10822248e-01 1.07416429e-01
-1.17522523e-01 -1.78744212e-01 1.08561136e-01 -1.01820245e-01
-4.79075789e-01 2.59004265e-01 -2.20389500e-01 9.38074827e-01
3.76409501e-01 6.27445817e-01 -7.90583491e-01 -2.70866126e-01
6.13297462e-01 8.64728272e-01 -3.43171269e-01 2.36502916e-01
5.78937493e-02 6.76095128e-01 -2.28104979e-01 4.77803707e-01
1.13730609e+00 3.23958024e-02 -2.50708669e-01 -7.90055335e-01
-2.58164972e-01 1.27219945e-01 -9.22281802e-01 1.56843090e+00
-8.92637551e-01 1.11890900e+00 -7.48441666e-02 -7.98552930e-01
9.25920188e-01 4.39545698e-02 1.47723064e-01 -1.12933671e+00
2.09661782e-01 4.45796959e-02 6.35272115e-02 -7.54489541e-01
1.00958273e-01 -6.08979911e-02 4.35969114e-01 -9.08713862e-02
4.48134430e-02 -7.28617609e-02 8.08964223e-02 -3.14565212e-01
6.15028739e-01 1.61360204e-01 1.48330450e-01 -1.64872229e-01
7.63016522e-01 -6.70834720e-01 6.88770115e-01 5.77584565e-01
-2.30273008e-01 5.45129120e-01 2.10572276e-02 -4.68379468e-01
-4.78459120e-01 -1.05474794e+00 -4.07038361e-01 8.27205241e-01
6.53377533e-01 4.09009904e-02 -3.08853567e-01 -2.53511548e-01
-1.77370206e-01 5.57419896e-01 -3.47354770e-01 -5.07992864e-01
-3.73108685e-01 -8.39540958e-01 3.17977130e-01 1.70161635e-01
1.23312652e+00 -8.46625745e-01 -5.36354840e-01 -2.00611576e-01
-4.16620225e-01 -1.38547945e+00 -1.66063055e-01 -4.19339240e-02
-4.74474639e-01 -1.14194453e+00 -4.04598743e-01 -7.32873380e-01
5.05157351e-01 8.59688222e-01 7.92562723e-01 1.09048396e-01
-6.61805391e-01 8.10766667e-02 -3.91430229e-01 -4.45768028e-01
-7.01197237e-02 -2.80538827e-01 -1.64899290e-01 3.80678207e-01
3.65613699e-02 -6.16546333e-01 -1.15186870e+00 2.30093792e-01
-9.88857329e-01 1.33953005e-01 6.63530171e-01 8.09475780e-01
5.31942904e-01 3.82451206e-01 5.28340936e-01 -6.54494166e-01
3.34357679e-01 -1.91990688e-01 -4.58677500e-01 4.37562503e-02
-7.55337715e-01 -7.82301873e-02 8.49972546e-01 -3.97440344e-01
-1.32559097e+00 -2.23068401e-01 1.18813112e-01 -4.06804651e-01
-1.41103759e-01 1.21569507e-01 -4.09343243e-01 -5.25318801e-01
7.82326818e-01 2.97115862e-01 -2.74846703e-01 -3.13859135e-01
2.23856106e-01 6.52288675e-01 5.89252651e-01 -2.01566786e-01
9.19005692e-01 7.43343711e-01 3.61192435e-01 -9.02512014e-01
-1.14496851e+00 -4.05338675e-01 -1.20703064e-01 -4.67945933e-01
6.99764311e-01 -1.30956185e+00 -7.03863800e-01 5.53792238e-01
-9.68545854e-01 -2.69614249e-01 -7.98386559e-02 5.17982125e-01
-3.25677067e-01 3.29412282e-01 -6.16172612e-01 -4.54913318e-01
-2.11072788e-01 -1.21015942e+00 7.88565040e-01 7.38683939e-01
6.05755627e-01 -7.79789925e-01 -1.63698301e-01 2.81239837e-01
6.03277743e-01 8.49465355e-02 9.18752313e-01 2.54770160e-01
-7.31884241e-01 5.96067943e-02 -7.38009393e-01 7.39222229e-01
3.24375421e-01 -3.32214236e-02 -1.62694347e+00 -5.47236681e-01
2.23081082e-01 -1.25609860e-01 1.07894647e+00 4.78864044e-01
1.50270593e+00 -1.29104316e-01 -5.08506969e-02 1.18861151e+00
1.76057875e+00 -1.23208249e-02 1.09603548e+00 2.44472370e-01
6.15446806e-01 4.04639482e-01 6.18179560e-01 1.80576384e-01
1.46548161e-02 5.77424109e-01 7.65673339e-01 -5.41124582e-01
-9.21888590e-01 1.13475867e-01 2.99430370e-01 5.09546936e-01
-2.38351569e-01 -5.01115441e-01 -4.42214757e-01 4.77028787e-01
-1.43230247e+00 -9.74672794e-01 -2.33957797e-01 1.95389795e+00
7.67140508e-01 -6.58029541e-02 -5.84677339e-01 1.89817593e-01
7.27634668e-01 5.37146807e-01 -5.79690814e-01 -1.55641302e-01
-4.09569472e-01 3.58774215e-01 6.29898131e-01 3.13259184e-01
-9.13158476e-01 5.95702171e-01 6.22134399e+00 7.38641262e-01
-1.67902291e+00 6.82319030e-02 5.53177416e-01 -1.34537280e-01
-1.39242634e-02 -1.01045758e-01 -5.97152889e-01 4.85392243e-01
5.68054497e-01 5.52165471e-02 6.27470851e-01 3.44009221e-01
4.35623437e-01 -9.35633853e-02 -8.56894433e-01 1.27217793e+00
3.10913384e-01 -9.58102763e-01 4.42722514e-02 -4.81808670e-02
9.39464033e-01 2.28594244e-01 3.78815413e-01 1.61314040e-01
-2.57425249e-01 -1.06863654e+00 2.26896301e-01 8.38447809e-01
1.02795362e+00 -6.35189116e-01 6.21466577e-01 1.63641140e-01
-1.06398928e+00 -1.74520910e-01 -6.97762609e-01 -1.59092098e-01
-2.61302829e-01 7.65461981e-01 -3.71937484e-01 7.09351540e-01
8.84959340e-01 1.23388493e+00 -6.33263052e-01 1.10731578e+00
-4.73065734e-01 5.59996784e-01 2.10350245e-01 2.95376033e-01
-2.81743314e-02 -2.09339499e-01 3.30561966e-01 1.17972600e+00
2.76210368e-01 4.26148698e-02 1.48715079e-01 9.05058026e-01
-2.13596091e-01 -2.51728594e-01 -5.88885427e-01 2.50850201e-01
4.24995758e-02 1.35797131e+00 -3.13888460e-01 7.94936270e-02
-6.66584611e-01 8.02391469e-01 -7.33211711e-02 7.95810103e-01
-7.42203534e-01 -5.97253740e-01 6.62344933e-01 2.11926460e-01
8.50170329e-02 -1.96775272e-01 1.63534619e-02 -1.16705811e+00
5.50118089e-02 -6.69133902e-01 1.10713241e-03 -1.44044185e+00
-1.42928803e+00 4.80356097e-01 -1.15568921e-01 -1.38081253e+00
2.94929832e-01 -8.37748230e-01 -5.63156486e-01 1.10040927e+00
-2.41327000e+00 -1.03122413e+00 -9.46225405e-01 8.29691768e-01
5.15617907e-01 -3.64702642e-02 4.56431627e-01 5.01492500e-01
-4.32219088e-01 3.18064660e-01 2.08276361e-01 1.39365252e-02
9.42740202e-01 -7.14157641e-01 -1.76178947e-01 1.12547815e+00
-1.19915709e-01 4.04158533e-01 6.43139124e-01 -1.28011718e-01
-1.33682680e+00 -1.38680911e+00 2.41330191e-01 2.19330087e-01
1.62972689e-01 -1.35404125e-01 -9.86954749e-01 3.04449558e-01
3.63153636e-01 6.44974351e-01 3.18531215e-01 -2.24817872e-01
-5.81451416e-01 -7.34974146e-01 -1.10002887e+00 4.67011005e-01
9.37438905e-01 -7.01735795e-01 -3.08860958e-01 3.78219694e-01
5.54001033e-01 -1.69942737e-01 -4.73236859e-01 5.21570265e-01
3.74015033e-01 -1.28583050e+00 1.17458534e+00 -1.07567720e-02
4.70529675e-01 -5.05282104e-01 -2.65700161e-01 -1.62280226e+00
-4.44918066e-01 -2.61354029e-01 1.42998919e-01 1.17039859e+00
-1.48528263e-01 -9.84721363e-01 2.46242061e-01 -1.20576680e-01
-2.43945062e-01 -5.98849416e-01 -7.49726117e-01 -6.43800676e-01
-6.10200465e-02 -9.51208025e-02 4.53961641e-01 7.00215995e-01
-5.98200798e-01 3.52062494e-01 -2.42729902e-01 5.28814852e-01
6.97010100e-01 3.13461483e-01 5.95998645e-01 -1.23091531e+00
-1.27713397e-01 -1.43917114e-01 -5.84280252e-01 -1.05359530e+00
3.23154122e-01 -8.05330992e-01 2.69207418e-01 -1.68569636e+00
1.38991699e-01 -3.38190436e-01 -5.90858400e-01 3.95811707e-01
-1.21213980e-01 6.01536751e-01 1.59726605e-01 3.16044927e-01
-4.00891244e-01 7.24292338e-01 1.57144845e+00 -3.98466408e-01
1.06342234e-01 -1.28065288e-01 -7.23645687e-01 6.51398957e-01
6.28184378e-01 -2.06837699e-01 -6.79417729e-01 -6.46380723e-01
4.08964545e-01 -1.92773417e-01 4.81150180e-01 -1.27960157e+00
1.16535919e-02 -2.93535560e-01 7.27613389e-01 -3.61241490e-01
3.35252672e-01 -7.35612929e-01 -1.59439221e-01 3.19811344e-01
-2.91549861e-01 -4.93314952e-01 2.72364795e-01 6.88158691e-01
-3.63016665e-01 -1.09464690e-01 1.21126747e+00 -4.92799133e-02
-6.90036595e-01 3.72588336e-01 -1.41440168e-01 3.38604152e-02
7.42223084e-01 -2.00298354e-01 -9.12079692e-01 -3.26962024e-01
-2.41876349e-01 -1.22085050e-01 4.49500620e-01 3.39529663e-01
7.25079834e-01 -1.30634677e+00 -6.24587357e-01 5.10120988e-01
3.48186195e-01 -1.98751353e-02 6.32293701e-01 7.82776058e-01
-6.33570075e-01 1.39649123e-01 -6.47688866e-01 -6.00562871e-01
-9.70117211e-01 4.81289178e-01 6.27635717e-01 1.31525785e-01
-6.02266610e-01 5.08747935e-01 6.16480649e-01 4.74916361e-02
-4.09252681e-02 -3.62554729e-01 -4.12529886e-01 -3.46510202e-01
6.58759058e-01 3.05375576e-01 2.97534168e-01 -7.37872839e-01
-1.89077944e-01 8.39510798e-01 1.35755002e-01 4.44447994e-01
1.34410799e+00 -4.34884310e-01 -2.08339207e-02 3.82480949e-01
1.49630773e+00 -3.11450567e-02 -1.41688752e+00 -3.43106419e-01
-8.90091479e-01 -9.76488471e-01 6.23405516e-01 -6.60284936e-01
-1.47015512e+00 1.30942440e+00 1.21973789e+00 -1.53191518e-02
1.64901459e+00 -3.88893902e-01 6.44659281e-01 3.96198004e-01
2.03304157e-01 -8.28179419e-01 2.63150483e-01 2.53405303e-01
8.64504397e-01 -1.30102134e+00 1.61791965e-01 -5.19324005e-01
-2.29249686e-01 1.19133449e+00 5.35538137e-01 -3.05297047e-01
5.42155683e-01 -1.46961128e-02 8.15131962e-02 -1.06715307e-01
-4.86075580e-01 -3.95093620e-01 4.55430001e-01 9.02251184e-01
4.41595793e-01 -2.65727401e-01 -2.34541632e-02 2.20104471e-01
2.13071764e-01 -7.46872053e-02 6.57294035e-01 3.66903931e-01
-4.60400403e-01 -4.73769665e-01 -2.20726311e-01 2.77258545e-01
-4.72462893e-01 -5.99828549e-02 1.46933168e-01 6.05522096e-01
6.41212523e-01 1.31602287e+00 4.65238132e-02 -1.50996849e-01
3.75306398e-01 -3.93226355e-01 6.83806598e-01 -5.32402575e-01
-2.14856490e-01 2.13221703e-02 -3.10799927e-01 -7.18223214e-01
-7.93530285e-01 -1.94840983e-01 -1.13372493e+00 -1.85303781e-02
-4.00417060e-01 -4.95202482e-01 6.06426299e-01 6.97809935e-01
3.35541904e-01 8.55857491e-01 9.63501990e-01 -1.14430702e+00
-1.35774687e-01 -8.37184131e-01 -6.97822928e-01 6.62746787e-01
6.85016155e-01 -1.00080764e+00 -6.82593107e-01 -2.39259079e-02] | [10.834203720092773, -2.5671029090881348] |
3beda96a-b795-4e29-a1b6-88a764750d74 | cross-domain-3d-equivariant-image-embeddings | 1812.02716 | null | https://arxiv.org/abs/1812.02716v2 | https://arxiv.org/pdf/1812.02716v2.pdf | Cross-Domain 3D Equivariant Image Embeddings | Spherical convolutional networks have been introduced recently as tools to learn powerful feature representations of 3D shapes. Spherical CNNs are equivariant to 3D rotations making them ideally suited to applications where 3D data may be observed in arbitrary orientations. In this paper we learn 2D image embeddings with a similar equivariant structure: embedding the image of a 3D object should commute with rotations of the object. We introduce a cross-domain embedding from 2D images into a spherical CNN latent space. This embedding encodes images with 3D shape properties and is equivariant to 3D rotations of the observed object. The model is supervised only by target embeddings obtained from a spherical CNN pretrained for 3D shape classification. We show that learning a rich embedding for images with appropriate geometric structure is sufficient for tackling varied applications, such as relative pose estimation and novel view synthesis, without requiring additional task-specific supervision. | ['Ameesh Makadia', 'Avneesh Sud', 'Zhengyi Luo', 'Kostas Daniilidis', 'Carlos Esteves'] | 2018-12-06 | null | null | null | null | ['3d-shape-retrieval'] | ['computer-vision'] | [-3.27357836e-02 4.31060702e-01 -2.64530182e-01 -5.43817699e-01
-1.31379411e-01 -9.72243071e-01 9.46653008e-01 -2.71177381e-01
-9.24269110e-02 -2.67651957e-02 4.03359652e-01 -9.54744890e-02
1.77387983e-01 -5.68196058e-01 -1.05991793e+00 -7.22253561e-01
-5.27156815e-02 8.27907979e-01 -1.16852187e-01 -1.50197059e-01
3.09328213e-02 1.33171260e+00 -1.26476336e+00 -2.05259733e-02
1.00537436e-02 7.00555563e-01 -1.24170728e-01 8.04191053e-01
1.90794006e-01 1.01681516e-01 -2.29704127e-01 -7.97587633e-02
6.51637912e-01 -9.18665994e-03 -5.90142369e-01 5.95374465e-01
6.08587384e-01 -2.85345942e-01 -5.31176209e-01 7.32926190e-01
3.50914747e-01 -7.29290694e-02 1.15482545e+00 -1.17342663e+00
-1.00795937e+00 -4.06502523e-02 -1.89648375e-01 -3.24844062e-01
3.91497940e-01 -2.55524307e-01 1.14754736e+00 -1.18320441e+00
1.02350128e+00 1.42357993e+00 4.19961065e-01 5.41627765e-01
-1.60277343e+00 1.60847530e-01 -2.10616607e-02 -3.14892888e-01
-1.23474538e+00 -4.26903993e-01 1.07425356e+00 -5.45071959e-01
1.08434153e+00 8.84554535e-02 5.98825634e-01 1.18526888e+00
2.88358629e-01 6.33528948e-01 6.12322032e-01 -3.37292373e-01
6.05528094e-02 9.50705782e-02 -5.45488656e-01 7.10938513e-01
2.36687109e-01 -1.08074233e-01 -2.45554596e-01 1.56217456e-01
1.50378346e+00 3.39591920e-01 2.32917722e-02 -1.55517304e+00
-1.44413245e+00 9.74995077e-01 7.30355978e-01 1.23621419e-01
-1.09603830e-01 1.04261704e-01 8.98102373e-02 3.09841484e-01
6.44506872e-01 6.30551815e-01 -6.46378815e-01 1.20632529e-01
-2.33466551e-01 2.94802159e-01 7.54963219e-01 1.37569487e+00
5.97946465e-01 5.04280210e-01 4.22130197e-01 5.39461017e-01
3.41219008e-01 8.47384453e-01 2.78314441e-01 -1.09537745e+00
1.30270168e-01 7.27304935e-01 2.68449485e-02 -1.01434255e+00
-5.23177624e-01 -2.35979423e-01 -8.21440160e-01 3.80827039e-01
2.79336035e-01 3.12499881e-01 -8.98773849e-01 1.73108602e+00
1.99701130e-01 -2.87389100e-01 1.20170571e-01 7.28128493e-01
6.63230598e-01 3.59031200e-01 -7.25053132e-01 2.84080505e-01
1.27558362e+00 -4.61361557e-01 -1.79048985e-01 -6.14470802e-03
5.64190030e-01 -8.51920605e-01 8.64441097e-01 -8.97479802e-02
-1.20034063e+00 -3.86308372e-01 -1.16752839e+00 -4.26117331e-01
-4.85907167e-01 8.83188322e-02 2.86438107e-01 4.16649848e-01
-1.07435513e+00 2.03508183e-01 -8.29417586e-01 -5.83929598e-01
3.16901773e-01 4.63947952e-01 -9.22825098e-01 1.37136593e-01
-3.96951318e-01 1.03816664e+00 9.61101577e-02 6.50418783e-03
-9.78984535e-01 -5.85384727e-01 -1.43427396e+00 -8.62405375e-02
1.00064009e-01 -9.59289730e-01 1.18238318e+00 -6.40387356e-01
-1.46042526e+00 1.51177132e+00 -1.99492648e-01 -1.84556946e-01
4.65728670e-01 -8.47204924e-02 -1.76997203e-02 8.66665095e-02
5.90999611e-02 6.15109921e-01 1.38969445e+00 -1.21892667e+00
3.06810260e-01 -7.98870027e-01 4.37344551e-01 3.66463542e-01
3.43801640e-02 -3.01743358e-01 -1.20161623e-01 -7.21687436e-01
7.49831736e-01 -1.22421360e+00 -2.21319363e-01 3.50771606e-01
-3.94991875e-01 4.05930988e-02 1.16928601e+00 -4.21422534e-02
-2.23408937e-02 -2.00993228e+00 6.21634901e-01 9.63076577e-02
3.48487616e-01 -1.61098108e-01 -2.92629153e-01 3.75455737e-01
-4.77175891e-01 7.56473765e-02 9.85219777e-02 -4.76023972e-01
2.46570900e-01 4.11592782e-01 -3.94508243e-01 9.30377960e-01
6.86433256e-01 1.17595840e+00 -7.16773748e-01 1.06182791e-01
3.48566502e-01 7.90830493e-01 -7.95622170e-01 5.50954759e-01
-1.69343367e-01 4.58318979e-01 -3.86820853e-01 3.68133873e-01
6.05691373e-01 -3.14222097e-01 -6.14158027e-02 -4.37499285e-01
8.82928595e-02 2.85917848e-01 -1.10763752e+00 1.64625454e+00
-6.60145462e-01 5.47563851e-01 5.39873540e-02 -1.14159799e+00
1.15897000e+00 3.68743271e-01 3.53682458e-01 -2.02252045e-01
2.31557474e-01 1.79126635e-01 -3.03259909e-01 -3.15297425e-01
3.45259279e-01 -2.97197938e-01 -1.17968835e-01 4.59927976e-01
4.71066564e-01 -7.54254341e-01 -5.05017340e-01 9.84536037e-02
6.27623498e-01 2.54707307e-01 3.25207740e-01 -3.66418839e-01
3.69914114e-01 -5.36198974e-01 1.14640735e-01 2.11273611e-01
2.52927601e-01 8.65750670e-01 5.53295374e-01 -8.62693310e-01
-1.77231085e+00 -1.71109307e+00 -2.01160520e-01 5.55025280e-01
2.23052856e-02 -9.45933461e-02 -3.95159572e-01 -6.21750951e-01
1.12392828e-01 8.79192278e-02 -7.81177163e-01 -2.12950230e-01
-6.22232795e-01 -1.57210734e-02 1.39654309e-01 7.17939317e-01
7.13221952e-02 -6.77134752e-01 -4.92616773e-01 -1.26808822e-01
3.22135538e-01 -1.53155816e+00 -8.52648079e-01 3.30732912e-01
-1.14284039e+00 -1.07897115e+00 -8.76742542e-01 -1.07559919e+00
1.30818772e+00 5.92318058e-01 9.50913250e-01 -4.78571683e-01
-2.95313746e-01 9.56328750e-01 -2.23548591e-01 -3.55593234e-01
-3.06705266e-01 9.05493349e-02 5.67100644e-01 1.31963357e-01
2.03989968e-01 -8.80152643e-01 -2.56153673e-01 3.85146260e-01
-9.51188922e-01 6.58146665e-02 3.22369367e-01 8.39858055e-01
5.50798893e-01 -6.12788856e-01 1.24102302e-01 -7.07573652e-01
1.31495655e-01 -8.65159333e-02 -6.61454558e-01 -1.18259907e-01
8.37951377e-02 4.25969929e-01 7.05242276e-01 -3.74861777e-01
-5.32058120e-01 3.28353822e-01 3.19361910e-02 -7.11400151e-01
-3.24941903e-01 -6.73738644e-02 -4.09314990e-01 -1.39328614e-01
6.59583390e-01 2.30507195e-01 3.08366984e-01 -5.88183880e-01
6.48149014e-01 1.83744907e-01 3.61444801e-01 -5.06654918e-01
1.33550692e+00 8.67697120e-01 5.22091031e-01 -1.24391913e+00
-7.80727386e-01 -2.48928472e-01 -1.42694938e+00 2.14175463e-01
1.12183702e+00 -9.78671610e-01 -7.30462015e-01 2.71889120e-01
-1.32191622e+00 -1.43970074e-02 -6.35674179e-01 4.61428642e-01
-1.21948946e+00 1.42188430e-01 -2.07159922e-01 -2.26641953e-01
1.12653166e-01 -1.34839320e+00 1.53858197e+00 -1.77629605e-01
-2.13811964e-01 -1.27887154e+00 -1.06830094e-02 -7.91645050e-02
1.35899425e-01 3.76664132e-01 1.06884432e+00 -6.03795767e-01
-7.14045882e-01 -2.97500789e-01 -1.19126543e-01 4.13860768e-01
2.85279065e-01 -1.26821950e-01 -1.04537702e+00 -4.23537701e-01
1.95125431e-01 -1.94201976e-01 4.51557487e-01 5.23506820e-01
1.02941930e+00 -2.77919710e-01 5.85916936e-02 1.13220835e+00
1.16108143e+00 -1.68300480e-01 1.11484341e-01 1.24301687e-01
9.04156268e-01 6.05776608e-01 -9.35659781e-02 2.36512005e-01
2.42647335e-01 8.17548990e-01 7.75795281e-01 -1.32782489e-01
1.85977742e-01 -4.49882627e-01 3.91107470e-01 8.52363646e-01
-2.53811866e-01 1.68805435e-01 -6.85564399e-01 4.98976916e-01
-1.40754390e+00 -5.22424757e-01 2.19342023e-01 2.14916611e+00
3.35093915e-01 1.86944991e-01 3.43675092e-02 -8.29290822e-02
3.16154599e-01 5.02509832e-01 -6.34458184e-01 -6.18776441e-01
-4.04659003e-01 1.75820887e-01 4.27985102e-01 5.99996567e-01
-1.15217996e+00 8.73387992e-01 6.52519369e+00 1.38735309e-01
-1.24797058e+00 -2.25287318e-01 2.06406236e-01 1.80512577e-01
-7.31052279e-01 -3.14858668e-02 -6.79566920e-01 -3.04676086e-01
4.19034243e-01 -9.62232724e-02 1.92372948e-01 8.56689692e-01
2.21180841e-02 5.73617756e-01 -1.60469615e+00 1.03299344e+00
3.48509192e-01 -1.55966473e+00 5.37584841e-01 3.14576656e-01
9.15497184e-01 4.99617010e-02 5.13170123e-01 -1.89541817e-01
1.48794562e-01 -1.21723640e+00 8.08032632e-01 2.61049569e-01
9.86116588e-01 -5.30823112e-01 4.77112144e-01 2.24476680e-01
-1.13794434e+00 3.59345406e-01 -5.44029713e-01 1.61561575e-02
3.87061425e-02 7.60971010e-02 -1.06251085e+00 4.17554706e-01
4.75585043e-01 1.14889801e+00 -3.22840184e-01 4.79231864e-01
-2.25350738e-01 -1.28571853e-01 -3.11477989e-01 2.18795329e-01
3.31622958e-01 -4.49498981e-01 1.00059175e+00 8.20518851e-01
4.26273733e-01 -1.53253675e-01 -1.11997403e-01 9.91958976e-01
-2.99744248e-01 -1.33585379e-01 -1.59425938e+00 -7.85840899e-02
3.01266033e-02 1.06675017e+00 -6.87774301e-01 1.74615476e-02
-3.80298644e-01 9.97938216e-01 1.19679615e-01 5.45746922e-01
-4.03995663e-01 -2.15974435e-01 1.01601100e+00 2.91085362e-01
7.19168484e-01 -8.97076190e-01 -2.42318973e-01 -1.43243885e+00
3.28985974e-02 -4.45588410e-01 -2.32592255e-01 -1.15032339e+00
-1.06288874e+00 5.37735581e-01 9.17266309e-02 -1.51344240e+00
-3.02857369e-01 -1.31964898e+00 -5.34502089e-01 7.43944883e-01
-1.16977191e+00 -1.49862146e+00 -1.16127387e-01 6.32130682e-01
4.63891476e-01 -4.09651786e-01 1.09685683e+00 -3.83817226e-01
1.80914477e-01 3.81685197e-01 -9.70038678e-03 2.84776032e-01
6.22440040e-01 -1.62417758e+00 6.20786369e-01 4.30193901e-01
5.06861925e-01 7.99127221e-01 6.26441240e-01 -5.18334806e-02
-1.84284556e+00 -1.02515984e+00 8.35464239e-01 -9.51091349e-01
4.75887865e-01 -8.82179499e-01 -4.58198547e-01 1.21499383e+00
-2.21678428e-02 3.25481743e-01 5.71229994e-01 1.11262372e-03
-9.13572967e-01 -4.90722694e-02 -8.28257680e-01 7.64463067e-01
1.16846669e+00 -1.05897295e+00 -6.05598152e-01 3.63043517e-01
9.38520968e-01 -6.96069896e-01 -1.03946817e+00 2.77898550e-01
6.43508792e-01 -5.74864149e-01 1.40836930e+00 -1.07555664e+00
2.95346051e-01 -2.86466092e-01 -4.98145372e-01 -1.42433262e+00
-2.44327560e-01 -5.19256413e-01 -4.60053869e-02 4.21134651e-01
3.05722058e-01 -6.01385236e-01 8.38600218e-01 1.25154108e-01
-2.37539485e-01 -6.08797908e-01 -9.23187315e-01 -8.53145897e-01
1.71323836e-01 -2.70800203e-01 5.55765629e-01 9.13372755e-01
-3.81641448e-01 4.33337599e-01 -1.43361270e-01 4.06719834e-01
4.97560799e-01 2.85056889e-01 1.10217035e+00 -1.24475491e+00
1.82082415e-01 -4.35757011e-01 -1.09442258e+00 -1.56720746e+00
3.41901064e-01 -1.26557541e+00 -3.02945167e-01 -9.74432111e-01
9.48924422e-02 -1.56273693e-02 2.79104471e-01 1.68616802e-01
5.78585744e-01 3.99871409e-01 1.87214971e-01 -9.44931060e-02
-2.70465732e-01 9.40172076e-01 1.59248209e+00 -2.80885436e-02
-2.95919157e-03 8.72311369e-02 -4.58173394e-01 9.56974030e-01
4.58164871e-01 -2.35060722e-01 -4.24299121e-01 -6.60292506e-01
4.34545368e-01 -2.31274784e-01 5.89316070e-01 -3.68676931e-01
-4.09974791e-02 -5.83943464e-02 5.68654120e-01 -6.18665516e-01
7.36385286e-01 -1.13880014e+00 1.53677613e-02 8.05790648e-02
-2.98170567e-01 2.55898625e-01 1.27343740e-02 4.82048154e-01
-2.25666061e-01 -6.80251867e-02 8.50000024e-01 -2.15268508e-01
-2.99557984e-01 6.96714282e-01 -1.38591215e-01 1.98076084e-01
9.49198425e-01 -6.46756947e-01 8.19154903e-02 -6.34209156e-01
-1.12420905e+00 -1.92881152e-01 1.00595903e+00 7.02527463e-01
9.42801774e-01 -1.80758905e+00 -4.80467826e-01 9.19081628e-01
3.80380929e-01 3.92035723e-01 -3.14836390e-02 5.45434833e-01
-7.80727804e-01 6.75824642e-01 -4.22607034e-01 -1.14753902e+00
-1.18092299e+00 4.71299917e-01 5.97071290e-01 2.64779866e-01
-6.66095793e-01 7.51197100e-01 7.95943141e-01 -1.27949882e+00
3.56558450e-02 -6.72621965e-01 7.50498772e-02 -1.29805714e-01
1.34167686e-01 -9.01460797e-02 9.46243256e-02 -1.15813017e+00
-2.22101629e-01 1.08359289e+00 1.22101948e-01 3.20062265e-02
1.73539639e+00 1.09373517e-01 -9.15726200e-02 6.66742384e-01
1.92648530e+00 1.76931955e-02 -1.34509575e+00 -4.52992380e-01
-2.28796989e-01 -4.93178397e-01 -2.23572791e-01 1.37852460e-01
-7.98183799e-01 1.13851428e+00 7.41966292e-02 1.30189195e-01
5.47733307e-01 4.73437577e-01 2.55281568e-01 5.78490376e-01
1.99849144e-01 -5.40606618e-01 5.05057991e-01 9.80123639e-01
1.42248452e+00 -1.22952390e+00 -1.69523329e-01 -1.72743961e-01
-3.79218668e-01 1.48594797e+00 2.37466007e-01 -6.56632900e-01
1.17919433e+00 -7.95417875e-02 -1.35920525e-01 -3.19273591e-01
-5.62474906e-01 5.41586094e-02 8.16598296e-01 8.52791965e-01
1.93348542e-01 1.85551211e-01 5.86054325e-01 5.73477754e-03
-3.07362318e-01 -7.91581452e-01 6.74834251e-01 6.63325429e-01
-2.25067332e-01 -1.00267553e+00 -2.64820933e-01 7.27226809e-02
-5.44581451e-02 3.35134089e-01 -4.66070265e-01 8.11773539e-01
-1.67815089e-01 1.10862581e-02 3.72037947e-01 -2.00220600e-01
6.22727096e-01 8.96151550e-03 9.75537181e-01 -9.79815423e-01
3.34817201e-01 1.61153182e-01 -3.48977000e-01 -5.50854027e-01
-5.67202806e-01 -7.71737933e-01 -9.36461747e-01 -7.36552849e-02
1.99649572e-01 -2.50486791e-01 7.20325291e-01 6.99326813e-01
2.49963060e-01 1.11696146e-01 9.14273322e-01 -1.41844487e+00
-6.25138700e-01 -5.73531091e-01 -7.82547414e-01 4.07713145e-01
8.06460321e-01 -9.29871559e-01 -4.81547207e-01 2.56624579e-01] | [8.805188179016113, 2.3417704105377197] |
a8bce9d7-7272-4ca8-940a-9a98602cef0f | neural-temporal-relation-extraction | null | null | https://aclanthology.org/E17-2118 | https://aclanthology.org/E17-2118.pdf | Neural Temporal Relation Extraction | We experiment with neural architectures for temporal relation extraction and establish a new state-of-the-art for several scenarios. We find that neural models with only tokens as input outperform state-of-the-art hand-engineered feature-based models, that convolutional neural networks outperform LSTM models, and that encoding relation arguments with XML tags outperforms a traditional position-based encoding. | ['Dmitriy Dligach', 'Guergana Savova', 'Chen Lin', 'Timothy Miller', 'Steven Bethard'] | 2017-04-01 | null | null | null | eacl-2017-4 | ['temporal-relation-extraction', 'temporal-information-extraction'] | ['natural-language-processing', 'natural-language-processing'] | [ 1.86219946e-01 8.51804972e-01 -8.27157736e-01 -5.84606886e-01
-8.04108322e-01 -3.86809647e-01 8.52880180e-01 5.74213803e-01
-5.23052037e-01 6.93359792e-01 4.03654814e-01 -9.36757386e-01
-4.26827967e-01 -1.19345176e+00 -9.33723152e-01 2.54932255e-01
-8.17517936e-01 7.69177437e-01 5.89419663e-01 -5.93852520e-01
-3.98877829e-01 2.62693346e-01 -1.04369330e+00 9.52293873e-01
5.65426759e-02 1.59018207e+00 -8.62972498e-01 5.92290401e-01
-4.26596761e-01 1.47083080e+00 -5.90402603e-01 -7.44452417e-01
-1.33250788e-01 1.77279949e-01 -1.49349475e+00 -9.15446877e-01
-5.85752819e-03 -3.37005973e-01 -1.19668043e+00 5.54134250e-01
1.48146734e-01 1.75301582e-02 2.21524447e-01 -9.53895390e-01
-1.18186772e+00 1.43593264e+00 -2.28003860e-02 5.67011654e-01
6.08523667e-01 -3.00145566e-01 1.43812466e+00 -3.18081200e-01
1.22037649e+00 1.35794592e+00 1.04215324e+00 1.61690280e-01
-1.43383110e+00 -2.75303692e-01 3.60235691e-01 3.84089679e-01
-8.46398473e-01 -5.39905667e-01 4.30368483e-01 -2.02523708e-01
2.59558463e+00 1.44230813e-01 6.32268667e-01 1.26197410e+00
6.20779932e-01 7.30306923e-01 3.18730652e-01 -4.66016710e-01
-2.72090435e-01 -5.59844553e-01 6.87249184e-01 7.18854606e-01
2.48743132e-01 6.46660268e-01 -6.86268151e-01 -4.84483063e-01
3.95140946e-01 -3.36255074e-01 3.13936114e-01 5.17654419e-02
-1.19257367e+00 7.88593411e-01 5.76965392e-01 7.01842785e-01
-4.40207392e-01 7.32162476e-01 9.85833704e-01 5.89980960e-01
6.71530902e-01 4.51652527e-01 -1.43660808e+00 -3.84191990e-01
-5.82508504e-01 4.32540387e-01 1.13590419e+00 1.05186939e+00
3.13183665e-01 -1.41271740e-01 -3.63875389e-01 5.36081612e-01
5.88448308e-02 -2.24320278e-01 5.12422740e-01 -8.88132453e-01
7.22710788e-01 7.08242834e-01 -3.37014794e-01 -4.56440926e-01
-6.29581511e-01 -3.02479327e-01 -3.55322957e-01 -2.50618130e-01
1.72117278e-01 -2.78150439e-01 -8.71196866e-01 1.67207122e+00
-7.43791759e-02 -9.99891758e-02 5.41490257e-01 7.80423507e-02
1.22067678e+00 5.76822042e-01 2.86228627e-01 -3.39074910e-01
1.55717123e+00 -7.81562507e-01 -1.00444472e+00 -5.45223951e-01
9.68650520e-01 -1.97805852e-01 3.76710236e-01 -5.68922935e-03
-1.37726927e+00 -3.35400492e-01 -1.17380512e+00 -2.98496842e-01
-9.30354714e-01 -2.27912307e-01 1.46058500e+00 3.04535598e-01
-8.78188789e-01 1.15148830e+00 -8.33448350e-01 -5.33219755e-01
5.28896153e-01 6.37110114e-01 -6.28893673e-01 5.93598127e-01
-1.93480730e+00 1.20354009e+00 1.21806121e+00 2.07499519e-01
-4.64312673e-01 -6.90812111e-01 -1.28628349e+00 1.28995344e-01
4.18970525e-01 -6.14404857e-01 1.79341841e+00 -1.48281381e-01
-1.31618392e+00 6.69083118e-01 -2.77032375e-01 -1.34549153e+00
-6.23332188e-02 -4.27026421e-01 -9.01072919e-01 -1.00181811e-01
-1.04238898e-01 4.60650265e-01 4.39853966e-02 -7.26439059e-01
-6.49542451e-01 1.81118906e-01 3.27137768e-01 -4.02083457e-01
-2.52725869e-01 5.36752403e-01 1.72975942e-01 -5.26719451e-01
3.64201777e-02 -4.39937830e-01 -2.39017338e-01 -3.42886597e-01
-6.27474427e-01 -7.37757564e-01 8.24183524e-01 -6.81964576e-01
1.54052591e+00 -1.74696922e+00 -3.38681847e-01 7.34733865e-02
1.42753333e-01 2.17310995e-01 -1.80034295e-01 6.63985789e-01
-5.51896513e-01 4.15960133e-01 2.09532797e-01 6.62379414e-02
3.56124938e-01 4.05972719e-01 -3.17567617e-01 -6.24021739e-02
7.77456820e-01 1.66717851e+00 -1.05488443e+00 -8.03340554e-01
-2.22839460e-01 3.35200518e-01 -2.51418680e-01 1.95241570e-01
-6.75942659e-01 -4.61630076e-01 -4.35454130e-01 6.91356778e-01
-8.14113691e-02 -3.81760508e-01 7.76313007e-01 -3.24374050e-01
1.92309782e-01 1.36542535e+00 -5.07251203e-01 1.64818370e+00
-3.31145257e-01 9.21211064e-01 -7.43068278e-01 -8.55764389e-01
7.87889183e-01 9.42755580e-01 6.35316253e-01 -8.00443649e-01
6.85298666e-02 3.20941359e-01 4.18858916e-01 -5.30005515e-01
6.48541152e-01 2.38791794e-01 -3.34465742e-01 4.45482075e-01
5.14425874e-01 3.21595073e-01 2.98011273e-01 9.23021957e-02
1.71980321e+00 5.20870566e-01 5.83736777e-01 1.31190673e-01
-2.03752108e-02 -5.65525405e-02 5.88733017e-01 5.20573616e-01
7.31941015e-02 -2.18422040e-01 8.49025071e-01 -1.24263930e+00
-6.62255824e-01 -6.43999696e-01 -2.66319364e-01 1.29390514e+00
-6.55478537e-01 -9.85360682e-01 -3.88427824e-01 -1.06641448e+00
1.25981539e-01 5.77572942e-01 -1.00976837e+00 -1.02736928e-01
-1.04620504e+00 -5.79737663e-01 1.05041552e+00 1.18261588e+00
2.94598550e-01 -1.48814046e+00 -7.77083457e-01 9.13548231e-01
-4.70048934e-02 -1.39949453e+00 2.69731432e-01 1.16055799e+00
-8.63412023e-01 -1.10747480e+00 8.89248177e-02 -7.84579515e-01
-1.63008377e-01 -8.50211203e-01 1.75953281e+00 -1.83687985e-01
9.58862677e-02 -4.45558280e-01 -3.90473515e-01 -1.99395388e-01
-2.97660500e-01 6.57440722e-01 -3.75018477e-01 -8.68252218e-01
7.01954067e-01 -6.69456780e-01 -1.28283501e-02 -1.97944120e-01
-6.35428607e-01 -4.58791107e-01 4.31744069e-01 8.43117237e-01
3.04111391e-01 9.89689678e-02 3.65387499e-01 -1.27881670e+00
8.81133616e-01 -3.73711258e-01 -2.85228640e-01 4.53693032e-01
-6.81405127e-01 5.94483852e-01 5.67484833e-02 -6.62384748e-01
-1.00886989e+00 -2.24913985e-01 -2.50123531e-01 -6.48770528e-03
-1.59877732e-01 9.39458549e-01 9.49271098e-02 2.25463375e-01
7.63045013e-01 -3.31683517e-01 -5.22362471e-01 -3.15403610e-01
7.95354247e-01 1.82805762e-01 7.11455405e-01 -1.01141036e+00
2.40260258e-01 7.90844560e-02 1.25014856e-01 -1.32109478e-01
-9.84713674e-01 2.54883081e-01 -7.77432024e-01 5.31028807e-01
9.98656154e-01 -4.44250733e-01 -9.88292754e-01 1.59352481e-01
-1.85045278e+00 -7.80591249e-01 -6.57604694e-01 2.11015224e-01
-4.81599689e-01 -1.51913494e-01 -1.57776928e+00 -6.82379007e-01
-5.30678153e-01 -7.12698281e-01 6.49234474e-01 -4.34351638e-02
-9.09656286e-01 -1.14576149e+00 1.77450404e-01 -4.38036025e-01
3.72053117e-01 4.59589660e-01 1.39203477e+00 -1.29971218e+00
-3.08047742e-01 -4.33490008e-01 -3.75143021e-01 -4.45073813e-01
6.45282045e-02 1.31347617e-02 -1.01505744e+00 2.78777987e-01
-6.89516842e-01 -4.48680818e-01 8.99008393e-01 4.33564633e-01
9.63389814e-01 -5.05090237e-01 -1.01919770e+00 5.59549332e-01
9.01311874e-01 6.75142169e-01 6.30287290e-01 8.81122530e-01
4.25579011e-01 5.04452229e-01 5.88313103e-01 6.47556484e-02
6.24521852e-01 6.13778532e-01 4.06840384e-01 6.77349940e-02
4.34419513e-02 -3.46647292e-01 -1.38945311e-01 2.31517047e-01
-2.37906426e-01 -4.11936551e-01 -9.48033094e-01 8.68507504e-01
-2.25562716e+00 -9.47763622e-01 -1.07845135e-01 1.30639148e+00
1.15658987e+00 8.83301198e-01 -3.50683704e-02 4.04792905e-01
2.76136756e-01 4.49694037e-01 -2.25311801e-01 -9.67124462e-01
-7.57603794e-02 8.46283436e-01 6.85586393e-01 3.33602399e-01
-1.61651468e+00 1.41454792e+00 8.20697212e+00 4.89523560e-01
-6.47329986e-01 1.06801167e-01 4.13639784e-01 2.63781279e-01
-3.29376787e-01 3.34790111e-01 -7.76822150e-01 -1.19752757e-01
1.73566806e+00 8.79837498e-02 1.28946587e-01 7.06610560e-01
-5.45630515e-01 4.24995035e-01 -1.60906124e+00 3.75247598e-01
-6.24265015e-01 -1.81606317e+00 -9.44006070e-02 2.91298896e-01
1.55408412e-01 1.43776417e-01 -2.36962423e-01 3.83917183e-01
1.03289950e+00 -1.16534913e+00 8.63396883e-01 4.84412044e-01
8.08544636e-01 -5.10770500e-01 9.56257284e-01 -3.56166601e-01
-1.62898135e+00 -2.41725687e-02 -1.12344198e-01 -2.42801592e-01
5.57669997e-01 4.70378518e-01 -7.58542418e-01 8.89490008e-01
7.18092084e-01 9.85080302e-01 -4.33411032e-01 4.15111691e-01
-4.84451056e-01 4.39074129e-01 -5.00838757e-01 2.85474230e-02
4.22187448e-01 7.46810913e-01 1.35633498e-01 1.55042958e+00
3.55147687e-03 1.46619320e-01 -1.66038916e-01 5.81538498e-01
-1.51207760e-01 -4.87528116e-01 -8.78191173e-01 -4.45290297e-01
6.79878175e-01 7.72042990e-01 -4.20003325e-01 -5.77732503e-01
-4.25202996e-01 4.24558997e-01 6.18989408e-01 5.08206844e-01
-6.94394171e-01 -6.85507298e-01 7.09042311e-01 -1.22317575e-01
6.38841331e-01 -1.96909279e-01 -1.31910622e-01 -7.02809393e-01
-1.48996100e-01 -6.30975127e-01 9.83441949e-01 -5.41890323e-01
-1.08332849e+00 1.31745708e+00 3.41895521e-01 -6.74488187e-01
-1.10436773e+00 -6.42982423e-01 -5.33488512e-01 6.75570905e-01
-1.29062605e+00 -1.42195845e+00 4.91286337e-01 2.14933127e-01
9.37145874e-02 -1.75103813e-01 1.42469800e+00 3.62124562e-01
-2.96291649e-01 7.81406522e-01 -6.92007124e-01 7.80450284e-01
2.21253753e-01 -1.14615571e+00 1.45296979e+00 5.55801392e-01
6.28821552e-01 9.49818194e-01 4.32399988e-01 -8.31218362e-01
-1.21446204e+00 -8.25128376e-01 1.59814775e+00 -3.50999475e-01
1.17340052e+00 -2.26996168e-02 -7.59216309e-01 1.55592787e+00
7.06113875e-01 6.44622624e-01 5.50335288e-01 8.36851656e-01
-7.31634676e-01 2.61508897e-02 -8.38608563e-01 2.35781014e-01
1.51190639e+00 -8.96548808e-01 -9.94612336e-01 1.36093214e-01
1.44642043e+00 -6.17630124e-01 -1.62437129e+00 7.58761942e-01
9.14912462e-01 -4.46033120e-01 1.17618823e+00 -1.37908638e+00
4.42908555e-01 1.71389267e-01 -1.97804987e-01 -1.01046157e+00
-3.93381685e-01 -9.37801003e-01 -1.17402709e+00 1.33077443e+00
1.19163585e+00 -9.34472144e-01 8.22711945e-01 5.80844879e-01
-7.97491148e-02 -1.10334384e+00 -7.49585271e-01 -7.84694910e-01
2.06212893e-01 -6.74275279e-01 1.01425147e+00 8.29372227e-01
6.02997720e-01 5.27055740e-01 2.17816979e-01 -1.93965048e-01
5.09776082e-03 5.73972352e-02 -2.92389262e-02 -1.32569528e+00
-4.20903802e-01 -5.89371204e-01 -4.40819293e-01 -7.75752902e-01
7.54635990e-01 -6.48679674e-01 -2.91560739e-01 -1.97542489e+00
-3.54072213e-01 -4.32333261e-01 -6.60411417e-01 1.16839755e+00
2.79895127e-01 -1.26854509e-01 -3.30367386e-01 -3.27722013e-01
-4.88991141e-01 1.18999943e-01 4.96401787e-01 -5.55547297e-01
-6.96266219e-02 -1.85285330e-01 -5.19046724e-01 4.12704617e-01
5.64298272e-01 -7.30412543e-01 -3.42932612e-01 -5.20547926e-01
8.77103329e-01 2.37721607e-01 -3.27843465e-02 -5.11255443e-01
2.04565108e-01 4.35073115e-02 4.00914788e-01 -7.45260715e-01
4.01076376e-01 -5.34927845e-01 1.05569243e-01 4.28733826e-01
-7.83123732e-01 2.59251505e-01 6.07293725e-01 3.20568025e-01
-3.47556472e-01 -3.21412012e-02 -1.83798060e-01 -2.59592563e-01
-5.87895274e-01 1.93935707e-01 -3.65016550e-01 1.51276454e-01
4.49692070e-01 1.12948284e-01 -8.70145440e-01 -2.93202639e-01
-7.67825425e-01 -3.86183755e-03 -1.90159470e-01 4.84904289e-01
4.66356069e-01 -1.47361362e+00 -4.05550122e-01 -4.03593220e-02
8.91930088e-02 -5.95418960e-02 -6.52566195e-01 4.15675581e-01
-3.08455139e-01 7.97745526e-01 1.89106748e-01 3.01203970e-02
-1.01550031e+00 7.68866003e-01 2.93571025e-01 -1.00264692e+00
-7.53436625e-01 9.94910002e-01 -5.41253865e-01 -3.25416833e-01
3.11889440e-01 -1.19793189e+00 -2.25795671e-01 6.66738674e-02
1.93041950e-01 -2.66356260e-01 3.78328055e-01 -1.20523281e-01
-7.41905332e-01 1.97186798e-01 -1.37031898e-01 -3.80943716e-01
1.55350959e+00 6.45201623e-01 -9.47357938e-02 5.11075377e-01
1.05847609e+00 -5.97329557e-01 -7.13737905e-01 -5.72217882e-01
8.50457609e-01 1.46469876e-01 -1.08518921e-01 -8.91482353e-01
-8.57016027e-01 5.41972399e-01 -1.39338290e-02 7.36773133e-01
6.26291037e-01 3.25293154e-01 1.21092963e+00 8.81722093e-01
3.50136042e-01 -7.43377686e-01 -3.25764775e-01 1.02281225e+00
4.13247615e-01 -6.87968194e-01 1.41320840e-01 -4.33275104e-01
-2.42219627e-01 1.25807166e+00 3.22427660e-01 -1.61249861e-01
1.01594937e+00 9.81827080e-01 -2.04929728e-02 -5.61325788e-01
-1.37663329e+00 -5.75583100e-01 3.67745042e-01 6.81562364e-01
1.07898176e+00 1.64809451e-02 -3.09383243e-01 9.40967500e-01
-4.29962844e-01 8.15515742e-02 -2.21920863e-01 1.47666430e+00
1.63718700e-01 -1.73045087e+00 3.78373474e-01 5.64122438e-01
-9.24414873e-01 -5.23487031e-01 -5.38585901e-01 1.09699833e+00
6.44557476e-02 9.39592242e-01 2.67770231e-01 -7.50524104e-01
5.62609017e-01 5.02515614e-01 6.69098914e-01 -6.07515037e-01
-1.21669674e+00 -2.26678178e-01 1.46108258e+00 -9.88605440e-01
-3.99893284e-01 -4.92214352e-01 -1.32123280e+00 -1.65441543e-01
-3.89757425e-01 3.16599220e-01 2.12671012e-01 1.04496300e+00
2.59034455e-01 1.06118333e+00 -1.79995000e-01 -6.11702740e-01
-1.07014939e-01 -1.29589725e+00 -1.52583523e-02 1.94107480e-02
4.00513947e-01 -5.97937346e-01 3.31876904e-01 2.45292988e-02] | [9.212636947631836, 8.963118553161621] |
85e2ca24-1828-446b-b27f-92cb5af02159 | a-triplet-loss-dilated-residual-network-for | 2303.08398 | null | https://arxiv.org/abs/2303.08398v1 | https://arxiv.org/pdf/2303.08398v1.pdf | A Triplet-loss Dilated Residual Network for High-Resolution Representation Learning in Image Retrieval | Content-based image retrieval is the process of retrieving a subset of images from an extensive image gallery based on visual contents, such as color, shape or spatial relations, and texture. In some applications, such as localization, image retrieval is employed as the initial step. In such cases, the accuracy of the top-retrieved images significantly affects the overall system accuracy. The current paper introduces a simple yet efficient image retrieval system with a fewer trainable parameters, which offers acceptable accuracy in top-retrieved images. The proposed method benefits from a dilated residual convolutional neural network with triplet loss. Experimental evaluations show that this model can extract richer information (i.e., high-resolution representations) by enlarging the receptive field, thus improving image retrieval accuracy without increasing the depth or complexity of the model. To enhance the extracted representations' robustness, the current research obtains candidate regions of interest from each feature map and applies Generalized-Mean pooling to the regions. As the choice of triplets in a triplet-based network affects the model training, we employ a triplet online mining method. We test the performance of the proposed method under various configurations on two of the challenging image-retrieval datasets, namely Revisited Paris6k (RPar) and UKBench. The experimental results show an accuracy of 94.54 and 80.23 (mean precision at rank 10) in the RPar medium and hard modes and 3.86 (recall at rank 4) in the UKBench dataset, respectively. | ['Hamidreza Mahyar', 'Hamidreza Pourreza', 'Saeideh Yousefzadeh'] | 2023-03-15 | null | null | null | null | ['content-based-image-retrieval'] | ['computer-vision'] | [ 1.11946337e-01 -6.23516858e-01 -2.36737549e-01 -1.48854569e-01
-1.25602579e+00 -2.77282566e-01 4.49954242e-01 6.10350408e-02
-5.46713412e-01 5.07902384e-01 -3.10710251e-01 6.41115084e-02
-6.44837379e-01 -7.15689242e-01 -6.63989484e-01 -1.05237067e+00
-1.31363928e-01 1.07290946e-01 1.95241690e-01 -2.62977511e-01
6.11869633e-01 7.79367566e-01 -2.07036400e+00 5.21761954e-01
8.61007512e-01 1.78389013e+00 3.94792855e-01 2.17612833e-01
3.84820178e-02 5.69701195e-01 -6.34573340e-01 -1.86898559e-01
5.29979706e-01 1.84630381e-03 -5.45850337e-01 1.06625564e-01
3.24754030e-01 -4.63459462e-01 -3.90461683e-01 8.86183798e-01
5.62930882e-01 1.57999501e-01 6.98437333e-01 -9.65171754e-01
-8.35431814e-01 -5.10949467e-04 -8.33153069e-01 1.31079197e-01
1.68085352e-01 -3.35654140e-01 7.99955249e-01 -1.29730570e+00
5.42837083e-01 1.02781332e+00 1.95650354e-01 2.96138488e-02
-7.60561347e-01 -9.07336116e-01 -1.92817837e-01 5.74530602e-01
-2.15342355e+00 -4.11855072e-01 8.33478332e-01 -1.70185510e-02
6.63953424e-01 4.97035623e-01 2.87619412e-01 4.83926475e-01
2.52880126e-01 8.65368366e-01 1.08527541e+00 -5.00345170e-01
-1.85600966e-01 3.41129124e-01 9.00865905e-03 7.22101510e-01
2.18124807e-01 -9.69969854e-02 -5.45593143e-01 -2.31970713e-01
6.45869374e-01 4.01911199e-01 -2.90924609e-01 -2.77598023e-01
-1.06208932e+00 7.04950273e-01 9.46233690e-01 3.53062630e-01
-3.76293302e-01 -1.39410645e-01 1.94581807e-01 4.64155406e-01
3.18210065e-01 2.69844532e-01 -2.42079794e-01 5.66081524e-01
-9.27546561e-01 9.97588411e-02 2.16928318e-01 9.85914230e-01
1.07497168e+00 -3.21852922e-01 -3.34350526e-01 1.21701992e+00
2.28490859e-01 7.08686709e-01 7.20783770e-01 -4.88037258e-01
4.68750298e-01 8.49948943e-01 2.49406651e-01 -1.52939284e+00
-8.20950568e-02 -4.85592157e-01 -9.49057221e-01 -2.15082124e-01
5.32328412e-02 4.82925773e-01 -1.24493480e+00 1.18429899e+00
1.27520323e-01 -3.21082890e-01 1.21765314e-02 1.25606179e+00
1.09644794e+00 7.42302120e-01 -2.58754790e-01 -1.46490648e-01
1.28082311e+00 -7.81266451e-01 -4.69466090e-01 6.57813847e-02
4.94872034e-01 -1.12101543e+00 7.52019584e-01 4.63073105e-01
-8.78016591e-01 -6.81213677e-01 -1.03730094e+00 2.15760320e-02
-8.10659289e-01 7.37560391e-01 4.44388449e-01 2.64179409e-01
-1.01738560e+00 4.54236537e-01 -1.80454493e-01 -2.52202183e-01
3.06118339e-01 4.76704925e-01 -6.00256503e-01 -5.09295642e-01
-1.24349785e+00 5.65731466e-01 5.87896466e-01 3.68788034e-01
-5.58843672e-01 -3.32267433e-01 -5.70379376e-01 6.04146458e-02
2.76740462e-01 -2.77017415e-01 4.93292302e-01 -9.49073255e-01
-1.01085865e+00 9.35247779e-01 1.02071829e-01 -2.80115902e-01
4.13661867e-01 -1.09922029e-01 -4.74723935e-01 6.93057954e-01
7.98950866e-02 9.04636562e-01 1.07194173e+00 -1.30883193e+00
-7.25509465e-01 -5.33634245e-01 -5.06562851e-02 3.79205853e-01
-5.79393268e-01 1.22716695e-01 -1.00667405e+00 -6.30192637e-01
4.46796864e-01 -1.08270609e+00 8.26344490e-02 5.30691706e-02
-2.34649152e-01 -3.15943748e-01 7.63720810e-01 -6.54216647e-01
1.08599830e+00 -2.24164057e+00 -2.04333261e-01 6.15282595e-01
-2.03859374e-01 2.56747514e-01 -2.56661206e-01 3.88505161e-01
1.15428478e-01 2.17236862e-01 1.79028101e-02 -4.68415059e-02
-3.42429638e-01 2.99982857e-02 -2.19846711e-01 5.48948348e-01
2.08990186e-01 8.88828874e-01 -3.18800390e-01 -8.98519933e-01
2.83968329e-01 6.26603842e-01 -1.24571010e-01 2.81314459e-02
2.92500883e-01 6.01832457e-02 -8.38314176e-01 1.06383348e+00
8.45226526e-01 -2.84634918e-01 -4.03881639e-01 -3.71085465e-01
6.61234632e-02 -3.90853673e-01 -1.15836978e+00 1.76531637e+00
-4.51414168e-01 5.55386245e-01 -2.03478143e-01 -8.68485451e-01
1.20777357e+00 -2.29270775e-02 5.29760778e-01 -1.41743755e+00
2.63292193e-02 3.41568381e-01 -1.87289611e-01 -5.11151850e-01
8.58049154e-01 5.71912229e-01 -2.93120947e-02 2.06333652e-01
-3.10583025e-01 3.99940550e-01 1.69037282e-01 2.37185493e-01
5.97261369e-01 -3.11448462e-02 -4.19575535e-02 -1.62567958e-01
6.04454458e-01 4.59509827e-02 2.28164777e-01 8.33647847e-01
7.37839490e-02 8.56414676e-01 -1.26756147e-01 -5.02581656e-01
-8.17587078e-01 -8.23115408e-01 -6.07684195e-01 8.99514735e-01
7.00067878e-01 -1.97278962e-01 -1.99355185e-01 -4.01536226e-01
7.58701041e-02 2.29290053e-02 -5.83818138e-01 -3.80982459e-01
-4.40161914e-01 -8.84346187e-01 4.30332184e-01 2.00131118e-01
9.72107768e-01 -1.24685037e+00 -5.26376665e-01 -1.76933527e-01
-4.22198594e-01 -9.55958605e-01 -3.02271754e-01 -1.02742009e-01
-7.03626931e-01 -1.05547154e+00 -1.01088262e+00 -1.04565990e+00
9.34459448e-01 8.74053419e-01 7.96635687e-01 6.03483677e-01
-5.79934835e-01 1.85001597e-01 -6.64037168e-01 -1.85613766e-01
2.51366287e-01 8.18110257e-02 -2.14231446e-01 3.88096005e-01
3.01326841e-01 -1.26611744e-03 -9.77136672e-01 6.12755120e-01
-1.13964844e+00 -2.93752491e-01 1.05158472e+00 1.04903328e+00
9.08719778e-01 3.55673522e-01 5.02525389e-01 -3.54471713e-01
6.33873403e-01 -3.21082681e-01 -5.35525680e-01 5.63183427e-01
-8.04533601e-01 -1.77040309e-01 4.90896463e-01 -4.29126024e-01
-9.31734979e-01 -9.10499692e-02 5.70703864e-01 -7.35049367e-01
1.96845755e-01 6.23628974e-01 4.77502681e-02 -4.90070909e-01
5.08875370e-01 6.90555453e-01 5.73963411e-02 -4.03739810e-01
-1.74615737e-02 9.06756997e-01 1.66288763e-01 -3.62882882e-01
7.56748796e-01 3.27252924e-01 -2.99741868e-02 -7.57707059e-01
-4.66658682e-01 -8.05749774e-01 -3.54294479e-01 -1.90705314e-01
2.77683407e-01 -1.13245475e+00 -5.37109554e-01 3.37505132e-01
-7.71437168e-01 4.36905265e-01 2.51092196e-01 4.31253254e-01
-1.01805225e-01 3.26756477e-01 -4.33334827e-01 -7.71837056e-01
-6.73220873e-01 -1.35350633e+00 1.40980840e+00 3.97627383e-01
4.34458762e-01 -2.15360552e-01 -6.38513744e-01 4.99436557e-01
4.33657855e-01 6.96478412e-02 8.75231683e-01 -5.87150276e-01
-9.43498313e-01 -6.03466988e-01 -7.51847208e-01 2.37706304e-01
-1.67431235e-02 -2.07755312e-01 -1.02902806e+00 -4.26445484e-01
-4.31281596e-01 -5.18338799e-01 1.16140366e+00 1.48167416e-01
1.54750586e+00 -2.26419926e-01 -3.53403777e-01 4.84791547e-01
1.65251994e+00 3.31579030e-01 9.92708266e-01 5.25196970e-01
4.61530894e-01 6.63043916e-01 1.13640893e+00 2.83733368e-01
-6.89634075e-03 7.08402812e-01 4.27725077e-01 -9.75227579e-02
2.53210333e-03 1.84207596e-02 7.24518076e-02 5.84441721e-01
-3.76268886e-02 -1.20797530e-01 -6.13013685e-01 6.96787953e-01
-1.84008908e+00 -7.42817700e-01 2.21341625e-01 2.32129025e+00
6.90311909e-01 -1.82266623e-01 -3.47772956e-01 1.40735492e-01
7.06369102e-01 -2.17068307e-02 -4.77969289e-01 -5.06494939e-02
-2.77028829e-01 1.34578377e-01 7.27597535e-01 -1.02411797e-02
-1.11928368e+00 7.78300464e-01 5.04503202e+00 1.46707106e+00
-1.14596200e+00 -3.18282366e-01 8.59145164e-01 -1.65864483e-01
1.51002795e-01 -2.47311264e-01 -6.89969599e-01 3.98144901e-01
4.16610777e-01 2.32778937e-01 4.99050468e-01 6.11412585e-01
-4.00198847e-02 -4.02450264e-01 -7.09418297e-01 1.35314870e+00
3.35258007e-01 -1.16933501e+00 5.52211046e-01 1.09256126e-01
6.39410794e-01 -1.02766432e-01 4.49320585e-01 2.61814296e-01
-4.76288855e-01 -1.16526723e+00 4.41533148e-01 7.84283459e-01
9.35048640e-01 -1.05868745e+00 9.40937281e-01 3.27618986e-01
-1.05347180e+00 -1.37553468e-01 -6.66732311e-01 5.62238812e-01
-5.50981045e-01 3.55356812e-01 -5.34418941e-01 8.51968527e-01
1.19975650e+00 5.38484633e-01 -1.02121532e+00 1.29045832e+00
2.84663141e-01 8.33278522e-02 -4.43523258e-01 -6.08284622e-02
2.03250796e-01 -1.95254907e-01 2.65728682e-01 9.67327297e-01
4.39569086e-01 3.35540697e-02 2.04472825e-01 5.29524088e-01
-2.23639235e-01 4.89697069e-01 -6.41285896e-01 2.74901778e-01
5.06998718e-01 1.50134838e+00 -7.69312263e-01 -1.54684946e-01
-7.44159743e-02 7.12186873e-01 4.89266217e-02 4.77838486e-01
-6.27252758e-01 -7.57551730e-01 1.05897740e-01 1.04775047e-02
4.61294264e-01 2.41417214e-01 7.68526420e-02 -8.38475823e-01
4.29082632e-01 -9.05895531e-01 4.41617399e-01 -1.02205217e+00
-9.61572111e-01 7.26815879e-01 7.01535270e-02 -1.62733352e+00
-1.61891326e-01 -5.12326896e-01 -4.79394048e-02 8.60587060e-01
-1.76238251e+00 -1.30772913e+00 -5.64198315e-01 7.65607715e-01
4.20210570e-01 -4.40678656e-01 7.13709116e-01 5.34202397e-01
-5.39035261e-01 8.84859741e-01 3.86905909e-01 1.45630673e-01
9.34883714e-01 -6.84810877e-01 -4.63321060e-01 4.95869070e-01
1.95627719e-01 7.47609973e-01 1.64957196e-01 -3.23436797e-01
-1.63649929e+00 -1.15497279e+00 5.03804445e-01 -2.34298296e-02
1.70057148e-01 -2.98066251e-02 -7.88746417e-01 -9.51904524e-03
-6.93838596e-02 3.19208324e-01 3.77043754e-01 -1.94741741e-01
-3.19917679e-01 -5.45278847e-01 -1.30524778e+00 5.21390915e-01
6.15804553e-01 -6.05150342e-01 -1.47402927e-01 3.16404074e-01
3.95152479e-01 -2.23096296e-01 -9.68136787e-01 6.28897071e-01
9.19618249e-01 -8.44709516e-01 1.22771215e+00 -2.74287164e-02
2.16609433e-01 -4.27870303e-01 -4.58617032e-01 -8.29075098e-01
-2.17829853e-01 -5.98232001e-02 1.73118219e-01 1.06406188e+00
2.77787477e-01 -5.34925580e-01 5.99063635e-01 3.04770827e-01
2.09054947e-01 -1.03021348e+00 -8.28156829e-01 -5.56286871e-01
-2.77569354e-01 1.52403787e-02 6.17099583e-01 6.09011710e-01
-6.45169318e-01 -4.79482710e-02 -2.34675482e-01 1.03310265e-01
5.85633278e-01 5.71981192e-01 6.28491759e-01 -1.29699373e+00
2.01802135e-01 -2.25442469e-01 -4.74671483e-01 -8.10613573e-01
-1.17068566e-01 -8.91464174e-01 -1.57641664e-01 -1.35172164e+00
4.33882594e-01 -7.10150123e-01 -8.96590412e-01 4.43618417e-01
7.90489018e-02 8.25377941e-01 1.60339862e-01 7.74748385e-01
-8.71130109e-01 5.69208562e-01 1.29982686e+00 -5.01078069e-01
-1.33212373e-01 -1.47918165e-01 -6.29337430e-01 3.98508877e-01
7.91174769e-01 -4.12504196e-01 -4.17235851e-01 -3.35657537e-01
1.18102245e-01 7.06332643e-03 5.47293484e-01 -7.97667503e-01
3.72919708e-01 1.40820161e-01 1.01193678e+00 -9.58370328e-01
5.35081625e-01 -9.75895107e-01 -5.44435484e-03 2.70797461e-01
-2.88002700e-01 8.37445781e-02 1.96609437e-01 6.18448615e-01
-6.09910727e-01 -2.27693006e-01 5.32890975e-01 -1.11259483e-01
-1.01641524e+00 3.24791551e-01 2.13571459e-01 -4.35745895e-01
9.61482108e-01 -5.52724540e-01 -3.74055356e-01 -2.96115607e-01
-2.52358735e-01 2.96087086e-01 3.25500578e-01 6.58851504e-01
1.08604705e+00 -1.53971040e+00 -5.92940807e-01 3.01660031e-01
7.26410687e-01 -3.07218023e-02 3.09778392e-01 7.13001668e-01
-4.63176161e-01 4.63802457e-01 -1.73238471e-01 -7.81848133e-01
-1.34790242e+00 4.66871530e-01 2.83028483e-01 -2.36902565e-01
-4.57646012e-01 4.44142461e-01 1.46903247e-01 -9.15458351e-02
2.90683180e-01 1.16281450e-01 -4.95425612e-01 2.22499728e-01
6.08393490e-01 2.09877253e-01 3.92912060e-01 -8.05659413e-01
-4.14412975e-01 9.19963062e-01 -5.88420749e-01 2.63081670e-01
1.10182524e+00 -1.16500132e-01 -2.47362152e-01 2.88391691e-02
1.51899338e+00 -2.57129163e-01 -6.15467012e-01 -3.54553938e-01
-1.46180794e-01 -6.68975115e-01 3.55436057e-01 -7.75837123e-01
-1.27646005e+00 4.85673279e-01 1.05574918e+00 3.77362780e-02
1.53040302e+00 -1.97136387e-01 7.40179539e-01 7.27723181e-01
6.73864484e-01 -1.22864115e+00 1.00731619e-01 7.19053075e-02
1.19549072e+00 -1.59266460e+00 2.75631517e-01 -1.46548674e-01
-3.26509386e-01 9.54772949e-01 5.60623467e-01 -2.73461640e-01
6.73898757e-01 -3.12223107e-01 -5.79675362e-02 -2.15975061e-01
-4.94525999e-01 -1.32330939e-01 7.10328400e-01 1.99646130e-01
3.68265621e-02 -9.67284963e-02 -3.82027864e-01 3.05751264e-01
7.09120780e-02 -2.53459245e-01 3.74876671e-02 8.92136574e-01
-4.99345064e-01 -7.13797927e-01 -6.59340203e-01 6.77921593e-01
-6.14472926e-01 -1.92377552e-01 -5.62803090e-01 9.41143930e-01
-5.94802760e-02 1.13777900e+00 3.57475169e-02 -4.62136775e-01
1.64115861e-01 -2.93781072e-01 4.30308998e-01 -6.69270903e-02
-4.81503367e-01 2.23266914e-01 -2.72475898e-01 -5.32717705e-01
-5.65460086e-01 -1.85547978e-01 -9.42396164e-01 -1.23023756e-01
-7.52192795e-01 2.53825903e-01 6.83651805e-01 7.63063788e-01
5.41345537e-01 2.35851169e-01 9.32598591e-01 -6.75282836e-01
-4.68105644e-01 -9.42331254e-01 -6.93625391e-01 4.62017059e-01
2.68173993e-01 -7.34236121e-01 -2.18134105e-01 -2.64370710e-01] | [10.723499298095703, 0.6844955682754517] |
b35b9cad-8211-47cf-a618-9d652e0efbeb | probabilistic-semantic-retrieval-for | 1712.06204 | null | http://arxiv.org/abs/1712.06204v2 | http://arxiv.org/pdf/1712.06204v2.pdf | Probabilistic Semantic Retrieval for Surveillance Videos with Activity Graphs | We present a novel framework for finding complex activities matching
user-described queries in cluttered surveillance videos. The wide diversity of
queries coupled with unavailability of annotated activity data limits our
ability to train activity models. To bridge the semantic gap we propose to let
users describe an activity as a semantic graph with object attributes and
inter-object relationships associated with nodes and edges, respectively. We
learn node/edge-level visual predictors during training and, at test-time,
propose to retrieve activity by identifying likely locations that match the
semantic graph. We formulate a novel CRF based probabilistic activity
localization objective that accounts for mis-detections, mis-classifications
and track-losses, and outputs a likelihood score for a candidate grounded
location of the query in the video. We seek groundings that maximize overall
precision and recall. To handle the combinatorial search over all
high-probability groundings, we propose a highest precision subgraph matching
algorithm. Our method outperforms existing retrieval methods on benchmarked
datasets. | ['Yu-Ting Chen', 'Gregory Castañón', 'Venkatesh Saligrama', 'Yannan Bai', 'Joseph Wang'] | 2017-12-17 | null | null | null | null | ['semantic-retrieval'] | ['natural-language-processing'] | [ 4.62449372e-01 2.54221782e-02 -5.94926775e-01 -3.24607074e-01
-9.86164927e-01 -8.82325768e-01 5.04362166e-01 4.48743612e-01
-3.48812640e-01 5.04392028e-01 4.72508758e-01 7.78138787e-02
-3.54638010e-01 -6.25780880e-01 -9.23405111e-01 -2.49076381e-01
-5.50362051e-01 4.80930299e-01 8.09698761e-01 5.92673242e-01
5.25343865e-02 4.89502817e-01 -1.47296429e+00 2.97879487e-01
5.80467284e-01 1.05251396e+00 3.19115907e-01 7.05772877e-01
2.50454605e-01 8.90422821e-01 -6.54577136e-01 -3.59809637e-01
9.80624110e-02 -3.55018705e-01 -7.27994204e-01 5.15177250e-01
7.94099391e-01 -2.63461709e-01 -6.62220478e-01 8.12604070e-01
1.28341541e-01 3.69626880e-01 6.65229857e-01 -1.74997067e+00
-1.38885677e-01 -1.50105774e-01 -4.71716404e-01 6.44764662e-01
1.07805014e+00 -1.21358797e-01 1.11345887e+00 -6.69884741e-01
9.64537799e-01 9.71547365e-01 6.07426703e-01 3.12517017e-01
-1.02969301e+00 -2.79480726e-01 5.59729040e-01 1.18017040e-01
-1.73762798e+00 -2.97351062e-01 4.54834521e-01 -5.74527800e-01
1.13149476e+00 1.19611688e-01 9.42560136e-01 1.31121051e+00
-4.87545654e-02 9.60874915e-01 3.25998157e-01 -1.21863291e-01
3.04024458e-01 -1.47270724e-01 -3.67463417e-02 1.35230601e+00
5.22179663e-01 -9.50297341e-02 -1.00759828e+00 -7.08718359e-01
6.60372615e-01 4.14966971e-01 -2.95923322e-01 -9.90814447e-01
-1.27743030e+00 5.77672362e-01 3.50755662e-01 1.25201032e-01
-4.25797701e-01 5.00934303e-01 -5.60758561e-02 -1.07253879e-01
2.85898656e-01 3.35285187e-01 -3.39643002e-01 -1.20963581e-01
-1.13460863e+00 2.32049361e-01 8.96123290e-01 1.41016626e+00
8.46309960e-01 -6.46636963e-01 -7.32854724e-01 4.20054257e-01
4.10340071e-01 5.90941966e-01 -2.01815099e-01 -9.03849781e-01
4.65260953e-01 8.34528089e-01 4.13834780e-01 -1.25583434e+00
-1.79879546e-01 -4.11034673e-01 -1.38661504e-01 -5.34366667e-01
3.36154044e-01 1.70560479e-01 -9.22284782e-01 1.83100379e+00
3.97959799e-01 6.80883765e-01 -1.49547681e-01 8.46054614e-01
4.11785424e-01 5.56831360e-01 6.47377849e-01 -2.36921459e-01
1.47144234e+00 -9.75627959e-01 -5.98325610e-01 -5.24389386e-01
7.26596415e-01 -3.94029737e-01 7.02031195e-01 -2.42286026e-02
-7.64661968e-01 -2.02347234e-01 -8.44569206e-01 3.39657098e-01
-3.85539472e-01 2.09883362e-01 5.33481836e-01 3.95110577e-01
-1.09984708e+00 2.69894630e-01 -1.05163729e+00 -8.04867983e-01
5.30809999e-01 2.35856578e-01 -4.74215299e-01 -2.56478935e-01
-8.25468242e-01 7.45892584e-01 4.02005494e-01 -3.04865479e-01
-1.37805092e+00 -3.23556781e-01 -1.02868223e+00 3.04852016e-02
1.01505888e+00 -8.63995373e-01 1.11798537e+00 -7.58482814e-01
-4.19684470e-01 9.19528186e-01 -5.64139426e-01 -3.99302572e-01
1.73741013e-01 -2.40394592e-01 -5.11331975e-01 5.23552775e-01
4.95484203e-01 7.66391456e-01 5.95719218e-01 -1.09448016e+00
-9.40210640e-01 -3.61649305e-01 9.86020118e-02 3.60978663e-01
-1.08809240e-01 8.08151381e-04 -1.10825765e+00 -6.99412286e-01
1.79097950e-01 -8.66467834e-01 -1.60747886e-01 3.57278615e-01
-2.40166679e-01 -3.73756528e-01 8.37299168e-01 -6.73898280e-01
1.22627044e+00 -1.96393144e+00 -7.03502893e-02 6.00896180e-01
2.59424955e-01 -2.54654855e-01 -2.26862848e-01 3.80079567e-01
6.58727825e-01 -6.23184219e-02 8.32272402e-04 -2.65505403e-01
-1.92660019e-01 3.60613227e-01 -5.38135953e-02 5.32743156e-01
3.12680036e-01 1.01316583e+00 -1.37773299e+00 -1.03322721e+00
-1.16868816e-01 3.19632977e-01 -5.84632754e-01 5.05402088e-01
-5.61171055e-01 1.64111584e-01 -8.81062746e-01 1.06411278e+00
1.51043966e-01 -7.12957978e-01 2.50368416e-01 -3.01276743e-01
2.24073261e-01 6.81075305e-02 -1.09850430e+00 2.15480733e+00
1.08940974e-01 5.71123004e-01 -1.63020581e-01 -7.91472316e-01
5.95896065e-01 2.30416283e-01 8.90971899e-01 -4.06645328e-01
-3.24270099e-01 -1.36779338e-01 -8.36528242e-01 -6.49154365e-01
4.65255857e-01 6.78808689e-01 -3.26227218e-01 2.94167697e-01
4.59854126e-01 5.37291527e-01 2.29320094e-01 5.99743605e-01
1.65660393e+00 3.43873382e-01 4.58613247e-01 -1.23734875e-02
3.29970866e-02 4.38475460e-01 5.56652725e-01 1.27812135e+00
-2.05116078e-01 3.09444070e-01 3.70086789e-01 -3.69514227e-01
-5.04240453e-01 -1.44945157e+00 3.77435267e-01 1.36573112e+00
5.96391737e-01 -6.39970958e-01 -5.94178498e-01 -1.05351901e+00
-2.53837675e-01 2.04190895e-01 -4.93475765e-01 -1.07014589e-01
-4.27100748e-01 -2.99950153e-01 5.00196576e-01 5.87154448e-01
4.28057641e-01 -9.33315694e-01 -9.70342040e-01 1.79372042e-01
-5.45568109e-01 -1.48023891e+00 -8.81711066e-01 -9.26880725e-03
-7.60557532e-01 -1.63082016e+00 -5.78480959e-01 -9.06042695e-01
9.63325441e-01 4.71195459e-01 1.46479237e+00 2.44593978e-01
-5.78669250e-01 1.23256588e+00 -4.15360302e-01 -2.85043940e-02
1.06987879e-01 -4.96620871e-03 -1.10078424e-01 1.68947913e-02
7.29228556e-01 -4.14001122e-02 -8.14194679e-01 5.11906266e-01
-6.98162317e-01 -1.79390147e-01 5.27401268e-01 3.50800097e-01
8.20591807e-01 -3.07095438e-01 5.95417805e-02 -3.85908097e-01
3.25700492e-01 -5.87411165e-01 -7.27407336e-01 9.84095514e-01
-3.45939398e-01 2.38190353e-01 -3.08498770e-01 -5.75135529e-01
-7.68072248e-01 5.25052190e-01 5.95722258e-01 -5.33206999e-01
-2.32857198e-01 3.51400137e-01 -2.13289380e-01 -2.94295996e-02
6.28876150e-01 1.87120870e-01 -6.48995697e-01 -2.09436208e-01
1.98463753e-01 9.50092375e-02 6.04490459e-01 -5.39531291e-01
5.18984199e-01 4.16630596e-01 -1.76819339e-02 -7.34850228e-01
-1.03153968e+00 -9.93339360e-01 -4.92463410e-01 -5.29712975e-01
1.03393209e+00 -1.16729033e+00 -5.38781762e-01 -8.72918516e-02
-1.26194572e+00 -1.80446044e-01 -6.64519742e-02 6.05553091e-01
-6.29129469e-01 3.16653550e-01 -5.97929433e-02 -9.66255605e-01
6.99700117e-02 -8.88751686e-01 1.55310524e+00 1.50477707e-01
-4.21339065e-01 -7.93405473e-01 2.62028724e-01 2.19853014e-01
3.71589027e-02 4.15118456e-01 3.00161123e-01 -8.51740777e-01
-1.24037421e+00 -3.51364136e-01 -2.34817967e-01 -4.38729733e-01
-4.66366298e-02 -3.26327920e-01 -7.32626140e-01 -3.52831095e-01
-7.40210474e-01 -1.62788779e-01 7.32800961e-01 4.23382223e-01
1.12040567e+00 -6.84449971e-01 -1.19810283e+00 3.98096830e-01
1.19556963e+00 2.15500176e-01 3.05399030e-01 -5.69092706e-02
4.41484869e-01 5.14500856e-01 8.95846903e-01 6.04151070e-01
2.44960263e-01 9.08608794e-01 3.71923983e-01 6.10243119e-02
1.57123223e-01 -7.96626210e-01 2.49354690e-01 -2.78975308e-01
8.15000385e-02 -8.45419884e-01 -8.63967419e-01 8.18089008e-01
-2.10325313e+00 -1.16642261e+00 3.70975345e-01 2.36704516e+00
2.69419670e-01 -2.30914112e-02 3.02011073e-01 -4.96664703e-01
7.10104167e-01 2.94802070e-01 -2.89456934e-01 6.28904045e-01
7.68800750e-02 -2.43464977e-01 7.02675939e-01 4.64674592e-01
-1.44085538e+00 9.16345775e-01 6.45872021e+00 5.00626206e-01
-2.24791780e-01 2.00247645e-01 1.85828611e-01 -2.20297575e-01
-8.08500648e-02 7.53493235e-03 -8.07951331e-01 2.30634540e-01
6.39728367e-01 2.20432863e-01 4.03230697e-01 1.00686359e+00
1.34968579e-01 -4.37218100e-01 -1.27280712e+00 1.05432785e+00
4.26041156e-01 -1.29297388e+00 1.06335990e-01 -1.17726274e-01
6.71751082e-01 7.48591125e-02 -4.13836539e-01 1.39662493e-02
4.24610347e-01 -6.44826949e-01 6.24973178e-01 7.91439772e-01
6.15073681e-01 -3.02328140e-01 2.54164398e-01 2.77469277e-01
-1.63720441e+00 -2.83824652e-02 9.19279829e-02 5.34463108e-01
2.56655842e-01 -1.70857720e-02 -1.04267073e+00 2.12071568e-01
8.84829462e-01 8.57941687e-01 -6.35676086e-01 1.51390290e+00
-1.88803434e-01 3.92542094e-01 -5.07437170e-01 -2.65798658e-01
2.22706050e-01 8.76030028e-02 7.51240015e-01 1.49330354e+00
4.41855103e-01 1.32453814e-01 8.52135301e-01 6.49350345e-01
-1.06131025e-01 -1.53426185e-01 -9.91261780e-01 -1.01797052e-01
8.06510389e-01 8.95295918e-01 -1.04097593e+00 -5.62826097e-01
-4.74978864e-01 1.29902005e+00 3.13102752e-01 6.24830663e-01
-9.25270557e-01 -3.44694033e-02 5.52982926e-01 4.52160656e-01
1.42079905e-01 -2.74400353e-01 4.79559600e-01 -1.22306991e+00
3.66483897e-01 -4.67732579e-01 1.01530063e+00 -8.58261704e-01
-9.41635072e-01 3.03369164e-01 6.02577806e-01 -1.18738317e+00
-2.90175557e-01 -3.10207576e-01 -3.45959187e-01 1.91002801e-01
-9.82148051e-01 -1.19705093e+00 -7.75083423e-01 8.17242682e-01
5.03658652e-01 9.44219902e-02 7.33674586e-01 3.12302828e-01
-1.65632993e-01 2.80392766e-01 -5.86130619e-01 3.56663942e-01
3.85108858e-01 -9.13576543e-01 1.85056806e-01 9.80285645e-01
6.91223562e-01 4.58161801e-01 5.78216612e-01 -1.00049222e+00
-1.15123129e+00 -1.27365613e+00 1.17453015e+00 -8.37104380e-01
4.57688093e-01 -5.35302043e-01 -6.55061841e-01 9.38602149e-01
-2.72292435e-01 2.13520750e-01 6.61442399e-01 -7.76996911e-02
-4.90134031e-01 2.26815745e-01 -1.10804307e+00 4.37556356e-01
1.82841372e+00 -6.68837309e-01 -5.85886955e-01 5.97739756e-01
5.59412658e-01 -1.29497528e-01 -5.71531653e-01 3.83984864e-01
5.45851707e-01 -5.25733471e-01 1.07808518e+00 -8.48912120e-01
-4.20951843e-01 -4.98382866e-01 -2.15152979e-01 -7.83186138e-01
-1.97400734e-01 -6.03847086e-01 -5.26397645e-01 6.36388004e-01
4.69345570e-01 -1.13669261e-02 1.10225296e+00 6.62808537e-01
7.14971125e-02 -3.81116390e-01 -1.01629210e+00 -8.40950191e-01
-1.27091324e+00 -2.95658976e-01 1.96351930e-01 6.51934743e-01
7.82116950e-02 -3.89010436e-03 -5.23465395e-01 6.94255054e-01
7.94195354e-01 2.91101206e-02 5.71322799e-01 -1.28924561e+00
-3.80629122e-01 -3.85018103e-02 -6.99888408e-01 -1.41767728e+00
4.04170118e-02 -5.87570846e-01 1.28323346e-01 -1.80456424e+00
5.60812950e-01 -2.14345232e-01 -1.38857886e-01 6.46219194e-01
5.62606230e-02 1.76952243e-01 -1.67395696e-01 2.57205993e-01
-1.74053478e+00 1.80041298e-01 5.92469096e-01 -2.54727155e-01
-3.36036175e-01 4.23076749e-02 -1.40741751e-01 4.46134478e-01
4.87988293e-01 -8.53444934e-01 -6.02341413e-01 -3.60601693e-01
1.24089360e-01 4.76461321e-01 8.72630239e-01 -1.21049428e+00
5.66603243e-01 -3.29926550e-01 5.31123996e-01 -7.10609198e-01
5.40679634e-01 -1.07211292e+00 3.49144518e-01 3.11244547e-01
-6.10430181e-01 8.58379304e-02 -2.25639090e-01 1.31048930e+00
-1.31425530e-01 -8.46001357e-02 2.17450589e-01 -1.20994546e-01
-1.00393116e+00 7.09322810e-01 -5.80737054e-01 2.39164904e-01
1.27573001e+00 -4.03625190e-01 -1.66751266e-01 -4.38153446e-01
-9.68784869e-01 4.35647845e-01 4.98883545e-01 6.87546372e-01
6.89884365e-01 -1.18535030e+00 -1.98511556e-01 1.67188253e-02
8.02800536e-01 -2.05162764e-01 -2.48863064e-02 5.50585091e-01
-2.53684878e-01 3.88357788e-01 1.71258539e-01 -6.48716033e-01
-1.44603181e+00 5.22471368e-01 4.01257306e-01 -2.88068414e-01
-2.89168328e-01 7.28224754e-01 4.44017462e-02 2.11152151e-01
7.99683392e-01 -2.08353534e-01 -6.85626641e-02 -1.97359905e-01
4.17031199e-01 3.81533593e-01 -2.08436444e-01 -6.21605039e-01
-7.27459431e-01 5.14381826e-01 3.38662744e-01 -1.01260148e-01
9.00580764e-01 7.74982795e-02 4.17654544e-01 -6.83080703e-02
1.10074508e+00 -3.02172691e-01 -1.59875286e+00 -4.29681242e-01
5.13917983e-01 -6.62317872e-01 -1.72216713e-01 -6.77670836e-01
-7.88151503e-01 3.12279910e-01 7.08008051e-01 -7.22339749e-02
9.36123431e-01 7.85749495e-01 3.04166079e-01 7.91442275e-01
6.26717806e-01 -9.43297088e-01 5.92475355e-01 1.97659787e-02
5.63607097e-01 -1.30219567e+00 6.71179919e-03 -5.20553291e-01
-5.40058672e-01 6.44172013e-01 7.25134075e-01 2.46344339e-02
5.18227935e-01 1.22041047e-01 -4.36300069e-01 -7.37896979e-01
-7.32439637e-01 -4.61267501e-01 6.59880579e-01 8.80196571e-01
1.11124933e-01 -1.35299668e-01 1.01801671e-01 1.16926372e-01
6.26178622e-01 5.86744733e-02 -1.69570029e-01 1.29110622e+00
-7.71647692e-01 -6.01282835e-01 -2.87369877e-01 5.92340648e-01
-3.76530170e-01 1.96537137e-01 -5.50485313e-01 6.15156054e-01
-8.19576606e-02 1.08968592e+00 3.26295048e-01 -1.88912645e-01
4.44878608e-01 1.10170513e-01 4.63228643e-01 -5.28421462e-01
-1.53256208e-01 8.68111849e-02 2.87879795e-01 -1.01459503e+00
-7.92054296e-01 -6.67081714e-01 -9.76538777e-01 4.20203894e-01
-3.32229376e-01 3.08200061e-01 2.64800429e-01 8.54110181e-01
5.88398993e-01 7.11635873e-02 9.00933594e-02 -4.96658087e-01
-1.09063424e-01 -6.42000616e-01 -2.81301528e-01 6.07275307e-01
2.51124084e-01 -9.56616640e-01 -8.65865946e-02 4.23291415e-01] | [9.687668800354004, 0.7021060585975647] |
c17872d0-e1e0-48a8-9f32-9cb847bce11c | aanet-adaptive-aggregation-network-for | 2004.09548 | null | https://arxiv.org/abs/2004.09548v1 | https://arxiv.org/pdf/2004.09548v1.pdf | AANet: Adaptive Aggregation Network for Efficient Stereo Matching | Despite the remarkable progress made by learning based stereo matching algorithms, one key challenge remains unsolved. Current state-of-the-art stereo models are mostly based on costly 3D convolutions, the cubic computational complexity and high memory consumption make it quite expensive to deploy in real-world applications. In this paper, we aim at completely replacing the commonly used 3D convolutions to achieve fast inference speed while maintaining comparable accuracy. To this end, we first propose a sparse points based intra-scale cost aggregation method to alleviate the well-known edge-fattening issue at disparity discontinuities. Further, we approximate traditional cross-scale cost aggregation algorithm with neural network layers to handle large textureless regions. Both modules are simple, lightweight, and complementary, leading to an effective and efficient architecture for cost aggregation. With these two modules, we can not only significantly speed up existing top-performing models (e.g., $41\times$ than GC-Net, $4\times$ than PSMNet and $38\times$ than GA-Net), but also improve the performance of fast stereo models (e.g., StereoNet). We also achieve competitive results on Scene Flow and KITTI datasets while running at 62ms, demonstrating the versatility and high efficiency of the proposed method. Our full framework is available at https://github.com/haofeixu/aanet . | ['Juyong Zhang', 'Haofei Xu'] | 2020-04-20 | aanet-adaptive-aggregation-network-for-1 | http://openaccess.thecvf.com/content_CVPR_2020/html/Xu_AANet_Adaptive_Aggregation_Network_for_Efficient_Stereo_Matching_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Xu_AANet_Adaptive_Aggregation_Network_for_Efficient_Stereo_Matching_CVPR_2020_paper.pdf | cvpr-2020-6 | ['scene-flow-estimation'] | ['computer-vision'] | [-1.47149965e-01 -5.21046281e-01 -5.28980307e-02 -3.71280253e-01
-4.25904423e-01 -1.30349293e-01 1.78684667e-01 -1.75108284e-01
-6.02796495e-01 5.48354626e-01 -6.35925904e-02 -4.59864199e-01
1.42907172e-01 -1.15294695e+00 -7.91916251e-01 -4.16783810e-01
2.79369392e-02 2.03560635e-01 4.66244608e-01 -1.11276805e-01
2.41387755e-01 4.44682986e-01 -1.79259455e+00 1.75119475e-01
1.25966763e+00 1.38136292e+00 2.03795701e-01 4.45464253e-01
-1.60019472e-01 7.50014722e-01 2.08383109e-02 -4.64851558e-01
5.43307185e-01 2.44121570e-02 -6.48398936e-01 -1.27028212e-01
1.02800500e+00 -8.29444349e-01 -7.96333075e-01 1.08593392e+00
6.09610379e-01 1.26763314e-01 1.27850384e-01 -1.03535140e+00
-1.42832503e-01 2.58243173e-01 -7.61268318e-01 1.92614302e-01
2.18366161e-02 3.28465432e-01 8.07539403e-01 -9.24531400e-01
3.61106843e-01 1.30645144e+00 8.28738391e-01 2.34063491e-01
-1.10482609e+00 -1.07013178e+00 1.90096200e-01 4.84981120e-01
-1.45012736e+00 -5.23936391e-01 6.87993109e-01 -1.82539552e-01
1.08737636e+00 1.24433972e-01 8.46106350e-01 6.53744698e-01
3.48292105e-02 8.16271007e-01 1.08057570e+00 -1.62279904e-02
1.38473511e-02 -4.62196857e-01 -5.37235811e-02 9.69704807e-01
1.93741024e-01 3.36105138e-01 -6.32886350e-01 6.39797375e-02
1.07409668e+00 1.58783406e-01 -3.20481718e-01 -2.47294128e-01
-1.05212903e+00 6.89747930e-01 8.25949669e-01 -8.53351802e-02
-2.07672954e-01 6.06769919e-01 4.89335001e-01 1.15404271e-01
5.95671535e-01 -1.79998636e-01 -4.06622797e-01 -2.52951562e-01
-1.05032158e+00 4.05769348e-01 4.35028076e-01 1.04357529e+00
1.20789683e+00 1.09645635e-01 1.01702906e-01 9.12018478e-01
2.73948945e-02 6.20289505e-01 9.39834639e-02 -1.28733087e+00
6.26393378e-01 6.21518850e-01 -1.56060755e-01 -1.02618551e+00
-3.80580306e-01 -4.48788553e-01 -1.14694333e+00 4.00082976e-01
3.55649590e-01 1.36885177e-02 -8.53360832e-01 1.48840463e+00
3.43701839e-01 6.59056544e-01 -3.98598492e-01 9.41849113e-01
7.91179061e-01 6.04647219e-01 4.78157289e-02 2.79540926e-01
1.19761002e+00 -1.16883481e+00 -1.73917621e-01 -2.84997314e-01
5.76207697e-01 -9.39599395e-01 9.58136857e-01 1.52773812e-01
-1.34166348e+00 -7.01021969e-01 -9.64376807e-01 -4.47503537e-01
-7.49087110e-02 -1.35450423e-01 1.23581600e+00 4.13021564e-01
-1.12434804e+00 8.18572283e-01 -8.65851462e-01 -4.40220125e-02
6.87914908e-01 5.67399979e-01 -9.00747404e-02 -4.37501639e-01
-9.29107547e-01 4.80375499e-01 2.34764233e-01 8.48040208e-02
-5.00090420e-01 -1.14000499e+00 -1.01550519e+00 1.66495770e-01
2.53830105e-01 -9.88648295e-01 1.23911726e+00 -7.03966379e-01
-1.39123726e+00 8.93562317e-01 -4.34884578e-01 -6.25641584e-01
6.95504785e-01 -3.67126197e-01 -7.36236274e-02 6.15060255e-02
7.52624273e-02 1.07150722e+00 5.40341496e-01 -7.81841218e-01
-9.91855741e-01 -2.90134609e-01 2.65226275e-01 2.81125486e-01
-3.39170873e-01 -1.26468033e-01 -8.24181139e-01 -7.34280884e-01
2.69047618e-01 -7.69678295e-01 -3.83465946e-01 5.88760316e-01
-4.79699820e-02 -1.82530303e-02 5.32516479e-01 -5.30362010e-01
1.06908572e+00 -2.11946344e+00 -3.04174125e-01 2.99653411e-02
3.88813019e-01 4.34596986e-01 4.98260697e-03 -5.44603094e-02
1.73472553e-01 -1.82207182e-01 -2.84531713e-01 -4.95201141e-01
-1.07538074e-01 3.19851972e-02 -1.00947469e-01 4.87516582e-01
2.39990503e-02 9.33454812e-01 -7.62916327e-01 -5.71211994e-01
6.97550535e-01 6.13017380e-01 -9.44930792e-01 1.67308797e-04
4.42742258e-02 2.67117411e-01 -2.49362350e-01 7.97847986e-01
1.18578553e+00 -3.38524908e-01 -1.26195297e-01 -4.71713096e-01
-4.01316434e-01 4.70025301e-01 -1.38049400e+00 2.01636314e+00
-6.97334588e-01 6.59721851e-01 1.40579984e-01 -9.77674007e-01
7.61257231e-01 -9.68402624e-02 4.84198242e-01 -1.13221478e+00
1.71599150e-01 5.77853322e-01 -1.12387873e-01 -6.71850890e-02
5.39493740e-01 6.78292215e-02 2.08891645e-01 9.99058783e-02
-1.82252288e-01 -3.90171707e-02 2.08064392e-01 -6.58261627e-02
9.69929516e-01 2.39578456e-01 -5.41023612e-02 -5.50411880e-01
5.97656131e-01 3.12785711e-03 8.44650626e-01 5.17097354e-01
-2.40422651e-01 6.93735778e-01 5.30141033e-02 -8.34711671e-01
-9.79995549e-01 -1.02655363e+00 -2.35573813e-01 8.11510980e-01
5.96232593e-01 -2.67865598e-01 -6.31404161e-01 -2.03067839e-01
2.25159720e-01 5.11677302e-02 -2.09413394e-01 1.35443524e-01
-1.03008211e+00 -5.91147006e-01 4.96765971e-01 8.08261335e-01
1.28095126e+00 -6.49147034e-01 -6.85238421e-01 4.32214111e-01
-2.93418765e-01 -1.36402249e+00 -6.17002487e-01 -3.04914296e-01
-1.29893267e+00 -9.41566229e-01 -7.45487154e-01 -8.50565016e-01
5.46797156e-01 6.35075986e-01 1.31638944e+00 3.57430607e-01
-3.41853619e-01 -2.45331392e-01 4.81165275e-02 -2.15472028e-01
3.03513050e-01 7.54937381e-02 -1.48535892e-01 -2.66256690e-01
3.89320940e-01 -8.58915508e-01 -1.20390880e+00 3.72759908e-01
-6.83438599e-01 5.18033862e-01 4.32787031e-01 7.46039510e-01
7.21724272e-01 -1.89403780e-02 -1.95421260e-02 -5.99923015e-01
-5.96614815e-02 1.18008284e-02 -9.88497078e-01 -1.69141322e-01
-5.07785320e-01 -9.64836851e-02 6.67062759e-01 -1.78707212e-01
-1.09613991e+00 4.19271365e-03 -4.43977177e-01 -5.61970770e-01
-5.47725856e-02 1.96733803e-01 1.48545638e-01 -5.26220083e-01
3.08488578e-01 2.01239392e-01 -6.81430474e-02 -5.12853503e-01
3.20139378e-02 3.41436595e-01 5.33089280e-01 -6.48997545e-01
7.71650255e-01 1.01393020e+00 1.20579809e-01 -6.49599493e-01
-7.28560805e-01 -3.75831723e-01 -3.35493714e-01 -1.89698428e-01
5.67464411e-01 -1.39953780e+00 -9.78016257e-01 9.41575706e-01
-1.11788118e+00 -5.22366643e-01 8.03134125e-03 5.54984510e-01
-4.98101741e-01 5.34820974e-01 -8.82996380e-01 -3.76313508e-01
-6.17400169e-01 -1.25897539e+00 1.02457416e+00 3.84794176e-01
1.41484171e-01 -8.69800508e-01 -2.96605140e-01 5.17030180e-01
7.07312286e-01 5.17163202e-02 6.33927405e-01 3.52904350e-01
-1.11694694e+00 1.67395696e-01 -9.51490343e-01 2.02371806e-01
-8.04069936e-02 -1.46155447e-01 -1.04012930e+00 -3.07690203e-01
-2.09144250e-01 -1.99925482e-01 1.08743823e+00 6.82371497e-01
1.54594982e+00 1.17756061e-01 -2.16180727e-01 1.40076840e+00
1.69126368e+00 1.29752100e-01 7.72806227e-01 4.41187024e-01
7.53852248e-01 3.92638832e-01 5.68333328e-01 4.67499405e-01
6.19214594e-01 7.32707441e-01 4.88775879e-01 -4.62737679e-01
-4.50866282e-01 -2.40244091e-01 -9.59986262e-03 9.18725669e-01
-3.63924772e-01 6.35656267e-02 -9.46655035e-01 4.40255940e-01
-1.92622566e+00 -7.60533333e-01 -2.50602186e-01 2.17315435e+00
6.36841893e-01 3.37812424e-01 -3.27758640e-02 1.06948232e-02
6.94992840e-01 2.92092025e-01 -6.31247222e-01 -2.66593933e-01
-2.23817587e-01 6.45532429e-01 8.39310110e-01 5.67655146e-01
-1.23363578e+00 1.02556491e+00 5.38141632e+00 1.10586810e+00
-1.14094663e+00 -6.90359920e-02 8.79651248e-01 -3.33018690e-01
-1.84437051e-01 1.05895437e-02 -8.02277327e-01 5.42617857e-01
5.17269313e-01 -1.65452957e-01 4.80998158e-01 8.72838080e-01
3.12100679e-01 -2.13027939e-01 -8.88848007e-01 1.43552423e+00
-2.12381855e-01 -1.79210663e+00 -1.35050910e-02 1.91477776e-01
7.66472161e-01 3.86054546e-01 -1.25018090e-01 9.16621760e-02
1.34051740e-01 -8.97465765e-01 6.55422091e-01 2.93027032e-02
1.03043246e+00 -8.57068002e-01 7.72788525e-01 1.09858133e-01
-1.70565164e+00 8.50169584e-02 -6.50433421e-01 -4.89519596e-01
2.81174183e-01 8.80558431e-01 8.16833451e-02 5.46335220e-01
1.03606570e+00 9.37440038e-01 -2.09610462e-01 1.41052938e+00
-2.18358524e-02 2.38560289e-01 -5.82826555e-01 2.43778512e-01
4.22442406e-01 -3.04119647e-01 2.09318712e-01 1.17362499e+00
4.55430925e-01 1.56046748e-01 8.98567662e-02 7.45308936e-01
-3.19793403e-01 -1.66258048e-02 -5.21228433e-01 6.34394646e-01
4.42758352e-01 1.02037346e+00 -6.23837888e-01 -5.45740545e-01
-6.97862148e-01 9.73465323e-01 4.38791573e-01 1.88453510e-01
-1.08436465e+00 -4.20461029e-01 1.19287789e+00 1.37548134e-01
3.49575728e-01 -3.31419259e-01 -6.31508291e-01 -1.37333643e+00
3.12720865e-01 -6.74267709e-01 2.55237818e-01 -4.30617601e-01
-1.06624234e+00 4.90202904e-01 -3.33001137e-01 -1.43411040e+00
3.41498137e-01 -6.33305609e-01 -6.06766224e-01 8.57046723e-01
-2.06932235e+00 -8.65471840e-01 -7.87707329e-01 7.72754312e-01
6.74054027e-01 2.90097475e-01 4.14987564e-01 9.50347781e-01
-5.24429321e-01 6.74559593e-01 -5.36013627e-03 1.46980941e-01
6.26143694e-01 -8.83903205e-01 7.38982439e-01 8.90239775e-01
-1.75719634e-01 4.32533294e-01 1.87537163e-01 -3.23862404e-01
-1.42781770e+00 -1.09884214e+00 7.63675928e-01 1.35881886e-01
5.47813773e-01 -2.61570454e-01 -8.00566673e-01 3.19186568e-01
-7.55764917e-02 2.15242714e-01 2.38665149e-01 -9.07060411e-03
-4.63767558e-01 -5.20720065e-01 -1.00528979e+00 7.14145243e-01
1.66648030e+00 -3.58648032e-01 -1.72051907e-01 1.19299889e-01
6.15904629e-01 -8.50749373e-01 -6.95007324e-01 6.76295519e-01
6.03899479e-01 -1.55363619e+00 1.29264569e+00 -1.02920815e-01
7.30505764e-01 -2.70981759e-01 -1.67189315e-01 -7.61325717e-01
-3.23349357e-01 -4.85625267e-01 -1.94699869e-01 8.55996132e-01
3.18534411e-02 -8.60616565e-01 1.12822688e+00 6.42766416e-01
-3.67044002e-01 -8.98420513e-01 -1.19587469e+00 -7.63170063e-01
-4.43572849e-02 -6.28484905e-01 6.52454674e-01 7.23006427e-01
-4.48519498e-01 -8.24531317e-02 -3.79750490e-01 9.74647999e-02
1.03810906e+00 4.72335458e-01 8.35347176e-01 -1.01930308e+00
-2.21953824e-01 -6.89967334e-01 -5.98167181e-01 -1.72952032e+00
-8.40210319e-02 -6.10587120e-01 -2.28496194e-01 -1.40258956e+00
1.72426477e-01 -8.54122996e-01 -2.26984233e-01 2.99808234e-01
-2.65540093e-01 6.51980877e-01 2.37537578e-01 1.85220063e-01
-3.73532355e-01 5.57435334e-01 1.33543408e+00 -5.85000180e-02
-1.11743454e-02 -2.05290020e-01 -4.18915838e-01 9.84015226e-01
8.43778729e-01 -1.51824892e-01 -2.26714566e-01 -1.04908454e+00
-4.30878513e-02 -5.31857833e-02 5.45834541e-01 -1.35519171e+00
2.55278498e-01 -7.57120401e-02 2.16821000e-01 -6.86060905e-01
4.96291995e-01 -7.26292968e-01 1.28831238e-01 7.23909795e-01
2.57701188e-01 6.84786364e-02 4.22152996e-01 1.74680904e-01
-4.95109171e-01 2.38918185e-01 9.99964654e-01 -8.75679702e-02
-1.08379138e+00 8.52962434e-01 1.36084557e-01 -4.12037149e-02
8.37792099e-01 -4.10055459e-01 -4.27633017e-01 -1.61864251e-01
-1.34993568e-01 3.73385280e-01 6.40061438e-01 2.07670361e-01
5.63389778e-01 -1.32716346e+00 -5.90180099e-01 2.34185934e-01
-1.00414418e-01 4.06118572e-01 7.54156828e-01 8.39290857e-01
-1.13458014e+00 5.17046571e-01 -3.34242404e-01 -7.35453248e-01
-1.13251615e+00 2.21829474e-01 3.36951494e-01 -1.12456195e-01
-8.82747650e-01 9.98004854e-01 4.35769260e-01 -2.29279488e-01
3.21175098e-01 -3.52409720e-01 3.58062565e-01 -2.50269562e-01
5.17494261e-01 6.97341859e-01 1.27499476e-01 -2.94517398e-01
-3.67038846e-01 9.64075089e-01 9.55439508e-02 3.91098917e-01
1.29188168e+00 -3.90136614e-02 -7.04029948e-02 -1.72498435e-01
1.29188943e+00 -2.58350372e-01 -1.55321956e+00 -3.52320582e-01
-4.95242208e-01 -9.33356464e-01 2.97302037e-01 -2.46073946e-01
-1.53289735e+00 1.08798969e+00 7.39849389e-01 -3.00266892e-01
1.31884992e+00 -2.99529284e-01 1.44224143e+00 1.51740581e-01
6.51587188e-01 -1.04532778e+00 -3.62554640e-01 4.69109446e-01
3.73914778e-01 -1.38474178e+00 7.87871778e-02 -8.57438564e-01
3.93941347e-03 1.05856431e+00 9.25772369e-01 -3.30436468e-01
6.25183105e-01 3.57295632e-01 3.18516493e-02 -6.87288418e-02
-4.66686428e-01 -3.35306287e-01 1.29434988e-01 3.52068067e-01
4.26887125e-01 -4.40503173e-02 -3.71602148e-01 -1.52338892e-01
-8.37387443e-02 2.78082162e-01 1.26510441e-01 8.90166581e-01
-2.70045727e-01 -1.04817450e+00 -9.86612290e-02 6.11958086e-01
-3.35694909e-01 -4.97984946e-01 3.94007266e-01 6.42890990e-01
1.62717134e-01 6.76690817e-01 3.97513002e-01 -4.01963830e-01
4.46042329e-01 -4.42319781e-01 3.80782187e-01 -1.63155288e-01
-4.40275937e-01 2.12052744e-03 6.49805441e-02 -1.08588982e+00
-5.93738914e-01 -4.08318967e-01 -1.06235981e+00 -1.03974104e+00
-1.72106065e-02 -3.80694389e-01 4.82032657e-01 5.86573303e-01
6.19960666e-01 2.94446379e-01 5.50304890e-01 -1.08045959e+00
-1.71690732e-01 -5.30370057e-01 -2.55995184e-01 3.58799070e-01
6.98908940e-02 -7.65398860e-01 -2.76066661e-01 -1.41207755e-01] | [8.843924522399902, -2.2401058673858643] |
7692a43a-86a1-4fa7-8110-bc41b093a28b | attention-based-dynamic-subspace-learners-for | 2206.09068 | null | https://arxiv.org/abs/2206.09068v1 | https://arxiv.org/pdf/2206.09068v1.pdf | Attention-based Dynamic Subspace Learners for Medical Image Analysis | Learning similarity is a key aspect in medical image analysis, particularly in recommendation systems or in uncovering the interpretation of anatomical data in images. Most existing methods learn such similarities in the embedding space over image sets using a single metric learner. Images, however, have a variety of object attributes such as color, shape, or artifacts. Encoding such attributes using a single metric learner is inadequate and may fail to generalize. Instead, multiple learners could focus on separate aspects of these attributes in subspaces of an overarching embedding. This, however, implies the number of learners to be found empirically for each new dataset. This work, Dynamic Subspace Learners, proposes to dynamically exploit multiple learners by removing the need of knowing apriori the number of learners and aggregating new subspace learners during training. Furthermore, the visual interpretability of such subspace learning is enforced by integrating an attention module into our method. This integrated attention mechanism provides a visual insight of discriminative image features that contribute to the clustering of image sets and a visual explanation of the embedding features. The benefits of our attention-based dynamic subspace learners are evaluated in the application of image clustering, image retrieval, and weakly supervised segmentation. Our method achieves competitive results with the performances of multiple learners baselines and significantly outperforms the classification network in terms of clustering and retrieval scores on three different public benchmark datasets. Moreover, our attention maps offer a proxy-labels, which improves the segmentation accuracy up to 15% in Dice scores when compared to state-of-the-art interpretation techniques. | ['Herve Lombaert', 'Jose Dolz', 'Sukesh Adiga V'] | 2022-06-18 | null | null | null | null | ['image-clustering'] | ['computer-vision'] | [ 3.55272532e-01 1.14663221e-01 -3.56072873e-01 -3.85038137e-01
-7.87123799e-01 -7.38448620e-01 4.05673385e-01 4.20289963e-01
-5.41780651e-01 2.03599244e-01 1.72381520e-01 -1.20967224e-01
-5.84066093e-01 -3.32166910e-01 -7.50092328e-01 -9.65618551e-01
-1.03650495e-01 4.89513993e-01 1.43131077e-01 1.12148382e-01
1.19180225e-01 4.05484855e-01 -1.45807183e+00 4.93134201e-01
9.97408628e-01 9.84073937e-01 4.63778347e-01 4.03418303e-01
-1.20513655e-01 1.88518077e-01 -3.59353215e-01 -3.55186105e-01
7.89117739e-02 -2.34536424e-01 -6.93053186e-01 2.52729654e-01
5.80338538e-01 -7.38507733e-02 -1.36091456e-01 1.27750337e+00
4.00750607e-01 2.79350094e-02 8.64471734e-01 -1.06270313e+00
-9.22941506e-01 5.71017444e-01 -6.26789689e-01 4.13376212e-01
8.51312354e-02 1.55395031e-01 1.06562138e+00 -1.03533328e+00
6.29821122e-01 9.66717541e-01 2.90243149e-01 5.20964146e-01
-1.34893978e+00 -4.04783249e-01 1.64612412e-01 4.43576217e-01
-1.30365813e+00 -7.60209337e-02 9.07866836e-01 -5.59454262e-01
3.64630640e-01 4.85362649e-01 6.29152715e-01 1.03125632e+00
9.44250375e-02 9.11464870e-01 1.00157392e+00 -1.85366318e-01
3.09490949e-01 4.51436907e-01 3.22589368e-01 7.74169028e-01
2.54416078e-01 -2.71579742e-01 -6.07812285e-01 -3.15974019e-02
5.80431223e-01 3.21087509e-01 -3.00692916e-01 -8.05228889e-01
-1.42782414e+00 8.00812542e-01 7.69288123e-01 4.17365253e-01
-2.80644864e-01 6.83457451e-03 2.68807948e-01 6.46310672e-02
2.87955850e-01 7.05226481e-01 -1.94757029e-01 2.60715246e-01
-6.94111347e-01 -2.65082121e-01 2.33229086e-01 5.90078413e-01
5.95954776e-01 -2.16308892e-01 -2.45694861e-01 6.78669155e-01
7.91589692e-02 3.85652065e-01 6.85407817e-01 -9.50796843e-01
2.94175535e-01 9.80026424e-01 -4.53157067e-01 -1.08083189e+00
-4.84327525e-01 -6.88120484e-01 -9.61239457e-01 1.90939456e-01
2.11469322e-01 2.02232242e-01 -9.43607330e-01 1.75890958e+00
3.82359266e-01 2.82002896e-01 3.39930318e-02 1.07461429e+00
7.54978120e-01 2.89739579e-01 -3.61924060e-02 -1.15752898e-01
1.41776097e+00 -9.44765091e-01 -6.04602933e-01 -9.40780714e-02
5.88665783e-01 -3.78038585e-01 1.31970417e+00 4.20441359e-01
-7.61251152e-01 -6.32562459e-01 -1.05366921e+00 4.73653898e-02
-3.50811601e-01 1.46862656e-01 6.79996073e-01 5.21616280e-01
-9.47087049e-01 4.93704736e-01 -8.75237942e-01 -2.59406596e-01
6.78726792e-01 5.72005928e-01 -5.24052918e-01 -9.06532183e-02
-7.37302899e-01 5.36460578e-01 3.60824674e-01 -5.66908009e-02
-7.06472158e-01 -8.85673523e-01 -7.25372016e-01 1.12552717e-01
3.39970708e-01 -6.88009083e-01 4.23746258e-01 -1.18053722e+00
-9.53260124e-01 8.45960140e-01 -3.20334546e-02 -2.40018547e-01
4.05052036e-01 -8.49399194e-02 -3.31373066e-01 5.36371052e-01
9.49960500e-02 8.09224665e-01 9.75440979e-01 -1.46906185e+00
-4.07772303e-01 -6.77815557e-01 1.53092578e-01 4.00705010e-01
-9.33064103e-01 -4.29319322e-01 -5.67917645e-01 -6.63634121e-01
3.41477633e-01 -1.04293895e+00 -3.02914858e-01 4.41061825e-01
-2.72341907e-01 -8.11104476e-02 9.13876712e-01 -4.62059468e-01
1.02392220e+00 -2.47036791e+00 5.60292244e-01 4.40377802e-01
4.54230815e-01 7.66592324e-02 -1.47160158e-01 -2.59889603e-01
-2.29864627e-01 1.59980252e-01 -3.50744516e-01 -1.73734218e-01
-2.53899813e-01 2.74179012e-01 -1.28440991e-01 4.73462254e-01
1.04313470e-01 9.17752326e-01 -1.03521693e+00 -6.67427242e-01
2.90952206e-01 4.96560663e-01 -6.26561642e-01 -4.34550829e-02
1.91590980e-01 5.83973110e-01 -3.37963223e-01 5.09058297e-01
3.57972652e-01 -5.67580879e-01 3.63337964e-01 -4.62524503e-01
4.57814604e-01 -2.63466626e-01 -1.11496568e+00 2.17488456e+00
-2.83598781e-01 4.08049554e-01 -2.61091322e-01 -1.20554316e+00
5.47801971e-01 2.03357130e-01 7.58011997e-01 -3.49354297e-01
-6.41816705e-02 1.80405393e-01 2.78308779e-01 -6.12742662e-01
6.66177049e-02 4.30714898e-02 7.40186572e-02 5.56644619e-01
2.39836410e-01 3.52768779e-01 9.28392187e-02 3.81406039e-01
8.44197452e-01 -2.80742735e-01 1.33692890e-01 -3.54750276e-01
6.24223411e-01 -3.22588772e-01 4.03790236e-01 6.07803822e-01
-1.38768867e-01 7.30661690e-01 5.20425320e-01 -4.79180336e-01
-7.55412996e-01 -1.25762296e+00 -4.06041682e-01 1.06350040e+00
4.25650448e-01 -2.82504171e-01 -6.14802837e-01 -1.14637852e+00
1.27960024e-02 5.57146788e-01 -1.06405914e+00 -5.33958316e-01
-3.08465809e-01 -8.02503407e-01 2.80703396e-01 6.36213481e-01
7.53288791e-02 -9.51166749e-01 -7.39953160e-01 -2.20162749e-01
-1.54505894e-01 -1.03773642e+00 -6.65660083e-01 3.27203184e-01
-1.02442062e+00 -1.13175619e+00 -5.02778649e-01 -7.40759075e-01
1.19786990e+00 3.66878331e-01 6.74463451e-01 2.01985851e-01
-4.62507993e-01 6.87797308e-01 -2.45922446e-01 -1.70051709e-01
-1.88693076e-01 1.04872204e-01 1.21284448e-01 3.89951468e-01
1.54298022e-01 -5.30757189e-01 -8.39342117e-01 1.86751887e-01
-1.19297051e+00 -2.16873158e-02 7.43904769e-01 1.03275168e+00
6.82193279e-01 -9.39260572e-02 4.06220257e-01 -9.65195298e-01
2.76935130e-01 -5.22455871e-01 -1.82121769e-01 5.44727385e-01
-6.81788266e-01 4.40573454e-01 5.78989208e-01 -6.30058289e-01
-6.82491004e-01 2.10555181e-01 3.84995997e-01 -7.75935531e-01
-7.57557601e-02 4.32961613e-01 -2.01635540e-01 -4.50347289e-02
6.05455101e-01 2.08025441e-01 2.46094659e-01 -2.78981984e-01
4.33872730e-01 3.14635038e-01 4.90464270e-01 -5.04492164e-01
8.06726992e-01 7.52538562e-01 3.90589572e-02 -6.38152957e-01
-8.77779722e-01 -7.04185784e-01 -9.90429938e-01 -2.61951447e-01
1.01332760e+00 -6.47697568e-01 -4.96609092e-01 2.60447506e-02
-8.36830616e-01 1.18810251e-01 -3.42642218e-01 4.36999887e-01
-3.68452072e-01 6.17499828e-01 -2.41748109e-01 -3.48081648e-01
-2.33264998e-01 -1.49923384e+00 9.89252985e-01 4.81126979e-02
-2.49715284e-01 -1.17721105e+00 -2.43880376e-01 5.48388541e-01
3.74949537e-02 2.25197539e-01 1.27476525e+00 -7.74245799e-01
-5.67902625e-01 -5.49709536e-02 -3.96583304e-02 1.80808485e-01
2.22769633e-01 -3.27451617e-01 -9.42160368e-01 -4.74366784e-01
-6.34857938e-02 -2.73631483e-01 1.02336538e+00 3.92850488e-01
1.82964444e+00 -2.94101715e-01 -4.48886424e-01 8.96605432e-01
1.35285521e+00 9.48079899e-02 2.51163393e-01 1.71440661e-01
9.17389691e-01 8.22014451e-01 3.71119857e-01 2.25753456e-01
2.22419187e-01 6.54338419e-01 5.40849745e-01 -3.50589514e-01
-7.04129338e-02 1.03367195e-02 2.46736780e-01 7.08241165e-01
-9.67628062e-02 1.43940568e-01 -9.23914075e-01 5.36698401e-01
-1.81247830e+00 -7.28317142e-01 1.63582683e-01 2.22959161e+00
7.01457322e-01 1.21493600e-02 -8.17650855e-02 8.58660266e-02
6.18920147e-01 -1.92283019e-02 -9.33480501e-01 -1.27212241e-01
-1.48768798e-01 -3.98236839e-03 3.97455275e-01 2.64045715e-01
-1.18794990e+00 6.03380263e-01 5.21904898e+00 7.51944721e-01
-1.25488782e+00 2.30245918e-01 6.80017948e-01 -2.92732894e-01
-6.30878925e-01 -2.19470873e-01 -3.28597814e-01 4.30884212e-01
4.66785342e-01 -1.62762739e-02 4.39958364e-01 7.00242639e-01
-7.73190632e-02 1.42063245e-01 -1.34063494e+00 1.05124319e+00
5.92670202e-01 -1.34652364e+00 5.79971969e-01 2.72034407e-01
7.52478957e-01 -2.21616372e-01 7.31057644e-01 2.63135768e-02
-2.33588114e-01 -1.13879311e+00 4.65253115e-01 4.26748574e-01
7.85132229e-01 -7.07888842e-01 6.73529804e-01 1.83292791e-01
-9.43100572e-01 -3.18105221e-01 -2.57643908e-01 6.13825858e-01
-2.45080248e-01 1.68583140e-01 -8.64787042e-01 5.31890154e-01
7.85506308e-01 8.18239033e-01 -1.03116465e+00 8.78293276e-01
-4.94347885e-02 3.55866820e-01 -5.66771217e-02 1.99456170e-01
3.41892570e-01 -1.28984779e-01 5.08495152e-01 8.96996677e-01
1.36015221e-01 9.14068967e-02 1.25642776e-01 7.56488621e-01
7.30202207e-03 1.71743348e-01 -5.27038991e-01 1.79385066e-01
1.28428981e-01 1.43256748e+00 -9.53393638e-01 -3.11875224e-01
-2.85261095e-01 1.09796095e+00 2.72301584e-01 4.24640566e-01
-7.92951882e-01 2.03299262e-02 5.90801835e-01 2.03945309e-01
3.33171993e-01 -1.31568061e-02 -5.34192979e-01 -1.13685763e+00
1.35909051e-01 -9.33387458e-01 7.01914966e-01 -4.29719627e-01
-1.17228878e+00 4.86228049e-01 -6.17360184e-03 -1.43087411e+00
1.13662379e-02 -6.93407834e-01 -3.80755186e-01 3.30852389e-01
-1.27059853e+00 -1.19653225e+00 -4.32678431e-01 7.10982502e-01
6.10122502e-01 -3.29751253e-01 7.86080420e-01 2.39250481e-01
-3.98349136e-01 8.18371594e-01 2.86361128e-01 1.30782247e-01
8.06149006e-01 -1.45475948e+00 -2.90021956e-01 6.84573650e-01
6.70419455e-01 7.19169438e-01 4.22889918e-01 -2.42305607e-01
-1.39727199e+00 -1.01249814e+00 3.29497844e-01 -5.91027439e-01
5.57544291e-01 -3.67109448e-01 -1.14407146e+00 4.11977410e-01
-1.74059551e-02 2.72554517e-01 1.13405943e+00 1.72088444e-01
-4.64856237e-01 -3.58998090e-01 -8.56090724e-01 5.71251392e-01
9.59807038e-01 -5.12345254e-01 -4.62256074e-01 3.53566647e-01
6.15202248e-01 -1.14826553e-01 -8.96595418e-01 4.39620882e-01
7.07745850e-01 -7.96785712e-01 1.17865002e+00 -8.30146670e-01
4.48095828e-01 -3.99513483e-01 -1.69259667e-01 -1.14690804e+00
-3.79260361e-01 -1.59614999e-02 -3.80701758e-02 9.08164561e-01
4.53760654e-01 -3.28703195e-01 6.69352710e-01 5.09065449e-01
-1.83574393e-01 -1.07669187e+00 -9.19586062e-01 -6.05575264e-01
9.23336819e-02 -2.80652285e-01 3.61162424e-01 1.18839860e+00
-4.43603694e-02 3.23965579e-01 -1.33665159e-01 3.30902725e-01
8.99922013e-01 1.87418163e-01 4.83805269e-01 -1.24351060e+00
-3.75478834e-01 -5.50897300e-01 -7.51700222e-01 -6.17015183e-01
3.33358377e-01 -1.44642937e+00 -3.80350351e-01 -1.31306875e+00
6.61207736e-01 -4.08583492e-01 -8.03178787e-01 4.89597410e-01
-3.62343132e-01 3.18894416e-01 2.44168326e-01 4.45420176e-01
-8.37537646e-01 4.67754304e-01 1.21292019e+00 -5.90540528e-01
-1.26587272e-01 -2.36969590e-01 -8.11838210e-01 7.67254174e-01
5.77300549e-01 -4.60689485e-01 -5.85476220e-01 -5.46176314e-01
6.50287941e-02 -2.15096876e-01 5.83106101e-01 -8.59607875e-01
3.95389020e-01 1.56684607e-01 6.46929502e-01 -3.68617147e-01
2.39386663e-01 -9.70814347e-01 -7.63794929e-02 6.42318606e-01
-5.31751812e-01 -8.51473063e-02 -7.47255981e-02 8.63290489e-01
-2.09804744e-01 -1.45236611e-01 7.64018059e-01 -2.83204541e-02
-8.01162183e-01 3.98818791e-01 3.61428373e-02 8.00163001e-02
1.16337597e+00 -4.67277437e-01 -5.18582314e-02 -7.02928528e-02
-9.30000305e-01 2.52891690e-01 4.22046095e-01 6.43537343e-01
9.18196619e-01 -1.29712713e+00 -5.20001590e-01 2.83854127e-01
4.53395844e-01 -4.07907367e-02 4.26512778e-01 1.09988558e+00
-2.43149862e-01 1.50426701e-01 -3.06733519e-01 -1.15741551e+00
-1.32898653e+00 7.91900814e-01 2.38742471e-01 -2.10222110e-01
-6.27061069e-01 6.13895416e-01 7.49632657e-01 -2.38641486e-01
3.23078543e-01 -2.56195962e-01 -4.21855569e-01 1.61029413e-01
3.35270166e-01 1.65858567e-01 -4.04831357e-02 -5.19039929e-01
-4.70404178e-01 8.64201784e-01 -1.35902777e-01 4.75351997e-02
1.24212849e+00 -6.29446357e-02 -4.88169752e-02 6.84238017e-01
1.32675469e+00 -2.01796860e-01 -1.10888028e+00 -3.49443197e-01
1.42366169e-02 -4.09556746e-01 -7.94926472e-03 -7.03049839e-01
-1.22033489e+00 1.03713143e+00 9.99254525e-01 -3.03667098e-01
1.09427059e+00 4.56018686e-01 5.17541826e-01 3.71322989e-01
1.14572503e-01 -9.72552657e-01 5.66262424e-01 -4.66411896e-02
8.37107420e-01 -1.67469645e+00 4.60267030e-02 -2.77436495e-01
-8.30783427e-01 1.10572171e+00 6.52322590e-01 6.37101382e-02
6.61813438e-01 -5.71583249e-02 1.63419202e-01 -3.44822317e-01
-5.26475728e-01 -6.38064593e-02 9.03640866e-01 4.80892688e-01
2.67150939e-01 4.14412096e-02 -2.23783046e-01 5.39554358e-01
4.49390151e-02 -6.79675758e-01 1.64176866e-01 4.01449710e-01
-2.54251271e-01 -8.93396795e-01 -2.36976430e-01 4.93380755e-01
-3.85408133e-01 -4.45040017e-02 -3.31651658e-01 6.00102723e-01
3.46228212e-01 5.25955319e-01 1.41929448e-01 -3.57257992e-01
4.92234863e-02 -8.85838047e-02 3.45871776e-01 -6.79062605e-01
-4.33068931e-01 1.33645805e-02 -5.89380860e-01 -5.05029976e-01
-4.59853321e-01 -7.96114385e-01 -1.35820997e+00 3.59052420e-01
-2.08445922e-01 7.67987892e-02 5.37549794e-01 8.79297554e-01
2.73314238e-01 5.95473111e-01 6.48151517e-01 -4.42204416e-01
-4.49923366e-01 -5.06675541e-01 -5.29784024e-01 8.18887532e-01
2.57777959e-01 -7.70283639e-01 -3.52055132e-01 1.55601799e-01] | [14.710687637329102, -2.158618211746216] |
8165008c-5af9-4286-995b-8be3eafbe104 | semantic2graph-graph-based-multi-modal | 2209.05653 | null | https://arxiv.org/abs/2209.05653v4 | https://arxiv.org/pdf/2209.05653v4.pdf | Semantic2Graph: Graph-based Multi-modal Feature Fusion for Action Segmentation in Videos | Video action segmentation and recognition tasks have been widely applied in many fields. Most previous studies employ large-scale, high computational visual models to understand videos comprehensively. However, few studies directly employ the graph model to reason about the video. The graph model provides the benefits of fewer parameters, low computational cost, a large receptive field, and flexible neighborhood message aggregation. In this paper, we present a graph-based method named Semantic2Graph, to turn the video action segmentation and recognition problem into node classification of graphs. To preserve fine-grained relations in videos, we construct the graph structure of videos at the frame-level and design three types of edges: temporal, semantic, and self-loop. We combine visual, structural, and semantic features as node attributes. Semantic edges are used to model long-term spatio-temporal relations, while the semantic features are the embedding of the label-text based on the textual prompt. A Graph Neural Networks (GNNs) model is used to learn multi-modal feature fusion. Experimental results show that Semantic2Graph achieves improvement on GTEA and 50Salads, compared to the state-of-the-art results. Multiple ablation experiments further confirm the effectiveness of semantic features in improving model performance, and semantic edges enable Semantic2Graph to capture long-term dependencies at a low cost. | ['Meng-Hsun Tsai', 'Pei-Hsuan Tsai', 'Junbin Zhang'] | 2022-09-13 | null | null | null | null | ['action-segmentation'] | ['computer-vision'] | [ 1.60160393e-01 -1.12142928e-01 -5.90536714e-01 -3.75670940e-01
-1.55614242e-01 -2.73223579e-01 4.56003398e-01 -2.33938396e-02
-3.06978405e-01 2.44780988e-01 6.52923882e-01 1.19879417e-01
-1.38108104e-01 -5.78727663e-01 -6.21721625e-01 -6.32171750e-01
-2.02448741e-01 -2.16854289e-02 6.91736221e-01 2.40983292e-02
6.31887764e-02 9.94820595e-02 -1.40556014e+00 6.94285572e-01
5.89934111e-01 1.27853382e+00 2.56225079e-01 3.98572534e-01
-1.23714887e-01 1.39803517e+00 -2.99835563e-01 -3.60284656e-01
1.93091631e-02 -6.33322179e-01 -8.57920945e-01 4.79980707e-01
2.90234953e-01 -3.57451618e-01 -9.52076077e-01 1.12487876e+00
1.72192380e-01 3.31733137e-01 4.25779819e-01 -1.63426125e+00
-8.32467616e-01 6.72442377e-01 -6.48507416e-01 2.73917258e-01
6.21972978e-01 1.87496915e-01 1.20559514e+00 -5.66726983e-01
8.06822181e-01 1.57034445e+00 6.03879571e-01 3.17285180e-01
-9.18438971e-01 -3.27953875e-01 6.03027940e-01 9.12540019e-01
-1.36115170e+00 -1.35089576e-01 8.33971500e-01 -3.67469728e-01
1.07910979e+00 1.98363602e-01 1.07153058e+00 1.32984328e+00
3.65573585e-01 1.11549997e+00 8.69627237e-01 -1.05416536e-01
6.00331798e-02 -4.62403059e-01 2.68559396e-01 1.13144994e+00
-6.29393160e-02 -1.31083071e-01 -7.01264083e-01 1.66504934e-01
7.72973359e-01 4.68558162e-01 -3.55415165e-01 -3.69527489e-01
-1.00222695e+00 8.85066092e-01 6.70598209e-01 3.23910683e-01
-4.51330304e-01 4.24048483e-01 6.16766691e-01 3.50774154e-02
4.25469548e-01 -1.84209142e-02 -1.33902669e-01 -1.59250051e-01
-5.26766479e-01 -1.14241734e-01 5.50516307e-01 1.08083725e+00
5.80615103e-01 4.68365885e-02 -5.64455152e-01 6.67127848e-01
4.64825034e-01 4.25084352e-01 5.90182066e-01 -1.12325823e+00
3.52549374e-01 1.04190254e+00 -7.85164058e-01 -1.56137276e+00
-4.64044839e-01 -1.26811035e-03 -8.11807036e-01 -3.45016420e-01
-9.03202519e-02 2.69651562e-01 -1.16879129e+00 1.53422081e+00
1.90494835e-01 5.60474455e-01 -1.15854375e-01 9.99852300e-01
1.11456704e+00 5.84044039e-01 3.40426475e-01 -1.44160137e-01
1.35251009e+00 -1.33397579e+00 -9.61637497e-01 -3.70560884e-01
8.29775214e-01 -2.61680067e-01 9.03334498e-01 -7.19444156e-02
-8.20023954e-01 -6.00400031e-01 -6.88497782e-01 -8.91485214e-02
-4.26592112e-01 -8.49535093e-02 7.20699370e-01 1.28691122e-01
-1.10538077e+00 3.15016448e-01 -9.28661585e-01 -5.78277946e-01
7.09666610e-01 9.08929780e-02 -4.74725604e-01 -2.82194108e-01
-1.33388221e+00 4.74798441e-01 6.78135872e-01 -1.48025183e-02
-9.59080338e-01 -1.42732501e-01 -1.26603281e+00 1.03401423e-01
8.32030475e-01 -7.51832008e-01 6.17261469e-01 -1.13935733e+00
-9.61281002e-01 6.98184311e-01 -2.88387328e-01 -4.08109874e-01
1.12281799e-01 -3.14272009e-02 -4.24255699e-01 7.60760009e-01
1.42091051e-01 7.47564256e-01 8.27433407e-01 -1.03904533e+00
-6.45149171e-01 -5.16384900e-01 3.98474216e-01 4.61032629e-01
-5.52325964e-01 6.19263435e-03 -1.09437835e+00 -8.26333821e-01
1.85897544e-01 -9.61021543e-01 -1.33491531e-01 -2.28779137e-01
-3.81643236e-01 -3.46420974e-01 1.14326894e+00 -8.05317700e-01
1.45737028e+00 -2.31711102e+00 4.30800736e-01 1.48717016e-01
5.04977226e-01 1.09606549e-01 -4.31508303e-01 3.34125519e-01
4.85191345e-02 1.25880346e-01 1.33274812e-02 -6.82089180e-02
-6.30412474e-02 4.65138882e-01 2.14676801e-02 3.89091194e-01
-4.37074862e-02 1.42145574e+00 -8.76883924e-01 -8.56835485e-01
3.14712971e-01 5.00285983e-01 -5.92053175e-01 1.65446818e-01
-1.96832329e-01 1.98948488e-01 -9.18610990e-01 6.65211439e-01
4.63965945e-02 -6.80590212e-01 3.53364736e-01 -5.66124976e-01
4.37661171e-01 -1.32215112e-01 -8.82776737e-01 2.00030994e+00
7.97692686e-02 5.48470318e-01 -1.18808016e-01 -1.28091741e+00
5.10177493e-01 -2.99027376e-02 7.93974578e-01 -8.83554161e-01
2.71329343e-01 -4.17470366e-01 -3.53635639e-01 -8.67590487e-01
3.12447369e-01 4.05663997e-01 -5.18974103e-02 8.36756378e-02
2.72551566e-01 3.57160360e-01 4.39969629e-01 6.64475083e-01
1.21548605e+00 4.25381750e-01 1.30307898e-01 -1.63704976e-02
4.00244206e-01 -1.26481265e-01 6.07221305e-01 4.91629064e-01
-3.71573776e-01 4.64711696e-01 6.75400198e-01 -2.92885303e-01
-3.73771966e-01 -7.68767178e-01 5.44203997e-01 1.30887651e+00
7.83849299e-01 -9.93457198e-01 -9.22059774e-01 -1.00719285e+00
-9.40608904e-02 4.58237171e-01 -7.26851702e-01 -5.45647919e-01
-5.56396842e-01 -3.02827924e-01 4.75798994e-01 7.99796939e-01
8.40946078e-01 -1.15609062e+00 -3.11898500e-01 -1.58025131e-01
-5.86050928e-01 -1.59855926e+00 -7.35415876e-01 -3.53091627e-01
-6.88479483e-01 -1.30198157e+00 -2.63344169e-01 -9.03653145e-01
5.74990809e-01 4.90996629e-01 9.78992343e-01 2.50024617e-01
-2.74003774e-01 9.42462444e-01 -8.44589829e-01 2.77489990e-01
5.05444258e-02 -2.73509949e-01 -2.08731577e-01 3.32593918e-01
5.11763692e-01 -3.73717785e-01 -5.55914164e-01 4.15805817e-01
-1.11433744e+00 1.60444185e-01 6.04386270e-01 6.89141929e-01
6.90575004e-01 2.38050073e-01 1.88991398e-01 -7.63617456e-01
3.21108460e-01 -3.74224544e-01 -5.79651408e-02 5.14156759e-01
-5.16185105e-01 -6.18662350e-02 4.53441173e-01 -3.73832256e-01
-8.86258662e-01 1.58393215e-02 2.09441081e-01 -8.88818562e-01
-1.59229502e-01 7.12736428e-01 -3.62578809e-01 -9.93233323e-02
1.93240136e-01 3.86148244e-01 1.20576154e-02 -1.27874002e-01
6.04181230e-01 3.48618448e-01 2.92847693e-01 -2.89535969e-01
3.65676880e-01 5.32910287e-01 1.06562734e-01 -8.44006836e-01
-1.01783252e+00 -5.96947908e-01 -6.94164991e-01 -6.33265257e-01
1.34500825e+00 -9.85644937e-01 -6.36114657e-01 4.85561788e-01
-8.49543452e-01 -3.39525133e-01 -1.74395934e-01 5.80629587e-01
-7.49250114e-01 7.85507858e-01 -7.20573664e-01 -2.94224739e-01
4.15903553e-02 -1.16918766e+00 1.13180399e+00 1.58001900e-01
2.72441539e-04 -1.26060808e+00 -4.23534513e-01 6.64305389e-01
-6.19685315e-02 4.12903905e-01 7.50605941e-01 -5.58532596e-01
-7.69166529e-01 7.23965792e-03 -4.72486019e-01 3.39989632e-01
8.32439214e-02 -1.87128946e-01 -5.99236250e-01 -1.88464880e-01
-2.22872943e-01 -4.11569089e-01 1.24520695e+00 6.01591229e-01
1.42559993e+00 -3.97864133e-01 -6.73802555e-01 6.87948883e-01
1.12174404e+00 1.34864375e-01 6.25664949e-01 1.21309809e-01
1.45839632e+00 4.96216625e-01 6.60713255e-01 3.43685359e-01
7.39843547e-01 5.75289488e-01 4.94377971e-01 -7.11074322e-02
-5.28724432e-01 -4.79904562e-01 6.70321226e-01 8.85271728e-01
-2.13597521e-01 -4.33882236e-01 -7.46438146e-01 3.58075261e-01
-2.34485149e+00 -1.16418576e+00 -1.75051272e-01 1.51084793e+00
4.46326315e-01 1.08438926e-02 7.70174339e-02 -2.09648788e-01
8.00210297e-01 6.58740282e-01 -4.40103084e-01 1.84396163e-01
-2.90915608e-01 -4.86711472e-01 4.98067647e-01 2.97748834e-01
-1.32184696e+00 1.24557984e+00 5.74473143e+00 1.10233271e+00
-7.94631958e-01 1.01559581e-02 5.73277831e-01 9.99924913e-02
-1.11823939e-01 2.97979899e-02 -5.02811849e-01 4.27981287e-01
6.63559437e-01 1.29279211e-01 5.80275893e-01 7.02728033e-01
1.35301143e-01 -1.29981907e-02 -1.10778511e+00 1.29247546e+00
6.12647176e-01 -1.28680348e+00 4.49149251e-01 -1.29527390e-01
5.80867708e-01 -9.37689915e-02 -1.94365501e-01 2.53336310e-01
1.98436439e-01 -8.67667019e-01 6.95592165e-01 7.38828719e-01
5.96413612e-01 -4.49606240e-01 5.38334727e-01 2.33557541e-02
-1.74016631e+00 -2.28715658e-01 -7.76927397e-02 3.37859988e-02
4.02404159e-01 1.11856297e-01 -2.12868184e-01 8.26458871e-01
8.98387492e-01 1.65845978e+00 -8.31230402e-01 6.57781184e-01
-3.40038985e-01 6.46818101e-01 2.77762320e-02 -4.36801687e-02
5.72264850e-01 -3.53706807e-01 4.38136786e-01 1.15019345e+00
-3.07054017e-02 3.78790587e-01 8.27764690e-01 6.21736884e-01
-2.42275419e-03 -3.27537395e-02 -6.15352809e-01 -4.40197408e-01
6.79251924e-02 1.07234645e+00 -9.54941630e-01 -4.61453259e-01
-7.42761195e-01 1.25498939e+00 2.54892468e-01 7.42873549e-01
-1.06308877e+00 -1.81558028e-01 4.61051047e-01 -6.54187575e-02
3.77301246e-01 -2.74762630e-01 -5.41766407e-03 -1.24158859e+00
7.02327564e-02 -7.90220141e-01 8.85547936e-01 -8.67056727e-01
-1.19403481e+00 5.08308232e-01 1.82650462e-01 -1.07489789e+00
-1.12787142e-01 -5.59106350e-01 -2.00667292e-01 1.49489731e-01
-1.16616082e+00 -1.46212077e+00 -5.62308073e-01 9.93820131e-01
6.93606675e-01 -1.50962904e-01 4.75903183e-01 2.23094612e-01
-5.92126310e-01 2.77991205e-01 -3.22956711e-01 6.02328897e-01
2.52879530e-01 -8.92342091e-01 -7.09586442e-02 8.64199817e-01
4.87385184e-01 2.41229862e-01 2.03123361e-01 -9.23683703e-01
-1.54833555e+00 -1.36823571e+00 5.54917336e-01 -3.77338111e-01
8.31810951e-01 -2.87845463e-01 -7.72905529e-01 8.72021973e-01
1.84498608e-01 3.32124949e-01 5.28080583e-01 -2.49050800e-02
-4.43818539e-01 3.72729227e-02 -6.05612516e-01 5.70604146e-01
1.68045819e+00 -7.48268902e-01 -5.36905825e-01 2.76680708e-01
9.67572629e-01 -1.69333503e-01 -8.33860278e-01 4.65189636e-01
4.58791018e-01 -8.97282243e-01 1.03139889e+00 -7.91201353e-01
3.90300244e-01 -2.35233173e-01 -2.85535932e-01 -1.14718986e+00
-5.95385134e-01 -3.81795436e-01 -2.85237372e-01 1.10364819e+00
1.22637779e-01 -4.64784205e-01 6.02822185e-01 4.25087094e-01
-2.10314646e-01 -7.60186851e-01 -7.22129941e-01 -8.11911523e-01
-5.64849198e-01 -3.90105367e-01 2.21124738e-01 1.06895065e+00
7.46220127e-02 6.01942658e-01 -4.79649723e-01 5.63893467e-03
4.48416829e-01 2.24970996e-01 4.05604750e-01 -9.36369061e-01
-1.27692267e-01 -5.43139696e-01 -9.98947740e-01 -1.21064651e+00
6.15652561e-01 -1.02352130e+00 -1.48245126e-01 -1.88971961e+00
6.07735813e-01 1.16988786e-01 -4.41900194e-01 8.26855600e-01
-2.84130514e-01 2.28489369e-01 3.31541538e-01 1.12638056e-01
-1.45789325e+00 9.22610223e-01 1.36846471e+00 -4.31107372e-01
-3.62468213e-02 -4.92129654e-01 -4.17316735e-01 9.02802646e-01
5.10395348e-01 -2.48654932e-01 -8.05405676e-01 -5.10589659e-01
-4.67194468e-02 -6.93010092e-02 5.10576904e-01 -8.41828287e-01
3.12743574e-01 -3.05931568e-01 3.35427225e-01 -2.47951552e-01
4.31895733e-01 -9.41002131e-01 -1.35512175e-02 2.12421536e-01
-4.65564430e-01 -1.60810471e-01 -1.27950877e-01 1.06530797e+00
-5.02831876e-01 3.14844787e-01 4.32812810e-01 -6.41417205e-02
-1.34863126e+00 7.56717622e-01 -2.25593731e-01 1.69486269e-01
1.27450252e+00 -4.39822108e-01 -3.51931900e-01 -6.09926224e-01
-9.47330713e-01 4.86264884e-01 4.09394443e-01 8.98231030e-01
9.82813060e-01 -1.65739107e+00 -4.28972661e-01 8.91900361e-02
3.01303059e-01 -3.95019412e-01 5.61842799e-01 9.99561727e-01
-2.43587345e-01 3.08240294e-01 -1.69895887e-01 -7.84991384e-01
-1.46785665e+00 7.11426377e-01 9.76570547e-02 -9.12858397e-02
-9.05487061e-01 8.52126658e-01 6.98557556e-01 1.24084018e-01
3.36473733e-01 -3.07206601e-01 -5.04924357e-01 1.39508694e-01
4.00559753e-01 3.56111825e-01 -4.25383329e-01 -1.19341516e+00
-5.25131285e-01 7.15447664e-01 2.46367380e-01 2.48902053e-01
1.11869800e+00 -3.08553278e-01 -1.99982092e-01 3.92176628e-01
1.35160708e+00 -5.23988068e-01 -1.35097778e+00 -4.20312047e-01
-2.04879671e-01 -5.49943745e-01 2.14628235e-01 -5.67389131e-01
-1.51947010e+00 7.03057468e-01 3.51572216e-01 3.71575952e-01
1.31546712e+00 4.18452263e-01 8.42620552e-01 1.70292705e-01
2.09968343e-01 -1.17273152e+00 4.87124920e-01 5.20773292e-01
8.25610220e-01 -1.07181728e+00 1.46337105e-02 -6.87813222e-01
-1.09153032e+00 9.63289857e-01 7.91605115e-01 1.73089409e-03
6.47812068e-01 -1.39069885e-01 -6.35795742e-02 -5.76846361e-01
-6.13931298e-01 -5.64314246e-01 7.18491852e-01 4.95578587e-01
-2.08315831e-02 5.76969124e-02 -1.68863833e-01 6.27385020e-01
4.10008252e-01 -1.69822887e-01 1.45436123e-01 9.29171622e-01
-3.31052631e-01 -7.19615042e-01 2.06023410e-01 5.71435213e-01
-2.51089752e-01 -6.42117485e-02 -6.37809813e-01 6.34523869e-01
5.73654240e-03 1.12344754e+00 5.44805638e-02 -8.20528865e-01
2.07455456e-01 -1.78902984e-01 3.50326151e-01 -3.78986508e-01
-2.02514499e-01 1.46168947e-01 2.00147629e-01 -1.40305436e+00
-9.38421786e-01 -4.93123442e-01 -1.65287900e+00 -1.26654789e-01
-1.34887844e-01 -1.28513211e-02 1.65442973e-01 1.11787391e+00
6.59810424e-01 7.16093183e-01 4.30442393e-01 -5.50039411e-01
-4.54627797e-02 -7.70657420e-01 -7.14082658e-01 9.40781474e-01
2.32600849e-02 -8.51439357e-01 -2.06615254e-01 4.60159212e-01] | [9.739890098571777, 0.8117493391036987] |
9e3601d7-4b3d-4e5c-adbb-623be18460aa | frame-subtitle-self-supervision-for-multi | 2209.03609 | null | https://arxiv.org/abs/2209.03609v1 | https://arxiv.org/pdf/2209.03609v1.pdf | Frame-Subtitle Self-Supervision for Multi-Modal Video Question Answering | Multi-modal video question answering aims to predict correct answer and localize the temporal boundary relevant to the question. The temporal annotations of questions improve QA performance and interpretability of recent works, but they are usually empirical and costly. To avoid the temporal annotations, we devise a weakly supervised question grounding (WSQG) setting, where only QA annotations are used and the relevant temporal boundaries are generated according to the temporal attention scores. To substitute the temporal annotations, we transform the correspondence between frames and subtitles to Frame-Subtitle (FS) self-supervision, which helps to optimize the temporal attention scores and hence improve the video-language understanding in VideoQA model. The extensive experiments on TVQA and TVQA+ datasets demonstrate that the proposed WSQG strategy gets comparable performance on question grounding, and the FS self-supervision helps improve the question answering and grounding performance on both QA-supervision only and full-supervision settings. | ['Weike Jin', 'Zhou Zhao', 'Jiong Wang'] | 2022-09-08 | null | null | null | null | ['video-question-answering'] | ['computer-vision'] | [ 1.31540313e-01 1.03671677e-01 -7.84968361e-02 -6.38090670e-01
-1.25109828e+00 -6.24850035e-01 4.45643097e-01 -1.52362540e-01
-3.07852805e-01 5.48437178e-01 4.61659193e-01 -2.34537765e-01
-1.37519268e-02 -5.02393961e-01 -8.59807789e-01 -4.82396036e-01
2.11851999e-01 3.36962134e-01 8.66160393e-01 -2.22055450e-01
-3.64906974e-02 -3.98151070e-01 -1.41348159e+00 8.53340924e-01
9.58507001e-01 1.19288039e+00 3.97703797e-01 1.00037897e+00
-3.35240573e-01 1.60431659e+00 -5.22722065e-01 -6.86711133e-01
-2.40847468e-04 -8.89668167e-01 -1.43388700e+00 1.99003771e-01
7.81308174e-01 -5.34301221e-01 -7.10413456e-01 8.77348244e-01
2.24085256e-01 4.63431031e-01 7.48073459e-02 -1.20158517e+00
-8.17747355e-01 2.05726624e-01 -4.49336857e-01 7.55261004e-01
8.31722498e-01 3.45200062e-01 1.36838114e+00 -7.15110123e-01
5.78191578e-01 1.44976938e+00 2.69448459e-01 5.95566690e-01
-6.16947353e-01 -3.26139957e-01 6.09058559e-01 9.35846865e-01
-1.12850416e+00 -4.90576893e-01 6.77517414e-01 -2.07554415e-01
7.10496426e-01 3.50685149e-01 4.23707038e-01 9.58770752e-01
-5.86563833e-02 1.16162431e+00 7.17505455e-01 -2.30647400e-01
2.44210605e-02 -4.58620340e-01 3.38820338e-01 1.13593113e+00
-8.08192432e-01 -3.20899755e-01 -8.47904444e-01 1.22650683e-01
6.68933094e-01 -1.09192871e-01 -5.12908757e-01 1.09032830e-02
-1.42162764e+00 7.58590877e-01 2.86444098e-01 2.26892039e-01
-3.40082854e-01 3.54067475e-01 4.86894101e-01 4.31632847e-01
4.39534992e-01 1.48944825e-01 -4.71205741e-01 -2.91007251e-01
-8.00157011e-01 2.21078321e-01 2.61614114e-01 1.09788144e+00
7.81965792e-01 -3.83555502e-01 -9.06244934e-01 6.78559840e-01
1.59645334e-01 3.50955963e-01 4.48135734e-01 -1.49056721e+00
1.04378259e+00 7.10414112e-01 3.13920230e-01 -9.75412071e-01
-1.33062363e-01 -7.50475330e-03 -3.71006638e-01 -6.96527660e-01
6.82135105e-01 -2.17172112e-02 -9.77330685e-01 1.55247211e+00
5.20882428e-01 5.35825789e-01 -5.10175228e-02 1.22288072e+00
9.95048821e-01 1.06039107e+00 2.32451841e-01 -3.45868021e-01
1.65722668e+00 -1.56577647e+00 -1.24122393e+00 -1.86600924e-01
7.98423588e-01 -4.95165467e-01 1.62344325e+00 2.71064322e-02
-1.21519876e+00 -7.45802701e-01 -5.96055448e-01 -5.74834287e-01
2.63747722e-01 8.38323236e-02 2.47010455e-01 1.24839544e-01
-8.60546887e-01 1.78424433e-01 -8.87325585e-01 4.11007786e-03
4.96772647e-01 5.26210777e-02 -1.67033315e-01 -5.31740785e-01
-1.53005922e+00 5.76697230e-01 3.02010328e-01 2.23603562e-01
-1.26203167e+00 -4.92077678e-01 -9.11866188e-01 -1.58799123e-02
8.22301209e-01 -7.54450023e-01 1.60228574e+00 -1.28441155e+00
-1.56498563e+00 8.40872288e-01 -7.67288625e-01 -5.77869833e-01
4.45360631e-01 -3.61265659e-01 -3.69450092e-01 9.42482889e-01
2.44893536e-01 8.21756661e-01 9.29680526e-01 -7.56230354e-01
-9.56432521e-01 -1.97188139e-01 8.55279207e-01 4.44419414e-01
-1.58347785e-01 9.06615481e-02 -1.00104332e+00 -4.08831835e-01
1.09218933e-01 -6.51570320e-01 -3.18956114e-02 -5.99150099e-02
3.67785878e-02 -6.01644218e-01 9.85310018e-01 -1.12476265e+00
1.46334136e+00 -1.94172347e+00 2.71491021e-01 -4.74616617e-01
3.95907424e-02 1.11815721e-01 -4.11088437e-01 7.53491223e-02
-2.29090005e-02 -2.39189863e-01 -2.06812426e-01 -1.85245231e-01
-3.21871370e-01 6.19771302e-01 -4.20706481e-01 2.11737439e-01
3.92250806e-01 1.32609558e+00 -1.18776619e+00 -9.28008258e-01
4.20685709e-02 3.59966722e-03 -6.84824169e-01 6.12032473e-01
-8.95726681e-01 5.53368211e-01 -6.98944390e-01 5.71283519e-01
2.25530043e-01 -6.85453176e-01 -2.00948998e-01 -3.51649910e-01
2.49719903e-01 4.31824446e-01 -7.27297068e-01 2.11774993e+00
-4.17625457e-01 6.14059746e-01 -1.12387896e-01 -9.41279709e-01
4.36317116e-01 6.82192147e-01 3.23649019e-01 -1.17081881e+00
-1.23213641e-01 -1.88225731e-01 -2.11590573e-01 -1.26715374e+00
4.04194236e-01 -5.68053573e-02 1.21167488e-01 3.03979486e-01
3.01054120e-01 1.03966473e-02 3.67685646e-01 4.85775620e-01
1.03465295e+00 4.94562715e-01 -7.80263618e-02 2.05303449e-02
1.02839100e+00 1.27387106e-01 5.59489369e-01 5.14565945e-01
-6.07270062e-01 8.52582633e-01 7.12210655e-01 -4.96131778e-01
-6.92655385e-01 -6.54601634e-01 5.19359171e-01 1.67722261e+00
5.01766205e-01 -5.29288113e-01 -9.29043114e-01 -1.21952951e+00
-6.62811697e-01 5.22780120e-01 -7.09383547e-01 -1.89803064e-01
-9.75495338e-01 -2.30388030e-01 4.12740231e-01 4.60689932e-01
7.18315184e-01 -1.03034890e+00 -4.65777725e-01 1.99313283e-01
-1.21880376e+00 -1.40410805e+00 -9.31890786e-01 -3.98989826e-01
-7.79627800e-01 -1.24983656e+00 -8.62067521e-01 -8.20107281e-01
4.83412713e-01 4.55058426e-01 1.38354623e+00 4.26341057e-01
3.58489752e-01 6.30729198e-01 -8.16400051e-01 6.79854453e-02
-8.70569982e-03 -9.21234488e-02 -5.05669653e-01 3.01470786e-01
2.33658969e-01 -6.91522136e-02 -8.52604091e-01 5.95140219e-01
-9.92783368e-01 1.49549380e-01 1.04912765e-01 9.20122802e-01
6.33444071e-01 -3.75940710e-01 6.79282010e-01 -6.64397657e-01
3.15141648e-01 -3.88008505e-01 -4.34597224e-01 6.64740622e-01
-4.75225486e-02 1.53400525e-01 5.07230699e-01 -3.91363382e-01
-1.33665991e+00 -1.25769839e-01 -3.22069257e-01 -6.15494192e-01
-8.03007931e-03 4.11747068e-01 -2.22034588e-01 1.42394915e-01
4.06591654e-01 1.72585726e-01 -3.42053503e-01 -1.95614338e-01
4.61179733e-01 2.51389861e-01 6.96887314e-01 -6.66975796e-01
5.52320063e-01 5.76416552e-01 -4.03903127e-01 -4.73974079e-01
-1.75242245e+00 -6.92597806e-01 -5.49252927e-01 -4.46825296e-01
1.40251887e+00 -9.01653767e-01 -7.30779111e-01 9.31544509e-03
-1.46988916e+00 -3.71497899e-01 -1.83859393e-01 2.27121606e-01
-7.30924428e-01 7.41779327e-01 -6.59507632e-01 -7.63011634e-01
-1.73390791e-01 -1.16218841e+00 1.25708425e+00 2.96393842e-01
-3.68308835e-02 -8.10236871e-01 -6.82943836e-02 1.13940370e+00
-1.23563401e-01 -1.25742957e-01 7.27104485e-01 -3.29356849e-01
-9.06328201e-01 2.50221699e-01 -4.19643134e-01 1.65341467e-01
-1.45604387e-01 -3.63964051e-01 -9.14671659e-01 -8.25301036e-02
1.99983016e-01 -6.53466403e-01 7.65944958e-01 2.87648112e-01
1.34852254e+00 -3.55152100e-01 1.45661578e-01 3.49418789e-01
8.76201630e-01 2.45817855e-01 6.64677203e-01 2.37044573e-01
7.11094558e-01 7.63236046e-01 1.25220633e+00 2.98002604e-02
7.92290509e-01 7.46117294e-01 6.42436624e-01 8.39597806e-02
-2.20555946e-01 -3.51341218e-01 5.30624926e-01 9.07784641e-01
-4.35087383e-02 -4.40372080e-01 -7.93992579e-01 8.47912014e-01
-2.19498491e+00 -1.08919382e+00 -3.18128645e-01 1.76375973e+00
9.06625926e-01 -1.93378404e-01 1.61176860e-01 -1.40098527e-01
7.28079140e-01 4.29048389e-01 -4.06900644e-01 1.74462155e-01
-4.99506854e-02 -1.16058014e-01 -7.69077837e-02 7.53600597e-01
-9.99891341e-01 1.03995156e+00 6.13937759e+00 8.17799747e-01
-6.09614551e-01 5.31248271e-01 7.37438202e-01 -1.63013056e-01
-3.41215760e-01 1.01306895e-02 -5.66194355e-01 4.46668774e-01
1.06595504e+00 1.45742789e-01 2.29987755e-01 4.04131055e-01
4.47625309e-01 -1.68313548e-01 -1.09123194e+00 9.91081059e-01
2.47546673e-01 -1.34005606e+00 1.95193768e-01 -5.78817129e-01
6.92651570e-01 -2.44993687e-01 -7.06822649e-02 4.54878896e-01
-7.16883540e-02 -7.60087729e-01 9.10872936e-01 5.72532237e-01
5.59596121e-01 -5.02859473e-01 9.17873383e-01 4.24183935e-01
-1.34860134e+00 -2.26247594e-01 -2.16490701e-01 -3.15888971e-02
7.16196477e-01 7.07303500e-03 -4.05085415e-01 7.80539036e-01
9.08177614e-01 6.77917778e-01 -6.08620226e-01 6.74507320e-01
-3.88602823e-01 8.91665936e-01 5.70642799e-02 1.73528895e-01
5.90201557e-01 -2.32891962e-01 3.02968353e-01 8.06121290e-01
9.71747339e-02 6.56229556e-01 1.09790735e-01 4.86334234e-01
-8.99062231e-02 -7.41320252e-02 1.54984280e-01 -1.04451776e-02
9.09965336e-02 8.73221874e-01 -5.08533001e-01 -5.00731945e-01
-7.31687248e-01 1.25511730e+00 2.07284629e-01 5.96047521e-01
-1.20873797e+00 -1.64551660e-01 4.23339993e-01 1.35288313e-01
4.45489258e-01 -4.60045785e-02 2.08557427e-01 -1.53113365e+00
2.49304846e-01 -1.10056341e+00 1.07463539e+00 -1.15748847e+00
-9.49806511e-01 6.33946538e-01 4.50756447e-03 -1.13968623e+00
-2.08692506e-01 -2.83491760e-01 -3.98417473e-01 4.81632739e-01
-1.79598773e+00 -1.10495555e+00 -5.80721855e-01 9.86941040e-01
1.02500165e+00 3.35238934e-01 3.12566251e-01 4.52698469e-01
-5.73079407e-01 3.86486053e-01 -4.60471302e-01 1.41200259e-01
7.75041580e-01 -1.03365338e+00 8.93944502e-02 9.99334335e-01
3.95814747e-01 2.74851829e-01 6.69796407e-01 -3.49342197e-01
-1.14147055e+00 -1.09806240e+00 8.56931210e-01 -7.99827337e-01
7.26718426e-01 -2.26690825e-02 -1.25043643e+00 7.71468103e-01
4.04886484e-01 2.30233684e-01 3.32736582e-01 -2.24110439e-01
-2.89330453e-01 -8.76895562e-02 -7.54125118e-01 4.25589323e-01
1.05431604e+00 -1.02703059e+00 -1.07834911e+00 6.19554937e-01
1.32692337e+00 -5.44079542e-01 -6.52212560e-01 1.19950809e-01
1.87172726e-01 -9.98835266e-01 9.67009485e-01 -7.71035910e-01
6.10274732e-01 -6.03898466e-01 -5.18042035e-02 -6.74582005e-01
-4.85781915e-02 -6.96174800e-01 -4.96299177e-01 1.30239534e+00
2.78618455e-01 6.58303723e-02 9.05927360e-01 5.55479586e-01
-3.47334594e-01 -6.44898117e-01 -1.10349238e+00 -4.63007063e-01
-2.71508187e-01 -3.46388012e-01 5.12927353e-01 8.62689972e-01
-1.73637882e-01 5.09445131e-01 -5.41659772e-01 2.83872157e-01
1.55432776e-01 2.42932513e-01 5.62137544e-01 -6.58497870e-01
-3.63880724e-01 5.68846352e-02 -1.63951667e-03 -1.83091402e+00
2.49245450e-01 -4.17309374e-01 2.81212240e-01 -1.56854498e+00
7.34235644e-02 2.04104424e-01 -1.63569272e-01 3.37261707e-01
-8.10718477e-01 2.31564119e-01 1.44020215e-01 2.21452191e-01
-1.56498635e+00 7.04210281e-01 1.65102386e+00 -1.58266619e-01
-8.90253559e-02 -2.21147463e-01 -2.06052378e-01 6.81260407e-01
2.03857243e-01 -3.40188950e-01 -7.28276074e-01 -1.01206958e+00
2.65580863e-01 6.59146011e-01 3.52918297e-01 -6.79584920e-01
2.92626828e-01 -3.51823658e-01 -2.60293931e-02 -7.29396522e-01
3.91666293e-01 -6.71062291e-01 -3.19086462e-01 1.75160751e-01
-5.96447706e-01 2.50573844e-01 -1.68234501e-02 8.81864071e-01
-6.97012603e-01 -3.49299669e-01 5.23288429e-01 -2.13279113e-01
-9.20233727e-01 6.92938030e-01 -1.25562698e-01 5.24737597e-01
1.04795718e+00 -8.68988037e-02 -3.79748195e-01 -7.90177345e-01
-7.09474564e-01 8.32932651e-01 3.82322855e-02 4.48491663e-01
6.61444128e-01 -1.41586721e+00 -5.25029659e-01 -3.37296396e-01
2.65579462e-01 2.59896964e-01 7.26624966e-01 9.75341022e-01
-4.80464935e-01 6.04160488e-01 9.03833881e-02 -8.01375389e-01
-1.24285102e+00 6.51869893e-01 4.73623276e-01 -3.64302009e-01
-6.04084611e-01 1.28377748e+00 6.30042136e-01 -1.67462248e-02
3.07749271e-01 -3.74498397e-01 -4.26843733e-01 8.34261551e-02
7.70506322e-01 2.76909083e-01 -9.37950015e-02 -6.90475464e-01
-1.93984076e-01 5.05827665e-01 -6.96915388e-02 -1.76373079e-01
9.48432684e-01 -7.21492112e-01 1.95057616e-02 3.45312715e-01
1.01708353e+00 -3.57143551e-01 -1.64757669e+00 -4.33287114e-01
1.92822322e-01 -5.46648502e-01 -7.64960945e-02 -5.17466009e-01
-9.56997097e-01 1.14271295e+00 2.96442956e-01 2.84432650e-01
1.24848020e+00 2.91378558e-01 1.24275696e+00 3.69345695e-01
1.47922114e-01 -8.61797154e-01 8.15463722e-01 6.47464275e-01
9.65207875e-01 -1.27660191e+00 -4.19562429e-01 -4.58714813e-01
-9.01247442e-01 9.16120589e-01 1.00894487e+00 8.77089500e-02
-4.77407128e-03 -4.65268850e-01 2.00221658e-01 -2.84182459e-01
-1.00036514e+00 -3.81911546e-01 4.88965601e-01 4.10687119e-01
3.53102654e-01 -5.80574930e-01 -1.47343382e-01 6.80334628e-01
1.55105308e-01 5.82088530e-02 2.90982693e-01 6.08577430e-01
-3.57534170e-01 -8.18460226e-01 -4.24938053e-01 1.98382244e-01
-5.65948546e-01 -4.61298712e-02 -1.38324857e-01 4.39424157e-01
9.29810405e-02 1.18484056e+00 8.16104487e-02 -1.11786760e-01
3.98253679e-01 2.77717203e-01 3.91147047e-01 -4.66077447e-01
-5.43510914e-01 2.03146096e-02 2.27043226e-01 -9.01847661e-01
-8.14998448e-01 -4.40631717e-01 -1.39179862e+00 8.80877152e-02
-5.45205712e-01 6.24098957e-01 -2.21532896e-01 1.41128528e+00
3.53533179e-01 6.94903612e-01 4.68864530e-01 -2.81764120e-01
-3.05554092e-01 -8.46501291e-01 -9.36165452e-03 6.43628418e-01
6.88938141e-01 -5.10348737e-01 -2.40533695e-01 5.95923364e-01] | [10.411940574645996, 1.039218783378601] |
59418cbd-6cb2-438c-b018-2803a8b0e7e0 | emotalk-speech-driven-emotional | 2303.11089 | null | https://arxiv.org/abs/2303.11089v1 | https://arxiv.org/pdf/2303.11089v1.pdf | EmoTalk: Speech-driven emotional disentanglement for 3D face animation | Speech-driven 3D face animation aims to generate realistic facial expressions that match the speech content and emotion. However, existing methods often neglect emotional facial expressions or fail to disentangle them from speech content. To address this issue, this paper proposes an end-to-end neural network to disentangle different emotions in speech so as to generate rich 3D facial expressions. Specifically, we introduce the emotion disentangling encoder (EDE) to disentangle the emotion and content in the speech by cross-reconstructed speech signals with different emotion labels. Then an emotion-guided feature fusion decoder is employed to generate a 3D talking face with enhanced emotion. The decoder is driven by the disentangled identity, emotional, and content embeddings so as to generate controllable personal and emotional styles. Finally, considering the scarcity of the 3D emotional talking face data, we resort to the supervision of facial blendshapes, which enables the reconstruction of plausible 3D faces from 2D emotional data, and contribute a large-scale 3D emotional talking face dataset (3D-ETF) to train the network. Our experiments and user studies demonstrate that our approach outperforms state-of-the-art methods and exhibits more diverse facial movements. We recommend watching the supplementary video: https://ziqiaopeng.github.io/emotalk | ['Zhaoxin Fan', 'Jun He', 'Hongyan Liu', 'Xiangyu Zhu', 'Hao Xu', 'Zhenbo Song', 'HaoYu Wu', 'Ziqiao Peng'] | 2023-03-20 | null | null | null | null | ['3d-face-animation'] | ['computer-vision'] | [-1.83532998e-01 3.38186949e-01 1.18413754e-01 -6.93698049e-01
-5.24936318e-01 -3.10705632e-01 5.66503823e-01 -1.05885446e+00
1.25946879e-01 3.35120201e-01 5.84537804e-01 1.65557146e-01
4.14068222e-01 -2.98680961e-01 -4.50433910e-01 -6.94181859e-01
1.80057794e-01 1.34829178e-01 -9.20079589e-01 -3.61322075e-01
-4.42435741e-01 5.79007030e-01 -1.86664128e+00 2.28159666e-01
5.98788202e-01 1.23140967e+00 -2.21147850e-01 5.72998643e-01
-9.66792479e-02 6.56040788e-01 -5.95346689e-01 -7.21034110e-01
1.81702361e-01 -7.12248862e-01 -2.45993286e-01 5.16785026e-01
3.13286185e-01 -4.35367048e-01 -4.47619587e-01 1.05043113e+00
9.16296184e-01 8.91332179e-02 5.85530460e-01 -1.62162018e+00
-9.19426918e-01 1.28292382e-01 -4.91462946e-01 -5.25922418e-01
6.38342917e-01 2.30686009e-01 7.77135909e-01 -1.22639191e+00
7.01215804e-01 1.57417905e+00 1.25140518e-01 1.34706700e+00
-1.16021359e+00 -1.03501928e+00 8.12920555e-02 -2.15652045e-02
-1.39153886e+00 -1.07886028e+00 1.36981440e+00 -2.30571002e-01
4.76924837e-01 3.47383082e-01 9.64309096e-01 1.94013143e+00
-2.33017325e-01 8.30306590e-01 1.05751657e+00 -8.79147798e-02
-1.01673841e-01 2.10458323e-01 -6.63109243e-01 7.99786031e-01
-6.42303288e-01 4.34820056e-01 -6.97499216e-01 -1.77240372e-02
5.71384966e-01 -2.64813215e-01 -3.52456152e-01 -1.25449583e-01
-9.75702763e-01 7.35528767e-01 4.81084362e-02 1.12209678e-01
-5.40838540e-01 -8.48276839e-02 4.95268822e-01 1.80579305e-01
7.59932280e-01 5.87740317e-02 -2.17562526e-01 -2.53656566e-01
-6.35034740e-01 2.90479392e-01 6.52010143e-01 1.00947297e+00
5.18388450e-01 6.13647223e-01 -2.65559033e-02 1.16537440e+00
4.78792161e-01 7.72251248e-01 2.10425243e-01 -1.29474676e+00
2.00950038e-02 2.84648508e-01 -2.03100786e-01 -1.09901881e+00
-2.35996932e-01 -8.94506928e-03 -9.62456524e-01 3.95493001e-01
-1.46147549e-01 -4.38301831e-01 -6.89652562e-01 2.22327900e+00
5.42977154e-01 2.83467889e-01 2.51582533e-01 1.27896309e+00
1.21358263e+00 6.51251972e-01 1.02344066e-01 -2.27441609e-01
1.41497648e+00 -7.03502953e-01 -1.17435741e+00 -6.71700612e-02
1.94789022e-01 -6.56230569e-01 1.15400529e+00 2.65250623e-01
-1.29421234e+00 -6.80805922e-01 -6.82603836e-01 -7.35039636e-02
-4.54411916e-02 4.84902442e-01 4.10305351e-01 5.42718649e-01
-9.50299084e-01 1.35966599e-01 -4.61922109e-01 9.78893489e-02
5.37639916e-01 1.37353703e-01 -7.56900966e-01 3.03801805e-01
-1.34894204e+00 8.43822181e-01 -3.76065299e-02 4.40702230e-01
-7.96156645e-01 -5.60372829e-01 -1.26649833e+00 -1.47802517e-01
2.24776655e-01 -7.73678005e-01 1.20027673e+00 -1.62561429e+00
-2.21430850e+00 1.17463756e+00 -2.70992011e-01 3.06281835e-01
4.24363852e-01 6.63003745e-03 -7.11338341e-01 1.67649850e-01
-2.35624552e-01 9.85478282e-01 1.13633144e+00 -1.44395947e+00
-3.87263149e-02 -3.81247163e-01 -2.48298228e-01 2.99066097e-01
-2.65573829e-01 2.71203220e-01 -5.20746529e-01 -7.77402937e-01
-2.67289430e-01 -9.65049744e-01 2.22235993e-01 3.27328622e-01
-4.63104516e-01 -2.66385842e-02 7.34939337e-01 -6.50680125e-01
7.25637794e-01 -2.41371131e+00 6.63251638e-01 -4.99921851e-02
3.52806270e-01 9.59412903e-02 -4.48212326e-01 5.25015295e-02
-5.08529484e-01 6.33418858e-02 3.17158177e-02 -1.02995050e+00
3.02106082e-01 1.23855174e-01 -1.43463194e-01 5.07199287e-01
5.35594881e-01 9.07115042e-01 -6.83303118e-01 -5.22513211e-01
2.75104105e-01 1.13164461e+00 -6.52495503e-01 5.88873088e-01
-1.79691166e-01 8.45562518e-01 -3.37135941e-01 7.45660424e-01
7.02697933e-01 1.95841715e-01 2.37030312e-02 -5.39409041e-01
1.85579360e-01 1.38397599e-02 -7.81251550e-01 1.79300535e+00
-6.93920791e-01 6.04931176e-01 4.65892792e-01 -6.26271248e-01
1.21237147e+00 7.78696477e-01 5.55124938e-01 -6.89714372e-01
7.23141730e-01 1.52768284e-01 -2.88979620e-01 -8.39533210e-01
1.07733309e-01 -6.39848769e-01 -2.68776149e-01 1.13725059e-01
3.84350985e-01 -3.19073707e-01 -4.81637865e-01 -1.55227304e-01
3.86219800e-01 3.37841511e-01 -1.16144612e-01 2.62339473e-01
6.00214720e-01 -8.40012193e-01 5.15783429e-01 -3.33969712e-01
-3.20277572e-01 6.22868478e-01 6.15449309e-01 -1.15104847e-01
-7.61067629e-01 -1.00498891e+00 1.81886241e-01 8.96871805e-01
-3.84994708e-02 -2.48396724e-01 -8.97874534e-01 -6.95112765e-01
-2.36447975e-01 7.52149403e-01 -9.27126050e-01 -3.52012426e-01
-2.45923072e-01 -1.95105791e-01 7.15772867e-01 1.50436357e-01
2.27510765e-01 -1.09120464e+00 -1.70084760e-01 -2.42368102e-01
-4.36202168e-01 -1.32711899e+00 -6.40496731e-01 -3.42835695e-01
-1.54751554e-01 -7.17287183e-01 -7.96303093e-01 -6.46533906e-01
6.88383996e-01 -1.31748348e-01 8.45257580e-01 -2.10599259e-01
-1.22011125e-01 3.78033072e-01 -3.37341517e-01 -3.05248439e-01
-5.68133414e-01 -4.91252214e-01 3.63633424e-01 5.71625769e-01
3.01665157e-01 -8.29468966e-01 -4.19544071e-01 2.08160922e-01
-7.02395260e-01 4.19376671e-01 1.71584263e-01 8.39598060e-01
4.06521201e-01 -3.80997032e-01 7.29580164e-01 -3.82205188e-01
7.72164881e-01 -5.09613156e-01 -1.48919076e-01 -1.58888444e-01
5.09083048e-02 -1.14653587e-01 6.73618674e-01 -8.30630839e-01
-1.33960140e+00 8.37068036e-02 -7.60823727e-01 -1.25542498e+00
-4.44028556e-01 3.39528620e-02 -8.52388561e-01 2.13118732e-01
2.98146874e-01 1.63228393e-01 4.47161704e-01 -1.80066183e-01
7.38420546e-01 9.04511511e-01 6.09018445e-01 -5.22221506e-01
5.70857048e-01 3.64207417e-01 -1.95312053e-01 -8.73112917e-01
-5.49392462e-01 3.28829467e-01 -3.61712098e-01 -7.63053894e-01
1.08280766e+00 -1.18140662e+00 -1.03586447e+00 4.37604547e-01
-1.31655943e+00 -2.33173668e-01 -2.16324791e-01 4.19600248e-01
-9.32459831e-01 9.20844078e-02 -5.59723914e-01 -9.52839553e-01
-3.66736948e-01 -1.29465044e+00 1.39167941e+00 1.14654087e-01
-4.85394299e-01 -6.08478069e-01 -1.56129733e-01 4.50383037e-01
6.32425249e-02 7.41309702e-01 5.02293468e-01 -2.57509381e-01
1.13757536e-01 -1.38863578e-01 1.83936879e-02 5.14299273e-01
3.35832864e-01 2.87103146e-01 -1.19241893e+00 1.36671975e-01
3.56662199e-02 -5.72884083e-01 2.78457433e-01 1.44096091e-01
1.07308817e+00 -4.71153349e-01 2.20732272e-01 8.55016410e-01
6.08532071e-01 7.06322193e-02 5.78447759e-01 -5.23383915e-01
7.51396656e-01 1.16695976e+00 4.78231639e-01 6.36335313e-01
4.53450263e-01 8.03314865e-01 4.43552643e-01 -1.86881498e-01
-2.47549251e-01 -5.40085018e-01 5.26656926e-01 9.76287603e-01
5.53994812e-02 -3.25377703e-01 -3.56058300e-01 1.57905638e-01
-1.39345014e+00 -9.56050694e-01 3.16079259e-01 1.46944916e+00
8.66781592e-01 -3.32104385e-01 2.11969182e-01 -1.37094229e-01
6.20617807e-01 3.73586327e-01 -7.27305830e-01 -4.83478218e-01
-2.35873029e-01 9.01982635e-02 -3.81598026e-01 5.74255466e-01
-7.16589510e-01 1.09489417e+00 4.45940447e+00 7.75020957e-01
-1.47932982e+00 8.65921974e-02 7.19817460e-01 -5.60144722e-01
-6.61513627e-01 -4.92088675e-01 -4.32213247e-01 5.25586963e-01
9.17621136e-01 -3.18733137e-03 5.60376763e-01 7.54325926e-01
6.03387415e-01 6.38790071e-01 -1.02319419e+00 1.42764711e+00
3.75180095e-01 -9.13807392e-01 8.18488747e-02 4.38635089e-02
2.75823653e-01 -4.91256535e-01 2.49473661e-01 2.92801738e-01
-1.97126553e-03 -1.07889986e+00 1.04431224e+00 8.29880178e-01
1.39460015e+00 -8.87950897e-01 1.99777737e-01 2.53916591e-01
-9.40117657e-01 2.10263923e-01 2.08929598e-01 3.30780089e-01
5.16924262e-01 1.16160974e-01 -4.30375934e-01 2.48055413e-01
5.49818933e-01 7.37735033e-01 1.85203597e-01 3.38456035e-02
-4.44605827e-01 1.02612294e-01 -2.87463237e-02 -6.40061647e-02
-1.53393164e-01 -2.71062106e-01 7.67478943e-01 1.01884294e+00
5.43305099e-01 4.28795278e-01 -3.04836243e-01 1.20340669e+00
-3.88178974e-01 2.25757286e-02 -7.47233152e-01 -3.16056639e-01
3.64723355e-01 1.36424255e+00 1.97960138e-02 -8.77149105e-02
-3.00143629e-01 1.37277901e+00 1.94401413e-01 4.40750659e-01
-1.08598614e+00 -1.51040703e-01 1.41733336e+00 -2.73845792e-01
-7.07205012e-03 -7.12427348e-02 1.64642900e-01 -1.36425877e+00
-8.60157907e-02 -1.16596150e+00 -2.91426390e-01 -1.24282169e+00
-1.29912424e+00 1.18570888e+00 -1.88809589e-01 -1.34572732e+00
-4.90830749e-01 -4.77169186e-01 -4.27083910e-01 9.70337689e-01
-1.22320712e+00 -1.34127104e+00 -3.80759358e-01 7.16315925e-01
5.33653319e-01 -1.59240082e-01 9.20833886e-01 4.65136319e-01
-7.64037669e-01 9.47511315e-01 -6.57929778e-01 1.19932167e-01
7.57152855e-01 -6.63013935e-01 2.49847412e-01 3.28016192e-01
-7.83319399e-02 2.35291034e-01 7.89455593e-01 -3.04056883e-01
-1.60174668e+00 -9.94768977e-01 7.42551088e-01 -2.42402896e-01
3.64186049e-01 -7.74970710e-01 -6.93260610e-01 3.73948395e-01
2.87357360e-01 2.52898276e-01 9.75066781e-01 -3.94395143e-01
-5.54435730e-01 -2.42777122e-03 -1.22989869e+00 1.10432935e+00
1.30444098e+00 -8.12214673e-01 -1.72619194e-01 -1.62203074e-01
8.37706983e-01 -4.40431684e-01 -9.23000813e-01 4.67347860e-01
8.14581990e-01 -1.10111701e+00 8.12298357e-01 -6.66014969e-01
7.15486407e-01 -6.81920052e-02 -2.97808826e-01 -1.57017267e+00
8.57851282e-02 -9.20866132e-01 -1.87315196e-01 1.44178379e+00
2.34147340e-01 -4.50521618e-01 7.00679183e-01 6.54039502e-01
-1.20640092e-01 -9.01772559e-01 -8.83021533e-01 -2.72848457e-01
-1.01220578e-01 -5.95174611e-01 8.96610796e-01 1.02495313e+00
7.56895691e-02 4.09984291e-01 -8.54355097e-01 7.56109878e-02
3.74736667e-01 8.50503594e-02 8.96217167e-01 -9.50897098e-01
-4.65156101e-02 -5.89919984e-01 -1.97378978e-01 -9.72693980e-01
1.16639471e+00 -8.74108076e-01 1.12563364e-01 -9.16850150e-01
-1.59975693e-01 -1.16507016e-01 3.37491095e-01 4.28950757e-01
-3.17763351e-02 2.82603323e-01 2.98548341e-01 -3.43933880e-01
-1.17897213e-01 1.40070379e+00 1.70572591e+00 1.11933433e-01
-6.03126884e-02 -2.24594086e-01 -8.71973753e-01 7.23726392e-01
6.48293138e-01 -1.77733734e-01 -4.43432838e-01 -2.83753783e-01
1.44508295e-02 6.24156892e-01 5.06821752e-01 -3.92670095e-01
-2.58684039e-01 -1.04947925e-01 4.33923393e-01 -2.23375097e-01
1.08824885e+00 -1.00128901e+00 2.25859091e-01 -1.14613675e-01
-4.55708146e-01 -1.68185726e-01 2.86883920e-01 2.59466887e-01
-2.62376547e-01 3.23625535e-01 7.96860158e-01 1.13132372e-01
-2.07400396e-01 6.62708461e-01 -3.90519470e-01 1.38446148e-02
8.93845737e-01 -2.37933517e-01 7.74821639e-02 -8.54292989e-01
-1.04984832e+00 9.23624188e-02 3.76653880e-01 8.47728193e-01
1.02187014e+00 -1.84038484e+00 -9.50579941e-01 6.98507190e-01
9.02583301e-02 -3.14256936e-01 7.20245361e-01 6.51403129e-01
5.29214665e-02 -1.92958429e-01 -4.27819282e-01 -3.84053469e-01
-1.61444867e+00 4.34683412e-01 6.07129574e-01 5.27029335e-01
-3.25464189e-01 8.83540273e-01 1.89500213e-01 -7.38354206e-01
2.18204558e-01 1.59426108e-01 -8.54231343e-02 3.59534800e-01
4.49722469e-01 1.20609701e-02 -3.62500817e-01 -1.37726653e+00
-2.02652097e-01 6.06730759e-01 4.91205513e-01 -3.19206268e-01
1.26111913e+00 -3.03592145e-01 4.34263274e-02 3.23089689e-01
1.79913819e+00 2.28400931e-01 -1.48168337e+00 4.43807840e-02
-7.49000132e-01 -5.08739293e-01 4.21812758e-03 -5.69705486e-01
-1.43200874e+00 1.06171298e+00 5.06862700e-01 -1.85507014e-01
1.53961599e+00 1.35854989e-01 9.31309283e-01 -1.66636914e-01
3.95980030e-02 -6.64086521e-01 2.74758816e-01 3.32987338e-01
1.39224279e+00 -1.13210738e+00 -6.44422889e-01 -4.63024139e-01
-1.17295110e+00 1.00450671e+00 6.75767779e-01 6.99479133e-02
7.16576457e-01 4.44605142e-01 3.63685578e-01 -1.46958530e-01
-8.87571275e-01 -1.30988568e-01 3.67909759e-01 6.41519427e-01
3.55558872e-01 1.22907117e-01 1.82806745e-01 9.24520612e-01
-5.51227033e-01 5.23404498e-03 1.44875050e-01 2.31937304e-01
1.96426690e-01 -9.23928916e-01 -1.41535178e-01 -1.10877827e-01
-3.37941885e-01 1.12903610e-01 -6.04997158e-01 5.23116112e-01
2.23704219e-01 9.38075721e-01 9.87601280e-02 -8.17253649e-01
5.59305072e-01 2.56393373e-01 5.88755548e-01 -2.12548107e-01
-2.92149365e-01 2.82917649e-01 3.49955648e-01 -6.65315151e-01
-3.62675726e-01 -4.80971724e-01 -1.21983659e+00 -3.45028728e-01
-1.33660249e-02 4.86784764e-02 6.89939618e-01 5.92898607e-01
7.22690523e-01 4.68879730e-01 1.14068162e+00 -1.40743613e+00
-1.56181678e-01 -8.08803260e-01 -5.78704476e-01 5.91774106e-01
5.91570795e-01 -8.16970944e-01 -6.18704736e-01 1.72714554e-02] | [13.141063690185547, -0.3889904022216797] |
4e264709-ec72-49b5-b03c-b1c0bc776ba4 | digital-twin-framework-for-time-to-failure | 2205.03513 | null | https://arxiv.org/abs/2205.03513v1 | https://arxiv.org/pdf/2205.03513v1.pdf | Digital Twin Framework for Time to Failure Forecasting of Wind Turbine Gearbox: A Concept | Wind turbine is a complex machine with its rotating and non-rotating equipment being sensitive to faults. Due to increased wear and tear, the maintenance aspect of a wind turbine is of critical importance. Unexpected failure of wind turbine components can lead to increased O\&M costs which ultimately reduces effective power capture of a wind farm. Fault detection in wind turbines is often supplemented with SCADA data available from wind farm operators in the form of time-series format with a 10-minute sample interval. Moreover, time-series analysis and data representation has become a powerful tool to get a deeper understating of the dynamic processes in complex machinery like wind turbine. Wind turbine SCADA data is usually available in form of a multivariate time-series with variables like gearbox oil temperature, gearbox bearing temperature, nacelle temperature, rotor speed and active power produced. In this preprint, we discuss the concept of a digital twin for time to failure forecasting of the wind turbine gearbox where a predictive module continuously gets updated with real-time SCADA data and generates meaningful insights for the wind farm operator. | ['Harsh S. Dhiman', 'Sakshi Deshmukh', 'Mili Wadhwani'] | 2022-04-28 | null | null | null | null | ['fault-detection'] | ['miscellaneous'] | [-4.63130713e-01 -6.06357276e-01 4.17628914e-01 2.24061191e-01
4.81206059e-01 -8.09388399e-01 5.26698604e-02 -5.34516387e-02
3.60089272e-01 8.56115103e-01 -4.37897354e-01 -3.88797253e-01
-6.16157770e-01 -8.06003928e-01 8.03733394e-02 -6.68317676e-01
-2.74635911e-01 9.52954441e-02 2.06318125e-01 -4.39730167e-01
1.93836391e-02 9.34016049e-01 -2.10659933e+00 -3.65108997e-01
5.40524423e-01 9.85667169e-01 3.37248802e-01 7.82144010e-01
3.20006460e-01 2.95468032e-01 -1.11520255e+00 3.43199193e-01
2.84072131e-01 -1.60570562e-01 -4.03829694e-01 2.50050247e-01
-4.29254562e-01 -4.57876861e-01 5.28855510e-02 8.09432089e-01
3.86260152e-01 1.53975144e-01 2.81537265e-01 -1.20666003e+00
1.49712592e-01 8.45775306e-02 -2.37158939e-01 6.50599003e-01
-8.43948349e-02 -1.10232644e-02 5.43883264e-01 -7.32196867e-01
2.35465124e-01 6.53849602e-01 6.48170590e-01 -1.89385161e-01
-7.34561384e-01 9.41141695e-02 -5.52595913e-01 2.73773819e-02
-8.72780383e-01 7.65290260e-02 7.58190513e-01 -6.29785538e-01
1.12956381e+00 1.91809669e-01 8.34411204e-01 1.54947728e-01
8.59933436e-01 -5.00297956e-02 8.14849913e-01 -2.45734692e-01
4.51929182e-01 -4.61482048e-01 -1.17255121e-01 6.62257820e-02
6.24959052e-01 2.95127124e-01 -3.69335860e-01 1.55339837e-01
7.47071207e-01 4.30474907e-01 -3.70499581e-01 5.22684194e-02
-6.65726483e-01 4.28316474e-01 -1.33108303e-01 5.08517742e-01
-4.90786165e-01 1.31327594e-02 9.09990489e-01 8.58721972e-01
6.13785923e-01 4.33526039e-01 -9.74197745e-01 -7.13274002e-01
-7.31523097e-01 8.13480318e-02 5.34404695e-01 3.89148891e-01
7.94547498e-02 9.66245055e-01 8.83207798e-01 3.82294089e-01
4.58558723e-02 8.03928196e-01 9.06859875e-01 -4.66560125e-01
-4.69792038e-02 2.21581742e-01 4.69872922e-01 -8.09021711e-01
-4.95154381e-01 -5.99630237e-01 -7.07452536e-01 6.93697751e-01
-1.46741405e-01 -5.12548029e-01 -3.26219589e-01 8.58177423e-01
4.64394182e-01 1.69051141e-02 -1.05894469e-01 8.23725402e-01
-3.77915017e-02 5.54530978e-01 -5.73410630e-01 -6.29123390e-01
1.46359622e+00 -1.44545659e-01 -1.25760543e+00 1.94081008e-01
6.55906022e-01 -1.04019761e+00 4.65737373e-01 8.66684616e-01
-7.55817473e-01 -8.99459779e-01 -1.34605813e+00 5.66117048e-01
-5.93881488e-01 4.88113612e-01 2.42449895e-01 5.23910761e-01
-6.92439258e-01 1.16445112e+00 -1.11196959e+00 -1.12880126e-01
-1.48160324e-01 -2.95021348e-02 -4.99129832e-01 2.88175374e-01
-1.14237726e+00 1.47551799e+00 2.28148252e-01 5.09443581e-01
-6.27900362e-01 -8.46317053e-01 -6.83758855e-01 3.77253965e-02
2.96870381e-01 -4.13634419e-01 1.32900345e+00 -5.63736148e-02
-1.46669197e+00 -2.55245090e-01 -5.64544974e-03 -5.39488137e-01
1.37573451e-01 -5.89448988e-01 -1.05111790e+00 3.64656657e-01
-5.81661835e-02 -7.83759594e-01 1.47778094e+00 -3.59099627e-01
-8.96550894e-01 -3.62424791e-01 -6.60337627e-01 -2.35483915e-01
-1.38586476e-01 1.54438494e-02 1.11819613e+00 -5.94775379e-01
-8.93456191e-02 -6.42022789e-01 4.88987267e-02 -5.06432354e-01
1.45925745e-01 -3.09254646e-01 1.78838265e+00 -8.25394332e-01
1.27848399e+00 -2.29161048e+00 -1.13235034e-01 -7.99924359e-02
-2.31274128e-01 3.64422262e-01 6.91436529e-01 1.11012745e+00
-6.19962871e-01 1.96263269e-02 2.34726369e-01 4.20380890e-01
-3.32338691e-01 6.21739388e-01 -7.34067380e-01 5.03161609e-01
7.14074612e-01 -1.18335495e-02 -8.47881973e-01 3.76299471e-01
7.21031308e-01 2.43290603e-01 2.34108075e-01 1.36122018e-01
1.47922352e-01 5.57158068e-02 -3.06054741e-01 3.72635245e-01
5.10032594e-01 2.98493743e-01 -1.38871685e-01 -2.22367540e-01
-6.98921680e-01 -3.73122841e-02 -1.12233651e+00 1.05284560e+00
-8.09679449e-01 6.22124016e-01 2.40416467e-01 -1.05675280e+00
1.19026017e+00 1.12032628e+00 5.46494007e-01 -1.71956047e-01
-1.03850946e-01 2.62800336e-01 4.25447859e-02 -7.21980870e-01
6.87612951e-01 -1.67061105e-01 3.64368141e-01 3.07195693e-01
1.18774012e-01 -6.43503666e-01 4.79964107e-01 -2.64364213e-01
1.05778408e+00 1.94960266e-01 1.85786024e-01 -4.58536774e-01
2.92404324e-01 8.81704167e-02 6.75336421e-01 -4.53340530e-01
2.31964678e-01 1.65637080e-02 5.52632987e-01 -6.75012529e-01
-1.26274633e+00 -6.67987645e-01 -3.79963785e-01 1.74968347e-01
-3.76794934e-01 -4.99671131e-01 -1.53017551e-01 9.08969622e-03
2.33480856e-01 4.92948860e-01 -3.48476559e-01 -2.62280315e-01
-3.17196637e-01 -7.63765395e-01 -1.65431231e-01 7.25113273e-01
1.95351541e-01 -5.64910471e-01 -1.30925107e+00 6.89283192e-01
5.72496235e-01 -7.75968552e-01 1.01286069e-01 6.33686662e-01
-1.51269674e+00 -1.29728866e+00 1.28383413e-02 -3.48863631e-01
6.88284159e-01 3.55729848e-01 9.18802023e-01 4.05857712e-01
-7.88927615e-01 -3.87309231e-02 -7.38344967e-01 -7.31579065e-01
-3.07823926e-01 -5.45758307e-01 6.09360635e-01 -4.22809362e-01
-3.06396604e-01 -7.15628743e-01 -4.49917287e-01 5.58894515e-01
-8.66109371e-01 -4.46142107e-01 8.86140391e-02 1.25390208e+00
1.49745032e-01 1.57152700e+00 1.05385244e+00 -4.31795597e-01
5.13769686e-01 -6.97068751e-01 -1.05515766e+00 -1.60278067e-01
-8.65674138e-01 -3.22047353e-01 1.34193897e+00 3.98446023e-02
-9.42684591e-01 -2.83782482e-01 4.74786788e-01 -3.61327291e-01
-1.78529516e-01 1.10774660e+00 2.15354532e-01 5.19953728e-01
2.12412730e-01 -2.41653368e-01 4.28407162e-01 -9.69617248e-01
-3.55887935e-02 6.14471853e-01 7.07425892e-01 -3.35771963e-02
1.12011683e+00 3.01526040e-01 4.58540320e-01 -1.16506422e+00
-1.32203400e-01 -5.70223331e-01 -7.46465743e-01 -5.55406094e-01
2.60640502e-01 -8.16409469e-01 -5.84261775e-01 7.64521658e-01
-8.52778435e-01 1.22384280e-02 -5.09115398e-01 6.23852551e-01
-7.72685707e-02 2.44609371e-01 -6.10363066e-01 -9.29309845e-01
-4.22414392e-01 -7.39326954e-01 8.47301245e-01 2.47722641e-01
-1.59671351e-01 -1.21507430e+00 -1.10184886e-01 -1.50211960e-01
4.44212526e-01 5.82278550e-01 6.95367813e-01 -1.52614966e-01
1.55147538e-01 -8.25258851e-01 4.76638824e-01 7.54396737e-01
7.37455308e-01 8.76959145e-01 -6.49627626e-01 -5.03380597e-01
4.97466117e-01 2.49779835e-01 -2.71106303e-01 2.80442774e-01
5.91025054e-01 1.84932992e-01 4.02666554e-02 1.04422681e-01
1.64567840e+00 3.62105757e-01 2.82302111e-01 -5.40266074e-02
1.42879775e-02 3.98884475e-01 1.28704119e+00 8.30770731e-01
-5.02430856e-01 4.31981683e-01 5.53331256e-01 2.03941241e-01
3.74570727e-01 7.07800835e-02 5.10671020e-01 1.18336201e+00
-3.86743009e-01 3.30996402e-02 -5.13642013e-01 8.28650534e-01
-1.34702480e+00 -1.13264430e+00 -5.68570375e-01 2.13306141e+00
2.60943502e-01 1.86948746e-01 -1.53114572e-01 1.27322292e+00
7.09980369e-01 -1.45951018e-01 -2.24297807e-01 -7.75099635e-01
1.29271328e-01 6.35920763e-01 4.17336166e-01 5.24034463e-02
-8.07149172e-01 -1.97129883e-02 5.39195395e+00 5.03291070e-01
-1.36459613e+00 -1.50828376e-01 -1.47022128e-01 2.27686405e-01
3.87559772e-01 7.82845467e-02 -3.07345778e-01 8.67357671e-01
1.32965815e+00 -5.84212005e-01 2.16847748e-01 1.22422802e+00
8.99278641e-01 -6.76073432e-01 -6.35573924e-01 5.64742625e-01
-6.31893873e-01 -9.46174979e-01 -6.65455341e-01 1.37993887e-01
5.66879928e-01 -3.79727110e-02 -4.91468936e-01 -2.88965046e-01
-4.63221222e-02 -6.31821394e-01 4.82443601e-01 5.55720508e-01
7.47521520e-01 -7.33424425e-01 1.34734488e+00 1.61132142e-01
-1.56934536e+00 -2.99400061e-01 -4.21364188e-01 -6.93094730e-01
7.65478015e-01 1.39998543e+00 -1.12402689e+00 1.21930802e+00
1.05473351e+00 9.74392891e-01 -1.19149067e-01 9.40195739e-01
-2.56295979e-01 8.80742431e-01 -5.19236684e-01 2.58920491e-01
-4.51790959e-01 -3.97749364e-01 4.56474453e-01 -2.44998978e-03
8.19371164e-01 -5.69367371e-02 -1.21752024e-01 1.92186132e-01
7.22634733e-01 -4.24648166e-01 -8.55626106e-01 -2.33745798e-01
9.39119697e-01 1.52594554e+00 -6.05314434e-01 -1.12637050e-01
-1.62044957e-01 3.10843587e-01 -6.35350227e-01 5.90824150e-02
-1.98661551e-01 -8.03434193e-01 9.60010469e-01 2.08085746e-01
4.96437728e-01 -7.31664717e-01 -2.79996514e-01 -4.94724303e-01
2.27554679e-01 -2.78465301e-01 1.20543025e-01 -1.23881853e+00
-1.31636691e+00 3.41136187e-01 -4.30230759e-02 -1.79356587e+00
-8.23377252e-01 -9.77368891e-01 -1.20562220e+00 1.20937097e+00
-1.09937882e+00 -4.15685534e-01 -6.80888742e-02 5.58218919e-02
7.36977637e-01 -2.86261231e-01 9.27319229e-01 6.95954710e-02
-8.32003534e-01 -5.49156308e-01 4.08043504e-01 -1.10291123e-01
1.41969472e-01 -1.49923134e+00 3.13914567e-01 1.11331809e+00
-4.84327108e-01 2.73528993e-01 1.05385935e+00 -1.00392008e+00
-1.93309474e+00 -8.38954151e-01 5.46696305e-01 -3.08790565e-01
1.14615738e+00 3.80894661e-01 -8.65722537e-01 1.79486752e-01
2.77827084e-01 3.68149728e-01 4.64910924e-01 -3.26621950e-01
6.95840955e-01 -4.15041983e-01 -1.06418014e+00 4.49768901e-02
1.36132538e-01 -5.66603124e-01 -8.77679527e-01 1.99976444e-01
2.16434225e-01 -3.11374217e-01 -1.63129044e+00 2.94472456e-01
7.21918121e-02 -4.84331340e-01 4.29303139e-01 -4.91309673e-01
2.54453063e-01 -8.51593554e-01 5.39959192e-01 -2.05113506e+00
-3.34021002e-02 -9.87952054e-01 -8.40895697e-02 1.03895772e+00
-2.11958274e-01 -9.03549194e-01 -4.89518195e-02 -6.83142766e-02
-5.10258198e-01 -5.69936216e-01 -1.16919231e+00 -9.26502585e-01
-4.54667062e-01 -1.75801128e-01 7.24804699e-01 9.48962629e-01
4.22544062e-01 1.85212538e-01 -2.16331899e-01 6.86153829e-01
2.41491303e-01 1.42380986e-02 4.45356667e-01 -1.73913491e+00
1.37979746e-01 1.28391474e-01 -6.08315349e-01 1.08312458e-01
-4.21606988e-01 1.01648783e-02 -2.25838304e-01 -1.21726799e+00
-1.13608265e+00 9.97795537e-02 2.32262671e-01 2.15195119e-01
7.73053616e-02 -2.10733369e-01 -2.91600347e-01 9.46897489e-04
7.66097665e-01 3.95417005e-01 9.37117457e-01 4.69821870e-01
2.70526767e-01 3.49081248e-01 5.40898263e-01 3.92084092e-01
8.92919421e-01 -1.94774121e-01 -7.79112518e-01 -2.04499453e-01
5.58224440e-01 6.18720829e-01 3.32285434e-01 -1.05177927e+00
-5.61938621e-03 1.27667505e-02 3.51188928e-01 -1.03547454e+00
-1.60431325e-01 -1.27538443e+00 8.93332779e-01 8.28246176e-01
8.18090081e-01 1.02765203e+00 1.23544991e-01 5.08148909e-01
-7.05666900e-01 -4.17205423e-01 1.57588139e-01 1.80586472e-01
-5.67723691e-01 -1.83037207e-01 -6.65126383e-01 -6.00477993e-01
1.00150502e+00 -2.54199982e-01 -7.52301037e-01 1.33855969e-01
-6.29229665e-01 2.57399529e-01 4.06860113e-01 4.93755192e-01
4.94822145e-01 -8.85810375e-01 -4.95082378e-01 5.59385240e-01
-4.88339812e-01 1.97961897e-01 5.89928985e-01 9.04788017e-01
-9.17829931e-01 3.47463936e-01 -4.32530850e-01 -4.42727357e-01
-1.06378472e+00 4.88025188e-01 3.54162812e-01 1.88350081e-01
-6.63254976e-01 2.63079911e-01 -9.51708198e-01 4.93142396e-01
-8.14541161e-01 -6.88284934e-01 -1.99409649e-01 4.40173388e-01
5.45757174e-01 7.36833155e-01 1.01303673e+00 -2.00642243e-01
-1.38082787e-01 3.37969929e-01 3.82155657e-01 2.01317921e-01
1.54016435e+00 -1.73277721e-01 -4.11875993e-01 7.95908749e-01
6.31188810e-01 -6.54058084e-02 -1.13594902e+00 4.81653184e-01
1.36304670e-03 -6.03188396e-01 4.45053548e-01 -5.85280836e-01
-1.23303008e+00 7.30636060e-01 4.54421073e-01 1.07344937e+00
1.10047293e+00 -7.54961371e-01 5.90887964e-01 1.16562821e-01
6.81788981e-01 -1.49009228e+00 -3.14105272e-01 3.73424560e-01
7.70224273e-01 -6.62693024e-01 2.86132455e-01 -1.16483331e-01
-2.81876236e-01 1.70169675e+00 1.18757240e-01 -1.75990209e-01
1.17770970e+00 7.59769142e-01 6.94146454e-02 -1.98977247e-01
-1.14017224e+00 1.81462973e-01 -3.04554611e-01 6.01647317e-01
4.37946141e-01 2.41308138e-01 -4.66563374e-01 5.13596892e-01
-6.53460145e-01 1.88943908e-01 7.87021637e-01 1.33450949e+00
-4.71740723e-01 -9.98549640e-01 -7.16958880e-01 7.10183680e-01
-5.12351394e-01 3.70916367e-01 6.29360914e-01 6.74059033e-01
1.11197405e-01 1.21208990e+00 3.76485527e-01 -3.52963775e-01
4.92334843e-01 2.62557060e-01 -1.42636195e-01 -5.63382566e-01
-4.83113825e-01 -7.88112730e-02 1.91303834e-01 -6.44601285e-02
-1.91201344e-01 -9.55055296e-01 -1.23759866e+00 -3.40846360e-01
-8.70251060e-01 6.86777115e-01 1.15320373e+00 1.00660253e+00
2.95516551e-01 9.12398219e-01 1.32070148e+00 -7.02988446e-01
-7.94208586e-01 -1.17574191e+00 -1.38587105e+00 2.22741645e-02
3.03973138e-01 -1.11459279e+00 -6.75146699e-01 3.11006576e-01] | [6.5703864097595215, 2.429513931274414] |
1a4279ba-fd25-44b2-b253-2cbb5557e51b | multi-fidelity-hierarchical-neural-processes | 2206.04872 | null | https://arxiv.org/abs/2206.04872v1 | https://arxiv.org/pdf/2206.04872v1.pdf | Multi-fidelity Hierarchical Neural Processes | Science and engineering fields use computer simulation extensively. These simulations are often run at multiple levels of sophistication to balance accuracy and efficiency. Multi-fidelity surrogate modeling reduces the computational cost by fusing different simulation outputs. Cheap data generated from low-fidelity simulators can be combined with limited high-quality data generated by an expensive high-fidelity simulator. Existing methods based on Gaussian processes rely on strong assumptions of the kernel functions and can hardly scale to high-dimensional settings. We propose Multi-fidelity Hierarchical Neural Processes (MF-HNP), a unified neural latent variable model for multi-fidelity surrogate modeling. MF-HNP inherits the flexibility and scalability of Neural Processes. The latent variables transform the correlations among different fidelity levels from observations to latent space. The predictions across fidelities are conditionally independent given the latent states. It helps alleviate the error propagation issue in existing methods. MF-HNP is flexible enough to handle non-nested high dimensional data at different fidelity levels with varying input and output dimensions. We evaluate MF-HNP on epidemiology and climate modeling tasks, achieving competitive performance in terms of accuracy and uncertainty estimation. In contrast to deep Gaussian Processes with only low-dimensional (< 10) tasks, our method shows great promise for speeding up high-dimensional complex simulations (over 7000 for epidemiology modeling and 45000 for climate modeling). | ['Rose Yu', 'Yi-An Ma', 'Alessandro Vespignani', 'Matteo Chinazzi', 'Dongxia Wu'] | 2022-06-10 | null | null | null | null | ['epidemiology'] | ['medical'] | [-3.30052763e-01 -2.80710608e-01 1.97787017e-01 -2.49281868e-01
-9.24763083e-01 -3.40256065e-01 8.14702809e-01 5.41883260e-02
-3.06967854e-01 8.40710521e-01 2.98326761e-01 -5.62931061e-01
-3.16003293e-01 -1.03692257e+00 -7.55674005e-01 -8.38572562e-01
-3.80100489e-01 9.21177745e-01 -1.08289711e-01 3.04187417e-01
-4.60562885e-01 4.20244306e-01 -1.28547239e+00 -1.00359946e-01
1.02902937e+00 4.58845496e-01 6.34255260e-02 9.18736875e-01
1.59844980e-01 3.58468384e-01 -3.49940956e-01 -1.02882341e-01
6.74079433e-02 -3.94747853e-02 -1.15527146e-01 -5.50369799e-01
-9.93921384e-02 -4.53622729e-01 -3.67006302e-01 7.76538610e-01
6.29073858e-01 1.67909831e-01 1.09770846e+00 -1.33702397e+00
-5.60310304e-01 1.55896008e-01 -4.10303980e-01 -6.30868003e-02
-1.19095862e-01 6.55322790e-01 3.96239609e-01 -9.10004556e-01
1.00198708e-01 1.71454608e+00 1.27596557e+00 2.98398703e-01
-1.76374066e+00 -8.00438762e-01 1.35705233e-01 -7.01174855e-01
-1.45482731e+00 -8.31607655e-02 -8.52839127e-02 -1.13551438e+00
1.11325681e+00 -8.66033211e-02 5.10818541e-01 1.66379380e+00
8.65227997e-01 3.02303523e-01 1.13077867e+00 4.78683919e-01
5.77479482e-01 -1.51033670e-01 1.96263030e-01 4.94989812e-01
3.38445038e-01 6.20662570e-01 -3.50957036e-01 -9.03507471e-01
1.20930851e+00 4.56360221e-01 -2.21346617e-02 -5.59501015e-02
-1.25411654e+00 9.77503896e-01 1.07235231e-01 -2.96715170e-01
-6.70451224e-01 7.94525564e-01 2.00698242e-01 -4.22130153e-02
8.00565302e-01 1.95267707e-01 -4.69212949e-01 -3.24576259e-01
-1.29874456e+00 5.17191768e-01 7.88179338e-01 8.78986120e-01
3.74281406e-01 3.27451497e-01 -6.60608470e-01 5.34624636e-01
6.15248024e-01 1.15984809e+00 -7.46588632e-02 -9.35219228e-01
3.96544635e-01 -7.88436085e-02 4.80461627e-01 -8.73927474e-01
-6.83139205e-01 -6.53061271e-01 -1.55098844e+00 1.00369275e-01
3.39848325e-02 -6.47085607e-01 -1.32089758e+00 1.95614541e+00
3.16059381e-01 5.49006522e-01 -7.95939378e-03 4.34937865e-01
4.46749449e-01 1.15279055e+00 5.39283752e-01 6.56975582e-02
1.25275040e+00 -8.11023474e-01 -6.32355988e-01 -5.44333234e-02
2.79008120e-01 -4.58034456e-01 1.00413847e+00 1.01657227e-01
-9.79978144e-01 -5.18608451e-01 -5.94422579e-01 2.31183782e-01
-2.53004313e-01 -8.81246552e-02 7.10802853e-01 6.89367771e-01
-1.22786343e+00 7.59058475e-01 -1.51133728e+00 1.68146312e-01
3.50942343e-01 3.21802288e-01 -1.08131051e-01 7.58037567e-02
-1.48752248e+00 7.30135798e-01 7.70303607e-03 1.46917298e-01
-1.57724643e+00 -1.27611935e+00 -8.10715616e-01 2.99007386e-01
-1.55648381e-01 -1.38528454e+00 1.10660326e+00 -1.03848971e-01
-1.42236865e+00 -1.05454102e-02 -1.15348086e-01 -4.00066376e-01
8.00938487e-01 -3.14389080e-01 -4.69828367e-01 -1.87562793e-01
1.50570840e-01 4.78132486e-01 7.96374142e-01 -9.96594071e-01
-1.52407676e-01 -4.84469682e-02 -5.40021658e-01 5.45764230e-02
1.02655381e-01 -1.15411617e-01 -4.97802913e-01 -6.65529549e-01
-1.69484794e-01 -1.02674878e+00 -8.24790239e-01 -1.44639763e-03
-1.01351626e-01 1.77720532e-01 3.91027033e-01 -6.56599045e-01
1.12029672e+00 -1.83193076e+00 1.16499744e-01 1.19261272e-01
4.90271449e-01 1.41347973e-02 -6.31471351e-02 5.46371520e-01
1.98370010e-01 2.10281357e-01 -5.90181589e-01 -7.43814766e-01
1.83589116e-01 4.02644098e-01 -4.20263708e-01 5.14837921e-01
4.54427987e-01 9.92928803e-01 -1.07369649e+00 -3.09442937e-01
1.21987283e-01 9.70481038e-01 -5.54276288e-01 2.66454667e-01
-9.62857679e-02 5.86702287e-01 -3.82912189e-01 4.32419151e-01
7.47335315e-01 -7.85534561e-01 -2.47986540e-01 3.09696555e-01
1.91983022e-02 7.57363066e-02 -1.23452866e+00 1.26702011e+00
-6.73156381e-01 1.84183717e-01 8.07275996e-02 -2.72758871e-01
6.34691656e-01 6.06088877e-01 4.82958257e-01 -6.81846365e-02
-1.48031414e-01 9.69572961e-02 -2.69855827e-01 -4.75474261e-02
4.78128582e-01 -6.30111933e-01 -4.65881288e-01 4.45316970e-01
5.22261634e-02 -4.18141633e-01 -2.24829733e-01 8.37334618e-03
1.02855229e+00 2.84140468e-01 -4.07520756e-02 -3.42845887e-01
-2.86037534e-01 -1.87760264e-01 6.59114063e-01 1.23501897e+00
-6.36701100e-03 4.54618245e-01 5.39505243e-01 -2.25445017e-01
-1.32838488e+00 -1.73525357e+00 -4.78183389e-01 8.22727203e-01
-3.44645143e-01 -2.91454941e-01 -4.63891745e-01 -1.36915952e-01
3.96758974e-01 5.91607749e-01 -8.02421987e-01 -1.49902150e-01
-1.81903064e-01 -1.49428308e+00 9.08410609e-01 8.11489463e-01
1.17044754e-01 -6.28140867e-01 -2.96701461e-01 5.18129885e-01
1.00522302e-01 -7.61084557e-01 -1.66494057e-01 4.15559351e-01
-1.02475846e+00 -3.33530903e-01 -9.58356440e-01 2.02235337e-02
4.47214276e-01 -1.69108093e-01 1.15074074e+00 -6.31340086e-01
-3.19515169e-03 7.04995915e-02 4.89859581e-01 -4.91924882e-01
-5.56063771e-01 -1.08167514e-01 5.30221760e-01 -4.34167415e-01
3.41901295e-02 -7.23356903e-01 -5.76409757e-01 2.59962708e-01
-9.19310987e-01 1.93635538e-01 4.86085594e-01 1.03743744e+00
6.66188896e-01 7.18777394e-03 3.98609400e-01 -7.41601586e-01
8.76021266e-01 -1.01827931e+00 -7.16548860e-01 -1.96881387e-02
-4.74884212e-01 3.25546473e-01 6.10282362e-01 -5.64163983e-01
-9.89675045e-01 -2.29952589e-01 -7.47903809e-02 -7.07009554e-01
7.40914643e-02 6.29266977e-01 2.12364823e-01 5.82626998e-01
6.47418499e-01 -3.90616432e-02 -5.14876246e-02 -5.27671218e-01
3.97498906e-01 3.15925092e-01 4.59594995e-01 -8.26745868e-01
7.16515720e-01 5.95253289e-01 2.98542887e-01 -4.09388006e-01
-5.24027705e-01 -1.56886578e-01 -2.69275814e-01 4.35023978e-02
9.07875538e-01 -1.52213204e+00 -7.09972858e-01 7.04030275e-01
-9.79868889e-01 -6.66536450e-01 -2.80593514e-01 8.74724567e-01
-6.88589215e-01 3.17780152e-02 -1.07644904e+00 -9.68676507e-01
-3.57443929e-01 -1.30961883e+00 1.38394690e+00 -1.19733587e-01
-2.72220165e-01 -1.31261432e+00 4.78878051e-01 -3.99486840e-01
6.52115524e-01 3.41463149e-01 9.24577951e-01 -1.28249109e-01
-3.94548327e-01 -7.78988078e-02 -2.00711489e-01 1.23145066e-01
-3.01223118e-02 1.40235320e-01 -9.21853244e-01 -3.28114897e-01
1.09113045e-02 -8.39762688e-02 9.32700038e-01 8.75142336e-01
1.05638480e+00 -1.69083178e-01 -4.91960943e-01 8.63534391e-01
1.26345801e+00 -2.08999544e-01 2.99994469e-01 -1.43211499e-01
7.77607918e-01 4.29990530e-01 1.97252259e-01 5.01282573e-01
3.44918668e-01 2.51439154e-01 6.99397326e-02 -5.26851833e-01
2.95242250e-01 -5.26746213e-01 4.51985985e-01 8.85067701e-01
1.07459538e-01 -2.09345818e-01 -1.55068636e+00 4.86048788e-01
-2.00961423e+00 -1.02016199e+00 -4.95462686e-01 2.15849400e+00
8.71230185e-01 2.02983633e-01 -7.53250644e-02 -6.53870523e-01
6.41938925e-01 -8.05384070e-02 -7.98400521e-01 -2.85344869e-01
-4.71181646e-02 1.21860154e-01 7.31798351e-01 7.19804525e-01
-1.18097270e+00 5.20624518e-01 7.64000559e+00 1.03525853e+00
-8.23733807e-01 4.72096115e-01 8.18878293e-01 -3.65919530e-01
-4.82078940e-01 -2.02531829e-01 -1.10636604e+00 6.75347626e-01
1.70902717e+00 -5.85343987e-02 -4.12866063e-02 8.32178175e-01
6.72792256e-01 -4.85713640e-03 -1.04859245e+00 8.10679078e-01
-6.70096874e-01 -1.32713652e+00 1.33442758e-02 3.01082402e-01
1.11832130e+00 4.88111496e-01 4.48677957e-01 6.37169719e-01
1.24983644e+00 -1.56731594e+00 4.57404912e-01 9.88185465e-01
8.81140947e-01 -7.88728595e-01 6.13497853e-01 5.20275950e-01
-1.03103375e+00 1.76776081e-01 -3.85009259e-01 4.62132543e-02
6.65787101e-01 8.72684479e-01 -5.35986781e-01 3.47865134e-01
7.26451218e-01 4.12344038e-01 -3.19456756e-02 8.36733282e-01
8.46925601e-02 7.68857419e-01 -8.25858355e-01 2.71041453e-01
5.57612717e-01 -2.61549592e-01 4.27920401e-01 1.38388073e+00
6.14575148e-01 -3.57838303e-01 1.96081683e-01 1.20391560e+00
3.45760375e-01 -4.58533078e-01 -5.95796585e-01 -2.13007107e-02
4.25861597e-01 7.98517466e-01 -4.60756123e-01 -5.86636126e-01
-1.91623554e-01 7.07446158e-01 3.79596688e-02 6.96910858e-01
-1.24094856e+00 1.64847195e-01 1.09254849e+00 1.67555228e-01
1.92815468e-01 -6.29460931e-01 -3.43538821e-01 -1.04563010e+00
-4.11088645e-01 -6.85606658e-01 2.26986572e-01 -8.18016529e-01
-1.66245699e+00 6.69564366e-01 2.35287547e-01 -1.15279174e+00
-4.52232242e-01 -4.63393927e-01 -2.97021627e-01 1.78007376e+00
-1.30445409e+00 -1.19471192e+00 4.96773981e-02 3.59545618e-01
1.51223078e-01 1.62768915e-01 1.07258725e+00 1.69920996e-01
-4.96184528e-01 1.78083733e-01 8.70653689e-01 -2.79119462e-01
5.56413949e-01 -1.27770913e+00 1.15224040e+00 6.64654016e-01
-5.25797725e-01 9.54571009e-01 7.74377048e-01 -1.13621497e+00
-1.02536619e+00 -1.48900485e+00 4.80666667e-01 -8.32834423e-01
8.87727618e-01 -5.62760115e-01 -1.07257307e+00 4.53276634e-01
-4.21903551e-01 -1.10612862e-01 8.99286628e-01 3.35201561e-01
-3.66021365e-01 4.47545171e-01 -9.01135147e-01 6.24176443e-01
6.86097860e-01 -8.11827123e-01 -2.11153045e-01 3.33050907e-01
9.89808381e-01 -3.27019632e-01 -1.25335073e+00 5.83502650e-01
4.13544267e-01 -3.46846759e-01 1.10640681e+00 -8.81493092e-01
4.76448327e-01 -4.00319189e-01 -9.44116712e-02 -1.39149690e+00
-6.10596359e-01 -5.91675520e-01 -5.28253376e-01 7.37950563e-01
4.33028579e-01 -6.26348794e-01 4.56435591e-01 9.27869916e-01
9.46928263e-02 -8.05583000e-01 -7.83889532e-01 -9.31872129e-01
5.78758955e-01 -7.52329946e-01 8.85404706e-01 7.66598761e-01
-7.17274904e-01 2.40698546e-01 -7.44685352e-01 7.26441741e-01
9.37193155e-01 -1.55038852e-02 6.42902076e-01 -1.32627702e+00
-8.53822827e-01 -4.11359459e-01 7.82688428e-03 -9.41773772e-01
-1.19114503e-01 -4.72613603e-01 1.01622619e-01 -1.66594684e+00
3.09155524e-01 -6.45492375e-01 -1.91185579e-01 2.71893501e-01
-5.98770201e-01 -7.76300877e-02 -1.02462828e-01 4.24021959e-01
-1.20887034e-01 9.35051382e-01 1.05171490e+00 8.26563761e-02
-2.69664556e-01 1.15278639e-01 -3.01737547e-01 7.75496483e-01
5.84524214e-01 -7.62440562e-01 -4.40948278e-01 -4.27221656e-01
2.21033201e-01 4.87253457e-01 6.50243402e-01 -1.03364134e+00
1.47374004e-01 -4.31283295e-01 7.09072292e-01 -8.39258254e-01
5.74679255e-01 -5.08536816e-01 9.59382653e-01 5.81273079e-01
-1.17674373e-01 3.62085938e-01 4.60336566e-01 1.04671252e+00
8.95102620e-02 2.63895631e-01 7.66361535e-01 -1.07859962e-01
1.95419788e-03 7.44183064e-01 -9.26679373e-01 -6.73466772e-02
7.26463675e-01 1.36815995e-01 -2.62906134e-01 -4.96449769e-01
-9.78114665e-01 5.38136959e-01 4.17784810e-01 3.53471756e-01
3.38763446e-01 -1.22671449e+00 -7.92915821e-01 1.06584385e-01
-1.23349212e-01 1.96083650e-01 6.55488610e-01 7.17782974e-01
-4.40976948e-01 3.69850159e-01 -6.31369799e-02 -9.23338771e-01
-6.76332831e-01 5.60263991e-01 5.78838110e-01 -6.87987506e-01
-5.01008868e-01 6.98841929e-01 6.31877899e-01 -7.47611761e-01
-4.01298329e-02 -4.70330566e-01 4.22091961e-01 -9.19439793e-02
4.59841520e-01 4.28628922e-01 -1.46429569e-01 -4.11692888e-01
-1.55962920e-02 2.08649606e-01 2.98276544e-01 -4.77328002e-01
1.47878575e+00 5.67127876e-02 1.52875304e-01 8.72579396e-01
1.05810547e+00 -4.24034745e-01 -1.87015200e+00 -8.81149098e-02
-3.53092074e-01 -1.67610735e-01 2.29875177e-01 -6.16876304e-01
-3.33060265e-01 1.24030328e+00 6.26646817e-01 -1.86278209e-01
5.86534262e-01 -3.19082707e-01 5.25323153e-01 1.69455454e-01
3.99547786e-01 -8.98172736e-01 -1.28559753e-01 6.52365565e-01
7.60373414e-01 -8.88519943e-01 -5.64479642e-02 -2.93776626e-03
-5.99297822e-01 4.98567343e-01 3.32211107e-01 -8.50997865e-03
1.07595575e+00 6.56740427e-01 -1.94323212e-01 -2.89524764e-01
-1.03647387e+00 1.82872772e-01 1.69741318e-01 6.20707273e-01
1.83786497e-01 4.77673441e-01 3.78412575e-01 6.49833441e-01
-6.09928966e-02 5.86187951e-02 2.03281432e-01 6.70757115e-01
-1.93558469e-01 -8.99192274e-01 -6.63892806e-01 7.38513350e-01
-1.92902759e-01 -4.43938345e-01 5.47264636e-01 4.54853475e-01
-1.58755839e-01 6.20577574e-01 3.31680954e-01 -2.15941034e-02
5.52477092e-02 1.02627888e-01 -1.35971038e-02 -5.72259665e-01
-7.51784503e-01 2.14661092e-01 -1.02946810e-01 -5.00266850e-01
3.75812799e-02 -6.45883918e-01 -9.17149842e-01 -7.84194112e-01
9.70850885e-02 1.46434113e-01 7.07089126e-01 5.32819033e-01
6.06444955e-01 7.53493726e-01 2.14080766e-01 -8.89490604e-01
-8.65087032e-01 -9.21510041e-01 -5.31442881e-01 7.82122910e-02
4.85321879e-01 -8.62133265e-01 -2.94775635e-01 -1.01298362e-01] | [6.944258689880371, 3.833388090133667] |
32be9120-d7b5-436b-82e9-97cd40df8766 | question-context-alignment-and-answer-context | 2306.02196 | null | https://arxiv.org/abs/2306.02196v1 | https://arxiv.org/pdf/2306.02196v1.pdf | Question-Context Alignment and Answer-Context Dependencies for Effective Answer Sentence Selection | Answer sentence selection (AS2) in open-domain question answering finds answer for a question by ranking candidate sentences extracted from web documents. Recent work exploits answer context, i.e., sentences around a candidate, by incorporating them as additional input string to the Transformer models to improve the correctness scoring. In this paper, we propose to improve the candidate scoring by explicitly incorporating the dependencies between question-context and answer-context into the final representation of a candidate. Specifically, we use Optimal Transport to compute the question-based dependencies among sentences in the passage where the answer is extracted from. We then represent these dependencies as edges in a graph and use Graph Convolutional Network to derive the representation of a candidate, a node in the graph. Our proposed model achieves significant improvements on popular AS2 benchmarks, i.e., WikiQA and WDRASS, obtaining new state-of-the-art on all benchmarks. | ['Thuy Vu', 'Ankit Chadha', 'Thien Huu Nguyen', 'Toan Nguyen', 'Kishan Kc', 'Minh Van Nguyen'] | 2023-06-03 | null | null | null | null | ['open-domain-question-answering'] | ['natural-language-processing'] | [ 2.10091516e-01 4.17482018e-01 6.71093836e-02 -5.54911315e-01
-1.20880961e+00 -7.44556904e-01 2.54392922e-01 6.97738171e-01
-1.88577965e-01 4.97994632e-01 5.78728855e-01 -4.82212275e-01
-1.08365603e-01 -1.16918480e+00 -7.80459166e-01 1.68750659e-01
3.17109376e-01 4.94432062e-01 1.03569055e+00 -5.44247150e-01
4.99973267e-01 -2.06459820e-01 -1.25366354e+00 9.26524937e-01
1.01428437e+00 1.06544745e+00 1.56345144e-01 8.85477960e-01
-6.64924383e-01 1.32171166e+00 -6.27560973e-01 -7.71691442e-01
-1.36648908e-01 -7.87262321e-01 -1.65959311e+00 -2.19404653e-01
7.02006042e-01 -2.11097017e-01 -3.32484752e-01 7.97960103e-01
2.04693824e-01 1.82928398e-01 2.56584704e-01 -6.84558809e-01
-7.29225159e-01 8.53955746e-01 -1.68996602e-01 5.22524297e-01
8.50006104e-01 3.76110189e-02 1.77636385e+00 -1.03251076e+00
7.95251012e-01 1.19758141e+00 3.22722018e-01 4.16181505e-01
-8.93278301e-01 -1.17801577e-01 2.53961146e-01 7.43838727e-01
-9.80099440e-01 -3.96605432e-01 7.92632580e-01 1.09422795e-01
1.33323443e+00 5.64655125e-01 3.70603323e-01 5.02749383e-01
-5.63918687e-02 1.01677012e+00 7.32148051e-01 -6.13773465e-01
2.60875169e-02 -2.12310806e-01 8.95021021e-01 1.15393579e+00
-2.09776804e-01 -6.97759151e-01 -5.49159229e-01 -3.95071656e-01
-1.76342353e-01 -3.10323566e-01 -3.15164536e-01 1.42022371e-01
-8.01608682e-01 9.71894979e-01 6.78651512e-01 2.95251105e-02
-2.92254120e-01 1.93188101e-01 4.20459986e-01 7.43269205e-01
3.54724586e-01 6.24777257e-01 -5.88047981e-01 1.74855694e-01
-5.86239398e-01 5.04114151e-01 1.01995444e+00 7.48937547e-01
8.80233586e-01 -6.80459201e-01 -1.04486823e+00 8.41765285e-01
2.94596672e-01 1.30238116e-01 3.01551312e-01 -9.36556756e-01
9.60985065e-01 1.27771151e+00 -8.54402855e-02 -9.45708811e-01
-3.85234952e-01 -3.75787377e-01 -9.49677378e-02 -5.39298415e-01
3.28255028e-01 2.86378134e-02 -8.18582773e-01 1.53521872e+00
6.39071822e-01 1.56369656e-01 4.31872271e-02 8.56094182e-01
1.55214667e+00 6.86817944e-01 -5.69092594e-02 1.72151417e-01
1.47833383e+00 -1.48996937e+00 -3.55029702e-01 -3.53778541e-01
8.70601118e-01 -7.06995547e-01 1.28467011e+00 -2.41481900e-01
-9.97657716e-01 -3.80394548e-01 -8.06091130e-01 -5.31889617e-01
-1.16932057e-01 1.00775220e-01 1.40005529e-01 1.35764731e-02
-1.02186620e+00 5.31127989e-01 -3.27499926e-01 -2.86641359e-01
1.23541862e-01 2.25852147e-01 -1.02919772e-01 -1.79799542e-01
-1.53151107e+00 8.39882135e-01 9.59729627e-02 -3.39420766e-01
-4.51779097e-01 -5.80421209e-01 -9.54415441e-01 3.34303588e-01
4.90380973e-01 -1.06710935e+00 1.51897407e+00 -5.77990949e-01
-1.26499319e+00 9.39184308e-01 -6.38092697e-01 -6.50920033e-01
-1.51061863e-01 -1.41804740e-01 -4.11067516e-01 6.28912807e-01
2.39419878e-01 3.52088511e-01 6.60731912e-01 -8.72568011e-01
-6.29953265e-01 -4.10073429e-01 7.44090259e-01 1.56993985e-01
-2.49080762e-01 2.91131943e-01 -7.40367591e-01 -1.42692119e-01
3.26220274e-01 -7.71693587e-01 -1.43856168e-01 -3.49155337e-01
-6.20526791e-01 -9.72001672e-01 7.50046909e-01 -9.47564960e-01
1.71381760e+00 -1.62775648e+00 2.52568796e-02 1.69889018e-01
5.51284730e-01 1.78997159e-01 -4.81402993e-01 5.53091288e-01
2.45960906e-01 1.06813617e-01 -2.18960389e-01 -1.16348729e-01
9.24668685e-02 2.57400796e-03 -4.33321416e-01 -4.24217060e-02
4.05950487e-01 1.18633163e+00 -1.08573902e+00 -6.48226023e-01
-3.80249023e-01 -3.06942016e-01 -6.54070020e-01 3.45784903e-01
-8.30498815e-01 2.01645941e-02 -8.53796005e-01 3.69399697e-01
4.17266488e-01 -6.17115021e-01 -5.64387739e-02 -6.22336902e-02
3.66168767e-01 1.39377129e+00 -7.43933320e-01 1.49650300e+00
-6.38754904e-01 2.91413337e-01 -1.33513168e-01 -6.87705994e-01
1.01411223e+00 7.62057900e-02 -7.38606900e-02 -9.20378804e-01
-2.59950578e-01 4.14175063e-01 1.01811111e-01 -8.42784703e-01
6.38772964e-01 3.58287036e-01 -6.26219958e-02 5.25350749e-01
2.82667994e-01 -1.02729097e-01 6.40193462e-01 7.46712208e-01
1.77814555e+00 8.28802437e-02 4.26978990e-02 -1.21326037e-01
1.00394952e+00 8.85307267e-02 1.63738787e-01 7.82262564e-01
2.05275327e-01 6.71222806e-01 7.96437323e-01 -1.98361754e-01
-5.02107680e-01 -8.79186094e-01 4.10050929e-01 1.06320977e+00
1.58676170e-02 -8.53923202e-01 -7.60292232e-01 -1.35399270e+00
-6.20459393e-02 8.69799972e-01 -5.58863878e-01 -3.69410247e-01
-1.14350295e+00 -3.21231559e-02 3.52004349e-01 4.31974053e-01
3.75928134e-01 -1.12747812e+00 -4.04328793e-01 2.85930485e-01
-7.85802364e-01 -1.14682603e+00 -7.64234483e-01 3.13594155e-02
-7.89428949e-01 -1.41598678e+00 -1.79198295e-01 -1.08339465e+00
4.61751968e-01 1.65961549e-01 1.83785617e+00 7.53366113e-01
3.38978231e-01 4.01033223e-01 -6.06344700e-01 1.12211600e-01
-2.60513097e-01 4.62582469e-01 -7.14358807e-01 1.03522345e-01
5.85556865e-01 -3.17902356e-01 -7.73293614e-01 1.42911077e-01
-6.08659387e-01 -2.53740519e-01 2.90535651e-02 7.59928703e-01
5.78611612e-01 -4.09547508e-01 1.06795943e+00 -1.08115852e+00
1.06931210e+00 -7.24779665e-01 -3.76247585e-01 7.23162591e-01
-3.39565903e-01 4.33355272e-01 9.92022038e-01 2.13899300e-01
-9.25723732e-01 -2.60601223e-01 -4.71261144e-01 1.02203593e-01
3.03956807e-01 8.16310763e-01 8.26506987e-02 7.14182630e-02
7.25968719e-01 1.55571606e-02 -4.36972260e-01 -3.73079538e-01
3.66082311e-01 5.80865324e-01 2.32549518e-01 -5.74154198e-01
7.30175495e-01 4.35361415e-02 -1.15078762e-01 -3.23316723e-01
-1.47830236e+00 -7.91076720e-01 -2.66131908e-01 -1.11371174e-01
7.20028639e-01 -4.31916833e-01 -6.06109142e-01 -1.40539631e-01
-1.51424825e+00 9.72479731e-02 -2.95515925e-01 7.52762705e-02
-1.44012034e-01 3.57617199e-01 -6.48149967e-01 -4.66661930e-01
-8.31627607e-01 -9.24634397e-01 9.76464152e-01 2.63150603e-01
-3.79190385e-01 -7.67411649e-01 2.02707499e-01 7.75668502e-01
2.88909018e-01 -7.21254125e-02 1.26982248e+00 -1.02484608e+00
-7.38874853e-01 -4.11712825e-01 -2.42131874e-01 2.39260927e-01
-9.11712125e-02 -2.27892593e-01 -6.82331622e-01 1.01356365e-01
3.89301069e-02 -5.75164199e-01 1.24653625e+00 -1.89933449e-01
1.05465364e+00 -5.24194717e-01 -2.34774888e-01 9.24855564e-03
1.15759766e+00 -3.21281940e-01 5.05616963e-01 1.86544269e-01
5.12762487e-01 8.12810242e-01 4.59375858e-01 1.10058248e-01
8.73048186e-01 5.05874455e-01 4.87239331e-01 4.06626344e-01
-5.56392372e-01 -5.56334794e-01 2.45971084e-01 1.12241709e+00
5.77595711e-01 -4.60456431e-01 -7.65949428e-01 8.42495978e-01
-1.74898028e+00 -8.10827911e-01 -8.16660166e-01 2.05821705e+00
1.01761258e+00 2.80385971e-01 -2.06169412e-01 -2.52991896e-02
6.36157155e-01 3.30030471e-01 -4.11886483e-01 -3.49600524e-01
-7.73146190e-03 7.62310982e-01 1.80081502e-02 7.77716994e-01
-7.89226234e-01 9.28801477e-01 5.29279804e+00 6.77568972e-01
-5.05175173e-01 1.45421609e-01 4.49748218e-01 1.98586822e-01
-8.98731172e-01 4.86700565e-01 -9.53218341e-01 1.64741203e-01
9.55967963e-01 -1.85328960e-01 2.27190286e-01 4.55294132e-01
-1.02209218e-01 -1.12300850e-01 -1.16983593e+00 1.47902474e-01
4.27517146e-01 -1.58272266e+00 1.10986240e-01 -7.60759890e-01
5.49463034e-01 1.00217961e-01 -2.78955221e-01 6.41099215e-01
2.36909717e-01 -6.37953997e-01 5.42711794e-01 4.22945172e-01
3.48065108e-01 -5.74824810e-01 7.39256144e-01 2.44150043e-01
-1.38354266e+00 -9.16116126e-03 -3.09482455e-01 -1.50789488e-02
6.80901185e-02 4.78267372e-01 -9.00365829e-01 6.09319806e-01
5.62681317e-01 3.45911175e-01 -1.03845334e+00 1.05381477e+00
-8.07120562e-01 1.07561874e+00 -2.83857912e-01 -5.34322023e-01
4.84337509e-01 9.30742025e-02 5.80628693e-01 1.01932013e+00
1.12894550e-01 1.47280991e-01 1.74882948e-01 8.62607300e-01
-7.21422434e-01 3.11690003e-01 -2.68643886e-01 2.20040560e-01
3.87728453e-01 1.23564076e+00 -4.13272083e-01 -3.79734725e-01
-6.39229536e-01 9.67516840e-01 9.03822958e-01 1.74126342e-01
-6.95219457e-01 -7.58191109e-01 9.04292315e-02 1.88860282e-01
4.10379559e-01 1.46557674e-01 -2.78140344e-02 -1.17516351e+00
6.24056816e-01 -9.30218101e-01 9.14979994e-01 -7.92259455e-01
-1.12255132e+00 6.59201086e-01 -4.92860645e-01 -9.54759479e-01
-1.28300130e-01 -2.41758332e-01 -9.99306977e-01 9.53572750e-01
-1.76952922e+00 -1.01591516e+00 -2.02503383e-01 5.13694882e-01
7.04750657e-01 1.40340954e-01 8.18463862e-01 2.69011110e-01
-2.33677521e-01 6.35490239e-01 -3.37682545e-01 2.21040055e-01
4.19962227e-01 -1.38721824e+00 8.51300299e-01 9.39823031e-01
5.43390810e-01 7.01803088e-01 3.13988686e-01 -5.60834467e-01
-1.44681084e+00 -1.10735631e+00 1.71352494e+00 -5.73325038e-01
9.01652515e-01 -1.41417101e-01 -1.02631307e+00 5.35019279e-01
4.86673832e-01 8.27624574e-02 4.35598075e-01 1.15101151e-01
-5.74523449e-01 4.54300381e-02 -9.52686787e-01 5.60225546e-01
8.82570326e-01 -8.31114709e-01 -9.54481721e-01 6.01636112e-01
1.45869744e+00 -5.03435671e-01 -4.20503348e-01 2.51246125e-01
-1.53584406e-02 -7.06223190e-01 7.72337615e-01 -9.86921728e-01
8.01022351e-01 -4.36475515e-01 -3.02884504e-02 -1.18498659e+00
-2.00156912e-01 -4.67052370e-01 -5.93616962e-01 1.22450006e+00
1.09724343e+00 -3.40619355e-01 8.46636236e-01 4.05808330e-01
-4.80015337e-01 -1.19469833e+00 -9.87058163e-01 -2.80877322e-01
-2.25139130e-02 -1.53786525e-01 8.31614494e-01 4.07988995e-01
2.23001212e-01 1.03822744e+00 1.36947751e-01 7.37474263e-02
2.29256123e-01 6.63100302e-01 4.46786612e-01 -1.02203250e+00
-3.77470642e-01 -2.24723697e-01 1.25138581e-01 -1.52131569e+00
3.53522032e-01 -1.44787955e+00 -1.66193359e-02 -2.26963496e+00
8.59315321e-02 -9.13780183e-02 -2.42743477e-01 2.60951787e-01
-7.47819304e-01 -6.21328130e-02 1.72100410e-01 -1.86134756e-01
-1.25741136e+00 6.16009116e-01 1.41295683e+00 -4.69785303e-01
6.55199364e-02 2.25374371e-01 -7.92442262e-01 6.52516007e-01
7.93014348e-01 -7.27611482e-01 -4.07910734e-01 -8.40088248e-01
5.98299980e-01 3.96369964e-01 1.70390740e-01 -6.76102757e-01
4.86472726e-01 1.39432043e-01 -1.57786652e-01 -7.08476067e-01
4.68178838e-02 -5.36680996e-01 -7.61504114e-01 2.96513259e-01
-9.04551864e-01 2.88505048e-01 -2.38001853e-01 5.87094963e-01
-4.61386144e-01 -8.45338047e-01 3.81045550e-01 -2.85684347e-01
-4.16013330e-01 1.52635813e-01 -1.29011460e-02 8.33547771e-01
3.69778514e-01 2.80972093e-01 -6.94592893e-01 -4.84635174e-01
-4.22153443e-01 7.95095623e-01 -1.17336877e-01 4.55833942e-01
1.02183640e+00 -1.24013615e+00 -1.03855753e+00 -3.07025313e-01
3.94082189e-01 -1.55749768e-01 1.37861624e-01 6.76382661e-01
-4.35547620e-01 3.38819742e-01 6.13940716e-01 -1.19325221e-01
-1.40566027e+00 2.42026076e-01 3.70926887e-01 -8.23564410e-01
-4.87483293e-01 1.20044434e+00 -2.79232860e-01 -7.94815242e-01
3.79506312e-02 -5.86603224e-01 -7.17818618e-01 -1.54302105e-01
4.55071330e-01 2.49617919e-01 3.91818792e-01 -2.68953055e-01
-5.23453116e-01 3.40186775e-01 -7.42745772e-02 1.43992841e-01
9.98143196e-01 -1.16331235e-01 -7.45044351e-01 3.72844823e-02
1.44245422e+00 2.66590148e-01 -4.76784319e-01 -5.96565187e-01
6.16841853e-01 -2.15025276e-01 -1.73356250e-01 -8.49062502e-01
-9.29300129e-01 5.59511781e-01 -2.51512993e-02 3.06865990e-01
9.14956927e-01 3.99531215e-01 1.41279650e+00 8.68155837e-01
8.68461579e-02 -9.97106969e-01 4.13233101e-01 1.08962131e+00
1.01412809e+00 -1.10845029e+00 -2.48898461e-01 -4.56703424e-01
-3.69974822e-01 1.22693610e+00 8.97094131e-01 -2.00290695e-01
4.45970118e-01 -2.35326499e-01 -1.85293749e-01 -5.45187950e-01
-1.08785534e+00 -3.59165221e-01 6.59439445e-01 7.28706121e-02
5.82183480e-01 -1.29283965e-01 -6.65638268e-01 7.75869906e-01
-2.23076373e-01 -2.20329031e-01 5.02913415e-01 7.71580994e-01
-8.13973963e-01 -1.06163347e+00 4.29990105e-02 8.82138669e-01
-5.40616393e-01 -5.79329848e-01 -6.70904756e-01 -3.77692357e-02
-2.42652550e-01 1.59019423e+00 -2.66687483e-01 -5.21771073e-01
6.45703077e-01 2.77744085e-01 3.38831872e-01 -9.59925175e-01
-1.22121394e+00 -7.31446981e-01 7.62520432e-01 -4.38566297e-01
-6.30393252e-02 -2.82899141e-01 -1.66264749e+00 6.68279221e-03
-5.95681608e-01 6.76438212e-01 1.62370726e-01 1.04283154e+00
5.97216010e-01 6.16702914e-01 6.99732244e-01 4.55270678e-01
-7.92989433e-01 -9.10340846e-01 1.20546915e-01 3.51610005e-01
3.89873207e-01 -6.47620037e-02 -4.14984673e-01 -2.97685534e-01] | [11.148985862731934, 8.009674072265625] |
2b0b572e-d5b2-4869-a5cd-96a4cb67c69c | modeling-the-sequence-of-brain-volumes-by | 1603.01067 | null | http://arxiv.org/abs/1603.01067v1 | http://arxiv.org/pdf/1603.01067v1.pdf | Modeling the Sequence of Brain Volumes by Local Mesh Models for Brain Decoding | We represent the sequence of fMRI (Functional Magnetic Resonance Imaging)
brain volumes recorded during a cognitive stimulus by a graph which consists of
a set of local meshes. The corresponding cognitive process, encoded in the
brain, is then represented by these meshes each of which is estimated assuming
a linear relationship among the voxel time series in a predefined locality.
First, we define the concept of locality in two neighborhood systems, namely,
the spatial and functional neighborhoods. Then, we construct spatially and
functionally local meshes around each voxel, called seed voxel, by connecting
it either to its spatial or functional p-nearest neighbors. The mesh formed
around a voxel is a directed sub-graph with a star topology, where the
direction of the edges is taken towards the seed voxel at the center of the
mesh. We represent the time series recorded at each seed voxel in terms of
linear combination of the time series of its p-nearest neighbors in the mesh.
The relationships between a seed voxel and its neighbors are represented by the
edge weights of each mesh, and are estimated by solving a linear regression
equation. The estimated mesh edge weights lead to a better representation of
information in the brain for encoding and decoding of the cognitive tasks. We
test our model on a visual object recognition and emotional memory retrieval
experiments using Support Vector Machines that are trained using the mesh edge
weights as features. In the experimental analysis, we observe that the edge
weights of the spatial and functional meshes perform better than the
state-of-the-art brain decoding models. | ['Itir Onal', 'Mete Ozay', 'Fatos T. Yarman Vural', 'Ilke Oztekin', 'Eda Mizrak'] | 2016-03-03 | null | null | null | null | ['brain-decoding', 'brain-decoding'] | ['medical', 'miscellaneous'] | [-3.85010727e-02 6.28959984e-02 2.31809244e-02 -3.97269398e-01
2.00559422e-01 -1.90743655e-01 4.33377177e-01 6.36369467e-01
-3.76916856e-01 3.39299232e-01 2.32046604e-01 3.21120471e-01
-3.28663051e-01 -1.10357726e+00 -7.54641891e-01 -6.45114064e-01
-8.09351563e-01 4.63722557e-01 9.72513855e-02 1.42386481e-02
3.03422868e-01 6.06313229e-01 -1.43820047e+00 5.57944179e-01
5.28988659e-01 1.32409561e+00 2.48757631e-01 1.71069428e-01
-2.18857020e-01 4.31464970e-01 -2.79100806e-01 -6.09839633e-02
-1.38144223e-02 -4.08072323e-01 -7.97809422e-01 -8.14095587e-02
8.68979394e-02 2.02433437e-01 -4.08319861e-01 1.17870235e+00
1.35427848e-01 3.74217808e-01 9.04215932e-01 -9.30657148e-01
-7.96741068e-01 3.77322704e-01 -7.40706861e-01 5.64026058e-01
2.35225916e-01 -3.08547348e-01 7.02493012e-01 -1.00255263e+00
8.39214623e-01 1.10598648e+00 4.04981643e-01 1.69479832e-01
-1.38190019e+00 -4.57776725e-01 7.43261203e-02 3.84412557e-01
-1.64574420e+00 -3.12630385e-02 8.06315064e-01 -7.42685676e-01
1.11502075e+00 1.81187868e-01 1.15317786e+00 7.04264998e-01
1.11530137e+00 5.48787974e-02 1.12128103e+00 -1.86741322e-01
5.24796784e-01 -2.37676173e-01 3.62111747e-01 1.02270257e+00
-1.47039592e-01 -4.27681655e-02 -7.31436253e-01 -4.02011007e-01
1.13422954e+00 1.35031223e-01 -4.57287282e-01 -3.19004774e-01
-1.24984741e+00 7.81769514e-01 9.79557514e-01 8.17687154e-01
-6.44490242e-01 1.74995601e-01 3.03425312e-01 2.22886667e-01
5.42983830e-01 6.41517863e-02 -2.20306255e-02 5.99600971e-01
-1.07357705e+00 -3.20819259e-01 5.83842814e-01 2.69155592e-01
8.76713574e-01 -1.37069657e-01 -1.08077325e-01 7.78153658e-01
2.92074323e-01 4.30974811e-02 7.45883167e-01 -6.82381630e-01
1.02706661e-03 6.57432139e-01 -3.39468837e-01 -1.55651248e+00
-6.31606698e-01 -5.54656148e-01 -1.12157965e+00 1.09318517e-01
1.24557726e-01 3.62796545e-01 -6.13707185e-01 1.77836609e+00
1.54186293e-01 7.33114779e-01 -5.26025057e-01 9.69082296e-01
6.73273504e-01 6.64793313e-01 3.34810048e-01 -6.10088050e-01
1.64689314e+00 -6.94871366e-01 -6.71902001e-01 1.19110150e-02
2.84916639e-01 -1.29445180e-01 7.62561977e-01 8.25981572e-02
-1.20809281e+00 -6.84551299e-01 -9.18873847e-01 6.79656565e-02
-4.60206658e-01 -2.35612690e-01 3.37040246e-01 -7.69574046e-02
-1.24710095e+00 7.54287660e-01 -9.60680723e-01 -1.96578011e-01
7.46158212e-02 2.11450562e-01 -6.47257745e-01 2.54312217e-01
-1.26560903e+00 8.58159602e-01 5.13088219e-02 1.22714750e-01
-7.64613926e-01 -6.73766196e-01 -8.75319362e-01 2.46245235e-01
-3.15173388e-01 -6.39019549e-01 1.64116815e-01 -1.13022101e+00
-8.13607275e-01 1.24820876e+00 -4.49929118e-01 -2.67461896e-01
5.76130599e-02 6.12203419e-01 -5.52565992e-01 3.03225756e-01
9.67133231e-03 8.77257288e-01 1.12864625e+00 -1.16165626e+00
5.01003601e-02 -6.63200438e-01 -2.13128582e-01 8.83391500e-02
-4.39513534e-01 -5.19009866e-02 -3.08897465e-01 -5.98834574e-01
5.74208200e-01 -5.67887962e-01 -1.18491128e-01 2.39019260e-01
-4.27926220e-02 -3.07794005e-01 3.78158480e-01 -8.97569537e-01
1.21191180e+00 -2.34458828e+00 5.83973169e-01 6.97415411e-01
5.64504504e-01 -5.57263494e-01 -1.43981680e-01 1.38287827e-01
-5.83303452e-01 3.83851863e-02 -4.27875340e-01 1.20713733e-01
-3.98757368e-01 2.49071922e-02 -9.73362941e-03 8.66296887e-01
-6.19634464e-02 8.94751906e-01 -1.01821768e+00 -4.99584287e-01
3.69761139e-02 7.88574696e-01 -2.96304405e-01 4.39507738e-02
1.10955499e-01 4.24831748e-01 -3.73799086e-01 1.64610550e-01
5.01894414e-01 -1.88307956e-01 1.99574724e-01 -5.57858586e-01
-1.41771197e-01 -2.01104671e-01 -9.23386395e-01 1.62459016e+00
-2.65327513e-01 3.08854818e-01 8.33411049e-03 -1.35333014e+00
1.15294039e+00 3.37610006e-01 6.57783151e-01 -9.86650944e-01
2.32754782e-01 -2.23159380e-02 8.25492963e-02 -4.67812032e-01
-1.38266906e-01 -7.67981261e-02 1.60932541e-01 4.21383023e-01
9.29842293e-02 2.02069908e-01 1.71768647e-02 4.83520366e-02
1.19461298e+00 -3.05336386e-01 2.40543783e-01 -8.29782188e-01
6.02853656e-01 -5.39842486e-01 3.08677047e-01 4.47742164e-01
-1.86123207e-01 2.48303473e-01 6.02214396e-01 -6.01177156e-01
-9.10088062e-01 -1.34713471e+00 -5.58569133e-01 9.05731916e-01
-2.70478688e-02 -3.72714281e-01 -9.62049425e-01 3.70090492e-02
7.14775175e-02 4.95990366e-01 -1.07977879e+00 -4.19350207e-01
-4.65849191e-01 -4.61596787e-01 1.78414598e-01 3.01025748e-01
4.51556563e-01 -1.36010337e+00 -8.28732669e-01 1.80561930e-01
-1.69971228e-01 -6.56944275e-01 -5.52263677e-01 1.67165577e-01
-1.15496588e+00 -8.22561026e-01 -3.96498799e-01 -1.14296317e+00
1.09734511e+00 -1.25910550e-01 9.93764639e-01 2.86162615e-01
-5.01763105e-01 4.33489740e-01 4.28734645e-02 2.58070469e-01
1.30458415e-01 -4.73386019e-01 9.95624214e-02 4.86101121e-01
3.11558247e-02 -8.75184655e-01 -7.90850997e-01 1.81819782e-01
-9.77596223e-01 6.48011416e-02 6.04334362e-02 7.24290907e-01
1.13178587e+00 9.77505222e-02 3.18161726e-01 -4.67418253e-01
6.51496708e-01 -8.84333193e-01 -2.68125087e-01 3.03095311e-01
-4.00114208e-02 1.96580708e-01 4.64213431e-01 -5.10962129e-01
-4.93240088e-01 -9.17400792e-03 2.81769156e-01 -5.07704079e-01
1.06672838e-01 8.97281528e-01 1.24265559e-01 -2.55362123e-01
6.16884410e-01 2.04972491e-01 -7.72132501e-02 -3.49062145e-01
2.80667871e-01 1.50455639e-01 5.06837904e-01 -6.03179932e-01
9.38847736e-02 4.92717832e-01 2.53416270e-01 -8.85294199e-01
-1.82367146e-01 -3.41214724e-02 -8.79600585e-01 -7.56747067e-01
1.15253818e+00 -4.66074646e-01 -8.97767842e-01 5.34271561e-02
-1.04155898e+00 -9.71734002e-02 -1.83978945e-01 4.32177782e-01
-3.90211493e-01 2.46072114e-01 -7.65760720e-01 -4.57098395e-01
-1.93455949e-01 -9.21774566e-01 8.61511350e-01 -9.65961888e-02
-5.55281103e-01 -1.22100520e+00 4.05571237e-02 -6.91565424e-02
2.67055601e-01 3.57107043e-01 1.67544997e+00 -2.15504929e-01
-2.02877030e-01 -1.61756482e-02 -1.33583993e-01 -4.02860641e-02
-4.30342071e-02 -1.20274633e-01 -5.93390524e-01 -3.13836664e-01
6.56546116e-01 1.20087013e-01 8.94939303e-01 6.05909407e-01
1.56967938e+00 -1.08153276e-01 -5.79719067e-01 4.20930088e-01
1.39716256e+00 2.04103142e-01 4.89418864e-01 -8.26741382e-02
4.16802138e-01 7.41535842e-01 -1.76182985e-01 2.33540252e-01
1.28473029e-01 3.71708244e-01 2.63415396e-01 7.12252781e-03
1.08128199e-02 1.85068008e-02 1.83082581e-01 1.24234962e+00
-3.37433219e-01 2.05274880e-01 -1.12627077e+00 5.61746836e-01
-1.79719818e+00 -8.90605211e-01 -2.16427952e-01 2.44974542e+00
7.40661681e-01 -7.72144198e-02 -3.78521353e-01 -5.41618429e-02
8.33688676e-01 3.22676092e-01 -5.77650011e-01 -4.31446612e-01
-3.92881334e-02 5.58676600e-01 -1.67518798e-02 6.36367619e-01
-5.85507274e-01 4.40047145e-01 6.52658844e+00 5.93379855e-01
-1.10641396e+00 4.28273529e-01 7.93336511e-01 -2.58468628e-01
-3.90938222e-01 -2.14628264e-01 7.24535435e-02 4.52418238e-01
1.14592874e+00 -2.20680818e-01 9.30574656e-01 4.06946480e-01
2.09971786e-01 -2.24702463e-01 -1.33415973e+00 9.48379338e-01
1.30509198e-01 -1.21936214e+00 2.25049332e-01 -4.39788811e-02
4.50036496e-01 -5.24262041e-02 6.13774210e-02 -7.26087615e-02
-3.10680836e-01 -1.28038490e+00 8.23416114e-01 1.03208196e+00
7.22588897e-01 -5.15017271e-01 1.75946310e-01 3.54438782e-01
-1.56701553e+00 1.50018573e-01 -6.04325533e-01 -1.28491491e-01
-1.22654781e-01 8.43414843e-01 -1.22945681e-02 1.21106766e-01
7.13636577e-01 4.77197647e-01 -2.54801005e-01 9.18195546e-01
-1.25985721e-03 3.76544207e-01 -8.37127566e-02 2.04225838e-01
9.31101367e-02 -5.96990347e-01 3.57519597e-01 1.07201695e+00
3.47754806e-01 4.12323862e-01 1.47156373e-01 1.15340304e+00
-2.22399965e-01 3.56733054e-01 -7.20818102e-01 2.23616436e-01
4.54045326e-01 1.31800258e+00 -1.10706341e+00 -3.35824400e-01
-3.95301670e-01 9.63640988e-01 8.55863333e-01 5.03535748e-01
-7.67380714e-01 -3.40838581e-02 5.59736907e-01 4.10614371e-01
-1.35645941e-01 -3.80839139e-01 -3.49769324e-01 -8.09991241e-01
2.08322182e-02 -2.91014642e-01 3.49049658e-01 -1.07434487e+00
-1.43790710e+00 8.31330359e-01 -1.98236834e-02 -6.40780270e-01
2.30588675e-01 -4.57194000e-01 -6.30968571e-01 1.20210016e+00
-8.40023994e-01 -4.82680768e-01 -3.06494772e-01 9.04919446e-01
5.90444319e-02 2.11817518e-01 1.08903563e+00 1.55553833e-01
-2.04220727e-01 -2.16760132e-02 -4.47663069e-02 1.27203897e-01
2.00970322e-01 -6.46212637e-01 8.35908204e-02 1.50377840e-01
1.95948526e-01 7.25841999e-01 2.93153554e-01 -1.03208697e+00
-1.20571268e+00 -9.67506766e-01 1.12851083e+00 -6.03242256e-02
8.82510185e-01 -4.56551850e-01 -1.29576397e+00 7.80809343e-01
2.02957943e-01 3.72094423e-01 5.74724793e-01 -1.73391253e-01
-2.37708539e-01 -5.70704974e-02 -1.32601988e+00 4.18540865e-01
1.30483198e+00 -9.15368199e-01 -7.14854062e-01 5.40055990e-01
3.17593277e-01 -4.56801876e-02 -1.21941602e+00 2.18828082e-01
5.08188307e-01 -7.20707536e-01 1.02336752e+00 -6.08083844e-01
3.74631554e-01 -1.29055545e-01 8.70338604e-02 -1.53289962e+00
-9.06222463e-01 3.17890286e-01 1.03665352e-01 7.36065209e-01
1.22818723e-01 -5.22430956e-01 3.99804771e-01 5.10654867e-01
1.14347912e-01 -9.40419674e-01 -1.38094139e+00 -5.57069898e-01
3.59242074e-02 -2.10484669e-01 3.91366065e-01 1.16829133e+00
3.55414420e-01 5.26995435e-02 1.50244117e-01 -3.00614089e-02
6.89486802e-01 -3.89388353e-02 -5.12598991e-01 -1.38617373e+00
5.46130687e-02 -4.85341460e-01 -7.27155864e-01 -4.01251286e-01
4.96589184e-01 -1.59194636e+00 -2.16894686e-01 -1.56777704e+00
3.75577509e-01 -1.50749549e-01 -6.29343092e-01 4.14135575e-01
2.92106658e-01 2.60611236e-01 -7.61909559e-02 2.75986522e-01
-3.89976427e-02 4.33348805e-01 1.21980441e+00 -2.34957889e-01
-4.92804535e-02 -7.34420419e-01 -1.76797267e-02 8.64885092e-01
5.21519482e-01 -5.77879369e-01 -3.08449686e-01 -4.37868595e-01
2.14397848e-01 5.98043561e-01 4.57559556e-01 -1.08942342e+00
3.91629785e-01 7.28979185e-02 7.82942355e-01 -1.68174013e-01
4.17258650e-01 -9.31617856e-01 7.30853975e-01 6.76139534e-01
-3.35402995e-01 3.15084964e-01 3.10098361e-02 3.94990444e-01
-2.22441122e-01 -1.48186848e-01 7.75609255e-01 -2.51100063e-01
-4.35712874e-01 4.03469115e-01 -5.32344520e-01 -2.59701669e-01
1.05350411e+00 -1.55855209e-01 3.67842726e-02 -1.09345019e-01
-1.22976470e+00 -4.01424803e-02 9.86365527e-02 2.51627982e-01
9.88649070e-01 -1.66617846e+00 -4.76198226e-01 5.86244524e-01
-1.32545695e-01 -6.50120378e-01 3.37035567e-01 1.03264773e+00
-2.65020728e-01 1.76092759e-01 -8.47391725e-01 -4.99816179e-01
-9.67165947e-01 7.03679144e-01 5.84984601e-01 -1.18863285e-01
-6.39492452e-01 7.48042881e-01 4.57461447e-01 -1.38392817e-04
1.39499173e-01 -4.79348332e-01 -5.37349880e-01 2.81980425e-01
4.72585022e-01 3.06461543e-01 1.11273468e-01 -9.73242939e-01
-5.55780947e-01 8.49810362e-01 4.28643078e-01 -1.95595428e-01
1.32903337e+00 1.93649068e-01 -1.00281119e+00 7.95672178e-01
1.40369117e+00 -2.92683542e-01 -7.29357541e-01 -2.59111762e-01
-1.73136234e-01 -3.50608468e-01 2.45255992e-01 -4.80691969e-01
-1.43863773e+00 8.28758240e-01 6.96943998e-01 1.66134685e-01
1.04269016e+00 1.85035214e-01 4.33782011e-01 1.44001395e-02
6.89306200e-01 -7.65475154e-01 -7.90132303e-03 4.95685041e-01
1.22893965e+00 -5.22751927e-01 -3.29142600e-01 -4.02787715e-01
-2.72643000e-01 1.21339548e+00 3.33239675e-01 -5.62685847e-01
1.27674317e+00 1.98174626e-01 -5.37866890e-01 -6.97524786e-01
-5.36665618e-01 3.48224014e-01 7.21824229e-01 2.51676828e-01
5.74565351e-01 3.95978451e-01 -6.63232088e-01 6.92765772e-01
-3.17223780e-02 -9.31633264e-02 -1.44928917e-01 5.12180686e-01
-6.24813914e-01 -7.15617597e-01 -4.55435455e-01 7.34207928e-01
1.21454529e-01 -2.08627254e-01 -2.76702911e-01 1.71024457e-01
3.24581802e-01 5.66824734e-01 8.20174694e-01 -2.91371942e-01
2.03994319e-01 3.83938015e-01 7.73367107e-01 -4.23748344e-01
-7.11917520e-01 -7.62130991e-02 -3.19205433e-01 -8.01311672e-01
-1.72443092e-01 -6.18235290e-01 -1.96130598e+00 -3.39577883e-01
2.77162850e-01 1.89340442e-01 6.16336405e-01 8.48534822e-01
2.30469674e-01 6.03047967e-01 3.45900834e-01 -9.68713641e-01
8.31468329e-02 -7.39262581e-01 -9.80252385e-01 6.11418188e-01
-1.82767585e-01 -8.75333369e-01 -1.70282274e-01 1.32183120e-01] | [12.523965835571289, 3.3765509128570557] |
36871990-dee8-4d0a-aaad-2cf91a8f59b6 | reliable-amortized-variational-inference-with | 2207.11640 | null | https://arxiv.org/abs/2207.11640v3 | https://arxiv.org/pdf/2207.11640v3.pdf | Reliable amortized variational inference with physics-based latent distribution correction | Bayesian inference for high-dimensional inverse problems is computationally costly and requires selecting a suitable prior distribution. Amortized variational inference addresses these challenges via a neural network that approximates the posterior distribution not only for one instance of data, but a distribution of data pertaining to a specific inverse problem. During inference, the neural network -- in our case a conditional normalizing flow -- provides posterior samples at virtually no cost. However, the accuracy of amortized variational inference relies on the availability of high-fidelity training data, which seldom exists in geophysical inverse problems due to the Earth's heterogeneity. In addition, the network is prone to errors if evaluated over out-of-distribution data. As such, we propose to increase the resilience of amortized variational inference in the presence of moderate data distribution shifts. We achieve this via a correction to the latent distribution that improves the posterior distribution approximation for the data at hand. The correction involves relaxing the standard Gaussian assumption on the latent distribution and parameterizing it via a Gaussian distribution with an unknown mean and (diagonal) covariance. These unknowns are then estimated by minimizing the Kullback-Leibler divergence between the corrected and the (physics-based) true posterior distributions. While generic and applicable to other inverse problems, by means of a linearized seismic imaging example, we show that our correction step improves the robustness of amortized variational inference with respect to changes in the number of seismic sources, noise variance, and shifts in the prior distribution. This approach provides a seismic image with limited artifacts and an assessment of its uncertainty at approximately the same cost as five reverse-time migrations. | ['Felix J. Herrmann', 'Rafael Orozco', 'Gabrio Rizzuti', 'Ali Siahkoohi'] | 2022-07-24 | null | null | null | null | ['seismic-imaging'] | ['miscellaneous'] | [ 2.22181916e-01 -2.79664733e-02 3.88541549e-01 -1.65354192e-01
-9.97608900e-01 -4.45172966e-01 5.99316955e-01 -5.75835109e-02
-6.54267371e-01 7.95462012e-01 1.64536029e-01 -2.45104134e-01
-5.90786457e-01 -9.53191876e-01 -9.24437702e-01 -1.00596595e+00
1.08799534e-02 8.71945441e-01 1.39279887e-01 3.26460078e-02
2.29899928e-01 6.66115165e-01 -1.32733917e+00 -4.12418723e-01
7.74182141e-01 8.63873959e-01 2.46282279e-01 5.06363809e-01
9.46585611e-02 3.75820339e-01 -4.83301342e-01 -1.10915974e-01
1.89728394e-01 -3.06731910e-01 -5.53435504e-01 2.20925342e-02
1.72214761e-01 -6.42248809e-01 -1.17181785e-01 1.01481140e+00
4.34554994e-01 4.87936169e-01 1.05676401e+00 -7.86303878e-01
-1.72821581e-01 4.00845438e-01 -5.44360459e-01 1.23250335e-01
-1.83289334e-01 1.01002693e-01 7.12170541e-01 -8.56031775e-01
6.00353181e-01 9.34951365e-01 1.06801808e+00 4.29903977e-02
-1.84293795e+00 -2.56111741e-01 -3.24398614e-02 -1.65510505e-01
-1.56428099e+00 -5.64596295e-01 7.25140393e-01 -6.83569908e-01
6.25506699e-01 -8.01503360e-02 4.55831379e-01 9.85521913e-01
4.21372652e-02 1.36431903e-01 6.12300932e-01 -2.33197063e-01
7.02732503e-01 -7.86152333e-02 -2.03211814e-01 2.34105438e-01
3.73281389e-01 3.16801146e-02 -4.16321784e-01 -3.44731033e-01
8.00550044e-01 -2.83324361e-01 -4.32335287e-01 -3.14973533e-01
-9.53179717e-01 7.84668982e-01 9.82116088e-02 1.39564574e-01
-6.60458684e-01 3.62062931e-01 1.89167455e-01 -1.49859518e-01
5.62872171e-01 4.22352284e-01 -3.54493141e-01 -1.68177247e-01
-1.33411753e+00 4.21627700e-01 9.05695617e-01 4.24599797e-01
1.04111099e+00 3.36351186e-01 2.00803950e-01 7.21567512e-01
5.24549484e-01 9.95518208e-01 -4.88137640e-02 -1.34655988e+00
3.40568393e-01 -1.33363560e-01 3.70864302e-01 -1.17098498e+00
-1.96718663e-01 -6.36052728e-01 -8.85602772e-01 4.22108114e-01
7.55105615e-01 -3.23554128e-01 -7.73453355e-01 2.00623965e+00
3.61483634e-01 1.23386420e-01 9.88453999e-02 8.74223232e-01
7.53706023e-02 7.44665444e-01 -1.58366770e-01 -2.70494968e-01
1.05185354e+00 -1.35409087e-02 -6.73509240e-01 -3.30842435e-01
1.99057102e-01 -6.14349723e-01 9.63448584e-01 2.42957816e-01
-1.06767547e+00 -1.06221266e-01 -9.97612417e-01 1.16196267e-01
8.37076381e-02 -2.54707605e-01 9.84134376e-02 4.49521095e-01
-8.03451419e-01 9.08867657e-01 -1.18601787e+00 1.47330184e-02
7.16029480e-02 2.18014792e-01 -2.04365492e-01 1.26260966e-01
-1.05150020e+00 9.35154915e-01 1.07101157e-01 4.53570038e-01
-7.98504710e-01 -1.14767659e+00 -6.68740034e-01 3.10616851e-01
3.72293413e-01 -5.47260463e-01 1.05252647e+00 -7.24606097e-01
-1.52603698e+00 1.10176742e-01 -8.08906406e-02 -1.70805544e-01
7.64579475e-01 1.61923990e-02 7.65928850e-02 1.87922955e-01
2.47427136e-01 1.31220847e-01 8.55489433e-01 -1.32889128e+00
1.19890608e-02 -3.50230426e-01 -2.19779938e-01 -5.01185916e-02
5.05206808e-02 -6.08640432e-01 -4.01332170e-01 -4.86614138e-01
5.28283775e-01 -8.67287219e-01 -2.63663083e-01 9.85757709e-02
-2.31598645e-01 4.87347782e-01 5.27212739e-01 -8.78423035e-01
7.44355142e-01 -2.19722009e+00 2.21569732e-01 6.88529968e-01
-2.57224515e-02 -2.60224551e-01 9.86165479e-02 3.52586448e-01
7.14296699e-02 1.95809477e-03 -1.00295937e+00 -5.12970626e-01
5.17332815e-02 4.11780536e-01 -2.61947632e-01 7.42263436e-01
1.70295045e-01 2.44685873e-01 -6.87790275e-01 -2.60167420e-01
1.22915462e-01 8.49801779e-01 -7.98855901e-01 1.07241280e-01
-1.63626820e-01 5.93471408e-01 -4.27550524e-01 4.37125154e-02
8.98955464e-01 -2.01697007e-01 2.48508766e-01 -2.99928963e-01
-1.78828686e-01 6.28827289e-02 -1.62197566e+00 1.54846680e+00
-6.19356215e-01 5.39719939e-01 4.55715746e-01 -1.22043967e+00
8.80575955e-01 1.98835090e-01 4.03002262e-01 -4.12855983e-01
-1.28780395e-01 4.67606872e-01 1.04552857e-03 -5.12029529e-01
5.13767362e-01 -6.76679194e-01 1.56964332e-01 5.63152015e-01
1.98119450e-02 -4.74718481e-01 4.57084626e-02 2.02479154e-01
7.30357289e-01 3.82495016e-01 -2.08316982e-01 -5.48416734e-01
1.17448039e-01 -1.15134127e-01 6.75488591e-01 1.04099083e+00
4.03523415e-01 9.06636775e-01 6.98387921e-01 3.97561938e-02
-1.23174095e+00 -1.22760201e+00 -5.43047309e-01 4.35615987e-01
-2.88490087e-01 8.88264105e-02 -6.34514987e-01 2.72744931e-02
6.40043840e-02 9.61685896e-01 -4.52854574e-01 -1.31805360e-01
-5.07985771e-01 -1.09555626e+00 5.16554952e-01 3.01809132e-01
2.66250104e-01 -5.14815032e-01 -7.07291842e-01 4.33739871e-01
-3.93857092e-01 -9.73692536e-01 -7.51126483e-02 3.38483185e-01
-9.41609085e-01 -7.05379307e-01 -6.91897154e-01 2.06063259e-02
6.28719687e-01 -4.19682443e-01 1.01018250e+00 -3.05494726e-01
3.11542042e-02 3.70665133e-01 1.60528749e-01 -5.35784811e-02
-5.38694680e-01 -1.89633012e-01 1.00200571e-01 1.63845018e-01
-1.94524229e-01 -1.05027819e+00 -4.55684066e-01 -2.97245998e-02
-9.36078608e-01 -3.86461288e-01 2.25200340e-01 9.20057833e-01
5.12122393e-01 2.43318483e-01 2.25551620e-01 -5.25209188e-01
4.25325751e-01 -6.61618710e-01 -1.16836441e+00 -7.45934620e-02
-4.84055549e-01 5.64990163e-01 4.80205834e-01 -4.71592635e-01
-1.46204889e+00 -1.44542545e-01 -1.72793597e-01 -3.62237781e-01
2.22762469e-02 9.07235563e-01 7.38587305e-02 6.99937344e-02
7.58171260e-01 3.21729109e-02 2.98576176e-01 -6.53762937e-01
1.57876551e-01 3.55831653e-01 7.39442527e-01 -9.10521030e-01
6.13768518e-01 6.02541029e-01 5.66418052e-01 -1.03110504e+00
-6.60559177e-01 9.90395322e-02 -5.34176111e-01 -1.86085790e-01
7.52280056e-01 -7.19009936e-01 -7.29129374e-01 4.78260875e-01
-1.05796683e+00 -4.27775979e-01 -4.63305801e-01 8.39872837e-01
-5.03258348e-01 4.35837507e-01 -4.96532828e-01 -1.11963737e+00
1.29215896e-01 -1.33658540e+00 8.31525028e-01 -1.07246466e-01
-1.58525556e-01 -1.15506554e+00 8.47499594e-02 -1.10019788e-01
5.72952151e-01 3.67624700e-01 9.32465792e-01 -1.45334542e-01
-5.80250502e-01 -1.34002119e-01 -5.33078946e-02 4.22086418e-01
-1.23433158e-01 1.98260665e-01 -9.43676353e-01 -1.19039062e-02
4.02895898e-01 5.34292720e-02 8.77015591e-01 8.89208615e-01
8.26788068e-01 -2.47629583e-01 8.22818186e-03 6.61994398e-01
1.56868029e+00 -5.94068915e-02 4.15370435e-01 2.12664962e-01
4.51745182e-01 7.66061664e-01 2.55249795e-02 6.52742624e-01
1.18821271e-01 4.72577810e-01 4.63423789e-01 1.89628243e-01
2.02918753e-01 6.38251007e-03 -6.41693026e-02 5.70572913e-01
-2.31356934e-01 -1.53712004e-01 -1.24205565e+00 6.50525212e-01
-1.66618609e+00 -1.00421929e+00 -1.93273902e-01 2.44182801e+00
8.21074307e-01 7.55406320e-02 -3.30081075e-01 7.51116797e-02
5.75177670e-01 1.18787758e-01 -6.02558851e-01 4.07762788e-02
-7.75205791e-02 8.74969214e-02 6.56406343e-01 1.01746380e+00
-7.37287104e-01 2.26330355e-01 6.60500813e+00 5.95322132e-01
-1.09456491e+00 1.34972498e-01 3.17632109e-01 -1.09953202e-01
-6.08722329e-01 1.83126643e-01 -4.40327704e-01 6.11533701e-01
9.66403365e-01 1.01485111e-01 5.84072173e-01 4.95196640e-01
3.91325951e-01 -4.89091694e-01 -8.81305397e-01 6.30235434e-01
-2.10411713e-01 -1.48301065e+00 -1.90672457e-01 2.58677632e-01
4.52034831e-01 1.93524778e-01 -2.71257516e-02 -1.53527826e-01
2.06808895e-01 -7.50263751e-01 1.02736652e+00 8.95219982e-01
6.08458638e-01 -7.12089300e-01 7.68753052e-01 3.97317022e-01
-6.69635952e-01 1.16752170e-01 -2.63342649e-01 -1.04067028e-01
5.64752817e-01 1.13394058e+00 -3.30384970e-01 2.84523189e-01
6.33083701e-01 2.63444692e-01 1.58696994e-01 6.84684396e-01
-1.17468871e-01 7.11508334e-01 -9.41581368e-01 3.27318668e-01
2.08564714e-01 -7.07966983e-01 7.60934591e-01 9.41560984e-01
6.19689465e-01 1.13315195e-01 -8.62950459e-02 1.37022221e+00
2.43475124e-01 -3.59534800e-01 -3.83968771e-01 1.28341317e-01
6.57119393e-01 8.37951839e-01 -6.71797037e-01 -1.74745128e-01
-1.54228017e-01 3.93230498e-01 3.71638536e-02 9.08372819e-01
-4.80720490e-01 -1.80603161e-01 5.30392230e-01 2.59793907e-01
4.44748819e-01 -3.99094701e-01 -3.52247417e-01 -8.61138701e-01
2.07343638e-01 -5.15789151e-01 1.01744607e-01 -7.68663466e-01
-1.11421430e+00 1.37236163e-01 4.22020167e-01 -8.76941204e-01
-5.52053511e-01 -4.58248615e-01 -5.01603544e-01 1.20508003e+00
-1.17267036e+00 -6.17657006e-01 -1.03959724e-01 2.26199478e-01
-4.78075482e-02 1.26225486e-01 6.36459768e-01 2.95541465e-01
-3.74599129e-01 1.54471584e-02 6.52327955e-01 -2.00545132e-01
4.44357246e-01 -1.12594020e+00 1.02465406e-01 9.40826416e-01
-4.38091576e-01 6.65911496e-01 1.34662557e+00 -6.49899423e-01
-1.34154201e+00 -6.32469475e-01 5.04090667e-01 -3.59379165e-02
8.92057836e-01 -1.76589921e-01 -1.28842700e+00 7.65524924e-01
-2.53846437e-01 1.10898733e-01 2.78326899e-01 1.57039747e-01
-5.83877563e-02 -1.49089903e-01 -1.15980399e+00 3.46227378e-01
4.31195289e-01 -6.83101594e-01 -3.81707758e-01 1.40329197e-01
2.95770109e-01 -3.54124039e-01 -1.08619380e+00 3.38088155e-01
6.27088249e-01 -9.33545649e-01 1.00117064e+00 -8.66521448e-02
5.08736074e-01 -4.86140609e-01 -5.01406729e-01 -1.27738404e+00
-1.16118319e-01 -4.60674256e-01 1.62827834e-01 1.18042147e+00
4.65314239e-01 -7.44355738e-01 5.85520208e-01 8.64362776e-01
9.20065027e-03 -3.11030388e-01 -1.10605967e+00 -7.27564633e-01
3.97225022e-01 -7.60429800e-01 4.51533854e-01 7.52440333e-01
-5.94879270e-01 8.13905820e-02 -3.73554349e-01 6.23431921e-01
1.09420621e+00 -1.61799863e-01 5.67503631e-01 -1.27149034e+00
-6.26958251e-01 -2.00049996e-01 -4.33800556e-03 -8.93463016e-01
5.90465032e-02 -5.18130600e-01 4.41368788e-01 -1.23683345e+00
-1.81088820e-02 -5.10170221e-01 1.81391418e-01 7.28237554e-02
1.49156794e-01 1.12740740e-01 -2.09504124e-02 2.19934687e-01
3.60157549e-01 7.23123848e-01 8.32546294e-01 7.34168664e-02
-5.12926951e-02 -1.79659635e-01 -1.31810918e-01 8.59785616e-01
4.83414441e-01 -9.10687089e-01 -3.62200022e-01 -7.30433285e-01
5.48902571e-01 3.75714391e-01 6.13148570e-01 -8.59701574e-01
2.95209467e-01 -2.03860521e-01 3.15903932e-01 -5.91966569e-01
5.85405469e-01 -6.98465586e-01 7.04000950e-01 2.59012967e-01
-9.25528258e-02 -2.62212723e-01 1.34197101e-01 5.23520529e-01
-1.00575328e-01 -6.18097305e-01 9.12536979e-01 -1.13935620e-01
-2.31546685e-01 1.48145789e-02 -5.60482264e-01 6.40932173e-02
2.24788159e-01 -6.48833960e-02 3.83462943e-03 -6.28257513e-01
-8.54931474e-01 -1.31570980e-01 4.53488678e-01 -3.92426699e-01
3.05112422e-01 -9.35934663e-01 -7.72005618e-01 2.74003059e-01
-3.38233888e-01 3.59507740e-01 4.77268308e-01 1.06575441e+00
-5.03127813e-01 -2.36535251e-01 1.11110955e-01 -8.61322403e-01
-4.93857950e-01 6.85446411e-02 7.45432496e-01 1.89581774e-02
-6.76694810e-01 6.68878615e-01 8.24763477e-02 -5.54978371e-01
-1.14694402e-01 -1.22289404e-01 1.83880076e-01 1.42167151e-01
2.35711575e-01 5.15381455e-01 -1.13011226e-02 -6.41202271e-01
-2.92300731e-01 6.00182593e-01 4.35902387e-01 -6.48107350e-01
1.42046571e+00 -2.23055795e-01 -2.39664093e-01 6.84559345e-01
1.16132534e+00 6.80843508e-03 -1.66634667e+00 -3.48126590e-02
-2.37890482e-01 -3.15746933e-01 5.05408823e-01 -4.38605428e-01
-1.03288710e+00 8.23266029e-01 4.45785105e-01 7.07582235e-02
7.50254452e-01 -2.23669052e-01 2.31250539e-01 4.27941084e-01
-1.24965301e-02 -1.13803518e+00 -2.67264873e-01 5.50230742e-01
9.01569247e-01 -8.92523110e-01 3.17988276e-01 1.05696879e-01
-2.36892000e-01 1.07896471e+00 -2.75619570e-02 -1.64496481e-01
9.57184315e-01 5.36861122e-01 3.48798856e-02 -1.61706522e-01
-2.64731705e-01 2.63585627e-01 -2.74211913e-02 3.99595380e-01
8.98986608e-02 -2.06274360e-01 -5.73768606e-03 1.68334633e-01
-2.10371614e-01 -1.21304579e-01 6.60912454e-01 8.54571640e-01
-2.09428012e-01 -6.95927620e-01 -6.51976109e-01 2.90216714e-01
-3.90994698e-01 -9.68024135e-02 4.11644489e-01 8.39266360e-01
-9.92348194e-02 6.51338995e-01 4.85830158e-01 2.70393938e-01
1.88825414e-01 9.10127908e-03 2.31751785e-01 -1.63311183e-01
9.98466164e-02 4.15660679e-01 7.34988526e-02 -3.80761355e-01
-4.82256234e-01 -1.02005792e+00 -1.25387216e+00 -4.59932894e-01
-3.46509218e-01 2.58560598e-01 9.24546301e-01 1.17670095e+00
1.29202172e-01 5.24634957e-01 1.35995030e-01 -1.30010796e+00
-8.56967568e-01 -9.26701307e-01 -8.61613393e-01 2.27066889e-01
5.90382516e-01 -7.75806785e-01 -8.58276665e-01 -7.49206096e-02] | [6.824497699737549, 3.6165761947631836] |
c6f94d1f-3a70-4d48-ae7a-da5f6bcaf1e6 | the-recent-advances-in-automatic-term | 2301.06767 | null | https://arxiv.org/abs/2301.06767v1 | https://arxiv.org/pdf/2301.06767v1.pdf | The Recent Advances in Automatic Term Extraction: A survey | Automatic term extraction (ATE) is a Natural Language Processing (NLP) task that eases the effort of manually identifying terms from domain-specific corpora by providing a list of candidate terms. As units of knowledge in a specific field of expertise, extracted terms are not only beneficial for several terminographical tasks, but also support and improve several complex downstream tasks, e.g., information retrieval, machine translation, topic detection, and sentiment analysis. ATE systems, along with annotated datasets, have been studied and developed widely for decades, but recently we observed a surge in novel neural systems for the task at hand. Despite a large amount of new research on ATE, systematic survey studies covering novel neural approaches are lacking. We present a comprehensive survey of deep learning-based approaches to ATE, with a focus on Transformer-based neural models. The study also offers a comparison between these systems and previous ATE approaches, which were based on feature engineering and non-neural supervised learning algorithms. | ['Senja Pollak', 'Antoine Doucet', 'Jaya Caporusso', 'Matej Martinc', 'Hanh Thi Hong Tran'] | 2023-01-17 | null | null | null | null | ['feature-engineering', 'term-extraction'] | ['methodology', 'natural-language-processing'] | [ 4.44005996e-01 1.20251682e-02 -5.37165582e-01 -3.49235654e-01
-9.85526562e-01 -6.75065815e-01 7.07571208e-01 5.76068401e-01
-4.57879514e-01 6.08588934e-01 7.30515271e-02 -5.23615479e-01
-1.80402294e-01 -7.17086196e-01 -4.67350662e-01 -5.75753987e-01
-4.16207165e-02 5.31937957e-01 -3.34856242e-01 -4.40852731e-01
4.24427450e-01 3.72360498e-01 -1.56393433e+00 3.17922831e-01
8.63198102e-01 1.44082081e+00 5.52620515e-02 1.34371698e-01
-7.25590289e-01 6.33679152e-01 -6.74783826e-01 -4.80204523e-01
-7.81249031e-02 5.65564670e-02 -1.03788698e+00 -4.59826022e-01
2.48904526e-01 7.58355260e-02 -1.46039844e-01 1.13188839e+00
2.85714120e-01 4.21373844e-02 6.73897088e-01 -9.81049299e-01
-8.69376540e-01 7.27980852e-01 -2.85110354e-01 1.48887739e-01
9.48287621e-02 -5.16071200e-01 1.43033719e+00 -1.22943401e+00
7.22142041e-01 9.06266212e-01 8.21014404e-01 4.45859104e-01
-9.24797356e-01 -5.88301301e-01 2.97322452e-01 8.13967139e-02
-1.18279040e+00 -2.55643845e-01 8.15627098e-01 -3.96128178e-01
1.65988922e+00 1.01501271e-01 5.52310169e-01 1.01762474e+00
4.03351992e-01 8.29377770e-01 7.46817291e-01 -8.73476863e-01
1.79417014e-01 3.69066447e-01 5.19520283e-01 5.27291000e-01
2.97715455e-01 -2.05108434e-01 -6.62688792e-01 -3.50425094e-01
4.24287677e-01 -7.53915831e-02 -2.61978388e-01 -1.40009532e-02
-1.03737915e+00 1.08830011e+00 2.20349550e-01 9.43585455e-01
-5.46532452e-01 3.29220258e-02 6.30405128e-01 4.39322650e-01
8.13480198e-01 1.11498940e+00 -1.09065950e+00 -2.06103126e-04
-1.13429880e+00 2.53316402e-01 9.77090716e-01 1.03238130e+00
7.84843743e-01 8.58655497e-02 -2.16038406e-01 1.00740123e+00
1.48718372e-01 2.89771020e-01 9.71769035e-01 -3.84965152e-01
3.89401853e-01 9.30553675e-01 -1.71528429e-01 -1.12238216e+00
-5.46265602e-01 -7.50090003e-01 -7.50803113e-01 -2.95144111e-01
-2.99796183e-02 -5.87017797e-02 -1.09962261e+00 1.64851880e+00
-6.11446798e-02 -5.35045803e-01 -1.23175092e-01 4.29164112e-01
1.15134907e+00 8.73016953e-01 4.24553715e-02 -1.63893834e-01
1.56989574e+00 -1.05119193e+00 -9.05702889e-01 -4.89007354e-01
8.04468274e-01 -7.34964073e-01 6.66118622e-01 5.88851333e-01
-8.92508149e-01 -9.01220888e-02 -8.23799610e-01 -5.81939459e-01
-1.34054613e+00 2.82469541e-01 9.76433277e-01 5.79119086e-01
-1.13959754e+00 5.95879674e-01 -4.90271389e-01 -4.30631280e-01
5.05318820e-01 5.61353624e-01 -2.71627009e-01 1.32248804e-01
-1.68586361e+00 1.15043831e+00 4.66487169e-01 1.59395233e-01
-7.27395833e-01 -6.90587938e-01 -8.70291173e-01 4.12195414e-01
3.11156064e-01 -6.35526776e-01 1.46399081e+00 -1.09220064e+00
-1.11226404e+00 1.11765385e+00 -4.29650962e-01 -8.75139356e-01
-4.37138140e-01 -5.61894417e-01 -4.17541713e-01 -4.53410186e-02
1.53682619e-01 6.12483084e-01 7.41634965e-01 -7.46127844e-01
-7.26325035e-01 -1.99330077e-01 1.55044854e-01 1.95793033e-01
-9.06886995e-01 4.39990222e-01 -2.14375287e-01 -8.05975914e-01
-1.84384614e-01 -6.79557502e-01 -1.52514622e-01 -8.08936805e-02
-3.52374554e-01 -7.05623090e-01 4.32157546e-01 -6.10018909e-01
1.39861798e+00 -1.72000527e+00 -8.86670724e-02 1.19848251e-01
2.86139786e-01 4.01333839e-01 1.27644688e-01 6.74185276e-01
-2.45967478e-01 4.93813038e-01 -2.04901591e-01 -2.33642310e-01
2.15154011e-02 -2.44970500e-01 -7.90765047e-01 3.44635770e-02
4.12555933e-01 1.07522058e+00 -9.09060419e-01 -3.77023220e-01
9.61344130e-03 5.87053716e-01 -8.49300995e-02 -2.30635088e-02
-5.42062044e-01 -1.03007317e-01 -5.70219457e-01 7.70380616e-01
1.80602461e-01 -5.05974293e-01 -1.28021762e-01 -7.67695680e-02
-2.58828104e-01 1.04101408e+00 -5.35358608e-01 1.83617699e+00
-6.77333772e-01 1.08784294e+00 -1.62609398e-01 -1.19387877e+00
9.16242003e-01 6.44545734e-01 5.62840819e-01 -5.58809876e-01
4.88936603e-01 5.42638838e-01 -2.79369831e-01 -3.90663534e-01
7.51893580e-01 -3.18812802e-02 -5.13223596e-02 4.09287989e-01
3.90024662e-01 1.03131972e-01 3.05158645e-01 4.06997465e-02
1.05367720e+00 1.10687062e-01 4.94504184e-01 -4.35177088e-01
4.46432531e-01 5.52035332e-01 2.77997345e-01 6.00818038e-01
3.29657272e-02 4.87848550e-01 1.26140341e-01 -6.97121143e-01
-8.20078194e-01 -4.30818707e-01 -3.54313135e-01 1.35687268e+00
-1.71224520e-01 -5.47027171e-01 -6.78134739e-01 -4.74751025e-01
-1.93441316e-01 6.92667961e-01 -7.23609447e-01 -1.49975896e-01
-5.23020387e-01 -7.93767810e-01 7.38338888e-01 4.26361561e-01
4.27676678e-01 -1.31702948e+00 -4.37574178e-01 3.85803312e-01
-3.81375521e-01 -9.90246534e-01 -2.70557493e-01 8.94330323e-01
-9.12028551e-01 -6.19714797e-01 -7.97463238e-01 -1.37074220e+00
4.84746516e-01 3.20727944e-01 1.45657158e+00 -4.72443439e-02
-2.79767692e-01 8.57195780e-02 -3.21318626e-01 -8.66647065e-01
6.54327625e-04 6.71946466e-01 2.46424060e-02 -3.24433237e-01
1.16685569e+00 -5.25862098e-01 -4.62886453e-01 -2.29045168e-01
-8.31258833e-01 -1.23555303e-01 7.23419785e-01 1.00551033e+00
4.89632964e-01 5.30041493e-02 8.09351146e-01 -7.70922899e-01
9.81413484e-01 -5.10361135e-01 -4.46043611e-01 4.39536750e-01
-1.08661413e+00 2.33980626e-01 4.24607754e-01 -3.00400466e-01
-1.08233547e+00 -3.66666526e-01 -1.31054685e-01 -8.59310254e-02
-2.76428461e-01 1.15480983e+00 2.25175843e-01 -4.21323329e-02
9.15613472e-01 3.90435129e-01 -4.62201566e-01 -5.49835682e-01
1.67385221e-01 8.32348406e-01 1.80628344e-01 -4.72936034e-01
4.79319245e-01 2.40890071e-01 -2.70389855e-01 -7.46247113e-01
-1.10546982e+00 -6.94427550e-01 -4.61811155e-01 1.36492774e-01
5.83999574e-01 -8.38153601e-01 -1.98570490e-01 3.30442071e-01
-1.52546406e+00 2.01217607e-01 -1.14605017e-01 2.77965575e-01
-2.22111158e-02 -9.67544615e-02 -5.98107219e-01 -7.73997426e-01
-1.06347442e+00 -8.31755042e-01 1.29114163e+00 3.26026976e-01
-6.36317790e-01 -1.30400836e+00 2.07257673e-01 1.09519005e-01
7.15979457e-01 -2.45469548e-02 1.20747542e+00 -1.17146778e+00
-1.55957893e-01 -5.35001278e-01 -1.71641678e-01 2.98655987e-01
2.93170452e-01 -3.34548801e-01 -1.40329731e+00 6.21678345e-02
1.72906488e-01 -2.97628433e-01 9.87205327e-01 4.18905139e-01
9.23510194e-01 -2.81816214e-01 -7.00745881e-01 4.45669532e-01
1.24155331e+00 4.44199771e-01 2.38128841e-01 8.18198323e-01
5.63981056e-01 8.12589765e-01 4.41032529e-01 -1.39163032e-01
2.23483264e-01 4.18982685e-01 1.88069910e-01 -2.22280681e-01
8.94996375e-02 4.31060605e-02 2.05944449e-01 6.61424577e-01
8.97535235e-02 -3.13962966e-01 -1.16373003e+00 1.00228846e+00
-1.63111579e+00 -5.99467993e-01 1.08619660e-01 1.75618422e+00
1.23251200e+00 2.10587561e-01 -4.54049289e-01 1.33157030e-01
6.56569183e-01 -3.28042805e-02 -5.44657469e-01 -5.99082947e-01
-5.75376451e-02 5.40140092e-01 2.17703238e-01 -2.70168274e-03
-1.27478874e+00 1.14020276e+00 6.46890926e+00 9.51276958e-01
-1.40187514e+00 -1.36235595e-01 5.65548658e-01 3.34348828e-02
-3.13649833e-01 -1.31877452e-01 -9.74672616e-01 -5.24588414e-02
1.04528093e+00 -4.11072284e-01 4.16546762e-01 1.18541455e+00
-2.68844038e-01 1.17537789e-01 -1.39245725e+00 9.17797804e-01
1.79981872e-01 -1.45110643e+00 3.51098418e-01 -5.39899850e-03
7.61931121e-01 2.87419796e-01 1.71331674e-01 5.14207244e-01
1.03039064e-01 -1.10226429e+00 3.41997623e-01 4.60296609e-02
7.17157245e-01 -6.77250445e-01 8.72343540e-01 1.60566017e-01
-1.05976737e+00 6.65344149e-02 -1.35933504e-01 -1.39670625e-01
-4.43046764e-02 1.11658514e+00 -6.85385048e-01 5.16278625e-01
9.07946825e-01 5.55848718e-01 -2.26383910e-01 9.19752777e-01
-2.87126333e-01 6.73410118e-01 -3.32036734e-01 -3.10581118e-01
5.21035492e-01 2.15936229e-01 5.87303877e-01 1.37644958e+00
1.60311878e-01 -3.75948429e-01 -1.41964614e-01 9.75367665e-01
-3.92141372e-01 4.29363251e-01 -7.17819214e-01 -6.29682243e-01
3.45930517e-01 1.45826733e+00 -7.43345916e-01 -3.77442181e-01
-5.26480436e-01 5.49847305e-01 4.33575273e-01 3.58074576e-01
-4.21856970e-01 -9.00092483e-01 5.34622073e-01 -2.84750074e-01
2.56112844e-01 -1.70466945e-01 -6.08615041e-01 -1.31894600e+00
8.66297334e-02 -9.06039357e-01 2.45363608e-01 -5.05728841e-01
-1.24304664e+00 9.57384706e-01 -1.16549343e-01 -1.13643765e+00
-5.23591697e-01 -8.79738390e-01 -3.96993130e-01 9.54263985e-01
-1.77599168e+00 -1.00727201e+00 2.01374799e-01 1.43980056e-01
7.09096968e-01 -2.10490912e-01 1.30102420e+00 4.44056213e-01
-3.88640463e-01 4.37953025e-01 2.77868122e-01 2.35555217e-01
6.94349349e-01 -1.13504136e+00 7.56613851e-01 5.11873543e-01
4.64124560e-01 1.38559079e+00 5.87853491e-01 -4.61576104e-01
-1.30844796e+00 -9.89140809e-01 1.73791718e+00 -3.07155281e-01
7.51995504e-01 -6.20768845e-01 -8.76080573e-01 6.04875922e-01
5.00417888e-01 -4.63860184e-01 8.63707542e-01 5.64522505e-01
-3.80376309e-01 3.14699523e-02 -8.41550410e-01 4.81546193e-01
7.20559418e-01 -7.55277693e-01 -8.50300372e-01 5.05852222e-01
8.32541585e-01 -2.96758682e-01 -7.49419987e-01 3.72969061e-01
6.37325883e-01 -2.83268929e-01 8.02297771e-01 -6.60522759e-01
5.07363319e-01 4.08425704e-02 2.64929622e-01 -1.34201694e+00
-2.52419233e-01 -6.61826849e-01 -2.99776316e-01 1.29383814e+00
9.31514323e-01 -5.39907575e-01 4.51016188e-01 5.80071807e-01
-2.00868651e-01 -1.06642735e+00 -7.07402587e-01 -7.35720754e-01
3.69547933e-01 -5.57424724e-01 5.22782147e-01 1.08099163e+00
3.79782140e-01 7.67064631e-01 -1.78297162e-02 -2.34444872e-01
5.76237477e-02 3.87013197e-01 1.28141925e-01 -1.50060129e+00
1.38040751e-01 -8.34318519e-01 -5.68946898e-02 -1.10481143e+00
1.06872581e-01 -1.13888597e+00 2.73381889e-01 -1.93194056e+00
1.40887154e-02 -1.45761535e-01 -5.54688275e-01 6.22566998e-01
3.89127396e-02 -1.05652083e-02 -5.24917066e-01 1.50256962e-01
-3.87289137e-01 4.77690637e-01 6.76631749e-01 -5.64373136e-01
-2.22007290e-01 -1.88068271e-01 -1.11861145e+00 8.14166188e-01
7.66181171e-01 -7.67088830e-01 -2.29769960e-01 -9.01176929e-01
9.67304826e-01 -6.12199605e-01 -9.45525691e-02 -6.12469554e-01
3.84611338e-01 1.12778954e-01 2.10832775e-01 -5.61887443e-01
1.66227579e-01 -7.65387475e-01 -5.62920213e-01 9.57493037e-02
-6.43397331e-01 8.37647170e-02 3.74055356e-01 2.07284227e-01
-5.53425789e-01 -4.62040931e-01 1.75247237e-01 -3.39464784e-01
-7.28676438e-01 1.45902589e-01 -6.58608973e-01 -2.07879208e-02
4.47372675e-01 1.06618598e-01 -3.09865355e-01 -3.55358154e-01
-4.23987329e-01 2.16340363e-01 -2.33658895e-01 6.55417800e-01
4.61628973e-01 -1.16078377e+00 -5.81479549e-01 -1.08395092e-01
3.09361041e-01 2.17098981e-01 -3.72095823e-01 5.86803257e-01
-1.65200636e-01 1.16482294e+00 2.45509282e-01 -2.88944036e-01
-9.08056200e-01 7.39344716e-01 2.32135922e-01 -7.83316135e-01
-4.66746658e-01 1.08015990e+00 3.11360091e-01 -3.60070318e-01
5.55828273e-01 -5.33428371e-01 -7.35291600e-01 1.68833330e-01
7.49541759e-01 5.06628491e-02 6.72447205e-01 -4.03026283e-01
-4.69014376e-01 4.77829069e-01 -6.49074316e-01 -2.66105346e-02
1.49038446e+00 5.13133220e-03 -2.47231603e-01 6.36807859e-01
1.26271248e+00 -5.34640312e-01 -6.04073219e-02 -5.03250957e-01
4.24792320e-01 2.21513391e-01 5.40962517e-01 -8.61299396e-01
-9.10446942e-01 1.11830807e+00 3.50087494e-01 6.37620509e-01
1.19039261e+00 -1.09453008e-01 1.04297519e+00 9.80836153e-01
3.94187212e-01 -1.23061109e+00 -1.93730414e-01 9.60264266e-01
9.00154948e-01 -1.13406289e+00 -8.32033753e-02 -3.43500882e-01
-1.29466191e-01 1.29439151e+00 3.68463397e-01 -4.93090786e-02
8.50198567e-01 2.86806434e-01 3.68831456e-02 -6.22643113e-01
-8.14862013e-01 -2.18258306e-01 6.44693732e-01 4.13975775e-01
9.38501477e-01 -5.65009534e-01 -4.41061914e-01 8.63575399e-01
-2.82625407e-01 1.28355682e-01 -1.55847177e-01 1.18932247e+00
-5.10003746e-01 -1.06974435e+00 -1.97798595e-01 6.77506745e-01
-9.03436601e-01 -9.06094432e-01 -7.77393758e-01 4.28174824e-01
7.91690778e-03 1.08659708e+00 -1.37571707e-01 -1.93790928e-01
-6.21901713e-02 5.71986377e-01 1.84723496e-01 -8.85739326e-01
-1.08412027e+00 -4.24981713e-02 4.49633628e-01 -4.12928574e-02
-4.88875955e-01 -4.56138134e-01 -1.13660955e+00 2.04575837e-01
-9.03509021e-01 4.65096533e-01 8.95640910e-01 1.35187852e+00
5.19145072e-01 6.85061872e-01 3.60625982e-01 -6.87390387e-01
-1.87525824e-01 -1.27755642e+00 -2.61855155e-01 -5.72464895e-03
4.25218314e-01 -6.08377993e-01 -3.50852162e-01 2.21845120e-01] | [10.271455764770508, 8.608400344848633] |
32935de6-2b80-46ee-ade5-40e43cb358f0 | machine-learning-based-forward-solver-an | 2111.12148 | null | https://arxiv.org/abs/2111.12148v1 | https://arxiv.org/pdf/2111.12148v1.pdf | Machine Learning Based Forward Solver: An Automatic Framework in gprMax | General full-wave electromagnetic solvers, such as those utilizing the finite-difference time-domain (FDTD) method, are computationally demanding for simulating practical GPR problems. We explore the performance of a near-real-time, forward modeling approach for GPR that is based on a machine learning (ML) architecture. To ease the process, we have developed a framework that is capable of generating these ML-based forward solvers automatically. The framework uses an innovative training method that combines a predictive dimensionality reduction technique and a large data set of modeled GPR responses from our FDTD simulation software, gprMax. The forward solver is parameterized for a specific GPR application, but the framework can be extended in a straightforward manner to different electromagnetic problems. | ['Antonios Giannopoulos', 'Craig Warren', 'Iraklis Giannakis', 'Utsav Akhaury'] | 2021-11-23 | null | null | null | null | ['gpr', 'gpr'] | ['computer-vision', 'miscellaneous'] | [ 3.05885002e-02 -2.33326122e-01 1.00768781e+00 -3.60273421e-01
-1.13331640e+00 -3.25551420e-03 3.02511781e-01 -2.91713178e-01
4.43494171e-02 6.81709111e-01 -6.35106191e-02 -8.88196945e-01
-4.94581938e-01 -1.12243700e+00 -2.20433250e-01 -5.26031494e-01
-4.07958657e-01 8.61981511e-01 -1.28281683e-01 -4.54680920e-01
2.78775662e-01 9.31306422e-01 -1.41593039e+00 5.28183818e-01
7.34101176e-01 1.03724992e+00 3.58812690e-01 6.91082180e-01
-1.80026598e-03 7.73166299e-01 -1.52018398e-01 3.47807437e-01
4.78645377e-02 -1.12948664e-01 -8.04882765e-01 -4.94758338e-01
-4.54344660e-01 -1.24835879e-01 -1.74097121e-01 2.62589574e-01
9.88499999e-01 3.26249748e-01 7.16097355e-01 -7.06687808e-01
-1.08057715e-01 3.44940335e-01 -2.67298758e-01 7.38160908e-02
9.36681569e-01 -1.79153442e-01 2.40885168e-01 -1.15284169e+00
2.32427776e-01 1.01390660e+00 1.56604326e+00 3.58614922e-01
-1.31956530e+00 -6.07810795e-01 -6.09468937e-01 -1.37827948e-01
-1.42820072e+00 -3.87811780e-01 7.25372076e-01 -6.32897794e-01
1.57547462e+00 4.13041979e-01 7.32270598e-01 9.23992813e-01
4.75568891e-01 1.69022232e-01 1.33914518e+00 -7.04860568e-01
7.67800927e-01 8.67813081e-03 1.24183139e-02 3.75010282e-01
2.16206047e-03 5.96636236e-01 -3.67491040e-03 -9.12229955e-01
8.52818370e-01 -6.54906154e-01 -4.44387905e-02 -3.17419171e-01
-3.11819792e-01 1.16548026e+00 -7.80677646e-02 2.91641057e-01
-5.53242922e-01 -1.04413077e-01 4.46022779e-01 3.44686627e-01
7.89283454e-01 7.75117517e-01 -9.69780087e-01 -2.29044765e-01
-1.27171230e+00 8.13107133e-01 1.54527617e+00 6.02895319e-01
6.09434962e-01 3.19206148e-01 2.62703568e-01 8.35704088e-01
7.01743066e-01 2.88881570e-01 2.77024448e-01 -7.30022192e-01
5.52599411e-03 6.32787403e-03 3.90336871e-01 -8.34949315e-01
-7.72630692e-01 -4.63462472e-01 -4.18848515e-01 2.84356564e-01
-1.89488113e-01 -7.49397576e-01 -6.91356301e-01 1.01480019e+00
5.66190064e-01 3.88890862e-01 4.80746031e-01 6.82853580e-01
1.33930159e+00 8.24842334e-01 2.80569762e-01 -7.78974891e-02
8.83741796e-01 -5.57729542e-01 -2.49456331e-01 -2.93686152e-01
1.39233947e+00 -8.18070292e-01 5.00070512e-01 6.03682995e-01
-1.11544466e+00 -3.85653704e-01 -8.81299496e-01 3.47646832e-01
-2.88333148e-01 7.96205848e-02 1.21820784e+00 9.16566014e-01
-1.18500388e+00 9.58856881e-01 -9.11464095e-01 -2.19479918e-01
4.67723645e-02 2.78892279e-01 -8.96719843e-02 -2.72087783e-01
-1.31891263e+00 1.22710764e+00 -1.10338237e-02 -7.12506622e-02
-4.34382498e-01 -1.25486708e+00 -1.18455374e+00 -1.39122158e-01
-5.95295012e-01 -8.44605148e-01 1.49510992e+00 -2.82222033e-01
-1.84472632e+00 6.22620583e-01 2.27335036e-01 -1.74142361e-01
5.47383130e-01 -1.05858579e-01 -7.77513027e-01 -5.58232404e-02
1.84388198e-02 -1.45084172e-01 6.42040908e-01 -1.29945755e+00
-1.45510614e-01 1.20568238e-01 -2.93677300e-01 -8.59807357e-02
2.64022708e-01 2.44246468e-01 1.99753046e-01 -2.56691754e-01
3.66557956e-01 -7.66756177e-01 -9.22511816e-01 -3.69786292e-01
3.46966390e-03 5.54593682e-01 8.24847460e-01 -7.88984001e-01
9.05519128e-01 -1.71097016e+00 -5.07924139e-01 5.18987238e-01
-3.54718924e-01 4.55314405e-02 1.21227905e-01 9.17236328e-01
-6.34134412e-01 -4.65654671e-01 -3.44525963e-01 -3.75240594e-01
-2.82700956e-01 2.41116360e-02 -2.97755003e-01 5.82553864e-01
8.94090608e-02 2.97883213e-01 -6.71975553e-01 -7.80375451e-02
5.08123279e-01 7.53147781e-01 -7.72318482e-01 4.12084222e-01
9.21695530e-02 8.52891386e-01 -8.55756462e-01 5.17014384e-01
1.27158093e+00 -1.61011904e-01 -7.85813630e-02 -1.92415252e-01
-4.90011513e-01 3.63031507e-01 -1.55716848e+00 1.64738762e+00
-1.31623781e+00 5.67989871e-02 3.70754451e-01 -1.46604538e+00
1.25711751e+00 4.13047016e-01 7.87627041e-01 -6.57862306e-01
1.72339216e-01 4.62206602e-01 -4.37934130e-01 -8.51604640e-01
3.11424226e-01 -4.76338953e-01 -1.68891512e-02 6.09673619e-01
-1.72078013e-02 -8.03728700e-01 -1.92898914e-01 -3.04629713e-01
1.40465045e+00 4.63109374e-01 3.52847040e-01 -6.52156293e-01
4.25747186e-01 5.13720155e-01 -6.57301843e-02 5.84073722e-01
8.04262578e-01 4.97906417e-01 -2.81378925e-01 -4.57207978e-01
-7.68463552e-01 -8.79944086e-01 -8.33317995e-01 8.30870628e-01
-4.68494087e-01 -7.32923001e-02 -3.63186866e-01 3.55338514e-01
1.03761382e-01 1.04899287e+00 -2.31473759e-01 -4.80740219e-02
-4.26707029e-01 -1.25660288e+00 6.27566636e-01 5.31493008e-01
8.54683071e-02 -9.98172283e-01 -8.48319590e-01 6.77414477e-01
2.49221534e-01 -1.08713579e+00 7.94003546e-01 5.99235535e-01
-1.17161417e+00 -6.81863368e-01 -4.92168367e-01 -5.96969187e-01
3.12576443e-01 -3.12484652e-01 1.24615562e+00 -1.00761868e-01
-7.07855463e-01 6.70777023e-01 -6.69279516e-01 -3.98083568e-01
-7.65099168e-01 -3.18697959e-01 -1.05183631e-01 -5.89964986e-01
9.41902101e-02 -1.01705062e+00 -1.13283366e-01 -9.80915055e-02
-4.00610209e-01 9.53934193e-02 4.11923468e-01 7.02075064e-01
3.11984718e-01 3.02540779e-01 6.12313867e-01 -8.40999126e-01
8.44456613e-01 -7.99197078e-01 -6.17870390e-01 -2.97221810e-01
-2.08077908e-01 -1.77895635e-01 6.07739568e-01 -4.19167757e-01
-1.44512272e+00 1.63740963e-01 -1.18284822e+00 -1.20748155e-01
-3.89410883e-01 1.22645462e+00 6.60489202e-02 -9.57351983e-01
8.24219704e-01 8.25905204e-02 -2.31528819e-01 -9.46951032e-01
5.52463625e-03 6.61905110e-01 1.58308759e-01 -8.46531093e-01
7.05719292e-01 3.72576684e-01 2.42812276e-01 -1.16701889e+00
-2.48448774e-01 -4.25679147e-01 -1.60318717e-01 4.12946343e-02
4.56385268e-03 -9.55098689e-01 -3.45036805e-01 5.68553448e-01
-1.06645429e+00 -6.32186949e-01 -1.73139736e-01 8.98519456e-01
-7.21899450e-01 1.56812280e-01 -6.87970281e-01 -9.72718537e-01
-5.28139293e-01 -7.39564002e-01 1.01406848e+00 -7.82452673e-02
-5.59727490e-01 -1.40750206e+00 5.83545327e-01 -3.90295893e-01
9.32416320e-01 4.45147365e-01 8.62947285e-01 -6.76153958e-01
2.21506953e-01 -3.42970818e-01 -9.90388170e-02 2.72072628e-02
5.60031179e-03 -6.02295510e-02 -1.13591397e+00 -4.02882427e-01
6.76390231e-01 -3.22962403e-01 4.14294869e-01 7.59878576e-01
9.79013920e-01 1.43954449e-03 -9.02455389e-01 1.06325316e+00
1.87112010e+00 7.85312727e-02 6.96911097e-01 3.58631313e-01
3.49081039e-01 4.08319771e-01 6.45966649e-01 1.10416508e+00
1.21943459e-01 3.51750076e-01 2.01382302e-02 -6.02762640e-01
4.90247905e-01 1.89752474e-01 -1.62517160e-01 6.51127517e-01
-8.41591954e-02 6.83774799e-02 -1.22862494e+00 2.93961883e-01
-1.24129725e+00 -7.69603193e-01 -4.52316284e-01 1.68431425e+00
5.44442713e-01 -1.91934079e-01 -4.33273435e-01 4.42934543e-01
2.22218394e-01 -4.25752461e-01 -8.33662525e-02 -9.51381505e-01
4.22352463e-01 1.08487952e+00 2.60213286e-01 2.98236847e-01
-8.96750271e-01 4.04310197e-01 7.79753542e+00 6.10998750e-01
-1.33585238e+00 5.43681048e-02 2.06977889e-01 4.63514149e-01
-2.04402640e-01 -4.14802460e-03 -5.81144392e-01 1.78623572e-02
1.38665056e+00 -1.65363342e-01 1.97423771e-01 1.28148103e+00
5.90018094e-01 -2.33430475e-01 -1.08768427e+00 8.91832530e-01
-3.47423464e-01 -1.57607317e+00 -4.18871373e-01 -2.92479843e-01
3.61201733e-01 2.03505129e-01 -1.84046581e-01 6.20907128e-01
5.11476934e-01 -1.14757073e+00 3.02705258e-01 4.08391684e-01
9.39693093e-01 -6.45058334e-01 4.54320878e-01 5.24924397e-01
-1.25755310e+00 -1.61783710e-01 -2.82845587e-01 -5.89602649e-01
5.77669144e-01 1.02500606e+00 -7.40164459e-01 8.21827114e-01
9.32702482e-01 4.25316036e-01 2.67454702e-02 1.22648168e+00
5.37596107e-01 4.54606891e-01 -5.21487117e-01 4.44888175e-01
1.09166667e-01 -6.12104349e-02 3.32993180e-01 1.53046203e+00
1.03804171e+00 4.42521930e-01 8.76409709e-02 1.00228322e+00
9.22022581e-01 1.21080950e-01 -7.44423211e-01 3.39630991e-01
3.63490105e-01 1.42403805e+00 -3.69705200e-01 1.60372779e-01
-3.79544050e-01 4.43006277e-01 -7.77583420e-02 2.36166075e-01
-9.04455960e-01 -2.99580663e-01 5.31381369e-01 3.73356789e-01
4.34177935e-01 -2.01608330e-01 -3.14806104e-01 -6.11620247e-01
-4.83386785e-01 -7.49025226e-01 2.33446300e-01 -1.14705229e+00
-1.71849561e+00 8.30788970e-01 3.36960167e-01 -1.13940871e+00
-6.00842416e-01 -7.51330256e-01 -9.83215570e-01 1.38558853e+00
-1.56824875e+00 -1.10224664e+00 -3.69267881e-01 7.40315855e-01
-5.39929010e-02 -1.28485700e-02 1.52077758e+00 5.58329821e-01
1.22487679e-01 1.74007297e-01 3.10565472e-01 -2.34379709e-01
1.04489192e-01 -6.27854586e-01 6.97616100e-01 1.47773162e-01
-6.47928953e-01 3.78754318e-01 1.28124428e+00 -7.07328618e-01
-1.53311026e+00 -1.15186894e+00 6.21356368e-01 7.07887262e-02
6.92258120e-01 -1.09526388e-01 -1.05448914e+00 4.88636911e-01
-4.46786970e-01 2.10620970e-01 7.38805354e-01 1.60691172e-01
3.10723841e-01 4.19642985e-01 -1.70916998e+00 1.14039972e-01
7.14609563e-01 -3.25559378e-01 -6.09325290e-01 4.21190351e-01
2.36021325e-01 -9.31269169e-01 -1.15569460e+00 7.24682391e-01
2.22547770e-01 -8.31688464e-01 1.29338682e+00 -2.07513735e-01
2.83488810e-01 2.48351365e-01 -1.26698330e-01 -1.50581574e+00
-4.24666673e-01 -5.26073039e-01 -1.71162128e-01 8.73262167e-01
2.95759022e-01 -9.81300890e-01 6.48005366e-01 7.08999336e-01
-3.54633480e-01 -9.07216132e-01 -1.05312133e+00 -6.89170420e-01
3.19117635e-01 -9.49232757e-01 5.15337884e-01 9.74859774e-01
2.19956547e-01 -1.46759912e-01 -6.76571429e-02 5.14915705e-01
4.16269839e-01 2.11736500e-01 4.00251269e-01 -1.64593041e+00
-6.74963474e-01 2.07048178e-01 -1.21962428e-01 -7.21911609e-01
1.30335644e-01 -6.07549608e-01 1.43877685e-01 -1.71066499e+00
-3.85580778e-01 -1.24150419e+00 2.28303596e-01 1.87890023e-01
5.17215550e-01 -2.53559556e-02 -7.77959406e-01 -1.54124632e-01
3.67063850e-01 2.60848939e-01 8.19046497e-01 4.50807035e-01
-5.00664055e-01 1.61414415e-01 -3.78788322e-01 8.91961515e-01
8.42092931e-01 -7.02225983e-01 -2.93477893e-01 -2.79105604e-01
9.38640907e-02 5.48314869e-01 4.62889105e-01 -1.46316969e+00
1.84237659e-01 7.21431822e-02 5.97544134e-01 -4.74143982e-01
6.59795344e-01 -5.40925980e-01 6.34638309e-01 4.19747531e-01
1.94268674e-01 -1.00972736e-02 9.21301723e-01 -4.86818440e-02
2.37036347e-01 -2.68834531e-01 1.07269931e+00 -5.33488274e-01
-5.75925648e-01 1.59824528e-02 -8.40839863e-01 -6.72940165e-02
7.95167923e-01 -4.29507107e-01 1.52315438e-01 -1.79170400e-01
-7.92674363e-01 -4.05596524e-01 3.04529369e-01 -2.92510033e-01
6.85992956e-01 -9.72245812e-01 -9.41299200e-01 5.92999995e-01
-2.26501748e-01 -3.14717919e-01 6.93572581e-01 8.20552349e-01
-7.02437639e-01 3.33065152e-01 -3.09662193e-01 -3.52165580e-01
-9.33852911e-01 3.46032947e-01 8.29506814e-01 -6.07942343e-01
-1.11374462e+00 9.36937630e-01 -2.21156552e-01 -8.75805140e-01
-7.08882868e-01 -8.12288970e-02 -3.58780891e-01 -4.08829153e-01
5.01961470e-01 4.49708223e-01 7.90919900e-01 -6.00965202e-01
-3.91520679e-01 4.84162539e-01 4.07407761e-01 -1.11340329e-01
1.85944772e+00 2.08326489e-01 7.62860030e-02 -9.84297842e-02
1.14188433e+00 -1.37292475e-01 -7.32037961e-01 5.88147342e-02
-2.42159858e-01 -3.17544997e-01 5.49242198e-01 -7.20728755e-01
-7.87260413e-01 5.43767095e-01 4.90474343e-01 -2.51198076e-02
1.05854809e+00 -2.50894994e-01 6.88391447e-01 5.09312749e-01
6.99175298e-01 -9.73629236e-01 -3.90343457e-01 7.47813642e-01
7.76110590e-01 -6.57294989e-01 3.05098176e-01 -6.05030239e-01
-1.70979381e-01 1.25556350e+00 2.12095946e-01 -3.90183359e-01
1.30300796e+00 1.38314986e+00 1.40031036e-02 -4.67897087e-01
-6.70050442e-01 3.16248029e-01 -1.99583501e-01 1.09784162e+00
5.85623503e-01 -3.18699032e-01 -1.31712258e-01 1.04527175e+00
-1.65425763e-01 4.37918842e-01 4.06258285e-01 1.27757359e+00
-2.97666878e-01 -1.01335144e+00 -7.49668002e-01 5.34478664e-01
-3.00449401e-01 -3.42378080e-01 4.13684011e-01 8.12951803e-01
-1.25963390e-01 9.39238787e-01 -8.43707994e-02 -2.53996104e-01
2.72464693e-01 -2.19616637e-01 7.64540493e-01 -7.44373083e-01
-4.18073267e-01 -1.87531099e-01 5.39615393e-01 -3.87525380e-01
-1.89933956e-01 -6.36108279e-01 -1.42575216e+00 -4.75808084e-01
-3.74767929e-01 4.35285836e-01 7.24263132e-01 1.24922407e+00
8.94000307e-02 5.44941843e-01 5.47722876e-01 -1.58603621e+00
-4.70973909e-01 -8.03767443e-01 -1.05305099e+00 -2.13539675e-01
-3.03834051e-01 -1.01317215e+00 -4.85202938e-01 -3.67366672e-01] | [6.826801300048828, 2.000822067260742] |
6b97a35c-6a5f-472f-bedf-ec4e26bd6d46 | a-monte-carlo-language-model-pipeline-for | 2305.15051 | null | https://arxiv.org/abs/2305.15051v1 | https://arxiv.org/pdf/2305.15051v1.pdf | A Monte Carlo Language Model Pipeline for Zero-Shot Sociopolitical Event Extraction | We consider dyadic zero-shot event extraction (EE) to identify actions between pairs of actors. The \emph{zero-shot} setting allows social scientists or other non-computational researchers to extract any customized, user-specified set of events without training, resulting in a \emph{dyadic} event database, allowing insight into sociopolitical relational dynamics among actors and the higher level organizations or countries they represent. Unfortunately, we find that current zero-shot EE methods perform poorly for the task, with issues including word sense ambiguity, modality mismatch, and efficiency. Straightforward application of large language model prompting typically performs even worse. We address these challenges with a new fine-grained, multi-stage generative question-answer method, using a Monte Carlo approach to exploit and overcome the randomness of generative outputs. It performs 90\% fewer queries than a previous approach, with strong performance on the widely-used Automatic Content Extraction dataset. Finally, we extend our method to extract affiliations of actor arguments and demonstrate our method and findings on a dyadic international relations case study. | ["Brendan O'Connor", 'Erica Cai'] | 2023-05-24 | null | null | null | null | ['event-extraction'] | ['natural-language-processing'] | [ 2.10951865e-01 4.86335754e-01 -2.36672953e-01 -5.34385554e-02
-1.46002066e+00 -9.80115712e-01 1.21025717e+00 4.63202447e-01
-5.72956800e-01 6.69198215e-01 1.00960147e+00 -5.07509589e-01
-1.05372965e-01 -9.80081141e-01 -4.47279751e-01 -3.57951880e-01
1.84236199e-01 8.50329459e-01 1.14763692e-01 -2.97514200e-01
3.07274133e-01 1.22512281e-01 -1.19816840e+00 4.90565211e-01
4.51949090e-01 4.67723191e-01 -3.33174258e-01 6.66478157e-01
-2.08572373e-01 1.11780643e+00 -8.51220667e-01 -9.91402864e-01
5.62463216e-02 -6.18577421e-01 -9.81330991e-01 -2.77275145e-01
-8.56702626e-02 -1.21382006e-01 -5.37048317e-02 6.35715842e-01
7.01354504e-01 8.54666457e-02 8.47643554e-01 -9.36550200e-01
-4.54722077e-01 1.35711491e+00 -3.94726396e-01 6.77489936e-01
7.47264743e-01 -1.34404534e-02 1.52646112e+00 -1.03259122e+00
1.39014733e+00 1.32923651e+00 6.74962163e-01 -9.93439406e-02
-1.30049431e+00 -6.06011212e-01 -8.20840895e-02 -6.70010522e-02
-1.08054864e+00 -5.13082922e-01 7.41420805e-01 -5.02889752e-01
1.07476223e+00 3.26936364e-01 5.61883271e-01 1.66105664e+00
-2.20756426e-01 4.57486689e-01 1.04544628e+00 -3.90815794e-01
9.83327255e-02 -9.16179344e-02 1.77975208e-01 3.63650650e-01
1.44619554e-01 -3.35429907e-01 -7.30213702e-01 -8.26083958e-01
4.15350109e-01 -1.81985140e-01 6.30550086e-02 2.98494667e-01
-1.49330378e+00 1.08102787e+00 -7.17735589e-02 6.64621949e-01
-5.58337510e-01 1.16998874e-01 4.16730404e-01 1.78733632e-01
6.73970759e-01 8.46610785e-01 -2.29696363e-01 -5.92513621e-01
-9.75307226e-01 6.02497220e-01 1.40303385e+00 6.88422501e-01
5.28847992e-01 -4.09868091e-01 -5.69764853e-01 5.11214554e-01
-2.04625521e-02 2.25910455e-01 -5.00111356e-02 -1.18063939e+00
7.42116094e-01 5.34060478e-01 2.39202306e-01 -1.27628291e+00
-3.05603981e-01 -3.04586589e-01 -3.51165563e-01 -4.24919963e-01
7.32646108e-01 -4.44405586e-01 -4.91549462e-01 1.68562555e+00
5.47438860e-01 1.22193165e-01 -1.87277533e-02 5.96185148e-01
9.52344716e-01 7.84739852e-01 3.58244389e-01 -3.56397897e-01
1.89471710e+00 -3.79115850e-01 -6.88136399e-01 -1.81418672e-01
5.87655604e-01 -9.11498785e-01 1.05806756e+00 1.82849646e-01
-1.26420045e+00 2.11794734e-01 -6.45862281e-01 -1.34958759e-01
-3.38714242e-01 -3.08932930e-01 7.46711493e-01 4.88474965e-01
-5.29332876e-01 3.92836154e-01 -6.21738553e-01 -4.44005221e-01
4.11217541e-01 -7.78682306e-02 1.65364686e-02 2.04596043e-01
-1.55498254e+00 7.46567130e-01 3.39873210e-02 -4.46233481e-01
-5.17153323e-01 -8.96620810e-01 -7.05043435e-01 1.64172187e-01
8.89830768e-01 -5.05522430e-01 1.31714833e+00 -5.93140423e-01
-1.16525531e+00 1.02270794e+00 -1.79311588e-01 -3.09167743e-01
3.80282521e-01 -7.21853301e-02 -4.60599780e-01 1.74791321e-01
5.42566538e-01 2.78168432e-02 3.60365182e-01 -1.15933359e+00
-2.85518408e-01 1.81722566e-02 1.56165004e-01 1.40785992e-01
-3.10719460e-01 7.83315957e-01 -2.35498831e-01 -8.51894557e-01
-9.25287604e-02 -6.39925003e-01 -2.62687355e-01 -9.53693271e-01
-5.17354965e-01 -6.21390045e-01 4.03843254e-01 -8.15021992e-01
1.59846044e+00 -1.90915370e+00 3.32511514e-02 4.05079633e-01
3.16689163e-01 -2.12319955e-01 1.52719170e-01 1.19793868e+00
1.26085967e-01 3.97194594e-01 -1.74304754e-01 -3.28956276e-01
1.47600293e-01 1.11346647e-01 -6.78847432e-01 1.95793018e-01
1.44698814e-01 1.10186720e+00 -1.09013522e+00 -8.24735045e-01
-2.50030339e-01 2.96946853e-01 -6.72648549e-01 -1.64385527e-01
-2.82617122e-01 2.00225413e-01 -6.36105478e-01 7.09014475e-01
-1.06127419e-01 -7.12919772e-01 6.66022718e-01 -9.69651267e-02
-2.16962725e-01 6.34691179e-01 -1.11498427e+00 1.52763343e+00
-2.98291206e-01 7.34639287e-01 -8.67061317e-03 -7.23183990e-01
5.39000750e-01 4.60500032e-01 6.04235291e-01 -3.29726636e-01
2.15922922e-01 1.51436657e-01 -1.33725688e-01 -4.81027126e-01
6.99517488e-01 -1.03375129e-01 -8.63210738e-01 1.15552771e+00
1.04228415e-01 9.03044045e-02 2.59969294e-01 9.06761646e-01
1.45815587e+00 -1.10258818e-01 5.99361300e-01 -3.45351137e-02
-2.99464583e-01 4.24554735e-01 4.77202892e-01 1.20759988e+00
8.49495307e-02 5.92743456e-01 1.09281456e+00 -2.85643667e-01
-1.11877048e+00 -9.60018992e-01 4.12664004e-02 1.25477600e+00
-1.34802341e-01 -8.23763371e-01 -6.56079054e-01 -6.51098073e-01
-1.05505183e-01 8.83513689e-01 -6.32637441e-01 3.26138169e-01
-9.49741721e-01 -9.98476088e-01 7.82817125e-01 1.98731288e-01
1.60647221e-02 -1.14277840e+00 -5.97553849e-01 5.76602280e-01
-6.38706565e-01 -1.29455733e+00 -2.26467520e-01 -1.46136701e-01
-1.07464574e-01 -1.12857556e+00 -3.51780891e-01 -3.38098168e-01
3.22199196e-01 -2.65424550e-01 1.52330101e+00 -3.59490573e-01
-2.62076706e-01 3.91207308e-01 -3.96629989e-01 -2.90038049e-01
-6.33571863e-01 1.87034890e-01 -3.57571632e-01 -1.87401026e-02
5.45991361e-01 -7.40827918e-01 -5.74520648e-01 4.71788310e-02
-8.40531528e-01 -1.35784775e-01 3.11985254e-01 6.13401294e-01
1.18389681e-01 -4.06716853e-01 8.40823412e-01 -1.37523103e+00
1.13743937e+00 -1.05287945e+00 -5.33920303e-02 3.03409606e-01
-6.27817333e-01 -1.46110147e-01 2.39436299e-01 -5.95026195e-01
-1.13675082e+00 -5.27320504e-01 6.16423748e-02 1.46550819e-01
1.57031238e-01 7.62464762e-01 2.09711399e-02 6.82367980e-01
9.78347242e-01 -2.97640145e-01 -3.80432546e-01 -2.85487115e-01
6.60768151e-01 6.00425780e-01 6.00882173e-01 -6.83426559e-01
6.16723835e-01 6.42116845e-01 -3.53309810e-01 -4.14845645e-01
-1.04446459e+00 -4.57464069e-01 -2.81353425e-02 -2.93585300e-01
8.30919683e-01 -1.00711346e+00 -7.51817822e-01 -2.17861831e-02
-1.34978461e+00 -1.48499280e-01 -5.40059388e-01 4.75787252e-01
-3.72834325e-01 8.53396878e-02 -8.20355833e-01 -6.45298958e-01
-2.69109547e-01 -7.40074813e-01 1.09084404e+00 -3.81756830e-03
-9.30861652e-01 -9.77883041e-01 4.19800550e-01 4.70341086e-01
2.32709721e-01 7.69027829e-01 8.95783782e-01 -1.24520934e+00
-4.88346487e-01 -1.82502419e-01 -1.00146390e-01 -4.13715333e-01
-2.67966360e-01 5.80607587e-03 -5.76159716e-01 2.13311166e-01
-2.44799092e-01 -3.27408671e-01 7.43963599e-01 1.23795852e-01
4.85873938e-01 -7.27037966e-01 -5.49087167e-01 2.55692154e-02
1.11152005e+00 -1.34333119e-01 3.92992646e-01 2.01494828e-01
4.09607708e-01 6.07871652e-01 3.98651689e-01 8.39159846e-01
7.48700500e-01 5.77742338e-01 -8.69499147e-02 8.13769922e-02
4.87200692e-02 -4.77475613e-01 2.41545111e-01 5.52700758e-01
-3.10142249e-01 -5.40732384e-01 -1.18183744e+00 8.96983445e-01
-1.88284647e+00 -1.48243618e+00 -2.32740864e-01 1.53535783e+00
1.03845298e+00 2.37232551e-01 2.44401425e-01 -1.96259335e-01
5.97947776e-01 5.05045116e-01 -3.60063650e-02 -2.99278200e-01
-3.04723889e-01 5.67763805e-01 4.70053047e-01 3.68421733e-01
-7.14697421e-01 1.07780457e+00 6.67261839e+00 1.00021970e+00
-5.02604306e-01 5.17773569e-01 5.37867188e-01 -3.98253322e-01
-8.90804470e-01 4.55391914e-01 -8.39311242e-01 4.36426699e-01
1.06699669e+00 -3.35362285e-01 3.42615098e-01 3.94773662e-01
1.92775309e-01 -2.90715039e-01 -9.17355835e-01 7.46575713e-01
2.38772154e-01 -1.86077464e+00 -2.84500390e-01 2.49548078e-01
8.21726382e-01 -6.56943694e-02 -2.88978457e-01 2.90482521e-01
1.06800973e+00 -8.79688859e-01 7.81109691e-01 4.67390567e-01
5.85975707e-01 -5.83578169e-01 3.34505856e-01 1.97965711e-01
-8.96091580e-01 -1.68806672e-01 2.57099718e-01 -7.52976984e-02
8.87533128e-01 8.80516708e-01 -1.03288925e+00 5.41335285e-01
5.27602553e-01 3.82270038e-01 -2.55118191e-01 2.00959161e-01
-2.35898882e-01 1.02984393e+00 -4.56004292e-01 -1.95680991e-01
2.10811004e-01 4.85295542e-02 9.59227562e-01 1.36275876e+00
4.05965388e-01 6.12627387e-01 4.33591455e-02 9.14625704e-01
-4.47771907e-01 1.03173494e-01 -7.00448275e-01 -4.17135805e-01
7.60600924e-01 1.42137718e+00 -1.01014328e+00 -5.83275318e-01
-2.98990577e-01 5.60963333e-01 4.14201289e-01 4.87369508e-01
-8.02838504e-01 -2.74387866e-01 2.97857136e-01 3.73930126e-01
2.68593729e-01 -2.16692790e-01 -2.04214677e-01 -1.19364774e+00
-1.80826724e-01 -1.08996558e+00 7.79681325e-01 -5.55138469e-01
-1.19723594e+00 3.10395837e-01 3.00561816e-01 -5.83306193e-01
-7.07730055e-01 -4.86345850e-02 -7.31814981e-01 5.97816586e-01
-5.90637743e-01 -1.22693586e+00 1.23552412e-01 2.56661177e-01
3.69181514e-01 1.71641279e-02 5.73993981e-01 2.76653945e-01
-3.92865032e-01 2.73744524e-01 -3.08649898e-01 3.50902379e-01
7.36206889e-01 -1.00942767e+00 4.74319637e-01 7.99026728e-01
4.01674181e-01 6.32103324e-01 8.62945557e-01 -8.82838249e-01
-1.24045920e+00 -6.20996058e-01 1.66175127e+00 -9.55951810e-01
1.03842413e+00 -7.66519129e-01 -4.34777349e-01 8.27498972e-01
5.08236647e-01 -4.16582823e-01 9.60855424e-01 4.48877096e-01
-3.76118928e-01 3.76348168e-01 -8.23425949e-01 8.56128097e-01
1.34363747e+00 -7.38887131e-01 -8.70945692e-01 5.71064770e-01
8.58537793e-01 -1.47213981e-01 -1.08366394e+00 2.55024999e-01
4.71018016e-01 -8.21698368e-01 9.47529078e-01 -8.81583929e-01
7.13465393e-01 1.29555538e-01 -9.86480117e-02 -9.44594204e-01
-1.82055622e-01 -1.20370626e+00 -2.06665605e-01 1.70891631e+00
6.78075433e-01 -5.99911451e-01 3.08452159e-01 7.07952261e-01
1.45564005e-01 -5.52183986e-01 -7.82086968e-01 -2.42364511e-01
-1.70829579e-01 -5.38534880e-01 6.31114483e-01 1.34428787e+00
2.00928047e-01 6.31480098e-01 -3.07150185e-01 -1.05925076e-01
4.24232364e-01 4.65813935e-01 7.44264185e-01 -1.07060099e+00
-5.57606876e-01 -4.31624323e-01 1.33595854e-01 -5.15605569e-01
1.34913206e-01 -8.53742421e-01 -3.43167514e-01 -1.65261304e+00
3.79528195e-01 -3.62096041e-01 9.79169086e-02 2.75795162e-01
-2.04444900e-01 2.03405052e-01 -3.01962756e-02 3.40892583e-01
-7.21000016e-01 1.18998408e-01 9.52885747e-01 1.91795975e-01
-3.24557066e-01 -2.92589992e-01 -9.97555614e-01 5.77014029e-01
4.13495272e-01 -7.09335625e-01 -9.58844125e-02 -1.10502735e-01
9.87424910e-01 2.13870704e-01 6.16900742e-01 -3.90193552e-01
3.04753423e-01 -1.28268242e-01 -3.62351313e-02 -5.50402105e-01
1.14288822e-01 -8.34423825e-02 5.42156100e-01 1.28761334e-02
-5.93749285e-01 2.51623005e-01 -4.10110295e-01 5.71108699e-01
-1.93463489e-01 -5.44199795e-02 2.16414228e-01 -4.10479754e-01
-1.46028548e-01 -7.54080340e-02 -4.73704159e-01 6.77475274e-01
1.03134966e+00 3.49997804e-02 -4.91706520e-01 -6.41541481e-01
-8.78881037e-01 -1.04928933e-01 3.21194977e-01 2.14841053e-01
1.15518883e-01 -1.07187438e+00 -9.84643161e-01 -3.38841647e-01
-8.76350552e-02 -1.62583098e-01 6.81165457e-02 1.00045478e+00
-1.27966836e-01 1.16307072e-01 5.15512407e-01 4.93180880e-04
-9.08957362e-01 2.71339148e-01 -1.32433891e-01 -8.64774764e-01
-6.65949702e-01 7.28304267e-01 -1.37723640e-01 -1.03040338e-01
-2.62511373e-01 -1.52499788e-02 -1.81274891e-01 7.75721431e-01
2.08055824e-01 3.95273656e-01 -2.88645893e-01 -4.71351266e-01
-2.71193683e-01 4.80598845e-02 2.11001262e-01 -8.35634708e-01
1.49169612e+00 -1.44958168e-01 -1.12342320e-01 5.44140816e-01
9.55332696e-01 6.50406420e-01 -8.46242309e-01 -3.38632226e-01
2.60220855e-01 -2.92534987e-03 -3.82144600e-01 -7.00650692e-01
-3.85230452e-01 3.18875700e-01 -4.55221683e-01 8.12372804e-01
5.03374934e-01 6.94747627e-01 8.96963954e-01 1.61973312e-02
1.77613392e-01 -1.22468889e+00 1.34645805e-01 4.19568717e-01
6.98934793e-01 -9.24938321e-01 9.13303420e-02 -3.54774058e-01
-8.15203309e-01 6.56799018e-01 3.38500254e-02 -1.17501363e-01
5.49629152e-01 3.64437968e-01 -1.09998219e-01 -6.87965631e-01
-8.64628077e-01 -1.24589011e-01 5.56484237e-03 -1.30521297e-01
3.23777378e-01 5.00934683e-02 -6.88931227e-01 9.03668880e-01
-4.58215356e-01 -3.76170784e-01 4.24466252e-01 9.14023042e-01
-3.47500034e-02 -1.25241220e+00 -2.37600729e-01 4.76264089e-01
-9.92093384e-01 -4.17272836e-01 -6.19558573e-01 9.05821383e-01
3.73259671e-02 9.99612927e-01 2.22659335e-01 -4.66052853e-02
1.75141826e-01 2.25752279e-01 2.06807479e-01 -6.61255181e-01
-1.22655678e+00 3.84612046e-02 9.30672884e-01 -5.66248715e-01
-5.94158053e-01 -1.22860241e+00 -1.09939158e+00 -3.41949731e-01
-2.70391908e-02 3.66912901e-01 5.09139657e-01 1.04004920e+00
7.13286221e-01 2.38123059e-01 6.55811250e-01 -4.45310175e-01
-4.13826525e-01 -9.73653138e-01 -1.41006365e-01 6.49608254e-01
-1.03667565e-01 -5.62449157e-01 -4.42372113e-01 1.49602816e-01] | [9.059900283813477, 9.367782592773438] |
3c6ebab9-f34b-4fe2-a629-90e7ba132745 | object-detection-in-aerial-images-with | 2208.10781 | null | https://arxiv.org/abs/2208.10781v2 | https://arxiv.org/pdf/2208.10781v2.pdf | Object Detection in Aerial Images with Uncertainty-Aware Graph Network | In this work, we propose a novel uncertainty-aware object detection framework with a structured-graph, where nodes and edges are denoted by objects and their spatial-semantic similarities, respectively. Specifically, we aim to consider relationships among objects for effectively contextualizing them. To achieve this, we first detect objects and then measure their semantic and spatial distances to construct an object graph, which is then represented by a graph neural network (GNN) for refining visual CNN features for objects. However, refining CNN features and detection results of every object are inefficient and may not be necessary, as that include correct predictions with low uncertainties. Therefore, we propose to handle uncertain objects by not only transferring the representation from certain objects (sources) to uncertain objects (targets) over the directed graph, but also improving CNN features only on objects regarded as uncertain with their representational outputs from the GNN. Furthermore, we calculate a training loss by giving larger weights on uncertain objects, to concentrate on improving uncertain object predictions while maintaining high performances on certain objects. We refer to our model as Uncertainty-Aware Graph network for object DETection (UAGDet). We then experimentally validate ours on the challenging large-scale aerial image dataset, namely DOTA, that consists of lots of objects with small to large sizes in an image, on which ours improves the performance of the existing object detection network. | ['Sung Ju Hwang', 'Jinheon Baek', 'Jongha Kim'] | 2022-08-23 | null | null | null | null | ['object-detection-in-aerial-images'] | ['computer-vision'] | [ 7.73310140e-02 3.35820854e-01 2.03209013e-01 -5.74595332e-01
-3.24976236e-01 -5.24345696e-01 4.34022695e-01 3.48893285e-01
-1.09607860e-01 4.17333007e-01 -1.98628843e-01 3.80897447e-02
-5.25721312e-01 -9.92697060e-01 -8.86429965e-01 -5.29019535e-01
-3.44930112e-01 3.51437062e-01 6.55128658e-01 4.45819236e-02
1.47605669e-02 5.57279289e-01 -1.62232983e+00 1.45471662e-01
9.16285038e-01 1.57413411e+00 3.70584905e-01 2.06983343e-01
9.19788256e-02 7.57636726e-01 -5.29299021e-01 -4.75472271e-01
1.74013674e-01 -1.84241869e-02 -7.29434729e-01 4.21139956e-01
4.51353371e-01 -4.17063266e-01 -4.21297073e-01 1.50517344e+00
1.20144472e-01 4.10593212e-01 7.18659222e-01 -1.48861611e+00
-1.24331415e+00 1.04119647e+00 -6.49041235e-01 2.55950093e-02
-1.28767833e-01 1.66888684e-01 1.11250579e+00 -1.01249301e+00
3.13900203e-01 1.37951910e+00 3.39012235e-01 3.99563700e-01
-8.91381562e-01 -5.15965044e-01 8.92934978e-01 3.70266587e-01
-1.81319630e+00 -7.77808428e-02 7.58938074e-01 -2.83902079e-01
5.44086695e-01 2.01295018e-01 4.96230751e-01 6.30997598e-01
-1.15799256e-01 9.21593547e-01 3.28688294e-01 1.29072219e-01
3.26697618e-01 -1.02285212e-02 7.68990740e-02 7.84143686e-01
4.06529814e-01 -5.78680597e-02 -1.69927925e-02 1.96267620e-01
4.37899441e-01 3.28033835e-01 -4.53241378e-01 -5.65675199e-01
-1.15044475e+00 7.22541809e-01 1.28604388e+00 8.86032544e-03
-5.33924937e-01 3.69140387e-01 1.18778281e-01 -1.49003446e-01
4.67779666e-01 3.16674292e-01 -3.90349746e-01 8.26775551e-01
-5.19505501e-01 4.82588708e-02 1.86415792e-01 1.45865691e+00
7.09578395e-01 -4.13218476e-02 -6.18187904e-01 5.69247365e-01
6.11564457e-01 3.86803955e-01 4.07732185e-03 -6.41780019e-01
5.02065182e-01 9.65798259e-01 1.48655832e-01 -1.53691709e+00
-4.59652841e-01 -6.25997603e-01 -1.02330291e+00 1.32553458e-01
8.22067782e-02 -1.21261682e-02 -1.26557755e+00 1.72423816e+00
4.51985687e-01 4.79070842e-01 4.33658101e-02 1.13713157e+00
1.00547194e+00 7.62558401e-01 1.79811627e-01 2.53087822e-02
1.24174798e+00 -9.32187736e-01 -3.73960912e-01 -4.17417973e-01
4.17952150e-01 -4.03063208e-01 7.36430168e-01 2.13401660e-01
-7.35203505e-01 -6.07572436e-01 -9.68968511e-01 1.30658641e-01
-5.65849602e-01 3.43624502e-01 5.10066748e-01 1.42659426e-01
-7.45295167e-01 8.92860353e-01 -5.92129827e-01 -7.50143304e-02
9.28401768e-01 3.49522531e-01 -2.52087772e-01 -2.73083538e-01
-1.07640350e+00 5.65422595e-01 1.12002254e+00 5.48575938e-01
-1.20324755e+00 -3.62005919e-01 -1.22917783e+00 3.37192774e-01
8.74818683e-01 -5.83717465e-01 7.40139723e-01 -7.67033517e-01
-7.05526114e-01 3.48468035e-01 4.54188615e-01 -3.81814539e-01
3.08567792e-01 -5.98318502e-02 -4.71974730e-01 -3.99209596e-02
7.03075379e-02 1.03464556e+00 1.09860873e+00 -1.53650630e+00
-1.16661286e+00 -5.06701767e-01 4.88553733e-01 1.46803662e-01
-3.62785906e-01 -2.66963094e-01 -7.04384685e-01 -6.31060898e-01
6.46897376e-01 -7.59339452e-01 -3.45066339e-01 3.34770471e-01
-7.86501706e-01 -4.43505347e-01 9.86654460e-01 -4.24366921e-01
1.08067667e+00 -2.09450746e+00 2.71749675e-01 4.03935641e-01
4.45735127e-01 1.20418370e-01 -3.91247422e-01 -3.82801414e-01
8.32214020e-03 1.50767639e-01 -5.48552215e-01 -1.61782593e-01
1.01257995e-01 3.03796768e-01 -3.35546076e-01 3.29180419e-01
6.37151480e-01 1.05836546e+00 -1.16915834e+00 -4.28367794e-01
2.86459476e-01 3.02736610e-01 -1.87031835e-01 1.45871565e-01
-5.88797808e-01 -2.23317631e-02 -8.46460104e-01 9.95843470e-01
9.97506917e-01 -4.28935379e-01 -1.57410011e-01 -5.04364014e-01
2.84543395e-01 -4.19847995e-01 -1.43354154e+00 1.39056754e+00
-5.88180497e-02 1.88400030e-01 -1.07662186e-01 -9.71094728e-01
1.08614278e+00 -1.65268347e-01 1.46067157e-01 -3.00940394e-01
1.89821795e-01 -7.50864372e-02 -8.00246745e-02 -2.39724368e-01
4.80046391e-01 4.13549930e-01 -4.10751719e-03 -1.53281569e-01
1.14195682e-01 -1.80037111e-01 7.77013674e-02 5.47173202e-01
8.03870499e-01 -1.06151244e-02 3.29336748e-02 -2.05632567e-01
5.34253538e-01 1.82301982e-03 5.38721323e-01 7.11029470e-01
-3.69166344e-01 7.18256831e-01 4.32165682e-01 -2.63731241e-01
-6.33604944e-01 -1.11796236e+00 -3.72953974e-02 7.76391029e-01
8.07011127e-01 -2.35192239e-01 -5.13332427e-01 -1.23948860e+00
3.19524944e-01 7.00081587e-01 -7.90582776e-01 -5.90315521e-01
-6.59209862e-02 -5.91489255e-01 4.00338210e-02 8.47260237e-01
5.02569139e-01 -1.08935726e+00 -2.00446054e-01 1.35984018e-01
-2.07439382e-02 -1.00592804e+00 -5.51919818e-01 1.86263308e-01
-6.60986483e-01 -1.24503279e+00 -5.57042360e-01 -7.68073797e-01
9.27216291e-01 4.00690407e-01 1.06462407e+00 2.64413327e-01
-3.05143923e-01 3.25937271e-01 -4.55354244e-01 -6.51042819e-01
1.86172396e-01 -1.50043324e-01 1.30292237e-01 2.73603112e-01
4.96686026e-02 -1.82898283e-01 -7.44522452e-01 5.37322819e-01
-1.04328692e+00 -2.39564329e-01 6.42256260e-01 7.49463618e-01
9.19068694e-01 5.57051897e-01 5.65486252e-01 -7.05052555e-01
1.78396687e-01 -7.08641529e-01 -9.27867591e-01 5.43744981e-01
-4.60197955e-01 7.82347564e-03 5.29269278e-01 -4.55319941e-01
-9.12857473e-01 2.80227572e-01 2.42247149e-01 -8.25573802e-01
1.15731405e-02 5.33178627e-01 -4.45562780e-01 -2.03830957e-01
6.68294072e-01 -4.35108058e-02 -5.16003549e-01 -1.12891216e-02
6.22164249e-01 3.42667907e-01 6.49192095e-01 -3.84822398e-01
1.07820940e+00 3.67666781e-01 2.00203747e-01 -2.07921892e-01
-1.15153205e+00 -2.87441730e-01 -5.57561934e-01 -2.42513284e-01
6.29965127e-01 -9.09729898e-01 -6.45885050e-01 2.30673149e-01
-9.63120520e-01 6.71199709e-02 -4.51002270e-01 5.90195239e-01
-2.74252176e-01 2.00613707e-01 -2.22826600e-01 -8.02473068e-01
-1.42074764e-01 -1.04925919e+00 1.19675934e+00 4.06757474e-01
3.84800673e-01 -7.07492828e-01 -6.33212388e-01 -8.55509490e-02
2.15774462e-01 4.56142455e-01 8.56583536e-01 -5.97624302e-01
-1.04585564e+00 -1.03574932e-01 -8.53453219e-01 5.03635466e-01
1.06941879e-01 1.08955048e-01 -8.59978139e-01 -2.63921052e-01
-4.23278302e-01 -3.21778566e-01 1.23996198e+00 4.65899348e-01
1.67278755e+00 -2.37488344e-01 -5.89125693e-01 5.97471476e-01
1.32168913e+00 -9.03947931e-03 4.50367481e-01 -8.42909217e-02
1.02832103e+00 7.41553605e-01 1.12595093e+00 6.01576567e-01
2.82523215e-01 3.13110173e-01 1.33119726e+00 4.27702721e-03
-2.34452169e-02 -2.76249379e-01 -5.78138186e-03 1.54503509e-01
1.31460562e-01 -6.50405407e-01 -6.73784137e-01 8.06359589e-01
-2.18183661e+00 -5.09640992e-01 -9.27118361e-02 1.94331169e+00
2.92550951e-01 9.42791551e-02 -2.36696839e-01 -2.14089885e-01
1.08704591e+00 2.82780617e-01 -8.30578506e-01 3.44597846e-01
-1.61781982e-01 -3.47354442e-01 4.47159052e-01 9.08765867e-02
-1.44311273e+00 9.92103755e-01 4.91203117e+00 9.81421947e-01
-8.05228889e-01 -2.97324240e-01 7.75695920e-01 9.50547867e-03
-3.20986569e-01 -8.70432854e-02 -7.54589736e-01 3.39724511e-01
2.07065970e-01 -1.84899285e-01 3.71174037e-01 1.34344232e+00
-3.09901834e-01 1.78919837e-01 -1.18856668e+00 8.58289063e-01
4.56123911e-02 -1.19652474e+00 5.04025459e-01 -3.81745011e-01
7.78244972e-01 -1.91800758e-01 2.26086304e-01 4.19050336e-01
5.12393951e-01 -9.64827240e-01 9.69340086e-01 5.93319356e-01
4.17555213e-01 -9.30232704e-01 8.51189971e-01 2.59041905e-01
-1.58243728e+00 -3.71656120e-01 -1.04050756e+00 5.16469121e-01
-1.94743365e-01 7.05313087e-01 -6.64430916e-01 8.15093815e-01
8.70355248e-01 9.57291901e-01 -8.61472130e-01 1.18119693e+00
-3.47172678e-01 2.64610171e-01 -3.76358360e-01 4.28813659e-02
4.81567115e-01 -1.16802745e-01 3.81865829e-01 6.78530514e-01
7.15902030e-01 5.17106175e-01 4.11062032e-01 1.29049504e+00
-3.39980841e-01 -1.09908372e-01 -5.20380080e-01 1.14557534e-01
5.05129397e-01 1.55284214e+00 -9.29808557e-01 -3.82184595e-01
-1.41671672e-01 7.90807486e-01 5.52864671e-01 2.03481078e-01
-1.01850867e+00 -5.73047519e-01 5.54961681e-01 -3.09535414e-01
5.09101570e-01 1.20182388e-01 8.53662496e-04 -9.96018112e-01
1.41505405e-01 -3.39882582e-01 7.92380929e-01 -1.15833664e+00
-1.44206226e+00 8.02335322e-01 4.59608510e-02 -1.30402482e+00
3.09188902e-01 -7.44855106e-01 -5.63314915e-01 7.97608554e-01
-1.54019010e+00 -1.38346636e+00 -6.62708104e-01 7.80587912e-01
3.46403629e-01 1.10231787e-01 2.59990931e-01 1.06178015e-01
-5.60727656e-01 3.05524617e-01 -2.57421136e-01 -6.46155269e-04
4.64986712e-01 -1.28725374e+00 2.95033336e-01 9.88947392e-01
2.14011893e-01 2.86048710e-01 4.01975632e-01 -8.60795379e-01
-1.01865435e+00 -2.01985288e+00 3.85607094e-01 -5.69740474e-01
6.11117303e-01 -2.49299288e-01 -1.10973024e+00 7.45462596e-01
-3.07338089e-01 6.77263021e-01 -1.68404296e-01 -3.10751200e-02
-3.83487791e-01 -1.04334302e-01 -1.31040382e+00 6.01209819e-01
1.38003826e+00 -2.59446770e-01 -4.70078796e-01 6.49890304e-01
1.19859052e+00 -4.15171504e-01 -7.07795560e-01 8.49606037e-01
1.01162791e-01 -5.52219093e-01 1.07814991e+00 -6.48290932e-01
2.31779024e-01 -9.06880260e-01 -2.65354574e-01 -1.61573112e+00
-6.70244515e-01 1.52701750e-01 -1.96647033e-01 1.33086276e+00
5.21081090e-01 -3.85517448e-01 5.52579582e-01 5.88835120e-01
-4.65800732e-01 -6.06505394e-01 -6.19648695e-01 -9.28128541e-01
-4.47401404e-01 -5.08245289e-01 1.03808701e+00 7.77776718e-01
-4.08119529e-01 -4.16247845e-02 -2.80217260e-01 1.07164967e+00
7.21576989e-01 3.34251314e-01 1.45842955e-01 -1.31597459e+00
8.95343199e-02 -2.88955629e-01 -7.83276618e-01 -7.90455699e-01
5.35884760e-02 -8.92675161e-01 3.37521911e-01 -1.47070014e+00
1.84494615e-01 -5.65102696e-01 -7.06142366e-01 7.07087636e-01
-4.46300983e-01 2.66039371e-01 2.29921773e-01 3.70304175e-02
-1.04190314e+00 8.86222005e-01 1.32296836e+00 -8.34854722e-01
-8.24185386e-02 1.63320273e-01 -8.00008595e-01 7.51456022e-01
5.16285062e-01 -4.90470111e-01 -6.74865305e-01 -6.25828683e-01
1.42957926e-01 -2.37533733e-01 7.39774466e-01 -1.05092633e+00
3.13607186e-01 -1.78410605e-01 7.41565347e-01 -8.20323706e-01
1.52457550e-01 -1.17816389e+00 -3.31954435e-02 2.41584674e-01
-3.26808065e-01 -2.95054883e-01 1.43935710e-01 9.92394447e-01
-2.42553487e-01 -2.51517236e-01 7.00150549e-01 -1.00331612e-01
-1.23516846e+00 1.04036462e+00 2.57141382e-01 -2.07598448e-01
1.48042178e+00 1.60682991e-01 -3.84664893e-01 -8.92009586e-02
-9.78084385e-01 8.08067024e-01 2.75328755e-01 6.19341195e-01
1.00935340e+00 -1.52382660e+00 -6.02410972e-01 8.29502940e-03
7.15208948e-01 5.98079920e-01 4.65569019e-01 2.43098512e-01
-1.83507323e-01 2.42570698e-01 -9.87147763e-02 -8.96060824e-01
-1.09418225e+00 1.15667987e+00 3.79007757e-01 1.97389811e-01
-3.79741639e-01 1.22545052e+00 6.27442062e-01 -4.62238282e-01
4.10549641e-01 -6.51619434e-01 -5.71078300e-01 4.40962911e-02
4.29451823e-01 1.35182232e-01 -1.62948351e-02 -5.94038367e-01
-5.44314146e-01 5.09639561e-01 3.32256630e-02 5.78303456e-01
1.09252954e+00 -2.86108762e-01 -2.92071663e-02 3.12874503e-02
7.82203496e-01 -5.09695649e-01 -1.49887466e+00 -7.21529305e-01
-8.10625479e-02 -5.51740408e-01 2.72802472e-01 -6.81202531e-01
-1.53274310e+00 7.19857991e-01 7.96044350e-01 2.24279299e-01
1.26170242e+00 3.73915851e-01 3.93289447e-01 7.09431767e-01
4.85305518e-01 -9.29347038e-01 6.72419220e-02 3.34801495e-01
1.08999920e+00 -1.49632359e+00 6.18017605e-03 -6.72508061e-01
-6.90810382e-01 9.47344124e-01 9.93528068e-01 -4.78763841e-02
7.53622472e-01 -1.71129048e-01 -4.03126389e-01 -3.98237944e-01
-5.01470625e-01 -6.76097333e-01 6.85929120e-01 7.33841658e-01
-3.88434500e-01 3.19883764e-01 3.66520345e-01 9.05103683e-01
4.11719441e-01 -2.56281525e-01 1.25578016e-01 5.87657571e-01
-7.57771850e-01 -3.50523531e-01 -2.53349543e-01 5.66504359e-01
3.99724655e-02 -1.35206535e-01 -3.16755176e-01 6.74514651e-01
4.14026588e-01 9.66579139e-01 1.20501138e-01 -6.50593638e-01
5.83919823e-01 -4.61899519e-01 2.56776474e-02 -7.91488707e-01
-1.14360616e-01 -5.16533136e-01 -2.21709684e-01 -5.41431427e-01
-1.92431778e-01 -2.88976520e-01 -1.31824553e+00 6.00961000e-02
-7.55011201e-01 4.64761071e-02 4.79000568e-01 7.98208714e-01
3.83584559e-01 8.64767253e-01 8.20589483e-01 -7.10804760e-01
-5.64748526e-01 -6.98824465e-01 -9.95269239e-01 4.06699389e-01
1.44471884e-01 -8.38072479e-01 -3.26083899e-01 -1.58737212e-01] | [10.013474464416504, 1.62588632106781] |
0b0d9423-d5a1-4b93-a9f7-d68e031e8942 | h-gan-the-power-of-gans-in-your-hands | 2103.15017 | null | https://arxiv.org/abs/2103.15017v2 | https://arxiv.org/pdf/2103.15017v2.pdf | H-GAN: the power of GANs in your Hands | We present HandGAN (H-GAN), a cycle-consistent adversarial learning approach implementing multi-scale perceptual discriminators. It is designed to translate synthetic images of hands to the real domain. Synthetic hands provide complete ground-truth annotations, yet they are not representative of the target distribution of real-world data. We strive to provide the perfect blend of a realistic hand appearance with synthetic annotations. Relying on image-to-image translation, we improve the appearance of synthetic hands to approximate the statistical distribution underlying a collection of real images of hands. H-GAN tackles not only the cross-domain tone mapping but also structural differences in localized areas such as shading discontinuities. Results are evaluated on a qualitative and quantitative basis improving previous works. Furthermore, we relied on the hand classification task to claim our generated hands are statistically similar to the real domain of hands. | ['Antonis Argyros', 'Jose Garcia-Rodriguez', 'Aggeliki Tsoli', 'Alberto Garcia-Garcia', 'Iason Oikonomidis', 'Sergio Orts-Escolano', 'Nikolaos Kyriazis', 'Pablo Martinez-Gonzalez', 'Giorgos Karvounas', 'Sergiu Oprea'] | 2021-03-27 | null | null | null | null | ['tone-mapping'] | ['computer-vision'] | [ 6.07450366e-01 1.99810177e-01 1.92096472e-01 -3.10119893e-02
-1.04156172e+00 -1.03998137e+00 6.28015339e-01 -8.51064622e-01
3.57668996e-02 9.28014457e-01 4.85631898e-02 2.77847588e-01
4.30848867e-01 -7.69750118e-01 -1.04688418e+00 -7.47177362e-01
3.04957747e-01 6.10108614e-01 9.73008722e-02 -3.01582694e-01
-2.52765864e-01 4.63251859e-01 -1.23730922e+00 6.32422805e-01
5.37982523e-01 8.17901969e-01 -1.85953408e-01 1.10651124e+00
3.03566039e-01 7.56611943e-01 -9.89012182e-01 -7.50569046e-01
6.88151062e-01 -1.00225639e+00 -7.23237216e-01 2.96071887e-01
1.00152111e+00 -5.22060990e-01 -3.72445673e-01 9.97927845e-01
8.28228414e-01 -3.27382267e-01 1.01262379e+00 -1.22924566e+00
-9.52602208e-01 2.23300129e-01 -6.31205440e-01 -4.59053040e-01
6.89384699e-01 7.19036162e-01 8.43661070e-01 -6.58360481e-01
1.01912045e+00 1.31747639e+00 8.83466721e-01 7.16950178e-01
-1.62303507e+00 -5.97047746e-01 -3.49361390e-01 -3.93272787e-01
-1.38874757e+00 -1.78474516e-01 8.73107731e-01 -5.30696392e-01
-5.52028883e-03 4.06652361e-01 7.51003444e-01 1.96596086e+00
1.87128380e-01 5.59905946e-01 1.81785369e+00 -6.82360113e-01
6.32509068e-02 1.90159023e-01 -7.79531419e-01 6.59227788e-01
-1.21831611e-01 4.91021544e-01 -5.26494145e-01 -4.00427543e-02
1.31331730e+00 -3.83134544e-01 -4.77228492e-01 -3.92202467e-01
-1.34918582e+00 4.76630986e-01 4.45758849e-01 7.81289339e-02
-5.82069635e-01 5.52341938e-01 1.37753775e-02 2.57469267e-01
1.50052473e-01 4.38873082e-01 -8.91420618e-02 1.33458540e-01
-8.61275136e-01 5.91140985e-01 7.55326450e-01 1.16214132e+00
5.46518564e-01 3.21316123e-01 -6.61162674e-01 4.53364909e-01
-1.18838742e-01 1.17474389e+00 5.60793048e-03 -1.33313406e+00
2.73228139e-01 1.97028160e-01 3.48835707e-01 -7.87766635e-01
4.17482644e-01 -2.10173532e-01 -7.77367711e-01 8.39201510e-01
9.69996095e-01 -2.80486763e-01 -1.09171379e+00 1.70308387e+00
1.97745953e-02 -1.60334736e-01 -1.65014789e-01 1.02455926e+00
4.31991756e-01 3.26998115e-01 -3.38976383e-02 3.75403613e-01
1.14323962e+00 -7.24834740e-01 -5.73571622e-01 1.93281025e-02
-4.66016591e-01 -1.01473856e+00 1.65259087e+00 3.95244181e-01
-1.36694443e+00 -9.43526864e-01 -8.33692610e-01 2.66471088e-01
-1.76874608e-01 1.70701161e-01 2.28079349e-01 1.07191551e+00
-1.18890822e+00 5.73822141e-01 -3.83997500e-01 -1.02143556e-01
5.65501034e-01 -2.70766933e-02 -4.40851003e-01 4.84624691e-02
-1.11225832e+00 6.78128123e-01 6.67689219e-02 -1.33308962e-01
-1.13302875e+00 -9.00145411e-01 -6.36820257e-01 -3.90830129e-01
1.02829069e-01 -6.98770821e-01 1.15722382e+00 -1.65528882e+00
-1.62543714e+00 1.26644564e+00 8.63924772e-02 -4.77152467e-02
1.27639306e+00 8.04478899e-02 -1.58304706e-01 2.24614561e-01
-1.96901709e-01 6.85478330e-01 1.47512972e+00 -2.01483822e+00
-1.73060477e-01 -7.95548111e-02 -8.71603861e-02 -1.10240296e-01
2.90143937e-01 -2.11107373e-01 -3.25157553e-01 -1.25678420e+00
-6.52654946e-01 -1.10137761e+00 1.00801796e-01 3.50630999e-01
-7.27709532e-01 5.70881724e-01 7.22066224e-01 -1.19846773e+00
5.10144591e-01 -1.87919331e+00 2.45252490e-01 4.06756282e-01
2.78783709e-01 8.77480358e-02 -4.61829692e-01 3.46252143e-01
-1.01057447e-01 -1.23816334e-01 -4.10581052e-01 -5.18235862e-01
2.02454269e-01 5.81790097e-02 -5.97707331e-01 5.08756816e-01
4.74974900e-01 1.24551737e+00 -7.88574755e-01 -5.36543250e-01
1.59436315e-01 8.07031691e-01 -3.76657277e-01 5.37178099e-01
-5.26410580e-01 1.08177555e+00 -1.04380469e-03 7.25765288e-01
7.69914925e-01 1.29662365e-01 1.11535884e-01 -3.30294371e-01
4.51320350e-01 -4.59699303e-01 -9.00942981e-01 1.75907397e+00
-5.98577499e-01 6.96556211e-01 -2.09886562e-02 -2.36668810e-01
4.73114848e-01 3.52518618e-01 1.08324729e-01 -6.96685553e-01
1.29795372e-01 2.44908884e-01 7.48639926e-03 -7.96044171e-02
2.44164586e-01 -3.76270235e-01 -1.45281866e-01 5.55927217e-01
1.02493800e-01 -6.23871982e-01 -3.57583433e-01 -9.78397727e-02
8.08073580e-01 4.88486767e-01 -1.20579332e-01 -1.09321125e-01
1.94708899e-01 -1.46265686e-01 3.72894085e-03 8.55214059e-01
5.20960689e-02 1.51036417e+00 5.87441921e-01 -2.05900162e-01
-1.69669116e+00 -1.77531040e+00 2.91746348e-01 6.53961658e-01
-9.39079933e-03 3.81537378e-01 -1.45130873e+00 -8.69177222e-01
7.61981085e-02 4.90154773e-01 -1.08008862e+00 6.89720586e-02
-5.18425643e-01 -1.18139461e-01 9.43652689e-01 5.65746605e-01
5.38510859e-01 -1.29266119e+00 -5.60256243e-01 -9.01782587e-02
-3.58016640e-02 -1.20875764e+00 -8.86328399e-01 -1.33874595e-01
5.70446849e-02 -1.20557427e+00 -1.56435227e+00 -6.32645786e-01
7.39065289e-01 -3.91988873e-01 1.59453881e+00 -3.01836520e-01
-8.65707636e-01 6.44390404e-01 -3.08854848e-01 -3.38082701e-01
-1.04097223e+00 -2.42374942e-01 -7.58014992e-02 1.92558810e-01
-3.40205133e-01 -6.02455497e-01 -8.52383137e-01 4.64546889e-01
-8.22148860e-01 -1.29408628e-01 5.93082130e-01 1.06612718e+00
6.37690544e-01 -8.32154304e-02 1.98765218e-01 -9.80859578e-01
5.94099939e-01 2.48435497e-01 -3.62954706e-01 3.93083841e-01
-1.42602861e-01 -8.37434456e-02 7.62178004e-01 -7.58757830e-01
-1.31930435e+00 1.70362722e-02 1.88172050e-02 -6.79524958e-01
-3.77107948e-01 -4.10490274e-01 -3.20782632e-01 -2.22420737e-01
9.30401325e-01 3.79556566e-01 -4.89340685e-02 -1.24482185e-01
7.08575010e-01 5.48758030e-01 9.13276732e-01 -1.02025616e+00
1.08487165e+00 6.80772185e-01 -1.42815724e-01 -6.04689598e-01
-4.96442080e-01 2.68354446e-01 -6.84432268e-01 -2.96322733e-01
9.84378636e-01 -6.92102373e-01 -6.17093146e-01 8.12444270e-01
-1.28859377e+00 -1.14572668e+00 -9.47389960e-01 -1.24421716e-01
-1.03764474e+00 1.62514731e-01 -5.53491116e-01 -8.02894294e-01
-1.16678238e-01 -8.15298378e-01 1.61453009e+00 -2.17215996e-02
-3.78190011e-01 -1.06403542e+00 1.73823237e-01 8.64939243e-02
4.29170698e-01 8.07189047e-01 6.70671821e-01 1.47134721e-01
-4.29629415e-01 -1.37350097e-01 -1.29277751e-01 7.71436572e-01
3.46482426e-01 1.59134567e-01 -1.42104983e+00 -5.36226034e-01
-3.33976835e-01 -4.26223308e-01 4.89616781e-01 2.95496821e-01
1.06641221e+00 -4.43716258e-01 1.65325135e-01 3.85049909e-01
1.39277828e+00 -1.45344734e-01 1.02034080e+00 -3.17778885e-01
7.36429989e-01 8.12585115e-01 3.44069660e-01 2.66350359e-01
-2.00648591e-01 8.38861763e-01 2.77667284e-01 -7.11011648e-01
-1.11824107e+00 -6.07721269e-01 2.25246325e-01 -8.23229402e-02
-4.44570839e-01 -3.91337425e-01 -7.08777010e-01 5.17423630e-01
-1.21087766e+00 -1.02467108e+00 1.29586875e-01 2.18796515e+00
1.05835557e+00 -4.56370860e-02 6.10410690e-01 2.00993299e-01
8.03525984e-01 -1.52517622e-02 -4.34443563e-01 -1.18740253e-01
-3.32343400e-01 9.26845372e-01 5.75490832e-01 4.45058197e-01
-8.70832264e-01 9.04682994e-01 7.02348471e+00 9.15149808e-01
-1.14721918e+00 2.45305791e-01 6.36656642e-01 1.57420576e-01
-4.38153177e-01 -3.64300579e-01 -1.70399159e-01 6.31062388e-01
3.70545685e-01 3.76154959e-01 8.62798214e-01 5.54995835e-01
-1.91501573e-01 2.55876094e-01 -1.02932966e+00 1.01728201e+00
8.20894614e-02 -1.12708116e+00 1.96928009e-01 1.49550349e-01
1.21178794e+00 -5.88708878e-01 7.74731636e-01 -2.10623935e-01
6.11883283e-01 -1.51157856e+00 1.19164312e+00 7.30506063e-01
1.72338724e+00 -5.78896523e-01 7.10198700e-01 -1.36810914e-01
-1.01339889e+00 4.94322509e-01 2.05582738e-01 3.96526843e-01
6.20970465e-02 2.25483969e-01 -7.02419937e-01 3.54388922e-01
5.55393398e-01 1.11033954e-01 -5.99625707e-01 2.02049360e-01
-3.84593755e-01 3.36266786e-01 -1.23049868e-02 3.90154600e-01
-1.41866401e-01 -4.62336279e-02 3.83284837e-01 1.25487185e+00
2.00283110e-01 -4.38105673e-01 -1.26353502e-01 1.41114402e+00
-1.70337275e-01 -2.62507468e-01 -6.71716630e-01 -2.73195982e-01
1.96065322e-01 7.18236327e-01 -5.40958583e-01 -1.67035565e-01
-6.79317564e-02 1.81831717e+00 -6.32948875e-02 5.75581670e-01
-9.35833395e-01 -5.35778463e-01 5.70458651e-01 4.04564291e-01
3.06418687e-01 2.85014287e-02 -4.23021078e-01 -8.43931198e-01
2.30588004e-01 -1.26615226e+00 -1.58273086e-01 -9.55692947e-01
-1.59106433e+00 6.97795868e-01 -3.15267861e-01 -1.09360826e+00
-3.74210984e-01 -6.92481160e-01 -5.83736360e-01 1.28694820e+00
-1.14572239e+00 -1.95347309e+00 -7.16722250e-01 7.04640269e-01
3.37820351e-01 -3.10560375e-01 1.06356084e+00 3.70581746e-02
2.88367510e-01 9.97684479e-01 -1.14681095e-01 3.25147659e-01
9.73252296e-01 -1.52881229e+00 7.73407042e-01 6.47231162e-01
-1.33433482e-02 3.58992040e-01 9.76452649e-01 -4.86556977e-01
-1.15219235e+00 -1.08731318e+00 7.49395508e-03 -8.36463034e-01
2.75119573e-01 -3.45020711e-01 -6.62477136e-01 5.75142026e-01
3.68209481e-01 2.45697916e-01 2.58491546e-01 -5.57413578e-01
-6.21292293e-01 1.11483395e-01 -1.56353259e+00 6.18753612e-01
9.99064147e-01 -9.96067584e-01 -3.06290358e-01 8.14502910e-02
1.93667948e-01 -6.86050832e-01 -9.41653907e-01 1.06285870e-01
1.04354286e+00 -1.25562680e+00 1.23654735e+00 -4.28473592e-01
8.00098598e-01 -2.55896360e-01 -2.75435358e-01 -1.43468475e+00
1.48461848e-01 -7.85737395e-01 7.18439743e-02 1.28195167e+00
8.05174187e-02 -4.04478967e-01 1.09208500e+00 2.19664484e-01
4.41480368e-01 -2.89832294e-01 -6.51007533e-01 -9.91541386e-01
3.67852271e-01 -1.34835154e-01 5.57588398e-01 7.80377090e-01
-6.91031158e-01 -2.49219052e-02 -6.20962203e-01 -6.47885427e-02
1.13461936e+00 1.10971898e-01 1.18135893e+00 -7.64766693e-01
-7.95837760e-01 -3.08139592e-01 -3.58100057e-01 -6.01818502e-01
2.70704508e-01 -4.46552217e-01 1.71428338e-01 -1.08968818e+00
2.16388464e-01 -4.24406379e-01 1.22673370e-01 1.26451641e-01
-9.02376473e-02 1.10617840e+00 1.13909997e-01 2.98050418e-02
-1.64056867e-02 3.55200112e-01 1.77265799e+00 -3.56525451e-01
7.63213262e-02 -9.04290355e-04 -4.04387981e-01 5.24233282e-01
6.70847535e-01 -2.08272710e-01 -2.61875659e-01 -1.77297682e-01
-9.06008407e-02 -4.83961552e-02 9.16061223e-01 -9.67578650e-01
-3.99375349e-01 -1.01899959e-01 7.45809138e-01 -1.42653435e-02
4.19534564e-01 -9.14304793e-01 5.42861760e-01 5.68533003e-01
-4.62096334e-01 -3.64697635e-01 6.81829127e-03 6.17199838e-01
-9.48569179e-03 2.01101124e-01 1.31176519e+00 -2.45670900e-01
-3.11448961e-01 1.71248391e-01 1.33300677e-01 3.16761017e-01
1.00311589e+00 -3.13542396e-01 -7.03928322e-02 -7.54095793e-01
-6.67830408e-01 -5.18497467e-01 1.07274532e+00 1.64102405e-01
4.11973804e-01 -1.43242741e+00 -8.42242777e-01 3.49497199e-01
1.04145139e-01 -2.80577578e-02 1.69498652e-01 3.70765895e-01
-8.93598735e-01 6.89372793e-02 -5.68419456e-01 -6.10875428e-01
-1.20704699e+00 4.42963332e-01 6.26035571e-01 -6.05117343e-02
-5.52258909e-01 8.91150057e-01 4.04870778e-01 -4.69596863e-01
2.45769322e-01 -1.98589697e-01 7.14291215e-01 -4.48784232e-01
3.24774265e-01 1.65241987e-01 -2.02434823e-01 -5.28126717e-01
-5.46554253e-02 7.47165442e-01 5.32016516e-01 -4.65625077e-01
8.14698815e-01 2.97560394e-01 1.15998209e-01 1.95852876e-01
1.03177476e+00 5.25671184e-01 -1.83132207e+00 -5.97724468e-02
-6.41313195e-01 -7.91763544e-01 -4.56378311e-01 -1.18941295e+00
-1.11030746e+00 9.28868234e-01 1.04130280e+00 1.50898481e-02
8.54602218e-01 9.66359153e-02 8.16623211e-01 -1.60891190e-01
6.08248830e-01 -7.91422188e-01 5.51451504e-01 -1.84542760e-02
1.23024654e+00 -1.05139136e+00 -2.37829506e-01 -6.33338094e-01
-8.29485536e-01 8.06123197e-01 3.80094051e-01 -2.71027952e-01
2.08026081e-01 5.57523012e-01 3.58667940e-01 1.62364244e-01
2.75636651e-02 -2.49762177e-01 4.22653586e-01 1.34564388e+00
1.68321729e-01 1.75964907e-01 3.92424315e-01 2.36072674e-01
-3.97070736e-01 -1.06452368e-02 1.59349293e-01 5.58771193e-01
3.41309816e-01 -1.26184106e+00 -6.44899249e-01 -9.84731689e-03
-4.58631754e-01 1.34847492e-01 -7.12801516e-01 9.34454679e-01
3.08669776e-01 5.21689355e-01 1.44398706e-02 -4.13290918e-01
4.98824120e-01 -1.68033913e-01 1.32314122e+00 -3.92792970e-01
-5.11636794e-01 -1.25509650e-01 -1.53024033e-01 -4.68115658e-01
-3.03913057e-01 -2.41694078e-01 -6.91221654e-01 -5.15791118e-01
1.28068820e-01 -2.78725207e-01 4.79341269e-01 5.62387407e-01
-5.43999933e-02 8.12602043e-01 5.23951352e-01 -1.27949846e+00
-6.48193300e-01 -7.62412310e-01 -8.93677652e-01 8.60283613e-01
4.80046660e-01 -4.52558607e-01 -3.33111912e-01 6.24178231e-01] | [11.770124435424805, -0.5404829382896423] |
654687c1-7f82-4d05-ae78-89d8de429b5c | xwikigen-cross-lingual-summarization-for | 2303.12308 | null | https://arxiv.org/abs/2303.12308v2 | https://arxiv.org/pdf/2303.12308v2.pdf | XWikiGen: Cross-lingual Summarization for Encyclopedic Text Generation in Low Resource Languages | Lack of encyclopedic text contributors, especially on Wikipedia, makes automated text generation for low resource (LR) languages a critical problem. Existing work on Wikipedia text generation has focused on English only where English reference articles are summarized to generate English Wikipedia pages. But, for low-resource languages, the scarcity of reference articles makes monolingual summarization ineffective in solving this problem. Hence, in this work, we propose XWikiGen, which is the task of cross-lingual multi-document summarization of text from multiple reference articles, written in various languages, to generate Wikipedia-style text. Accordingly, we contribute a benchmark dataset, XWikiRef, spanning ~69K Wikipedia articles covering five domains and eight languages. We harness this dataset to train a two-stage system where the input is a set of citations and a section title and the output is a section-specific LR summary. The proposed system is based on a novel idea of neural unsupervised extractive summarization to coarsely identify salient information followed by a neural abstractive model to generate the section-specific text. Extensive experiments show that multi-domain training is better than the multi-lingual setup on average. | ['Vasudeva Varma', 'Manish Gupta', 'Shivansh Subramanian', 'Anupam Patil', 'Shivprasad Sagare', 'Dhaval Taunk'] | 2023-03-22 | null | null | null | null | ['unsupervised-extractive-summarization', 'multi-document-summarization', 'extractive-summarization', 'document-summarization'] | ['natural-language-processing', 'natural-language-processing', 'natural-language-processing', 'natural-language-processing'] | [ 1.83763430e-01 4.08713609e-01 -4.73893851e-01 1.42702863e-01
-1.47871315e+00 -5.66931546e-01 9.26116824e-01 1.54779613e-01
-3.71519059e-01 1.48364317e+00 8.81880641e-01 -2.48289511e-01
2.92333275e-01 -7.59328187e-01 -9.47485626e-01 -1.50055960e-01
5.05614698e-01 5.46791792e-01 -7.35667720e-02 -5.14901578e-01
5.63786507e-01 -1.76137820e-01 -1.28310800e+00 4.59623665e-01
1.78899729e+00 1.67136595e-01 5.01813650e-01 7.99953818e-01
-7.21936584e-01 6.94358766e-01 -9.66821492e-01 -4.68948007e-01
5.61324917e-02 -8.50059152e-01 -9.26489711e-01 -8.07664245e-02
7.72281647e-01 -2.79722273e-01 -3.06551307e-01 9.01290536e-01
8.21525931e-01 2.91897859e-02 1.10238695e+00 -7.58346856e-01
-9.22419012e-01 1.41458046e+00 -6.89234376e-01 2.28541881e-01
5.53993464e-01 -7.13600591e-02 1.18288100e+00 -8.90380442e-01
1.22359109e+00 1.16797447e+00 3.18533212e-01 7.01503992e-01
-1.00418365e+00 -2.27054656e-01 -1.13766886e-01 -2.08331674e-01
-1.11853325e+00 -6.14182055e-01 8.02404523e-01 -2.26360232e-01
1.13284755e+00 3.13145667e-02 2.87215799e-01 1.29670787e+00
3.52524430e-01 9.94349360e-01 6.99788988e-01 -6.68534279e-01
-2.81897858e-02 3.94456744e-01 1.30757419e-02 3.53941768e-01
8.89437020e-01 -6.27417326e-01 -5.91396689e-01 2.31015757e-02
1.39139101e-01 -6.50789261e-01 -3.07749212e-01 2.67380208e-01
-1.52321577e+00 8.77676129e-01 4.24081646e-02 2.07544595e-01
-6.45577133e-01 -3.80208522e-01 7.06851363e-01 2.38124818e-01
6.24658585e-01 8.10290456e-01 -1.29532024e-01 1.34239167e-01
-1.30932999e+00 7.41530776e-01 1.18562651e+00 1.36499846e+00
5.67373931e-01 2.74991512e-01 -4.77875888e-01 1.10876691e+00
5.58873035e-05 7.79838800e-01 8.82854342e-01 -4.69738513e-01
1.37060475e+00 6.72350049e-01 9.93351117e-02 -9.30305719e-01
-2.33685538e-01 -2.87267357e-01 -1.04802835e+00 -4.67571229e-01
-1.19752042e-01 -7.86357880e-01 -5.51697493e-01 1.43924022e+00
9.34279989e-03 -6.30708933e-01 8.74656439e-01 5.39118826e-01
1.45447350e+00 1.17887723e+00 -1.59721047e-01 -3.75020236e-01
1.24172056e+00 -1.28290904e+00 -7.42404938e-01 -3.28444570e-01
6.31833434e-01 -9.61163938e-01 8.30418885e-01 5.19877672e-02
-1.30896437e+00 -4.87513393e-01 -1.35939133e+00 -5.67129910e-01
-6.69619262e-01 7.55315900e-01 -2.46007647e-02 -7.49870837e-02
-8.46603572e-01 3.83930296e-01 -1.10637784e-01 -5.71010232e-01
1.45886734e-01 -3.09020102e-01 -3.79818738e-01 -3.15711275e-02
-1.45227635e+00 9.37438786e-01 9.95706916e-01 -1.79854453e-01
-5.03943741e-01 -5.62433302e-01 -8.85750532e-01 -2.00619787e-01
3.78003478e-01 -1.03546417e+00 1.19523883e+00 -6.11073315e-01
-1.40139771e+00 7.79535830e-01 -5.45273386e-02 -6.28622949e-01
6.95014954e-01 -3.43212992e-01 -3.37211460e-01 2.83049315e-01
6.83924079e-01 7.08404899e-01 4.89412189e-01 -1.22427702e+00
-8.67926955e-01 2.74379384e-02 -7.91716725e-02 6.27041996e-01
-2.11461231e-01 -9.20862928e-02 -4.80804712e-01 -9.03106570e-01
-5.33004224e-01 -6.42408609e-01 -6.52828440e-02 -1.07759094e+00
-1.20857036e+00 -4.18826848e-01 3.81629199e-01 -1.20828378e+00
1.55054355e+00 -1.50655997e+00 3.11620891e-01 -2.38796175e-01
3.45765837e-02 1.01250984e-01 -2.89595783e-01 9.95753169e-01
2.54373580e-01 3.75759244e-01 -2.72873700e-01 -1.42279208e-01
1.47534654e-01 -3.59897643e-01 -6.53595090e-01 -2.21116006e-01
3.10095042e-01 8.77923787e-01 -1.00856948e+00 -8.70244086e-01
-3.47784847e-01 4.15321812e-02 -3.17960232e-01 2.26982072e-01
-5.21284878e-01 7.01963678e-02 -6.60370409e-01 3.38841110e-01
4.16577965e-01 1.22557670e-01 6.31324435e-03 -1.13226429e-01
-5.32624483e-01 6.49312019e-01 -8.68038237e-01 1.81435239e+00
-6.37323737e-01 7.60218441e-01 -3.48849684e-01 -6.43393159e-01
9.44618165e-01 2.94471771e-01 5.61314374e-02 -6.40685081e-01
-8.18095356e-02 5.50425708e-01 -2.47250453e-01 -4.68451679e-01
1.33452535e+00 3.22597593e-01 -6.70664907e-01 7.81938374e-01
2.15634182e-01 -5.12097776e-01 1.19089282e+00 8.60796332e-01
8.36864829e-01 2.66158521e-01 5.56165338e-01 -2.34186113e-01
6.15669847e-01 5.47774494e-01 2.33892843e-01 8.63089502e-01
4.56297040e-01 5.97249329e-01 5.93552411e-01 1.44445151e-02
-1.45981777e+00 -8.96555841e-01 1.73855394e-01 8.54922712e-01
-1.57471523e-01 -6.60133481e-01 -1.11981249e+00 -6.77361190e-01
-1.24403268e-01 1.15825987e+00 -1.80494338e-01 -8.27564076e-02
-8.08522642e-01 -7.88789213e-01 7.72054791e-01 1.08724676e-01
6.64548993e-01 -1.39958739e+00 -2.50703543e-01 3.75606686e-01
-8.28846097e-01 -1.08124697e+00 -6.21950686e-01 -1.92365319e-01
-5.88517785e-01 -7.56063223e-01 -1.18270540e+00 -9.99177933e-01
7.13981688e-01 1.42536640e-01 1.37944579e+00 -5.19822776e-01
-1.82361379e-01 1.38828322e-01 -4.08069342e-01 -7.06286430e-01
-1.00541186e+00 9.05585110e-01 7.54812211e-02 -4.81978685e-01
2.15259284e-01 1.36213535e-02 -1.56118765e-01 -5.37601292e-01
-8.41226399e-01 5.66491067e-01 1.11742520e+00 7.54613638e-01
5.82688749e-01 -2.44025320e-01 1.29977310e+00 -1.16064298e+00
1.40548694e+00 -7.15590417e-01 -2.54420877e-01 6.03130460e-01
-2.79208094e-01 4.07590449e-01 1.11078846e+00 -1.45081818e-01
-1.51945126e+00 -3.68874282e-01 1.28039196e-01 4.71884787e-01
2.37154618e-01 9.74874556e-01 -1.26597211e-01 7.33385921e-01
9.12191093e-01 7.58449554e-01 -2.57075638e-01 -4.04682219e-01
6.73937738e-01 1.12982202e+00 7.97245622e-01 -7.00958073e-01
7.48159051e-01 -2.72637278e-01 -4.98141676e-01 -1.09474564e+00
-9.14447129e-01 -2.78905541e-01 -7.04897940e-01 -2.29778960e-01
7.43757606e-01 -1.22183013e+00 2.47633725e-01 2.81979740e-01
-1.41537750e+00 -1.88660935e-01 -3.29173207e-01 2.41233662e-01
-5.00586033e-01 3.40426117e-01 -4.28494662e-01 -2.25029707e-01
-1.19649732e+00 -6.33968532e-01 1.09567213e+00 5.69703758e-01
-3.87352675e-01 -8.16975534e-01 4.02527511e-01 4.51077402e-01
1.69032037e-01 3.13980162e-01 9.16062415e-01 -9.07668471e-01
-3.92181844e-01 -2.33637571e-01 -2.20364884e-01 3.81478816e-01
1.43888429e-01 2.60366201e-01 -2.64685035e-01 -5.04355207e-02
-5.14222383e-01 -6.02120459e-01 9.83504117e-01 2.74392039e-01
6.62112236e-01 -7.26182580e-01 -2.06188992e-01 1.43410712e-01
1.32625270e+00 -2.39893124e-01 5.02234280e-01 5.40033281e-01
8.82355094e-01 6.15336537e-01 6.21737063e-01 4.30477172e-01
9.36864972e-01 3.59174639e-01 -3.73693109e-01 -3.44757177e-02
-4.06379819e-01 -4.67194170e-01 6.90300703e-01 1.60720003e+00
-5.09670749e-02 -5.81325173e-01 -8.05895388e-01 8.05106342e-01
-1.62112010e+00 -1.23607326e+00 -1.71184570e-01 1.94698501e+00
1.50469589e+00 7.56882876e-02 -2.35622134e-02 -5.33938527e-01
7.57444918e-01 3.44216138e-01 -5.27206063e-01 -6.53584599e-01
-4.68634605e-01 -1.92185909e-01 4.60494310e-01 3.83505821e-01
-9.47183728e-01 1.13991773e+00 5.32370090e+00 9.89392400e-01
-1.02421534e+00 -2.31826857e-01 3.41837823e-01 -1.63532734e-01
-5.08183122e-01 -3.20013314e-01 -1.28444374e+00 5.14578819e-01
1.04202139e+00 -1.23412251e+00 9.30165276e-02 8.35950196e-01
3.88883919e-01 -9.36915204e-02 -7.65445590e-01 5.89314759e-01
6.06218338e-01 -1.51413858e+00 7.97676086e-01 -3.14057410e-01
1.28277040e+00 1.19443767e-01 -2.62360901e-01 6.62617624e-01
4.81620222e-01 -6.57033026e-01 7.92494953e-01 5.72082162e-01
9.52323198e-01 -8.53103280e-01 6.71009898e-01 5.94959915e-01
-8.35421264e-01 4.99026656e-01 -6.72473192e-01 3.67664665e-01
2.73017079e-01 4.81542587e-01 -1.02899230e+00 9.85455632e-01
1.68873370e-01 7.26001620e-01 -5.99958301e-01 7.72318780e-01
-4.08102155e-01 2.94333756e-01 1.39706032e-02 -3.68626267e-01
4.93081778e-01 -1.78178206e-01 7.50973642e-01 1.74925911e+00
5.74705660e-01 -1.61002889e-01 2.39040792e-01 7.56110251e-01
-7.98940659e-01 6.39272153e-01 -9.38832462e-01 -3.36215705e-01
4.93330896e-01 1.43651378e+00 -4.72551942e-01 -7.42736280e-01
-2.91156769e-01 1.05218494e+00 4.21669126e-01 3.33791375e-01
-5.28148532e-01 -1.22227836e+00 -7.77846575e-02 -3.05978894e-01
1.43242970e-01 3.73244509e-02 -1.51092842e-01 -1.49851561e+00
1.46203563e-01 -1.05982351e+00 2.83427328e-01 -7.37576067e-01
-1.06978810e+00 7.71946490e-01 1.23761734e-02 -1.21048272e+00
-7.76963234e-01 -6.00472838e-02 -8.04863036e-01 1.00685549e+00
-1.60435319e+00 -1.25820124e+00 -9.11921337e-02 3.17970403e-02
1.23962176e+00 -6.39006376e-01 5.13539672e-01 1.20987989e-01
-8.37162375e-01 5.30253530e-01 6.08135104e-01 1.57195330e-01
1.09788561e+00 -1.56349754e+00 5.95769644e-01 1.25871813e+00
-5.97289465e-02 7.62325406e-01 8.31973374e-01 -1.15098524e+00
-1.31963527e+00 -1.49484587e+00 1.41948903e+00 -3.04607749e-01
5.90883911e-01 -2.74895251e-01 -5.96187472e-01 4.72645134e-01
9.40803170e-01 -9.75389719e-01 5.68107307e-01 -2.92104870e-01
6.15897328e-02 -2.73285899e-02 -8.00058007e-01 9.88836050e-01
7.08769619e-01 -1.97603256e-01 -9.89883006e-01 4.84654456e-01
8.75633419e-01 -5.21845520e-01 -7.92045593e-01 8.08704924e-03
2.39266425e-01 -3.34068149e-01 6.50681317e-01 -3.96622390e-01
1.31057811e+00 -3.13570589e-01 3.76312703e-01 -1.89538503e+00
3.19108292e-02 -6.03223681e-01 -9.18294638e-02 1.72877395e+00
8.53161931e-01 -2.89624304e-01 2.74157912e-01 7.96551525e-04
-5.25791287e-01 -4.12613750e-01 -4.44982708e-01 -4.52642143e-01
3.16504210e-01 2.69554585e-01 2.67353475e-01 7.19463289e-01
2.02467188e-01 1.10376787e+00 -4.51964766e-01 -3.93534809e-01
5.42595267e-01 1.66127682e-01 1.08645225e+00 -9.54598844e-01
3.51827949e-01 -6.22661650e-01 4.00140136e-01 -9.17272985e-01
3.86717290e-01 -1.14914441e+00 3.69045854e-01 -2.17401505e+00
4.54262555e-01 8.51333141e-02 4.61506903e-01 1.66001290e-01
-4.62750167e-01 -1.27410546e-01 -4.46134657e-02 2.61579216e-01
-7.95478284e-01 5.69710910e-01 1.11454713e+00 -2.55737573e-01
-3.95978659e-01 -3.17957729e-01 -1.21712160e+00 4.49681908e-01
7.29419589e-01 -2.15662539e-01 -4.12751645e-01 -4.61639225e-01
3.70193094e-01 4.76824567e-02 -2.71942616e-01 -1.04799199e+00
3.49491805e-01 2.30716076e-03 2.94900000e-01 -9.69042599e-01
-3.93721908e-01 1.23894595e-01 -2.74435014e-01 1.98842525e-01
-7.44063258e-01 1.50798008e-01 2.18893662e-01 2.60547429e-01
-4.35663849e-01 -4.84943062e-01 4.24830705e-01 -4.27404732e-01
-6.54564798e-01 8.63415375e-02 -5.52806854e-01 7.26491272e-01
6.76014721e-01 6.10287674e-02 -8.59556258e-01 -1.70495138e-01
8.04463103e-02 4.49075818e-01 4.09866303e-01 5.58723569e-01
4.67981935e-01 -1.30281901e+00 -1.53799164e+00 -3.55805516e-01
2.86486566e-01 1.74768642e-01 2.42732838e-01 4.28974420e-01
-8.77530813e-01 9.06248212e-01 -3.26641858e-01 1.50210513e-02
-8.43937039e-01 2.08033338e-01 -2.91049778e-01 -6.94820344e-01
-6.23377025e-01 3.92401993e-01 -4.05031852e-02 -5.14536202e-01
-1.54162034e-01 -1.77261576e-01 -7.08907843e-01 5.95075369e-01
7.45314956e-01 3.73512745e-01 -3.28044519e-02 -7.12971509e-01
2.06871286e-01 3.02966654e-01 -5.96936107e-01 -2.99927682e-01
1.30500913e+00 -2.94686258e-01 -2.94603974e-01 5.07189631e-01
9.69460368e-01 4.42336351e-01 -5.44026494e-01 -1.48870945e-01
1.79789066e-01 3.66076082e-01 -1.22719049e-01 -8.97393465e-01
-4.66878951e-01 4.92187858e-01 -4.16398644e-01 1.91949427e-01
7.44531274e-01 -1.07428551e-01 1.00693488e+00 7.51825154e-01
2.72592120e-02 -1.66118467e+00 -7.44026676e-02 8.42159629e-01
1.32886517e+00 -1.28948486e+00 1.90836176e-01 7.78044090e-02
-1.05188692e+00 1.25155866e+00 8.62203360e-01 -5.27687706e-02
-1.72927096e-01 -1.81607231e-01 -4.65467386e-02 4.01018262e-02
-8.60685110e-01 -6.75001293e-02 4.49913472e-01 4.68425542e-01
6.30193174e-01 -5.69764599e-02 -7.32558310e-01 6.56139255e-01
-6.96236014e-01 -1.31795317e-01 1.17510426e+00 6.31981492e-01
-5.35322785e-01 -7.96194911e-01 -1.91512913e-01 7.88785577e-01
-7.23607659e-01 -3.72088313e-01 -5.79055846e-01 7.99767435e-01
-3.51059556e-01 7.03439116e-01 -8.31948444e-02 -4.53077219e-02
2.63500482e-01 -1.61711371e-03 1.01579204e-01 -9.34964776e-01
-4.30458963e-01 -7.07218722e-02 5.54125249e-01 1.44747555e-01
-2.01839939e-01 -7.20220208e-01 -1.18448150e+00 -3.36863190e-01
-2.89483834e-02 4.35259819e-01 8.16150904e-01 7.48525798e-01
5.35248041e-01 6.62231565e-01 4.98444170e-01 -8.62915814e-01
-5.49194872e-01 -1.34638727e+00 -2.43157610e-01 2.10897550e-01
1.26221046e-01 -6.19865321e-02 -2.16058508e-01 4.22277093e-01] | [12.342965126037598, 9.517804145812988] |
6ddb1fda-8b6b-4db3-9232-e703f2834909 | yolo3d-end-to-end-real-time-3d-oriented | 1808.02350 | null | http://arxiv.org/abs/1808.02350v1 | http://arxiv.org/pdf/1808.02350v1.pdf | YOLO3D: End-to-end real-time 3D Oriented Object Bounding Box Detection from LiDAR Point Cloud | Object detection and classification in 3D is a key task in Automated Driving
(AD). LiDAR sensors are employed to provide the 3D point cloud reconstruction
of the surrounding environment, while the task of 3D object bounding box
detection in real time remains a strong algorithmic challenge. In this paper,
we build on the success of the one-shot regression meta-architecture in the 2D
perspective image space and extend it to generate oriented 3D object bounding
boxes from LiDAR point cloud. Our main contribution is in extending the loss
function of YOLO v2 to include the yaw angle, the 3D box center in Cartesian
coordinates and the height of the box as a direct regression problem. This
formulation enables real-time performance, which is essential for automated
driving. Our results are showing promising figures on KITTI benchmark,
achieving real-time performance (40 fps) on Titan X GPU. | ['Mahmoud Zidan', 'Sherif Abdelkarim', 'Mohamed Zahran', 'Waleed Ali', 'Ahmad El Sallab'] | 2018-08-07 | null | null | null | null | ['3d-point-cloud-reconstruction', 'point-cloud-reconstruction'] | ['computer-vision', 'computer-vision'] | [-1.05099723e-01 -3.12628210e-01 1.17177129e-01 -4.33734596e-01
-6.60787761e-01 -4.19448078e-01 4.13656771e-01 -8.54594111e-02
-7.44090796e-01 3.22725147e-01 -6.25875592e-01 -5.64575911e-01
2.43621081e-01 -7.77781308e-01 -8.84729624e-01 -5.37077665e-01
-9.41216722e-02 7.90493369e-01 5.64348161e-01 -4.76500422e-01
3.79476041e-01 1.18506086e+00 -2.02273250e+00 -2.01539412e-01
4.45798308e-01 1.19532490e+00 1.77920371e-01 9.14137423e-01
-2.07439467e-01 3.14617723e-01 -4.08510089e-01 -8.29412714e-02
6.28442764e-01 4.15970802e-01 -1.91609129e-01 1.04126751e-01
7.21259356e-01 -3.43503147e-01 -7.93947838e-03 6.59769952e-01
4.59804922e-01 1.19494922e-01 6.10332131e-01 -1.59469223e+00
4.26306218e-01 -2.72588521e-01 -6.89554989e-01 1.21533766e-01
-5.85241169e-02 2.72167504e-01 5.68643689e-01 -1.17357886e+00
4.45316881e-01 1.19319963e+00 5.85637987e-01 3.00924331e-01
-1.08362198e+00 -7.64907777e-01 7.91957602e-02 3.34217399e-01
-1.53037786e+00 -2.66852975e-01 7.43059397e-01 -6.85600042e-01
1.10040641e+00 2.12494984e-01 6.02232277e-01 5.75770438e-01
4.91777092e-01 5.10726750e-01 1.10906923e+00 -2.14762479e-01
2.68107116e-01 2.15960249e-01 1.86746150e-01 6.11487031e-01
2.33194545e-01 4.94132638e-01 -4.57790345e-01 1.53237000e-01
4.45858717e-01 -8.89037997e-02 3.04554284e-01 -9.15980279e-01
-8.88032615e-01 1.02624571e+00 5.15902340e-01 -4.19080168e-01
-4.82382439e-02 4.80564415e-01 4.13392097e-01 1.38777256e-01
6.59668148e-01 5.35345189e-02 -2.86037177e-01 -3.43938172e-01
-7.92632043e-01 6.40105665e-01 4.66070354e-01 1.36945283e+00
8.98058236e-01 1.04983084e-01 1.90201923e-01 1.58289611e-01
1.33031115e-01 8.26059043e-01 2.34645996e-02 -9.55485821e-01
5.58397114e-01 6.90871656e-01 2.62925208e-01 -5.99022150e-01
-4.89691138e-01 -4.24552411e-01 -2.95233637e-01 1.14067030e+00
2.16085419e-01 -4.50167619e-02 -7.86791623e-01 8.49216759e-01
8.57383490e-01 8.65529031e-02 -1.54895499e-01 9.11054313e-01
6.51051104e-01 5.75434506e-01 -3.28953117e-01 1.73608184e-01
1.27568674e+00 -8.87423635e-01 -1.34633705e-01 -6.04329288e-01
7.11386859e-01 -6.33905053e-01 6.95394516e-01 5.03913939e-01
-9.13002193e-01 -9.14531410e-01 -1.19466317e+00 -4.99572098e-01
-4.89988685e-01 1.50888652e-01 3.71566594e-01 5.95022202e-01
-7.69228637e-01 1.06355026e-01 -8.54359210e-01 9.24673006e-02
4.72673297e-01 4.66222346e-01 -4.43469107e-01 5.95787242e-02
-4.53812629e-01 1.17819428e+00 2.13064551e-01 9.31702107e-02
-6.48392737e-01 -8.14782619e-01 -8.78661931e-01 -1.98515773e-01
6.98243737e-01 -4.29474562e-01 1.11977899e+00 -6.88064992e-02
-1.28626192e+00 1.10331821e+00 -3.43790203e-01 -8.71958852e-01
9.85808969e-01 -6.10339105e-01 2.15954348e-01 -2.09875256e-01
1.25454694e-01 9.23320115e-01 1.06323564e+00 -1.19488823e+00
-9.84998882e-01 -7.29394317e-01 -3.28069419e-01 1.88385800e-01
4.61274832e-01 -2.37146199e-01 -2.88727164e-01 8.27952847e-02
2.27326035e-01 -1.01242459e+00 -4.20826137e-01 2.44645879e-01
-1.36920333e-01 -2.60785401e-01 1.18695939e+00 -1.54074863e-01
3.94083649e-01 -2.30493903e+00 -1.57306001e-01 -8.32744688e-02
1.42302722e-01 2.29682941e-02 3.95228684e-01 -4.79302742e-02
1.17933549e-01 -2.80083120e-01 -1.55352727e-01 -6.16715789e-01
5.02293035e-02 9.00167078e-02 -7.30558753e-01 6.60868883e-01
5.97688735e-01 8.00406814e-01 -5.81290841e-01 -5.00888586e-01
5.52040517e-01 5.91990709e-01 -3.11539173e-01 1.32493198e-01
-1.30348682e-01 2.84219801e-01 -4.63497132e-01 4.82714444e-01
9.55058753e-01 3.34338784e-01 -6.33479834e-01 7.20756799e-02
-7.16560662e-01 3.25817376e-01 -1.19893348e+00 1.48141003e+00
-5.95055699e-01 8.24777603e-01 1.34919420e-01 -6.55625820e-01
1.41537714e+00 -3.28571349e-01 5.71106791e-01 -6.26627445e-01
2.42600694e-01 1.82917133e-01 -1.51079863e-01 -5.57606295e-02
9.12691772e-01 -5.21723591e-02 -2.46562093e-01 7.05779120e-02
-5.28743804e-01 -1.01539576e+00 2.52956413e-02 -1.43526673e-01
7.98024535e-01 5.54656982e-01 1.58247128e-01 -1.01630628e-01
5.34445286e-01 7.35883892e-01 1.50985718e-01 5.32464802e-01
-1.49554461e-01 6.08420610e-01 5.32071292e-01 -6.31080925e-01
-1.40977049e+00 -9.02207196e-01 -5.55533648e-01 5.97678900e-01
2.61329055e-01 -1.76308200e-01 -4.24416751e-01 -5.30960143e-01
3.81016821e-01 8.25947881e-01 -4.23407823e-01 8.60052854e-02
-9.74826694e-01 -3.97357941e-01 9.23176706e-02 6.53304517e-01
3.57008159e-01 -6.03227258e-01 -1.66592789e+00 2.90511221e-01
3.53033960e-01 -1.42423201e+00 -1.26761803e-02 6.39331639e-01
-9.98563290e-01 -9.91776347e-01 -1.77133769e-01 -2.93583512e-01
3.16194952e-01 5.06353676e-01 1.00374234e+00 -3.21671695e-01
-8.34636807e-01 1.60798311e-01 -6.48269728e-02 -9.66873705e-01
-1.22730814e-01 -6.04492798e-02 1.84164047e-02 -3.46362472e-01
5.49893260e-01 -2.03945354e-01 -3.76212806e-01 5.45860767e-01
-2.87402481e-01 8.50730017e-02 1.87745556e-01 3.95323366e-01
9.51297402e-01 -8.63148943e-02 3.10966279e-02 -4.46070641e-01
-2.38156989e-01 -1.15588382e-01 -1.50571668e+00 -6.53586566e-01
-4.54696059e-01 -3.03468972e-01 1.72847614e-01 -1.77782148e-01
-5.44462919e-01 7.56196260e-01 -1.52653962e-01 -5.98284602e-01
-7.45789260e-02 -2.18965292e-01 1.02348648e-01 -2.75099069e-01
5.25079668e-01 1.85712501e-02 2.04112262e-01 -3.79054457e-01
3.30502242e-01 4.64122564e-01 5.68785727e-01 -3.93944889e-01
1.05494487e+00 8.94506037e-01 6.75797820e-01 -8.73751581e-01
-7.06293881e-01 -8.46150577e-01 -1.20224738e+00 -3.50772411e-01
8.48654091e-01 -1.15722001e+00 -9.36803222e-01 7.98419770e-03
-1.29283738e+00 -3.51269156e-01 -4.55401123e-01 3.37862730e-01
-7.81020403e-01 -3.05346283e-03 2.65878171e-01 -1.08916771e+00
-2.82393903e-01 -1.34204531e+00 1.56618571e+00 -7.04928488e-02
1.86497435e-01 -2.92578757e-01 -2.59348303e-02 4.35442716e-01
1.64172828e-01 4.83525962e-01 4.42090064e-01 -1.65648788e-01
-9.94118512e-01 -5.27886033e-01 -3.20170105e-01 2.01284692e-01
-4.35525358e-01 1.28931299e-01 -1.17798102e+00 -1.16019323e-01
2.48805478e-01 -1.07189648e-01 8.39466035e-01 3.16843212e-01
1.01439548e+00 5.43746233e-01 -5.35418928e-01 7.73147583e-01
1.45869434e+00 -2.25026738e-02 3.69951487e-01 4.23299849e-01
5.93228161e-01 7.66993999e-01 1.26245630e+00 3.29442799e-01
4.18441087e-01 1.03982759e+00 9.27601099e-01 -8.52819011e-02
-7.24677369e-02 -4.00233688e-03 3.55280370e-01 2.02248350e-01
1.12175994e-01 1.23352654e-01 -1.02044988e+00 3.91876996e-01
-1.62195611e+00 -5.73134363e-01 -8.14790368e-01 2.37896872e+00
1.19503044e-01 5.98929346e-01 3.24591279e-01 2.02873290e-01
4.01875436e-01 -1.31860644e-01 -6.48726940e-01 -5.49511850e-01
1.62604302e-01 1.45203114e-01 9.65303481e-01 6.16687119e-01
-1.07254863e+00 1.07815278e+00 5.65159416e+00 6.00614309e-01
-1.17304909e+00 1.64927498e-01 3.33301544e-01 -5.00630498e-01
4.09534067e-01 -9.08305794e-02 -1.65912282e+00 3.16173762e-01
9.39545035e-01 -7.71877840e-02 -6.13876842e-02 1.29796362e+00
2.56320775e-01 -3.94067019e-01 -1.06547821e+00 9.76058125e-01
2.58780271e-02 -1.07434487e+00 -4.76841956e-01 5.04602492e-01
2.46663406e-01 5.60216129e-01 -6.49010837e-02 5.02455235e-01
3.95015068e-02 -9.67320561e-01 9.72271323e-01 2.02809736e-01
9.08185363e-01 -9.04029787e-01 5.84199131e-01 7.84658611e-01
-1.19525421e+00 -2.50922590e-02 -5.98804891e-01 -2.16976449e-01
2.95776874e-01 4.92115229e-01 -1.50482488e+00 2.54315615e-01
7.37977028e-01 3.71060878e-01 -6.32283390e-01 1.13995910e+00
-4.41431776e-02 2.45932013e-01 -7.08607137e-01 -9.26150233e-02
3.64203900e-01 -3.22262913e-01 7.56485343e-01 1.00642622e+00
3.68747234e-01 -1.22058623e-01 2.19347492e-01 8.70722830e-01
2.80392557e-01 -3.53354365e-02 -9.26919580e-01 5.46747029e-01
6.09769560e-02 1.34950185e+00 -7.47378170e-01 -1.94624700e-02
-1.85886979e-01 3.35014015e-01 2.71425635e-01 -2.17744246e-01
-1.02678072e+00 -2.46274278e-01 9.06547189e-01 5.79731345e-01
6.06053233e-01 -8.53776634e-01 -5.98192036e-01 -5.79187572e-01
3.25425982e-01 -9.65929329e-02 -2.57983893e-01 -9.00055945e-01
-6.15209877e-01 5.57797432e-01 1.72099650e-01 -1.40838003e+00
-2.00731263e-01 -1.09891129e+00 -2.41545215e-01 8.57425153e-01
-1.89350879e+00 -1.00626504e+00 -6.04936421e-01 2.50031769e-01
7.90914655e-01 1.42408431e-01 4.35708731e-01 2.01662734e-01
-2.09713459e-01 2.04067919e-02 -1.52564794e-01 -2.43757308e-01
4.44478571e-01 -1.21371627e+00 9.27978873e-01 5.78464568e-01
5.30931614e-02 -3.70687246e-02 7.44487524e-01 -6.73783064e-01
-1.68997025e+00 -1.29665291e+00 6.60200536e-01 -1.01222324e+00
4.59095776e-01 -7.85548449e-01 -7.27454185e-01 5.80771804e-01
-4.20969069e-01 3.90318096e-01 9.21606421e-02 -3.95485222e-01
-1.52768135e-01 -3.78139019e-01 -1.06289244e+00 3.00611556e-01
1.01625896e+00 -4.14172709e-02 -2.85719723e-01 3.37243527e-01
8.11570942e-01 -1.08619249e+00 -4.33584929e-01 5.20153522e-01
3.29743415e-01 -9.66984510e-01 1.12244904e+00 -3.76341790e-01
1.52181268e-01 -6.60369098e-01 -1.11611441e-01 -6.10718906e-01
1.69827640e-01 -4.92997110e-01 2.22704243e-02 5.98750114e-01
2.48301089e-01 -6.00895941e-01 1.05477798e+00 2.82482356e-01
-4.62559640e-01 -6.31258428e-01 -1.43571472e+00 -7.78043091e-01
-5.22595830e-02 -1.17711091e+00 5.60615182e-01 9.37655196e-02
-6.89763725e-01 3.46328020e-01 -2.36711912e-02 3.07552516e-01
6.81059599e-01 2.93959945e-01 1.50970697e+00 -1.45419276e+00
1.44389778e-01 -3.85781527e-01 -8.91054034e-01 -1.08661950e+00
1.45522684e-01 -6.89449847e-01 2.54348248e-01 -1.02815366e+00
-3.82355005e-01 -5.89441121e-01 4.34532642e-01 8.04917812e-02
3.26902956e-01 4.65395629e-01 2.30936378e-01 -3.72322625e-03
-1.87925056e-01 5.02035320e-01 9.64467108e-01 -9.91321653e-02
-8.48124772e-02 3.89616132e-01 -2.03651767e-02 6.98963344e-01
5.64111531e-01 -6.79799616e-01 2.02139840e-03 -4.72174913e-01
3.35638940e-01 -1.41951248e-01 6.70422614e-01 -1.16768754e+00
1.93482950e-01 -8.66890922e-02 2.67593384e-01 -1.63941765e+00
1.14030337e+00 -1.08620381e+00 -2.39111289e-01 5.34180701e-01
2.40970120e-01 1.28286377e-01 5.39625168e-01 3.53906065e-01
3.08204964e-02 -2.31645942e-01 8.38492215e-01 5.18373214e-02
-6.77986026e-01 3.26663613e-01 -4.49815057e-02 -1.66214168e-01
1.36581933e+00 -7.58092105e-01 -8.33691508e-02 1.60498634e-01
-3.99578303e-01 4.50677693e-01 6.18674874e-01 4.11069721e-01
7.92944431e-01 -1.03189969e+00 -9.33198988e-01 4.80734497e-01
2.12021172e-01 7.00368822e-01 -7.51119107e-02 7.16095209e-01
-8.92295778e-01 6.06163144e-01 -1.22783586e-01 -1.26943731e+00
-1.43613148e+00 5.45963943e-01 3.59678894e-01 2.18229532e-01
-9.35596943e-01 8.83138239e-01 6.09284081e-02 -4.17230576e-01
2.23319188e-01 -4.76124108e-01 3.26087140e-02 -1.71946585e-01
4.78468657e-01 4.71178472e-01 5.46509862e-01 -7.18787014e-01
-4.29864436e-01 8.94030213e-01 6.08099177e-02 -1.34986088e-01
1.24879265e+00 8.35169014e-03 2.59794593e-01 6.25132203e-01
1.03728592e+00 1.06578313e-01 -1.53279102e+00 7.18262941e-02
-4.03869571e-03 -6.02391183e-01 3.56499016e-01 -2.16985568e-01
-6.70808613e-01 1.30678558e+00 9.43765879e-01 -6.54469952e-02
4.32222247e-01 -1.10015064e-01 6.44124508e-01 4.90352333e-01
5.37165880e-01 -8.85230482e-01 -2.18412727e-01 7.03616798e-01
8.74220252e-01 -1.43298900e+00 4.00480121e-01 -5.21584272e-01
-3.85799378e-01 1.31238019e+00 6.92457736e-01 -3.90957952e-01
3.34463775e-01 5.61282873e-01 1.08819467e-03 -2.86700159e-01
-6.12184405e-01 -3.37203264e-01 2.54750937e-01 5.52288175e-01
2.12551502e-04 8.31889138e-02 1.07839346e-01 2.77174618e-02
-6.79499447e-01 -3.35664630e-01 5.13723731e-01 1.05280900e+00
-6.85264230e-01 -7.87560642e-01 -6.74458563e-01 4.06339206e-02
4.48380969e-02 3.25987786e-01 -2.22262353e-01 1.02473056e+00
3.25475156e-01 7.38451004e-01 4.75721538e-01 -1.00186616e-01
5.65471768e-01 6.52561113e-02 4.93584067e-01 -7.41695285e-01
-3.79662454e-01 -4.05201949e-02 -7.14761615e-02 -6.04092121e-01
1.97351471e-01 -6.74457490e-01 -1.35015821e+00 1.79161970e-02
-4.29461956e-01 -1.45952895e-01 1.63059914e+00 7.73224413e-01
2.78500617e-01 2.67017454e-01 7.45417416e-01 -1.31894064e+00
-5.32384276e-01 -5.96289814e-01 -2.56714761e-01 -2.65567511e-01
6.18919611e-01 -7.38900781e-01 -3.00574452e-01 -1.68367028e-01] | [7.768596172332764, -2.6304521560668945] |
6b53273c-a9ec-49ac-bce9-ab2d0986e49a | two-hand-global-3d-pose-estimation-using | 2006.01320 | null | https://arxiv.org/abs/2006.01320v4 | https://arxiv.org/pdf/2006.01320v4.pdf | Two-hand Global 3D Pose Estimation Using Monocular RGB | We tackle the challenging task of estimating global 3D joint locations for both hands via only monocular RGB input images. We propose a novel multi-stage convolutional neural network based pipeline that accurately segments and locates the hands despite occlusion between two hands and complex background noise and estimates the 2D and 3D canonical joint locations without any depth information. Global joint locations with respect to the camera origin are computed using the hand pose estimations and the actual length of the key bone with a novel projection algorithm. To train the CNNs for this new task, we introduce a large-scale synthetic 3D hand pose dataset. We demonstrate that our system outperforms previous works on 3D canonical hand pose estimation benchmark datasets with RGB-only information. Additionally, we present the first work that achieves accurate global 3D hand tracking on both hands using RGB-only inputs and provide extensive quantitative and qualitative evaluation. | ['Fanqing Lin', 'Tony Martinez', 'Connor Wilhelm'] | 2020-06-01 | null | null | null | null | ['3d-canonical-hand-pose-estimation'] | ['computer-vision'] | [-3.98273200e-01 -3.35621625e-01 -7.84466341e-02 1.44378990e-02
-7.28725791e-01 -8.22143316e-01 2.66758949e-01 -6.40625358e-01
-8.44865501e-01 4.21730161e-01 1.28227681e-01 6.75219148e-02
2.69451499e-01 -2.09274501e-01 -6.42842412e-01 -3.41097593e-01
3.00993204e-01 1.08852410e+00 1.36899814e-01 1.29313841e-01
1.65059522e-01 1.14106393e+00 -1.17700183e+00 -2.34613195e-01
-2.25388005e-01 8.32646966e-01 -1.55949280e-01 1.09942305e+00
5.13643444e-01 3.25209916e-01 -6.23760223e-01 -4.41867620e-01
7.72311807e-01 -7.42684752e-02 -8.54512334e-01 1.83787897e-01
8.62918735e-01 -1.12692869e+00 -6.87476695e-01 3.89808476e-01
1.20509458e+00 -1.40012249e-01 4.22676861e-01 -1.13784385e+00
-9.10280868e-02 1.03414580e-02 -9.16030943e-01 -3.25035214e-01
1.00117755e+00 6.00043535e-01 8.50882828e-01 -1.05812049e+00
1.14646423e+00 1.37173414e+00 1.00385213e+00 5.32338381e-01
-1.12639177e+00 -4.15404677e-01 1.56002253e-01 -1.38226509e-01
-1.83044052e+00 7.00534582e-02 7.40763843e-01 -4.22919035e-01
1.13523662e+00 7.57043362e-02 1.22083724e+00 1.38542438e+00
8.52888301e-02 9.04461801e-01 1.15143991e+00 -6.14374816e-01
-6.31225407e-02 -5.80878973e-01 -1.67971775e-01 1.07493508e+00
-5.81540540e-02 2.46908218e-02 -7.83104777e-01 -8.05415809e-02
1.52281296e+00 3.26146990e-01 -2.21023470e-01 -6.14602327e-01
-1.59175503e+00 3.70526649e-02 7.16746628e-01 -8.68454874e-02
-6.19420826e-01 7.12623298e-01 -8.76920074e-02 -2.13651791e-01
-1.47140756e-01 9.85562801e-02 -8.37190807e-01 -3.33261997e-01
-9.25767004e-01 6.73862576e-01 7.61528492e-01 1.31152987e+00
2.99132884e-01 -4.94465590e-01 -2.87155420e-01 1.20354690e-01
4.92989570e-01 7.70233333e-01 -2.46091768e-01 -1.31820369e+00
4.02802110e-01 5.86580694e-01 4.53316867e-01 -4.81001079e-01
-6.65577233e-01 -3.40458751e-01 -2.57351726e-01 4.92721856e-01
1.16739285e+00 -1.09837063e-01 -1.23468685e+00 1.42107427e+00
4.68838096e-01 -4.68416542e-01 -7.09639549e-01 1.29017341e+00
5.72985888e-01 -1.83156505e-01 -2.07610399e-01 3.45535696e-01
1.39709306e+00 -9.42607820e-01 -4.12811846e-01 -1.26374438e-01
4.42262217e-02 -9.26784277e-01 1.19441438e+00 5.21473706e-01
-1.15015841e+00 -3.45168680e-01 -6.91137731e-01 -3.80979329e-01
-1.50011033e-01 6.21743858e-01 6.03869796e-01 6.33432925e-01
-9.77187097e-01 3.96154493e-01 -1.32913852e+00 -3.97147059e-01
3.13152194e-01 5.59920251e-01 -6.25120103e-01 -5.59593663e-02
-5.96108854e-01 9.54170704e-01 -5.36086038e-02 5.97982228e-01
-8.00741851e-01 -4.46010172e-01 -5.80477774e-01 -4.62439060e-01
4.50598687e-01 -9.70678449e-01 1.31504333e+00 -1.21735081e-01
-1.67321980e+00 1.17028856e+00 -2.51710236e-01 2.40058526e-01
1.12681413e+00 -6.94283545e-01 8.37463021e-01 2.14830846e-01
-1.36114255e-01 8.37080359e-01 7.34333694e-01 -1.14273047e+00
-1.31911725e-01 -1.10480452e+00 2.48245597e-02 1.45244882e-01
2.33122140e-01 -5.56851067e-02 -1.09112954e+00 -6.85036421e-01
5.30172586e-01 -1.28465378e+00 -1.09166034e-01 7.32393622e-01
-9.21184480e-01 -1.03832744e-01 4.64988202e-01 -1.00923908e+00
4.53013062e-01 -1.49472737e+00 6.96111023e-01 4.17021573e-01
3.58816862e-01 -1.76721364e-01 1.46094203e-01 5.00630215e-02
3.07193816e-01 -3.99064481e-01 7.24156201e-02 -8.61921251e-01
1.98176414e-01 -4.52041030e-02 -6.82017906e-03 6.48102224e-01
-2.40062907e-01 1.30181360e+00 -7.18663216e-01 -5.65663278e-01
2.78582215e-01 1.14289284e+00 -2.60502070e-01 3.82521063e-01
-1.22639783e-01 7.88641095e-01 -1.26958355e-01 1.28708625e+00
5.25247455e-01 -1.31672725e-01 2.60567904e-01 -4.95147139e-01
2.32123256e-01 -3.66897359e-02 -1.33322585e+00 2.41680694e+00
-4.60168302e-01 3.47872376e-01 7.80930743e-02 2.36810982e-01
3.45512688e-01 5.83914101e-01 2.97502458e-01 -4.92988266e-02
5.66881180e-01 3.54713380e-01 -3.17422330e-01 2.65627932e-02
-1.86919083e-03 -6.80745319e-02 2.31003404e-01 8.41023564e-01
2.26409927e-01 -3.86745960e-01 -5.17028153e-01 -3.22834030e-02
9.12981987e-01 8.70324969e-01 1.94587499e-01 3.46650690e-01
-8.01218301e-03 -2.76520938e-01 1.69715043e-02 5.79192102e-01
-2.50815541e-01 1.18118811e+00 5.38851738e-01 -7.38155544e-01
-1.15702248e+00 -1.36571801e+00 2.30808064e-01 8.69495988e-01
-1.19020410e-01 -2.81122595e-01 -8.90976131e-01 -7.27483273e-01
4.55676764e-01 -1.13065019e-01 -8.53018403e-01 5.69100916e-01
-8.00258815e-01 -1.08450465e-01 6.32761598e-01 1.21515691e+00
3.05365562e-01 -9.28725064e-01 -9.95313823e-01 -1.70593947e-01
-1.36278858e-02 -1.25853801e+00 -6.59817696e-01 8.21435973e-02
-6.59092367e-01 -1.38407183e+00 -1.42569709e+00 -6.74715519e-01
7.38110065e-01 -1.41488478e-01 8.38730812e-01 -7.49948919e-02
-8.05144072e-01 6.29219830e-01 2.30882820e-02 -1.57631546e-01
3.98454905e-01 3.61249804e-01 2.05449760e-01 -5.70224941e-01
2.10056886e-01 -4.73678559e-01 -1.04916739e+00 4.70705450e-01
9.43637341e-02 -1.11224011e-01 4.54772651e-01 4.43866521e-01
6.40270531e-01 -7.82644689e-01 -2.55233079e-01 -8.89505595e-02
3.82453173e-01 6.11247420e-01 -5.54851115e-01 3.09524685e-01
3.92118655e-02 -3.90426405e-02 1.17323332e-01 -4.35294449e-01
-8.96467745e-01 8.64358187e-01 -3.08695614e-01 -6.74352586e-01
-2.72512794e-01 -3.30000103e-01 -1.28625765e-01 -3.12405974e-01
6.65334821e-01 -1.10973865e-02 -1.20673120e-01 -4.82158065e-01
6.06438816e-01 5.61916351e-01 7.57347584e-01 -7.20662057e-01
6.71937048e-01 9.27905262e-01 3.22023332e-01 -3.91848117e-01
-7.09022522e-01 -2.86166161e-01 -1.82968748e+00 -3.19843173e-01
9.18553591e-01 -8.59276652e-01 -1.61523175e+00 9.84336793e-01
-1.61938679e+00 -4.38617915e-01 -8.77251178e-02 5.75766444e-01
-8.00473154e-01 3.06485564e-01 -7.77324498e-01 -9.96115625e-01
-5.85451126e-01 -1.40903163e+00 1.96767998e+00 -2.43861318e-01
-6.49586976e-01 -6.66153133e-01 6.08804896e-02 1.58818707e-01
-1.29522204e-01 5.41985095e-01 2.36685112e-01 1.12425700e-01
-8.38438034e-01 -8.33977401e-01 -6.23085313e-02 -4.70761731e-02
2.02553179e-02 5.46734259e-02 -1.20839190e+00 -4.96234953e-01
-5.05898952e-01 -4.46645528e-01 4.46150273e-01 5.49350619e-01
8.40230286e-01 1.36918992e-01 -3.05706233e-01 7.94860840e-01
1.00151968e+00 -3.63914222e-01 1.92023650e-01 2.68520087e-01
1.10995412e+00 5.35777748e-01 3.64879876e-01 4.76347685e-01
3.17310512e-01 8.00668001e-01 3.61207992e-01 1.40264973e-01
-3.76227349e-01 -2.23242685e-01 -1.68461487e-01 3.02592099e-01
-8.91004980e-01 2.24685416e-01 -1.04741478e+00 3.33928049e-01
-1.44075096e+00 -1.91754833e-01 1.81743518e-01 2.17533875e+00
8.81718755e-01 1.61003605e-01 6.40780330e-01 2.92375177e-01
3.67855817e-01 -1.69285029e-01 -7.62374103e-01 4.35826570e-01
1.78179085e-01 6.70038402e-01 5.19478083e-01 6.56974256e-01
-8.63180459e-01 1.15662444e+00 6.74667740e+00 5.00763254e-03
-9.02157724e-01 1.44750834e-01 -1.61202952e-01 -7.02692688e-01
2.72658020e-01 -2.25567877e-01 -7.36033440e-01 -2.20655307e-01
7.62415072e-03 8.19519401e-01 4.30611521e-01 7.38200188e-01
-3.04668218e-01 -1.49311885e-01 -1.37862277e+00 1.31778061e+00
3.38356607e-02 -7.65078604e-01 -2.53976017e-01 2.21519992e-01
5.64458311e-01 -1.73867330e-01 8.56365636e-02 -4.87146288e-01
1.18599184e-01 -9.79672432e-01 1.05091369e+00 7.00172126e-01
1.03748226e+00 -5.59484720e-01 5.77572584e-01 2.14280814e-01
-1.33974946e+00 2.43317619e-01 8.37213695e-02 -1.97525412e-01
2.12097749e-01 -4.85100895e-02 -9.37566996e-01 -1.28810322e-02
7.47518778e-01 5.15299559e-01 -6.32546723e-01 7.29192376e-01
-8.54246855e-01 -2.56171703e-01 -5.97881079e-01 4.21596617e-02
-1.87568069e-01 3.29180628e-01 4.59449649e-01 9.56570387e-01
-6.30198345e-02 1.01873362e-02 -1.54040262e-01 1.03468525e+00
-8.80560353e-02 -3.51374835e-01 -2.24132046e-01 2.14519739e-01
3.78624946e-01 1.12852085e+00 -1.05459583e+00 7.20084012e-02
1.16080724e-01 1.56548107e+00 2.40493476e-01 4.73469973e-01
-5.41843474e-01 -5.87582886e-01 5.89740872e-01 9.86880735e-02
1.86696842e-01 -9.54712749e-01 -4.59124476e-01 -1.17395413e+00
5.60782194e-01 -4.33401018e-01 -2.72546057e-02 -1.03328288e+00
-9.06408966e-01 4.48387921e-01 -2.72112459e-01 -7.34989464e-01
-3.97184134e-01 -1.13307250e+00 -1.19713947e-01 1.22145057e+00
-9.49631512e-01 -1.72671628e+00 -7.76468456e-01 8.50912511e-01
1.43882260e-01 1.29841521e-01 9.49497998e-01 -1.94268852e-01
-1.86325178e-01 7.61243999e-01 -4.76579279e-01 5.58416307e-01
8.81183863e-01 -1.53185892e+00 1.04216635e+00 4.08718050e-01
6.65151849e-02 1.09774053e+00 3.22573185e-01 -6.35221660e-01
-1.83002400e+00 -4.27407771e-01 4.87527460e-01 -1.52877355e+00
1.14405721e-01 -7.08379507e-01 -9.88509431e-02 1.17917609e+00
-2.13397399e-01 1.76119953e-01 2.47190058e-01 2.09056318e-01
-6.91951275e-01 1.53655574e-01 -1.28611434e+00 4.55187857e-01
1.50546622e+00 -8.85505199e-01 -5.75177550e-01 4.95607764e-01
3.77526224e-01 -1.05152881e+00 -1.00930595e+00 -7.08120614e-02
1.64059150e+00 -8.74046385e-01 1.43700671e+00 -5.82985342e-01
6.87643364e-02 -3.73057991e-01 -1.58964291e-01 -7.88787901e-01
1.64312050e-01 -5.14144003e-01 -4.68467027e-01 4.83589053e-01
-1.38590500e-01 -6.78589195e-02 1.40391731e+00 7.32153833e-01
7.32009053e-01 -7.56301880e-01 -1.07661605e+00 -4.76290166e-01
-1.89326018e-01 -5.28101802e-01 3.05729836e-01 2.22398743e-01
-3.12363386e-01 -1.12774864e-01 -3.03951472e-01 1.95036143e-01
1.06113112e+00 2.74665445e-01 1.17668056e+00 -1.13631451e+00
-6.64879307e-02 -2.47740597e-01 -6.38187468e-01 -1.21383202e+00
1.35449573e-01 -4.64070648e-01 8.04183483e-02 -1.50620353e+00
2.61472851e-01 1.02542169e-01 2.65426606e-01 6.05679512e-01
-7.02549890e-03 6.44427419e-01 2.61495531e-01 3.13923746e-01
-3.33154112e-01 9.10557061e-02 1.52296996e+00 2.40741625e-01
-2.11885959e-01 1.71283901e-01 6.88638398e-03 8.30015361e-01
3.00631970e-01 -1.78251520e-01 1.96527466e-01 -5.24032772e-01
2.97952265e-01 6.80203503e-03 9.82903719e-01 -8.31204891e-01
3.03373992e-01 1.44858971e-01 1.16199505e+00 -1.05801880e+00
5.94622433e-01 -8.70529532e-01 4.73265462e-02 7.09108233e-01
-1.83638066e-01 1.44413590e-01 -8.45651999e-02 3.07371765e-01
5.73176682e-01 3.55831683e-01 6.79383874e-01 -5.97271442e-01
-3.14845949e-01 4.43168938e-01 9.66583490e-02 -2.57785290e-01
9.50105369e-01 -2.99507171e-01 2.26516709e-01 -3.27297032e-01
-1.22967899e+00 -6.63644373e-02 7.38001704e-01 2.03380570e-01
8.03281426e-01 -1.29158342e+00 -3.66157889e-01 3.23338687e-01
-1.62126139e-01 4.41469789e-01 -8.65512639e-02 8.81768525e-01
-1.15031087e+00 7.06049979e-01 -3.98260206e-01 -9.68769848e-01
-1.38890243e+00 2.84625590e-01 7.43158996e-01 3.14594388e-01
-5.12955904e-01 1.20795882e+00 -4.59720403e-01 -1.01686263e+00
7.58868635e-01 -5.53881764e-01 7.56548107e-01 -1.57088161e-01
4.07927275e-01 7.48534799e-01 1.06689699e-01 -7.09196270e-01
-6.83323085e-01 1.14077294e+00 2.25291431e-01 -4.28342909e-01
1.24584544e+00 8.47130790e-02 -1.50240362e-01 2.17994124e-01
1.26641035e+00 2.07301183e-03 -1.75004685e+00 -2.58558422e-01
-4.77565974e-01 -9.17771995e-01 -2.52507299e-01 -1.01095629e+00
-1.10711849e+00 1.39229274e+00 7.77619958e-01 -7.62766957e-01
6.34111345e-01 3.52018118e-01 6.94311202e-01 6.93011343e-01
6.41014278e-01 -9.13368881e-01 3.19752723e-01 4.04591531e-01
1.10156929e+00 -9.62960541e-01 2.64578640e-01 -3.48860264e-01
-2.18917608e-01 1.28577459e+00 7.75888383e-01 -6.90203607e-02
6.30330384e-01 3.96784037e-01 2.47067615e-01 -2.70551383e-01
1.14548959e-01 -3.38512570e-01 4.51515973e-01 7.10138381e-01
4.94989425e-01 1.29036009e-01 4.01698053e-01 2.70324111e-01
-3.45377445e-01 1.54771045e-01 -1.20395184e-01 1.47329509e+00
1.55908376e-01 -1.06684256e+00 -7.27472484e-01 2.09263712e-02
-3.26254815e-01 3.16075355e-01 -9.32140648e-01 1.02150989e+00
1.01775564e-01 1.42911270e-01 -8.79069716e-02 -4.98989642e-01
8.37211311e-01 2.46404856e-01 1.53944886e+00 -5.24098396e-01
-4.92099404e-01 3.09012324e-01 -5.96106172e-01 -8.70554090e-01
-3.98798466e-01 -4.86722469e-01 -1.13488197e+00 -3.81433249e-01
-2.82203794e-01 -7.96115696e-01 8.45403314e-01 9.11957502e-01
1.42129019e-01 2.98161298e-01 3.23313065e-02 -1.91295993e+00
-6.56174123e-01 -9.06707585e-01 -8.43082249e-01 2.43353983e-03
5.03180861e-01 -1.00913739e+00 -6.43831640e-02 -9.67191681e-02] | [6.579189777374268, -0.7757436633110046] |
6dffe18d-773a-42e2-bdcc-a5562e644eb9 | sportscap-monocular-3d-human-motion-capture | 2104.11452 | null | https://arxiv.org/abs/2104.11452v4 | https://arxiv.org/pdf/2104.11452v4.pdf | SportsCap: Monocular 3D Human Motion Capture and Fine-grained Understanding in Challenging Sports Videos | Markerless motion capture and understanding of professional non-daily human movements is an important yet unsolved task, which suffers from complex motion patterns and severe self-occlusion, especially for the monocular setting. In this paper, we propose SportsCap -- the first approach for simultaneously capturing 3D human motions and understanding fine-grained actions from monocular challenging sports video input. Our approach utilizes the semantic and temporally structured sub-motion prior in the embedding space for motion capture and understanding in a data-driven multi-task manner. To enable robust capture under complex motion patterns, we propose an effective motion embedding module to recover both the implicit motion embedding and explicit 3D motion details via a corresponding mapping function as well as a sub-motion classifier. Based on such hybrid motion information, we introduce a multi-stream spatial-temporal Graph Convolutional Network(ST-GCN) to predict the fine-grained semantic action attributes, and adopt a semantic attribute mapping block to assemble various correlated action attributes into a high-level action label for the overall detailed understanding of the whole sequence, so as to enable various applications like action assessment or motion scoring. Comprehensive experiments on both public and our proposed datasets show that with a challenging monocular sports video input, our novel approach not only significantly improves the accuracy of 3D human motion capture, but also recovers accurate fine-grained semantic action attributes. | ['Jingyi Yu', 'Lan Xu', 'Yuexin Ma', 'Wei Yang', 'Anqi Pang', 'Xin Chen'] | 2021-04-23 | null | null | null | null | ['markerless-motion-capture', 'action-assessment'] | ['computer-vision', 'computer-vision'] | [ 5.58403581e-02 -5.63761830e-01 -5.67539692e-01 -1.44989923e-01
-7.27414489e-01 -5.46306729e-01 2.78166294e-01 -4.82949972e-01
-3.20115060e-01 3.20693105e-01 8.59729409e-01 2.49358878e-01
-1.00141749e-01 -4.84707087e-01 -6.93481505e-01 -6.13098681e-01
8.51841867e-02 3.19789827e-01 4.85156685e-01 -1.20454662e-01
7.94023499e-02 2.08879799e-01 -1.60051227e+00 4.60073888e-01
4.63591218e-01 8.55619371e-01 1.86278701e-01 9.16211963e-01
1.54933080e-01 1.06582606e+00 -2.73710698e-01 -3.96419130e-02
3.25592101e-01 -3.74496639e-01 -7.68165410e-01 3.61848116e-01
5.61508238e-01 -5.15095055e-01 -9.15672362e-01 6.34388506e-01
5.85754395e-01 3.64199579e-01 4.06746477e-01 -1.31276548e+00
-2.55073369e-01 3.48094031e-02 -4.56809908e-01 3.67384911e-01
7.49114573e-01 8.48058641e-01 8.11103284e-01 -6.14733756e-01
8.88303161e-01 1.34557831e+00 5.02030730e-01 6.27981663e-01
-7.51057625e-01 -5.72760224e-01 2.77455360e-01 7.35003829e-01
-1.13916981e+00 -2.52489537e-01 9.32355642e-01 -5.87256968e-01
7.84925818e-01 1.43838495e-01 1.27542996e+00 1.39591312e+00
1.88903049e-01 1.25779498e+00 7.33170688e-01 2.12051883e-01
-7.01344609e-02 -7.49389291e-01 -2.68147532e-02 7.51605809e-01
5.50253727e-02 1.64316222e-01 -8.63099158e-01 1.91640303e-01
1.16953397e+00 2.74962366e-01 -3.34804416e-01 -5.68954170e-01
-1.68207884e+00 4.62059975e-01 2.90289104e-01 8.57598484e-02
-5.54785311e-01 6.18602872e-01 4.34441447e-01 -3.80343832e-02
4.11494017e-01 -7.56527185e-02 -2.74157315e-01 -8.03843617e-01
-7.26682723e-01 6.14246666e-01 3.19635242e-01 8.55376840e-01
6.57989025e-01 6.40876815e-02 -5.09192050e-01 4.84749407e-01
9.31357071e-02 5.59334993e-01 3.73447746e-01 -1.35701191e+00
8.26224625e-01 8.07823718e-01 1.33517772e-01 -1.24207044e+00
-4.01152045e-01 -3.26967210e-01 -6.69076145e-01 3.10612060e-02
4.31471139e-01 2.26596519e-01 -8.72078419e-01 1.75148392e+00
6.69307709e-01 7.07682848e-01 -3.17006171e-01 1.44310343e+00
7.24706173e-01 4.42692012e-01 3.49970788e-01 2.12621599e-01
1.49576521e+00 -1.12104452e+00 -6.42420828e-01 -5.10517240e-01
8.40470731e-01 -3.36369932e-01 1.16008210e+00 8.46690312e-02
-9.98469293e-01 -8.29354644e-01 -9.40080762e-01 -4.11941677e-01
1.51286244e-01 1.32433757e-01 5.04486501e-01 2.00889871e-01
-5.45250773e-01 3.71786863e-01 -1.14086688e+00 -5.15654087e-02
5.16129255e-01 1.43654823e-01 -7.81239688e-01 -5.67565560e-01
-1.20647728e+00 6.64132953e-01 3.17561567e-01 2.51903623e-01
-1.00525284e+00 -7.18259513e-01 -1.28848684e+00 -2.31569871e-01
5.82187235e-01 -1.26300204e+00 8.20759475e-01 -5.46332538e-01
-1.30489743e+00 8.85342836e-01 -2.14751780e-01 -1.21593937e-01
6.97128356e-01 -7.25282311e-01 -9.87316817e-02 5.34160674e-01
3.26961070e-01 6.34747446e-01 6.50418639e-01 -8.25529695e-01
-6.99867666e-01 -5.55898488e-01 1.63570881e-01 7.17090786e-01
-1.38492376e-01 -3.56617540e-01 -9.39942718e-01 -9.35506403e-01
1.00867361e-01 -1.09576631e+00 -2.19414830e-01 2.13714108e-01
-1.78437442e-01 -3.45290415e-02 6.80698693e-01 -9.05596852e-01
1.25057912e+00 -2.06160831e+00 7.39408016e-01 -2.94991791e-01
2.81121582e-01 1.59968480e-01 -1.37332112e-01 1.25501320e-01
1.44068480e-01 -4.17198092e-01 -2.96976388e-01 -2.24388033e-01
9.47439671e-03 4.52131152e-01 5.73926270e-02 5.05833805e-01
2.45219678e-01 1.38730562e+00 -1.33758366e+00 -6.29798651e-01
5.84039867e-01 4.94173557e-01 -6.54726148e-01 3.92252535e-01
-4.38855924e-02 8.99948537e-01 -8.83538485e-01 8.68793547e-01
2.77523398e-01 -1.73338532e-01 -2.63526320e-01 -4.06866014e-01
3.13336730e-01 -2.47578397e-02 -1.38215458e+00 2.59643769e+00
-2.29984194e-01 4.13428754e-01 -1.71254769e-01 -9.35891330e-01
6.11574054e-01 1.76307037e-01 1.00863397e+00 -6.59365118e-01
-5.26047088e-02 8.49884823e-02 -3.38273585e-01 -8.92545938e-01
4.24497008e-01 -1.31500736e-02 -3.95560831e-01 4.08767968e-01
-5.55775240e-02 2.34335244e-01 -9.47199315e-02 1.26755491e-01
1.29174852e+00 8.54730427e-01 1.20268315e-01 1.63666025e-01
6.28779352e-01 3.32243174e-01 9.88213301e-01 2.94603258e-01
-5.29064536e-01 8.93973887e-01 1.71116829e-01 -7.48160839e-01
-1.08064210e+00 -9.90424454e-01 6.38367474e-01 8.19685638e-01
6.00698113e-01 -2.56736904e-01 -6.13901079e-01 -6.50769293e-01
7.42904097e-02 6.43796772e-02 -5.70737541e-01 -5.53614438e-01
-1.18144655e+00 -4.55483735e-01 4.85647559e-01 1.02128339e+00
6.85666919e-01 -1.17399132e+00 -7.14644790e-01 3.79645795e-01
-8.33852589e-01 -1.57378888e+00 -9.65667844e-01 -5.49763203e-01
-7.09494293e-01 -1.24844432e+00 -9.31093395e-01 -5.65020382e-01
1.64054856e-01 4.98248249e-01 1.02412724e+00 5.94680868e-02
-4.35232341e-01 5.00342548e-01 -4.82701242e-01 3.11078310e-01
3.17575261e-02 -9.11895558e-02 -1.43205356e-02 3.65513295e-01
5.26559055e-01 -5.40254354e-01 -1.21008372e+00 4.43321735e-01
-8.80773783e-01 2.85728872e-01 6.22708917e-01 7.43315935e-01
6.34107172e-01 -4.83775049e-01 1.84504718e-01 -4.65402693e-01
1.56446099e-02 -5.39512992e-01 7.92223364e-02 6.49703890e-02
4.95303534e-02 -7.62437135e-02 9.58206505e-02 -5.93184531e-01
-8.33799958e-01 3.71559113e-01 -1.08943954e-01 -1.01484025e+00
-2.81439722e-01 1.70283690e-01 -5.13175845e-01 1.50957480e-01
3.85958642e-01 5.01156390e-01 -3.56855430e-02 -4.96194929e-01
5.50782323e-01 3.49462092e-01 1.10680735e+00 -5.54790378e-01
8.39858234e-01 8.72811377e-01 2.05031008e-01 -4.05526817e-01
-9.23201561e-01 -8.95929754e-01 -1.03388441e+00 -5.44380307e-01
1.49177516e+00 -1.39540005e+00 -7.82795846e-01 7.31302440e-01
-9.36248422e-01 -3.98995101e-01 -1.50079578e-01 8.65987003e-01
-1.09282005e+00 6.47172511e-01 -7.75546908e-01 -4.01892543e-01
-1.47403449e-01 -1.16764057e+00 1.69339907e+00 -7.07387105e-02
-2.82782018e-01 -7.83675671e-01 2.24203900e-01 9.27301526e-01
-2.63075680e-01 8.61029565e-01 5.20533264e-01 4.33029346e-02
-9.67720628e-01 -1.86638892e-01 2.02392805e-02 1.06421322e-01
4.79492620e-02 -5.99513412e-01 -5.52826047e-01 -1.78966269e-01
-2.55716681e-01 -2.96924829e-01 9.10047710e-01 5.14445364e-01
1.03005028e+00 7.69533515e-02 -2.57807910e-01 1.07939029e+00
1.04361522e+00 -3.04530054e-01 6.56561434e-01 4.80606467e-01
1.47802222e+00 6.71663046e-01 1.23816597e+00 4.82346088e-01
7.71939278e-01 9.76982653e-01 3.59568268e-01 2.17019506e-02
-4.32373911e-01 -5.42405665e-01 4.60041434e-01 8.98994565e-01
-3.22555751e-01 8.87920558e-02 -7.60419071e-01 7.41598904e-01
-2.17050433e+00 -1.20407689e+00 -2.29426831e-01 1.92808354e+00
6.01976931e-01 -9.09361541e-02 4.64370131e-01 1.96506917e-01
6.43270731e-01 5.67033291e-01 -7.32052207e-01 3.51735502e-01
-1.66637763e-01 -1.30029097e-01 3.52317661e-01 3.24483782e-01
-1.17236793e+00 1.00061548e+00 4.86506653e+00 8.10814083e-01
-8.18349898e-01 1.85598850e-01 1.50881976e-01 -4.64253634e-01
-2.74987429e-01 -3.72571424e-02 -5.21496892e-01 4.98194218e-01
4.20076638e-01 6.56118467e-02 -6.34983256e-02 5.45737445e-01
4.57492560e-01 1.30570382e-01 -9.78378475e-01 1.20717192e+00
1.10794872e-01 -1.33856738e+00 2.62169063e-01 2.32114106e-01
6.19998276e-01 -3.64503056e-01 -2.36098319e-01 1.87201962e-01
9.89595205e-02 -9.36418235e-01 8.60026896e-01 8.30203950e-01
8.01662266e-01 -6.16473973e-01 4.03416842e-01 5.20570695e-01
-1.72619283e+00 -1.71930075e-01 -1.24912195e-01 -2.18421996e-01
6.31358564e-01 1.81360617e-01 -1.75915454e-02 8.44385266e-01
6.80136800e-01 1.52633095e+00 -2.88014352e-01 9.19248760e-01
-1.28766149e-01 2.42002919e-01 -1.35133890e-02 3.13678652e-01
4.00540143e-01 -7.15874434e-02 7.62807190e-01 9.40062582e-01
2.71180004e-01 5.31498611e-01 4.97268617e-01 4.84945089e-01
3.74642164e-01 -1.07852742e-01 -4.56075907e-01 2.00314522e-02
2.48486295e-01 1.05517912e+00 -4.48136747e-01 -4.24325168e-01
-4.20644403e-01 1.31433022e+00 2.55270481e-01 3.28027934e-01
-8.80810618e-01 1.32390290e-01 1.35121644e+00 1.96886078e-01
2.50207543e-01 -6.05451167e-01 -1.66271627e-01 -1.63963342e+00
3.90522778e-01 -8.34953547e-01 6.09413028e-01 -7.90912628e-01
-8.82323444e-01 2.56391525e-01 5.46320491e-02 -1.82585418e+00
-4.76523221e-01 -5.09108484e-01 -4.19312119e-01 6.99179828e-01
-1.35607946e+00 -1.51255858e+00 -8.58798802e-01 1.00983191e+00
8.18196595e-01 1.16107278e-02 3.46283287e-01 5.39387465e-01
-4.66395229e-01 2.14303851e-01 -5.47447860e-01 2.80542254e-01
6.81601703e-01 -9.89370584e-01 5.15365839e-01 7.82221615e-01
2.00875312e-01 1.07612573e-02 4.97062773e-01 -8.41953754e-01
-1.69690406e+00 -1.28502607e+00 5.55264473e-01 -1.06616521e+00
4.19451535e-01 -1.95947036e-01 -9.32678878e-01 6.54427648e-01
-4.90596145e-01 2.68589258e-01 5.10473192e-01 -5.10023236e-01
-5.75131364e-02 9.07964930e-02 -4.78337228e-01 5.58420420e-01
1.94166696e+00 -5.38792133e-01 -7.29183376e-01 2.08959252e-01
6.49865270e-01 -6.49322629e-01 -1.02275491e+00 5.49710572e-01
7.74013698e-01 -7.71667242e-01 1.48015881e+00 -9.36721206e-01
7.73974180e-01 -4.98491734e-01 -1.89380735e-01 -9.51886714e-01
-4.20619041e-01 -4.84466761e-01 -4.57831115e-01 7.77478516e-01
-4.29499418e-01 1.41589478e-01 1.20554793e+00 4.11096811e-01
-3.63695085e-01 -7.61086524e-01 -9.49619293e-01 -5.87407291e-01
-2.73523360e-01 -7.61505067e-01 4.37808007e-01 8.34564686e-01
-3.34807515e-01 -6.67588273e-03 -9.54005420e-01 7.16069564e-02
8.13534617e-01 3.10001552e-01 1.28898728e+00 -8.52818787e-01
-3.71069700e-01 -1.35681421e-01 -1.10782027e+00 -1.60299122e+00
3.38449240e-01 -6.72628939e-01 -2.73861792e-02 -1.51497948e+00
1.73151731e-01 1.10952035e-02 -2.51877487e-01 2.06116065e-01
-6.03547871e-01 5.29001296e-01 2.83144653e-01 4.33007240e-01
-6.92532361e-01 8.35412681e-01 1.70727253e+00 -7.68380761e-02
-3.29444920e-05 -1.32861570e-01 -2.21387744e-01 7.28738964e-01
1.30940408e-01 -4.52961683e-01 -5.50281584e-01 -6.55447245e-01
-9.37792286e-02 4.84590918e-01 9.26656425e-01 -1.14723456e+00
1.87514961e-01 -5.04379332e-01 3.97729427e-01 -7.48703420e-01
7.41132617e-01 -5.78236103e-01 3.04721087e-01 5.47212005e-01
-1.29711866e-01 9.42177773e-02 -1.65706724e-01 9.89755511e-01
-4.18258071e-01 4.65620130e-01 3.10698271e-01 -3.14498633e-01
-1.51888704e+00 9.88097906e-01 9.06254351e-02 3.50313425e-01
1.19064879e+00 -6.76072180e-01 -4.84676026e-02 -4.59147573e-01
-1.11162901e+00 5.69476247e-01 5.50816536e-01 9.06968832e-01
8.48202407e-01 -1.82296705e+00 -8.73751104e-01 1.54614359e-01
4.19355541e-01 -5.88533515e-03 8.77096951e-01 9.24268723e-01
-6.84920490e-01 2.30843112e-01 -6.62972808e-01 -9.52092886e-01
-1.10644746e+00 2.80947238e-01 1.63753092e-01 -2.22026378e-01
-1.17059731e+00 5.72029591e-01 5.10302067e-01 -1.10168166e-01
-8.90494287e-02 -1.82793826e-01 -3.20251673e-01 -2.36002550e-01
4.87907499e-01 5.11177659e-01 -4.70840633e-01 -1.17118478e+00
-3.93513232e-01 1.21674180e+00 5.08153975e-01 -1.40920714e-01
1.16165495e+00 -4.57237214e-01 4.98181641e-01 5.46436727e-01
1.37136090e+00 -4.46584404e-01 -1.98474622e+00 -4.30573374e-01
-2.20891118e-01 -1.05023468e+00 -2.11154491e-01 -2.38822967e-01
-1.04448545e+00 1.21232581e+00 2.78361350e-01 -6.04789972e-01
1.10953331e+00 1.28246099e-02 1.33607125e+00 -8.88336375e-02
4.74718094e-01 -9.84325588e-01 4.83759671e-01 1.83435887e-01
7.21375823e-01 -1.12285137e+00 -3.04687116e-02 -6.01561368e-01
-1.01100886e+00 8.74615073e-01 8.04068267e-01 -2.96805382e-01
4.09815758e-01 -3.06394666e-01 1.64651141e-01 -3.17675918e-01
-7.54950583e-01 -3.59090894e-01 6.32427275e-01 6.62823975e-01
-3.09509188e-02 -2.68825889e-02 -1.47501737e-01 7.02091813e-01
-1.39220199e-02 2.23844856e-01 1.81045637e-01 8.70959461e-01
-2.23575294e-01 -8.88424575e-01 -1.47858873e-01 7.14117289e-02
-2.53036320e-01 3.89376402e-01 -1.86144099e-01 8.38224769e-01
2.42254689e-01 5.80932260e-01 7.72626251e-02 -7.54179537e-01
6.56944096e-01 -3.07936855e-02 4.61037427e-01 -5.53604186e-01
-4.32392508e-01 5.85415214e-02 1.71799287e-01 -1.24529314e+00
-6.45469129e-01 -7.54510581e-01 -1.17098510e+00 -2.30321795e-01
4.55874413e-01 -3.11511040e-01 1.78793624e-01 1.13441241e+00
3.96878600e-01 5.90375483e-01 3.51310551e-01 -1.09449291e+00
-4.44393158e-01 -7.52621710e-01 -5.50356925e-01 1.17201364e+00
3.65569085e-01 -1.12650573e+00 -4.82348353e-02 3.04322511e-01] | [7.295968055725098, -0.4924871325492859] |
c04d3266-0bd1-4e7f-8fd2-5e9481324f91 | adviser-networks-learning-what-question-to | 1802.01666 | null | http://arxiv.org/abs/1802.01666v3 | http://arxiv.org/pdf/1802.01666v3.pdf | Adviser Networks: Learning What Question to Ask for Human-In-The-Loop Viewpoint Estimation | Humans have an unparalleled visual intelligence and can overcome visual
ambiguities that machines currently cannot. Recent works have shown that
incorporating guidance from humans during inference for monocular
viewpoint-estimation can help overcome difficult cases in which the
computer-alone would have otherwise failed. These hybrid intelligence
approaches are hence gaining traction. However, deciding what question to ask
the human at inference time remains an unknown for these problems.
We address this question by formulating it as an Adviser Problem: can we
learn a mapping from the input to a specific question to ask the human to
maximize the expected positive impact to the overall task? We formulate a
solution to the adviser problem for viewpoint estimation using a deep network
where the question asks for the location of a keypoint in the input image. We
show that by using the Adviser Network's recommendations, the model and the
human outperforms the previous hybrid-intelligence state-of-the-art by 3.7%,
and the computer-only state-of-the-art by 5.28% absolute. | ['Mohamed El Banani', 'Jason J. Corso'] | 2018-02-05 | null | null | null | null | ['viewpoint-estimation'] | ['computer-vision'] | [ 1.01373784e-01 7.23256111e-01 1.80859968e-01 -5.80798626e-01
-3.76871943e-01 -6.31807327e-01 5.57841063e-01 -3.04136515e-01
-5.67101300e-01 6.17159128e-01 8.17524865e-02 -4.64617789e-01
-2.66786128e-01 -4.60321367e-01 -6.71190858e-01 -3.55062395e-01
6.13617063e-01 1.06454182e+00 1.50023788e-01 -1.71683326e-01
6.82987034e-01 3.91026646e-01 -1.63056600e+00 3.45708191e-01
8.07643652e-01 9.05473769e-01 3.40151101e-01 9.06097412e-01
2.05848545e-01 1.12659323e+00 -5.91871083e-01 -3.95319819e-01
5.31170547e-01 -2.62130439e-01 -1.02136612e+00 8.35852027e-02
8.79476607e-01 -7.14285254e-01 -1.59065321e-01 1.07786393e+00
2.14851782e-01 2.91374713e-01 6.65542126e-01 -1.30797946e+00
-7.24183738e-01 -1.25129288e-02 -3.14257860e-01 3.17368418e-01
6.06840789e-01 3.49141568e-01 9.57675397e-01 -7.76465535e-01
8.27142000e-01 1.32440102e+00 1.01009749e-01 4.87104684e-01
-9.01421845e-01 -4.12360542e-02 5.25271535e-01 6.56249881e-01
-1.15583658e+00 -3.65736336e-01 5.85371196e-01 -5.27765512e-01
1.23912585e+00 2.48641059e-01 4.20779556e-01 6.97668552e-01
-1.01478919e-01 8.45869482e-01 1.12080395e+00 -3.14416349e-01
1.47760406e-01 4.25389946e-01 -2.23442301e-01 8.06130648e-01
2.34317079e-01 2.57528722e-01 -6.79884493e-01 2.93493241e-01
9.97209430e-01 -1.79498866e-01 -2.57993966e-01 -4.42191601e-01
-1.30880105e+00 6.98039830e-01 8.14475417e-01 4.89551350e-02
-6.42402053e-01 -1.58808790e-02 -2.76192009e-01 4.39731359e-01
2.88851291e-01 1.27716064e+00 -5.12263715e-01 1.12644963e-01
-7.26581514e-01 5.05733490e-01 1.09186459e+00 6.26951873e-01
4.88392293e-01 -1.89744845e-01 3.07823300e-01 2.45001152e-01
4.08128351e-01 2.68901736e-01 -2.13031352e-01 -1.73295891e+00
2.72584319e-01 7.92302787e-01 6.70618713e-01 -1.25078106e+00
-5.39764762e-01 -4.95614350e-01 -5.31146467e-01 8.12763214e-01
8.93527269e-01 -2.49538943e-01 -7.73779392e-01 1.44620204e+00
4.40584809e-01 -3.28500897e-01 -1.13756366e-01 1.43154347e+00
5.10698318e-01 5.43958068e-01 -3.89696538e-01 1.41757429e-01
1.27576649e+00 -1.22947919e+00 -4.04729337e-01 -8.41654241e-01
1.79795086e-01 -7.24752069e-01 7.42335141e-01 9.34640884e-01
-1.21945167e+00 -8.24813426e-01 -1.13169682e+00 -3.87040704e-01
-3.12139779e-01 2.50971287e-01 6.16787851e-01 6.56848848e-02
-1.49963880e+00 4.98449028e-01 -4.79906261e-01 -5.68715692e-01
6.81716055e-02 4.54804152e-01 -2.67641068e-01 -1.98301300e-01
-9.26065862e-01 1.49553609e+00 9.14258882e-02 7.58467689e-02
-8.20613027e-01 -3.58235061e-01 -5.69527507e-01 1.03208341e-01
1.09680116e+00 -1.30624807e+00 1.50443995e+00 -1.17405045e+00
-1.31701636e+00 9.98872757e-01 -2.25522593e-01 -5.28100550e-01
7.12760627e-01 -4.80296552e-01 2.03825995e-01 4.51280326e-01
2.37137705e-01 1.14018464e+00 7.30991006e-01 -1.29404569e+00
-8.93925548e-01 -6.39376223e-01 7.95337617e-01 6.57231271e-01
3.46644640e-01 -2.90824145e-01 -4.72604334e-01 -1.94510832e-01
1.50390685e-01 -1.06654656e+00 -3.31697822e-01 5.14824629e-01
-1.11906014e-01 -4.80513573e-01 3.84748042e-01 -7.51552403e-01
5.19094348e-01 -1.71178615e+00 4.35015768e-01 -4.79628630e-02
5.96848488e-01 1.72624350e-01 9.45217237e-02 1.82476968e-01
2.48595864e-01 -7.51958713e-02 3.96314055e-01 -1.81638911e-01
-8.29979125e-03 7.21716881e-02 -1.64770603e-01 3.47780704e-01
4.58417721e-02 8.72816741e-01 -9.22266304e-01 -2.88380891e-01
3.49061519e-01 3.16278249e-01 -7.78238118e-01 8.20522785e-01
-4.62343961e-01 3.63236129e-01 -2.80255467e-01 3.53525162e-01
4.09233987e-01 -7.35786021e-01 2.92155921e-01 -1.76650271e-01
-1.07150868e-01 3.14255476e-01 -1.10201061e+00 1.48093879e+00
-3.11187565e-01 7.58182466e-01 2.27233320e-01 -7.81170964e-01
6.35422826e-01 7.65533447e-02 -1.86516970e-01 -5.71653783e-01
-1.33909341e-02 -2.64618956e-02 2.89043218e-01 -6.47234142e-01
3.46539199e-01 -3.72471102e-02 3.22311580e-01 5.20392060e-01
1.83052626e-02 -2.21036896e-01 -6.27921373e-02 3.85622680e-01
1.00870585e+00 4.00928855e-01 4.60566878e-01 -1.62425749e-02
4.12300646e-01 1.64371610e-01 2.11165443e-01 1.15745926e+00
-3.52370530e-01 5.14645457e-01 4.87572849e-01 -9.55522299e-01
-8.77277911e-01 -9.46620107e-01 6.75058424e-01 1.22319973e+00
3.42410505e-01 2.94749998e-02 -8.06544363e-01 -7.39616930e-01
-4.21092547e-02 1.10047662e+00 -7.30457187e-01 2.66137332e-01
-3.26930702e-01 1.13208406e-02 -2.52698481e-01 6.55161202e-01
6.04551375e-01 -9.35326040e-01 -1.20271671e+00 -3.46946791e-02
-3.83738816e-01 -1.18527615e+00 -3.18778694e-01 -1.32574752e-01
-5.45601010e-01 -1.07399750e+00 -5.50479770e-01 -3.85317594e-01
9.30491805e-01 5.21890044e-01 1.56724441e+00 4.62514460e-01
3.24935652e-02 5.68921268e-01 -9.40825418e-02 -3.73547912e-01
-3.73026505e-02 -9.02073383e-02 5.13013043e-02 -2.45689034e-01
3.75734299e-01 -4.83100384e-01 -9.57012951e-01 6.19717181e-01
-3.71164978e-01 5.65643728e-01 7.13072002e-01 5.52393973e-01
1.75010830e-01 -2.73196906e-01 4.26752359e-01 -8.68582249e-01
2.80044764e-01 -1.81604683e-01 -8.61182332e-01 2.74034858e-01
-6.80982590e-01 2.99094856e-01 3.46505731e-01 -2.56151468e-01
-1.25494468e+00 1.71585649e-01 1.40226483e-01 -3.39199305e-01
-3.33836079e-01 1.88510224e-01 -3.10540069e-02 -1.46891013e-01
7.85524905e-01 -9.46101695e-02 -2.72999424e-02 -2.43003488e-01
4.86591429e-01 3.49944621e-01 8.10691953e-01 -2.91655779e-01
7.38870800e-01 5.16185224e-01 -4.05077599e-02 -5.65076768e-01
-1.40029776e+00 -5.61050415e-01 -7.71852493e-01 -4.37711537e-01
1.07227194e+00 -8.55615139e-01 -1.27632368e+00 1.48257902e-02
-1.79060459e+00 -4.07583788e-02 -2.59789918e-03 1.89642727e-01
-7.31007814e-01 1.58718377e-01 1.58260167e-02 -1.05421090e+00
6.12707064e-02 -1.01119363e+00 9.09919500e-01 4.10995632e-01
-4.72589850e-01 -7.82466948e-01 -2.01064825e-01 1.08549690e+00
4.86949325e-01 -9.27815437e-02 7.20409274e-01 -8.97634566e-01
-1.24253285e+00 -5.62024601e-02 -5.88295281e-01 -3.08125485e-02
-4.20973778e-01 -4.48072493e-01 -9.78574216e-01 6.90931678e-02
2.76733786e-01 -2.79190630e-01 6.38511896e-01 2.52680689e-01
8.26939464e-01 -5.09458482e-01 -2.84174055e-01 4.37469363e-01
1.13430977e+00 2.18650579e-01 4.95940447e-01 3.14719319e-01
5.73306203e-01 8.17794025e-01 6.49402499e-01 9.89569649e-02
9.19256151e-01 6.53468966e-01 8.65334451e-01 5.33862459e-03
1.42642185e-01 -1.67340368e-01 -9.39446390e-02 1.06149361e-01
-4.60071206e-01 -4.14126366e-01 -8.35475743e-01 6.23410165e-01
-2.04703641e+00 -9.52425063e-01 4.00146604e-01 2.04489088e+00
3.96056831e-01 3.99656981e-01 7.90753402e-03 -1.23595610e-01
3.50734055e-01 1.09118529e-01 -8.08536351e-01 -3.08731526e-01
3.91312808e-01 -4.71932262e-01 1.09993398e-01 7.69202471e-01
-9.92112339e-01 9.59911466e-01 6.26657009e+00 1.69773195e-02
-7.07244277e-01 -1.14562027e-01 7.12836504e-01 -1.47641137e-01
-3.07296459e-02 1.97210968e-01 -7.79324710e-01 -8.28586966e-02
6.52752340e-01 6.75106198e-02 9.02881861e-01 8.72276247e-01
1.27777427e-01 -6.44478142e-01 -1.56018209e+00 9.80186164e-01
5.92547536e-01 -1.25182152e+00 3.29116695e-02 2.09156349e-01
6.25839233e-01 -5.35920262e-02 5.67748174e-02 1.60517275e-01
4.41936463e-01 -1.21569741e+00 4.29830819e-01 1.00594294e+00
3.40341717e-01 -6.47832096e-01 7.28660643e-01 9.74814057e-01
-6.15151703e-01 -1.84473470e-01 -2.81456739e-01 -6.26963854e-01
1.42315492e-01 2.33049586e-01 -1.27898431e+00 3.31805706e-01
5.20464659e-01 2.70902485e-01 -5.56588590e-01 8.33419442e-01
-6.06302857e-01 -3.46067883e-02 -2.73375750e-01 -2.31660739e-01
2.92539805e-01 6.83026165e-02 6.29383683e-01 5.36265790e-01
-3.73837724e-02 4.53532577e-01 2.07758427e-01 1.07963383e+00
7.74167925e-02 -4.47280318e-01 -7.48182833e-01 -4.77645434e-02
2.13484377e-01 1.13687110e+00 -5.49852550e-01 -3.57863039e-01
-2.24780202e-01 1.12392128e+00 6.62652135e-01 3.48270357e-01
-3.76460552e-01 -2.59948015e-01 5.46381056e-01 3.45757067e-01
2.80880719e-01 -2.03629330e-01 -2.98547685e-01 -1.02253985e+00
-9.49682072e-02 -1.16647089e+00 3.16422343e-01 -1.72323883e+00
-1.06367671e+00 6.32266462e-01 -3.14889029e-02 -8.43655825e-01
-7.73518205e-01 -1.08446705e+00 -4.36727226e-01 7.93549478e-01
-1.35187566e+00 -1.05425894e+00 -4.47033525e-01 1.05749056e-01
7.11425722e-01 -6.52936175e-02 5.58204949e-01 -3.48929614e-01
8.06232914e-02 1.39063403e-01 -7.03724205e-01 -5.67773543e-02
6.59003377e-01 -1.45900071e+00 4.95735407e-01 8.62106264e-01
2.00721711e-01 8.01411033e-01 1.07207751e+00 -4.42138225e-01
-1.48487806e+00 -3.82183671e-01 1.12775278e+00 -1.15505934e+00
3.25764596e-01 -9.13029313e-02 -7.13121533e-01 8.67251933e-01
4.73435342e-01 -2.30739102e-01 1.94421396e-01 7.24884942e-02
-4.01564658e-01 1.12419307e-01 -1.19038653e+00 8.28834116e-01
1.04779851e+00 -5.75910985e-01 -1.16645384e+00 6.14344895e-01
7.36041486e-01 -4.59207177e-01 -3.51064324e-01 7.78999776e-02
6.60302818e-01 -1.34480774e+00 1.11768329e+00 -6.76240325e-01
6.29287481e-01 -4.96191531e-01 -1.26155064e-01 -1.28325605e+00
-4.06978816e-01 -4.99265552e-01 -1.64882436e-01 3.71709883e-01
4.94758487e-01 -4.86780524e-01 7.81010985e-01 8.83412063e-01
3.49528193e-01 -6.03278279e-01 -7.66616762e-01 -2.03513518e-01
-4.63267118e-01 -1.02919616e-01 2.02198043e-01 5.51048875e-01
-5.66157475e-02 7.90655971e-01 -4.83189851e-01 3.55837345e-01
7.14146852e-01 2.61599451e-01 1.01940036e+00 -1.43953609e+00
-4.41841245e-01 -1.86364740e-01 -3.17005277e-01 -1.46896005e+00
3.35543826e-02 -4.21871662e-01 2.68993437e-01 -1.90703487e+00
1.09285437e-01 4.37204719e-01 -1.03373546e-03 3.54138047e-01
-2.61209399e-01 -4.34630625e-02 5.08448720e-01 1.62227079e-02
-1.12568140e+00 -6.24611825e-02 1.35153019e+00 3.97993997e-02
2.25779399e-01 1.36804700e-01 -1.19450641e+00 1.17649281e+00
5.90395749e-01 -1.69552162e-01 -5.37621677e-01 -6.45153582e-01
9.25288022e-01 2.90652812e-01 7.11314321e-01 -9.88048971e-01
9.09568191e-01 -2.87433535e-01 6.77038252e-01 -7.54201591e-01
5.67460716e-01 -1.00194144e+00 -6.64664507e-02 3.05100799e-01
-4.59607095e-01 2.15771973e-01 -1.23411894e-01 5.99279463e-01
1.77346036e-01 -7.12978393e-02 6.23532176e-01 -6.19857430e-01
-8.00258815e-01 -3.38463997e-03 -4.06628907e-01 1.86532855e-01
7.67056882e-01 -1.32762581e-01 -4.97380763e-01 -1.08950329e+00
-7.14975178e-01 3.37912470e-01 4.75390106e-01 2.02413082e-01
7.49556720e-01 -7.55066395e-01 -6.00842535e-01 1.55909183e-02
-1.16813565e-02 -8.78334045e-02 8.84142146e-02 6.84982181e-01
-3.68472785e-01 6.72596872e-01 -2.77392715e-01 -4.58466142e-01
-1.15296435e+00 5.87562799e-01 5.96978366e-01 -2.34214962e-01
-1.73294321e-01 1.06480181e+00 4.52108979e-01 -5.68286240e-01
2.32543945e-01 -7.54047558e-02 -3.26464623e-01 -1.04150109e-01
5.79794347e-01 4.00390476e-01 -2.71699995e-01 -3.49801093e-01
-3.79508346e-01 4.85569537e-01 -2.22109869e-01 -3.75004947e-01
1.19217896e+00 -3.71211857e-01 7.69351900e-04 2.65055329e-01
5.91433764e-01 -3.75342995e-01 -1.42934930e+00 -2.89026171e-01
-1.97308421e-01 -7.42869377e-01 6.46155104e-02 -1.54794097e+00
-7.59434879e-01 1.03540051e+00 2.18118533e-01 1.95142806e-01
7.27744460e-01 7.91378319e-02 2.13885158e-01 1.11542141e+00
4.07278121e-01 -1.05394650e+00 3.25493395e-01 6.03486001e-01
1.33162987e+00 -1.59578931e+00 3.57871056e-01 -8.43572468e-02
-1.05471635e+00 9.29965675e-01 1.15151846e+00 -1.72124758e-01
3.67548943e-01 -4.69492972e-02 4.19888705e-01 -3.85515034e-01
-1.26499319e+00 3.85518819e-02 6.25569582e-01 6.26942992e-01
1.88855499e-01 -1.22266255e-01 3.91914696e-01 3.05730402e-01
-3.17109227e-01 7.50824958e-02 3.98791999e-01 6.78272426e-01
-7.73374677e-01 -3.67939085e-01 -3.91752064e-01 4.73244905e-01
-6.30397424e-02 8.65234211e-02 -9.30965126e-01 6.67102695e-01
1.13402896e-01 1.16856098e+00 9.48434025e-02 -2.61316210e-01
2.93981105e-01 1.40714329e-02 6.39955938e-01 -6.16834760e-01
-5.06058514e-01 -1.52074441e-01 2.00917751e-01 -8.39150965e-01
-3.97596687e-01 -3.64855379e-01 -8.12824130e-01 -8.12304989e-02
-1.95994303e-01 -1.57224208e-01 5.59789062e-01 1.23951936e+00
4.95854139e-01 3.60849649e-01 2.22813904e-01 -1.24702334e+00
-8.11352134e-01 -7.67647326e-01 6.36666194e-02 -1.32388324e-02
4.08032209e-01 -6.01747870e-01 -6.81930184e-01 -1.26600608e-01] | [10.739184379577637, 1.9580763578414917] |
3e9eb429-f297-4ad8-811f-dc0090225252 | cylin-painting-seamless-360deg-panoramic | 2204.08563 | null | https://arxiv.org/abs/2204.08563v1 | https://arxiv.org/pdf/2204.08563v1.pdf | Cylin-Painting: Seamless 360° Panoramic Image Outpainting and Beyond with Cylinder-Style Convolutions | Image outpainting gains increasing attention since it can generate the complete scene from a partial view, providing a valuable solution to construct 360{\deg} panoramic images. As image outpainting suffers from the intrinsic issue of unidirectional completion flow, previous methods convert the original problem into inpainting, which allows a bidirectional flow. However, we find that inpainting has its own limitations and is inferior to outpainting in certain situations. The question of how they may be combined for the best of both has as yet remained under-explored. In this paper, we provide a deep analysis of the differences between inpainting and outpainting, which essentially depends on how the source pixels contribute to the unknown regions under different spatial arrangements. Motivated by this analysis, we present a Cylin-Painting framework that involves meaningful collaborations between inpainting and outpainting and efficiently fuses the different arrangements, with a view to leveraging their complementary benefits on a consistent and seamless cylinder. Nevertheless, directly applying the cylinder-style convolution often generates visually unpleasing results as it could discard important positional information. To address this issue, we further present a learnable positional embedding strategy and incorporate the missing component of positional encoding into the cylinder convolution, which significantly improves the panoramic results. Note that while developed for image outpainting, the proposed solution can be effectively extended to other panoramic vision tasks, such as object detection, depth estimation, and image super resolution. | ['Yao Zhao', 'Yunchao Wei', 'Wenqi Ren', 'Chunyu Lin', 'Xiangyu Xu', 'Kang Liao'] | 2022-04-18 | null | null | null | null | ['image-outpainting'] | ['computer-vision'] | [ 6.47050858e-01 1.36191308e-01 -1.01788223e-01 -8.67802948e-02
-4.37977344e-01 -5.54144204e-01 4.27392423e-01 -2.94074625e-01
-7.70915300e-02 7.60406911e-01 2.06447244e-01 -1.05753578e-01
-1.36482671e-01 -9.04215276e-01 -7.61434197e-01 -8.22538972e-01
4.70444798e-01 -1.20042548e-01 1.14608943e-01 -2.06956148e-01
2.48359501e-01 6.92936420e-01 -1.56631148e+00 2.24504292e-01
9.70180035e-01 8.37088227e-01 3.63246262e-01 4.87411588e-01
-1.68112844e-01 9.61478472e-01 -3.75343829e-01 -6.06472433e-01
5.10379612e-01 -5.66431701e-01 -5.28607070e-01 2.71124244e-01
6.67738438e-01 -7.21225798e-01 -5.58572710e-01 1.00456119e+00
1.63270772e-01 -1.47953942e-01 3.20219070e-01 -1.16661239e+00
-5.93554914e-01 2.49981791e-01 -8.51211607e-01 -2.33176574e-01
2.81632781e-01 5.65509088e-02 8.40017855e-01 -7.60081470e-01
6.86087310e-01 9.70568657e-01 5.97426534e-01 4.24454719e-01
-1.19323099e+00 -5.55671334e-01 -1.68468952e-02 1.14112705e-01
-1.11029136e+00 -3.97814691e-01 1.26312423e+00 -2.95771927e-01
3.39618832e-01 5.75655282e-01 7.17161536e-01 9.17330503e-01
1.37254685e-01 8.24818850e-01 1.22474194e+00 -5.01281083e-01
1.06897123e-01 8.79639667e-03 -3.11236084e-01 5.49576819e-01
2.20313087e-01 3.01369816e-01 -4.82036173e-01 2.32722685e-02
1.22816789e+00 2.20687255e-01 -7.66969800e-01 -5.87330699e-01
-1.06505251e+00 7.27278233e-01 5.52575469e-01 7.56363571e-02
-2.64181107e-01 1.45786420e-01 2.49420460e-02 2.96311706e-01
4.72952276e-01 5.38731277e-01 -2.52067655e-01 4.39875424e-02
-1.15719748e+00 2.85670489e-01 4.76236790e-01 7.76403606e-01
9.53457832e-01 3.44212428e-02 -8.20551813e-02 7.84823895e-01
-2.82798707e-02 6.15222305e-02 7.09892735e-02 -1.30471337e+00
3.77931058e-01 4.42609012e-01 3.05048227e-01 -1.08806825e+00
1.86487997e-03 -2.62498200e-01 -8.12523007e-01 6.69577360e-01
5.41193247e-01 2.29722112e-02 -5.63153982e-01 1.71645999e+00
2.55701691e-01 1.41014993e-01 -1.27799943e-01 1.00772142e+00
4.55450535e-01 6.12717092e-01 -3.28162193e-01 -3.02799672e-01
1.28062057e+00 -1.22290754e+00 -9.18426216e-01 -2.89058834e-01
1.01619601e-01 -9.35911179e-01 1.06414294e+00 4.26165879e-01
-1.28586626e+00 -3.36776495e-01 -1.14439499e+00 -5.68661809e-01
-7.41472021e-02 -1.52969565e-02 7.87212610e-01 6.30598426e-01
-1.03149939e+00 5.47431350e-01 -5.93560517e-01 -1.32301375e-01
3.16735953e-01 -6.25583809e-03 -4.61133212e-01 -3.79616946e-01
-8.95567238e-01 7.98436224e-01 5.68373762e-02 4.07099398e-03
-3.68216187e-01 -8.05528283e-01 -9.16827202e-01 1.65711850e-01
5.24800897e-01 -8.39893281e-01 1.00974703e+00 -1.08160627e+00
-1.61652827e+00 6.22691572e-01 -1.73735052e-01 -3.07531238e-01
8.16018879e-01 -2.44793653e-01 -5.00298738e-02 4.07968163e-01
-3.62650044e-02 7.44491398e-01 1.08364022e+00 -1.39014363e+00
-4.80577737e-01 -3.50015730e-01 2.02284768e-01 3.47591311e-01
-4.19502527e-01 -2.20219478e-01 -7.16901124e-01 -1.21179545e+00
1.31268278e-01 -5.75381100e-01 -1.22600913e-01 6.50723040e-01
-2.91603565e-01 3.95576715e-01 1.16787982e+00 -7.02356994e-01
1.14193356e+00 -2.08501053e+00 4.09828484e-01 -2.22760320e-01
2.13617936e-01 5.13879135e-02 -7.72107765e-02 5.86067319e-01
-6.44819662e-02 -2.19105631e-01 -6.54407442e-01 -5.95487416e-01
-4.45169717e-01 2.80443519e-01 -6.31799698e-01 3.99022371e-01
2.58404076e-01 8.46400321e-01 -9.24267471e-01 -3.70139480e-01
5.25247097e-01 8.51684570e-01 -5.77878952e-01 1.93602815e-01
-1.69275135e-01 5.83439529e-01 -3.52782607e-01 7.11106539e-01
1.03116918e+00 8.34801719e-02 1.17471190e-02 -4.16927248e-01
-3.31604958e-01 -1.22267365e-01 -1.02911663e+00 1.86353338e+00
-5.83643138e-01 8.56377065e-01 4.83739763e-01 -7.81718135e-01
7.96823680e-01 2.60412414e-02 5.40735483e-01 -7.31755137e-01
-7.38393366e-02 2.11452618e-01 -4.73182976e-01 -4.18780893e-01
7.66045392e-01 -2.19583735e-01 2.97267050e-01 3.93613726e-01
-2.30827913e-01 -3.84142250e-01 -2.60678064e-02 5.83218709e-02
8.36296320e-01 2.94248492e-01 2.26311848e-01 1.82667434e-01
3.27183992e-01 -2.42774308e-01 2.47216851e-01 5.65349400e-01
2.09679827e-02 1.28509402e+00 5.18437028e-01 -3.17308903e-01
-1.06070673e+00 -1.00256205e+00 -2.03828096e-01 4.49651182e-01
5.14528811e-01 -4.03920203e-01 -8.15553784e-01 -6.14880800e-01
-1.67023733e-01 6.77491426e-01 -6.00186527e-01 -5.60047850e-02
-6.99613869e-01 -5.37761867e-01 2.31993899e-01 4.02250975e-01
5.85447609e-01 -8.76349568e-01 -9.17841911e-01 1.60140425e-01
-4.28445578e-01 -1.15043509e+00 -5.92605829e-01 2.28871442e-02
-1.01625657e+00 -1.12203372e+00 -1.18110228e+00 -5.62795341e-01
5.88787973e-01 7.43538976e-01 8.10965955e-01 -5.47183044e-02
-1.94107339e-01 1.93997517e-01 -2.92533666e-01 -5.85677028e-02
-3.54311436e-01 -2.14694694e-01 -4.02224988e-01 1.76305950e-01
-1.07016347e-01 -1.01694643e+00 -9.30468261e-01 1.89625621e-01
-1.41853535e+00 6.31610096e-01 7.33408630e-01 8.73678982e-01
3.84088248e-01 6.63362667e-02 1.36082292e-01 -8.15858722e-01
3.78904432e-01 -1.83370084e-01 -5.16592085e-01 1.95710391e-01
-5.11283338e-01 7.21035302e-02 7.24114180e-01 -2.81855822e-01
-1.23248696e+00 8.57857689e-02 -2.27472782e-01 -7.57045627e-01
2.08891146e-02 4.24035043e-02 -2.10701466e-01 -2.81146705e-01
2.57424146e-01 4.14699107e-01 1.71306044e-01 -6.14550173e-01
6.61060452e-01 4.10138905e-01 5.58139622e-01 -3.48907024e-01
8.31814706e-01 1.04390955e+00 5.24164774e-02 -8.84521067e-01
-7.33665764e-01 -4.24819924e-02 -3.54875535e-01 -1.35253772e-01
6.82773292e-01 -7.34957933e-01 -4.40764099e-01 4.92356777e-01
-1.22102666e+00 -2.10466713e-01 -5.81015289e-01 1.58272788e-01
-6.17963552e-01 8.87431026e-01 -6.88576996e-01 -5.97567618e-01
-6.84781671e-02 -1.23097813e+00 1.17577720e+00 1.91123813e-01
3.56514491e-02 -7.20829964e-01 -2.81162485e-02 4.06380296e-01
3.74500662e-01 3.52434158e-01 9.18394983e-01 3.70105743e-01
-9.61532772e-01 7.32024014e-03 -4.34292138e-01 3.89822543e-01
3.30327034e-01 -1.61834314e-01 -1.09949458e+00 -2.35439658e-01
3.62635046e-01 -4.76815365e-02 1.06863523e+00 3.96412641e-01
1.29880178e+00 -3.78308684e-01 -1.23672538e-01 1.05935276e+00
1.49043679e+00 1.43557787e-02 1.06837904e+00 4.57456768e-01
7.91912138e-01 9.04239178e-01 4.59025413e-01 2.95491695e-01
1.43992037e-01 9.39904392e-01 7.15763748e-01 -3.23331892e-01
-4.92283672e-01 -5.06314397e-01 3.17351848e-01 3.43552411e-01
-8.92514288e-02 -2.04817578e-01 -2.68755585e-01 2.37519279e-01
-1.62928605e+00 -9.39902365e-01 5.40887825e-02 2.25475931e+00
9.34049666e-01 -2.30560482e-01 -1.82634056e-01 2.85165250e-01
5.47481060e-01 5.28871775e-01 -4.84508812e-01 -2.09673554e-01
-3.07486415e-01 3.41595292e-01 5.10551512e-01 7.71939993e-01
-8.15702081e-01 7.87739158e-01 5.86588955e+00 1.15806603e+00
-1.36171544e+00 6.30678535e-02 7.18575358e-01 1.62824877e-02
-6.42570674e-01 2.73058832e-01 -3.79719466e-01 4.75407004e-01
1.20155357e-01 2.32906908e-01 5.00386059e-01 5.96640348e-01
1.86587378e-01 -2.89632171e-01 -9.10793483e-01 1.16284740e+00
1.64735571e-01 -1.45133531e+00 2.82877088e-01 -5.90007007e-03
7.03786433e-01 -4.87160653e-01 1.55067012e-01 -2.99966812e-01
-2.29727611e-01 -9.59776640e-01 8.58923733e-01 3.26815277e-01
9.65824783e-01 -7.01631427e-01 2.78813809e-01 3.48264605e-01
-8.87706935e-01 -1.02138229e-01 -2.26680458e-01 -2.70822257e-01
4.62244362e-01 8.24409664e-01 -3.02682132e-01 8.63842189e-01
6.16745412e-01 7.29075193e-01 -4.31218356e-01 9.19664919e-01
-3.65324765e-01 1.19243599e-01 -1.66383788e-01 6.26923501e-01
2.12229174e-02 -6.36774123e-01 6.14038587e-01 7.89428174e-01
5.13493240e-01 8.50742031e-03 -4.44354862e-01 1.15837216e+00
8.23102072e-02 -1.14086695e-01 -7.06948042e-01 1.98027655e-01
2.62820333e-01 1.21914589e+00 -7.64759481e-01 -7.77402744e-02
-5.20522773e-01 1.39067900e+00 2.26707563e-01 5.93956888e-01
-7.55144596e-01 -3.44361007e-01 6.39962137e-01 4.00815517e-01
3.36055368e-01 -1.33921862e-01 -5.98417819e-01 -1.34876144e+00
3.01051229e-01 -6.51811242e-01 -1.12544540e-02 -9.17642236e-01
-9.69339371e-01 5.29066563e-01 -6.86191395e-02 -1.34800136e+00
-4.38516103e-02 -5.02492666e-01 -6.15661323e-01 6.59889221e-01
-1.90927041e+00 -1.23549604e+00 -4.53781128e-01 5.07204771e-01
6.81290507e-01 3.96816045e-01 5.19782305e-01 4.68929708e-01
-3.49128842e-01 4.67747509e-01 1.07550062e-01 -2.32687175e-01
7.84153283e-01 -9.26039815e-01 5.51443994e-02 1.01546919e+00
1.00348197e-01 3.81318778e-01 8.42790544e-01 -3.79489899e-01
-1.36289680e+00 -1.02289820e+00 7.16123223e-01 -3.43961626e-01
1.81326956e-01 -2.56992280e-01 -7.79457808e-01 4.32481140e-01
3.08557630e-01 -6.04514517e-02 5.80653511e-02 -5.88331163e-01
-2.89830565e-01 -1.26686454e-01 -1.05395854e+00 9.19166148e-01
9.71304953e-01 -4.43856925e-01 -3.78910869e-01 1.17054032e-02
7.20151782e-01 -3.07082355e-01 -4.13983583e-01 3.98947507e-01
5.94302893e-01 -1.67035711e+00 1.27591634e+00 1.23169176e-01
8.62147391e-01 -4.35390532e-01 -3.74650843e-02 -1.04480290e+00
5.44517189e-02 -8.11052203e-01 -2.45605007e-01 1.18752015e+00
-1.52453154e-01 -5.79626679e-01 8.31351936e-01 4.37192947e-01
-1.64114967e-01 -9.45440888e-01 -8.02238703e-01 -4.69933689e-01
6.13939166e-02 -4.30552572e-01 4.60763544e-01 1.02295327e+00
-3.79481316e-01 -5.01229875e-02 -8.31708968e-01 -3.38124260e-02
6.16625369e-01 3.85129869e-01 7.05018997e-01 -8.47128689e-01
-5.69790959e-01 -3.74415070e-01 -5.87479025e-03 -1.55477035e+00
-8.77995640e-02 -4.88394976e-01 -1.44731775e-01 -1.28251731e+00
1.25970200e-01 -4.14860040e-01 1.62364841e-01 1.76619351e-01
-1.29570484e-01 6.68185234e-01 3.15620422e-01 3.99278373e-01
1.50077969e-01 6.90063119e-01 1.68384933e+00 2.06768643e-02
-1.49521828e-01 -9.29539129e-02 -9.13814902e-01 6.73358440e-01
5.78352809e-01 -1.89385444e-01 -5.75871408e-01 -6.52699590e-01
1.87316760e-01 3.86421144e-01 5.87421536e-01 -8.40293705e-01
1.66113809e-01 -1.15138896e-01 3.74935657e-01 -4.35468942e-01
7.16377556e-01 -8.64040732e-01 2.94422001e-01 2.13138342e-01
-1.17157526e-01 -1.44714922e-01 7.56259859e-02 6.43509984e-01
-5.16866148e-01 -1.81230947e-01 8.08224678e-01 -2.06497580e-01
-5.46655357e-01 2.06149906e-01 -5.63707910e-02 -2.62730062e-01
9.56348896e-01 -6.70972586e-01 -1.78030267e-01 -5.52247703e-01
-3.47125351e-01 -1.38424128e-01 1.02755678e+00 3.48958284e-01
8.09285462e-01 -1.14070761e+00 -3.76925707e-01 5.03195167e-01
-1.77334651e-01 1.53781369e-01 5.46088994e-01 9.19130027e-01
-7.84080565e-01 1.61280513e-01 -3.51828635e-01 -4.49199110e-01
-8.94139826e-01 6.02495611e-01 1.32536590e-01 -2.11790606e-01
-8.97550464e-01 6.36748970e-01 7.80301213e-01 -8.56892616e-02
2.01025352e-01 -2.75632888e-01 -3.70837972e-02 -2.03115530e-02
5.38620889e-01 3.42025846e-01 2.06339601e-02 -4.33982998e-01
3.14612687e-02 8.87613118e-01 -8.49342160e-03 -1.54150620e-01
1.25286138e+00 -3.76100302e-01 -1.80419922e-01 1.57478809e-01
1.06590307e+00 4.43788648e-01 -1.77230310e+00 1.09296344e-01
-4.31861222e-01 -8.88743103e-01 8.89724642e-02 -5.01082778e-01
-1.37694645e+00 1.07564151e+00 3.43735516e-01 -4.50643199e-03
1.39777219e+00 -2.16888890e-01 1.11085820e+00 -2.01445520e-01
2.54071414e-01 -7.28103817e-01 1.09010749e-01 8.39831829e-02
1.00530457e+00 -1.09823740e+00 6.75445199e-02 -8.51992011e-01
-3.27211320e-01 1.37562513e+00 4.44670916e-01 -2.12148130e-01
3.20348650e-01 3.85663599e-01 -5.87219885e-03 -3.84864546e-02
-3.12845290e-01 9.45524350e-02 2.09664311e-02 5.44538856e-01
2.49945268e-01 -1.79555133e-01 -4.60231572e-01 2.48651654e-01
-9.21171829e-02 3.18531832e-03 6.26665533e-01 7.97882199e-01
-9.31188539e-02 -1.30155241e+00 -5.43039680e-01 8.82755071e-02
-3.22855443e-01 -1.36717156e-01 -3.24452817e-01 9.00608122e-01
1.57853425e-01 5.99097729e-01 -4.87542972e-02 -4.85115722e-02
1.63160399e-01 -3.02803636e-01 6.58797264e-01 -2.53331721e-01
-1.66083276e-01 1.83397770e-01 -2.75480002e-01 -7.46807575e-01
-4.51213419e-01 -3.07500064e-01 -7.08950281e-01 -1.31275430e-01
-1.96032003e-01 -1.98114827e-01 3.93786967e-01 5.78802705e-01
3.44217032e-01 3.52512717e-01 5.19360840e-01 -1.20196164e+00
-3.97068322e-01 -5.82275569e-01 -6.05506897e-01 4.58724201e-01
6.12981260e-01 -5.66028059e-01 -5.92958450e-01 -6.07141890e-02] | [11.250763893127441, -1.1583794355392456] |
609e22f8-6c0d-41b3-a323-2cd646fff14e | aligning-to-social-norms-and-values-in | 2205.01975 | null | https://arxiv.org/abs/2205.01975v2 | https://arxiv.org/pdf/2205.01975v2.pdf | Aligning to Social Norms and Values in Interactive Narratives | We focus on creating agents that act in alignment with socially beneficial norms and values in interactive narratives or text-based games -- environments wherein an agent perceives and interacts with a world through natural language. Such interactive agents are often trained via reinforcement learning to optimize task performance, even when such rewards may lead to agent behaviors that violate societal norms -- causing harm either to the agent itself or other entities in the environment. Social value alignment refers to creating agents whose behaviors conform to expected moral and social norms for a given context and group of people -- in our case, it means agents that behave in a manner that is less harmful and more beneficial for themselves and others. We build on the Jiminy Cricket benchmark (Hendrycks et al. 2021), a set of 25 annotated interactive narratives containing thousands of morally salient scenarios covering everything from theft and bodily harm to altruism. We introduce the GALAD (Game-value ALignment through Action Distillation) agent that uses the social commonsense knowledge present in specially trained language models to contextually restrict its action space to only those actions that are aligned with socially beneficial values. An experimental study shows that the GALAD agent makes decisions efficiently enough to improve state-of-the-art task performance by 4% while reducing the frequency of socially harmful behaviors by 25% compared to strong contemporary value alignment approaches. | ['Yejin Choi', 'Hannaneh Hajishirzi', 'Maarten Sap', 'Liwei Jiang', 'Prithviraj Ammanabrolu'] | 2022-05-04 | null | https://aclanthology.org/2022.naacl-main.439 | https://aclanthology.org/2022.naacl-main.439.pdf | naacl-2022-7 | ['text-based-games'] | ['playing-games'] | [ 3.85492653e-01 9.64081466e-01 3.30884196e-02 -2.21766129e-01
-8.09409991e-02 -4.22882587e-01 1.10147250e+00 2.37254664e-01
-8.01123977e-01 1.20145988e+00 8.83746564e-01 2.43804350e-01
-1.58837020e-01 -9.65080619e-01 -2.73491442e-01 -6.21051908e-01
-1.46129340e-01 8.30425143e-01 -8.69957581e-02 -9.38598871e-01
8.57428387e-02 -9.25711393e-02 -1.37038589e+00 1.40538216e-01
6.52004838e-01 4.44869071e-01 -3.65110278e-01 8.11239004e-01
5.37864625e-01 1.72475946e+00 -7.45691299e-01 -1.02753699e+00
3.58560622e-01 -9.21670496e-01 -1.07677662e+00 -2.64581740e-01
-1.84233651e-01 -6.01535797e-01 -1.23948574e-01 1.07113540e+00
4.20789659e-01 6.62741482e-01 9.22348142e-01 -1.49604630e+00
-8.32225025e-01 1.30664480e+00 -2.44646385e-01 -2.84115404e-01
6.28977060e-01 7.20301569e-01 1.38816607e+00 1.99203923e-01
1.02634871e+00 1.46629071e+00 4.37443346e-01 1.34518385e+00
-1.19087982e+00 -2.64032513e-01 -1.28509626e-01 1.35313019e-01
-5.30684352e-01 -3.40307713e-01 7.02191770e-01 -5.00275433e-01
1.40685058e+00 4.90775675e-01 1.05394542e+00 1.54053652e+00
3.60668629e-01 3.91277343e-01 8.21921349e-01 8.48497450e-02
5.11680961e-01 -2.57941872e-01 -4.74440992e-01 3.61922979e-01
8.07277709e-02 4.73957628e-01 -5.92683971e-01 -3.51851523e-01
3.78847301e-01 -5.53642452e-01 3.97673249e-01 -3.44011545e-01
-1.64834905e+00 8.94567490e-01 5.15837133e-01 2.74892122e-01
-7.90541172e-01 8.02169561e-01 7.40938246e-01 4.13929135e-01
3.11664075e-01 1.51646721e+00 -6.46717995e-02 -6.75590515e-01
6.91222847e-02 1.01552582e+00 8.54667306e-01 4.54102546e-01
2.34531701e-01 1.33992583e-01 -3.40435266e-01 6.73446894e-01
5.78703992e-02 4.18606311e-01 2.17507154e-01 -1.75032866e+00
1.22256041e-01 6.29670799e-01 3.31495285e-01 -1.12601209e+00
-6.13096714e-01 2.04197615e-02 -4.90293205e-01 5.08219421e-01
5.07096410e-01 -4.97423023e-01 -7.20700547e-02 2.13550878e+00
4.38345820e-01 -1.99492216e-01 5.45129180e-01 8.82555246e-01
7.63236582e-01 5.12868881e-01 5.81921101e-01 -2.46066034e-01
1.17470562e+00 -6.03561699e-01 -7.53125072e-01 -4.47847486e-01
1.16860962e+00 -2.41522312e-01 1.21269739e+00 1.87952250e-01
-1.20069277e+00 3.28582942e-01 -7.20789433e-01 2.02151807e-03
-6.44645169e-02 -9.21034575e-01 7.20429242e-01 4.69268739e-01
-9.72634733e-01 7.29462326e-01 -2.26370230e-01 -3.43945205e-01
5.80828726e-01 2.38070294e-01 -4.16171789e-01 6.19454443e-01
-1.49485612e+00 1.33435369e+00 5.38088918e-01 -6.32925093e-01
-1.15842021e+00 -5.70929825e-01 -9.49556410e-01 5.86738437e-02
8.69512200e-01 -7.78520167e-01 1.40749741e+00 -1.20135581e+00
-1.75598800e+00 1.23186183e+00 7.42552638e-01 -8.75992239e-01
9.94842291e-01 -2.25162044e-01 4.40465882e-02 -5.41631468e-02
2.72551060e-01 1.01363957e+00 3.58383060e-01 -1.07680571e+00
-4.79751706e-01 9.57162827e-02 9.22220051e-01 8.63559783e-01
-2.22053051e-01 4.18625951e-01 7.09278107e-01 -6.96857452e-01
-8.50225925e-01 -1.13478780e+00 -5.50930679e-01 -1.37665287e-01
-5.42530715e-01 -4.11859870e-01 2.65127182e-01 -3.50162596e-01
5.95367789e-01 -1.71187341e+00 5.59374571e-01 -5.78055494e-02
6.72753870e-01 -8.66188481e-02 -1.63202778e-01 4.10788953e-01
9.86305922e-02 1.76062435e-01 -3.37047547e-01 -3.36209461e-02
4.52757955e-01 3.10536385e-01 9.83453169e-02 4.91870522e-01
1.14527130e-02 1.13604391e+00 -1.45965433e+00 -5.93058288e-01
2.44687796e-01 4.51587066e-02 -1.19686258e+00 4.79709595e-01
-5.78854084e-01 3.94697666e-01 -2.97458321e-01 1.83311645e-02
-1.63126469e-01 1.67874560e-01 3.09044182e-01 5.28384805e-01
2.76759326e-01 4.64985847e-01 -6.98659241e-01 1.10636961e+00
-1.32405713e-01 6.00378454e-01 -2.46843487e-01 -5.18127441e-01
6.36308193e-01 2.00076535e-01 6.01417363e-01 -6.41773462e-01
6.33439600e-01 -1.12827674e-01 5.17005086e-01 -5.28529882e-01
6.51466548e-01 -6.35641634e-01 -6.44294798e-01 1.04080415e+00
-2.33109728e-01 -7.74220169e-01 2.33695552e-01 4.81565803e-01
1.24397349e+00 2.52631694e-01 8.68525326e-01 -4.00392205e-01
3.22095960e-01 2.25187838e-01 5.75591266e-01 7.75125325e-01
-6.07397974e-01 -2.02522993e-01 1.15731490e+00 -6.49954677e-01
-1.13574624e+00 -9.88056302e-01 5.39860487e-01 1.74226618e+00
2.92548746e-01 -2.74965584e-01 -9.81038034e-01 -7.59324014e-01
-1.55981511e-01 1.42946088e+00 -1.07170427e+00 -6.19693279e-01
-6.42495513e-01 -4.24510539e-01 8.22003126e-01 1.76608950e-01
4.71126080e-01 -1.82469606e+00 -1.29549241e+00 1.45694405e-01
-2.48875931e-01 -9.36152816e-01 -1.97276443e-01 -7.82479122e-02
-1.54211611e-01 -1.17970896e+00 -9.79925692e-02 -1.80965230e-01
4.64506298e-01 -3.63720447e-01 1.24599779e+00 3.74610037e-01
1.51650950e-01 3.03595990e-01 -6.38469756e-01 -5.43314099e-01
-9.90746677e-01 -3.39840412e-01 3.59877110e-01 -2.96906352e-01
2.16434896e-01 -5.29421031e-01 -3.45047653e-01 2.35802516e-01
-6.18331790e-01 4.52220708e-01 -1.72418118e-01 7.20230162e-01
-2.33396545e-01 -3.64138186e-01 6.89878762e-01 -1.00083780e+00
1.14166605e+00 -5.77994049e-01 -5.44172116e-02 -1.29633188e-01
-3.52447182e-01 -2.42331717e-02 8.39323223e-01 -4.81735110e-01
-1.17966735e+00 -2.62987792e-01 1.38758153e-01 4.48833019e-01
5.94138214e-03 1.92831196e-02 -1.44898370e-01 2.46104047e-01
1.24035561e+00 -2.09214360e-01 -1.07806828e-02 2.88404793e-01
7.04641461e-01 3.19650710e-01 5.76225400e-01 -1.06154299e+00
7.08974481e-01 3.72810423e-01 3.23931389e-02 -4.59561348e-01
-5.46289027e-01 3.18377048e-01 -1.20588534e-01 -9.63025630e-01
1.26223338e+00 -6.12868607e-01 -1.35856152e+00 4.90264356e-01
-9.89267886e-01 -9.08044517e-01 -8.14350963e-01 3.13679487e-01
-1.20445716e+00 -2.97727827e-02 -5.29228568e-01 -9.17250514e-01
-1.29969940e-01 -9.22008157e-01 5.66665113e-01 1.80193022e-01
-1.04869878e+00 -8.45843196e-01 7.29739517e-02 5.56465089e-01
4.18958664e-01 8.77716601e-01 7.96875060e-01 -9.73069251e-01
5.22232093e-02 1.59873098e-01 1.53051585e-01 1.32029340e-01
3.19726095e-02 -1.02756791e-01 -5.94805717e-01 2.52670258e-01
-9.52338278e-02 -9.64298189e-01 1.72105536e-01 1.27146721e-01
4.44810152e-01 -9.46166098e-01 1.14146203e-01 4.70766611e-02
7.71862268e-01 5.36958575e-01 7.15652943e-01 6.25634074e-01
6.05686188e-01 1.27861369e+00 7.20590413e-01 8.32538188e-01
6.36710227e-01 6.05473578e-01 8.17745388e-01 3.00064504e-01
3.30077261e-01 -2.22313195e-01 5.34042954e-01 -1.24036007e-01
-6.47266388e-01 -1.90187365e-01 -9.22633767e-01 5.56316853e-01
-2.17887592e+00 -1.54266977e+00 7.90405497e-02 1.61450040e+00
1.18576646e+00 2.38646716e-01 4.82471704e-01 -2.83566654e-01
5.15046358e-01 6.28781915e-01 -7.14139938e-01 -9.57727134e-01
-1.12540931e-01 -2.94705898e-01 1.95483878e-01 7.69369781e-01
-8.62077475e-01 1.33263409e+00 5.86392498e+00 5.88629723e-01
-3.59606951e-01 9.42456946e-02 9.61122692e-01 -7.38674283e-01
-5.09463906e-01 -2.18406007e-01 1.67387441e-01 1.26873970e-01
6.67978764e-01 -8.17993820e-01 9.18313205e-01 9.16674495e-01
4.64385986e-01 -1.53356627e-01 -1.38431132e+00 7.00137377e-01
-1.09500200e-01 -1.31381953e+00 -3.53918612e-01 1.38325259e-01
9.04661238e-01 -2.28821144e-01 -1.81507438e-01 4.62462783e-01
1.40798879e+00 -1.31733680e+00 1.20575655e+00 2.89356709e-01
4.30435270e-01 -8.60603034e-01 7.08629489e-01 4.54717904e-01
-5.12675643e-01 -1.36577502e-01 1.85599737e-02 -8.65632832e-01
2.30848685e-01 -1.76170155e-01 -9.05229449e-01 -2.46629506e-01
4.35042590e-01 6.59342945e-01 -1.12672903e-01 7.98666850e-02
-4.51634437e-01 1.72295049e-01 -3.98300625e-02 -6.84089303e-01
4.84245837e-01 -2.12054387e-01 9.64062691e-01 8.58659923e-01
-1.31582081e-01 5.26930749e-01 -1.45056501e-01 9.84940588e-01
-3.82963866e-01 1.13266982e-01 -1.04300725e+00 -2.02957615e-01
1.92919508e-01 9.44597006e-01 -5.11388421e-01 -4.10201848e-01
1.40268624e-01 6.89592957e-01 2.69211203e-01 -8.83326605e-02
-1.15832543e+00 -3.73591259e-02 1.24185550e+00 1.91311575e-02
-6.58603668e-01 1.84710070e-01 -3.89045298e-01 -7.97249854e-01
-4.47595030e-01 -1.26412964e+00 2.91475803e-01 -9.18748796e-01
-1.20834506e+00 3.66623282e-01 1.63499787e-01 -8.40670824e-01
-7.34173238e-01 -2.58689493e-01 -5.53157032e-01 2.46623203e-01
-5.92098236e-01 -1.01813781e+00 9.72505957e-02 3.21206778e-01
2.90861428e-01 -3.91896784e-01 6.04895532e-01 -4.41623718e-01
-4.45539467e-02 1.83259174e-01 -3.99800628e-01 1.51964203e-01
5.41735768e-01 -1.29569888e+00 5.40464878e-01 5.51307082e-01
-4.08840090e-01 3.44755381e-01 1.19931614e+00 -8.51505995e-01
-8.14198911e-01 -7.55107105e-01 8.66093218e-01 -6.92644596e-01
9.10729408e-01 -2.89068222e-01 -4.35429215e-01 7.30236530e-01
5.19242346e-01 -4.85556304e-01 6.53888643e-01 5.91790164e-03
-3.05413842e-01 3.26236397e-01 -1.63511360e+00 1.61418629e+00
1.69947743e+00 -3.94391358e-01 -1.00187206e+00 4.84185189e-01
1.04819524e+00 -2.91962266e-01 -6.09760165e-01 9.93522536e-03
4.03743953e-01 -1.21957088e+00 9.10883665e-01 -1.29116404e+00
1.18197942e+00 -2.12681107e-02 -2.78618306e-01 -1.70091891e+00
-2.56290227e-01 -9.97342646e-01 3.15568447e-01 8.97188783e-01
1.23984933e-01 -6.30176485e-01 7.06615508e-01 1.25500679e+00
7.02442368e-03 -4.87550974e-01 -1.04461873e+00 -5.26441276e-01
2.12761745e-01 -5.52345097e-01 8.69924426e-01 1.35192561e+00
6.92200124e-01 3.75418097e-01 -7.46477127e-01 -6.39122963e-01
7.38916874e-01 -5.70421875e-01 8.35988104e-01 -9.22394812e-01
-3.14512193e-01 -9.18496192e-01 -2.81147182e-01 -2.19081059e-01
5.36015987e-01 -7.59212315e-01 2.73144245e-01 -1.54331493e+00
3.38589162e-01 -1.72064036e-01 2.30209485e-01 7.12980866e-01
-1.57619134e-01 9.34541151e-02 6.93900526e-01 -1.47333428e-01
-1.00088406e+00 7.08908498e-01 1.46415913e+00 -1.79678649e-01
-2.13246107e-01 -5.69956541e-01 -1.25574100e+00 1.10161316e+00
8.35955083e-01 -6.37136400e-01 -4.74980086e-01 5.75474091e-02
9.22841668e-01 1.62787020e-01 3.94434363e-01 -6.80585325e-01
-2.13721886e-01 -1.20386291e+00 -2.51042545e-01 6.83300912e-01
3.16676468e-01 -6.50737286e-01 2.27861553e-01 9.20107126e-01
-1.16669738e+00 -2.24327311e-01 -2.81910300e-01 3.04726303e-01
3.50083232e-01 -1.31348446e-01 9.60522652e-01 -3.07339936e-01
-6.11881256e-01 -3.39143664e-01 -7.67741859e-01 4.63596255e-01
1.73168480e+00 -5.82767650e-02 -8.56477201e-01 -9.84549344e-01
-4.68885690e-01 4.34775889e-01 7.07203269e-01 4.57535803e-01
4.05395776e-01 -1.38427842e+00 -1.22151506e+00 -4.96321261e-01
2.70114124e-01 -3.30683500e-01 2.45402992e-01 3.11962485e-01
-7.88809597e-01 -1.81436285e-01 -8.08727145e-01 1.13043740e-01
-1.23395348e+00 2.58916229e-01 6.86426163e-01 -3.73727322e-01
-5.92916608e-01 1.06374526e+00 4.10566390e-01 -6.70898557e-01
-1.48645550e-01 -1.35776708e-02 -6.11855865e-01 1.03597626e-01
4.69157606e-01 5.14118791e-01 -6.58141673e-01 -1.01289177e+00
-4.68637258e-01 -1.25486121e-01 2.88771987e-01 -4.06073451e-01
1.54708934e+00 2.51449585e-01 -2.66647667e-01 1.17888577e-01
6.62604928e-01 -8.08622837e-02 -1.30078888e+00 8.65837187e-02
-1.60327964e-02 -6.16835535e-01 -4.84125316e-01 -1.02604508e+00
-5.53784907e-01 1.46956354e-01 -1.63869649e-01 4.52714413e-01
7.54675508e-01 1.05178498e-01 5.66600800e-01 6.25973642e-01
6.21786952e-01 -1.57129860e+00 6.35173678e-01 7.02338278e-01
1.22655153e+00 -1.10880351e+00 -4.64020930e-02 1.76007766e-02
-1.68427479e+00 5.53974271e-01 9.84339535e-01 -3.58456552e-01
2.57654991e-02 1.33481890e-01 5.60310669e-03 -2.49112800e-01
-1.13930368e+00 -1.13176584e-01 -2.44298652e-01 1.00321662e+00
2.74786472e-01 7.55206406e-01 -5.06639421e-01 5.66478014e-01
-7.73399711e-01 -3.10711324e-01 1.08261502e+00 6.47919774e-01
-4.01429772e-01 -7.89921045e-01 -2.39291728e-01 6.23468637e-01
-2.92697906e-01 3.16940024e-02 -1.10099256e+00 6.09866023e-01
2.17195481e-01 1.15451765e+00 -4.44167480e-02 -6.49443805e-01
3.34569156e-01 -2.64718413e-01 2.30307758e-01 -6.14511728e-01
-1.33556926e+00 -4.90518481e-01 9.36538160e-01 -9.08778429e-01
-5.44503987e-01 -8.09725881e-01 -1.75341177e+00 -8.38444471e-01
4.85651374e-01 -1.32431611e-01 -8.30240026e-02 1.03333437e+00
-1.27637088e-01 5.64695597e-01 2.61290997e-01 -6.42430246e-01
-5.25841355e-01 -6.24567032e-01 -5.00989854e-01 9.65116203e-01
5.36769070e-02 -6.54187322e-01 -2.57833481e-01 -1.62793661e-03] | [11.641067504882812, 8.019150733947754] |
3cc4cb15-e3e6-4c9c-a937-37238ab069d3 | culturebert-fine-tuning-transformer-based | 2212.00509 | null | https://arxiv.org/abs/2212.00509v2 | https://arxiv.org/pdf/2212.00509v2.pdf | CultureBERT: Fine-Tuning Transformer-Based Language Models for Corporate Culture | This paper introduces supervised machine learning to the literature measuring corporate culture from text documents. We compile a unique data set of employee reviews that were labeled by human evaluators with respect to the information the reviews reveal about the firms' corporate culture. Using this data set, we fine-tune state-of-the-art transformer-based language models to perform the same classification task. In out-of-sample predictions, our language models classify 16 to 28 percent points more of employee reviews in line with human evaluators than traditional approaches of text classification. We make our models publicly available. | ['Stefan Pasch', 'Sebastian Koch'] | 2022-12-01 | null | null | null | null | ['culture'] | ['speech'] | [-1.52161986e-01 3.53535831e-01 -9.08506989e-01 -7.22907424e-01
-1.13778222e+00 -7.30599642e-01 8.26180398e-01 4.99539942e-01
-4.85386431e-01 5.84039092e-01 5.53032219e-01 -7.20842481e-01
4.02590007e-01 -6.71393514e-01 -3.29741210e-01 7.92191625e-02
7.08567321e-01 7.11901128e-01 -4.88752276e-01 -3.49354625e-01
7.56933033e-01 -6.07318394e-02 -9.04099941e-01 9.33065772e-01
5.24202704e-01 8.82767439e-01 -2.74501145e-01 7.05939770e-01
-5.11821568e-01 1.65271592e+00 -5.23153186e-01 -1.10828674e+00
-1.70504913e-01 -1.73214316e-01 -9.44495142e-01 2.08816111e-01
5.68583071e-01 1.28023759e-01 1.18511736e-01 9.16791260e-01
-5.22894040e-02 -5.52799284e-01 9.62526679e-01 -6.23943388e-01
-1.34210420e+00 1.00299895e+00 -5.23652673e-01 4.62480396e-01
2.46512726e-01 -3.16174328e-01 1.70539713e+00 -1.17261851e+00
6.83076620e-01 1.26043308e+00 7.48175979e-01 1.30019277e-01
-1.16620886e+00 -7.18816340e-01 2.59579092e-01 -4.43421155e-01
-7.87670076e-01 -4.76881921e-01 8.50316167e-01 -1.10462677e+00
1.24966514e+00 -1.87707946e-01 5.17461658e-01 1.24611092e+00
8.97981584e-01 2.74640203e-01 1.33910692e+00 -7.34583557e-01
-7.21037239e-02 9.50389028e-01 6.52393997e-01 8.59471381e-01
3.87891233e-01 -3.90379757e-01 -5.96441567e-01 -3.89753222e-01
1.94640920e-01 -2.27821574e-01 4.96531636e-01 2.74653938e-02
-9.22581315e-01 1.16515350e+00 -1.07471757e-01 3.95200610e-01
-2.37071067e-01 8.89552459e-02 5.71848214e-01 5.66452801e-01
1.25559580e+00 8.81401479e-01 -7.50986576e-01 -3.80648583e-01
-6.14414513e-01 -1.25457898e-01 1.15129888e+00 8.43337178e-01
6.47729337e-01 2.71342159e-03 2.47633144e-01 1.03723550e+00
3.68969232e-01 7.37967491e-01 6.39570177e-01 -1.02461541e+00
6.94572210e-01 9.08483863e-01 4.94720675e-02 -1.22563028e+00
-3.01738709e-01 -9.82609510e-01 -4.06568676e-01 -4.17046577e-01
-2.06515074e-01 -2.63432682e-01 -3.57383281e-01 8.86049688e-01
-6.65242672e-01 -7.14930117e-01 2.29321178e-02 2.20555678e-01
6.67442083e-01 4.42926794e-01 9.51756313e-02 -3.61908555e-01
1.37505507e+00 -1.08663297e+00 -8.74188721e-01 -8.43059242e-01
1.39662218e+00 -8.47926319e-01 1.18807387e+00 6.48081541e-01
-7.90632486e-01 -5.79299808e-01 -9.03700769e-01 -1.04781903e-01
-5.13860524e-01 3.98805827e-01 7.62324035e-01 8.86107206e-01
-9.01179790e-01 2.68353224e-01 -3.83046865e-01 -2.19714597e-01
2.56914675e-01 3.59849967e-02 -2.34630525e-01 2.81125724e-01
-9.16465282e-01 1.10287893e+00 -2.52530247e-01 -2.82131433e-01
-4.64433253e-01 -4.68413532e-01 -8.38795602e-01 -1.26175225e-01
1.06667588e-02 -1.68910623e-01 1.92194831e+00 -1.18145394e+00
-9.89281774e-01 1.49346340e+00 -5.04966140e-01 -3.61992896e-01
-1.46941282e-02 -3.21560651e-01 -6.50622070e-01 -3.50379437e-01
5.43795168e-01 -1.49452537e-01 5.58293283e-01 -1.31796658e+00
-7.14327812e-01 -4.34835851e-01 -2.16532096e-01 -4.01962787e-01
-1.02270794e+00 4.46848720e-01 4.62121330e-02 -5.06934881e-01
-3.80270153e-01 -8.85583639e-01 -1.63754076e-01 -1.00846279e+00
-4.28315431e-01 -2.08585843e-01 4.27074105e-01 -5.52565873e-01
1.55989087e+00 -1.72236490e+00 -4.09958482e-01 2.58236855e-01
4.53648597e-01 -3.13245468e-02 -1.49823204e-01 5.43754697e-01
-2.51737505e-01 8.51402581e-01 6.01019263e-01 -7.00778484e-01
2.70860512e-02 -2.53665805e-01 -7.55213082e-01 1.84208632e-01
2.62286693e-01 8.62413883e-01 -6.95229292e-01 -5.96356511e-01
-3.97809744e-01 -6.92102835e-02 -4.53742862e-01 -1.36678725e-01
1.02102101e-01 -2.19858944e-01 -6.31081402e-01 7.76071846e-01
-1.62674949e-01 -6.27487302e-01 5.40770888e-01 1.32449955e-01
-2.11889341e-01 8.83039892e-01 -4.18936601e-03 8.83645177e-01
-6.94628656e-01 9.61811125e-01 -2.61226684e-01 -1.08149290e+00
1.56429243e+00 1.30472630e-01 1.70489132e-01 -1.00934017e+00
2.51691908e-01 3.25273156e-01 -2.48790719e-02 -2.34610572e-01
9.15408254e-01 -6.02863014e-01 -7.34463573e-01 7.13919997e-01
-1.22240454e-01 -1.67576522e-01 1.59377202e-01 7.62381628e-02
1.04822862e+00 -4.86469030e-01 4.90916103e-01 -5.08440197e-01
5.13048351e-01 4.79096830e-01 4.81712937e-01 6.17767453e-01
-2.97914352e-02 1.18341640e-01 8.66826355e-01 -6.37233436e-01
-1.19690537e+00 -3.57364714e-01 2.20859330e-02 1.52154791e+00
-8.01138282e-01 -7.84252226e-01 -4.92772818e-01 -7.60452330e-01
1.32020310e-01 8.08415413e-01 -8.42574179e-01 -8.22992176e-02
-3.34390432e-01 -4.88119751e-01 3.79518151e-01 7.01038539e-01
-1.54925976e-02 -7.28351533e-01 -1.36783533e-02 2.63773441e-01
-4.57016751e-02 -1.28912830e+00 -5.27011991e-01 5.99025846e-01
-7.38857746e-01 -1.03421235e+00 -8.73066187e-02 -1.04837477e+00
4.14485931e-01 1.27641112e-01 1.81178880e+00 -1.49762258e-01
1.02757096e-01 4.29501265e-01 -2.48324051e-01 -1.05500114e+00
-7.88464189e-01 7.09507942e-01 1.15049649e-02 -3.70678037e-01
1.30389380e+00 -5.81647828e-02 3.75953168e-01 -3.90369706e-02
-1.42876968e-01 -2.84692168e-01 4.46135372e-01 8.88744831e-01
3.52469295e-01 1.49432644e-01 8.67895067e-01 -1.75472212e+00
1.27234328e+00 -4.12165552e-01 -3.92616540e-01 3.07003438e-01
-1.27025867e+00 7.00246468e-02 6.88533485e-01 -2.88638681e-01
-9.48808968e-01 -3.35701197e-01 3.66877884e-01 6.36941331e-05
4.94746774e-01 1.13215268e+00 7.36855984e-01 2.91596442e-01
8.26195240e-01 -2.22289205e-01 -1.51134074e-01 -3.00109863e-01
-4.54559512e-02 1.09800231e+00 2.92563468e-01 -5.08994222e-01
7.45751262e-01 9.71524194e-02 -6.16373122e-01 -3.83585125e-01
-1.75610948e+00 -5.49092948e-01 -9.00941253e-01 -2.02441812e-01
9.00044560e-01 -1.22877359e+00 -4.41208750e-01 6.67282194e-02
-8.95104706e-01 -1.01365656e-01 -2.74639321e-03 4.22553360e-01
-4.04987633e-01 -2.17864245e-01 -1.03910959e+00 -1.10126615e+00
-6.14933312e-01 -7.26841211e-01 1.17661965e+00 -1.24541871e-01
-8.44487250e-01 -1.43930411e+00 4.13980454e-01 9.40715075e-01
1.66217044e-01 -2.32870132e-01 1.21884882e+00 -1.01307929e+00
3.20690304e-01 -5.79329908e-01 5.46042882e-02 7.17190087e-01
2.59772748e-01 2.41914585e-01 -9.08598185e-01 -1.18267365e-01
3.71764004e-02 -7.85177350e-01 8.11881006e-01 1.01490408e-01
8.19557428e-01 -3.30921650e-01 -3.31072927e-01 -1.17029082e-02
1.25661576e+00 2.24760354e-01 2.12358221e-01 8.97025645e-01
6.29128397e-01 8.22912157e-01 7.63996422e-01 6.51565790e-01
6.41981363e-01 6.74496219e-02 -2.21693024e-01 -9.92530212e-03
6.24716163e-01 -4.01779503e-01 5.63947618e-01 1.46839154e+00
1.06166266e-01 -4.56725154e-03 -1.28812695e+00 6.26485527e-01
-1.52385545e+00 -8.94237339e-01 -1.44462883e-01 1.37952793e+00
9.39182758e-01 9.04256642e-01 1.08513802e-01 -5.13507938e-03
5.68013728e-01 2.48199597e-01 -1.17634952e-01 -1.15800965e+00
-2.07626633e-02 1.77944005e-01 7.05387294e-01 4.77337509e-01
-1.10241878e+00 9.36344445e-01 7.65624332e+00 1.36142895e-01
-9.86139834e-01 -3.13867703e-02 1.11914635e+00 -9.52382162e-02
-5.58481753e-01 5.13670407e-03 -1.27721930e+00 2.63567120e-01
1.58426964e+00 -5.14609098e-01 1.46651715e-01 1.23711205e+00
3.07841927e-01 2.97558844e-01 -1.21603060e+00 7.14107692e-01
4.78578717e-01 -1.42043304e+00 1.06119877e-02 4.77803528e-01
9.70434666e-01 -7.68346414e-02 3.28118473e-01 7.38088310e-01
5.93137503e-01 -1.17196083e+00 8.36512923e-01 5.48968554e-01
6.93431377e-01 -8.41506839e-01 1.15824878e+00 5.08250773e-01
-7.44012713e-01 -5.56729972e-01 -4.18750435e-01 -4.97025907e-01
-1.33088186e-01 8.54867399e-01 -1.10161614e+00 -1.00549035e-01
6.40987098e-01 1.06977952e+00 -1.03675222e+00 -2.01475501e-01
2.16175526e-01 1.03307950e+00 5.75478613e-01 -3.92317533e-01
2.46890292e-01 -2.58838236e-01 -3.53619069e-01 1.54149818e+00
2.80200839e-01 -7.71937445e-02 2.14726210e-01 7.96436131e-01
-5.81600726e-01 3.21772605e-01 -1.07858193e+00 -1.01953733e+00
4.52440679e-01 1.29145038e+00 -2.90464371e-01 -5.13615966e-01
-6.74415350e-01 1.94734842e-01 5.75968206e-01 3.65179442e-02
-2.89381653e-01 -4.36051160e-01 3.14214557e-01 3.25404018e-01
1.80875286e-02 -1.94722444e-01 -8.23022008e-01 -9.63776469e-01
-3.95757854e-02 -1.14857924e+00 9.87495333e-02 -7.79817700e-01
-1.66329694e+00 4.44041312e-01 -5.57904363e-01 -8.61335695e-01
-5.55632532e-01 -1.13078570e+00 -5.10524869e-01 7.72028029e-01
-1.29884470e+00 -1.04230344e+00 -1.13605903e-02 -9.99533385e-02
4.29717034e-01 -9.00292039e-01 9.92245972e-01 1.54519990e-01
-3.78053755e-01 5.06454170e-01 -1.41784400e-01 6.17585957e-01
1.09368169e+00 -1.29471219e+00 5.32647550e-01 2.21995533e-01
2.05409285e-02 1.00477219e+00 4.74149346e-01 -7.18938410e-01
-1.11137414e+00 -1.04304874e+00 1.77791691e+00 -9.75969374e-01
1.25661969e+00 -3.46260667e-01 -4.03764635e-01 1.12603116e+00
3.49208772e-01 -4.69632119e-01 1.36474013e+00 7.84705102e-01
-7.56854057e-01 -1.72739252e-01 -4.56549078e-01 2.59128958e-01
5.63206434e-01 -1.32835829e+00 -9.56891477e-01 4.97966617e-01
6.37406111e-01 1.89005390e-01 -1.23290467e+00 9.25967991e-02
5.68990648e-01 -5.47921240e-01 3.82071614e-01 -1.14857149e+00
1.27635586e+00 4.57677305e-01 -1.76226243e-01 -1.21138823e+00
-8.56074989e-01 -1.32462204e-01 3.14234644e-01 1.24841201e+00
1.02614343e+00 -5.84697008e-01 5.45491755e-01 4.98592705e-01
-1.70468286e-01 -8.09674621e-01 -1.83119312e-01 -4.08844709e-01
5.33126295e-01 -3.82192522e-01 1.28558174e-01 1.28537071e+00
4.44155872e-01 1.14622247e+00 -1.19897403e-01 -6.95820689e-01
2.26560012e-01 1.12927340e-01 8.18946958e-01 -1.66556942e+00
-6.96238205e-02 -4.36323464e-01 2.84077898e-02 -4.80510235e-01
8.49515259e-01 -9.21701312e-01 -2.65150398e-01 -1.67255592e+00
6.89990401e-01 -1.18186064e-01 -2.59791493e-01 5.10730624e-01
2.00744852e-01 1.41287863e-01 -6.50986135e-02 2.18761042e-01
-6.87541366e-01 1.36703804e-01 1.04291785e+00 -5.06402075e-01
1.01741873e-01 -1.34548515e-01 -1.73955750e+00 8.96604240e-01
9.07056928e-01 -4.76769626e-01 -3.03460389e-01 -4.83674467e-01
7.55787075e-01 -6.10111319e-02 -7.28063658e-02 -4.05197442e-01
-5.37526496e-02 -2.95819521e-01 6.10903203e-01 -5.70672393e-01
2.12024271e-01 -3.20866734e-01 -4.46938843e-01 2.49416023e-01
-1.09138775e+00 8.34264576e-01 -8.80173221e-03 3.17712098e-01
-4.52456772e-01 -3.43105197e-01 4.96299028e-01 -4.19512600e-01
1.87629312e-02 -4.37255025e-01 -8.81761849e-01 4.05204356e-01
5.11631846e-01 1.71822339e-01 -9.34423327e-01 -7.45847523e-01
-3.47082555e-01 -1.80380166e-01 3.17525238e-01 5.82290828e-01
1.38982743e-01 -1.15259039e+00 -8.14693749e-01 -1.86318204e-01
4.61878151e-01 -8.70291650e-01 -5.61248243e-01 6.21400893e-01
-1.17370009e-01 1.34853911e+00 1.54627651e-01 -9.93867517e-02
-1.22153080e+00 4.76884574e-01 8.80714804e-02 -6.88794911e-01
7.17748031e-02 6.30555212e-01 -5.41549586e-02 -6.09955311e-01
-2.14614019e-01 -4.31591332e-01 -4.66494799e-01 4.75479633e-01
3.70899439e-01 -2.45905388e-02 1.53611436e-01 -4.95437741e-01
-2.69895703e-01 3.94902706e-01 -5.10158360e-01 -2.25397944e-01
1.30452359e+00 2.82578263e-02 -4.35767025e-01 1.36037791e+00
1.55752742e+00 4.99817312e-01 -1.25892088e-01 -4.45876807e-01
6.15784526e-01 -1.69181004e-01 1.33517146e-01 -8.22458923e-01
-6.95130825e-01 8.26757789e-01 7.53363594e-02 5.14219403e-01
4.64357018e-01 -3.36480290e-02 2.93904722e-01 9.99721646e-01
2.92603701e-01 -1.84188676e+00 5.32812953e-01 1.00088525e+00
5.17423689e-01 -1.45512831e+00 1.60779208e-01 -2.47976646e-01
-1.05570960e+00 1.03794634e+00 7.50776410e-01 -1.79831967e-01
8.71779680e-01 4.47277278e-01 4.49204892e-01 -5.13375819e-01
-1.51642442e+00 4.03502584e-01 3.13426465e-01 2.62200773e-01
1.33668828e+00 1.88671142e-01 -3.39629591e-01 1.04111230e+00
-8.62219751e-01 -1.42070845e-01 8.99398506e-01 7.27273405e-01
-6.77983046e-01 -1.07387555e+00 -1.32069170e-01 9.62172985e-01
-1.01083279e+00 -4.72830743e-01 -1.03963423e+00 6.56347871e-01
-4.54552382e-01 1.26970923e+00 1.24506377e-01 -8.01123321e-01
2.79052675e-01 3.49074036e-01 -1.71213314e-01 -1.27476180e+00
-8.07204306e-01 1.19369067e-01 9.72341895e-01 -2.82036271e-02
-4.85739648e-01 -8.38081479e-01 -1.06358445e+00 -1.72516122e-01
-1.30894363e-01 3.84487867e-01 5.24745405e-01 1.07549047e+00
2.15108186e-01 6.07624948e-01 9.35050130e-01 -1.23179339e-01
-8.53966773e-01 -1.17966294e+00 -5.15255988e-01 1.10381298e-01
8.80859792e-02 -2.36107185e-01 -4.51643080e-01 4.23287898e-01] | [11.062344551086426, 7.194669246673584] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.