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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
20d448d7-37bf-4092-a7ab-2768c9b7b997 | trajectory-flow-map-graph-based-approach-to | 2212.02927 | null | https://arxiv.org/abs/2212.02927v1 | https://arxiv.org/pdf/2212.02927v1.pdf | Trajectory Flow Map: Graph-based Approach to Analysing Temporal Evolution of Aggregated Traffic Flows in Large-scale Urban Networks | This paper proposes a graph-based approach to representing spatio-temporal trajectory data that allows an effective visualization and characterization of city-wide traffic dynamics. With the advance of sensor, mobile, and Internet of Things (IoT) technologies, vehicle and passenger trajectories are being increasingly collected on a massive scale and are becoming a critical source of insight into traffic pattern and traveller behaviour. To leverage such trajectory data to better understand traffic dynamics in a large-scale urban network, this study develops a trajectory-based network traffic analysis method that converts individual trajectory data into a sequence of graphs that evolve over time (known as dynamic graphs or time-evolving graphs) and analyses network-wide traffic patterns in terms of a compact and informative graph-representation of aggregated traffic flows. First, we partition the entire network into a set of cells based on the spatial distribution of data points in individual trajectories, where the cells represent spatial regions between which aggregated traffic flows can be measured. Next, dynamic flows of moving objects are represented as a time-evolving graph, where regions are graph vertices and flows between them are treated as weighted directed edges. Given a fixed set of vertices, edges can be inserted or removed at every time step depending on the presence of traffic flows between two regions at a given time window. Once a dynamic graph is built, we apply graph mining algorithms to detect change-points in time, which represent time points where the graph exhibits significant changes in its overall structure and, thus, correspond to change-points in city-wide mobility pattern throughout the day (e.g., global transition points between peak and off-peak periods). | ['Marty Papamanolis', 'Sanghyung Ahn', 'Jonathan Corcoran', 'Kai Zheng', 'Jiwon Kim'] | 2022-12-06 | null | null | null | null | ['graph-mining'] | ['graphs'] | [-2.58732438e-01 -2.07581565e-01 -4.04112279e-01 2.86685582e-02
-3.15044299e-02 -7.45414317e-01 6.12934947e-01 7.31901228e-01
1.27472296e-01 5.16305327e-01 3.10995698e-01 -7.39001811e-01
-6.33424640e-01 -1.63196397e+00 -2.94380724e-01 -4.11631554e-01
-7.43048131e-01 4.73995090e-01 6.89099252e-01 -1.98993325e-01
-9.65966582e-02 9.33232129e-01 -1.38872433e+00 -6.77974150e-02
9.26412344e-01 5.56139767e-01 -1.63282290e-01 5.20545065e-01
-4.04255241e-01 3.85437608e-01 -4.23478723e-01 -1.39858514e-01
1.54517055e-01 -2.18548417e-01 -4.11966860e-01 2.90358901e-01
-2.40543395e-01 2.76373357e-01 -8.55453432e-01 8.82675767e-01
-7.91314915e-02 4.59962994e-01 2.38589406e-01 -1.74515700e+00
-2.72948325e-01 4.22976613e-01 -5.02026200e-01 6.57651365e-01
3.21425050e-01 2.75870800e-01 7.68202007e-01 -2.43354991e-01
6.61802411e-01 1.17693949e+00 3.17449838e-01 -1.65016130e-01
-1.21276188e+00 -4.14606601e-01 5.44958949e-01 3.96112293e-01
-1.49887538e+00 -4.24161218e-02 9.93150413e-01 -6.21031702e-01
6.91091299e-01 6.56296074e-01 1.21822882e+00 4.21086162e-01
2.09931239e-01 3.21540594e-01 4.07637298e-01 2.05687001e-01
2.70954311e-01 -3.66406798e-01 2.48434603e-01 6.02612257e-01
3.73478353e-01 -1.16916887e-01 -8.53724256e-02 -1.74614251e-01
2.09300280e-01 5.45540810e-01 -3.16665545e-02 -2.69111693e-01
-1.00271237e+00 5.72034061e-01 9.03158486e-01 4.10211504e-01
-5.47049999e-01 9.32048783e-02 3.55429381e-01 2.71727920e-01
6.31717026e-01 -2.19558999e-01 -4.16524597e-02 -2.56312460e-01
-8.71876955e-01 2.78735101e-01 4.18434530e-01 7.21875131e-01
1.13484967e+00 -9.55747440e-02 -1.41978383e-01 4.35113549e-01
-2.72312220e-02 6.58284366e-01 4.88274805e-02 -4.79578882e-01
6.78634524e-01 1.57987618e+00 -1.13221124e-01 -1.64711845e+00
-5.63774943e-01 -1.98817044e-01 -8.56389701e-01 -2.97822356e-01
4.46102083e-01 -2.77332123e-03 -6.48210883e-01 1.55787313e+00
7.28859007e-01 6.04451716e-01 -3.00648808e-01 6.03249550e-01
5.51429093e-01 7.94029593e-01 9.31746066e-02 -2.54452527e-01
1.20765424e+00 -3.57453376e-01 -4.93031532e-01 1.48095757e-01
7.60262311e-01 5.40490113e-02 7.37208486e-01 -4.26271528e-01
-8.09763968e-01 -2.94992208e-01 -5.15160263e-01 5.20118058e-01
-9.74242985e-01 -5.76003909e-01 2.15237096e-01 4.98127967e-01
-7.38303363e-01 4.34703022e-01 -7.54761398e-01 -5.02634585e-01
4.56461579e-01 4.60132398e-02 -2.14697912e-01 -1.11912109e-01
-1.12693107e+00 3.44055772e-01 3.07755411e-01 -2.25097854e-02
-4.23877507e-01 -8.54082286e-01 -7.45572984e-01 1.71906009e-01
4.84415025e-01 -4.11587775e-01 5.09757221e-01 -4.74518210e-01
-6.84560537e-01 5.47113657e-01 -4.07894135e-01 -3.61906022e-01
3.83178860e-01 7.28658080e-01 -1.45937765e+00 -6.39802516e-02
1.45713940e-01 -1.27708279e-02 3.70098293e-01 -1.08704257e+00
-1.02137887e+00 -5.28842092e-01 2.42212698e-01 -1.73776403e-01
-2.56450742e-01 -1.55050486e-01 -6.84370875e-01 -3.36373150e-01
1.15308456e-01 -9.97360528e-01 -2.16336355e-01 -5.13260424e-01
-5.64436257e-01 -3.40315759e-01 1.36931884e+00 -3.97224575e-01
2.24666667e+00 -2.01331687e+00 -1.28658429e-01 9.24808502e-01
6.48560345e-01 2.23653466e-01 -9.12400484e-02 8.76724184e-01
1.34090036e-01 3.18167448e-01 -3.23251933e-01 2.56762862e-01
-1.92025840e-01 3.28666508e-01 -5.77900633e-02 4.22355056e-01
-1.96295902e-01 1.11697245e+00 -1.35788047e+00 -1.27499178e-01
6.49569154e-01 1.74020916e-01 -1.56252354e-01 -2.03888804e-01
-2.79898047e-01 5.25941372e-01 -7.03906119e-01 4.16956633e-01
5.79776943e-01 -1.35823876e-01 1.82416812e-01 1.50660664e-01
-6.61105812e-01 1.57788679e-01 -1.08404708e+00 8.33638132e-01
-4.26889509e-01 9.50265110e-01 -2.36849532e-01 -1.03436768e+00
9.95429993e-01 1.19447351e-01 8.65014613e-01 -1.00788963e+00
-1.09536894e-01 -8.69937614e-02 8.15902930e-03 -5.26995182e-01
4.88862097e-01 5.03327727e-01 -2.95189559e-01 5.03839314e-01
-8.30924213e-01 5.22860050e-01 8.01170528e-01 2.18084931e-01
1.61883152e+00 -8.56081605e-01 1.29476562e-01 -1.94603786e-01
6.64556265e-01 2.65282452e-01 4.66245234e-01 2.00743064e-01
-1.93033859e-01 1.02206441e-02 6.69372261e-01 -7.88502157e-01
-8.83468211e-01 -1.25021410e+00 1.41296938e-01 8.93075347e-01
2.19862744e-01 -5.96267998e-01 -3.62766415e-01 -4.88797963e-01
4.11263645e-01 4.48741287e-01 -6.89630926e-01 -2.03841284e-01
-8.68299961e-01 -7.12768435e-01 8.19379836e-02 9.95082110e-02
5.39327145e-01 -8.28222275e-01 -5.21169007e-01 3.29722941e-01
-7.35940039e-02 -9.53301013e-01 -5.29117584e-01 -5.88840187e-01
-8.04399312e-01 -1.44919991e+00 -1.37084484e-01 -3.53197336e-01
8.86869192e-01 8.18214357e-01 9.69489515e-01 1.26293644e-01
-1.00000344e-01 3.92687738e-01 -2.80572921e-01 -1.57716513e-01
-2.71051347e-01 8.00084546e-02 -1.91981778e-01 6.14891410e-01
3.23075205e-01 -9.12657380e-01 -7.60377109e-01 5.98223150e-01
-6.80704236e-01 -1.36068568e-01 -1.26445860e-01 -1.27827302e-01
7.01761842e-01 5.96700490e-01 5.06408334e-01 -7.79618084e-01
8.15582752e-01 -1.34003687e+00 -7.78625727e-01 2.35025018e-01
-5.03013968e-01 -3.18705797e-01 6.43823206e-01 -2.44675994e-01
-4.68335927e-01 -4.91440058e-01 4.78053093e-01 -3.08005810e-01
-9.19368044e-02 9.05137002e-01 -2.53428996e-01 3.32777858e-01
3.36489677e-01 1.71055853e-01 -3.20287079e-01 -1.97699636e-01
5.60047984e-01 6.20969892e-01 3.16271216e-01 -1.56652316e-01
9.48375642e-01 8.07048261e-01 2.95300186e-01 -9.98779655e-01
-1.02414824e-01 -7.85097659e-01 -6.78321123e-01 -8.72062683e-01
5.73335528e-01 -4.42207068e-01 -1.01845562e+00 3.85554247e-02
-5.44271767e-01 -3.71223986e-01 -4.89752352e-01 2.04746723e-01
-1.08113755e-02 1.14604048e-01 7.50028528e-03 -7.96535313e-01
2.90747643e-01 -7.61074007e-01 5.59622407e-01 2.61530429e-01
-3.48992854e-01 -1.56183684e+00 4.15524811e-01 -1.45025060e-01
3.41138273e-01 7.61631846e-01 1.14867544e+00 -3.11462253e-01
-7.82666624e-01 -5.63921511e-01 -2.97569752e-01 -5.18966019e-01
5.59363306e-01 2.80113846e-01 -1.97239026e-01 -2.28544116e-01
-7.94342399e-01 9.53946531e-01 3.78388047e-01 4.97500837e-01
1.18303978e+00 -5.83313942e-01 -9.41623688e-01 4.31037635e-01
1.07246363e+00 5.07639408e-01 5.28099477e-01 1.85591415e-01
9.27858710e-01 7.51902640e-01 1.43197387e-01 2.73830771e-01
7.01873899e-01 8.07026982e-01 5.12642562e-01 9.28290710e-02
-1.08219065e-01 -5.41165709e-01 1.86042972e-02 7.00222552e-01
-2.32541710e-01 -5.66347718e-01 -1.30355752e+00 9.48716164e-01
-2.16181636e+00 -1.34146571e+00 -6.48915172e-01 2.49773002e+00
-2.17737943e-01 1.99071243e-01 7.93193400e-01 2.16933593e-01
1.16307843e+00 5.45849323e-01 -7.65828550e-01 -2.86460578e-01
1.98027529e-02 -3.54884028e-01 6.96284413e-01 4.98450100e-01
-5.52953303e-01 7.26635456e-01 5.51387548e+00 5.46980739e-01
-1.12723553e+00 -6.90377727e-02 4.00473326e-01 -1.29161432e-01
-6.96561873e-01 3.83154079e-02 -2.60185719e-01 8.67600441e-01
1.34741366e+00 -7.25901425e-01 5.71535468e-01 4.70315576e-01
9.06143427e-01 -4.90442403e-02 -6.55242503e-01 8.26562166e-01
-6.05541050e-01 -1.65539658e+00 1.68043077e-01 6.82166159e-01
7.51238644e-01 1.89962655e-01 -2.22467333e-02 7.83316493e-02
5.96202016e-01 -7.43771076e-01 4.48024005e-01 6.82037473e-01
6.52050316e-01 -8.83768439e-01 -3.31334621e-02 3.57232988e-01
-2.17368698e+00 -3.76466125e-01 6.20905198e-02 9.00674518e-03
6.22949183e-01 9.88594055e-01 -5.11029661e-01 8.06503177e-01
4.57304507e-01 9.99139369e-01 -4.32103485e-01 1.17660856e+00
7.37228766e-02 8.02264869e-01 -3.06717336e-01 3.03115901e-02
3.96727949e-01 -8.65846753e-01 9.06838596e-01 1.10432458e+00
4.14282054e-01 1.07738100e-01 2.40832880e-01 8.67368162e-01
-1.10760316e-01 -2.78530329e-01 -1.11511505e+00 -1.81808714e-02
8.44547570e-01 1.26562500e+00 -1.05998063e+00 -2.44570374e-01
-5.09586036e-01 2.64894158e-01 2.37390190e-01 7.06891000e-01
-7.62355804e-01 -2.62606472e-01 1.15838993e+00 7.83255398e-01
7.10860193e-02 -6.40370011e-01 7.00905919e-02 -7.33492911e-01
-7.70552456e-03 -7.34930858e-02 5.26719213e-01 -3.25740516e-01
-1.01329052e+00 3.37882370e-01 1.63026392e-01 -1.36293030e+00
-8.22099950e-03 6.06077462e-02 -1.24881029e+00 5.14049649e-01
-1.20830202e+00 -8.54803681e-01 -4.72798765e-01 9.07621324e-01
6.76523149e-02 1.39182582e-01 2.48271629e-01 3.65944058e-01
-7.72128105e-01 1.08235680e-01 2.98719794e-01 2.66733766e-01
-5.48377573e-01 -8.92340124e-01 8.72218490e-01 9.13995564e-01
3.29055637e-01 4.02084410e-01 2.62743413e-01 -1.04828405e+00
-1.26343906e+00 -1.61129248e+00 1.11168742e+00 -3.81118238e-01
1.11195171e+00 -3.88841718e-01 -8.29409838e-01 6.69486761e-01
-3.75828505e-01 3.52979153e-01 4.89912212e-01 7.17961192e-02
-4.36600074e-02 -4.52928960e-01 -1.01212728e+00 8.82499278e-01
1.38907743e+00 -6.37994409e-01 -8.77303183e-02 3.58882129e-01
5.59951067e-01 9.10315141e-02 -7.18764305e-01 1.80738151e-01
3.04875135e-01 -7.09399879e-01 8.11749458e-01 -6.89232230e-01
-3.45999211e-01 -5.47538161e-01 8.77139941e-02 -1.48491216e+00
-5.71523011e-01 -8.32973838e-01 -2.55097449e-01 1.09157491e+00
2.15379074e-01 -8.42986226e-01 7.52339065e-01 4.89488453e-01
-5.07875830e-02 -6.51594043e-01 -1.12912917e+00 -8.41193795e-01
-3.37570608e-01 -6.58652186e-01 1.35191453e+00 8.09912264e-01
2.07002431e-01 3.85374539e-02 1.07612856e-01 1.60806820e-01
4.77467120e-01 2.79459149e-01 8.56915116e-01 -1.45515108e+00
6.65673554e-01 -8.03795934e-01 -7.44933546e-01 -6.83318377e-01
-1.77368317e-02 -1.10457540e+00 -7.17999578e-01 -1.77950835e+00
-1.80126265e-01 -6.01911724e-01 2.54831053e-02 -3.11382711e-02
2.01722547e-01 -1.18384473e-01 4.05554734e-02 2.44397879e-01
-4.94182438e-01 3.15104663e-01 9.66830373e-01 -1.59936160e-01
-6.51352644e-01 3.74366045e-01 -7.80845061e-02 4.78421092e-01
7.32792974e-01 -2.73283780e-01 -6.93630397e-01 5.77214547e-02
3.89491826e-01 3.08274806e-01 3.58831942e-01 -9.50946748e-01
3.04148287e-01 -4.12409931e-01 -3.39374006e-01 -8.45402062e-01
-9.93919000e-02 -1.20109296e+00 8.80286515e-01 5.61535537e-01
-2.37120539e-02 5.70786476e-01 1.15751587e-01 9.94016051e-01
-2.43462011e-01 6.91800296e-01 3.58584285e-01 1.07086927e-01
-7.21978068e-01 9.18345034e-01 -7.89483190e-01 6.65837675e-02
1.51892257e+00 -7.23570764e-01 -3.51761162e-01 -3.46027464e-01
-8.47961128e-01 6.09332621e-01 4.64016974e-01 6.87769175e-01
4.83778983e-01 -1.67290676e+00 -4.29605186e-01 1.49856016e-01
3.21903080e-01 -2.29287133e-01 5.22385836e-01 7.46192217e-01
-2.05855653e-01 5.40715814e-01 -5.50969318e-02 -6.08577728e-01
-1.11171412e+00 6.51077688e-01 3.60296279e-01 -3.86547863e-01
-9.83041942e-01 9.41278711e-02 -5.40966988e-02 -1.99980289e-01
-2.40166366e-01 -5.91844201e-01 -3.40114474e-01 3.52908999e-01
4.07757074e-01 1.13862813e+00 1.79671068e-02 -1.16320419e+00
-4.13498729e-01 6.26281440e-01 6.11577988e-01 2.14913651e-01
1.09490216e+00 -5.77196240e-01 -6.97989836e-02 7.75384367e-01
1.22314942e+00 9.06516463e-02 -1.11850750e+00 -1.80243596e-01
1.87633172e-01 -6.46537781e-01 -2.15897635e-01 -2.31423855e-01
-1.56889021e+00 5.12119114e-01 2.05438405e-01 1.05748975e+00
9.62654948e-01 1.68547168e-01 1.19562984e+00 -1.02237679e-01
2.70523936e-01 -7.89251328e-01 -3.00517917e-01 4.42476600e-01
3.60193610e-01 -7.49952018e-01 -3.43151957e-01 -4.87640202e-01
-1.99252844e-01 9.60660696e-01 1.68551773e-01 -2.77070183e-04
1.14257538e+00 -1.48136780e-01 -5.87343156e-01 -7.40886450e-01
-4.42582995e-01 -6.48951590e-01 3.98005128e-01 6.04401588e-01
-2.57443815e-01 6.97935164e-01 -7.69585520e-02 -4.80796732e-02
-3.72350901e-01 -2.53653765e-01 2.68696487e-01 6.10574782e-01
-5.57690442e-01 -8.48554254e-01 -1.17438592e-01 6.16217196e-01
1.98656201e-01 4.05924022e-01 -2.38307193e-01 8.49828482e-01
1.29017085e-01 1.16970813e+00 6.82182908e-01 -6.20824575e-01
6.79241538e-01 -1.55335873e-01 -3.65533978e-02 -3.62682790e-01
-3.10748994e-01 -7.39822924e-01 8.99636969e-02 -6.98497593e-01
-1.91817999e-01 -8.29812706e-01 -1.46076083e+00 -1.03191018e+00
1.61501825e-01 1.67591423e-01 3.85024011e-01 8.69357884e-01
7.51632571e-01 5.85904539e-01 9.18464124e-01 -4.60132152e-01
8.42511773e-01 -3.68606985e-01 -7.19359636e-01 6.83669984e-01
4.16781366e-01 -6.38469458e-01 -1.93565607e-01 -2.86979139e-01] | [6.406552791595459, 1.99032461643219] |
779f2f3a-ea44-4ca4-9019-3377d606c1ea | po-elic-perception-oriented-efficient-learned | 2205.14501 | null | https://arxiv.org/abs/2205.14501v1 | https://arxiv.org/pdf/2205.14501v1.pdf | PO-ELIC: Perception-Oriented Efficient Learned Image Coding | In the past years, learned image compression (LIC) has achieved remarkable performance. The recent LIC methods outperform VVC in both PSNR and MS-SSIM. However, the low bit-rate reconstructions of LIC suffer from artifacts such as blurring, color drifting and texture missing. Moreover, those varied artifacts make image quality metrics correlate badly with human perceptual quality. In this paper, we propose PO-ELIC, i.e., Perception-Oriented Efficient Learned Image Coding. To be specific, we adapt ELIC, one of the state-of-the-art LIC models, with adversarial training techniques. We apply a mixture of losses including hinge-form adversarial loss, Charbonnier loss, and style loss, to finetune the model towards better perceptual quality. Experimental results demonstrate that our method achieves comparable perceptual quality with HiFiC with much lower bitrate. | ['Yan Wang', 'Hongwei Qin', 'Xinjie Shi', 'Chenjian Gao', 'Yuan Chen', 'Jixiang Luo', 'Tongda Xu', 'Hongjiu Yu', 'Ziming Yang', 'Dailan He'] | 2022-05-28 | null | null | null | null | ['ms-ssim'] | ['computer-vision'] | [ 5.26008248e-01 -3.80607128e-01 -8.62520635e-02 -1.55987486e-01
-7.64027357e-01 -3.61321390e-01 2.61294156e-01 -5.30696154e-01
-1.56177729e-01 8.70369494e-01 1.93883732e-01 2.50073522e-02
7.87307546e-02 -6.60048008e-01 -9.16818261e-01 -7.66933680e-01
-3.77805810e-03 -3.86698157e-01 1.10059336e-01 2.24252623e-02
1.37507960e-01 1.75892666e-01 -1.29392827e+00 4.62064505e-01
1.19355237e+00 1.47219896e+00 1.71482667e-01 6.61113381e-01
2.60582775e-01 1.13423049e+00 -5.40553451e-01 -8.93873751e-01
3.75184089e-01 -6.20689511e-01 -3.58287036e-01 3.85149159e-02
5.57777047e-01 -6.12826705e-01 -7.50314832e-01 1.40826976e+00
7.07159638e-01 -2.18861088e-01 4.30893451e-01 -1.10792160e+00
-9.24172997e-01 2.43100554e-01 -6.17281139e-01 -1.03194371e-01
2.07006991e-01 4.99281764e-01 5.66413820e-01 -8.72893095e-01
3.87859583e-01 1.38238752e+00 8.06151330e-01 4.89559948e-01
-1.34093642e+00 -8.05301011e-01 -2.67283618e-01 5.29399693e-01
-1.22318888e+00 -5.66634953e-01 9.36216831e-01 -3.27866189e-02
2.99273074e-01 4.41202909e-01 3.56069863e-01 1.17906165e+00
4.80034739e-01 7.58567870e-01 1.42189288e+00 -3.39753419e-01
2.43020505e-01 -1.00897551e-02 -8.17437887e-01 5.24057567e-01
2.10114166e-01 3.63929361e-01 -5.62640548e-01 2.31711298e-01
1.05658031e+00 -9.38235000e-02 -7.14596331e-01 -3.13856214e-01
-1.03383005e+00 3.45691711e-01 6.89569771e-01 -6.48341998e-02
-2.58819312e-01 5.84212542e-01 4.45652753e-01 4.34724778e-01
3.43879879e-01 2.55023777e-01 -2.62753755e-01 -1.66502535e-01
-1.04626429e+00 -4.42700796e-02 4.12245929e-01 9.99813139e-01
4.25041914e-01 3.26627165e-01 -4.18846697e-01 1.23804939e+00
3.50607187e-01 4.83629286e-01 4.97372150e-01 -1.49192476e+00
4.69880432e-01 -1.55413285e-01 -6.89939857e-02 -1.15058875e+00
3.81807685e-01 -6.94297671e-01 -1.27233803e+00 4.57486928e-01
-8.72203857e-02 2.23677009e-01 -8.95803630e-01 1.68561769e+00
-2.06148833e-01 3.91088605e-01 5.76077737e-02 1.04477334e+00
5.26721716e-01 7.32540369e-01 1.97147369e-01 -2.81463116e-01
8.54219913e-01 -1.12774456e+00 -1.02867544e+00 -1.97545990e-01
-3.26182634e-01 -9.97327745e-01 1.06881166e+00 7.23240197e-01
-1.52427220e+00 -9.43662703e-01 -1.28965211e+00 -9.60955992e-02
1.75618917e-01 -1.22731425e-01 5.15435696e-01 7.42865562e-01
-9.14946079e-01 7.77148724e-01 -6.32213771e-01 9.45987403e-02
6.53761864e-01 -3.45748365e-02 1.09053627e-02 -4.98488009e-01
-1.16460299e+00 5.67866385e-01 1.56432793e-01 1.39196785e-02
-1.14489639e+00 -7.51388609e-01 -6.96537673e-01 -6.48653181e-03
2.29926258e-01 -7.61593878e-01 9.32663918e-01 -1.14491427e+00
-1.63386762e+00 6.00850523e-01 1.01717472e-01 -6.21182799e-01
8.64037693e-01 -3.86379212e-01 -6.20113790e-01 2.78568447e-01
-2.47326434e-01 7.08322465e-01 1.22916591e+00 -1.66963851e+00
-2.33481169e-01 -7.42264315e-02 -1.77559674e-01 8.74562785e-02
-2.47303203e-01 -1.66437015e-01 -9.33066249e-01 -1.09578037e+00
-1.43778860e-03 -5.60569644e-01 1.54674975e-02 5.87727964e-01
-1.94694176e-01 5.52518308e-01 9.34066713e-01 -9.64658320e-01
1.31494606e+00 -2.40624785e+00 5.93262799e-02 -1.69622108e-01
1.63087491e-02 3.83394331e-01 -2.47228906e-01 1.00735091e-01
1.57818571e-01 1.81100965e-01 -4.39077467e-01 -3.12776655e-01
-9.15983915e-02 3.53566766e-01 -4.35477734e-01 4.57268476e-01
3.36515233e-02 9.61281359e-01 -7.56124914e-01 -5.56791961e-01
4.91951257e-01 6.96609497e-01 -8.01333487e-01 3.72786701e-01
-6.65037557e-02 5.15648723e-01 -1.04644671e-01 7.41277933e-01
1.08157372e+00 -3.69055495e-02 -7.52917230e-02 -6.30401611e-01
2.61464775e-01 -6.51526079e-02 -7.22261250e-01 1.85432482e+00
-7.05415249e-01 6.91429019e-01 2.34660745e-01 -7.48059809e-01
7.77824879e-01 1.21003702e-01 2.34822154e-01 -1.00857508e+00
-6.50520325e-02 3.04985255e-01 -2.52751559e-01 -4.96356428e-01
2.90861428e-01 -1.09529749e-01 2.89825559e-01 -3.44030559e-01
4.68911417e-02 -3.47153038e-01 -2.37412825e-01 2.86573917e-02
1.00269794e+00 2.65226185e-01 4.84985765e-03 1.52312383e-01
5.83738208e-01 -6.88511491e-01 7.65071929e-01 5.89427590e-01
-4.98104811e-01 1.09218848e+00 2.61117190e-01 -9.64167193e-02
-1.43441451e+00 -1.42328084e+00 -3.45553219e-01 5.62951028e-01
5.08770108e-01 -2.18823522e-01 -7.52282679e-01 -2.08279878e-01
-1.06429651e-01 6.64659441e-01 -1.25586510e-01 -4.22927529e-01
-5.69717228e-01 -5.66819966e-01 6.47941530e-01 2.57624954e-01
1.23531079e+00 -1.05382800e+00 -3.34926426e-01 3.14109981e-01
-3.96015197e-01 -1.26021087e+00 -5.38231730e-01 -1.95365325e-01
-8.49257171e-01 -6.98477805e-01 -1.04920220e+00 -7.39699304e-01
4.77156103e-01 2.14228369e-02 1.23934221e+00 -2.43291855e-02
-4.99332696e-01 2.13404730e-01 -4.31775302e-01 -2.28809267e-02
-4.48408514e-01 -5.69432616e-01 -2.31613412e-01 1.58452526e-01
-2.93662131e-01 -7.87562013e-01 -1.16013241e+00 3.75410199e-01
-1.17739046e+00 1.06671736e-01 7.58671522e-01 1.08764052e+00
8.10245037e-01 4.37282622e-01 2.91974038e-01 -5.22402227e-01
5.01547873e-01 -2.42052168e-01 -3.03181976e-01 1.06942207e-01
-8.52995038e-01 -1.27724379e-01 9.11345601e-01 -4.15111661e-01
-1.16870940e+00 -3.57734859e-01 -1.94303304e-01 -7.22194195e-01
2.35782415e-02 2.20391840e-01 -3.04777652e-01 -4.07569200e-01
4.31931466e-01 6.88795209e-01 -8.52039158e-02 -4.93314266e-01
2.16136381e-01 6.60599828e-01 1.03467715e+00 -5.23233116e-01
8.75535011e-01 4.30228412e-01 -1.26453206e-01 -4.54989314e-01
-6.63957655e-01 -4.04297654e-03 3.49534452e-02 -4.00875747e-01
6.12195134e-01 -1.26596391e+00 -5.06883681e-01 7.30999649e-01
-9.86859977e-01 -5.13803244e-01 -2.22472548e-01 3.64003688e-01
-9.17304516e-01 6.47742987e-01 -9.03379798e-01 -6.46397173e-01
-4.17530179e-01 -1.14000499e+00 8.50197494e-01 1.57483220e-01
4.81168866e-01 -5.01896560e-01 -2.91365117e-01 5.60342371e-01
8.24529707e-01 3.73638719e-01 6.60890698e-01 6.53587878e-01
-8.87560427e-01 7.06675500e-02 -5.46139657e-01 1.02511168e+00
6.07942380e-02 -3.29865009e-01 -9.37952995e-01 -6.04184568e-01
1.52988210e-01 -4.60345745e-01 1.05286670e+00 3.86470944e-01
2.02854705e+00 -5.78791976e-01 9.32205021e-02 1.42092538e+00
1.76807499e+00 2.54240155e-01 1.42621195e+00 2.81990707e-01
5.68922818e-01 6.31112531e-02 5.02895236e-01 3.96679163e-01
6.95917243e-03 7.13987529e-01 7.01317370e-01 -1.50271460e-01
-8.49804699e-01 -5.03511548e-01 5.32782137e-01 8.37067187e-01
-7.95809552e-02 -3.99463624e-01 -3.65973264e-01 3.07669878e-01
-1.56752503e+00 -9.77192640e-01 3.80829275e-01 2.06273222e+00
1.18151534e+00 1.91313550e-01 -4.78039712e-01 2.93089539e-01
5.86782038e-01 4.53032523e-01 -7.60061085e-01 -1.70762330e-01
-3.79044503e-01 1.99856400e-01 9.01350796e-01 2.66551197e-01
-8.92398655e-01 7.64595747e-01 6.25879574e+00 1.19862401e+00
-1.13971019e+00 1.96207315e-01 1.04651892e+00 -2.00185850e-02
-3.28499317e-01 -2.27004707e-01 -4.83676605e-02 9.88035917e-01
7.19026506e-01 6.24295250e-02 9.48938787e-01 7.67190099e-01
1.09873295e-01 6.40392601e-02 -7.46946216e-01 1.32871664e+00
2.17145860e-01 -1.09842145e+00 1.84788182e-01 -2.64221057e-02
8.95868659e-01 -2.65952349e-01 5.43525815e-01 1.86815053e-01
1.49190664e-01 -1.20468366e+00 9.28377390e-01 6.50804996e-01
1.37863350e+00 -7.86734819e-01 6.38335645e-01 -1.62673574e-02
-1.00859737e+00 -6.48217946e-02 -5.31927288e-01 3.43254060e-01
2.46959195e-01 7.01217830e-01 5.20695038e-02 4.94852662e-01
9.56432998e-01 8.45768332e-01 -4.73649740e-01 1.33006132e+00
-3.45860153e-01 7.98732758e-01 8.35583732e-02 4.13868666e-01
6.01950064e-02 -1.34672657e-01 4.98051912e-01 1.03877962e+00
4.25525934e-01 1.40000985e-03 -2.29304180e-01 8.97199512e-01
-3.54919523e-01 -1.64957553e-01 -3.63055319e-01 3.36104631e-01
4.07941669e-01 7.29374230e-01 3.15758116e-05 -1.30957142e-01
-3.32904547e-01 1.60352683e+00 -1.27903968e-01 5.57946086e-01
-9.94632125e-01 -3.15417528e-01 7.93850839e-01 6.52063116e-02
4.16689754e-01 -1.14474587e-01 -4.01242971e-01 -1.14910805e+00
1.95255429e-01 -1.07226574e+00 -1.89957052e-01 -1.03661799e+00
-1.29458737e+00 3.70453894e-01 -3.82925540e-01 -1.43143654e+00
3.16149116e-01 -4.13775980e-01 -3.77543032e-01 7.14549720e-01
-2.01878452e+00 -8.54716182e-01 -5.28322697e-01 6.26634240e-01
6.83799744e-01 -1.39932096e-01 4.29444283e-01 7.65926957e-01
-3.99286151e-01 9.38483953e-01 4.76998448e-01 -3.06859128e-02
1.00208342e+00 -9.51486945e-01 2.45440528e-01 8.65028918e-01
-2.29731381e-01 2.68292993e-01 7.82380164e-01 -3.77798408e-01
-1.53739381e+00 -1.40648806e+00 2.48089135e-01 2.48920545e-01
3.21133763e-01 -9.29154307e-02 -8.60034525e-01 2.42568463e-01
3.20945323e-01 2.74620384e-01 2.69488275e-01 -6.77054524e-01
-5.65288544e-01 -5.81351757e-01 -1.47107613e+00 6.50739551e-01
1.08672416e+00 -5.70153832e-01 -7.55712539e-02 9.84765291e-02
1.01068068e+00 -3.87878329e-01 -8.62895072e-01 6.12278223e-01
6.07536256e-01 -1.33448613e+00 1.45900822e+00 1.42714277e-01
8.62463951e-01 -4.29046363e-01 -4.80803519e-01 -1.24887431e+00
-4.26098049e-01 -7.31969237e-01 -2.10183144e-01 1.19267046e+00
-5.63961342e-02 -3.96043330e-01 4.65243667e-01 7.08882734e-02
-2.05128968e-01 -6.74406767e-01 -9.57174540e-01 -9.79250073e-01
-6.86221421e-02 -3.80841136e-01 5.30561209e-01 6.60092056e-01
-3.69128615e-01 -2.83618450e-01 -9.48770046e-01 -3.03641222e-02
1.14975536e+00 -2.48324469e-01 5.74399531e-01 -5.81032276e-01
-6.10046387e-01 -3.22994590e-01 -5.32526553e-01 -1.33770049e+00
-1.78238988e-01 -3.91675651e-01 9.64819863e-02 -1.23667562e+00
2.54171282e-01 -5.42713761e-01 -5.49991608e-01 1.25644326e-01
-1.63122401e-01 6.68703735e-01 4.18309122e-01 4.50789541e-01
-6.21225297e-01 8.89280438e-01 1.65177023e+00 -4.60820794e-01
2.42919460e-01 -4.45340812e-01 -5.93250513e-01 5.51269114e-01
7.74181664e-01 -1.91227987e-01 -3.64930481e-01 -6.26983762e-01
1.19901426e-01 1.35196641e-01 6.01238072e-01 -1.58846855e+00
2.25660875e-02 -6.75829798e-02 6.93101823e-01 -3.31384897e-01
4.62162554e-01 -9.20463800e-01 3.52113485e-01 6.12090468e-01
-4.79004204e-01 -4.07754183e-01 2.78527532e-02 7.63456404e-01
-5.85269868e-01 1.63387686e-01 1.25447619e+00 -5.55484407e-02
-6.74440026e-01 2.98042148e-01 1.93124250e-01 1.11029763e-02
7.98527539e-01 -2.55017698e-01 -3.31668824e-01 -5.94652414e-01
-1.27332032e-01 -1.34036466e-01 6.57197833e-01 3.56372267e-01
8.66774082e-01 -1.43503797e+00 -7.98614681e-01 1.87553391e-01
-3.00494302e-02 -1.75523221e-01 3.87083501e-01 5.64995646e-01
-8.33995521e-01 1.26367509e-01 -4.31023359e-01 -3.84136468e-01
-1.02410853e+00 6.08312786e-01 2.17025608e-01 -1.07734255e-01
-6.21262133e-01 8.22388232e-01 2.01538846e-01 7.32617686e-03
5.63206673e-01 4.02339175e-02 3.49338472e-01 -6.84020936e-01
4.86087531e-01 3.89816195e-01 -2.27740735e-01 -3.28659147e-01
-6.70744628e-02 6.17111206e-01 8.70826691e-02 1.51119167e-02
1.20809722e+00 -4.43151981e-01 -5.50111420e-02 7.89047256e-02
1.38287473e+00 -1.17180459e-01 -1.75966525e+00 -1.98576882e-01
-7.55401254e-01 -1.00229013e+00 4.10703808e-01 -1.21016502e+00
-1.36249602e+00 7.09631264e-01 1.08086514e+00 -1.42551154e-01
1.72497845e+00 -2.97908068e-01 1.27987027e+00 -3.13738644e-01
5.86394429e-01 -1.00262165e+00 3.31871480e-01 9.82830673e-03
1.18682468e+00 -1.24543858e+00 1.14205718e-01 -4.76923674e-01
-4.93321478e-01 7.06794083e-01 5.30808449e-01 -1.65412143e-01
3.99388701e-01 3.06271613e-01 -4.40482646e-02 4.39895540e-01
-6.43098354e-01 1.82745412e-01 2.05718607e-01 6.76616669e-01
2.90031701e-01 2.89090332e-02 -4.94374245e-01 1.32737771e-01
-1.20874755e-01 2.24369138e-01 3.08881819e-01 5.14759839e-01
-2.35190496e-01 -8.98421288e-01 -5.37720621e-01 2.17814356e-01
-6.99740767e-01 -3.49920779e-01 1.59019187e-01 3.09622794e-01
3.64371270e-01 9.11059320e-01 -1.69170395e-01 -4.59833890e-01
2.43104517e-01 -4.97603267e-01 5.71371853e-01 1.76163003e-01
-2.43657157e-01 1.20920025e-01 -2.75660545e-01 -8.55332792e-01
-3.56444836e-01 -2.90793717e-01 -7.33798087e-01 -4.93683279e-01
-7.01660290e-03 -3.73240352e-01 7.43809223e-01 3.58962029e-01
3.59131038e-01 7.50259399e-01 8.89885426e-01 -7.23495305e-01
-5.12872636e-01 -7.89740562e-01 -5.80259681e-01 7.26424754e-01
5.15734851e-01 -3.45507264e-01 -5.69926023e-01 3.34166318e-01] | [11.334895133972168, -1.7851194143295288] |
f3eb8591-f178-4662-863d-912b5b674725 | unconditional-image-text-pair-generation-with | 2204.07537 | null | https://arxiv.org/abs/2204.07537v2 | https://arxiv.org/pdf/2204.07537v2.pdf | Unconditional Image-Text Pair Generation with Multimodal Cross Quantizer | Although deep generative models have gained a lot of attention, most of the existing works are designed for unimodal generation. In this paper, we explore a new method for unconditional image-text pair generation. We design Multimodal Cross-Quantization VAE (MXQ-VAE), a novel vector quantizer for joint image-text representations, with which we discover that a joint image-text representation space is effective for semantically consistent image-text pair generation. To learn a multimodal semantic correlation in a quantized space, we combine VQ-VAE with a Transformer encoder and apply an input masking strategy. Specifically, MXQ-VAE accepts a masked image-text pair as input and learns a quantized joint representation space, so that the input can be converted to a unified code sequence, then we perform unconditional image-text pair generation with the code sequence. Extensive experiments show the correlation between the quantized joint space and the multimodal generation capability on synthetic and real-world datasets. In addition, we demonstrate the superiority of our approach in these two aspects over several baselines. The source code is publicly available at: https://github.com/ttumyche/MXQ-VAE. | ['Edward Choi', 'Joonseok Lee', 'Sungjin Park', 'Hyungyung Lee'] | 2022-04-15 | null | null | null | null | ['multimodal-generation'] | ['natural-language-processing'] | [ 3.00575554e-01 -1.56826019e-01 -1.19286239e-01 -2.39724547e-01
-1.17564976e+00 -4.81676638e-01 8.36863279e-01 -4.48613286e-01
3.32583179e-04 7.00895190e-01 4.22096640e-01 -2.27133468e-01
3.46262872e-01 -9.04399812e-01 -7.71169305e-01 -8.34668875e-01
5.36108613e-01 1.32676139e-01 -1.09763168e-01 -2.36959502e-01
1.26105711e-01 -2.51540631e-01 -1.44201684e+00 4.61676538e-01
8.40473354e-01 9.53138232e-01 5.39647222e-01 7.53251791e-01
-2.55264968e-01 5.94095826e-01 -5.36079705e-01 -6.94467306e-01
1.24378547e-01 -1.03398287e+00 -7.02274919e-01 1.93173409e-01
2.58012116e-01 -3.91151249e-01 -5.86986840e-01 1.24046850e+00
5.84989965e-01 -1.59632340e-01 9.90711391e-01 -1.76558602e+00
-1.27504516e+00 6.21550262e-01 -4.35171187e-01 -4.37985331e-01
3.11988175e-01 2.58740276e-01 1.07074952e+00 -1.05957282e+00
5.69612145e-01 1.43706322e+00 1.93863064e-01 7.44767487e-01
-1.31980538e+00 -8.34453285e-01 -2.91709155e-01 4.66832854e-02
-1.62969637e+00 -3.37134272e-01 7.72190452e-01 -4.73783851e-01
4.75927323e-01 2.49326050e-01 3.48223597e-01 1.30075479e+00
3.71990710e-01 9.57441866e-01 8.88150573e-01 -2.76908398e-01
1.02705613e-01 1.96408197e-01 -6.76768005e-01 7.52164125e-01
-1.93478197e-01 -7.58120343e-02 -6.37968957e-01 -2.99484516e-03
9.88508999e-01 1.12376906e-01 -4.88227278e-01 -2.76509702e-01
-1.47138441e+00 1.11129737e+00 5.29998660e-01 3.07641089e-01
-1.12124540e-01 3.67476821e-01 3.89536113e-01 2.99991995e-01
1.91644743e-01 3.33789526e-03 1.98318928e-01 -8.87944922e-02
-7.89446354e-01 3.41466129e-01 2.72364885e-01 1.00161052e+00
7.92449415e-01 9.13479831e-03 -4.99825418e-01 9.29850817e-01
5.41453898e-01 8.58893752e-01 8.33134890e-01 -7.96116590e-01
7.18282521e-01 3.79178405e-01 -6.37114868e-02 -9.89913881e-01
4.17890996e-01 8.29034206e-03 -1.16415560e+00 5.42625040e-02
-2.27794528e-01 -1.28526643e-01 -8.29282939e-01 1.86943173e+00
-9.00677033e-03 1.26358783e-02 4.49597418e-01 8.92186642e-01
9.43758368e-01 1.11072826e+00 -9.41977873e-02 -3.61172631e-02
1.29749346e+00 -8.38557959e-01 -9.77858424e-01 5.18044084e-03
3.63004893e-01 -1.00095379e+00 1.20804155e+00 3.99584696e-02
-1.08865905e+00 -7.16663301e-01 -9.57710505e-01 -2.09705755e-01
-2.84701765e-01 4.26791340e-01 1.90306991e-01 4.52668279e-01
-1.01397562e+00 1.66612029e-01 -5.84616244e-01 -1.71020508e-01
3.43521535e-01 7.50923753e-02 -4.18443620e-01 -1.25025392e-01
-1.35013223e+00 3.50324720e-01 5.93330085e-01 -1.57962620e-01
-1.04858613e+00 -1.43819556e-01 -1.17475367e+00 2.15777587e-02
-5.49112214e-04 -9.39954519e-01 1.31252182e+00 -1.06857610e+00
-1.49346817e+00 7.22238243e-01 -4.17412549e-01 -1.91330388e-01
4.75155950e-01 1.95974231e-01 -4.13209140e-01 2.22288668e-01
4.40228373e-01 1.18289304e+00 1.07894635e+00 -1.55674791e+00
-4.07766730e-01 -5.89458980e-02 -4.12328959e-01 2.81768024e-01
-3.95832032e-01 -2.58296043e-01 -5.14458835e-01 -9.84989941e-01
3.59855257e-02 -7.89124906e-01 -1.56000732e-02 -1.13908388e-01
-6.20613635e-01 -2.57412404e-01 8.31822515e-01 -6.34854436e-01
1.28553152e+00 -2.28844547e+00 3.61134797e-01 -3.36484425e-02
1.01723209e-01 -1.27113974e-02 -3.13010663e-01 8.33338499e-01
-3.92549038e-02 3.00305456e-01 -6.01564229e-01 -7.68443525e-01
2.46841282e-01 2.61867344e-01 -7.85053134e-01 1.06156506e-01
2.97652185e-01 1.29787254e+00 -7.78165340e-01 -6.89465761e-01
2.01594487e-01 5.77347040e-01 -4.01139110e-01 5.47798514e-01
-3.19616526e-01 2.70704120e-01 -3.61717850e-01 5.78476012e-01
7.34901786e-01 -3.83900195e-01 2.70399339e-02 -1.76329747e-01
2.20836461e-01 -1.41748771e-01 -8.54532599e-01 1.93484092e+00
-3.47925574e-01 7.58621693e-01 -3.05791229e-01 -7.05047548e-01
1.04607975e+00 4.48670357e-01 3.00490428e-02 -7.14828968e-01
2.92421877e-01 3.10241878e-01 -5.43836117e-01 -3.99692267e-01
7.99435675e-01 -4.67288733e-01 -2.25299165e-01 3.60279381e-01
2.20833570e-01 -4.17241067e-01 3.16610187e-02 6.82871044e-01
4.66606647e-01 9.62112471e-02 7.20838904e-02 1.65062666e-01
4.78036284e-01 -3.63645136e-01 1.48044974e-01 4.65529382e-01
7.07789436e-02 1.05906093e+00 5.23146033e-01 1.78055167e-01
-1.19001150e+00 -1.47210526e+00 -2.61376332e-02 7.76279688e-01
4.23813403e-01 -6.68824613e-01 -8.20934117e-01 -5.55097640e-01
-2.97298521e-01 7.10362375e-01 -6.15026891e-01 -3.76503050e-01
-9.55377296e-02 -4.63075250e-01 5.07827163e-01 4.37236726e-01
7.68245995e-01 -9.49121714e-01 -1.19882263e-01 1.84537154e-02
-7.29053676e-01 -9.72117841e-01 -8.60532641e-01 -3.30648273e-01
-5.73039591e-01 -5.84010005e-01 -1.03197527e+00 -1.12351739e+00
5.75579405e-01 2.92350888e-01 9.79699850e-01 -1.26380742e-01
-2.88834244e-01 3.15926909e-01 -5.69450736e-01 -1.03925839e-02
-7.40173101e-01 -2.27109909e-01 -2.71147877e-01 3.03140163e-01
6.97548166e-02 -4.36593235e-01 -6.99894249e-01 1.38040543e-01
-1.36558282e+00 5.29142559e-01 6.74007237e-01 1.00953746e+00
5.96312940e-01 -1.90374982e-02 5.37023485e-01 -3.06487530e-01
8.68424833e-01 -5.82132697e-01 -3.47600818e-01 2.07540482e-01
-2.67369002e-01 2.70939738e-01 5.28142810e-01 -3.01036388e-01
-9.91891623e-01 3.34858783e-02 -1.40085474e-01 -7.41262257e-01
2.01993361e-02 4.74625140e-01 -4.56769317e-01 3.86925757e-01
3.94230276e-01 8.53716969e-01 1.32055908e-01 -1.80471674e-01
7.87628412e-01 1.08722198e+00 8.11693132e-01 -4.85524029e-01
8.23080003e-01 3.82095158e-01 -2.71861583e-01 -5.07342696e-01
-2.57056355e-01 -1.95814632e-02 -2.36912325e-01 -1.19668106e-02
1.34264743e+00 -1.08784902e+00 -4.69857126e-01 5.78432679e-01
-1.24943197e+00 -2.45198041e-01 1.68735683e-02 3.63896877e-01
-7.42476285e-01 5.86960375e-01 -5.77818453e-01 -7.33007133e-01
-3.51874381e-01 -1.41544485e+00 1.56881392e+00 1.94378242e-01
9.50352103e-02 -9.27372038e-01 1.47884011e-01 3.32211941e-01
7.05975518e-02 2.29984030e-01 7.38582253e-01 -8.69967788e-02
-7.75485933e-01 3.47696170e-02 -4.44682062e-01 4.95687068e-01
2.54048407e-01 -6.53318986e-02 -7.91836023e-01 -3.04162890e-01
-2.23948970e-01 -6.18243098e-01 9.89447176e-01 2.04363093e-01
1.15492296e+00 -3.75100046e-01 -1.68785870e-01 6.37211800e-01
1.54125452e+00 1.55192375e-01 9.04861093e-01 -4.73429337e-02
7.08079159e-01 2.73786068e-01 4.81919527e-01 5.23362398e-01
6.60703540e-01 6.49653792e-01 5.14506876e-01 -1.50373459e-01
-1.15808532e-01 -8.15330625e-01 5.20328820e-01 9.67306077e-01
3.68457496e-01 -5.81099451e-01 -6.46562159e-01 6.46273851e-01
-1.97529948e+00 -9.89830017e-01 1.74589574e-01 1.89514196e+00
1.03318036e+00 -2.91802883e-01 -5.72999492e-02 -5.18909916e-02
9.25164878e-01 1.58427656e-01 -2.92261183e-01 -1.27216756e-01
-2.77519137e-01 2.34504901e-02 1.66336522e-01 4.46678668e-01
-1.02560997e+00 9.52961981e-01 5.60142803e+00 1.19432855e+00
-1.20474017e+00 1.10719837e-01 7.07412481e-01 9.21443999e-02
-6.86622620e-01 -1.40846208e-01 -4.63614076e-01 7.32824087e-01
6.33049726e-01 -2.73274988e-01 3.37697834e-01 6.01276815e-01
-1.31259680e-01 1.65651947e-01 -9.33590055e-01 1.60732579e+00
2.75142431e-01 -1.42927992e+00 6.07804596e-01 8.68298262e-02
9.22783017e-01 -3.40991408e-01 5.14121294e-01 2.33939752e-01
2.82639384e-01 -1.25358880e+00 8.47823739e-01 4.50611860e-01
1.21027792e+00 -6.84567392e-01 6.58990562e-01 1.59740791e-01
-1.47020602e+00 1.17169365e-01 -3.63034040e-01 3.42494875e-01
2.87904620e-01 3.18062752e-01 -6.89838946e-01 8.81971538e-01
4.38352674e-01 7.37538576e-01 -5.99524677e-01 5.35657763e-01
-2.32434288e-01 4.05676156e-01 2.02047497e-01 -4.97914478e-02
3.17778260e-01 -3.23773980e-01 2.40232244e-01 1.15300560e+00
7.33478963e-01 -6.18658103e-02 1.02770992e-01 1.24715734e+00
-2.35326216e-01 2.16245502e-01 -7.36281335e-01 -3.84007275e-01
4.70339984e-01 1.07994652e+00 -4.93395448e-01 -6.14864886e-01
-2.58212924e-01 1.61831689e+00 -7.72968121e-03 5.17270803e-01
-1.02859676e+00 -6.97527170e-01 5.65088630e-01 -3.73699576e-01
4.45728689e-01 -1.46100849e-01 -9.87716764e-02 -1.45067441e+00
7.46801496e-02 -8.33755672e-01 1.06660083e-01 -1.23459637e+00
-1.24604952e+00 6.64462507e-01 1.05044888e-02 -1.42709875e+00
-4.43437308e-01 -3.40426564e-01 -6.10650837e-01 1.06128168e+00
-1.16238248e+00 -1.38133740e+00 -3.97149742e-01 8.96015584e-01
4.99789625e-01 -4.23185408e-01 7.61743069e-01 1.44667357e-01
-2.74225354e-01 7.51129270e-01 3.00605953e-01 4.21419263e-01
6.52615905e-01 -1.23836541e+00 4.72398192e-01 7.60337949e-01
2.80100346e-01 5.58227479e-01 5.07445812e-01 -6.12756193e-01
-1.28436327e+00 -1.29083526e+00 6.78114295e-01 -4.05386478e-01
5.06259263e-01 -6.21379137e-01 -7.49651194e-01 7.35047638e-01
7.93163121e-01 -3.13044995e-01 6.82247400e-01 -7.66469598e-01
-4.78149921e-01 3.23641971e-02 -8.86021078e-01 8.52333605e-01
7.43724108e-01 -9.41761315e-01 -5.60307026e-01 2.09482566e-01
1.18514490e+00 -2.40350321e-01 -8.05352867e-01 1.74260974e-01
3.43926102e-01 -1.03177190e+00 7.63314188e-01 -6.76193237e-02
1.00156713e+00 -5.16317785e-01 -5.67619085e-01 -1.30372989e+00
-9.02702585e-02 -5.33333659e-01 3.15526128e-01 1.42156005e+00
1.56484261e-01 -5.80637753e-01 4.15598989e-01 -2.90991049e-02
-2.03297883e-02 -7.43511975e-01 -8.35572243e-01 -6.48622632e-01
2.68101186e-01 -2.35040456e-01 7.94903517e-01 7.57713497e-01
-9.04863849e-02 5.08544624e-01 -7.28969634e-01 2.01156065e-02
5.84772944e-01 3.37797672e-01 8.87622118e-01 -4.40889806e-01
-3.82627100e-01 -5.39723992e-01 -4.96394128e-01 -1.31424129e+00
1.58979997e-01 -1.11737561e+00 1.94096267e-01 -1.65599716e+00
4.33757335e-01 1.14921138e-01 1.63173191e-02 3.55088949e-01
-3.44909489e-01 5.28826594e-01 2.99350828e-01 3.06169093e-01
-4.86154199e-01 1.21177340e+00 1.35121059e+00 -2.12056205e-01
1.04743466e-01 -4.71636832e-01 -6.69148564e-01 5.37489280e-02
8.49810839e-01 -1.61523297e-01 -6.75357997e-01 -3.76042128e-01
3.97015475e-02 3.00123066e-01 6.75989687e-01 -8.45674217e-01
-1.40438937e-02 -1.41442284e-01 3.83446872e-01 -5.81268013e-01
3.92626673e-01 -4.62396711e-01 2.74169505e-01 4.93061632e-01
-4.19709653e-01 -2.68256124e-02 -1.52146975e-02 5.45545936e-01
-6.84657454e-01 -1.35502711e-01 6.94099665e-01 -2.79295947e-02
-5.34045875e-01 4.11139190e-01 -1.57934323e-01 -4.40455191e-02
8.71347487e-01 -8.90026838e-02 -3.19305688e-01 -6.48288250e-01
-4.34425354e-01 1.73558921e-01 5.63205540e-01 6.31237328e-01
1.04491663e+00 -1.97264791e+00 -8.23374689e-01 3.56129885e-01
3.41616869e-01 -1.22817382e-01 3.68951082e-01 4.14745092e-01
-3.92616630e-01 3.61551195e-01 -1.86724722e-01 -7.22451091e-01
-1.26293075e+00 5.09524941e-01 2.34467432e-01 8.93759429e-02
-2.87764013e-01 8.34499300e-01 6.89166665e-01 -2.70135462e-01
-1.11292871e-02 -1.82231084e-01 1.37511328e-01 -2.06072599e-01
5.69384754e-01 -1.81071669e-01 -4.45536673e-01 -8.62383723e-01
1.51743740e-02 5.98240614e-01 1.69318646e-01 -6.16400361e-01
7.87723541e-01 -3.42731744e-01 -3.04952651e-01 4.23133850e-01
1.68322802e+00 -1.94529608e-01 -1.07941651e+00 -1.97943285e-01
-6.10722303e-01 -6.28241003e-01 -1.12238988e-01 -3.96287113e-01
-1.03503752e+00 1.08849275e+00 7.42861152e-01 1.25347570e-01
1.24776649e+00 3.20537090e-01 1.07200849e+00 1.03162646e-01
7.34995157e-02 -6.31798863e-01 6.62311018e-01 2.25671321e-01
1.14390838e+00 -1.30532598e+00 -4.08878207e-01 -1.77542567e-01
-1.08542275e+00 8.82683814e-01 5.36198974e-01 1.16660625e-01
3.76239419e-01 -1.54114455e-01 1.20018385e-01 -2.69359048e-03
-7.97027946e-01 -2.03096986e-01 2.05793038e-01 6.72752976e-01
4.71821696e-01 2.45810643e-01 -1.19238280e-01 6.45163119e-01
-5.41825354e-01 -2.88367197e-02 3.72846127e-01 6.25525117e-01
-3.14893812e-01 -1.30069327e+00 -4.12585706e-01 7.11476654e-02
-1.17320046e-01 -3.46920043e-01 -3.62062871e-01 3.39640111e-01
1.98186576e-01 8.58115375e-01 1.00822099e-01 -6.64855003e-01
-4.28427570e-02 7.98621252e-02 3.87707502e-01 -4.98553813e-01
-1.35088041e-01 3.74166667e-01 -4.90654796e-01 -3.56491297e-01
-2.89073735e-01 -3.84460837e-01 -1.32538235e+00 -2.07508817e-01
-4.63427231e-02 2.72483826e-01 5.63867688e-01 5.20775139e-01
4.12572950e-01 4.43478882e-01 8.19483280e-01 -7.32058048e-01
-3.96204025e-01 -8.02098274e-01 -6.46538079e-01 7.38419354e-01
3.83341968e-01 -4.33973461e-01 -2.40549117e-01 5.25129497e-01] | [11.237918853759766, 0.47684305906295776] |
51b6feed-1b81-4eb5-bdf6-46487e079d52 | bayesian-models-for-unit-discovery-on-a-very | 1802.06053 | null | http://arxiv.org/abs/1802.06053v2 | http://arxiv.org/pdf/1802.06053v2.pdf | Bayesian Models for Unit Discovery on a Very Low Resource Language | Developing speech technologies for low-resource languages has become a very
active research field over the last decade. Among others, Bayesian models have
shown some promising results on artificial examples but still lack of in situ
experiments. Our work applies state-of-the-art Bayesian models to unsupervised
Acoustic Unit Discovery (AUD) in a real low-resource language scenario. We also
show that Bayesian models can naturally integrate information from other
resourceful languages by means of informative prior leading to more consistent
discovered units. Finally, discovered acoustic units are used, either as the
1-best sequence or as a lattice, to perform word segmentation. Word
segmentation results show that this Bayesian approach clearly outperforms a
Segmental-DTW baseline on the same corpus. | ['François Yvon', 'Emmanuel Dupoux', 'Mark Hasegawa-Johnson', 'Laurent Besacier', 'Sanjeev Khudanpur', 'Odette Scharenborg', 'Lucas Ondel', 'Elin Larsen', 'Pierre Godard', 'Lukas Burget'] | 2018-02-16 | null | null | null | null | ['acoustic-unit-discovery'] | ['speech'] | [ 1.02895260e-01 2.31857479e-01 -1.26185209e-01 -6.24184847e-01
-1.29835975e+00 -3.79103571e-01 6.63209260e-01 -1.68444067e-01
-7.17216671e-01 6.60287261e-01 5.17392576e-01 -3.72395277e-01
1.62448540e-01 -4.60200280e-01 -6.03890240e-01 -5.97093701e-01
1.60812274e-01 1.01559305e+00 8.62262249e-01 -6.68793097e-02
1.74816906e-01 1.03343111e-02 -1.40877712e+00 2.05489799e-01
7.29443669e-01 2.74486572e-01 8.43253911e-01 6.53749466e-01
-5.50498188e-01 2.15665683e-01 -7.97293961e-01 -2.92196069e-02
-9.85851735e-02 -1.55365720e-01 -8.36075544e-01 -1.47000970e-02
2.04389930e-01 -2.11185843e-01 1.61944628e-01 9.42623913e-01
5.77419758e-01 3.14369619e-01 5.11863053e-01 -2.98355788e-01
-4.51720282e-02 1.58973575e+00 -1.95398748e-01 2.65065223e-01
2.37590373e-01 -8.79233927e-02 1.28340912e+00 -1.05896652e+00
6.12499237e-01 1.71972466e+00 3.70221674e-01 4.87130314e-01
-1.27962506e+00 -4.80768025e-01 2.59425312e-01 2.30921730e-01
-1.35634089e+00 -8.07670832e-01 7.40948200e-01 -3.50170910e-01
1.51484358e+00 5.82497306e-02 4.70617771e-01 1.16134727e+00
-4.91127580e-01 1.13414478e+00 1.03730404e+00 -9.43793356e-01
4.17695612e-01 2.47447994e-02 3.55159521e-01 4.67664003e-01
2.77371109e-02 -1.66598782e-01 -9.14819777e-01 4.77972142e-02
5.08734047e-01 -6.29215777e-01 6.87351599e-02 3.24054062e-01
-1.11359763e+00 8.02278399e-01 -3.00766438e-01 8.37898433e-01
-3.91045660e-01 2.42553875e-01 1.71885327e-01 -3.15845221e-01
8.13042283e-01 2.49583676e-01 -5.27589142e-01 -5.73338270e-01
-1.28767312e+00 -5.38282283e-02 8.56706858e-01 8.90669823e-01
8.33130598e-01 4.72193271e-01 5.51076345e-02 1.29876328e+00
9.29056287e-01 6.12841725e-01 5.46095729e-01 -7.94639111e-01
3.21244299e-01 -8.19830820e-02 -1.98379606e-01 -1.95732728e-01
-1.02871051e-02 -1.74926341e-01 -8.63035768e-02 -3.56430471e-01
3.57219189e-01 -1.23504743e-01 -1.18611658e+00 1.62280047e+00
2.65606403e-01 4.48042303e-01 7.30737746e-02 5.58393359e-01
5.89762449e-01 1.12814212e+00 1.24404527e-01 -2.39251599e-01
1.48580050e+00 -7.95313418e-01 -8.35386634e-01 -6.45274818e-01
3.21443290e-01 -1.01336324e+00 9.76841986e-01 7.27162302e-01
-1.09654748e+00 -4.80771929e-01 -8.80980074e-01 1.73890740e-01
-2.98403680e-01 -4.13844511e-02 5.66633463e-01 1.16915584e+00
-9.89584684e-01 2.38900200e-01 -1.23129809e+00 -5.81126511e-01
3.71397510e-02 1.62450626e-01 8.38065743e-02 -8.68752971e-02
-1.23501587e+00 5.84103107e-01 7.70947576e-01 -3.75570841e-02
-1.15251589e+00 -4.63922024e-01 -8.15707684e-01 -1.61598280e-01
6.07411444e-01 -1.00100935e-01 1.51913905e+00 -4.13171589e-01
-1.89684641e+00 7.32041776e-01 -4.51222926e-01 -6.20253146e-01
6.28980696e-02 -4.72152382e-01 -2.60584444e-01 -2.41368879e-02
-1.51589364e-01 7.74886191e-01 6.23052359e-01 -1.38075495e+00
-7.72851229e-01 -2.23888397e-01 -5.22659659e-01 8.92539620e-02
-2.54483819e-01 4.34537441e-01 -8.44513595e-01 -7.33762980e-01
2.26799533e-01 -7.64529824e-01 -2.27887958e-01 -9.03592348e-01
-4.95542318e-01 -6.15240574e-01 4.38645989e-01 -9.55824554e-01
1.36706185e+00 -2.04640913e+00 1.19587600e-01 3.13203961e-01
-4.78810370e-01 2.42225885e-01 1.69886842e-01 2.85423219e-01
5.18318295e-01 2.66274244e-01 -3.35251808e-01 -7.30466306e-01
1.33955866e-01 8.17911685e-01 -2.88953900e-01 1.15283549e-01
1.63445976e-02 5.14192283e-01 -8.29530716e-01 -7.52198637e-01
3.08807701e-01 5.31436741e-01 -3.97180945e-01 1.95572942e-01
-4.87431914e-01 2.94510424e-01 -2.78818965e-01 6.01207316e-01
2.64234722e-01 4.43090916e-01 5.21673083e-01 1.28409252e-01
-4.44141626e-01 8.44268143e-01 -1.26993179e+00 1.94343889e+00
-5.26023924e-01 6.18072152e-01 1.32654205e-01 -9.79908884e-01
1.05320632e+00 4.00050104e-01 1.77158862e-01 -2.56094545e-01
2.93509290e-02 5.29554248e-01 4.30696815e-01 -2.68598855e-01
6.48916721e-01 -1.92750439e-01 -8.77987668e-02 2.61490762e-01
5.02403557e-01 -5.67075908e-01 3.12834799e-01 -1.43269571e-02
8.88752162e-01 1.90478176e-01 1.73423126e-01 -3.78193289e-01
2.59822041e-01 -3.31112929e-02 6.47847414e-01 9.89631593e-01
-5.74475192e-02 7.85504580e-01 -7.52852857e-02 1.32969454e-01
-8.79934371e-01 -1.16047907e+00 -2.97362059e-01 1.42628789e+00
-4.14607227e-01 -4.81984735e-01 -9.15840089e-01 -2.83679873e-01
-4.46179450e-01 1.12014365e+00 -3.05495467e-02 4.56822932e-01
-8.95258188e-01 -7.41859138e-01 8.05061698e-01 3.63745332e-01
2.85516769e-01 -1.38531303e+00 -3.76021117e-01 7.32873023e-01
-3.13514829e-01 -1.25672483e+00 -1.93587273e-01 4.62309539e-01
-7.40561247e-01 -2.43853658e-01 -7.82163203e-01 -8.90207231e-01
1.19136557e-01 -3.10347099e-02 1.21826422e+00 -2.80122161e-01
-8.56727064e-02 2.99104124e-01 -6.06041133e-01 -7.27632701e-01
-6.22923076e-01 2.49559909e-01 2.08628237e-01 -1.01274326e-01
5.59014499e-01 -4.10865366e-01 -5.99634312e-02 3.05053413e-01
-7.52688289e-01 -3.19047838e-01 4.68393564e-01 6.35834217e-01
7.70345867e-01 -1.17412537e-01 5.97846806e-01 -9.01751280e-01
4.16476965e-01 -2.89094537e-01 -7.05507874e-01 1.93291217e-01
-3.63797218e-01 2.16354460e-01 1.72616560e-02 -4.42343652e-01
-1.54529905e+00 1.51513323e-01 -6.36190295e-01 2.49285344e-02
-4.72410202e-01 8.19314003e-01 -1.87706456e-01 6.19796097e-01
5.35549521e-01 2.54997790e-01 -4.54629213e-01 -9.35537279e-01
4.99228746e-01 1.08619070e+00 5.96745670e-01 -1.02354574e+00
3.00939918e-01 2.88504273e-01 -6.57837868e-01 -1.51451921e+00
-6.97174311e-01 -8.67271602e-01 -8.69345009e-01 -1.96355879e-01
9.47998285e-01 -8.30451012e-01 3.28194536e-02 4.42739725e-01
-1.41233456e+00 -5.10505199e-01 -2.56530732e-01 8.82747471e-01
-4.06179756e-01 4.74392265e-01 -5.67897797e-01 -1.34049642e+00
-1.59762681e-01 -1.18310761e+00 1.24400902e+00 1.70091674e-01
-4.64806437e-01 -9.30568278e-01 3.97014767e-01 5.20893395e-01
1.71676427e-01 -6.04124308e-01 8.14855039e-01 -1.03100002e+00
-5.80702126e-01 5.26597686e-02 2.72555590e-01 4.13783997e-01
-9.76373907e-03 2.66915947e-01 -1.26382756e+00 7.98492283e-02
-2.03575924e-01 -8.26735944e-02 9.13262308e-01 6.85995579e-01
5.29728174e-01 7.54519254e-02 -4.02674466e-01 1.87992588e-01
8.86724234e-01 4.17718440e-01 5.55772960e-01 -5.03103137e-02
6.11216843e-01 7.92037725e-01 6.72488153e-01 2.37276301e-01
2.49020681e-01 6.99208379e-01 2.24582795e-02 3.07633877e-01
-3.44402134e-01 -1.43320963e-01 7.08512485e-01 1.79355848e+00
3.66758741e-02 -2.95392543e-01 -1.47589159e+00 1.10066247e+00
-1.80837488e+00 -7.14836121e-01 1.18563492e-02 2.12596226e+00
1.11989319e+00 5.31494737e-01 8.77495408e-02 -3.08368970e-02
7.66463816e-01 2.37315893e-01 -7.73061439e-02 -4.10459518e-01
-2.69886881e-01 4.93473023e-01 2.55579889e-01 7.51746297e-01
-8.73207033e-01 1.60779190e+00 6.95276737e+00 1.26934266e+00
-6.27846301e-01 4.95251566e-01 4.52214301e-01 2.78981924e-01
-4.68572259e-01 3.66638452e-01 -1.25763166e+00 3.40701729e-01
1.31419146e+00 3.19047570e-01 3.42895210e-01 8.54711771e-01
2.34598562e-01 -4.82957512e-01 -8.41513932e-01 7.02089190e-01
9.83803347e-02 -1.19118488e+00 -1.05765060e-01 1.40886262e-01
6.42351747e-01 5.94191611e-01 -3.05064440e-01 2.87069827e-01
8.29776824e-01 -8.80330682e-01 8.52670550e-01 3.54697973e-01
5.83418250e-01 -6.33168161e-01 7.30702102e-01 5.12209952e-01
-1.18612015e+00 4.76759076e-01 -4.50349569e-01 2.65823036e-01
6.26003921e-01 5.93954325e-01 -1.46670008e+00 3.25706303e-01
7.80849099e-01 3.85987848e-01 -2.63664454e-01 1.12207127e+00
-4.83734876e-01 1.68190229e+00 -8.43732595e-01 -3.13998461e-01
4.73921210e-01 -2.52560735e-01 8.79534543e-01 1.78560460e+00
2.71309674e-01 -6.90382347e-02 3.66464406e-01 7.34726131e-01
2.73174196e-01 3.29015225e-01 -3.75205308e-01 -1.21754222e-01
6.81622982e-01 8.89759362e-01 -1.10363913e+00 -3.08863997e-01
-3.33132237e-01 5.15054762e-01 8.31946805e-02 1.70276776e-01
-3.62132460e-01 -1.37886107e-01 4.19489801e-01 -2.26200610e-01
5.60770810e-01 -6.41760230e-01 -1.94810227e-01 -7.26611137e-01
-3.02172750e-01 -6.89270854e-01 2.37150148e-01 -5.79875946e-01
-1.17344189e+00 6.29598856e-01 3.63277197e-01 -5.20133615e-01
-7.63531387e-01 -5.39860487e-01 -4.44065809e-01 8.92276943e-01
-1.26034522e+00 -9.99833524e-01 2.74659425e-01 1.22941010e-01
1.24165297e+00 -4.15414959e-01 9.21454728e-01 1.22765400e-01
-5.94151616e-01 2.48500556e-01 1.43615037e-01 2.41804078e-01
6.20274603e-01 -1.44332969e+00 6.62850201e-01 1.16306186e+00
9.03592646e-01 6.19027495e-01 8.23463380e-01 -9.15470719e-01
-1.01378536e+00 -5.07001698e-01 8.00438166e-01 -4.78233546e-01
8.79867554e-01 -6.89678729e-01 -1.05395198e+00 5.38914323e-01
5.42729557e-01 -4.98372346e-01 6.05999112e-01 4.65053231e-01
-1.71415821e-01 9.61455926e-02 -6.52529597e-01 5.83550274e-01
8.74697089e-01 -5.74682534e-01 -1.02532387e+00 1.06190681e-01
1.02679932e+00 -2.76848208e-02 -6.67566001e-01 1.48128226e-01
4.33790445e-01 -6.87141657e-01 8.17005038e-01 -2.58514076e-01
-3.01649034e-01 -2.40929917e-01 -5.57965815e-01 -1.41218197e+00
9.96936783e-02 -1.00738108e+00 1.13481507e-01 1.75586104e+00
8.19095135e-01 -2.94814378e-01 6.25138819e-01 1.43531859e-01
-6.31223381e-01 -2.00642183e-01 -1.32060623e+00 -9.05210078e-01
7.04532266e-02 -1.12648237e+00 2.70896256e-01 5.35467029e-01
-5.50043285e-02 4.36082780e-01 -2.24543527e-01 9.75293741e-02
5.39204299e-01 -2.87105054e-01 5.83164215e-01 -1.21038651e+00
-6.91679120e-01 -4.61981416e-01 -1.47989288e-01 -1.30383861e+00
4.82341588e-01 -6.59265339e-01 1.01543427e+00 -1.51400805e+00
-1.59517780e-01 -5.39088249e-01 -2.06822261e-01 4.63569909e-01
-1.40593827e-01 1.21685686e-02 -5.50950132e-02 3.98355238e-02
-3.18424791e-01 3.31604183e-01 5.44855058e-01 -1.20066546e-01
-3.46364617e-01 -4.00740537e-04 -9.17094648e-02 8.89257789e-01
6.89017653e-01 -7.28301346e-01 -2.00670630e-01 -5.15455246e-01
8.01745355e-02 -2.20118180e-01 -4.91989404e-01 -8.79565299e-01
2.13672906e-01 -9.79573578e-02 -3.83148104e-01 -8.81929696e-01
6.10580981e-01 -4.95053142e-01 -1.42207295e-01 1.19255669e-01
3.75192352e-02 -3.43229264e-01 2.11052179e-01 4.60676849e-01
-3.55896324e-01 -8.69521379e-01 5.24939239e-01 -2.63993084e-01
-9.89412725e-01 -2.12753221e-01 -9.11061585e-01 1.20171465e-01
4.35148746e-01 -2.27377132e-01 1.48458526e-01 -1.58390462e-01
-6.67818129e-01 -2.25629643e-01 -1.16715133e-01 3.31251502e-01
3.61344695e-01 -7.07282126e-01 -9.71697986e-01 -8.69187787e-02
3.87324169e-02 2.26264372e-01 -1.83860987e-01 4.15919691e-01
-4.28177893e-01 4.97767597e-01 3.42537701e-01 -1.05024850e+00
-1.23288202e+00 -3.55484411e-02 2.89425775e-02 -2.92433858e-01
-4.38099355e-01 1.10377634e+00 -1.04149021e-01 -5.39725900e-01
4.93963838e-01 -3.80245417e-01 -2.18551666e-01 3.05401295e-01
2.69242078e-01 2.52536118e-01 1.08368352e-01 -8.12076569e-01
-4.88021433e-01 3.64961088e-01 -6.91161528e-02 -1.11891997e+00
1.36504471e+00 -2.69079894e-01 6.87183216e-02 1.01327205e+00
5.24473608e-01 3.71505529e-01 -1.09839094e+00 -4.98003513e-01
5.92627466e-01 -1.53316170e-01 2.32863590e-01 -6.74034655e-01
-4.84554738e-01 1.06330645e+00 4.19259369e-01 2.09525406e-01
4.73571241e-01 3.65985423e-01 6.02884233e-01 5.57227552e-01
5.13862848e-01 -1.55262268e+00 -3.95545773e-02 9.77499604e-01
4.77969140e-01 -1.20547462e+00 -2.34478444e-01 -5.71930349e-01
-5.72904587e-01 9.54353571e-01 2.90539563e-01 1.30629674e-01
7.82984018e-01 5.71544170e-01 1.45467162e-01 8.59926715e-02
-6.31898880e-01 -7.69257128e-01 3.49705547e-01 7.80398846e-01
6.61183596e-01 3.12223911e-01 -1.92896962e-01 5.50431848e-01
-3.86540860e-01 -6.44527733e-01 3.75972688e-01 7.73821890e-01
-8.78431678e-01 -1.41366696e+00 -5.85509598e-01 1.34834349e-01
-5.69931507e-01 -5.76240957e-01 -2.77305543e-01 6.21602476e-01
-5.33058718e-02 1.22401023e+00 1.38347059e-01 -8.25980082e-02
8.77367258e-02 6.32904887e-01 3.45905334e-01 -1.24857914e+00
-4.57279235e-01 1.07402861e+00 5.82153380e-01 -1.41847789e-01
-4.95134652e-01 -1.15045857e+00 -1.35731459e+00 4.19080228e-01
-5.76749027e-01 2.16183245e-01 1.06329179e+00 1.22691214e+00
-1.22628972e-01 6.03412628e-01 9.23328027e-02 -9.79808748e-01
-2.63325721e-01 -1.50225759e+00 -5.16958535e-01 -1.54976711e-01
-1.74721107e-01 -4.33178812e-01 -2.51237601e-01 2.41521135e-01] | [14.468788146972656, 6.720249176025391] |
c6e5124b-a05d-44fb-82ac-cf5ee6deea17 | multi-head-attention-neural-network-for | 2205.08069 | null | https://arxiv.org/abs/2205.08069v1 | https://arxiv.org/pdf/2205.08069v1.pdf | Multi-Head Attention Neural Network for Smartphone Invariant Indoor Localization | Smartphones together with RSSI fingerprinting serve as an efficient approach for delivering a low-cost and high-accuracy indoor localization solution. However, a few critical challenges have prevented the wide-spread proliferation of this technology in the public domain. One such critical challenge is device heterogeneity, i.e., the variation in the RSSI signal characteristics captured across different smartphone devices. In the real-world, the smartphones or IoT devices used to capture RSSI fingerprints typically vary across users of an indoor localization service. Conventional indoor localization solutions may not be able to cope with device-induced variations which can degrade their localization accuracy. We propose a multi-head attention neural network-based indoor localization framework that is resilient to device heterogeneity. An in-depth analysis of our proposed framework across a variety of indoor environments demonstrates up to 35% accuracy improvement compared to state-of-the-art indoor localization techniques. | ['Sudeep Pasricha', 'Danish Gufran', 'Saideep Tiku'] | 2022-05-17 | null | null | null | null | ['indoor-localization'] | ['computer-vision'] | [-1.78555325e-02 -7.84502506e-01 -4.37944084e-02 -5.82472265e-01
-1.14041460e+00 -5.88618338e-01 7.25774318e-02 1.20534688e-01
-1.54274851e-01 7.23309577e-01 2.02764601e-01 -4.05860275e-01
-3.53678674e-01 -7.17866659e-01 -9.22418594e-01 -5.67923844e-01
9.99719836e-03 -5.38003072e-02 -5.57784401e-02 2.14889899e-01
1.37663200e-01 4.36166435e-01 -1.41706157e+00 -1.64826944e-01
8.36691141e-01 1.42360342e+00 2.23769516e-01 5.55111527e-01
1.19465783e-01 3.42712522e-01 -1.06650698e+00 1.14871882e-01
-2.28751808e-01 8.78290907e-02 -1.56911954e-01 -7.44287550e-01
4.61659312e-01 -3.19474310e-01 -2.65258491e-01 6.56693280e-01
9.50922430e-01 -4.10380587e-02 2.00894356e-01 -1.27669132e+00
-5.32728970e-01 3.02558988e-01 -3.13233554e-01 2.10111290e-01
8.33718419e-01 -6.55334592e-01 3.36053282e-01 -4.93330181e-01
-7.83323422e-02 5.83185136e-01 1.34441423e+00 -1.02230879e-02
-8.94451499e-01 -8.23988438e-01 -4.78494428e-02 -3.17831971e-02
-1.76103985e+00 -7.18828619e-01 7.96935856e-01 -3.67805474e-02
6.74030364e-01 4.60459799e-01 3.57464999e-01 1.47044659e+00
6.91523910e-01 4.15622473e-01 7.23251164e-01 -1.94072053e-01
4.18372601e-01 9.21725780e-02 -8.50202814e-02 1.03910439e-01
5.54938197e-01 -3.15345377e-01 -6.63809001e-01 -3.50010544e-01
3.46282989e-01 6.81138575e-01 -1.88647375e-01 -2.50716716e-01
-1.26090288e+00 1.15343384e-01 7.87319362e-01 8.24256420e-01
-3.76736015e-01 5.69895446e-01 4.85844463e-02 -2.01390758e-01
3.08770120e-01 4.98251677e-01 -5.86614370e-01 -4.54847962e-01
-1.18056583e+00 -5.68154231e-02 6.30646169e-01 1.20292485e+00
5.27239144e-01 -1.69782370e-01 9.52400640e-02 7.99569190e-01
4.58026379e-01 1.00799477e+00 5.81620991e-01 -8.53705525e-01
5.35478234e-01 9.28279236e-02 5.23128211e-01 -1.38894987e+00
-6.63343668e-01 -8.08418751e-01 -7.76790380e-01 -6.64754868e-01
4.09501255e-01 -3.83105397e-01 -2.52097309e-01 1.85166848e+00
9.72371548e-02 6.20485902e-01 -3.97779644e-01 1.89194798e-01
5.15100539e-01 3.01588297e-01 -1.75944373e-01 1.82431310e-01
8.58899653e-01 -7.47891128e-01 -6.30519092e-01 -2.75154293e-01
1.45809650e-01 -6.64914489e-01 9.37708318e-01 1.46870278e-02
-6.45603955e-01 -6.68703198e-01 -1.38477230e+00 3.07064801e-01
-6.49320662e-01 -2.25189514e-02 6.61414027e-01 1.34259701e+00
-9.97455418e-01 1.07055560e-01 -8.62568259e-01 -4.72122282e-01
1.89372048e-01 7.79542983e-01 -1.61343440e-01 -2.46838614e-01
-8.88400793e-01 2.36020938e-01 -4.20809090e-01 2.12905213e-01
-2.68366694e-01 -7.93577433e-01 -6.24114692e-01 1.50001884e-01
9.46499556e-02 -4.96588498e-01 1.33556616e+00 -4.47175652e-01
-1.62484491e+00 -2.18529236e-02 -7.49486029e-01 -1.53904051e-01
2.37659454e-01 -3.05720896e-01 -8.61758947e-01 -6.95814371e-01
5.53288221e-01 -1.76913753e-01 3.51925939e-01 -1.22367191e+00
-7.00511694e-01 -3.93163770e-01 -8.90247077e-02 -1.78294554e-01
-2.37791955e-01 -2.72483259e-01 -4.44526166e-01 -2.11030543e-01
3.29876930e-01 -9.79838967e-01 -2.07668431e-02 -6.62218690e-01
-5.69176733e-01 3.47378969e-01 7.54795909e-01 -2.78547317e-01
1.32723439e+00 -2.05163026e+00 -8.55075657e-01 3.54045123e-01
-1.60509229e-01 -1.98385805e-01 2.82466561e-01 3.76816034e-01
4.02871311e-01 8.10151175e-02 3.92368317e-01 -8.42949390e-01
1.67360529e-01 -7.82924145e-02 -2.46668339e-01 6.22634828e-01
-7.51590312e-01 7.63650894e-01 -1.03815639e+00 -3.97797041e-02
5.12018800e-01 8.77526641e-01 -2.87125915e-01 1.94102407e-01
5.86825907e-01 1.00809515e+00 -5.91709852e-01 9.97410655e-01
7.52487540e-01 -4.26700562e-01 3.28523889e-02 5.34442142e-02
-2.30068892e-01 4.49992299e-01 -1.19380236e+00 1.85500503e+00
-1.20862198e+00 6.07357681e-01 -1.54949605e-01 -5.26605964e-01
5.71470678e-01 1.81423441e-01 6.38862431e-01 -9.27812517e-01
1.70768470e-01 5.59169292e-01 -3.47377330e-01 -1.38686016e-01
6.28146231e-01 5.77552617e-01 -5.83045781e-01 2.94709623e-01
-3.27384502e-01 5.67348480e-01 -5.07407904e-01 -3.21506619e-01
1.38165081e+00 -1.47684768e-01 2.67067909e-01 -7.24783987e-02
4.36909437e-01 -7.31719136e-01 3.84841353e-01 1.40199780e+00
-3.26406509e-01 4.74983424e-01 -4.76134092e-01 -1.56796679e-01
-3.76486361e-01 -1.10416734e+00 -4.72372979e-01 1.24124396e+00
4.41627413e-01 -1.99349627e-01 -7.50656545e-01 -3.24436247e-01
1.04372256e-01 4.71263856e-01 -4.90634680e-01 1.73774570e-01
-3.24637443e-01 -7.40461409e-01 8.63796175e-01 5.04020751e-01
9.69285905e-01 -3.85870546e-01 -5.29747188e-01 4.68637288e-01
-4.42449182e-01 -1.35399783e+00 -4.18069392e-01 1.83004588e-01
-2.92876571e-01 -5.73018014e-01 -6.68689489e-01 -5.18302739e-01
5.55974901e-01 8.35588872e-01 9.81950700e-01 2.26403009e-02
1.87310681e-01 6.81610346e-01 2.98319827e-03 -3.96409690e-01
3.07628691e-01 6.16609216e-01 3.90123546e-01 1.87787309e-01
2.40740404e-01 -8.17679107e-01 -8.97031188e-01 7.11449802e-01
-1.48200482e-01 -7.02032745e-01 2.53584772e-01 4.03738022e-01
6.01152241e-01 3.07500660e-01 7.89062858e-01 -5.63694775e-01
5.59105098e-01 -1.00753689e+00 -4.31488961e-01 2.65079468e-01
-5.66505253e-01 -6.01312697e-01 8.56855869e-01 -2.68500060e-01
-7.76092172e-01 -6.79824874e-02 -4.66325194e-01 5.08496225e-01
-4.80162412e-01 2.57295072e-01 -7.48989522e-01 -5.44049382e-01
3.52327585e-01 1.80156976e-01 -9.84732807e-01 -4.98361468e-01
7.64219649e-03 1.23283100e+00 6.29098296e-01 -3.81334126e-01
4.85707402e-01 4.77955520e-01 -7.39713013e-02 -8.78108621e-01
-6.19248927e-01 -5.95665991e-01 -3.28787148e-01 2.75767464e-02
3.13700348e-01 -9.72015738e-01 -1.10919476e+00 5.96836090e-01
-9.98231649e-01 -3.29670668e-01 4.39008057e-01 2.16681749e-01
-2.84563303e-01 8.60074069e-03 -2.25798041e-02 -9.78169799e-01
-1.32821739e-01 -1.41799498e+00 1.32758844e+00 6.92182481e-01
-2.28256553e-01 -9.47751403e-01 1.10775478e-01 3.49956125e-01
1.24367821e+00 1.95285216e-01 3.46059829e-01 -4.41520154e-01
-7.37888515e-01 -8.20912480e-01 -9.79674608e-02 -3.92868876e-01
5.85447371e-01 -4.90341395e-01 -1.25756574e+00 -2.90090352e-01
-4.25140224e-02 4.09366608e-01 1.17589673e-02 8.56637597e-01
1.46907377e+00 -1.34356886e-01 -8.32531452e-01 1.15154552e+00
1.45394087e+00 5.51324248e-01 4.99814421e-01 5.33369243e-01
6.62676275e-01 -2.00976476e-01 2.23430172e-01 2.15590209e-01
7.37752140e-01 1.08765090e+00 2.57860571e-01 -7.91211054e-02
1.29799962e-01 -5.59848368e-01 -1.47133246e-01 4.49993670e-01
2.58544892e-01 -7.57181764e-01 -8.83756042e-01 4.82342184e-01
-1.59010911e+00 -7.89377689e-01 8.89350548e-02 2.58112478e+00
1.49783030e-01 -1.59452483e-02 -5.74633442e-02 2.53930867e-01
5.70062697e-01 1.45181939e-01 -5.38087487e-01 -6.90500215e-02
2.04368740e-01 -9.15230736e-02 9.58115101e-01 5.07367253e-01
-1.10441196e+00 4.41651642e-01 6.52117014e+00 3.71690661e-01
-1.53872287e+00 4.38123912e-01 5.33534825e-01 4.98344228e-02
-3.41206044e-02 -6.21077478e-01 -7.92478800e-01 1.23518932e+00
1.48711526e+00 3.25297415e-01 5.55653930e-01 1.19095540e+00
2.49954179e-01 -2.79396445e-01 -9.60814536e-01 1.29966986e+00
1.32677034e-02 -1.43401337e+00 -9.35451746e-01 2.07063034e-01
9.27015543e-01 3.58941495e-01 6.00412071e-01 2.32096046e-01
-2.66023934e-01 -8.86663496e-01 5.84116042e-01 5.35525978e-01
9.14250076e-01 -7.90839732e-01 1.03519213e+00 1.01894438e-01
-1.49815786e+00 -2.01224074e-01 6.78983692e-04 -1.08731993e-01
1.88341498e-01 6.62980437e-01 -6.66963339e-01 3.66583854e-01
1.14758551e+00 3.67355108e-01 -7.39579737e-01 1.30732226e+00
-2.75979973e-02 4.80253398e-01 -6.60767198e-01 -3.03322554e-01
-1.13039970e-01 2.23053515e-01 -2.96517294e-02 1.09301543e+00
8.51450205e-01 -6.25924885e-01 -1.43025056e-01 4.06486928e-01
-2.93759912e-01 -1.53144479e-01 -8.21326554e-01 5.49856663e-01
1.23672187e+00 8.57766867e-01 -5.48046947e-01 7.16534108e-02
-4.79397386e-01 1.10288358e+00 -2.27614239e-01 7.04576671e-01
-1.06032193e+00 -6.32317722e-01 7.95636773e-01 1.48723722e-01
1.87918574e-01 -4.75412458e-01 -4.70575988e-01 -8.22937012e-01
2.88605671e-02 -3.87772024e-01 -3.80269170e-01 -6.88533366e-01
-1.04666686e+00 4.25887495e-01 -4.10374850e-01 -9.83655751e-01
-2.84036666e-01 -2.01349467e-01 -6.86270237e-01 7.75594711e-01
-1.35597646e+00 -1.35614932e+00 -7.57161677e-01 6.28055394e-01
1.95353523e-01 2.41044443e-02 1.08024693e+00 1.01335037e+00
-6.42828405e-01 1.16823697e+00 1.06813693e+00 -6.07174858e-02
7.53583968e-01 -1.13073564e+00 6.61460936e-01 8.74110043e-01
2.86437422e-01 1.34982622e+00 6.39019191e-01 -2.13948712e-01
-1.68537271e+00 -1.18372941e+00 1.04588962e+00 -6.77778900e-01
3.01075369e-01 -5.87969184e-01 -3.12981427e-01 7.06242859e-01
-2.76073456e-01 2.38269061e-01 1.11634219e+00 4.67789739e-01
-1.29142910e-01 -7.05499589e-01 -1.33744442e+00 6.11223698e-01
9.54380691e-01 -7.69157827e-01 1.15068838e-01 4.20456827e-02
4.20614749e-01 -4.88607943e-01 -7.86952138e-01 1.55694947e-01
1.16254985e+00 -1.11682558e+00 1.25807190e+00 4.10623789e-01
-4.41526860e-01 -4.41149592e-01 -7.23535955e-01 -1.02806866e+00
-2.13234156e-01 -7.88838267e-01 -2.88349926e-01 1.40273213e+00
2.65759081e-01 -1.08803546e+00 6.61782920e-01 7.03785419e-01
1.63381219e-01 -6.51152432e-01 -1.25911784e+00 -7.58679211e-01
-4.57304090e-01 -9.14479077e-01 1.43331909e+00 3.70722562e-01
-3.71280879e-01 3.11896540e-02 -4.45325285e-01 6.39653802e-01
5.27962625e-01 -1.38205752e-01 1.04238188e+00 -1.03891146e+00
-1.61019385e-01 -1.70040652e-01 -4.26721066e-01 -1.43488026e+00
-1.38041586e-01 -2.46057421e-01 2.72184491e-01 -1.52745819e+00
-2.39000544e-01 -9.49536860e-01 -7.79235482e-01 1.40111655e-01
2.15569824e-01 6.81488454e-01 -3.05132687e-01 5.69124706e-02
-1.03496945e+00 1.65463150e-01 1.05402768e-01 -1.83163851e-01
-3.80117834e-01 6.71278059e-01 -9.02882099e-01 6.89167261e-01
8.44758391e-01 -6.05050445e-01 -3.05981129e-01 -6.70962095e-01
2.53371030e-01 -1.51002273e-01 2.10734963e-01 -1.85630262e+00
5.33334136e-01 1.01870745e-01 7.50097752e-01 -4.04837400e-01
3.50727051e-01 -1.26236844e+00 4.15996581e-01 6.41325638e-02
1.29801825e-01 1.16614237e-01 1.37134701e-01 7.51559973e-01
1.01107895e-01 3.58910143e-01 3.34959686e-01 4.14278030e-01
-2.75036961e-01 1.66955665e-01 -4.47785467e-01 -2.94450462e-01
6.00913286e-01 -4.29342091e-01 -3.46835434e-01 -6.52620792e-01
1.52657896e-01 -2.85202324e-01 7.37412155e-01 4.73261118e-01
1.52319744e-01 -1.40232086e+00 4.41880614e-01 2.84414828e-01
7.81529173e-02 -3.37098509e-01 3.96467030e-01 8.46005082e-01
-1.49328664e-01 8.49385798e-01 6.01029955e-02 -6.09466970e-01
-9.10838485e-01 8.50602239e-02 6.16582930e-01 -8.22031721e-02
-7.97634795e-02 8.36608827e-01 -4.12702382e-01 -5.21840870e-01
7.03973174e-01 -6.18812025e-01 3.64678055e-01 -4.78587270e-01
6.71515167e-01 6.36853695e-01 2.57664889e-01 -7.66093314e-01
-9.56510484e-01 9.33758259e-01 5.75968623e-01 5.45309372e-02
1.01648974e+00 -7.34941721e-01 2.23589107e-01 7.11997628e-01
1.38880420e+00 6.69855118e-01 -9.45752501e-01 2.66761780e-02
-3.20478566e-02 -5.84071100e-01 4.77344058e-02 -1.03796804e+00
-6.11571848e-01 4.66863453e-01 1.15099132e+00 5.01031458e-01
9.23367858e-01 -2.27565512e-01 1.18496919e+00 6.02258682e-01
1.27604926e+00 -7.13538110e-01 -4.45026904e-01 3.50430191e-01
2.48650596e-01 -1.45447278e+00 -2.55187064e-01 1.95560247e-01
2.17416301e-01 6.35610580e-01 3.68112996e-02 1.79141328e-01
8.34165394e-01 3.18660438e-01 2.82065254e-02 2.41083562e-01
3.52768898e-01 2.18818456e-01 2.38121822e-01 8.95869792e-01
6.60769999e-01 6.64696619e-02 3.07280540e-01 9.73915279e-01
-2.65508950e-01 -3.75811644e-02 4.86754961e-02 8.26748908e-01
-1.21722572e-01 -1.06089509e+00 -5.80986261e-01 2.38106370e-01
-7.20130920e-01 2.17965648e-01 1.58798754e-01 5.08266687e-01
5.32812953e-01 1.61900997e+00 -1.68702587e-01 -6.33876979e-01
3.27322572e-01 -1.70859545e-01 1.87148318e-01 -1.53219000e-01
-4.26948369e-01 -3.66689265e-01 -3.60585541e-01 -8.60127926e-01
-1.88514560e-01 -4.37862515e-01 -9.51542139e-01 -5.00171304e-01
-2.55024940e-01 4.11487669e-02 1.35696208e+00 9.32471395e-01
8.46119404e-01 7.87837803e-01 7.56528616e-01 -1.23769093e+00
1.00994587e-01 -6.17547572e-01 -4.83490229e-01 -7.93204382e-02
6.74600065e-01 -8.24052036e-01 -3.19598794e-01 -5.35402477e-01] | [6.409774303436279, 0.9103546142578125] |
94f4dc22-7891-4b92-87b7-e86a9a18c894 | bicubic-slim-slimmer-slimmest-designing-an | 2305.02126 | null | https://arxiv.org/abs/2305.02126v1 | https://arxiv.org/pdf/2305.02126v1.pdf | Bicubic++: Slim, Slimmer, Slimmest -- Designing an Industry-Grade Super-Resolution Network | We propose a real-time and lightweight single-image super-resolution (SR) network named Bicubic++. Despite using spatial dimensions of the input image across the whole network, Bicubic++ first learns quick reversible downgraded and lower resolution features of the image in order to decrease the number of computations. We also construct a training pipeline, where we apply an end-to-end global structured pruning of convolutional layers without using metrics like magnitude and gradient norms, and focus on optimizing the pruned network's PSNR on the validation set. Furthermore, we have experimentally shown that the bias terms take considerable amount of the runtime while increasing PSNR marginally, hence we have also applied bias removal to the convolutional layers. Our method adds ~1dB on Bicubic upscaling PSNR for all tested SR datasets and runs with ~1.17ms on RTX3090 and ~2.9ms on RTX3070, for 720p inputs and 4K outputs, both in FP16 precision. Bicubic++ won NTIRE 2023 RTSR Track 2 x3 SR competition and is the fastest among all competitive methods. Being almost as fast as the standard Bicubic upsampling method, we believe that Bicubic++ can set a new industry standard. | ['Mustafa Ayazoglu', 'Bahri Batuhan Bilecen'] | 2023-05-03 | null | null | null | null | ['image-super-resolution'] | ['computer-vision'] | [ 3.58758688e-01 -3.24281771e-03 -1.05421670e-01 -4.14484501e-01
-8.25896144e-01 -2.16507539e-01 2.61633635e-01 -4.89954621e-01
-7.38936663e-01 7.68648386e-01 2.81765938e-01 -1.70561939e-01
1.73807055e-01 -7.25732684e-01 -1.06946290e+00 -3.51370960e-01
-1.66988626e-01 -1.95751444e-01 5.76468527e-01 -3.18895727e-01
3.17898840e-01 6.73994958e-01 -1.56311989e+00 8.10759008e-01
3.88950706e-01 1.34504139e+00 2.10831612e-01 9.48689044e-01
2.74431288e-01 1.13448203e+00 -5.40558934e-01 -3.20478618e-01
7.82410502e-01 -4.91888523e-02 -8.16312730e-01 -5.68082213e-01
1.08584654e+00 -6.67066336e-01 -7.65091956e-01 1.00690830e+00
8.61928105e-01 -7.94912800e-02 -1.92364410e-01 -4.94332045e-01
-5.05833268e-01 8.96949947e-01 -9.46874261e-01 7.34521568e-01
-1.71901345e-01 3.25027764e-01 8.37785542e-01 -1.18512213e+00
9.09384787e-01 1.41623878e+00 8.08035970e-01 6.29926920e-01
-1.33236694e+00 -9.96474087e-01 -2.43202612e-01 3.83393675e-01
-1.51464808e+00 -6.86223686e-01 2.66561627e-01 3.05566698e-01
1.02746117e+00 2.73542643e-01 3.23902071e-01 7.58803070e-01
1.50094271e-01 4.64227170e-01 1.38549387e+00 -9.77321565e-02
9.80707482e-02 -2.83189356e-01 1.22446477e-01 4.30136830e-01
7.14486614e-02 2.53429532e-01 -9.38602924e-01 1.44559518e-01
1.19997525e+00 -1.82985961e-01 -4.07780677e-01 2.22479790e-01
-1.18192196e+00 4.35990989e-01 1.03294301e+00 3.06999423e-02
-2.81949192e-01 7.76901662e-01 6.67349279e-01 3.84712785e-01
3.03754359e-01 2.32411325e-01 -4.77823287e-01 -1.30810589e-01
-1.30435252e+00 2.33154744e-01 1.98516294e-01 8.44779134e-01
6.91577792e-01 2.87473083e-01 -1.25861809e-01 9.66547847e-01
-4.00681466e-01 4.12076175e-01 3.60318094e-01 -1.52467763e+00
5.01856625e-01 3.09302658e-01 -7.78881982e-02 -9.31297243e-01
-1.94335014e-01 -6.56021714e-01 -1.12207353e+00 5.53340673e-01
2.11761281e-01 9.74497125e-02 -1.02646387e+00 1.47976422e+00
-1.47342579e-02 3.17172706e-01 5.99438660e-02 1.21364152e+00
9.68564689e-01 7.80258417e-01 -3.38423640e-01 4.07555401e-02
1.40787435e+00 -1.00713158e+00 -2.83142984e-01 -1.65564418e-01
1.85018912e-01 -8.08785558e-01 1.21439528e+00 7.53226280e-01
-1.61313355e+00 -9.17046905e-01 -1.45029259e+00 -6.81663215e-01
2.00556666e-01 3.28335226e-01 4.99282271e-01 3.84547323e-01
-1.43930960e+00 1.13230932e+00 -8.22798371e-01 1.99132636e-01
8.88277531e-01 6.23930275e-01 -1.74369842e-01 -4.50262994e-01
-9.19860780e-01 5.65177143e-01 1.46616548e-01 8.32609739e-03
-8.10232818e-01 -1.23798180e+00 -4.24149990e-01 3.20550174e-01
1.69370741e-01 -3.28853458e-01 1.04937863e+00 -9.22819138e-01
-1.51453733e+00 5.81672847e-01 -1.43871486e-01 -1.07054162e+00
4.53344733e-01 -4.63419616e-01 -5.55138648e-01 4.15662706e-01
-2.83000588e-01 1.02053952e+00 9.56344604e-01 -8.28363359e-01
-7.93591321e-01 -3.39804202e-01 -1.25355527e-01 7.82654900e-03
-1.93743169e-01 2.05828607e-01 -5.51792264e-01 -6.43267632e-01
2.70815283e-01 -7.44625449e-01 -2.86589772e-01 3.30000490e-01
5.35405912e-02 3.87564927e-01 9.19011652e-01 -7.39825904e-01
1.03652465e+00 -2.38406825e+00 -3.03698331e-01 -1.36853307e-02
4.80800152e-01 4.88427192e-01 -1.52050272e-01 -2.79636204e-01
-1.26630634e-01 1.11748263e-01 -1.97370470e-01 -8.74319598e-02
-5.49408138e-01 -4.98756999e-03 -5.98876953e-01 5.04810810e-01
2.12099642e-01 7.28076339e-01 -5.12783408e-01 -2.43383527e-01
-5.52748404e-02 1.02857351e+00 -7.55311370e-01 -4.92567927e-01
2.49364793e-01 1.07272074e-01 8.86251256e-02 3.47025722e-01
1.11362922e+00 -3.38357568e-01 1.22417189e-01 -5.04750252e-01
-2.20618486e-01 5.50046802e-01 -1.25117171e+00 1.77590716e+00
-8.06292117e-01 1.03987002e+00 1.24244459e-01 -4.43705529e-01
9.82001305e-01 -1.23925932e-01 1.10886648e-01 -1.11925530e+00
-8.47052336e-02 4.95660841e-01 -1.20562963e-01 2.85003245e-01
8.65381122e-01 1.72408506e-01 3.90228152e-01 2.18786076e-01
3.09129264e-02 1.17183812e-01 7.15258792e-02 2.88655341e-01
1.30870700e+00 1.25630975e-01 -6.16211258e-03 -5.23403049e-01
5.52879930e-01 -1.12893939e-01 5.75898349e-01 6.39674544e-01
3.57526820e-03 8.49301338e-01 3.72568071e-01 -8.55608881e-01
-1.30092490e+00 -1.04613304e+00 -3.36547285e-01 1.00479150e+00
1.83585748e-01 -4.24728364e-01 -4.04205263e-01 -3.01750183e-01
-3.58744144e-01 2.99794167e-01 -3.26054573e-01 1.77443653e-01
-1.11312151e+00 -7.83104956e-01 8.45409870e-01 6.58573747e-01
1.15552807e+00 -8.80649626e-01 -8.85275364e-01 1.76118270e-01
9.93826836e-02 -1.37008858e+00 -5.45446157e-01 8.13708007e-02
-1.24635518e+00 -7.61588395e-01 -6.72281921e-01 -6.21301591e-01
3.16266328e-01 3.80464345e-01 1.27773404e+00 1.61106810e-01
-6.40231371e-01 -4.41448510e-01 3.70069928e-02 1.98124170e-01
-1.45357504e-01 -1.44682489e-02 -1.75187752e-01 -2.86842972e-01
-3.96774849e-03 -9.38449502e-01 -1.21079755e+00 1.57184914e-01
-6.89891636e-01 7.52365291e-02 7.58896828e-01 7.64053822e-01
8.77009332e-01 -8.47311765e-02 3.59465986e-01 -7.72383928e-01
1.60836428e-02 1.67175785e-01 -8.29955995e-01 -3.73712867e-01
-6.86013222e-01 1.20457016e-01 8.04338038e-01 -4.45380419e-01
-9.86510515e-01 1.41261190e-01 -1.31253913e-01 -5.98554254e-01
4.72294629e-01 -1.96240112e-01 2.62318522e-01 -3.87335777e-01
1.12765992e+00 2.56262451e-01 -1.31655172e-01 -4.89881009e-01
4.95703965e-01 4.50313210e-01 1.15812886e+00 -2.13249102e-01
6.82233751e-01 9.25202072e-01 1.50720701e-01 -8.08834791e-01
-5.31534851e-01 -1.76784828e-01 -2.35854581e-01 2.30993256e-01
5.03518105e-01 -1.47720289e+00 -8.77651453e-01 2.38625929e-01
-9.43917274e-01 -4.93098587e-01 -4.41152930e-01 2.14538306e-01
-4.00599152e-01 2.33662084e-01 -1.07998753e+00 -3.37210119e-01
-9.74553168e-01 -1.10475934e+00 9.30776477e-01 2.12744892e-01
1.28008008e-01 -1.07619531e-01 -4.32092309e-01 1.90719351e-01
9.21285450e-01 -6.84168711e-02 3.90962481e-01 -5.78032359e-02
-9.61936831e-01 2.86950022e-01 -1.01502872e+00 6.77566051e-01
-3.58878016e-01 -2.48668611e-01 -1.20770133e+00 -4.66062248e-01
-3.98229696e-02 -3.35424423e-01 1.34152079e+00 3.20722103e-01
1.40106785e+00 -4.04877067e-01 1.50160536e-01 1.14175415e+00
1.87081957e+00 -2.25066885e-01 1.16074502e+00 2.74812579e-01
5.24122775e-01 7.13827014e-02 5.06071448e-01 3.91493469e-01
-1.54805724e-02 7.05818892e-01 3.50868493e-01 -1.36359468e-01
-7.75725722e-01 2.50983853e-02 4.90256101e-01 3.81430358e-01
-2.41263196e-01 3.13300580e-01 -4.68543440e-01 3.90722454e-01
-1.14268816e+00 -9.61134970e-01 -2.01477662e-01 2.29569650e+00
1.10934079e+00 4.32489455e-01 -6.60732612e-02 2.28058398e-01
6.29187524e-01 4.46789831e-01 -6.05834126e-01 -5.39243400e-01
-4.15860564e-01 1.08892548e+00 1.10431325e+00 5.18191278e-01
-7.25110471e-01 1.02128351e+00 5.71186733e+00 1.08319104e+00
-1.63308227e+00 1.62933677e-01 1.01060653e+00 -6.79213345e-01
2.67931223e-01 -9.81931612e-02 -1.03654385e+00 1.74970254e-01
1.19262886e+00 4.82210405e-02 7.29175091e-01 1.00886488e+00
1.38667688e-01 -2.14750730e-02 -7.79922247e-01 1.06034851e+00
-2.45882735e-01 -1.83969307e+00 -5.65576926e-02 5.62465116e-02
8.25658977e-01 6.25947237e-01 1.41162619e-01 1.67105913e-01
2.98906505e-01 -1.28602576e+00 7.01430798e-01 6.30176589e-02
1.51334870e+00 -1.11411655e+00 5.52057326e-01 -1.60803482e-01
-1.02884531e+00 -1.83157429e-01 -7.94660985e-01 3.83533463e-02
-2.52478808e-01 7.61724591e-01 -5.85702002e-01 1.65987730e-01
1.32981920e+00 5.19178152e-01 -6.88834727e-01 6.32984161e-01
3.07590812e-02 5.58749437e-01 -4.64050442e-01 4.96859938e-01
3.73489931e-02 3.81680399e-01 3.23255211e-01 1.14147305e+00
3.92222464e-01 1.77646637e-01 -4.91678745e-01 6.69898808e-01
-5.61942339e-01 -1.79534793e-01 5.82256056e-02 4.25875515e-01
5.02928853e-01 1.27675700e+00 -6.62461817e-01 -4.90325689e-01
-2.00389162e-01 1.23246658e+00 -2.13673394e-02 6.19367249e-02
-7.79073775e-01 -4.11580443e-01 7.99406886e-01 3.25614393e-01
7.42282867e-01 -5.14936671e-02 -6.41835928e-01 -9.60743129e-01
1.13059454e-01 -1.13430130e+00 7.49175474e-02 -7.59014964e-01
-5.33597946e-01 9.57398415e-01 -5.49342334e-01 -1.07870471e+00
1.78129509e-01 -4.76632833e-01 -2.73873657e-01 9.55622554e-01
-1.90168452e+00 -7.07367301e-01 -5.20662129e-01 5.67822337e-01
6.31629646e-01 1.79373577e-01 2.90871292e-01 5.34582436e-01
-1.71954378e-01 7.03896046e-01 8.49011391e-02 -1.11692175e-01
7.47375250e-01 -9.14950192e-01 9.17827368e-01 9.37308848e-01
-6.16633669e-02 6.69212222e-01 6.95904672e-01 -3.39264154e-01
-1.40217328e+00 -1.15902984e+00 6.02114797e-01 1.17511287e-01
4.65300322e-01 -3.12105924e-01 -1.05388665e+00 4.52963740e-01
2.38245614e-02 8.24807942e-01 -1.73014939e-01 -2.91151106e-01
-8.37935030e-01 -6.23824716e-01 -1.25282979e+00 5.81491530e-01
1.28259206e+00 -3.42193663e-01 4.73245643e-02 1.06110096e-01
1.08148634e+00 -8.20264578e-01 -9.39354658e-01 4.07450259e-01
5.02282381e-01 -1.21154714e+00 1.41763246e+00 -9.18123722e-02
9.35892045e-01 -3.36374134e-01 -3.35557938e-01 -6.76629245e-01
-2.65899211e-01 -7.46458232e-01 -1.18583307e-01 8.22488785e-01
4.40150172e-01 -4.20939922e-01 9.70652163e-01 -5.05830385e-02
-5.71245626e-02 -8.44243050e-01 -1.07406223e+00 -8.28119040e-01
-2.07208693e-02 -4.27510977e-01 5.26469350e-01 4.25715089e-01
-6.49900496e-01 2.11151093e-01 -4.32470888e-01 3.30078751e-01
6.99268460e-01 -6.07455848e-03 6.54976845e-01 -8.28405917e-01
-3.64693969e-01 -3.21486086e-01 -3.36855114e-01 -1.27545679e+00
-4.77023602e-01 -6.75036788e-01 -1.55985028e-01 -8.22175205e-01
1.47543624e-01 -4.19731945e-01 -2.90772706e-01 4.54757392e-01
8.76106024e-02 1.04212511e+00 2.60816097e-01 2.44719759e-01
-3.59267622e-01 2.29151160e-01 1.26022673e+00 1.00520864e-01
-2.77576089e-01 -3.65435600e-01 -7.45677233e-01 6.34378850e-01
6.88015401e-01 -3.56922865e-01 -2.52205133e-01 -5.85089028e-01
2.19337702e-01 5.32510765e-02 6.29319072e-01 -1.31081486e+00
9.50491503e-02 1.82115674e-01 5.99538982e-01 -6.85338199e-01
4.76633847e-01 -4.25209403e-01 3.34507786e-02 5.35866737e-01
-4.24730092e-01 4.54269312e-02 3.36929023e-01 1.73433587e-01
1.09468484e-02 2.35994354e-01 1.50877082e+00 6.98029064e-04
-8.88307869e-01 2.32174680e-01 2.21795321e-01 1.41785964e-01
4.24991220e-01 -1.97044387e-01 -6.39016092e-01 -1.04061909e-01
-4.18886989e-01 -1.51360050e-01 5.33864260e-01 1.48909971e-01
8.17969620e-01 -9.55213964e-01 -9.06721711e-01 5.88660762e-02
-5.33212960e-01 1.45749271e-01 5.46708107e-01 5.81298232e-01
-1.10138607e+00 3.43041033e-01 -4.47342783e-01 -5.04378796e-01
-1.44813955e+00 4.12657648e-01 3.04997653e-01 -4.80016619e-01
-1.25761116e+00 9.92890000e-01 4.70225774e-02 1.17004499e-01
2.00532958e-01 -3.00991535e-01 1.01547718e-01 -3.06562096e-01
1.09935260e+00 6.82834506e-01 2.09717497e-01 -4.05225575e-01
-3.71438742e-01 4.53464091e-01 -2.44962975e-01 -1.61479861e-01
1.60942185e+00 1.28822736e-02 -1.06226951e-01 -2.02995360e-01
1.37157989e+00 5.38021177e-02 -1.53088951e+00 -3.87234092e-01
-1.09587021e-01 -5.68837821e-01 5.34530163e-01 -8.74768913e-01
-1.59060335e+00 8.08822989e-01 9.70645905e-01 -5.47864079e-01
1.61530101e+00 -3.54275763e-01 1.19175804e+00 2.85268426e-01
3.91702294e-01 -9.03007925e-01 1.26711009e-02 3.30999136e-01
1.10269511e+00 -9.37385559e-01 5.76883972e-01 -4.64276910e-01
-4.21510011e-01 1.08817816e+00 5.13285875e-01 -6.56746864e-01
2.39344627e-01 6.14817262e-01 -1.12019114e-01 1.05285555e-01
-9.48224843e-01 1.88541770e-01 -5.68863824e-02 2.72079676e-01
2.47714356e-01 -2.64919698e-01 -3.71445239e-01 2.40390912e-01
-4.83338505e-01 3.94990981e-01 7.99343944e-01 5.75185597e-01
-4.03010190e-01 -8.19172740e-01 -3.26885581e-01 2.80966848e-01
-9.46876645e-01 -6.12737656e-01 2.68569201e-01 6.92551792e-01
4.61320346e-03 4.42910850e-01 2.12411836e-01 -3.80814552e-01
2.45043233e-01 -5.80747187e-01 3.10157299e-01 -2.47462973e-01
-8.19269121e-01 -1.32016063e-01 -1.35041995e-03 -1.19233477e+00
-1.31647103e-02 -2.48328179e-01 -1.49580956e+00 -7.28211164e-01
3.08044832e-02 -4.44034159e-01 1.02156556e+00 2.84678131e-01
7.04886913e-01 5.63992679e-01 6.47805870e-01 -9.74872708e-01
-6.05762362e-01 -7.41509497e-01 -2.18982577e-01 -1.58238202e-01
3.81932974e-01 -2.54108608e-02 -3.59609187e-01 4.57169414e-02] | [10.99025821685791, -1.906261682510376] |
4ba37db7-6be9-48f9-afb6-f54a5e051341 | stepwise-extractive-summarization-and | 2010.02744 | null | https://arxiv.org/abs/2010.02744v1 | https://arxiv.org/pdf/2010.02744v1.pdf | Stepwise Extractive Summarization and Planning with Structured Transformers | We propose encoder-centric stepwise models for extractive summarization using structured transformers -- HiBERT and Extended Transformers. We enable stepwise summarization by injecting the previously generated summary into the structured transformer as an auxiliary sub-structure. Our models are not only efficient in modeling the structure of long inputs, but they also do not rely on task-specific redundancy-aware modeling, making them a general purpose extractive content planner for different tasks. When evaluated on CNN/DailyMail extractive summarization, stepwise models achieve state-of-the-art performance in terms of Rouge without any redundancy aware modeling or sentence filtering. This also holds true for Rotowire table-to-text generation, where our models surpass previously reported metrics for content selection, planning and ordering, highlighting the strength of stepwise modeling. Amongst the two structured transformers we test, stepwise Extended Transformers provides the best performance across both datasets and sets a new standard for these challenges. | ['Ryan Mcdonald', 'Blaž Bratanič', 'Daniele Pighin', 'Jakub Adamek', 'Joshua Maynez', 'Shashi Narayan'] | 2020-10-06 | null | https://aclanthology.org/2020.emnlp-main.339 | https://aclanthology.org/2020.emnlp-main.339.pdf | emnlp-2020-11 | ['table-to-text-generation'] | ['natural-language-processing'] | [ 5.44018447e-01 6.40053511e-01 -2.75927186e-01 -1.70299057e-02
-1.10886109e+00 -7.55497217e-01 1.01222408e+00 4.82578009e-01
-4.87758666e-01 8.99986088e-01 1.09867144e+00 -1.41924515e-01
-4.06888835e-02 -6.12313330e-01 -8.89627159e-01 -1.20279945e-01
1.00158058e-01 8.38486135e-01 1.83978513e-01 -5.83649874e-01
4.20244843e-01 2.12329030e-01 -1.36191332e+00 7.85476208e-01
1.20445263e+00 6.99794590e-01 2.75043339e-01 8.62982750e-01
-3.18952471e-01 1.10997057e+00 -8.69333684e-01 -6.70999229e-01
5.83210140e-02 -6.41409695e-01 -1.23772919e+00 4.05297279e-02
7.10506141e-01 -3.94202143e-01 -2.51867175e-01 5.29757679e-01
7.52777278e-01 -4.07588296e-02 5.83609164e-01 -7.91764796e-01
-4.35628414e-01 1.63742673e+00 -1.38210535e-01 3.54541391e-01
6.19856238e-01 2.62171715e-01 1.52180004e+00 -6.64871275e-01
1.07477629e+00 1.00289571e+00 5.95665395e-01 4.86722797e-01
-1.23753035e+00 -1.65976301e-01 2.64742494e-01 7.52845705e-02
-6.85006797e-01 -8.39141190e-01 5.01169622e-01 -6.91148788e-02
1.58232582e+00 7.82401085e-01 6.35499835e-01 1.16264093e+00
2.86288649e-01 1.30425704e+00 6.93191111e-01 -1.14377230e-01
-5.37696527e-03 -1.93690017e-01 1.80129763e-02 6.05980217e-01
3.13502550e-01 -4.42419320e-01 -9.46706057e-01 1.14931799e-01
1.66981399e-01 -4.28391129e-01 -2.29536414e-01 1.18853219e-01
-1.67727137e+00 4.73741084e-01 2.33873889e-01 2.21956030e-01
-6.07987761e-01 2.35254765e-01 9.38804805e-01 3.74061614e-01
6.03231966e-01 1.14915192e+00 -4.77921695e-01 -4.19804573e-01
-1.58314502e+00 6.14435256e-01 9.77704346e-01 1.43258011e+00
3.03255588e-01 2.68772572e-01 -1.10287845e+00 4.86870527e-01
-3.92612368e-01 3.10556889e-01 4.98388439e-01 -8.26619327e-01
1.09234762e+00 6.93253696e-01 -1.34574786e-01 -4.93874341e-01
-5.26694775e-01 -1.04203141e+00 -9.54165637e-01 -4.14612532e-01
4.71702330e-02 -3.24478857e-02 -9.26213086e-01 1.30138361e+00
-2.19420433e-01 -3.68997306e-01 1.34695306e-01 3.61360431e-01
1.37454224e+00 7.54378855e-01 -1.79674461e-01 -3.56355190e-01
1.35627556e+00 -1.27274954e+00 -7.53591776e-01 -4.19988304e-01
6.72825873e-01 -5.77458799e-01 1.05498815e+00 4.75561768e-01
-1.69004118e+00 -2.14438677e-01 -1.03455925e+00 -4.77211386e-01
-1.39765903e-01 3.06335092e-01 4.58761722e-01 1.28384262e-01
-1.30294788e+00 8.90499055e-01 -6.14780307e-01 -3.08338761e-01
5.07146776e-01 1.91722646e-01 -3.63129765e-01 3.46632689e-01
-9.12305832e-01 9.45572197e-01 6.35217607e-01 -2.09122345e-01
-9.10258770e-01 -9.50999081e-01 -7.92675674e-01 4.63514477e-01
5.21988332e-01 -1.20687354e+00 1.68758237e+00 -3.22618693e-01
-1.46689701e+00 6.38266802e-01 -2.05496266e-01 -1.11514390e+00
8.35978031e-01 -5.47940195e-01 7.78902844e-02 1.10300310e-01
1.84081271e-01 7.52800763e-01 6.47060454e-01 -9.27541554e-01
-5.49846709e-01 1.27638116e-01 6.05689473e-02 3.57819408e-01
-2.65725881e-01 2.16243956e-02 -2.86953986e-01 -7.91885614e-01
-2.61095107e-01 -4.61183161e-01 -2.95276254e-01 -8.15045059e-01
-1.26907408e+00 -1.44179046e-01 4.46041375e-01 -9.19027388e-01
1.61836672e+00 -1.47323227e+00 5.66719830e-01 -3.94010633e-01
3.54142994e-01 2.40182474e-01 -3.81876439e-01 1.13931346e+00
1.85137823e-01 4.39263821e-01 -3.79836410e-01 -8.38769257e-01
2.61104912e-01 -1.38772294e-01 -6.50450945e-01 -1.83372602e-01
4.86109167e-01 1.32751989e+00 -1.00119078e+00 -6.64074361e-01
-6.80599436e-02 -1.28322721e-01 -6.78067446e-01 1.76626876e-01
-7.36460686e-01 1.87408045e-01 -2.13569239e-01 2.87164718e-01
2.47535124e-01 -1.37699023e-01 -5.63119538e-02 -1.84436798e-01
-2.29811490e-01 1.00895631e+00 -6.31445944e-01 1.89600551e+00
-5.10470808e-01 7.94054806e-01 -2.25737855e-01 -5.75158715e-01
5.84204733e-01 2.50519514e-01 2.63028145e-01 -7.19062388e-01
8.82018544e-03 2.07142636e-01 -2.18606353e-01 -3.72128010e-01
1.35902405e+00 1.88691333e-01 -2.81450480e-01 5.68079293e-01
3.85465741e-01 -4.21964735e-01 8.75515223e-01 8.03320467e-01
1.45685625e+00 2.72747651e-02 4.66188997e-01 -3.02934766e-01
2.31742844e-01 1.20142810e-01 3.05665284e-01 1.02542961e+00
5.72312415e-01 9.00937080e-01 8.81154716e-01 -1.08712047e-01
-1.18560016e+00 -9.42266285e-01 4.37246710e-01 1.11338985e+00
-3.32255125e-01 -1.13226044e+00 -8.51575196e-01 -9.09907758e-01
-3.18064600e-01 1.20558739e+00 -4.78186220e-01 -5.12748770e-02
-7.06682861e-01 -4.98079866e-01 8.19691479e-01 5.04928648e-01
3.66986275e-01 -1.29050624e+00 -8.15525830e-01 3.96452904e-01
-6.08872771e-01 -1.23046970e+00 -6.68165505e-01 2.49169722e-01
-1.05353355e+00 -7.51705825e-01 -6.06693327e-01 -3.66942495e-01
2.97330469e-01 -1.71259403e-01 1.65345562e+00 -7.09014460e-02
1.41200587e-01 2.73150057e-01 -4.63757277e-01 -5.52101791e-01
-7.95839608e-01 9.93000686e-01 -4.07357365e-01 -3.21120232e-01
-5.12941062e-01 -6.85163736e-01 -5.07659495e-01 -3.04487616e-01
-9.69183624e-01 5.59011281e-01 1.00035441e+00 6.56691551e-01
3.36222887e-01 -3.20648909e-01 7.68506229e-01 -1.13434041e+00
1.17344809e+00 -2.08845377e-01 -7.23331720e-02 2.50472844e-01
-4.98063952e-01 4.49055642e-01 9.74844515e-01 -1.24752922e-02
-9.83248889e-01 -2.03205749e-01 -2.30882600e-01 1.23703800e-01
1.55289292e-01 6.95706069e-01 2.47302162e-03 8.47288847e-01
7.86929131e-01 6.79198921e-01 -2.05458224e-01 -6.21699393e-01
5.72758138e-01 3.08739424e-01 6.67340219e-01 -3.49845082e-01
7.59020984e-01 2.63630092e-01 -1.41417393e-02 -7.87268400e-01
-8.37582111e-01 -2.30774567e-01 -5.18982410e-01 1.69233248e-01
7.62933731e-01 -7.55260646e-01 -4.21560377e-01 3.07951659e-01
-1.51804006e+00 -3.28366399e-01 -8.92343998e-01 -1.29985332e-01
-8.08927536e-01 4.27972674e-01 -7.39895523e-01 -4.97035831e-01
-1.25631118e+00 -9.95096445e-01 1.40386617e+00 -1.64320111e-01
-5.80862701e-01 -8.31229985e-01 -1.00001186e-01 1.51757747e-01
5.58602750e-01 1.97887465e-01 1.00311744e+00 -1.02210164e+00
-7.60074437e-01 -1.94501802e-02 -3.80226560e-02 3.42106432e-01
-1.58042192e-01 -5.96803240e-03 -6.74037218e-01 -1.09936990e-01
-4.83826071e-01 -3.33275169e-01 1.41998434e+00 1.89904347e-01
9.42142010e-01 -8.86658192e-01 -6.31454214e-02 4.05641884e-01
8.82479429e-01 -2.69797653e-01 7.36257255e-01 4.07397985e-01
5.65829754e-01 4.29368496e-01 3.87261629e-01 5.71568012e-01
5.32081544e-01 5.83789051e-01 5.82921922e-01 7.00673833e-03
-3.79524022e-01 -5.65290987e-01 6.27328813e-01 1.02777112e+00
-7.16605186e-02 -5.92838466e-01 -6.74820960e-01 7.04641163e-01
-1.85643446e+00 -1.25138462e+00 -2.59983689e-01 1.89566123e+00
1.14420938e+00 2.97248095e-01 2.41244078e-01 1.84518039e-01
7.65421912e-02 5.42216837e-01 -2.57695764e-01 -6.47978187e-01
-2.34608963e-01 4.08053547e-01 3.93143088e-01 3.68700594e-01
-9.73574221e-01 1.14314616e+00 6.40158319e+00 9.27634418e-01
-8.45529497e-01 -3.78401913e-02 3.43272537e-01 -4.81528759e-01
-7.34407663e-01 -5.79630844e-02 -9.09466743e-01 3.52474064e-01
8.69152546e-01 -5.27704060e-01 2.03715384e-01 5.10880649e-01
3.01411450e-01 -1.83250420e-02 -1.32620811e+00 4.67025787e-01
8.01600814e-02 -1.87520301e+00 6.69314146e-01 -1.21743657e-01
6.45324767e-01 9.07307044e-02 1.27423434e-02 4.36872959e-01
3.63901138e-01 -1.00344038e+00 1.24472332e+00 4.55629796e-01
7.77875423e-01 -5.75821102e-01 5.47066450e-01 5.25133789e-01
-8.06203187e-01 4.47416641e-02 -1.54362902e-01 1.48487568e-01
4.66291517e-01 7.01399386e-01 -1.12231624e+00 8.96100104e-01
2.62497514e-01 8.35151196e-01 -9.52895939e-01 1.00904977e+00
-3.42915207e-01 7.39048958e-01 -1.73882425e-01 -5.27982898e-02
3.75986099e-01 2.18762279e-01 1.04792607e+00 1.75595844e+00
2.96369582e-01 -4.70773071e-01 -6.87625632e-02 7.65303791e-01
-5.22248983e-01 1.81100786e-01 -4.89426285e-01 -2.12998629e-01
3.42077732e-01 1.15092170e+00 -5.45466006e-01 -6.17258251e-01
1.95540249e-01 1.08510351e+00 3.19333434e-01 9.15893540e-02
-6.62170589e-01 -5.02850235e-01 2.06646457e-01 2.01502636e-01
3.88825268e-01 -1.16163544e-01 -4.65367556e-01 -1.30639637e+00
1.98056981e-01 -1.22208822e+00 2.60658026e-01 -5.09666920e-01
-6.76023543e-01 9.59049761e-01 2.90857702e-01 -1.10136843e+00
-6.50180280e-01 5.83334379e-02 -7.77690589e-01 4.79574531e-01
-1.43232977e+00 -1.26057887e+00 -9.28171948e-02 1.22441687e-01
1.13581932e+00 -1.14718921e-01 6.42594635e-01 -8.63876119e-02
-5.16561389e-01 5.89164257e-01 -1.16486579e-01 -2.00750872e-01
4.96321827e-01 -1.62540877e+00 1.06164110e+00 1.14629233e+00
2.67314196e-01 5.36225140e-01 1.12759459e+00 -6.84362710e-01
-1.37387431e+00 -1.24948514e+00 1.22564447e+00 -5.30980527e-01
4.16526526e-01 -5.83519757e-01 -4.91986841e-01 7.60708869e-01
9.66027439e-01 -8.78007412e-01 2.53433585e-01 1.20078050e-01
-3.16745013e-01 -2.12888792e-02 -7.12198913e-01 8.25679719e-01
1.47214282e+00 -1.06393903e-01 -6.94371641e-01 5.56318700e-01
1.21931839e+00 -6.25743330e-01 -6.90147340e-01 2.88610548e-01
3.41932863e-01 -1.10533178e+00 7.52731740e-01 -7.26891220e-01
1.18596697e+00 1.26903221e-01 3.12136292e-01 -1.68268740e+00
-3.08254123e-01 -1.11000466e+00 -4.38191205e-01 1.46300685e+00
8.35230172e-01 -3.91666859e-01 7.68176019e-01 4.69836826e-03
-7.12651253e-01 -9.73765671e-01 -6.84314728e-01 -9.28220332e-01
-4.13628556e-02 -2.93540806e-01 7.81015098e-01 4.54860926e-01
1.87875569e-01 8.88188124e-01 -3.76760393e-01 -5.37027895e-01
4.09315437e-01 1.29708514e-01 8.76221180e-01 -8.48138154e-01
-2.79374331e-01 -8.80368888e-01 1.09981112e-01 -1.19754708e+00
1.54855207e-01 -1.23079312e+00 -7.42053911e-02 -2.40419841e+00
3.09998482e-01 4.50203829e-02 2.69549519e-01 5.75454772e-01
-2.42824256e-01 -5.36471196e-02 6.10675573e-01 -3.24493535e-02
-8.85098159e-01 8.54159832e-01 1.28945482e+00 -4.07841712e-01
-3.12352806e-01 1.57660782e-01 -1.16027701e+00 2.48059049e-01
7.20453739e-01 -3.90699536e-01 -5.95406830e-01 -6.70649350e-01
6.04813993e-01 1.38450459e-01 8.58789459e-02 -1.11813080e+00
3.10443074e-01 4.20925468e-02 -8.17917213e-02 -8.62913132e-01
1.97311506e-01 -3.18813503e-01 5.00925034e-02 4.00186211e-01
-7.88859665e-01 3.28860849e-01 2.03557447e-01 3.09765100e-01
-1.67956829e-01 -2.49003753e-01 2.58658260e-01 -2.29637071e-01
-2.52520442e-01 2.39702657e-01 -6.68684781e-01 4.59575891e-01
3.89749497e-01 -9.61748958e-02 -7.48869419e-01 -7.48441994e-01
-4.42425728e-01 2.64902622e-01 2.86357254e-01 3.66556674e-01
5.74344933e-01 -7.90876389e-01 -1.20785916e+00 -2.29735121e-01
9.56175402e-02 2.33212963e-01 4.49424982e-02 9.75179493e-01
-5.71552098e-01 7.44389355e-01 -4.56544347e-02 -3.33815604e-01
-1.15185606e+00 3.40593815e-01 -1.64855830e-03 -1.15095794e+00
-7.39559114e-01 6.74232185e-01 -1.59939155e-01 -3.22537303e-01
4.60893214e-02 -8.44289780e-01 -1.73950359e-01 4.29536432e-01
4.46474075e-01 5.24400473e-01 5.54575205e-01 -2.58951515e-01
-7.73320496e-02 -1.20399646e-01 -4.19870377e-01 -2.20609725e-01
1.43741035e+00 1.30223826e-01 -2.54738182e-01 2.48692632e-01
8.73166978e-01 2.94541687e-01 -8.89638424e-01 -1.43624157e-01
2.08305910e-01 1.05982788e-01 -2.80133754e-01 -1.07761824e+00
-6.92010462e-01 7.46052921e-01 -6.28702343e-01 2.92577118e-01
1.10536623e+00 6.92478148e-03 9.86837208e-01 6.03225231e-01
1.03081048e-01 -8.98309410e-01 1.39259502e-01 9.16773498e-01
1.46106327e+00 -5.88252485e-01 2.84519166e-01 -3.55056435e-01
-9.94038165e-01 9.07079279e-01 3.09628397e-01 1.30139070e-03
-1.98712289e-01 2.47761250e-01 -4.87329334e-01 -1.69898272e-01
-1.42318642e+00 -3.37657690e-01 4.75371450e-01 4.74169642e-01
4.63885069e-01 -1.62493050e-01 -4.25069183e-01 5.15443981e-01
-9.04942691e-01 -1.88046217e-01 8.60500872e-01 7.54276931e-01
-3.71273607e-01 -1.01438081e+00 2.45545194e-01 9.33456719e-01
-5.50527632e-01 -5.45797944e-01 -7.79641926e-01 7.15822875e-01
-5.53489506e-01 8.75792563e-01 -1.29761413e-01 -4.41924542e-01
6.91537201e-01 9.67290178e-02 5.45334458e-01 -9.49523687e-01
-1.38636017e+00 -2.28473708e-01 7.02454746e-01 -5.19341707e-01
5.92134185e-02 -7.44239330e-01 -1.05886400e+00 -2.97788024e-01
-9.55204070e-02 2.58599192e-01 4.98799950e-01 7.66526878e-01
7.93037236e-01 8.90521109e-01 2.89336562e-01 -8.98110688e-01
-7.95475543e-01 -1.32241297e+00 -1.04250155e-01 1.88688442e-01
4.72854346e-01 1.33619189e-01 -2.72110421e-02 1.04652911e-01] | [12.379104614257812, 9.354162216186523] |
9b5ce76c-2716-499b-9f72-99c214bf32d1 | object-topological-character-acquisition-by | 2306.10664 | null | https://arxiv.org/abs/2306.10664v1 | https://arxiv.org/pdf/2306.10664v1.pdf | Object Topological Character Acquisition by Inductive Learning | Understanding the shape and structure of objects is undoubtedly extremely important for object recognition, but the most common pattern recognition method currently used is machine learning, which often requires a large number of training data. The problem is that this kind of object-oriented learning lacks a priori knowledge. The amount of training data and the complexity of computations are very large, and it is hard to extract explicit knowledge after learning. This is typically called "knowing how without knowing why". We adopted a method of inductive learning, hoping to derive conceptual knowledge of the shape of an object and its formal representation based on a small number of positive examples. It is clear that implementing object recognition is not based on simple physical features such as colors, edges, textures, etc., but on their common geometry, such as topologies, which are stable, persistent, and essential to recognition. In this paper, a formal representation of topological structure based on object's skeleton (RTS) was proposed and the induction process of "seeking common ground" is realized. This research helps promote the method of object recognition from empiricism to rationalism. | ['Yiran Wei', 'Liping Yu', 'Wei Hui'] | 2023-06-19 | null | null | null | null | ['object-recognition'] | ['computer-vision'] | [ 2.45973960e-01 1.25026718e-01 -1.96936965e-01 -4.38479602e-01
1.82396725e-01 -3.53543460e-01 4.99837667e-01 3.88698190e-01
-1.43371612e-01 7.04093695e-01 -3.27241004e-01 -5.18291056e-01
-6.97673559e-01 -1.15576994e+00 -5.09291947e-01 -6.61996186e-01
-1.01226479e-01 5.44281542e-01 2.80802339e-01 -3.05000961e-01
8.48643720e-01 1.04758906e+00 -1.88515258e+00 2.26007387e-01
6.64300919e-01 9.59999621e-01 1.89550757e-01 3.52721334e-01
-6.33435488e-01 8.12793255e-01 -2.92506963e-01 -2.48759553e-01
-4.82595451e-02 -3.60130012e-01 -1.23043549e+00 3.26616824e-01
-1.87248230e-01 2.23400667e-02 1.16531529e-01 9.22717154e-01
-4.82038707e-02 -4.14093100e-02 1.07335114e+00 -1.06553495e+00
-8.51498783e-01 2.33734205e-01 -2.64713943e-01 -2.24061102e-01
2.85671234e-01 -2.30928585e-01 9.67224300e-01 -8.06872308e-01
4.25977558e-01 1.09098279e+00 5.77363908e-01 3.44014496e-01
-1.09792066e+00 -1.17264822e-01 -3.91650349e-02 5.97537994e-01
-1.50294447e+00 2.33459305e-02 1.03058624e+00 -7.22740531e-01
3.93675417e-01 4.33972955e-01 9.57008719e-01 3.47528011e-01
6.71673119e-02 7.14697063e-01 1.29850352e+00 -9.72804546e-01
4.58157927e-01 5.12871087e-01 2.15577081e-01 9.08060133e-01
5.01711249e-01 1.23194568e-01 -1.41073257e-01 1.92191720e-01
1.00769305e+00 2.05584511e-01 3.57509069e-02 -5.93828559e-01
-7.93371379e-01 7.90513515e-01 4.13395256e-01 8.99462938e-01
-1.70692742e-01 -2.37417012e-01 3.74044590e-02 3.29185247e-01
8.98747891e-03 6.83553696e-01 -5.58242023e-01 8.55062455e-02
-4.42714572e-01 -2.58903712e-01 9.08322990e-01 5.87518454e-01
1.18438756e+00 -8.47469047e-02 7.29186893e-01 5.50561249e-01
3.56719553e-01 5.52215934e-01 4.93704230e-01 -6.27226353e-01
-3.52192581e-01 1.27758312e+00 -2.70471931e-01 -1.49504304e+00
-2.76564419e-01 -2.53495514e-01 -7.51218379e-01 6.32891178e-01
3.69351953e-01 2.19606414e-01 -6.96558356e-01 1.33402658e+00
2.76069552e-01 -1.31957814e-01 -1.52147608e-02 6.98124588e-01
7.99416602e-01 5.12233436e-01 -1.83749035e-01 -3.43566865e-01
1.00192082e+00 -3.85143489e-01 -4.62585777e-01 2.18948379e-01
7.62242794e-01 -5.96921623e-01 1.00662303e+00 6.72405243e-01
-5.90673745e-01 -6.04143381e-01 -8.33096325e-01 1.71543717e-01
-8.85200679e-01 3.21482629e-01 1.07391012e+00 7.15165615e-01
-6.43282235e-01 7.62771547e-01 -4.07254606e-01 -4.40766752e-01
2.68434644e-01 4.31570351e-01 -4.51984823e-01 1.74246147e-01
-7.01996684e-01 1.12923157e+00 7.65480459e-01 3.04059774e-01
-3.49827200e-01 -1.57792956e-01 -5.63814223e-01 -2.17711687e-01
4.26186204e-01 -3.99912447e-01 7.76415765e-01 -1.30007923e+00
-1.29235899e+00 9.06836390e-01 -1.65416583e-01 -1.11358643e-01
2.50099748e-01 -9.29810330e-02 -5.79757631e-01 2.55156219e-01
-3.50306004e-01 1.46372139e-01 1.02407300e+00 -1.82924283e+00
-1.66589260e-01 -4.75285918e-01 3.69061455e-02 -1.45497099e-01
-3.62746716e-01 -1.45367831e-01 2.18157843e-01 -3.13835055e-01
7.21716344e-01 -6.08051002e-01 -1.51240304e-01 1.33115962e-01
-2.08109528e-01 -5.02755523e-01 9.17492449e-01 -3.17459971e-01
1.13640583e+00 -2.07479215e+00 -1.79290295e-01 9.10628676e-01
2.31439367e-01 2.53698230e-01 1.97751984e-01 4.41787094e-01
-2.00067624e-01 2.00558275e-01 -3.01594108e-01 8.78612995e-01
-1.57134280e-01 4.58945096e-01 -3.20924670e-01 1.65997520e-01
3.05841357e-01 5.80905735e-01 -9.36243773e-01 -7.24658072e-01
3.16655457e-01 1.28817156e-01 -3.08117062e-01 2.14296356e-02
-1.49719507e-01 1.65721804e-01 -8.27290475e-01 7.31046140e-01
4.30575997e-01 -4.94331032e-01 4.32069182e-01 -3.52394879e-01
-2.86961854e-01 8.35871175e-02 -1.42977667e+00 1.00207484e+00
-1.99754223e-01 5.88207483e-01 -4.20652628e-01 -1.68481553e+00
1.39144135e+00 2.38126904e-01 4.98494565e-01 -4.88508672e-01
3.13571393e-01 3.87336820e-01 1.75379932e-01 -9.11537468e-01
7.05356821e-02 -4.04334635e-01 3.05466533e-01 5.21440923e-01
-3.00497830e-01 -4.03176606e-01 2.06140235e-01 -1.26343384e-01
6.60035729e-01 2.40946785e-02 4.54606473e-01 -4.38119799e-01
7.58826792e-01 1.05653152e-01 2.63981968e-01 2.76358545e-01
4.39278722e-01 1.32393479e-01 3.28681052e-01 -8.72786283e-01
-9.10419345e-01 -1.07300174e+00 -4.59955245e-01 5.62596619e-01
2.47304231e-01 -2.07992196e-01 -5.77880025e-01 -4.31956291e-01
-3.08927055e-02 5.47610700e-01 -6.55241013e-01 -1.91510528e-01
-4.69666779e-01 -4.62397754e-01 3.42098773e-02 2.99212456e-01
4.60645527e-01 -1.28506732e+00 -6.56329453e-01 2.11282253e-01
2.12025017e-01 -4.56537426e-01 5.14072418e-01 5.79022281e-02
-1.50317347e+00 -1.51073408e+00 -2.51556933e-01 -8.66723716e-01
1.23698437e+00 2.62745291e-01 9.27227736e-01 7.46061027e-01
-5.69752872e-01 4.25686330e-01 -5.19440174e-01 -6.83189213e-01
-3.29367638e-01 -2.49676332e-01 1.27183720e-01 1.90550521e-01
3.24869961e-01 -6.51791036e-01 -2.53848255e-01 3.70096058e-01
-9.90758240e-01 -3.71455587e-02 1.07786787e+00 6.66696846e-01
5.40326774e-01 6.07502222e-01 7.04600096e-01 -6.19061887e-01
3.25442463e-01 4.33927923e-02 -2.71400243e-01 6.02345288e-01
-5.71411252e-01 1.92791030e-01 6.21008098e-01 -3.78788203e-01
-1.03271461e+00 4.41595912e-02 6.93559721e-02 5.29426113e-02
-4.72683400e-01 6.03670180e-01 -2.58488417e-01 -2.82698065e-01
6.19563818e-01 5.71768582e-01 2.93691039e-01 -5.42300224e-01
2.28492126e-01 5.12556672e-01 2.84516007e-01 -8.98097038e-01
7.87290514e-01 4.56674814e-01 3.87513638e-01 -1.42012441e+00
-7.05904603e-01 -3.67135286e-01 -1.00583899e+00 -4.19875622e-01
4.18203026e-01 -4.26039360e-02 -7.91496158e-01 2.53794491e-01
-1.04272330e+00 -5.75855724e-04 -5.19407868e-01 8.32419217e-01
-6.25238657e-01 5.07072091e-01 4.08805571e-02 -1.07560587e+00
1.28035381e-01 -5.34152567e-01 4.08764720e-01 1.72985837e-01
-1.78861409e-01 -1.22725034e+00 -4.04835083e-02 1.74451411e-01
2.03048345e-02 1.40613168e-01 1.35177922e+00 -6.47080898e-01
-6.98403478e-01 -2.51035303e-01 -2.45014757e-01 4.85028982e-01
4.82394993e-01 2.99561143e-01 -6.31463647e-01 5.02568968e-02
3.50431979e-01 -3.89780015e-01 4.86310750e-01 4.88247536e-02
1.39708817e+00 -3.63311023e-01 -2.86644727e-01 3.62015925e-02
1.60520363e+00 5.22200048e-01 9.61462200e-01 8.90916735e-02
4.60157782e-01 9.50723946e-01 5.55324018e-01 1.92324862e-01
-3.65796462e-02 2.67326415e-01 2.24595308e-01 -1.43837392e-01
7.21715763e-02 -5.26664741e-02 -9.82270166e-02 1.15990090e+00
-6.89123869e-01 5.02984405e-01 -1.04508352e+00 3.28369796e-01
-1.61213613e+00 -1.19309950e+00 -2.84186244e-01 2.26360917e+00
8.25862467e-01 2.69531757e-01 -8.33318830e-02 7.49885023e-01
6.40065670e-01 -6.28679752e-01 -1.92524612e-01 -4.20063376e-01
-9.36580077e-02 1.77672073e-01 -8.80815685e-02 3.65862995e-01
-7.62108743e-01 6.60488248e-01 6.55686951e+00 7.66915560e-01
-1.46667016e+00 -5.67713499e-01 3.40487301e-01 8.55993867e-01
-2.38693073e-01 2.29689896e-01 -5.57357907e-01 1.91966891e-01
3.70200276e-01 -6.68982267e-02 1.72501743e-01 9.70519900e-01
-3.83546390e-02 -2.74180621e-01 -1.18910599e+00 8.48190367e-01
9.29684117e-02 -1.30091977e+00 4.56731170e-01 6.33939356e-02
4.24892068e-01 -7.98223794e-01 -3.26376617e-01 -4.63740565e-02
1.72704123e-02 -1.14612222e+00 5.23192883e-01 8.52970839e-01
4.66015697e-01 -4.42615718e-01 6.42530620e-01 5.92638850e-01
-1.20527136e+00 -7.95873404e-02 -5.74689388e-01 -3.94059658e-01
-3.06530088e-01 7.07542360e-01 -8.93958509e-01 5.53426147e-01
3.50916028e-01 4.65105325e-01 -7.24012434e-01 1.26596057e+00
-2.59282142e-01 4.93993133e-01 -4.94953871e-01 -5.75355113e-01
-4.85154241e-02 -3.62943769e-01 1.95207745e-01 9.86183584e-01
2.96648920e-01 5.64863324e-01 6.09911159e-02 8.70253325e-01
3.69017839e-01 5.49029112e-01 -8.04962516e-01 -1.20631978e-01
2.27162868e-01 1.07993841e+00 -1.19585967e+00 -4.01807427e-01
-2.66277254e-01 1.88423261e-01 4.34251353e-02 5.92089184e-02
-3.41678351e-01 -4.99427885e-01 -1.44689202e-01 2.39027560e-01
2.73174822e-01 -2.89724171e-01 -5.98572671e-01 -9.59173381e-01
1.28270090e-02 -7.30300128e-01 9.74247158e-02 -5.96948922e-01
-1.30481088e+00 1.78487748e-01 5.03954366e-02 -1.19114518e+00
1.43774142e-02 -1.04494762e+00 -8.42830658e-01 5.46277761e-01
-9.96346831e-01 -1.05285394e+00 -7.45298415e-02 6.58591688e-01
1.63135260e-01 -2.08253130e-01 9.12890494e-01 -7.64959902e-02
-1.58248484e-01 2.23551422e-01 -4.83768387e-03 3.63885999e-01
1.79377459e-02 -1.06757951e+00 -5.19525588e-01 1.81149140e-01
3.68237078e-01 6.65839434e-01 5.68949103e-01 -4.86506850e-01
-1.52210748e+00 -4.35146213e-01 1.13631678e+00 -2.58804590e-01
6.69036925e-01 4.83664311e-02 -1.15780258e+00 2.03834340e-01
-4.03961271e-01 -1.09351374e-01 7.86148429e-01 4.05094415e-01
-3.15256715e-01 -4.21677798e-01 -1.04879057e+00 5.15269279e-01
8.15625668e-01 -4.05752867e-01 -1.18204975e+00 3.30835462e-01
2.83244904e-03 4.79101479e-01 -8.40842485e-01 5.30788958e-01
6.63739920e-01 -1.03271580e+00 9.45310175e-01 -8.57541621e-01
2.34704912e-01 -5.04014194e-01 -3.08444463e-02 -8.55462551e-01
-4.10666645e-01 -7.81275630e-02 1.20935902e-01 9.51145649e-01
5.44684231e-01 -7.20789433e-01 9.23967004e-01 3.95337373e-01
4.58925515e-02 -8.79057884e-01 -5.43986738e-01 -1.12047851e+00
-5.07771298e-02 -3.58246684e-01 3.93435061e-01 1.35132682e+00
1.87916026e-01 4.72673833e-01 1.38194770e-01 -1.79788053e-01
6.32492602e-01 5.89182079e-01 6.37054741e-01 -1.93310988e+00
-1.55684024e-01 -5.58211088e-01 -8.11422646e-01 -6.00665152e-01
-4.95180823e-02 -7.78278410e-01 -2.53086448e-01 -1.62692881e+00
7.65886232e-02 -7.89969563e-01 -2.21119642e-01 4.28184748e-01
3.33223671e-01 -1.30082950e-01 -1.19517902e-02 3.84411603e-01
-3.26746643e-01 4.46853220e-01 1.49604583e+00 -5.42378835e-02
-2.05339998e-01 6.34863088e-03 -5.61303616e-01 1.34927177e+00
9.14697647e-01 -2.42885947e-01 -5.68773687e-01 -4.55711596e-02
4.36153740e-01 -1.38278946e-01 4.05448973e-01 -9.48605657e-01
1.94844186e-01 -5.71980715e-01 6.46646678e-01 -4.58209574e-01
1.33397639e-01 -1.15960038e+00 2.18306839e-01 8.39589298e-01
-2.01211482e-01 -3.09332132e-01 -1.96686178e-01 2.95049965e-01
-1.39954939e-01 -8.02663743e-01 7.34027147e-01 -2.83742279e-01
-9.73479271e-01 -1.03332743e-01 -2.09161013e-01 -1.22584090e-01
1.03590012e+00 -7.23293722e-01 7.10300207e-02 -7.18439743e-02
-8.05633545e-01 -3.41860682e-01 3.92630100e-01 1.34898890e-02
1.02513742e+00 -1.28034079e+00 -3.91775966e-01 2.49101132e-01
-3.23265940e-02 -9.63058323e-02 -8.36043060e-02 5.56814790e-01
-6.94438457e-01 4.22604561e-01 -5.12899339e-01 -5.56311905e-01
-1.13477707e+00 7.30209410e-01 -5.93417464e-03 1.79555371e-01
-6.01397753e-01 4.73589778e-01 1.41518146e-01 -2.89177805e-01
1.35907933e-01 -1.22269996e-01 -4.52771068e-01 -4.00369428e-02
4.36812341e-01 4.89796579e-01 -1.09562300e-01 -5.34697235e-01
-1.96129531e-01 1.03120792e+00 8.03930610e-02 2.43873686e-01
1.32736051e+00 1.62942663e-01 -6.33694649e-01 6.56974733e-01
1.01646304e+00 -1.44084215e-01 -5.71925461e-01 -2.00750649e-01
3.24350059e-01 -5.35173774e-01 -3.14031065e-01 -5.66757500e-01
-7.10733294e-01 1.07269573e+00 3.31360340e-01 6.28046989e-01
1.10679185e+00 1.18435368e-01 2.39214152e-01 1.17869556e+00
6.13557398e-01 -1.32950962e+00 4.52219874e-01 4.21150982e-01
1.11157727e+00 -1.17188299e+00 3.02811295e-01 -6.86776340e-01
-1.97161049e-01 1.57922006e+00 5.14355183e-01 -1.05494410e-01
1.06329644e+00 4.75398637e-02 -2.16962561e-01 -4.39617664e-01
-4.95658487e-01 -1.75588638e-01 4.86691147e-01 5.54888248e-01
4.74732786e-01 1.31861255e-01 -5.13993382e-01 1.71701252e-01
-3.53775352e-01 1.56861484e-01 2.14312360e-01 1.24933600e+00
-1.09122860e+00 -1.16117036e+00 -5.55107415e-01 4.71323937e-01
-1.61186486e-01 3.01763207e-01 -7.06987321e-01 9.22647595e-01
5.34773409e-01 6.59464359e-01 5.12599945e-02 -3.62020522e-01
1.41403496e-01 -7.41896778e-02 8.63915026e-01 -4.43883479e-01
1.52253538e-01 -3.86364996e-01 -4.83695492e-02 -1.30897939e-01
-7.76693583e-01 -4.31936592e-01 -1.47578371e+00 -3.69056016e-01
-3.97429228e-01 5.26795387e-01 8.54369342e-01 1.24396002e+00
-2.42358834e-01 6.11036376e-04 7.97576547e-01 -3.99978161e-01
-4.89423305e-01 -3.88448983e-01 -6.97370291e-01 4.73787010e-01
-8.65801647e-02 -8.59770298e-01 -3.33902597e-01 3.80700916e-01] | [10.123737335205078, -0.5827431082725525] |
42e2c0a5-b124-4533-971c-97e7e02891a3 | treepiece-faster-semantic-parsing-via-tree | 2303.17161 | null | https://arxiv.org/abs/2303.17161v1 | https://arxiv.org/pdf/2303.17161v1.pdf | TreePiece: Faster Semantic Parsing via Tree Tokenization | Autoregressive (AR) encoder-decoder neural networks have proved successful in many NLP problems, including Semantic Parsing -- a task that translates natural language to machine-readable parse trees. However, the sequential prediction process of AR models can be slow. To accelerate AR for semantic parsing, we introduce a new technique called TreePiece that tokenizes a parse tree into subtrees and generates one subtree per decoding step. On TopV2 benchmark, TreePiece shows 4.6 times faster decoding speed than standard AR, and comparable speed but significantly higher accuracy compared to Non-Autoregressive (NAR). | ['Sasha Livshits', 'Akshat Shrivastava', 'Sid Wang'] | 2023-03-30 | null | null | null | null | ['semantic-parsing'] | ['natural-language-processing'] | [ 4.04092252e-01 6.26402915e-01 -7.98556507e-02 -7.53084481e-01
-1.49735272e+00 -6.17407739e-01 2.47552752e-01 3.89911793e-02
-1.38584360e-01 4.03051257e-01 5.24362206e-01 -9.73975062e-01
7.54230201e-01 -1.01343465e+00 -8.90487790e-01 -1.81611091e-01
2.06406981e-01 8.21586609e-01 -2.92912349e-02 -9.80717242e-02
-2.32524231e-01 1.64832532e-01 -9.21993792e-01 5.29044032e-01
5.24144411e-01 7.75532961e-01 4.07921582e-01 1.27265894e+00
-9.07211661e-01 9.22697425e-01 -5.93577921e-01 -4.72119808e-01
-1.75044369e-02 -4.21105355e-01 -1.05144906e+00 -5.87969542e-01
1.44921631e-01 -2.52199292e-01 -2.98108160e-01 8.15692782e-01
1.07461117e-01 -9.70821828e-03 4.37145054e-01 -5.75739920e-01
-1.00071788e+00 1.12586129e+00 -4.87165183e-01 2.03963265e-01
4.18011934e-01 -4.99367625e-01 1.26905453e+00 -8.91167283e-01
3.11037898e-01 1.89875424e+00 5.40689588e-01 7.64992893e-01
-1.07454252e+00 -6.24965787e-01 1.31217644e-01 -2.44596750e-01
-7.77579784e-01 -4.82762575e-01 5.12619257e-01 -1.09542161e-01
1.59239423e+00 5.05616181e-02 1.15591712e-01 1.26007187e+00
5.47678351e-01 1.00199366e+00 8.12515259e-01 -4.86435801e-01
2.26360694e-01 -4.12021816e-01 7.12068915e-01 7.64225185e-01
-1.52249292e-01 -3.37031156e-01 -2.60920435e-01 -2.39141539e-01
6.90095603e-01 -3.17583054e-01 4.91302341e-01 3.90534014e-01
-7.99137533e-01 1.23125923e+00 2.44306430e-01 -1.55895099e-01
-6.11790538e-01 5.64880252e-01 6.01177752e-01 4.38028544e-01
5.99437356e-01 2.84245849e-01 -7.90938735e-01 -5.97522140e-01
-4.53195482e-01 -4.93014306e-02 8.14193308e-01 9.77165699e-01
5.24260998e-01 6.87415302e-01 -5.57336584e-02 1.11542070e+00
4.04630303e-01 7.32694685e-01 5.99744558e-01 -1.01992679e+00
7.31043696e-01 4.17867184e-01 -3.25772405e-01 -4.61604655e-01
-2.94036686e-01 -5.12806140e-03 -9.62706029e-01 -2.48536095e-01
-3.80931906e-02 -2.30546698e-01 -1.29353857e+00 1.55480254e+00
7.04986975e-02 -8.23584870e-02 5.79393029e-01 5.91523826e-01
1.14347661e+00 1.30968773e+00 6.50470138e-01 -5.24463579e-02
1.53940189e+00 -1.02650988e+00 -7.43851185e-01 -7.08106399e-01
8.36722851e-01 -5.63714802e-01 1.32191813e+00 2.80979574e-01
-1.32108188e+00 -4.78767276e-01 -6.78340554e-01 -7.37897575e-01
-2.05440208e-01 2.18785316e-01 1.05659163e+00 6.34317279e-01
-1.19746721e+00 3.28283489e-01 -1.12325931e+00 -1.52199073e-02
2.70406455e-01 3.03001016e-01 -3.42374951e-01 3.47802672e-03
-1.07393157e+00 6.72272563e-01 2.98452288e-01 1.33389995e-01
-8.96879256e-01 -4.90981817e-01 -1.39553130e+00 2.49172643e-01
1.86561830e-02 -8.92137885e-01 1.80618250e+00 -8.62083614e-01
-1.99748003e+00 6.75564110e-01 -1.01319516e+00 -8.11620235e-01
-1.18814178e-01 -7.43923962e-01 -2.08767042e-01 -1.47784561e-01
1.75712079e-01 8.98167551e-01 7.23453045e-01 -7.55376875e-01
-3.87157351e-01 -4.84971642e-01 -3.66349928e-02 1.39303699e-01
1.50202751e-01 4.37777370e-01 -5.43057203e-01 -4.66254503e-01
3.81582677e-01 -7.34131634e-01 -6.01072848e-01 -8.58563244e-01
-5.13900161e-01 -4.79905427e-01 5.25126219e-01 -1.10501885e+00
9.97845769e-01 -1.95205986e+00 2.63136953e-01 -3.35513443e-01
1.75340608e-01 1.23427972e-01 -4.51694310e-01 7.01074004e-02
-3.31248999e-01 2.88710684e-01 -3.36127073e-01 -7.20113456e-01
-1.72327042e-01 6.67027056e-01 -7.68437147e-01 -1.24841325e-01
5.00972629e-01 1.27516079e+00 -8.05607319e-01 -3.79537672e-01
-4.04253835e-03 5.50423920e-01 -5.61767638e-01 2.30493754e-01
-4.93530035e-01 2.51274526e-01 -6.65784240e-01 6.95749402e-01
3.23239356e-01 -2.89861768e-01 8.85948911e-02 4.19589907e-01
2.86547184e-01 9.58967090e-01 -3.55746359e-01 1.62432766e+00
-1.08068800e+00 8.89700055e-01 -1.63004369e-01 -8.66072714e-01
1.28103709e+00 1.80150837e-01 -2.12967861e-02 -7.36276090e-01
1.45159826e-01 -7.46477917e-02 -3.73869777e-01 1.44267862e-03
8.48250210e-01 -1.69156883e-02 -4.77476627e-01 1.93272308e-01
-8.15920308e-02 -1.47112086e-01 -1.72874227e-01 9.27469581e-02
1.36017299e+00 1.46920800e-01 2.09391519e-01 2.04862475e-01
2.28947416e-01 2.48426467e-01 5.89173377e-01 8.59394312e-01
3.74356329e-01 6.45577729e-01 8.14935327e-01 -7.47396529e-01
-1.20142055e+00 -1.24186683e+00 2.87162185e-01 1.57856560e+00
-3.62458885e-01 -5.45324743e-01 -9.13715303e-01 -6.36238337e-01
-2.98035651e-01 1.26706696e+00 -3.06705356e-01 4.02801707e-02
-8.88375521e-01 -7.24268794e-01 7.94300795e-01 1.03126681e+00
2.63221055e-01 -1.32060921e+00 -3.27133477e-01 4.84258175e-01
-2.05893382e-01 -1.50769067e+00 2.19836086e-02 3.99915397e-01
-1.22152412e+00 -5.80269098e-01 -3.43343943e-01 -9.18212831e-01
4.45839435e-01 1.02669872e-01 1.54150140e+00 -3.14235300e-01
3.82490531e-02 -2.55883448e-02 -4.80866492e-01 -4.91510570e-01
-1.03731704e+00 2.55421162e-01 -2.71307260e-01 -6.93764091e-01
5.55714250e-01 -3.44983280e-01 6.41633943e-02 -2.34716684e-01
-5.63525856e-01 3.51641715e-01 6.36776209e-01 7.67370284e-01
1.00391579e+00 -2.96885252e-01 3.46207231e-01 -1.26533616e+00
7.51611531e-01 -5.19747376e-01 -6.90154970e-01 1.55305043e-01
-3.66489232e-01 2.78688848e-01 1.09377432e+00 -1.32807285e-01
-1.14792323e+00 3.01826596e-01 -7.22979069e-01 -3.41110855e-01
-9.50111672e-02 5.01793087e-01 8.24135635e-03 6.34828150e-01
4.09171790e-01 1.78349778e-01 -3.03915113e-01 -8.19620371e-01
5.83278596e-01 7.09231794e-01 7.91615367e-01 -5.02844036e-01
5.30788481e-01 3.25722992e-02 -1.16715126e-01 -8.28698516e-01
-1.06368697e+00 -2.92630494e-01 -5.07723451e-01 5.66902518e-01
1.18518865e+00 -1.10876513e+00 -4.41951901e-01 7.46202692e-02
-1.76667380e+00 -5.52994907e-01 -1.83429778e-01 2.80484140e-01
-7.30365932e-01 2.31476560e-01 -1.08955967e+00 -8.46450329e-01
-1.08668697e+00 -1.04236066e+00 1.61033630e+00 1.66332096e-01
-3.61406833e-01 -9.02163744e-01 5.43927923e-02 4.36392993e-01
2.23597571e-01 -1.89739034e-01 1.36719906e+00 -8.49210501e-01
-2.40835845e-01 -2.30088159e-01 -3.82070243e-01 2.51130998e-01
-2.20169127e-01 -7.20174462e-02 -1.01123619e+00 1.90798461e-01
-2.61285126e-01 -1.54510727e-02 9.69437122e-01 6.13772154e-01
1.39160657e+00 -3.68552029e-01 -1.59274057e-01 6.73328161e-01
1.16115630e+00 2.59799212e-01 7.34671175e-01 2.13319972e-01
7.52494276e-01 1.45024508e-01 5.09865940e-01 1.33047337e-02
5.82701504e-01 4.15685773e-01 1.94312394e-01 -2.23662943e-01
-7.48977512e-02 -6.53050542e-01 9.44357574e-01 1.25269258e+00
1.95465162e-01 -4.29823816e-01 -1.09743679e+00 3.73382568e-01
-1.61567140e+00 -3.17443937e-01 -6.01306856e-01 1.93363953e+00
4.85747397e-01 2.73466080e-01 -3.01208109e-01 -3.15860242e-01
4.84606594e-01 2.26897016e-01 -3.89872283e-01 -1.49462581e+00
-7.49288723e-02 6.63941860e-01 6.69108272e-01 7.74094880e-01
-1.12149346e+00 1.90235174e+00 7.68363857e+00 5.55564642e-01
-6.82810485e-01 3.52922350e-01 8.76373589e-01 3.27042490e-01
-4.31245387e-01 1.26485959e-01 -1.12186623e+00 8.42493922e-02
1.78432727e+00 3.48420851e-02 3.62095356e-01 1.41969466e+00
2.50200570e-01 1.48329973e-01 -9.61824000e-01 8.23280036e-01
-2.19428375e-01 -1.29872894e+00 4.82279420e-01 -2.24204585e-01
2.19708174e-01 5.05076170e-01 -1.64639458e-01 8.77712011e-01
9.97369349e-01 -1.30216634e+00 2.08840162e-01 7.48248398e-02
9.21885431e-01 -7.10091114e-01 7.53112197e-01 2.14043379e-01
-1.32661390e+00 2.80614980e-02 -9.59403753e-01 -1.46597534e-01
4.14849848e-01 5.31988680e-01 -1.08673489e+00 2.18043357e-01
7.08603144e-01 4.28179055e-01 -3.25552344e-01 1.17638648e-01
-5.89979708e-01 1.15870488e+00 -1.64185792e-01 4.24900949e-02
4.68868345e-01 -2.35562176e-01 3.92808795e-01 1.40940273e+00
1.85557067e-01 2.84207404e-01 -1.03375111e-02 7.79652536e-01
-1.80859372e-01 9.50252786e-02 -7.09139705e-01 -4.25932348e-01
4.78050649e-01 1.03131354e+00 -4.92933661e-01 -6.81492150e-01
-3.75032783e-01 1.19087470e+00 4.12361890e-01 2.27646276e-01
-8.74222159e-01 -1.90893203e-01 7.59715497e-01 -4.04437274e-01
7.72041678e-02 -6.07227147e-01 -5.56560397e-01 -8.65989923e-01
-6.41528890e-02 -7.38047719e-01 5.18025935e-01 -1.09642577e+00
-7.58944273e-01 9.28384364e-01 -2.77224809e-01 -4.78695422e-01
-8.67789626e-01 -6.10544860e-01 -5.54008126e-01 9.66912508e-01
-1.55290842e+00 -1.11068928e+00 -8.53091627e-02 7.13026002e-02
1.42895353e+00 -1.71538308e-01 1.38690901e+00 -1.36855036e-01
-7.29217887e-01 5.26151121e-01 -1.28314726e-03 3.11580718e-01
3.97704951e-02 -1.50514424e+00 1.68329632e+00 8.27744901e-01
3.56705904e-01 4.80251729e-01 5.92167020e-01 -5.76323628e-01
-1.85313785e+00 -1.26032090e+00 1.32611513e+00 -4.87643033e-01
6.79969072e-01 -4.24772859e-01 -1.24398482e+00 1.10480940e+00
-9.17742401e-02 -3.30800474e-01 4.30318713e-01 4.19073701e-01
-6.08110130e-01 3.06313485e-01 -6.28951430e-01 5.65217018e-01
8.39258730e-01 -6.18298888e-01 -7.40948558e-01 3.95340085e-01
1.42989028e+00 -6.70274436e-01 -8.72352779e-01 2.09722519e-01
4.29677665e-01 -3.91918033e-01 1.00427318e+00 -1.08295727e+00
9.14704561e-01 2.44562373e-01 -2.74223149e-01 -1.21073878e+00
-3.04020166e-01 -7.84540534e-01 -1.94751367e-01 1.03018141e+00
7.36151159e-01 -5.88484943e-01 1.01096582e+00 6.24702513e-01
-3.53607267e-01 -5.14055192e-01 -9.07704353e-01 -5.99063575e-01
3.46978456e-01 -7.55252779e-01 4.83009517e-01 4.12331820e-01
-4.46944445e-01 8.13093066e-01 -3.17184746e-01 3.67070735e-01
1.76265761e-01 9.20837149e-02 8.06072712e-01 -1.19507241e+00
-2.63099611e-01 -2.59259403e-01 -1.66176736e-01 -1.49791825e+00
6.93623900e-01 -8.66176128e-01 2.94399261e-01 -1.72868860e+00
8.69328231e-02 -2.88532346e-01 -4.34708521e-02 9.21151876e-01
-2.91513622e-01 -1.10853374e-01 -5.10118864e-02 9.03316066e-02
-4.52421278e-01 4.90921378e-01 8.13215196e-01 2.93529825e-03
-3.32191706e-01 -7.98379704e-02 -6.57064378e-01 8.43224347e-01
9.01779771e-01 -7.81267643e-01 -2.25325167e-01 -9.08958375e-01
3.59544337e-01 6.34238064e-01 -1.98516503e-01 -4.87497926e-01
-9.79274362e-02 1.62519306e-01 1.51404709e-01 -6.92199111e-01
5.55523336e-01 -2.11091459e-01 -1.08454563e-01 2.21021995e-01
-4.80426162e-01 6.84688509e-01 2.66034991e-01 6.02539837e-01
-2.89962858e-01 -2.21913666e-01 4.90155995e-01 -2.30306834e-01
-8.38623226e-01 4.84169759e-02 -6.46776915e-01 9.25855115e-02
3.93242180e-01 -1.28985196e-02 -2.82604605e-01 -5.05537271e-01
-7.22830057e-01 6.72819763e-02 4.65601794e-02 8.17133665e-01
7.49217570e-01 -8.31763744e-01 -9.29838479e-01 2.35922128e-01
-3.03776383e-01 4.87140268e-01 3.29545699e-02 3.06516290e-01
-6.94389641e-01 6.26779020e-01 2.79706538e-01 -5.31228364e-01
-1.41207480e+00 3.50984693e-01 8.48771706e-02 -4.16571319e-01
-1.11903501e+00 8.80536079e-01 3.86690378e-01 -4.88268673e-01
-1.19175855e-02 -4.25019115e-01 -2.96130747e-01 -4.79246348e-01
5.21830559e-01 9.34180990e-02 3.22731435e-02 -4.66398507e-01
-1.04595609e-01 4.53902900e-01 -3.22517961e-01 6.96675852e-02
1.38827467e+00 8.21527541e-02 -1.93607792e-01 6.23198271e-01
1.08823025e+00 -1.70368508e-01 -8.24325323e-01 -1.02145679e-01
2.46110633e-01 9.05824825e-02 1.37453541e-01 -5.47034085e-01
-7.95100927e-01 1.12035179e+00 1.88454255e-01 1.70593992e-01
8.61393213e-01 2.64649987e-01 1.30987298e+00 7.61794031e-01
2.21202970e-01 -7.97023833e-01 -4.33683991e-01 1.11914968e+00
6.64253950e-01 -1.14525747e+00 -4.85707760e-01 -6.60891414e-01
-7.99041271e-01 1.25911832e+00 3.08735639e-01 -1.70949653e-01
1.66171342e-01 4.97525305e-01 1.90260902e-01 -6.42701611e-02
-1.21180212e+00 5.97999766e-02 -9.00404304e-02 5.00313461e-01
6.94852889e-01 4.90445048e-01 6.56808838e-02 8.85789633e-01
-7.59797394e-01 -3.00718904e-01 5.59590816e-01 5.68584859e-01
-4.87343520e-01 -1.09495080e+00 -2.05511525e-01 3.94399166e-01
-7.46280968e-01 -5.12064576e-01 -4.89070773e-01 4.64609534e-01
-9.09893870e-01 1.06854701e+00 3.99702549e-01 -2.92128712e-01
6.79844394e-02 2.72766411e-01 1.98510196e-02 -9.20892894e-01
-5.39580345e-01 -3.17115970e-02 5.23902655e-01 -8.35717499e-01
3.13931435e-01 -3.09456229e-01 -1.77034497e+00 -1.01048067e-01
4.76054586e-02 2.51866281e-01 1.16679788e+00 8.40326011e-01
5.93874931e-01 7.43029118e-01 4.60690677e-01 -2.84545749e-01
-5.08300364e-01 -8.95639181e-01 -3.09228078e-02 -1.08278632e-01
6.10911176e-02 -2.01374963e-01 -4.97442968e-02 9.33735520e-02] | [10.4653959274292, 9.18059253692627] |
0e2153ac-90b0-4017-830b-915a8c89da1f | honestbait-forward-references-for-attractive | 2306.14828 | null | https://arxiv.org/abs/2306.14828v1 | https://arxiv.org/pdf/2306.14828v1.pdf | HonestBait: Forward References for Attractive but Faithful Headline Generation | Current methods for generating attractive headlines often learn directly from data, which bases attractiveness on the number of user clicks and views. Although clicks or views do reflect user interest, they can fail to reveal how much interest is raised by the writing style and how much is due to the event or topic itself. Also, such approaches can lead to harmful inventions by over-exaggerating the content, aggravating the spread of false information. In this work, we propose HonestBait, a novel framework for solving these issues from another aspect: generating headlines using forward references (FRs), a writing technique often used for clickbait. A self-verification process is included during training to avoid spurious inventions. We begin with a preliminary user study to understand how FRs affect user interest, after which we present PANCO1, an innovative dataset containing pairs of fake news with verified news for attractive but faithful news headline generation. Automatic metrics and human evaluations show that our framework yields more attractive results (+11.25% compared to human-written verified news headlines) while maintaining high veracity, which helps promote real information to fight against fake news. | ['Lun-Wei Ku', 'Dennis Wu', 'Chih-Yao Chen'] | 2023-06-26 | null | null | null | null | ['headline-generation'] | ['natural-language-processing'] | [ 9.15876999e-02 3.75650734e-01 -3.79130244e-01 -3.06946874e-01
-8.03091824e-01 -6.39548421e-01 9.87910211e-01 3.69075574e-02
-1.84983119e-01 1.05109656e+00 4.65798706e-01 -1.47852644e-01
3.38111192e-01 -8.71412575e-01 -9.00692701e-01 5.43219130e-03
3.27369958e-01 1.64110824e-01 3.22625041e-01 -6.02698028e-01
7.19965875e-01 -9.52462405e-02 -1.43894851e+00 7.21828341e-01
1.03935516e+00 9.48330879e-01 -3.08894157e-01 6.60456002e-01
-6.38445914e-02 1.07684684e+00 -1.33143365e+00 -9.74750757e-01
2.03963026e-01 -6.16931677e-01 -4.38395500e-01 -2.64097244e-01
8.97808433e-01 -5.77773452e-01 -2.07163557e-01 9.81189013e-01
3.97167593e-01 -2.77857929e-01 5.17810047e-01 -1.35356450e+00
-1.17167461e+00 9.24633265e-01 -5.36796689e-01 5.58306128e-02
6.31675601e-01 1.32168278e-01 1.32889605e+00 -7.28908241e-01
8.02590668e-01 1.25333989e+00 5.58576882e-01 6.32307053e-01
-1.17995381e+00 -8.22046936e-01 -2.34278087e-02 -2.68803030e-01
-9.44364488e-01 -4.68949407e-01 6.63285673e-01 -5.30097485e-01
1.98003232e-01 9.18327689e-01 6.26237810e-01 1.83182704e+00
4.48307581e-03 1.12385976e+00 1.28077066e+00 -1.23083353e-01
1.34085163e-01 8.47029984e-01 1.72889620e-01 3.39888304e-01
6.12071633e-01 2.81188637e-01 -9.04347599e-01 -3.17490399e-01
6.08293176e-01 -2.16044515e-01 -4.39936697e-01 2.80024439e-01
-1.07322001e+00 1.15077031e+00 6.28093839e-01 8.90605599e-02
-3.11602086e-01 -4.39874791e-02 8.02356824e-02 4.43809420e-01
6.11471295e-01 1.35133851e+00 -2.14407608e-01 -3.00385594e-01
-1.09131837e+00 7.17673600e-01 1.04885781e+00 9.57923949e-01
3.44598204e-01 -9.93400440e-02 -8.31447303e-01 7.13931382e-01
-5.40887490e-02 6.86955750e-01 7.07807004e-01 -6.66676700e-01
4.03059393e-01 4.10301387e-01 6.02379084e-01 -1.61657012e+00
2.61187345e-01 -8.10878396e-01 -3.29495251e-01 4.87409979e-02
4.77287322e-01 -1.87409222e-01 -4.71036285e-01 1.55672705e+00
-2.82146633e-01 -3.05529743e-01 -3.84090364e-01 8.08124661e-01
7.27728426e-01 6.01583898e-01 -2.84698457e-01 -2.24094048e-01
1.25721872e+00 -1.07740581e+00 -1.05872548e+00 -1.48309201e-01
6.28908336e-01 -1.15652597e+00 1.55433619e+00 5.58077037e-01
-1.18473125e+00 -2.59580821e-01 -9.42716599e-01 -3.63124115e-03
-6.16388857e-01 2.83640921e-01 4.64875340e-01 8.59476686e-01
-5.38078725e-01 6.00407779e-01 9.70940143e-02 1.01702765e-01
5.89078903e-01 -2.23827198e-01 8.11744034e-02 4.00632530e-01
-1.63195062e+00 8.18458140e-01 -1.38825253e-01 -5.91030180e-01
-3.74274522e-01 -8.94894898e-01 -2.82826275e-01 -5.54339625e-02
3.19591969e-01 -4.75490093e-01 1.48736322e+00 -1.02142680e+00
-1.32435727e+00 4.53321636e-01 2.64132628e-03 -6.94195449e-01
1.22884524e+00 -8.22751701e-01 -4.27089304e-01 -1.24036074e-01
4.57879335e-01 6.28587067e-01 1.34020126e+00 -1.29677129e+00
-5.97033083e-01 -7.31705502e-02 5.48613109e-02 -7.75984377e-02
-5.62496960e-01 -3.24146837e-01 -9.36755985e-02 -1.24481511e+00
-2.86765754e-01 -8.51650357e-01 1.75397366e-01 -6.59761131e-02
-9.99706864e-01 -5.01787737e-02 9.39927280e-01 -8.55059266e-01
1.78601646e+00 -1.79733002e+00 -5.55450678e-01 1.75820127e-01
5.33439279e-01 3.09260994e-01 6.09010793e-02 4.59488273e-01
3.34656566e-01 6.46448135e-01 4.02699023e-01 -2.82962918e-02
1.37868049e-02 -4.31119025e-01 -8.85697365e-01 -1.05167113e-01
8.30435306e-02 1.12657118e+00 -9.59568679e-01 -1.99975461e-01
-4.60243076e-01 2.27733210e-01 -5.79557657e-01 5.78022189e-02
-3.67872208e-01 9.20945480e-02 -3.79654318e-01 4.78648305e-01
6.05163455e-01 -4.78841394e-01 -6.84532523e-01 -1.90809369e-02
-1.26055792e-01 5.71753800e-01 -6.52987599e-01 7.28041351e-01
-4.48118240e-01 8.72671187e-01 -5.26050925e-01 1.51059413e-02
8.74553025e-01 6.71130791e-02 -6.62337244e-02 -8.61853957e-01
6.13034563e-03 3.27352911e-01 -4.65954453e-01 -3.27618212e-01
1.00743866e+00 2.12493956e-01 9.61694494e-03 7.57308960e-01
-5.25953829e-01 -1.64970264e-01 1.78555265e-01 7.84607053e-01
9.84625399e-01 -3.42824429e-01 5.61035164e-02 2.38339007e-01
1.31502628e-01 -1.83165073e-03 -5.81103228e-02 1.30257285e+00
2.59894449e-02 6.40936852e-01 7.76085258e-01 -1.95529073e-01
-1.11810040e+00 -8.79204631e-01 1.30504206e-01 1.09266353e+00
9.49967206e-02 -6.71858132e-01 -8.31277192e-01 -1.01992679e+00
3.64281889e-03 1.40866923e+00 -7.41127551e-01 -3.54686946e-01
-1.32986367e-01 -4.00929064e-01 6.82159245e-01 5.97668141e-02
4.19043988e-01 -7.71993399e-01 -2.29417428e-01 2.21171722e-01
-6.27375305e-01 -8.00273001e-01 -8.72156918e-01 -6.02493405e-01
-4.72589195e-01 -8.16078782e-01 -1.13919270e+00 -9.37022194e-02
7.04038799e-01 5.16159117e-01 1.08383286e+00 1.21510588e-01
-5.20746000e-02 -3.62574719e-02 -4.32408392e-01 -7.05550194e-01
-6.36179209e-01 3.83024812e-02 -1.48934186e-01 8.85436758e-02
2.69455969e-01 -7.07015321e-02 -5.89236975e-01 6.32482350e-01
-9.23660278e-01 1.40010327e-01 6.53089404e-01 1.08292592e+00
-1.46973163e-01 -2.00813040e-01 7.34597027e-01 -1.46436524e+00
1.37067997e+00 -4.38863635e-01 -5.45871973e-01 1.93317190e-01
-9.14228320e-01 7.90606812e-02 5.05812764e-01 -7.26469934e-01
-1.07336175e+00 -3.23101223e-01 1.42463207e-01 7.05354884e-02
2.61569291e-01 2.15907276e-01 4.64173824e-01 -1.28147565e-02
1.33523357e+00 1.96875006e-01 4.06639501e-02 -3.56094509e-01
5.87940931e-01 8.60412359e-01 1.95818871e-01 -1.07666813e-01
1.03840339e+00 4.02383864e-01 -6.53682768e-01 -7.85646498e-01
-1.25588679e+00 -4.04098183e-01 5.64099140e-02 -3.08586061e-01
2.41797388e-01 -6.73290312e-01 -6.45960748e-01 2.66744792e-01
-1.34522402e+00 2.96510637e-01 -2.96044737e-01 1.98664948e-01
-3.65693718e-01 1.65847510e-01 -5.08211076e-01 -8.39967191e-01
-2.69459397e-01 -8.73270810e-01 9.29675996e-01 -3.66704655e-04
-8.03039312e-01 -7.23688245e-01 1.49304597e-02 7.45859742e-01
8.89147878e-01 1.83674037e-01 5.12995541e-01 -9.52962637e-01
-6.82812452e-01 -9.42906976e-01 -3.95321012e-01 2.96995729e-01
2.96851117e-02 1.18098609e-01 -1.04289544e+00 6.59618750e-02
-1.51432320e-01 -4.23286885e-01 7.71506846e-01 1.01046532e-01
8.97942603e-01 -1.37477791e+00 -2.47179523e-01 -9.43542272e-02
9.30456817e-01 -3.22339237e-01 7.67605722e-01 3.92765135e-01
4.53711629e-01 7.10748792e-01 7.03409076e-01 8.32505941e-01
7.59351104e-02 8.45700502e-01 3.43177021e-01 -7.11183771e-02
-1.15152344e-01 -1.02574277e+00 4.20115560e-01 3.24138135e-01
6.47977293e-02 -4.94679242e-01 -1.07149094e-01 1.19760878e-01
-1.51920629e+00 -1.37693155e+00 -4.14261311e-01 2.44499087e+00
1.19398701e+00 4.72110808e-01 3.69585812e-01 1.48631573e-01
7.78412998e-01 1.96694434e-01 -1.61638603e-01 -1.34048596e-01
-3.12062889e-01 -1.83106408e-01 7.56176353e-01 6.10762417e-01
-6.47632360e-01 9.37433600e-01 6.32603359e+00 9.37536895e-01
-1.26814115e+00 -6.68839514e-02 8.02163720e-01 -2.12616846e-01
-7.12591290e-01 -2.75446236e-01 -9.25146639e-01 9.52817678e-01
5.47303975e-01 -4.19569433e-01 3.10098350e-01 1.11034119e+00
6.05386674e-01 5.82487471e-02 -8.64692092e-01 7.67460883e-01
1.85735792e-01 -1.61869252e+00 3.41999143e-01 7.78452232e-02
9.43487048e-01 -4.34663713e-01 6.50498807e-01 2.91150004e-01
4.41181540e-01 -8.00939918e-01 9.33584571e-01 3.92232299e-01
4.98217314e-01 -3.01263154e-01 6.15688205e-01 4.11612630e-01
-1.34227782e-01 1.54671803e-01 -4.06185612e-02 -3.55255716e-02
3.26500326e-01 9.52447474e-01 -1.26847601e+00 -2.09944591e-01
3.45536679e-01 4.48391020e-01 -9.48008239e-01 8.75829875e-01
-4.00257081e-01 6.75102234e-01 -1.13832645e-01 -7.49521136e-01
1.88648179e-01 4.12569582e-01 8.23529601e-01 1.16042864e+00
3.78556997e-01 -3.46571058e-01 -4.29825783e-01 1.25571287e+00
-3.79208446e-01 2.09031045e-01 -7.75241137e-01 -5.32276452e-01
4.81253535e-01 1.16689372e+00 -2.15904459e-01 -5.55506289e-01
-1.41581073e-01 1.18090272e+00 4.55601104e-02 2.90159166e-01
-1.04619873e+00 -5.96857250e-01 9.41760689e-02 7.91520059e-01
-2.31659822e-02 2.72785962e-01 -5.75293243e-01 -1.22667587e+00
8.99376720e-02 -1.20298183e+00 -4.46575545e-02 -8.79375219e-01
-1.46069574e+00 5.65465569e-01 -4.68215853e-01 -1.40773773e+00
-8.60542431e-02 -1.95936546e-01 -6.84008539e-01 6.49284780e-01
-1.32240140e+00 -8.73105884e-01 -3.58576149e-01 1.06527008e-01
6.42489195e-01 -1.76851973e-01 2.91240633e-01 2.21730888e-01
-1.18617967e-01 1.00415528e+00 -1.07483536e-01 -1.61062688e-01
1.20798159e+00 -1.28900361e+00 5.68428099e-01 5.12693584e-01
2.87719280e-01 9.54427481e-01 1.20402038e+00 -9.12427902e-01
-7.76718736e-01 -8.85341406e-01 1.35775018e+00 -7.98037350e-01
8.39149654e-01 -2.19283104e-01 -7.68594980e-01 2.37922892e-01
1.29644528e-01 -4.76739258e-01 7.17634141e-01 6.90040588e-02
-6.40391052e-01 2.58857995e-01 -1.15480304e+00 9.82592702e-01
7.19466507e-01 -2.09976926e-01 -4.61993545e-01 6.93528116e-01
8.83729279e-01 -1.36790916e-01 -2.80091703e-01 -7.81373903e-02
6.91965878e-01 -1.27268970e+00 6.85619235e-01 -7.69146860e-01
1.01284671e+00 1.59321278e-02 3.07306230e-01 -1.46786225e+00
-2.93650955e-01 -1.08837688e+00 -1.95000976e-01 1.21104991e+00
1.02662027e+00 -5.11652827e-01 8.80862951e-01 6.80397391e-01
2.49936998e-01 -4.74931806e-01 -2.48108774e-01 -8.34174991e-01
-4.56242591e-01 -1.57523185e-01 3.56261283e-01 1.06254399e+00
1.63811207e-01 7.12378919e-01 -9.84017491e-01 -2.83247620e-01
4.98422503e-01 -3.40077072e-01 1.04891479e+00 -1.01930547e+00
-3.11900973e-01 -5.49416959e-01 5.10519966e-02 -1.32281697e+00
-5.38990378e-01 -3.39599103e-01 -1.36312142e-01 -8.79264474e-01
-5.05392179e-02 -3.06702882e-01 2.01013595e-01 1.74069345e-01
-2.38528714e-01 3.49913090e-01 9.75915119e-02 4.91658896e-01
-5.62589347e-01 3.89366269e-01 1.64873230e+00 -9.37054083e-02
-3.48363757e-01 6.22254312e-01 -1.11770940e+00 5.53265154e-01
6.27412915e-01 -3.91732305e-01 -3.83198917e-01 -5.95499873e-02
6.66470647e-01 -2.08807841e-01 4.62981611e-01 -8.09794545e-01
3.93490866e-02 -8.23009461e-02 1.89501777e-01 -5.35630822e-01
3.45694095e-01 -4.50791687e-01 -2.69635797e-01 4.75933701e-01
-1.14589608e+00 2.47849990e-02 -3.72533858e-01 9.00044143e-01
-2.25154951e-01 -3.17637444e-01 6.45766675e-01 -3.37347925e-01
1.52674911e-03 -1.87377613e-02 -3.76026988e-01 3.16019356e-01
8.34924400e-01 -1.21684484e-01 -8.31874847e-01 -1.17472398e+00
-4.03655082e-01 -1.75383955e-01 3.61150622e-01 7.34206975e-01
6.19988143e-01 -1.35082555e+00 -7.41956413e-01 1.31379873e-01
4.32279378e-01 -8.69570851e-01 -8.77302662e-02 6.36098146e-01
-4.59947288e-01 4.50269878e-01 2.04420127e-02 -1.11690305e-01
-1.17508888e+00 3.95130247e-01 -2.44132698e-01 -1.13538504e-01
-1.41240597e-01 1.06245399e+00 8.18282366e-02 1.41430587e-01
3.00660908e-01 -3.18071339e-03 -1.70859754e-01 2.94266015e-01
8.25769007e-01 6.78289533e-01 -8.05425793e-02 -4.55367714e-01
2.79345930e-01 -1.72551796e-01 -6.64016485e-01 -4.22680467e-01
8.10285985e-01 -5.49787767e-02 2.19588533e-01 1.68138117e-01
1.11197293e+00 7.23585904e-01 -9.32463229e-01 -1.50849238e-01
-6.36541750e-03 -9.81515825e-01 1.89094227e-02 -1.28106999e+00
-7.50126779e-01 6.23021841e-01 3.25035453e-02 8.65051270e-01
3.97323191e-01 -2.25030899e-01 1.25548565e+00 2.76798248e-01
1.17119543e-01 -1.33305216e+00 7.46310949e-01 1.89351261e-01
1.29034674e+00 -1.47960055e+00 -4.73983698e-02 -5.19546688e-01
-1.13425803e+00 9.60614204e-01 6.18859410e-01 2.58302510e-01
2.47929767e-01 -1.49546370e-01 1.25310332e-01 -6.07431978e-02
-7.03897595e-01 3.57837737e-01 4.69711512e-01 4.32567596e-01
6.08642340e-01 -8.22134763e-02 -6.70697153e-01 6.81177855e-01
-7.12610900e-01 8.31748039e-05 7.93945491e-01 6.19554400e-01
-6.66551769e-01 -7.20811069e-01 -2.65558869e-01 7.32476413e-01
-7.47105658e-01 -2.17537403e-01 -9.87978995e-01 4.66491282e-01
-2.28391916e-01 8.23423207e-01 -3.52076352e-01 -6.07053518e-01
2.82819927e-01 -3.74303043e-01 -1.19860761e-01 -4.99351919e-01
-6.18941486e-01 -1.04951605e-01 5.06056011e-01 -4.82323468e-01
1.64099500e-01 -2.42967412e-01 -4.78732437e-01 -5.69536865e-01
-8.93290162e-01 3.50891650e-01 7.38362432e-01 4.42163378e-01
4.29041803e-01 1.73893318e-01 9.03504014e-01 -3.25059056e-01
-1.03786874e+00 -9.12505746e-01 -3.03769737e-01 8.34105074e-01
3.37396502e-01 -4.37114269e-01 -8.55463743e-01 3.11017931e-02] | [12.033865928649902, 9.04726791381836] |
6abb2762-5f6a-41a1-9c0e-d292a6286113 | large-language-models-are-effective-table-to | 2305.14987 | null | https://arxiv.org/abs/2305.14987v1 | https://arxiv.org/pdf/2305.14987v1.pdf | Large Language Models are Effective Table-to-Text Generators, Evaluators, and Feedback Providers | Large language models (LLMs) have shown remarkable ability on controllable text generation. However, the potential of LLMs in generating text from structured tables remains largely under-explored. In this paper, we study the capabilities of LLMs for table-to-text generation tasks, particularly aiming to investigate their performance in generating natural language statements that can be logically entailed by a provided table. First, we investigate how LLMs compare to state-of-the-art table-to-text fine-tuned models, and demonstrate that LLMs can generate statements with higher faithfulness compared with previous state-of-the-art fine-tuned models. Given this finding, we next explore whether LLMs can serve as faithfulness-level automated evaluation metrics. Through human evaluation, we show that evaluation metrics adopted from LLMs correlates better with human judgments compared with existing faithfulness-level metrics. Finally, we demonstrate that LLMs using chain-of-thought prompting can generate high-fidelity natural language feedback for other table-to-text models' generations, provide insights for future work regarding the distillation of text generation capabilities from LLMs to smaller models. | ['Arman Cohan', 'Xiangru Tang', 'Linyong Nan', 'Shengyun Si', 'Haowei Zhang', 'Yilun Zhao'] | 2023-05-24 | null | null | null | null | ['table-to-text-generation'] | ['natural-language-processing'] | [ 2.71523654e-01 1.05225825e+00 -8.62456337e-02 -3.15803796e-01
-1.03294063e+00 -5.30322254e-01 1.09230614e+00 5.46509206e-01
1.93822280e-01 1.20624328e+00 7.83412695e-01 -4.06600922e-01
2.28291214e-01 -1.19968545e+00 -6.96095049e-01 2.04981774e-01
2.17454195e-01 7.78934777e-01 -6.58779219e-02 -6.47100687e-01
4.74030912e-01 1.21250533e-01 -1.39076328e+00 9.73246455e-01
1.21402860e+00 6.18654370e-01 -8.59306157e-02 6.74497962e-01
-3.30907285e-01 1.38462186e+00 -8.67677212e-01 -7.46679366e-01
-9.76515934e-02 -6.38983428e-01 -1.19236767e+00 -2.48130366e-01
4.85556722e-01 -3.71660084e-01 1.78598911e-01 7.80166745e-01
2.94588655e-01 7.35094324e-02 8.00076723e-01 -1.17066288e+00
-1.07426596e+00 1.33120191e+00 1.44950122e-01 -1.08316056e-01
1.05264866e+00 3.33929062e-01 1.17069376e+00 -9.14390206e-01
8.46214235e-01 1.79168475e+00 5.36498606e-01 8.42002988e-01
-1.41937530e+00 -4.74838316e-01 -3.26612704e-02 -1.69940725e-01
-1.09325647e+00 -5.78712761e-01 5.06537139e-01 -4.19517279e-01
1.34033644e+00 4.27180797e-01 5.24385810e-01 1.26222706e+00
4.96284753e-01 7.22718537e-01 1.23496008e+00 -7.06521749e-01
2.52307206e-01 5.99386275e-01 -1.80194706e-01 8.18215251e-01
5.93880653e-01 1.41902119e-02 -9.86083210e-01 -2.14263308e-03
5.73261201e-01 -7.13901222e-01 1.04245925e-02 2.00196519e-01
-1.49361682e+00 9.40334439e-01 2.61646390e-01 2.07835868e-01
-3.55869055e-01 7.30668604e-02 2.70199627e-01 1.79321691e-01
4.87751871e-01 1.24905920e+00 -1.90756813e-01 -4.07540619e-01
-1.15075946e+00 7.64558554e-01 1.17325211e+00 1.26865888e+00
5.41765928e-01 2.24376470e-01 -1.04642224e+00 4.02927816e-01
2.88336456e-01 5.46962321e-01 6.11804187e-01 -7.92750955e-01
7.44853854e-01 8.02441537e-01 3.23305398e-01 -1.03822601e+00
-1.84712112e-01 -2.20549092e-01 -8.65038216e-01 5.17799146e-02
6.81305081e-02 -2.28458270e-01 -4.52523142e-01 1.52297354e+00
-1.06811605e-01 -7.95350850e-01 4.25720423e-01 4.75021034e-01
8.59425664e-01 7.74677932e-01 -1.17997035e-01 -3.54599446e-01
1.25041544e+00 -8.72358084e-01 -8.43333125e-01 -2.53953815e-01
7.27659643e-01 -6.20228887e-01 1.61321557e+00 4.32928801e-01
-1.49740493e+00 -7.47976422e-01 -9.65265930e-01 -8.43667835e-02
-3.92646700e-01 1.39294863e-01 6.75545633e-01 7.48484313e-01
-1.35733700e+00 6.58326030e-01 -5.59254408e-01 -3.91879141e-01
1.59581944e-01 -9.42514166e-02 -9.05538350e-02 2.87640780e-01
-1.56343257e+00 1.20379627e+00 6.18772984e-01 -1.86943561e-01
-9.31245029e-01 -6.94080770e-01 -9.85387146e-01 5.83965071e-02
3.63872677e-01 -1.04787278e+00 1.69999814e+00 -1.93815872e-01
-1.58628428e+00 6.27421200e-01 -3.33040208e-01 -7.63129592e-01
5.87361038e-01 -2.47756824e-01 -3.58602107e-01 5.47730690e-03
4.83126760e-01 1.12047136e+00 3.64818454e-01 -1.40321887e+00
-2.45013893e-01 2.51471221e-01 2.52415568e-01 7.42070153e-02
-3.86811852e-01 -1.96021825e-01 2.15746418e-01 -8.19585383e-01
-3.07925135e-01 -6.20270550e-01 -1.72994241e-01 -2.16354668e-01
-9.40401375e-01 -4.03761268e-01 7.24676475e-02 -3.12800437e-01
1.65711749e+00 -1.27875137e+00 -3.11718613e-01 -1.04839116e-01
1.65697727e-02 1.25205085e-01 -2.09380850e-01 1.17630458e+00
1.56407788e-01 7.20725477e-01 1.80388503e-02 -2.22257316e-01
5.15103102e-01 -1.37110367e-01 -7.23903775e-01 -6.16474807e-01
6.62660241e-01 1.26808071e+00 -1.02797818e+00 -8.76812696e-01
1.08535640e-01 -1.45212440e-02 -7.29299307e-01 2.92664796e-01
-7.43044019e-01 -6.37794510e-02 -2.84474462e-01 3.84847760e-01
1.24315239e-01 -4.63151604e-01 -9.15860459e-02 -7.67262951e-02
6.16889074e-02 8.28630447e-01 -1.02706528e+00 1.50489819e+00
-5.48840284e-01 5.46358287e-01 -7.23326683e-01 -1.27241060e-01
1.16745365e+00 2.41062179e-01 -3.87563705e-01 -6.28402293e-01
-2.65385181e-01 8.39744434e-02 -1.29827008e-01 -3.26938599e-01
1.12220252e+00 -5.26629388e-01 -2.44364962e-01 7.72941709e-01
-1.58921871e-02 -9.91791368e-01 8.66012633e-01 6.90243363e-01
7.33147442e-01 2.09647208e-01 5.06652653e-01 -3.69758755e-01
5.17357826e-01 3.39101315e-01 -1.79674029e-01 1.08008492e+00
2.29591250e-01 5.32683313e-01 6.07944310e-01 -1.66437924e-01
-8.97220671e-01 -9.42677259e-01 3.98225151e-02 8.62942815e-01
-3.41817468e-01 -1.03144825e+00 -1.07550669e+00 -5.10159016e-01
-9.20821205e-02 1.66819131e+00 -7.96174765e-01 -2.72102952e-01
-8.90180841e-02 -4.64750141e-01 8.83564413e-01 5.67589164e-01
4.34869826e-01 -1.42137206e+00 -5.60074270e-01 3.69629949e-01
-5.27948081e-01 -1.06303644e+00 -3.49600196e-01 -1.25878096e-01
-1.02492487e+00 -6.20169282e-01 -2.96109587e-01 -4.41451043e-01
5.64869583e-01 -2.41739079e-01 1.76833117e+00 -1.65600061e-01
1.77380532e-01 1.04898781e-01 -4.94986653e-01 -7.45343566e-01
-1.37497914e+00 1.61295369e-01 6.24652281e-02 -5.32337904e-01
5.34903035e-02 -2.03588992e-01 -2.83438057e-01 8.86696354e-02
-1.17451811e+00 7.52149880e-01 6.24782801e-01 6.48640811e-01
2.01991826e-01 -5.73940650e-02 8.19543779e-01 -1.12026358e+00
1.58500874e+00 4.28539515e-02 -1.34735316e-01 4.69725400e-01
-1.00927579e+00 7.47873724e-01 9.20843959e-01 -1.12943836e-01
-1.27835369e+00 -6.05121136e-01 1.62457116e-02 2.60889620e-01
-1.68098863e-02 7.10368335e-01 2.62393244e-02 4.92279708e-01
1.19977880e+00 3.37435126e-01 -2.25051582e-01 1.35089174e-01
8.09319913e-01 4.72475797e-01 3.89147341e-01 -1.01849139e+00
8.16509604e-01 1.29251644e-01 -2.91783605e-02 -1.83229893e-01
-1.02685487e+00 3.25224757e-01 -4.35753644e-01 -1.66949242e-01
7.07924187e-01 -7.84345567e-01 -5.69142699e-01 -1.06710248e-01
-1.24802327e+00 -5.32044172e-01 -6.11419141e-01 6.86342549e-03
-6.99335635e-01 1.31512329e-01 -8.38321507e-01 -1.05137050e+00
-7.62663126e-01 -8.31073105e-01 1.27254045e+00 9.12522674e-02
-1.10072398e+00 -1.21084559e+00 1.70974433e-01 4.16510522e-01
5.00576019e-01 2.84979671e-01 1.08442032e+00 -5.91221213e-01
-4.90099788e-01 -2.20810533e-01 5.94989434e-02 1.38196528e-01
1.47239268e-01 1.45762727e-01 -8.40930462e-01 -3.47400494e-02
-1.43144950e-01 -8.59881639e-01 4.76559848e-01 -1.43388696e-02
7.58549333e-01 -9.23979461e-01 -1.13384597e-01 1.05774608e-02
1.02294159e+00 -1.06295027e-01 6.21916413e-01 3.09136659e-01
3.96308541e-01 6.88312232e-01 8.91243160e-01 6.67523205e-01
6.12235725e-01 4.00280327e-01 7.75888190e-02 1.47915825e-01
-2.93183863e-01 -1.04147363e+00 5.34970403e-01 9.27282214e-01
1.85320556e-01 -4.05144244e-01 -8.13017190e-01 3.68813962e-01
-1.54466033e+00 -1.04503763e+00 -2.20803276e-01 1.85401559e+00
1.39823008e+00 6.26953423e-01 -1.60220310e-01 1.37919813e-01
3.60453129e-01 2.85078913e-01 -2.50902504e-01 -8.85205865e-01
-1.42998770e-01 3.11861724e-01 -2.50323743e-01 6.29450262e-01
-5.22620440e-01 1.06293213e+00 6.98782063e+00 9.44300592e-01
-6.14821911e-01 -3.71007144e-01 8.80957663e-01 -7.22998241e-03
-9.54552472e-01 9.84875932e-02 -9.59103465e-01 7.21247569e-02
1.19246817e+00 -6.64121449e-01 3.40179890e-01 7.34465301e-01
5.67090392e-01 1.10431701e-01 -1.61296642e+00 5.51367939e-01
2.26596594e-01 -1.57967913e+00 1.02266204e+00 -1.80359542e-01
1.13631880e+00 -8.13189268e-01 -3.75167936e-01 6.39002979e-01
6.08084738e-01 -1.30107343e+00 1.08880305e+00 7.39930511e-01
8.80320370e-01 -5.94914258e-01 6.46381795e-01 5.18797159e-01
-1.03499579e+00 3.78176749e-01 -2.92019546e-01 -3.89577806e-01
2.06016049e-01 6.43170297e-01 -1.57793474e+00 7.00799048e-01
1.49554253e-01 2.23617375e-01 -1.01504278e+00 1.86555579e-01
-5.33009410e-01 5.04552782e-01 2.10583463e-01 -6.71382308e-01
3.27657014e-02 2.42878556e-01 2.28359714e-01 1.46685839e+00
3.18003982e-01 -1.56498581e-01 8.54572281e-02 1.68223250e+00
-9.36845988e-02 1.14793621e-01 -6.61421537e-01 -4.26412612e-01
4.66495633e-01 1.13592780e+00 -5.26751757e-01 -8.24549794e-01
2.01078743e-01 8.46003652e-01 4.70960110e-01 7.11635947e-02
-5.25062084e-01 -3.75823855e-01 1.01790167e-01 2.44765267e-01
-2.26431608e-01 4.05161679e-02 -7.32446373e-01 -1.11774099e+00
1.13017663e-01 -1.33666790e+00 5.35640642e-02 -1.46120584e+00
-1.21490610e+00 9.57556486e-01 2.63986886e-01 -1.09307420e+00
-1.10657108e+00 -3.58020931e-01 -6.49143577e-01 1.00859201e+00
-1.01317239e+00 -1.10110354e+00 -3.04192513e-01 1.54296681e-01
7.23775685e-01 -5.18280938e-02 1.13031638e+00 -6.81935728e-01
-1.72149256e-01 6.68725073e-01 -5.58846533e-01 -1.99422128e-02
6.73374772e-01 -1.52500236e+00 1.08164763e+00 9.93312120e-01
2.54101545e-01 1.01655400e+00 8.41952384e-01 -9.81484652e-01
-1.22798896e+00 -1.18868840e+00 1.46144319e+00 -8.17875743e-01
6.54993832e-01 -5.00406444e-01 -7.27026999e-01 4.86627758e-01
5.75517833e-01 -6.31715178e-01 5.99146008e-01 2.11316533e-02
-2.07025409e-01 1.85229573e-02 -1.04997325e+00 9.88281012e-01
9.21714723e-01 -6.80222273e-01 -8.53415370e-01 4.73352581e-01
1.02122068e+00 -4.96855944e-01 -8.82596016e-01 3.05672318e-01
3.30963731e-01 -1.07169461e+00 5.87767899e-01 -7.49200225e-01
1.36164284e+00 -2.46708646e-01 -1.75472386e-02 -1.57152283e+00
-3.86782110e-01 -1.05118060e+00 -1.29384860e-01 1.49497771e+00
8.71896327e-01 -3.54423046e-01 6.34365678e-01 7.91696489e-01
-2.23432407e-01 -8.05611849e-01 -3.79121304e-01 -6.95222497e-01
3.87901694e-01 -4.66084212e-01 6.86504543e-01 6.81994617e-01
6.49285316e-01 7.68927753e-01 -1.48951322e-01 -3.75463188e-01
2.34655097e-01 9.22382027e-02 8.80344450e-01 -9.23175395e-01
-3.77558321e-01 -6.67185009e-01 2.02162087e-01 -7.98696756e-01
1.66408904e-02 -1.13868153e+00 2.02159733e-01 -2.18472385e+00
3.44029546e-01 -3.28045227e-02 3.17413628e-01 4.47669894e-01
-5.58094442e-01 -2.86672544e-02 5.97606838e-01 -4.16803621e-02
-5.81243813e-01 6.10044837e-01 1.55256867e+00 -1.39884889e-01
-8.80790278e-02 -3.16195965e-01 -1.34003603e+00 2.92384684e-01
6.98809445e-01 -2.56588280e-01 -6.34396017e-01 -5.18775806e-02
6.61150455e-01 3.49261642e-01 1.52058080e-01 -1.18625665e+00
8.72085914e-02 -4.00161415e-01 3.73038113e-01 -5.18412828e-01
6.14434807e-03 -9.53127444e-03 -8.98373686e-03 5.26272953e-01
-1.09179461e+00 3.36988658e-01 4.33474630e-01 1.70598283e-01
-2.56594807e-01 -3.87598515e-01 2.53311753e-01 -4.21484590e-01
-1.00900196e-01 -2.55706042e-01 -6.30223334e-01 4.56152678e-01
5.48126876e-01 -1.78557951e-02 -5.35351396e-01 -7.97840118e-01
-1.66867211e-01 8.29816461e-02 3.56385261e-01 4.83563870e-01
7.16391444e-01 -1.42673135e+00 -1.07436395e+00 -1.12951137e-02
3.84829521e-01 -3.82992774e-02 -2.10560098e-01 2.19935387e-01
-5.08581400e-01 1.03827298e+00 -5.02205528e-02 -3.16384107e-01
-7.14332998e-01 5.96599817e-01 -3.05378474e-02 -9.49798703e-01
-1.52218401e-01 5.90371311e-01 5.80581985e-02 -5.37257314e-01
-6.16591200e-02 -1.01383793e+00 5.47790937e-02 -1.75277814e-01
5.33204496e-01 1.98463634e-01 2.30524480e-01 8.79607499e-02
-1.57340094e-01 -4.92769703e-02 -6.44113198e-02 -6.54058218e-01
7.88507164e-01 2.08575636e-01 -5.64302132e-02 7.23610580e-01
4.77360815e-01 1.68189257e-01 -7.81211317e-01 1.42940179e-01
8.38752929e-03 -9.39011425e-02 -5.33509910e-01 -1.30608153e+00
-1.73319981e-01 8.66255462e-01 -2.08839670e-01 4.29426998e-01
8.04586351e-01 -1.63702443e-01 6.20361090e-01 7.73666263e-01
4.27376360e-01 -9.01556373e-01 5.64880490e-01 5.74315310e-01
1.45892525e+00 -1.31123292e+00 -3.32171172e-02 -3.13946992e-01
-1.10328245e+00 1.17518938e+00 7.52436101e-01 1.60453185e-01
-6.15703389e-02 3.52945387e-01 6.74034609e-03 -5.07663749e-02
-1.53852355e+00 1.22937135e-01 5.24954677e-01 6.42092824e-01
1.00809491e+00 2.38651589e-01 -1.70549169e-01 5.70517302e-01
-1.28087556e+00 2.43533388e-01 9.09229100e-01 7.27317333e-01
-4.32099700e-01 -1.04534614e+00 -4.78594571e-01 6.56193316e-01
-1.24257728e-01 -7.16498554e-01 -8.89198244e-01 7.40943491e-01
-3.52206796e-01 1.47673929e+00 -1.99624091e-01 -3.96299005e-01
3.41082186e-01 2.62752831e-01 5.01878321e-01 -9.45722163e-01
-9.38690782e-01 -5.01972198e-01 7.59674013e-01 -2.81051785e-01
-1.10162407e-01 -4.61439788e-01 -1.30926275e+00 -3.79056394e-01
-1.38124987e-01 4.35482472e-01 2.61105031e-01 8.11883032e-01
3.93286705e-01 6.88930809e-01 2.74938077e-01 -5.60569823e-01
-9.24767017e-01 -1.45188427e+00 -1.53461054e-01 6.20135665e-01
3.84166390e-02 -1.14803508e-01 -1.09640382e-01 3.22557926e-01] | [11.519046783447266, 8.80410099029541] |
85831ec2-a59f-4a5c-8ddb-c62b0c16db4d | an-empirical-survey-of-data-augmentation-for-2 | null | null | https://openreview.net/forum?id=n3MFoq1WOXU | https://openreview.net/pdf?id=n3MFoq1WOXU | An Empirical Survey of Data Augmentation \\for Limited Data Learning in NLP |
NLP has achieved great progress in the past decade through the use of neural models and large labeled datasets. The dependence on abundant data prevents NLP models from being applied to low-resource settings or novel tasks where significant time, money, or expertise is required to label massive amounts of textual data. Recently, data augmentation methods have been explored as a means of improving data efficiency in NLP. To date, there has been no systematic empirical overview of data augmentation for NLP in the limited labeled data setting, making it difficult to understand which methods work in which settings. In this paper, we provide an empirical survey of recent progress on data augmentation for NLP in the limited labeled data setting, summarizing the landscape of methods (including token-level augmentations, sentence-level augmentations, adversarial augmentations, and hidden-space augmentations) and carrying out experiments on 11 datasets covering topics/news classification, inference tasks, paraphrasing tasks, and single-sentence tasks. Based on the results, we draw several conclusions to help practitioners choose appropriate augmentations in different settings and discuss the current challenges and future directions for limited data learning in NLP. | ['Anonymous'] | 2021-08-17 | null | null | null | acl-arr-august-2021-8 | ['news-classification'] | ['natural-language-processing'] | [ 6.05846882e-01 2.73870528e-01 -7.28715837e-01 -5.30770242e-01
-9.61687624e-01 -7.34906137e-01 6.25329852e-01 4.18790758e-01
-6.69875622e-01 9.64610338e-01 5.92269182e-01 -5.28123498e-01
2.73173511e-01 -5.45161307e-01 -5.71652353e-01 -4.11680281e-01
3.42235833e-01 7.64210165e-01 -4.53435302e-01 -3.31775546e-01
1.20849617e-01 4.21708047e-01 -1.17287862e+00 4.38879877e-01
9.20890868e-01 7.39276588e-01 -9.33556631e-02 4.89122748e-01
-6.91176951e-01 7.16688871e-01 -7.45052338e-01 -5.09714842e-01
3.27952325e-01 -3.50688517e-01 -1.15663302e+00 -1.82952374e-01
5.07063687e-01 -2.88867187e-02 -2.27068633e-01 7.15133965e-01
7.47876465e-01 3.54552001e-01 5.12276411e-01 -1.30288839e+00
-1.01496947e+00 8.54955256e-01 -5.80930889e-01 5.28997421e-01
3.04633677e-01 3.04312520e-02 1.11594343e+00 -1.04677939e+00
5.80605149e-01 1.18256009e+00 7.66667068e-01 7.89268374e-01
-1.29463792e+00 -5.25461793e-01 2.29248747e-01 3.34125124e-02
-7.45335579e-01 -4.80839878e-01 7.07684755e-01 -1.51313707e-01
1.27309859e+00 1.53707087e-01 2.71188617e-01 1.61941803e+00
-3.18717927e-01 1.19473135e+00 1.13553166e+00 -8.38383436e-01
1.53668910e-01 2.68437833e-01 4.36452448e-01 1.75802529e-01
6.29475266e-02 -7.60414898e-02 -6.67228580e-01 -2.56334037e-01
2.99090415e-01 -2.80646414e-01 -1.76314920e-01 -4.40014899e-03
-1.32212758e+00 1.13786638e+00 3.25446337e-01 3.61608177e-01
-2.82762349e-01 -9.09102336e-02 7.68088043e-01 3.87549609e-01
1.07941854e+00 1.09208977e+00 -1.01259458e+00 -2.74852633e-01
-9.15569067e-01 3.72231662e-01 9.11252081e-01 9.28409755e-01
4.09878016e-01 3.32837939e-01 -4.61305022e-01 1.13148654e+00
-2.07277730e-01 3.49732786e-01 6.76741123e-01 -7.45826840e-01
1.19190145e+00 5.01501918e-01 7.13730827e-02 -7.03243554e-01
-6.59987569e-01 -2.63841599e-01 -1.11127782e+00 -2.84768701e-01
3.99054527e-01 -4.85868573e-01 -9.87394392e-01 1.84483945e+00
1.51072249e-01 -5.38091175e-02 3.00915956e-01 4.96418267e-01
8.73054862e-01 9.13246751e-01 2.49097303e-01 -2.52553284e-01
1.37461257e+00 -1.19734144e+00 -1.14948010e+00 -7.06007123e-01
1.07043362e+00 -7.22872257e-01 1.63036859e+00 5.82334213e-03
-1.01379883e+00 -4.51871693e-01 -8.16094518e-01 -5.17912924e-01
-7.92709529e-01 1.23473406e-01 8.73064280e-01 7.19164848e-01
-6.49283230e-01 4.71405894e-01 -6.75968885e-01 -4.87345546e-01
6.69471025e-01 2.91495919e-01 -2.82780528e-01 -2.46489227e-01
-1.59056258e+00 1.17642927e+00 5.55536389e-01 6.21146858e-02
-3.65959227e-01 -1.07589364e+00 -9.42229569e-01 -1.39986351e-02
3.23736161e-01 -7.10973382e-01 1.26326895e+00 -6.33414745e-01
-1.33556056e+00 6.73940003e-01 -1.25021830e-01 -8.01724434e-01
2.69483894e-01 -3.97927999e-01 -3.07353139e-01 -1.96071178e-01
-2.96022724e-02 9.15001810e-01 4.82911199e-01 -8.21858644e-01
-4.65596825e-01 -3.15400034e-01 6.47455640e-03 5.70365012e-01
-7.08189070e-01 3.00282598e-01 1.09741297e-02 -9.76535201e-01
-2.24226907e-01 -8.48760843e-01 -4.78885412e-01 -3.49264592e-01
-5.68310380e-01 -3.42826545e-01 7.37144768e-01 -7.21889377e-01
9.83929217e-01 -2.04771376e+00 8.35022405e-02 -3.00438583e-01
-1.01466604e-01 4.44213986e-01 -3.17525655e-01 5.16645551e-01
-1.40098885e-01 5.62933385e-01 -3.96045387e-01 -6.47862554e-01
-1.42405359e-02 4.02449548e-01 -7.09788203e-01 -4.49882969e-02
3.72938275e-01 1.22064257e+00 -7.95468748e-01 -4.76003379e-01
1.29170612e-01 3.17541689e-01 -4.06529158e-01 1.76100016e-01
-4.73624438e-01 5.40060580e-01 -3.14601302e-01 5.70328474e-01
4.37452585e-01 -3.72172922e-01 -2.47165143e-01 1.12393670e-01
9.20663774e-02 8.00100207e-01 -7.11147010e-01 1.84362280e+00
-7.18532562e-01 8.04891944e-01 -1.56303868e-01 -1.13359988e+00
8.15474808e-01 5.44344604e-01 2.52658099e-01 -5.41383922e-01
2.33522356e-02 1.66056119e-02 -1.14793167e-01 -5.14605820e-01
7.06384480e-01 -3.41405779e-01 -3.15398067e-01 7.27380991e-01
1.53403088e-01 -1.69397667e-01 4.07188565e-01 2.87420243e-01
1.03278053e+00 -1.98960751e-01 2.71882385e-01 1.14248395e-01
2.41652951e-01 2.21033946e-01 3.42321366e-01 9.50021386e-01
-2.00781133e-02 6.35806382e-01 3.54699492e-01 -4.35579330e-01
-1.36011088e+00 -5.44590950e-01 -1.74629495e-01 1.44527781e+00
-6.61933184e-01 -8.95923600e-02 -5.02596438e-01 -9.74588275e-01
-4.46838029e-02 8.92679334e-01 -6.99549615e-01 -1.05865449e-01
-6.70040011e-01 -1.28133345e+00 7.54972875e-01 9.29652154e-01
5.68163693e-01 -1.46036124e+00 1.49736345e-01 1.00337759e-01
-6.11402810e-01 -1.54799747e+00 -3.06381106e-01 3.67972761e-01
-1.00281656e+00 -5.81160903e-01 -5.86313903e-01 -9.40429509e-01
6.37492955e-01 1.15419909e-01 1.16602278e+00 -4.35118705e-01
-1.79420076e-02 1.95816100e-01 -5.95594823e-01 -7.63164461e-01
-5.12918055e-01 5.80289185e-01 2.19512701e-01 -3.68996739e-01
5.96911073e-01 -3.35682392e-01 -7.89345652e-02 -1.08131438e-01
-7.84441769e-01 1.80040877e-02 8.13612759e-01 1.09869611e+00
3.64518911e-01 -2.98039258e-01 1.02267540e+00 -1.26100600e+00
1.00488687e+00 -5.78493595e-01 -2.19067320e-01 2.63052672e-01
-4.95670766e-01 8.34154263e-02 7.30566859e-01 -6.60927832e-01
-1.22153950e+00 -1.41245469e-01 -4.24793780e-01 -6.71794564e-02
-2.79122978e-01 8.86194050e-01 -2.27161974e-01 4.35469374e-02
9.24770117e-01 9.03400034e-02 -4.91001643e-02 -7.22765625e-01
7.85058141e-01 8.18848610e-01 3.35811764e-01 -5.46121418e-01
6.86267912e-01 1.79547369e-01 -3.36453378e-01 -9.72264111e-01
-1.55867183e+00 -3.38327944e-01 -8.34979117e-01 4.02328044e-01
6.36419475e-01 -7.06181765e-01 -6.36428595e-02 1.66760594e-01
-1.18913627e+00 -3.64335448e-01 -9.34617579e-01 4.18536425e-01
-4.39550906e-01 3.88162673e-01 -8.17179441e-01 -5.07194340e-01
-7.27203548e-01 -7.78124213e-01 6.58977211e-01 -3.30902003e-02
-4.59125817e-01 -1.17964947e+00 2.32299939e-01 9.18078899e-01
4.98663396e-01 6.58068508e-02 1.15866208e+00 -1.39591074e+00
-6.35567084e-02 -5.48504472e-01 -1.42301753e-01 6.48689866e-01
7.11385012e-02 -4.11444217e-01 -1.06261158e+00 -1.55615389e-01
3.00880242e-02 -9.01265025e-01 6.59818351e-01 3.20728451e-01
1.39637434e+00 -6.23325288e-01 -1.91358745e-01 3.82138997e-01
8.40573967e-01 -1.20104678e-01 4.15250123e-01 2.21056283e-01
6.87411308e-01 5.96901834e-01 6.40536487e-01 1.42984152e-01
1.35403633e-01 4.76533085e-01 6.65357485e-02 -2.58046061e-01
-1.29668921e-01 -3.03351402e-01 1.25075176e-01 8.89276266e-01
2.19115928e-01 -6.40510917e-01 -1.00006557e+00 7.27259576e-01
-1.60962689e+00 -6.98638022e-01 8.58270526e-02 1.94323945e+00
1.16016626e+00 8.19456726e-02 -1.39174104e-01 1.47282943e-01
6.89040005e-01 3.29318464e-01 -6.05188847e-01 -4.86036450e-01
-3.28664154e-01 2.76320606e-01 2.64543682e-01 3.31684232e-01
-1.28442192e+00 1.04010475e+00 7.06267405e+00 7.43629932e-01
-7.13646293e-01 2.89290071e-01 7.42086828e-01 -1.62139207e-01
-1.75591916e-01 -8.53788033e-02 -1.06494153e+00 4.32068855e-01
1.16119790e+00 -1.10139072e-01 2.10172370e-01 8.35751534e-01
8.26845225e-03 2.84473002e-01 -1.27780926e+00 6.86701655e-01
1.05668247e-01 -1.52786171e+00 1.93068266e-01 -8.62491950e-02
9.39157486e-01 4.06303883e-01 3.33176196e-01 6.81436419e-01
4.45606291e-01 -1.03715599e+00 -4.93017435e-02 -2.18403965e-01
7.17656851e-01 -5.85565686e-01 1.07404816e+00 5.84664226e-01
-5.44433594e-01 -1.24156550e-01 -4.17146057e-01 -2.10736305e-01
3.44292760e-01 4.89593595e-01 -1.05780268e+00 2.77389854e-01
4.42859173e-01 5.09325624e-01 -3.42897713e-01 6.04270399e-01
-2.82003224e-01 9.59769070e-01 -3.90575677e-01 -1.74187377e-01
4.72788811e-01 1.09496735e-01 3.07321817e-01 1.38245869e+00
-1.11646004e-01 9.27104279e-02 2.40656212e-01 6.34168386e-01
-5.78149080e-01 4.78994489e-01 -6.31807625e-01 -5.10902345e-01
6.09167635e-01 1.18872976e+00 -5.17760813e-01 -3.30908656e-01
-6.29001200e-01 7.41978645e-01 5.12582541e-01 4.23902303e-01
-5.45686305e-01 -2.57044792e-01 4.32292134e-01 -6.77588060e-02
-2.19979778e-01 -7.34724924e-02 -6.95372105e-01 -1.37481236e+00
-1.02582732e-02 -9.90571678e-01 6.61678314e-01 -6.78750515e-01
-1.62391520e+00 5.26354790e-01 1.04048222e-01 -8.33571196e-01
-2.78840065e-01 -5.18181384e-01 -3.04622740e-01 8.95803571e-01
-1.51588774e+00 -1.11965537e+00 1.29542537e-02 2.59318799e-01
8.67814422e-01 -4.00614768e-01 1.13529587e+00 1.99220613e-01
-6.01623416e-01 6.26680911e-01 3.27999234e-01 4.84262615e-01
8.18316281e-01 -1.24174535e+00 8.70835125e-01 7.12359786e-01
3.80951136e-01 5.27996778e-01 6.03109717e-01 -5.63992500e-01
-1.01821911e+00 -9.52117920e-01 1.21789265e+00 -7.77927756e-01
7.82438755e-01 -7.45062590e-01 -1.19350302e+00 1.03005087e+00
4.09534991e-01 2.10658386e-01 1.10347557e+00 5.23054123e-01
-3.19564819e-01 2.24188507e-01 -1.30914986e+00 5.30309379e-01
7.59521186e-01 -4.52442348e-01 -8.52806449e-01 8.83834541e-01
1.17371547e+00 -4.70479965e-01 -8.13464224e-01 5.61891735e-01
-5.70213869e-02 -5.59352525e-02 1.05622089e+00 -1.22621095e+00
5.52556753e-01 4.10334498e-01 8.56769532e-02 -1.47901905e+00
-1.15557820e-01 -5.20314991e-01 -3.24512243e-01 1.41872752e+00
7.96925187e-01 -6.31954134e-01 1.17073691e+00 7.65075266e-01
-1.83667392e-01 -9.09743011e-01 -8.56196284e-01 -7.18287528e-01
4.64075834e-01 -5.48602104e-01 4.36223209e-01 1.40660381e+00
1.78062215e-01 8.84641349e-01 -5.17743170e-01 -1.62790984e-01
2.16971993e-01 -4.89269234e-02 6.89852595e-01 -1.09027183e+00
1.46574778e-02 -1.37349799e-01 6.67287111e-02 -1.02452731e+00
6.23257220e-01 -1.17599463e+00 -1.91198856e-01 -1.70576549e+00
1.50173202e-01 -5.86645544e-01 -1.18647188e-01 9.58916128e-01
-2.94636220e-01 1.99815154e-01 1.85381055e-01 1.92280754e-01
-2.75761515e-01 7.30072796e-01 1.02008462e+00 -2.49408215e-01
-3.74718755e-01 1.18369564e-01 -8.94787788e-01 7.67638326e-01
1.11529148e+00 -5.08807123e-01 -4.52898890e-01 -6.30951464e-01
5.06420493e-01 -2.81320065e-01 -6.05602227e-02 -5.66835582e-01
1.13881506e-01 -9.80167687e-02 3.56063515e-01 -7.61564612e-01
5.67346394e-01 -7.54758894e-01 -7.21384645e-01 1.83589593e-01
-9.54501808e-01 2.76288427e-02 4.90464896e-01 5.30808985e-01
-2.58323163e-01 -5.15435517e-01 6.96629584e-01 -3.07476521e-01
-4.81361628e-01 3.25933754e-01 -3.15384835e-01 8.00744534e-01
7.38119483e-01 1.32972986e-01 -3.69321644e-01 -4.09614950e-01
-9.13468361e-01 4.45106208e-01 -2.85136532e-02 7.23331571e-01
4.01941061e-01 -1.18482459e+00 -1.02371168e+00 1.55742407e-01
1.36373356e-01 1.44420266e-01 3.20760399e-01 4.76548553e-01
-1.45599723e-01 7.05022216e-01 -5.05959578e-02 -1.52545482e-01
-1.14674616e+00 7.85447061e-01 -9.10193771e-02 -7.35567629e-01
-5.70444584e-01 9.30525362e-01 -8.09910968e-02 -9.50730622e-01
2.95453697e-01 -2.69086659e-01 -3.62069726e-01 2.92303592e-01
6.07888103e-01 3.52946043e-01 3.05873632e-01 -2.91479349e-01
-6.78768829e-02 5.28826639e-02 -5.50030351e-01 -1.67961299e-01
1.43626678e+00 4.12218682e-02 1.75339393e-02 5.80949903e-01
1.15751982e+00 -2.51385123e-01 -7.64989734e-01 -6.63261175e-01
6.74406812e-02 -2.03147665e-01 -1.00734115e-01 -1.10014749e+00
-8.06833684e-01 1.15793145e+00 8.47523808e-02 1.75250337e-01
8.98040116e-01 -3.21275629e-02 9.08314705e-01 7.74272978e-01
9.54311620e-03 -1.25733244e+00 3.20801079e-01 8.28842044e-01
1.05800903e+00 -1.42719030e+00 8.45909789e-02 -2.88331658e-01
-9.59368765e-01 7.13114500e-01 6.58904731e-01 2.59097368e-01
5.35216331e-01 4.00687486e-01 -5.55130932e-03 1.18342377e-01
-7.41069198e-01 1.47975355e-01 7.70315900e-02 6.63950920e-01
4.49920326e-01 -2.59185255e-01 -1.74397558e-01 4.01039541e-01
-4.65023667e-01 -1.37281865e-01 5.61404526e-01 9.47622418e-01
-1.72034025e-01 -1.22862899e+00 -1.61430985e-01 7.36319661e-01
-8.04086983e-01 -5.70871770e-01 -5.78891456e-01 9.30805445e-01
-2.30892166e-01 8.38860512e-01 1.34734243e-01 1.58738673e-01
2.71464348e-01 7.20052958e-01 1.74224421e-01 -1.03397608e+00
-7.52339363e-01 -4.89800245e-01 3.90844673e-01 -9.95089188e-02
-5.03297150e-01 -7.46917307e-01 -1.06237209e+00 -2.08490759e-01
-4.37604785e-01 3.04194182e-01 8.35184097e-01 1.00158715e+00
5.48077404e-01 3.34566921e-01 3.05441499e-01 -6.71154082e-01
-7.42659807e-01 -1.48329234e+00 -2.09051073e-01 3.67324442e-01
2.14831665e-01 -2.93681264e-01 -4.63049382e-01 -8.46259147e-02] | [10.77418041229248, 8.266281127929688] |
06b35f40-dc78-4b75-a3a0-cbc290ed6cfd | one-class-support-measure-machines-for-group-1 | 1408.2064 | null | http://arxiv.org/abs/1408.2064v1 | http://arxiv.org/pdf/1408.2064v1.pdf | One-Class Support Measure Machines for Group Anomaly Detection | We propose one-class support measure machines (OCSMMs) for group anomaly
detection which aims at recognizing anomalous aggregate behaviors of data
points. The OCSMMs generalize well-known one-class support vector machines
(OCSVMs) to a space of probability measures. By formulating the problem as
quantile estimation on distributions, we can establish an interesting
connection to the OCSVMs and variable kernel density estimators (VKDEs) over
the input space on which the distributions are defined, bridging the gap
between large-margin methods and kernel density estimators. In particular, we
show that various types of VKDEs can be considered as solutions to a class of
regularization problems studied in this paper. Experiments on Sloan Digital Sky
Survey dataset and High Energy Particle Physics dataset demonstrate the
benefits of the proposed framework in real-world applications. | ['Krikamol Muandet', 'Bernhard Schoelkopf'] | 2014-08-09 | null | null | null | null | ['group-anomaly-detection'] | ['methodology'] | [-1.32317305e-01 1.17729023e-01 -5.62465966e-01 -6.26447558e-01
-5.32280564e-01 -3.11528355e-01 4.36969370e-01 4.98378813e-01
-2.60766566e-01 8.07050467e-01 -3.08432430e-01 -6.17630303e-01
-3.60951692e-01 -7.97881126e-01 -5.61749518e-01 -8.11533332e-01
-2.97752649e-01 4.30954069e-01 3.89498264e-01 1.37750283e-02
4.00478482e-01 6.44572318e-01 -1.39184237e+00 1.45907179e-01
1.12423944e+00 1.34734797e+00 -6.05836630e-01 6.37197137e-01
-3.63740057e-01 6.32308125e-01 -5.82072198e-01 -1.46592185e-01
2.11652309e-01 -6.62289634e-02 -6.98549747e-01 1.02596931e-01
6.57593489e-01 1.26145199e-01 -1.80422932e-01 1.34066474e+00
-1.91750064e-01 3.50527912e-01 1.35322821e+00 -2.08751488e+00
-6.31214380e-01 6.56925067e-02 -6.93909228e-01 6.60747111e-01
1.15575202e-01 -3.61349016e-01 1.08700156e+00 -8.51303756e-01
-1.40722498e-01 1.13489103e+00 6.36589706e-01 3.21172535e-01
-1.37627006e+00 -4.67603564e-01 1.36575103e-01 4.02451217e-01
-1.27299058e+00 4.00595851e-02 4.64428335e-01 -8.34216416e-01
7.37147570e-01 5.14553189e-01 2.17282936e-01 9.56586719e-01
4.42876726e-01 9.03908193e-01 8.90084386e-01 -3.40576857e-01
7.38108873e-01 4.21493798e-01 8.74756396e-01 5.39578795e-01
7.50309050e-01 -1.33126751e-01 -2.23820955e-01 -9.95682597e-01
5.83877861e-01 2.66331285e-01 -9.15401950e-02 -6.58098876e-01
-7.87935376e-01 1.25558317e+00 -1.47520915e-01 2.01354384e-01
-1.78363293e-01 -8.03007260e-02 5.86413741e-01 5.85718393e-01
8.37002754e-01 4.78441454e-02 -4.53422725e-01 2.18749493e-01
-6.74437642e-01 1.15288466e-01 8.84684265e-01 8.68296385e-01
7.33843386e-01 2.81646132e-01 -1.93357989e-01 5.90314567e-01
2.08008140e-01 5.64488053e-01 4.84694421e-01 -3.84208173e-01
4.89016354e-01 6.20215237e-01 2.72172630e-01 -9.17599678e-01
-3.56497973e-01 -1.07326575e-01 -9.10748124e-01 2.45445639e-01
6.12506270e-01 1.40936702e-01 -4.36564744e-01 1.30045116e+00
2.71519542e-01 6.91890597e-01 -1.07557461e-01 6.40131712e-01
1.70447037e-01 4.99735296e-01 8.38924423e-02 -3.19256902e-01
6.96900725e-01 -4.61187363e-01 -5.65847099e-01 1.56947836e-01
9.07367289e-01 -7.62076378e-02 1.13105237e+00 5.90859175e-01
-4.30610180e-01 -3.52186412e-01 -1.13164377e+00 5.40826678e-01
-3.06219190e-01 -8.91024992e-02 7.22271860e-01 7.96196043e-01
-7.30065346e-01 9.06813920e-01 -1.05361342e+00 -2.41714731e-01
5.66089809e-01 2.66398996e-01 -3.60660732e-01 3.36202770e-01
-8.04926813e-01 3.93104315e-01 3.51035476e-01 -3.32038224e-01
-4.25638497e-01 -7.63896346e-01 -8.82788897e-01 2.13324148e-02
2.16412723e-01 -4.45149988e-02 1.00225234e+00 -1.11593425e+00
-1.07827508e+00 9.11555648e-01 2.83104237e-02 -7.17309356e-01
3.46479446e-01 -3.02139550e-01 -7.65306950e-01 -1.95469316e-02
3.87739018e-02 -5.02210140e-01 1.10792148e+00 -5.49555540e-01
-7.60566652e-01 -2.73766845e-01 -4.95538205e-01 -5.19132555e-01
-5.78450441e-01 2.60783117e-02 6.11023903e-01 -6.01192474e-01
1.90578878e-01 -7.47824967e-01 -2.01150298e-01 -3.37457396e-02
-7.20520973e-01 -7.02452362e-01 1.09226537e+00 -2.89680839e-01
1.37806201e+00 -2.20288491e+00 1.90598249e-01 7.35062122e-01
2.12922439e-01 1.70496926e-01 2.15107769e-01 1.85462311e-01
-3.03826243e-01 -8.23837072e-02 -5.10286987e-01 -1.80992886e-01
6.42226636e-02 4.95090276e-01 -8.02687764e-01 1.17117751e+00
1.73981428e-01 4.73793894e-01 -6.23454630e-01 -3.55061859e-01
9.88431051e-02 -2.23642290e-01 -2.55126894e-01 1.02975614e-01
-2.02680320e-01 3.02713662e-01 -6.33582771e-01 7.77364731e-01
8.43227565e-01 -2.64585853e-01 -3.12917054e-01 3.92542392e-01
2.30713859e-01 -3.03597152e-01 -1.41390753e+00 9.68674660e-01
3.80326509e-02 5.04953146e-01 -1.75633490e-01 -1.72392392e+00
1.10507631e+00 -4.88737226e-02 5.22902489e-01 -1.46091133e-01
5.79179823e-02 2.53274530e-01 -1.77175790e-01 -3.35173905e-01
3.78377497e-01 -3.31542015e-01 -2.78185934e-01 3.31534863e-01
2.62139648e-01 4.08438534e-01 2.69028693e-02 2.29329929e-01
9.98601913e-01 -5.07082701e-01 6.88442528e-01 -7.96131492e-01
8.87467742e-01 -3.17402691e-01 6.42005384e-01 9.75355744e-01
-3.82822484e-01 2.15643585e-01 8.56120884e-01 -5.64701259e-01
-9.03097451e-01 -1.67247295e+00 -7.48506188e-01 9.73403931e-01
-1.68496341e-01 -3.51746947e-01 -4.39368516e-01 -1.02441442e+00
5.88915348e-01 8.33229780e-01 -5.76431990e-01 -2.51374274e-01
-2.51645446e-01 -8.25630069e-01 4.54531491e-01 6.96553826e-01
-1.77061073e-02 -6.33041203e-01 -1.73764661e-01 -1.39814541e-01
3.47563833e-01 -9.60581064e-01 -2.24972397e-01 3.61668058e-02
-9.84413385e-01 -1.37929034e+00 -4.52689886e-01 -3.89491886e-01
4.04693127e-01 -8.94609988e-02 1.18214929e+00 -7.33336210e-02
-5.13732076e-01 6.98502660e-01 -2.02553436e-01 -7.52012849e-01
-4.86225784e-01 -1.23752996e-01 8.69340181e-01 4.00343597e-01
7.68682301e-01 -7.20497549e-01 -2.32265592e-01 4.76829171e-01
-8.51146340e-01 -7.64779747e-01 2.24291533e-01 7.05478370e-01
7.34733284e-01 5.84839173e-02 1.04106867e+00 -1.02089608e+00
5.88386476e-01 -1.10981810e+00 -9.66568351e-01 3.15361500e-01
-7.60029435e-01 1.09416537e-01 7.46601522e-01 -6.16632283e-01
-4.92621303e-01 -3.17134231e-01 3.96976680e-01 -6.03901923e-01
-4.02182102e-01 1.75619721e-01 -5.82843199e-02 -6.70253038e-02
6.88945234e-01 3.27813178e-01 7.47208372e-02 -4.85465944e-01
5.78104344e-04 8.08726847e-01 6.06597424e-01 -7.97084630e-01
8.18383813e-01 5.69564223e-01 4.75242376e-01 -1.20771062e+00
-1.12535906e+00 -8.50596488e-01 -6.45390987e-01 2.08923921e-01
7.97123373e-01 -5.18313706e-01 -7.46960640e-01 3.40619713e-01
-5.58301270e-01 5.97680882e-02 -5.66811323e-01 5.58073401e-01
-8.33781838e-01 6.83374047e-01 -5.37062407e-01 -1.26832926e+00
-2.03843683e-01 -3.99639249e-01 9.26441193e-01 2.62068212e-01
-4.17146757e-02 -1.37298608e+00 2.98062652e-01 -2.78195143e-01
2.92739093e-01 4.54661518e-01 1.21320617e+00 -1.33979106e+00
-1.01686493e-01 -5.97158134e-01 4.16876050e-03 7.96733797e-01
7.55328089e-02 -3.38759385e-02 -7.39086628e-01 -5.11402726e-01
1.61712423e-01 -1.12042353e-01 8.93660128e-01 5.20933568e-01
1.56237328e+00 -2.81536490e-01 -3.96963537e-01 4.88918841e-01
1.39528465e+00 -2.66465336e-01 4.86220062e-01 3.51452351e-01
4.27422732e-01 1.59861773e-01 7.28576362e-01 7.38724887e-01
-1.26134902e-01 4.34571475e-01 3.58457059e-01 5.89580871e-02
9.16708827e-01 -2.38698628e-02 3.68216246e-01 5.71700692e-01
2.51847893e-01 1.44364923e-01 -8.07739735e-01 3.58340919e-01
-2.21538019e+00 -1.02925658e+00 -5.77758968e-01 2.55873299e+00
4.27712649e-01 1.88003942e-01 6.70413017e-01 3.09913397e-01
9.05199349e-01 2.55986629e-03 -7.14754760e-01 -5.29573917e-01
-2.56701350e-01 3.20610523e-01 5.09132802e-01 3.10293168e-01
-1.49271393e+00 4.01760757e-01 6.18125296e+00 9.30488646e-01
-7.58958280e-01 -9.74296778e-02 5.61562359e-01 6.16773404e-03
1.01519972e-01 -4.96206358e-02 -7.70036936e-01 6.40347123e-01
1.10126841e+00 -5.49856067e-01 -1.18885256e-01 1.46430326e+00
-2.98909932e-01 -1.50357917e-01 -1.36684418e+00 1.08021462e+00
-1.65803164e-01 -1.00849974e+00 -1.45336375e-01 1.35389760e-01
5.24077833e-01 -1.20203868e-02 1.86840951e-01 3.99691999e-01
1.21328041e-01 -1.18308198e+00 2.30824694e-01 7.73292303e-01
6.55718327e-01 -1.02879655e+00 7.74508655e-01 4.91900265e-01
-9.12959456e-01 -3.36488515e-01 -9.69881535e-01 -6.86636269e-02
-1.94846652e-02 9.94225800e-01 -6.14936113e-01 4.58521813e-01
6.12058520e-01 7.00314224e-01 -5.05671918e-01 1.20932889e+00
1.28576264e-01 9.63611305e-01 -3.47394377e-01 3.05406544e-02
3.27416323e-02 -5.30269504e-01 7.35211015e-01 1.09248984e+00
1.99277386e-01 -3.98808271e-02 3.21233541e-01 7.07497239e-01
2.82793045e-01 2.53113210e-01 -8.66583228e-01 4.27298881e-02
2.26850837e-01 1.23223031e+00 -5.89137614e-01 -3.24740738e-01
-7.27888703e-01 8.03852141e-01 4.80063647e-01 2.91424006e-01
-7.47947991e-01 -4.70772862e-01 1.13948214e+00 3.00918043e-01
1.97398901e-01 -1.60521537e-01 -1.82524219e-01 -1.54128468e+00
-2.25773342e-02 -4.44869667e-01 8.88920546e-01 4.93306592e-02
-2.06589389e+00 2.29144752e-01 1.66862398e-01 -1.52871406e+00
-8.73153806e-02 -9.85251665e-01 -9.73656297e-01 5.34946442e-01
-1.14391589e+00 -3.98444772e-01 8.78522322e-02 9.06467617e-01
1.78422064e-01 -5.93030572e-01 8.27985823e-01 -5.34344539e-02
-5.68444908e-01 6.86751544e-01 4.78804380e-01 8.25670362e-02
4.48991269e-01 -1.71265459e+00 2.28606120e-01 5.92072248e-01
4.20111090e-01 2.27824315e-01 8.74434769e-01 -5.58354676e-01
-1.08346450e+00 -1.15362406e+00 3.21017206e-01 -6.49328172e-01
1.06560969e+00 -2.86746353e-01 -1.36585355e+00 7.30457842e-01
-3.25600922e-01 6.90840960e-01 1.15639699e+00 3.70831877e-01
-4.11268651e-01 1.40628487e-01 -1.43799222e+00 -1.67989638e-02
7.11601794e-01 -4.18987572e-01 -5.80897689e-01 6.47276282e-01
3.38149071e-01 1.43503852e-03 -8.23695540e-01 2.67606020e-01
8.50578099e-02 -8.37549448e-01 8.72493565e-01 -1.65566516e+00
-3.26642655e-02 -6.77242354e-02 -5.93408763e-01 -1.31249619e+00
-1.85552329e-01 -4.35929358e-01 -6.07741237e-01 8.67424250e-01
9.08853188e-02 -9.92111921e-01 7.46830165e-01 5.50966918e-01
3.98537777e-02 -7.92336047e-01 -1.37066555e+00 -1.26367855e+00
4.31474268e-01 -5.64965904e-01 2.89700449e-01 1.16292286e+00
2.91395336e-01 -9.41480324e-02 -4.24834371e-01 4.93071586e-01
1.07163608e+00 1.14363305e-01 6.94521844e-01 -1.90684700e+00
-5.87152898e-01 -2.57547498e-01 -1.18131995e+00 -5.43038726e-01
7.71414697e-01 -1.18902135e+00 -4.38850999e-01 -7.87325382e-01
4.51010093e-02 -3.85988951e-01 -6.23541832e-01 2.14970902e-01
-3.16788815e-02 -2.06151053e-01 -3.45652312e-01 1.43734217e-01
-8.09130490e-01 6.77916110e-01 3.32245111e-01 9.44933146e-02
-1.27283826e-01 7.75360942e-01 -3.85498613e-01 1.06494915e+00
9.05477464e-01 -5.51115453e-01 -3.32726926e-01 4.14593518e-01
-5.79032395e-03 -5.92404455e-02 4.81864423e-01 -9.34988379e-01
9.00211409e-02 -6.19258463e-01 4.22616065e-01 -4.99435157e-01
-3.76692340e-02 -5.48811018e-01 -5.12003958e-01 3.79015684e-01
-2.51517773e-01 1.12516671e-01 1.37351304e-01 1.12964344e+00
-2.19290629e-01 -2.64376879e-01 1.09441555e+00 3.28445792e-01
-5.29702485e-01 4.70409870e-01 -3.66988719e-01 3.39090705e-01
1.38856006e+00 -6.29962459e-02 -3.36440295e-01 -2.61614025e-01
-9.64102268e-01 1.95157781e-01 1.48105189e-01 2.53721356e-01
6.39268100e-01 -1.52324927e+00 -6.10361338e-01 4.70309436e-01
4.54254061e-01 -3.31138372e-01 8.69219825e-02 1.04214895e+00
-5.97272925e-02 3.49597275e-01 -2.01309443e-01 -9.06317890e-01
-1.08553398e+00 5.84328115e-01 3.47854942e-01 -3.13446447e-02
-8.16071749e-01 6.42691314e-01 2.83135206e-01 -1.72197729e-01
2.40959421e-01 -3.92823011e-01 -3.54169831e-02 -2.31306478e-01
7.37963915e-01 7.31461227e-01 3.99505869e-02 -4.95683432e-01
-4.28002268e-01 8.74219388e-02 -3.63978863e-01 3.01424980e-01
1.32592714e+00 1.62620053e-01 -1.17579885e-01 9.50615823e-01
1.25909185e+00 -1.28039390e-01 -9.49565887e-01 -4.51334625e-01
8.20763946e-01 -5.08163095e-01 -4.06773239e-01 2.50085164e-02
-6.58797026e-01 6.02716684e-01 6.22055054e-01 6.05136454e-01
8.48259747e-01 3.93307120e-01 5.32242596e-01 6.62584424e-01
4.47577655e-01 -1.36577022e+00 1.12427488e-01 4.29696113e-01
6.57897115e-01 -1.28953779e+00 -1.26147166e-01 -4.01437521e-01
-6.25937641e-01 1.33353424e+00 6.02920234e-01 -8.28637481e-01
1.10520041e+00 2.39282414e-01 -3.67718518e-01 6.30205171e-03
-6.79685771e-01 1.04177825e-01 4.99053478e-01 7.07990527e-01
5.67418411e-02 3.50009292e-01 -2.51048058e-01 1.01861680e+00
-4.84746024e-02 -3.89547735e-01 5.97886384e-01 7.35875368e-01
-8.74788582e-01 -9.83761132e-01 -3.69159549e-01 1.06338811e+00
-4.43037689e-01 2.79612184e-01 -1.94346115e-01 6.63417578e-01
-4.66892034e-01 7.40940094e-01 4.11068499e-01 -1.46819979e-01
2.79502392e-01 4.24144626e-01 2.06293553e-01 -5.55699587e-01
1.33709118e-01 -5.26750565e-01 -1.73915029e-01 -7.70014048e-01
8.01218953e-03 -7.21844375e-01 -1.11013615e+00 -2.30341092e-01
-2.81332731e-01 2.65577734e-01 3.06653708e-01 1.11352742e+00
1.62013136e-02 -2.19863076e-02 8.29678237e-01 -2.72533774e-01
-1.33663344e+00 -9.52023149e-01 -1.44356585e+00 5.93541443e-01
4.86445665e-01 -8.75887573e-01 -8.48181725e-01 -5.31413794e-01] | [7.637380599975586, 2.4987175464630127] |
37309ea0-8e25-4e98-93ad-940e2e41384d | deep-learning-algorithms-for-coronary-artery | 1912.06417 | null | https://arxiv.org/abs/1912.06417v1 | https://arxiv.org/pdf/1912.06417v1.pdf | Deep Learning Algorithms for Coronary Artery Plaque Characterisation from CCTA Scans | Analysing coronary artery plaque segments with respect to their functional significance and therefore their influence to patient management in a non-invasive setup is an important subject of current research. In this work we compare and improve three deep learning algorithms for this task: A 3D recurrent convolutional neural network (RCNN), a 2D multi-view ensemble approach based on texture analysis, and a newly proposed 2.5D approach. Current state of the art methods utilising fluid dynamics based fractional flow reserve (FFR) simulation reach an AUC of up to 0.93 for the task of predicting an abnormal invasive FFR value. For the comparable task of predicting revascularisation decision, we are able to improve the performance in terms of AUC of both existing approaches with the proposed modifications, specifically from 0.80 to 0.90 for the 3D-RCNN, and from 0.85 to 0.90 for the multi-view texture-based ensemble. The newly proposed 2.5D approach achieves comparable results with an AUC of 0.90. | ['Andreas Maier', 'Axel Schmermund', 'Michael Sühling', 'Anika Reidelshöfer', 'Katharina Breininger', 'Michael Wels', 'Joachim Eckert', 'Felix Denzinger'] | 2019-12-13 | null | null | null | null | ['texture-classification'] | ['computer-vision'] | [-6.14807084e-02 -1.05832852e-01 6.12779558e-02 -1.03862204e-01
-8.83410752e-01 -5.52423120e-01 4.44368124e-01 3.33615035e-01
-3.08687568e-01 8.83784950e-01 3.09756994e-01 -7.23444760e-01
-3.98521274e-01 -8.11584532e-01 -1.88998595e-01 -7.50375509e-01
-4.16145802e-01 5.86482525e-01 5.22083104e-01 -1.66977897e-01
2.22475871e-01 9.56570566e-01 -1.33977306e+00 4.88648415e-01
5.46768546e-01 1.31329286e+00 -2.91501820e-01 1.04999900e+00
2.55739659e-01 7.84344554e-01 -1.75079674e-01 -9.58499089e-02
3.25183183e-01 -2.28436232e-01 -7.92067945e-01 -4.95305628e-01
3.58662516e-01 -2.77146637e-01 -6.93946034e-02 2.83093959e-01
1.15288579e+00 -4.04316075e-02 6.07129395e-01 -1.09738559e-01
-5.29948510e-02 2.62497723e-01 -2.63037682e-01 7.34016776e-01
-2.01173842e-01 6.57729656e-02 5.05706429e-01 -6.15543187e-01
6.93092585e-01 9.39471841e-01 9.19256687e-01 2.16523200e-01
-1.16944003e+00 9.15149674e-02 -2.39246398e-01 7.37302974e-02
-9.80659723e-01 -3.52460802e-01 5.29692292e-01 -6.20405555e-01
8.65771115e-01 2.89724171e-01 6.69032753e-01 8.61672282e-01
6.42549515e-01 2.78854012e-01 1.45695782e+00 -2.00592577e-01
1.97976887e-01 3.18435580e-02 1.55404195e-01 4.64328378e-01
5.97125471e-01 3.16159785e-01 6.81400374e-02 -3.29907954e-01
1.06460428e+00 -2.46476650e-01 1.28358930e-01 -4.84483659e-01
-1.14003956e+00 8.07855189e-01 2.91545242e-01 6.87200904e-01
-5.91330051e-01 5.42586632e-02 1.05193615e+00 3.37381035e-01
9.85403776e-01 3.24402362e-01 -5.78883350e-01 -1.79423615e-01
-9.36417997e-01 3.38580132e-01 6.77165329e-01 -2.15632319e-01
-2.06709616e-02 -3.68293151e-02 -1.74636111e-01 8.34136665e-01
3.10665488e-01 3.86112332e-01 5.74429452e-01 -9.40983951e-01
2.35563800e-01 6.62954211e-01 2.27072909e-01 -8.69028926e-01
-8.57304573e-01 -1.04246473e+00 -1.01168835e+00 6.06602669e-01
6.83255851e-01 -1.91079497e-01 -7.66007185e-01 1.21499240e+00
3.19142461e-01 1.49762839e-01 3.61878604e-01 1.06607389e+00
8.57190788e-01 2.59464444e-03 2.61371315e-01 -1.83945000e-01
1.32141459e+00 -6.52380466e-01 -1.90235302e-01 1.84437260e-02
1.08127689e+00 -7.55934358e-01 5.22389472e-01 3.42119485e-01
-1.18993366e+00 -5.27567506e-01 -8.05433631e-01 2.92406261e-01
-9.96914878e-02 3.68524581e-01 2.87165850e-01 8.53568494e-01
-1.15204632e+00 1.16318929e+00 -1.01279283e+00 -2.15427816e-01
6.41650200e-01 2.21037820e-01 -3.35637957e-01 3.81160937e-02
-1.11961782e+00 1.22082055e+00 1.70687567e-02 3.71019304e-01
-3.31273377e-01 -1.12145233e+00 -3.43973666e-01 -1.57982111e-01
-1.68134168e-01 -1.05868411e+00 9.36999321e-01 -6.32462084e-01
-1.52595806e+00 1.07673430e+00 -5.34192547e-02 -8.33664894e-01
1.16953480e+00 -3.40757519e-01 -2.75612235e-01 5.98967314e-01
1.53083429e-01 -5.30812740e-02 2.39610866e-01 -1.12347436e+00
-3.61980528e-01 -7.03221440e-01 -2.24970151e-02 -1.35836303e-01
3.45854223e-01 -2.23686144e-01 2.99825758e-01 -5.88462949e-01
3.60302292e-02 -1.12888265e+00 -6.70888782e-01 -9.71813276e-02
-5.15838601e-02 -7.67606590e-03 5.52612901e-01 -9.25239503e-01
1.14656937e+00 -1.54760003e+00 1.74898162e-01 2.68044204e-01
7.83414245e-01 8.84445548e-01 2.34415516e-01 7.08694384e-02
-2.83996463e-01 3.32176924e-01 -2.47773528e-01 1.89831436e-01
-5.67016363e-01 -1.84985444e-01 2.58142501e-01 5.49544096e-01
2.75236040e-01 9.87510502e-01 -6.51560724e-01 -3.30548167e-01
5.89820921e-01 5.85637748e-01 -2.13074058e-01 -1.47461668e-01
2.96211690e-01 7.45294392e-01 -7.43915439e-01 1.42150953e-01
6.88349962e-01 -3.15199554e-01 4.16600913e-01 -2.38590792e-01
-3.13077420e-01 -1.36093438e-01 -9.37367558e-01 1.30304563e+00
-8.50138843e-01 6.64789200e-01 -4.29530025e-01 -1.40764487e+00
1.38129056e+00 6.53968632e-01 1.03951514e+00 -6.23778701e-01
1.99171856e-01 7.84983277e-01 5.43559074e-01 -5.49776971e-01
-2.07537398e-01 -3.67445260e-01 3.55616748e-01 4.85258073e-01
-2.78189659e-01 6.79874957e-01 -1.20271957e-02 -2.19264746e-01
1.15008259e+00 2.95441657e-01 2.63570279e-01 -8.08982134e-01
1.14720750e+00 -2.05330521e-01 -3.07066017e-03 6.51372135e-01
-5.05506933e-01 6.52604043e-01 8.87044668e-01 -1.19700384e+00
-1.24892080e+00 -8.96822274e-01 -4.45798397e-01 3.20446134e-01
-1.20947301e-01 1.14159159e-01 -4.22090948e-01 -4.71516073e-01
9.47596654e-02 1.87400311e-01 -9.39565837e-01 -3.02069604e-01
-1.16988277e+00 -9.83710885e-01 4.66541499e-01 6.35632157e-01
4.91209447e-01 -7.54116833e-01 -1.26124096e+00 6.09668612e-01
-2.04068422e-01 -9.57671165e-01 3.55794728e-01 -5.63177578e-02
-1.36364603e+00 -1.03832328e+00 -1.34208596e+00 -4.04167891e-01
-7.99589604e-02 -2.10019037e-01 1.44196749e+00 7.81804696e-02
-3.39105487e-01 1.12699963e-01 -4.63529736e-01 -1.49638414e-01
-7.52990842e-01 3.53949845e-01 -3.66568744e-01 2.24027097e-01
-1.30061703e-02 -7.97781050e-01 -1.34929979e+00 3.56826097e-01
-3.61412168e-01 -2.89393842e-01 6.99045300e-01 3.84611219e-01
6.36825562e-01 -7.64799595e-01 1.02220309e+00 -1.05642200e+00
4.61842507e-01 -4.36669528e-01 -7.93436244e-02 1.06705755e-01
-5.94206095e-01 4.10326757e-02 4.66888785e-01 -1.71184003e-01
-8.90150011e-01 -5.62103018e-02 -1.91030324e-01 -3.63234013e-01
-2.70299405e-01 4.05821949e-01 4.72605109e-01 -1.68372273e-01
1.18959188e+00 -1.28186136e-01 4.62492645e-01 -5.17050624e-01
1.41247392e-01 3.82272929e-01 -1.76936798e-02 -4.63513196e-01
1.15752026e-01 6.96751177e-01 7.12607205e-01 -5.37514448e-01
-6.56430423e-01 -6.01692557e-01 -1.00585401e+00 -5.21302402e-01
7.70870507e-01 -6.21579468e-01 -7.05725849e-01 3.78953546e-01
-8.31450403e-01 -2.24416137e-01 -4.14527774e-01 5.68438292e-01
-8.59251857e-01 5.22752285e-01 -4.50551808e-01 -7.28030622e-01
-9.32802498e-01 -1.25249374e+00 1.07140899e+00 -1.30953506e-01
-1.14005186e-01 -1.31161320e+00 3.84620786e-01 1.62860408e-01
1.04278719e+00 1.16402721e+00 9.76131022e-01 -6.13210797e-01
1.40798241e-02 -3.75379860e-01 -3.39300215e-01 1.49955109e-01
2.11340655e-02 -2.58517593e-01 -8.40453088e-01 -1.69323877e-01
-1.77282423e-01 1.18619241e-01 1.12169170e+00 1.10387135e+00
7.06542015e-01 4.68021750e-01 -4.06731516e-01 2.70030424e-02
1.57650888e+00 2.05997929e-01 9.64687526e-01 5.02433479e-01
2.55902559e-01 4.96408701e-01 1.27543405e-01 3.82500529e-01
-2.40579948e-01 7.07770169e-01 2.72259206e-01 -1.74559176e-01
-4.18814421e-01 6.38790250e-01 -6.36191741e-02 1.25121206e-01
-8.54842365e-01 -6.32017525e-03 -1.11528373e+00 3.81535292e-01
-1.92434740e+00 -8.21650207e-01 -7.95714915e-01 2.20555973e+00
1.66246772e-01 3.46576244e-01 4.46420342e-01 7.72864074e-02
7.68539786e-01 8.23061466e-02 -6.75722539e-01 -7.03295529e-01
-9.06639397e-02 6.06311500e-01 3.19116324e-01 3.73741686e-01
-1.29251814e+00 4.71233606e-01 5.38617277e+00 3.64305109e-01
-1.36475515e+00 7.65863061e-02 1.09730697e+00 1.17675655e-01
2.90314823e-01 -1.95935950e-01 -3.65094066e-01 1.03155501e-01
1.57223034e+00 1.98939919e-01 -2.26054356e-01 3.01746577e-01
5.95905840e-01 -2.20185071e-01 -4.78032082e-01 6.57817185e-01
-1.14632428e-01 -1.55525577e+00 -8.60737711e-02 2.23230392e-01
4.17117894e-01 2.38880083e-01 -5.04045673e-02 8.39328580e-03
-2.78889775e-01 -9.86413479e-01 1.48590475e-01 1.04255664e+00
8.81057680e-01 -4.52984750e-01 1.17834353e+00 -9.22580510e-02
-1.17496622e+00 1.14852376e-01 -3.41101736e-02 1.54271320e-01
4.67521191e-01 9.45398510e-01 -4.73230273e-01 8.19742084e-01
5.93020201e-01 9.26409781e-01 -4.85106766e-01 1.10907948e+00
5.49382627e-01 6.04649365e-01 -2.15386182e-01 2.27066576e-01
3.33805859e-01 9.66534242e-02 7.61778057e-01 1.02558959e+00
3.12999666e-01 -1.81609213e-01 -7.56981522e-02 5.85371912e-01
5.59531748e-01 4.26772594e-01 -3.65493834e-01 5.57325840e-01
-1.95560768e-01 1.26430559e+00 -1.01883674e+00 -5.41807830e-01
-1.07630461e-01 5.36836624e-01 7.89598841e-03 4.50149067e-02
-6.55242920e-01 -1.77745968e-01 2.95278162e-01 5.14216125e-01
2.91865945e-01 3.38180065e-02 -5.07807136e-01 -1.01651704e+00
6.84696808e-02 -3.11926126e-01 4.50997829e-01 -4.25940692e-01
-1.31352925e+00 9.75229919e-01 -2.07877532e-01 -1.33413005e+00
-2.40018964e-01 -1.00742817e+00 -6.02556646e-01 1.28372002e+00
-1.58786607e+00 -9.33212042e-01 -1.51675224e-01 1.29198253e-01
3.61091226e-01 -3.32321107e-01 1.01225233e+00 2.50342369e-01
-1.86900273e-01 1.49726719e-01 4.20628823e-02 -7.49025270e-02
5.18954754e-01 -1.26847875e+00 2.81922340e-01 3.64716083e-01
-5.95698297e-01 1.27364397e-01 4.34117526e-01 -5.94265938e-01
-6.68416679e-01 -1.15413773e+00 7.45468080e-01 -5.85995197e-01
4.69674349e-01 4.24363911e-01 -7.26664245e-01 1.80205211e-01
-2.31138319e-01 6.57701135e-01 7.32090712e-01 -6.10363074e-02
-1.02003150e-01 1.94389895e-02 -1.07362950e+00 2.39959687e-01
8.24504018e-01 9.01035150e-04 -1.35576442e-01 3.58572334e-01
-7.90683844e-04 -3.89493108e-01 -1.72524130e+00 7.59376705e-01
1.14352262e+00 -1.35611439e+00 1.19440079e+00 -9.80567694e-01
6.22352183e-01 9.03638452e-02 -1.97791886e-02 -9.67338681e-01
-5.15497684e-01 -2.96986014e-01 5.81328049e-02 3.95257801e-01
5.48823655e-01 -9.71723080e-01 8.44169497e-01 1.66331321e-01
-1.61303684e-01 -1.33715999e+00 -1.18605316e+00 -2.96058953e-01
7.62229800e-01 -5.26978150e-02 5.37653305e-02 4.01891977e-01
-4.60731536e-01 1.61128938e-01 -6.49712384e-02 -1.81324929e-01
2.71127969e-01 3.38587523e-01 3.72446775e-01 -1.88451970e+00
-3.70192714e-02 -6.88386858e-01 -6.85194373e-01 -2.29184240e-01
-6.43287003e-02 -1.12933326e+00 -7.73836851e-01 -1.59285641e+00
6.76929653e-02 -6.49992704e-01 -5.81848621e-01 -2.14446530e-01
-4.72031720e-02 4.05953079e-01 -1.40684634e-01 3.80547762e-01
-8.82213563e-02 -1.17805697e-01 1.51076663e+00 3.91237587e-01
-4.04164404e-01 4.85965490e-01 -6.03016078e-01 5.54634094e-01
1.42475581e+00 -8.11832547e-02 -1.19559281e-01 -6.19478561e-02
6.65035918e-02 5.54893911e-01 8.31793070e-01 -1.07055128e+00
-5.62419891e-01 4.32165861e-01 6.62622035e-01 -2.29192451e-01
-4.18597832e-02 -4.61305529e-01 7.43981227e-02 9.81525719e-01
-4.41006452e-01 3.86908278e-02 2.98040450e-01 4.40961778e-01
1.48358122e-01 7.09613319e-03 9.95174885e-01 -5.72750211e-01
-2.30034262e-01 1.41909972e-01 -6.01958334e-01 2.13796832e-02
8.43275547e-01 -3.92781556e-01 -3.29525352e-01 -1.30020261e-01
-1.54993272e+00 -4.88848597e-01 -2.64083177e-01 2.53842324e-01
2.18302682e-01 -1.03908241e+00 -1.25070882e+00 -3.08754481e-02
-1.11890882e-01 -4.56111699e-01 7.54501462e-01 1.54203427e+00
-9.86400127e-01 6.22830868e-01 -4.72607821e-01 -8.51614773e-01
-1.05765021e+00 6.83081672e-02 1.22393012e+00 -8.63255680e-01
-1.00367987e+00 1.37572914e-01 -1.41767070e-01 -2.36113146e-01
-5.52166224e-01 -3.01979244e-01 -6.89462781e-01 1.21366102e-02
2.82204300e-01 8.06813538e-01 7.27964699e-01 -7.72076905e-01
-2.91140229e-01 1.02165449e+00 -2.22557075e-02 2.10111037e-01
1.53483236e+00 4.64207307e-02 -1.19092442e-01 3.87760669e-01
1.23225999e+00 -3.55142653e-01 -7.50991285e-01 -3.52924950e-02
-1.28449559e-01 -1.76702626e-02 4.86634761e-01 -1.15791357e+00
-1.35008204e+00 9.89227593e-01 1.38407850e+00 2.41481423e-01
9.19605076e-01 -1.48295194e-01 7.82193005e-01 -1.10277340e-01
-5.54416403e-02 -3.99681926e-01 -2.52009660e-01 2.03866825e-01
8.36185277e-01 -1.01584637e+00 1.16440877e-01 -4.42302734e-01
-3.77266735e-01 1.74779689e+00 9.95391384e-02 -5.73725164e-01
1.12841940e+00 -6.38622418e-03 3.71165723e-01 -2.90367454e-01
-6.02529109e-01 -2.70044714e-01 2.12359518e-01 4.29156423e-01
7.79617906e-01 1.78414360e-01 -7.40068913e-01 2.77334362e-01
2.47264042e-01 2.36250475e-01 3.72377008e-01 6.63891196e-01
-4.22979534e-01 -1.03082407e+00 6.53954744e-02 7.65496016e-01
-7.67711282e-01 2.30353981e-01 -1.76612422e-01 8.42709482e-01
-1.77895561e-01 5.33602059e-01 -1.10115901e-01 -7.44233876e-02
5.95026195e-01 1.64729029e-01 3.67032439e-01 -1.07667148e-01
-8.80460739e-01 2.88943678e-01 2.74220794e-01 -6.42856419e-01
-9.44652915e-01 -9.42097068e-01 -8.03116202e-01 -1.15254216e-01
-1.79421142e-01 -6.36094585e-02 3.34502310e-01 9.85121727e-01
6.79080188e-01 7.61828482e-01 5.05768776e-01 -9.57000732e-01
-2.84774452e-01 -9.22128797e-01 -6.07711375e-01 3.02249014e-01
3.60555947e-01 -7.96922624e-01 -1.23208396e-01 -4.56324071e-02] | [14.163189888000488, -2.4597086906433105] |
502c656e-3566-425c-98aa-bf650274accc | few-shot-adversarial-learning-of-realistic | 1905.08233 | null | https://arxiv.org/abs/1905.08233v2 | https://arxiv.org/pdf/1905.08233v2.pdf | Few-Shot Adversarial Learning of Realistic Neural Talking Head Models | Several recent works have shown how highly realistic human head images can be obtained by training convolutional neural networks to generate them. In order to create a personalized talking head model, these works require training on a large dataset of images of a single person. However, in many practical scenarios, such personalized talking head models need to be learned from a few image views of a person, potentially even a single image. Here, we present a system with such few-shot capability. It performs lengthy meta-learning on a large dataset of videos, and after that is able to frame few- and one-shot learning of neural talking head models of previously unseen people as adversarial training problems with high capacity generators and discriminators. Crucially, the system is able to initialize the parameters of both the generator and the discriminator in a person-specific way, so that training can be based on just a few images and done quickly, despite the need to tune tens of millions of parameters. We show that such an approach is able to learn highly realistic and personalized talking head models of new people and even portrait paintings. | ['Victor Lempitsky', 'Aliaksandra Shysheya', 'Egor Zakharov', 'Egor Burkov'] | 2019-05-20 | few-shot-adversarial-learning-of-realistic-1 | http://openaccess.thecvf.com/content_ICCV_2019/html/Zakharov_Few-Shot_Adversarial_Learning_of_Realistic_Neural_Talking_Head_Models_ICCV_2019_paper.html | http://openaccess.thecvf.com/content_ICCV_2019/papers/Zakharov_Few-Shot_Adversarial_Learning_of_Realistic_Neural_Talking_Head_Models_ICCV_2019_paper.pdf | iccv-2019-10 | ['talking-head-generation'] | ['computer-vision'] | [ 2.74935395e-01 5.74705005e-01 5.26240766e-01 -5.28077543e-01
-7.18691289e-01 -3.27634424e-01 7.25592911e-01 -6.59325421e-01
-4.35592383e-01 7.15647936e-01 2.14031637e-01 3.34925890e-01
4.01553631e-01 -6.77736938e-01 -9.64613557e-01 -7.57845283e-01
1.46310970e-01 1.09041107e+00 3.20707023e-01 -2.57855147e-01
-3.36864889e-01 4.03522760e-01 -1.99596250e+00 1.74570903e-01
2.69715786e-01 6.58749104e-01 3.07366580e-01 1.15348792e+00
2.72384048e-01 7.42763340e-01 -8.33725929e-01 -8.40851128e-01
2.86669552e-01 -4.78310019e-01 -5.61377108e-01 4.34722483e-01
7.88066268e-01 -6.15274727e-01 -6.11074150e-01 7.21188247e-01
1.17074776e+00 2.13180348e-01 4.10654515e-01 -1.24559247e+00
-3.52865815e-01 8.70325685e-01 1.11638028e-02 -9.43366513e-02
4.41300869e-01 8.57415199e-01 4.59559679e-01 -7.34358609e-01
8.13449681e-01 1.24434316e+00 5.47930300e-01 1.30015159e+00
-9.83377576e-01 -6.76282287e-01 -2.03362629e-01 3.08114082e-01
-1.31880856e+00 -8.64257872e-01 5.30043304e-01 -1.84616327e-01
5.84632277e-01 1.29735753e-01 8.81823778e-01 1.81138992e+00
-2.67508626e-01 6.44004464e-01 8.14474285e-01 -3.09500664e-01
3.02900404e-01 3.48226905e-01 -4.29752082e-01 5.57317138e-01
-8.70914236e-02 -5.83330207e-02 -4.21809494e-01 6.53112009e-02
4.46247995e-01 -5.29805152e-03 -3.34502399e-01 -1.98213741e-01
-1.01685834e+00 8.44341218e-01 4.22954172e-01 1.75234810e-01
-3.27063233e-01 2.44387671e-01 4.53085899e-01 -1.94202829e-02
-7.22985491e-02 3.16021323e-01 1.84037521e-01 8.17199275e-02
-1.09109128e+00 3.60084563e-01 1.04721379e+00 1.03883135e+00
5.82976997e-01 1.15346052e-01 -2.99671352e-01 6.03983939e-01
-4.25396651e-01 8.08655858e-01 7.91845143e-01 -1.01427317e+00
1.65448293e-01 -4.47888039e-02 1.35719255e-01 -6.38928175e-01
-3.99190575e-01 -1.70935348e-01 -9.58485961e-01 1.62580028e-01
4.30383444e-01 -4.30005699e-01 -1.15168452e+00 1.95005083e+00
3.86627465e-01 4.96833622e-01 5.89338988e-02 7.74102330e-01
8.49889815e-01 6.97106242e-01 -1.08452037e-01 -1.12021789e-01
1.36498988e+00 -1.02825522e+00 -4.64707166e-01 -6.99560523e-01
8.48435238e-02 -3.81008595e-01 1.06591284e+00 2.83181995e-01
-1.61817181e+00 -6.49290025e-01 -7.57749856e-01 3.90495472e-02
-3.23955834e-01 -3.61588478e-01 3.49601954e-01 7.25743473e-01
-1.16200292e+00 5.32108366e-01 -4.09340054e-01 -5.40509462e-01
5.27359664e-01 6.20685101e-01 -5.01237273e-01 -3.48514408e-01
-1.23184133e+00 9.61725533e-01 3.61379176e-01 -1.38403028e-01
-1.43602324e+00 -5.61278403e-01 -7.70681143e-01 3.10266137e-01
2.46778503e-01 -1.30374658e+00 1.37409210e+00 -1.13706541e+00
-1.80978847e+00 1.00657761e+00 1.10366203e-01 -7.91587472e-01
8.87751997e-01 9.05193761e-02 -1.83743611e-01 3.61108452e-01
-3.12973768e-01 1.04867613e+00 1.42181075e+00 -1.26238358e+00
-4.43593383e-01 -3.89360785e-01 7.26583302e-02 1.05499640e-01
-3.88955623e-01 4.91049029e-02 -6.62626803e-01 -3.25585186e-01
-5.78990459e-01 -1.20449352e+00 -1.39005244e-01 -1.50102466e-01
-4.80051607e-01 5.83126135e-02 7.71765113e-01 -5.89902759e-01
4.59988683e-01 -1.99905658e+00 1.48181662e-01 -1.26870200e-01
2.53830373e-01 4.41655755e-01 -1.96900055e-01 2.56207228e-01
-2.76947673e-02 -2.09402710e-01 -1.70860484e-01 -8.11778009e-01
1.33570924e-01 2.95608521e-01 -2.43313372e-01 4.53497678e-01
-1.24837756e-01 1.15862119e+00 -8.81067157e-01 -5.20118594e-01
2.98712909e-01 7.73498416e-01 -5.46149254e-01 6.18358314e-01
-1.91824719e-01 6.07473016e-01 -3.18790674e-02 7.18652532e-02
4.98944581e-01 -1.01972513e-01 7.90152922e-02 1.10241175e-01
4.75818038e-01 -1.81370482e-01 -1.00350630e+00 1.50338566e+00
-7.95067251e-01 6.07466996e-01 3.20661403e-02 -9.53302026e-01
5.81399083e-01 6.43209100e-01 2.71989048e-01 -4.78267014e-01
3.79572123e-01 -5.96806630e-02 -2.11739734e-01 -6.16151929e-01
2.76305437e-01 -6.19738162e-01 -3.17669332e-01 6.51169479e-01
4.35148507e-01 -2.77932853e-01 7.44634867e-02 6.55443296e-02
9.19952989e-01 -3.87625247e-01 -7.66618624e-02 3.77346635e-01
4.96375591e-01 -4.88370925e-01 3.41308892e-01 9.02657509e-01
1.02245040e-01 1.06504750e+00 1.27297282e-01 -6.01137221e-01
-1.57952523e+00 -9.64348972e-01 2.37646461e-01 1.24693251e+00
-2.82476783e-01 -3.34243402e-02 -1.33381617e+00 -2.95256972e-01
-2.97846884e-01 8.50652575e-01 -7.10791826e-01 -3.93119037e-01
-9.22463238e-01 -6.23570144e-01 6.98337674e-01 4.48646277e-01
4.46072012e-01 -1.32857859e+00 -7.59410143e-01 1.55917719e-01
-1.39626097e-02 -1.28522217e+00 -4.55960274e-01 -6.63736239e-02
-4.12261069e-01 -6.82462275e-01 -1.28220546e+00 -9.41031694e-01
7.22292185e-01 -1.76869296e-02 1.26435137e+00 5.29782772e-02
-5.19768357e-01 4.48734134e-01 -9.28205699e-02 -3.58736783e-01
-6.83692276e-01 7.93541446e-02 2.39659593e-01 2.67257512e-01
-4.84029651e-02 -9.33379829e-01 -5.94500542e-01 1.94141716e-01
-8.83973181e-01 7.88749084e-02 4.69200581e-01 9.25106585e-01
1.66291505e-01 -1.74498186e-01 5.45605063e-01 -1.05615497e+00
3.25560540e-01 -3.50224912e-01 -4.73912954e-01 2.03431055e-01
-3.44530977e-02 -1.24843724e-01 1.08697736e+00 -8.48986566e-01
-1.11151373e+00 3.05908293e-01 -3.17624956e-01 -6.58493519e-01
-3.72116804e-01 -3.02224427e-01 -3.54987770e-01 8.78270864e-02
8.19504142e-01 4.17831719e-01 -1.19571209e-01 -3.09804291e-01
8.85709107e-01 6.38530076e-01 1.17980278e+00 -3.27328146e-01
1.02119935e+00 6.20837927e-01 -1.74284503e-01 -6.85470700e-01
-5.67837238e-01 -4.92790267e-02 -8.34449708e-01 -3.15992564e-01
8.71815503e-01 -8.63906682e-01 -8.56180072e-01 6.38605416e-01
-9.56864953e-01 -5.96540391e-01 -6.35035455e-01 6.75088093e-02
-8.57650220e-01 4.37419489e-03 -4.79592025e-01 -5.46779454e-01
-4.43431199e-01 -1.01440346e+00 1.07397056e+00 3.73968571e-01
-1.07281797e-01 -8.81379664e-01 -6.93473220e-02 3.08595926e-01
5.41406453e-01 1.96657345e-01 5.77421188e-01 -7.38151252e-01
-6.92176998e-01 -6.37540817e-01 4.20119345e-01 3.25270116e-01
-2.76587605e-01 -2.07902253e-01 -1.41916072e+00 -5.66100478e-01
3.49583551e-02 -5.46623290e-01 6.43574178e-01 1.65956929e-01
1.14188814e+00 -7.71553695e-01 -3.04247618e-01 7.15632975e-01
1.00238812e+00 -1.43197805e-01 8.22821856e-01 -8.70820805e-02
6.32589698e-01 6.28153503e-01 -1.88010082e-01 4.37467515e-01
3.89541626e-01 8.61527979e-01 3.72834772e-01 7.21070468e-02
-3.13065648e-01 -3.20667595e-01 2.91966349e-01 4.74966884e-01
-1.11757159e-01 -4.58789200e-01 -6.87007248e-01 5.14175415e-01
-1.37032461e+00 -1.45850992e+00 5.56480229e-01 2.37726188e+00
7.30293155e-01 2.10744560e-01 5.13880372e-01 1.30925819e-01
9.07468915e-01 4.67934161e-02 -7.50736654e-01 -2.43333369e-01
-8.82657990e-02 3.19046825e-01 2.46797591e-01 3.07275832e-01
-8.10042143e-01 1.12029397e+00 6.89651680e+00 4.27271128e-01
-1.41582930e+00 2.15949371e-01 5.22866666e-01 -5.74921191e-01
-5.85837476e-02 -2.63414085e-01 -8.08492422e-01 5.93744993e-01
1.38163185e+00 -4.89248127e-01 6.08782172e-01 1.05158675e+00
-9.45452377e-02 2.63838470e-01 -1.18653440e+00 1.34139943e+00
4.71021652e-01 -1.17869115e+00 1.82851061e-01 3.42512168e-02
6.95431650e-01 4.96473769e-03 2.61066467e-01 4.62696940e-01
2.29613990e-01 -1.24432981e+00 7.08967447e-01 7.81687558e-01
7.26185322e-01 -8.06073666e-01 6.71920955e-01 7.53266454e-01
-4.81233567e-01 -3.47920239e-01 -6.34571373e-01 3.28762084e-01
5.20870388e-01 3.16794664e-01 -1.09671926e+00 -1.27450258e-01
6.24766886e-01 1.85964536e-02 -6.40462935e-01 9.64472115e-01
-1.22108974e-01 2.21915424e-01 -3.88830572e-01 1.32351011e-01
-6.85765445e-02 4.09959942e-01 3.32886219e-01 1.08107030e+00
4.57802534e-01 1.88312575e-01 -1.10187642e-01 5.55360377e-01
-4.71774250e-01 -2.21313506e-01 -7.40441084e-01 3.13866317e-01
3.38703960e-01 1.36663747e+00 -2.01286793e-01 -7.40105212e-01
-1.41731650e-01 1.25886154e+00 3.27632844e-01 3.77912745e-02
-9.48990345e-01 -1.12688430e-01 2.71305442e-01 4.08301592e-01
4.66627568e-01 1.21842645e-01 2.78465211e-01 -1.12396419e+00
-2.08596677e-01 -1.07949412e+00 1.75783977e-01 -1.13705778e+00
-1.08869803e+00 9.09979820e-01 -6.85531199e-02 -9.25939322e-01
-9.94013786e-01 -3.51909250e-01 -9.91080403e-01 7.41819143e-01
-8.93385530e-01 -1.36522412e+00 -6.16605341e-01 9.71054614e-01
4.66827571e-01 -2.71725893e-01 8.93963158e-01 1.85906470e-01
-5.10149062e-01 8.04336727e-01 -1.33232281e-01 8.83763880e-02
7.35955417e-01 -1.06968582e+00 8.32259178e-01 4.92925137e-01
1.02501422e-01 3.94191414e-01 1.06411850e+00 -1.54935300e-01
-1.25769222e+00 -9.19264615e-01 7.79004514e-01 -5.17694652e-01
1.18700124e-01 -6.73336744e-01 -8.89346004e-01 8.88783276e-01
3.65640879e-01 2.22522452e-01 4.72513765e-01 -3.62502515e-01
-6.53235093e-02 -1.84660748e-01 -1.24123335e+00 5.99511445e-01
1.00743115e+00 -4.48948294e-01 -6.31211281e-01 6.74504995e-01
3.98249745e-01 -5.09141207e-01 -6.81950271e-01 1.04697019e-01
5.80218494e-01 -1.19683218e+00 1.14768398e+00 -6.93347514e-01
2.19582677e-01 6.63880557e-02 2.41264224e-01 -1.33906281e+00
-6.84207007e-02 -8.38583589e-01 -2.53446877e-01 1.12375486e+00
2.12153107e-01 -3.64670038e-01 7.78327644e-01 7.60718048e-01
9.26682353e-02 -3.01678717e-01 -9.05123115e-01 -7.19939530e-01
5.15021086e-02 -1.83635995e-01 9.62666452e-01 4.74860638e-01
-3.06022197e-01 6.46850884e-01 -9.73391473e-01 7.16887414e-02
6.88728392e-01 -1.00330681e-01 1.34537768e+00 -1.14296281e+00
-5.35860240e-01 -1.01136968e-01 -4.81443942e-01 -8.39976013e-01
4.07325983e-01 -6.83320463e-01 1.83496624e-01 -1.16842639e+00
3.94191056e-01 -1.14844888e-01 1.71938479e-01 2.95683503e-01
-2.17841938e-01 4.37394261e-01 5.09908199e-01 1.97892681e-01
-4.70548958e-01 3.92908990e-01 1.24216092e+00 -2.16616899e-01
1.04542337e-01 3.41609567e-01 -4.78143007e-01 8.70296359e-01
5.27205765e-01 -4.45594490e-01 -3.88680279e-01 -3.96022618e-01
-4.54069898e-02 1.75012618e-01 7.79855549e-01 -1.45575333e+00
4.19046372e-01 1.78815112e-01 7.41810262e-01 -2.51316220e-01
9.33626413e-01 -5.77973425e-01 4.03845966e-01 2.78095603e-01
-1.91881731e-01 -1.90743551e-01 -5.60948439e-02 3.96947205e-01
1.68942943e-01 -3.43508244e-01 1.19395399e+00 -6.81381345e-01
-6.70249999e-01 4.44437057e-01 -1.50819302e-01 6.55351356e-02
1.06379259e+00 -5.10590859e-02 -1.42820165e-01 -1.01982725e+00
-1.18194032e+00 8.97281896e-03 7.29929805e-01 3.99079502e-01
4.88422573e-01 -1.06288207e+00 -7.16023684e-01 2.55751848e-01
-1.03816964e-01 7.76054859e-02 6.52869046e-01 2.51879245e-01
-3.31170708e-01 1.15426622e-01 -4.93331909e-01 -4.89792019e-01
-1.15973699e+00 1.00754416e+00 5.01438320e-01 -2.33862121e-02
-1.03146327e+00 8.92635882e-01 3.86879474e-01 -1.88225135e-01
2.92348504e-01 3.00641358e-01 5.48449997e-03 -8.53696167e-02
9.64744627e-01 2.75427550e-02 -1.53398708e-01 -9.56537247e-01
-1.25842080e-01 4.17438179e-01 2.47749873e-02 -3.66117984e-01
1.43318498e+00 -6.87003136e-02 4.00421530e-01 1.83019087e-01
1.16439116e+00 -8.65218267e-02 -1.59464252e+00 1.92909893e-02
-5.59300482e-01 -3.27007234e-01 -3.31044048e-01 -5.43720424e-01
-1.12717497e+00 1.21221709e+00 6.00918889e-01 6.46885037e-02
1.13577449e+00 1.77978799e-01 1.26509011e+00 5.66264391e-01
4.73540723e-01 -9.39658761e-01 3.50345790e-01 1.69761389e-01
8.08907747e-01 -1.00758672e+00 -2.48295322e-01 4.25551906e-02
-8.22107017e-01 9.73722219e-01 3.19914341e-01 -7.57416040e-02
2.90079743e-01 2.41032407e-01 1.37523971e-02 4.76767719e-02
-7.64573455e-01 -1.79617822e-01 6.86263219e-02 8.77411544e-01
-1.53915510e-01 -4.09741029e-02 4.15845066e-01 4.25230026e-01
-8.25616956e-01 9.13213268e-02 7.06541777e-01 5.83945453e-01
-4.28646654e-01 -1.07802260e+00 -4.12003309e-01 6.56971484e-02
-3.97364110e-01 1.72884576e-02 -1.95214599e-01 6.90834165e-01
1.11909702e-01 6.12862825e-01 1.62884861e-01 -3.92489433e-01
3.09514642e-01 4.02638167e-01 5.97266972e-01 -7.04451442e-01
-4.70266849e-01 -2.51782060e-01 -1.57110363e-01 -2.60057420e-01
-1.09326415e-01 -4.24226820e-01 -7.67090917e-01 -5.82354486e-01
-3.06523852e-02 -1.30395681e-01 5.68169177e-01 9.78887141e-01
1.92317098e-01 2.81887382e-01 8.39529216e-01 -1.50536633e+00
-5.95295846e-01 -1.05974412e+00 -6.28122807e-01 8.26774061e-01
3.29872698e-01 -4.16919798e-01 -3.82538140e-01 4.55816120e-01] | [12.920309066772461, -0.2884344458580017] |
3e4c5447-2436-4cc3-a6f3-96ab1bb050cf | local-geometric-indexing-of-high-resolution | 1903.00119 | null | https://arxiv.org/abs/1903.00119v2 | https://arxiv.org/pdf/1903.00119v2.pdf | Local Geometric Indexing of High Resolution Data for Facial Reconstruction from Sparse Markers | When considering sparse motion capture marker data, one typically struggles to balance its overfitting via a high dimensional blendshape system versus underfitting caused by smoothness constraints. With the current trend towards using more and more data, our aim is not to fit the motion capture markers with a parameterized (blendshape) model or to smoothly interpolate a surface through the marker positions, but rather to find an instance in the high resolution dataset that contains local geometry to fit each marker. Just as is true for typical machine learning applications, this approach benefits from a plethora of data, and thus we also consider augmenting the dataset via specially designed physical simulations that target the high resolution dataset such that the simulation output lies on the same so-called manifold as the data targeted. | ['Ronald Fedkiw', 'Matthew Cong', 'Lana Lan'] | 2019-03-01 | null | null | null | null | ['physical-simulations'] | ['miscellaneous'] | [ 5.48173822e-02 5.40664345e-02 7.12003484e-02 2.02700615e-01
-8.88082266e-01 -3.46169621e-01 5.88997364e-01 1.09869391e-01
-2.96206594e-01 6.81461632e-01 -3.56432423e-02 -2.33793594e-02
-2.86339074e-01 -6.90933585e-01 -5.42268217e-01 -6.41280055e-01
-1.38393208e-01 7.55789578e-01 3.34553510e-01 -3.35215889e-02
1.95140049e-01 7.87457764e-01 -1.20985341e+00 -3.21933389e-01
8.56725335e-01 6.28615260e-01 -1.06377445e-01 8.60506952e-01
1.87801167e-01 1.24698162e-01 -2.09115162e-01 -1.78447038e-01
2.20116466e-01 -2.57364124e-01 -4.99720842e-01 4.86356951e-02
4.92654353e-01 1.89776599e-01 -1.74084231e-01 9.62105751e-01
2.41740659e-01 3.63524616e-01 7.28375673e-01 -9.40421522e-01
1.59997597e-01 -2.54794478e-01 -5.65653682e-01 -9.71840397e-02
3.91142964e-01 2.45883420e-01 5.37454188e-01 -8.38979661e-01
9.15889382e-01 9.96613920e-01 1.21177840e+00 3.27380449e-01
-1.54839504e+00 8.01396370e-02 -2.76715755e-01 -4.75572824e-01
-1.50478828e+00 -2.69831926e-01 1.04693520e+00 -7.49989629e-01
4.99895781e-01 3.95280987e-01 8.55337083e-01 9.21066880e-01
3.43686461e-01 1.44373134e-01 8.26201022e-01 -2.79865354e-01
3.81434888e-01 -1.07489750e-02 7.07575455e-02 4.66299325e-01
3.38874459e-01 1.92996636e-01 -8.82042944e-02 -6.11223698e-01
1.44780552e+00 -7.47641250e-02 -4.94747490e-01 -1.16536260e+00
-1.48218930e+00 8.71572435e-01 2.36537889e-01 2.98848063e-01
-5.46640456e-01 2.37861976e-01 2.21014231e-01 -1.68337286e-01
3.37457240e-01 7.81096220e-01 -3.76757532e-01 -4.85265404e-01
-1.25277841e+00 6.12522721e-01 6.99928820e-01 6.30672336e-01
1.00352025e+00 6.95600733e-02 2.25942776e-01 4.12398934e-01
1.93044618e-01 3.27538550e-01 2.40726694e-01 -1.17172563e+00
3.47260028e-01 5.76624155e-01 4.99556005e-01 -1.11180222e+00
-6.66730821e-01 -3.01954418e-01 -9.18798089e-01 5.32633722e-01
7.12328672e-01 -4.12162185e-01 -7.16601193e-01 1.50025511e+00
6.80089474e-01 4.10898775e-01 -4.80173267e-02 1.13893557e+00
1.10385612e-01 6.69189215e-01 9.70079303e-02 -1.27494991e-01
8.70126545e-01 -3.15955490e-01 -4.51546460e-01 -1.88084751e-01
7.30574489e-01 -7.38665700e-01 1.06477082e+00 2.41882399e-01
-1.13394046e+00 -5.32501101e-01 -1.03404093e+00 -3.67675796e-02
-8.64175707e-02 -9.74877253e-02 4.22280490e-01 2.21783012e-01
-7.40265608e-01 9.98356640e-01 -1.02527606e+00 -2.57185757e-01
1.29427895e-01 2.50037223e-01 -3.88951033e-01 1.79684341e-01
-8.53025854e-01 1.07946932e+00 -7.35550607e-03 9.51867327e-02
-4.52982098e-01 -1.10306227e+00 -7.82355487e-01 -2.49870047e-01
2.67753452e-01 -8.36958766e-01 7.46824265e-01 -8.58255208e-01
-1.38313067e+00 7.09754109e-01 -6.73569646e-03 -1.19894713e-01
9.21447933e-01 -2.14038476e-01 -2.25986138e-01 5.76560036e-04
-1.85644835e-01 5.04442573e-01 7.69878805e-01 -1.52030611e+00
-1.69530824e-01 -1.95638418e-01 -3.61983031e-01 3.09248157e-02
2.96047658e-01 -2.25336313e-01 -4.66419786e-01 -6.64949179e-01
3.06919783e-01 -1.15720558e+00 -7.12023318e-01 1.45949170e-01
-3.66345316e-01 3.35676968e-01 1.19277358e+00 -5.69462419e-01
1.16061914e+00 -2.05466914e+00 4.20498878e-01 5.79916120e-01
2.45466217e-01 2.31443599e-01 2.04729766e-01 4.57626104e-01
-7.70144984e-02 -1.10786952e-01 -6.28391445e-01 -3.26292187e-01
-1.96123272e-01 1.66246727e-01 -1.87955976e-01 8.68287504e-01
2.95537561e-01 6.98237717e-01 -7.72509336e-01 -4.33329701e-01
3.52619648e-01 6.76575005e-01 -5.28385818e-01 1.01509191e-01
-2.47782558e-01 6.21110857e-01 -6.80282593e-01 2.34656557e-01
7.08249688e-01 -2.48703226e-01 -1.67049155e-01 -7.26758018e-02
-2.96370953e-01 -1.22070402e-01 -1.56524622e+00 1.77011883e+00
-6.40798211e-02 4.74198937e-01 2.03605533e-01 -9.18974519e-01
1.16684115e+00 2.51636982e-01 1.13844597e+00 -2.41193861e-01
-6.99500069e-02 2.46751651e-01 -1.77777976e-01 -5.41222274e-01
6.88319743e-01 -3.18886936e-01 -3.14066380e-01 1.13844544e-01
-4.77493227e-01 -2.55680263e-01 -5.25603831e-01 -1.12947330e-01
9.33782816e-01 5.35359919e-01 2.70616803e-02 -4.40906912e-01
2.46887699e-01 6.67250216e-01 4.93552357e-01 3.14613253e-01
5.85908331e-02 1.09814906e+00 5.01118600e-01 -5.08181810e-01
-1.57244253e+00 -9.88641202e-01 -3.91584903e-01 -6.87812045e-02
1.76471904e-01 -1.99917778e-01 -9.21817541e-01 -6.08634315e-02
9.20723453e-02 2.49027476e-01 -5.99167466e-01 -8.39749351e-02
-9.09257054e-01 -5.75954735e-01 2.74202645e-01 2.25186765e-01
2.49964997e-01 -6.43427193e-01 -1.07915711e+00 4.01977777e-01
2.84022659e-01 -9.29405272e-01 -3.81469339e-01 2.07519636e-01
-1.10707378e+00 -1.09580410e+00 -8.06320906e-01 -3.50088954e-01
5.69312871e-01 -1.49690136e-01 1.05385518e+00 3.87041550e-03
-3.25205684e-01 4.03553069e-01 2.66675018e-02 1.49793938e-01
-3.40575516e-01 -4.35930528e-02 5.42774945e-02 -3.90506275e-02
-5.30349873e-02 -5.71434975e-01 -5.44045687e-01 2.29711592e-01
-8.19249272e-01 1.43686950e-01 2.15549320e-01 5.27957201e-01
8.03156972e-01 -7.61622488e-02 1.85369194e-01 -6.71674728e-01
3.52543145e-01 -5.90637505e-01 -7.93899417e-01 -8.64241570e-02
-1.17499359e-01 -6.66163862e-02 5.85457146e-01 -7.65759945e-01
-5.96058786e-01 5.61867237e-01 -9.96219888e-02 -8.50147605e-01
-2.95195222e-01 5.42984784e-01 -1.42528072e-01 -2.59237736e-01
6.73682690e-01 -1.92710325e-01 2.18895927e-01 -6.38146579e-01
2.92809665e-01 1.68043077e-01 7.26940811e-01 -7.00293124e-01
7.90195942e-01 5.12214005e-01 6.41680241e-01 -1.03727877e+00
-2.02366143e-01 -2.20468313e-01 -9.76930499e-01 -3.15044254e-01
1.00093913e+00 -4.12596554e-01 -5.54962158e-01 1.10429853e-01
-1.02649736e+00 -4.63642359e-01 -5.67728102e-01 6.35884225e-01
-9.61926699e-01 3.04190159e-01 -1.57035008e-01 -8.69016767e-01
1.21628866e-01 -1.18955636e+00 1.07971156e+00 3.65352146e-02
-5.72799146e-01 -1.25731838e+00 5.00300944e-01 -1.32584020e-01
2.59501189e-01 9.99480784e-01 8.58114004e-01 -1.89107046e-01
-6.53181076e-01 -6.55506670e-01 1.68883547e-01 -1.38151556e-01
6.95342049e-02 4.47498202e-01 -7.38746822e-01 -1.97758645e-01
1.48206666e-01 1.33335516e-01 1.57623306e-01 6.73096180e-01
6.52530730e-01 -2.51842849e-02 -5.87104023e-01 8.11144054e-01
1.68296885e+00 -1.07371025e-01 6.64727509e-01 2.13048697e-01
8.71998727e-01 5.74815810e-01 6.34677589e-01 2.23964542e-01
1.48706183e-01 1.37679160e+00 3.41493934e-01 -3.17981541e-01
1.29005104e-01 -3.67773414e-01 3.63413170e-02 3.72595280e-01
1.41124818e-02 3.64078909e-01 -1.19767666e+00 5.18417776e-01
-2.21521068e+00 -8.11971307e-01 -6.41822636e-01 2.58028746e+00
5.41822076e-01 7.60364756e-02 5.17864048e-01 6.08095452e-02
4.96726334e-01 -1.45232126e-01 -5.80523252e-01 -2.59648830e-01
8.33593756e-02 -1.26119808e-03 4.66076016e-01 6.86373174e-01
-9.80131984e-01 4.88522589e-01 7.23055172e+00 6.10997260e-01
-1.35688508e+00 -2.93769807e-01 3.87335718e-01 2.93622147e-02
-1.71306148e-01 2.38634154e-01 -5.80236077e-01 5.65608382e-01
1.11887276e+00 -1.77232489e-01 2.22784325e-01 6.37243688e-01
6.52945459e-01 -3.69090706e-01 -9.37101960e-01 7.26175666e-01
-3.93254310e-01 -1.45894015e+00 -2.60478854e-01 3.64271581e-01
5.95338225e-01 -7.99436942e-02 -8.54120478e-02 -2.79090911e-01
3.89804095e-02 -1.16208398e+00 6.00419104e-01 1.05870605e+00
7.39162564e-01 -5.99725664e-01 1.82250202e-01 7.63657868e-01
-1.10105896e+00 5.25681615e-01 -2.99200356e-01 1.08302720e-02
3.81694824e-01 5.44379115e-01 -5.83732426e-01 5.24416566e-01
2.78548151e-01 5.38438380e-01 -2.72923172e-01 1.48395932e+00
4.76249844e-01 3.54443431e-01 -5.65911591e-01 3.01353991e-01
1.67769819e-01 -7.23876059e-01 7.39130080e-01 9.72149193e-01
4.32112902e-01 -7.27354735e-02 1.85982019e-01 8.55054557e-01
5.03694415e-01 -9.73120034e-02 -9.27234113e-01 3.83340418e-01
2.29591191e-01 1.22786498e+00 -6.42095268e-01 -1.12726860e-01
-3.42575580e-01 5.67704022e-01 2.03443527e-01 5.15965521e-01
-6.41140103e-01 4.03770171e-02 9.88008976e-01 6.95734680e-01
2.49601379e-01 -8.44923615e-01 -5.99063158e-01 -9.06250477e-01
6.74461052e-02 -6.10861957e-01 1.48884997e-01 -8.79772902e-01
-9.94462967e-01 3.33600760e-01 2.16098458e-01 -1.56360936e+00
-4.30882335e-01 -4.58777040e-01 -6.77133441e-01 1.14133084e+00
-1.12196815e+00 -7.75620997e-01 -1.63203463e-01 4.39673781e-01
-6.86316844e-03 4.31103528e-01 7.17015684e-01 2.97696024e-01
-2.65426725e-01 9.35872495e-02 2.27979422e-01 -1.75557360e-01
2.73703724e-01 -1.00441825e+00 2.74596840e-01 6.17074907e-01
1.67425163e-02 5.35636961e-01 1.11139286e+00 -8.27531278e-01
-1.63547707e+00 -9.72168684e-01 4.60307956e-01 -7.47504115e-01
7.00263917e-01 -2.05856979e-01 -1.39835477e+00 5.67249060e-01
-3.37102026e-01 4.14285183e-01 2.79195569e-02 -3.61535430e-01
3.74808878e-01 2.31919333e-01 -1.24412525e+00 6.78543270e-01
6.78524613e-01 -3.14565301e-01 -2.48121306e-01 -3.54791433e-02
2.82331616e-01 -8.13954949e-01 -1.18625975e+00 4.64517146e-01
2.87909687e-01 -6.31126821e-01 1.10724223e+00 -6.88567698e-01
3.35791886e-01 -6.00189328e-01 3.16606350e-02 -1.39330614e+00
-1.57393917e-01 -9.54085231e-01 -9.60248411e-02 9.54606116e-01
1.48071468e-01 -3.32242787e-01 1.26980639e+00 1.01396573e+00
-1.50419816e-01 -8.92048359e-01 -1.32249987e+00 -7.35151410e-01
4.17390823e-01 -4.77596015e-01 5.30759275e-01 9.63803530e-01
-7.19954148e-02 -1.29170135e-01 -2.61869162e-01 2.98779517e-01
7.67469645e-01 -4.80526546e-03 1.04357922e+00 -1.26630807e+00
-1.56532422e-01 -3.67298067e-01 -4.16223943e-01 -1.08761942e+00
-2.23542973e-01 -4.50586557e-01 1.00563176e-01 -1.40648711e+00
-2.72602230e-01 -9.52713728e-01 5.39038301e-01 -2.24504963e-01
-1.19910955e-01 2.50800494e-02 -6.45595565e-02 3.75279486e-01
-9.20938328e-02 4.09289837e-01 1.31023622e+00 5.28486669e-01
-2.58533120e-01 1.84311971e-01 -1.82722420e-01 7.99402297e-01
3.94277692e-01 -3.43542457e-01 -1.35874197e-01 -1.37337685e-01
2.14378551e-01 6.85204923e-01 7.69697785e-01 -1.16294408e+00
1.99859485e-01 -3.68042439e-01 3.49311829e-01 -3.97138059e-01
6.08138859e-01 -1.28093171e+00 9.90692556e-01 1.47964641e-01
-2.00521663e-01 2.63830841e-01 2.61601776e-01 3.66378725e-01
2.20083091e-02 -3.73399079e-01 1.00695252e+00 7.65042156e-02
-2.71587014e-01 3.08749080e-01 -2.65674978e-01 7.12687671e-02
1.06933856e+00 -6.25146449e-01 -1.17236719e-01 -2.01021358e-01
-8.78770411e-01 6.40336610e-03 1.24140036e+00 -2.10529147e-03
4.23306137e-01 -1.57747316e+00 -4.06814814e-01 2.21638501e-01
6.25806674e-03 4.69138801e-01 3.67332287e-02 8.03553164e-01
-5.51937878e-01 1.37217999e-01 -2.24997013e-04 -7.75858998e-01
-7.75987804e-01 5.01550496e-01 7.82799125e-01 -1.39556035e-01
-1.28219533e+00 2.06977382e-01 -1.72496036e-01 -1.28163382e-01
-1.21787399e-01 -2.53320992e-01 1.70296907e-01 -2.98831701e-01
6.43828360e-04 5.11313856e-01 -4.48011011e-02 -9.16573107e-01
-1.45750657e-01 9.65223730e-01 6.30548000e-01 -2.06440002e-01
1.27455473e+00 1.32308364e-01 2.54016638e-01 5.65203488e-01
1.30402613e+00 1.05994433e-01 -2.07510567e+00 1.06792577e-01
2.89007753e-01 -5.94906628e-01 2.24813502e-02 -2.08517760e-01
-8.88869345e-01 6.55591786e-01 3.28502893e-01 4.22110111e-01
5.14819682e-01 -1.69567451e-01 5.64100266e-01 -1.39008909e-01
3.16234708e-01 -9.01013017e-01 -1.66194260e-01 2.28799015e-01
8.23731601e-01 -9.91692126e-01 1.72235772e-01 -4.44964439e-01
-6.14318848e-01 1.14587295e+00 3.11032474e-01 -6.75036132e-01
5.80493033e-01 5.22758126e-01 4.80418727e-02 -4.69593436e-01
-3.76748145e-01 2.22042739e-01 4.39189941e-01 7.14224398e-01
2.62295604e-01 -8.38830769e-02 -4.27266173e-02 1.42883226e-01
-2.28363380e-01 1.51368722e-01 5.80094457e-01 8.08885157e-01
-4.88857836e-01 -1.02248943e+00 -7.39112079e-01 4.22771782e-01
-1.04366302e-01 3.95893216e-01 -2.81419139e-02 1.01226604e+00
-1.79475501e-01 4.43700939e-01 3.49648029e-01 5.20939976e-02
5.88892043e-01 3.84761319e-02 2.08913296e-01 -4.86751080e-01
-4.51529324e-01 2.45046154e-01 1.01587661e-02 -5.06861985e-01
-1.76813513e-01 -1.02222514e+00 -1.44751203e+00 -3.94608796e-01
4.47359569e-02 2.16674075e-01 7.00682878e-01 8.25870991e-01
2.56332308e-01 2.46396512e-01 5.21876037e-01 -1.21968150e+00
-2.56167889e-01 -5.54852724e-01 -4.88948435e-01 5.35121322e-01
6.74143076e-01 -8.39622438e-01 -3.16254854e-01 -7.92849958e-02] | [8.092839241027832, -2.6308419704437256] |
2419904a-a313-48bc-b87e-53570c87f61e | scmhl5-at-trac-2-shared-task-on-aggression | null | null | https://aclanthology.org/2020.trac-1.10 | https://aclanthology.org/2020.trac-1.10.pdf | Scmhl5 at TRAC-2 Shared Task on Aggression Identification: Bert Based Ensemble Learning Approach | This paper presents a system developed during our participation (team name: scmhl5) in the TRAC-2 Shared Task on aggression identification. In particular, we participated in English Sub-task A on three-class classification ({`}Overtly Aggressive{'}, {`}Covertly Aggressive{'} and {`}Non-aggressive{'}) and English Sub-task B on binary classification for Misogynistic Aggression ({`}gendered{'} or {`}non-gendered{'}). For both sub-tasks, our method involves using the pre-trained Bert model for extracting the text of each instance into a 768-dimensional vector of embeddings, and then training an ensemble of classifiers on the embedding features. Our method obtained accuracy of 0.703 and weighted F-measure of 0.664 for Sub-task A, whereas for Sub-task B the accuracy was 0.869 and weighted F-measure was 0.851. In terms of the rankings, the weighted F-measure obtained using our method for Sub-task A is ranked in the 10th out of 16 teams, whereas for Sub-task B the weighted F-measure is ranked in the 8th out of 15 teams. | ['Pete Burnap', 'Matthew Williams', 'Wafa Alorainy', 'Han Liu'] | 2020-05-01 | null | null | null | lrec-2020-5 | ['aggression-identification'] | ['natural-language-processing'] | [-3.45890939e-01 1.87372014e-01 1.13400914e-01 -2.06896260e-01
-6.86130166e-01 -3.03888708e-01 5.15318871e-01 4.88312334e-01
-1.05437708e+00 7.61583090e-01 2.26971984e-01 -1.47219226e-01
-8.38291764e-01 -5.17971158e-01 1.31765679e-01 -5.36948919e-01
-2.59658635e-01 7.91291177e-01 9.73993167e-02 -4.12285119e-01
6.75446570e-01 5.53372920e-01 -1.47702599e+00 3.57439607e-01
3.93145144e-01 9.87210691e-01 -3.73753905e-01 9.34091568e-01
9.65469182e-02 9.15443122e-01 -8.30668747e-01 -1.14760172e+00
8.47641155e-02 -1.47246584e-01 -1.22300994e+00 -4.60536152e-01
4.82932806e-01 -4.64759737e-01 -4.24814641e-01 9.34831500e-01
3.61034274e-01 3.11933398e-01 1.11475027e+00 -1.38073254e+00
-2.88089573e-01 7.01064050e-01 -5.48436046e-01 5.41098475e-01
4.22464997e-01 1.68961331e-01 1.42200232e+00 -7.26468265e-01
4.08477455e-01 1.14648783e+00 6.86807275e-01 9.44377899e-01
-1.12106681e+00 -9.49284434e-01 -1.29919618e-01 2.62874305e-01
-1.22510815e+00 -4.02009159e-01 4.36922938e-01 -8.84008110e-01
1.08600700e+00 5.11589348e-01 6.17083132e-01 1.55186296e+00
-1.12889878e-01 4.42672908e-01 1.23314106e+00 6.65604845e-02
1.75713837e-01 1.57028466e-01 6.13408327e-01 5.86016774e-01
1.97194472e-01 3.17203522e-01 -6.13965034e-01 -4.65453148e-01
2.50603050e-01 -5.22973239e-01 3.84143084e-01 5.06668091e-01
-6.97325289e-01 1.02858281e+00 3.78596149e-02 4.43624288e-01
-2.25525022e-01 -2.64356304e-02 8.57506454e-01 2.68841118e-01
3.46984923e-01 7.80333877e-01 -1.58961713e-01 -9.13333774e-01
-6.99289501e-01 8.12302709e-01 6.04656041e-01 3.66395622e-01
1.22750431e-01 2.35503707e-02 -2.36012787e-01 1.34909713e+00
-7.25335926e-02 -3.09154242e-02 8.08239400e-01 -1.00573957e+00
5.41779399e-01 6.69891834e-01 4.50511985e-02 -6.94025755e-01
-1.00310445e+00 -4.32988822e-01 -4.49313819e-01 3.77105147e-01
7.20045090e-01 1.61326658e-02 -3.01175892e-01 1.88066161e+00
2.19566934e-02 -4.00023073e-01 -3.74816693e-02 7.15312064e-01
8.64801466e-01 1.30990058e-01 2.69726843e-01 -5.86163700e-02
1.33083665e+00 -8.89427364e-01 -2.42757618e-01 -5.53217530e-02
8.22742283e-01 -5.28840184e-01 7.83315003e-01 6.10030711e-01
-1.22449148e+00 -4.16283786e-01 -7.86406994e-01 1.27102107e-01
-4.75443512e-01 -1.49798647e-01 4.11569387e-01 8.90198052e-01
-9.09450769e-01 5.28575718e-01 -4.70013648e-01 -2.05978900e-01
2.31165424e-01 6.47864401e-01 -9.37674284e-01 4.05339360e-01
-1.00686240e+00 1.05473924e+00 2.77327627e-01 -3.53793114e-01
-8.73272002e-01 -7.01353490e-01 -6.31323397e-01 -4.50347178e-02
-2.89947465e-02 -1.44940928e-01 9.92442071e-01 -4.46694791e-01
-1.29584301e+00 1.39637578e+00 3.44436705e-01 -6.22457787e-02
6.24214351e-01 -1.91121757e-01 -5.74897468e-01 -7.25035891e-02
3.94901663e-01 3.44847947e-01 4.52296436e-01 -8.98380280e-01
-1.01416314e+00 -7.68939674e-01 2.85936832e-01 1.44775197e-01
-5.35702169e-01 7.36320019e-01 2.01020822e-01 -4.41093624e-01
-1.66786522e-01 -9.51734185e-01 1.13610461e-01 -3.31069708e-01
-4.61164445e-01 -7.96448171e-01 3.67580861e-01 -7.73943305e-01
1.67521524e+00 -2.29483438e+00 6.07400775e-01 6.04920238e-02
6.36545122e-01 3.63047183e-01 1.46588743e-01 5.93573511e-01
-5.62815726e-01 1.89495161e-01 -1.56532213e-01 -5.73506474e-01
2.98237383e-01 1.02143712e-01 1.02545589e-01 3.92224580e-01
-7.71393999e-02 1.50297105e-01 -8.97119343e-01 -2.24561691e-01
-9.27282348e-02 -2.15607643e-01 -7.49838412e-01 4.14544702e-01
6.89987242e-01 3.62376332e-01 -2.77562410e-01 2.89089531e-01
2.48994231e-01 7.36165881e-01 -2.96317399e-01 3.48152995e-01
-4.24421072e-01 1.96791902e-01 -8.99796724e-01 9.41000640e-01
-5.13494015e-01 5.83607495e-01 2.37552971e-01 -7.02316463e-01
8.68809640e-01 9.70186889e-02 4.44835931e-01 -4.56429154e-01
2.80206174e-01 5.69965541e-01 6.15115404e-01 -8.30961645e-01
6.00283504e-01 -3.23527902e-01 -5.61915874e-01 5.04824817e-01
2.23502163e-02 -1.06243044e-01 4.64744955e-01 2.53682792e-01
1.60143185e+00 -4.13349450e-01 1.27712563e-01 -1.54716760e-01
6.71416223e-01 -3.91646445e-01 2.63606936e-01 6.86595142e-01
-5.92119515e-01 4.03350264e-01 1.04365146e+00 -5.31564236e-01
-1.10028374e+00 -9.97546554e-01 -5.07076979e-01 1.40068996e+00
-5.09663761e-01 -7.18839347e-01 -7.69161463e-01 -7.96193421e-01
8.76204371e-02 1.06116796e+00 -1.05556679e+00 -4.25298661e-01
-4.68706846e-01 -6.22892320e-01 1.22763312e+00 4.75908905e-01
9.74819809e-02 -1.06811178e+00 -5.35160303e-01 -3.28883007e-02
-7.74789974e-02 -9.09600317e-01 -1.41512215e-01 2.89539069e-01
-8.76389444e-02 -1.07952285e+00 -4.51780260e-01 -2.58759081e-01
2.03108519e-01 -5.15314937e-01 6.40144646e-01 2.40377456e-01
-3.95954490e-01 1.56579316e-01 -5.59390545e-01 -1.92573756e-01
-1.65011823e-01 4.16761562e-02 6.38498724e-01 3.84268090e-02
4.71360773e-01 -5.23045421e-01 -1.27333060e-01 2.58489519e-01
-5.95429838e-01 -3.91988248e-01 9.51474383e-02 9.60408568e-01
-2.47814164e-01 -1.43167719e-01 4.07171428e-01 -5.92761576e-01
1.06069338e+00 -5.66555083e-01 -1.49920672e-01 -4.21144158e-01
-1.57105684e-01 -5.69655776e-01 6.97682083e-01 -3.08448017e-01
-4.63217258e-01 -5.24895251e-01 -7.32960165e-01 -2.19480827e-01
-4.42628741e-01 1.77962333e-01 3.80634129e-01 3.36142540e-01
6.91057265e-01 -1.54176876e-01 -3.09456259e-01 -4.62933421e-01
-3.17528874e-01 1.21644950e+00 4.99202818e-01 -8.12235177e-01
7.44288683e-01 -4.24685962e-02 9.71150845e-02 -9.08970237e-01
-9.41192746e-01 -3.18840563e-01 -6.81742191e-01 -4.05788839e-01
1.22625124e+00 -5.05144000e-01 -1.10910785e+00 5.63588381e-01
-7.48653471e-01 -3.01398784e-01 -1.31936282e-01 7.64350772e-01
-5.99357128e-01 1.30782157e-01 -4.98754919e-01 -1.20096755e+00
-2.89711446e-01 -1.30167770e+00 1.10779941e+00 -3.67442146e-02
-1.13331914e+00 -6.97171032e-01 4.36073840e-01 9.20815110e-01
1.75641179e-01 9.90928710e-02 1.24996698e+00 -1.48345160e+00
8.44238579e-01 -6.20186925e-01 -2.44532883e-01 4.70729172e-01
-1.65266350e-01 1.30867526e-01 -9.82233882e-01 -5.65693602e-02
-3.00074279e-01 -7.39523768e-01 5.92823446e-01 -1.70391858e-01
1.40490901e+00 -1.48996010e-01 7.70957991e-02 3.05395335e-01
5.94146132e-01 3.49830598e-01 5.04587531e-01 4.54925597e-01
6.03307366e-01 9.02789652e-01 4.56774354e-01 6.18361771e-01
3.88201147e-01 1.28690410e+00 3.98935407e-01 6.00019217e-01
1.67397365e-01 -7.28528351e-02 4.40514058e-01 4.30144161e-01
-4.36260998e-01 2.23412916e-01 -8.37515652e-01 6.24802172e-01
-1.57163179e+00 -1.25970805e+00 -4.36068118e-01 2.18301821e+00
6.05785191e-01 2.55846739e-01 9.90296364e-01 5.97940743e-01
4.16287214e-01 4.58117351e-02 -1.25051811e-01 -1.39593756e+00
2.60903150e-01 4.92059737e-01 -1.08257122e-01 6.45099998e-01
-1.01686144e+00 9.57671463e-01 5.40363455e+00 1.05491769e+00
-7.39525020e-01 5.45938313e-02 3.59789699e-01 -5.22148192e-01
3.21673363e-01 -3.72585744e-01 -7.98152626e-01 7.03009069e-01
1.12497950e+00 -2.08631769e-01 2.96888351e-01 7.90971994e-01
-2.10390046e-01 -1.64270222e-01 -1.24435890e+00 9.97869551e-01
8.08181614e-02 -5.19835830e-01 -1.72751337e-01 4.56514865e-01
2.11775512e-01 -4.19859767e-01 3.04348767e-01 9.17479217e-01
3.83645743e-01 -1.46608472e+00 8.20261061e-01 3.15534115e-01
8.12213004e-01 -9.70204890e-01 9.54718769e-01 4.41470742e-01
-6.15255058e-01 -5.13169408e-01 -2.17072248e-01 -6.21025085e-01
-1.24897599e-01 1.45727322e-01 -3.50685686e-01 4.94715124e-01
1.10738075e+00 2.05467045e-01 -4.67911988e-01 7.26355135e-01
-2.53255397e-01 6.99020207e-01 -2.70427972e-01 -2.43636906e-01
4.34359640e-01 -2.10903108e-01 8.63534689e-01 1.23117816e+00
9.09946412e-02 2.02248752e-01 -1.24533430e-01 2.41970479e-01
6.35231063e-02 2.54682481e-01 -4.84220088e-01 -1.34205118e-01
1.56187072e-01 1.30846441e+00 -1.69334665e-01 -9.37710926e-02
9.65499133e-02 6.68030322e-01 9.57342088e-01 -2.85282552e-01
-8.45465720e-01 -6.27676785e-01 8.97554338e-01 7.06688687e-02
-1.95714101e-01 2.19018180e-02 -3.75919878e-01 -5.62768519e-01
-2.75160521e-01 -7.58963287e-01 5.95774055e-01 -4.52763349e-01
-1.61841893e+00 8.55420351e-01 2.57972598e-01 -9.02745128e-01
-2.36861691e-01 -9.51209366e-01 -6.30851924e-01 7.32481420e-01
-3.88973236e-01 -9.22352076e-01 1.79326609e-02 3.38639528e-01
3.86028290e-01 -7.86939025e-01 1.03813374e+00 5.25382638e-01
-9.74507987e-01 9.34132814e-01 -2.49075711e-01 2.43340939e-01
5.59771657e-01 -1.25440240e+00 -2.69909710e-01 7.31110647e-02
-1.92953661e-01 5.47923267e-01 8.22990060e-01 -3.25883478e-01
-5.34609079e-01 -8.02725196e-01 1.14847648e+00 -6.63724124e-01
1.23732054e+00 -3.40061337e-01 -5.73336840e-01 5.40966928e-01
-1.07655950e-01 -5.06418765e-01 1.23490286e+00 4.52817500e-01
-5.60471416e-01 2.75252074e-01 -1.30374908e+00 6.99110687e-01
1.20785463e+00 -5.03044307e-01 -8.12612295e-01 5.70688903e-01
-8.47792774e-02 -4.47747827e-01 -1.02398026e+00 1.39968827e-01
9.42288101e-01 -1.35222590e+00 8.18687439e-01 -1.23128104e+00
1.14339316e+00 3.44510615e-01 -3.75305474e-01 -1.30395603e+00
-5.36584556e-01 -1.33594334e-01 4.68726248e-01 9.50381100e-01
4.18273300e-01 -6.45352185e-01 4.73019183e-01 5.53101301e-01
-3.82596701e-01 -8.53510380e-01 -1.69187081e+00 -7.00152576e-01
6.37079120e-01 -5.87487638e-01 1.98311672e-01 9.78271365e-01
4.38353032e-01 7.27635846e-02 -4.50259209e-01 -1.84491351e-01
4.69801694e-01 -5.59710920e-01 9.36720133e-01 -1.30643523e+00
-2.27395609e-01 -1.04976320e+00 -9.73363221e-01 -8.52941945e-02
5.64558685e-01 -1.05368400e+00 -5.94194293e-01 -1.17682171e+00
3.89426380e-01 -2.66078711e-01 -2.69423336e-01 5.65727711e-01
-8.52959752e-02 5.89980602e-01 6.56031370e-02 -1.64294928e-01
-3.00907850e-01 3.49127412e-01 5.41682363e-01 9.50274691e-02
1.99926436e-01 1.20346844e-01 -7.57906020e-01 9.17595267e-01
7.69137740e-01 -6.77172899e-01 9.38788429e-02 2.71265232e-03
2.56418735e-01 -6.18136227e-02 2.42182940e-01 -1.07241356e+00
-1.24736279e-01 -2.73361534e-01 1.66408196e-01 -3.09111588e-02
9.74532068e-01 -4.55231071e-01 -6.24977536e-02 4.37078089e-01
-4.22309190e-01 9.02720541e-02 1.94874227e-01 -1.05773941e-01
-1.72388807e-01 -9.40335214e-01 7.67685831e-01 1.83340788e-01
-9.44011807e-02 -3.62763815e-02 -7.10502505e-01 2.25817487e-01
1.26028752e+00 -1.64855257e-01 -4.53491569e-01 -3.19170922e-01
-1.05012250e+00 1.01045571e-01 1.64072663e-01 5.05576670e-01
4.54999387e-01 -1.02326095e+00 -9.82738793e-01 8.39077905e-02
3.93231928e-01 -7.70023346e-01 4.44683701e-01 9.78833556e-01
-4.20507789e-01 1.85072049e-01 -5.98061323e-01 -2.56717708e-02
-1.68525696e+00 -9.30143446e-02 2.37245768e-01 -4.77659941e-01
-1.55921906e-01 1.30165482e+00 -2.41010878e-02 -7.38801777e-01
2.07396969e-03 6.56203687e-01 -5.02180576e-01 3.81000519e-01
5.68568587e-01 1.03724146e+00 6.20259196e-02 -1.46313286e+00
-4.31998938e-01 5.87990046e-01 -4.31597799e-01 -4.01575625e-01
1.41787243e+00 6.19943500e-01 -3.70013863e-01 5.66296697e-01
1.28451467e+00 4.40670937e-01 -2.38062859e-01 4.31239039e-01
-6.94769397e-02 -5.62228560e-01 -6.56430572e-02 -7.98281133e-01
-6.89849854e-01 9.36841369e-01 2.33140886e-01 3.99128914e-01
5.04620552e-01 -3.41738686e-02 3.36915553e-01 2.15688825e-01
3.45492572e-01 -1.27642918e+00 1.43019870e-01 8.57599080e-01
1.21577287e+00 -7.70196974e-01 2.01253250e-01 -6.70302212e-02
-9.23320591e-01 1.06378829e+00 9.01982069e-01 -1.80234477e-01
2.69665629e-01 -9.48573947e-02 -4.65242937e-02 -5.42867243e-01
-7.28761315e-01 -9.59436670e-02 1.58735275e-01 4.87728328e-01
3.56061518e-01 3.88522118e-01 -7.34531164e-01 1.10500574e+00
-1.12811565e+00 -5.61891139e-01 2.79651642e-01 5.08743644e-01
-1.14840113e-01 -9.46658611e-01 -2.39517197e-01 9.29783821e-01
-6.66636407e-01 3.68898630e-01 -9.52752531e-01 8.76339793e-01
3.53859484e-01 1.03942955e+00 2.52455384e-01 -9.95972395e-01
5.37873387e-01 2.45539531e-01 4.99911934e-01 -6.64134622e-01
-1.25118935e+00 -5.69098413e-01 7.49424696e-01 -4.04126197e-01
3.51125032e-01 -8.29118013e-01 -9.69451070e-01 -4.22356874e-01
9.92879346e-02 2.49350056e-01 4.88714308e-01 9.24058914e-01
-4.65289235e-01 4.17237043e-01 5.23511052e-01 -8.19329321e-01
-7.93009341e-01 -1.22645128e+00 -9.24312830e-01 7.38911450e-01
-9.07411873e-02 -1.11970031e+00 -9.90013003e-01 -9.05890882e-01] | [8.809158325195312, 10.758642196655273] |
2c707316-e46e-4a36-9070-efda0764a716 | bayesian-inference-for-the-mixed-conditional | null | null | https://academic.oup.com/ectj/article-abstract/10/2/408/5062603?login=false | https://academic.oup.com/ectj/article-abstract/10/2/408/5062603?login=false | Bayesian inference for the mixed conditional heteroskedasticity model | We estimate by Bayesian inference the mixed conditional heteroskedasticity model of Haas et al. (2004a Journal of Financial Econometrics 2, 211–50). We construct a Gibbs sampler algorithm to compute posterior and predictive densities. The number of mixture components is selected by the marginal likelihood criterion.We apply the model to the SP500 daily returns. | ['L. BAUWENS and J.V.K. ROMBOUTS'] | 2007-02-01 | null | null | null | econometrics-journal-2007-2 | ['econometrics'] | ['miscellaneous'] | [-5.43682814e-01 -6.45091459e-02 -2.06379369e-01 -3.37042630e-01
-8.96172106e-01 -4.80894893e-01 1.00094140e+00 -5.85944235e-01
-2.64486849e-01 1.10522592e+00 8.37177038e-02 -7.47015774e-01
-9.78154615e-02 -9.31336522e-01 -2.73515910e-01 -6.64588511e-01
-2.33019561e-01 5.92901111e-01 3.79634239e-02 5.69326520e-01
4.90216047e-01 2.36964062e-01 -7.93214679e-01 -2.59763122e-01
7.19356477e-01 7.15290070e-01 1.18070394e-01 9.14892912e-01
1.05992697e-01 1.00265622e+00 -2.67881602e-01 -5.93488395e-01
2.13120133e-01 -6.14384115e-01 -5.17974377e-01 1.61589175e-01
-5.94453096e-01 -5.07826149e-01 1.04792796e-01 1.17118204e+00
-1.28712952e-01 2.75649488e-01 1.45790279e+00 -8.71993780e-01
-3.66519302e-01 5.76332748e-01 -6.35587096e-01 6.94333911e-01
-1.00668326e-01 4.52869684e-02 1.03190017e+00 -1.19731665e+00
2.20195562e-01 1.41027117e+00 3.25338423e-01 -2.80301541e-01
-1.58584321e+00 -6.11346960e-01 -3.27046841e-01 -1.08150408e-01
-1.48418891e+00 -3.77253562e-01 4.93560791e-01 -7.80800819e-01
8.76881480e-01 -1.80793554e-01 7.36062884e-01 9.08253551e-01
5.82559168e-01 7.14028597e-01 1.38558245e+00 -4.44227368e-01
6.82160616e-01 7.51288310e-02 2.20411271e-01 2.82102060e-02
5.56482732e-01 1.57942623e-01 -2.14730814e-01 -7.35599518e-01
1.11876678e+00 -7.02850968e-02 3.34882140e-01 -1.27778456e-01
-7.31122732e-01 1.48925257e+00 -6.21557176e-01 2.35607326e-01
-8.50316346e-01 4.14951831e-01 -1.29985005e-01 2.60552734e-01
6.46916509e-01 -1.65777817e-01 -5.64846218e-01 -2.50408024e-01
-1.13435614e+00 5.26274920e-01 1.14673710e+00 6.84113801e-01
8.55103016e-01 2.48498276e-01 8.28221813e-02 6.71173632e-01
9.06515837e-01 1.08342493e+00 3.74526441e-01 -1.32929826e+00
2.47248355e-03 -3.65157962e-01 4.36406076e-01 -5.51398277e-01
1.49515226e-01 -7.06648007e-02 -5.03489912e-01 8.76138359e-02
4.55058426e-01 -4.23112988e-01 -5.38427770e-01 1.33656001e+00
-5.75979566e-03 2.58523196e-01 3.17183375e-01 3.71975303e-01
-1.06108360e-01 9.18788791e-01 8.09660330e-02 -7.68528461e-01
1.10581183e+00 -4.38224733e-01 -1.05263925e+00 4.45581339e-02
-4.39216942e-02 -9.04791534e-01 2.63514459e-01 4.54416156e-01
-1.25600004e+00 -2.40582332e-01 -4.22569394e-01 6.56431437e-01
1.06982499e-01 -2.92006016e-01 6.29259765e-01 8.71558428e-01
-9.97998536e-01 6.96925819e-01 -1.35895360e+00 1.06532440e-01
-3.40188593e-02 3.68079245e-02 4.05735821e-01 3.97533536e-01
-1.02951360e+00 8.04154098e-01 5.94970644e-01 -1.17208034e-01
-8.78707767e-01 -1.32844865e-01 -6.74800873e-01 6.14936016e-02
-4.49019596e-02 -5.18661976e-01 1.70568371e+00 -8.31463397e-01
-1.79466498e+00 5.10525644e-01 -1.65068075e-01 -6.57595038e-01
2.94293731e-01 -2.57317424e-02 -3.85096639e-01 1.86386451e-01
1.77854329e-01 -1.64328650e-01 1.10603368e+00 -7.67073870e-01
-5.63706398e-01 -1.75219849e-01 -8.21893036e-01 -1.93871126e-01
3.64606261e-01 4.23953414e-01 -1.50947571e-01 -8.07044864e-01
3.31950545e-01 -8.48358393e-01 -2.86135256e-01 -1.14339852e+00
-1.31682441e-01 -2.92330295e-01 1.42873809e-01 -9.76726532e-01
1.22353101e+00 -2.11266112e+00 -3.33783537e-01 8.71869981e-01
-3.26196492e-01 -8.59269440e-01 4.11516130e-01 3.57178420e-01
-3.17032456e-01 -2.14124452e-02 -5.28315127e-01 -2.55927086e-01
2.44645283e-01 2.50163525e-01 -4.60109085e-01 6.63969278e-01
-5.39467717e-03 5.30960023e-01 -4.95223373e-01 -3.17580551e-01
-4.47473675e-03 4.75600213e-02 -4.81165051e-01 2.33280763e-01
2.50464886e-01 -3.48289683e-02 -5.24447739e-01 4.82993245e-01
8.69385064e-01 -4.55241323e-01 6.41978264e-01 4.66694027e-01
-4.90029961e-01 5.07212102e-01 -1.66219938e+00 7.36051798e-01
1.04051247e-01 3.61034602e-01 -3.19909826e-02 -9.74142194e-01
7.80943215e-01 5.78426182e-01 3.39555323e-01 -1.38824299e-01
-8.05458240e-03 2.55395204e-01 -1.60962105e-01 -3.19237411e-01
3.96290898e-01 -6.27820253e-01 -2.07787961e-01 5.56151390e-01
1.14114605e-01 -4.24259514e-01 2.78672367e-01 1.11631513e-01
8.96041214e-01 6.50530010e-02 7.04444885e-01 -8.11294913e-01
-8.22989047e-02 -1.44868642e-01 4.91514951e-01 1.17246306e+00
6.50423467e-02 2.82430381e-01 8.78262043e-01 2.55229194e-02
-9.48384583e-01 -1.52472448e+00 -6.74078226e-01 8.79318774e-01
-8.77647758e-01 9.02607292e-02 -5.14023423e-01 8.39156210e-02
2.59405881e-01 1.02542412e+00 -3.66222262e-01 3.57545912e-01
-1.06089540e-01 -1.40925646e+00 9.67561901e-02 3.11504692e-01
1.87680796e-01 -1.03189278e+00 -6.23609662e-01 3.58131528e-01
1.28797114e-01 -6.17411613e-01 4.58676554e-02 3.90216023e-01
-1.05246854e+00 -9.97514307e-01 -8.32408011e-01 -2.54475981e-01
1.33045197e-01 -2.57353604e-01 1.23252010e+00 -6.60646200e-01
3.00073385e-01 4.22993869e-01 -5.75335883e-02 -1.20646916e-01
-4.97784615e-01 -5.33696413e-01 -8.66191229e-04 1.58831969e-01
5.11677504e-01 -4.85553116e-01 -2.97723055e-01 -9.71918851e-02
-7.49121666e-01 -3.33913147e-01 3.98003787e-01 8.13310683e-01
4.35495764e-01 2.42022038e-01 5.30994534e-01 -7.41480112e-01
5.99460602e-01 -7.38889992e-01 -1.24532890e+00 -3.20346393e-02
-6.61722064e-01 -1.72409207e-01 -1.43539831e-01 -3.16249847e-01
-1.59788096e+00 -3.97689611e-01 3.35227698e-01 -5.74016385e-02
-2.56421685e-01 1.01858819e+00 2.73373395e-01 6.20524228e-01
-1.40847743e-01 6.77904189e-02 -1.16557442e-01 -6.85921907e-01
4.14492153e-02 5.33385038e-01 4.11562383e-01 -4.70582366e-01
4.76653904e-01 4.34845179e-01 -5.77275902e-02 -6.47791862e-01
-5.63034475e-01 -2.33198166e-01 -2.84683079e-01 -4.95943613e-02
1.12761271e+00 -1.19559681e+00 -5.31441808e-01 5.48078120e-01
-9.79056895e-01 -5.14868855e-01 -2.92612016e-01 1.29777789e+00
-8.33989620e-01 1.45734236e-01 -1.20666456e+00 -1.61192369e+00
2.31142282e-01 -8.23495567e-01 4.59911942e-01 1.61532134e-01
-2.80055761e-01 -1.15741122e+00 6.89115822e-01 -2.37393558e-01
8.86069685e-02 1.23473272e-01 6.75733805e-01 -7.80306578e-01
-3.66111785e-01 -2.37466395e-01 -4.81839515e-02 2.00796768e-01
1.04504891e-01 5.88200808e-01 -7.68180132e-01 -6.24247193e-02
5.21966696e-01 7.49795288e-02 8.67432952e-01 1.11582673e+00
7.26275086e-01 -3.48462641e-01 2.81617083e-02 1.88594222e-01
1.57388937e+00 6.67525232e-01 6.51593506e-01 2.70533890e-01
-1.97999179e-01 6.07980847e-01 3.98865163e-01 1.04413271e+00
3.15162301e-01 1.04709506e-01 -1.24174967e-01 5.70642769e-01
9.17907417e-01 -1.42373189e-01 3.65245879e-01 9.66603398e-01
-1.00497521e-01 -6.89745247e-02 -1.17226374e+00 7.03715026e-01
-1.84908247e+00 -1.31816590e+00 -5.33790350e-01 2.14767480e+00
9.25567389e-01 3.67315024e-01 4.60984558e-01 -1.76646188e-01
8.62813652e-01 -7.04028904e-02 1.56436801e-01 -2.20969871e-01
-1.61819905e-01 5.76604187e-01 6.45568728e-01 8.37823689e-01
-9.41102982e-01 7.17909575e-01 8.17473984e+00 7.87275732e-01
-1.10142559e-01 -4.24451940e-02 8.53827238e-01 -7.52727687e-02
-3.49418700e-01 6.45559669e-01 -7.93633759e-01 8.64502072e-01
1.42705977e+00 -4.03260499e-01 3.16856503e-01 8.01957667e-01
3.39762419e-01 -8.59824896e-01 -6.02906466e-01 5.66789627e-01
-2.21449926e-01 -8.11239123e-01 -4.45224226e-01 5.96808732e-01
1.06091535e+00 -8.16865638e-02 2.18145788e-01 3.79197776e-01
1.32859612e+00 -8.49212408e-01 6.62236154e-01 1.04940271e+00
-4.04414460e-02 -1.21163070e+00 8.28024209e-01 3.58795583e-01
-7.80787230e-01 -3.48443463e-02 -5.30603290e-01 -2.96749532e-01
4.88914818e-01 1.02641237e+00 -5.36651194e-01 4.21257019e-02
8.49681616e-01 4.67580050e-01 -1.68750778e-01 9.18416679e-01
-4.17889282e-02 1.26080132e+00 -4.70611304e-01 1.73707288e-02
1.76793665e-01 -1.07932472e+00 2.14863226e-01 1.03016007e+00
4.43805546e-01 3.22052211e-01 -2.17587262e-01 1.12479126e+00
4.56312656e-01 -1.02175534e-01 -4.28866506e-01 -9.07805637e-02
5.14821053e-01 6.88559949e-01 -1.08971572e+00 -8.78679991e-01
-7.71559119e-01 3.90118092e-01 -1.73620969e-01 7.92535067e-01
-4.82334465e-01 -5.72862960e-02 1.27569035e-01 -2.75108963e-01
7.83082068e-01 -3.97725999e-01 -3.04610878e-01 -1.27709031e+00
-4.87002701e-01 -7.35752344e-01 6.09466672e-01 -5.70293546e-01
-1.59424663e+00 -2.18388006e-01 5.79169810e-01 -6.57718241e-01
-7.66834497e-01 -6.44285917e-01 -9.02799189e-01 1.23278379e+00
-1.17110646e+00 -2.70710289e-01 7.18696773e-01 3.44264627e-01
2.35141382e-01 -2.01780483e-01 5.35040617e-01 -2.85096198e-01
-5.81057489e-01 -4.59897846e-01 8.73988032e-01 1.83411196e-01
2.97351241e-01 -1.45791256e+00 5.79053938e-01 7.31588960e-01
-2.06768408e-01 7.73738325e-01 8.32888901e-01 -1.07176495e+00
-8.41020107e-01 -3.49626601e-01 9.50238526e-01 -2.16829672e-01
1.15701985e+00 5.28517738e-02 -7.87953436e-01 1.06191874e+00
6.38391256e-01 -5.02103567e-01 8.32204700e-01 2.33400837e-02
9.38629657e-02 4.49303299e-01 -9.17993009e-01 2.20941141e-01
-2.94571877e-01 -5.11180997e-01 -7.87716150e-01 1.65658846e-01
2.80351758e-01 3.65716875e-01 -1.01359367e+00 2.64034212e-01
5.37684023e-01 -1.17812252e+00 6.93301082e-01 -3.81940961e-01
1.44813299e-01 1.02092057e-01 -4.48452830e-01 -8.90631557e-01
-7.14724958e-01 -7.69750535e-01 -2.60320362e-02 1.23417473e+00
3.47350091e-01 -9.14964437e-01 3.01939815e-01 7.98545182e-01
5.71299314e-01 -3.77131626e-02 -1.22557199e+00 -9.76248443e-01
5.15973330e-01 -6.08436584e-01 4.92909551e-01 9.29152369e-01
1.30632505e-01 -1.61366776e-01 -4.54571247e-01 -1.65562123e-01
1.05874324e+00 2.35978395e-01 4.78492588e-01 -1.40481675e+00
-8.63908589e-01 -5.16349494e-01 4.39015895e-01 -1.03870416e+00
3.55294973e-01 -2.63221294e-01 1.10995680e-01 -9.41957474e-01
5.13653040e-01 3.14472318e-01 -2.07090497e-01 -1.74497947e-01
7.67207444e-02 -1.65535405e-01 -2.25611478e-01 2.59682298e-01
-3.52005035e-01 5.42423546e-01 4.92083669e-01 3.92181784e-01
-8.38366076e-02 2.76051670e-01 -3.96025106e-02 9.53278840e-01
8.06078613e-01 -8.45185518e-01 1.12405814e-01 3.15186411e-01
3.45516264e-01 9.48431849e-01 5.76562881e-01 -2.45859444e-01
-1.60063252e-01 -7.37481475e-01 4.40061539e-01 -9.93253469e-01
1.73151016e-01 -4.56526756e-01 5.56249857e-01 5.10077536e-01
-8.49170536e-02 2.96072632e-01 -1.52368098e-01 7.24770784e-01
-3.33335996e-01 -8.13380957e-01 8.75162005e-01 -3.34376931e-01
1.95280910e-01 -1.60122037e-01 -1.42420435e+00 1.07012816e-01
6.86751366e-01 2.02084675e-01 3.56759250e-01 -5.66754937e-01
-9.36782539e-01 -4.95006181e-02 3.11042190e-01 -4.72485006e-01
4.18150216e-01 -1.38292336e+00 -7.48702288e-01 1.67346865e-01
-5.97720563e-01 -5.55913746e-01 -1.02816276e-01 1.10199952e+00
-4.92405355e-01 4.85402852e-01 2.79192507e-01 -2.23437786e-01
-4.77220953e-01 4.34307516e-01 2.51011014e-01 -3.49315822e-01
-2.76683897e-01 5.12810826e-01 6.50416687e-02 1.77884847e-01
-3.75094891e-01 -1.01526238e-01 -2.39548579e-01 1.97916687e-01
4.25551534e-01 8.86925936e-01 -5.88407755e-01 -3.44370723e-01
-2.96835482e-01 1.14127427e-01 2.86530644e-01 -1.09987783e+00
1.35482264e+00 -3.21672201e-01 -5.24180532e-01 9.99007940e-01
9.44288015e-01 -9.21012014e-02 -1.22760820e+00 -2.94245910e-02
6.68759584e-01 -4.45648611e-01 3.20794508e-02 3.55719924e-02
-5.75266421e-01 4.88276213e-01 1.54142722e-01 7.93689489e-01
6.85020685e-01 -6.34710863e-02 9.39222574e-02 9.69891697e-02
1.88589748e-02 -1.44166887e+00 -1.81327611e-01 4.97780681e-01
4.05470848e-01 -9.21409190e-01 1.37423888e-01 1.78157672e-01
-7.63063252e-01 1.09265268e+00 -1.46127909e-01 -7.86603570e-01
1.48126364e+00 3.66864949e-01 -3.45805228e-01 -1.58677205e-01
-9.85785365e-01 2.21337210e-02 2.41611749e-01 1.09323010e-01
4.63826180e-01 3.47911745e-01 -5.68938494e-01 7.74304330e-01
-1.75553828e-01 -1.51375771e-01 6.76936269e-01 8.76271665e-01
-5.95200658e-01 -7.69136667e-01 -7.53006995e-01 6.29169762e-01
-1.01862085e+00 -2.46830732e-01 6.94482476e-02 6.15369558e-01
-6.22330368e-01 1.19626915e+00 6.95189178e-01 4.70849395e-01
-2.06132576e-01 5.73404551e-01 1.07554689e-01 -4.23038244e-01
-3.16832438e-02 1.28095305e+00 -1.57277748e-01 4.49353307e-02
-6.23846531e-01 -1.53989649e+00 -6.89947367e-01 -3.67473036e-01
-5.00901937e-01 4.80532020e-01 2.55793482e-01 1.00862467e+00
-3.88190389e-01 1.06934220e-01 8.55072677e-01 -4.63839322e-01
-9.22100902e-01 -1.28095925e+00 -1.48942804e+00 -1.51232928e-01
-7.80049304e-04 -4.36340690e-01 -1.01611412e+00 3.32588255e-01] | [6.26272439956665, 4.0137104988098145] |
f323661b-cfd3-4a3f-a78f-241d83cf6a93 | distributed-cpu-scheduling-subject-to | 2208.14059 | null | https://arxiv.org/abs/2208.14059v1 | https://arxiv.org/pdf/2208.14059v1.pdf | Distributed CPU Scheduling Subject to Nonlinear Constraints | This paper considers a network of collaborating agents for local resource allocation subject to nonlinear model constraints. In many applications, it is required (or desirable) that the solution be anytime feasible in terms of satisfying the sum-preserving global constraint. Motivated by this, sufficient conditions on the nonlinear mapping for anytime feasibility (or non-asymptotic feasibility) are addressed in this paper. For the two proposed distributed solutions, one converges over directed weight-balanced networks and the other one over undirected networks. In particular, we elaborate on uniform quantization and discuss the notion of {\epsilon}-accurate solution, which gives an estimate of how close we can get to the exact optimizer subject to different quantization levels. This work, further, handles general (possibly non-quadratic) strictly convex objective functions with application to CPU allocation among a cloud of data centers via distributed solutions. The results can be used as a coordination mechanism to optimally balance the tasks and CPU resources among a group of networked servers while addressing quantization or limited server capacity. Index Terms: multi-agent systems, sum-preserving resource allocation, distributed optimization, anytime feasibility | ['Themistoklis Charalambous', 'Karl H. Johansson', 'Christoforos N. Hadjicostis', 'Evangelia Kalyvianaki', 'Andreas Grammenos', 'Apostolos I. Rikos', 'Alireza Aghasi', 'Mohammadreza Doostmohammadian'] | 2022-08-30 | null | null | null | null | ['distributed-optimization'] | ['methodology'] | [-2.47024968e-01 2.31049061e-01 -1.76814139e-01 -1.94128662e-01
-3.78700316e-01 -6.65856183e-01 -2.30608672e-01 1.10783130e-01
-1.61977693e-01 1.07458365e+00 -3.69037449e-01 -1.36941597e-01
-9.99770403e-01 -6.48027182e-01 -2.45726094e-01 -1.18916571e+00
-2.80930042e-01 9.83294129e-01 -2.76878953e-01 -1.48383588e-01
3.47967625e-01 7.57258952e-01 -1.25436532e+00 -4.14248139e-01
7.43660748e-01 1.23841274e+00 5.01849949e-01 7.47769415e-01
8.48607421e-02 6.29854798e-01 -7.26825058e-01 -4.83493060e-02
7.60383904e-01 -1.66901931e-01 -8.15351069e-01 4.80682045e-01
1.61602333e-01 -1.74106643e-01 1.81286439e-01 1.40001309e+00
6.06810451e-01 3.36416423e-01 5.64879537e-01 -1.97452176e+00
-8.24667156e-01 5.01621902e-01 -4.86896694e-01 3.54248255e-01
-3.64408761e-01 -8.06470215e-02 9.36140239e-01 -4.41005290e-01
4.84337419e-01 8.84026885e-01 2.12941647e-01 3.27097565e-01
-1.01917803e+00 -2.29602128e-01 4.83686291e-02 4.13325578e-01
-1.48128879e+00 -4.39104766e-01 4.72067803e-01 -1.85176298e-01
8.64606202e-01 5.55942714e-01 2.57006347e-01 -1.72958672e-01
1.68228686e-01 -8.64689201e-02 6.69013858e-01 -6.24233186e-01
4.67533797e-01 3.12327445e-01 -7.25685013e-03 5.70227861e-01
5.96622825e-01 -5.41059434e-01 -1.39890626e-01 -3.97148371e-01
7.47555554e-01 -2.95395702e-01 -4.17502671e-01 -4.13218886e-01
-1.01505554e+00 9.02890027e-01 1.11511551e-01 4.58624542e-01
-8.36754739e-01 -5.80755644e-04 3.51383924e-01 5.47813892e-01
6.60427332e-01 5.24821639e-01 -3.13016653e-01 2.70450264e-01
-7.69795239e-01 1.23505935e-01 1.08656693e+00 1.35249889e+00
5.93936622e-01 3.09345484e-01 -2.77550846e-01 5.81362665e-01
1.61270052e-01 6.57050252e-01 -1.07665598e-01 -1.55608439e+00
4.73335981e-01 -2.19007824e-02 4.92409885e-01 -1.19825613e+00
-5.24759173e-01 -4.95465666e-01 -1.02897441e+00 1.29658461e-01
2.99166650e-01 -3.89189452e-01 1.83989927e-01 1.77207828e+00
4.55150843e-01 -1.02153212e-01 -8.80564228e-02 1.37298799e+00
-1.58558145e-01 9.24524903e-01 -4.21306461e-01 -1.20552886e+00
1.12333477e+00 -1.28313696e+00 -1.14208448e+00 5.40030956e-01
3.84172350e-01 -8.54036510e-01 3.88186544e-01 3.79214346e-01
-1.47659373e+00 7.44220689e-02 -8.27719033e-01 3.65655988e-01
-1.69852376e-01 -1.67923674e-01 1.69938326e-01 6.06575191e-01
-1.76375544e+00 3.80955905e-01 -3.34696174e-01 -4.52336997e-01
-2.44297400e-01 6.53782904e-01 7.47062415e-02 -5.85147701e-02
-9.87803638e-01 9.75498021e-01 2.30292991e-01 4.14106876e-01
-5.34244180e-01 -5.39441526e-01 -1.78283080e-01 4.13252562e-01
6.11547172e-01 -8.39616537e-01 9.76581216e-01 -8.53646815e-01
-1.56762421e+00 3.48961890e-01 1.66319862e-01 -1.82942882e-01
5.14416337e-01 7.29456186e-01 -4.65606228e-02 4.15371805e-01
1.61289200e-01 -1.05302846e-02 6.75153315e-01 -1.08116019e+00
-5.60602069e-01 -4.92847204e-01 2.55105555e-01 6.83003724e-01
-9.69453573e-01 2.17234507e-01 -1.70065790e-01 -2.45807022e-01
-3.81762981e-01 -8.64482701e-01 -2.96149582e-01 9.11853462e-02
-3.79529059e-01 -6.03758752e-01 8.15995693e-01 -4.70004737e-01
1.11874723e+00 -1.78837788e+00 4.74113554e-01 4.92409319e-01
2.84311593e-01 -1.25481546e-01 -1.97523475e-01 6.38993204e-01
2.45552108e-01 1.18793428e-01 2.06152126e-01 -4.81591851e-01
4.16193902e-01 3.21440101e-01 2.10972577e-01 1.05935943e+00
-2.47800589e-01 2.52112538e-01 -6.89357877e-01 -5.36237359e-01
1.95906132e-01 2.15320140e-01 -3.54702324e-01 2.21724167e-01
-9.64537635e-02 2.89807498e-01 -6.35244906e-01 4.84042197e-01
7.69952834e-01 -4.33420807e-01 5.39335668e-01 -2.10079834e-01
-3.02862436e-01 -6.38798654e-01 -1.39625609e+00 1.20450139e+00
-6.51719689e-01 5.96024156e-01 1.31797206e+00 -1.49602771e+00
6.60282493e-01 6.09521925e-01 1.04108322e+00 -3.94410312e-01
2.17689097e-01 2.48966336e-01 -1.89972863e-01 -5.22504807e-01
5.11151731e-01 -2.90449505e-04 2.96198517e-01 7.40783572e-01
-2.37141877e-01 2.21543796e-02 5.46483457e-01 6.79646358e-02
8.38177383e-01 -8.91582251e-01 5.58751822e-02 -1.03646410e+00
6.86302364e-01 -2.37860188e-01 4.43687737e-01 5.70203125e-01
-5.85376084e-01 5.24124503e-02 6.47832930e-01 -1.96163610e-01
-1.45530415e+00 -4.84354675e-01 -1.55541068e-02 1.20158267e+00
3.83974314e-01 3.14591229e-01 -8.67477596e-01 -2.22961623e-02
-6.72025606e-03 5.39566100e-01 -2.71793664e-01 1.80320725e-01
-4.28738624e-01 -6.85199738e-01 4.03412245e-02 -1.41821265e-01
2.94800639e-01 -6.60693347e-01 -4.77884293e-01 3.75764340e-01
-1.74286500e-01 -1.25926852e+00 -1.03639233e+00 1.82775483e-01
-5.18372715e-01 -9.92077589e-01 -1.02213466e+00 -9.04291630e-01
9.24978256e-01 5.17756402e-01 9.82609570e-01 1.26826301e-01
-6.27165437e-02 8.59150290e-01 5.37578389e-02 -8.00813660e-02
-5.40649407e-02 7.00142384e-02 5.57391346e-01 2.70211846e-01
-2.19209433e-01 -4.07929063e-01 -5.47010422e-01 5.32528400e-01
-8.91577482e-01 -3.28176767e-01 1.38281405e-01 3.77278477e-01
6.88029289e-01 5.59392154e-01 8.94849181e-01 -2.77062625e-01
1.11889684e+00 -6.91701889e-01 -9.87334788e-01 6.37539208e-01
-8.89767170e-01 -7.33359903e-02 1.30392742e+00 -1.46929994e-01
-5.71042717e-01 -3.77613723e-01 8.35184395e-01 -7.29301810e-01
3.63130748e-01 1.13000743e-01 -1.55755088e-01 -4.38019276e-01
1.48266226e-01 7.62014324e-03 3.22094887e-01 1.29274174e-03
4.57867742e-01 8.18304777e-01 1.23623431e-01 -8.65770698e-01
3.36917698e-01 3.16075265e-01 5.39900899e-01 -7.87895083e-01
-5.07241726e-01 -4.75311488e-01 -3.10794324e-01 -5.59006929e-01
8.47864270e-01 -2.87908792e-01 -1.49128437e+00 1.65491015e-01
-1.35227394e+00 -1.48620337e-01 -2.73253679e-01 3.67005080e-01
-9.63716328e-01 2.77644634e-01 -4.83682722e-01 -1.15329683e+00
-5.21318853e-01 -1.11068177e+00 5.19157350e-01 2.54902601e-01
4.77362543e-01 -1.37644994e+00 -2.83911899e-02 3.75407755e-01
1.01169121e+00 8.22559446e-02 7.15137362e-01 -4.59294736e-01
-6.84401989e-01 1.54457092e-01 -3.53543788e-01 4.13451105e-01
2.18184203e-01 7.45344684e-02 -1.78179264e-01 -8.12300205e-01
2.46451840e-01 -1.81903020e-01 -9.06808153e-02 6.78277075e-01
1.06898618e+00 -9.73858654e-01 -4.29048296e-03 2.88546860e-01
1.86679220e+00 6.43773526e-02 -1.00464247e-01 2.79640138e-01
2.77792662e-01 6.61481798e-01 4.64183748e-01 9.79240656e-01
4.99941796e-01 8.11314940e-01 9.71707106e-01 2.36634105e-01
2.92939305e-01 8.94583344e-01 -1.01533916e-03 1.26414359e+00
-2.77655393e-01 -8.28992248e-01 -7.00853646e-01 6.86751902e-01
-1.99232519e+00 -8.31118822e-01 2.29556598e-02 2.19927216e+00
4.27776515e-01 -4.99635786e-01 2.18093008e-01 -2.17593253e-01
1.40739465e+00 -8.92728195e-02 -7.59813368e-01 -8.93704474e-01
-1.67580619e-01 -4.93177265e-01 1.02411437e+00 6.45356655e-01
-6.47630334e-01 9.90006253e-02 5.92287779e+00 8.52272868e-01
-1.00238562e+00 4.23947841e-01 6.73391223e-01 -4.95229185e-01
3.73279899e-02 -4.86870170e-01 -4.68349010e-01 6.31842911e-01
1.18566227e+00 -9.16852534e-01 1.20286286e+00 7.13822305e-01
7.99675345e-01 2.71569546e-02 -8.29354465e-01 1.01050925e+00
1.12758487e-01 -1.19821095e+00 -3.16823214e-01 3.44088525e-01
1.20093930e+00 -1.75173447e-01 1.51505873e-01 -3.84289384e-01
2.24189728e-01 -6.88769042e-01 4.40586478e-01 6.20074153e-01
4.97869492e-01 -1.07256138e+00 8.88466299e-01 4.44537967e-01
-1.15719533e+00 -2.50505686e-01 -6.30470216e-01 4.51261029e-02
4.01226968e-01 6.74418688e-01 -4.93881434e-01 9.06398356e-01
5.33020854e-01 2.17849538e-02 2.32849345e-01 9.91534829e-01
5.88761926e-01 -1.34662464e-01 -5.35971403e-01 -2.67078817e-01
3.75907034e-01 -6.03842974e-01 5.06418586e-01 7.61381865e-01
4.67919886e-01 3.51053715e-01 7.00818181e-01 7.29400277e-01
-1.31759614e-01 4.13668185e-01 -3.95732194e-01 9.29486752e-02
8.53964031e-01 1.62507749e+00 -9.26800311e-01 -3.55209976e-01
-2.76977599e-01 7.84866869e-01 3.88893992e-01 5.08339524e-01
-8.07564497e-01 -5.74369848e-01 8.74554038e-01 -2.56629676e-01
9.41717029e-02 -4.15085167e-01 -3.24532330e-01 -7.27745533e-01
1.69039309e-01 -3.46862435e-01 3.10727358e-01 -5.37417948e-01
-1.44718814e+00 5.63467205e-01 -1.01239637e-01 -8.88140678e-01
-2.03033239e-01 -3.46298695e-01 -4.71061349e-01 1.06898963e+00
-1.53110290e+00 -6.65727139e-01 -1.26615778e-01 7.51089633e-01
4.15230036e-01 -2.34475404e-01 5.19839227e-01 5.69400489e-01
-7.44695544e-01 3.04436713e-01 6.13374829e-01 -6.01788342e-01
4.41437334e-01 -1.24592710e+00 -9.71949756e-01 6.41887903e-01
-5.49217701e-01 1.84636086e-01 1.00429010e+00 -1.14594847e-01
-1.55625987e+00 -1.27503538e+00 6.73336267e-01 2.08141938e-01
8.43840301e-01 4.97230664e-02 -5.70612550e-01 3.02571714e-01
7.54925966e-01 2.62262344e-01 3.51707131e-01 -4.69691306e-01
5.18925250e-01 -2.36487836e-01 -1.52610338e+00 1.53781936e-01
5.99910855e-01 -1.57744437e-01 3.76617938e-01 1.25148618e+00
6.58679903e-01 -6.15172945e-02 -1.13428068e+00 -1.13300174e-01
-3.65284756e-02 -5.56397498e-01 6.96011961e-01 -7.80206978e-01
-1.86458871e-01 -2.17123657e-01 -3.99225056e-01 -1.44342232e+00
-6.06240988e-01 -9.14535046e-01 -9.59648117e-02 1.05794001e+00
4.65097539e-02 -7.57392943e-01 6.70238972e-01 8.36360395e-01
-1.79634392e-01 -7.18966901e-01 -1.32279634e+00 -1.27217591e+00
9.24157649e-02 2.26167351e-01 4.55051064e-01 1.28250062e+00
3.51838261e-01 -3.23525555e-02 -4.23893660e-01 5.60020387e-01
9.97395277e-01 7.20476657e-02 1.09651625e-01 -8.10838282e-01
-1.22597352e-01 -6.14936709e-01 -3.05692255e-01 -7.46019661e-01
5.20822823e-01 -8.37785482e-01 -2.44151857e-02 -1.60468686e+00
2.51322631e-02 -8.61341178e-01 -3.44127238e-01 2.57822752e-01
4.48240638e-01 5.31587228e-02 5.67156017e-01 4.78922844e-01
-1.16256821e+00 4.11063939e-01 1.32208896e+00 -6.75342083e-02
-5.49969934e-02 2.81067770e-02 -3.35753322e-01 1.28772169e-01
1.01489437e+00 -3.48204702e-01 -6.41123235e-01 -6.20974243e-01
1.63364932e-01 5.25210500e-01 -2.95473002e-02 -4.80875909e-01
7.68337965e-01 -6.48952603e-01 -5.36189675e-01 -1.43249780e-01
2.28211820e-01 -1.41823530e+00 1.53546199e-01 3.36550981e-01
-3.49046528e-01 5.02301931e-01 -3.97972733e-01 3.74269664e-01
6.28582016e-02 -6.28261626e-01 7.81021118e-01 2.50815973e-02
-2.75189191e-01 4.43003535e-01 -3.88933480e-01 2.55876966e-02
1.65929890e+00 2.13937402e-01 -5.55877030e-01 -6.76414967e-01
-9.84196246e-01 7.84103692e-01 1.83550179e-01 -2.36214492e-02
6.64850930e-03 -1.35916984e+00 -7.85745978e-01 -2.76453704e-01
-7.04626262e-01 -2.57783741e-01 2.22919643e-01 1.16423595e+00
-7.14578688e-01 8.46795321e-01 -2.31811687e-01 -4.92088586e-01
-9.76789176e-01 9.07974839e-01 5.19914508e-01 -2.39118814e-01
3.28802951e-02 4.95556951e-01 -3.21331620e-01 -2.51906186e-01
2.76979208e-01 -1.04038483e-02 1.37843847e-01 8.75286832e-02
2.45155215e-01 1.05343080e+00 2.47414008e-01 -4.27377522e-01
-3.04240853e-01 4.77375448e-01 2.72811294e-01 -8.21211040e-02
1.34066486e+00 -7.42006123e-01 -8.44881833e-01 8.76974165e-02
1.26906610e+00 -4.63620961e-01 -8.93640995e-01 -2.38191307e-01
-3.09312046e-01 -3.80221188e-01 2.27315545e-01 -3.97013515e-01
-1.71237648e+00 3.81395519e-01 4.89882737e-01 1.24498737e+00
1.22735941e+00 -1.88530892e-01 2.16589972e-01 4.10163313e-01
7.36310065e-01 -1.49347234e+00 -1.90788433e-01 3.12585026e-01
1.00684309e+00 -8.85824144e-01 -1.70325831e-01 -5.11882842e-01
-2.66839653e-01 1.53245616e+00 5.84056258e-01 -3.21251065e-01
6.46582663e-01 2.99390674e-01 -2.96674967e-01 -2.89201364e-02
-1.02343714e+00 1.24247380e-01 -2.75596529e-01 5.71842194e-01
1.40435085e-01 2.24059820e-01 -6.26157761e-01 -1.68880653e-02
1.26254961e-01 -2.48434097e-01 8.59968305e-01 5.67547619e-01
-4.96742696e-01 -6.34776592e-01 -6.80320561e-01 3.62071276e-01
-3.45365345e-01 1.71836480e-01 1.53579354e-01 5.33277512e-01
-9.59077999e-02 1.26398337e+00 3.86216283e-01 4.64509815e-01
2.78159082e-01 -2.59300679e-01 3.38858694e-01 -2.33062059e-01
-3.42187971e-01 -2.36351173e-02 7.97567517e-03 -4.48941976e-01
-5.07929742e-01 -3.75617087e-01 -9.99589145e-01 -8.81790102e-01
-6.14491999e-01 5.27220964e-01 8.41687918e-01 7.03723788e-01
4.56182480e-01 7.63100088e-01 1.07562244e+00 -6.89852059e-01
-1.14380431e+00 -7.05212951e-01 -1.08569467e+00 -2.30841279e-01
2.14348644e-01 -2.31396362e-01 -5.57923973e-01 -1.48727566e-01] | [6.180289268493652, 4.917590141296387] |
76c5c793-7fc6-46c6-886e-1816ea52e422 | camelira-an-arabic-multi-dialect | 2211.16807 | null | https://arxiv.org/abs/2211.16807v1 | https://arxiv.org/pdf/2211.16807v1.pdf | Camelira: An Arabic Multi-Dialect Morphological Disambiguator | We present Camelira, a web-based Arabic multi-dialect morphological disambiguation tool that covers four major variants of Arabic: Modern Standard Arabic, Egyptian, Gulf, and Levantine. Camelira offers a user-friendly web interface that allows researchers and language learners to explore various linguistic information, such as part-of-speech, morphological features, and lemmas. Our system also provides an option to automatically choose an appropriate dialect-specific disambiguator based on the prediction of a dialect identification component. Camelira is publicly accessible at http://camelira.camel-lab.com. | ['Nizar Habash', 'Go Inoue', 'Ossama Obeid'] | 2022-11-30 | null | null | null | null | ['dialect-identification', 'morphological-disambiguation'] | ['natural-language-processing', 'natural-language-processing'] | [-7.81279564e-01 -4.90805477e-01 -2.57781427e-02 -4.29075301e-01
-9.61223066e-01 -1.34225237e+00 3.73803347e-01 5.48648298e-01
-2.70622730e-01 5.82013249e-01 8.89787227e-02 -7.17939019e-01
2.18558963e-02 -8.95028234e-01 1.19214617e-01 -5.39634943e-01
-8.73689950e-02 8.04670334e-01 -9.36257094e-03 -1.39083529e+00
3.46129477e-01 7.19911218e-01 -1.20424974e+00 3.51759791e-01
1.32200634e+00 2.15319902e-01 1.49188548e-01 4.81770158e-01
-2.61614770e-01 3.42458077e-02 -6.40378952e-01 -5.77312171e-01
1.25119403e-01 -3.85775387e-01 -9.55960810e-01 -3.53492707e-01
3.37325901e-01 -9.21696126e-02 1.55345276e-01 1.24446392e+00
5.79181612e-01 -1.33064762e-01 5.06417990e-01 -9.16335166e-01
-7.21313715e-01 1.07614636e+00 -6.45280004e-01 2.98371255e-01
7.12407410e-01 -1.31859735e-01 1.13450599e+00 -1.26980758e+00
7.38138497e-01 1.53441823e+00 5.85958183e-01 1.04873747e-01
-6.62142098e-01 -4.64933515e-01 1.10897169e-01 3.17636371e-01
-1.47768259e+00 -3.55932444e-01 4.50925350e-01 -1.68786302e-01
7.60891557e-01 2.43717983e-01 4.60753202e-01 3.52342606e-01
1.55817851e-01 8.83389413e-01 1.41949391e+00 -9.44518924e-01
-1.89183950e-01 4.88002934e-02 6.26465023e-01 1.16889322e+00
1.58330560e-01 -5.60212910e-01 -5.70958436e-01 -4.38123733e-01
5.26875913e-01 -7.24900007e-01 -2.17432424e-01 2.70257771e-01
-9.43338811e-01 8.90763283e-01 -7.25772381e-02 6.61431730e-01
-1.48282811e-01 -9.22052860e-01 2.95200884e-01 6.08416975e-01
1.70569703e-01 5.58340549e-01 -9.17452097e-01 -1.43235013e-01
-4.15423691e-01 2.61967480e-01 8.95521164e-01 7.27815509e-01
9.82095122e-01 1.03306875e-01 6.24008834e-01 1.46687031e+00
5.60463130e-01 8.71711969e-01 7.10836649e-01 -4.28059578e-01
2.51983494e-01 6.60316944e-01 6.21242076e-02 -5.92600405e-01
-5.02220809e-01 1.43167853e-01 -1.69178843e-02 2.15616971e-01
8.52048576e-01 -5.06563902e-01 -7.03022182e-01 1.08563292e+00
4.13098991e-01 -8.98297131e-01 2.14284822e-01 9.16105747e-01
8.66648972e-01 6.60832107e-01 -2.94234961e-01 -5.33280857e-02
1.79352891e+00 -5.82583249e-01 -6.92199111e-01 -2.61042327e-01
5.06685257e-01 -1.37121677e+00 1.25036693e+00 6.02719247e-01
-1.02362227e+00 6.59686178e-02 -1.19476795e+00 4.02932540e-02
-1.02890098e+00 1.39797315e-01 6.55851185e-01 1.01681054e+00
-9.76192415e-01 6.41118437e-02 -7.00831950e-01 -5.70373774e-01
-3.99662822e-01 2.98293650e-01 -5.52191734e-01 -3.95717807e-02
-1.44348454e+00 1.21855283e+00 3.39252949e-01 -5.01845963e-03
-3.00408989e-01 -2.63885409e-01 -1.28183591e+00 -4.64828849e-01
1.72520336e-02 3.03075224e-01 1.24317658e+00 -1.04015851e+00
-1.46909606e+00 1.11342096e+00 -1.26645148e-01 1.91134468e-01
-2.50030935e-01 -2.27130920e-01 -1.15107465e+00 2.06198528e-01
2.16322005e-01 1.71724707e-01 2.44539306e-01 -1.21691477e+00
-8.68280470e-01 -7.06146479e-01 1.39460236e-01 4.32917863e-01
-1.81029007e-01 7.95460284e-01 -3.25168222e-01 -8.65033090e-01
1.43043503e-01 -8.53534997e-01 1.69263080e-01 -6.22917354e-01
-4.23947675e-03 -3.22788924e-01 5.49787760e-01 -1.33245850e+00
1.25919724e+00 -1.85829914e+00 -2.41099507e-01 4.74827051e-01
-2.92855680e-01 5.33919573e-01 -3.74502689e-01 8.76514196e-01
3.25411111e-02 1.13372564e-01 -1.24816649e-01 4.74741459e-01
1.38731627e-02 4.23836559e-02 9.07350332e-02 4.68724936e-01
2.52039135e-01 4.50633705e-01 -1.16496420e+00 -4.45772499e-01
-2.74350098e-03 3.81512344e-01 6.66991770e-02 -1.34203538e-01
-1.10552303e-01 1.11174183e-02 2.72037880e-03 1.33618987e+00
9.50200915e-01 4.40074503e-01 6.98867321e-01 9.99206770e-03
-5.52901447e-01 6.66967332e-01 -1.21067750e+00 1.40203643e+00
-6.53557301e-01 5.05652010e-01 3.59310538e-01 -4.32127714e-01
1.00939703e+00 2.15526611e-01 4.27778289e-02 -4.19942588e-01
2.94402719e-01 7.55959272e-01 1.56523690e-01 -1.68008193e-01
9.59281027e-01 2.43857816e-01 -2.26867393e-01 7.28984416e-01
1.66721746e-01 3.88334394e-02 8.89558852e-01 -2.02491246e-02
4.00247306e-01 2.37047985e-01 7.23815858e-01 -6.79822624e-01
7.15775549e-01 1.51406273e-01 5.03816605e-01 2.34894708e-01
-2.17170388e-01 1.59870580e-01 1.45527765e-01 -1.92318439e-01
-3.91805470e-01 -1.33459806e+00 -4.76093680e-01 1.58175516e+00
6.35548234e-02 -6.66260004e-01 -7.39645243e-01 -9.09216344e-01
-3.46280038e-02 5.88929653e-01 -2.56434232e-01 4.76909816e-01
-9.03660834e-01 -9.29230928e-01 8.53754044e-01 6.10802369e-03
3.87028903e-01 -1.00789022e+00 7.14525441e-03 2.20852401e-02
-5.58134317e-01 -4.65014756e-01 -7.98544526e-01 9.53297410e-03
-3.50035578e-01 -1.16440129e+00 -5.27715385e-01 -1.53717303e+00
5.02520144e-01 2.45327726e-01 1.06184006e+00 1.38474777e-01
-3.92323434e-02 2.38899961e-01 -7.24844873e-01 -2.38201141e-01
-6.68043137e-01 1.61811769e-01 5.24000525e-02 -2.07010686e-01
7.79207528e-01 -1.73554271e-02 -1.27536759e-01 2.79968858e-01
-7.11700559e-01 -6.80201948e-01 1.84211880e-01 5.85978925e-01
2.15788424e-01 -4.74562263e-03 7.43986726e-01 -1.02270567e+00
7.25475669e-01 -5.37756085e-01 -6.59922600e-01 8.18264961e-01
-5.08901834e-01 -3.73577215e-02 4.77713943e-01 -9.06908065e-02
-1.12527883e+00 1.48314297e-01 -6.12001956e-01 8.10385585e-01
-2.76990443e-01 1.04700363e+00 -6.32230103e-01 -2.42043868e-01
6.79599106e-01 1.06711790e-01 6.77969754e-02 -5.72270930e-01
7.54052520e-01 1.29356968e+00 4.11360115e-01 -6.08666539e-01
4.19967830e-01 -8.05154722e-03 -7.89811373e-01 -1.20509994e+00
-3.05730522e-01 -6.45883262e-01 -1.00419641e+00 -3.15719306e-01
3.41241986e-01 -8.31122398e-01 -5.64435780e-01 9.33360398e-01
-8.50877047e-01 -9.73315686e-02 4.94464517e-01 2.51329452e-01
2.27190964e-02 5.88689864e-01 -1.00108445e+00 -4.33869511e-01
-4.53800738e-01 -1.07530594e+00 5.23015440e-01 5.30615687e-01
-3.89368623e-01 -1.23854399e+00 1.08919390e-01 6.62455037e-02
3.74211483e-02 -1.06729299e-01 1.17962015e+00 -9.22914803e-01
1.99700981e-01 1.38939396e-01 1.48250088e-01 2.31135581e-02
7.13471889e-01 4.30381298e-01 -5.14539838e-01 -2.66357601e-01
-5.00498831e-01 -3.29970926e-01 2.08794981e-01 -2.68608537e-02
-4.73412732e-03 -3.64758700e-01 2.12243617e-01 2.21824124e-01
1.44458127e+00 5.39488554e-01 2.91945070e-01 7.25870728e-01
2.35505179e-01 6.80163920e-01 9.71674562e-01 3.93442869e-01
8.05589736e-01 4.21848536e-01 -9.85176265e-02 1.57752171e-01
7.01580644e-02 1.38686746e-01 8.33849013e-01 1.34604990e+00
3.00800264e-01 -1.63253082e-03 -1.43046129e+00 7.72839487e-01
-1.37736320e+00 -4.80102152e-01 -2.67462909e-01 2.00193810e+00
1.36985683e+00 -4.27908272e-01 6.11376464e-01 1.42033994e-02
8.40259016e-01 -1.69585124e-01 -3.73931713e-02 -7.62569964e-01
-6.56802773e-01 5.19230306e-01 4.48204815e-01 1.20838094e+00
-1.17172527e+00 1.46506715e+00 6.24278545e+00 8.22872460e-01
-1.24759555e+00 -7.22510442e-02 2.00807750e-01 5.73940992e-01
-4.19213325e-01 1.17399283e-02 -1.07076752e+00 -3.27582285e-02
9.04528081e-01 -4.17914391e-01 5.31139433e-01 3.47226143e-01
5.08329496e-02 -2.95002759e-01 -2.29756191e-01 6.32215977e-01
3.83241981e-01 -1.01922977e+00 3.36884469e-01 -2.30261177e-01
2.88591594e-01 7.95505941e-02 -1.12439200e-01 1.57776810e-02
7.47436941e-01 -6.74370110e-01 7.01223195e-01 -8.82617235e-02
7.35662520e-01 -1.18574381e+00 5.62145650e-01 -2.41664007e-01
-1.21054053e+00 2.74670690e-01 -1.36386022e-01 2.60007530e-01
6.64575100e-02 6.79674596e-02 -9.09958541e-01 6.76457345e-01
5.71091890e-01 4.82629895e-01 -9.26296175e-01 9.44284439e-01
-4.72838253e-01 9.69052255e-01 -8.08747634e-02 -1.13471456e-01
9.42234024e-02 -7.07329273e-01 5.18575072e-01 1.55189347e+00
2.51406938e-01 5.30239269e-02 3.78537923e-01 -1.29299507e-01
4.16808426e-01 8.65063727e-01 -2.22834900e-01 -4.87902202e-02
8.61390233e-01 1.48328900e+00 -1.03016174e+00 -2.07569432e-02
-3.93260300e-01 8.34703505e-01 1.59248561e-01 1.21316753e-01
-2.92401344e-01 -1.29897952e+00 8.11365128e-01 -7.39764720e-02
1.90384105e-01 -6.86112165e-01 -2.99268037e-01 -1.12825763e+00
-4.74971056e-01 -1.68339610e+00 9.24079776e-01 -5.14790893e-01
-1.19866633e+00 8.25442493e-01 -1.95248470e-01 -9.80108619e-01
-4.30273890e-01 -1.12264180e+00 -2.35827416e-01 1.06400633e+00
-1.25949240e+00 -1.40114665e+00 3.15808445e-01 6.28887475e-01
4.78390545e-01 -6.40128374e-01 1.29146791e+00 3.79631430e-01
-6.04947448e-01 8.56831312e-01 2.10267022e-01 7.28674591e-01
1.17913187e+00 -1.54403591e+00 3.46346051e-01 8.93713236e-01
2.93164372e-01 8.38251829e-01 5.83727837e-01 -6.74368262e-01
-1.38374043e+00 -6.72645986e-01 1.29688990e+00 -2.81134367e-01
1.07762325e+00 -2.15964347e-01 -7.54147947e-01 6.75477862e-01
8.02777052e-01 -8.04602563e-01 1.23932588e+00 4.20148551e-01
-5.77770531e-01 -1.87486589e-01 -1.12066221e+00 9.49505389e-01
2.97969222e-01 -5.20291984e-01 -7.03417599e-01 3.04395139e-01
2.31694758e-01 -3.89008760e-01 -1.14776194e+00 -1.77078083e-01
6.80177748e-01 -4.91611481e-01 5.49498796e-01 -2.03960627e-01
-1.08084060e-01 -4.10260022e-01 -2.35761791e-01 -1.77640009e+00
-1.25356361e-01 -9.69418705e-01 5.85301995e-01 1.42284858e+00
9.53757763e-01 -9.32600439e-01 1.06897786e-01 1.91676132e-02
-2.42707178e-01 -2.05323964e-01 -5.05501151e-01 -6.34855211e-01
6.00695133e-01 1.52873993e-03 8.54381859e-01 1.21381569e+00
7.50688851e-01 4.45951730e-01 1.33101940e-01 5.45964301e-01
1.79743871e-01 3.88073087e-01 1.25967562e-01 -9.95995879e-01
-8.12879503e-02 -5.78388631e-01 -2.30489429e-02 -6.87030256e-01
2.02119932e-01 -1.07550585e+00 9.07774866e-02 -1.31573880e+00
-6.96341217e-01 -5.99359572e-01 -9.40799415e-02 9.93617952e-01
-2.93577880e-01 5.08812547e-01 1.77523687e-01 6.41538799e-02
-2.11746916e-01 6.21730201e-02 8.46064508e-01 -3.29773948e-02
-4.85950261e-01 -1.31865978e-01 -9.39748108e-01 7.39714801e-01
1.34005296e+00 -1.45878300e-01 1.18606538e-01 -5.23550272e-01
2.62902826e-01 -1.01622306e-01 -6.32838309e-01 -3.77384782e-01
-8.79911929e-02 -4.57053363e-01 5.25862463e-02 -5.56257844e-01
5.98138869e-02 -1.00552395e-01 -4.62185949e-01 5.15157163e-01
1.86502889e-01 9.59495425e-01 4.95066613e-01 -5.33205628e-01
-3.43229324e-01 -5.61098576e-01 8.09543312e-01 -1.85233042e-01
-1.00478935e+00 6.00357279e-02 -1.20992756e+00 3.00305575e-01
6.64225638e-01 2.00425118e-01 -5.94244540e-01 -2.48369709e-01
-6.13086164e-01 3.25205684e-01 5.13513982e-01 6.53886557e-01
5.85127175e-01 -1.10069966e+00 -1.07894754e+00 4.61976320e-01
3.62942904e-01 -9.44080532e-01 -3.69820118e-01 2.87172884e-01
-1.17346990e+00 1.58468574e-01 -6.17892742e-01 1.36736989e-01
-1.61106968e+00 4.73965053e-03 2.91364312e-01 7.16614276e-02
1.15738595e-02 5.70776343e-01 -5.58727086e-01 -1.19297624e+00
-7.82257095e-02 3.75609517e-01 -5.99186659e-01 4.70595479e-01
7.07915306e-01 3.64359409e-01 4.24327910e-01 -1.29437649e+00
-7.02609301e-01 3.06753397e-01 -2.62400687e-01 -6.43114269e-01
9.94837582e-01 -4.22872514e-01 -8.79235268e-01 5.67928970e-01
5.55867851e-01 1.03197300e+00 -4.05758321e-01 5.44241555e-02
5.30859709e-01 -2.48373702e-01 -3.30196470e-01 -1.38234627e+00
-6.60101116e-01 4.62682068e-01 5.47784686e-01 8.72544497e-02
1.08442426e+00 -2.70033687e-01 7.70426214e-01 4.02681023e-01
5.05074084e-01 -1.30896533e+00 -5.05368352e-01 1.34965718e+00
6.74959064e-01 -8.98810267e-01 -2.62806535e-01 -5.46658218e-01
-6.49191439e-01 1.26382005e+00 4.91985589e-01 1.85266584e-01
1.22675371e+00 2.22259462e-01 1.21523786e+00 5.85587956e-02
-2.45670944e-01 -5.63020766e-01 2.00815529e-01 9.47255909e-01
1.06413662e+00 2.29460701e-01 -9.22027886e-01 6.60995841e-01
-6.87425077e-01 -7.79722214e-01 8.58649969e-01 1.14026046e+00
-5.87822735e-01 -1.86385334e+00 -7.75308788e-01 1.99996069e-01
-6.07125700e-01 -5.33868253e-01 -6.12604260e-01 8.56484473e-01
6.90509081e-02 1.25756109e+00 -1.24735624e-01 -2.15599313e-01
2.87892856e-02 2.55636185e-01 6.59818888e-01 -6.71642065e-01
-1.06375837e+00 2.05161139e-01 5.06557107e-01 1.64896205e-01
-1.73463359e-01 -1.04302502e+00 -1.62769794e+00 -6.02753282e-01
-3.29428464e-01 5.30653954e-01 5.79364479e-01 7.22081602e-01
1.00022398e-01 -5.32204449e-01 6.18173420e-01 -4.80254233e-01
-1.75329477e-01 -9.65069771e-01 -8.35037827e-01 5.41577348e-03
-9.38661490e-03 -2.91424632e-01 1.13981962e-01 -1.10905720e-02] | [10.329154968261719, 10.467994689941406] |
5e4c7aee-d580-4e3e-b3af-500f5340464e | textless-speech-to-speech-translation-on-real | 2112.08352 | null | https://arxiv.org/abs/2112.08352v2 | https://arxiv.org/pdf/2112.08352v2.pdf | Textless Speech-to-Speech Translation on Real Data | We present a textless speech-to-speech translation (S2ST) system that can translate speech from one language into another language and can be built without the need of any text data. Different from existing work in the literature, we tackle the challenge in modeling multi-speaker target speech and train the systems with real-world S2ST data. The key to our approach is a self-supervised unit-based speech normalization technique, which finetunes a pre-trained speech encoder with paired audios from multiple speakers and a single reference speaker to reduce the variations due to accents, while preserving the lexical content. With only 10 minutes of paired data for speech normalization, we obtain on average 3.2 BLEU gain when training the S2ST model on the VoxPopuli S2ST dataset, compared to a baseline trained on un-normalized speech target. We also incorporate automatically mined S2ST data and show an additional 2.0 BLEU gain. To our knowledge, we are the first to establish a textless S2ST technique that can be trained with real-world data and works for multiple language pairs. Audio samples are available at https://facebookresearch.github.io/speech_translation/textless_s2st_real_data/index.html . | ['Wei-Ning Hsu', 'Juan Pino', 'Yossi Adi', 'Jiatao Gu', 'Sravya Popuri', 'Changhan Wang', 'Peng-Jen Chen', 'Holger Schwenk', 'Paul-Ambroise Duquenne', 'Hongyu Gong', 'Ann Lee'] | 2021-12-15 | null | https://aclanthology.org/2022.naacl-main.63 | https://aclanthology.org/2022.naacl-main.63.pdf | naacl-2022-7 | ['speech-to-speech-translation'] | ['speech'] | [ 4.44983393e-01 4.04465556e-01 -5.17261997e-02 -4.45154130e-01
-1.56174386e+00 -6.64775133e-01 6.34865344e-01 -1.47795558e-01
-4.13564265e-01 5.53765953e-01 5.84245145e-01 -5.52733600e-01
5.83999336e-01 -2.24685282e-01 -7.67631531e-01 -4.55782145e-01
6.02813244e-01 7.61339426e-01 2.26018608e-01 -6.54282272e-01
-4.20837671e-01 8.69488344e-03 -1.17008793e+00 2.78129309e-01
6.55534089e-01 6.48892105e-01 3.75742406e-01 1.06387603e+00
-1.44940764e-01 6.42564476e-01 -8.38757217e-01 -4.72251147e-01
4.84808505e-01 -7.36095726e-01 -8.08526993e-01 -2.22086444e-01
6.63032234e-01 -2.35403538e-01 -4.48046744e-01 8.50843310e-01
1.08213508e+00 2.17741638e-01 2.62069464e-01 -1.16847682e+00
-4.83245164e-01 1.15649986e+00 5.68460748e-02 2.76332587e-01
4.13408428e-01 -4.84770164e-03 8.19581330e-01 -8.21260214e-01
5.32889783e-01 1.38440990e+00 3.54429275e-01 9.35265064e-01
-1.23121846e+00 -7.06520617e-01 -2.95796365e-01 -4.01890799e-02
-1.32022345e+00 -1.45233476e+00 5.02587616e-01 -1.90329537e-01
1.14602292e+00 6.33397520e-01 2.28761420e-01 1.58176637e+00
-3.90695214e-01 7.89414823e-01 8.21556270e-01 -7.93500483e-01
7.68900663e-02 2.73974687e-01 -3.67177427e-01 2.51209646e-01
-5.70371807e-01 1.01645470e-01 -8.95809889e-01 2.07977414e-01
3.34959686e-01 -5.39039433e-01 -3.06257337e-01 1.87834471e-01
-1.42628527e+00 5.72594583e-01 -1.06465086e-01 5.87709427e-01
7.61048868e-02 -2.74139866e-02 6.35801613e-01 7.56440818e-01
7.26690769e-01 1.62189186e-01 -7.07570553e-01 -5.90609133e-01
-1.38163066e+00 -1.84936494e-01 9.16304529e-01 1.35707223e+00
3.85093540e-01 3.25434387e-01 -2.09625632e-01 1.15352321e+00
1.55433238e-01 1.04560220e+00 9.97204006e-01 -9.14423704e-01
8.92118990e-01 -1.15727931e-02 -3.39268744e-01 -2.54123747e-01
1.25795990e-01 -4.44799572e-01 -6.58530891e-01 -2.35703886e-01
3.12622666e-01 -3.54702413e-01 -9.11948919e-01 1.91524863e+00
1.02749035e-01 3.46127376e-02 2.92110741e-01 5.33545494e-01
9.42051589e-01 8.99028063e-01 -4.33595359e-01 -3.56257498e-01
1.03877342e+00 -1.41813672e+00 -8.69568288e-01 -4.07725930e-01
5.95598459e-01 -1.36895752e+00 1.37583327e+00 -3.34455073e-02
-1.27594018e+00 -4.40237910e-01 -9.52792525e-01 -1.59014836e-01
-2.82566994e-01 1.01262983e-02 -1.25186607e-01 8.18906784e-01
-1.49386239e+00 2.64440268e-01 -1.00410390e+00 -7.09610343e-01
-1.50281891e-01 1.60051018e-01 -4.87765193e-01 1.59504250e-01
-1.29035449e+00 8.56218934e-01 1.71926290e-01 -4.16769177e-01
-8.72689903e-01 -6.42341495e-01 -8.43861461e-01 -2.05790643e-02
4.51627195e-01 -4.55532849e-01 1.99400449e+00 -1.21793175e+00
-2.09208941e+00 6.91509068e-01 -5.31530440e-01 -5.02949595e-01
5.75248241e-01 2.70950608e-02 -6.96389318e-01 4.59588394e-02
1.93089411e-01 5.41945815e-01 9.09496844e-01 -7.87990749e-01
-3.67005020e-01 -2.69835852e-02 -4.77887750e-01 2.93488234e-01
-4.95409489e-01 5.01466870e-01 -5.59422493e-01 -9.63694096e-01
-1.15415223e-01 -9.35455620e-01 1.94450319e-01 -3.73601705e-01
-5.68181276e-01 1.06653713e-01 6.82948232e-01 -1.08875608e+00
1.18629336e+00 -2.17964101e+00 1.01369835e-01 -1.57990083e-01
-2.11067989e-01 4.85002667e-01 -5.36706924e-01 6.61788344e-01
-1.05046548e-01 1.53972536e-01 -3.52836490e-01 -9.22694862e-01
1.35250941e-01 1.22222103e-01 -2.88254708e-01 1.10422589e-01
-1.15088701e-01 8.45617235e-01 -7.40782142e-01 -3.50308895e-01
2.41040558e-01 6.59709632e-01 -4.16954190e-01 4.10866112e-01
2.93761715e-02 5.55477500e-01 2.05949262e-01 3.59670728e-01
4.37692553e-01 2.56927252e-01 -6.37503201e-03 1.14303194e-01
-1.90163314e-01 1.16508234e+00 -9.93809938e-01 1.87060583e+00
-8.74734461e-01 9.42248642e-01 4.93825436e-01 -7.33260453e-01
8.52847695e-01 9.07351971e-01 2.94466496e-01 -8.56153131e-01
1.49368450e-01 5.28364301e-01 -1.64178967e-01 -1.40731871e-01
4.07638669e-01 -2.27258369e-01 -1.27455946e-02 5.33144593e-01
5.77287257e-01 -5.30283868e-01 1.26396701e-01 1.80036768e-01
1.17101789e+00 -3.18143159e-01 1.31268293e-01 7.73521364e-02
5.37667692e-01 -3.38253915e-01 4.42926645e-01 5.62115133e-01
-2.59752303e-01 8.47232759e-01 -4.19646092e-02 1.78177416e-01
-1.21520090e+00 -1.08415639e+00 2.42421567e-01 1.59035528e+00
-5.48667252e-01 -5.86163104e-01 -1.13632321e+00 -4.20870423e-01
-5.24626255e-01 1.03732669e+00 -1.08889595e-01 -1.17579056e-02
-7.11329758e-01 -2.21209228e-01 1.12366796e+00 1.19265378e-01
2.12216452e-01 -8.75997245e-01 2.89854288e-01 3.92475873e-01
-7.16398120e-01 -1.44068766e+00 -1.16126764e+00 2.54046261e-01
-4.77508485e-01 -3.92439842e-01 -8.55613470e-01 -9.24640477e-01
1.93011761e-01 1.94011509e-01 1.10013664e+00 -3.53131264e-01
2.05342054e-01 2.90027857e-01 -4.82084244e-01 -3.45198840e-01
-1.26966906e+00 5.98464489e-01 2.20355302e-01 -1.02824308e-01
2.31485814e-01 -6.89286530e-01 -4.55077887e-02 3.49808753e-01
-6.06230438e-01 2.05394581e-01 5.17023683e-01 7.73386419e-01
3.40801179e-01 -2.90783525e-01 5.26290119e-01 -4.47260082e-01
4.57914621e-01 -2.84151345e-01 -4.77961004e-01 1.87356159e-01
-4.07603711e-01 1.36305198e-01 6.01969600e-01 -3.89023781e-01
-9.70546544e-01 7.17995018e-02 -6.92128778e-01 -3.15399557e-01
-3.06180120e-01 1.43641964e-01 -4.07945216e-01 2.47221082e-01
7.19175160e-01 5.28156698e-01 1.45058529e-02 -7.31993675e-01
5.45136154e-01 1.44543576e+00 7.87211955e-01 -2.66008019e-01
9.19350803e-01 -7.46697094e-03 -6.89403236e-01 -1.08214915e+00
-6.51689112e-01 -5.93140721e-01 -7.26179004e-01 1.39759228e-01
6.07880235e-01 -1.02160072e+00 -2.27746770e-01 3.51972848e-01
-1.29154646e+00 -6.43530369e-01 -3.77751976e-01 5.96623778e-01
-7.66989589e-01 2.21400887e-01 -6.04905486e-01 -7.00276971e-01
-6.73783004e-01 -1.12550902e+00 1.20568943e+00 -3.45934451e-01
-2.66980588e-01 -7.94735551e-01 3.96676511e-01 7.77664840e-01
7.38301873e-01 -5.03144264e-01 2.74051398e-01 -1.08468854e+00
-1.57722548e-01 1.82884131e-02 2.77640134e-01 8.24486852e-01
4.30304170e-01 -1.24748290e-01 -1.26930511e+00 -4.71310914e-01
2.83966251e-02 -1.95805535e-01 5.43502688e-01 7.36292899e-02
3.92565727e-01 -7.91335940e-01 1.46360219e-01 5.97072542e-01
6.73366845e-01 1.02546550e-01 2.75336772e-01 7.64571205e-02
5.86488664e-01 3.53732437e-01 2.94865340e-01 8.89840275e-02
4.81720895e-01 1.05696511e+00 4.54729870e-02 -2.34401852e-01
-7.20740259e-01 -5.21562696e-01 9.52214777e-01 1.81841862e+00
3.60485464e-01 -4.86311495e-01 -8.90928566e-01 8.01761448e-01
-1.41079116e+00 -9.94115353e-01 1.71701148e-01 2.23733258e+00
1.28014219e+00 3.82203655e-03 2.49030441e-01 2.31028557e-01
9.05103445e-01 6.49844408e-02 -2.22066686e-01 -4.91186112e-01
-1.97938308e-01 3.38136911e-01 5.24521530e-01 1.14122391e+00
-9.55044985e-01 1.32991099e+00 5.85170460e+00 1.06531835e+00
-1.49730110e+00 6.37956381e-01 3.45084459e-01 -5.64159095e-01
-2.86733150e-01 -3.27431038e-02 -6.52647853e-01 5.29157817e-01
1.78137410e+00 -4.40783262e-01 8.91242146e-01 6.32561862e-01
5.36467433e-01 5.95413506e-01 -1.01718020e+00 1.13000357e+00
2.18815669e-01 -9.80866373e-01 -1.02884598e-01 -1.76934853e-01
5.76349258e-01 7.53340065e-01 -5.10965884e-02 2.33621970e-01
4.54910100e-01 -8.73516798e-01 1.07966089e+00 8.55944213e-03
1.20268619e+00 -5.96117377e-01 6.41040623e-01 4.00204688e-01
-1.10337651e+00 3.20511311e-01 5.25799114e-03 1.92553252e-01
3.60636294e-01 4.41638708e-01 -1.52964485e+00 5.48969924e-01
5.71409702e-01 5.47956169e-01 -3.28710198e-01 5.91696441e-01
-4.08289254e-01 1.20531595e+00 -6.33439064e-01 1.81556106e-01
-3.58072817e-02 1.66553333e-01 9.29129541e-01 1.48616898e+00
6.80525124e-01 -4.38046575e-01 -7.44007453e-02 3.38439137e-01
-2.55941242e-01 5.85215807e-01 -5.07678866e-01 -2.55002856e-01
8.61353099e-01 9.92205262e-01 -2.13149488e-01 -4.41786081e-01
-4.19081956e-01 1.35402536e+00 3.22559848e-02 2.29538098e-01
-4.52431917e-01 -4.39022005e-01 7.69071996e-01 1.69213131e-01
1.64707348e-01 -2.95814902e-01 1.78509243e-02 -1.29482961e+00
1.11461334e-01 -1.22213614e+00 9.58207920e-02 -6.62074745e-01
-1.09027398e+00 1.04560685e+00 -3.23446184e-01 -1.11636353e+00
-7.80018330e-01 -2.13015422e-01 -3.72651100e-01 1.04719269e+00
-1.29381037e+00 -1.26103687e+00 3.46058719e-02 6.04639113e-01
9.09692883e-01 -4.88103896e-01 1.01721811e+00 6.17935896e-01
-5.49612105e-01 1.03702331e+00 4.01328981e-01 2.88700134e-01
1.27828646e+00 -1.15483153e+00 1.06985807e+00 1.02856374e+00
4.64082301e-01 4.51300085e-01 8.53469074e-01 -4.30648923e-01
-1.03555024e+00 -1.08939350e+00 1.52153254e+00 -5.00096738e-01
7.36690581e-01 -9.21735525e-01 -6.12389088e-01 8.80611360e-01
5.98107636e-01 -1.86456591e-01 8.60577583e-01 -1.53567180e-01
-3.66217345e-01 -1.94887668e-01 -9.08131301e-01 6.98929071e-01
1.14879251e+00 -8.77557993e-01 -6.60186112e-01 3.13163698e-01
1.31991196e+00 -5.04684210e-01 -8.77532780e-01 -4.88748178e-02
3.27530205e-01 -5.35718441e-01 7.59445250e-01 -2.69408435e-01
-1.36693880e-01 -1.92933530e-01 -5.47149420e-01 -1.78976166e+00
1.80598259e-01 -1.30614257e+00 2.50703216e-01 1.60421324e+00
8.39592099e-01 -7.35266984e-01 3.24179828e-01 1.93324983e-01
-6.03335261e-01 1.25526786e-01 -1.49568820e+00 -1.16077185e+00
1.89213276e-01 -6.62076890e-01 7.88121700e-01 9.59227860e-01
5.20943478e-02 6.67850614e-01 -4.97263163e-01 6.97325915e-02
3.73482764e-01 -4.16101128e-01 9.46906865e-01 -8.40831339e-01
-5.00195742e-01 -3.49909186e-01 -1.94765061e-01 -1.11394835e+00
2.04761818e-01 -1.18657446e+00 2.07426712e-01 -1.30523658e+00
-2.36506730e-01 -2.29810208e-01 4.07799482e-02 8.80564809e-01
8.47986713e-02 3.15531880e-01 3.12059343e-01 1.83079485e-02
-3.03237677e-01 7.97974944e-01 9.67215836e-01 -2.52554566e-01
-3.91148537e-01 8.39733556e-02 -7.06607878e-01 2.87789255e-01
1.05094409e+00 -6.57394528e-01 -4.99669999e-01 -6.67193711e-01
-1.73163489e-01 -2.74139736e-02 -1.17445946e-01 -9.67099428e-01
3.50974411e-01 -5.54006780e-03 -3.33627671e-01 -3.47407728e-01
4.73445028e-01 -6.11499369e-01 6.17097318e-02 1.20827548e-01
-4.47092652e-01 -3.16393040e-02 3.23293716e-01 -9.53194574e-02
-4.08988744e-01 -4.98754438e-03 7.74730384e-01 2.50958763e-02
-1.37872666e-01 1.83845162e-02 -6.04971111e-01 3.56642365e-01
4.06855881e-01 2.46966660e-01 -2.63786554e-01 -8.52519274e-01
-7.13925481e-01 -1.35712028e-01 3.81364524e-01 7.40813971e-01
2.15805382e-01 -1.41168654e+00 -1.18149757e+00 4.39947844e-01
7.85237104e-02 -2.35078782e-01 -9.64844078e-02 6.74356818e-01
-1.50893122e-01 5.63255966e-01 8.94529745e-02 -5.18156111e-01
-1.54134607e+00 2.48202309e-01 4.11022246e-01 1.36338964e-01
-3.41946959e-01 7.70710528e-01 -2.02607736e-01 -1.13000500e+00
2.12862313e-01 -3.26117188e-01 3.58072698e-01 -1.13709765e-02
5.30049026e-01 3.81778985e-01 6.20467424e-01 -1.10422826e+00
-4.04426813e-01 1.72393203e-01 3.56265008e-02 -9.56809223e-01
1.37656593e+00 -4.42154616e-01 4.34970483e-03 6.59965277e-01
1.28180015e+00 5.12472928e-01 -8.91713262e-01 -4.72001553e-01
-2.23218277e-01 -1.59359738e-01 5.85801788e-02 -9.09621060e-01
-9.39193130e-01 8.91646564e-01 4.95663375e-01 1.33848608e-01
1.06494033e+00 1.24507874e-01 1.04588890e+00 4.30156767e-01
7.04998448e-02 -1.26614535e+00 -9.35041904e-02 8.15608561e-01
1.03484130e+00 -1.14587128e+00 -7.81047761e-01 -3.74868751e-01
-5.40555954e-01 8.22231174e-01 1.18807323e-01 5.61927140e-01
4.41382051e-01 6.46591127e-01 7.37348199e-01 4.64507580e-01
-8.48873556e-01 -3.10984552e-01 1.56994149e-01 7.35275924e-01
6.83897555e-01 2.46943578e-01 1.27353951e-01 4.54990715e-01
-1.08605182e+00 -2.97880739e-01 4.63665009e-01 6.98138237e-01
-3.02773386e-01 -1.57896304e+00 -4.48092639e-01 6.23159334e-02
-4.62444186e-01 -5.83954871e-01 -6.42260432e-01 2.54240125e-01
-2.77269483e-01 1.44088221e+00 -4.70606089e-02 -3.94832224e-01
4.29804772e-01 4.58375484e-01 2.02415496e-01 -8.30422163e-01
-5.74205518e-01 4.51923996e-01 3.22793722e-01 -4.52717632e-01
-1.11480154e-01 -6.81250393e-01 -9.72181320e-01 -4.47389424e-01
-9.42872465e-02 1.97006822e-01 1.02916908e+00 7.72241175e-01
5.07716835e-01 5.27978241e-01 7.75732040e-01 -8.49454522e-01
-4.51646537e-01 -1.42222345e+00 -1.42066970e-01 1.58597112e-01
7.74864852e-01 -3.91232669e-02 -6.05407298e-01 2.38811612e-01] | [14.562471389770508, 6.977108478546143] |
fd97a7a0-dd71-49ba-9dd6-2b46a326849a | interpreting-a-recurrent-neural-network-model | 1905.09865 | null | https://arxiv.org/abs/1905.09865v4 | https://arxiv.org/pdf/1905.09865v4.pdf | Interpreting a Recurrent Neural Network's Predictions of ICU Mortality Risk | Deep learning has demonstrated success in many applications; however, their use in healthcare has been limited due to the lack of transparency into how they generate predictions. Algorithms such as Recurrent Neural Networks (RNNs) when applied to Electronic Medical Records (EMR) introduce additional barriers to transparency because of the sequential processing of the RNN and the multi-modal nature of EMR data. This work seeks to improve transparency by: 1) introducing Learned Binary Masks (LBM) as a method for identifying which EMR variables contributed to an RNN model's risk of mortality (ROM) predictions for critically ill children; and 2) applying KernelSHAP for the same purpose. Given an individual patient, LBM and KernelSHAP both generate an attribution matrix that shows the contribution of each input feature to the RNN's sequence of predictions for that patient. Attribution matrices can be aggregated in many ways to facilitate different levels of analysis of the RNN model and its predictions. Presented are three methods of aggregations and analyses: 1) over volatile time periods within individual patient predictions, 2) over populations of ICU patients sharing specific diagnoses, and 3) across the general population of critically ill children. | ['Melissa D. Aczon', 'David Ledbetter', 'Randall Wetzel', 'Long V. Ho'] | 2019-05-23 | null | null | null | null | ['icu-mortality'] | ['medical'] | [ 4.65168446e-01 4.44072038e-01 4.21838425e-02 -4.32122678e-01
-5.67479312e-01 -2.50385821e-01 3.84467304e-01 4.02729064e-01
-3.16870004e-01 6.77652478e-01 8.87712598e-01 -8.01966071e-01
-4.20925707e-01 -5.62923372e-01 -4.15636599e-01 -4.47212905e-01
-6.68773949e-02 5.80076873e-01 -6.07648194e-01 1.37511641e-01
-1.01295568e-01 5.76187909e-01 -1.08401334e+00 7.71636069e-01
5.42186499e-01 7.42809653e-01 -2.23913401e-01 7.07171500e-01
4.08868082e-02 1.64792144e+00 -8.23982358e-01 -2.94722110e-01
1.49085045e-01 -5.63849866e-01 -7.68113554e-01 -4.05438989e-01
-6.69153873e-03 -3.64548415e-01 -5.05494356e-01 6.06988490e-01
6.33560181e-01 3.85569818e-02 9.48060453e-01 -9.62476969e-01
-5.45211613e-01 1.11788666e+00 5.06228916e-02 2.13684335e-01
1.70337111e-01 2.68380195e-01 6.74079776e-01 -4.71826106e-01
3.74018133e-01 9.94610965e-01 9.38670218e-01 8.69971693e-01
-1.38952422e+00 -6.44448996e-01 -2.06492811e-01 -1.43815652e-01
-1.07082999e+00 -3.52822304e-01 3.46249223e-01 -1.03385127e+00
1.37195516e+00 3.93738866e-01 7.43533671e-01 1.22361445e+00
4.88675892e-01 3.78430039e-01 8.46866250e-01 1.14005491e-01
4.07958254e-02 1.39349774e-01 6.50598109e-02 3.51285994e-01
5.84656857e-02 3.14058989e-01 -5.82732499e-01 -6.11134470e-01
6.77705646e-01 5.73837817e-01 -3.53376091e-01 1.69843495e-01
-1.44229734e+00 8.31351697e-01 1.43148258e-01 5.46702929e-02
-6.91955924e-01 1.50778770e-01 5.56691766e-01 2.79622912e-01
4.96230125e-01 6.43620193e-01 -7.55329311e-01 -2.12178439e-01
-9.44302380e-01 1.55249098e-02 6.95076108e-01 5.48849642e-01
3.95759910e-01 2.85212159e-01 -2.62265384e-01 7.69191325e-01
7.91242570e-02 1.65600210e-01 7.93420494e-01 -6.48873866e-01
5.34276843e-01 6.37263060e-01 9.51344892e-03 -8.45396042e-01
-8.69636416e-01 -2.15419516e-01 -1.19646382e+00 4.44735512e-02
1.25991285e-01 -5.43734550e-01 -9.55054581e-01 1.67276978e+00
-4.09462526e-02 -3.91774066e-02 5.05688071e-01 5.62115252e-01
1.06899881e+00 6.16548419e-01 1.81809515e-01 -1.85073335e-02
1.10577440e+00 -6.27589047e-01 -5.69671869e-01 -9.65487137e-02
1.01425970e+00 -4.64808047e-01 3.83877575e-01 1.84067339e-01
-1.12349486e+00 -1.20450161e-01 -4.89354461e-01 1.22880787e-01
-3.48435789e-01 2.50851344e-02 3.50750089e-01 3.57978463e-01
-1.17444360e+00 8.34699690e-01 -8.02673578e-01 -1.11147687e-01
4.63510305e-01 5.75999856e-01 -3.52833241e-01 3.94915700e-01
-1.22174025e+00 1.36750209e+00 2.05268785e-01 1.69149458e-01
-7.79558539e-01 -1.04183984e+00 -8.74976754e-01 2.45749265e-01
-2.54361719e-01 -1.03923488e+00 1.01496780e+00 -9.77743804e-01
-1.10783422e+00 6.63220286e-01 6.64567202e-02 -7.62250543e-01
6.45988405e-01 -2.22421646e-01 -3.08147162e-01 -1.11417927e-01
-2.17976883e-01 3.49097937e-01 6.32338107e-01 -7.47804582e-01
-4.83756810e-01 -2.63314694e-01 -4.46407199e-01 2.12733403e-01
3.68750021e-02 2.44388193e-01 3.19228739e-01 -9.11997080e-01
-6.72767833e-02 -7.46113181e-01 -5.57803869e-01 -5.94607294e-01
-6.47001326e-01 9.23646614e-02 3.16939861e-01 -9.56687808e-01
1.39339650e+00 -2.07540965e+00 1.28334716e-01 2.39323467e-01
7.07734704e-01 2.65987903e-01 1.27707973e-01 5.18209815e-01
-6.27372086e-01 2.76047885e-01 -2.38964587e-01 -3.63203168e-01
-4.01097298e-01 -1.25420215e-02 -5.93286812e-01 3.79499346e-01
2.93301702e-01 1.04321909e+00 -7.39390254e-01 -1.99546352e-01
3.53824109e-01 8.38212073e-01 -4.81661022e-01 4.64414269e-01
1.87217847e-01 5.03974497e-01 -1.10781305e-01 2.97187299e-01
-1.16786450e-01 -3.99052143e-01 1.43923253e-01 1.31428257e-01
-4.61734310e-02 6.52090669e-01 -6.32878363e-01 7.81670809e-01
-3.71220767e-01 6.32269144e-01 -3.67245555e-01 -6.90379441e-01
9.10777509e-01 8.05485785e-01 7.86077261e-01 -1.43768648e-02
1.60770625e-01 5.07512204e-02 2.83426344e-01 -5.06131351e-01
1.30718067e-01 -3.53826731e-01 1.22358598e-01 9.18097079e-01
-2.13977516e-01 2.26523221e-01 -4.62277681e-01 1.56675652e-02
1.56166923e+00 -3.58131737e-01 4.24595863e-01 -6.34405017e-02
-3.33123654e-02 -1.13125607e-01 7.24905968e-01 8.28382730e-01
3.37440707e-02 8.82100999e-01 4.45704550e-01 -1.03758883e+00
-1.08853328e+00 -9.16821420e-01 -2.26060078e-01 6.22355819e-01
-8.08713019e-01 -2.15422437e-01 -5.25551021e-01 -4.43579853e-01
2.12749586e-01 9.34069693e-01 -1.03210986e+00 -3.88364732e-01
-5.03415644e-01 -1.08017647e+00 7.56895959e-01 7.09322095e-01
-3.06618065e-01 -1.47647250e+00 -1.35888219e+00 4.82892722e-01
-1.35082319e-01 -6.88213885e-01 -3.75109613e-01 6.35755002e-01
-1.10857618e+00 -1.16878819e+00 -5.84125459e-01 -3.61469537e-01
7.99233079e-01 -3.86934012e-01 1.07723725e+00 -7.05731809e-02
-5.06369889e-01 4.60373938e-01 -1.34738654e-01 -6.08953476e-01
-8.73505175e-01 -6.34899512e-02 3.70124936e-01 -1.63880542e-01
5.07683575e-01 -4.31256920e-01 -6.07536554e-01 -5.28128631e-02
-9.58485663e-01 2.49420032e-01 5.13180733e-01 9.83115852e-01
3.06941420e-01 -2.78730273e-01 6.31447673e-01 -1.29216564e+00
8.99739444e-01 -8.13686490e-01 -3.02059770e-01 2.35237122e-01
-8.58904541e-01 -8.32920149e-03 7.36817658e-01 -4.91492540e-01
-5.41650772e-01 -4.38274667e-02 7.14344308e-02 -6.97910964e-01
-1.55403942e-01 4.42124397e-01 5.51967561e-01 6.21248126e-01
6.10689580e-01 1.45839140e-01 2.20204324e-01 -4.66012329e-01
-3.93780246e-02 6.64546728e-01 3.52410465e-01 -9.18941665e-03
3.82951230e-01 1.52935972e-02 -2.02198893e-01 -3.44173819e-01
-5.78643799e-01 -5.97069003e-02 -5.01060545e-01 -5.95134087e-02
1.13535917e+00 -8.96445990e-01 -8.69326532e-01 2.88847238e-01
-1.04565740e+00 -5.37481546e-01 -5.11619329e-01 6.96130097e-01
-5.11168122e-01 -2.13125065e-01 -9.20310676e-01 -8.90574753e-01
-8.35979640e-01 -1.24449837e+00 4.68972087e-01 -8.97974595e-02
-9.30690229e-01 -1.07765734e+00 1.95460960e-01 2.07175389e-01
5.14961481e-01 3.59607428e-01 1.38825023e+00 -1.19056523e+00
-9.33647305e-02 -2.74683446e-01 -5.36590777e-02 4.81836975e-01
5.01143157e-01 1.21636800e-01 -9.13896859e-01 -1.00170918e-01
1.39724940e-01 -2.03485098e-02 5.82720816e-01 5.96073508e-01
1.17070007e+00 -8.05196702e-01 -2.23618820e-01 6.40941262e-01
1.08485866e+00 4.71211225e-01 4.07392055e-01 1.50822252e-01
8.49400878e-01 8.51840317e-01 -3.60471644e-02 5.69741189e-01
2.66561329e-01 2.01005012e-01 2.77756035e-01 -3.29378515e-01
2.81505793e-01 -2.35948622e-01 3.24532628e-01 8.96948576e-01
4.54964787e-02 1.12726897e-01 -1.34889495e+00 5.21631181e-01
-1.98149312e+00 -8.92918885e-01 -3.46946605e-02 2.24521732e+00
7.09860861e-01 -1.62428200e-01 9.66474041e-03 -2.02768832e-01
5.76396167e-01 -2.03130413e-02 -7.54356384e-01 -8.75792146e-01
-1.45611987e-01 3.88787985e-02 4.80365306e-01 3.24410558e-01
-8.20917249e-01 4.04518694e-01 7.50024223e+00 3.36335190e-02
-9.35268521e-01 -4.25786115e-02 1.14309883e+00 -4.67969894e-01
-3.42052251e-01 -3.89956266e-01 -4.62832272e-01 5.60050428e-01
1.38329637e+00 -2.03042831e-02 5.01002669e-01 6.41144097e-01
4.11228329e-01 2.64098465e-01 -1.52936125e+00 1.03103602e+00
-1.34513795e-01 -1.47334111e+00 1.67741701e-01 3.89651842e-02
6.80224717e-01 3.78928363e-01 2.62732893e-01 -3.29387337e-02
7.72427380e-01 -1.55843103e+00 4.51523393e-01 8.51417184e-01
9.08228219e-01 -7.33436048e-01 9.49886322e-01 2.12232381e-01
-5.26465237e-01 -2.51999885e-01 -1.19469330e-01 -8.42402801e-02
5.36920391e-02 5.60628772e-01 -1.31895137e+00 5.67601807e-02
8.37024689e-01 6.22041702e-01 -1.54598996e-01 5.44861197e-01
6.26243129e-02 7.67417908e-01 8.55777040e-03 2.00363025e-01
1.08779907e-01 -1.01289496e-01 4.56536263e-01 1.31467128e+00
1.69638067e-01 1.65317476e-01 -4.42237407e-01 1.07258725e+00
-5.23803234e-02 7.18311444e-02 -9.21627462e-01 -2.43039936e-01
2.75407016e-01 9.05560255e-01 -2.24090099e-01 -4.69843477e-01
-3.19678575e-01 5.56484759e-01 3.01437795e-01 4.22207505e-01
-5.66667020e-01 -1.50859550e-01 1.08428597e+00 1.82602465e-01
-7.34830946e-02 3.85277003e-01 -5.11320651e-01 -8.97984445e-01
-3.60091776e-01 -1.10383582e+00 6.00444138e-01 -8.18443954e-01
-1.31021380e+00 8.11247826e-01 -3.49763662e-01 -1.08737159e+00
-6.27366483e-01 -4.24626231e-01 -4.53579158e-01 1.33135557e+00
-9.46553767e-01 -5.13637483e-01 2.24912733e-01 4.61819768e-01
2.64999986e-01 -4.05265361e-01 1.23112977e+00 1.06343843e-01
-8.15673769e-01 4.96876061e-01 1.82997644e-01 4.74071681e-01
5.49149394e-01 -1.15173495e+00 3.90270382e-01 4.66395468e-01
-1.60947859e-01 9.53254342e-01 5.98888040e-01 -8.12452078e-01
-8.83991122e-01 -1.38377571e+00 9.83422458e-01 -1.03076160e+00
5.24698436e-01 2.41458137e-02 -9.43430603e-01 1.17146838e+00
-1.73971042e-01 -4.28702235e-01 1.29453111e+00 2.09303536e-02
-3.08809280e-01 1.24072120e-01 -1.20706260e+00 6.56347871e-01
4.94958043e-01 -7.43058145e-01 -7.99668372e-01 2.44857371e-01
8.26830745e-01 -4.48432982e-01 -1.20007145e+00 6.70807004e-01
5.67952275e-01 -8.94855142e-01 7.27622569e-01 -1.23976445e+00
9.56326246e-01 2.79150605e-02 -2.96554994e-02 -1.25730848e+00
-4.47472543e-01 -6.28041506e-01 -2.38730878e-01 7.82618165e-01
7.30352044e-01 -1.13568950e+00 3.96777302e-01 1.27971494e+00
-4.92642187e-02 -1.38727582e+00 -7.05820441e-01 -2.87926756e-02
-2.13901792e-03 -4.93625730e-01 7.83252954e-01 1.07927573e+00
1.10654704e-01 9.04727131e-02 -5.29541433e-01 2.29914948e-01
1.21094562e-01 -1.42644778e-01 3.08395833e-01 -1.10022211e+00
-3.19078833e-01 -5.52436292e-01 -4.14134860e-01 -1.69715241e-01
-1.28866479e-01 -8.84094536e-01 -1.91185638e-01 -1.58287168e+00
4.36228186e-01 -4.37806129e-01 -6.30260050e-01 8.05412948e-01
-1.40260756e-01 -7.32433051e-02 2.05824375e-01 5.27516723e-01
5.70301041e-02 6.01272993e-02 4.31618273e-01 2.73642000e-02
-3.31957281e-01 1.43229067e-01 -7.62323022e-01 6.84682786e-01
8.36689353e-01 -7.57069290e-01 -2.20136747e-01 -4.26266342e-01
3.67635638e-01 4.21585351e-01 2.98704743e-01 -7.14735687e-01
1.92970149e-02 -3.98763455e-02 6.74998283e-01 -5.66401303e-01
2.28712723e-01 -6.93026721e-01 6.96042061e-01 7.64641166e-01
-7.53237188e-01 6.29974961e-01 2.82412201e-01 2.65190095e-01
-1.80277340e-02 6.04090132e-02 6.52199447e-01 -2.30600551e-01
1.60564810e-01 2.31805608e-01 -7.65180588e-01 4.56358902e-02
7.40046561e-01 -1.19054243e-01 -1.56692028e-01 -4.68945682e-01
-8.73236656e-01 1.44726411e-01 3.55898023e-01 4.83719409e-01
8.02955627e-01 -8.98358285e-01 -6.99226141e-01 2.85772234e-01
-7.69486427e-02 9.30816457e-02 2.90103316e-01 9.29306805e-01
-4.55225646e-01 5.20908713e-01 -7.09576309e-02 -2.69200504e-01
-1.32051337e+00 5.89297473e-01 6.51443243e-01 -4.27914679e-01
-9.94947672e-01 8.20457578e-01 4.38796997e-01 -6.03195071e-01
2.32628062e-01 -3.10013205e-01 -2.84722596e-01 1.24879979e-01
6.72548831e-01 4.51349288e-01 -2.84905732e-02 -7.08317816e-01
-3.24317932e-01 7.50077516e-02 -1.86558917e-01 1.26246393e-01
1.55694902e+00 2.54796147e-01 -3.21066469e-01 8.78335953e-01
1.03506052e+00 -4.54185933e-01 -9.54435885e-01 -5.26590422e-02
-1.80034727e-01 1.18579127e-01 -2.38027275e-01 -1.00905347e+00
-1.03060520e+00 8.64425480e-01 3.27765942e-01 8.95971879e-02
1.02766252e+00 -1.44059315e-01 6.04824603e-01 7.02396184e-02
3.18722352e-02 -7.57188976e-01 -3.91949266e-01 3.83248866e-01
6.29568756e-01 -9.26519632e-01 -2.55626887e-01 4.24791634e-01
-9.29415286e-01 1.19993973e+00 1.95653275e-01 2.67321736e-01
6.51512504e-01 2.64522940e-01 6.30616426e-01 -2.76514381e-01
-1.13557994e+00 4.61833119e-01 2.29510233e-01 3.02745193e-01
6.77019417e-01 4.94911015e-01 1.80537224e-01 7.19410062e-01
-4.44821298e-01 -3.90543975e-02 6.10668421e-01 6.48399472e-01
1.19420767e-01 -6.83150589e-01 -4.06554163e-01 1.31353819e+00
-7.75319278e-01 -4.68773335e-01 -3.26907486e-01 2.28416592e-01
-1.42994210e-01 7.16272116e-01 9.28304568e-02 -6.91477418e-01
2.11771697e-01 4.58672047e-01 -1.70026854e-01 -8.42735231e-01
-1.14509022e+00 -2.99345911e-01 -3.50636132e-02 -5.37988961e-01
7.48202950e-02 -7.68590331e-01 -1.35614908e+00 -1.37788087e-01
1.08177871e-01 1.07064180e-01 4.95961189e-01 8.21158171e-01
6.03261113e-01 8.82487714e-01 4.02994961e-01 -5.18579304e-01
-6.91883624e-01 -9.85254705e-01 -4.91560906e-01 4.51485753e-01
7.50525773e-01 -1.77073807e-01 -5.56276321e-01 -2.83760261e-02] | [8.031935691833496, 6.146010398864746] |
03b3dc2d-4f14-4772-959d-07a65ba69b35 | a-geometrically-constrained-point-matching | 2211.03007 | null | https://arxiv.org/abs/2211.03007v1 | https://arxiv.org/pdf/2211.03007v1.pdf | A Geometrically Constrained Point Matching based on View-invariant Cross-ratios, and Homography | In computer vision, finding point correspondence among images plays an important role in many applications, such as image stitching, image retrieval, visual localization, etc. Most of the research worksfocus on the matching of local feature before a sampling method is employed, such as RANSAC, to verify initial matching results via repeated fitting of certain global transformation among the images. However, incorrect matches may still exist, while careful examination of such problems is often skipped. Accordingly, a geometrically constrained algorithm is proposed in this work to verify the correctness of initially matched SIFT keypoints based on view-invariant cross-ratios (CRs). By randomly forming pentagons from these keypoints and matching their shape and location among images with CRs, robust planar region estimation can be achieved efficiently for the above verification, while correct and incorrect matches of keypoints can be examined easily with respect to those shape and location matched pentagons. Experimental results show that satisfactory results can be obtained for various scenes with single as well as multiple planar regions. | ['Jen-Hui Chuang', 'Chen-Tao Hsu', 'Ching-Huai Yang', 'Yueh-Cheng Huang'] | 2022-11-06 | null | null | null | null | ['image-stitching', 'visual-localization'] | ['computer-vision', 'computer-vision'] | [ 2.08913743e-01 -6.97964489e-01 -9.18867141e-02 -1.49076238e-01
-5.78653157e-01 -7.55837500e-01 6.39425397e-01 1.83489904e-01
-3.15153480e-01 2.74735212e-01 -3.37627262e-01 -1.85846686e-01
-3.98250401e-01 -6.57209516e-01 -6.31693721e-01 -6.35312259e-01
1.47183135e-01 5.09809315e-01 3.81169438e-01 5.83680440e-03
7.63615727e-01 9.91999149e-01 -1.52926588e+00 -3.30127388e-01
7.61933863e-01 8.56298447e-01 3.51121396e-01 4.00275677e-01
-1.51101733e-02 5.13814799e-02 -4.11271542e-01 -3.89977604e-01
6.26246393e-01 -1.76712707e-01 -4.77969646e-01 6.22577608e-01
5.02161264e-01 -4.47668701e-01 -1.27897426e-01 1.52927852e+00
1.55996028e-02 1.97775051e-01 5.54029167e-01 -1.36431766e+00
-2.38477200e-01 -5.85624166e-02 -1.09372175e+00 -1.85672402e-01
6.96829736e-01 -2.10253388e-01 6.83487535e-01 -1.11288106e+00
5.75854480e-01 1.11334908e+00 6.19212627e-01 -2.95576662e-01
-9.67765450e-01 -7.35390663e-01 -2.32280031e-01 3.85695398e-01
-1.88288438e+00 -3.27210426e-01 1.00591660e+00 -9.59812924e-02
3.25165004e-01 5.05526006e-01 6.10637426e-01 -1.69898383e-02
3.34242761e-01 4.92302895e-01 9.30773020e-01 -7.06791937e-01
-1.66619405e-01 6.33975118e-02 -4.19705845e-02 8.35469246e-01
3.75289768e-01 7.85926729e-02 -1.22021278e-02 -2.80311733e-01
1.25169611e+00 5.75595737e-01 -4.29673016e-01 -7.90896297e-01
-1.57693851e+00 6.44713163e-01 6.65880203e-01 3.83096904e-01
-3.69315147e-01 -1.85276106e-01 2.86739077e-02 3.42797264e-02
-1.57260999e-01 2.99511671e-01 -9.16642845e-02 2.54953653e-01
-8.33286762e-01 1.74713165e-01 1.65808260e-01 1.44074607e+00
1.19343102e+00 -3.05008262e-01 3.79463524e-01 6.78050339e-01
5.86035550e-02 8.27087641e-01 3.40871155e-01 -7.86085308e-01
4.40898448e-01 7.17393100e-01 3.48222792e-01 -1.80997813e+00
-1.77808329e-01 -1.54145015e-03 -1.06266093e+00 1.67267516e-01
3.66627038e-01 3.85858029e-01 -6.59015715e-01 7.74046242e-01
6.54863894e-01 1.68236405e-01 -1.64347768e-01 9.03729200e-01
4.88310426e-01 6.25360727e-01 -5.84495723e-01 -3.04873437e-01
1.49537480e+00 -6.88840091e-01 -4.58533734e-01 -3.13634239e-02
1.74725696e-01 -1.41551459e+00 6.09616280e-01 3.65554184e-01
-8.58949661e-01 -7.70324111e-01 -9.54215169e-01 9.51747149e-02
-9.36588645e-02 2.77127117e-01 3.73234808e-01 2.50942409e-01
-7.95134008e-01 3.51985216e-01 -5.10164380e-01 -3.03169876e-01
-2.08837509e-01 3.41949522e-01 -7.04502583e-01 -4.45204139e-01
-6.90651953e-01 6.24770522e-01 3.18797022e-01 2.92633086e-01
-2.59275675e-01 -3.40327799e-01 -8.93108606e-01 -1.77348420e-01
2.66852707e-01 -4.18510288e-01 8.28818440e-01 -7.91619301e-01
-9.58314002e-01 8.45603824e-01 -4.93302912e-01 -1.11748651e-01
4.65023845e-01 1.41951680e-01 -3.82679880e-01 3.24689329e-01
4.19928253e-01 2.75493622e-01 1.01655459e+00 -1.35396278e+00
-8.61296713e-01 -5.15583754e-01 -1.75597697e-01 3.28095675e-01
1.78787872e-01 1.72475263e-01 -6.19023383e-01 -3.44293326e-01
9.20994043e-01 -9.50823307e-01 -3.75321180e-01 2.10306704e-01
-4.83318061e-01 -9.35563371e-02 9.81023192e-01 -5.19977033e-01
8.11823308e-01 -2.22689462e+00 -3.07112128e-01 8.37770164e-01
-6.78036883e-02 5.93900457e-02 6.19461946e-02 4.27713186e-01
-1.50094643e-01 -1.98546514e-01 6.37404695e-02 1.07119434e-01
-4.13931876e-01 -1.25010595e-01 -2.48409852e-01 9.39968169e-01
-1.06106706e-01 4.46583986e-01 -7.88994849e-01 -6.55459285e-01
7.15681791e-01 2.77408093e-01 -2.47799098e-01 1.47534549e-01
4.61186469e-01 3.39194387e-01 -6.87726021e-01 7.82636344e-01
1.33144426e+00 1.55066982e-01 -1.91274107e-01 -6.06151819e-01
-4.97483045e-01 -4.15430158e-01 -1.61377203e+00 1.14368379e+00
-3.10999572e-01 3.45969707e-01 -6.07076217e-04 -8.08596551e-01
1.14664042e+00 1.09845303e-01 7.37058043e-01 -5.12957573e-01
2.29964808e-01 2.82813609e-01 -1.81213826e-01 -3.12904656e-01
7.59076238e-01 3.12230706e-01 1.49579883e-01 1.71948850e-01
-6.74122214e-01 -4.33679074e-01 1.85430553e-02 -4.49139401e-02
5.48398793e-01 -2.45244116e-01 7.77112782e-01 -2.07367986e-01
9.94116426e-01 3.69623065e-01 3.40308040e-01 3.73104841e-01
-5.19830873e-03 7.99102128e-01 -2.36094460e-01 -5.10790467e-01
-1.28300798e+00 -8.33536267e-01 -3.46112698e-01 8.79437700e-02
1.00745690e+00 -2.73234844e-01 -5.35377264e-01 -2.34718785e-01
-1.12901367e-02 5.15996888e-02 -1.36961088e-01 5.46731353e-02
-6.30982041e-01 -8.80071893e-02 -2.28455916e-04 2.14176774e-01
9.93304968e-01 -7.44244933e-01 -7.45724082e-01 7.81409591e-02
-1.78254947e-01 -1.10072744e+00 -7.12018669e-01 -3.18369240e-01
-7.70814300e-01 -1.52810526e+00 -8.24803472e-01 -1.05733383e+00
1.22374237e+00 1.52323914e+00 5.24081767e-01 5.27340174e-01
-3.54641527e-01 3.19414169e-01 -4.17074561e-01 -4.41188887e-02
-1.75668880e-01 -5.27592242e-01 1.26189172e-01 1.70028478e-01
1.78894669e-01 -1.77906334e-01 -7.70012438e-01 1.13855207e+00
-9.33576286e-01 -1.24788783e-01 5.41750610e-01 6.71678722e-01
8.06238711e-01 6.47256970e-01 -9.62185562e-02 -2.49461904e-01
1.60774678e-01 1.34577781e-01 -1.00758934e+00 4.04776812e-01
-2.15677425e-01 -2.10742578e-01 5.78898132e-01 -3.08044851e-01
-6.78020835e-01 3.05651605e-01 1.17771305e-01 -5.78150749e-01
-1.42981037e-01 3.26309025e-01 -1.12453461e-01 -8.00382018e-01
3.27174067e-01 5.13319492e-01 1.91006333e-01 8.27399362e-03
3.08953047e-01 7.16314971e-01 7.26123273e-01 -3.21792960e-01
1.32033587e+00 7.65805304e-01 3.03370714e-01 -1.13081896e+00
-1.01886369e-01 -1.05403399e+00 -8.45683932e-01 -3.01710010e-01
6.22242093e-01 -6.77610397e-01 -6.38522625e-01 5.16114414e-01
-1.21462083e+00 7.50150383e-01 2.94465989e-01 7.44594514e-01
-5.55988967e-01 8.86905849e-01 -3.70071828e-02 -6.18468821e-01
-2.80896276e-01 -1.47557640e+00 1.04367721e+00 2.99047053e-01
-9.80525538e-02 -8.89756441e-01 -1.73896089e-01 1.21812634e-01
1.60666019e-01 4.22835201e-02 7.40963161e-01 -2.63613760e-01
-1.03988242e+00 -7.45877743e-01 -3.63987565e-01 -9.74782035e-02
4.56984401e-01 4.16751385e-01 -4.16334540e-01 -4.50472862e-01
5.83958775e-02 2.57002562e-01 1.08058505e-01 4.51621711e-01
7.63848305e-01 -8.95776227e-02 -6.37476504e-01 6.62038386e-01
1.58317482e+00 4.75124508e-01 7.84458518e-01 6.46259725e-01
5.72359622e-01 5.16296983e-01 1.18828845e+00 3.40735525e-01
1.71481088e-01 6.47737205e-01 5.04733562e-01 -2.06716061e-01
3.73557419e-01 -1.97099134e-01 1.12596038e-03 4.94731486e-01
-7.37738609e-02 2.27753311e-01 -6.62337959e-01 4.47519958e-01
-1.57073224e+00 -9.22916234e-01 -4.93279248e-01 2.70261955e+00
3.43264908e-01 -2.47505620e-01 -1.26502842e-01 3.79302621e-01
1.26951945e+00 2.66460776e-02 -2.16744319e-01 7.54455477e-03
1.37606636e-01 -2.82142133e-01 6.46825194e-01 5.37506461e-01
-9.69533980e-01 8.07449222e-01 5.60389471e+00 9.54041064e-01
-1.17870319e+00 -5.28603733e-01 1.55864313e-01 8.14699769e-01
-2.65438437e-01 4.65313017e-01 -9.45123672e-01 2.94234663e-01
-2.75957674e-01 -1.53703049e-01 4.30843353e-01 8.13298523e-01
7.47905225e-02 -5.92449963e-01 -6.53506815e-01 1.46005106e+00
2.59575397e-01 -1.08590364e+00 1.38694033e-01 8.87169316e-02
8.56170356e-01 -6.01237535e-01 3.62245291e-02 -4.04869437e-01
-1.08097605e-01 -4.77214485e-01 5.68573058e-01 3.18321973e-01
6.73038423e-01 -9.93473828e-01 6.54773593e-01 4.82897550e-01
-1.55864179e+00 2.46545434e-01 -8.26807320e-01 3.07810694e-01
2.40987595e-02 4.93500859e-01 -9.49662983e-01 8.68181229e-01
4.60189670e-01 5.77100992e-01 -4.75932866e-01 1.54403019e+00
-6.73304487e-04 -3.48053426e-01 -4.91784096e-01 6.13638163e-02
1.16462722e-01 -7.64069319e-01 5.75068176e-01 4.77795213e-01
7.87132919e-01 1.37413666e-01 3.75944853e-01 5.84194839e-01
3.57647806e-01 4.97557998e-01 -8.28457296e-01 5.66750705e-01
6.70364976e-01 1.40677977e+00 -1.05330992e+00 -2.92397946e-01
-2.92817265e-01 9.51994061e-01 -2.21466750e-01 2.27085665e-01
-8.13005030e-01 -5.81078291e-01 3.01957577e-01 2.19322756e-01
1.93472847e-01 -5.29117584e-01 8.88350233e-02 -1.01952720e+00
6.61937669e-02 -8.37091446e-01 2.08921671e-01 -9.75673497e-01
-7.96490490e-01 5.10104597e-01 1.63945153e-01 -1.99895549e+00
-3.21316481e-01 -2.54095674e-01 -7.23101139e-01 7.61775136e-01
-1.33462441e+00 -1.11615777e+00 -5.11427522e-01 1.07273054e+00
4.81951654e-01 1.60730556e-02 3.90964389e-01 -3.48842293e-02
-1.64931953e-01 4.99745786e-01 3.24882805e-01 1.43645003e-01
7.36155748e-01 -6.94093525e-01 2.06453100e-01 1.03416181e+00
1.96373805e-01 8.63610446e-01 7.13661194e-01 -6.82402253e-01
-1.46340144e+00 -8.15510571e-01 8.03994358e-01 -4.52075526e-02
5.12915969e-01 1.49739772e-01 -7.71750867e-01 4.87729847e-01
8.47291276e-02 8.47235769e-02 -7.11462498e-02 -5.43239236e-01
-4.45075221e-02 -1.43292680e-01 -1.17213082e+00 7.21686304e-01
6.37141764e-01 -3.30715239e-01 -3.54993850e-01 4.17812854e-01
1.58756286e-01 -6.23355329e-01 -7.16721117e-01 4.92397249e-01
6.03075564e-01 -1.04319429e+00 1.25344062e+00 7.48189464e-02
1.51781021e-02 -8.89047563e-01 -1.10322841e-01 -1.11769783e+00
-3.02604795e-01 -3.21670383e-01 8.30132484e-01 1.26346147e+00
-2.13235542e-01 -7.10467160e-01 6.71712875e-01 2.87827224e-01
1.03025749e-01 -2.64641672e-01 -8.70470822e-01 -8.36891413e-01
-6.62440777e-01 -5.14977984e-03 7.18529403e-01 9.79179323e-01
-2.22119600e-01 -1.52326822e-01 -3.03221136e-01 7.14961946e-01
8.34996760e-01 6.72787368e-01 1.31789935e+00 -1.06502187e+00
2.19696984e-01 -2.12655112e-01 -8.93600345e-01 -1.23971236e+00
-1.31553516e-01 -5.07686317e-01 6.06667586e-02 -1.26879394e+00
9.93826240e-02 -7.43382335e-01 2.54948765e-01 9.08673392e-04
-1.05239689e-01 4.19391155e-01 6.92905560e-02 5.30674160e-01
-2.82692909e-01 2.83266723e-01 1.30023491e+00 1.96074042e-03
-6.42741323e-02 2.85700649e-01 -2.91390151e-01 9.59036231e-01
5.43677449e-01 -1.88853234e-01 -2.50322461e-01 -1.76335305e-01
-1.19468495e-01 3.21073294e-01 4.68930006e-01 -1.07647026e+00
5.70371807e-01 -3.59937429e-01 4.88164663e-01 -1.17181098e+00
2.63216078e-01 -1.45687747e+00 5.44897497e-01 4.13245499e-01
1.51836142e-01 6.63254201e-01 -1.08195722e-01 4.26297605e-01
-4.07256633e-01 -6.20804727e-01 7.62157679e-01 -1.50497675e-01
-8.68086755e-01 4.03446555e-01 -2.02261657e-02 -5.12584329e-01
1.43128121e+00 -8.62288415e-01 -1.21944591e-01 -4.65731829e-01
-1.50933236e-01 7.21020177e-02 9.04855847e-01 9.05211568e-02
1.18275201e+00 -1.57601762e+00 -4.71084595e-01 7.72458553e-01
4.00032640e-01 2.22643331e-01 3.42926025e-01 8.75995219e-01
-1.00008547e+00 4.23756510e-01 -1.20354883e-01 -9.91629601e-01
-1.72330022e+00 8.07353437e-01 1.48198873e-01 1.60407335e-01
-4.65823263e-01 3.91932666e-01 1.05319254e-01 -1.70274034e-01
-5.56224994e-02 -4.00073975e-01 -1.14406437e-01 -2.14369252e-01
5.77612638e-01 3.51945132e-01 8.84063840e-02 -1.05299568e+00
-3.58662874e-01 1.55148983e+00 -1.45812616e-01 1.66714787e-01
7.37615347e-01 -3.47208053e-01 -2.48143464e-01 -2.18261018e-01
1.36110616e+00 5.21349192e-01 -9.24870491e-01 -4.81716931e-01
-3.45108658e-01 -1.25057149e+00 -2.34992534e-01 5.50596267e-02
-8.66400540e-01 5.57947278e-01 3.82036537e-01 2.25354552e-01
1.04782224e+00 -2.28768140e-01 6.50496721e-01 2.78539628e-01
9.21148598e-01 -5.01831710e-01 -8.64493325e-02 2.85064220e-01
7.94922531e-01 -1.24209845e+00 3.44557554e-01 -7.64627337e-01
-3.77867550e-01 1.37345803e+00 4.03527170e-01 -5.02023816e-01
4.76689875e-01 -9.72814113e-02 -7.41675273e-02 -2.34700833e-02
4.56494130e-02 -6.85076937e-02 3.52563888e-01 5.60375333e-01
-2.38625612e-02 -2.32916161e-01 -1.64244831e-01 -3.33558291e-01
-2.20294401e-01 -4.86394048e-01 5.18906057e-01 9.64522421e-01
-6.65249646e-01 -9.06331897e-01 -1.16284859e+00 6.00257888e-02
-1.06191024e-01 -1.51749868e-02 7.32218055e-03 1.10862529e+00
-2.36049622e-01 8.67916584e-01 2.64060646e-01 -4.25317645e-01
4.08894300e-01 -7.67990112e-01 5.45865059e-01 -1.15000255e-01
-1.70414329e-01 3.71635228e-01 -4.53117311e-01 -4.08234239e-01
-5.16237199e-01 -8.04082096e-01 -1.21929097e+00 -3.43674570e-01
-7.26501584e-01 3.60855103e-01 6.75294101e-01 7.67505825e-01
-1.01094775e-01 -3.29201996e-01 1.15885341e+00 -9.24343586e-01
-5.40980577e-01 -2.91651189e-01 -6.96693063e-01 3.98949623e-01
3.43887389e-01 -6.31709874e-01 -4.88827169e-01 -9.84548591e-03] | [8.130913734436035, -2.3181512355804443] |
4c60a79c-2736-41c6-b02b-dd6a04878bd7 | 4d-millimeter-wave-radar-in-autonomous | 2306.04242 | null | https://arxiv.org/abs/2306.04242v2 | https://arxiv.org/pdf/2306.04242v2.pdf | 4D Millimeter-Wave Radar in Autonomous Driving: A Survey | The 4D millimeter-wave (mmWave) radar, capable of measuring the range, azimuth, elevation, and velocity of targets, has attracted considerable interest in the autonomous driving community. This is attributed to its robustness in extreme environments and outstanding velocity and elevation measurement capabilities. However, despite the rapid development of research related to its sensing theory and application, there is a notable lack of surveys on the topic of 4D mmWave radar. To address this gap and foster future research in this area, this paper presents a comprehensive survey on the use of 4D mmWave radar in autonomous driving. Reviews on the theoretical background and progress of 4D mmWave radars are presented first, including the signal processing flow, resolution improvement ways, extrinsic calibration process, and point cloud generation methods. Then it introduces related datasets and application algorithms in autonomous driving perception and localization and mapping tasks. Finally, this paper concludes by predicting future trends in the field of 4D mmWave radar. To the best of our knowledge, this is the first survey specifically for the 4D mmWave radar. | ['Jianqiang Wang', 'Shaobing Xu', 'Lei He', 'Shuocheng Yang', 'Zikun Xu', 'Jiahao Wang', 'Zeyu Han'] | 2023-06-07 | null | null | null | null | ['point-cloud-generation'] | ['computer-vision'] | [ 1.98013276e-01 -3.07774395e-01 2.60152936e-01 -6.39386773e-01
-4.49778110e-01 -5.65977991e-01 5.25522232e-01 -4.86524850e-01
-2.86104530e-01 5.74458957e-01 -1.90075755e-01 -4.04721588e-01
-3.78023148e-01 -1.12976825e+00 6.51063956e-03 -8.21659982e-01
-3.13310415e-01 3.57113093e-01 -1.27071798e-01 -3.00767928e-01
3.56796682e-01 1.13639772e+00 -1.54223585e+00 -5.47805130e-01
7.49400020e-01 1.15111172e+00 2.88771927e-01 5.41513860e-01
1.07638165e-01 7.00824782e-02 -4.95193422e-01 -8.31874013e-02
7.19848871e-01 -1.69172317e-01 4.09560025e-01 -3.61061871e-01
6.54305160e-01 -2.74440438e-01 -6.56993330e-01 8.29417109e-01
1.01314461e+00 -1.19529024e-01 8.40734363e-01 -1.22371674e+00
-2.59961545e-01 -1.46929361e-02 -5.92157841e-01 5.05995989e-01
8.92801583e-02 6.21949583e-02 4.27304953e-01 -1.01777709e+00
1.73816562e-01 7.20472038e-01 5.89283526e-01 1.31893367e-01
-4.09998357e-01 -1.04607415e+00 -3.44108045e-01 2.54321873e-01
-1.37814355e+00 -4.29284811e-01 6.58939838e-01 -6.27911806e-01
8.43218982e-01 -1.58402398e-02 6.47431850e-01 6.94163442e-01
8.50284874e-01 5.40904095e-03 1.22578824e+00 -5.05911171e-01
2.78512776e-01 -8.85834396e-02 3.83711994e-01 3.20125163e-01
9.70721304e-01 7.93957591e-01 -6.48788869e-01 2.01637715e-01
4.06339228e-01 -1.69536784e-01 -2.77509123e-01 -5.41274130e-01
-6.58747196e-01 8.35262597e-01 1.69676855e-01 4.74040926e-01
-6.36050165e-01 2.57382721e-01 -4.09962535e-01 6.11480713e-01
3.59116137e-01 6.09503925e-01 -2.00018048e-01 -9.59372148e-02
-9.60310400e-01 6.23348176e-01 4.84864473e-01 1.20131290e+00
9.11605716e-01 5.48497677e-01 2.53044009e-01 4.86597419e-01
6.58656657e-01 1.76321638e+00 -2.21288905e-01 -5.16075075e-01
3.86673927e-01 1.84673026e-01 1.14960290e-01 -9.70102370e-01
-6.98402405e-01 -1.30332816e+00 -5.99594057e-01 5.57253420e-01
-2.15302765e-01 -6.84410810e-01 -1.23042810e+00 1.14384460e+00
-4.31114845e-02 2.54973173e-01 4.58895802e-01 7.88523972e-01
1.07782996e+00 4.55903798e-01 -4.44416434e-01 -3.17139655e-01
1.08116984e+00 -1.30367130e-01 -7.46218920e-01 -9.88163710e-01
1.90159202e-01 -8.71670485e-01 -1.58957228e-01 2.65372015e-04
-6.54638946e-01 -5.42051613e-01 -1.30593872e+00 6.68012142e-01
-3.80557090e-01 -2.37794980e-01 6.36522055e-01 1.35509861e+00
-5.94395697e-01 -4.20565397e-01 -6.43702030e-01 -5.58915496e-01
2.61968017e-01 2.09259372e-02 1.29451722e-01 -4.96397048e-01
-1.26544416e+00 1.36069715e+00 -2.58029193e-01 3.23202074e-01
-2.42199585e-01 -1.17650986e+00 -7.02158868e-01 -4.38881546e-01
-3.14630687e-01 -4.93559510e-01 1.03779352e+00 4.13649172e-01
-1.05783713e+00 7.77038813e-01 -3.08037311e-01 -7.31813192e-01
-6.72317743e-02 -2.92603225e-01 -9.68741596e-01 -9.75818262e-02
3.75494838e-01 2.61408836e-01 5.73169053e-01 -1.04541540e+00
-1.06477916e+00 -7.75966585e-01 -3.73539954e-01 6.06379583e-02
5.64972699e-01 -5.27939558e-01 3.35980743e-01 -3.05050135e-01
8.11546326e-01 -8.87430131e-01 -3.52650821e-01 -4.95002747e-01
3.06946725e-01 2.38558978e-01 7.22668529e-01 1.93344846e-01
8.40943575e-01 -2.30198431e+00 -4.44709212e-01 4.58272994e-01
1.10917009e-01 -8.77375752e-02 5.38756466e-03 8.39685559e-01
3.02191287e-01 -5.19314468e-01 -3.79656821e-01 1.97744936e-01
-5.34286015e-02 -5.08098025e-03 -6.86979234e-01 5.99590182e-01
-2.00980678e-01 9.70389664e-01 -3.39810133e-01 2.16252118e-01
3.95911038e-01 2.09792390e-01 7.48362392e-02 -1.76534146e-01
3.25647980e-01 1.77837536e-01 -7.90200293e-01 9.68456507e-01
1.60492718e+00 6.75449431e-01 -5.81631720e-01 -5.96264452e-02
-8.76732171e-01 5.83558977e-02 -1.15176833e+00 1.21531975e+00
-2.86220342e-01 1.03762424e+00 1.99475691e-01 -9.79502738e-01
1.78285348e+00 4.72457968e-02 6.27094686e-01 -1.06276166e+00
-1.00355767e-01 5.90499163e-01 1.24229349e-01 -6.45964146e-01
4.31884974e-01 -5.38078845e-01 -2.35887587e-01 3.53139043e-01
-5.38656890e-01 -1.01074851e+00 -1.50579244e-01 -1.33651316e-01
1.17767584e+00 -2.74664044e-01 1.87228516e-01 -1.69378534e-01
3.93187970e-01 2.74600446e-01 3.99650753e-01 9.08910275e-01
-3.44367057e-01 1.22557446e-01 -4.85227078e-01 -3.42111439e-01
-4.87552434e-01 -1.22707653e+00 -6.14070475e-01 2.63289720e-01
5.61370373e-01 5.11501618e-02 1.18640356e-01 9.77846421e-03
7.64848650e-01 7.42495656e-01 -2.26564124e-01 8.03630352e-02
-4.61958468e-01 -1.09994805e+00 4.92980897e-01 2.27612585e-01
7.73712695e-01 -4.33842897e-01 -1.13305295e+00 2.48765260e-01
-1.72019094e-01 -1.20881426e+00 6.10731184e-01 3.28998566e-01
-1.10938168e+00 -8.72629225e-01 -2.16134086e-01 -6.61630213e-01
-5.09479605e-02 1.16010273e+00 8.40540469e-01 -6.76562011e-01
-4.21028018e-01 7.44321764e-01 -4.81450111e-01 -1.20534289e+00
3.13724756e-01 -1.19246580e-01 2.14058936e-01 -2.23978579e-01
1.26673889e+00 -9.38472211e-01 -5.09245217e-01 2.31378213e-01
-9.76205915e-02 -4.38471556e-01 1.19090319e+00 -4.34133224e-02
2.74453133e-01 1.00281328e-01 7.60057747e-01 -5.31212807e-01
3.43617886e-01 -3.70615572e-01 -8.83221388e-01 -5.10551453e-01
-8.17876756e-01 -2.37879500e-01 -1.70394287e-01 6.37047887e-01
-7.20787048e-01 8.27975720e-02 -8.61611292e-02 1.04129903e-01
-3.90708864e-01 4.94323403e-01 -1.03774562e-03 -7.54233897e-01
8.82833421e-01 3.40588957e-01 -7.02391565e-02 -2.70768851e-01
1.22345626e-01 8.90918612e-01 4.36486125e-01 2.20166281e-01
1.67795765e+00 7.70576358e-01 6.62359595e-01 -1.48158562e+00
-6.29661858e-01 -8.81437838e-01 -6.20554864e-01 -6.04580820e-01
6.18297994e-01 -1.02305698e+00 -1.08644977e-01 2.62445152e-01
-9.17801142e-01 2.05817983e-01 1.03900942e-03 1.20610952e+00
-5.28309882e-01 -1.56108290e-01 2.14443460e-01 -1.09763384e+00
-2.09935248e-01 -7.02139258e-01 6.94868147e-01 1.64551809e-01
-4.42263819e-02 -5.73118925e-01 4.86937195e-01 4.26806062e-01
8.39711368e-01 4.07828301e-01 4.75046664e-01 3.14434677e-01
-1.13469076e+00 -5.71639299e-01 -1.14836849e-01 -3.66996020e-01
-5.93547635e-02 -6.82894647e-01 -1.09821165e+00 -1.87133685e-01
2.43530273e-01 3.36370796e-01 9.53077674e-01 1.17770565e+00
-1.03518419e-01 5.26786447e-01 -7.91821957e-01 1.02283990e+00
1.76807833e+00 7.55040109e-01 6.88737035e-01 4.36372817e-01
-8.64844471e-02 5.58527648e-01 1.15006244e+00 3.10723692e-01
-3.61698084e-02 3.82028729e-01 6.62515700e-01 2.36745760e-01
-1.64598390e-01 2.72476166e-01 7.81926066e-02 2.68419564e-01
-4.46813971e-01 -1.90022215e-01 -7.66081750e-01 2.84277320e-01
-1.42223871e+00 -1.35859907e+00 -5.93091071e-01 1.64975572e+00
-5.21194696e-01 3.39458883e-02 -4.74922895e-01 2.01362148e-01
4.05034184e-01 5.82295299e-01 -3.91307861e-01 -1.35698766e-01
-2.69257456e-01 5.88688433e-01 1.16379309e+00 7.49860644e-01
-9.46807563e-01 6.76564455e-01 6.53211355e+00 2.31506079e-01
-1.08395779e+00 -1.37582421e-01 -4.70596939e-01 9.99369621e-02
-3.00921053e-01 -9.38971490e-02 -1.27212965e+00 4.28188629e-02
5.88986099e-01 -2.51517296e-01 -1.81547552e-01 6.05054617e-01
4.04932588e-01 -4.54809010e-01 -5.60194850e-01 1.39592981e+00
1.68121815e-01 -1.25879562e+00 -3.43191922e-01 4.17373478e-01
5.55801272e-01 4.45298642e-01 2.79857367e-01 2.28004321e-01
-9.26245749e-02 -8.62861216e-01 4.68538851e-01 5.51763475e-01
6.48687720e-01 -1.02222121e+00 1.04666245e+00 3.81696820e-01
-1.28921056e+00 -1.12344511e-02 -7.34851122e-01 -5.87930202e-01
4.17350411e-01 1.20724332e+00 -8.43738019e-01 7.11243093e-01
5.46365857e-01 8.28038692e-01 -6.38060495e-02 1.29276383e+00
-3.12628478e-01 4.28578138e-01 -2.05790505e-01 -3.64385486e-01
5.91320097e-01 -4.84297067e-01 9.25930381e-01 1.20685375e+00
9.82814372e-01 6.14274621e-01 2.47329324e-02 6.17802203e-01
6.13339365e-01 -2.84862280e-01 -1.25195670e+00 3.10492307e-01
8.54140162e-01 1.20192492e+00 -2.61263475e-02 3.49765986e-01
-4.97394204e-01 1.02311121e-02 -6.80130661e-01 5.50766766e-01
-4.91944730e-01 -9.81710672e-01 1.20366645e+00 5.24913132e-01
2.46078953e-01 -1.10189462e+00 -8.32393646e-01 -2.57483363e-01
-1.21232845e-01 -8.72924924e-02 2.03900076e-02 -7.37212598e-01
-9.22590911e-01 3.01240206e-01 7.59014115e-02 -1.58842480e+00
-1.77816078e-01 -5.89588165e-01 -3.73927683e-01 1.08104515e+00
-2.03871179e+00 -7.16337144e-01 -9.64354455e-01 1.40360162e-01
3.43696415e-01 -6.64688528e-01 6.77589655e-01 1.93063483e-01
3.42771173e-01 1.18395155e-02 1.27065942e-01 -2.36950874e-01
8.17081869e-01 -6.33029222e-01 5.94127953e-01 8.51916432e-01
-2.10624442e-01 1.88763022e-01 1.10917079e+00 -8.07581007e-01
-1.85629892e+00 -1.17854905e+00 1.06364071e+00 -3.56534272e-01
3.23660016e-01 -3.91735196e-01 -5.15667498e-02 6.24525785e-01
3.70703526e-02 -5.08886993e-01 7.33110487e-01 -1.27696693e-02
3.24074328e-02 -5.73940635e-01 -1.21950758e+00 3.51254016e-01
1.33354974e+00 5.19464500e-02 -7.29792535e-01 2.36841455e-01
-2.60495935e-02 -3.38750660e-01 -4.98855948e-01 8.96813989e-01
9.10791516e-01 -1.03934455e+00 1.00863385e+00 2.91990280e-01
-7.30430558e-02 -4.72118169e-01 -6.52535319e-01 -1.26517713e+00
-9.54935491e-01 -2.21635535e-01 1.85484052e-01 4.65365648e-01
2.94541866e-01 -1.04974139e+00 1.28507864e+00 -3.64495248e-01
-4.76971835e-01 -3.72878581e-01 -1.17529082e+00 -9.05252099e-01
-1.07957453e-01 -8.98677945e-01 5.39881945e-01 3.36877257e-01
-2.00446829e-01 6.21050656e-01 -1.00047007e-01 1.06743252e+00
1.40849209e+00 6.88058197e-01 1.07691038e+00 -1.75311220e+00
2.75619984e-01 -2.45008305e-01 -1.02607238e+00 -1.42793477e+00
-1.50130093e-01 -8.26809227e-01 2.02469990e-01 -2.00079751e+00
-5.99682570e-01 -8.00369382e-01 3.71587813e-01 -2.97229737e-01
4.44815785e-01 5.32551527e-01 -1.41536772e-01 2.03757033e-01
3.90580505e-01 3.37199926e-01 1.05204606e+00 -1.95384189e-01
-3.22142661e-01 7.36272931e-01 -5.34430444e-01 4.13999498e-01
1.22628295e+00 -3.93873215e-01 -4.71835703e-01 -5.12008548e-01
2.14267045e-01 1.60028771e-01 1.31248102e-01 -1.59942007e+00
6.23773277e-01 -4.16459888e-01 5.55914402e-01 -1.58485055e+00
7.13900864e-01 -8.74526978e-01 7.76185989e-02 7.33072519e-01
6.95305645e-01 -1.11759342e-01 -4.61944286e-03 6.82794213e-01
-2.49452978e-01 -2.57939875e-01 1.04475427e+00 -7.98942596e-02
-1.15107882e+00 6.81188703e-01 -1.00091863e+00 -1.49091050e-01
1.33115613e+00 -8.22226405e-01 -3.41141552e-01 -5.29065251e-01
-2.32757673e-01 1.87010884e-01 -1.13173857e-01 2.84971565e-01
9.66551244e-01 -1.14066315e+00 -1.15120327e+00 6.64162695e-01
4.98068780e-01 -4.67211515e-01 2.80235648e-01 4.97301310e-01
-4.28497672e-01 1.01769102e+00 -3.46547633e-01 -6.92257464e-01
-9.60619748e-01 -9.63556522e-04 5.07596374e-01 4.62866336e-01
-6.41772866e-01 6.44374371e-01 -2.60662101e-02 -2.16711685e-01
-5.60490154e-02 6.75143674e-03 -2.68373102e-01 -1.39723241e-01
4.69772965e-01 4.58616853e-01 3.52593035e-01 -4.33546335e-01
-9.81777310e-01 1.52418005e+00 3.47777963e-01 -5.00897527e-01
1.32088625e+00 -2.99298465e-01 1.00406684e-01 2.57904500e-01
5.32562017e-01 5.83697259e-01 -7.62071371e-01 1.37429208e-01
-2.36154765e-01 -6.18238688e-01 4.13195282e-01 -5.93032718e-01
-7.74037302e-01 8.66937399e-01 7.26375520e-01 1.56478569e-01
9.59545434e-01 -1.32322252e-01 6.35318339e-01 6.91653073e-01
7.82772064e-01 -8.26298416e-01 -4.21002060e-01 1.02886307e+00
6.85177386e-01 -9.15000081e-01 2.55321473e-01 -5.80283403e-01
9.43680406e-02 1.12275898e+00 2.37090781e-01 -3.03174019e-01
1.19366622e+00 7.93233871e-01 6.00496948e-01 -4.98479694e-01
-2.97381788e-01 -5.17800152e-01 -1.74131930e-01 1.30964351e+00
3.38040888e-01 1.45014092e-01 -3.93894762e-01 8.68471116e-02
-7.14557886e-01 -2.59299606e-01 4.90056276e-01 1.07542157e+00
-1.48899066e+00 -8.23053658e-01 -6.79174662e-01 5.53340256e-01
2.16410890e-01 -1.25550069e-02 -1.64477706e-01 8.37367535e-01
2.12590605e-01 1.29658461e+00 6.57022521e-02 -3.50459784e-01
7.94391096e-01 3.45766470e-02 7.57577956e-01 -4.38976318e-01
1.78421155e-01 -3.51192892e-01 1.74901877e-02 -1.92310810e-01
-3.84005308e-01 -6.47215664e-01 -7.83508062e-01 -2.19021901e-01
-1.67251512e-01 4.04796571e-01 1.18188798e+00 5.69383562e-01
3.17541510e-01 2.98078030e-01 7.40639985e-01 -4.37314928e-01
-2.76271582e-01 -9.35493946e-01 -1.16001868e+00 -7.03490794e-01
4.21172172e-01 -1.02710736e+00 -7.03576386e-01 -4.86157030e-01] | [6.679361820220947, 0.7396187782287598] |
70db2538-e90f-4891-8092-131bffe4fed6 | tanet-a-new-paradigm-for-global-face-super | 2109.08174 | null | https://arxiv.org/abs/2109.08174v1 | https://arxiv.org/pdf/2109.08174v1.pdf | TANet: A new Paradigm for Global Face Super-resolution via Transformer-CNN Aggregation Network | Recently, face super-resolution (FSR) methods either feed whole face image into convolutional neural networks (CNNs) or utilize extra facial priors (e.g., facial parsing maps, facial landmarks) to focus on facial structure, thereby maintaining the consistency of the facial structure while restoring facial details. However, the limited receptive fields of CNNs and inaccurate facial priors will reduce the naturalness and fidelity of the reconstructed face. In this paper, we propose a novel paradigm based on the self-attention mechanism (i.e., the core of Transformer) to fully explore the representation capacity of the facial structure feature. Specifically, we design a Transformer-CNN aggregation network (TANet) consisting of two paths, in which one path uses CNNs responsible for restoring fine-grained facial details while the other utilizes a resource-friendly Transformer to capture global information by exploiting the long-distance visual relation modeling. By aggregating the features from the above two paths, the consistency of global facial structure and fidelity of local facial detail restoration are strengthened simultaneously. Experimental results of face reconstruction and recognition verify that the proposed method can significantly outperform the state-of-the-art methods. | ['Jiayi Ma', 'Zhongyuan Wang', 'JiaMing Wang', 'Junjun Jiang', 'Yanduo Zhang', 'Tao Lu', 'Yuanzhi Wang'] | 2021-09-16 | null | null | null | null | ['face-reconstruction'] | ['computer-vision'] | [-1.68367587e-02 1.33783579e-01 -2.07511205e-02 -6.21602178e-01
-3.63311052e-01 4.93096709e-02 3.76071155e-01 -6.60803378e-01
1.02432445e-01 4.78727251e-01 5.88302076e-01 3.88737977e-01
-1.05116211e-01 -1.03808892e+00 -7.61912525e-01 -8.05598617e-01
5.28401494e-01 -2.19855636e-01 1.54339105e-01 -3.39194864e-01
-3.83587442e-02 7.51163721e-01 -1.78644884e+00 3.93226624e-01
7.07097352e-01 1.23527920e+00 1.86363667e-01 -1.94770291e-01
-3.76418501e-01 7.43696809e-01 -6.52705207e-02 -3.74796569e-01
1.05108306e-01 -3.48820925e-01 -5.58018386e-01 2.72213250e-01
4.17781800e-01 -7.22473860e-01 -6.80983484e-01 1.24493933e+00
4.23736513e-01 6.50497973e-02 1.17465653e-01 -7.92455852e-01
-1.12421882e+00 3.37447792e-01 -7.33179450e-01 8.17317069e-02
2.04294354e-01 4.19668779e-02 6.85863078e-01 -1.22937858e+00
5.84748209e-01 1.68308663e+00 5.37538886e-01 7.84677267e-01
-9.16950643e-01 -1.05637717e+00 4.73827004e-01 1.94800466e-01
-1.59079027e+00 -9.57575858e-01 1.07150555e+00 -4.37424444e-02
4.22214925e-01 3.36776972e-02 4.49513406e-01 7.44465053e-01
-6.76532090e-02 3.36727232e-01 9.17892635e-01 1.36799673e-02
-1.82075217e-01 -7.65179545e-02 -2.36568809e-01 1.04799449e+00
1.02598369e-01 2.14625642e-01 -6.50433481e-01 1.78159863e-01
1.30255258e+00 3.56815934e-01 -3.88095409e-01 -1.18662519e-02
-6.89519405e-01 4.27061826e-01 7.24496067e-01 4.13076848e-01
-5.60900509e-01 2.63037309e-02 6.26928434e-02 -8.52090344e-02
6.33016586e-01 -3.53162944e-01 -2.38762960e-01 4.91119772e-01
-8.42585981e-01 -8.62374995e-03 -8.14073682e-02 8.68595243e-01
1.10450923e+00 3.54904890e-01 -3.00968051e-01 9.53397751e-01
6.33987725e-01 1.42396420e-01 3.91527981e-01 -9.37559247e-01
2.59805202e-01 7.21367359e-01 -1.88779667e-01 -1.21471953e+00
-1.68315023e-01 -4.30790633e-01 -1.15351987e+00 2.83551384e-02
-5.25957458e-02 2.27768943e-01 -9.41982985e-01 1.95293033e+00
6.32490456e-01 3.97173941e-01 -1.55104727e-01 1.01485944e+00
1.27500594e+00 6.59366786e-01 5.79732694e-02 -3.31793576e-01
1.41569448e+00 -8.77587318e-01 -9.27312851e-01 7.52026588e-02
-1.17343470e-01 -6.41910374e-01 7.46489763e-01 -2.67795920e-02
-1.36697567e+00 -9.67244208e-01 -8.09212387e-01 -5.24975717e-01
1.49034232e-01 1.75389826e-01 3.27787638e-01 2.28817865e-01
-1.19154334e+00 6.60393059e-01 -7.56594539e-01 1.67107210e-02
9.00308430e-01 3.63758236e-01 -7.33475089e-01 -4.65647876e-01
-9.89249110e-01 5.13646662e-01 1.43310260e-02 6.78079307e-01
-1.04525661e+00 -7.78097570e-01 -9.46078837e-01 3.77810419e-01
1.51144534e-01 -6.16955698e-01 6.94102168e-01 -9.79979455e-01
-1.65776360e+00 5.82973719e-01 -4.36437994e-01 3.11921179e-01
2.09119946e-01 -2.71358918e-02 -4.27826643e-01 2.25646570e-01
-8.30437914e-02 6.19661629e-01 9.14266765e-01 -1.30042517e+00
-4.77369279e-01 -7.07006991e-01 -1.29123718e-01 5.40556619e-03
-3.98362756e-01 2.33478457e-01 -5.93854606e-01 -6.00319147e-01
4.11979795e-01 -3.54446352e-01 2.22632457e-02 4.10766274e-01
-1.78500488e-01 -3.39818805e-01 9.46694434e-01 -1.05335307e+00
1.07129455e+00 -2.23744321e+00 2.37898663e-01 1.66983828e-01
3.74061197e-01 2.59511024e-01 -4.24583048e-01 -2.57615428e-02
-3.52602988e-01 8.01630914e-02 -9.70729366e-02 -5.01936257e-01
-3.61891031e-01 3.28080118e-01 -2.18794227e-01 4.88894373e-01
4.50101644e-01 9.52211678e-01 -5.64595819e-01 -5.99736810e-01
1.66425526e-01 1.13017583e+00 -6.58298135e-01 3.42516631e-01
2.07729228e-02 7.13454425e-01 -6.53595984e-01 8.65088046e-01
1.16500473e+00 -1.90379232e-01 1.21137835e-01 -7.01881886e-01
-1.83731437e-01 1.83075503e-01 -1.06041658e+00 1.74423599e+00
-4.26806986e-01 1.76581927e-02 4.12706971e-01 -7.71556079e-01
1.21409452e+00 1.99663386e-01 3.58879477e-01 -9.84031856e-01
1.29565954e-01 7.59196188e-03 -3.40399474e-01 -3.75345826e-01
1.05689973e-01 -2.81187594e-01 6.21743977e-01 2.37684458e-01
2.35824585e-01 6.40407503e-01 -3.70045096e-01 -1.38758034e-01
6.14753067e-01 3.42186630e-01 1.12357922e-02 -2.75279045e-01
8.27625155e-01 -8.17251146e-01 9.31705058e-01 -9.92971435e-02
-7.12856948e-02 6.76682413e-01 2.65351027e-01 -7.02185869e-01
-9.69283938e-01 -8.46716404e-01 -7.24365786e-02 1.03597605e+00
2.55926967e-01 -3.54826778e-01 -8.22966099e-01 -4.47447836e-01
-2.02185616e-01 5.18520586e-02 -8.64195287e-01 -3.73477638e-01
-7.88579464e-01 -4.13872987e-01 3.33748043e-01 6.99332952e-01
1.01477754e+00 -1.18711901e+00 -5.10075204e-02 9.57060903e-02
-3.24120492e-01 -1.17656052e+00 -5.70592105e-01 -5.55180967e-01
-9.14902806e-01 -8.51325393e-01 -5.99180460e-01 -8.56124938e-01
9.64782059e-01 4.56711382e-01 7.29876578e-01 5.80765307e-01
-2.19100341e-01 -5.05058207e-02 -2.00483367e-01 2.89930075e-01
7.23571330e-02 -3.91380757e-01 -1.10101424e-01 6.16955101e-01
-3.37241963e-02 -9.52478230e-01 -7.43730009e-01 2.97233105e-01
-9.14490700e-01 1.09467819e-01 7.48197436e-01 8.53571594e-01
9.13855553e-01 3.38718981e-01 3.97197127e-01 -7.43814826e-01
1.57167345e-01 -4.53613669e-01 -4.77594197e-01 3.31402689e-01
-4.01127189e-01 2.02486776e-02 6.20154321e-01 -3.53557348e-01
-1.61620390e+00 -6.65822020e-03 -3.55020970e-01 -7.45596528e-01
-7.13013411e-02 2.08402053e-01 -9.04564679e-01 -3.83142143e-01
1.39296399e-02 6.23130500e-01 1.05289653e-01 -8.09627891e-01
2.17523590e-01 2.79172838e-01 6.53624356e-01 -6.80661440e-01
7.95092225e-01 6.42574668e-01 1.74594745e-01 -4.79574800e-01
-9.35310721e-01 1.15511984e-01 -5.39731503e-01 -1.10526346e-01
8.58052731e-01 -1.07647121e+00 -7.73288012e-01 4.67036217e-01
-1.16814327e+00 1.46011218e-01 -5.44332825e-02 1.32795244e-01
-2.50864208e-01 2.99118370e-01 -8.49636555e-01 -6.16727769e-01
-3.54825914e-01 -1.11423910e+00 1.29070520e+00 5.36511064e-01
5.80406129e-01 -5.41194141e-01 -3.11984360e-01 3.88858408e-01
5.69429696e-01 2.70133018e-01 8.03819597e-01 1.02348983e-01
-9.36620891e-01 2.44520232e-01 -7.87097216e-01 4.39012140e-01
4.18613136e-01 -2.77227582e-03 -1.23699725e+00 -1.04457751e-01
4.88409512e-02 1.06047583e-03 8.17422330e-01 2.63134718e-01
1.58445692e+00 -6.42339110e-01 -1.23839982e-01 8.80843878e-01
1.34398544e+00 -6.69506192e-02 1.06693411e+00 -9.16662887e-02
9.76044714e-01 8.98460269e-01 2.77188361e-01 3.86467338e-01
5.56456208e-01 5.71419954e-01 5.31596065e-01 -1.28168121e-01
-5.55634856e-01 -6.45552933e-01 2.68945098e-01 6.93018436e-01
-5.47858179e-01 3.85884017e-01 -2.69830465e-01 2.06618175e-01
-1.62136579e+00 -1.02612090e+00 3.68423074e-01 1.86051202e+00
8.11567426e-01 -3.87940735e-01 -3.50527078e-01 -1.31312042e-01
9.05078650e-01 3.26285660e-01 -5.35189211e-01 1.45160198e-01
-1.83673710e-01 2.75999963e-01 -1.94710612e-01 4.35025930e-01
-6.48078382e-01 9.88992751e-01 5.15855598e+00 9.47560012e-01
-9.76299226e-01 1.86747923e-01 7.47158289e-01 1.15710244e-01
-5.21838844e-01 -9.85377058e-02 -8.31152797e-01 4.88424152e-01
4.11635101e-01 1.35653585e-01 6.55471087e-01 7.05864906e-01
9.50301066e-02 4.70648855e-01 -7.81324744e-01 1.07166314e+00
6.95218816e-02 -1.38947666e+00 5.14169276e-01 1.51766330e-01
5.64466476e-01 -4.19486731e-01 1.45116642e-01 7.21103549e-02
-4.35848832e-02 -1.24135244e+00 6.27558053e-01 9.05910075e-01
1.16705370e+00 -9.35781419e-01 6.30110800e-01 7.06689581e-02
-1.64465845e+00 -7.94221014e-02 -6.21232092e-01 3.04686397e-01
-4.11884710e-02 4.62489843e-01 1.69434175e-01 7.25684345e-01
1.05175424e+00 9.11580026e-01 -4.80496496e-01 4.41741288e-01
-1.21955842e-01 1.52997956e-01 -1.18134663e-01 6.61646605e-01
-2.85465959e-02 -2.13684216e-01 1.45216033e-01 7.79368579e-01
3.18599015e-01 6.62185013e-01 -4.84142788e-02 1.19670260e+00
-3.68820131e-01 -4.84573189e-03 -2.95144141e-01 2.41806790e-01
5.55156469e-01 1.57059431e+00 -3.17450315e-01 -4.88719791e-02
-5.59033632e-01 7.74648368e-01 6.24640107e-01 3.62091720e-01
-6.47373736e-01 -7.28076175e-02 9.11825359e-01 3.34834218e-01
4.61716652e-01 2.23158095e-02 -8.17404166e-02 -1.10804784e+00
3.34583431e-01 -6.69107318e-01 1.29994228e-01 -8.44423175e-01
-1.19301343e+00 9.87835705e-01 -3.03794414e-01 -8.93698037e-01
3.00602645e-01 -3.00228924e-01 -6.41581357e-01 1.10388553e+00
-2.02158380e+00 -1.54951799e+00 -7.30422199e-01 1.12284994e+00
3.76435101e-01 4.54657199e-03 5.59924364e-01 5.21552682e-01
-8.55370283e-01 7.73922205e-01 -4.71044332e-01 1.91148683e-01
4.40915525e-01 -3.88428748e-01 5.46821319e-02 8.30478966e-01
-2.43229747e-01 9.14607584e-01 1.20084226e-01 -6.46964073e-01
-1.34098864e+00 -1.35206366e+00 6.26652837e-01 -1.07435295e-02
2.13422403e-01 -2.12987691e-01 -1.31810200e+00 6.31716907e-01
-2.04548523e-01 6.40912473e-01 3.58020365e-01 -1.06230937e-01
-8.74138594e-01 -5.41323364e-01 -1.40819621e+00 3.80057573e-01
1.32605445e+00 -6.39742792e-01 -3.25789392e-01 -1.39322832e-01
8.53696167e-01 -2.45716736e-01 -9.48204637e-01 6.93081379e-01
6.72559619e-01 -1.18615949e+00 1.03745484e+00 -5.62688112e-01
6.55277550e-01 -4.79284942e-01 -3.38442177e-01 -8.45310926e-01
-7.44009852e-01 -3.80158037e-01 -9.29288417e-02 1.71119499e+00
-1.60365894e-01 -6.18916571e-01 5.55904210e-01 5.48837483e-01
-2.70077199e-01 -8.83188307e-01 -1.04488945e+00 -3.06068867e-01
-1.57497033e-01 1.44682869e-01 1.18369746e+00 8.91037881e-01
-5.82237720e-01 8.70257914e-02 -4.44810748e-01 3.59403133e-01
7.34038353e-01 1.48984015e-01 2.57932574e-01 -1.23559666e+00
1.58795446e-01 -4.41625118e-01 -2.42670476e-01 -9.51453090e-01
4.70865339e-01 -6.41767681e-01 -1.65917248e-01 -1.34539640e+00
4.49819475e-01 -3.96619141e-01 -4.03995544e-01 6.67907238e-01
-1.73908263e-01 3.78903866e-01 2.40242220e-02 2.66504973e-01
-4.66886759e-01 9.86117005e-01 1.69605780e+00 2.56345928e-01
1.04966797e-01 -4.09233898e-01 -1.03178382e+00 7.67779529e-01
4.98580366e-01 -1.13190129e-01 -2.69448489e-01 -6.82864070e-01
-1.33632824e-01 1.77876070e-01 6.29360139e-01 -6.38045371e-01
4.01380867e-01 -1.80665284e-01 8.71688068e-01 -3.19817573e-01
4.12788898e-01 -8.37315381e-01 2.06306711e-01 7.32612312e-02
-1.70415699e-01 -7.49999583e-02 1.05674207e-01 5.92610836e-01
-4.00413364e-01 4.55995023e-01 1.17555702e+00 -2.18714744e-01
-5.10065854e-01 1.04469705e+00 3.22653890e-01 -3.77666324e-01
5.90552926e-01 -9.73960310e-02 -3.15906614e-01 -8.85084122e-02
-5.69625080e-01 5.00887707e-02 4.01140124e-01 5.32268763e-01
1.10417700e+00 -1.62500072e+00 -8.11566949e-01 6.93211913e-01
-8.56018290e-02 3.20679486e-01 9.32944655e-01 5.95274210e-01
-2.19744742e-01 2.08002523e-01 -6.51176631e-01 -3.10521841e-01
-1.10274446e+00 6.24063015e-01 5.32492042e-01 -4.48589958e-02
-8.57311964e-01 9.10859466e-01 7.65513957e-01 -2.35552102e-01
1.11681543e-01 5.62768355e-02 -4.66813117e-01 -2.04892933e-01
1.01439369e+00 1.27462253e-01 -7.72798583e-02 -1.05386686e+00
-4.09338295e-01 1.04604220e+00 -2.37353981e-01 3.04910630e-01
1.54375660e+00 -4.40210819e-01 -7.16729760e-01 -2.19909698e-01
1.10766494e+00 -8.30261037e-03 -1.47578597e+00 -5.81154644e-01
-5.24388373e-01 -7.58259177e-01 3.45721513e-01 -4.77252126e-01
-1.81183636e+00 9.41473842e-01 4.81652409e-01 -4.36832666e-01
1.65395653e+00 6.92283735e-02 8.30709696e-01 -2.15382934e-01
4.59676504e-01 -6.02814496e-01 8.87492374e-02 1.17094122e-01
1.06551123e+00 -9.78443801e-01 -1.02072567e-01 -6.78384542e-01
-3.32781762e-01 1.17631185e+00 9.68390584e-01 -2.76853174e-01
9.11694765e-01 8.75753015e-02 -3.77288222e-01 -1.90644950e-01
-7.19158351e-01 -7.81929120e-02 3.99518639e-01 6.04804218e-01
2.89086044e-01 -1.90444753e-01 4.33450975e-02 1.06853139e+00
8.18206817e-02 1.01752751e-01 5.15652113e-02 4.77762699e-01
-3.83304298e-01 -7.73652792e-01 -2.22747415e-01 4.41054292e-02
-4.49293405e-01 -1.67222738e-01 -7.66360983e-02 5.37949443e-01
4.69693363e-01 7.59122908e-01 1.55322313e-01 -4.68819946e-01
3.73412758e-01 -3.27909172e-01 6.12896442e-01 -4.70670044e-01
-2.23923936e-01 2.34931305e-01 -3.07679802e-01 -1.02535343e+00
-5.27940333e-01 -3.54468167e-01 -1.09380150e+00 -5.80129027e-01
-1.63953364e-01 -1.42781943e-01 3.54264021e-01 9.01706100e-01
5.99135160e-01 6.32473409e-01 8.35581958e-01 -8.71269226e-01
-2.00639561e-01 -9.14626420e-01 -7.40808547e-01 2.72878967e-02
5.33978879e-01 -8.65068138e-01 -1.06040955e-01 7.34289959e-02] | [12.87299919128418, 0.014749184250831604] |
f648a6a2-f83d-43fa-9e3d-f28b6bdd40f0 | parameter-free-dynamic-graph-embedding-for | 2210.08189 | null | https://arxiv.org/abs/2210.08189v2 | https://arxiv.org/pdf/2210.08189v2.pdf | Parameter-free Dynamic Graph Embedding for Link Prediction | Dynamic interaction graphs have been widely adopted to model the evolution of user-item interactions over time. There are two crucial factors when modelling user preferences for link prediction in dynamic interaction graphs: 1) collaborative relationship among users and 2) user personalized interaction patterns. Existing methods often implicitly consider these two factors together, which may lead to noisy user modelling when the two factors diverge. In addition, they usually require time-consuming parameter learning with back-propagation, which is prohibitive for real-time user preference modelling. To this end, this paper proposes FreeGEM, a parameter-free dynamic graph embedding method for link prediction. Firstly, to take advantage of the collaborative relationships, we propose an incremental graph embedding engine to obtain user/item embeddings, which is an Online-Monitor-Offline architecture consisting of an Online module to approximately embed users/items over time, a Monitor module to estimate the approximation error in real time and an Offline module to calibrate the user/item embeddings when the online approximation errors exceed a threshold. Meanwhile, we integrate attribute information into the model, which enables FreeGEM to better model users belonging to some under represented groups. Secondly, we design a personalized dynamic interaction pattern modeller, which combines dynamic time decay with attention mechanism to model user short-term interests. Experimental results on two link prediction tasks show that FreeGEM can outperform the state-of-the-art methods in accuracy while achieving over 36X improvement in efficiency. All code and datasets can be found in https://github.com/FudanCISL/FreeGEM. | ['Ning Gu', 'Peng Zhang', 'Tun Lu', 'Hansu Gu', 'Dongsheng Li', 'Jiahao Liu'] | 2022-10-15 | null | null | null | null | ['dynamic-graph-embedding'] | ['graphs'] | [-5.07388532e-01 -1.25990167e-01 -3.93193036e-01 -1.97806418e-01
-5.83235435e-02 -3.73331875e-01 1.70091376e-01 2.97874302e-01
-1.39342353e-01 2.78761655e-01 2.30515152e-01 -5.12036324e-01
-3.59964848e-01 -9.30213153e-01 -3.82513434e-01 -4.14941818e-01
-5.10208428e-01 6.02064550e-01 2.39342868e-01 -3.05157602e-01
-3.37083608e-01 -2.22801752e-02 -1.07491684e+00 5.88724390e-02
1.06600392e+00 1.01906526e+00 -3.44321467e-02 6.93498909e-01
-2.58698940e-01 4.94470626e-01 3.42429653e-02 -7.89836049e-01
1.63942844e-01 -2.17720002e-01 -5.19433260e-01 -1.67403147e-01
-2.00011939e-01 -3.50513458e-01 -8.32876503e-01 8.83741498e-01
7.05937147e-01 4.45222884e-01 4.07210976e-01 -1.63623619e+00
-9.95217204e-01 1.03035569e+00 -4.55529958e-01 2.43014261e-01
6.11745000e-01 6.60163462e-02 1.32037675e+00 -7.70341814e-01
1.23933308e-01 1.38547325e+00 8.19650531e-01 2.05425560e-01
-1.26628613e+00 -6.73193157e-01 7.05621243e-01 5.87572336e-01
-1.45582938e+00 -4.12126109e-02 1.05424929e+00 -3.27759713e-01
8.30391765e-01 4.44567025e-01 9.91917372e-01 1.00314093e+00
5.10375015e-02 9.29714262e-01 2.73129046e-01 -1.97888399e-03
-1.62046969e-01 3.88238221e-01 3.66146505e-01 5.53943813e-01
8.21959749e-02 -7.84064457e-02 -1.81667238e-01 -4.86165345e-01
5.80210388e-01 4.21342641e-01 -4.60091770e-01 -3.98684382e-01
-7.52267241e-01 7.45557427e-01 8.46536636e-01 1.36106297e-01
-2.80099362e-01 3.18792500e-02 5.12601376e-01 6.60564125e-01
6.11868799e-01 -1.22548984e-02 -7.15747416e-01 -5.83255589e-02
-2.24281147e-01 -1.96984429e-02 9.07135606e-01 1.06575084e+00
6.42600775e-01 -3.09694529e-01 -1.12205148e-01 1.06666040e+00
7.19070554e-01 -9.94438212e-03 6.31569982e-01 -1.48555502e-01
5.20405233e-01 9.33584392e-01 2.02624165e-02 -1.58713448e+00
-3.62320244e-01 -6.84720993e-01 -7.85869896e-01 -6.25755310e-01
1.21892546e-03 -3.29554193e-02 -4.02385503e-01 1.59828842e+00
5.73761106e-01 5.81064582e-01 -3.64776403e-01 8.59212101e-01
8.01347673e-01 8.39173913e-01 1.95935369e-01 -4.13392067e-01
1.12161326e+00 -1.26835036e+00 -7.28497446e-01 8.85897782e-03
9.06021714e-01 -5.09543180e-01 1.32627141e+00 5.49504608e-02
-7.22607136e-01 -5.72559237e-01 -8.95281017e-01 2.38683224e-02
-5.00012457e-01 -1.06461123e-01 8.32764924e-01 4.81830776e-01
-8.07324231e-01 8.68241191e-01 -7.30478585e-01 -2.19036788e-01
2.43470997e-01 6.34716392e-01 -2.57330686e-02 5.23018949e-02
-1.76623130e+00 4.52872038e-01 5.05832136e-01 3.37230265e-01
-9.94971246e-02 -9.02575850e-01 -7.00319707e-01 2.98110187e-01
3.63432407e-01 -7.59731650e-01 9.88083243e-01 -7.94117212e-01
-1.44742060e+00 2.69801050e-01 1.15952030e-01 -1.82190374e-01
7.15151608e-01 -1.18873008e-01 -1.02770221e+00 -4.70866650e-01
-4.90399659e-01 -1.57350421e-01 5.50171196e-01 -9.57012177e-01
-4.60478425e-01 -2.56016195e-01 2.75974363e-01 4.01312947e-01
-9.44621861e-01 -3.79447371e-01 -1.07483852e+00 -5.91520607e-01
-2.18928382e-01 -9.88735735e-01 -9.64391306e-02 -1.62978277e-01
-2.40843445e-01 -5.48837543e-01 6.83812976e-01 -7.37698674e-01
2.09104395e+00 -2.05224371e+00 2.56579727e-01 4.28424835e-01
3.35432202e-01 3.21708947e-01 -1.97725162e-01 6.58756375e-01
-1.63532481e-01 1.22540735e-01 3.23357612e-01 -4.06246096e-01
2.04715014e-01 9.00572911e-02 -5.71169630e-02 2.18496442e-01
-5.00696838e-01 1.25384951e+00 -1.17223692e+00 -3.68646085e-01
1.86257616e-01 6.77032709e-01 -5.96203387e-01 3.02603692e-01
-1.03622124e-01 1.38901934e-01 -5.60884118e-01 2.71259069e-01
6.64030671e-01 -5.18162608e-01 5.37281394e-01 -5.38227618e-01
4.17852938e-01 3.19028497e-01 -1.27965868e+00 1.36739433e+00
-6.03393435e-01 1.97059289e-01 -1.62356362e-01 -6.84185743e-01
6.12886310e-01 2.67337918e-01 5.65354466e-01 -3.23696882e-01
3.87059152e-01 3.92502211e-02 6.02394715e-02 -5.09696722e-01
2.27722362e-01 4.95920718e-01 9.09926966e-02 3.39372218e-01
-1.04244918e-01 6.37504637e-01 3.95256020e-02 4.19548631e-01
9.38095331e-01 -1.35865867e-01 1.22428626e-01 4.98290025e-02
6.42870665e-01 -7.05197811e-01 4.83278632e-01 1.78639427e-01
-1.51321709e-01 1.03175357e-01 4.58029866e-01 -4.66694832e-01
-5.09207666e-01 -8.15364420e-01 1.85458183e-01 1.33375120e+00
3.52884382e-01 -8.02061081e-01 -2.33903825e-01 -8.26668620e-01
4.35120553e-01 5.50746739e-01 -7.30770111e-01 -6.21005595e-01
-2.51205385e-01 -9.21202064e-01 -8.21135491e-02 5.39428234e-01
1.30085155e-01 -7.60779321e-01 4.60290372e-01 5.22316217e-01
-6.46073222e-02 -6.89077497e-01 -1.19930768e+00 -3.09138298e-01
-7.57452369e-01 -1.05874336e+00 -3.81950349e-01 -7.18653262e-01
5.98025799e-01 3.69904876e-01 9.17826653e-01 4.81958807e-01
6.63552210e-02 5.00080407e-01 -4.76038218e-01 -1.62317175e-02
9.61510912e-02 2.89685756e-01 3.31650138e-01 4.13618594e-01
5.21939039e-01 -9.28853393e-01 -9.16226625e-01 5.39512753e-01
-6.56578660e-01 -1.99767966e-02 3.81473720e-01 8.57047498e-01
3.41289192e-01 1.65933535e-01 5.42209208e-01 -1.15121603e+00
1.04138112e+00 -8.99007499e-01 -3.35091949e-01 4.07606065e-01
-1.19581008e+00 -2.60281682e-01 7.99859345e-01 -9.26152110e-01
-6.38100922e-01 -2.80841291e-01 -1.21594653e-01 -5.97163320e-01
5.20702660e-01 9.97054517e-01 -4.73503798e-01 -2.58089788e-02
3.12751442e-01 1.23249836e-01 -1.35613069e-01 -5.22609472e-01
5.30592084e-01 7.73710549e-01 3.90977524e-02 -1.54612228e-01
8.08254898e-01 -1.82229459e-01 -5.43589413e-01 -3.85311991e-01
-5.87514400e-01 -6.54130042e-01 -4.46383119e-01 -1.92374468e-01
1.83345526e-01 -8.88769329e-01 -8.19342554e-01 3.13319951e-01
-7.30777025e-01 -4.08551276e-01 1.38745233e-01 5.19414544e-01
-9.73358899e-02 5.82553387e-01 -9.10131991e-01 -7.24585414e-01
-6.04121327e-01 -6.46734178e-01 2.87494779e-01 3.88945371e-01
-1.15340047e-01 -1.57073843e+00 3.03677027e-03 2.35133797e-01
4.54753995e-01 -7.45489523e-02 7.52476037e-01 -7.46029258e-01
-2.72319853e-01 -6.02969825e-01 -1.44587517e-01 7.01814443e-02
3.70549887e-01 -1.02685019e-02 -6.35859370e-01 -6.63868010e-01
-5.19832313e-01 1.85370982e-01 5.84534585e-01 1.46836475e-01
1.38403606e+00 -5.78651309e-01 -7.60504425e-01 7.18013644e-01
1.18221283e+00 2.32703805e-01 4.73993838e-01 -3.57911922e-02
1.12253881e+00 3.11546117e-01 5.67200601e-01 4.70506728e-01
9.43590999e-01 8.27760637e-01 3.35745513e-01 7.61162937e-02
1.51957437e-01 -7.16049075e-01 2.50677317e-01 1.47461128e+00
-4.03635055e-02 -6.62670732e-01 -5.36440492e-01 5.23345053e-01
-2.28818846e+00 -6.92023158e-01 -2.79095858e-01 2.22304702e+00
7.36399114e-01 2.74064183e-01 5.04182756e-01 5.42341061e-02
6.53474927e-01 6.75403178e-02 -7.51628339e-01 -1.84583917e-01
4.68447506e-01 -5.12855232e-01 3.88134867e-01 6.54221833e-01
-9.50436175e-01 8.52635741e-01 4.51086473e+00 8.65563810e-01
-9.66373265e-01 2.47367457e-01 4.32691902e-01 -7.97756463e-02
-5.57987034e-01 -1.86572269e-01 -5.47028601e-01 9.65235174e-01
9.58076358e-01 -5.59552908e-01 7.23240316e-01 8.10540497e-01
2.10642934e-01 7.51993895e-01 -9.93217230e-01 1.20689595e+00
-1.19950086e-01 -8.22365642e-01 -3.78846365e-04 1.37387589e-01
3.91611665e-01 -9.07834619e-02 -3.10835261e-02 7.15647936e-01
2.75526762e-01 -6.62631035e-01 1.90189425e-02 5.54723501e-01
4.63362724e-01 -8.59861135e-01 7.65829206e-01 2.75593460e-01
-1.76455522e+00 -2.95817375e-01 -5.51252007e-01 3.02442200e-02
2.80302644e-01 7.76325405e-01 -6.90035582e-01 8.86435330e-01
5.70747077e-01 1.03877497e+00 -5.27958572e-01 1.28942704e+00
-1.06952339e-01 6.91694558e-01 -4.07175779e-01 -2.24346697e-01
-1.26304582e-01 -4.39219713e-01 3.90394837e-01 1.01326084e+00
3.78143489e-01 2.04126641e-01 4.30938631e-01 3.16277593e-01
-2.33033434e-01 5.32565355e-01 -3.82440627e-01 -2.24675804e-01
7.15853453e-01 1.48358583e+00 -2.97059655e-01 -8.99806619e-02
-5.15592635e-01 1.32065022e+00 4.49148953e-01 4.74875718e-01
-1.14880836e+00 -5.21490335e-01 7.20336676e-01 2.62219489e-01
1.95293605e-01 -7.92037770e-02 2.15854466e-01 -1.33845878e+00
8.84567797e-02 -4.36899036e-01 7.81585336e-01 -4.08859700e-01
-1.54594469e+00 6.32369459e-01 -2.65712619e-01 -1.33505261e+00
7.36523867e-02 -5.25578916e-01 -7.69173265e-01 7.88387179e-01
-1.16225576e+00 -1.32140768e+00 -4.31670576e-01 5.17786741e-01
2.61444718e-01 -1.09136896e-02 7.30851710e-01 8.03643048e-01
-8.70035350e-01 1.20478177e+00 6.65976107e-02 7.42845312e-02
5.68131506e-01 -1.13200736e+00 5.59373617e-01 3.01000774e-01
1.94126934e-01 7.19893515e-01 5.59824467e-01 -6.79211617e-01
-1.49350834e+00 -1.30769229e+00 9.18049395e-01 -4.79669213e-01
9.49604690e-01 -4.17953193e-01 -1.19063199e+00 8.68491948e-01
-7.94152245e-02 1.53160021e-01 1.04781377e+00 5.98648727e-01
-2.66951323e-01 -3.59764785e-01 -7.85905123e-01 7.52109706e-01
1.45281422e+00 -4.96385604e-01 -2.00566188e-01 4.72090662e-01
9.69087899e-01 -3.29449654e-01 -1.34305298e+00 9.85460654e-02
7.65590549e-01 -4.83110875e-01 1.09630060e+00 -6.36287749e-01
-1.78320929e-01 -2.77606219e-01 2.04809964e-01 -1.73310101e+00
-8.50599229e-01 -9.42709863e-01 -9.54704583e-01 1.42705750e+00
5.67044258e-01 -9.65573788e-01 6.73575342e-01 8.51806045e-01
9.81116444e-02 -1.21050501e+00 -5.62780023e-01 -6.59593940e-01
-3.78430396e-01 -3.07821363e-01 9.77602601e-01 1.17271113e+00
3.50326419e-01 6.00867629e-01 -7.13314056e-01 1.19523853e-01
3.54708254e-01 5.99782579e-02 6.51280403e-01 -1.51871181e+00
-5.67573071e-01 -4.23076928e-01 -4.32901561e-01 -1.27660072e+00
4.47336435e-02 -1.13673949e+00 -4.80373144e-01 -1.63050282e+00
-1.40114361e-02 -6.25437140e-01 -8.07040095e-01 3.45444322e-01
-5.29489100e-01 -1.35222360e-01 -1.04672529e-01 2.78244823e-01
-7.03529894e-01 9.07756984e-01 1.09739578e+00 -1.51086733e-01
-7.07233787e-01 2.66417146e-01 -8.86330187e-01 4.35127199e-01
7.83618271e-01 -1.48713127e-01 -9.79897559e-01 -3.34082872e-01
5.57976305e-01 -1.96728036e-01 -7.92530924e-02 -6.72183812e-01
3.18106562e-01 -1.06825365e-03 2.63336986e-01 -2.89500386e-01
2.69087672e-01 -1.19018710e+00 4.61167008e-01 2.81664312e-01
-1.61364630e-01 1.45795316e-01 1.50017753e-01 1.12807357e+00
1.46882012e-01 2.58787006e-01 2.38561273e-01 3.06237459e-01
-6.28614485e-01 1.08296084e+00 1.38430595e-01 -3.11919391e-01
9.94722962e-01 7.13067427e-02 -2.02658623e-01 -6.19222820e-01
-9.11157668e-01 9.81035173e-01 2.97517806e-01 8.26152027e-01
5.41964889e-01 -1.89740014e+00 -3.89072716e-01 5.73810600e-02
2.76909024e-01 -4.18460697e-01 5.00493646e-01 9.80853736e-01
-1.21208459e-01 3.34641784e-02 3.40875864e-01 -1.79780692e-01
-1.25268853e+00 8.92644942e-01 2.55043685e-01 -4.57194418e-01
-4.17230636e-01 1.04127169e+00 -2.19877988e-01 -7.84742057e-01
2.98699468e-01 -9.87549201e-02 -4.39727008e-01 3.45109373e-01
4.61952001e-01 4.52113986e-01 -1.51832998e-01 -4.46524739e-01
-3.20919931e-01 4.42549229e-01 -2.83392072e-01 5.63012421e-01
1.14851284e+00 -5.56902468e-01 -2.86714751e-02 6.29111409e-01
1.49540734e+00 -1.64034695e-03 -1.01111686e+00 -4.49217975e-01
-2.67589301e-01 -6.60081327e-01 1.71361610e-01 -6.10758901e-01
-1.24253738e+00 5.95157743e-01 6.28482759e-01 6.84543431e-01
1.06004703e+00 -2.72633582e-01 1.24059844e+00 4.09616008e-02
3.65919352e-01 -9.98384118e-01 -2.13394508e-01 2.71691144e-01
8.89189422e-01 -1.21887338e+00 3.71494368e-02 -6.48718476e-01
-5.99391520e-01 8.56118023e-01 7.89494097e-01 6.33098260e-02
1.29054177e+00 -3.71928662e-01 -8.79885852e-02 6.65735155e-02
-9.79401231e-01 5.65131046e-02 7.76140273e-01 2.88105965e-01
4.96180356e-01 2.94385791e-01 -4.13526803e-01 1.03004456e+00
-3.24509777e-02 -2.17613783e-02 1.57361533e-02 3.29052120e-01
-2.53393371e-02 -1.36097229e+00 3.12253296e-01 8.14565182e-01
-9.40013900e-02 -1.24655753e-01 -2.61737645e-01 4.88034755e-01
-1.16068169e-01 8.98659885e-01 -1.24235980e-01 -1.18175447e+00
5.14682472e-01 5.77181987e-02 1.15222655e-01 -4.04290229e-01
-5.07888973e-01 -3.86392847e-02 6.93022385e-02 -5.17048955e-01
2.03676909e-01 -4.36365366e-01 -1.00862527e+00 -7.00683773e-01
-6.65518701e-01 4.01485205e-01 1.24512896e-01 4.21215445e-01
7.28811920e-01 6.36643171e-01 8.80566716e-01 -6.96901679e-01
-4.41867769e-01 -1.07817125e+00 -6.62665725e-01 5.64813972e-01
2.27848310e-02 -7.38068163e-01 -5.02562106e-01 -4.29095924e-01] | [10.184409141540527, 5.620010852813721] |
ce88de6b-9ba5-4d8d-85f4-2f8f2ea8d536 | argument-component-classification-for-1 | 1909.03022 | null | https://arxiv.org/abs/1909.03022v1 | https://arxiv.org/pdf/1909.03022v1.pdf | Argument Component Classification for Classroom Discussions | This paper focuses on argument component classification for transcribed spoken classroom discussions, with the goal of automatically classifying student utterances into claims, evidence, and warrants. We show that an existing method for argument component classification developed for another educationally-oriented domain performs poorly on our dataset. We then show that feature sets from prior work on argument mining for student essays and online dialogues can be used to improve performance considerably. We also provide a comparison between convolutional neural networks and recurrent neural networks when trained under different conditions to classify argument components in classroom discussions. While neural network models are not always able to outperform a logistic regression model, we were able to gain some useful insights: convolutional networks are more robust than recurrent networks both at the character and at the word level, and specificity information can help boost performance in multi-task training. | ['Luca Lugini', 'Diane Litman'] | 2019-09-06 | argument-component-classification-for | https://aclanthology.org/W18-5208 | https://aclanthology.org/W18-5208.pdf | ws-2018-11 | ['component-classification'] | ['natural-language-processing'] | [ 3.17192435e-01 5.10830939e-01 -3.32985878e-01 -3.76575738e-01
-1.10245144e+00 -7.09191263e-01 7.56958902e-01 7.47811675e-01
-2.43583441e-01 7.87390113e-01 7.83233821e-01 -1.13726771e+00
-3.80256563e-01 -5.70373416e-01 -5.10349095e-01 -2.46730849e-01
5.50851703e-01 2.75979370e-01 1.56483412e-01 -3.92663866e-01
5.56071341e-01 4.88479203e-03 -1.63024533e+00 6.30704284e-01
9.25869644e-01 4.16669548e-01 -4.47009325e-01 1.03763247e+00
-5.74084222e-01 1.67476547e+00 -1.07860017e+00 -4.94919181e-01
-5.31887054e-01 -6.24898970e-01 -1.54424798e+00 -2.14076146e-01
7.17299461e-01 -2.25246966e-01 5.09533845e-02 5.71433067e-01
5.41630149e-01 6.94322661e-02 6.78285718e-01 -8.34054112e-01
-6.38163030e-01 1.44029856e+00 -2.21787374e-02 3.29455882e-01
5.00952363e-01 -4.45780665e-01 1.22581875e+00 -4.68145102e-01
5.29443979e-01 1.29330993e+00 8.01172733e-01 4.01282609e-01
-1.08384562e+00 -6.40934408e-01 2.57299483e-01 2.96936780e-01
-2.09071655e-02 -4.37067717e-01 8.31444919e-01 -4.74965036e-01
9.87043560e-01 3.06903064e-01 8.32380474e-01 1.40044439e+00
-2.86062896e-01 1.35570157e+00 1.34809673e+00 -6.53685510e-01
-1.65515304e-01 3.57856303e-02 9.42367435e-01 6.54036760e-01
-1.60339952e-01 -4.33061093e-01 -5.55331409e-01 -3.46236080e-01
1.17078535e-01 -4.42443490e-01 -1.07085727e-01 5.70386695e-03
-9.59878922e-01 1.40304911e+00 -1.21769771e-01 4.26579595e-01
7.41865709e-02 4.62673865e-02 7.86847293e-01 8.57451081e-01
6.83366537e-01 5.85544229e-01 -6.40001655e-01 -8.39240611e-01
-8.18576336e-01 4.09636885e-01 1.36354947e+00 4.05470312e-01
-1.04866996e-01 9.05311331e-02 -2.08681628e-01 1.22757387e+00
5.70429005e-02 -4.61735800e-02 7.00637162e-01 -9.72864628e-01
6.49210036e-01 6.68928087e-01 -4.08922970e-01 -4.48885649e-01
-4.18569982e-01 -3.18265498e-01 4.17148769e-02 5.75520331e-03
8.20455313e-01 -5.62936127e-01 -4.12387997e-01 1.50276268e+00
2.11422086e-01 -2.07561523e-01 5.33273935e-01 2.71459341e-01
1.58072102e+00 4.01453018e-01 4.61535081e-02 -7.43456855e-02
1.45044339e+00 -1.17818987e+00 -8.63159895e-01 -5.06415293e-02
1.27348483e+00 -1.10942972e+00 9.45011020e-01 3.56574863e-01
-1.25711298e+00 -7.46924654e-02 -9.75822747e-01 -2.31582850e-01
-3.70711833e-01 5.45223169e-02 6.71292484e-01 9.22195911e-01
-5.99529624e-01 4.06208009e-01 -2.77016759e-01 -2.67253984e-02
3.92708749e-01 -9.48965847e-02 9.15657058e-02 2.55256146e-01
-9.96297300e-01 8.77774119e-01 -1.25135556e-01 -3.95473599e-01
-4.35205728e-01 -6.03587747e-01 -8.93531680e-01 1.66666105e-01
3.62319618e-01 -1.48880035e-02 1.79999614e+00 -7.48859048e-01
-1.75543714e+00 8.03388119e-01 2.07957432e-01 -5.65593243e-01
4.18836921e-01 -2.79482037e-01 1.48466006e-01 -3.51657420e-02
-1.06225414e-02 2.16853634e-01 3.81407529e-01 -5.36629200e-01
-3.55275482e-01 5.91654219e-02 2.03924745e-01 2.41568342e-01
-5.29081762e-01 4.45327431e-01 6.00886047e-01 -4.84102100e-01
1.52887613e-01 -7.66248763e-01 1.80100575e-01 -5.75900137e-01
-4.57937539e-01 -1.21419179e+00 1.02825856e+00 -5.22273123e-01
8.75161767e-01 -1.72437036e+00 -1.94772393e-01 -2.12417975e-01
1.25730321e-01 4.87088561e-01 -1.40049979e-01 5.19283354e-01
-3.94685000e-01 2.95655340e-01 2.99305558e-01 -3.11701233e-03
-4.20592204e-02 -1.38660343e-02 -3.88903767e-01 2.28706494e-01
1.56196296e-01 8.85058522e-01 -8.05059433e-01 -5.14834881e-01
-3.24682705e-02 2.68192977e-01 -3.17440301e-01 1.92816526e-01
-2.70566344e-01 3.31342146e-02 -5.86003363e-01 2.61685789e-01
-5.30456156e-02 -2.83753783e-01 3.03266466e-01 5.25172949e-01
-2.19108298e-01 1.51604295e+00 -7.50861466e-01 1.20460904e+00
-7.22450733e-01 1.31684399e+00 1.32743046e-01 -1.51132178e+00
8.35601091e-01 7.74600267e-01 2.08427295e-01 -6.40557587e-01
3.55395496e-01 2.25128770e-01 4.88259614e-01 -5.63062370e-01
4.29972798e-01 2.14031979e-01 -3.46946009e-02 1.20826614e+00
-1.19485438e-01 -3.81333381e-01 2.28258565e-01 2.24817038e-01
1.11947680e+00 -7.78270736e-02 -1.54720303e-02 -3.39251906e-01
4.32496190e-01 7.61705786e-02 -3.19298804e-02 7.66759515e-01
6.98024184e-02 3.11874241e-01 9.96101975e-01 -4.82610494e-01
-7.10793853e-01 -4.40748304e-01 -3.72394830e-01 1.65181327e+00
-6.72377348e-01 -3.16443741e-01 -7.50011444e-01 -1.07643557e+00
-1.27463028e-01 6.97469234e-01 -5.17370880e-01 3.97203147e-01
-6.77787006e-01 -5.12751698e-01 7.82590389e-01 5.40423393e-01
3.41229379e-01 -1.10198045e+00 -9.28848445e-01 3.63746375e-01
-1.42920718e-01 -1.00611436e+00 2.50818022e-02 8.38319063e-01
-6.98370218e-01 -1.45184553e+00 -3.70584577e-01 -1.01536417e+00
2.15677798e-01 1.90579832e-01 1.17770863e+00 6.96739316e-01
1.82882808e-02 6.62401199e-01 -4.25251722e-01 -9.77670908e-01
-8.68382215e-01 3.39555413e-01 -4.77433026e-01 -9.08832610e-01
2.51394421e-01 -3.71622652e-01 1.09761149e-01 -1.66850626e-01
-5.46250105e-01 6.28985316e-02 -4.97802102e-04 1.26063216e+00
-3.65234017e-01 -3.29429328e-01 8.93660843e-01 -1.33174133e+00
1.46443439e+00 -3.21353346e-01 -2.15484127e-01 2.25675359e-01
-7.13467300e-01 2.81850040e-01 3.75536799e-01 -6.69243395e-01
-8.59566033e-01 -5.93823433e-01 -5.87036073e-01 2.66727924e-01
-1.55965388e-01 6.41695678e-01 5.54204643e-01 2.07738560e-02
6.56714797e-01 -1.76280379e-01 2.74656057e-01 -3.97936732e-01
3.62421513e-01 6.93921983e-01 3.10679283e-02 -9.61147249e-01
2.43114352e-01 -3.67414266e-01 -2.50957161e-01 -9.00411069e-01
-1.36266327e+00 -1.59022406e-01 -3.16565901e-01 -8.95738974e-02
7.20196366e-01 -9.31041539e-01 -1.07225490e+00 3.40837874e-02
-1.13814163e+00 -7.25523412e-01 -3.88706148e-01 4.70494390e-01
-3.80354166e-01 2.14098662e-01 -1.19736278e+00 -9.39433873e-01
-2.38225162e-01 -1.21993995e+00 4.83337134e-01 3.55684817e-01
-5.31561673e-01 -1.34617841e+00 9.22621042e-02 1.06937909e+00
3.39994431e-01 1.89203932e-03 1.30501342e+00 -1.66292036e+00
1.57210335e-01 -2.50029601e-02 -1.85465720e-02 3.25178713e-01
-1.50121137e-01 3.19974661e-01 -1.00923800e+00 2.47707993e-01
1.14507690e-01 -1.13081110e+00 1.07590485e+00 2.22350582e-01
1.07219565e+00 -6.07384324e-01 1.16458341e-01 -1.50585756e-01
4.64732140e-01 -6.51000626e-03 1.57122642e-01 7.15076387e-01
4.55170155e-01 9.02454972e-01 2.08478197e-01 -1.60558984e-01
6.46684945e-01 2.09039405e-01 1.47674218e-01 -3.64197828e-02
-2.33337149e-01 -1.19470090e-01 3.02806675e-01 9.12471890e-01
2.35930219e-01 -3.57332885e-01 -1.08753431e+00 7.50470281e-01
-1.94809330e+00 -1.16148603e+00 -6.24573469e-01 1.37724984e+00
1.08574057e+00 3.00614834e-01 4.47656423e-01 5.67187488e-01
3.80460113e-01 3.10244471e-01 -3.92818116e-02 -1.30443072e+00
-1.00370049e-01 6.18071496e-01 7.09326789e-02 4.77854818e-01
-1.06101084e+00 7.06800461e-01 7.17353344e+00 6.33595288e-01
-8.83320749e-01 9.09716785e-02 7.29405761e-01 8.29469487e-02
-6.36347115e-01 -1.58700839e-01 -6.64319813e-01 -1.54388873e-02
1.29415357e+00 9.60747376e-02 -1.18261375e-01 7.42287338e-01
-2.42475986e-01 1.96327548e-02 -9.84775126e-01 2.67607719e-01
7.71118253e-02 -1.71913099e+00 -4.13334519e-01 4.97449525e-02
7.36881733e-01 3.55442241e-02 1.97281390e-01 5.78492522e-01
1.05612910e+00 -1.13171744e+00 5.24496853e-01 -2.07561061e-01
1.47233903e-02 -6.21406436e-01 7.89240062e-01 5.69951355e-01
-4.89388108e-01 -2.16630489e-01 -3.85428071e-02 -7.23189771e-01
-4.66076523e-01 2.54229039e-01 -1.21825957e+00 6.51428103e-02
4.01191860e-01 6.18409157e-01 -6.78469300e-01 5.97907305e-01
-5.93079567e-01 1.28833103e+00 -1.81513101e-01 -7.48471439e-01
6.75828218e-01 1.11618536e-02 3.85509849e-01 1.26082432e+00
-2.01820910e-01 1.42989606e-01 2.02753246e-01 3.81479204e-01
-8.37859735e-02 5.09692311e-01 -6.38437271e-01 -1.61264226e-01
2.65159369e-01 1.38879371e+00 -7.42551684e-01 -4.15966988e-01
-4.67230499e-01 3.30905467e-02 5.52750051e-01 8.33017528e-02
-5.02740264e-01 -6.92860484e-02 5.71610093e-01 -2.30826691e-01
2.75368661e-01 -1.31775901e-01 -3.13499093e-01 -1.03453124e+00
-4.41270024e-01 -1.19448638e+00 6.44596100e-01 -5.39617598e-01
-1.21401489e+00 2.86375493e-01 -5.06631844e-02 -5.38882136e-01
-8.33865821e-01 -7.01897085e-01 -1.08973682e+00 5.24652302e-01
-1.43204379e+00 -9.59299564e-01 1.97262913e-01 2.49498755e-01
9.83436465e-01 -2.37335995e-01 9.66232479e-01 -2.54138321e-01
-6.99964762e-01 5.94132304e-01 9.86486822e-02 5.20275295e-01
6.91769660e-01 -1.37318683e+00 1.40927643e-01 5.14694452e-01
5.69089532e-01 4.62165147e-01 6.97615087e-01 -3.31510961e-01
-1.07598817e+00 -4.59099799e-01 9.47420776e-01 -5.93157291e-01
1.05357099e+00 -1.31820396e-01 -9.62557316e-01 5.85970640e-01
7.63001978e-01 -3.78675848e-01 1.06349289e+00 8.81553352e-01
-6.86934948e-01 8.30135107e-01 -7.55629659e-01 5.24788499e-01
6.83859766e-01 -9.04475093e-01 -1.15385628e+00 6.01570904e-01
5.92394710e-01 -7.38685608e-01 -9.24302697e-01 1.78682134e-02
6.32072151e-01 -8.18117440e-01 8.52167606e-01 -1.13562417e+00
1.07913482e+00 5.62453568e-01 2.69744694e-01 -1.11159670e+00
2.30370671e-01 -3.96873325e-01 2.95347329e-02 1.41511083e+00
7.28117585e-01 -5.54015458e-01 9.50455844e-01 5.70403516e-01
-2.29211211e-01 -1.01433814e+00 -7.82953799e-01 -3.33678901e-01
9.24530208e-01 -4.14865762e-01 3.63962054e-01 1.17026234e+00
5.42156339e-01 1.05091691e+00 2.09339023e-01 -4.97609377e-01
2.93800920e-01 5.07174194e-01 6.88709319e-01 -1.59977031e+00
2.09589958e-01 -1.00312364e+00 7.24395216e-02 -9.35266018e-01
9.61178124e-01 -9.23419535e-01 -1.03599280e-01 -1.75022471e+00
6.35871664e-02 -3.84860724e-01 1.31276220e-01 7.14322090e-01
-2.81081080e-01 -6.14748802e-03 -2.50430722e-02 -1.98060095e-01
-3.97621095e-01 2.04347849e-01 1.07154024e+00 -2.72457272e-01
-5.06202653e-02 1.43575773e-01 -7.89126873e-01 9.01091218e-01
8.03824127e-01 -4.71092999e-01 -5.03019750e-01 -3.49890709e-01
3.79422933e-01 1.79690599e-01 1.13340020e-01 -6.18469536e-01
2.24900544e-01 -6.13413826e-02 2.39287794e-01 -6.32065952e-01
6.69643879e-02 -2.10326776e-01 -7.60667145e-01 2.92240858e-01
-1.17286980e+00 -8.75033885e-02 2.64388353e-01 1.50183156e-01
-2.51100451e-01 -8.37702811e-01 3.22158039e-01 -1.33220643e-01
1.70673192e-01 -5.19435704e-01 -8.65174472e-01 3.92606944e-01
4.07721549e-01 -9.64198932e-02 -8.88787031e-01 -7.83969581e-01
-3.99077535e-01 3.63616943e-01 -6.20453879e-02 5.03252268e-01
4.45313007e-01 -8.22966993e-01 -9.85216379e-01 4.08994630e-02
-3.20489973e-01 -2.83598453e-01 -2.59901136e-01 7.49666154e-01
-1.93843290e-01 5.63183606e-01 3.01573783e-01 -6.13865316e-01
-1.89561939e+00 -1.05546370e-01 5.22874296e-02 -5.27785003e-01
-5.88278770e-01 9.71479774e-01 -4.55921382e-01 -6.89003110e-01
4.19710875e-01 -5.07744610e-01 -8.78691792e-01 6.25101805e-01
3.95039052e-01 2.07008153e-01 2.00562268e-01 -2.19640121e-01
1.22515544e-01 1.13174707e-01 -2.62011081e-01 -5.37853166e-02
1.74801338e+00 2.65706182e-01 1.89762741e-01 6.40165806e-01
1.03762341e+00 4.74781752e-01 -5.79604447e-01 -3.42205092e-02
3.37076515e-01 3.01961213e-01 4.10401486e-02 -7.21332133e-01
-5.75706780e-01 9.82180417e-01 3.93622033e-02 9.64588523e-01
3.59974921e-01 1.10343702e-01 6.06327951e-01 6.98558450e-01
-5.33653140e-01 -1.12730968e+00 3.65153998e-01 1.14717042e+00
4.33575392e-01 -1.24244046e+00 2.17304066e-01 -2.64369845e-01
-4.92366999e-01 1.67827404e+00 3.95406872e-01 8.71523917e-02
5.14829099e-01 4.41329122e-01 1.63765162e-01 -6.22368991e-01
-1.22293329e+00 -1.97387367e-01 2.00122714e-01 1.96679667e-01
1.40995026e+00 -1.16574019e-01 -4.82893705e-01 3.67309541e-01
-6.25570655e-01 -6.13445282e-01 1.02546179e+00 9.20612156e-01
-5.04063666e-01 -1.30400240e+00 -3.65577757e-01 6.06989801e-01
-9.46175694e-01 -2.43902817e-01 -9.81070399e-01 8.94171059e-01
-4.26775485e-01 1.33576453e+00 9.55670252e-02 6.04789192e-03
8.83361623e-02 7.93759823e-01 3.91679615e-01 -9.04834926e-01
-1.47546780e+00 -2.27414057e-01 1.06962490e+00 -1.61157295e-01
-7.71158159e-01 -8.18531752e-01 -1.03699481e+00 -2.48885527e-01
-6.86406970e-01 7.60238349e-01 8.78628731e-01 1.31870234e+00
-3.25028330e-01 7.31245756e-01 2.68480599e-01 -6.41898960e-02
-5.94501674e-01 -1.13200903e+00 1.97652638e-01 7.66347572e-02
5.07495105e-01 -4.12963957e-01 -5.52106917e-01 -3.41125697e-01] | [10.516480445861816, 9.42950439453125] |
76969844-9780-4aea-b068-50e7fd5793e8 | voicebox-text-guided-multilingual-universal | 2306.15687 | null | https://arxiv.org/abs/2306.15687v1 | https://arxiv.org/pdf/2306.15687v1.pdf | Voicebox: Text-Guided Multilingual Universal Speech Generation at Scale | Large-scale generative models such as GPT and DALL-E have revolutionized natural language processing and computer vision research. These models not only generate high fidelity text or image outputs, but are also generalists which can solve tasks not explicitly taught. In contrast, speech generative models are still primitive in terms of scale and task generalization. In this paper, we present Voicebox, the most versatile text-guided generative model for speech at scale. Voicebox is a non-autoregressive flow-matching model trained to infill speech, given audio context and text, trained on over 50K hours of speech that are neither filtered nor enhanced. Similar to GPT, Voicebox can perform many different tasks through in-context learning, but is more flexible as it can also condition on future context. Voicebox can be used for mono or cross-lingual zero-shot text-to-speech synthesis, noise removal, content editing, style conversion, and diverse sample generation. In particular, Voicebox outperforms the state-of-the-art zero-shot TTS model VALL-E on both intelligibility (5.9% vs 1.9% word error rates) and audio similarity (0.580 vs 0.681) while being up to 20 times faster. See voicebox.metademolab.com for a demo of the model. | ['Wei-Ning Hsu', 'Jay Mahadeokar', 'Yossi Adi', 'Vimal Manohar', 'Mary Williamson', 'Rashel Moritz', 'Leda Sari', 'Brian Karrer', 'Bowen Shi', 'Apoorv Vyas', 'Matthew Le'] | 2023-06-23 | null | null | null | null | ['text-to-speech-synthesis', 'speech-synthesis'] | ['speech', 'speech'] | [ 2.39686981e-01 1.20774530e-01 2.67005056e-01 -1.12105042e-01
-1.11189771e+00 -5.31276286e-01 8.43663156e-01 -4.55110461e-01
-1.43343538e-01 5.03750265e-01 5.96148312e-01 -5.07830322e-01
1.72816321e-01 -6.48301125e-01 -7.18847811e-01 -5.23381889e-01
4.01101440e-01 3.96664768e-01 -2.84435302e-02 -3.85706961e-01
-5.40979914e-02 1.86000541e-01 -1.67231655e+00 4.24156159e-01
8.95692050e-01 6.87825203e-01 7.27278113e-01 1.24011099e+00
-2.84587175e-01 5.08674085e-01 -8.63011181e-01 -3.76983762e-01
-1.08138025e-01 -6.66678667e-01 -4.68840450e-01 2.14288950e-01
6.61309421e-01 -2.41284117e-01 -5.67500651e-01 7.59825885e-01
1.06786931e+00 3.09301704e-01 5.43712199e-01 -1.07427323e+00
-1.08992660e+00 6.46747828e-01 -2.14965362e-02 1.56055078e-01
4.47050244e-01 4.46631879e-01 7.77689993e-01 -1.29889834e+00
5.82115412e-01 1.67812681e+00 4.62849408e-01 8.49460661e-01
-1.32753646e+00 -7.28943408e-01 -1.03539214e-01 -5.56778759e-02
-1.14851904e+00 -9.75690603e-01 5.09357333e-01 -3.76769632e-01
1.30161774e+00 4.68791217e-01 5.39226055e-01 1.56206501e+00
2.85578936e-01 7.05819547e-01 9.93647039e-01 -5.73083162e-01
1.86900288e-01 -1.40577136e-02 -5.93032360e-01 5.10449469e-01
-5.30449808e-01 4.51828271e-01 -8.98180485e-01 1.47718102e-01
7.92333066e-01 -4.88760054e-01 -2.19443068e-01 2.69924074e-01
-1.34695780e+00 9.33232725e-01 1.29047525e-03 1.22896574e-01
-1.77263439e-01 3.66515934e-01 3.67117882e-01 6.14541590e-01
6.73095882e-01 5.15170813e-01 -3.35381031e-01 -4.82037812e-01
-1.20946181e+00 3.48227441e-01 7.75809288e-01 1.24619782e+00
4.34377998e-01 9.98324513e-01 -4.63463843e-01 1.19600964e+00
1.70569628e-01 9.36366141e-01 8.06266785e-01 -1.01131320e+00
3.86881471e-01 -3.11873585e-01 -3.68539900e-01 -4.62949455e-01
4.93020006e-03 -4.12110776e-01 -8.29593122e-01 2.04152003e-01
4.59859818e-02 -4.49669331e-01 -1.30428958e+00 1.58235407e+00
-1.81400089e-03 3.25838953e-01 1.66726574e-01 5.56657791e-01
1.14391005e+00 1.19391072e+00 -2.22118944e-02 -1.63836905e-03
1.23989737e+00 -1.18167186e+00 -9.41091597e-01 -4.97045517e-01
1.77759722e-01 -1.31403530e+00 1.45376682e+00 4.23311800e-01
-1.27722359e+00 -8.46751273e-01 -7.53105879e-01 -1.50338650e-01
-3.93766224e-01 8.97246823e-02 2.95509517e-01 8.71655166e-01
-1.48146737e+00 5.24689257e-01 -5.14570355e-01 -4.88968670e-01
2.13372916e-01 -7.64054731e-02 -7.64905214e-02 4.76334281e-02
-1.14981616e+00 7.38437057e-01 1.60238296e-01 -4.47501928e-01
-1.06640446e+00 -1.01148605e+00 -9.30877566e-01 5.10604419e-02
3.64745885e-01 -9.24518764e-01 1.59042799e+00 -7.33040392e-01
-2.06126356e+00 5.54702699e-01 -3.01561445e-01 -6.08161211e-01
4.77648735e-01 -3.89943689e-01 -8.31892967e-01 7.02010766e-02
2.47086119e-02 1.09430814e+00 1.33503950e+00 -7.95116603e-01
-5.53473532e-01 2.76384860e-01 -4.35098767e-01 3.92446905e-01
-6.18455261e-02 1.31158471e-01 -4.52152967e-01 -1.21700442e+00
-2.52336800e-01 -7.53802657e-01 -6.05493672e-02 -5.29070981e-02
-2.62307644e-01 -6.54476210e-02 1.19396651e+00 -8.84415030e-01
1.08267725e+00 -2.21549606e+00 1.63657051e-02 -3.14817399e-01
-1.54333100e-01 4.97972459e-01 -4.82196748e-01 5.36609888e-01
-3.62140266e-03 3.35493207e-01 5.49223693e-03 -5.88316023e-01
8.29715654e-02 1.43536448e-01 -5.93258500e-01 -1.66617796e-01
4.38254446e-01 1.07443082e+00 -8.21118414e-01 -2.80360729e-01
6.30357742e-01 8.32332969e-01 -8.21226239e-01 1.05117992e-01
-2.93646663e-01 4.92594481e-01 1.58173189e-01 3.60349953e-01
1.26994252e-01 1.47082105e-01 -2.88277626e-01 7.61268958e-02
-1.25412777e-01 4.55490947e-01 -1.02534676e+00 1.97121167e+00
-9.15929735e-01 1.17184985e+00 4.12435830e-02 -5.02641916e-01
1.13239324e+00 7.26183057e-01 9.03394744e-02 -7.34508991e-01
4.96285595e-03 2.51764059e-01 -8.63018855e-02 -4.42313254e-01
8.03642392e-01 -2.51305997e-01 1.08441569e-01 2.18396217e-01
5.24594367e-01 -8.66154432e-01 3.77257347e-01 2.11829960e-01
7.39365637e-01 7.06150606e-02 1.27323568e-02 -1.05439007e-01
1.54848233e-01 -4.73933697e-01 1.29420459e-01 8.37553263e-01
1.69310838e-01 9.57423151e-01 1.68831572e-02 1.65472195e-01
-1.35154533e+00 -1.35147715e+00 9.87484120e-04 1.38748515e+00
-5.33150673e-01 -6.50657535e-01 -9.00921524e-01 1.56672765e-02
-1.91997245e-01 1.10907376e+00 -1.10475115e-01 -2.23257422e-01
-4.86859143e-01 -1.88256964e-01 7.32784808e-01 4.34437841e-01
2.67088532e-01 -1.40217793e+00 -1.12471186e-01 4.89478588e-01
-3.04151535e-01 -1.23337507e+00 -9.71558273e-01 -6.38893619e-02
-7.26464570e-01 -2.63443172e-01 -9.55704808e-01 -8.16701531e-01
1.33413792e-01 1.00960001e-01 1.23609328e+00 -2.96658039e-01
-3.45579892e-01 5.21181405e-01 -4.75478530e-01 -6.79452419e-01
-1.03467715e+00 7.98481107e-02 1.07678421e-01 -1.52007729e-01
-7.32449889e-02 -6.96117640e-01 -2.42356777e-01 1.92376181e-01
-8.58320653e-01 1.75689965e-01 5.64321578e-01 9.42550838e-01
5.17691851e-01 -1.78374708e-01 9.57530856e-01 -5.18271685e-01
1.03929472e+00 -9.73872319e-02 -3.07609648e-01 8.92647803e-02
-5.26747942e-01 -7.50397891e-02 6.28658950e-01 -7.42334247e-01
-1.22170353e+00 -2.81737953e-01 -5.01208186e-01 -7.15747595e-01
-2.72643209e-01 3.75863016e-01 -1.63808465e-01 3.07383627e-01
8.44688296e-01 4.29149747e-01 2.40363907e-02 -4.69986558e-01
8.70351732e-01 8.63102436e-01 8.19310904e-01 -4.99778658e-01
7.77019858e-01 -9.10690129e-02 -4.40423340e-01 -1.43178773e+00
-3.49064291e-01 -1.12368181e-01 -1.97843641e-01 -6.09095804e-02
8.90265167e-01 -1.01556730e+00 -1.46220267e-01 5.33595562e-01
-1.07773018e+00 -6.20974004e-01 -7.08216012e-01 3.73473287e-01
-9.23416972e-01 1.03877440e-01 -7.04227567e-01 -7.78986394e-01
-5.65534770e-01 -1.05273306e+00 1.19845879e+00 2.06488341e-01
-3.91253471e-01 -9.63437319e-01 -9.99486074e-02 1.48427978e-01
8.38703215e-01 -3.15616012e-01 6.32140100e-01 -3.98264885e-01
-3.34673792e-01 5.97013757e-02 1.66607648e-01 6.01136327e-01
2.19629169e-01 2.16977298e-01 -1.28490376e+00 -3.32888603e-01
-1.76542357e-01 -1.83381364e-01 6.86120450e-01 5.47654271e-01
1.06654775e+00 -4.20226097e-01 4.84811477e-02 7.54321754e-01
7.34089851e-01 5.36423028e-01 8.02435696e-01 -1.96223736e-01
5.79584897e-01 4.07093972e-01 2.99863368e-01 2.89725870e-01
-3.51324044e-02 5.81446707e-01 -8.67842138e-02 -1.89842746e-01
-8.68764818e-01 -6.32913113e-01 6.44979775e-01 1.06405830e+00
2.34484628e-01 -5.02838016e-01 -6.57799482e-01 5.65514505e-01
-1.37611341e+00 -1.19650614e+00 1.82143256e-01 2.00014710e+00
9.54181314e-01 1.68839619e-01 4.30958383e-02 1.11488625e-01
7.80632734e-01 2.46604592e-01 -3.61606598e-01 -7.09202528e-01
-2.40106404e-01 6.99703336e-01 6.76976964e-02 6.69202983e-01
-8.90209138e-01 1.39412975e+00 6.65615511e+00 1.29280424e+00
-1.29077947e+00 1.39152333e-01 5.31308949e-01 -3.00007612e-01
-2.75793701e-01 -2.70837218e-01 -7.03043938e-01 4.70398337e-01
1.30345857e+00 -4.25954998e-01 7.34661698e-01 7.29783595e-01
3.87768567e-01 2.51421422e-01 -1.07587028e+00 1.15258408e+00
9.87820253e-02 -1.50043774e+00 2.59128004e-01 -3.07607986e-02
7.63331056e-01 5.00546880e-02 4.17526335e-01 6.10040188e-01
3.21148515e-01 -1.20372701e+00 9.85670626e-01 3.70236456e-01
1.34281945e+00 -7.09614456e-01 1.57552063e-01 2.95742214e-01
-1.21344352e+00 1.74584672e-01 -2.33822927e-01 2.22685501e-01
3.94908160e-01 3.82677108e-01 -1.00395453e+00 3.01753014e-01
6.24797225e-01 3.74139667e-01 -3.04092199e-01 9.31470096e-01
-3.39790434e-01 8.30347359e-01 -2.41897836e-01 5.53080030e-02
1.59005418e-01 -5.90163982e-03 8.37918878e-01 1.54144418e+00
8.30105126e-01 -1.74794972e-01 7.98260272e-02 8.72737288e-01
-4.32618745e-02 1.29236743e-01 -6.19281948e-01 -3.65482599e-01
6.81808889e-01 1.04161966e+00 -3.19185972e-01 -5.31777442e-01
-2.59970933e-01 1.14365125e+00 -2.53623277e-01 6.79917037e-01
-7.42660463e-01 -6.00019753e-01 7.77320683e-01 1.96805447e-01
5.19628227e-01 -2.49984637e-01 8.72662757e-04 -1.04517853e+00
-3.51501644e-01 -1.23227620e+00 -2.33488500e-01 -1.18986273e+00
-1.28227794e+00 8.54973078e-01 -1.53536305e-01 -1.04585803e+00
-7.21772134e-01 -5.70409298e-01 -7.85690367e-01 1.18875921e+00
-1.14410615e+00 -1.16565919e+00 -2.39914302e-02 6.17520928e-01
1.35547268e+00 -5.82111895e-01 9.07273173e-01 2.61336058e-01
-2.59117335e-01 6.60444021e-01 1.70505587e-02 -9.65456385e-03
8.76976907e-01 -1.12713516e+00 1.24485636e+00 1.00202942e+00
4.78974611e-01 4.44398135e-01 8.80524039e-01 -6.73015952e-01
-1.20528126e+00 -1.30266166e+00 9.81130779e-01 -3.16600174e-01
6.94421232e-01 -5.92244804e-01 -7.47614801e-01 5.39645433e-01
5.92572808e-01 -1.45585075e-01 4.47201788e-01 -3.36808637e-02
-3.12034130e-01 1.13090212e-02 -9.62903619e-01 9.66886342e-01
1.10891509e+00 -8.38351011e-01 -4.77851182e-01 1.19651712e-01
1.30782866e+00 -6.46908581e-01 -7.02440143e-01 -5.05585633e-02
3.25324982e-01 -7.82505810e-01 1.00518572e+00 -4.84986573e-01
3.21673661e-01 -4.64244336e-02 -3.09235841e-01 -1.72149038e+00
-3.24656427e-01 -1.36055851e+00 -1.13712372e-02 1.51568627e+00
5.90946674e-01 -6.19713604e-01 4.78603154e-01 8.56593400e-02
-6.51795805e-01 -3.77380252e-01 -8.26583385e-01 -1.02710843e+00
2.21601129e-01 -7.95885503e-01 5.99393547e-01 6.89673543e-01
-2.43787304e-01 6.63003087e-01 -5.88260114e-01 -1.55899271e-01
2.89777398e-01 -2.88905352e-01 8.98137391e-01 -9.42770839e-01
-5.77511013e-01 -5.58028519e-01 2.06809095e-03 -1.17231190e+00
1.32123949e-02 -8.32375646e-01 1.83176830e-01 -1.52308297e+00
-4.47004586e-01 -6.16434291e-02 3.24152589e-01 3.32243264e-01
-1.15577810e-01 9.61740613e-02 5.15222251e-01 -1.97577149e-01
7.93939456e-02 7.82719672e-01 1.46360350e+00 -1.24312997e-01
-4.02239352e-01 1.78638864e-02 -6.30667925e-01 4.74804431e-01
9.17474091e-01 -2.11655110e-01 -6.97612882e-01 -4.33735728e-01
-1.92729056e-01 2.50696510e-01 3.59490514e-01 -1.10062313e+00
1.29894435e-01 -8.98785368e-02 1.68468550e-01 -4.80221748e-01
6.68214202e-01 -2.87423253e-01 1.88910797e-01 1.23286612e-01
-4.15604472e-01 7.80281052e-02 4.68550742e-01 3.70194644e-01
-2.05200583e-01 -1.26646265e-01 6.59213245e-01 -1.20578639e-01
-6.27422571e-01 3.09031785e-01 -7.71798611e-01 2.54838526e-01
5.20647287e-01 -2.49608308e-01 -5.03126919e-01 -9.18711245e-01
-8.29108119e-01 -2.83800036e-01 1.26074150e-01 7.80205905e-01
9.10899222e-01 -1.34010422e+00 -1.03653610e+00 4.81731325e-01
-5.45061268e-02 -1.94705844e-01 4.02884871e-01 4.08039898e-01
-1.59161896e-01 4.63724196e-01 1.75245218e-02 -5.70858598e-01
-1.27683067e+00 4.20271784e-01 2.27706835e-01 1.91294923e-01
-6.75219774e-01 9.38988090e-01 3.45537245e-01 -3.81650746e-01
1.41967222e-01 -3.41174543e-01 2.60069638e-01 -5.04168198e-02
6.16616011e-01 2.73733407e-01 1.10367946e-01 -4.94419247e-01
9.05986205e-02 3.45729351e-01 3.88968103e-02 -7.38292694e-01
1.12713051e+00 -1.79080337e-01 4.55804497e-01 4.45814401e-01
9.09295559e-01 6.58096895e-02 -1.34437931e+00 -2.25791112e-01
-4.48730171e-01 -3.59975487e-01 1.19126424e-01 -8.85566533e-01
-8.35770369e-01 1.10855341e+00 5.10419905e-01 2.02205390e-01
1.04453611e+00 1.21500660e-02 8.48546267e-01 2.71600544e-01
6.57946011e-03 -1.15005875e+00 4.86102939e-01 8.83642495e-01
1.35083091e+00 -8.61392379e-01 -4.42520350e-01 -2.92259634e-01
-8.41771066e-01 8.59454572e-01 3.64590704e-01 2.03912500e-02
3.91261160e-01 6.05770588e-01 1.41432926e-01 1.95549816e-01
-1.09629273e+00 -2.27876872e-01 4.71233606e-01 1.03191090e+00
6.88618183e-01 6.33462220e-02 2.29475200e-01 1.95121661e-01
-9.72373128e-01 -2.23708466e-01 4.44951147e-01 5.79987645e-01
-6.75178230e-01 -1.03898013e+00 -5.82400024e-01 3.19598734e-01
-2.78730869e-01 -6.57918572e-01 -1.94011450e-01 5.25142312e-01
-1.84632897e-01 1.34450495e+00 9.72520262e-02 -4.22241539e-01
3.71580034e-01 3.79870892e-01 4.25284326e-01 -7.59311616e-01
-5.29585421e-01 7.16769457e-01 2.06545755e-01 -2.71555185e-01
4.85932641e-02 -5.96281946e-01 -9.84154224e-01 -4.32796270e-01
-3.65173936e-01 -4.70658168e-02 8.73838603e-01 6.64503098e-01
4.44035381e-01 9.75750983e-01 4.27495807e-01 -1.00050390e+00
-5.20415187e-01 -1.19587302e+00 -3.33259434e-01 5.83723933e-02
3.73737663e-01 -1.87491998e-01 -3.31392586e-01 4.32498634e-01] | [15.189194679260254, 6.366114139556885] |
08bf005d-b4fb-4366-8779-f8f8afa38e67 | capturing-the-motion-of-every-joint-3d-human | 2303.00298 | null | https://arxiv.org/abs/2303.00298v1 | https://arxiv.org/pdf/2303.00298v1.pdf | Capturing the motion of every joint: 3D human pose and shape estimation with independent tokens | In this paper we present a novel method to estimate 3D human pose and shape from monocular videos. This task requires directly recovering pixel-alignment 3D human pose and body shape from monocular images or videos, which is challenging due to its inherent ambiguity. To improve precision, existing methods highly rely on the initialized mean pose and shape as prior estimates and parameter regression with an iterative error feedback manner. In addition, video-based approaches model the overall change over the image-level features to temporally enhance the single-frame feature, but fail to capture the rotational motion at the joint level, and cannot guarantee local temporal consistency. To address these issues, we propose a novel Transformer-based model with a design of independent tokens. First, we introduce three types of tokens independent of the image feature: \textit{joint rotation tokens, shape token, and camera token}. By progressively interacting with image features through Transformer layers, these tokens learn to encode the prior knowledge of human 3D joint rotations, body shape, and position information from large-scale data, and are updated to estimate SMPL parameters conditioned on a given image. Second, benefiting from the proposed token-based representation, we further use a temporal model to focus on capturing the rotational temporal information of each joint, which is empirically conducive to preventing large jitters in local parts. Despite being conceptually simple, the proposed method attains superior performances on the 3DPW and Human3.6M datasets. Using ResNet-50 and Transformer architectures, it obtains 42.0 mm error on the PA-MPJPE metric of the challenging 3DPW, outperforming state-of-the-art counterparts by a large margin. Code will be publicly available at https://github.com/yangsenius/INT_HMR_Model | ['Gang Yu', 'Wankou Yang', 'Guozhong Luo', 'Gang Liu', 'Wen Heng', 'Sen yang'] | 2023-03-01 | null | null | null | null | ['3d-human-pose-estimation', '3d-human-pose-and-shape-estimation'] | ['computer-vision', 'computer-vision'] | [-5.67909367e-02 -2.29736418e-01 -1.37937576e-01 -2.62645006e-01
-5.27006030e-01 -2.69060612e-01 3.76748860e-01 -4.65281665e-01
-5.40892124e-01 4.73870963e-01 2.22847834e-01 4.25809443e-01
1.78538933e-01 -2.68042117e-01 -8.50095510e-01 -7.12361634e-01
-8.79923925e-02 3.35411280e-01 1.93695933e-01 -9.12254527e-02
-1.62095398e-01 2.63663411e-01 -1.35416186e+00 -1.75847039e-01
5.35013199e-01 1.18476343e+00 1.73614070e-01 6.06323242e-01
4.07191247e-01 5.65801620e-01 -4.19642001e-01 -2.36499369e-01
4.10927415e-01 -2.67431915e-01 -5.86238980e-01 3.67589444e-01
7.19549358e-01 -6.55246139e-01 -7.70285189e-01 8.03470135e-01
8.36611688e-01 1.48422897e-01 2.27857575e-01 -1.06674230e+00
-3.64059657e-01 5.36415912e-02 -6.70893848e-01 4.53555919e-02
6.73229456e-01 4.23293412e-01 6.85864687e-01 -9.65336323e-01
7.76993394e-01 1.18597698e+00 6.72482669e-01 4.94591027e-01
-1.00081658e+00 -5.86321175e-01 3.15514207e-01 3.10523510e-01
-1.56248748e+00 -4.91409510e-01 9.63596821e-01 -4.05109107e-01
7.91728795e-01 3.26031968e-02 1.07188308e+00 1.16188538e+00
2.20915839e-01 1.02285028e+00 7.57379234e-01 -1.76571757e-01
-1.61746249e-01 -4.55445319e-01 -2.23749131e-01 8.21543276e-01
3.41990255e-02 2.64641464e-01 -8.61618996e-01 1.17858537e-01
1.07552338e+00 1.21386498e-01 -5.53962648e-01 -8.16125870e-01
-1.34006155e+00 3.18353385e-01 5.44743299e-01 8.65520090e-02
-3.36620778e-01 5.14252424e-01 2.75799960e-01 2.57354341e-02
3.20378721e-01 -7.41472244e-02 -5.61468959e-01 -3.14304411e-01
-7.48475432e-01 4.03150439e-01 3.71957004e-01 1.09259987e+00
5.55525601e-01 1.35950834e-01 -1.58670023e-01 6.95696056e-01
4.35711205e-01 8.58851492e-01 4.47955370e-01 -9.66794252e-01
5.92587531e-01 3.17735434e-01 2.45186225e-01 -1.15009665e+00
-7.32596397e-01 -5.26308119e-01 -7.79546201e-01 -1.31332278e-01
4.82439667e-01 -4.54665534e-02 -1.09698725e+00 1.82216024e+00
6.75399184e-01 1.52545661e-01 -4.47356254e-01 1.36581755e+00
7.42174506e-01 3.55825275e-01 -2.80126184e-01 6.87675029e-02
1.29597497e+00 -9.97458279e-01 -5.34066617e-01 -3.98488849e-01
3.53354931e-01 -6.82487428e-01 6.42697871e-01 3.32213044e-01
-1.20031929e+00 -8.04939091e-01 -9.91250157e-01 -1.73433393e-01
1.24586739e-01 4.36060786e-01 3.62454414e-01 4.18012172e-01
-9.12278891e-01 4.38030124e-01 -1.17743206e+00 -3.27763140e-01
-3.25599797e-02 4.48634952e-01 -5.15331686e-01 -1.50654688e-01
-1.17639303e+00 8.47974896e-01 9.98406559e-02 6.37061417e-01
-7.13366807e-01 -4.66115445e-01 -1.21274889e+00 -1.81298941e-01
5.38697124e-01 -9.97879982e-01 1.11559880e+00 -6.33828282e-01
-1.72955489e+00 7.44677842e-01 -1.99715883e-01 -3.96127671e-01
7.81114936e-01 -6.46128476e-01 -5.29366173e-02 4.55196053e-01
-1.40877828e-01 8.26314390e-01 1.15976965e+00 -1.03098416e+00
-3.66595417e-01 -4.80063587e-01 -3.99409793e-02 4.76655960e-01
-9.62323844e-02 -2.54878938e-01 -1.11264741e+00 -7.64392614e-01
3.85075599e-01 -1.26630747e+00 -1.92553475e-01 2.79698133e-01
-2.30708882e-01 1.55375019e-01 4.41361129e-01 -9.22904968e-01
9.37690139e-01 -2.11700368e+00 5.05150259e-01 2.20199749e-02
1.35256246e-01 1.48476556e-01 -2.19310224e-02 8.17670971e-02
9.09586847e-02 -4.16940093e-01 5.00787348e-02 -5.98486364e-01
-2.48941854e-02 1.54914364e-01 1.34734064e-01 8.53540242e-01
7.39456713e-02 1.04661977e+00 -6.90777004e-01 -4.93139058e-01
6.32595420e-01 8.30045044e-01 -7.41319776e-01 1.30501583e-01
1.98595934e-02 6.71711266e-01 -4.03812528e-01 7.50841081e-01
6.90220594e-01 -3.38039726e-01 1.58973724e-01 -5.21704555e-01
1.25576809e-01 5.55327199e-02 -1.32965362e+00 2.17983532e+00
-1.66794896e-01 2.83893794e-01 1.43314570e-01 -9.69669402e-01
7.01912582e-01 4.13186967e-01 9.01564360e-01 -6.51742399e-01
3.50955278e-01 -2.08206717e-02 -2.04592898e-01 -2.85448253e-01
5.81752002e-01 2.52089679e-01 -2.27245875e-02 2.05675303e-03
1.22833967e-01 2.26856582e-02 -2.22652461e-02 -1.01246178e-01
8.40947688e-01 7.83592343e-01 2.05907896e-01 7.05027580e-02
4.18760091e-01 -4.09184307e-01 8.35827529e-01 4.32641655e-01
-4.46634442e-01 1.03457069e+00 9.62717757e-02 -5.20048738e-01
-1.01082420e+00 -1.17250121e+00 3.72181647e-02 6.05303407e-01
3.71176571e-01 -4.71439600e-01 -4.96443003e-01 -3.16699117e-01
7.44300857e-02 -1.09733507e-01 -5.57237327e-01 -1.49957359e-01
-8.83467674e-01 -5.46973825e-01 3.24151188e-01 6.13041043e-01
6.88151419e-01 -6.28191829e-01 -9.23717558e-01 1.66719094e-01
-6.02204680e-01 -1.42187035e+00 -8.99194241e-01 -1.60305351e-01
-7.91696131e-01 -1.03147435e+00 -1.04259050e+00 -5.29877424e-01
7.04002678e-01 2.30323195e-01 8.45132828e-01 -2.73598284e-02
-4.43379998e-01 5.36680043e-01 -3.06643307e-01 1.78770751e-01
2.77258635e-01 1.68482754e-02 2.74227232e-01 1.39140626e-02
3.87494117e-02 -5.71468532e-01 -9.85482097e-01 5.50201356e-01
-6.20186567e-01 2.67230719e-01 5.70526183e-01 9.77537572e-01
5.22653461e-01 -4.06084001e-01 -2.82520093e-02 -5.67691363e-02
-1.50028363e-01 4.76416051e-02 -4.78727758e-01 -2.22105533e-02
-2.00623408e-01 -8.61110017e-02 3.05482477e-01 -6.20368004e-01
-9.25882638e-01 4.47690755e-01 -1.25784636e-01 -8.66939247e-01
1.26507366e-02 3.44254315e-01 -5.55201583e-02 -4.85322326e-02
4.10696834e-01 3.29025924e-01 2.09515452e-01 -3.70649725e-01
2.43205696e-01 1.64277643e-01 8.51212859e-01 -7.08935022e-01
9.10602510e-01 6.85544670e-01 -5.89146093e-02 -8.23688269e-01
-7.13208795e-01 -6.21290863e-01 -8.39984953e-01 -3.85947168e-01
8.69835079e-01 -1.36919045e+00 -8.11004281e-01 8.18422258e-01
-1.05917096e+00 -2.65213788e-01 -1.07725233e-01 8.26438010e-01
-7.72311270e-01 6.87291861e-01 -8.69855106e-01 -5.53056657e-01
-3.64354163e-01 -1.08027136e+00 1.38653910e+00 -4.90514701e-03
-2.23584846e-01 -5.80093980e-01 -1.61051288e-01 4.71754938e-01
1.12696536e-01 3.68939608e-01 2.01528758e-01 9.92520154e-02
-6.79657340e-01 -3.23137969e-01 -2.20581517e-03 3.46639603e-01
1.66996628e-01 -2.60901690e-01 -7.18280077e-01 -6.93569720e-01
2.57380810e-02 -3.10945958e-01 5.93326986e-01 7.30150163e-01
8.69385839e-01 -7.12670311e-02 -2.13328913e-01 8.97417843e-01
1.02482843e+00 -1.83911517e-01 5.56423128e-01 3.39165837e-01
9.64776278e-01 3.58028173e-01 7.41864800e-01 7.46273398e-01
5.85431576e-01 1.09958196e+00 2.79896528e-01 -3.88969183e-02
-2.58811027e-01 -2.92540044e-01 5.86149633e-01 8.82062078e-01
-3.58540028e-01 1.36166513e-01 -6.21226847e-01 3.96460354e-01
-1.94767320e+00 -8.86879385e-01 3.32946211e-01 2.23970056e+00
7.86748648e-01 8.16322938e-02 1.84361070e-01 -1.02850676e-01
5.81439316e-01 3.26394677e-01 -6.19353831e-01 3.76876801e-01
-4.02358770e-02 -1.26392975e-01 5.62758923e-01 4.08817619e-01
-1.14398670e+00 9.48010266e-01 5.69715071e+00 4.97639030e-01
-1.12751007e+00 -1.62600949e-01 2.55092591e-01 -4.99373436e-01
1.46781340e-01 -1.59851789e-01 -9.84612763e-01 4.29339647e-01
3.62902850e-01 1.70993343e-01 3.04406255e-01 5.38098991e-01
2.84755081e-01 -1.65062711e-01 -1.13266981e+00 1.23783660e+00
3.03038150e-01 -8.45940650e-01 -9.56197605e-02 8.55494067e-02
6.17644846e-01 5.39410822e-02 4.70823944e-02 3.43412347e-02
2.94000171e-02 -7.74180353e-01 1.09545434e+00 5.75579405e-01
7.35920072e-01 -5.38562953e-01 5.67433417e-01 2.78582275e-01
-1.46463764e+00 2.80617736e-02 -2.83897489e-01 -1.37392253e-01
2.99143612e-01 5.32575667e-01 -4.69936132e-01 7.63031244e-01
9.82823014e-01 9.37670529e-01 -3.74280721e-01 9.00032163e-01
-2.94329226e-01 1.46329269e-01 -6.04451597e-01 4.11350518e-01
-1.50481135e-01 1.05467001e-02 5.09850144e-01 8.24115872e-01
3.34310353e-01 8.80616978e-02 4.94381130e-01 6.27775609e-01
1.05323277e-01 -1.70749202e-01 -3.31353754e-01 3.07616621e-01
2.02526644e-01 1.09654617e+00 -4.62358028e-01 -2.07973078e-01
-4.86544967e-01 1.19889092e+00 2.07249448e-01 4.14473951e-01
-1.03819692e+00 7.06950426e-02 6.54821694e-01 7.77852014e-02
6.71070814e-01 -6.70361817e-01 1.11088064e-02 -1.56441867e+00
5.78295708e-01 -1.00172520e+00 2.80373782e-01 -7.80942619e-01
-9.74747479e-01 4.01030511e-01 1.89014092e-01 -1.50157070e+00
-6.02470696e-01 -5.24056673e-01 9.31643695e-03 5.90160728e-01
-1.37914217e+00 -1.38971937e+00 -4.96256590e-01 7.69384861e-01
4.38084543e-01 2.80642837e-01 4.20532644e-01 3.54400963e-01
-3.66142780e-01 6.99352980e-01 -1.91804141e-01 3.99794012e-01
1.01876056e+00 -9.21643078e-01 5.93390763e-01 8.33686948e-01
4.30304408e-02 5.83244503e-01 7.85915136e-01 -6.32481277e-01
-1.68764377e+00 -7.04980195e-01 5.91534436e-01 -5.58010578e-01
3.50968033e-01 -4.31925088e-01 -4.38587666e-01 7.77413309e-01
-1.78956091e-01 3.22535813e-01 1.07633680e-01 -3.96988243e-02
-2.18271926e-01 -1.88829049e-01 -8.65079880e-01 5.97078085e-01
1.31007850e+00 -4.26690698e-01 -5.36327064e-01 6.18983060e-02
6.15069807e-01 -1.00496113e+00 -1.00036311e+00 5.27092934e-01
8.96879017e-01 -8.60937953e-01 1.25084865e+00 -7.29406904e-03
2.80339032e-01 -5.32588124e-01 -1.33320287e-01 -1.13553214e+00
-2.98357397e-01 -6.91797853e-01 -3.11061710e-01 8.10160875e-01
-1.63094655e-01 -4.14401710e-01 9.86517370e-01 5.79762042e-01
8.57844483e-03 -6.73451006e-01 -1.15920889e+00 -8.26002896e-01
-1.89855099e-01 -3.77688289e-01 3.83940458e-01 5.77047765e-01
-8.51742700e-02 6.55629635e-02 -9.67825055e-01 2.20939651e-01
7.72047400e-01 5.90073876e-02 1.19331539e+00 -7.67845094e-01
-4.78620172e-01 -5.31539395e-02 -6.24330640e-01 -1.69711566e+00
-5.83335869e-02 -4.51981544e-01 2.21050620e-01 -1.08319211e+00
1.06622584e-01 -2.81657130e-02 -1.00917704e-01 4.35001194e-01
-1.51379243e-01 3.70480418e-01 4.80649710e-01 2.38506168e-01
-5.40449083e-01 8.92180085e-01 1.38734806e+00 -1.29691154e-01
-6.93497760e-03 -1.02173306e-01 -1.08834736e-01 7.88862765e-01
4.40254509e-01 -1.54925257e-01 -2.66568452e-01 -5.98683119e-01
-3.68684158e-02 2.78939813e-01 8.14894617e-01 -1.16842258e+00
4.03692901e-01 4.53121066e-02 8.18548083e-01 -8.76855195e-01
6.90870881e-01 -7.30249524e-01 2.75621861e-01 5.35095572e-01
-7.33358832e-03 2.16150507e-01 3.31560709e-02 5.30199826e-01
-1.98238298e-01 2.77408123e-01 6.47338569e-01 -1.57717124e-01
-7.18359053e-01 6.87170863e-01 -5.15233837e-02 7.10556880e-02
7.17896760e-01 -4.13951933e-01 -3.68514750e-03 -5.12209117e-01
-7.67229199e-01 3.56267095e-01 7.84160435e-01 6.26456022e-01
7.51084149e-01 -1.42973113e+00 -5.81661999e-01 3.63798231e-01
8.19845945e-02 2.70856827e-01 5.69141030e-01 1.00225186e+00
-4.24987793e-01 4.57094133e-01 -2.58183897e-01 -1.02662373e+00
-1.21511400e+00 3.95453840e-01 4.37536627e-01 -2.70836383e-01
-8.67646337e-01 6.90182328e-01 3.16300452e-01 -3.79700214e-01
3.14759970e-01 -4.02203232e-01 1.91411749e-01 -2.57057220e-01
2.98560202e-01 2.02665746e-01 -1.02614552e-01 -8.81325424e-01
-4.36304510e-01 1.05238843e+00 2.23562354e-03 -9.97277200e-02
1.04465008e+00 -4.58950758e-01 2.83074796e-01 2.39387840e-01
1.24084699e+00 -1.35849833e-01 -1.84932888e+00 -4.76024628e-01
-4.29749995e-01 -6.41207755e-01 -2.50691473e-01 -5.04920483e-01
-1.26020169e+00 7.16954410e-01 6.33104801e-01 -5.92309475e-01
1.03394282e+00 -7.63055757e-02 9.37362075e-01 2.27809787e-01
5.77165067e-01 -1.28494155e+00 2.59002864e-01 4.79121357e-01
8.68005931e-01 -1.14325750e+00 3.60819280e-01 -4.80006844e-01
-3.83319378e-01 9.62637722e-01 8.02340031e-01 -6.32388443e-02
5.09658873e-01 1.60868436e-01 8.84077176e-02 -4.09931503e-02
-5.44812143e-01 -1.59268573e-01 5.21298110e-01 4.82922971e-01
4.15815353e-01 -1.21691801e-01 -1.27726838e-01 3.08668733e-01
-2.32680172e-01 2.03603312e-01 9.96974781e-02 1.01073790e+00
-1.30494460e-01 -8.70964170e-01 -5.02517641e-01 -2.83689313e-02
-3.80795032e-01 1.32510126e-01 4.01207469e-02 8.25352967e-01
-5.50626719e-04 6.33133531e-01 -3.98254022e-02 -4.99724090e-01
4.32941020e-01 -2.12297067e-01 7.86178529e-01 -2.05603421e-01
-2.76192904e-01 4.33259547e-01 7.49911815e-02 -9.51119125e-01
-6.30466521e-01 -8.47110748e-01 -1.12998474e+00 -3.71210724e-01
-2.05979690e-01 -2.30884969e-01 3.87079597e-01 9.94165540e-01
3.28328609e-01 3.20885628e-01 3.77115220e-01 -1.30174077e+00
-7.31589079e-01 -8.46529841e-01 -4.56586629e-01 5.09442031e-01
3.28800321e-01 -9.14095223e-01 -1.93384647e-01 8.99707228e-02] | [7.191433906555176, -0.7443550229072571] |
e4edcd7e-75e5-4c09-a07e-75abc1475039 | xlda-cross-lingual-data-augmentation-for | 1905.11471 | null | https://arxiv.org/abs/1905.11471v1 | https://arxiv.org/pdf/1905.11471v1.pdf | XLDA: Cross-Lingual Data Augmentation for Natural Language Inference and Question Answering | While natural language processing systems often focus on a single language, multilingual transfer learning has the potential to improve performance, especially for low-resource languages. We introduce XLDA, cross-lingual data augmentation, a method that replaces a segment of the input text with its translation in another language. XLDA enhances performance of all 14 tested languages of the cross-lingual natural language inference (XNLI) benchmark. With improvements of up to $4.8\%$, training with XLDA achieves state-of-the-art performance for Greek, Turkish, and Urdu. XLDA is in contrast to, and performs markedly better than, a more naive approach that aggregates examples in various languages in a way that each example is solely in one language. On the SQuAD question answering task, we see that XLDA provides a $1.0\%$ performance increase on the English evaluation set. Comprehensive experiments suggest that most languages are effective as cross-lingual augmentors, that XLDA is robust to a wide range of translation quality, and that XLDA is even more effective for randomly initialized models than for pretrained models. | ['Caiming Xiong', 'Nitish Shirish Keskar', 'Richard Socher', 'Jasdeep Singh', 'Bryan McCann'] | 2019-05-27 | xlda-cross-lingual-data-augmentation-for-1 | https://openreview.net/forum?id=BJgAf6Etwr | https://openreview.net/pdf?id=BJgAf6Etwr | iclr-2020-1 | ['cross-lingual-natural-language-inference'] | ['natural-language-processing'] | [-2.47732952e-01 7.36137852e-02 -3.98621053e-01 -3.82556558e-01
-1.52613044e+00 -7.82867670e-01 7.15545595e-01 3.23806018e-01
-9.11964476e-01 1.00657201e+00 3.37085187e-01 -8.69796097e-01
3.68744165e-01 -6.73986912e-01 -8.90457690e-01 -1.69682801e-01
2.04473153e-01 8.14762890e-01 -2.09595457e-01 -4.87738848e-01
-3.21425200e-01 9.58493203e-02 -9.16301191e-01 4.97170269e-01
1.00988245e+00 5.41972876e-01 -1.82809889e-01 3.51885676e-01
-5.33597887e-01 6.48845732e-01 -6.45974100e-01 -8.00185621e-01
2.09920898e-01 -2.68753499e-01 -1.07552242e+00 -1.68870494e-01
8.08424532e-01 -3.22589666e-01 6.09755851e-02 6.91741705e-01
4.42355990e-01 -6.03931434e-02 4.68073219e-01 -8.78707647e-01
-9.84689116e-01 1.01124847e+00 -4.80078727e-01 1.83429584e-01
5.48778713e-01 1.60791501e-01 1.32702267e+00 -1.16769779e+00
6.15213811e-01 1.44884670e+00 8.95209730e-01 3.05004507e-01
-1.42706645e+00 -6.43209815e-01 1.22353479e-01 -2.35552102e-01
-1.19913638e+00 -4.11826432e-01 2.99876392e-01 -8.35101679e-02
1.50327921e+00 -9.58846062e-02 -3.19950990e-02 8.88726115e-01
1.45878136e-01 9.16145444e-01 1.30228651e+00 -8.66620421e-01
-2.33944505e-01 3.18382472e-01 2.19566286e-01 8.02680552e-01
1.76613972e-01 -5.35558984e-02 -3.86385411e-01 -1.71811610e-01
2.77221262e-01 -5.86183906e-01 3.61031257e-02 1.43541679e-01
-1.36929941e+00 8.41298163e-01 3.19045573e-01 3.93720955e-01
-1.02320142e-01 3.21082324e-02 6.99768245e-01 7.26837158e-01
6.71531200e-01 8.30254495e-01 -1.05623162e+00 -1.17094964e-01
-5.47319770e-01 2.41433561e-01 8.33814621e-01 9.11837339e-01
9.12001908e-01 1.81044519e-01 2.59994101e-02 1.02868259e+00
-3.29949819e-02 6.73241854e-01 6.61074221e-01 -8.04855704e-01
8.46671522e-01 6.59299731e-01 -5.54902069e-02 -3.65624517e-01
-2.34952390e-01 -3.44665170e-01 -4.76561636e-01 3.12155243e-02
7.88048446e-01 -4.95898277e-01 -7.63683081e-01 1.91338766e+00
7.88980201e-02 -4.89170015e-01 4.85951006e-01 2.68274635e-01
6.49051309e-01 9.55381632e-01 3.47772419e-01 -8.55267495e-02
1.32628977e+00 -1.11399841e+00 -6.75799549e-01 -6.02440655e-01
1.24134851e+00 -1.08844745e+00 1.78359187e+00 3.07986766e-01
-1.08322370e+00 -6.56287968e-01 -9.56687689e-01 -5.98427474e-01
-7.09680319e-01 2.68401414e-01 5.97728610e-01 6.42590106e-01
-1.02594078e+00 1.41603738e-01 -6.21793211e-01 -4.74223495e-01
-1.39428496e-01 1.54809445e-01 -4.69846129e-01 -5.64853430e-01
-1.32583272e+00 1.43628967e+00 4.87619281e-01 -3.69361281e-01
-3.31152350e-01 -9.59733367e-01 -1.17592478e+00 -3.15250099e-01
2.14020923e-01 -4.08223242e-01 1.46762741e+00 -7.88904369e-01
-1.05115914e+00 1.02618694e+00 -3.26855093e-01 -6.57737851e-01
2.96130866e-01 -5.14750719e-01 -5.22443771e-01 -3.50872040e-01
4.59126472e-01 9.11625981e-01 2.72117108e-01 -1.02976537e+00
-5.31354547e-01 -1.85769126e-01 1.56495497e-02 3.82724822e-01
-3.61227810e-01 2.76056260e-01 -4.47195172e-01 -7.17014849e-01
-2.90569425e-01 -9.19309199e-01 -1.09831139e-01 -2.17989489e-01
-1.01016261e-01 -8.07822466e-01 6.76280320e-01 -9.51107740e-01
1.11075127e+00 -1.72788489e+00 -1.20675571e-01 -1.88155025e-01
-3.11716706e-01 5.58297634e-01 -2.62321025e-01 4.44245696e-01
-5.53848334e-02 2.82545477e-01 -3.42267394e-01 -5.16297162e-01
7.40123689e-02 4.82031077e-01 -1.63813487e-01 3.69447768e-02
4.57011312e-01 1.13509738e+00 -8.11375976e-01 -3.92131299e-01
7.14714974e-02 2.05499455e-01 -6.07665181e-01 -2.16951985e-02
-4.74998146e-01 8.08629841e-02 -4.17492539e-02 5.50597191e-01
2.72611767e-01 4.68776077e-02 1.23305671e-01 -4.65542451e-02
-1.19093604e-01 8.74948144e-01 -6.46523178e-01 1.93097913e+00
-8.04099500e-01 5.97461224e-01 -3.26149553e-01 -5.38820446e-01
9.29511607e-01 3.88409317e-01 8.02080333e-02 -7.59562373e-01
-1.33371264e-01 6.71895921e-01 3.09761882e-01 -3.98094356e-01
5.35852969e-01 -3.20148200e-01 -4.50166315e-01 7.83859491e-01
1.91598952e-01 -3.03271472e-01 4.72285300e-01 2.13317394e-01
8.85572791e-01 2.53831893e-01 3.21448684e-01 -4.75778520e-01
4.34272110e-01 2.09900633e-01 2.65140414e-01 5.52897930e-01
7.32206255e-02 1.65788874e-01 2.19802275e-01 -3.42676729e-01
-1.19676507e+00 -9.59971964e-01 -2.68381804e-01 1.51484311e+00
-5.27558744e-01 -5.26625931e-01 -7.70175517e-01 -8.73623788e-01
1.39115721e-01 1.25259817e+00 -3.78318250e-01 -5.41628757e-03
-7.56957173e-01 -7.71485806e-01 1.05259311e+00 6.15237415e-01
5.63632071e-01 -1.18440998e+00 9.00114030e-02 2.73801774e-01
-3.68596971e-01 -1.38936245e+00 -5.27195394e-01 2.65374511e-01
-7.81945586e-01 -7.19635665e-01 -5.51086605e-01 -1.23011374e+00
5.02638280e-01 -1.04700118e-01 1.61536264e+00 -9.58001837e-02
1.58277497e-01 2.26912230e-01 -3.71081531e-01 -3.98882687e-01
-8.79778445e-01 4.20882195e-01 2.06556901e-01 -4.96186823e-01
8.55525732e-01 -8.52139518e-02 1.23971306e-01 -1.30954003e-02
-6.81249142e-01 -3.02787215e-01 5.12190580e-01 8.73684347e-01
5.44162333e-01 -2.45037973e-01 8.52580249e-01 -1.22557461e+00
8.45711529e-01 -4.39632952e-01 -2.40176558e-01 4.09217328e-01
-7.51864851e-01 3.03295553e-01 8.49951923e-01 -2.13063672e-01
-9.96330500e-01 -1.99678317e-01 -4.60995704e-01 9.73799452e-02
-2.53400296e-01 9.07715082e-01 -1.72096074e-01 2.87273824e-01
1.01019585e+00 -4.58240621e-02 7.87669211e-04 -6.76668167e-01
6.20475352e-01 5.51603556e-01 5.70465684e-01 -8.87119830e-01
7.69412279e-01 -1.04388632e-01 -6.39571905e-01 -8.35138559e-01
-9.61021662e-01 -2.42260665e-01 -7.73356736e-01 3.56541574e-01
6.86958849e-01 -1.11017656e+00 -1.39533490e-01 3.33033860e-01
-1.16229427e+00 -5.85814774e-01 -2.94142157e-01 4.13184077e-01
-6.85104653e-02 2.14721948e-01 -8.86305273e-01 -3.17386508e-01
-4.27326858e-01 -1.09695721e+00 9.31883037e-01 -2.57639557e-01
-4.65400904e-01 -1.34828436e+00 1.31141469e-01 4.65637207e-01
2.15357125e-01 -5.56052625e-02 1.30296600e+00 -8.62339020e-01
1.53712593e-02 -3.41779292e-01 -8.78288299e-02 7.85357654e-01
4.20303911e-01 -1.17495038e-01 -7.25683272e-01 -4.50146377e-01
-1.90799564e-01 -8.97098064e-01 5.74579239e-01 -7.84797892e-02
6.43720806e-01 -2.55389035e-01 3.72643471e-02 1.33128434e-01
1.30975163e+00 2.21376438e-02 4.14801687e-01 4.38074797e-01
6.06093347e-01 5.58093309e-01 5.73453307e-01 -3.54408294e-01
7.58299470e-01 5.44413984e-01 -1.20327763e-01 -3.96476984e-01
-2.46055037e-01 -4.27394658e-01 7.09992886e-01 1.20690429e+00
4.34749573e-01 -2.02095881e-01 -1.26054037e+00 7.31849134e-01
-1.41147304e+00 -4.29679215e-01 -1.91688508e-01 2.14708018e+00
1.28433597e+00 3.17202508e-01 1.06418036e-01 -9.11150053e-02
4.28647220e-01 -1.55928046e-01 -4.24761176e-01 -7.24631011e-01
-4.58782434e-01 6.18710101e-01 3.62019628e-01 8.27631831e-01
-9.91766095e-01 1.49070477e+00 7.15418959e+00 6.85476363e-01
-9.46259797e-01 2.45212868e-01 6.73769772e-01 2.67610729e-01
-4.67727005e-01 -1.57313153e-01 -9.79044318e-01 1.96987346e-01
1.17057514e+00 -2.83443391e-01 3.17334861e-01 6.28277063e-01
-2.36863717e-02 -5.43462560e-02 -1.29642177e+00 6.49446666e-01
2.50657439e-01 -1.15117431e+00 2.35489234e-01 -1.26445830e-01
9.43273365e-01 5.46176374e-01 -7.84746334e-02 8.41531336e-01
5.94734728e-01 -1.05748653e+00 3.33469719e-01 -2.80928969e-01
9.21924949e-01 -1.01780796e+00 7.86223710e-01 4.97406334e-01
-8.89606655e-01 3.75357360e-01 -3.57924819e-01 -1.62535995e-01
1.71045691e-01 3.44288610e-02 -8.92690837e-01 4.06807542e-01
5.96824586e-01 4.49615538e-01 -7.54505813e-01 4.28400099e-01
-4.93742555e-01 9.91438985e-01 -3.73272359e-01 1.13462023e-01
6.39055490e-01 -9.46724489e-02 1.82243153e-01 1.63434982e+00
2.43641958e-02 -6.72561675e-02 3.82487714e-01 3.71388495e-01
-4.47169334e-01 5.69462478e-01 -8.61276090e-01 -1.35197744e-01
4.01829332e-01 1.01700854e+00 -8.54403898e-02 -6.64920390e-01
-7.43336260e-01 8.37681592e-01 6.19436264e-01 2.68412769e-01
-7.07578719e-01 -4.40536231e-01 4.54712629e-01 3.06227449e-02
-3.59539129e-02 -3.52751106e-01 -2.87019610e-01 -1.25671625e+00
1.38694078e-01 -1.35926187e+00 5.11291027e-01 -7.82108963e-01
-1.39078212e+00 1.00673878e+00 -1.56333327e-01 -9.98667240e-01
-6.47876322e-01 -8.19023490e-01 -2.85406828e-01 1.14754641e+00
-1.48431885e+00 -1.30164313e+00 2.70012408e-01 5.07721305e-01
7.44459510e-01 -3.69746029e-01 1.21983361e+00 6.12115026e-01
-3.65245730e-01 1.10480642e+00 6.99537322e-02 4.74837244e-01
1.05219936e+00 -1.35037792e+00 7.89997518e-01 9.65677977e-01
4.80284005e-01 7.19831288e-01 4.89465803e-01 -4.47285354e-01
-1.30993581e+00 -1.16007686e+00 1.43087494e+00 -7.15046585e-01
1.01623869e+00 -2.62768537e-01 -1.21279550e+00 1.23734593e+00
7.88932323e-01 -2.85449386e-01 7.66114950e-01 5.12719214e-01
-6.22421503e-01 -5.99570833e-02 -1.04399765e+00 6.50296032e-01
4.87377167e-01 -8.76760960e-01 -9.46832836e-01 5.57996035e-01
9.10360813e-01 -3.72179389e-01 -1.16263986e+00 3.55398118e-01
1.77017853e-01 -3.45723450e-01 6.50799811e-01 -1.03266931e+00
4.90400106e-01 -2.08851323e-01 -2.43017629e-01 -1.54723394e+00
-5.84231839e-02 -4.96387690e-01 2.61054218e-01 1.38661218e+00
1.07072818e+00 -7.80062079e-01 4.42124546e-01 5.23229539e-01
-3.52202594e-01 -6.41681433e-01 -8.12014163e-01 -8.04432034e-01
9.20040369e-01 -6.58530295e-01 3.10213417e-01 1.28834796e+00
1.09459497e-01 9.81102884e-01 -1.97065979e-01 -2.66114213e-02
4.23398674e-01 -1.99616417e-01 8.09714437e-01 -6.35923207e-01
-3.04986715e-01 -2.73829788e-01 9.15412605e-02 -1.10655105e+00
4.49677169e-01 -1.21849990e+00 -1.42827500e-02 -1.50787067e+00
-1.87549651e-01 -6.84233069e-01 -1.34960756e-01 8.81155849e-01
-4.04566586e-01 5.24254382e-01 3.94005626e-02 -1.01964176e-01
-1.99797347e-01 3.50025147e-01 1.04680467e+00 -1.14272371e-01
-1.27231106e-01 -3.19824994e-01 -8.76368880e-01 7.67925084e-01
8.60192657e-01 -2.50933558e-01 -2.34682247e-01 -1.21790648e+00
7.88929239e-02 -2.64148653e-01 -2.48121127e-01 -6.28708124e-01
-7.64092579e-02 1.96837917e-01 2.09515676e-01 -4.31857288e-01
2.48390719e-01 -4.04792607e-01 -4.57535326e-01 4.30060208e-01
-4.62883890e-01 6.46122873e-01 8.55879307e-01 -1.24672085e-01
-3.54257762e-01 -1.60172790e-01 7.05416143e-01 -2.21004590e-01
-6.01692736e-01 -4.18908596e-02 -4.74810779e-01 5.59546232e-01
4.50380057e-01 1.49354279e-01 -3.99133950e-01 -2.77249038e-01
-4.00828958e-01 4.10624683e-01 2.60212302e-01 6.78500295e-01
1.19235432e-02 -1.37531531e+00 -1.03566563e+00 1.32487819e-01
2.78686970e-01 -8.64039361e-02 -4.88299668e-01 5.06040573e-01
-6.43961489e-01 6.59365594e-01 9.03225839e-02 -4.34419543e-01
-8.88408720e-01 3.47730339e-01 4.24847871e-01 -4.85173255e-01
-1.22519970e-01 8.73612046e-01 -8.97260681e-02 -1.20997119e+00
9.21961963e-02 -4.62026000e-01 1.69501305e-01 -2.51946282e-02
3.64353061e-01 1.59012049e-01 3.76490563e-01 -6.70879126e-01
-2.04301044e-01 2.71601379e-01 -4.88629848e-01 -3.09197724e-01
1.10440445e+00 1.05886295e-01 -3.38024020e-01 7.01571047e-01
1.26911974e+00 3.45525742e-01 -7.03834116e-01 -6.46888614e-01
2.64002413e-01 5.78141697e-02 -1.30247533e-01 -1.19310343e+00
-6.17438734e-01 7.06167042e-01 1.40291050e-01 -1.12315491e-01
8.40697825e-01 3.30082304e-03 9.07401562e-01 7.91552722e-01
4.05575424e-01 -9.34347749e-01 -1.30805746e-01 9.23158050e-01
7.91575193e-01 -1.50196576e+00 -2.51700860e-02 -7.46318474e-02
-6.61752641e-01 7.79774368e-01 8.64646256e-01 -2.82365214e-02
4.62065101e-01 2.90612310e-01 4.52105880e-01 1.79139733e-01
-8.39296997e-01 -1.21548489e-01 2.69887954e-01 2.57669598e-01
1.12052965e+00 1.05721392e-01 -3.64475638e-01 2.09973812e-01
-5.18346667e-01 -2.99911708e-01 1.91072211e-01 8.40591252e-01
-2.25034282e-01 -1.60023546e+00 -2.69594014e-01 4.36745852e-01
-5.70548832e-01 -6.36989832e-01 -4.10540521e-01 1.28150046e+00
1.84249982e-01 9.62692618e-01 1.68866366e-01 -9.47503746e-02
4.98598754e-01 5.58452189e-01 5.01560807e-01 -8.87992382e-01
-8.74180734e-01 1.26756519e-01 4.90601748e-01 -1.85649037e-01
-2.10264683e-01 -7.46419966e-01 -1.29020703e+00 -3.39944631e-01
-2.03784049e-01 4.22143608e-01 6.03677511e-01 1.19168115e+00
9.34711322e-02 2.48502791e-01 2.81920820e-01 -1.75727159e-01
-5.64485073e-01 -1.28464746e+00 -7.58501664e-02 2.50883639e-01
1.68681368e-01 -2.02926576e-01 -2.45777324e-01 9.13613439e-02] | [11.032797813415527, 9.805609703063965] |
9af58851-6507-4f08-84cf-22de706492ce | coverhunter-cover-song-identification-with | 2306.09025 | null | https://arxiv.org/abs/2306.09025v1 | https://arxiv.org/pdf/2306.09025v1.pdf | CoverHunter: Cover Song Identification with Refined Attention and Alignments | Abstract: Cover song identification (CSI) focuses on finding the same music with different versions in reference anchors given a query track. In this paper, we propose a novel system named CoverHunter that overcomes the shortcomings of existing detection schemes by exploring richer features with refined attention and alignments. CoverHunter contains three key modules: 1) A convolution-augmented transformer (i.e., Conformer) structure that captures both local and global feature interactions in contrast to previous methods mainly relying on convolutional neural networks; 2) An attention-based time pooling module that further exploits the attention in the time dimension; 3) A novel coarse-to-fine training scheme that first trains a network to roughly align the song chunks and then refines the network by training on the aligned chunks. At the same time, we also summarize some important training tricks used in our system that help achieve better results. Experiments on several standard CSI datasets show that our method significantly improves over state-of-the-art methods with an embedding size of 128 (2.3% on SHS100K-TEST and 17.7% on DaTacos). | ['Xintong Han', 'Yinan Xu', 'Deyi Tuo', 'Feng Liu'] | 2023-06-15 | null | null | null | null | ['cover-song-identification'] | ['music'] | [ 1.76197827e-01 -4.15439904e-01 -3.29875231e-01 -1.19006649e-01
-1.10916531e+00 -7.15604067e-01 2.60275126e-01 -5.17124534e-02
-2.55424768e-01 3.63771290e-01 4.52728689e-01 5.39857261e-02
-2.47676954e-01 -6.55910611e-01 -8.07967186e-01 -6.71970725e-01
-4.12396073e-01 1.91266432e-01 2.93483317e-01 -3.40041965e-01
2.52273351e-01 -5.10948263e-02 -1.59255469e+00 4.79425520e-01
4.37148184e-01 1.39037240e+00 -4.58924137e-02 5.16075373e-01
1.12723053e-01 5.98145843e-01 -6.35895967e-01 -2.45241642e-01
4.73100007e-01 -3.89471650e-01 -7.51031637e-01 -2.01462001e-01
4.36108857e-01 -2.32727751e-01 -6.11567020e-01 8.21229994e-01
6.94839954e-01 4.09299172e-02 1.17085896e-01 -9.87690806e-01
-7.39815116e-01 1.15812039e+00 -7.49099255e-01 5.51022470e-01
3.00139338e-01 -1.10627957e-01 1.39048600e+00 -8.44074547e-01
2.32173398e-01 9.90486681e-01 9.25775647e-01 2.11320341e-01
-1.20642436e+00 -1.01372325e+00 1.94842070e-01 6.13664865e-01
-1.68253362e+00 -2.89402962e-01 8.20410490e-01 -2.11570337e-01
1.04445493e+00 4.15960878e-01 7.32147753e-01 1.10464883e+00
-6.47455379e-02 9.27933991e-01 7.18722820e-01 -3.79432946e-01
-9.59447771e-02 -1.20134056e-01 1.65661752e-01 2.18793452e-01
6.28634840e-02 2.49416068e-01 -9.33777392e-01 -2.30801895e-01
6.42635167e-01 -8.48575160e-02 -4.92895424e-01 -1.14484824e-01
-1.62523973e+00 7.70269513e-01 6.49930716e-01 4.81232285e-01
-2.89329112e-01 4.08530444e-01 5.19337416e-01 4.21774864e-01
3.43339652e-01 5.57387829e-01 -4.63052303e-01 -1.93761110e-01
-1.16917539e+00 2.36980230e-01 6.88253105e-01 9.06556129e-01
5.35168707e-01 -4.47419137e-02 -5.70394874e-01 6.87719166e-01
3.00011151e-02 1.30670190e-01 6.16466522e-01 -4.36587989e-01
6.37505531e-01 4.30289745e-01 -1.08483499e-02 -7.38986611e-01
-1.05642557e-01 -1.40516913e+00 -7.40434945e-01 -2.28360221e-01
-3.08277346e-02 1.11013673e-01 -7.14869499e-01 1.77824807e+00
1.85154244e-01 7.18056619e-01 -1.00597903e-01 9.53788280e-01
6.96371198e-01 5.18838167e-01 -3.26684386e-01 1.59889728e-01
1.52412915e+00 -1.36276650e+00 -7.24304795e-01 -3.39616984e-02
2.66486704e-01 -7.83263505e-01 9.18826938e-01 2.22809955e-01
-9.14610624e-01 -8.58987451e-01 -1.44552231e+00 -1.09188966e-02
-4.51905817e-01 3.04875672e-01 5.12151301e-01 3.90460551e-01
-9.83064353e-01 8.23168099e-01 -5.05838573e-01 -8.38061944e-02
3.19915026e-01 5.66672564e-01 -9.36384574e-02 3.28607261e-01
-1.46289718e+00 2.88034856e-01 5.16435742e-01 3.81374955e-02
-8.48273039e-01 -7.48796463e-01 -6.96089149e-01 5.30569375e-01
5.22418678e-01 -4.04340684e-01 1.24830163e+00 -9.13485527e-01
-1.41569865e+00 6.28633320e-01 -3.20179909e-02 -7.50566721e-01
1.79515854e-01 -5.71054220e-01 -5.17032444e-01 -1.95829086e-02
1.31847456e-01 5.04668593e-01 8.04500520e-01 -6.75152957e-01
-7.60671258e-01 -1.29463211e-01 1.17433541e-01 2.17549399e-01
-4.86504078e-01 1.09355683e-02 -8.35725963e-01 -1.19289792e+00
2.74220929e-02 -8.35172117e-01 4.99352105e-02 -3.40067446e-01
-4.92495954e-01 -1.72502786e-01 7.08721459e-01 -6.35666072e-01
1.76664412e+00 -2.35560822e+00 2.38152757e-01 1.78011566e-01
2.01246981e-02 3.14880997e-01 -3.69731873e-01 7.27358162e-01
-1.74413934e-01 -7.06625655e-02 -1.08368605e-01 -3.75025660e-01
1.31801218e-01 -3.19398731e-01 -7.05718756e-01 4.18287098e-01
7.29811043e-02 9.76203024e-01 -8.36934388e-01 -2.04800650e-01
-1.53209895e-01 4.59789753e-01 -5.90541959e-01 9.64497253e-02
-8.28792527e-02 4.06372130e-01 -1.86190262e-01 8.12680542e-01
4.71964240e-01 -2.54117042e-01 -4.13574800e-02 -3.52549821e-01
-2.18312114e-01 6.99943721e-01 -1.29858005e+00 2.01986384e+00
-7.56072327e-02 7.02722251e-01 -2.20735416e-01 -9.26065207e-01
8.76749694e-01 3.56594771e-01 5.62680602e-01 -5.08721411e-01
1.10013939e-01 2.84044564e-01 -6.20187409e-02 -1.97456881e-01
7.16150939e-01 3.57145458e-01 -2.18740344e-01 4.45984215e-01
2.16079950e-01 6.62835598e-01 9.98445507e-03 -5.23870066e-02
1.14415860e+00 1.28871247e-01 1.42414078e-01 -1.84518635e-01
5.16221941e-01 -3.21378499e-01 7.56516993e-01 8.46406817e-01
-4.26303968e-02 4.87353742e-01 2.13698819e-01 -4.92406160e-01
-8.62433553e-01 -6.88453138e-01 -5.52476197e-02 1.22314894e+00
4.21970487e-01 -9.26291406e-01 -4.85168755e-01 -4.79659766e-01
1.30521521e-01 3.48241031e-01 -9.37386036e-01 -1.67371154e-01
-4.27390814e-01 -5.51471889e-01 9.15313005e-01 8.29295695e-01
7.39161789e-01 -1.23091722e+00 -5.39764524e-01 5.12048006e-01
-3.28526914e-01 -7.96694517e-01 -1.03238559e+00 3.80758494e-01
-6.81975782e-01 -1.01228797e+00 -5.96717536e-01 -8.19639683e-01
5.58315665e-02 4.58348751e-01 1.07277322e+00 1.56763420e-02
-1.98908567e-01 1.64526943e-02 -5.98222554e-01 -5.51305950e-01
2.98829138e-01 6.43412650e-01 -3.94834206e-02 2.60980755e-01
5.82431197e-01 -7.31384397e-01 -6.48489416e-01 3.61282945e-01
-7.76627421e-01 8.60683694e-02 6.36067629e-01 1.09338748e+00
7.76802242e-01 -4.95745204e-02 6.80674076e-01 -5.49563289e-01
4.23147112e-01 -4.73675847e-01 -5.09684980e-01 2.86872447e-01
-5.68136454e-01 1.05807908e-01 2.44411036e-01 -7.13147640e-01
-4.32076871e-01 7.11340532e-02 5.88322058e-03 -5.62211394e-01
1.86852157e-01 4.46142912e-01 -1.86944976e-02 -4.66386043e-02
4.79250610e-01 5.42558908e-01 -4.70170110e-01 -8.99732471e-01
1.71438724e-01 8.09299469e-01 6.94345653e-01 -4.22817916e-01
1.13043189e+00 5.25289655e-01 -4.68821526e-01 -2.99125701e-01
-1.13884377e+00 -7.80980170e-01 -5.20328641e-01 2.54913997e-02
5.16280293e-01 -1.22967088e+00 -7.52172887e-01 3.78119022e-01
-8.17546308e-01 -1.35918632e-01 -5.23110271e-01 6.59753382e-01
-4.90231514e-01 1.07314683e-01 -6.77618921e-01 -5.63313723e-01
-7.33451605e-01 -8.48521531e-01 1.31844294e+00 2.63695300e-01
-8.98190215e-02 -3.22678685e-01 3.81583273e-01 4.30091955e-02
6.14458442e-01 1.81618944e-01 4.91804838e-01 -6.79024100e-01
-7.19942570e-01 -1.51720732e-01 -1.39471367e-01 7.18626976e-02
5.66896833e-02 -4.29992199e-01 -1.29254544e+00 -6.12841487e-01
-2.48984113e-01 -1.70500740e-01 1.26827776e+00 2.52751976e-01
1.46109200e+00 -3.21870148e-01 -3.60586315e-01 1.02770305e+00
1.45574713e+00 1.89051941e-01 6.33026659e-01 6.53155804e-01
4.95649874e-01 1.28478900e-01 7.36240149e-01 3.83218914e-01
2.04638824e-01 1.03994524e+00 5.92893124e-01 -1.16594493e-01
-5.81158064e-02 -4.33597744e-01 4.64111358e-01 8.49584579e-01
-3.22749168e-02 -1.80970639e-01 -2.39090666e-01 8.85758042e-01
-2.02142549e+00 -1.16791785e+00 2.39044830e-01 2.16396213e+00
8.74951780e-01 2.21975029e-01 2.23368675e-01 5.46004355e-01
7.91843712e-01 4.64112043e-01 -7.04621792e-01 -1.58424228e-01
-1.20911695e-01 3.75410557e-01 5.74587166e-01 1.02093250e-01
-1.47830975e+00 8.87200773e-01 5.82041788e+00 1.09285772e+00
-1.02527869e+00 1.44187525e-01 8.09767544e-02 -3.59567404e-01
-7.84995407e-02 -4.97020073e-02 -8.92269492e-01 4.76279914e-01
6.94703877e-01 -1.29355818e-01 7.30159879e-01 7.30054080e-01
-3.13766181e-01 5.47983170e-01 -1.12998366e+00 1.10247421e+00
1.57854989e-01 -1.61141014e+00 -1.38810709e-01 7.94451833e-02
6.43569171e-01 2.41975278e-01 1.21020503e-01 5.20910859e-01
-2.23667204e-01 -8.30167055e-01 1.12201369e+00 3.98096293e-01
9.88001108e-01 -7.51837790e-01 9.18905020e-01 -5.15493713e-02
-1.86535645e+00 -1.27266213e-01 -2.95145184e-01 4.67977487e-03
-8.20692480e-02 3.03078502e-01 -6.56385303e-01 8.29479158e-01
9.81603503e-01 9.12862420e-01 -5.14903605e-01 1.55625081e+00
-1.71491757e-01 7.37092137e-01 -2.58316040e-01 1.39244601e-01
3.15676749e-01 3.04441780e-01 7.55676031e-01 1.29814792e+00
4.80614185e-01 -2.54378915e-01 4.32972275e-02 7.58988202e-01
-2.60957509e-01 -4.60919403e-02 -2.40945786e-01 -5.10671437e-02
7.10490465e-01 1.15833461e+00 -4.45631534e-01 -2.76401043e-01
-3.67663026e-01 1.12482977e+00 1.87648475e-01 2.57329047e-01
-9.66177106e-01 -8.38310421e-01 7.73469746e-01 -2.98847968e-04
1.01489186e+00 9.20813233e-02 1.21660203e-01 -1.20246065e+00
5.63846566e-02 -9.22839224e-01 6.17186129e-01 -4.58966315e-01
-1.08599555e+00 8.33975732e-01 -3.93798232e-01 -1.74997854e+00
-1.01883858e-01 -2.29276672e-01 -7.49503195e-01 8.82697165e-01
-1.65395105e+00 -1.25081432e+00 -3.20347995e-01 6.78243399e-01
5.53396940e-01 -2.08138660e-01 1.02866983e+00 5.76706231e-01
-4.22786951e-01 1.15596890e+00 2.48755455e-01 3.48021537e-01
9.37891424e-01 -1.25688195e+00 7.36438811e-01 6.09660149e-01
7.89987445e-01 5.28949380e-01 3.34239781e-01 -2.81508386e-01
-1.24778032e+00 -1.21851194e+00 9.32447135e-01 -1.34845227e-01
5.67597747e-01 -4.19511735e-01 -7.90587723e-01 6.59788966e-01
2.59110093e-01 -9.74197686e-02 9.84372258e-01 3.82215887e-01
-7.24804044e-01 -4.66435969e-01 -6.26113892e-01 3.15567642e-01
1.11109066e+00 -7.86602616e-01 -6.10337019e-01 5.82854375e-02
1.00482082e+00 -6.77181244e-01 -8.07612538e-01 3.75915974e-01
9.57887709e-01 -7.55652130e-01 9.88637924e-01 -6.28675401e-01
-6.51291385e-02 -6.63229346e-01 -3.56846869e-01 -9.39644992e-01
-6.86574399e-01 -8.80400419e-01 -5.39390802e-01 1.19967377e+00
2.47358695e-01 -3.36383849e-01 5.40667713e-01 -3.59109908e-01
-2.71376252e-01 -9.78640437e-01 -9.91738796e-01 -9.03846681e-01
-5.24438441e-01 -4.90628034e-01 1.05818701e+00 8.65930140e-01
-2.87495162e-02 2.96134263e-01 -6.68878794e-01 3.73511642e-01
4.45185363e-01 6.77221537e-01 7.46714711e-01 -1.23709118e+00
-6.68644965e-01 -3.68209451e-01 -5.08509994e-01 -1.14531898e+00
-1.49196863e-01 -8.65508795e-01 -1.13948159e-01 -1.06838453e+00
3.62702668e-01 -1.74650773e-01 -1.02105319e+00 7.01532960e-01
-1.51359988e-02 6.08402491e-01 1.19454272e-01 4.51633036e-01
-9.49147701e-01 7.04189897e-01 8.17297816e-01 -4.09634173e-01
-4.02867734e-01 1.15434736e-01 -1.05398345e+00 3.99376929e-01
7.72819042e-01 -5.41402578e-01 -8.10501724e-02 -4.52427596e-01
6.53487295e-02 -2.65661627e-01 5.05757809e-01 -1.32731736e+00
2.61637032e-01 3.47965688e-01 3.49380493e-01 -1.09687459e+00
3.67231518e-01 -7.04854786e-01 2.74250239e-01 5.39008439e-01
-5.51363707e-01 -4.63292040e-02 3.95462811e-01 8.20601165e-01
-4.40227598e-01 5.35333343e-02 4.47514743e-01 3.04470867e-01
-5.86596847e-01 3.96485031e-01 1.07037902e-01 -2.65899032e-01
7.56607711e-01 -1.42128289e-01 -2.34748289e-01 -4.72741574e-01
-5.55479527e-01 3.72924745e-01 8.13803151e-02 8.01548898e-01
3.03596735e-01 -1.87257659e+00 -7.63150036e-01 3.19223821e-01
4.27930862e-01 -4.19304013e-01 1.98998332e-01 7.52649188e-01
-1.35171548e-01 6.90329254e-01 -6.73869550e-02 -5.14057457e-01
-1.31819701e+00 6.13524735e-01 2.92148650e-01 -5.37295163e-01
-7.37980306e-01 1.06222677e+00 1.72643080e-01 3.76905091e-02
7.40931809e-01 -4.20825213e-01 -2.41820738e-01 8.75526890e-02
8.23712409e-01 1.35443836e-01 -2.42804065e-02 -6.27634525e-01
-5.13588846e-01 7.19057024e-01 -2.30085403e-01 -1.79881677e-02
1.40489280e+00 -7.26339296e-02 1.63060918e-01 4.47930127e-01
1.07008219e+00 1.97964057e-01 -1.27222228e+00 -7.31185019e-01
5.40996622e-03 -5.12494028e-01 2.31103525e-02 -7.78719425e-01
-1.15068877e+00 7.66718268e-01 8.07856321e-01 3.47457021e-01
1.47020257e+00 2.47695576e-03 9.96257186e-01 1.99522585e-01
2.81489402e-01 -9.37571287e-01 8.38469341e-03 3.95573974e-01
1.07298136e+00 -8.83111715e-01 -1.99454099e-01 -1.14850529e-01
-1.57879010e-01 1.01212990e+00 4.27880615e-01 -2.06974864e-01
5.73916435e-01 3.95533070e-02 -2.63715625e-01 -2.51336366e-01
-7.49652386e-01 -3.98374438e-01 6.36539698e-01 2.43297338e-01
3.27098459e-01 -6.39995886e-03 -1.64950460e-01 1.23737180e+00
-3.04386377e-01 -4.70028743e-02 -2.44715791e-02 6.21171534e-01
-3.50541413e-01 -1.15815306e+00 -3.99325073e-01 3.77554834e-01
-6.95966601e-01 -2.18540668e-01 -3.67861181e-01 6.52025998e-01
4.64181930e-01 8.27183187e-01 5.98111115e-02 -1.02200150e+00
2.82370985e-01 -1.50647148e-01 2.13549867e-01 -4.92302865e-01
-9.75832164e-01 4.06676799e-01 -1.57518432e-01 -8.03079545e-01
-5.08270562e-01 -3.89964223e-01 -7.19024897e-01 -1.83295846e-01
-7.07114935e-01 4.02306348e-01 3.47644359e-01 7.09679425e-01
6.50487423e-01 9.31871295e-01 7.42127895e-01 -9.23843443e-01
-5.30655205e-01 -1.19117224e+00 -7.69735038e-01 1.02378383e-01
6.16660595e-01 -5.48474014e-01 -1.85436070e-01 -6.78160340e-02] | [15.737329483032227, 5.215534210205078] |
d5c6b729-ed3f-4d5c-96cf-03b723e2913f | earthnet2021-a-large-scale-dataset-and | 2104.10066 | null | https://arxiv.org/abs/2104.10066v1 | https://arxiv.org/pdf/2104.10066v1.pdf | EarthNet2021: A large-scale dataset and challenge for Earth surface forecasting as a guided video prediction task | Satellite images are snapshots of the Earth surface. We propose to forecast them. We frame Earth surface forecasting as the task of predicting satellite imagery conditioned on future weather. EarthNet2021 is a large dataset suitable for training deep neural networks on the task. It contains Sentinel 2 satellite imagery at 20m resolution, matching topography and mesoscale (1.28km) meteorological variables packaged into 32000 samples. Additionally we frame EarthNet2021 as a challenge allowing for model intercomparison. Resulting forecasts will greatly improve (>x50) over the spatial resolution found in numerical models. This allows localized impacts from extreme weather to be predicted, thus supporting downstream applications such as crop yield prediction, forest health assessments or biodiversity monitoring. Find data, code, and how to participate at www.earthnet.tech | ['Joachim Denzler', 'Jakob Runge', 'Markus Reichstein', 'Vitus Benson', 'Christian Requena-Mesa'] | 2021-04-16 | null | null | null | null | ['video-forensics', 'crop-yield-prediction', 'crop-yield-prediction', 'earth-surface-forecasting'] | ['computer-vision', 'computer-vision', 'miscellaneous', 'time-series'] | [-1.83888618e-02 7.89947137e-02 -2.44524524e-01 -4.37793374e-01
-3.06202620e-01 -6.04200721e-01 9.13011968e-01 -9.72338095e-02
-8.50251466e-02 9.48514223e-01 3.91674757e-01 -9.39910531e-01
2.33962968e-01 -1.30210209e+00 -6.62408769e-01 -6.20383263e-01
-9.52112496e-01 7.32130408e-02 -1.46047398e-01 -7.03338027e-01
-8.71110782e-02 7.59773970e-01 -1.77196145e+00 2.54931599e-01
7.85829484e-01 1.15266562e+00 5.45088112e-01 9.50946093e-01
7.56764412e-02 4.26401764e-01 -9.59481820e-02 2.87109643e-01
5.52391589e-01 8.10569599e-02 -7.46820033e-01 -3.54460657e-01
6.39638722e-01 -4.90818977e-01 -2.59825196e-02 1.09441364e+00
3.85442495e-01 -4.67841849e-02 3.28731716e-01 -1.03231716e+00
-2.94780016e-01 2.47509241e-01 -4.50166613e-01 4.34090137e-01
-5.97623177e-02 2.89191395e-01 5.16020954e-01 -7.08799064e-01
6.95772111e-01 1.19105291e+00 1.00222409e+00 5.36905080e-02
-1.07610464e+00 -6.40828311e-01 9.68510658e-02 -7.56191984e-02
-1.27479053e+00 -6.71711147e-01 -6.67275535e-03 -8.36790144e-01
1.19270205e+00 7.06593215e-01 6.03294790e-01 7.18167901e-01
6.59569263e-01 3.01540852e-01 9.77422774e-01 5.42215817e-02
2.30976954e-01 -3.00412178e-01 -2.70582408e-01 1.51587546e-01
5.97729199e-02 6.91396236e-01 -1.98129028e-01 1.42639214e-02
5.87077141e-01 1.01248644e-01 -3.60870481e-01 4.08244938e-01
-1.30628622e+00 5.76433539e-01 1.12299740e+00 1.25325590e-01
-1.11140394e+00 1.54310003e-01 2.48871058e-01 5.15810609e-01
1.19206023e+00 3.25204402e-01 -8.52662921e-01 2.24790379e-01
-1.33428586e+00 5.55465519e-01 4.74312305e-01 6.47988617e-01
7.55149066e-01 4.22361404e-01 4.99129683e-01 4.18633342e-01
2.79737473e-01 1.13735271e+00 2.30911672e-01 -1.17679560e+00
2.03667149e-01 3.12814534e-01 7.06758618e-01 -1.28286278e+00
-7.40502238e-01 -3.80665749e-01 -1.44922698e+00 4.49783474e-01
-1.56900346e-01 -6.38027728e-01 -8.89547884e-01 1.37692404e+00
3.59855801e-01 3.83861870e-01 9.77124944e-02 9.53996360e-01
5.58925986e-01 1.35501385e+00 2.96954155e-01 1.40099805e-02
1.05713880e+00 -3.97629946e-01 -5.82092881e-01 -3.89085203e-01
9.57260013e-01 -4.37379152e-01 4.04238135e-01 1.39385134e-01
-5.06501973e-01 -6.43142819e-01 -7.16859519e-01 6.20446503e-01
-1.08215308e+00 1.67999431e-01 7.79928446e-01 3.83367315e-02
-1.75701225e+00 9.42489803e-01 -9.20330107e-01 -6.27313614e-01
1.57942131e-01 -1.42965123e-01 -2.84955829e-01 3.48378927e-01
-1.71421659e+00 1.15458345e+00 6.68443143e-01 5.98088264e-01
-8.37486804e-01 -9.46405888e-01 -9.75839913e-01 2.31317207e-01
-5.48716843e-01 -3.12608421e-01 9.59231496e-01 -9.05829489e-01
-1.07011580e+00 9.25471604e-01 -5.46939783e-02 -8.91452432e-01
2.63675302e-01 -1.70973927e-01 -9.67216790e-01 -3.22353721e-01
4.12136137e-01 1.08834052e+00 2.36046985e-01 -9.35551941e-01
-9.33435559e-01 -5.59462786e-01 -3.34893256e-01 1.63163051e-01
-5.55537865e-02 2.32181191e-01 3.17902505e-01 -5.41836858e-01
4.81573865e-02 -1.00457239e+00 -7.57126153e-01 1.26946107e-01
-1.91330671e-01 3.29108864e-01 8.04341495e-01 -1.23616874e+00
6.94853246e-01 -1.98901522e+00 -3.82411420e-01 -2.36464426e-01
-2.13002771e-01 3.98222238e-01 -6.01358652e-01 5.72597384e-01
-3.75674188e-01 5.66613197e-01 -2.67259151e-01 4.83496487e-02
-1.89728275e-01 6.77980632e-02 -1.05109799e+00 5.20043850e-01
3.18374813e-01 7.52163827e-01 -7.08113194e-01 1.97364360e-01
4.48801249e-01 4.21434641e-01 6.97175488e-02 2.02677816e-01
-5.85009873e-01 6.60476565e-01 -2.17283592e-01 3.88114959e-01
1.42701864e+00 -3.09414603e-02 2.49990970e-01 9.78437886e-02
-8.44905555e-01 2.90645450e-01 -1.05247247e+00 1.18401659e+00
-6.62974834e-01 8.54783714e-01 5.25079906e-01 -7.01204419e-01
1.10712624e+00 2.53108084e-01 2.92809784e-01 -6.95795417e-01
-3.47527862e-01 1.93509638e-01 -2.22754091e-01 -3.75007331e-01
6.96340382e-01 -1.95396125e-01 2.11030945e-01 1.19943336e-01
-2.93590188e-01 -3.50746624e-02 -6.78701773e-02 -1.48781881e-01
3.99603426e-01 2.62655854e-01 2.14830413e-01 -8.45904946e-01
1.10360213e-01 4.01196539e-01 5.25950968e-01 4.57681209e-01
-8.75950232e-02 1.59983158e-01 3.12958002e-01 -1.34248173e+00
-1.06765938e+00 -6.93538070e-01 -6.27775133e-01 1.03369069e+00
-3.66898298e-01 1.76395383e-02 -1.03748262e-01 -2.15492371e-04
2.92446017e-01 7.63283312e-01 -7.29543328e-01 4.15682495e-01
-1.62225701e-02 -1.26755750e+00 6.94771111e-01 4.40873265e-01
7.80874252e-01 -9.94882107e-01 -6.40096903e-01 2.10046828e-01
-2.76933014e-01 -8.46525967e-01 6.00833297e-01 1.60874084e-01
-1.12485695e+00 -7.49228835e-01 -7.32206643e-01 -1.47363096e-01
-3.56985964e-02 1.23462632e-01 1.39759910e+00 -1.54236853e-01
-5.73186576e-02 -2.15406656e-01 -6.17197528e-02 -3.98367763e-01
-2.58750319e-01 2.13357896e-01 4.89967823e-01 -3.37150753e-01
1.57675743e-01 -8.43294680e-01 -7.13383019e-01 1.48746356e-01
-8.02612662e-01 2.62176126e-01 2.56355703e-01 3.24871600e-01
3.37536246e-01 -6.50021210e-02 4.87307578e-01 -2.00934291e-01
1.14770710e-01 -1.06911564e+00 -1.15033674e+00 1.48680821e-01
-3.95814121e-01 -4.18604046e-01 4.53637332e-01 3.77191544e-01
-1.18844926e+00 3.62380922e-01 -1.91006929e-01 1.37106121e-01
-9.47849274e-01 1.23228824e+00 1.13931775e-01 6.61458224e-02
7.67455101e-01 4.35130030e-01 -2.38795951e-01 -6.78690076e-01
3.60602319e-01 8.93266082e-01 7.34352529e-01 4.02827784e-02
4.12392825e-01 5.44373870e-01 -3.00987456e-02 -1.21698785e+00
-7.07474113e-01 -3.48445624e-01 -7.06020653e-01 -4.59695697e-01
7.05997646e-01 -1.58430636e+00 -4.16372001e-01 6.66622698e-01
-1.22884679e+00 -7.80670941e-01 2.41901040e-01 5.14709949e-01
-1.88242018e-01 -8.33165944e-02 -2.78076828e-01 -8.45310688e-01
-6.13730073e-01 -2.91417629e-01 9.24462795e-01 1.76443830e-01
-4.96478332e-03 -1.09625959e+00 6.28317893e-01 -4.45425123e-01
1.18624854e+00 8.14467728e-01 1.32391170e-01 -5.77881746e-02
-4.03783888e-01 -2.28102326e-01 -4.41481411e-01 8.90954137e-02
2.76085556e-01 3.12147737e-01 -1.02961087e+00 -3.74346226e-01
-2.63820797e-01 -3.02293003e-01 1.35509932e+00 7.28614032e-01
1.06926501e+00 -4.99544710e-01 -4.56158638e-01 1.14362252e+00
1.54836285e+00 -1.76442876e-01 7.45619118e-01 6.97219312e-01
1.37535587e-01 8.07621479e-01 5.68852127e-01 6.37971818e-01
2.97866017e-01 2.33444214e-01 1.13451648e+00 -3.90654474e-01
2.59942383e-01 1.59508869e-01 2.39069447e-01 1.63337573e-01
-1.47379518e-01 -1.23450428e-01 -1.84499800e+00 8.64893079e-01
-1.65508008e+00 -1.29824388e+00 -6.49038136e-01 2.04659677e+00
5.60958683e-01 -2.67628014e-01 -3.32397282e-01 -4.96030271e-01
6.30322278e-01 8.00000131e-01 -6.86128974e-01 -2.31020316e-01
-3.83116394e-01 -7.31817633e-02 1.17816865e+00 7.38508761e-01
-1.58336914e+00 1.21683884e+00 6.52114010e+00 1.80305108e-01
-1.74622309e+00 -9.85018611e-02 8.67711365e-01 1.47612467e-01
-4.05540794e-01 3.10113337e-02 -1.01512754e+00 4.55956876e-01
1.67195380e+00 -1.54667065e-01 2.21689895e-01 8.11893106e-01
8.27416897e-01 -3.00746650e-01 -1.56093180e-01 2.44359463e-01
-7.86406517e-01 -1.94153416e+00 5.46114594e-02 1.92054823e-01
9.37540412e-01 1.24529111e+00 2.98046097e-02 2.81282037e-01
9.03190196e-01 -1.14961946e+00 1.53501526e-01 9.11358714e-01
1.11548567e+00 -5.51978230e-01 7.56190598e-01 6.95343137e-01
-1.25203443e+00 1.80392131e-01 -6.84502661e-01 -6.54194355e-01
1.48207799e-01 1.01458013e+00 -5.96281528e-01 6.22215867e-01
1.17390120e+00 1.23821259e+00 -3.86134952e-01 9.67923582e-01
4.92469817e-02 7.67852306e-01 -6.36438310e-01 5.79682529e-01
4.33921099e-01 -1.69311002e-01 4.48374689e-01 1.04656148e+00
7.99394667e-01 2.85086632e-01 1.00521185e-01 6.43902004e-01
2.12456077e-01 -1.04199894e-01 -1.13151050e+00 1.65616721e-01
5.00011027e-01 1.21276295e+00 -3.03683609e-01 -4.63029325e-01
5.35849109e-02 7.42076516e-01 -1.13212734e-01 3.87651086e-01
-3.01901430e-01 -1.75211310e-01 1.35518420e+00 1.15009844e-01
2.28960559e-01 -4.03500199e-01 -1.88903809e-01 -1.28750849e+00
-4.33231086e-01 -6.01826191e-01 4.05291421e-03 -1.20665753e+00
-7.62128651e-01 6.88900352e-01 -3.16252530e-01 -1.48847163e+00
-5.14064133e-01 -7.41852820e-01 -1.08577657e+00 1.39192045e+00
-2.11326051e+00 -1.21227968e+00 -3.62441629e-01 -2.06260309e-02
3.39909196e-02 1.24561064e-01 1.34473419e+00 -2.05101475e-01
-4.14303869e-01 -5.32969236e-01 8.39292765e-01 -1.33071363e-01
5.34596920e-01 -1.17886853e+00 1.38329327e+00 9.93739426e-01
-6.35001540e-01 2.22461671e-01 8.08647454e-01 -9.40814734e-01
-9.39586163e-01 -2.00014114e+00 1.40844667e+00 1.04423292e-01
9.10930216e-01 1.94152907e-01 -8.49185288e-01 8.95774245e-01
2.44972110e-01 3.68781447e-01 4.10682917e-01 1.63975000e-01
-4.17641699e-02 -4.28666383e-01 -9.45180595e-01 3.00080299e-01
4.15314943e-01 -4.08576220e-01 9.39912163e-03 6.18610799e-01
5.01164138e-01 -2.15843156e-01 -1.25398469e+00 7.01584518e-01
5.94525754e-01 -1.00791192e+00 5.88531077e-01 -1.02175641e+00
8.74433994e-01 -2.91749984e-01 -5.95448613e-01 -1.85407841e+00
-6.48753524e-01 -1.24014311e-01 1.98574021e-01 8.25299203e-01
3.41789424e-01 -6.55301154e-01 5.01832545e-01 4.73691463e-01
-5.36994301e-02 -3.25228758e-02 -9.75827754e-01 -8.87062848e-01
6.43383205e-01 -2.69456655e-01 1.01927865e+00 1.27439368e+00
-4.02794927e-01 -4.45557773e-01 -4.73623037e-01 9.77442801e-01
4.97633576e-01 4.96027619e-01 7.14703441e-01 -1.66782892e+00
4.77708399e-01 -3.40461493e-01 -2.28241950e-01 -8.48250151e-01
2.28892155e-02 -5.46789229e-01 -1.08885668e-01 -1.30517054e+00
-4.02974516e-01 -3.58543664e-01 -1.44824401e-01 9.30064023e-01
-3.36086378e-02 2.47226924e-01 -4.32332270e-02 2.69422114e-01
9.32152495e-02 8.23827147e-01 6.48746073e-01 -1.32003754e-01
1.66161463e-01 -5.79647627e-03 -1.20752685e-01 6.34651423e-01
1.27820849e+00 -3.39894056e-01 4.02178347e-01 -7.48079777e-01
5.27555585e-01 4.80308771e-01 7.22858429e-01 -1.37873471e+00
-2.73907602e-01 -6.59285307e-01 6.34730160e-01 -1.12513626e+00
-6.54982254e-02 -6.38646662e-01 6.13743126e-01 7.03432024e-01
-2.60202974e-01 4.40041609e-02 7.97853827e-01 1.98465332e-01
-4.31406021e-01 4.14787143e-01 5.29526055e-01 -3.52473170e-01
-1.16124606e+00 4.90493178e-01 -8.89931679e-01 -5.13457775e-01
6.46757543e-01 3.66485476e-01 -5.06415606e-01 -4.59544182e-01
-9.95278656e-01 6.40889108e-01 6.48643911e-01 4.12124246e-01
9.91888903e-03 -1.26031935e+00 -1.26185930e+00 4.72851425e-01
-2.93292180e-02 -2.15677559e-01 5.32370031e-01 5.39993942e-01
-7.61956513e-01 7.01134145e-01 -6.03035390e-01 -4.91246879e-01
-7.62195349e-01 2.55118847e-01 8.43299150e-01 -1.65806398e-01
-4.07543778e-01 7.58907855e-01 -2.42919922e-01 -1.17814612e+00
-3.35821688e-01 -3.64193171e-01 -3.06000561e-01 2.92789370e-01
9.83362854e-01 1.00052267e-01 1.77345246e-01 -6.52616262e-01
-4.66269851e-01 -3.01717930e-02 5.83937168e-01 -4.50571217e-02
1.89168096e+00 -3.89657825e-01 -3.01027834e-01 5.64641178e-01
9.97352660e-01 -5.24704754e-01 -1.40697682e+00 -1.37955666e-01
1.32476449e-01 -3.33848029e-01 8.00648630e-01 -1.10659599e+00
-1.19478476e+00 1.13860393e+00 1.08118165e+00 4.72841144e-01
1.07204092e+00 -5.80741167e-01 6.10960722e-01 6.62744164e-01
1.25820443e-01 -6.59369409e-01 -1.17505300e+00 8.02247763e-01
1.33130825e+00 -1.47737443e+00 -5.78602683e-03 3.13119680e-01
-2.11237952e-01 1.17165518e+00 4.62388247e-01 -3.17764059e-02
1.12652159e+00 3.59182149e-01 3.67503017e-01 -8.07918459e-02
-1.03693223e+00 -3.65791917e-01 1.51605709e-02 5.93568742e-01
4.98532116e-01 7.12866783e-01 2.97705859e-01 2.13065282e-01
-2.21842542e-01 3.35094243e-01 4.48678792e-01 5.30934691e-01
-7.37224460e-01 -4.48549092e-01 -7.26973355e-01 5.69514215e-01
-3.78952213e-02 -6.67790830e-01 1.06681436e-02 2.52988577e-01
-1.37571424e-01 5.61753154e-01 4.27261204e-01 -2.41200179e-01
-1.26526237e-01 -4.24350686e-02 -4.75465596e-01 -2.44380862e-01
-1.56113103e-01 -5.52416265e-01 2.27661461e-01 -8.09896648e-01
-4.07173663e-01 -7.05407441e-01 -5.81745446e-01 -8.86024356e-01
9.66416895e-02 6.33936562e-03 1.00038576e+00 6.93308413e-01
8.72507572e-01 1.30387053e-01 8.13413262e-01 -1.72781432e+00
-5.04530251e-01 -1.42225909e+00 -9.19571936e-01 -3.33909005e-01
6.67923748e-01 -1.63542986e-01 -3.82382154e-01 9.07793716e-02] | [9.48739242553711, -1.5305424928665161] |
5c3c8d76-9e56-410e-8fc7-10986e8fa678 | label-aware-hyperbolic-embeddings-for-fine | 2306.14822 | null | https://arxiv.org/abs/2306.14822v1 | https://arxiv.org/pdf/2306.14822v1.pdf | Label-Aware Hyperbolic Embeddings for Fine-grained Emotion Classification | Fine-grained emotion classification (FEC) is a challenging task. Specifically, FEC needs to handle subtle nuance between labels, which can be complex and confusing. Most existing models only address text classification problem in the euclidean space, which we believe may not be the optimal solution as labels of close semantic (e.g., afraid and terrified) may not be differentiated in such space, which harms the performance. In this paper, we propose HypEmo, a novel framework that can integrate hyperbolic embeddings to improve the FEC task. First, we learn label embeddings in the hyperbolic space to better capture their hierarchical structure, and then our model projects contextualized representations to the hyperbolic space to compute the distance between samples and labels. Experimental results show that incorporating such distance to weight cross entropy loss substantially improves the performance with significantly higher efficiency. We evaluate our proposed model on two benchmark datasets and found 4.8% relative improvement compared to the previous state of the art with 43.2% fewer parameters and 76.9% less training time. Code is available at https: //github.com/dinobby/HypEmo. | ['Lun-Wei Ku', 'Yi-Li Hsu', 'Tun-Min Hung', 'Chih-Yao Chen'] | 2023-06-26 | null | null | null | null | ['emotion-classification', 'classification-1', 'text-classification', 'emotion-classification'] | ['computer-vision', 'methodology', 'natural-language-processing', 'natural-language-processing'] | [-1.96250424e-01 -1.90392435e-01 -8.71623214e-03 -6.12647176e-01
-7.48932183e-01 -4.78038847e-01 1.39398575e-01 4.55760717e-01
-4.12642270e-01 3.94148439e-01 3.15839201e-01 1.49627747e-02
1.18656931e-02 -6.15472198e-01 -3.85487191e-02 -7.27047384e-01
1.49706692e-01 1.61843598e-01 -3.70640792e-02 -2.73933169e-02
4.03111875e-01 1.89452648e-01 -1.28910100e+00 3.28860343e-01
8.97248745e-01 1.52468765e+00 -3.15746158e-01 3.51772040e-01
-1.44544141e-02 7.01733530e-01 -4.41826314e-01 -4.94037211e-01
-5.78810386e-02 -2.01578617e-01 -8.65365863e-01 -2.00180382e-01
2.84701511e-02 -1.71684384e-01 -1.90199658e-01 1.02139056e+00
4.59535420e-01 2.82327801e-01 9.75040317e-01 -1.47597456e+00
-9.43353951e-01 4.55393702e-01 -7.18924165e-01 4.35618684e-02
1.07798256e-01 -4.84380603e-01 1.29858077e+00 -1.13839412e+00
3.86232920e-02 1.15037453e+00 8.71158957e-01 4.46046621e-01
-9.11426902e-01 -9.10124838e-01 6.03676699e-02 4.51718390e-01
-1.75573194e+00 -1.19932830e-01 8.16857457e-01 -3.07206631e-01
5.12905121e-01 3.47352028e-01 4.00808901e-01 1.16588414e+00
-9.10011120e-04 6.31764412e-01 1.25692093e+00 -2.87464172e-01
3.08124542e-01 4.99094516e-01 6.07429028e-01 6.39730573e-01
3.55290622e-02 -3.02145392e-01 -2.74248332e-01 -2.27394044e-01
8.16106945e-02 2.43745506e-01 -3.27968627e-01 -2.53682546e-02
-8.12606931e-01 1.08415055e+00 7.35485256e-01 2.36465394e-01
-8.71550888e-02 -1.16357259e-01 5.83762944e-01 5.32911764e-03
6.82901025e-01 4.36411798e-01 -3.70272964e-01 -3.83609295e-01
-5.82278252e-01 -7.32386485e-02 6.89051151e-01 7.37146258e-01
5.94779670e-01 -3.66377950e-01 5.48602007e-02 1.17521477e+00
2.66036421e-01 -1.02791851e-02 5.64521134e-01 -9.01040614e-01
3.25284779e-01 6.25772715e-01 4.15412970e-02 -1.42896295e+00
-5.10239840e-01 -3.98831487e-01 -9.11465585e-01 -1.23476483e-01
2.27144793e-01 -2.15180710e-01 -4.01288390e-01 1.53546929e+00
4.96060133e-01 3.18525136e-01 5.29685393e-02 1.03181291e+00
7.09465981e-01 9.35979962e-01 1.43230334e-01 4.96765301e-02
1.55578434e+00 -1.21305287e+00 -7.54330993e-01 -1.16680615e-01
9.46778297e-01 -7.84290433e-01 1.41559696e+00 3.05414975e-01
-6.24356747e-01 -2.44885400e-01 -1.20354676e+00 -3.58683228e-01
-6.76984310e-01 1.66100070e-01 5.58071673e-01 7.74447083e-01
-6.58332288e-01 6.67456865e-01 -6.00569785e-01 -3.31396312e-01
4.95816618e-01 3.60631594e-03 -3.14959168e-01 -9.20453519e-02
-1.42670894e+00 7.05341101e-01 4.12125766e-01 6.22993633e-02
-1.97054058e-01 -6.55074656e-01 -9.19474125e-01 2.91645050e-01
2.69450158e-01 -9.67085268e-03 1.13755631e+00 -4.44814265e-01
-1.33881819e+00 6.48379803e-01 4.74777538e-03 1.02275135e-02
2.07562178e-01 -3.22147280e-01 -5.27445138e-01 8.21780041e-02
-8.59974474e-02 4.87918288e-01 5.19242048e-01 -1.15396297e+00
-5.20844042e-01 -4.43831980e-01 5.20984195e-02 4.63874847e-01
-9.30034578e-01 9.71916318e-02 -1.07689098e-01 -6.23644471e-01
8.60585868e-02 -1.02056479e+00 1.53022408e-01 1.53295413e-01
-3.26649010e-01 -5.49613178e-01 1.08830857e+00 -5.68516374e-01
1.59395075e+00 -2.45627093e+00 -2.70533621e-01 -8.71942565e-02
3.07696819e-01 1.50007233e-01 2.09883854e-01 3.88908863e-01
6.17484711e-02 4.62405354e-01 -1.95679501e-01 -4.35860068e-01
4.07119095e-01 7.82512575e-02 -2.20762208e-01 5.48613310e-01
1.01700805e-01 5.95653653e-01 -7.97208548e-01 -6.25569642e-01
-8.01064819e-02 6.70752883e-01 -5.41058481e-01 2.63850152e-01
3.65821391e-01 -7.43169244e-03 -5.49465954e-01 7.27852285e-01
7.94745922e-01 -4.49140847e-01 -9.91130024e-02 -3.41041952e-01
1.50290132e-01 2.73310304e-01 -1.06902218e+00 1.32997119e+00
-6.33980930e-01 6.01171076e-01 -1.81888521e-01 -1.05290055e+00
1.07274413e+00 2.90580094e-01 3.22944313e-01 -4.49992716e-01
4.61598784e-01 -2.34378055e-02 -2.36737564e-01 -4.90718365e-01
6.18140996e-01 -3.02290708e-01 -3.81199241e-01 5.63602567e-01
-1.37459129e-01 3.70097123e-02 -3.10248524e-01 -2.90926974e-02
8.49364519e-01 -2.35917285e-01 4.58328038e-01 -3.25361133e-01
2.98084736e-01 -3.92832756e-01 8.54576528e-01 1.73210353e-01
-6.89918578e-01 5.75418830e-01 6.32023871e-01 -2.43814945e-01
-7.99543619e-01 -1.01404703e+00 -4.71557379e-01 1.33341002e+00
3.66976917e-01 -6.95993781e-01 -7.33354807e-01 -9.18934047e-01
-1.86977029e-01 8.50462437e-01 -8.90801728e-01 -3.97095054e-01
-1.66832998e-01 -9.67692971e-01 5.52361786e-01 5.15921056e-01
4.94706184e-01 -7.62582779e-01 -5.39657414e-01 -7.18579516e-02
-2.90134907e-01 -1.00138664e+00 -6.26576483e-01 2.11768404e-01
-6.32830679e-01 -9.00212646e-01 -3.55310172e-01 -8.97734165e-01
4.33826208e-01 2.38174826e-01 8.86229038e-01 2.96194945e-02
-2.51236051e-01 1.70288831e-02 -8.36886287e-01 -2.92010158e-01
6.47050589e-02 1.24192938e-01 1.18038254e-02 2.30209291e-01
7.26486087e-01 -4.35457617e-01 -8.58033657e-01 5.36577106e-01
-8.27989519e-01 -1.41413152e-01 1.42706603e-01 1.04720557e+00
4.04638320e-01 2.83453733e-01 6.21208131e-01 -8.44204009e-01
7.07811594e-01 -8.82424116e-01 -2.32998244e-02 2.12106571e-01
-7.67446399e-01 -1.97783709e-01 8.70881677e-01 -4.28848803e-01
-9.11964118e-01 -2.91741788e-01 -2.14232609e-01 -2.14078620e-01
-1.99705273e-01 4.14152950e-01 -7.06414431e-02 9.37816426e-02
5.39462507e-01 -1.13739528e-01 -3.31112534e-01 -4.58964616e-01
3.24558139e-01 1.31137550e+00 1.56698734e-01 -7.14852214e-01
2.37133205e-01 3.28540713e-01 -3.16664279e-01 -5.45490086e-01
-1.16407061e+00 -6.36203349e-01 -4.02759850e-01 7.26319896e-03
8.58189881e-01 -6.57578170e-01 -6.75391853e-01 4.86045629e-02
-9.34040964e-01 -9.04838666e-02 6.70135990e-02 5.70787668e-01
-3.69758934e-01 3.38061571e-01 -7.67835319e-01 -6.80530667e-01
-3.81288826e-01 -9.75749552e-01 9.66247320e-01 4.06505018e-01
-3.56901199e-01 -1.10028970e+00 6.16788194e-02 3.87329817e-01
4.37094659e-01 3.36128324e-01 8.95308256e-01 -8.24276149e-01
2.22105205e-01 -3.50645542e-01 -5.39501071e-01 4.57482874e-01
2.02732742e-01 -6.12597354e-03 -1.19999444e+00 -1.51673898e-01
1.92136392e-01 -6.92317367e-01 6.48982763e-01 -1.52122438e-01
1.60307789e+00 -2.68796980e-01 -2.29761705e-01 6.59812212e-01
1.30363607e+00 1.86682820e-01 4.71405655e-01 1.49419054e-01
6.75474882e-01 6.20238721e-01 7.89915621e-01 7.95593441e-01
5.73795974e-01 6.70524478e-01 1.76647648e-01 6.33147135e-02
1.39513209e-01 -1.78588346e-01 3.16920102e-01 1.33657885e+00
2.97975838e-01 -2.07133666e-01 -9.33699071e-01 3.96444798e-01
-1.80157185e+00 -8.20504904e-01 5.15366979e-02 1.71364725e+00
8.68743122e-01 -1.05015501e-01 -1.61721379e-01 4.06945080e-01
7.97363818e-01 1.78281739e-01 -4.80283678e-01 -7.87915111e-01
2.13180155e-01 1.71334490e-01 9.76219997e-02 4.23327982e-01
-1.29241061e+00 8.10372889e-01 5.09654093e+00 1.02889836e+00
-1.05656028e+00 4.16866928e-01 9.29245412e-01 -7.94194341e-02
-5.26971407e-02 -2.23127648e-01 -6.73505723e-01 6.34798646e-01
9.71425116e-01 -3.63919020e-01 1.84612304e-01 9.21232045e-01
-3.74489427e-02 2.63127744e-01 -1.01565564e+00 1.24615896e+00
2.53513515e-01 -7.61454463e-01 -3.81372750e-01 -1.79855555e-01
6.14967167e-01 -3.45238060e-01 3.27122211e-01 6.25063360e-01
1.20948106e-01 -1.10232937e+00 4.72868115e-01 3.13735008e-01
6.46562696e-01 -9.45600331e-01 9.28157687e-01 2.48365283e-01
-1.27421427e+00 -1.90402418e-01 -6.77002609e-01 -1.32500112e-01
-6.76685944e-03 6.49811029e-01 -6.19235694e-01 3.00247997e-01
9.51535225e-01 8.45829606e-01 -5.61730385e-01 7.59623706e-01
3.51935104e-02 6.10556960e-01 -1.34178117e-01 -3.58056694e-01
3.50627750e-01 -2.78185695e-01 -4.37074304e-02 1.38846493e+00
6.30264044e-01 3.90040398e-01 1.77704439e-01 5.00305593e-01
-3.42535108e-01 5.61023295e-01 -4.67940956e-01 6.17893355e-04
6.69887185e-01 1.55281925e+00 -6.43479228e-01 -2.15354979e-01
-5.27977884e-01 1.20328569e+00 7.68314600e-01 1.20186202e-01
-1.16048384e+00 -8.48085701e-01 7.58679926e-01 -2.90926903e-01
1.03437491e-01 2.97104493e-02 -3.94328266e-01 -1.16889679e+00
-3.63666676e-02 -4.79193807e-01 7.27662981e-01 -6.83282316e-01
-1.67030466e+00 8.21382046e-01 -3.23006570e-01 -1.15440118e+00
2.25743234e-01 -6.67773545e-01 -4.20731395e-01 5.72507799e-01
-1.26099980e+00 -7.67997682e-01 -5.77894032e-01 4.00443792e-01
3.67533505e-01 2.53559381e-01 1.04969382e+00 5.15237868e-01
-6.95127428e-01 1.05386245e+00 4.59648073e-01 1.23652041e-01
1.03269970e+00 -1.32260191e+00 -2.32674524e-01 2.40529418e-01
6.98176473e-02 3.74883920e-01 3.54824841e-01 -7.38986656e-02
-7.05227196e-01 -1.23953247e+00 1.01458371e+00 -3.19410831e-01
7.73033857e-01 -6.19838417e-01 -1.09292924e+00 4.73219663e-01
1.73312262e-01 1.94634303e-01 1.29655087e+00 3.01608950e-01
-7.76102364e-01 -2.01204836e-01 -1.34123123e+00 7.47405767e-01
8.57133925e-01 -6.28926635e-01 -5.00009537e-01 2.50510246e-01
8.27533782e-01 -1.53927937e-01 -1.16124785e+00 2.65456825e-01
5.56427419e-01 -1.02438939e+00 6.57464325e-01 -4.52479511e-01
4.91107702e-01 -1.36379078e-01 -5.75617731e-01 -1.45561647e+00
-5.02827227e-01 -1.43921718e-01 3.52840498e-02 1.30001390e+00
3.55549276e-01 -7.48689115e-01 5.95629454e-01 6.31873608e-01
-1.38932616e-01 -1.35180640e+00 -8.31908286e-01 -6.55953884e-01
3.93574089e-01 -4.25905615e-01 7.68583536e-01 1.35039485e+00
5.74952364e-01 3.51585180e-01 -3.38672400e-01 1.89857870e-01
3.82122457e-01 1.29918218e-01 8.29051733e-02 -1.26865280e+00
7.51141533e-02 -5.83625257e-01 -4.68008488e-01 -7.76156187e-01
3.04986924e-01 -1.12490726e+00 3.28508280e-02 -1.26132333e+00
3.37688088e-01 -7.46504962e-01 -7.18595505e-01 4.85147983e-01
-3.44450951e-01 4.50643808e-01 9.64036509e-02 1.57974008e-02
-7.46435285e-01 9.75224376e-01 8.82970750e-01 -5.71376942e-02
9.41455215e-02 -3.94068658e-01 -1.02753389e+00 7.22254634e-01
1.40810084e+00 -6.58244252e-01 -4.11619812e-01 -2.51535535e-01
9.86789241e-02 -3.03432614e-01 5.29165864e-02 -9.74427462e-01
1.69133931e-01 -2.14010272e-02 2.45415375e-01 -5.58853820e-02
5.64347386e-01 -8.19730341e-01 -2.24052474e-01 2.00155284e-02
-6.08653247e-01 1.18019186e-01 7.10709542e-02 5.13074696e-01
-3.52609396e-01 -4.15400445e-01 8.94203663e-01 3.13610375e-01
-4.31365937e-01 4.37047243e-01 4.38107178e-02 2.53338128e-01
1.23578012e+00 1.28331795e-01 -4.32729214e-01 -2.40428939e-01
-6.56423867e-01 1.97313264e-01 3.94886762e-01 4.35883671e-01
6.54999435e-01 -1.75145209e+00 -5.08759320e-01 -2.17856318e-01
3.73251826e-01 -2.71583915e-01 3.84028196e-01 7.38489807e-01
-4.24978405e-01 2.02614740e-01 -2.58479901e-02 -2.51216352e-01
-1.22141242e+00 6.19704306e-01 2.09201947e-01 -1.81270301e-01
-5.57655513e-01 8.38132262e-01 1.78281784e-01 -6.85613155e-01
2.82797933e-01 -2.50388980e-01 -2.79026985e-01 3.10630590e-01
5.13754547e-01 5.09618580e-01 -8.75613391e-02 -7.79733360e-01
-4.14866567e-01 7.51179576e-01 -3.03480059e-01 8.11789110e-02
1.00428379e+00 -1.89025208e-01 1.34716006e-02 7.44766235e-01
1.94103229e+00 -1.10125631e-01 -1.07731688e+00 -2.99013257e-01
-5.20854704e-02 -5.73443234e-01 1.88299492e-01 -7.37636209e-01
-9.45085049e-01 1.26849496e+00 6.16837621e-01 3.77868891e-01
1.18941140e+00 3.57495621e-02 9.61318851e-01 2.50593275e-01
1.26438543e-01 -1.36512947e+00 2.48461336e-01 4.91182238e-01
7.39727259e-01 -1.27337754e+00 -1.35899514e-01 -4.73865151e-01
-9.01437461e-01 1.08538413e+00 6.02943718e-01 -3.35000195e-02
1.00242221e+00 -3.75536904e-02 1.56541362e-01 -8.73393416e-02
-7.92974234e-01 1.88743204e-01 1.04330488e-01 2.39024892e-01
6.20937645e-01 3.46024990e-01 -4.32314545e-01 1.12005663e+00
-3.79488587e-01 -3.86868566e-01 4.28179264e-01 7.73702502e-01
-4.70930189e-01 -8.17609489e-01 -2.10422054e-01 4.68102634e-01
-5.35871744e-01 -2.68764403e-02 -1.64990127e-01 5.10718346e-01
3.05692762e-01 1.13636374e+00 1.60932168e-01 -5.70924699e-01
1.76200718e-01 2.60517955e-01 6.59234524e-02 -4.05505568e-01
-2.56714344e-01 -2.45068252e-01 -1.65714009e-03 -4.23179507e-01
1.59369893e-02 -5.26607394e-01 -1.38474238e+00 -4.87180322e-01
-2.20410675e-01 4.05569732e-01 6.42264605e-01 5.10613263e-01
4.92412001e-01 3.57901633e-01 9.79009748e-01 -4.68933821e-01
-7.32321084e-01 -9.77336943e-01 -8.96421909e-01 5.70501447e-01
1.70546606e-01 -9.04033303e-01 -6.12286150e-01 -1.95595548e-01] | [10.42637825012207, 6.687601089477539] |
8251b69e-4dc0-4bc8-b017-b6f178ac93b9 | medai-at-semeval-2021-task-10-negation-aware | null | null | https://aclanthology.org/2021.semeval-1.183 | https://aclanthology.org/2021.semeval-1.183.pdf | MedAI at SemEval-2021 Task 10: Negation-aware Pre-training for Source-free Negation Detection Domain Adaptation | Due to the increasing concerns for data privacy, source-free unsupervised domain adaptation attracts more and more research attention, where only a trained source model is assumed to be available, while the labeled source data remain private. To get promising adaptation results, we need to find effective ways to transfer knowledge learned in source domain and leverage useful domain specific information from target domain at the same time. This paper describes our winning contribution to SemEval 2021 Task 10: Source-Free Domain Adaptation for Semantic Processing. Our key idea is to leverage the model trained on source domain data to generate pseudo labels for target domain samples. Besides, we propose Negation-aware Pre-training (NAP) to incorporate negation knowledge into model. Our method win the 1st place with F1-score of 0.822 on the official blind test set of Negation Detection Track. | ['Lei Zhang', 'Yu Wang', 'Qi Zhang', 'Jinquan Sun'] | 2021-08-01 | null | null | null | semeval-2021 | ['source-free-domain-adaptation', 'negation-detection'] | ['computer-vision', 'natural-language-processing'] | [ 3.02573770e-01 4.32790101e-01 -5.44875026e-01 -8.64744484e-01
-9.61010039e-01 -8.55289876e-01 5.35858274e-01 1.01660796e-01
-7.04154551e-01 1.10049427e+00 3.28535110e-01 -3.88507508e-02
3.67538661e-01 -7.21158028e-01 -7.97900558e-01 -3.49722654e-01
5.99051476e-01 4.91838068e-01 1.32843703e-01 -2.62588322e-01
-1.42890573e-01 9.18662027e-02 -7.34018266e-01 5.31199276e-01
1.12749660e+00 9.58678126e-01 -2.07249403e-01 2.57513523e-01
-2.72504836e-01 5.74472189e-01 -7.33129442e-01 -1.06294501e+00
3.51906121e-01 -6.55614078e-01 -8.43742371e-01 -1.78644553e-01
2.54648417e-01 -2.35069275e-01 -1.14216857e-01 1.53294265e+00
5.66038132e-01 -1.65483803e-01 9.04336452e-01 -1.27815604e+00
-1.04744208e+00 7.22942770e-01 -2.77040839e-01 -1.42062649e-01
-5.87403513e-02 5.04353084e-02 7.88794219e-01 -8.52681994e-01
9.58625257e-01 9.79781330e-01 3.44009846e-01 1.19895053e+00
-1.14986157e+00 -1.04881299e+00 -1.32027660e-02 2.77418941e-01
-1.13463283e+00 -6.16127253e-01 7.33867943e-01 -1.06286794e-01
6.16153896e-01 -4.72920895e-01 1.91728342e-02 1.65047920e+00
-3.00120592e-01 8.34984422e-01 1.02576387e+00 -6.71579778e-01
6.71520293e-01 7.04970360e-01 1.56244729e-02 2.55568564e-01
5.14631569e-01 2.11535729e-02 -7.18036056e-01 -2.62012005e-01
1.69113472e-01 -2.38470078e-01 -2.01760143e-01 -6.99976623e-01
-9.47211623e-01 9.23572779e-01 3.29471409e-01 -7.29843751e-02
-1.76576853e-01 -3.05218995e-01 5.39562821e-01 6.11114085e-01
5.70174396e-01 2.99899608e-01 -9.97251809e-01 1.30590901e-01
-7.03408241e-01 3.43534082e-01 9.12323594e-01 1.27178192e+00
6.91917479e-01 -2.22608760e-01 -1.06805772e-01 9.21371460e-01
2.11343139e-01 9.16643560e-01 5.12514174e-01 -8.19571018e-01
7.75642753e-01 5.91373086e-01 3.54945958e-01 -2.62138814e-01
4.49295938e-02 -3.63481611e-01 -7.13227034e-01 1.19063348e-01
6.70326650e-01 -5.42328656e-01 -1.01302993e+00 1.84260845e+00
1.81233823e-01 -2.74599660e-02 7.78605282e-01 6.98748231e-01
6.34081364e-01 3.46510559e-01 4.76509809e-01 1.73880681e-02
1.24609029e+00 -7.95837581e-01 -7.05775440e-01 -5.57283819e-01
7.89833426e-01 -5.43091238e-01 1.02790487e+00 2.57146358e-01
-6.02354825e-01 -1.27409786e-01 -1.08637202e+00 1.71571951e-02
-4.99451756e-01 3.05530578e-02 3.64035130e-01 8.88810992e-01
-6.06910944e-01 1.24723502e-01 -5.56593716e-01 -4.10525978e-01
9.79504347e-01 2.84136117e-01 -7.38621175e-01 -5.95754504e-01
-1.53663707e+00 6.72846138e-01 7.17144787e-01 -5.31642556e-01
-7.90614724e-01 -7.19716430e-01 -9.07108068e-01 -1.76121265e-01
4.11536455e-01 -7.14702487e-01 1.40743256e+00 -1.28234434e+00
-1.40207458e+00 1.28954315e+00 -3.19986552e-01 -9.17087495e-01
5.51187038e-01 -2.30178565e-01 -7.78541744e-01 1.06706709e-01
3.39186162e-01 6.96566284e-01 1.01820362e+00 -1.10081398e+00
-7.23012447e-01 -3.34062368e-01 -1.68958083e-01 1.30248442e-01
-7.09366262e-01 2.28583952e-03 -2.84214199e-01 -5.94234705e-01
-3.16415310e-01 -7.72146165e-01 -7.94564337e-02 -2.57218517e-02
-4.01918620e-01 -2.24645033e-01 7.08260357e-01 -7.94659197e-01
8.21488857e-01 -2.35199738e+00 -1.97773248e-01 1.83218658e-01
6.91960156e-02 7.12418854e-01 -1.68334946e-01 2.52250619e-02
-2.26239964e-01 8.41180980e-02 -5.20419300e-01 -2.80113041e-01
1.71439663e-01 2.01797202e-01 -6.61808789e-01 2.36472234e-01
4.15410548e-01 9.38280165e-01 -9.32794273e-01 -3.00436765e-01
-1.76680088e-01 8.39109346e-02 -8.23857844e-01 2.47029454e-01
-5.15625000e-01 3.26517671e-01 -4.66870159e-01 7.53170609e-01
1.11442304e+00 -7.53031522e-02 5.41694425e-02 -6.27704114e-02
5.98060191e-01 2.75900424e-01 -9.83248889e-01 1.81603122e+00
-1.81162879e-01 2.06569761e-01 1.25997260e-01 -1.12435186e+00
9.48045790e-01 4.47696209e-01 3.29788402e-02 -8.25489819e-01
2.66058147e-02 4.60659146e-01 -2.71070004e-01 -3.49299192e-01
1.12248152e-01 -4.99720871e-01 -3.45657796e-01 1.69478878e-01
5.15105307e-01 2.09237318e-02 -1.34758472e-01 2.99258828e-01
1.02022445e+00 1.86551921e-02 3.62519503e-01 2.54288279e-02
5.10395408e-01 3.00741673e-01 8.93797517e-01 4.65780973e-01
-4.88462567e-01 5.14992476e-01 5.33467054e-01 -3.83548513e-02
-9.27741647e-01 -1.21215963e+00 9.08141062e-02 8.35031033e-01
-6.70989081e-02 -2.11442187e-01 -8.07656586e-01 -1.48075473e+00
1.94548313e-02 1.05415428e+00 -3.95824552e-01 -5.71034193e-01
-4.26000029e-01 -4.70549822e-01 8.58008087e-01 4.44081336e-01
7.97645211e-01 -1.01570058e+00 1.15676060e-01 4.41837572e-02
-4.01460856e-01 -1.20704651e+00 -3.17043781e-01 4.18207079e-01
-9.16425228e-01 -9.82015371e-01 -8.79559875e-01 -8.68368983e-01
7.54813612e-01 -1.93510100e-01 9.84980583e-01 -8.43453348e-01
2.80935168e-01 2.82247573e-01 -4.79341954e-01 -8.67042899e-01
-5.16510546e-01 1.77939788e-01 8.98308232e-02 1.30289020e-02
1.20413256e+00 -3.60529184e-01 -2.89490223e-01 1.70767710e-01
-9.46545601e-01 -5.67095459e-01 6.79656267e-01 9.08221066e-01
6.61716104e-01 4.96680960e-02 1.01987970e+00 -1.39065039e+00
6.19679451e-01 -6.65466547e-01 -6.50631547e-01 3.04986805e-01
-6.69911385e-01 6.59620166e-02 7.69718230e-01 -3.56424898e-01
-1.42211342e+00 3.42943668e-01 -2.39738915e-02 -2.79970974e-01
-3.35468054e-01 2.77259231e-01 -9.19016540e-01 1.14185758e-01
1.17147553e+00 1.24244027e-01 -3.45610343e-02 -5.30650914e-01
6.50148630e-01 9.41868365e-01 6.54319048e-01 -4.37847674e-01
1.03644896e+00 5.44286788e-01 -5.12715399e-01 -4.88411576e-01
-1.15658343e+00 -5.07446885e-01 -5.23768723e-01 5.94588399e-01
7.91522086e-01 -1.28987551e+00 -6.42412379e-02 5.93040049e-01
-1.15247500e+00 -2.16099426e-01 -5.45508444e-01 4.03292805e-01
-3.02645892e-01 1.81896538e-01 -1.65274348e-02 -5.09676099e-01
-3.05014938e-01 -4.92125660e-01 7.07740843e-01 1.81522846e-01
-1.50197223e-01 -1.00534225e+00 1.06562041e-01 5.92736363e-01
3.60883743e-01 -7.83075690e-02 8.55613172e-01 -1.44220686e+00
-4.11381394e-01 -3.68243903e-01 -3.46855253e-01 1.15351260e+00
2.50291705e-01 -8.87027502e-01 -1.17806923e+00 -1.32077724e-01
2.90548682e-01 -5.62079608e-01 1.02423453e+00 8.31116959e-02
8.49274933e-01 -4.21809077e-01 -3.46943736e-01 5.25518477e-01
1.35000432e+00 -4.66897972e-02 5.72751403e-01 2.48609453e-01
3.86171132e-01 5.17981708e-01 7.41399646e-01 2.07253173e-01
3.35166246e-01 3.45832199e-01 2.74685085e-01 1.49774820e-01
-2.73641378e-01 -6.79360449e-01 5.26581407e-01 1.16131119e-01
5.46502292e-01 -3.45442951e-01 -8.58000636e-01 8.95718396e-01
-1.54384422e+00 -8.03071380e-01 2.58328885e-01 2.36778736e+00
1.20807457e+00 2.84282774e-01 -1.01491019e-01 -3.00945401e-01
6.95024490e-01 -3.65658551e-01 -8.51319492e-01 -8.89064446e-02
-4.51916993e-01 4.90021378e-01 7.98637033e-01 3.04236025e-01
-1.23336017e+00 1.27158773e+00 5.70847034e+00 8.77084434e-01
-9.39019561e-01 3.47507775e-01 4.00340229e-01 -9.15669203e-02
-3.66536200e-01 2.30232179e-01 -1.12870920e+00 5.22053301e-01
1.20177925e+00 -4.27765161e-01 1.14012606e-01 1.18721557e+00
-2.43662983e-01 6.74858838e-02 -1.10754180e+00 9.19061542e-01
1.03044488e-01 -1.11964381e+00 7.66900629e-02 6.19600974e-02
9.04749155e-01 2.45455593e-01 3.37663069e-02 6.10610485e-01
5.70831716e-01 -9.16829944e-01 3.45156282e-01 3.59678894e-01
1.04367983e+00 -7.27193415e-01 1.02762127e+00 3.88630271e-01
-4.70165044e-01 -5.21043837e-02 -6.47703886e-01 1.76402748e-01
5.51211797e-02 9.36194956e-01 -1.34849513e+00 4.54486310e-01
4.20133561e-01 6.60204351e-01 -4.02819693e-01 1.02295840e+00
-6.32669568e-01 9.52870011e-01 -2.81202048e-01 2.05607742e-01
-1.60766795e-01 1.50517732e-01 4.15798455e-01 1.19433808e+00
2.04816103e-01 -9.82878804e-02 -1.84552193e-01 6.87477827e-01
-7.72017002e-01 2.09083691e-01 -8.31515133e-01 -2.68043488e-01
6.96290672e-01 8.31811070e-01 -1.71200454e-01 -3.27436745e-01
-5.60427666e-01 1.39708924e+00 2.18908951e-01 5.22745728e-01
-6.12855077e-01 -6.06136739e-01 8.89697969e-01 5.52173369e-02
3.56129855e-01 2.08727010e-02 -4.80233938e-01 -1.45521343e+00
2.86759049e-01 -9.19682622e-01 7.51768053e-01 -5.09773254e-01
-1.78601444e+00 2.54774451e-01 -1.28297895e-01 -1.21653998e+00
-4.03766513e-01 -6.88798964e-01 -1.71972677e-01 1.17462802e+00
-1.94698560e+00 -1.16686952e+00 2.39381324e-02 1.02598023e+00
2.20753953e-01 -6.49632812e-01 1.16057408e+00 3.11652780e-01
-5.75403858e-04 1.19860542e+00 3.77346396e-01 6.16712093e-01
1.52702117e+00 -1.13510776e+00 3.65334034e-01 9.13425505e-01
-1.25211757e-02 7.11871684e-01 4.90590751e-01 -8.88803899e-01
-8.25273216e-01 -1.40127754e+00 1.43906569e+00 -8.85112226e-01
4.72622931e-01 -4.80164140e-01 -9.26481545e-01 7.87983894e-01
1.20028473e-01 2.39717439e-01 8.79219592e-01 -4.01819423e-02
-9.28490818e-01 -3.51282299e-01 -1.73988986e+00 3.17886919e-01
9.46154058e-01 -7.19009757e-01 -8.94963264e-01 3.58907372e-01
1.00832808e+00 -2.82839537e-01 -4.81531799e-01 1.92364633e-01
-1.90518439e-01 -5.80656350e-01 9.04195130e-01 -7.50402331e-01
3.74827415e-01 -3.51355582e-01 -3.68017703e-01 -1.35976505e+00
2.25449413e-01 -3.87275785e-01 3.45681049e-02 1.59496295e+00
7.00752199e-01 -8.13844383e-01 1.12881589e+00 8.92250657e-01
2.51515150e-01 1.99489549e-01 -9.95599091e-01 -9.23370898e-01
4.01656210e-01 -3.86946380e-01 4.38232124e-01 1.11215258e+00
-3.18140648e-02 5.57171464e-01 -2.40336269e-01 2.67007053e-01
7.97386169e-01 -1.80262372e-01 7.45353878e-01 -1.21632051e+00
-7.04636872e-02 1.97440162e-01 -2.81946003e-01 -9.68965828e-01
5.10183930e-01 -1.23575604e+00 -2.48953447e-01 -1.28179097e+00
6.67122751e-03 -3.39040607e-01 -4.44082737e-01 8.41822982e-01
4.22559492e-02 1.25899896e-01 -7.35643432e-02 4.05537598e-02
-6.14466190e-01 7.12426901e-01 7.52365530e-01 -2.99590439e-01
-4.29894291e-02 2.47507766e-01 -1.36204994e+00 6.32171512e-01
8.98752511e-01 -8.87498081e-01 -6.58658683e-01 -5.00062883e-01
6.81210235e-02 -4.68551725e-01 4.61946309e-01 -9.42751944e-01
1.74886137e-01 -1.79052055e-01 3.73717904e-01 -1.54222757e-01
8.80936086e-02 -1.08508313e+00 -4.51994151e-01 2.04028264e-01
-3.72822285e-01 -7.38963485e-01 1.54136419e-01 8.06413531e-01
-4.88119245e-01 -4.01232928e-01 9.02815163e-01 -1.96501106e-01
-1.03088391e+00 1.92451566e-01 1.51709497e-01 6.35324419e-01
8.81842673e-01 2.02641133e-02 -4.30523634e-01 -3.35103899e-01
-6.98299646e-01 2.14484349e-01 5.24462521e-01 5.37522078e-01
4.10292208e-01 -1.25921679e+00 -8.72906446e-01 2.93114096e-01
6.25928819e-01 1.34682924e-01 8.54984373e-02 6.02871776e-02
1.49687126e-01 3.84978175e-01 -1.76050007e-01 -2.37053588e-01
-8.38330984e-01 5.63209236e-01 1.25740290e-01 -1.15115985e-01
-2.86350876e-01 1.17752087e+00 1.60368904e-01 -8.43428612e-01
2.10009396e-01 2.87206899e-02 4.30005081e-02 1.83878106e-03
6.98957860e-01 2.41991114e-02 3.06718946e-01 -2.48603731e-01
-5.79183638e-01 8.88185725e-02 -3.15592110e-01 -3.08588058e-01
1.19871688e+00 -1.78277604e-02 -5.85877001e-02 -1.32287040e-01
1.29164660e+00 3.11249822e-01 -1.08790541e+00 -8.21696639e-01
2.88514256e-01 -3.40514213e-01 -2.59603143e-01 -1.53850079e+00
-7.99131870e-01 7.63382018e-01 6.86989844e-01 -4.48792845e-01
1.33159208e+00 7.29317870e-03 8.99097860e-01 5.93916714e-01
3.82773012e-01 -1.25665200e+00 -3.11330035e-02 7.07818151e-01
5.32724738e-01 -1.52416992e+00 -4.32246268e-01 -3.82593691e-01
-1.05162847e+00 7.71702230e-01 7.69025922e-01 1.03962131e-01
4.93591398e-01 6.35547191e-02 3.39934587e-01 1.97077990e-01
-4.16765243e-01 4.80553955e-02 6.98605627e-02 1.46091735e+00
1.04408853e-01 -8.74708071e-02 -5.96692525e-02 1.27493811e+00
-7.03418180e-02 4.30046618e-01 3.04925174e-01 6.80245042e-01
-3.31109732e-01 -1.77713728e+00 -1.86887652e-01 3.15492898e-01
-5.19195378e-01 -3.77094001e-01 -5.62191427e-01 4.55722272e-01
1.44098282e-01 8.43119085e-01 -4.41541106e-01 -1.36614561e-01
5.04365683e-01 6.93440020e-01 -2.59289332e-02 -8.28079283e-01
-1.67616323e-01 -3.88498783e-01 2.14207187e-01 -4.20518160e-01
-3.40428203e-01 -5.79430640e-01 -1.28837442e+00 -2.77665779e-02
6.01681210e-02 3.16950202e-01 6.29153132e-01 8.59787166e-01
7.20407605e-01 -3.67931426e-02 2.46905193e-01 1.11068971e-01
-9.30455029e-01 -7.50908852e-01 -4.66022581e-01 5.31624734e-01
3.55611384e-01 -2.76084334e-01 -2.57948041e-01 4.84886706e-01] | [10.38086986541748, 3.158299684524536] |
9d042afb-aa98-41e0-94ef-0ac79c0ff547 | affinity-aware-compression-and-expansion | 2008.10191 | null | https://arxiv.org/abs/2008.10191v1 | https://arxiv.org/pdf/2008.10191v1.pdf | Affinity-aware Compression and Expansion Network for Human Parsing | As a fine-grained segmentation task, human parsing is still faced with two challenges: inter-part indistinction and intra-part inconsistency, due to the ambiguous definitions and confusing relationships between similar human parts. To tackle these two problems, this paper proposes a novel \textit{Affinity-aware Compression and Expansion} Network (ACENet), which mainly consists of two modules: Local Compression Module (LCM) and Global Expansion Module (GEM). Specifically, LCM compresses parts-correlation information through structural skeleton points, obtained from an extra skeleton branch. It can decrease the inter-part interference, and strengthen structural relationships between ambiguous parts. Furthermore, GEM expands semantic information of each part into a complete piece by incorporating the spatial affinity with boundary guidance, which can effectively enhance the semantic consistency of intra-part as well. ACENet achieves new state-of-the-art performance on the challenging LIP and Pascal-Person-Part datasets. In particular, 58.1% mean IoU is achieved on the LIP benchmark. | ['Pengfei Xiong', 'Yunfeng Wang', 'Xinyan Zhang'] | 2020-08-24 | null | null | null | null | ['human-parsing'] | ['computer-vision'] | [ 3.30428064e-01 2.55154818e-01 -3.28576148e-01 -4.94252652e-01
-5.98121464e-01 -3.78625929e-01 -1.85547508e-02 3.14214230e-02
-3.49361092e-01 4.63837951e-01 4.78644818e-01 2.00622484e-01
5.61912842e-02 -5.66281378e-01 -6.12023592e-01 -5.35265148e-01
5.49939632e-01 5.76813459e-01 7.14387715e-01 -3.88572030e-02
1.77128576e-02 2.09058538e-01 -1.61300790e+00 4.08162683e-01
1.22868764e+00 1.08092797e+00 3.41150045e-01 2.61662751e-01
-7.50050724e-01 1.81988284e-01 -5.44109702e-01 -7.37750173e-01
-3.28239426e-02 -3.30739915e-01 -1.03549778e+00 1.04679905e-01
5.67498505e-01 -2.39946380e-01 -1.22210711e-01 1.41200078e+00
7.49458551e-01 -2.43484564e-02 2.29667693e-01 -1.11586392e+00
-3.17201346e-01 8.64581466e-01 -1.10915017e+00 -3.98513228e-02
9.50910747e-02 -2.03025267e-02 9.57344174e-01 -4.22797978e-01
5.61814070e-01 1.61766720e+00 8.68246078e-01 7.40444124e-01
-8.27455640e-01 -6.24730229e-01 3.76687706e-01 3.49401325e-01
-1.14934659e+00 -3.13740462e-01 7.85315633e-01 1.00418873e-01
6.81744695e-01 2.53271043e-01 7.43550181e-01 7.76129305e-01
-1.81144923e-01 1.32958984e+00 7.80301809e-01 -1.11910678e-01
-4.50587608e-02 -5.06097317e-01 3.42708886e-01 7.13565707e-01
3.36894155e-01 -2.92447180e-01 -4.65975195e-01 1.86017528e-01
5.90958774e-01 -8.24337900e-02 -3.79380405e-01 -3.17555159e-01
-8.04356396e-01 4.15182203e-01 4.63847607e-01 2.99558729e-01
-1.56695947e-01 1.10282768e-02 6.86147332e-01 -4.24912632e-01
1.16275951e-01 -1.55733675e-01 -7.97427177e-01 -3.16002756e-01
-1.04190862e+00 4.06598765e-03 5.70084751e-01 1.20881915e+00
6.15467191e-01 -5.37305355e-01 -2.47907072e-01 1.27337384e+00
4.26180929e-01 2.76515007e-01 6.98312581e-01 -1.20007706e+00
7.08049417e-01 7.89123833e-01 -6.06620371e-01 -8.22046280e-01
-4.67633754e-01 -4.57716703e-01 -1.07636535e+00 -3.66211504e-01
3.94278914e-01 2.15182211e-02 -1.15024376e+00 1.90585637e+00
7.20060408e-01 3.22865648e-03 -1.68425098e-01 9.48976040e-01
1.29928517e+00 3.52480829e-01 4.20662045e-01 -1.02492191e-01
1.78822720e+00 -1.49459243e+00 -8.65141332e-01 -4.29294854e-01
2.48249084e-01 -9.44723070e-01 1.04547524e+00 2.51734138e-01
-1.28175187e+00 -8.74078751e-01 -8.66925597e-01 -6.62681401e-01
-1.25370890e-01 1.88947096e-01 5.97867131e-01 3.66623640e-01
-7.01042950e-01 4.98025239e-01 -6.62061810e-01 -1.70872539e-01
7.26950228e-01 2.89575577e-01 -2.57853329e-01 -3.42722863e-01
-1.01776636e+00 2.08239123e-01 6.29409552e-01 2.28576586e-01
-1.18637932e-02 -7.01514900e-01 -1.06647837e+00 1.36465505e-01
6.53094947e-01 -8.68072629e-01 1.19984567e+00 -9.26011086e-01
-1.28217542e+00 9.31577682e-01 -2.49105588e-01 -1.35703236e-01
6.22902751e-01 -3.24793369e-01 -2.11523429e-01 3.61737341e-01
3.33385527e-01 1.25783646e+00 5.97991705e-01 -1.05192423e+00
-7.63691783e-01 -7.86724269e-01 -4.67609286e-01 4.42835718e-01
9.52627212e-02 -1.91401899e-01 -1.44355536e+00 -1.00930738e+00
6.78624272e-01 -6.67175889e-01 6.41240999e-02 1.50684834e-01
-6.65758252e-01 -4.00152922e-01 9.24617410e-01 -1.04384434e+00
1.33061886e+00 -2.20682573e+00 1.73332721e-01 -5.88422641e-04
1.66023791e-01 4.39588606e-01 -2.68772095e-01 -2.36873984e-01
1.79652333e-01 2.26158872e-01 -5.47726154e-01 -6.18106723e-01
2.45145839e-02 4.83339250e-01 2.85058588e-01 8.64320062e-03
-2.58380901e-02 1.06117201e+00 -6.43359542e-01 -1.18522859e+00
8.50818232e-02 5.76561391e-01 -4.20639336e-01 -6.69043511e-02
-2.41936252e-01 1.79554194e-01 -4.63499546e-01 1.04643476e+00
1.29783547e+00 -1.08307779e-01 1.81171864e-01 -5.55288017e-01
1.54038265e-01 7.43014589e-02 -1.29389191e+00 2.34301257e+00
8.17281008e-02 3.74092646e-02 3.58708203e-01 -6.70583189e-01
6.72332644e-01 -1.64018586e-01 4.74837184e-01 -1.00687242e+00
1.51636988e-01 5.68100512e-02 -1.36186942e-01 -4.35040683e-01
4.13392156e-01 1.51943684e-01 -1.77255925e-02 8.00317600e-02
7.61049315e-02 9.88583267e-02 2.18404710e-01 1.37303352e-01
8.44720840e-01 4.04446155e-01 1.45811602e-01 -2.91212231e-01
8.06869566e-01 -3.82177711e-01 1.11220944e+00 1.53911352e-01
-6.27985775e-01 7.39351034e-01 5.21976769e-01 -1.77982479e-01
-8.03052425e-01 -1.03694844e+00 -2.12279841e-01 9.36109960e-01
7.93089032e-01 -5.65955341e-01 -1.23821497e+00 -8.63501310e-01
-1.85169965e-01 2.36980870e-01 -5.39788544e-01 -1.02402255e-01
-8.38440180e-01 -5.90112090e-01 6.63760900e-01 9.42743063e-01
1.25594676e+00 -1.19667757e+00 -4.01978284e-01 1.74161211e-01
-7.38424599e-01 -1.23524749e+00 -9.92738247e-01 -1.03146665e-01
-8.82313907e-01 -1.21649265e+00 -9.25760448e-01 -9.66333807e-01
5.77627003e-01 1.63611799e-01 1.19970751e+00 3.17749947e-01
-4.20073360e-01 -1.58680320e-01 -3.60509872e-01 -1.18198343e-01
-6.78899139e-02 1.04688458e-01 -4.89997268e-01 -3.36481661e-01
3.79001588e-01 -3.69504243e-01 -9.35456336e-01 6.24792278e-01
-7.79306591e-01 2.80603915e-01 8.85535300e-01 6.97971225e-01
1.27502000e+00 8.93671289e-02 9.07152891e-02 -6.59616768e-01
2.35196337e-01 6.57941550e-02 -1.50819480e-01 4.61465478e-01
-4.63152826e-01 8.18776041e-02 3.24996293e-01 -2.23754793e-01
-1.26912189e+00 1.47666723e-01 -4.88903284e-01 -2.51772076e-01
-3.28999907e-01 -1.06685810e-01 -1.01792276e+00 1.68325976e-01
-8.78275931e-02 3.42318535e-01 -1.90175436e-02 -8.78535330e-01
3.60923976e-01 6.72645450e-01 1.04300725e+00 -8.17230463e-01
1.33137807e-01 2.53436089e-01 2.45970022e-02 -5.65685689e-01
-8.40905011e-01 -7.26873994e-01 -8.83013070e-01 7.66246617e-02
1.08890855e+00 -8.38858664e-01 -6.42788768e-01 8.11820447e-01
-1.25589764e+00 -1.79580897e-01 -3.22221816e-01 5.01701832e-02
-3.33867759e-01 9.98954475e-01 -9.68319178e-01 -2.62060583e-01
-7.42449284e-01 -1.17338645e+00 1.39181209e+00 7.36095905e-01
7.93210045e-02 -5.36113262e-01 -1.34591267e-01 9.19259012e-01
-2.12746710e-01 2.31581628e-02 8.53336453e-01 -4.01867002e-01
-3.59528005e-01 2.31100038e-01 -6.26628041e-01 5.36635041e-01
-4.73975241e-02 -2.99125407e-02 -7.65172541e-01 -2.65539205e-03
-3.35413605e-01 -2.34441251e-01 1.00425971e+00 4.95604753e-01
1.58401620e+00 -2.47226208e-01 -5.21218300e-01 8.89975965e-01
1.05178916e+00 -4.08001915e-02 7.69294679e-01 9.30448100e-02
8.20955098e-01 6.30794048e-01 7.50224650e-01 2.30739325e-01
5.51609159e-01 6.08270109e-01 3.49603295e-01 -7.62329996e-02
-6.44626498e-01 -5.40399551e-01 -9.81937647e-02 7.25057960e-01
-1.93519215e-03 -1.36384591e-01 -4.94405031e-01 3.33714515e-01
-1.86464489e+00 -8.28314543e-01 -1.02665842e-01 1.87059832e+00
9.94071305e-01 2.62045383e-01 2.16334149e-01 1.23639442e-01
1.07626987e+00 1.05718419e-01 -7.68411756e-01 -9.52752903e-02
-2.89556652e-01 2.63767242e-01 3.37085992e-01 2.64525950e-01
-1.12693226e+00 1.08529246e+00 5.15885878e+00 1.33844256e+00
-5.48038781e-01 1.59558684e-01 9.41981494e-01 3.41696948e-01
-1.55816913e-01 -3.44531208e-01 -1.09544170e+00 7.37704277e-01
9.98996049e-02 5.64300835e-01 8.30531791e-02 8.83153081e-01
-2.21681356e-01 -2.58190274e-01 -6.96550369e-01 1.17353356e+00
1.69863865e-01 -9.50173676e-01 1.55501068e-01 -9.20445696e-02
4.62495506e-01 -2.81400234e-01 -2.19580650e-01 1.89121038e-01
-3.34218919e-01 -8.08570743e-01 7.77507424e-01 3.05787683e-01
9.11637068e-01 -1.10283482e+00 9.34537649e-01 3.30400616e-01
-1.78116989e+00 2.44598135e-01 -4.37457025e-01 6.09612823e-01
2.37604335e-01 5.76402128e-01 -1.63637903e-02 6.95357621e-01
9.76371169e-01 5.39607644e-01 -5.24531484e-01 1.05850101e+00
-4.57602799e-01 2.35496089e-01 -5.46203136e-01 1.68359593e-01
1.72952965e-01 -2.42865011e-01 4.01757479e-01 1.39391470e+00
-9.23248231e-02 2.75823116e-01 1.59722000e-01 6.81606829e-01
-3.56479168e-01 6.96017519e-02 3.53778720e-01 3.84292901e-01
5.67882001e-01 1.48670006e+00 -8.82158101e-01 -3.57182473e-01
-2.96540439e-01 1.39101660e+00 2.64512002e-01 -7.58536011e-02
-9.37187076e-01 -4.49660391e-01 6.75775647e-01 -1.44147212e-02
4.76435840e-01 1.66751876e-01 -4.48309749e-01 -1.00842595e+00
3.80277991e-01 -9.09245014e-01 6.71565235e-01 -7.05322564e-01
-1.04550827e+00 3.92149389e-01 -2.02882037e-01 -7.64837861e-01
2.79072315e-01 -2.55388469e-01 -5.13105512e-01 6.62874162e-01
-1.51359999e+00 -1.35565352e+00 -6.14880085e-01 6.76882148e-01
8.22184980e-01 3.10644001e-01 4.02826995e-01 4.89035547e-01
-7.25332499e-01 1.11661780e+00 -4.85033661e-01 3.24042231e-01
6.18445396e-01 -1.17029142e+00 3.83829027e-01 7.69598365e-01
-8.60327035e-02 4.55842644e-01 1.55707181e-01 -9.55800831e-01
-7.86589742e-01 -1.08054888e+00 7.27974713e-01 2.21030593e-01
6.54952601e-02 -6.80245161e-02 -8.64620805e-01 2.98804998e-01
7.54622137e-03 1.02543771e-01 4.68760759e-01 -4.52890173e-02
-4.84348923e-01 -7.69437626e-02 -1.38645339e+00 4.68302220e-01
1.60845459e+00 -2.22718462e-01 -6.65300131e-01 8.14476758e-02
1.05216444e+00 -6.06160939e-01 -7.65814364e-01 7.80953407e-01
5.57655275e-01 -1.22713733e+00 1.06345582e+00 -2.57800788e-01
4.16040927e-01 -3.61078322e-01 -8.73436332e-02 -7.05020368e-01
-2.79498249e-01 -6.81330025e-01 -2.53401577e-01 1.53554773e+00
1.27275035e-01 -1.79045409e-01 1.04482663e+00 3.69848222e-01
-4.96959716e-01 -1.06781065e+00 -1.03170753e+00 -5.36571562e-01
-5.48613295e-02 -4.06177193e-01 8.17196965e-01 5.01712203e-01
-2.34453127e-01 3.46611917e-01 -3.86704095e-02 -5.13158664e-02
6.66438282e-01 1.52113214e-01 5.15241563e-01 -9.95881140e-01
-4.24665421e-01 -6.87144876e-01 -3.35593104e-01 -1.64794624e+00
9.39791128e-02 -6.04069948e-01 2.79698789e-01 -1.58623147e+00
5.86917162e-01 -4.91745472e-01 -4.13550176e-02 5.92249393e-01
-4.55264568e-01 2.14565694e-01 3.51671606e-01 1.52601212e-01
-8.41630101e-01 4.88411516e-01 1.73846316e+00 -1.66684344e-01
-1.97505489e-01 3.61293815e-02 -6.75761998e-01 1.07570338e+00
6.47675872e-01 -1.14946485e-01 -2.62804896e-01 -4.65962738e-01
-2.54966408e-01 -6.50853440e-02 2.78617829e-01 -1.15937889e+00
3.16232592e-01 1.25610471e-01 5.59230030e-01 -1.23517072e+00
2.86294639e-01 -6.07151926e-01 -1.02375180e-01 5.47197819e-01
-1.86725147e-02 5.24395965e-02 2.84064561e-01 5.96153915e-01
-3.15085620e-01 -2.52783179e-01 9.49689388e-01 -2.63182789e-01
-1.00576699e+00 6.23798668e-01 5.59052825e-01 5.17601073e-01
8.36638570e-01 -2.98176706e-01 -3.34127069e-01 2.39392772e-01
-7.58982003e-01 4.07468766e-01 3.38524282e-01 2.93830305e-01
6.06516361e-01 -1.15418410e+00 -4.43463653e-01 2.80959308e-01
-1.41938135e-01 6.60176039e-01 7.97628820e-01 6.85781300e-01
-5.72557271e-01 2.14774698e-01 -1.77733988e-01 -7.16379166e-01
-1.72588694e+00 4.14225668e-01 2.90541917e-01 -3.02452594e-01
-7.05842555e-01 1.25264204e+00 2.38830358e-01 -3.36298883e-01
4.55198705e-01 -3.70512336e-01 -5.42993993e-02 4.70215417e-02
5.03970861e-01 5.14349878e-01 -1.66328326e-01 -8.13540041e-01
-5.03701448e-01 1.25728047e+00 -1.37451962e-01 1.14527740e-01
8.62724245e-01 -2.50679970e-01 -2.76667774e-01 -2.79998481e-01
1.16754460e+00 6.52752519e-02 -1.35109639e+00 -1.96420029e-01
-3.86840105e-01 -2.63552636e-01 -1.83285013e-01 -1.15850985e+00
-1.37490177e+00 9.46316600e-01 5.53736925e-01 -3.47086817e-01
1.28814077e+00 3.14682662e-01 1.57339168e+00 -4.14246786e-03
8.98541734e-02 -1.41966629e+00 5.22602424e-02 3.84488463e-01
6.83150291e-01 -9.99601126e-01 7.73093775e-02 -9.82986093e-01
-6.06440842e-01 1.00793827e+00 8.54079604e-01 3.82503629e-01
5.72939992e-01 2.41787061e-01 -2.21611753e-01 8.95203277e-02
-1.10014245e-01 -3.20502371e-01 5.34555078e-01 5.32612860e-01
2.89387137e-01 7.12564066e-02 -8.09108615e-01 9.68934298e-01
-2.16362357e-01 -2.07431674e-01 -2.83460766e-01 6.95333302e-01
-5.06747067e-01 -1.23189878e+00 -1.78309485e-01 2.20240429e-01
-5.49544096e-01 -1.55708343e-01 -4.45454240e-01 9.00391102e-01
8.82456899e-01 8.20680201e-01 2.30965167e-01 -3.62120152e-01
3.06613177e-01 -9.82152019e-03 4.81631488e-01 -2.13297665e-01
-3.99676979e-01 5.26972890e-01 -3.45100425e-02 -8.81528616e-01
-4.00435805e-01 -8.35102499e-01 -1.83994591e+00 -3.32383603e-01
-2.25335687e-01 6.67499155e-02 5.76485872e-01 8.85342479e-01
4.92579252e-01 6.86348379e-01 2.68649571e-02 -4.81710672e-01
-4.48541522e-01 -9.54053760e-01 -4.44537491e-01 6.38128579e-01
-1.58175677e-01 -4.33197618e-01 -8.62904489e-02 -6.69947118e-02] | [8.884822845458984, 0.032328050583601] |
27a9adc9-eedf-48ed-894d-4f7591b1f620 | tiefake-title-text-similarity-and-emotion | 2304.09421 | null | https://arxiv.org/abs/2304.09421v1 | https://arxiv.org/pdf/2304.09421v1.pdf | TieFake: Title-Text Similarity and Emotion-Aware Fake News Detection | Fake news detection aims to detect fake news widely spreading on social media platforms, which can negatively influence the public and the government. Many approaches have been developed to exploit relevant information from news images, text, or videos. However, these methods may suffer from the following limitations: (1) ignore the inherent emotional information of the news, which could be beneficial since it contains the subjective intentions of the authors; (2) pay little attention to the relation (similarity) between the title and textual information in news articles, which often use irrelevant title to attract reader' attention. To this end, we propose a novel Title-Text similarity and emotion-aware Fake news detection (TieFake) method by jointly modeling the multi-modal context information and the author sentiment in a unified framework. Specifically, we respectively employ BERT and ResNeSt to learn the representations for text and images, and utilize publisher emotion extractor to capture the author's subjective emotion in the news content. We also propose a scale-dot product attention mechanism to capture the similarity between title features and textual features. Experiments are conducted on two publicly available multi-modal datasets, and the results demonstrate that our proposed method can significantly improve the performance of fake news detection. Our code is available at https://github.com/UESTC-GQJ/TieFake. | ['Zhouguo Chen', 'Ling Tian', 'Zhao Kang', 'Quanjiang Guo'] | 2023-04-19 | null | null | null | null | ['fake-news-detection'] | ['natural-language-processing'] | [-0.29155734 -0.36409757 -0.406553 -0.15857275 -0.7026211 -0.34732613
0.708653 0.03691232 -0.28321293 0.3874972 0.6339349 0.14087166
0.5142661 -0.42868313 -0.664934 -0.5407293 0.5076869 -0.36954352
0.07032479 -0.47545505 0.60475296 -0.02064411 -1.2793458 0.5383605
0.9063099 1.3131981 -0.0252277 0.12238649 -0.15021193 1.1628324
-0.758055 -0.6656755 -0.09385807 -0.60248643 -0.28470603 0.35123277
0.26762947 -0.46264324 -0.79306775 1.5411397 0.3010452 -0.0774936
0.393897 -1.2873648 -1.2324771 0.38819444 -0.9158876 0.5069071
0.2606919 0.03604486 0.933738 -1.207853 0.65562093 1.3338231
0.36047754 0.36665592 -0.42729953 -0.7342345 0.3459529 0.42417976
-1.3770313 -0.48767585 1.1612715 -0.44022503 0.38952884 0.3255977
0.83888113 1.5823357 0.39222553 1.2305768 1.2248139 -0.0750157
-0.20097293 0.48193172 -0.02571181 0.73435354 0.10271376 -0.1328861
-0.4986256 -0.3182882 0.47241703 0.35150847 -0.65444803 0.07314505
-1.2500806 1.2445445 0.5255396 0.35935715 -0.31605855 0.06847261
0.74693006 0.5191859 1.0431705 0.37443835 -0.0804196 -0.07033818
-0.5927008 0.1598852 0.566787 0.8618575 0.53090084 -0.1305351
-0.16679502 0.7697775 0.43629292 0.77698344 0.9014442 -0.4323561
0.6048911 0.5217242 0.40633422 -2.025509 0.05945653 -0.5647992
-0.52550966 -0.5740173 -0.18791673 -0.14523867 -0.4950722 1.3612909
0.23172371 0.3181636 -0.03957175 1.2780964 1.3068434 1.0311176
-0.24486275 -0.45470572 1.6242926 -1.2622846 -1.2599812 -0.5724161
0.62602097 -1.0913787 1.1130662 0.12143696 -0.73221385 -0.07366922
-1.0352476 -0.22247736 -0.60196143 0.39464182 0.32221457 0.3534157
-0.2813439 0.19158597 -0.41384396 0.03451955 0.7406039 -0.24255663
0.04348217 -0.13833337 -1.6269443 0.7545013 0.18499719 0.30069646
-0.44958866 -0.00516803 -0.7187414 -0.08471882 0.7288808 -0.2991549
1.0777434 -1.7069508 -1.3068827 0.7480133 -0.09914123 0.04646176
0.63923043 -0.33234826 -0.87048376 0.27404732 0.24397317 0.18893042
1.1909672 -1.1623038 -0.5424028 -0.22667873 -0.12610434 0.34031758
-0.6100015 0.3606963 -0.7302539 -1.1552814 0.12213606 -0.7852838
0.24318141 -0.13307077 -0.5308767 -0.05530018 1.129445 -0.9153155
1.1713918 -2.419549 -0.05861744 -0.22468235 0.37374344 0.1656522
-0.07384376 0.31607738 0.2611964 0.08467834 0.20775214 -0.2249966
-0.06838937 -0.09031161 -0.60388887 0.86472285 0.00792605 1.1075318
-1.0504414 -0.53775007 -0.19506754 0.42096177 -0.2528025 0.01452885
-0.07152952 0.18837102 -0.9400172 0.70174956 0.6648272 -0.40290925
-0.39962095 -0.4889196 -0.09200309 0.42547333 -0.598088 1.0727848
-0.18424457 0.9335857 0.06867697 -0.93105197 0.69672984 0.30822924
0.2189184 -0.84494734 0.6639265 0.29847053 -0.38092318 -0.831867
0.6059716 -0.22828676 -0.3025578 0.45779464 -0.28041285 0.07328024
-0.25996116 0.4436927 0.57311916 -0.35258737 0.16744508 0.04447532
0.40294817 -0.04984654 0.6672522 0.5847864 -0.5567974 0.27691
0.7068984 -0.3425317 -0.68565696 -0.37274134 0.14539048 0.9787213
0.8064957 -0.3339336 -0.3887793 -1.0458666 -0.11075744 0.5448052
-0.7242885 -0.4302104 -0.29269513 -0.84273434 0.24959058 -0.04518325
0.65081036 -0.93582827 -0.09862302 0.06906798 -0.78530616 -1.2760653
-0.99245566 -0.5698144 -0.36526325 -0.79930407 -0.67689884 -0.7610032
0.6484258 0.767301 0.6838647 0.27333978 0.099601 0.23264778
-0.6531129 -0.36776534 -0.21202947 -0.27921954 -0.03489437 0.45221263
0.5433217 0.04329703 -0.57948726 0.34325102 -0.9650111 0.21661112
0.5371826 1.0832987 0.3224359 0.07672763 0.6913308 -0.8965911
0.84019244 -0.94830644 -0.30702314 -0.04003195 -0.50906324 -0.47590002
0.5964657 -0.68738306 -0.98025787 -0.43066257 0.0421086 -0.672127
0.21972397 0.8639531 -0.09991671 0.02009974 0.21174857 0.4963647
-0.13817398 -0.2098921 0.07978135 1.0645988 0.2872883 -0.06460953
0.7307229 0.72771657 -0.68779683 -0.62913203 -1.4062955 -0.5863353
0.1671761 -0.30179533 0.44801357 -1.1902822 -0.37186667 0.71114594
-1.3025577 0.39615628 0.27400804 0.69937754 -0.07414073 0.78591704
-1.0887984 -0.5991426 -0.19445728 -1.168298 0.8718007 0.1972073
0.28476664 -0.8304204 -0.37667802 0.7336332 0.2583918 0.19825447
0.34748912 -0.7947293 -0.4400609 -0.5435026 -0.6467266 0.39146078
0.1471261 -0.12033417 -0.7106158 -0.2543762 0.3956179 -0.45319596
0.96998096 0.05539047 0.9277507 -0.9550519 -0.23949754 0.28201857
1.1068553 -0.12606463 0.51087606 0.36510557 0.9966422 0.61135244
0.9700185 0.6176833 0.45701185 0.63600856 0.5666522 0.0165873
0.2687682 -0.2959952 0.7459732 1.233534 0.2293605 -0.54157406
-0.46357962 0.5929381 -2.0299857 -1.1531719 -0.35240725 1.5776724
0.6421665 0.16735068 -0.09529873 -0.32114062 0.99740636 0.57365483
-0.4426041 -0.1635568 -0.4218625 -0.76830417 0.53807396 0.21938689
-1.2847472 1.0515813 4.505124 1.139113 -1.4956164 0.50281745
0.70039415 -0.28948668 -0.4094637 -0.2401637 -0.41413647 1.0663815
0.49913448 -0.17116125 0.37945282 0.93637943 0.4769011 0.23251265
-0.3566231 1.0983481 0.5769852 -1.3233799 -0.07379473 0.08210585
0.8348805 0.157865 0.42678437 0.39675188 -0.1856054 -0.6368387
0.9846578 0.46442857 0.37177184 -0.84386426 1.0819165 0.45303982
-0.73722076 -0.04763572 -0.3800812 -0.09711906 0.09215566 0.76948655
-0.4038681 0.28882492 0.6560163 1.2056834 -0.5386637 0.5907425
-0.32741407 0.5387349 -0.02484299 -0.4108964 0.5127355 -0.10600968
0.7597157 1.1482862 0.34497964 0.13510616 0.154751 0.75766295
-0.4695484 0.5169844 -0.3616568 -0.4650381 0.3788445 1.1477156
-0.66610754 -0.6047102 -0.81746805 1.2335123 0.19199498 0.24939741
-1.1429814 -0.3581772 0.34617805 -0.1219341 0.4701796 0.2822282
0.15710574 -1.924078 0.24542275 -1.015171 0.24529916 -0.9976417
-1.7245902 0.5099828 -0.5148169 -1.445227 0.4050399 -0.34045008
-0.47440466 0.5466726 -1.5752908 -1.1839585 -0.2818344 0.507872
0.58485395 -0.04455034 0.22405379 0.33032778 -0.7584634 0.45257196
0.3991422 0.3581348 0.8746497 -0.68835753 0.04464006 0.79472166
-0.05593983 0.46949658 0.85989034 -0.8330721 -1.3790171 -1.0936098
0.94701463 -0.30580956 0.9455379 -0.14889257 -0.9100746 0.63318545
0.24972963 0.27275226 0.63120085 -0.2903728 -0.6324439 0.275047
-1.0585455 0.73489004 0.5860763 -0.6825282 -0.6548474 0.5801301
0.63822037 -0.41974553 -0.48289946 -0.05055539 0.5622165 -0.9116277
0.6590053 -0.4658366 0.9787404 -0.22653727 0.02525398 -1.51744
-0.27402914 -0.36637577 -0.2694871 1.0057435 0.16720326 -0.80210924
0.33714736 0.1268456 -0.14431027 -0.8337953 -0.79975677 -0.43155867
-0.29951432 -0.17187566 0.47135335 1.5806253 0.23963548 0.49070767
-1.0165964 0.28166512 0.19686912 0.44910282 0.5085807 -0.8108834
-0.18327457 -0.554374 -0.28945956 -1.3100837 0.25511038 -0.47952932
-0.02445081 -1.120037 0.41130197 0.06914727 -0.32797793 0.15963538
-0.40060112 0.4949409 -0.0566018 0.7107265 -0.917852 0.8683824
1.7108377 -0.28040692 0.19237602 -0.25552908 -0.9697707 0.9050045
0.5905703 -0.7581682 -0.04412133 -0.3362999 0.59473056 0.07973676
0.6076274 -0.3190923 0.03915666 -0.36443526 0.27080488 -0.4772178
0.43740866 -0.6731163 -0.39065924 0.396262 -0.29826167 0.03847863
-0.18767415 1.0662924 -0.49291348 -0.21881483 0.70827985 -0.22061336
-0.5180737 0.32874045 -0.524969 0.14059189 0.8136579 0.13698238
-0.8573041 -0.8542364 -0.36412948 0.1544017 0.586402 0.6921588
0.76317424 -1.5843408 -0.8940139 -0.18327315 0.47712868 -0.56124425
0.47059852 1.153619 -0.34646672 0.18263556 0.11489242 -0.05020152
-1.0688677 0.8956277 0.10725257 0.09288365 -0.3897507 0.83570075
0.30420038 0.03315682 -0.21768637 0.01040319 -0.35015157 0.29226032
0.6466558 0.11372192 -0.24608749 -1.3339424 -0.28109747 0.12276249
-0.48467577 0.17459202 1.1906614 -0.660848 -0.15263912 0.47756946
1.4632941 0.5009768 -0.8527908 -0.56742346 -0.2380012 -0.8567666
0.6885921 -0.6153866 -1.3758396 0.57734364 0.2698119 0.42463082
0.8792769 0.16345726 1.3764567 0.26878104 -0.13166809 -1.3234528
0.4417145 0.24157062 1.045348 -1.6820158 0.14058992 -0.5504001
-1.001571 1.0099963 0.5487151 -0.22428362 0.54307926 -0.34505764
0.37977275 -0.56873155 -0.6078459 0.06458198 0.41472122 -0.0333252
0.28132793 -0.02561947 -0.5494075 0.9277413 0.10949584 -0.25920013
0.8194023 0.8402295 -0.36797926 -0.48074424 -0.38904378 0.5285901
-0.92744595 -0.23363765 -0.6374412 0.49460176 0.14196926 0.9135694
-0.1387949 -0.32654178 0.01806879 -0.3609619 -0.08772142 -0.3856142
-0.435068 0.46795928 0.15310562 -0.40372112 -0.53822184 -0.54533094
-0.73876256 -0.37781945 -0.8150152 0.29480192 0.55157006 1.1207247
0.54563797 0.2973555 1.0003798 -0.55531204 -0.43536845 -1.0070082
-0.58194244 0.57029223 0.57174796 -0.7892809 -0.85401756 -0.07895016] | [8.163406372070312, 10.288440704345703] |
c60d1b58-95e0-41dd-8ca4-64a15a72001d | sg-net-spatial-granularity-network-for-one | 2103.10284 | null | https://arxiv.org/abs/2103.10284v2 | https://arxiv.org/pdf/2103.10284v2.pdf | SG-Net: Spatial Granularity Network for One-Stage Video Instance Segmentation | Video instance segmentation (VIS) is a new and critical task in computer vision. To date, top-performing VIS methods extend the two-stage Mask R-CNN by adding a tracking branch, leaving plenty of room for improvement. In contrast, we approach the VIS task from a new perspective and propose a one-stage spatial granularity network (SG-Net). Compared to the conventional two-stage methods, SG-Net demonstrates four advantages: 1) Our method has a one-stage compact architecture and each task head (detection, segmentation, and tracking) is crafted interdependently so they can effectively share features and enjoy the joint optimization; 2) Our mask prediction is dynamically performed on the sub-regions of each detected instance, leading to high-quality masks of fine granularity; 3) Each of our task predictions avoids using expensive proposal-based RoI features, resulting in much reduced runtime complexity per instance; 4) Our tracking head models objects centerness movements for tracking, which effectively enhances the tracking robustness to different object appearances. In evaluation, we present state-of-the-art comparisons on the YouTube-VIS dataset. Extensive experiments demonstrate that our compact one-stage method can achieve improved performance in both accuracy and inference speed. We hope our SG-Net could serve as a strong and flexible baseline for the VIS task. Our code will be available. | ['Yingjie Chen', 'Wenbo Tan', 'Yiming Cui', 'Dongfang Liu'] | 2021-03-18 | null | http://openaccess.thecvf.com//content/CVPR2021/html/Liu_SG-Net_Spatial_Granularity_Network_for_One-Stage_Video_Instance_Segmentation_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Liu_SG-Net_Spatial_Granularity_Network_for_One-Stage_Video_Instance_Segmentation_CVPR_2021_paper.pdf | cvpr-2021-1 | ['head-detection', 'video-instance-segmentation'] | ['computer-vision', 'computer-vision'] | [-2.45381743e-02 6.66377619e-02 -4.97635692e-01 -2.24959403e-01
-7.62682140e-01 -4.41236407e-01 3.64224911e-01 -3.05870086e-01
-4.99494165e-01 3.20666462e-01 2.89310347e-02 -4.64357771e-02
4.35090959e-01 -3.38615417e-01 -9.29425597e-01 -5.50205767e-01
2.99737528e-02 4.06846076e-01 1.20702970e+00 1.71203390e-01
-1.50895312e-01 5.69481254e-01 -1.43731320e+00 1.88023597e-01
6.20893538e-01 1.25116026e+00 3.93451512e-01 5.17078698e-01
2.10778043e-01 5.68997562e-01 -4.21209484e-01 -6.00913346e-01
6.47417188e-01 -1.97807723e-03 -5.46718478e-01 2.62443870e-01
1.09567153e+00 -4.85911667e-01 -5.68207800e-01 1.01446295e+00
4.78053451e-01 6.29862621e-02 3.44889343e-01 -1.18595827e+00
-3.65718573e-01 5.70433378e-01 -9.60387766e-01 3.30424219e-01
-2.14963973e-01 5.57499230e-01 1.02489030e+00 -8.29154432e-01
5.44671416e-01 1.10992754e+00 9.29276407e-01 6.60665214e-01
-1.23981285e+00 -7.42303550e-01 7.47389793e-01 3.97221074e-02
-1.30072439e+00 -5.17505527e-01 4.29513931e-01 -4.85104918e-01
6.70427561e-01 2.40685821e-01 8.76208603e-01 9.91713583e-01
1.32579803e-01 1.40932989e+00 6.90319717e-01 9.12410319e-02
-5.09834327e-02 6.62183091e-02 1.44011766e-01 9.94279265e-01
2.91120470e-01 1.69796541e-01 -3.23383480e-01 2.23664671e-01
9.02954698e-01 2.19164133e-01 -3.35285783e-01 -5.60496867e-01
-1.35361350e+00 4.29049581e-01 5.65465212e-01 1.40116945e-01
-1.52790323e-01 5.18028259e-01 4.41706985e-01 -1.16419449e-01
5.96869111e-01 8.16764534e-02 -5.63167930e-01 -1.26500893e-02
-1.54732203e+00 2.11519480e-01 5.11590660e-01 1.17219388e+00
4.16593611e-01 8.62714276e-02 -7.25223541e-01 5.81034958e-01
3.11336458e-01 4.30550396e-01 1.54018134e-01 -9.80582774e-01
3.17596227e-01 5.02953589e-01 2.19170615e-01 -6.81477964e-01
-5.52410483e-01 -8.45812559e-01 -4.68296051e-01 2.88359642e-01
5.29579282e-01 -1.55608580e-01 -1.14084983e+00 1.62413168e+00
5.78058183e-01 6.99106693e-01 -7.26925313e-01 9.86294746e-01
1.03022957e+00 3.02386254e-01 1.12914190e-01 2.27496829e-02
1.41540289e+00 -1.59008896e+00 -5.49898505e-01 -3.49642605e-01
2.60631680e-01 -5.94412982e-01 8.19307506e-01 3.13941091e-01
-1.40894163e+00 -8.50876749e-01 -8.72995615e-01 -1.32035971e-01
-1.47646010e-01 3.21996808e-01 5.98335803e-01 6.77885950e-01
-1.20305848e+00 3.53972763e-01 -1.13138247e+00 -4.36014123e-02
9.64115262e-01 5.30575454e-01 -9.30911824e-02 5.57370521e-02
-6.22233331e-01 6.54416502e-01 2.67031193e-01 2.24280387e-01
-1.05027926e+00 -8.76264572e-01 -8.65573227e-01 1.48328751e-01
8.33984375e-01 -6.53680682e-01 1.27212846e+00 -7.68316686e-01
-1.21239030e+00 8.83037269e-01 -4.08841759e-01 -4.08997118e-01
9.14689660e-01 -2.67324924e-01 -8.63783509e-02 4.11158912e-02
1.89184815e-01 1.03257942e+00 9.64830339e-01 -1.24813950e+00
-9.98495936e-01 -3.08899313e-01 1.83464196e-02 -5.28772213e-02
-2.27782860e-01 1.02279671e-01 -1.39362335e+00 -9.27527428e-01
-9.17537659e-02 -9.96352196e-01 -3.34118843e-01 4.25212264e-01
-4.05878276e-01 -1.83801159e-01 9.18258250e-01 -5.38909912e-01
1.40284836e+00 -2.19448495e+00 1.27830490e-01 -1.80368230e-01
6.64265573e-01 5.61691940e-01 -1.16102219e-01 -3.02596092e-01
1.81109041e-01 -1.18799157e-01 -1.58514276e-01 -9.05599415e-01
-6.52291700e-02 -8.81912559e-02 -1.69010162e-01 7.22176313e-01
1.68339729e-01 1.38367295e+00 -8.57296646e-01 -7.37288594e-01
3.84084433e-01 3.45536560e-01 -5.35466671e-01 -1.87718347e-02
-3.85262996e-01 4.32919085e-01 -4.25742000e-01 9.16469693e-01
6.63310826e-01 -5.50622523e-01 -7.11271912e-02 -4.18922007e-01
-1.58282325e-01 -8.73815417e-02 -1.19784713e+00 1.60607994e+00
-8.70340541e-02 6.48594797e-01 2.36725375e-01 -7.45766938e-01
3.93889129e-01 1.25964224e-01 6.80536449e-01 -5.66587389e-01
2.58170426e-01 -6.91217929e-02 -4.11467999e-02 -3.12770426e-01
4.40512896e-01 3.76537263e-01 1.72481135e-01 2.33671293e-01
2.24075262e-02 3.13969612e-01 2.79231220e-01 9.81826559e-02
9.51536119e-01 5.51655471e-01 -1.01732567e-01 -1.53292924e-01
3.37038428e-01 -1.01095840e-01 8.97993684e-01 8.14221799e-01
-7.09042251e-01 7.90448308e-01 2.79598564e-01 -5.68697035e-01
-6.52483940e-01 -9.21926320e-01 -2.53296167e-01 1.16974854e+00
4.83719409e-01 -3.27356309e-01 -8.05133164e-01 -1.07183158e+00
2.08736151e-01 3.61676872e-01 -8.43270540e-01 2.86583543e-01
-1.02071702e+00 -4.85908628e-01 4.64682758e-01 9.63781297e-01
3.59159440e-01 -1.02412522e+00 -6.96980953e-01 3.17486972e-01
-1.06648542e-01 -1.40438306e+00 -1.01939487e+00 6.38162345e-02
-7.87338912e-01 -1.12250125e+00 -9.33329225e-01 -8.31511319e-01
6.18559182e-01 6.45179987e-01 1.11993361e+00 2.31951773e-01
-4.58516032e-01 3.54831219e-01 -1.26056790e-01 -4.55760360e-01
6.75073713e-02 5.69092557e-02 -1.88947439e-01 6.74694777e-02
2.10554212e-01 -1.38715655e-01 -9.85619009e-01 6.88019037e-01
-7.67680287e-01 1.12628527e-01 4.80421871e-01 4.75466907e-01
8.21201742e-01 -2.82388508e-01 2.40010366e-01 -8.75473559e-01
-1.53181016e-01 -2.31877849e-01 -9.44328964e-01 3.10253024e-01
-4.18009907e-01 -3.24264973e-01 3.22027683e-01 -5.36385834e-01
-9.19500053e-01 3.84380013e-01 -3.60254347e-02 -7.40645468e-01
7.40649004e-04 -1.49012730e-01 -4.63356972e-02 -2.87580460e-01
2.94610083e-01 1.76184475e-01 8.26357827e-02 -4.77049738e-01
5.10670245e-01 2.56906509e-01 5.85398734e-01 -3.71127218e-01
8.77612770e-01 7.76398122e-01 -1.14770480e-01 -3.68618071e-01
-1.09954619e+00 -6.80716038e-01 -7.30371118e-01 -4.18720186e-01
1.10883021e+00 -1.20131290e+00 -8.41640115e-01 4.64961141e-01
-9.42314386e-01 -6.34442866e-01 -2.72276253e-01 2.57764697e-01
-4.29980308e-01 3.90810221e-01 -6.76108778e-01 -5.65200686e-01
-1.94913596e-01 -1.47002852e+00 1.37582076e+00 2.62620956e-01
-1.30862156e-02 -8.56439948e-01 -3.67018163e-01 4.28834587e-01
3.27133298e-01 1.74398884e-01 2.33084604e-01 -4.10588443e-01
-1.05543256e+00 -1.21324591e-01 -5.03110707e-01 2.34909609e-01
-3.04253306e-02 -1.71619281e-02 -9.32364225e-01 -5.32963693e-01
-1.28315002e-01 -8.97573307e-02 1.20328832e+00 1.01691675e+00
1.60595310e+00 2.28964854e-02 -7.74060667e-01 8.87620687e-01
1.12383616e+00 -5.62028848e-02 4.06814426e-01 1.72080517e-01
1.04330122e+00 3.49254757e-01 8.62430096e-01 1.49253279e-01
4.51131076e-01 1.02270162e+00 5.93178213e-01 -5.61509788e-01
-6.07850075e-01 7.35009089e-02 3.96058768e-01 2.66602814e-01
-1.45882756e-01 -1.82277843e-01 -5.89140415e-01 6.12356126e-01
-2.19533110e+00 -9.10556495e-01 -1.68053895e-01 2.15590215e+00
5.59321821e-01 3.41799587e-01 6.74005806e-01 -3.74652535e-01
7.82051504e-01 3.61922234e-01 -7.17294216e-01 4.33757424e-01
1.89548910e-01 5.28031066e-02 7.22132862e-01 3.16256851e-01
-1.37658370e+00 1.17827237e+00 6.28907537e+00 8.76623750e-01
-1.16744924e+00 2.89452165e-01 7.92284966e-01 -5.89919150e-01
2.25806430e-01 -3.64293486e-01 -1.26641929e+00 6.52660072e-01
3.56629729e-01 3.63778621e-01 1.43365800e-01 8.16789865e-01
2.12913677e-01 1.49614643e-03 -1.28570008e+00 1.03215528e+00
1.03092976e-02 -1.52374411e+00 -1.80721670e-01 9.62426811e-02
6.44974768e-01 4.23935324e-01 1.61095649e-01 3.29427570e-01
1.29919469e-01 -7.83296287e-01 1.10228860e+00 3.22041035e-01
8.56955171e-01 -5.37198007e-01 4.39886361e-01 2.48123661e-01
-1.60764885e+00 -1.76142141e-01 -2.38822147e-01 4.31518883e-01
4.49117213e-01 3.29752445e-01 -3.87856305e-01 3.31894934e-01
9.73239779e-01 8.69265079e-01 -7.16111660e-01 1.42060995e+00
5.35171363e-04 6.49451971e-01 -2.06067801e-01 2.08939046e-01
3.28093499e-01 5.05716447e-03 5.19990981e-01 1.48635292e+00
-4.33592312e-02 -5.12224659e-02 5.57416975e-01 9.24622655e-01
-1.30317420e-01 -3.78119290e-01 -7.84718394e-02 3.02980363e-01
3.92397195e-01 1.38563538e+00 -1.10926294e+00 -5.03425360e-01
-6.75953031e-01 7.95429051e-01 3.11869770e-01 3.02607626e-01
-1.33755851e+00 9.52507257e-02 6.24024630e-01 3.66735280e-01
1.01861525e+00 -4.94122989e-02 -3.59251767e-01 -1.15008390e+00
9.85667631e-02 -8.09420228e-01 3.00960898e-01 -3.51427615e-01
-1.11354327e+00 5.26081383e-01 -1.40388265e-01 -1.33062887e+00
2.29605749e-01 -6.18403375e-01 -4.93477851e-01 5.43064415e-01
-1.67029989e+00 -1.40650678e+00 -3.75120074e-01 5.73185921e-01
9.16367888e-01 2.56003886e-02 1.95825964e-01 5.96229553e-01
-9.07652438e-01 8.87490332e-01 -2.41760477e-01 4.79740024e-01
7.05431938e-01 -1.16177630e+00 7.06909597e-01 9.81282353e-01
2.27074862e-01 4.61586267e-01 3.78834158e-01 -7.74013817e-01
-1.27712071e+00 -1.53051353e+00 3.60431790e-01 -7.81719863e-01
5.17680466e-01 -5.06362617e-01 -6.64805949e-01 1.00538445e+00
-4.01765481e-02 5.90211153e-01 3.92995119e-01 2.67126635e-02
-2.57418096e-01 -1.91378862e-01 -9.37594414e-01 6.96433783e-01
1.38962066e+00 1.18933022e-01 -2.25234747e-01 3.19987208e-01
9.05239403e-01 -7.78785467e-01 -7.20097601e-01 4.08177644e-01
6.69461131e-01 -1.03858244e+00 1.11222053e+00 -2.79275924e-01
1.70934811e-01 -6.33278787e-01 2.34548703e-01 -7.27595091e-01
-6.43814147e-01 -6.57779574e-01 -6.10684752e-01 9.80383933e-01
3.60198528e-01 -4.43889230e-01 9.68423486e-01 6.92403615e-01
-3.80127758e-01 -1.20637250e+00 -7.19836354e-01 -8.06907535e-01
-2.91245610e-01 -4.92447019e-01 5.30936062e-01 6.01317704e-01
-6.01601601e-01 1.06960468e-01 -4.11403000e-01 2.13059738e-01
7.51163900e-01 7.85030574e-02 9.15366828e-01 -1.13129461e+00
-6.47046685e-01 -7.16255009e-01 -2.00369373e-01 -1.71275640e+00
-2.81835217e-02 -8.25789452e-01 1.00695364e-01 -1.48144901e+00
3.71445030e-01 -5.84252834e-01 -3.86327118e-01 6.29647374e-01
-5.20425618e-01 6.60202980e-01 4.29795831e-01 2.78585166e-01
-1.12912154e+00 3.39701682e-01 1.51982474e+00 -9.56930146e-02
-1.35243952e-01 2.82349080e-01 -6.80609345e-01 8.51605773e-01
2.82623947e-01 -4.57075119e-01 -2.07784489e-01 -4.28863734e-01
-3.27474058e-01 -1.07094675e-01 4.86787200e-01 -9.86283660e-01
2.97713429e-01 -1.16716642e-02 6.06276751e-01 -8.78252625e-01
4.04786557e-01 -9.03352678e-01 -9.76252630e-02 5.18602729e-01
3.74090411e-02 -1.88155577e-01 3.22975010e-01 6.81900680e-01
2.35920846e-02 5.48280180e-02 9.00863886e-01 -1.20560132e-01
-7.92531908e-01 9.14896309e-01 7.34937936e-02 1.40550688e-01
1.22386563e+00 -5.94175637e-01 -1.85031340e-01 -2.03299113e-02
-6.92974091e-01 5.77596307e-01 5.47600567e-01 5.42610288e-01
3.46329927e-01 -9.69802439e-01 -5.53782761e-01 -6.59015924e-02
-8.79205540e-02 3.06857526e-01 1.71517611e-01 1.23065484e+00
-1.90346643e-01 2.41729349e-01 1.33781224e-01 -1.02195287e+00
-1.33380449e+00 5.57079434e-01 2.41583988e-01 -2.68880874e-01
-1.05192232e+00 1.16304934e+00 6.51500046e-01 -5.96410073e-02
6.50807202e-01 -3.53626192e-01 -1.45996690e-01 -3.24373581e-02
6.72794759e-01 4.47219253e-01 -7.58641213e-02 -6.82250023e-01
-3.69309366e-01 6.20997310e-01 -3.31923217e-01 2.00243354e-01
1.30578935e+00 -7.53181428e-02 1.38078168e-01 5.56022711e-02
8.84954274e-01 7.23018274e-02 -1.92729974e+00 -2.88218021e-01
-2.14795843e-01 -4.94621068e-01 7.97985792e-02 -7.57437706e-01
-1.65216863e+00 5.56445658e-01 5.55388331e-01 4.43488136e-02
9.35588539e-01 2.11061612e-01 1.09144521e+00 -1.40451491e-01
3.24217737e-01 -1.00768089e+00 3.43566500e-02 2.95376688e-01
5.73530138e-01 -1.29280806e+00 9.47342813e-02 -6.04299426e-01
-6.43622279e-01 8.38849843e-01 1.03394008e+00 -1.29937813e-01
5.41872919e-01 5.08783460e-01 -4.49655158e-03 -2.76501417e-01
-5.32004058e-01 -2.91509688e-01 6.95549011e-01 4.58760619e-01
3.59936893e-01 -8.58136714e-02 -9.02558938e-02 4.98574227e-01
8.96571577e-02 -2.62923911e-02 1.47729740e-01 6.93628609e-01
-4.38772529e-01 -8.75452340e-01 -2.68544376e-01 5.66731095e-01
-5.37669897e-01 7.81253725e-02 -1.12294711e-01 9.61739898e-01
3.13224643e-01 6.25376821e-01 8.21666345e-02 -1.57584324e-01
2.87882507e-01 -3.23535055e-01 6.32875800e-01 -6.89533174e-01
-7.46026695e-01 3.49122316e-01 -1.31496921e-01 -1.17244613e+00
-3.91339123e-01 -7.33863890e-01 -1.20348358e+00 -1.71811916e-02
-5.81656754e-01 -4.02694553e-01 3.90963644e-01 9.22641873e-01
4.99816120e-01 8.00607085e-01 1.76537752e-01 -1.32830298e+00
-3.89286906e-01 -5.80802023e-01 -3.40819091e-01 2.10159346e-01
4.45977181e-01 -9.14078712e-01 6.63699582e-02 2.14974638e-02] | [9.036554336547852, -0.14720898866653442] |
dd545144-14fd-4b72-bc6a-024bd62662f4 | collaborative-metric-learning-recommendation | 1803.00202 | null | http://arxiv.org/abs/1803.00202v1 | http://arxiv.org/pdf/1803.00202v1.pdf | Collaborative Metric Learning Recommendation System: Application to Theatrical Movie Releases | Product recommendation systems are important for major movie studios during
the movie greenlight process and as part of machine learning personalization
pipelines. Collaborative Filtering (CF) models have proved to be effective at
powering recommender systems for online streaming services with explicit
customer feedback data. CF models do not perform well in scenarios in which
feedback data is not available, in cold start situations like new product
launches, and situations with markedly different customer tiers (e.g., high
frequency customers vs. casual customers). Generative natural language models
that create useful theme-based representations of an underlying corpus of
documents can be used to represent new product descriptions, like new movie
plots. When combined with CF, they have shown to increase the performance in
cold start situations. Outside of those cases though in which explicit customer
feedback is available, recommender engines must rely on binary purchase data,
which materially degrades performance. Fortunately, purchase data can be
combined with product descriptions to generate meaningful representations of
products and customer trajectories in a convenient product space in which
proximity represents similarity. Learning to measure the distance between
points in this space can be accomplished with a deep neural network that trains
on customer histories and on dense vectorizations of product descriptions. We
developed a system based on Collaborative (Deep) Metric Learning (CML) to
predict the purchase probabilities of new theatrical releases. We trained and
evaluated the model using a large dataset of customer histories, and tested the
model for a set of movies that were released outside of the training window.
Initial experiments show gains relative to models that do not train on
collaborative preferences. | ['Miguel Campo', 'Abhinav Taliyan', 'Julie Rieger', 'JJ Espinoza'] | 2018-03-01 | null | null | null | null | ['product-recommendation'] | ['miscellaneous'] | [-5.98125458e-02 -2.59769976e-01 -9.54327062e-02 -8.60653281e-01
-6.56996727e-01 -8.38144600e-01 6.28918409e-01 4.74932224e-01
-3.70810598e-01 7.26541653e-02 6.00679994e-01 -3.19310188e-01
-3.42624098e-01 -8.28662992e-01 -4.13838536e-01 -2.36923873e-01
-2.74100691e-01 7.80530035e-01 -9.96636078e-02 -5.27073264e-01
2.59547353e-01 4.07914132e-01 -1.75285304e+00 8.86830270e-01
5.01326859e-01 9.50511813e-01 3.23096067e-01 7.21465409e-01
-2.18974784e-01 4.35904443e-01 -5.06353498e-01 -5.25959373e-01
5.97175777e-01 -1.72351912e-01 -2.80828893e-01 1.45495072e-01
3.76199901e-01 -3.27336192e-01 -2.54414469e-01 4.84891564e-01
2.93094307e-01 7.35776007e-01 6.65571809e-01 -8.40265095e-01
-9.76459622e-01 9.40228999e-01 -1.74066991e-01 4.63687271e-01
5.60710609e-01 1.63935214e-01 1.56734622e+00 -1.16431201e+00
6.01097286e-01 8.91178966e-01 7.93044269e-01 1.40432581e-01
-1.53917706e+00 -2.88396418e-01 4.04802531e-01 1.48416623e-01
-1.03065515e+00 -3.26352596e-01 4.00999427e-01 -5.43188691e-01
1.20011151e+00 3.03195328e-01 7.62727380e-01 1.05538487e+00
1.02897994e-01 7.96270549e-01 3.09846252e-01 4.80335616e-02
3.44923049e-01 5.26330113e-01 1.59045711e-01 1.10378917e-02
-2.33008519e-01 1.67189971e-01 -3.81193131e-01 -1.80322185e-01
3.12726915e-01 6.80688858e-01 -8.72260854e-02 -3.11086804e-01
-7.48424172e-01 1.26632786e+00 3.82149577e-01 3.35525453e-01
-6.31326675e-01 -1.88607231e-01 2.39836022e-01 6.72202229e-01
6.35345161e-01 1.05533040e+00 -6.05056405e-01 -2.26143047e-01
-1.26724017e+00 5.70451736e-01 9.32536721e-01 8.77886057e-01
4.18309480e-01 -3.33570540e-02 -1.91150308e-01 8.62813413e-01
1.44989684e-01 1.46343589e-01 6.38378322e-01 -8.59889627e-01
1.90352067e-01 1.43181846e-01 3.28309953e-01 -9.32812274e-01
-4.25262123e-01 -7.92133331e-01 -2.79001653e-01 1.50343892e-03
3.00971061e-01 -2.49000400e-01 -7.61787772e-01 1.29066277e+00
4.47075628e-02 8.11659768e-02 -5.11014313e-02 9.82144296e-01
4.51696455e-01 9.02735949e-01 -1.66706428e-01 -1.56130955e-01
9.04297650e-01 -7.06821620e-01 -3.64855886e-01 -5.23126647e-02
8.63643765e-01 -9.54069912e-01 1.23095727e+00 8.93417358e-01
-1.04774201e+00 -7.68795907e-01 -9.40115213e-01 1.56266004e-01
-3.57834488e-01 -2.80915052e-01 1.00417995e+00 5.78511178e-01
-8.85924339e-01 1.26402056e+00 -6.80865705e-01 -4.22555894e-01
4.55300838e-01 2.70307034e-01 -4.12039198e-02 -4.18442994e-01
-1.06963098e+00 7.23749280e-01 6.92666788e-03 -6.18487783e-02
-9.64731514e-01 -9.67052400e-01 -5.84371030e-01 4.18153316e-01
8.08883384e-02 -5.99858522e-01 1.37148154e+00 -1.24893534e+00
-1.30297720e+00 2.48349216e-02 3.77350301e-01 -5.18521488e-01
1.38765842e-01 -4.02922362e-01 -8.56851518e-01 -3.34078610e-01
-2.45977476e-01 3.59305799e-01 7.65295863e-01 -7.95471251e-01
-9.81246471e-01 -1.05063908e-01 2.83023417e-01 2.23099172e-01
-4.76781666e-01 1.70972288e-01 -1.81383379e-02 -4.73782510e-01
-2.00332239e-01 -9.23587084e-01 -4.70440626e-01 -4.86241013e-01
9.31811035e-02 -2.73655891e-01 3.10824722e-01 -3.71856302e-01
1.14938092e+00 -2.28469467e+00 -5.23427725e-02 3.56484920e-01
-1.35360304e-02 7.35590383e-02 -2.87722230e-01 7.56350279e-01
3.42380367e-02 -4.10551466e-02 4.75531846e-01 -4.96349394e-01
1.91678345e-01 8.42402950e-02 -4.30697560e-01 2.28406668e-01
-3.09514999e-02 5.53548276e-01 -9.92701292e-01 2.35627577e-01
-6.23084940e-02 2.71215767e-01 -1.08057022e+00 2.86033481e-01
-6.10961020e-01 1.46927938e-01 -1.31832659e-01 3.07466000e-01
2.74104208e-01 -2.04035074e-01 4.06669788e-02 1.78653598e-02
1.84162799e-02 5.73676884e-01 -1.17092180e+00 1.90752494e+00
-7.58951724e-01 4.04571593e-01 -2.22463056e-01 -7.23827720e-01
1.04350650e+00 1.64083034e-01 6.24630034e-01 -5.41475058e-01
4.22512293e-02 -4.62305136e-02 1.95744276e-01 -4.57344621e-01
1.04753602e+00 -3.08467716e-01 -6.19199593e-03 7.45778084e-01
2.95299858e-01 -3.25273424e-02 3.16559523e-01 4.63697612e-01
1.31350017e+00 2.04262771e-02 -3.58695984e-01 6.61639348e-02
-2.16961727e-01 -2.59542232e-03 4.48061109e-01 7.76053250e-01
2.66764820e-01 8.80266130e-01 1.30875871e-01 -5.97809613e-01
-1.10550714e+00 -1.20744908e+00 -8.32306966e-02 1.62580931e+00
-1.54547170e-01 -1.03992021e+00 4.65888269e-02 -6.20714962e-01
1.54010937e-01 1.22547424e+00 -3.71764809e-01 -4.62984681e-01
-2.01047555e-01 -7.04949975e-01 -1.79868758e-01 6.29602849e-01
-4.82735902e-01 -8.55328977e-01 -1.62898287e-01 6.39283895e-01
2.99713522e-01 -6.49103105e-01 -6.69044971e-01 2.80731738e-01
-7.57672191e-01 -5.72831213e-01 -4.53497380e-01 -3.99057746e-01
2.54149169e-01 2.77955979e-01 1.31239903e+00 -3.65516961e-01
2.13280357e-02 4.02283251e-01 -6.65177047e-01 -2.49302536e-01
-3.21209669e-01 1.35673493e-01 4.13877159e-01 2.88102686e-01
7.00707257e-01 -7.52427995e-01 -7.02789366e-01 4.61595863e-01
-8.09325933e-01 -3.59163404e-01 2.86153674e-01 8.31383228e-01
3.05845827e-01 1.87706918e-01 8.19873571e-01 -1.17828178e+00
9.12898362e-01 -1.05197310e+00 -3.38285148e-01 -2.25787070e-02
-9.59684730e-01 -1.16612375e-01 8.25794458e-01 -6.62186205e-01
-9.46563125e-01 -2.11419705e-02 -3.28880012e-01 -5.35850585e-01
-7.13930950e-02 7.98107207e-01 3.33489120e-01 7.46277928e-01
1.15469551e+00 -2.99786150e-01 -1.73606932e-01 -9.79583085e-01
7.53432691e-01 8.45463991e-01 3.67446065e-01 -9.11095515e-02
3.92846584e-01 1.29396645e-02 -6.11399174e-01 -3.45425695e-01
-8.89905155e-01 -7.09958017e-01 -4.54993576e-01 7.98544064e-02
4.50476676e-01 -8.81776452e-01 -3.45741957e-01 -3.31979990e-01
-5.41295469e-01 -1.81208000e-01 -7.82515705e-01 8.92707169e-01
-4.14243340e-01 -7.21691996e-02 -6.61158264e-01 -6.62137568e-01
-1.35137975e-01 -9.11827743e-01 5.02814174e-01 2.41821200e-01
-5.74340940e-01 -1.00874138e+00 2.58473873e-01 1.71252280e-01
7.20003307e-01 -2.15989143e-01 8.09248507e-01 -1.38650167e+00
-2.98088372e-01 -6.04946852e-01 4.33536530e-01 5.17233133e-01
7.71796256e-02 2.00919852e-01 -7.60996580e-01 -5.18501759e-01
-1.99093729e-01 -3.95945124e-02 7.83680379e-01 3.65624994e-01
8.93119633e-01 -3.04616213e-01 -1.73279434e-01 6.01220429e-01
1.05990541e+00 3.55753273e-01 3.39252144e-01 -9.05127004e-02
3.66513819e-01 6.59456372e-01 8.48853588e-01 7.62102425e-01
2.63271511e-01 7.60380447e-01 3.08055311e-01 1.57325938e-01
2.71185964e-01 -3.97525489e-01 5.23268521e-01 8.56329918e-01
2.33814701e-01 -3.80794346e-01 -4.82983112e-01 4.91470039e-01
-1.79786551e+00 -1.07017553e+00 1.34871691e-01 2.35613894e+00
4.51518089e-01 4.80474263e-01 4.81422484e-01 -2.31181279e-01
3.09299856e-01 -1.74745694e-01 -6.71042740e-01 -9.03651655e-01
1.34586155e-01 2.82978386e-01 3.47969860e-01 2.68951863e-01
-9.36431885e-01 6.35396302e-01 5.97967339e+00 5.34775555e-01
-8.98800910e-01 2.74222821e-01 4.92996305e-01 -9.23979759e-01
-7.68794894e-01 -1.69884693e-02 -7.57573783e-01 5.82791865e-01
1.56463492e+00 -1.32826045e-01 5.54993153e-01 1.02581632e+00
1.88934177e-01 1.73895717e-01 -1.60082889e+00 1.08893490e+00
1.17494538e-01 -1.62588525e+00 -2.63034135e-01 1.10675424e-01
7.68211067e-01 4.58913863e-01 3.46621752e-01 7.18062222e-01
8.95980239e-01 -9.76296723e-01 5.02089739e-01 5.93236506e-01
4.24419105e-01 -8.55958283e-01 8.18312466e-01 4.25471425e-01
-5.64711452e-01 -5.56284368e-01 -6.42860532e-01 -1.94755211e-01
5.10562003e-01 5.99622607e-01 -1.17847824e+00 1.98504582e-01
7.71143496e-01 1.11093080e+00 -4.20613587e-01 1.10658419e+00
3.67543638e-01 6.32918358e-01 -5.22121966e-01 -4.19431627e-02
1.77126795e-01 -3.38828444e-01 3.17487985e-01 1.04314601e+00
6.19833529e-01 2.97087748e-02 6.28013834e-02 8.05183649e-01
-2.46433541e-02 2.39736170e-01 -6.72612131e-01 -3.51902515e-01
3.35259020e-01 1.39565027e+00 -3.79877537e-01 -1.91030726e-01
-5.19879878e-01 8.14958036e-01 1.56152155e-02 2.90101290e-01
-5.89339554e-01 -1.73086256e-01 9.58260059e-01 6.26874030e-01
7.01624930e-01 -1.59201488e-01 1.08663969e-01 -1.23344791e+00
-3.90117288e-01 -8.46448600e-01 4.82663423e-01 -7.20914006e-01
-1.59116912e+00 7.36789346e-01 -2.15318948e-01 -1.28599489e+00
-5.48036218e-01 -1.56132728e-01 -6.30122542e-01 7.87342966e-01
-7.57051528e-01 -6.88062847e-01 1.63080364e-01 5.02567053e-01
7.03608096e-01 -5.35375655e-01 8.56647909e-01 5.52561522e-01
-2.06325576e-01 7.01411009e-01 6.22750103e-01 -4.41663414e-01
7.13651180e-01 -1.33030117e+00 6.67949975e-01 5.41565418e-01
9.05304134e-01 8.71228397e-01 8.93248439e-01 -4.03415084e-01
-1.33997178e+00 -9.42728758e-01 8.22178304e-01 -7.77653754e-01
7.02528894e-01 -6.38423562e-01 -7.95289993e-01 7.52658308e-01
-1.21617317e-03 -1.84688285e-01 1.34152770e+00 8.49299192e-01
-3.65602136e-01 -4.01229352e-01 -9.50627387e-01 3.86039585e-01
9.19054568e-01 -4.41105604e-01 -4.72256631e-01 8.08947146e-01
8.39740813e-01 3.67845148e-02 -1.33056390e+00 -6.04164042e-02
6.01158381e-01 -9.38124835e-01 8.36497366e-01 -1.05782425e+00
3.54678661e-01 -5.69197908e-02 -3.70629460e-01 -1.89866972e+00
-8.29076052e-01 -7.95958340e-01 -2.91064195e-03 1.27821732e+00
8.06725502e-01 -2.73188323e-01 5.90011001e-01 8.61479759e-01
-4.09429193e-01 -8.79382968e-01 -3.33377302e-01 -5.70593834e-01
-1.31411642e-01 -6.87105179e-01 8.64514887e-01 6.51793480e-01
2.74695218e-01 6.42934263e-01 -3.61866474e-01 -1.55484840e-01
-5.86438738e-02 3.91889155e-01 7.90895045e-01 -1.34229076e+00
-9.03426945e-01 -4.60865736e-01 -3.60816002e-01 -1.13891804e+00
-3.39892119e-01 -1.16575027e+00 -2.27686435e-01 -1.44318032e+00
-3.44896257e-01 -8.50866795e-01 -5.78889668e-01 7.49476328e-02
2.07054228e-01 1.30757704e-01 3.90373230e-01 3.41401994e-01
-6.08099043e-01 3.09921414e-01 7.77750432e-01 -3.63150686e-02
-5.19988298e-01 5.48314929e-01 -9.32274342e-01 3.95698369e-01
6.76398098e-01 -4.50897157e-01 -6.63022876e-01 -2.87139922e-01
8.57723892e-01 -4.34572250e-02 -4.17969584e-01 -7.96439588e-01
1.01103380e-01 -5.06683737e-02 4.68168050e-01 -5.14722109e-01
5.63145518e-01 -7.58100569e-01 6.11709535e-01 -1.41308561e-01
-8.35083604e-01 2.15957791e-01 -3.06436479e-01 6.78932905e-01
-1.68677943e-03 -2.43840471e-01 3.34630191e-01 -4.33336198e-02
-5.11826873e-01 4.77469474e-01 -4.92769927e-01 -2.97793597e-01
7.45489955e-01 -2.04869255e-01 6.58494011e-02 -7.09691346e-01
-1.28273284e+00 9.44253728e-02 4.36412245e-01 8.84961069e-01
4.58250314e-01 -1.25766778e+00 -7.07871854e-01 4.23821747e-01
2.64072478e-01 -3.11571956e-01 2.67338753e-01 6.84869111e-01
-1.86476380e-01 2.15432152e-01 -3.82339731e-02 -3.44709456e-01
-7.22416818e-01 1.00472248e+00 1.08419850e-01 -1.04813658e-01
-6.34222507e-01 1.07330263e+00 4.88561168e-02 -1.80103928e-01
9.84904766e-02 -5.30407965e-01 -5.21546364e-01 3.79753858e-01
7.04314113e-01 6.56012371e-02 4.26425755e-01 -3.76240551e-01
6.29374087e-02 -2.44618401e-01 -6.11584067e-01 -6.32583126e-02
1.66460931e+00 -1.53086662e-01 6.29148066e-01 6.77975893e-01
1.28844488e+00 -8.10406134e-02 -1.45814550e+00 -4.34747487e-01
7.09059788e-03 -7.61884451e-01 1.87176347e-01 -8.58290553e-01
-1.16873264e+00 7.19981134e-01 6.61528289e-01 5.76478183e-01
8.21989536e-01 -2.02356596e-02 9.89650488e-01 1.47516012e-01
3.59833539e-01 -1.11670256e+00 -2.70305555e-02 1.78873897e-01
6.75651193e-01 -1.14940727e+00 -1.16349850e-03 3.88796479e-01
-8.51662755e-01 1.01812303e+00 2.19367743e-01 -3.36679667e-01
1.10969579e+00 4.49514762e-02 -1.48533091e-01 -2.49572873e-01
-1.34144628e+00 -2.60668676e-02 3.03606153e-01 5.19046903e-01
4.24210072e-01 3.39391291e-01 3.71503755e-02 1.03568137e+00
-3.80357504e-01 -1.57963723e-01 6.19682610e-01 5.64729810e-01
-3.96082878e-01 -1.10533619e+00 1.45965815e-01 1.05566502e+00
-4.83034700e-01 -2.67260641e-01 -4.73875329e-02 7.15952590e-02
2.61568666e-01 1.09324527e+00 4.95399475e-01 -8.77331793e-01
3.46464992e-01 4.21493407e-03 2.08500698e-01 -1.18728352e+00
-1.05956388e+00 1.91425592e-01 3.97751987e-01 -7.04577267e-01
1.64226070e-01 -1.12146938e+00 -8.63841891e-01 -4.27871227e-01
-5.05385995e-01 4.15517002e-01 8.63543332e-01 6.83148146e-01
6.77831888e-01 3.33701998e-01 1.00277472e+00 -9.95946109e-01
-8.11448276e-01 -1.05174756e+00 -1.04879951e+00 6.35514677e-01
5.12920618e-02 -4.90216017e-01 -3.81073445e-01 6.59233257e-02] | [10.107243537902832, 5.7809157371521] |
388d9a5e-41b4-4e52-8a81-d14bdfa3e8ca | capitalization-cues-improve-dependency | null | null | https://aclanthology.org/W12-1903 | https://aclanthology.org/W12-1903.pdf | Capitalization Cues Improve Dependency Grammar Induction | null | ['Hiyan Alshawi', 'Valentin I. Spitkovsky', 'Daniel Jurafsky'] | 2012-06-01 | null | null | null | ws-2012-6 | ['dependency-grammar-induction'] | ['natural-language-processing'] | [-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01] | [-7.398244857788086, 3.6925456523895264] |
d70dc534-8e85-44a2-96fd-5f779b6d9958 | cadet-fully-self-supervised-anomaly-detection | 2210.01742 | null | https://arxiv.org/abs/2210.01742v3 | https://arxiv.org/pdf/2210.01742v3.pdf | CADet: Fully Self-Supervised Out-Of-Distribution Detection With Contrastive Learning | Handling out-of-distribution (OOD) samples has become a major stake in the real-world deployment of machine learning systems. This work explores the use of self-supervised contrastive learning to the simultaneous detection of two types of OOD samples: unseen classes and adversarial perturbations. First, we pair self-supervised contrastive learning with the maximum mean discrepancy (MMD) two-sample test. This approach enables us to robustly test whether two independent sets of samples originate from the same distribution, and we demonstrate its effectiveness by discriminating between CIFAR-10 and CIFAR-10.1 with higher confidence than previous work. Motivated by this success, we introduce CADet (Contrastive Anomaly Detection), a novel method for OOD detection of single samples. CADet draws inspiration from MMD, but leverages the similarity between contrastive transformations of a same sample. CADet outperforms existing adversarial detection methods in identifying adversarially perturbed samples on ImageNet and achieves comparable performance to unseen label detection methods on two challenging benchmarks: ImageNet-O and iNaturalist. Significantly, CADet is fully self-supervised and requires neither labels for in-distribution samples nor access to OOD examples. | ['Joao Monteiro', 'Ioannis Mitliagkas', 'David Vazquez', 'Pau Rodriguez', 'Charles Guille-Escuret'] | 2022-10-04 | null | null | null | null | ['self-supervised-anomaly-detection', 'supervised-anomaly-detection'] | ['computer-vision', 'computer-vision'] | [ 3.58443469e-01 2.49134842e-02 4.12201881e-02 -2.11376861e-01
-1.06526089e+00 -1.14289129e+00 9.83246863e-01 3.27381700e-01
-3.52536470e-01 4.33852494e-01 -1.13418207e-01 -3.50105196e-01
4.38169867e-01 -4.99102265e-01 -8.99634719e-01 -5.21206260e-01
-3.07420492e-01 4.73678350e-01 2.75863469e-01 1.54931527e-02
1.36971101e-02 6.57465279e-01 -1.37712860e+00 2.97453851e-01
6.32156730e-01 1.15308642e+00 -6.52904749e-01 8.56956005e-01
1.67203337e-01 8.53012621e-01 -1.06784987e+00 -5.37021041e-01
8.57959509e-01 -4.40684468e-01 -5.39039969e-01 7.44453892e-02
1.46261585e+00 -5.14962316e-01 -4.85163808e-01 1.36408472e+00
6.56501770e-01 -4.26973999e-02 1.15535176e+00 -1.88343251e+00
-7.53753662e-01 2.07965106e-01 -7.09942222e-01 8.93147409e-01
2.01680928e-01 5.13449848e-01 7.95213103e-01 -1.04132438e+00
4.08935189e-01 1.52612329e+00 6.74680173e-01 6.46621644e-01
-1.55728233e+00 -1.00310254e+00 1.10395662e-01 -1.40346497e-01
-1.18475294e+00 -6.06825769e-01 6.38854563e-01 -6.56293809e-01
6.13595068e-01 3.24443549e-01 1.48544803e-01 1.59473109e+00
-1.20873225e-03 9.81761754e-01 1.15504980e+00 -3.54787707e-01
4.60228086e-01 2.85423219e-01 -3.57992724e-02 5.00470757e-01
2.47845680e-01 4.63755071e-01 -3.01104009e-01 -4.25857425e-01
2.62705833e-01 -4.65336591e-02 -2.16389790e-01 -4.58460838e-01
-8.70430470e-01 8.80432129e-01 3.87443990e-01 -1.08364828e-01
3.73599981e-03 1.38257295e-01 6.26156747e-01 5.98946750e-01
7.01936662e-01 7.03476489e-01 -1.16921932e-01 2.18211904e-01
-7.91296661e-01 2.42431790e-01 7.66195834e-01 9.32443678e-01
5.83693624e-01 4.13097680e-01 -3.41622800e-01 5.88501692e-01
1.62461385e-01 8.19541812e-01 4.69823152e-01 -8.28126729e-01
2.72680640e-01 2.54817337e-01 -4.74334061e-02 -8.35702837e-01
6.00397140e-02 -5.91039598e-01 -6.94121480e-01 9.10648227e-01
6.81908011e-01 -1.21113740e-01 -1.17698324e+00 1.64637554e+00
5.11177719e-01 3.00608903e-01 2.17185855e-01 6.22565985e-01
6.09989285e-01 3.42027336e-01 1.01664558e-01 2.90277511e-01
1.10582745e+00 -8.36835802e-01 -2.38027066e-01 -5.48610508e-01
5.76069593e-01 -7.41013646e-01 1.03515065e+00 3.88038278e-01
-6.51938140e-01 -3.94838333e-01 -1.13270271e+00 3.99894059e-01
-5.11101007e-01 -5.15751123e-01 1.86866120e-01 9.25324857e-01
-6.68769836e-01 6.05446219e-01 -4.06550318e-01 -4.23406586e-02
9.88781214e-01 8.43492616e-03 -4.77410227e-01 -2.20737606e-01
-1.04994452e+00 6.27338052e-01 2.47386917e-01 -4.60079044e-01
-1.53854394e+00 -9.79567111e-01 -9.49311674e-01 -4.09048587e-01
3.29085559e-01 -6.01821803e-02 1.16953433e+00 -1.34460127e+00
-7.14613318e-01 1.51300442e+00 2.58599877e-01 -7.90985346e-01
1.11472392e+00 -2.88151413e-01 -7.20570743e-01 2.82528490e-01
3.19671810e-01 5.99174201e-01 1.60595477e+00 -1.30003870e+00
-5.76503217e-01 -2.20822483e-01 -1.31538898e-01 -1.21557422e-01
-4.10878420e-01 -3.63564193e-02 6.39195517e-02 -1.04985750e+00
-2.19220385e-01 -7.77958632e-01 1.67875230e-01 3.22983652e-01
-7.65378654e-01 -1.50885835e-01 1.06481767e+00 -2.54307002e-01
6.71722889e-01 -2.55727601e+00 -5.51835656e-01 3.57860625e-01
7.16002524e-01 4.58465993e-01 -2.70039380e-01 -7.73649365e-02
-4.53727096e-01 2.06521645e-01 -3.84387732e-01 -4.06893641e-01
2.68279612e-01 8.31726342e-02 -7.73603916e-01 8.34669530e-01
5.98400772e-01 7.11047649e-01 -1.17586744e+00 -4.31145281e-01
5.39228059e-02 9.38206166e-03 -4.50001746e-01 4.42711294e-01
-5.14649637e-02 1.80635944e-01 3.15300450e-02 1.00348866e+00
7.50605047e-01 1.70021370e-01 -5.67335427e-01 -1.02887444e-01
3.80122960e-01 -2.66127050e-01 -1.16856515e+00 9.31710303e-01
1.14964657e-01 7.40499079e-01 -1.09435715e-01 -8.38944137e-01
8.24162900e-01 1.01503886e-01 5.20150140e-02 -7.12275863e-01
9.53091308e-02 3.88886839e-01 -5.12703620e-02 -2.80150145e-01
5.40720634e-02 -1.37885988e-01 1.10980142e-02 2.91440159e-01
2.60216027e-01 -2.98122466e-01 1.68490723e-01 3.67142677e-01
1.48388159e+00 -3.20304602e-01 3.44440758e-01 -3.04982483e-01
2.22912967e-01 -1.64684489e-01 5.02562106e-01 1.40817165e+00
-8.40235412e-01 7.51078844e-01 5.47523260e-01 -3.44804436e-01
-1.06117761e+00 -1.76685131e+00 -3.38551849e-01 9.99017477e-01
8.49609450e-03 8.14010724e-02 -7.29965448e-01 -1.37004638e+00
5.44443607e-01 7.48987615e-01 -8.85226905e-01 -4.69646752e-01
-2.54629374e-01 -5.91469169e-01 9.67152119e-01 4.05252486e-01
3.96415532e-01 -1.06445312e+00 -5.10684550e-02 -3.08476053e-02
2.95714527e-01 -1.23568106e+00 -7.74029016e-01 4.60629851e-01
-4.32867408e-01 -1.32379138e+00 -5.61102569e-01 -7.61028588e-01
6.14237726e-01 1.39402822e-01 1.47838664e+00 -3.30869317e-01
-6.98918164e-01 8.22134435e-01 -2.15684950e-01 -8.94385278e-01
-1.12489021e+00 -2.29095817e-01 3.10200244e-01 2.13162512e-01
5.10108173e-01 -5.75582802e-01 -4.94519413e-01 4.27625954e-01
-1.10171139e+00 -8.82169127e-01 3.89858365e-01 6.03816628e-01
5.85078478e-01 4.33777012e-02 6.66713178e-01 -1.07028401e+00
4.96202350e-01 -7.77277291e-01 -5.78863442e-01 -6.65977150e-02
-5.25487661e-01 -2.91966274e-02 8.50960374e-01 -8.70905340e-01
-4.05497462e-01 -1.77919239e-01 -3.48168910e-02 -1.12637234e+00
-7.12655604e-01 -2.90954411e-01 -1.73324883e-01 -1.12630859e-01
1.26433003e+00 -3.72403376e-02 2.65485905e-02 -1.81545079e-01
2.72220105e-01 4.29263145e-01 1.04708779e+00 -5.44925094e-01
1.44363797e+00 5.86827695e-01 -1.75278053e-01 -8.21666002e-01
-1.03428113e+00 -5.35215437e-01 -3.51441175e-01 -1.19969375e-01
6.81800604e-01 -1.05016673e+00 -2.06179246e-01 8.69773507e-01
-6.76919520e-01 -5.10668695e-01 -7.79350281e-01 1.81096047e-01
-1.77651584e-01 5.06508589e-01 -4.87924546e-01 -7.45127976e-01
-2.74198234e-01 -9.03110743e-01 1.02397561e+00 -1.35351643e-01
-3.52495939e-01 -1.08488202e+00 1.07022569e-01 1.45931570e-02
2.41039664e-01 8.07779491e-01 6.42837226e-01 -1.45055485e+00
-1.49691850e-01 -6.60558105e-01 -3.47534567e-02 8.75743628e-01
1.81629762e-01 -4.02353667e-02 -1.55097890e+00 -6.36498511e-01
1.35994963e-02 -8.64862502e-01 7.81993866e-01 -7.62738436e-02
1.16662931e+00 -3.30516934e-01 4.56367768e-02 7.55725384e-01
1.20411599e+00 -1.15930684e-01 5.44882596e-01 4.50285047e-01
7.90163934e-01 3.44551772e-01 5.76674402e-01 2.36625880e-01
-2.43286878e-01 2.57190198e-01 8.57429981e-01 -3.18069756e-01
-2.60240495e-01 -2.17033297e-01 4.72991079e-01 6.28166050e-02
7.64906704e-01 -4.23444599e-01 -8.57686818e-01 5.66888630e-01
-1.24887252e+00 -1.20451963e+00 1.21568881e-01 2.26188731e+00
8.21797669e-01 6.93762660e-01 3.53596717e-01 3.32735866e-01
9.58777547e-01 2.41639569e-01 -1.11688268e+00 -2.44233385e-01
-3.55920196e-01 3.47204834e-01 6.63859189e-01 2.83187926e-01
-1.84650588e+00 6.59098208e-01 6.34144163e+00 9.96577859e-01
-9.12846625e-01 4.40805197e-05 9.85740960e-01 -1.12026490e-01
-2.25413162e-02 -5.04064679e-01 -8.05829465e-01 6.21874571e-01
6.04970396e-01 1.01688102e-01 2.30459392e-01 1.22387791e+00
-3.79506528e-01 2.33767048e-01 -1.48548770e+00 1.06410217e+00
4.15595651e-01 -1.00359464e+00 7.56619051e-02 1.08089224e-01
8.14734042e-01 3.36585820e-01 4.83666092e-01 5.76309502e-01
7.34954834e-01 -1.03126097e+00 9.01614010e-01 -1.69835314e-02
9.06065524e-01 -6.02295518e-01 5.91519296e-01 1.58009931e-01
-8.44205618e-01 -1.37519196e-01 -2.49647364e-01 3.85381132e-01
-3.58108759e-01 7.68200159e-01 -9.46001053e-01 -5.75821698e-02
8.44894528e-01 6.44929111e-01 -7.95010209e-01 9.85310435e-01
-2.01529622e-01 9.79963243e-01 -3.53766620e-01 4.87300843e-01
2.29273617e-01 3.15849662e-01 9.92362738e-01 1.21293902e+00
-1.27923131e-01 -5.57437181e-01 4.10629809e-01 9.47055757e-01
-5.53630948e-01 -2.54791081e-01 -8.15913141e-01 6.39387667e-02
6.59999847e-01 1.08108890e+00 -5.60770035e-01 -3.67174566e-01
-1.86710685e-01 1.22689676e+00 1.97295859e-01 2.93929309e-01
-9.79677081e-01 -4.95144993e-01 9.67875063e-01 9.14773643e-02
3.41162711e-01 2.78376043e-01 2.23010585e-01 -1.18303061e+00
2.08303705e-02 -1.37421346e+00 6.26488090e-01 -3.26461643e-01
-2.16098952e+00 6.00451291e-01 -1.41684219e-01 -1.44537997e+00
-1.67311251e-01 -8.06502044e-01 -9.29666102e-01 7.34359622e-01
-1.57426655e+00 -8.31047833e-01 -4.87434447e-01 7.14519024e-01
4.85811591e-01 -4.93975788e-01 8.48651409e-01 2.30424479e-01
-6.44140363e-01 1.18261349e+00 1.64184794e-01 7.49715447e-01
1.30648708e+00 -1.86090732e+00 9.28021252e-01 1.22271466e+00
1.89526871e-01 4.17432338e-02 7.28186250e-01 -5.74033499e-01
-8.10836315e-01 -1.61388993e+00 -1.85970571e-02 -7.83427715e-01
7.95121431e-01 -6.08358204e-01 -9.30624366e-01 6.82902753e-01
-1.03307709e-01 8.03352416e-01 7.06266344e-01 -4.34636086e-01
-9.49858546e-01 -1.61879942e-01 -1.51866007e+00 5.47982454e-01
9.53046858e-01 -7.59576380e-01 -7.36650765e-01 4.58486408e-01
6.71096325e-01 -3.97610635e-01 -7.48410523e-01 3.16419214e-01
6.13076575e-02 -9.70833242e-01 1.31227434e+00 -7.07039416e-01
3.08629125e-01 -3.12320411e-01 -2.39638925e-01 -1.34268987e+00
-1.97837085e-01 -7.48471916e-01 -4.50223953e-01 1.36231911e+00
1.21478215e-01 -9.20631588e-01 6.45555615e-01 3.29316050e-01
-1.80749029e-01 -3.81396860e-01 -7.02837050e-01 -1.29481125e+00
2.36106619e-01 -3.68773580e-01 3.91672999e-01 1.18574047e+00
-6.67869508e-01 1.38709461e-02 -1.16504170e-01 5.39629638e-01
1.27200150e+00 -3.44907999e-01 1.08335066e+00 -1.22836566e+00
-4.80766863e-01 -3.82840484e-01 -8.70441556e-01 -7.34147489e-01
3.03950846e-01 -1.07102096e+00 -4.67278585e-02 -6.98221385e-01
-1.58789799e-01 -4.33140606e-01 -4.38377798e-01 4.71459478e-01
-4.28412288e-01 9.29193497e-01 1.63437854e-02 2.88893491e-01
-6.28557980e-01 4.30183500e-01 6.50833428e-01 -5.30089855e-01
1.45116925e-01 -1.76956574e-03 -6.65323436e-01 8.64434183e-01
7.80951083e-01 -7.97602415e-01 -3.23395818e-01 -1.84674188e-02
-1.46531492e-01 -8.06342423e-01 7.37661779e-01 -1.27137458e+00
-2.73953974e-01 2.43294135e-01 7.21034884e-01 -3.87767792e-01
-1.24352045e-01 -6.57512069e-01 -6.45797372e-01 5.26021302e-01
-4.69210297e-01 2.55611707e-02 4.07669753e-01 9.41612244e-01
-1.07841447e-01 -3.64182562e-01 1.21218359e+00 -9.13159475e-02
-7.68607974e-01 5.41272700e-01 -3.97991359e-01 1.11562467e+00
1.23830116e+00 -1.79317355e-01 -5.17775655e-01 -2.19494537e-01
-5.03045678e-01 6.24224283e-02 4.19015199e-01 3.90096098e-01
5.35006344e-01 -1.29348743e+00 -7.60454714e-01 4.66414601e-01
5.40428042e-01 7.08877221e-02 -1.62032411e-01 4.24664706e-01
-5.20313740e-01 -4.08485681e-01 -8.57215226e-02 -7.86096632e-01
-1.19856679e+00 8.05506349e-01 6.58442080e-01 -2.92551313e-02
-4.38755155e-01 1.02662170e+00 4.84937847e-01 -4.94928151e-01
5.18667579e-01 7.97274858e-02 2.96605468e-01 -9.64179412e-02
7.50315964e-01 3.15498322e-01 2.52969656e-02 -2.78629601e-01
-2.58358061e-01 -1.61446020e-01 -3.71440530e-01 1.59261793e-01
9.29139376e-01 2.53920913e-01 1.92287236e-01 5.00935435e-01
1.55477107e+00 3.03054184e-01 -1.52894521e+00 -4.47054714e-01
-1.55701740e-02 -6.10175669e-01 -2.32721657e-01 -6.92698419e-01
-8.46736848e-01 7.41860986e-01 9.03542757e-01 3.94991964e-01
7.55842507e-01 2.94403601e-02 7.15700865e-01 2.68685549e-01
-2.11272269e-01 -7.72994995e-01 6.98196232e-01 2.46060714e-01
6.40750945e-01 -1.68310773e+00 -1.14016950e-01 -2.28678435e-01
-3.92735213e-01 9.20487821e-01 8.36793244e-01 -5.01253128e-01
5.78351617e-01 3.32729399e-01 2.92996943e-01 -1.78563401e-01
-2.43448928e-01 -2.48552533e-03 3.89822364e-01 8.92151177e-01
8.98047443e-03 -1.49047658e-01 7.50777841e-01 -4.73264307e-02
-2.14974836e-01 -7.71152854e-01 3.63540292e-01 9.57545638e-01
-2.24574834e-01 -7.09824979e-01 -6.34453535e-01 6.52774096e-01
-6.09192491e-01 -2.61006713e-01 -6.14613235e-01 8.82576883e-01
5.18120304e-02 8.46803844e-01 3.12880874e-01 -3.26476365e-01
3.02442908e-01 1.18872859e-01 3.00817192e-01 -5.38701713e-01
-4.83077884e-01 -1.75055519e-01 -2.29276851e-01 -4.19296443e-01
-7.83153847e-02 -6.49844527e-01 -6.89498663e-01 -1.73604012e-01
-2.13072866e-01 -3.50061394e-02 2.39447653e-01 8.68575692e-01
2.85354733e-01 4.75589514e-01 9.83703852e-01 -7.27092266e-01
-1.03457081e+00 -7.09872067e-01 -6.29715860e-01 1.08058786e+00
8.39441597e-01 -5.52426517e-01 -1.20686984e+00 -9.67538282e-02] | [8.275948524475098, 2.59401273727417] |
638d0a4f-bd16-4763-95ce-29256f604b96 | duluthnlp-at-semeval-2021-task-7-fine-tuning | null | null | https://aclanthology.org/2021.semeval-1.169 | https://aclanthology.org/2021.semeval-1.169.pdf | DuluthNLP at SemEval-2021 Task 7: Fine-Tuning RoBERTa Model for Humor Detection and Offense Rating | This paper presents the DuluthNLP submission to Task 7 of the SemEval 2021 competition on Detecting and Rating Humor and Offense. In it, we explain the approach used to train the model together with the process of fine-tuning our model in getting the results. We focus on humor detection, rating, and of-fense rating, representing three out of the four subtasks that were provided. We show that optimizing hyper-parameters for learning rate, batch size and number of epochs can increase the accuracy and F1 score for humor detection | ['Samuel Akrah'] | 2021-08-01 | null | null | null | semeval-2021 | ['humor-detection'] | ['natural-language-processing'] | [-3.90019089e-01 8.72058049e-03 -5.78330830e-03 -3.52681041e-01
-3.46750915e-01 -4.94396001e-01 6.33894026e-01 1.25350535e-01
-5.29491365e-01 6.01199925e-01 8.73054385e-01 -1.74889956e-02
2.75735259e-01 -3.30866188e-01 -1.72004893e-01 -6.07819073e-02
1.97033495e-01 3.11390966e-01 6.05725832e-02 -4.65430558e-01
8.70612621e-01 1.70048624e-01 -1.19588637e+00 8.84136319e-01
6.74678564e-01 5.44559717e-01 -2.46885255e-01 9.05524671e-01
4.13074493e-01 1.95391583e+00 -9.24699783e-01 -7.01960623e-01
-7.57692829e-02 -6.21141076e-01 -1.18486261e+00 -5.28945960e-02
5.32613873e-01 -3.46802384e-01 -5.61096430e-01 8.41254532e-01
5.78590870e-01 5.69728732e-01 7.67492235e-01 -1.17499852e+00
-8.88890207e-01 9.93604720e-01 -2.79674411e-01 5.02277792e-01
6.73372030e-01 2.82297015e-01 1.23087490e+00 -1.36217344e+00
5.59669256e-01 1.08383346e+00 7.14006841e-01 7.91525066e-01
-8.07232320e-01 -3.36963385e-01 -6.60584807e-01 7.97312260e-01
-8.42962444e-01 -5.92620790e-01 8.28383267e-01 -8.40931654e-01
1.42730045e+00 1.23805292e-01 4.61338162e-01 1.24407327e+00
-7.74399564e-02 1.20948374e+00 1.14621043e+00 -3.84588689e-01
9.18615907e-02 3.81653190e-01 6.40191317e-01 6.08756840e-01
-1.83599770e-01 -2.26045385e-01 -9.54864621e-01 -2.14138255e-01
4.68716413e-01 -5.92039466e-01 -2.85945505e-01 2.24607661e-01
-8.69087040e-01 1.08643389e+00 4.59928513e-01 2.77165860e-01
-1.40447930e-01 -6.62677884e-02 9.57453787e-01 4.19210345e-01
4.05334204e-01 1.06894898e+00 -7.29878247e-02 -3.46343994e-01
-1.11157596e+00 5.93273401e-01 1.04500067e+00 6.89859629e-01
-2.53228322e-02 1.72074765e-01 -6.76653743e-01 1.33004367e+00
-1.78572759e-01 2.25794688e-02 7.80471146e-01 -1.10015428e+00
4.73629504e-01 5.36926508e-01 3.19309235e-01 -9.33121979e-01
-8.17830443e-01 -3.12845558e-01 -3.59688073e-01 1.72554642e-01
4.65072632e-01 -9.95687097e-02 -5.83807409e-01 1.39491665e+00
-3.21214020e-01 -3.24215204e-01 -3.23692530e-01 1.16791654e+00
1.36438417e+00 4.98688787e-01 1.83801241e-02 -3.09318155e-01
1.28623784e+00 -1.36738253e+00 -8.70098114e-01 -4.69073206e-01
7.95547307e-01 -8.58697355e-01 1.47744274e+00 6.36263549e-01
-1.48256445e+00 -3.38303149e-01 -1.30273342e+00 -5.41909695e-01
-3.24270606e-01 1.88432522e-02 2.16637880e-01 4.47342366e-01
-6.69281781e-01 8.21354151e-01 -5.09673990e-02 -2.06717446e-01
2.78069764e-01 -1.62511453e-01 -1.02537453e-01 2.53282756e-01
-1.55575275e+00 1.80516648e+00 3.22856247e-01 -2.51680911e-01
-1.01151645e+00 -5.52599788e-01 -6.16824985e-01 4.57027070e-02
-1.24388613e-01 -3.64204675e-01 1.54632390e+00 -5.62377870e-01
-1.13246107e+00 1.37818623e+00 1.55866817e-01 -5.48070490e-01
4.95103180e-01 -4.55304176e-01 -1.73509553e-01 -1.33908823e-01
-6.47762120e-02 2.44148254e-01 7.68299639e-01 -6.62722230e-01
-3.16280454e-01 -1.16823852e-01 6.24459684e-02 2.05406874e-01
-7.27095187e-01 7.26150751e-01 3.95714670e-01 -2.34312356e-01
-5.47683299e-01 -5.09283781e-01 9.10541564e-02 -9.53345716e-01
-1.27906293e-01 -4.67107475e-01 5.12959659e-01 -9.30010915e-01
1.51317430e+00 -1.78899491e+00 2.50042349e-01 -3.55789423e-01
4.57505047e-01 2.84485459e-01 -1.03625804e-01 4.36613500e-01
2.00866461e-02 1.79665536e-02 8.87486339e-02 -3.38879138e-01
2.30266571e-01 -4.02644515e-01 -4.46402729e-01 3.53747606e-01
-3.31246734e-01 1.06855226e+00 -9.72388446e-01 -3.59262854e-01
1.53734595e-01 2.80339211e-01 -4.27583903e-01 5.69513202e-01
8.71042386e-02 -1.67071566e-01 2.84907371e-01 3.32967520e-01
1.96315601e-01 -3.29216987e-01 -2.91323423e-01 1.60181388e-01
1.94942519e-01 9.89216387e-01 -5.38799345e-01 9.24255252e-01
-5.18662512e-01 1.02371693e+00 -1.62213355e-01 -2.07489595e-01
1.05117023e+00 3.96174401e-01 3.96102145e-02 -7.55243242e-01
2.75154144e-01 2.19614714e-01 -2.14088008e-01 -6.34346485e-01
8.51469100e-01 -5.48877239e-01 -2.45948628e-01 5.89729428e-01
1.08194187e-01 -2.67780274e-01 4.82076645e-01 2.92933673e-01
1.24069846e+00 -3.30944568e-01 6.66784167e-01 -3.04098397e-01
8.69640350e-01 -6.32326901e-02 1.12291798e-01 7.04469085e-01
-5.35254061e-01 5.12798369e-01 7.57968068e-01 -6.66827261e-01
-1.53777552e+00 -5.49527228e-01 1.83697902e-02 1.65498102e+00
-5.31822562e-01 -6.84237897e-01 -7.08855689e-01 -6.73092008e-01
-1.45104721e-01 1.39866984e+00 -6.77449048e-01 -1.50374770e-01
-7.64705777e-01 -9.32683587e-01 7.20680892e-01 4.91231352e-01
2.21838370e-01 -1.73151886e+00 -9.69221175e-01 1.38712049e-01
-6.07379258e-01 -1.14276505e+00 -4.26801741e-01 3.79198939e-01
-7.04661012e-01 -1.17494071e+00 -2.34065980e-01 -4.60159004e-01
6.91632926e-02 1.54853567e-01 1.61289632e+00 4.38932478e-01
-1.36834621e-01 -1.85012855e-02 -5.32079875e-01 -9.11258832e-02
-4.84822512e-01 -3.48145813e-02 -2.00833216e-01 -7.38047481e-01
8.24796498e-01 -5.51457286e-01 -3.15847725e-01 4.67106774e-02
-5.06922245e-01 7.30442628e-03 -7.60450661e-02 9.52746153e-01
-3.83323729e-01 -2.79950857e-01 7.69935846e-01 -1.06694365e+00
1.47513866e+00 -3.82121652e-01 -4.49928977e-02 -3.14766705e-01
-8.94523442e-01 -4.50916409e-01 8.49967420e-01 -2.10382506e-01
-5.88186979e-01 -1.44433677e-01 -8.04989338e-02 -1.60215482e-01
1.32055774e-01 4.71844494e-01 3.51962030e-01 2.66176015e-01
1.41282833e+00 -9.64638144e-02 -5.16648710e-01 -4.43325967e-01
4.26044136e-01 9.24431324e-01 7.76880920e-01 -3.41081083e-01
4.91224825e-01 -1.05194002e-01 -3.31575453e-01 -4.16342109e-01
-1.51545131e+00 -7.26977706e-01 -7.66041458e-01 -4.96247232e-01
5.49281061e-01 -8.37485850e-01 -5.80698729e-01 5.43713927e-01
-1.59168017e+00 -4.44126517e-01 -2.02514067e-01 -1.78198926e-02
-8.11880291e-01 1.96961194e-01 -1.12263775e+00 -9.48738992e-01
-1.03843522e+00 -4.89733487e-01 3.23816746e-01 1.11382425e-01
-8.20837498e-01 -9.52701092e-01 4.18576062e-01 1.10188198e+00
4.85220373e-01 5.12493178e-02 1.01543903e+00 -1.13264990e+00
3.53941828e-01 -3.76604080e-01 -2.77214915e-01 5.89738369e-01
-5.08347929e-01 -4.00842667e-01 -1.15258038e+00 -1.13096416e-01
2.95440465e-01 -1.17576325e+00 1.03635800e+00 4.21906002e-02
9.39319193e-01 -8.00189197e-01 5.13703346e-01 1.51536211e-01
1.00241590e+00 -3.45873922e-01 9.10774112e-01 8.37567151e-01
6.23441517e-01 6.45910323e-01 3.28519881e-01 9.27044809e-01
2.29625374e-01 5.34670651e-01 5.21237195e-01 4.55438405e-01
-1.95685282e-01 -4.24716383e-01 7.99250007e-01 7.58325100e-01
-1.60486594e-01 1.63390767e-02 -9.01947260e-01 7.50933528e-01
-1.95155919e+00 -1.53087664e+00 -6.44596636e-01 2.06785297e+00
1.05559158e+00 1.19853429e-01 8.72020960e-01 2.27542251e-01
5.24429560e-01 4.33219582e-01 -1.57673791e-01 -8.70893538e-01
-1.57059669e-01 -8.97709057e-02 -4.19692658e-02 1.06175113e+00
-8.69051158e-01 1.27114654e+00 7.75974369e+00 5.30796945e-01
-4.73266870e-01 3.95461261e-01 3.46627593e-01 -6.71935499e-01
-1.15999609e-01 -2.50554323e-01 -6.30951166e-01 2.76486963e-01
1.20715070e+00 -3.23416531e-01 9.08364236e-01 1.14073992e+00
4.06167567e-01 1.07517950e-01 -1.12016118e+00 7.52040982e-01
5.89094758e-01 -8.75947475e-01 -3.86200279e-01 -4.04755414e-01
8.69701326e-01 1.74256429e-01 -2.91062117e-01 7.84501016e-01
3.02829862e-01 -1.49623501e+00 8.06709528e-01 2.65071750e-01
1.34620920e-01 -8.41925383e-01 9.82386410e-01 6.03655219e-01
-2.51052082e-01 -4.68151748e-01 -6.27257645e-01 -5.49041033e-01
1.14750341e-01 7.10016429e-01 -9.02154446e-01 -5.11443079e-01
5.74355721e-01 6.24890149e-01 -9.59406614e-01 1.23656070e+00
-9.58357453e-01 8.25769067e-01 2.39008740e-01 -2.77454257e-01
-4.20202799e-02 4.03601110e-01 5.90517879e-01 1.87209535e+00
-7.55245909e-02 8.00530985e-02 -4.98597212e-02 9.60593939e-01
-3.62692177e-01 2.38506407e-01 -1.74985886e-01 7.71732926e-02
5.59678376e-01 1.60699058e+00 -1.41760141e-01 -3.13801259e-01
9.53419283e-02 8.77817988e-01 7.70222902e-01 -1.73575863e-01
-8.48936141e-01 -6.06715381e-01 7.65742734e-02 3.39234173e-01
1.06181493e-02 -1.01060256e-01 -9.83082235e-01 -1.11621380e+00
-3.12985450e-01 -1.15527892e+00 6.15433395e-01 -1.09228551e+00
-1.58140194e+00 5.62032104e-01 -1.44773215e-01 -4.55659240e-01
-2.78397620e-01 -6.93884730e-01 -9.32225347e-01 7.06120849e-01
-1.08771348e+00 -9.06663299e-01 -4.93154526e-01 3.84604871e-01
6.39648974e-01 -2.73969531e-01 4.26817060e-01 9.61940214e-02
-3.63246530e-01 5.31303585e-01 -2.77503610e-01 1.26537308e-01
9.36339617e-01 -1.47851670e+00 3.11285764e-01 5.23381591e-01
-1.97688699e-01 4.02890205e-01 1.47131085e+00 -4.82483029e-01
-5.12070835e-01 -4.89902288e-01 1.66893947e+00 -9.90223408e-01
1.22781765e+00 -2.85294086e-01 -1.02666473e+00 4.50107962e-01
5.84137678e-01 -6.89323664e-01 4.54249442e-01 5.76964021e-01
-9.12073970e-01 5.37986934e-01 -1.19406295e+00 2.89780796e-01
6.51709378e-01 -7.39344001e-01 -9.57364559e-01 8.53522003e-01
3.86998475e-01 -4.60747033e-01 -7.63625979e-01 1.72927380e-01
4.72886711e-01 -1.23852694e+00 6.11946225e-01 -1.11712646e+00
1.46013343e+00 7.78488070e-02 -1.44328609e-01 -1.17836690e+00
-8.53931665e-01 -5.53992569e-01 -6.26405776e-01 9.92577493e-01
4.15827572e-01 2.53321916e-01 8.00838470e-01 5.53025126e-01
-6.27511889e-02 -6.70120239e-01 -7.84541309e-01 -5.39704204e-01
5.97222686e-01 -2.49822453e-01 1.80323292e-02 9.51088428e-01
9.77737546e-01 1.00841415e+00 -9.44573581e-01 -6.33846402e-01
5.87181926e-01 -2.38293380e-01 5.65158069e-01 -1.05001569e+00
-2.46648461e-01 -7.90773213e-01 -8.54889303e-02 -6.29609644e-01
2.34619290e-01 -9.97608304e-01 1.13211766e-01 -1.45718026e+00
1.02251124e+00 6.40569508e-01 -4.45504725e-01 8.45341921e-01
-3.89813006e-01 4.76207137e-01 6.66453898e-01 5.10065317e-01
-9.15874243e-01 4.17801768e-01 9.25789356e-01 -2.86958013e-02
-1.91281095e-01 -1.25702277e-01 -7.54674017e-01 7.10367024e-01
8.98212314e-01 -5.60645401e-01 -2.09528074e-01 -6.35503698e-03
7.26752162e-01 4.98831272e-02 6.54547095e-01 -9.14289355e-01
2.21760124e-01 -2.23016605e-01 4.01801080e-01 -5.54557025e-01
2.75876045e-01 -1.30653113e-01 -5.59362233e-01 2.22888991e-01
-9.25262690e-01 2.69429944e-03 1.35640649e-03 -1.71011448e-01
1.87070593e-01 -1.06441462e+00 1.25809896e+00 -3.10289979e-01
-2.21188724e-01 -5.74381351e-01 -6.93733573e-01 3.71246010e-01
5.77135682e-01 1.38583064e-01 -9.98957753e-01 -7.89384365e-01
-6.31264150e-01 2.30121568e-01 3.55330110e-01 6.28720164e-01
7.01547265e-01 -1.31845450e+00 -1.25445008e+00 -4.83288676e-01
5.49500473e-02 -1.03141057e+00 7.84045309e-02 8.71609807e-01
-1.84342951e-01 4.69627112e-01 -5.04510105e-01 2.24915937e-01
-1.54600418e+00 6.39171183e-01 5.89730620e-01 -6.82776749e-01
-4.70636010e-01 6.94872737e-01 -4.62480426e-01 -5.58057904e-01
2.90397435e-01 8.06971729e-01 -7.60777652e-01 1.69128150e-01
9.48842824e-01 1.03785229e+00 3.43001261e-02 -6.00042939e-01
-1.94758385e-01 -2.56575882e-01 -4.04468715e-01 -1.80790380e-01
1.38571715e+00 8.16654488e-02 -5.37330449e-01 7.07899928e-01
9.17168796e-01 -2.26951111e-02 -5.18382490e-01 -6.03367798e-02
3.78866792e-01 -4.17192250e-01 2.38422692e-01 -1.51675940e+00
-4.90768820e-01 8.91776264e-01 1.15462005e-01 2.50041127e-01
6.67459488e-01 -5.59082329e-02 9.40195918e-01 5.79951167e-01
1.78548858e-01 -1.70726252e+00 4.69945908e-01 1.24918544e+00
1.18510675e+00 -1.01474488e+00 4.30729359e-01 2.12644607e-01
-1.05666041e+00 1.26504278e+00 1.09045362e+00 -3.49874765e-01
-4.01600599e-02 -1.35118023e-01 6.60077250e-03 -5.27937651e-01
-1.25127399e+00 2.61996984e-01 5.41455448e-01 1.75246462e-01
1.24606526e+00 -3.71981077e-02 -8.23706925e-01 9.19598043e-01
-7.03455567e-01 -1.27994075e-01 1.07092559e+00 3.36735755e-01
-1.27089036e+00 -3.78972292e-01 -2.18620658e-01 4.17310148e-01
-5.62578678e-01 -3.45127314e-01 -1.31450224e+00 3.85930955e-01
-2.36904204e-01 1.12230206e+00 -6.42199636e-01 -9.05095339e-01
5.70734501e-01 3.67293537e-01 5.22777736e-01 -4.88570333e-01
-1.34500802e+00 -5.62641084e-01 6.64948702e-01 -4.34046417e-01
9.46631879e-02 -4.65231925e-01 -1.10556304e+00 -8.12149227e-01
-2.38294706e-01 1.96309194e-01 3.11628997e-01 1.10749674e+00
-1.85079902e-01 1.02175742e-01 7.01010168e-01 -6.03204787e-01
-9.83505964e-01 -1.50553334e+00 -5.25262892e-01 6.60291135e-01
1.13214664e-01 -2.68904775e-01 -9.37070787e-01 -2.25361794e-01] | [8.8696870803833, 11.08786392211914] |
0e238e27-dd06-4571-bde9-4f556d325d15 | image-seam-carving-by-controlling-positional | 1912.13214 | null | https://arxiv.org/abs/1912.13214v1 | https://arxiv.org/pdf/1912.13214v1.pdf | Image Seam-Carving by Controlling Positional Distribution of Seams | Image retargeting is a new image processing task that renders the change of aspect ratio in images. One of the most famous image-retargeting algorithms is seam-carving. Although seam-carving is fast and straightforward, it usually distorts the images. In this paper, we introduce a new seam-carving algorithm that not only has the simplicity of the original seam-carving but also lacks the usual unwanted distortion existed in the original method. The positional distribution of seams is introduced. We show that the proposed method outperforms the original seam-carving in terms of retargeted image quality assessment and seam coagulation measures. | ['Shadrokh Samavi', 'Nader Karimi', 'Mahdi Ahmadi'] | 2019-12-31 | null | null | null | null | ['image-retargeting'] | ['computer-vision'] | [ 3.51221859e-01 -1.09610714e-01 2.46204466e-01 -1.13794476e-01
-3.88507575e-01 -4.26817924e-01 5.02020061e-01 -1.11833975e-01
-4.24761921e-01 6.54178679e-01 1.15216784e-01 -2.29956239e-01
9.37498286e-02 -7.10011780e-01 -6.46719337e-01 -7.27121711e-01
3.81658316e-01 -1.75476953e-01 7.66012788e-01 -6.43817604e-01
6.43632948e-01 4.76517230e-01 -1.55115795e+00 -1.70614749e-01
1.07370186e+00 5.00806570e-01 4.29845095e-01 7.47334838e-01
-2.96003092e-02 1.56115264e-01 -5.65317273e-01 -5.03449738e-01
3.48576456e-01 -7.79101431e-01 -4.75750685e-01 3.69562089e-01
6.59022212e-01 -6.25933185e-02 -3.78134936e-01 1.53697932e+00
4.80360329e-01 2.92985588e-01 5.02949059e-01 -9.48822260e-01
-1.11608887e+00 1.02356322e-01 -1.14908278e+00 5.17279267e-01
8.94522369e-02 -1.25998393e-01 1.82282344e-01 -1.03395975e+00
5.61045051e-01 1.27391434e+00 6.36021435e-01 3.86107475e-01
-1.37228322e+00 -5.36117196e-01 -2.22715065e-01 4.39850777e-01
-1.70190179e+00 -1.87895745e-01 7.79897213e-01 -3.48193675e-01
2.33809724e-01 7.75323212e-01 4.36576098e-01 1.93527952e-01
8.21197748e-01 3.82093698e-01 1.19067776e+00 -7.88360596e-01
1.51026249e-01 9.65898857e-03 -1.02476321e-01 5.91785431e-01
1.78421810e-01 2.99792111e-01 -2.82397509e-01 1.71253234e-01
9.42910492e-01 -2.74092387e-02 -4.78802353e-01 -5.50775290e-01
-1.35169780e+00 5.81201613e-01 3.91564488e-01 4.52562541e-01
3.10467612e-02 1.85284972e-01 1.86334103e-01 3.19474310e-01
5.34781277e-01 4.24412400e-01 1.66166678e-01 1.58206627e-01
-1.03550518e+00 1.91636086e-01 -7.93039128e-02 9.65905547e-01
1.06978047e+00 4.25182879e-01 -2.67767370e-01 9.24564481e-01
-1.21509777e-02 3.88953686e-01 8.92278790e-01 -1.00010347e+00
-1.19013548e-01 4.39714760e-01 -5.33620007e-02 -1.33096957e+00
-2.22925931e-01 -8.67494494e-02 -8.23580384e-01 8.07824969e-01
-1.46768549e-02 1.29012823e-01 -1.05394387e+00 1.26962209e+00
3.89055103e-01 7.44940117e-02 -1.66031227e-01 9.84403193e-01
7.27071643e-01 7.31295884e-01 -2.32079133e-01 -2.55563200e-01
1.27186763e+00 -1.21548557e+00 -1.11905122e+00 -1.58404803e-03
5.01113795e-02 -1.37565327e+00 1.06416869e+00 2.50728726e-01
-1.29299700e+00 -7.94724226e-01 -1.33064055e+00 -2.08905578e-01
-3.63792956e-01 -2.31090441e-01 3.98005605e-01 1.12464595e+00
-1.39969778e+00 3.33625287e-01 -1.98487371e-01 -5.36242187e-01
-6.67823628e-02 1.25910655e-01 -5.50844014e-01 2.75009461e-02
-8.21268141e-01 1.09863281e+00 4.04795289e-01 -2.55140126e-01
-5.74167669e-01 -8.11233044e-01 -8.62511814e-01 -1.32761881e-01
3.65945309e-01 -5.29523432e-01 1.06301224e+00 -1.03503180e+00
-1.57293630e+00 8.46085191e-01 -2.23105162e-01 -6.66400194e-02
7.14861751e-01 -3.57528552e-02 -7.65146673e-01 6.89650252e-02
4.51924428e-02 7.25991428e-01 1.34909141e+00 -1.48149955e+00
-6.88356698e-01 -2.51088083e-01 -4.06773806e-01 2.89485335e-01
-1.32832423e-01 -1.78875417e-01 -7.56999135e-01 -1.45891130e+00
1.88625649e-01 -7.31135488e-01 -3.13876957e-01 1.00313932e-01
-3.88364077e-01 3.44621956e-01 1.34079719e+00 -5.07541597e-01
1.20958471e+00 -2.49010372e+00 6.53970893e-03 1.09495319e-01
2.11201012e-01 2.73141593e-01 -2.54019558e-01 5.45859635e-01
-4.60370690e-01 3.13267447e-02 -4.90163147e-01 -1.23355485e-01
-5.21219254e-01 6.32157624e-02 -2.68658489e-01 6.66792452e-01
-2.53153801e-01 8.94121110e-01 -7.42634952e-01 -6.85328484e-01
6.42514408e-01 6.15927160e-01 -3.90340954e-01 1.27305631e-02
2.38542825e-01 4.95470762e-01 3.36692214e-01 4.56349432e-01
1.18146265e+00 4.60863173e-01 -3.82633269e-01 -5.20242751e-01
-5.08752704e-01 -6.94108903e-01 -1.13630581e+00 1.71777809e+00
-2.58119345e-01 8.56754065e-01 -1.42853245e-01 -4.94403869e-01
1.10050464e+00 6.79921284e-02 3.72956842e-01 -7.83821881e-01
-8.81428793e-02 2.13247105e-01 -3.49881768e-01 -6.10153794e-01
1.17434919e+00 -4.72637087e-01 1.79498285e-01 3.43586057e-01
-1.88619629e-01 -5.27959943e-01 2.16544092e-01 1.19465813e-01
4.53887135e-01 -2.95449942e-02 5.55947244e-01 -7.57054687e-01
7.61397600e-01 -2.45635561e-03 4.12454665e-01 3.39463443e-01
-8.00756440e-02 1.35492110e+00 -7.62248561e-02 -5.06843984e-01
-1.35985029e+00 -1.27684438e+00 -1.88298032e-01 7.88087785e-01
8.38073194e-01 -1.32912472e-01 -9.73547876e-01 -4.00786608e-01
-2.17207670e-01 8.65885794e-01 -7.81217873e-01 -4.10087585e-01
-5.17578840e-01 -6.33585572e-01 2.70228744e-01 6.71994090e-02
8.39118183e-01 -6.27431393e-01 -5.04618764e-01 -7.99415261e-02
-1.62907004e-01 -6.22697175e-01 -1.42502081e+00 -3.97198111e-01
-9.31750894e-01 -1.00794160e+00 -1.16059279e+00 -9.78286982e-01
1.03920674e+00 1.00840092e+00 8.34461689e-01 2.43425444e-01
-5.78088760e-01 2.35158607e-01 -4.71679449e-01 -1.31692439e-01
-5.81781745e-01 -3.89427304e-01 -1.83078662e-01 3.15653801e-01
1.54998282e-03 -4.09020752e-01 -7.59094894e-01 6.07033670e-01
-1.45712304e+00 7.87664130e-02 5.09859860e-01 6.76597536e-01
7.58075833e-01 6.62726164e-01 2.84563482e-01 -6.79299712e-01
6.56455636e-01 2.30918646e-01 -4.98458803e-01 2.87296355e-01
-9.59768772e-01 1.06476722e-02 3.21925908e-01 -4.06599432e-01
-1.40367699e+00 -5.45188524e-02 2.78872736e-02 -2.88298637e-01
6.60276562e-02 -1.05403088e-01 -1.60090417e-01 -7.36467659e-01
6.42258942e-01 5.65580726e-01 2.55626023e-01 -7.00334191e-01
7.39002347e-01 3.38165224e-01 1.24699354e+00 -3.80833969e-02
1.08398187e+00 6.63374126e-01 -1.46943703e-02 -1.14251435e+00
4.32836032e-03 -2.28288695e-01 -5.66533804e-01 -4.62965012e-01
8.45261335e-01 -4.41922098e-01 -3.89174014e-01 4.61543888e-01
-1.03572106e+00 1.46266222e-01 -6.56099319e-01 2.03557312e-01
-5.96766949e-01 1.00480723e+00 -2.90139586e-01 -3.77290577e-01
-2.52076477e-01 -1.25970030e+00 8.26494396e-01 4.27983373e-01
2.71313414e-02 -8.53498816e-01 3.62726063e-01 3.98504436e-02
6.20555401e-01 2.06029907e-01 9.76776063e-01 2.05591425e-01
-2.39281356e-01 -3.91483456e-02 -2.84949362e-01 2.01002017e-01
5.99719822e-01 8.61997157e-03 -6.78000450e-01 -3.58444750e-01
2.31477290e-01 5.25703013e-01 7.87146986e-01 4.93787587e-01
9.73534167e-01 -2.79825270e-01 -1.78853601e-01 6.80953801e-01
1.69323218e+00 7.01255858e-01 1.24443746e+00 6.29119873e-01
4.53498214e-01 4.70919311e-01 6.12649560e-01 -1.99603047e-02
1.43379062e-01 9.68083978e-01 4.38773066e-01 -6.60182476e-01
-5.58411896e-01 -2.37369761e-01 7.36627057e-02 7.05866754e-01
-5.77440560e-02 1.28717735e-01 -3.95774096e-01 5.21833718e-01
-1.76788211e+00 -8.53139639e-01 -4.29524481e-01 2.27174616e+00
7.24644780e-01 -3.47891837e-01 -7.99321011e-02 2.06839815e-01
1.18365216e+00 9.98987556e-02 -1.65017128e-01 -7.35187888e-01
-3.92467678e-01 -6.08002096e-02 6.74384296e-01 9.14934576e-01
-1.05843532e+00 8.13437521e-01 7.64592505e+00 1.01719546e+00
-8.19742203e-01 2.03423396e-01 4.31838930e-01 2.67295152e-01
-4.93156433e-01 -6.43383637e-02 -3.16644251e-01 5.70667207e-01
1.32767215e-01 -7.01867580e-01 2.94654489e-01 5.50707519e-01
2.17958584e-01 -5.40356219e-01 -5.36746562e-01 1.24745941e+00
5.07281661e-01 -1.21413112e+00 4.26446229e-01 -1.96636692e-01
8.63234460e-01 -7.73738861e-01 4.56010103e-01 -3.72469574e-01
-7.23192543e-02 -9.09571469e-01 8.81917119e-01 5.21143198e-01
8.03929329e-01 -1.03825831e+00 7.85367668e-01 -1.86903805e-01
-1.12211692e+00 2.87255675e-01 -5.68324625e-01 3.86079490e-01
2.89583653e-01 7.75164187e-01 -3.12853515e-01 5.95302105e-01
6.91530049e-01 3.41030777e-01 -9.76434708e-01 1.48007298e+00
1.54013544e-01 -3.07978280e-02 1.84488341e-01 5.18467784e-01
-4.99300621e-02 -7.49212921e-01 7.84566998e-01 1.21996415e+00
7.81914651e-01 9.21123102e-03 -4.19933945e-01 8.16351831e-01
2.33505756e-01 2.78621584e-01 -7.25587249e-01 4.66380119e-01
2.73659647e-01 1.09886408e+00 -8.37465644e-01 -2.93293178e-01
-3.73444945e-01 1.56581354e+00 -5.19101024e-01 3.41755569e-01
-6.23854995e-01 -7.01354980e-01 5.02539635e-01 2.13115126e-01
2.52117127e-01 -1.69799939e-01 -3.73050243e-01 -8.40929925e-01
-2.63035297e-01 -7.23159850e-01 2.76380926e-01 -9.73791718e-01
-9.06724870e-01 6.25402629e-01 1.25262454e-01 -1.46617866e+00
1.03784256e-01 -1.73510626e-01 -7.33571529e-01 7.60622799e-01
-1.30719185e+00 -9.85669076e-01 -5.37525415e-01 7.72548020e-01
1.05627954e+00 -1.91343114e-01 4.13382441e-01 4.27057743e-01
-4.06570554e-01 6.42633557e-01 2.22009465e-01 -4.30441022e-01
9.53391194e-01 -1.11301410e+00 5.23783565e-01 1.36527050e+00
-6.42523617e-02 4.39254612e-01 1.29552352e+00 -6.13871753e-01
-1.11685371e+00 -1.05460012e+00 5.96244156e-01 2.67742258e-02
1.18583724e-01 1.40328079e-01 -9.58507478e-01 5.40851593e-01
7.28301406e-01 -2.68770814e-01 4.22803432e-01 -7.40955055e-01
-2.46836662e-01 -1.68065831e-01 -1.37249613e+00 9.03069556e-01
6.64801419e-01 -5.25156744e-02 -7.88463354e-01 -1.33416548e-01
7.93649793e-01 -3.12136769e-01 -5.46803951e-01 2.48958096e-01
4.22285408e-01 -1.36992764e+00 1.25011671e+00 2.35877439e-01
8.96872443e-05 -7.91445792e-01 -3.87033299e-02 -1.54399884e+00
-7.11180687e-01 -8.07893813e-01 3.74718457e-01 1.16396892e+00
2.57555321e-02 -5.51543474e-01 5.63424826e-01 2.74416387e-01
-3.17624241e-01 -8.16360787e-02 -1.05317354e+00 -1.06401920e+00
-1.03667349e-01 1.04746081e-01 6.99868917e-01 8.42773080e-01
1.73963625e-02 -2.96506975e-02 -6.48244023e-01 -9.82002392e-02
9.08473492e-01 2.50346703e-03 5.90632200e-01 -6.91109955e-01
-4.65456732e-02 -4.69780505e-01 -6.14701509e-01 -6.53872371e-01
-5.31718373e-01 -5.71972907e-01 7.81539455e-02 -1.45074069e+00
4.75296862e-02 5.60107902e-02 1.38387486e-01 1.54918984e-01
-4.57898587e-01 8.85969698e-01 1.85843512e-01 1.73805848e-01
-7.25227147e-02 7.39476621e-01 1.49008858e+00 -2.56379545e-01
-3.46229374e-01 -3.07813048e-01 -6.69232905e-01 5.02436519e-01
9.28934813e-01 -3.10550541e-01 -5.26324093e-01 -4.33068544e-01
-7.12265223e-02 -3.37976426e-01 2.35827282e-01 -1.05170310e+00
2.32955247e-01 -1.34513989e-01 1.41559407e-01 -7.59715319e-01
1.41267121e-01 -9.35519278e-01 6.83236241e-01 5.23062706e-01
-7.81307835e-03 5.10484874e-01 3.41209710e-01 5.72690785e-01
-3.45257193e-01 -3.99113148e-01 1.30659151e+00 -6.77848831e-02
-1.05367255e+00 -4.82172184e-02 -7.47732401e-01 -4.00293767e-01
1.48468935e+00 -7.60894299e-01 -3.36880684e-01 -2.97223002e-01
-3.93562257e-01 -3.46711934e-01 9.93745089e-01 6.47238970e-01
9.38409746e-01 -1.65444279e+00 -5.54229259e-01 3.36706877e-01
1.44460976e-01 -3.34383368e-01 5.67651510e-01 7.93249905e-01
-1.02697539e+00 1.96084782e-01 -7.08100438e-01 -2.55538881e-01
-1.40323579e+00 1.21758223e+00 3.17416817e-01 4.94885743e-01
-9.74056184e-01 5.97260594e-01 6.47735715e-01 -3.98266166e-02
-2.48445228e-01 2.38810509e-01 -1.78788275e-01 -2.83249021e-01
8.61736655e-01 8.12152445e-01 1.16389647e-01 -8.47363412e-01
-1.10771276e-01 1.14481747e+00 -5.35913184e-02 -2.18613952e-01
7.62675405e-01 -7.46693313e-01 -4.78319287e-01 2.76697099e-01
1.14077401e+00 9.08769816e-02 -9.74668860e-01 2.60537080e-02
-3.01207215e-01 -1.21221292e+00 9.73310843e-02 -3.90450388e-01
-1.35826468e+00 7.27652252e-01 1.09453952e+00 2.10740045e-01
1.42453289e+00 -3.90762806e-01 8.21234047e-01 -2.05786988e-01
3.64895821e-01 -1.08772445e+00 1.55274376e-01 -3.98915783e-02
1.12219858e+00 -8.74791503e-01 1.96224347e-01 -6.97411478e-01
-6.08005047e-01 1.08107769e+00 6.67755723e-01 -8.00969899e-02
5.95960379e-01 2.45114446e-01 3.73454034e-01 -4.89814915e-02
-4.04211506e-02 -1.01758592e-01 3.28253597e-01 1.01951563e+00
1.80527627e-01 -1.83525264e-01 -9.75781739e-01 -1.10565230e-01
-1.77280292e-01 -1.33896947e-01 8.87151778e-01 6.72831655e-01
-6.73595428e-01 -1.09377372e+00 -1.02463019e+00 -6.56445250e-02
-2.65433848e-01 -2.83385888e-02 -2.59793758e-01 6.93612874e-01
1.91711053e-01 1.05824494e+00 1.27015904e-01 -6.20850742e-01
5.09286284e-01 -3.57786626e-01 4.29440379e-01 -1.81668147e-01
-3.63561004e-01 1.10494964e-01 -4.55428392e-01 -6.24334514e-01
-4.42439020e-01 -1.55984595e-01 -9.89454329e-01 -5.03672302e-01
-2.99221992e-01 2.21993983e-01 6.36640847e-01 3.65610659e-01
3.10759753e-01 4.79316592e-01 6.24397933e-01 -9.33122694e-01
8.95430148e-02 -6.65347040e-01 -8.96696270e-01 3.99158388e-01
4.98135120e-01 -6.02111578e-01 -2.80693263e-01 5.22427857e-01] | [11.107629776000977, -1.202759027481079] |
38444511-3480-42b5-ae13-e508d849e072 | a-highly-effective-low-rank-compression-of | 2111.15179 | null | https://arxiv.org/abs/2111.15179v2 | https://arxiv.org/pdf/2111.15179v2.pdf | A Highly Effective Low-Rank Compression of Deep Neural Networks with Modified Beam-Search and Modified Stable Rank | Compression has emerged as one of the essential deep learning research topics, especially for the edge devices that have limited computation power and storage capacity. Among the main compression techniques, low-rank compression via matrix factorization has been known to have two problems. First, an extensive tuning is required. Second, the resulting compression performance is typically not impressive. In this work, we propose a low-rank compression method that utilizes a modified beam-search for an automatic rank selection and a modified stable rank for a compression-friendly training. The resulting BSR (Beam-search and Stable Rank) algorithm requires only a single hyperparameter to be tuned for the desired compression ratio. The performance of BSR in terms of accuracy and compression ratio trade-off curve turns out to be superior to the previously known low-rank compression methods. Furthermore, BSR can perform on par with or better than the state-of-the-art structured pruning methods. As with pruning, BSR can be easily combined with quantization for an additional compression. | ['Wonjong Rhee', 'Suhyun Kang', 'Moonjung Eo'] | 2021-11-30 | null | null | null | null | ['low-rank-compression'] | ['computer-code'] | [ 2.45768324e-01 -1.71079203e-01 -3.56214046e-01 -2.42927566e-01
-8.91568601e-01 -1.74737364e-01 3.38871062e-01 5.31360507e-01
-5.34870982e-01 6.82507753e-01 6.65674135e-02 -3.54643732e-01
-6.16605639e-01 -8.78965199e-01 -6.01025283e-01 -7.26134896e-01
-1.20335363e-01 6.81431293e-01 4.78463322e-01 -8.59773234e-02
4.05713499e-01 5.35027325e-01 -1.96323478e+00 1.33553296e-01
1.01656401e+00 1.41342807e+00 5.87296844e-01 4.76810724e-01
-1.20426476e-01 4.97485161e-01 -3.61644626e-01 -4.84651357e-01
4.08434093e-01 -1.73285589e-01 -6.66875482e-01 -3.91916126e-01
4.29899782e-01 -5.67966759e-01 -6.21139824e-01 1.08002055e+00
4.98883069e-01 1.47330135e-01 2.69444793e-01 -7.62403905e-01
-1.26652922e-02 9.06605244e-01 -7.04435170e-01 2.36088037e-01
1.38795758e-02 -4.16044980e-01 1.15251708e+00 -6.46947622e-01
5.07205784e-01 1.13045955e+00 6.26903117e-01 3.15504491e-01
-1.22054505e+00 -6.58428490e-01 -3.99755388e-01 5.44302642e-01
-1.62445652e+00 -5.18030584e-01 5.67381799e-01 -1.01976968e-01
8.16573083e-01 3.07117194e-01 7.87456691e-01 4.06019270e-01
9.72003415e-02 6.48152888e-01 9.15525794e-01 -5.13977170e-01
5.01455486e-01 -1.33532375e-01 1.29911855e-01 7.35454977e-01
6.38413429e-01 5.58101386e-02 -8.38843524e-01 -2.20010936e-01
6.41324103e-01 -4.07680869e-02 -2.71536618e-01 -3.92343074e-01
-8.87875617e-01 9.87420976e-01 2.92195231e-01 3.62434894e-01
-1.38383389e-01 5.28095841e-01 5.98692954e-01 4.67181176e-01
1.50759473e-01 3.48525405e-01 -3.76156926e-01 -5.30953825e-01
-1.69364440e+00 2.91954547e-01 7.16440558e-01 7.69107223e-01
5.48213124e-01 -3.42254564e-02 -1.71264395e-01 1.06611407e+00
1.17989607e-01 1.62063986e-01 5.76936364e-01 -1.01111007e+00
4.84282851e-01 4.32130218e-01 -3.72201234e-01 -1.20531154e+00
-2.30791286e-01 -8.10425699e-01 -1.28830874e+00 9.28147361e-02
-2.69665942e-03 4.18302894e-01 -6.91712439e-01 1.58656812e+00
1.26318231e-01 6.66904896e-02 2.77010575e-02 6.52967870e-01
7.64435351e-01 6.61026061e-01 -3.66298199e-01 -4.32326019e-01
1.19648182e+00 -8.65585089e-01 -8.04452717e-01 -1.32092193e-01
6.57145500e-01 -6.67981386e-01 1.01147521e+00 7.67812431e-01
-1.33070266e+00 -3.24255288e-01 -1.59509814e+00 -2.16727301e-01
-2.09164340e-02 2.43291005e-01 1.02231002e+00 6.80370331e-01
-1.10082698e+00 1.20641327e+00 -8.99795294e-01 -4.63896878e-02
5.32984495e-01 6.74967766e-01 -4.30828333e-01 -3.87022823e-01
-1.04834390e+00 5.73113024e-01 5.97671092e-01 -3.18126470e-01
-6.72675729e-01 -6.70771837e-01 -5.11939347e-01 5.00179529e-01
3.21018785e-01 -7.05732763e-01 1.15604603e+00 -3.71520787e-01
-1.51273727e+00 5.95847011e-01 -1.25127941e-01 -9.94119644e-01
4.93100017e-01 -5.09809852e-01 -6.18232153e-02 6.41507924e-01
-1.84073478e-01 5.50736725e-01 8.15629780e-01 -7.48250306e-01
-5.66078365e-01 -3.57226342e-01 -1.77922696e-01 2.51004070e-01
-9.05346811e-01 -1.21534161e-01 -7.48923898e-01 -8.83731902e-01
5.06589353e-01 -8.14366996e-01 -3.98057342e-01 9.09285173e-02
-6.65723905e-02 -9.09102857e-02 6.24472022e-01 -4.15396720e-01
1.89577103e+00 -2.11227083e+00 3.16580802e-01 2.33409941e-01
4.72851753e-01 4.53877509e-01 7.16149658e-02 3.15504044e-01
1.13869354e-01 1.85099527e-01 1.60346478e-02 -3.39647889e-01
-2.31180981e-01 1.70223445e-01 -2.70740509e-01 5.20463884e-01
-2.72154748e-01 4.16141897e-01 -6.83799922e-01 -6.21557474e-01
-1.03192581e-02 6.28900826e-01 -8.84627759e-01 -2.30897721e-02
1.74864545e-01 -1.82751507e-01 -2.22294807e-01 2.85628319e-01
6.86026514e-01 -3.47068638e-01 1.21155009e-01 -4.63250995e-01
1.09480195e-01 4.88055110e-01 -1.43758070e+00 1.74135327e+00
-3.43405813e-01 6.63012385e-01 2.11849421e-01 -9.98640716e-01
8.88553381e-01 7.02867359e-02 6.02890134e-01 -5.92927575e-01
2.82480597e-01 5.31414688e-01 -1.37262151e-01 8.95010233e-02
7.53993630e-01 1.98674984e-02 2.51769453e-01 4.40443844e-01
4.42216285e-02 5.41638173e-02 6.63410366e-01 5.83619595e-01
1.39168310e+00 -5.02697527e-01 2.97649056e-01 -4.88507539e-01
5.29963732e-01 -2.95305252e-01 5.53726315e-01 7.51876712e-01
2.50859380e-01 6.69397414e-01 3.26323718e-01 -3.05686712e-01
-1.12189186e+00 -6.73829436e-01 -3.87975007e-01 9.23355401e-01
4.86605465e-02 -1.06114900e+00 -6.31753862e-01 -3.20280530e-02
-2.03514546e-01 3.00406128e-01 -1.79668039e-01 -2.30103657e-01
-5.72635353e-01 -6.01090372e-01 4.62068111e-01 5.56423366e-01
5.55195749e-01 -5.66654265e-01 -8.94602418e-01 1.99767590e-01
7.26532266e-02 -1.02646673e+00 -1.43604517e-01 6.69216633e-01
-1.54003227e+00 -7.89807379e-01 -5.59583426e-01 -4.41909283e-01
4.42249507e-01 4.88581896e-01 1.20877099e+00 3.77368122e-01
-1.74272865e-01 -3.12521577e-01 -6.01051271e-01 -6.83390796e-02
-2.22940922e-01 3.53354692e-01 2.05851704e-01 -4.14709240e-01
4.49588038e-02 -1.02332175e+00 -8.46785486e-01 1.60533279e-01
-1.16301548e+00 1.70982972e-01 9.12336409e-01 9.56381619e-01
9.21060264e-01 5.69799721e-01 3.21016550e-01 -8.73587728e-01
6.34042561e-01 1.08469985e-01 -6.43429518e-01 7.86590565e-04
-1.00334466e+00 3.70570719e-01 6.76592290e-01 -3.37828040e-01
-3.35535258e-01 2.08129808e-01 -1.65694669e-01 -6.96279466e-01
3.62077028e-01 6.98385537e-01 -2.72650998e-02 -9.46036577e-02
6.57091975e-01 2.73747861e-01 -6.85564280e-02 -6.45421922e-01
3.85360092e-01 5.56589544e-01 5.25310040e-01 -4.85267699e-01
7.74620831e-01 3.70247006e-01 3.99014741e-01 -7.96667993e-01
-7.64234006e-01 -4.76264477e-01 -5.76090097e-01 1.20095119e-01
3.35734367e-01 -9.56552207e-01 -4.04626548e-01 1.10496357e-01
-7.44960010e-01 -7.86522180e-02 -3.46868008e-01 3.69048625e-01
-6.61000967e-01 5.30824900e-01 -7.25103140e-01 -5.39744377e-01
-8.12520385e-01 -1.15877473e+00 8.72505724e-01 -1.89039167e-02
2.06003096e-02 -3.84812593e-01 -2.28168786e-01 2.99383998e-01
4.84386235e-01 1.78694725e-03 1.07212222e+00 -4.63909835e-01
-6.30535901e-01 -5.05235016e-01 -3.37698847e-01 1.94296837e-01
-3.46547633e-01 -2.93023497e-01 -5.78123510e-01 -6.59287751e-01
5.79404049e-02 -4.39587355e-01 1.17228329e+00 3.36798131e-01
1.52152407e+00 -4.17270035e-01 -2.71915317e-01 8.85032117e-01
1.56582808e+00 5.84773757e-02 6.44775987e-01 4.16931659e-01
3.96658182e-01 -3.40661630e-02 6.52942061e-01 6.50303185e-01
-2.79977858e-01 7.11613655e-01 3.42792630e-01 1.99140817e-01
-1.31476879e-01 -1.60656959e-01 -4.24000211e-02 1.10526145e+00
2.19111424e-02 -9.83901136e-03 -7.82143354e-01 3.70451987e-01
-1.88666093e+00 -9.05456424e-01 7.05770552e-02 2.56307101e+00
9.53573823e-01 3.52209747e-01 -4.74508852e-02 1.01235580e+00
4.18899477e-01 1.52221709e-01 -4.70410228e-01 -2.78266966e-01
-4.89773676e-02 4.44738030e-01 8.11647296e-01 2.38769785e-01
-9.96229649e-01 7.61462450e-01 6.83013630e+00 1.36201620e+00
-8.42927217e-01 1.34555772e-01 6.07036829e-01 -5.24030805e-01
-1.89776644e-01 -8.43642801e-02 -1.08292234e+00 2.15123788e-01
1.13239563e+00 -2.56906390e-01 7.26452768e-01 1.11536121e+00
7.55845150e-03 -7.39843622e-02 -1.12041140e+00 1.56168401e+00
-3.15609165e-02 -1.62913620e+00 1.60305545e-01 2.41934538e-01
6.41688228e-01 -5.53235970e-03 1.28555655e-01 2.11891755e-02
-7.71764442e-02 -1.18081522e+00 5.72597384e-01 -4.99785990e-02
1.12509322e+00 -1.19615901e+00 7.49816477e-01 3.24516803e-01
-1.24738753e+00 -2.98095375e-01 -8.19047749e-01 1.11106247e-01
2.49819368e-01 1.20410502e+00 -1.75432697e-01 3.46231014e-01
9.54877317e-01 5.69833577e-01 -5.64800560e-01 1.33614528e+00
1.48041666e-01 5.47595620e-01 -7.19735563e-01 4.95592616e-02
1.08908564e-01 -1.38517767e-01 4.80403334e-01 1.16722286e+00
6.72021031e-01 1.63675636e-01 -5.20982035e-02 2.89457411e-01
-3.28829288e-01 2.40458265e-01 -3.57594341e-01 -2.89203972e-01
6.43080473e-01 1.05331981e+00 -8.62072289e-01 -4.01984543e-01
-1.01730198e-01 8.05684149e-01 3.84754092e-01 -3.53565097e-01
-6.26614451e-01 -4.61854964e-01 4.68126684e-01 4.59840178e-01
6.09732449e-01 -4.18314517e-01 -4.80358213e-01 -1.00479746e+00
2.56375819e-02 -9.16132629e-01 3.00208271e-01 -3.01269948e-01
-7.76457131e-01 7.39390314e-01 9.55682993e-02 -1.29555666e+00
7.09728599e-02 -5.11730134e-01 -1.83391455e-03 4.29082036e-01
-1.47399819e+00 -4.14912075e-01 -2.64533937e-01 4.06352431e-01
5.65621972e-01 -3.77336890e-01 7.72058308e-01 7.43607461e-01
-3.34037006e-01 1.00936532e+00 4.40374672e-01 -4.70828980e-01
3.40075195e-01 -9.64056253e-01 -2.62646936e-02 7.14450121e-01
9.11162570e-02 7.00649202e-01 6.92354500e-01 -4.56495851e-01
-1.64391971e+00 -8.95405114e-01 7.36718476e-01 3.78490955e-01
5.28982222e-01 -8.16882625e-02 -8.83643329e-01 -8.68351012e-03
-9.48054641e-02 -2.40231901e-02 8.56436729e-01 1.79301634e-01
-3.19527864e-01 -5.13886213e-01 -1.01229024e+00 5.50538957e-01
1.16015220e+00 -3.00788730e-01 -3.40105504e-01 4.94788557e-01
7.61354446e-01 -4.10591453e-01 -8.62785518e-01 6.24371529e-01
6.27159595e-01 -1.16559708e+00 1.13522458e+00 -1.79898351e-01
7.45923877e-01 -1.95141524e-01 -6.24486864e-01 -7.53906369e-01
-6.15415037e-01 -8.33688438e-01 -7.75257885e-01 9.95083392e-01
2.35712394e-01 -2.61529349e-02 1.17715800e+00 3.47043455e-01
3.16146947e-02 -1.15531623e+00 -1.14619458e+00 -1.00345910e+00
-2.06027508e-01 -4.91538525e-01 4.73542660e-01 3.75583857e-01
-1.00099765e-01 5.10711253e-01 -4.90764856e-01 -2.90735185e-01
6.76178396e-01 1.07417747e-01 5.89269221e-01 -1.53662527e+00
-6.70340836e-01 -6.65538967e-01 -5.31048000e-01 -1.33895695e+00
-2.83283383e-01 -9.35946047e-01 -2.41764680e-01 -1.56744707e+00
3.56699735e-01 -6.88497186e-01 -4.11635786e-01 3.32614034e-01
1.23312175e-01 3.53975326e-01 3.40366989e-01 4.03127879e-01
-5.95327377e-01 6.79957867e-01 1.05118287e+00 1.19201494e-02
-4.18292820e-01 -6.35999143e-02 -7.75729716e-01 5.39715886e-01
7.85441101e-01 -6.11740649e-01 -6.39642656e-01 -4.82232630e-01
6.49271727e-01 2.06646577e-01 -2.94111401e-01 -1.40353465e+00
4.15619403e-01 2.49922484e-01 1.14082851e-01 -1.00290716e+00
4.59570169e-01 -8.65764856e-01 1.84144288e-01 6.46953285e-01
-3.93470973e-01 3.52461129e-01 8.64165351e-02 7.14962721e-01
-3.20277900e-01 -6.03217602e-01 1.07976532e+00 -5.58221117e-02
-2.47147977e-01 3.64447236e-01 -2.02913120e-01 -1.92072496e-01
7.19991982e-01 -4.20655906e-01 -3.93707938e-02 -3.23501527e-01
-2.68928468e-01 -1.21880807e-01 3.51323724e-01 1.08610779e-01
8.61919641e-01 -1.28578329e+00 -4.65226233e-01 1.80590034e-01
-2.60731220e-01 1.85196593e-01 1.42163336e-01 6.46093607e-01
-9.02437210e-01 5.64543605e-01 -4.10534292e-02 -5.04016638e-01
-1.64166820e+00 6.04616106e-01 -2.47963354e-01 -7.47941971e-01
-9.23614621e-01 8.03544998e-01 -3.19427431e-01 3.02968413e-01
6.13951623e-01 -4.47409041e-03 -1.36003956e-01 5.11336029e-02
9.18477774e-01 5.55783749e-01 4.38660383e-01 -2.96941996e-01
-7.03237727e-02 5.55587471e-01 -1.95478946e-01 -7.56189674e-02
1.66728532e+00 1.34635076e-01 -3.04724872e-01 9.65655074e-02
1.23580432e+00 -1.87785447e-01 -7.77660370e-01 -3.03259134e-01
-2.58730263e-01 -7.35462248e-01 6.90308332e-01 -3.79088163e-01
-1.47773135e+00 8.92264307e-01 7.89978445e-01 1.01289876e-01
1.54479134e+00 -3.15398663e-01 1.15408242e+00 5.83304226e-01
4.72819746e-01 -1.10921812e+00 7.21119642e-02 4.77919132e-01
6.06543660e-01 -9.41160798e-01 6.86071932e-01 -4.52687323e-01
-9.67330933e-02 1.15925491e+00 2.00670421e-01 -1.52076855e-01
9.19086874e-01 4.75109041e-01 -5.05637169e-01 -1.26691848e-01
-8.94416332e-01 9.30975750e-02 2.95752525e-01 2.13455722e-01
4.26575452e-01 -5.91359613e-03 -1.01222670e+00 4.17560190e-01
-7.51645148e-01 -1.63922980e-02 2.10247576e-01 7.94273853e-01
-7.89281189e-01 -1.24850655e+00 -1.29972592e-01 1.01358485e+00
-6.39560878e-01 -2.51680017e-01 -6.16456266e-04 1.93240047e-01
-1.89667195e-01 8.70217204e-01 -6.77739233e-02 -7.22225010e-01
-3.48476544e-02 -2.57636547e-01 4.28591400e-01 -4.77706343e-01
-4.57428426e-01 1.20866790e-01 -1.71014652e-01 -9.38656926e-01
-2.21838385e-01 -3.92095059e-01 -1.03674281e+00 -6.31134629e-01
-5.48579514e-01 2.01375932e-01 7.17685699e-01 7.29065597e-01
4.31844711e-01 3.67204696e-01 3.65457714e-01 -7.86570191e-01
-6.38061225e-01 -6.52565181e-01 -6.06380701e-01 2.50899196e-01
8.43944475e-02 -6.87818289e-01 -3.02403003e-01 -3.32690805e-01] | [8.511474609375, 3.243786334991455] |
5ac06e6e-faa0-429e-9bf3-e28bcd35685b | causal-discovery-from-temporal-data-an | 2303.10112 | null | https://arxiv.org/abs/2303.10112v2 | https://arxiv.org/pdf/2303.10112v2.pdf | Causal Discovery from Temporal Data: An Overview and New Perspectives | Temporal data, representing chronological observations of complex systems, has always been a typical data structure that can be widely generated by many domains, such as industry, medicine and finance. Analyzing this type of data is extremely valuable for various applications. Thus, different temporal data analysis tasks, eg, classification, clustering and prediction, have been proposed in the past decades. Among them, causal discovery, learning the causal relations from temporal data, is considered an interesting yet critical task and has attracted much research attention. Existing casual discovery works can be divided into two highly correlated categories according to whether the temporal data is calibrated, ie, multivariate time series casual discovery, and event sequence casual discovery. However, most previous surveys are only focused on the time series casual discovery and ignore the second category. In this paper, we specify the correlation between the two categories and provide a systematical overview of existing solutions. Furthermore, we provide public datasets, evaluation metrics and new perspectives for temporal data casual discovery. | ['Jingping Bi', 'Wenbin Li', 'Chuzhe Zhang', 'Di Yao', 'Chang Gong'] | 2023-03-17 | null | null | null | null | ['causal-discovery'] | ['knowledge-base'] | [ 1.22912891e-01 -4.70867425e-01 -6.35634899e-01 -2.80546278e-01
-1.94020212e-01 -5.31213164e-01 9.84826982e-01 5.35387933e-01
2.94014625e-02 9.34300065e-01 4.44316626e-01 -5.46230495e-01
-1.00652313e+00 -6.88491523e-01 -2.78207362e-01 -9.04539943e-01
-6.58628106e-01 3.59234631e-01 4.59167480e-01 -3.48686911e-02
3.12476635e-01 2.57141143e-01 -1.49897683e+00 2.60716498e-01
8.16325307e-01 1.15492547e+00 -3.99761572e-02 7.39295408e-02
-1.11560084e-01 8.37971509e-01 -4.67174679e-01 -5.01932651e-02
-9.73075442e-03 -4.69494879e-01 -5.12238204e-01 -2.67546594e-01
-6.50532067e-01 7.99400955e-02 -4.74049330e-01 6.60061717e-01
1.07317030e-01 7.73986205e-02 5.62731862e-01 -1.73610210e+00
-5.71396470e-01 9.40646410e-01 -8.97576630e-01 6.77343667e-01
2.28775352e-01 8.73287171e-02 9.27844942e-01 -4.15539324e-01
4.65264916e-01 1.41699743e+00 5.09102762e-01 3.19332406e-02
-1.06251609e+00 -8.67654145e-01 3.59272212e-01 7.64733315e-01
-1.14528048e+00 4.48503196e-02 9.29912806e-01 -7.80618787e-01
2.51224756e-01 3.78807157e-01 4.99329120e-01 1.15736294e+00
3.20442975e-01 7.87162185e-01 1.23925829e+00 -1.04119651e-01
2.17371374e-01 -2.57004201e-01 7.73992315e-02 -1.30686909e-01
-4.30992395e-02 2.67615139e-01 -4.51652914e-01 -3.16736609e-01
6.54500127e-01 4.82217818e-01 -3.04339737e-01 -3.78920212e-02
-1.53162432e+00 6.94356799e-01 2.83114135e-01 6.40377581e-01
-4.74946558e-01 2.37316340e-02 6.29692018e-01 3.50539505e-01
5.39049208e-01 4.55181934e-02 -7.07724452e-01 -2.69142449e-01
-5.97521901e-01 4.46671695e-01 5.07359862e-01 7.39115059e-01
3.57676297e-02 -1.58847436e-01 -1.97525844e-01 5.58881104e-01
1.77304164e-01 4.14507687e-01 5.07867396e-01 -3.00584316e-01
3.87504309e-01 7.27011383e-01 1.84450001e-01 -1.22741389e+00
-5.29244423e-01 -3.04440409e-01 -1.28565562e+00 -5.22367299e-01
3.84308338e-01 -4.23024930e-02 -5.81908286e-01 1.70180929e+00
4.83530313e-01 7.33980000e-01 -1.72471046e-01 7.60808468e-01
6.42752111e-01 8.05869043e-01 2.11131781e-01 -9.85535502e-01
1.19935369e+00 -2.77962029e-01 -9.92142022e-01 4.81084228e-01
3.18689980e-02 -8.18530321e-01 5.20498574e-01 3.27568680e-01
-5.72075188e-01 -2.27920383e-01 -6.53318584e-01 3.89243960e-01
-5.09030223e-01 -3.67675155e-01 1.01298344e+00 2.49948129e-02
-5.24256885e-01 6.83953762e-01 -9.87359345e-01 -6.07110798e-01
1.16814084e-01 1.10582568e-01 -1.49550848e-02 -4.66761738e-02
-1.70004559e+00 4.52059448e-01 5.64903319e-01 1.69244148e-02
-8.43587160e-01 -7.69978881e-01 -3.24139744e-01 -2.23402858e-01
5.44345915e-01 -2.63526320e-01 1.16044915e+00 -2.27476671e-01
-6.55698061e-01 3.37930888e-01 -2.07967505e-01 -4.82202768e-01
5.45362771e-01 1.04133330e-01 -1.15108550e+00 -4.18056369e-01
2.25504771e-01 -4.03905660e-01 5.58872461e-01 -7.82571495e-01
-1.07793367e+00 -3.65684271e-01 -6.79617003e-02 -2.55481541e-01
-3.51569146e-01 2.37947017e-01 -2.43247330e-01 -9.20911729e-01
9.64095443e-02 -7.02828050e-01 -5.02391577e-01 -4.54009622e-01
-5.95807672e-01 -8.57298553e-01 1.04754543e+00 -3.52847248e-01
1.90484047e+00 -2.20136380e+00 1.38645291e-01 1.88256323e-01
2.34037295e-01 -2.75001794e-01 3.62010211e-01 8.77406836e-01
-5.26781261e-01 3.03131849e-01 -2.69040346e-01 1.77258089e-01
-1.95973665e-01 2.22970918e-01 -9.07048225e-01 4.25392419e-01
-9.13848877e-02 5.67356527e-01 -1.12397456e+00 -3.63599569e-01
2.92442381e-01 1.32047385e-01 6.94084242e-02 3.33942413e-01
-3.99724424e-01 8.58958662e-01 -7.60669768e-01 5.28196514e-01
3.23409170e-01 -5.35399973e-01 2.21510053e-01 -1.46882802e-01
-6.82572663e-01 9.45536718e-02 -9.61070538e-01 1.22734594e+00
-1.03398867e-01 3.92151117e-01 -5.15166938e-01 -1.10788989e+00
8.98633301e-01 7.52631068e-01 1.26483047e+00 -7.25814760e-01
4.57428508e-02 3.56489301e-01 2.18268573e-01 -7.18970001e-01
3.03418696e-01 -1.69616073e-01 -3.43749285e-01 5.99663794e-01
-6.87104046e-01 5.05463719e-01 4.96108532e-01 -9.67474878e-02
1.26139867e+00 -2.79932231e-01 5.95751882e-01 -1.61511987e-01
3.77298862e-01 3.32563043e-01 9.13910687e-01 1.92736030e-01
-1.59135804e-01 2.28332564e-01 6.04622185e-01 -7.43619025e-01
-8.47227931e-01 -1.12860262e+00 -4.73073035e-01 5.87040842e-01
3.44297200e-01 -4.57499206e-01 4.14547138e-02 -4.01545316e-01
2.27377057e-01 3.90375882e-01 -7.56399870e-01 -1.00121744e-01
-6.60554171e-01 -1.29929411e+00 2.81701297e-01 5.49839199e-01
3.78675222e-01 -1.09380984e+00 -3.60005915e-01 4.54663813e-01
-4.10157263e-01 -7.64785945e-01 -2.02333376e-01 7.47672021e-02
-8.76569152e-01 -1.26093173e+00 -4.46178973e-01 -2.73068756e-01
6.08162135e-02 3.26358765e-01 1.00806975e+00 -8.03876221e-02
-3.80553663e-01 -4.57001366e-02 -6.18423820e-01 -5.77499449e-01
7.27130026e-02 -4.42716554e-02 4.71489489e-01 3.07652503e-01
5.93205333e-01 -1.10185707e+00 -6.59489870e-01 6.13700390e-01
-9.35196519e-01 -1.50197580e-01 6.76010430e-01 5.65755308e-01
7.35623479e-01 7.01852739e-01 8.89381409e-01 -6.82090938e-01
5.31579375e-01 -1.22132552e+00 -4.85536456e-01 2.58466542e-01
-6.69948995e-01 2.49337424e-02 7.30506301e-01 -6.00421727e-01
-9.56590176e-01 -2.00565845e-01 2.60862857e-01 -3.88551116e-01
-1.88356549e-01 1.18438470e+00 -8.38244855e-02 6.51396692e-01
4.47072864e-01 2.22385213e-01 -4.76558417e-01 -6.63827419e-01
1.69859186e-01 5.35104394e-01 4.72011328e-01 -5.97144306e-01
9.57452714e-01 5.55002511e-01 9.47525054e-02 -4.90936667e-01
-7.20926404e-01 -8.38706017e-01 -7.13117003e-01 -2.76096880e-01
7.44764507e-01 -5.51034212e-01 -7.44530022e-01 2.37814009e-01
-1.25743222e+00 -4.57903557e-02 -7.00945854e-02 6.78535283e-01
-2.66213804e-01 -1.82986468e-01 -3.57874900e-01 -9.25281703e-01
1.81824088e-01 -6.97345257e-01 7.54034758e-01 1.15691476e-01
-4.00857329e-01 -1.23002231e+00 3.43087852e-01 -2.20412627e-01
3.63123715e-01 9.73184645e-01 1.16991735e+00 -4.70807016e-01
-7.69970357e-01 -5.84698096e-02 -3.55338126e-01 -5.88779509e-01
5.34673810e-01 2.82430556e-02 -6.06857061e-01 5.24931774e-02
-9.13420320e-02 5.34332879e-02 7.08358765e-01 4.18521881e-01
1.36782646e+00 -2.97790647e-01 -9.22944903e-01 2.36047059e-01
1.14172423e+00 7.92016804e-01 3.95234674e-01 2.28234783e-01
5.45329452e-01 8.07917535e-01 1.09394360e+00 8.50345671e-01
2.65209854e-01 6.94117069e-01 4.93040293e-01 -5.12774661e-02
4.55089033e-01 -2.12691411e-01 -2.87606232e-02 8.70161712e-01
-5.06482124e-01 -2.58959532e-01 -1.00953400e+00 7.79195189e-01
-2.43802619e+00 -1.37768841e+00 -7.16538429e-01 2.00820255e+00
1.03854120e+00 -1.73914921e-03 2.82728493e-01 5.27753174e-01
7.99246669e-01 2.96497047e-01 -7.21194625e-01 1.92514479e-01
-1.20491415e-01 -1.86110318e-01 3.65741372e-01 -6.34129643e-02
-1.28080297e+00 4.76659268e-01 6.14112997e+00 8.91152918e-01
-1.10771549e+00 7.99925923e-02 6.29786551e-01 7.61937490e-03
-2.39714429e-01 9.20287371e-02 -4.99805093e-01 7.42368340e-01
9.99796391e-01 -8.09742987e-01 1.34845093e-01 6.06349349e-01
8.89860630e-01 1.55684859e-01 -1.05315924e+00 1.01985371e+00
-6.03558362e-01 -1.16836786e+00 4.60760631e-02 1.39056385e-01
7.72359669e-01 -3.42777252e-01 9.34765711e-02 4.80591916e-02
5.65990031e-01 -1.13168931e+00 4.34068829e-01 6.77484393e-01
6.39313817e-01 -7.24174559e-01 6.72253370e-01 4.19109106e-01
-1.84094346e+00 -2.82436520e-01 1.52617134e-02 -2.96286285e-01
7.10479915e-01 1.25953376e+00 -2.52018213e-01 1.01147115e+00
1.17671955e+00 1.44929039e+00 -2.59152889e-01 1.30529594e+00
-8.69844779e-02 9.39670503e-01 -2.37243861e-01 2.89340988e-02
-2.51250658e-02 -1.53653607e-01 4.95121717e-01 6.96301341e-01
4.93856221e-01 2.52262831e-01 4.52250123e-01 5.58297157e-01
2.36872405e-01 -3.05142775e-02 -5.06551683e-01 -1.18069224e-01
6.70987189e-01 8.58251631e-01 -7.23801196e-01 -1.53147489e-01
-2.76948959e-01 4.01997149e-01 -1.19396403e-01 1.60168394e-01
-1.16402698e+00 -9.76145789e-02 7.00631618e-01 2.67248869e-01
-2.49328405e-01 -3.36530656e-01 -2.95714945e-01 -1.02657032e+00
5.68766519e-03 -6.54345870e-01 9.27694023e-01 -4.16265875e-01
-1.72229052e+00 4.18016702e-01 5.75736105e-01 -1.72414100e+00
-1.57074884e-01 -3.15982729e-01 -5.82726538e-01 7.76353776e-01
-1.22421420e+00 -8.96272957e-01 -3.88803989e-01 9.25249994e-01
4.78528351e-01 2.03148708e-01 4.55758303e-01 6.37374461e-01
-6.76159680e-01 -1.49075473e-02 2.75833607e-01 -9.48299281e-03
7.42786586e-01 -1.24565208e+00 6.29135072e-02 7.42668331e-01
-2.56887436e-01 9.13543999e-01 9.56732929e-01 -7.84334481e-01
-1.14547265e+00 -1.45793903e+00 1.05544317e+00 -3.60394925e-01
1.32776940e+00 -7.15214312e-02 -9.76412535e-01 6.19921565e-01
1.11801885e-01 -1.29560038e-01 5.01386464e-01 2.82247245e-01
-1.97539926e-01 -4.34056133e-01 -5.70393562e-01 6.63619220e-01
1.25061357e+00 -1.50809646e-01 -5.51737130e-01 5.14975011e-01
9.17339742e-01 -1.95784234e-02 -1.08921874e+00 5.13051987e-01
4.11212057e-01 -8.26491356e-01 8.81073177e-01 -6.15496635e-01
4.68925565e-01 -4.96657044e-01 7.67829195e-02 -1.20918941e+00
-4.78749692e-01 -5.90558946e-01 -2.56800234e-01 1.48369277e+00
2.01874480e-01 -6.31608784e-01 4.03632104e-01 2.65738696e-01
-4.65450324e-02 -5.93086243e-01 -9.61096287e-01 -1.06775594e+00
-2.93552130e-01 -5.16185582e-01 8.74033868e-01 1.40115821e+00
1.52881920e-01 3.69200528e-01 -7.49448895e-01 1.21237099e-01
6.04596317e-01 6.97056770e-01 6.09658360e-01 -1.68650007e+00
-8.67310166e-02 -6.23129249e-01 -3.82969141e-01 -1.02502751e+00
-3.98251355e-01 -5.65330327e-01 -2.95725763e-02 -1.47953033e+00
1.40278324e-01 -7.89915919e-01 -5.44292510e-01 4.92292434e-01
-1.02794133e-01 -1.88304752e-01 -5.75349867e-01 7.32412815e-01
-4.34119880e-01 5.19948125e-01 9.65297937e-01 -1.19805552e-01
-3.05106252e-01 2.81211495e-01 -3.89433652e-01 5.66507459e-01
6.96560323e-01 -5.13094962e-01 -7.11294651e-01 -2.11315528e-02
2.59102434e-01 4.03404534e-01 3.97810489e-01 -6.64661109e-01
4.01636720e-01 -9.69131172e-01 -1.26289874e-01 -9.50966418e-01
-1.19530685e-01 -1.07525516e+00 7.08869159e-01 5.70992291e-01
-2.98957616e-01 3.96883994e-01 6.70339912e-02 1.04729664e+00
-6.21150017e-01 5.34980357e-01 4.54439878e-01 7.08335638e-02
-1.00766921e+00 9.19959843e-01 -1.92187995e-01 -1.26803786e-01
1.33603418e+00 2.63642102e-01 -4.17314649e-01 -2.44807109e-01
-6.11266613e-01 5.28909624e-01 -2.06705585e-01 7.67939031e-01
4.26696151e-01 -1.57808137e+00 -7.32750893e-01 -4.67515588e-01
2.96625406e-01 -3.45184840e-02 2.54358530e-01 1.41908908e+00
1.08825505e-01 6.74654186e-01 1.63184047e-01 -7.28609383e-01
-1.08622956e+00 1.14455390e+00 -1.74980313e-01 -5.87799788e-01
-6.02966130e-01 1.76417530e-01 2.43176013e-01 5.91270290e-02
1.30475432e-01 -3.42150092e-01 -6.74894273e-01 3.08662295e-01
7.43134499e-01 4.17200595e-01 -2.61485845e-01 -3.80169272e-01
-5.63768387e-01 4.66893256e-01 1.59939244e-01 1.43091977e-01
1.44344926e+00 -1.21552520e-01 -3.83309156e-01 1.16156387e+00
8.59170556e-01 -2.28935301e-01 -8.64364624e-01 -4.20819461e-01
6.27328336e-01 -3.60251755e-01 -4.02942389e-01 -5.67540526e-01
-8.78806293e-01 7.50152588e-01 2.62444526e-01 9.61709023e-01
1.35526037e+00 2.88151830e-01 7.27421641e-01 -1.13993764e-01
6.18657470e-01 -6.58495665e-01 -6.51914328e-02 4.36563462e-01
9.93935227e-01 -1.20305061e+00 -1.44802127e-02 -6.55995429e-01
-7.38036335e-02 9.38279986e-01 2.72181928e-01 7.34196082e-02
1.11121261e+00 7.71590918e-02 -2.72216022e-01 -1.89105660e-01
-9.62626576e-01 -2.90766597e-01 3.96040082e-01 3.88978302e-01
6.36926472e-01 2.38207057e-01 -7.22325623e-01 8.25173259e-01
-2.29605421e-01 6.40897006e-02 7.68182799e-03 6.73300266e-01
-1.04595564e-01 -9.98443663e-01 -4.86694932e-01 6.66776955e-01
-4.49325562e-01 -1.30446807e-01 -2.20247537e-01 9.69542503e-01
2.82202154e-01 1.24995995e+00 1.15819864e-01 -5.41542470e-01
4.25858170e-01 -3.64225507e-01 -1.23468973e-01 -3.60339463e-01
-1.79456830e-01 -8.86753201e-02 -2.00149536e-01 -5.12170315e-01
-7.29526997e-01 -9.67803121e-01 -1.22094679e+00 -6.00601971e-01
-3.21505517e-02 3.80501807e-01 2.11915150e-01 1.13688517e+00
1.13475777e-01 7.90339231e-01 8.88830543e-01 -2.58996993e-01
-3.13484184e-02 -9.54156697e-01 -6.46582484e-01 4.05131757e-01
3.44795316e-01 -1.09290957e+00 -5.83165467e-01 3.67323428e-01] | [7.671210765838623, 5.13963508605957] |
b4fade84-d6d2-4465-be12-f5f2f20486fa | multi-criterion-evolutionary-design-of-deep | 1912.01369 | null | https://arxiv.org/abs/1912.01369v3 | https://arxiv.org/pdf/1912.01369v3.pdf | Multi-Objective Evolutionary Design of Deep Convolutional Neural Networks for Image Classification | Early advancements in convolutional neural networks (CNNs) architectures are primarily driven by human expertise and by elaborate design processes. Recently, neural architecture search was proposed with the aim of automating the network design process and generating task-dependent architectures. While existing approaches have achieved competitive performance in image classification, they are not well suited to problems where the computational budget is limited for two reasons: (1) the obtained architectures are either solely optimized for classification performance, or only for one deployment scenario; (2) the search process requires vast computational resources in most approaches. To overcome these limitations, we propose an evolutionary algorithm for searching neural architectures under multiple objectives, such as classification performance and floating-point operations (FLOPs). The proposed method addresses the first shortcoming by populating a set of architectures to approximate the entire Pareto frontier through genetic operations that recombine and modify architectural components progressively. Our approach improves computational efficiency by carefully down-scaling the architectures during the search as well as reinforcing the patterns commonly shared among past successful architectures through Bayesian model learning. The integration of these two main contributions allows an efficient design of architectures that are competitive and in most cases outperform both manually and automatically designed architectures on benchmark image classification datasets: CIFAR, ImageNet, and human chest X-ray. The flexibility provided from simultaneously obtaining multiple architecture choices for different compute requirements further differentiates our approach from other methods in the literature. Code is available at https://github.com/mikelzc1990/nsganetv1 | ['Vishnu Naresh Boddeti', 'Erik Goodman', 'Zhichao Lu', 'Yashesh Dhebar', 'Wolfgang Banzhaf', 'Kalyanmoy Deb', 'Ian Whalen'] | 2019-12-03 | null | null | null | null | ['pneumonia-detection'] | ['medical'] | [ 1.85254544e-01 -6.48577288e-02 -1.87655568e-01 -3.75983357e-01
-2.72673875e-01 -4.63268429e-01 3.98494482e-01 -2.85583902e-02
-6.44277573e-01 6.65004909e-01 -3.38150203e-01 -4.43068802e-01
-5.95357835e-01 -7.18481600e-01 -5.26641786e-01 -7.25107491e-01
2.06706285e-01 5.44365227e-01 1.27239570e-01 -1.01211905e-01
4.84195381e-01 6.65154338e-01 -1.86513293e+00 9.68325213e-02
9.01821971e-01 1.21003389e+00 3.82198125e-01 5.82040668e-01
-3.27086419e-01 4.08787519e-01 -5.37702978e-01 -4.09170836e-01
3.46764833e-01 -3.73515427e-01 -6.79061830e-01 -1.35940716e-01
1.16108609e-02 1.60974726e-01 8.42261594e-03 1.02832568e+00
5.87763488e-01 -7.28051588e-02 5.84713876e-01 -1.15550351e+00
-3.18077475e-01 7.94087946e-01 -3.22843373e-01 1.72892883e-01
-4.27600771e-01 1.08315520e-01 9.49781358e-01 -7.18920708e-01
1.41427919e-01 8.42894495e-01 5.48673272e-01 7.43284523e-01
-1.18797970e+00 -6.64697409e-01 2.43793949e-01 4.13079590e-01
-1.67457926e+00 -4.72807109e-01 8.46510172e-01 -5.22456050e-01
1.11796665e+00 3.36028904e-01 7.83448040e-01 9.87136304e-01
7.82482028e-02 6.40292227e-01 6.96446896e-01 -6.19016647e-01
6.63399756e-01 1.31390333e-01 3.52897197e-02 6.80093348e-01
4.92934346e-01 2.12146565e-01 -3.09712917e-01 -1.01887859e-01
5.37719011e-01 -1.26046389e-01 -2.19863251e-01 -4.73045260e-01
-8.16426516e-01 8.29062879e-01 4.18278396e-01 6.44501209e-01
-5.93092442e-01 2.67590016e-01 2.59466350e-01 2.30821759e-01
-3.03671025e-02 9.07928169e-01 -6.34184957e-01 5.48707088e-03
-1.17834628e+00 2.01120168e-01 6.75875902e-01 7.79863477e-01
8.37233722e-01 4.11074996e-01 6.59061503e-03 8.87662709e-01
2.90711284e-01 -5.54693900e-02 7.75914311e-01 -6.96382105e-01
2.28661969e-01 7.36994982e-01 -2.28848681e-01 -9.27407444e-01
-4.66297954e-01 -1.00753701e+00 -9.46994722e-01 2.90941417e-01
2.90501833e-01 -3.57444376e-01 -9.89813268e-01 1.67984533e+00
1.17192343e-01 -1.64435208e-01 -4.65091914e-02 7.21991420e-01
6.18810713e-01 3.53783786e-01 9.53559503e-02 -1.11199878e-02
1.40062606e+00 -9.80410874e-01 -4.24916923e-01 -3.82295072e-01
4.45668668e-01 -6.02682233e-01 7.05307007e-01 5.03959119e-01
-1.06600869e+00 -6.13410532e-01 -1.39293718e+00 4.08809990e-01
-2.97767252e-01 5.91067493e-01 4.56471801e-01 1.08849406e+00
-1.24726570e+00 5.17400563e-01 -8.54532719e-01 -1.09602042e-01
6.99542522e-01 7.31211662e-01 2.72417784e-01 3.13437998e-01
-8.55493009e-01 7.98776746e-01 8.26146007e-01 3.37517411e-01
-7.28972077e-01 -6.29418314e-01 -3.54352236e-01 4.24265355e-01
5.28945148e-01 -8.24212253e-01 1.07925045e+00 -1.53243029e+00
-1.78083205e+00 3.48866314e-01 1.03155665e-01 -7.54762232e-01
4.78432387e-01 1.04393266e-01 -2.58641869e-01 -1.71706364e-01
-6.29433870e-01 8.79775524e-01 9.95626569e-01 -1.09278512e+00
-6.46421552e-01 -1.63321570e-01 -9.54140276e-02 3.30183245e-02
-8.38280499e-01 -2.00154349e-01 -5.73181629e-01 -6.95826709e-01
4.29358073e-02 -1.07174551e+00 -4.82473373e-01 -1.91877589e-01
-3.20838749e-01 4.50066663e-02 4.53209460e-01 -1.14757814e-01
1.47388387e+00 -1.90192521e+00 3.58082056e-01 4.46214646e-01
5.38296290e-02 6.50280654e-01 -1.97914004e-01 5.92068918e-02
-1.48159191e-01 3.02342117e-01 -1.73185274e-01 -1.87040105e-01
-3.51576507e-02 1.49545133e-01 1.60924867e-02 6.04102090e-02
3.41044426e-01 9.52954829e-01 -4.27673578e-01 -4.54419971e-01
1.12881169e-01 5.87772667e-01 -7.42185831e-01 -2.09931266e-02
-3.01492512e-01 2.32407138e-01 -4.94076759e-01 5.60467303e-01
3.69957507e-01 -3.63209873e-01 4.35314715e-01 -3.13493013e-01
-2.20506877e-01 1.15757855e-02 -1.27202475e+00 1.48099148e+00
-4.51961458e-01 4.25030440e-01 -4.93758321e-02 -1.18860424e+00
1.12631500e+00 3.05523008e-01 3.73045892e-01 -5.23039401e-01
5.65968096e-01 3.56761813e-01 3.81503105e-01 -1.85763434e-01
4.16965753e-01 9.98338014e-02 2.12130621e-01 2.95360088e-01
1.81440666e-01 8.13732743e-02 2.13610515e-01 -5.22439241e-01
9.01830196e-01 1.03512341e-02 3.86738539e-01 -3.88598561e-01
5.74234009e-01 1.19263589e-01 6.34400904e-01 7.43153632e-01
-4.03812714e-03 5.56004941e-01 4.77567583e-01 -7.18919814e-01
-1.09249723e+00 -6.92811072e-01 -7.88060427e-02 9.55473840e-01
-3.04948092e-01 -1.89177886e-01 -8.44896257e-01 -4.57065284e-01
-4.91479635e-01 6.69684947e-01 -5.58489561e-01 -1.69345707e-01
-7.19638586e-01 -9.48899746e-01 6.03044748e-01 5.04942954e-01
3.77454042e-01 -9.78745043e-01 -1.24906003e+00 2.60292083e-01
2.11800113e-01 -7.36620665e-01 -2.33746264e-02 3.97284538e-01
-1.14886630e+00 -8.34664524e-01 -8.36293757e-01 -5.64947188e-01
7.75305033e-01 -1.65351838e-01 1.12428367e+00 4.78089482e-01
-5.78822196e-01 -7.43728057e-02 -2.34198213e-01 -5.48565090e-01
-3.23417544e-01 7.26665497e-01 -8.66231546e-02 7.18067065e-02
1.85412273e-01 -5.96001565e-01 -5.95633030e-01 3.10100138e-01
-9.55280006e-01 1.53534591e-01 1.02076840e+00 1.01855338e+00
5.65957248e-01 4.05610114e-01 6.78929150e-01 -6.53955042e-01
4.95014995e-01 -3.39221925e-01 -1.05389345e+00 4.22489613e-01
-1.01284587e+00 3.64182860e-01 5.42889893e-01 -5.06837964e-01
-9.49754536e-01 3.77679974e-01 -1.97592065e-01 -6.77997530e-01
-2.11057782e-01 4.77227539e-01 -2.24642605e-02 -1.59416050e-01
8.23462248e-01 1.50757402e-01 8.72592479e-02 -4.06704992e-01
6.15812652e-02 3.73918951e-01 2.45644838e-01 -5.91792405e-01
5.76927245e-01 9.04402062e-02 3.26493159e-02 -7.93518603e-01
-3.12982470e-01 6.88059554e-02 -6.21541977e-01 -1.97533056e-01
7.53416836e-01 -5.70597708e-01 -6.37201190e-01 3.01041424e-01
-9.92246330e-01 -1.33757904e-01 -1.60172880e-01 4.20989633e-01
-2.84714133e-01 -7.94984214e-03 -1.10281482e-01 -7.99265683e-01
-5.33446014e-01 -1.58118165e+00 4.48516250e-01 3.94952625e-01
-3.81896615e-01 -8.07278752e-01 -2.40288645e-01 1.11492477e-01
7.58457780e-01 7.05284998e-02 1.28615999e+00 -6.12608194e-01
-6.49382770e-01 -2.68378872e-02 -3.28861224e-03 3.97895753e-01
-1.12907283e-01 1.35143518e-01 -8.33284676e-01 -2.40327299e-01
-2.33314991e-01 -1.67256057e-01 6.84585512e-01 5.10631323e-01
1.37896562e+00 -3.11197490e-01 -3.21334720e-01 5.93768477e-01
1.57671845e+00 6.17060661e-01 6.97125852e-01 4.70858157e-01
4.25562024e-01 6.91848636e-01 1.97495073e-01 4.83459324e-01
-8.28703865e-02 7.61507273e-01 6.63100302e-01 5.84109537e-02
3.57454158e-02 2.76059330e-01 -1.42506287e-01 4.99836117e-01
-2.18664467e-01 -3.32003713e-01 -1.21426260e+00 5.79876184e-01
-1.77421045e+00 -6.45673215e-01 2.19524950e-01 2.03525519e+00
5.63461006e-01 3.02694440e-01 3.98592725e-02 2.66650975e-01
5.84095001e-01 -1.52867466e-01 -6.43817961e-01 -3.35071176e-01
7.34733418e-02 4.26623553e-01 4.65755194e-01 1.77653477e-01
-8.59275579e-01 6.18297338e-01 5.92248487e+00 7.77170777e-01
-1.32715666e+00 -8.66496712e-02 8.92796397e-01 -4.72929329e-01
-1.58555016e-01 -1.33334935e-01 -8.14176023e-01 2.84947693e-01
9.38156366e-01 2.74233613e-02 4.21855330e-01 9.63684499e-01
-1.10546798e-01 2.37657622e-01 -9.31917191e-01 1.00042820e+00
-7.59313256e-02 -1.77054048e+00 -2.00570910e-03 1.15457021e-01
7.23305643e-01 1.52820364e-01 2.20066383e-01 4.33146134e-02
5.32214567e-02 -1.24566424e+00 1.05452681e+00 4.42367107e-01
4.62205708e-01 -9.02109146e-01 7.49116600e-01 2.03265086e-01
-9.40996706e-01 -5.39532125e-01 -1.63187355e-01 1.96119532e-01
-1.71736136e-01 3.99561882e-01 -8.63024414e-01 4.62517351e-01
8.78947198e-01 3.27538736e-02 -6.30981326e-01 1.18680739e+00
-6.79835007e-02 4.75491703e-01 -2.26807281e-01 -5.01821756e-01
4.66502130e-01 7.86525980e-02 4.41551626e-01 9.49238062e-01
5.46317875e-01 -1.92293495e-01 -3.84707034e-01 1.06011689e+00
1.01888672e-01 3.00468784e-02 -1.41442716e-01 -5.13801202e-02
6.88222766e-01 1.18587375e+00 -1.08836269e+00 3.85337858e-03
-7.90215135e-02 3.50981534e-01 2.58239776e-01 2.04771772e-01
-8.71132731e-01 -3.90287429e-01 5.81285059e-01 1.76494047e-02
6.63882077e-01 -2.18385324e-01 -6.84671521e-01 -7.15636969e-01
-1.21348575e-01 -9.85460639e-01 2.67584890e-01 -3.77626628e-01
-6.06389225e-01 1.21234345e+00 9.94830057e-02 -1.04387355e+00
-4.06389654e-01 -7.11487889e-01 -3.89564186e-01 7.57111609e-01
-1.47311008e+00 -7.89543569e-01 -2.85436511e-01 2.18197182e-01
6.16992116e-01 -6.33275747e-01 6.80656135e-01 4.88044173e-01
-8.12225044e-01 7.50945270e-01 -9.65813100e-02 -2.83412158e-01
1.44164667e-01 -7.81965852e-01 2.52296120e-01 8.01961780e-01
1.26884595e-01 6.00998700e-01 6.37129664e-01 -2.77634889e-01
-1.17081320e+00 -8.84328663e-01 5.75986803e-01 7.13349506e-02
3.62414330e-01 -2.44525999e-01 -7.78085530e-01 1.17871635e-01
1.54076144e-01 -3.18146110e-01 5.93858540e-01 7.72268176e-02
-1.05832279e-01 -2.05733314e-01 -9.09459591e-01 7.65988946e-01
8.45432520e-01 2.22159922e-01 -1.61968783e-01 -1.00734413e-01
2.49510720e-01 -2.84630209e-01 -5.94661653e-01 5.88806629e-01
6.29554391e-01 -1.01305306e+00 8.95418465e-01 -3.97343874e-01
2.86699921e-01 -3.39194298e-01 -7.71798715e-02 -1.11515546e+00
-2.44726479e-01 -6.17155969e-01 -9.37303677e-02 9.62659001e-01
7.78451920e-01 -6.35982096e-01 9.80331779e-01 4.50849295e-01
-1.70986012e-01 -1.27270687e+00 -1.02033484e+00 -6.09932542e-01
-7.17128441e-02 -4.30600286e-01 9.85269845e-01 5.03857255e-01
-7.13975132e-01 1.97494149e-01 -1.11951962e-01 6.99392036e-02
4.49194700e-01 6.88406751e-02 4.39871520e-01 -1.25784850e+00
-5.92397094e-01 -1.10066342e+00 -3.83817285e-01 -7.03036666e-01
-4.35648374e-02 -6.13972127e-01 1.72129902e-03 -1.04380620e+00
-1.54848814e-01 -8.92059147e-01 -2.74531871e-01 6.40378654e-01
2.95659721e-01 1.18864983e-01 1.59714594e-01 2.52846032e-01
-1.53904155e-01 3.83285552e-01 7.96640813e-01 -1.32797733e-02
-1.98877916e-01 1.03867225e-01 -8.67653489e-01 8.27488422e-01
9.81736958e-01 -4.80690986e-01 -7.09921598e-01 -7.75313616e-01
4.83045638e-01 -2.83958465e-01 2.89537460e-01 -1.38706255e+00
3.60120118e-01 -6.00707196e-02 4.18988049e-01 -4.77477521e-01
2.30788022e-01 -1.02936006e+00 5.93881965e-01 7.93326914e-01
-3.89878035e-01 2.79416323e-01 2.46242389e-01 4.93115246e-01
-1.42380923e-01 -7.44629920e-01 1.06101573e+00 -1.31767914e-01
-7.61709213e-01 1.42901674e-01 -4.17856485e-01 -3.30929130e-01
1.16052902e+00 -4.02896732e-01 -6.56960085e-02 3.54434662e-02
-5.32440782e-01 7.62734339e-02 1.82478920e-01 5.09113729e-01
4.57484156e-01 -1.00542462e+00 -5.07363558e-01 2.87158668e-01
-7.59665594e-02 -4.82530259e-02 2.65765190e-01 6.59476280e-01
-7.55890191e-01 6.13970280e-01 -4.35662895e-01 -6.92923903e-01
-1.26870155e+00 3.91388685e-01 6.10013366e-01 -2.92164356e-01
-2.11920097e-01 9.73415732e-01 -5.15735373e-02 -1.44947931e-01
4.08573419e-01 -2.25603610e-01 -2.61940092e-01 1.12526946e-01
3.29650491e-01 3.01012367e-01 4.94615048e-01 -3.41961443e-01
-3.63130063e-01 4.41007674e-01 -1.39165953e-01 5.86159043e-02
1.50554752e+00 1.09544411e-01 6.62872493e-02 -7.54293203e-02
9.70735252e-01 -6.53465509e-01 -1.14581752e+00 -1.59264550e-01
1.88747019e-01 -1.86717108e-01 3.62533122e-01 -7.45530427e-01
-1.44085956e+00 7.35987246e-01 8.36025059e-01 9.98986959e-02
1.61665881e+00 -2.12692589e-01 4.09431040e-01 5.50823450e-01
3.16505879e-01 -1.06897211e+00 1.99675873e-01 4.27295208e-01
7.12928951e-01 -7.87677407e-01 8.38695914e-02 -8.36875215e-02
-4.93267506e-01 1.24423969e+00 6.69608831e-01 1.56788563e-03
6.79389596e-01 4.55019563e-01 1.61149912e-02 -2.25225389e-01
-9.41608667e-01 1.53150931e-02 3.88880849e-01 2.83851713e-01
3.33218604e-01 -8.39593336e-02 -4.29328948e-01 6.55303359e-01
-2.51181662e-01 -9.99868587e-02 1.70540348e-01 1.08735001e+00
-3.32760334e-01 -1.25931895e+00 -2.59268135e-01 4.12827075e-01
-5.37901461e-01 -7.86929876e-02 -2.16919214e-01 6.11847997e-01
3.11133295e-01 7.06006944e-01 1.25880927e-01 -4.45642918e-01
1.72414199e-01 9.54817459e-02 3.74502242e-01 -3.86881024e-01
-8.72060776e-01 6.60567358e-02 6.31416515e-02 -2.63189256e-01
-3.45236003e-01 -5.11050105e-01 -9.84498382e-01 1.79956798e-02
-5.27384877e-01 1.47060335e-01 1.07140231e+00 8.24343979e-01
6.40810490e-01 1.00623214e+00 3.50802004e-01 -9.19687748e-01
-5.89382708e-01 -5.79629481e-01 -8.31051990e-02 -3.60177517e-01
-1.55139603e-02 -7.82517493e-01 1.96264256e-02 -8.95505250e-02] | [8.360098838806152, 3.233081340789795] |
f3a24285-979f-4b8a-bf8e-cbaff06ba023 | development-and-evaluation-of-automated | 2211.02760 | null | https://arxiv.org/abs/2211.02760v1 | https://arxiv.org/pdf/2211.02760v1.pdf | Development and evaluation of automated localization and reconstruction of all fruits on tomato plants in a greenhouse based on multi-view perception and 3D multi-object tracking | Accurate representation and localization of relevant objects is important for robots to perform tasks. Building a generic representation that can be used across different environments and tasks is not easy, as the relevant objects vary depending on the environment and the task. Furthermore, another challenge arises in agro-food environments due to their complexity, and high levels of clutter and occlusions. In this paper, we present a method to build generic representations in highly occluded agro-food environments using multi-view perception and 3D multi-object tracking. Our representation is built upon a detection algorithm that generates a partial point cloud for each detected object. The detected objects are then passed to a 3D multi-object tracking algorithm that creates and updates the representation over time. The whole process is performed at a rate of 10 Hz. We evaluated the accuracy of the representation on a real-world agro-food environment, where it was able to successfully represent and locate tomatoes in tomato plants despite a high level of occlusion. We were able to estimate the total count of tomatoes with a maximum error of 5.08% and to track tomatoes with a tracking accuracy up to 71.47%. Additionally, we showed that an evaluation using tracking metrics gives more insight in the errors in localizing and representing the fruits. | ['Gert Kootstra', 'Eldert J. van Henten', 'David Rapado Rincon'] | 2022-11-04 | null | null | null | null | ['3d-multi-object-tracking'] | ['computer-vision'] | [ 7.69152492e-02 -2.70855308e-01 3.93113077e-01 5.89911751e-02
-3.21570158e-01 -9.87791777e-01 2.12959453e-01 1.04405630e+00
-7.74317458e-02 1.63467363e-01 -6.33858204e-01 1.08589754e-01
-3.32422405e-02 -9.15944815e-01 -8.84493113e-01 -5.05126834e-01
-2.95820415e-01 8.60109210e-01 8.99725497e-01 -2.83650011e-01
-2.06578504e-02 8.62192631e-01 -1.93864310e+00 2.43718415e-01
5.27322471e-01 8.23596597e-01 1.20192432e+00 1.03167439e+00
1.18149444e-03 9.24053192e-02 -7.84996986e-01 1.44324183e-01
2.26205960e-01 1.94427639e-01 -1.81954622e-01 2.83016115e-01
5.27080297e-01 -2.63680279e-01 5.23558021e-01 9.32872653e-01
3.03927630e-01 -2.40662411e-01 6.08307958e-01 -1.31220698e+00
-3.22336614e-01 2.52987742e-01 -5.38191438e-01 -2.36768335e-01
7.40939021e-01 -1.35498911e-01 2.68274486e-01 -7.78437734e-01
7.66321123e-01 1.37697005e+00 9.29294586e-01 7.82712847e-02
-1.37958217e+00 -3.51595998e-01 1.72460675e-01 -4.46552709e-02
-1.45050859e+00 -1.14189774e-01 1.89706206e-01 -5.66481173e-01
6.59968853e-01 2.77917922e-01 6.92937851e-01 6.65621102e-01
3.00524890e-01 3.70678842e-01 8.19298029e-01 -6.06397271e-01
3.26140910e-01 2.60667980e-01 -7.81344175e-02 5.06984055e-01
8.78352165e-01 -3.73283215e-02 -1.33045569e-01 -1.38295352e-01
9.37885344e-01 4.09958884e-02 1.07955560e-01 -1.10281444e+00
-1.34513879e+00 6.48293972e-01 5.76043844e-01 5.28292418e-01
-8.39711368e-01 2.77233481e-01 3.74414444e-01 -1.33881390e-01
1.82779029e-01 1.41391844e-01 -5.57393551e-01 2.96168447e-01
-5.25638342e-01 2.56879807e-01 9.63312924e-01 1.18734455e+00
5.89030802e-01 -3.23267847e-01 1.33234620e-01 3.87187153e-01
4.06041682e-01 1.03953218e+00 -1.10171720e-01 -8.47112238e-01
7.46285543e-02 6.59421146e-01 5.65713882e-01 -1.46976686e+00
-5.79646349e-01 -1.36966407e-01 -5.08655071e-01 7.64834642e-01
5.78473806e-01 8.11285973e-02 -8.22830617e-01 1.31456017e+00
7.96897054e-01 -4.06241536e-01 -3.37483734e-02 6.78801358e-01
5.98402679e-01 5.26066601e-01 2.42137626e-01 -1.13939550e-02
1.75616920e+00 -4.82717186e-01 -8.01086009e-01 -3.99081409e-01
4.90225613e-01 -1.13213754e+00 5.17675281e-01 4.24740791e-01
-9.67219591e-01 -8.77531409e-01 -1.05186033e+00 3.25343609e-01
-7.45893717e-01 6.19671166e-01 5.43921113e-01 5.92929304e-01
-9.64225292e-01 5.82928777e-01 -9.45707202e-01 -1.05364251e+00
8.07067677e-02 3.38863820e-01 -5.20096481e-01 -9.51866508e-02
-4.83817995e-01 1.35791397e+00 8.70672345e-01 1.58953667e-01
-9.51089501e-01 -3.79239321e-01 -7.58666873e-01 1.02884129e-01
3.95902753e-01 -5.78976989e-01 9.44000304e-01 -5.45386672e-01
-1.19132364e+00 7.16693163e-01 4.61874902e-03 -1.48020178e-01
4.15859997e-01 -4.21682835e-01 -1.29584223e-01 1.87542811e-01
3.94910365e-01 6.10049903e-01 4.84562248e-01 -1.72828436e+00
-9.38685477e-01 -7.34260082e-01 1.15590833e-01 1.85878500e-01
9.61721838e-02 -9.19631310e-03 -3.26340616e-01 -2.55106568e-01
5.37704706e-01 -1.08394980e+00 -2.82956660e-01 6.21904492e-01
1.78700775e-01 1.91319555e-01 1.09653938e+00 -4.81136024e-01
5.84456801e-01 -1.91185689e+00 1.01067424e-01 -2.28559285e-01
-2.57623047e-01 3.34838331e-01 -1.14497811e-01 5.98387659e-01
2.08920076e-01 -3.62568319e-01 1.85305029e-01 -1.82107866e-01
-1.41870454e-01 1.41919419e-01 8.06690231e-02 5.05020261e-01
1.64896026e-02 2.92169631e-01 -1.05095637e+00 -3.95606846e-01
7.52596617e-01 8.20631564e-01 -6.24654219e-02 -3.10738795e-02
-4.71463948e-01 8.90257359e-02 -5.09895444e-01 8.32915664e-01
1.12272978e+00 5.10759503e-02 3.46972913e-01 -4.36139345e-01
-5.10034680e-01 -4.75399077e-01 -1.79034567e+00 1.80368447e+00
-5.09399652e-01 2.94568509e-01 5.12186766e-01 -5.89241743e-01
1.16383171e+00 3.53480101e-01 5.54537773e-01 -1.81238145e-01
6.79116398e-02 3.39700222e-01 -4.27009702e-01 -3.77596557e-01
7.32561827e-01 5.16776323e-01 -1.01593751e-02 4.50196154e-02
-9.64236557e-02 -4.59934711e-01 5.04814327e-01 -8.06016028e-02
8.66565108e-01 7.38106012e-01 6.23535931e-01 -3.66744995e-01
3.56245607e-01 5.98218858e-01 1.48819521e-01 6.77724898e-01
-1.41506463e-01 4.44172323e-01 -4.19395007e-02 -6.14308596e-01
-9.52257037e-01 -1.10011435e+00 -1.44183576e-01 1.09505796e+00
5.32035768e-01 -3.24013114e-01 -3.55678022e-01 -3.27293783e-01
1.69690296e-01 5.87652445e-01 -5.09519279e-01 6.82684034e-02
-5.64860582e-01 -6.11014068e-01 3.51409465e-02 4.30650830e-01
3.03159505e-01 -1.04404759e+00 -1.68887520e+00 6.37678027e-01
-1.17609277e-01 -1.29663157e+00 2.10431397e-01 4.77285743e-01
-9.57911789e-01 -1.14819717e+00 -7.02041149e-01 -8.58019948e-01
6.77798629e-01 7.97313690e-01 1.16371715e+00 -1.42869562e-01
-6.18164420e-01 4.91312444e-01 -6.53347313e-01 -8.96463335e-01
-7.69498408e-01 -2.08350062e-01 -8.71864185e-02 -7.20910311e-01
3.00952077e-01 -1.53986230e-01 -3.32173645e-01 4.79462534e-01
-7.47067690e-01 -1.19320512e-01 4.91690159e-01 2.93765754e-01
7.73250759e-01 -4.93943542e-02 1.46620244e-01 -3.32667470e-01
5.33683486e-02 -2.94332385e-01 -1.15704203e+00 5.02034366e-01
1.27274796e-01 -1.94316998e-01 2.06012636e-01 -7.55017400e-01
-5.63423634e-01 1.03722489e+00 2.65041739e-01 -2.79963791e-01
-5.72934866e-01 4.12077047e-02 -1.92706287e-02 -1.97947145e-01
7.76609302e-01 -2.03816056e-01 -9.67268646e-02 -3.98712218e-01
5.36517620e-01 4.40800786e-01 3.81653011e-01 -2.30096549e-01
5.81587195e-01 5.30021071e-01 2.60256886e-01 -1.08060956e+00
-4.98675525e-01 -6.30420029e-01 -1.07577753e+00 -4.03864771e-01
8.58243108e-01 -9.59063888e-01 -1.00382233e+00 3.77987564e-01
-1.65724671e+00 -1.35197923e-01 -2.95397967e-01 6.38621628e-01
-5.55043221e-01 2.60912925e-01 -7.60288686e-02 -1.00688386e+00
-2.06220418e-01 -1.09959459e+00 1.52620101e+00 1.35331511e-01
2.73488034e-02 -6.79501593e-01 -5.44341747e-03 -2.73289770e-01
4.07467037e-01 7.18097806e-01 3.40097606e-01 -9.17514935e-02
-5.78275323e-01 -3.41413587e-01 -1.97283864e-01 -3.96863341e-01
3.51119161e-01 2.33495578e-01 -9.37741876e-01 -4.62548196e-01
-8.55463073e-02 1.47334523e-02 3.76957804e-01 7.12271810e-01
5.72487712e-01 5.00333607e-01 -9.38055992e-01 4.98182029e-02
1.76152313e+00 3.49229187e-01 4.91287634e-02 3.43925834e-01
3.77764821e-01 7.17630088e-01 1.19234025e+00 4.99099851e-01
2.94736266e-01 1.19109917e+00 1.09631705e+00 1.24305487e-04
-1.89154297e-01 1.20727725e-01 1.88791096e-01 1.96951598e-01
-1.03401653e-02 -3.65255952e-01 -7.47261405e-01 5.76543808e-01
-1.90791798e+00 -8.33075881e-01 -6.22096837e-01 2.30114460e+00
8.18065405e-02 -1.80469260e-01 1.78968862e-01 1.94590032e-01
1.01696968e+00 -3.98171782e-01 -3.08825612e-01 -1.91373587e-01
1.37100011e-01 -9.36469063e-02 8.26841712e-01 3.25432062e-01
-1.36401820e+00 9.01311219e-01 6.42618561e+00 2.33758703e-01
-8.50607514e-01 -5.47659621e-02 -3.40574443e-01 4.68625069e-01
4.90727156e-01 -1.26846313e-01 -1.01245987e+00 6.18128404e-02
5.94054818e-01 9.89569053e-02 7.13218898e-02 1.00564623e+00
2.23322045e-02 -6.47131085e-01 -8.82342815e-01 7.66607285e-01
1.80263832e-01 -7.11590350e-01 -3.48639548e-01 -1.17395967e-01
2.74642527e-01 -7.70633221e-02 -4.48038608e-01 -6.97256476e-02
4.86353874e-01 -5.56246281e-01 1.06437123e+00 2.62351573e-01
3.96224022e-01 -3.17507744e-01 5.45071542e-01 6.37123942e-01
-1.83880520e+00 -5.37063628e-02 -7.86773682e-01 2.91766822e-02
2.77434707e-01 5.27224302e-01 -1.13360834e+00 6.76229596e-01
8.56660545e-01 3.12093258e-01 -6.38202667e-01 1.38688254e+00
5.71906641e-02 -4.43425119e-01 -9.06504393e-01 -3.13912690e-01
-5.31160235e-02 -5.12649342e-02 5.64777434e-01 1.16778421e+00
8.92184615e-01 -2.56355524e-01 4.48602706e-01 7.91800618e-01
7.05492198e-01 5.86923510e-02 -8.39047313e-01 3.07146907e-01
6.12909198e-01 1.50682592e+00 -1.22997689e+00 -1.67511865e-01
4.38020900e-02 1.07837081e+00 3.20576737e-03 -1.54142842e-01
-8.37715149e-01 -2.73514211e-01 1.52597785e-01 -1.03434529e-02
7.71529913e-01 -5.67769110e-01 2.00340778e-01 -6.81291461e-01
5.09774163e-02 -4.33332086e-01 1.48994267e-01 -1.07353401e+00
-7.34616160e-01 6.47937298e-01 3.33061039e-01 -1.27553701e+00
-1.86703175e-01 -8.30795586e-01 -5.24846241e-02 8.84364247e-01
-9.84647036e-01 -1.55836415e+00 -9.17642891e-01 5.94016686e-02
5.54366469e-01 3.42811882e-01 1.43498015e+00 3.11710566e-01
-5.34582622e-02 -2.67169058e-01 1.75486207e-01 -6.15447164e-01
3.89225125e-01 -1.24126101e+00 2.78261662e-01 6.68390572e-01
2.15896577e-01 1.69200629e-01 1.06502271e+00 -7.89059758e-01
-1.51277685e+00 -1.02458918e+00 6.93802476e-01 -6.55732989e-01
2.21621603e-01 -5.03998458e-01 -7.35007286e-01 6.57929301e-01
-5.16575389e-02 2.52594143e-01 1.22193389e-01 -2.27829084e-01
8.00113752e-02 1.46855004e-02 -1.47051966e+00 6.47605285e-02
8.74251127e-01 8.35706741e-02 -3.48082095e-01 5.64911604e-01
5.51469326e-01 -7.75581837e-01 -8.66697311e-01 3.75784874e-01
8.07455659e-01 -9.35382307e-01 1.41031814e+00 2.00675517e-01
-3.34466159e-01 -7.14095831e-01 -4.24228549e-01 -1.32443678e+00
-3.52618963e-01 4.20792438e-02 2.41886191e-02 1.04394066e+00
-1.04697689e-01 -2.49593928e-01 5.96297443e-01 -7.78016299e-02
9.94059816e-02 5.69016933e-02 -5.10664403e-01 -8.58997464e-01
-4.54029322e-01 -8.10031444e-02 4.74716306e-01 4.48708683e-01
-4.86752659e-01 -2.98363622e-02 8.79765898e-02 8.90854597e-01
7.68964469e-01 4.59854782e-01 9.01670218e-01 -1.63064158e+00
-9.50184278e-03 8.15700963e-02 -5.91412306e-01 -6.63805425e-01
-3.78770530e-01 -5.97945631e-01 2.79201508e-01 -1.91961658e+00
-1.26302451e-01 -6.36989236e-01 3.54142755e-01 2.38954917e-01
1.30688936e-01 1.36771023e-01 6.76891208e-01 -6.01920113e-02
-5.12879431e-01 -2.95418315e-02 9.26897049e-01 2.06536874e-02
-3.39231193e-01 1.68630585e-01 -6.45083040e-02 8.60995770e-01
7.45146871e-01 -6.90414906e-01 -9.00267884e-02 -7.27534473e-01
-2.21035331e-02 -2.25239620e-02 5.98771393e-01 -1.53513050e+00
1.61230758e-01 7.86613002e-02 6.43298388e-01 -1.31160927e+00
8.39731932e-01 -1.47349215e+00 7.23705769e-01 1.10287035e+00
2.51380384e-01 2.48824731e-01 6.65363193e-01 5.17339706e-01
3.23622525e-01 -6.30990028e-01 6.33447826e-01 -5.25641799e-01
-7.93252945e-01 -2.23422676e-01 -2.41181105e-01 -8.04247975e-01
1.58814764e+00 -3.77268672e-01 -1.45072505e-01 1.65585905e-01
-1.04212165e+00 1.34107977e-01 7.07499266e-01 4.45883840e-01
3.31383228e-01 -1.17780709e+00 -6.50196552e-01 1.94751859e-01
2.09920272e-01 6.68491721e-02 4.51789312e-02 3.15151513e-01
-1.06957626e+00 4.59539980e-01 -6.15369916e-01 -1.18865573e+00
-1.63057983e+00 7.59566009e-01 1.24237359e-01 -2.83038188e-02
-4.29784238e-01 2.84881890e-01 1.31204762e-02 -3.73694003e-01
2.43240744e-01 -7.51790881e-01 -4.68395084e-01 1.77038148e-01
4.61928070e-01 5.98060131e-01 1.06284961e-01 -7.73404956e-01
-5.93804598e-01 1.17360413e+00 4.60005403e-01 1.58420354e-01
1.32779872e+00 -1.14328176e-01 2.12069124e-01 4.48965758e-01
4.08088863e-01 -1.53475106e-01 -1.17180133e+00 2.68560171e-01
2.57221032e-02 -6.06919825e-01 -1.68696120e-01 -6.87548995e-01
-5.97834349e-01 7.27249861e-01 1.20083308e+00 6.34243488e-01
8.42829525e-01 -5.97239509e-02 1.22249141e-01 4.01234806e-01
1.05861199e+00 -5.88021576e-01 -1.04107291e-01 3.20207775e-01
1.04876614e+00 -1.17089689e+00 2.51918226e-01 -8.33311260e-01
-3.21666658e-01 1.36041558e+00 5.28180599e-01 -8.95406008e-02
3.86596620e-01 6.90339923e-01 8.09327215e-02 -2.60233551e-01
-2.55940557e-01 -3.39320332e-01 -2.85221040e-01 1.25406766e+00
2.42359489e-01 1.67150378e-01 -1.00715160e-01 -1.78874850e-01
2.60140479e-01 1.19437717e-01 5.24800539e-01 1.20847046e+00
-7.83860505e-01 -9.60769534e-01 -9.75615978e-01 -9.68252644e-02
-9.05852467e-02 6.02207601e-01 -1.26391515e-01 9.79596794e-01
3.73443931e-01 1.03035831e+00 7.75190145e-02 1.91230208e-01
8.49530458e-01 -2.97824025e-01 7.70696521e-01 -7.27375150e-01
-6.98665261e-01 1.72378063e-01 -1.87885135e-01 -5.77401042e-01
-6.69140875e-01 -7.69485116e-01 -1.16774416e+00 8.46660435e-02
-8.16545725e-01 -1.83788374e-01 1.35767388e+00 2.43672639e-01
3.35798770e-01 6.65810466e-01 2.22893208e-01 -1.40629554e+00
-4.63778734e-01 -8.33914161e-01 -5.42915165e-01 3.30371410e-01
7.32620060e-02 -8.87833178e-01 -1.03798687e-01 2.11097851e-01] | [7.5482659339904785, -2.0398292541503906] |
41d9f5a2-d045-4db6-8f79-c2782f3a2031 | are-deep-neural-networks-adequate-behavioural | 2305.17023 | null | https://arxiv.org/abs/2305.17023v1 | https://arxiv.org/pdf/2305.17023v1.pdf | Are Deep Neural Networks Adequate Behavioural Models of Human Visual Perception? | Deep neural networks (DNNs) are machine learning algorithms that have revolutionised computer vision due to their remarkable successes in tasks like object classification and segmentation. The success of DNNs as computer vision algorithms has led to the suggestion that DNNs may also be good models of human visual perception. We here review evidence regarding current DNNs as adequate behavioural models of human core object recognition. To this end, we argue that it is important to distinguish between statistical tools and computational models, and to understand model quality as a multidimensional concept where clarity about modelling goals is key. Reviewing a large number of psychophysical and computational explorations of core object recognition performance in humans and DNNs, we argue that DNNs are highly valuable scientific tools but that as of today DNNs should only be regarded as promising -- but not yet adequate -- computational models of human core object recognition behaviour. On the way we dispel a number of myths surrounding DNNs in vision science. | ['Robert Geirhos', 'Felix A. Wichmann'] | 2023-05-26 | null | null | null | null | ['object-recognition'] | ['computer-vision'] | [ 3.72889370e-01 -1.85316727e-01 -3.68537568e-02 -3.71431082e-01
1.12103738e-01 -4.03549969e-01 7.02175200e-01 -6.89687505e-02
-1.01259363e+00 7.59278387e-02 1.15163699e-01 -6.38352156e-01
-3.61596912e-01 -3.43677670e-01 -3.72405350e-01 -7.46913910e-01
2.28931785e-01 2.98202246e-01 2.77630687e-01 -8.90783295e-02
4.47144270e-01 8.50029528e-01 -1.87203431e+00 2.38977253e-01
6.27960861e-01 1.03093994e+00 3.88860226e-01 1.05192506e+00
1.33607350e-02 6.57753646e-01 -3.64520967e-01 -6.41531825e-01
1.24084190e-01 -3.47661376e-01 -6.47382557e-01 1.38455659e-01
4.91976291e-01 -3.69755208e-01 -3.53617907e-01 1.11493266e+00
4.88762379e-01 -1.52487546e-01 5.35677612e-01 -1.07411647e+00
-1.01893282e+00 -3.55388857e-02 -1.69609651e-01 5.96575022e-01
-4.27885950e-01 6.38531744e-01 8.73625934e-01 -5.01471102e-01
2.95731932e-01 1.44591379e+00 7.71630287e-01 7.45759308e-01
-1.43643856e+00 -1.94869190e-01 -1.34071680e-02 4.59391028e-01
-8.41689050e-01 -5.00483692e-01 2.67654151e-01 -7.22629189e-01
1.20337176e+00 4.00932461e-01 1.22956729e+00 9.85710204e-01
3.94473583e-01 6.81507289e-01 1.31979978e+00 -5.45511842e-01
3.34732920e-01 8.67481902e-02 2.77934760e-01 1.87815726e-01
6.08423829e-01 5.92321575e-01 -1.58520445e-01 -3.25242542e-02
1.21293437e+00 -1.75155252e-01 5.31566702e-02 -2.43797228e-01
-1.00876141e+00 8.85260820e-01 5.45090556e-01 5.05226672e-01
-4.99471813e-01 2.88558483e-01 5.83829582e-01 5.67738637e-02
-2.99283415e-02 4.26301211e-01 -3.84621501e-01 -2.33141035e-01
-6.73987508e-01 4.86166239e-01 3.52316350e-01 2.99649388e-01
3.08781117e-01 3.03452492e-01 3.18927228e-01 1.10010648e+00
4.82230306e-01 4.34195966e-01 4.43481207e-01 -1.15213120e+00
-4.99334097e-01 6.54170573e-01 1.02288142e-01 -9.87595081e-01
-5.77013314e-01 -5.60531199e-01 -8.16925883e-01 7.56772339e-01
5.40114641e-01 2.67358631e-01 -1.15827334e+00 1.52878046e+00
-1.70738801e-01 -2.91094124e-01 -2.84693956e-01 1.04387891e+00
6.74394786e-01 1.38907671e-01 6.29652977e-01 -9.47105978e-03
1.33810115e+00 -4.42453116e-01 -2.70172089e-01 -6.07837498e-01
5.26016951e-01 -7.40570009e-01 7.20922828e-01 5.96581936e-01
-1.12143028e+00 -8.72953892e-01 -7.76867568e-01 -1.78850204e-01
-3.92514974e-01 -1.30012408e-01 1.10278153e+00 9.87333536e-01
-1.23811066e+00 5.31943619e-01 -9.36792910e-01 -7.45614886e-01
7.99864709e-01 3.84321153e-01 -5.80905080e-02 2.66744103e-02
-7.11715281e-01 1.40871143e+00 6.73110545e-01 4.71120745e-01
-8.82824838e-01 -3.15169156e-01 -5.80879092e-01 -2.69192725e-01
5.04669882e-02 -8.12019050e-01 1.37945104e+00 -1.38928688e+00
-7.83425212e-01 1.39218497e+00 -2.43097663e-01 -7.50332236e-01
1.54709026e-01 4.58681360e-02 -5.00626147e-01 2.12903112e-01
-2.61534125e-01 1.17367017e+00 5.83824217e-01 -1.50114977e+00
-7.44359553e-01 -5.29728174e-01 1.21403374e-02 7.85287619e-02
1.16892457e-01 3.65566075e-01 -1.61148384e-02 -3.97040099e-01
3.09056908e-01 -8.26856732e-01 -4.37361658e-01 5.23106754e-01
2.22362965e-01 -5.28946638e-01 5.53169966e-01 -5.23882151e-01
8.88316154e-01 -2.00777364e+00 -2.58262992e-01 -2.87688132e-02
7.29594946e-01 8.77058864e-01 -2.40638778e-01 -2.56332476e-02
-4.31250073e-02 1.02458693e-01 -9.61667597e-02 1.42075583e-01
2.05844507e-01 6.35770679e-01 -3.45638469e-02 5.59224367e-01
2.70327419e-01 1.48023653e+00 -6.91652298e-01 -1.84081048e-01
6.50839806e-01 6.04184687e-01 -2.36523092e-01 -3.00010949e-01
-2.71204352e-01 5.50030433e-02 -1.69422150e-01 5.77287078e-01
8.24806392e-01 -2.11090907e-01 1.81397554e-02 -4.21885960e-02
-2.72264272e-01 1.07358776e-01 -8.49336147e-01 1.16100883e+00
3.12850684e-01 1.15206325e+00 7.09314123e-02 -1.23053920e+00
7.45661318e-01 1.41844237e-02 4.96913604e-02 -1.21796882e+00
3.85168850e-01 3.69461060e-01 1.06812882e+00 -5.95636070e-01
1.86736926e-01 -5.30026495e-01 6.55107021e-01 1.41383320e-01
4.61305901e-02 -1.55169189e-01 -3.58949509e-03 -3.69301826e-01
5.74566901e-01 1.01796627e-01 2.62343585e-01 -3.71940523e-01
2.32809290e-01 2.09347412e-01 4.94523078e-01 9.95941579e-01
-5.96187353e-01 3.82764310e-01 4.47088480e-01 -8.55573893e-01
-1.52053082e+00 -1.07791936e+00 -5.31637907e-01 9.46752608e-01
-1.74534783e-01 1.08888850e-01 -7.89196551e-01 1.66324452e-01
-1.15536928e-01 6.29626453e-01 -8.48177016e-01 -2.23413765e-01
-2.45295107e-01 -9.85988379e-01 6.89636230e-01 7.83208966e-01
3.19796354e-01 -1.37573183e+00 -1.22142100e+00 1.47337571e-01
5.22219598e-01 -9.88081694e-01 6.22780144e-01 4.14099604e-01
-9.63042498e-01 -1.09650004e+00 -8.40482712e-01 -9.41292524e-01
4.48656827e-01 5.67630768e-01 1.09246933e+00 4.84292209e-01
-7.31099129e-01 4.67635095e-01 -1.67569950e-01 -8.61640394e-01
-6.00520968e-01 -4.90336418e-01 1.74067304e-01 -4.97757673e-01
1.10137773e+00 -4.54721540e-01 -7.43288040e-01 2.63700992e-01
-1.11831415e+00 1.31109640e-01 7.11177111e-01 6.42278254e-01
1.89751551e-01 -1.63183510e-01 5.19009471e-01 -3.70523185e-01
6.52830005e-01 -9.33076292e-02 -5.36824942e-01 1.26880437e-01
-5.61428428e-01 -4.94510353e-01 5.18039204e-02 -3.72754842e-01
-9.83853400e-01 -5.35877347e-01 -3.67631108e-01 -1.37414113e-01
-8.41047287e-01 3.14000279e-01 2.94847608e-01 -4.34841067e-01
8.77510726e-01 4.45608526e-01 2.08820760e-01 -4.56447959e-01
2.51296222e-01 5.81175387e-01 3.93763512e-01 -4.06861514e-01
3.43256861e-01 6.28516793e-01 2.38757730e-02 -1.30570364e+00
-5.76056659e-01 -3.91844541e-01 -7.05885351e-01 -1.65521294e-01
1.05405474e+00 -5.88417470e-01 -1.13278306e+00 8.64195049e-01
-1.19389153e+00 -4.39127445e-01 -1.47869855e-01 6.09526157e-01
-7.82541692e-01 2.54841924e-01 -4.90337193e-01 -1.04988825e+00
1.92987725e-01 -1.14765787e+00 5.88814795e-01 5.24042189e-01
-4.00810033e-01 -1.24631977e+00 -2.38758177e-02 4.49689269e-01
3.84545416e-01 2.97318436e-02 1.19350564e+00 -6.17748022e-01
-2.74305761e-01 -5.95763884e-02 -7.67774999e-01 6.51407003e-01
-5.84120937e-02 8.32898095e-02 -1.41038275e+00 -2.96474155e-02
3.52087200e-01 -6.08015418e-01 1.01572704e+00 8.76297235e-01
9.09417629e-01 -3.53380106e-03 -2.09160239e-01 5.67970455e-01
1.59148586e+00 7.63158619e-01 9.33588266e-01 4.98042375e-01
3.63174111e-01 8.42328489e-01 -7.31159300e-02 3.63333523e-02
1.43184677e-01 2.31425673e-01 3.00609529e-01 -2.97050416e-01
-3.64720784e-02 2.17840150e-01 -1.25551164e-01 4.41020310e-01
-2.79443562e-01 1.08273765e-02 -1.13800633e+00 6.94908261e-01
-1.62852418e+00 -1.19476569e+00 -3.98974925e-01 1.92347980e+00
3.86980563e-01 5.62286615e-01 4.68087465e-01 5.61263897e-02
5.25431931e-01 -7.84765854e-02 -7.98783541e-01 -6.91962123e-01
-4.95061308e-01 1.73538402e-01 2.08821446e-01 2.78005600e-01
-8.99855554e-01 7.30410695e-01 7.39072180e+00 5.31420469e-01
-9.74148333e-01 -9.07039568e-02 7.72750616e-01 4.85455617e-02
-5.65423220e-02 -8.97197202e-02 -6.66181505e-01 1.87524214e-01
9.71389115e-01 2.07527235e-01 2.13995382e-01 7.52382338e-01
4.10805404e-01 -5.92870235e-01 -1.06983662e+00 1.04235721e+00
-1.71571632e-03 -1.36351907e+00 8.76122713e-02 2.22444832e-01
4.72621202e-01 2.00659275e-01 3.80599260e-01 7.74836680e-03
2.53007203e-01 -1.48424792e+00 5.82389891e-01 3.92420202e-01
2.41174549e-01 -3.92145544e-01 6.39536381e-01 3.78783226e-01
-5.33191323e-01 -5.39065711e-02 -9.57634687e-01 -5.43786764e-01
-2.92325187e-02 5.43419681e-02 -3.92071426e-01 -3.25609535e-01
8.18085432e-01 3.84819567e-01 -8.40956151e-01 1.55564857e+00
2.70791978e-01 5.24113059e-01 -2.45041698e-01 -1.07522741e-01
6.54251516e-01 7.65008107e-02 3.24822217e-01 1.24081981e+00
-2.72445649e-01 1.24464162e-01 -4.12379265e-01 1.15118563e+00
6.77735329e-01 -3.21182281e-01 -5.02990782e-01 -3.48720610e-01
9.35770273e-02 1.15778983e+00 -1.06714511e+00 -9.61325392e-02
-8.41790438e-01 7.29644537e-01 1.24903075e-01 4.28112298e-01
-3.63391280e-01 3.25010307e-02 9.31726992e-01 6.70684427e-02
2.77729452e-01 -3.63181561e-01 -7.72220075e-01 -6.53048873e-01
-2.08416998e-01 -9.96009707e-01 -1.47321731e-01 -1.00155330e+00
-1.25091767e+00 4.38794523e-01 -5.56146912e-02 -7.60867715e-01
3.09788529e-02 -1.43895602e+00 -3.04792434e-01 9.16317821e-01
-1.19673944e+00 -1.22425258e+00 1.42378192e-02 1.20625533e-01
3.80424500e-01 3.42076570e-02 6.82207763e-01 -2.68203199e-01
-1.10413983e-01 9.36019868e-02 3.08370024e-01 2.54089922e-01
3.52578610e-02 -9.76833701e-01 7.76211619e-01 4.73339856e-01
3.10928673e-01 9.17107344e-01 7.25952923e-01 -4.27949995e-01
-8.97266328e-01 -4.52506483e-01 5.21741092e-01 -6.03283048e-01
4.67842698e-01 -2.00352326e-01 -1.06661749e+00 6.17229939e-01
2.29083076e-01 -2.29208261e-01 7.99138129e-01 -1.48469303e-02
1.30243702e-02 2.18954012e-01 -8.97563815e-01 7.86522031e-01
9.31654930e-01 -4.71439213e-01 -9.50928450e-01 6.75386516e-03
1.78179651e-01 7.28435144e-02 -5.83271027e-01 1.86343223e-01
8.87347877e-01 -1.49042940e+00 1.06966376e+00 -9.52974200e-01
1.10003144e-01 -1.13507904e-01 -1.67101741e-01 -9.55519795e-01
-6.69083953e-01 -2.07396597e-01 4.09167409e-01 6.54165745e-01
-5.47028370e-02 -7.17070699e-01 9.24087524e-01 1.00861061e+00
-4.48712185e-02 -7.15012133e-01 -1.15022147e+00 -7.54472136e-01
3.79570127e-01 -9.25221920e-01 -1.10652141e-01 4.89180118e-01
-3.04122955e-01 8.24846774e-02 1.19352840e-01 -2.84991562e-01
8.91084969e-01 -4.48522955e-01 3.95385414e-01 -1.79057932e+00
2.93663470e-03 -1.02987790e+00 -1.04311776e+00 -8.26721728e-01
-1.46586001e-01 -4.61252391e-01 -2.33599231e-01 -1.74143064e+00
5.02489805e-01 -9.82463658e-02 -3.63576651e-01 2.00122327e-01
2.78015405e-01 4.21880603e-01 3.00753176e-01 1.84951887e-01
-1.87789470e-01 4.22779098e-03 1.35004711e+00 2.47792557e-01
1.26117766e-01 -1.32062376e-01 -1.06907272e+00 1.17885816e+00
7.49778390e-01 -1.86930180e-01 -3.67125481e-01 -4.85445470e-01
4.55480367e-01 -4.60794628e-01 1.22978580e+00 -1.07732809e+00
3.48562896e-02 -4.02117401e-01 1.01398563e+00 -3.82794082e-01
3.47294122e-01 -8.44860911e-01 -1.01762876e-01 7.33431935e-01
-2.31085435e-01 -9.17857140e-02 6.82265401e-01 2.87805885e-01
1.33839235e-01 -3.58710140e-01 1.22279692e+00 -5.83770692e-01
-1.28482628e+00 -4.10817862e-02 -7.59003222e-01 5.57140214e-04
9.29850876e-01 -9.26326275e-01 -5.73096156e-01 -2.09920600e-01
-8.87856483e-01 -7.47495964e-02 6.47945225e-01 3.94909054e-01
7.88268745e-01 -1.11370099e+00 -4.34653610e-01 2.29417160e-01
-8.67937505e-02 -3.18057030e-01 2.86844879e-01 8.80417883e-01
-6.69447064e-01 1.11336076e+00 -6.96422458e-01 -6.61578596e-01
-1.10106277e+00 6.36938035e-01 5.87432086e-01 5.87277591e-01
-4.02759582e-01 1.03211153e+00 6.20071113e-01 2.88935881e-02
2.37235233e-01 -3.54115009e-01 -6.14427496e-03 -2.10820418e-02
5.72395623e-01 5.75626016e-01 -3.04840863e-01 -7.60010064e-01
-3.73464018e-01 5.65709829e-01 -2.84766376e-01 -9.78069007e-02
1.49052298e+00 -1.58760801e-01 -3.13271165e-01 5.50011635e-01
8.44928503e-01 -8.15447211e-01 -1.42955756e+00 -9.19936970e-02
9.99121517e-02 -4.15853649e-01 1.03014514e-01 -1.01581752e+00
-3.55797797e-01 1.34892905e+00 9.86543179e-01 4.62570399e-01
1.01603878e+00 2.50780255e-01 3.71720642e-01 2.91946292e-01
-5.27167879e-02 -1.21477461e+00 1.36756534e-02 2.56996363e-01
9.53981996e-01 -1.37211287e+00 -2.99820881e-02 1.82137728e-01
-6.26486123e-01 1.17717445e+00 7.90806949e-01 -2.69240379e-01
5.37306547e-01 -3.17045934e-02 2.44608313e-01 -2.91174948e-01
-7.14232922e-01 -5.85497737e-01 4.29635465e-01 1.13133240e+00
4.46982294e-01 1.15988366e-02 -2.96790034e-01 2.38407895e-01
-1.00533240e-01 2.39621714e-01 6.46518394e-02 6.17165208e-01
-9.88232255e-01 -8.76495540e-01 -4.77168977e-01 7.09532976e-01
-3.52518171e-01 -2.63244752e-02 -4.35406268e-01 1.02071607e+00
5.18542409e-01 7.55510330e-01 3.23644668e-01 -1.50066301e-01
4.79031391e-02 7.13581964e-02 8.23271930e-01 -4.41818953e-01
-4.89415884e-01 2.88773209e-01 -1.65509105e-01 -2.51821131e-01
-5.70411801e-01 -6.52724683e-01 -8.01438332e-01 -4.15905803e-01
-1.53656498e-01 -4.30955857e-01 6.20241225e-01 1.04269791e+00
-2.40243077e-01 2.31308222e-01 -1.86066374e-01 -8.72379541e-01
-4.88022864e-01 -8.96205485e-01 -7.87322938e-01 1.48784772e-01
5.10235071e-01 -5.68074584e-01 -1.70324624e-01 1.65736705e-01] | [9.861238479614258, 2.3656978607177734] |
e4f8b32c-90c2-4a21-83d1-7f000cbb2de2 | ncsu-sas-ning-candidate-generation-and | null | null | https://aclanthology.org/W15-4313 | https://aclanthology.org/W15-4313.pdf | NCSU-SAS-Ning: Candidate Generation and Feature Engineering for Supervised Lexical Normalization | null | ['Ning Jin'] | 2015-07-01 | null | null | null | ws-2015-7 | ['lexical-normalization'] | ['natural-language-processing'] | [-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01] | [-7.419112205505371, 3.731546640396118] |
c7311233-1a02-4c4f-9e46-7b8c4691eac8 | simfbo-towards-simple-flexible-and | 2305.19442 | null | https://arxiv.org/abs/2305.19442v2 | https://arxiv.org/pdf/2305.19442v2.pdf | SimFBO: Towards Simple, Flexible and Communication-efficient Federated Bilevel Learning | Federated bilevel optimization (FBO) has shown great potential recently in machine learning and edge computing due to the emerging nested optimization structure in meta-learning, fine-tuning, hyperparameter tuning, etc. However, existing FBO algorithms often involve complicated computations and require multiple sub-loops per iteration, each of which contains a number of communication rounds. In this paper, we propose a simple and flexible FBO framework named SimFBO, which is easy to implement without sub-loops, and includes a generalized server-side aggregation and update for improving communication efficiency. We further propose System-level heterogeneity robust FBO (ShroFBO) as a variant of SimFBO with stronger resilience to heterogeneous local computation. We show that SimFBO and ShroFBO provably achieve a linear convergence speedup with partial client participation and client sampling without replacement, as well as improved sample and communication complexities. Experiments demonstrate the effectiveness of the proposed methods over existing FBO algorithms. | ['Kaiyi Ji', 'Peiyao Xiao', 'Yifan Yang'] | 2023-05-30 | null | null | null | null | ['bilevel-optimization', 'edge-computing'] | ['methodology', 'time-series'] | [-5.43389618e-01 -4.87578154e-01 -3.84598881e-01 -1.91216737e-01
-8.70927513e-01 -4.19395596e-01 2.64688045e-01 2.91877985e-01
-7.65473545e-02 7.85201788e-01 3.42267752e-01 -3.41393173e-01
-5.65565765e-01 -8.48382592e-01 -7.01121032e-01 -1.03491163e+00
-3.18285495e-01 6.33056343e-01 3.14497888e-01 -1.65252350e-02
-2.00400665e-01 3.85206014e-01 -1.10415924e+00 2.89430887e-01
8.97698760e-01 1.29701364e+00 -1.35440364e-01 6.43813133e-01
-1.74814284e-01 8.68190587e-01 -2.08654836e-01 -3.63526404e-01
3.88947457e-01 -1.84586152e-01 -7.57161975e-01 1.53901666e-01
-1.08836163e-02 -2.36830309e-01 -3.63159269e-01 7.59254932e-01
9.09406960e-01 9.86359268e-02 -1.53398560e-03 -1.60137272e+00
-2.52833277e-01 8.96022558e-01 -5.91240346e-01 2.79009491e-01
7.80984610e-02 2.15900376e-01 9.01104808e-01 -9.85916734e-01
5.44930339e-01 9.80606258e-01 9.31926727e-01 1.41595960e-01
-1.33224273e+00 -5.41111469e-01 1.64258838e-01 4.78114307e-01
-1.45661151e+00 -5.44045985e-01 5.37901819e-01 -6.31663501e-02
6.98261023e-01 7.23148942e-01 6.67313397e-01 3.69192004e-01
-9.06636938e-02 1.13809192e+00 1.23969352e+00 -2.97095329e-01
4.69948113e-01 6.60315901e-02 1.02447063e-01 6.25570774e-01
4.13075686e-01 1.29762851e-02 -7.83582389e-01 -7.98926055e-01
5.39821148e-01 2.01447383e-01 -3.93266022e-01 -6.58567071e-01
-1.20456755e+00 7.17881799e-01 4.51583982e-01 -3.58542763e-02
-4.17386234e-01 5.13437569e-01 6.83151066e-01 4.41221654e-01
4.55522120e-01 -3.50098550e-01 -4.64306980e-01 -4.81103212e-02
-8.21564138e-01 3.83992255e-01 1.30869412e+00 1.22786880e+00
7.81126738e-01 -8.69172141e-02 -4.14366752e-01 8.66302073e-01
5.79159074e-02 2.35524893e-01 4.24823225e-01 -1.03213644e+00
5.18293738e-01 3.94997180e-01 1.44398510e-01 -9.37126637e-01
-3.37532908e-01 -6.15289092e-01 -9.19268548e-01 -4.29404899e-02
2.33034134e-01 -1.77268043e-01 -2.39744231e-01 1.40288711e+00
1.11804402e+00 3.73481840e-01 -1.69269308e-01 1.19802308e+00
7.53620446e-01 6.84941649e-01 -3.32466781e-01 -5.16065896e-01
1.16778445e+00 -1.64416385e+00 -4.90849286e-01 2.77361155e-01
1.03270268e+00 -7.81864226e-01 9.90203083e-01 2.37629905e-01
-1.27076375e+00 3.11241359e-01 -6.52477860e-01 2.85373796e-02
2.35002041e-02 -2.77119309e-01 1.21963334e+00 7.94736922e-01
-9.42165136e-01 2.27651015e-01 -7.74162889e-01 -7.74725229e-02
6.31309927e-01 4.24938947e-01 -7.80477449e-02 -2.42470935e-01
-5.69558024e-01 1.65997580e-01 2.39359960e-01 9.28505659e-02
-1.09519184e+00 -9.85771596e-01 -2.16605619e-01 4.75724697e-01
8.89004290e-01 -1.06869137e+00 1.12909448e+00 -7.05899477e-01
-1.59061790e+00 2.81669974e-01 -3.25853944e-01 -3.27976942e-01
8.82704914e-01 7.96131194e-02 -8.66337344e-02 -1.02194183e-01
-1.25231087e-01 -2.02572450e-01 4.47945952e-01 -9.93909359e-01
-8.39102924e-01 -6.85437858e-01 1.61229402e-01 4.57410574e-01
-4.97385234e-01 -2.16035303e-02 -7.62862921e-01 -7.28936076e-01
1.76583931e-01 -7.52110243e-01 -6.27274275e-01 1.68731645e-01
-1.42428592e-01 -3.77772927e-01 5.80252171e-01 -1.54980078e-01
1.44058669e+00 -2.00722361e+00 -6.22467212e-02 4.72439438e-01
6.01667106e-01 -6.70939907e-02 4.36192080e-02 5.87729931e-01
6.66996241e-01 -6.83844462e-02 5.04347347e-02 -1.79508254e-01
2.21106932e-01 3.13504636e-01 3.63626964e-02 7.46881604e-01
-8.79327357e-01 8.26457977e-01 -8.37782860e-01 -6.35582924e-01
-2.93422908e-01 2.72905648e-01 -9.12375391e-01 -7.96817988e-02
-3.74964505e-01 2.59328008e-01 -6.95050478e-01 1.03643560e+00
8.25220585e-01 -7.04126179e-01 3.51009458e-01 1.32173104e-02
7.11974828e-03 1.84137851e-01 -1.73046207e+00 1.40733862e+00
-7.63730645e-01 1.87347993e-01 7.68215477e-01 -1.03347409e+00
4.83511567e-01 4.20392096e-01 7.10425079e-01 -2.93721676e-01
1.35429621e-01 5.23436546e-01 -5.39785326e-01 -3.50827783e-01
6.62646666e-02 2.45107844e-01 3.31240505e-01 7.77465224e-01
-2.12719738e-01 4.98302341e-01 9.97471884e-02 2.72211999e-01
1.13064623e+00 -2.48735055e-01 2.68098354e-01 -4.00965542e-01
7.17176139e-01 9.98935327e-02 8.50978613e-01 8.14672947e-01
-1.83090717e-01 1.22770689e-01 3.05682987e-01 -7.57039070e-01
-8.46933782e-01 -6.52387798e-01 -7.46408303e-04 1.39524293e+00
4.33954090e-01 -7.96653688e-01 -6.54812157e-01 -5.02146900e-01
3.35424304e-01 -9.70740151e-03 -2.13441998e-01 4.05086815e-01
-6.09021187e-01 -1.22091234e+00 2.95345902e-01 3.29487652e-01
5.74101567e-01 -5.09513795e-01 -1.92056954e-01 5.04917562e-01
-1.22352481e-01 -8.22830200e-01 -1.08870435e+00 -1.14916898e-01
-1.19402635e+00 -9.16876197e-01 -5.07078767e-01 -6.83706641e-01
5.72919011e-01 5.93341529e-01 1.03858781e+00 3.60037833e-01
-4.19564575e-01 2.32811034e-01 -2.17665091e-01 6.43172860e-02
1.20964870e-01 1.65559649e-01 -6.81997091e-02 4.61191982e-01
-2.70881318e-02 -9.14938629e-01 -1.01186585e+00 5.73861957e-01
-8.86815965e-01 7.93121941e-03 3.31945449e-01 1.21292675e+00
9.27570343e-01 -2.54027396e-01 4.14586723e-01 -1.11143136e+00
7.21861064e-01 -6.82229638e-01 -8.83784711e-01 4.27852780e-01
-1.30042946e+00 1.83520149e-02 7.65228510e-01 -3.75010192e-01
-9.99564528e-01 -2.73440126e-02 1.65787712e-01 -3.88807327e-01
4.36923355e-01 5.42180121e-01 -1.42331764e-01 -5.95206320e-01
5.82223535e-01 3.28428358e-01 -2.02422757e-02 -7.09854960e-01
4.03660744e-01 7.79992700e-01 2.23991275e-01 -1.01219189e+00
3.78686190e-01 8.44308138e-01 6.67818487e-02 -3.12360644e-01
-5.32335758e-01 -6.76477253e-01 1.74424544e-01 -6.03683926e-02
-2.41577309e-02 -9.26902950e-01 -1.39712775e+00 1.34121791e-01
-7.38591433e-01 -3.62162679e-01 -1.38506591e-01 4.29911554e-01
-3.93552691e-01 4.13758487e-01 -5.60748100e-01 -6.28841400e-01
-8.86744618e-01 -1.24436808e+00 5.83857775e-01 1.68200985e-01
4.85528678e-01 -1.00758159e+00 -2.28292942e-02 6.31033659e-01
7.05988824e-01 2.52900869e-01 4.96703744e-01 -3.99003595e-01
-1.12887740e+00 -1.16261125e-01 -3.22630525e-01 -3.09691638e-01
-3.59292269e-01 -6.16639435e-01 -4.33982491e-01 -8.08896899e-01
-1.09115444e-01 -2.59136975e-01 4.63242173e-01 2.66782999e-01
1.40478253e+00 -8.99052858e-01 -5.74094951e-01 1.37666726e+00
1.74878120e+00 -2.83817887e-01 6.16729781e-02 5.27731419e-01
4.44898814e-01 7.18157291e-02 3.10609162e-01 1.09789419e+00
6.32407129e-01 7.53600001e-01 4.57043231e-01 6.44675642e-02
-4.90815891e-03 1.33600041e-01 2.03344926e-01 1.01251984e+00
-9.44290757e-02 -5.77957295e-02 -6.41374171e-01 7.54399180e-01
-2.24353266e+00 -6.69099927e-01 -4.56585646e-01 2.22009230e+00
8.28940153e-01 -3.86237174e-01 3.91607046e-01 -4.01177034e-02
6.18255913e-01 6.44013509e-02 -7.19236016e-01 -4.10356551e-01
-8.76338929e-02 4.11122181e-02 9.59735453e-01 5.12297094e-01
-5.46156406e-01 5.25202274e-01 5.94486046e+00 1.22353256e+00
-9.86755073e-01 8.21425974e-01 7.18709528e-01 -6.27767682e-01
-4.64099765e-01 1.95904687e-01 -6.86069489e-01 5.26901543e-01
6.40632927e-01 -5.70070565e-01 1.21428931e+00 1.18711281e+00
2.19871867e-02 4.49420884e-02 -8.20646167e-01 1.31509972e+00
-1.40454724e-01 -1.90021777e+00 -3.37061018e-01 1.37994856e-01
1.15336156e+00 2.92699456e-01 -3.97732884e-01 1.61456056e-02
5.52533269e-01 -8.13343883e-01 4.10795629e-01 -6.28394932e-02
4.86433059e-01 -8.39858532e-01 5.42460442e-01 3.68027896e-01
-1.39042628e+00 -3.67424309e-01 -2.88958520e-01 1.95239633e-02
2.48304874e-01 9.23265517e-01 -4.83972967e-01 7.48808861e-01
1.12171912e+00 5.39338253e-02 -2.03824997e-01 1.37703860e+00
3.89450580e-01 7.57085741e-01 -9.29791689e-01 -3.72160316e-01
2.56115824e-01 -2.45249301e-01 7.03333735e-01 8.88034821e-01
1.98453009e-01 1.01870529e-01 6.84302866e-01 3.13814044e-01
-4.20359462e-01 5.35742939e-01 8.99930373e-02 4.92498934e-01
9.18805003e-01 1.40483999e+00 -5.00280380e-01 -5.22550762e-01
-6.75758183e-01 7.20658541e-01 5.21263838e-01 2.57590026e-01
-8.84549379e-01 -3.68209928e-01 5.85331798e-01 6.18537590e-02
3.48070294e-01 1.38675449e-02 -5.56212664e-01 -1.29971242e+00
3.55438858e-01 -9.13350165e-01 9.20055330e-01 -8.26048777e-02
-1.24245369e+00 4.96922702e-01 -3.26967508e-01 -8.02128851e-01
3.86777997e-01 -1.02564245e-01 -4.64141130e-01 5.77433705e-01
-1.52767038e+00 -1.07980788e+00 -6.28668427e-01 1.02317023e+00
2.07077175e-01 -8.75665247e-03 4.85730886e-01 6.98994696e-01
-6.19163871e-01 1.00850236e+00 8.94574046e-01 -3.06540936e-01
5.31750500e-01 -1.10296714e+00 -2.34488353e-01 6.09239638e-01
-2.72560418e-02 3.82776767e-01 6.52835131e-01 -2.29903087e-01
-2.06268477e+00 -1.10761702e+00 8.30507457e-01 1.18654832e-01
5.56533813e-01 -1.33618489e-01 -5.17471015e-01 6.95226848e-01
2.41356855e-03 8.82171154e-01 7.49411106e-01 2.75481403e-01
-2.57205844e-01 -8.59552801e-01 -1.20720994e+00 6.01730287e-01
1.32322407e+00 -2.01341003e-01 1.81917086e-01 9.11016345e-01
4.80540454e-01 -7.58493721e-01 -1.22034681e+00 2.56002188e-01
4.86418873e-01 -8.92578721e-01 9.14973974e-01 -5.71895838e-01
-2.68533945e-01 -4.41112995e-01 -1.80661380e-01 -8.52046192e-01
-4.42564487e-01 -1.48032582e+00 -7.20858216e-01 1.02025342e+00
1.55819058e-01 -1.01821649e+00 1.08021891e+00 5.19761682e-01
-3.47449109e-02 -1.28852749e+00 -1.30214989e+00 -9.75481153e-01
-3.39021087e-01 -3.32241684e-01 1.25067377e+00 8.00528526e-01
4.31356914e-02 3.81538086e-03 -5.55566967e-01 3.48599106e-01
9.22808588e-01 5.72861552e-01 1.18858862e+00 -8.80450189e-01
-8.89173925e-01 -3.29425573e-01 -5.47336787e-03 -1.21545446e+00
-6.19592428e-01 -1.24217129e+00 -3.35124582e-01 -1.27141678e+00
4.22253639e-01 -9.36329126e-01 -2.95263320e-01 3.69472980e-01
-1.04906879e-01 3.21857661e-01 2.01615319e-01 5.91133833e-01
-9.36801851e-01 5.33947051e-01 1.20499277e+00 2.34309942e-01
-3.70034516e-01 1.05363823e-01 -3.95726919e-01 3.33805323e-01
4.49195951e-01 -6.12312555e-01 -1.36989430e-01 -5.77803731e-01
3.73646677e-01 4.55933243e-01 1.54289752e-01 -5.81241429e-01
5.67416608e-01 -2.21873596e-01 -1.04060389e-01 -3.05789709e-01
-4.58193086e-02 -9.47043538e-01 6.40240729e-01 6.88375533e-01
-8.36424306e-02 2.22112209e-01 -3.31456095e-01 7.69624650e-01
-7.70593435e-02 7.62853697e-02 5.36204278e-01 -9.43898689e-03
-2.26858497e-01 8.10727954e-01 -5.18239327e-02 1.48702979e-01
1.36508274e+00 4.12200466e-02 -3.20635408e-01 -2.85294324e-01
-7.30950356e-01 9.02423084e-01 2.65836686e-01 -1.30251989e-01
1.84869945e-01 -1.58326006e+00 -6.87116504e-01 -2.37898961e-01
-4.58296277e-02 1.33434281e-01 3.62564951e-01 1.54449022e+00
-9.15914416e-01 3.89691204e-01 3.26850802e-01 -6.19579852e-01
-1.41490483e+00 7.93007195e-01 2.49094158e-01 -6.55913830e-01
-8.95542562e-01 7.42671371e-01 -1.61337525e-01 -5.46862185e-01
4.71323758e-01 8.90603960e-02 6.35000050e-01 -2.64692336e-01
4.87741560e-01 1.09048665e+00 1.30631685e-01 -1.28400311e-01
-3.36795807e-01 7.29054511e-02 1.43855467e-01 2.25389436e-01
1.46088326e+00 -3.95450741e-01 -5.40655136e-01 -1.01565808e-01
1.04809177e+00 1.11742020e-01 -1.04420733e+00 -7.43373632e-01
-1.10372819e-01 -7.89547205e-01 3.26574057e-01 -6.75362110e-01
-1.36962807e+00 3.52700174e-01 3.55145752e-01 7.51210004e-02
1.30961573e+00 -8.69741514e-02 1.22306693e+00 3.89206052e-01
7.66271770e-01 -1.01117063e+00 -4.24078405e-01 -2.64779385e-02
5.86716890e-01 -9.50398743e-01 2.70799339e-01 -5.54139614e-01
-2.49523684e-01 1.08361149e+00 4.70645398e-01 -3.34161893e-02
8.72938633e-01 3.22690815e-01 -3.26298743e-01 -3.81859131e-02
-1.22047734e+00 6.42461404e-02 -1.80905282e-01 4.58905427e-03
2.85660829e-02 1.91756502e-01 -9.55310225e-01 8.15446436e-01
1.00674160e-01 4.17903125e-01 6.56049177e-02 8.66791010e-01
-1.99193776e-01 -1.16775322e+00 -6.91768885e-01 4.71265435e-01
-5.96408486e-01 -8.75225738e-02 2.24518135e-01 4.39530164e-01
-8.28292444e-02 6.25479221e-01 -3.77035439e-01 6.97710086e-04
-6.67721182e-02 -2.10828468e-01 2.27304578e-01 3.05045359e-02
-9.40589309e-01 2.66677320e-01 1.47003487e-01 -8.97250056e-01
-1.81691011e-03 -4.51103836e-01 -1.08481431e+00 -7.85977721e-01
-6.28254712e-01 5.45832336e-01 5.79112530e-01 8.12840223e-01
8.87161732e-01 3.25623870e-01 1.10338140e+00 -4.54662710e-01
-9.53063548e-01 -5.61940670e-01 -5.05286396e-01 1.68449417e-01
2.02439487e-01 -3.11962783e-01 -1.49802834e-01 -2.41610736e-01] | [6.229273319244385, 5.083626747131348] |
089a310a-2a3d-4741-9e73-b6aa7de07e07 | generating-dispatching-rules-for-the | 2302.02506 | null | https://arxiv.org/abs/2302.02506v1 | https://arxiv.org/pdf/2302.02506v1.pdf | Generating Dispatching Rules for the Interrupting Swap-Allowed Blocking Job Shop Problem Using Graph Neural Network and Reinforcement Learning | The interrupting swap-allowed blocking job shop problem (ISBJSSP) is a complex scheduling problem that is able to model many manufacturing planning and logistics applications realistically by addressing both the lack of storage capacity and unforeseen production interruptions. Subjected to random disruptions due to machine malfunction or maintenance, industry production settings often choose to adopt dispatching rules to enable adaptive, real-time re-scheduling, rather than traditional methods that require costly re-computation on the new configuration every time the problem condition changes dynamically. To generate dispatching rules for the ISBJSSP problem, a method that uses graph neural networks and reinforcement learning is proposed. ISBJSSP is formulated as a Markov decision process. Using proximal policy optimization, an optimal scheduling policy is learnt from randomly generated instances. Employing a set of reported benchmark instances, we conduct a detailed experimental study on ISBJSSP instances with a range of machine shutdown probabilities to show that the scheduling policies generated can outperform or are at least as competitive as existing dispatching rules with predetermined priority. This study shows that the ISBJSSP, which requires real-time adaptive solutions, can be scheduled efficiently with the proposed machine learning method when production interruptions occur with random machine shutdowns. | ['Kincho H. Law', 'Jinkyoo Park', 'Junyoung Park', 'Sang Hun Kim', 'Vivian W. H. Wong'] | 2023-02-05 | null | null | null | null | ['blocking'] | ['natural-language-processing'] | [ 5.64041495e-01 9.49260667e-02 -4.44771081e-01 -2.41651759e-01
-1.41461268e-01 -3.66350710e-01 2.71230102e-01 3.42319429e-01
-1.26781836e-01 1.08077002e+00 -3.69895786e-01 -8.51142645e-01
-8.80940199e-01 -6.14255071e-01 -8.23083401e-01 -7.62767971e-01
-5.14187157e-01 1.36977160e+00 -5.54078780e-02 -2.23512352e-01
4.21958238e-01 8.45226824e-01 -1.27010608e+00 1.62438508e-02
7.11772263e-01 9.54409122e-01 8.44450593e-01 7.38188148e-01
-1.76007763e-01 5.40227115e-01 -9.66118932e-01 4.79870498e-01
5.02141237e-01 1.63886458e-01 -9.25990045e-01 5.73062062e-01
-6.90365613e-01 -1.03828348e-02 -1.54398128e-01 6.28081083e-01
1.86087620e-02 7.54393220e-01 4.21078026e-01 -1.79109144e+00
-4.99654889e-01 6.02055192e-01 -6.16433620e-01 4.26867217e-01
4.44636852e-01 3.54661375e-01 4.33132768e-01 -1.16424546e-01
3.00262779e-01 1.25635099e+00 -4.34124697e-04 1.59163266e-01
-1.32271397e+00 -1.43469930e-01 5.02852440e-01 3.20020437e-01
-9.88961458e-01 2.90102005e-01 6.99513137e-01 -6.63576508e-03
1.45716691e+00 2.58678257e-01 4.05408800e-01 8.26286852e-01
1.08757806e+00 5.37423790e-01 1.16625929e+00 -6.26535416e-01
5.49864769e-01 -3.74435008e-01 -1.48089856e-01 2.06046283e-01
1.66586161e-01 3.77116382e-01 1.25130042e-01 3.65317659e-03
4.31689918e-01 3.80814612e-01 5.50722517e-02 -3.07115912e-01
-1.18667197e+00 6.86789215e-01 5.68456203e-02 -3.28785703e-02
-9.33269203e-01 -1.10388920e-01 5.78481436e-01 5.97353399e-01
-9.25539732e-02 9.92882848e-01 -8.68120015e-01 2.10168734e-02
-6.71479583e-01 1.68675125e-01 1.14455342e+00 1.35738838e+00
1.59692556e-01 3.70628506e-01 -4.59174335e-01 4.72221494e-01
-1.39157668e-01 3.56025994e-01 5.28565496e-02 -1.03991079e+00
5.03794670e-01 1.31824061e-01 7.98499823e-01 -7.17049181e-01
-7.77284026e-01 -2.12680385e-01 -8.63258243e-01 1.99848056e-01
-1.45503785e-02 -1.68792143e-01 -1.33963716e+00 1.10605514e+00
2.62688905e-01 1.56851287e-03 -8.04371387e-02 7.73524046e-01
-2.87727535e-01 1.13088334e+00 -9.11557153e-02 -9.48905289e-01
9.92225468e-01 -1.27309203e+00 -1.18289506e+00 -9.97214317e-02
1.48369372e-01 -8.04841101e-01 7.04635680e-01 6.82175398e-01
-1.13057148e+00 -3.80583316e-01 -1.02708793e+00 8.99848878e-01
-4.09550548e-01 -2.83024311e-01 7.43959427e-01 4.64023314e-02
-7.77393878e-01 8.18057775e-01 -9.28239465e-01 -1.06374696e-01
-2.71370381e-01 6.70018792e-01 -1.52184159e-01 -4.17871565e-01
-1.17890525e+00 1.15509868e+00 9.59163725e-01 5.68462074e-01
-1.22452462e+00 -4.10207123e-01 -8.97016764e-01 2.28873253e-01
1.31807172e+00 -3.65848809e-01 1.55220878e+00 -5.66234171e-01
-1.68441069e+00 -2.23351434e-01 3.06855947e-01 -3.26808482e-01
1.65053964e-01 2.58997202e-01 -7.68876076e-01 -8.95684063e-02
-8.94478559e-02 7.19753443e-04 9.32858467e-01 -1.38887978e+00
-7.96781898e-01 1.26770297e-02 1.01865314e-01 2.31010020e-01
4.20753747e-01 -8.99982452e-02 -4.20146845e-02 -2.07696676e-01
-1.23637058e-01 -1.36142814e+00 -7.86096811e-01 -1.38564634e+00
-7.85350263e-01 -3.02057415e-01 6.32756412e-01 -5.42927444e-01
1.17097175e+00 -1.67894900e+00 2.96691656e-01 4.60591882e-01
-6.68288291e-01 2.50122458e-01 -2.47125238e-01 1.10929167e+00
-2.40279421e-01 -1.22796312e-01 -2.76120491e-02 3.80234011e-02
2.29322270e-01 9.91488814e-01 7.89557397e-03 8.54778066e-02
-2.41248030e-02 3.17051351e-01 -1.04595089e+00 2.33323649e-02
5.08716285e-01 -5.10004044e-01 8.52100626e-02 4.71489727e-01
-5.68975031e-01 3.93217355e-01 -2.89112747e-01 9.66462076e-01
5.68313420e-01 -9.11873356e-02 5.12665272e-01 3.44534099e-01
3.14369872e-02 -3.11434537e-01 -1.26977360e+00 1.14771914e+00
-9.01170790e-01 -9.54657197e-02 2.65014052e-01 -1.50368094e+00
8.26546133e-01 9.89478156e-02 6.15066826e-01 -6.97945535e-01
-2.55484991e-02 -1.11957327e-01 3.60030383e-02 -5.74216843e-01
5.19599438e-01 -2.10066773e-02 -2.16791257e-01 2.07470760e-01
-2.72041589e-01 -1.79813936e-01 5.61318398e-01 -1.09788947e-01
1.47945511e+00 1.70714594e-02 2.32982695e-01 -2.94306606e-01
4.44347978e-01 2.29492098e-01 9.38019037e-01 9.62207615e-01
-1.30766645e-01 -8.63735285e-03 6.16966307e-01 -8.05663049e-01
-8.96056890e-01 -8.98150563e-01 3.83834958e-01 9.55404222e-01
4.10864890e-01 3.49504977e-01 -1.52317986e-01 -7.19751477e-01
4.29599941e-01 1.34719074e+00 -3.79012197e-01 -1.76193953e-01
-3.98685962e-01 -7.63182998e-01 -5.30399561e-01 3.92792791e-01
-7.54220411e-02 -1.38974500e+00 -4.08465594e-01 7.17223704e-01
3.02689075e-01 -1.16533947e+00 -6.13424480e-01 8.40818048e-01
-5.52384317e-01 -1.00985956e+00 -2.19038665e-01 -7.15366125e-01
1.02816188e+00 3.14752072e-01 1.09882593e+00 4.68156300e-02
-4.09909397e-01 1.96182758e-01 -2.22213328e-01 -4.99024630e-01
-7.71960676e-01 2.06331804e-01 3.52580816e-01 -3.47375363e-01
-6.60521835e-02 -2.22571582e-01 -3.38860661e-01 6.37521327e-01
-1.04541969e+00 5.43442219e-02 7.76322305e-01 1.22006643e+00
9.76504803e-01 1.06325495e+00 1.08947968e+00 -9.62499499e-01
7.46982992e-01 -5.45427561e-01 -9.81454194e-01 6.45634830e-01
-1.08378208e+00 2.45529026e-01 1.15509534e+00 -4.37628776e-01
-1.36255884e+00 2.65088286e-02 3.95148516e-01 -5.61325669e-01
-4.38615918e-01 7.73666441e-01 -1.16537176e-02 3.95036727e-01
-1.80789500e-01 2.52670050e-01 9.18660834e-02 -1.11260660e-01
-1.96221322e-01 3.59754860e-01 2.46902078e-01 -7.26623774e-01
5.91066897e-01 -2.14466467e-01 4.52181101e-01 -2.55900890e-01
-6.25873744e-01 -4.18574035e-01 -3.92421782e-01 -1.78126171e-01
4.55913931e-01 -1.71711832e-01 -1.20363116e+00 2.52921224e-01
-7.38150179e-01 -5.82564890e-01 -9.95767768e-03 3.23171347e-01
-9.85989809e-01 5.53692952e-02 -7.17623830e-01 -9.87246275e-01
7.94224292e-02 -1.20629072e+00 6.36139154e-01 2.52277166e-01
1.77059874e-01 -9.52377975e-01 -3.18745255e-01 4.04334217e-01
3.24257195e-01 5.09974003e-01 1.27958786e+00 -7.34331071e-01
-5.52319884e-01 -3.62978399e-01 4.11145657e-01 3.25020015e-01
7.00225353e-01 5.60814738e-02 1.01767384e-01 -9.34638560e-01
-2.04044849e-01 -7.47142965e-03 -5.24692833e-02 4.88600880e-01
1.49398065e+00 -6.44839525e-01 -4.71164107e-01 1.02162343e-02
1.41895080e+00 1.18975341e+00 3.22306037e-01 6.47738814e-01
2.23238930e-01 6.63798273e-01 1.47654259e+00 7.03357935e-01
1.67992562e-01 5.75260699e-01 9.00484681e-01 3.20749693e-02
7.27399528e-01 -5.71056977e-02 2.37947509e-01 7.58313358e-01
2.19091233e-02 -5.25573730e-01 -9.18475270e-01 4.16886747e-01
-2.03016496e+00 -8.22283804e-01 2.01654688e-01 2.47035646e+00
4.80178237e-01 7.57494748e-01 -1.13127559e-01 4.70956452e-02
9.82873082e-01 -1.68898255e-01 -6.74397886e-01 -1.39814854e+00
5.08649170e-01 -2.48015895e-01 9.01848674e-01 4.80429739e-01
-8.34709287e-01 6.27769709e-01 5.78078938e+00 5.11168778e-01
-7.40274131e-01 -4.19739455e-01 7.40014851e-01 -2.39470690e-01
2.35057339e-01 -1.69307943e-02 -6.30935848e-01 8.36246014e-01
1.41286242e+00 -1.10721670e-01 1.23919368e+00 7.54653156e-01
8.06297779e-01 -2.67298192e-01 -1.27417946e+00 5.27590752e-01
-4.11815107e-01 -1.22791815e+00 -2.37846106e-01 -5.48785925e-02
1.09188080e+00 -3.76191914e-01 -2.46065751e-01 5.94496071e-01
5.64175606e-01 -1.05015779e+00 3.40526223e-01 5.91973424e-01
1.71388611e-01 -1.38219404e+00 1.28247261e+00 5.72869241e-01
-8.30923378e-01 -7.45672345e-01 -2.26946354e-01 -2.18912423e-01
6.75956309e-01 5.83589196e-01 -1.36884737e+00 1.13988781e+00
4.12143528e-01 -1.40612975e-01 2.07248881e-01 8.72265100e-01
-6.35630041e-02 2.66185373e-01 -1.97182372e-01 -5.92739172e-02
5.03901243e-01 -2.68299848e-01 3.88858736e-01 8.08089495e-01
2.76661545e-01 -2.05010712e-01 1.15081763e+00 4.49468464e-01
5.84748507e-01 -4.39874083e-01 -5.78363478e-01 -2.27041408e-01
6.62113428e-01 1.25234437e+00 -1.21315610e+00 2.48795412e-02
-7.41570443e-02 8.54159713e-01 -1.06167525e-01 6.28174007e-01
-9.82211232e-01 -5.78395128e-01 3.67502898e-01 -5.09442575e-02
3.35758656e-01 -5.01430333e-01 1.12402454e-01 -2.08076522e-01
-1.29969731e-01 -7.30125070e-01 2.99643725e-01 -6.83665097e-01
-1.28366923e+00 3.89703006e-01 2.17626750e-01 -8.41622651e-01
-5.00881851e-01 -6.05159521e-01 -6.72098041e-01 7.06994772e-01
-1.67810869e+00 -6.54254615e-01 4.16575760e-01 2.66409725e-01
1.08530390e+00 -8.32028389e-02 5.73075712e-01 -1.01273216e-01
-9.12728131e-01 -1.11944251e-01 3.99093062e-01 -7.63939321e-01
2.51255423e-01 -1.59394014e+00 1.48404047e-01 6.21464193e-01
-6.63773358e-01 3.55446965e-01 1.16555536e+00 -8.87806654e-01
-2.07427931e+00 -1.14280593e+00 4.27086234e-01 1.20504268e-01
5.84103942e-01 -1.36233225e-01 -7.99520791e-01 7.52402663e-01
3.77818078e-01 -4.85918745e-02 2.48207837e-01 1.31081827e-02
8.04403603e-01 -3.44297260e-01 -1.06677818e+00 2.79983699e-01
6.91693306e-01 3.37593228e-01 -5.21542549e-01 8.77283454e-01
8.28697681e-01 -7.41342485e-01 -9.36837137e-01 5.15029609e-01
-2.98146755e-01 -1.43209472e-01 5.60792446e-01 -1.10270703e+00
3.15015838e-02 -2.29261518e-01 1.52466148e-01 -2.01680470e+00
-8.06278586e-01 -1.07153332e+00 -3.14714551e-01 1.06227779e+00
4.99088675e-01 -4.95647401e-01 3.98374617e-01 7.50241220e-01
-5.77368021e-01 -1.06302369e+00 -1.11221015e+00 -1.38773620e+00
-6.01540625e-01 4.79522422e-02 1.05812824e+00 7.40125000e-01
-2.03519776e-01 7.23931864e-02 -4.28965181e-01 6.54213130e-01
4.71097529e-01 4.48923945e-01 3.17856312e-01 -7.56742716e-01
-3.11150789e-01 2.25833952e-02 1.05112113e-01 -4.46332395e-01
5.45957029e-01 -4.80350584e-01 6.32028282e-01 -1.91126490e+00
-3.93408030e-01 -6.82590485e-01 -9.58328187e-01 5.84440708e-01
1.48533925e-01 -1.04578817e+00 1.85688302e-01 -1.98293522e-01
-8.03559780e-01 5.89676917e-01 1.60258543e+00 -2.75533557e-01
-5.88081121e-01 5.96362829e-01 -2.34580636e-01 3.00533831e-01
9.70752656e-01 -3.40289176e-01 -7.73131847e-01 -7.53456503e-02
1.71580762e-01 9.78163242e-01 -3.25202554e-01 -5.74428976e-01
1.52085707e-01 -1.21543086e+00 4.25923653e-02 -6.82072997e-01
3.40196714e-02 -1.29185724e+00 6.07337356e-01 7.29745924e-01
-3.78243059e-01 8.46336842e-01 2.59258181e-01 1.07944643e+00
6.37646392e-02 -3.58115524e-01 2.82035470e-01 -1.09147139e-01
-9.35384274e-01 3.08959544e-01 -7.54311860e-01 -5.10147035e-01
1.55715549e+00 -1.17706239e-01 -6.68518543e-02 -1.58098459e-01
-9.55234885e-01 1.07930946e+00 -1.09629132e-01 8.10582757e-01
4.41197067e-01 -1.11322546e+00 -2.97046989e-01 1.25806849e-03
-3.03861082e-01 1.11430071e-01 4.83830452e-01 6.58399045e-01
-3.67823780e-01 7.19997883e-01 -4.17989999e-01 -4.68495220e-01
-7.15966702e-01 1.49969530e+00 -1.22819290e-01 -8.06008339e-01
-4.68351156e-01 2.02269852e-01 -4.19338942e-01 -3.80725056e-01
8.17040950e-02 -4.51967806e-01 9.52218100e-02 -2.43944243e-01
2.47896329e-01 6.07140303e-01 4.25202608e-01 1.69897377e-01
-2.04167649e-01 -4.89846438e-01 -2.96390742e-01 3.70106965e-01
1.29764903e+00 -2.54029781e-01 -1.55577192e-03 3.89051020e-01
2.94380546e-01 -8.11571896e-01 -1.37506497e+00 6.82762265e-02
3.63894522e-01 -5.82137287e-01 3.01654171e-02 -1.29725850e+00
-1.04828811e+00 4.92902733e-02 3.39897245e-01 8.05107951e-01
1.18698871e+00 -3.91169399e-01 6.64368391e-01 5.52293777e-01
8.37022603e-01 -1.69881094e+00 -9.29611027e-02 6.62010133e-01
9.98522520e-01 -1.02437544e+00 -2.18824282e-01 -3.15181941e-01
-9.61509585e-01 9.37143207e-01 8.35888088e-01 6.10388406e-02
3.20749253e-01 3.94242048e-01 -3.74343812e-01 7.73902759e-02
-1.09744847e+00 3.02821308e-01 -3.77673388e-01 6.47789896e-01
-3.86705577e-01 6.19167924e-01 -2.82054394e-01 2.33356744e-01
2.82564580e-01 4.41740267e-02 8.07246029e-01 1.50984120e+00
-1.98950514e-01 -1.06478333e+00 -5.46659648e-01 7.11245298e-01
-2.06049547e-01 5.13762712e-01 3.77348334e-01 9.02887344e-01
-1.71046421e-01 1.07575822e+00 9.55168530e-02 5.79596646e-02
5.33334553e-01 -5.53550012e-02 3.77445489e-01 -7.98101306e-01
-2.38357037e-01 -7.62758106e-02 2.96985745e-01 -5.95661521e-01
3.81527729e-02 -4.10102755e-01 -1.60955322e+00 -2.25723147e-01
-4.54345047e-01 3.50858152e-01 8.98262084e-01 1.11045301e+00
4.21541393e-01 1.35332310e+00 1.21228445e+00 -9.81438100e-01
-1.16447234e+00 -6.48623109e-01 -9.31325495e-01 1.19382210e-01
1.12785093e-01 -1.18382800e+00 -1.73014954e-01 -3.56589496e-01] | [4.809714317321777, 2.4279680252075195] |
c560a10a-57bc-4960-97bf-934d2787007e | dan-net-dual-domain-adaptive-scaling-non | 2102.08003 | null | https://arxiv.org/abs/2102.08003v1 | https://arxiv.org/pdf/2102.08003v1.pdf | DAN-Net: Dual-Domain Adaptive-Scaling Non-local Network for CT Metal Artifact Reduction | Metal implants can heavily attenuate X-rays in computed tomography (CT) scans, leading to severe artifacts in reconstructed images, which significantly jeopardize image quality and negatively impact subsequent diagnoses and treatment planning. With the rapid development of deep learning in the field of medical imaging, several network models have been proposed for metal artifact reduction (MAR) in CT. Despite the encouraging results achieved by these methods, there is still much room to further improve performance. In this paper, a novel Dual-domain Adaptive-scaling Non-local network (DAN-Net) for MAR. We correct the corrupted sinogram using adaptive scaling first to preserve more tissue and bone details as a more informative input. Then, an end-to-end dual-domain network is adopted to successively process the sinogram and its corresponding reconstructed image generated by the analytical reconstruction layer. In addition, to better suppress the existing artifacts and restrain the potential secondary artifacts caused by inaccurate results of the sinogram-domain network, a novel residual sinogram learning strategy and nonlocal module are leveraged in the proposed network model. In the experiments, the proposed DAN-Net demonstrates performance competitive with several state-of-the-art MAR methods in both qualitative and quantitative aspects. | ['Yi Zhang', 'Jiliu Zhou', 'Hu Chen', 'Yan Liu', 'Huaiqiang Sun', 'Yongqiang Huang', 'Wenjun Xia', 'Tao Wang'] | 2021-02-16 | null | null | null | null | ['metal-artifact-reduction'] | ['medical'] | [ 4.06525135e-01 8.52997079e-02 1.52033418e-01 -3.35725456e-01
-9.51931894e-01 2.15156674e-01 1.31329238e-01 -8.34434014e-03
-3.07295024e-01 6.90453947e-01 3.63782197e-01 -5.05772494e-02
-2.86658913e-01 -7.61040151e-01 -5.08066356e-01 -9.85868931e-01
1.63487628e-01 2.27514759e-01 5.32250226e-01 8.42073187e-03
-1.55918896e-01 5.94596505e-01 -5.68505764e-01 4.48786110e-01
8.98401976e-01 9.71154809e-01 3.87066603e-01 8.71547684e-02
1.95652187e-01 1.11981225e+00 -2.38530934e-01 -1.27821937e-01
9.60925967e-02 -7.24689782e-01 -5.68484068e-01 -2.08354294e-01
5.86323142e-02 -4.31608230e-01 -7.26126015e-01 1.28822577e+00
9.26300347e-01 -4.16671811e-03 4.54645276e-01 -4.89542693e-01
-5.57806253e-01 7.60554552e-01 -7.54304647e-01 3.59800279e-01
-1.30077109e-01 2.96196580e-01 1.16127707e-01 -1.04814625e+00
5.42979181e-01 9.48424339e-01 1.09007883e+00 5.31749308e-01
-9.59133804e-01 -7.56265759e-01 -3.55370760e-01 1.70596689e-01
-1.13946414e+00 -1.79052845e-01 1.09171546e+00 -1.99618906e-01
5.50614655e-01 2.56380141e-01 7.02586949e-01 9.92513120e-01
8.81451011e-01 5.64298391e-01 9.70561862e-01 -2.02037945e-01
2.11975634e-01 -2.30099887e-01 -1.38405055e-01 6.77847624e-01
2.54121721e-01 1.72463015e-01 -2.33890846e-01 -9.36782584e-02
1.17098224e+00 4.20405775e-01 -5.83481848e-01 -2.16316015e-01
-1.02294993e+00 5.21149397e-01 1.09410715e+00 5.68158865e-01
-9.39347327e-01 3.03907871e-01 5.14225364e-01 -1.61858395e-01
5.79918325e-01 2.58282751e-01 -4.01873291e-02 3.25835258e-01
-1.02901208e+00 9.62543562e-02 5.53350523e-02 3.67249995e-01
1.04250811e-01 3.79579306e-01 -3.62730324e-01 9.09035683e-01
4.18403894e-01 4.43436921e-01 8.11247170e-01 -4.41335976e-01
3.50718021e-01 4.74947125e-01 -1.88928500e-01 -9.89429832e-01
-7.46091127e-01 -8.88989687e-01 -1.47220254e+00 3.06340933e-01
1.74940988e-01 2.05834001e-01 -1.34162211e+00 1.42593241e+00
5.05266368e-01 4.45825279e-01 -1.87357038e-01 1.36736023e+00
9.26972568e-01 5.02509892e-01 1.37362346e-01 -5.26891351e-01
1.20644176e+00 -8.18421364e-01 -9.56667185e-01 -2.17461005e-01
3.77793938e-01 -9.30855513e-01 8.10343027e-01 3.51557136e-01
-1.50894618e+00 -3.30809414e-01 -1.30103970e+00 6.31062835e-02
3.90586674e-01 -5.87235317e-02 3.21824223e-01 2.83129960e-01
-7.40318954e-01 9.68235254e-01 -1.43824637e+00 2.50858009e-01
8.42564821e-01 3.37594241e-01 8.20882767e-02 -4.67660189e-01
-1.08487833e+00 1.01089478e+00 -1.19445130e-01 5.91115355e-01
-8.15166116e-01 -1.03815389e+00 -6.94787800e-01 -2.56474793e-01
2.91112840e-01 -7.77619839e-01 1.37437689e+00 -8.79304349e-01
-1.54182172e+00 3.08667183e-01 2.43421763e-01 -2.57703543e-01
8.92817736e-01 -7.13678673e-02 -5.32682061e-01 1.76220685e-01
1.53181851e-01 1.16241984e-01 7.85285532e-01 -1.28499031e+00
-8.75579342e-02 -5.20043910e-01 -6.39478981e-01 1.47980049e-01
-5.81895299e-02 3.26055326e-02 -4.82819229e-01 -1.23813736e+00
7.56529391e-01 -8.44687939e-01 -6.00179732e-01 4.00758922e-01
-3.14407468e-01 2.31368408e-01 4.72540796e-01 -1.10189819e+00
1.16812420e+00 -2.05521798e+00 7.55916862e-03 3.31873327e-01
3.77698839e-01 1.88623965e-01 -6.86576217e-02 -2.33185351e-01
-5.28467894e-01 -3.00800562e-01 -7.67893016e-01 -2.57134259e-01
-8.29030156e-01 1.18913546e-01 1.51654258e-01 9.40869331e-01
-7.36576542e-02 7.82856345e-01 -9.03470218e-01 -4.18364495e-01
3.13893348e-01 8.30202997e-01 -3.34163874e-01 1.40373027e-02
2.39031076e-01 7.94656992e-01 -6.64929152e-01 5.71984231e-01
1.09190309e+00 -4.01251405e-01 4.71259244e-02 -5.52853644e-01
1.32059902e-02 1.86532348e-01 -9.86586571e-01 1.86922622e+00
-6.48662508e-01 2.04377070e-01 3.10547352e-01 -7.71938622e-01
3.49237233e-01 5.37144959e-01 9.42003191e-01 -1.13565731e+00
3.23772132e-01 5.60829401e-01 2.56826699e-01 -6.08495772e-01
-1.00517072e-01 -7.41864920e-01 4.54519659e-01 1.70211881e-01
-3.86978507e-01 -1.32275477e-01 -5.82302332e-01 -4.88325991e-02
1.18064106e+00 -1.31827563e-01 6.34000078e-02 -2.84680545e-01
5.82647860e-01 -4.92769480e-02 7.86848247e-01 4.04734910e-01
-2.38234326e-01 1.05417454e+00 -1.31612286e-01 -6.97456121e-01
-1.00596511e+00 -1.16687310e+00 -4.11242157e-01 2.01058879e-01
3.27100068e-01 1.74887985e-01 -6.25167072e-01 -5.50263226e-01
-3.59386891e-01 5.25820911e-01 -6.15907609e-01 -3.68060529e-01
-1.19443369e+00 -9.86915231e-01 3.11413348e-01 8.84069741e-01
5.72104692e-01 -1.26230896e+00 -4.28381503e-01 6.15324318e-01
-2.85114706e-01 -8.09621632e-01 -6.14699185e-01 2.93864876e-01
-1.32318079e+00 -9.26337302e-01 -1.03063452e+00 -7.26438165e-01
7.61407852e-01 2.65558749e-01 8.15447211e-01 2.62167633e-01
-2.91041017e-01 -2.81438172e-01 -1.69744223e-01 -2.18418509e-01
-6.63968563e-01 -3.64905238e-01 -1.91957250e-01 -2.14319453e-01
-3.81857574e-01 -6.79206431e-01 -1.00806475e+00 2.37852752e-01
-1.27096033e+00 3.01822960e-01 8.45819950e-01 1.03739405e+00
7.65063107e-01 1.61371052e-01 5.57975173e-01 -8.57227743e-01
4.34988618e-01 -4.17928725e-01 -3.99053246e-01 9.11877975e-02
-6.10452890e-01 -1.97001733e-02 8.56856525e-01 -2.24602520e-01
-1.34654260e+00 9.91939083e-02 -5.02133727e-01 -3.95032525e-01
3.44924927e-01 7.27703154e-01 1.79160863e-01 -3.09117317e-01
8.33701909e-01 2.65256643e-01 1.72423646e-01 -3.29755485e-01
-2.27126062e-01 1.80659398e-01 6.03269756e-01 -1.18762851e-02
8.09376061e-01 6.51515126e-01 2.69106448e-01 -5.59602082e-01
-8.62804711e-01 -2.54254192e-01 -1.65196121e-01 -2.41261765e-01
1.04777288e+00 -8.38517904e-01 -3.19719970e-01 6.62766993e-01
-1.19624054e+00 -1.15067057e-01 -2.25846171e-01 7.61653066e-01
-2.69789189e-01 5.00511885e-01 -1.04778373e+00 -3.74704897e-01
-8.19646418e-01 -1.60019791e+00 6.48711324e-01 2.63144672e-01
3.31672020e-02 -6.86128736e-01 -5.72549589e-02 2.07073748e-01
7.83833683e-01 2.60082543e-01 9.08620298e-01 -2.36019239e-01
-5.76192915e-01 -1.01868615e-01 -3.12582880e-01 4.56674784e-01
2.76731074e-01 -6.67029381e-01 -7.95405269e-01 -3.04022342e-01
7.25216925e-01 -4.23083603e-02 7.10540771e-01 9.84486401e-01
1.36858177e+00 -1.52401272e-02 -2.34023750e-01 8.87360394e-01
1.58959496e+00 4.13244963e-01 7.62407899e-01 3.41741949e-01
7.56044030e-01 2.11029164e-02 2.79603720e-01 3.16356480e-01
-1.41356379e-01 5.42872787e-01 5.80101788e-01 -3.90845120e-01
-6.49180055e-01 4.36541922e-02 -6.99460283e-02 8.75724554e-01
-1.29010037e-01 7.40256160e-03 -9.29864287e-01 4.17855620e-01
-1.76296401e+00 -7.39742637e-01 -4.70204562e-01 1.96400416e+00
8.69209230e-01 9.84709933e-02 -3.72180730e-01 -2.78044697e-02
6.33362412e-01 3.40726003e-02 -6.79863274e-01 2.82130122e-01
-2.28736494e-02 5.80459416e-01 5.80239952e-01 3.46619904e-01
-9.00255680e-01 2.43916720e-01 5.37171936e+00 9.66945350e-01
-1.37483454e+00 4.72647607e-01 7.02075541e-01 4.22673747e-02
-2.85846084e-01 -3.95265073e-01 3.12128980e-02 4.50402796e-01
5.34333110e-01 2.38894090e-01 2.70271689e-01 6.56551421e-01
5.68220019e-01 -1.93708614e-01 -7.32715905e-01 9.70429897e-01
-6.64656907e-02 -1.53017902e+00 -2.50885576e-01 -2.11890385e-01
7.79540837e-01 1.96456954e-01 8.97147320e-03 -2.86505185e-02
-4.06368375e-02 -9.10899222e-01 5.79587162e-01 4.55049068e-01
7.51807570e-01 -7.80128419e-01 1.01641428e+00 9.10876617e-02
-9.18952703e-01 7.56217316e-02 -2.71696955e-01 4.75050628e-01
3.18236083e-01 8.97710443e-01 -8.00929785e-01 8.68891299e-01
6.44952774e-01 5.74663997e-01 -3.11086625e-01 1.28547966e+00
-2.16673002e-01 6.17487907e-01 -2.61092305e-01 4.07161057e-01
2.52435863e-01 -3.56959775e-02 5.59446096e-01 8.74921679e-01
3.79729986e-01 4.55953509e-01 -1.13730565e-01 7.98306942e-01
-2.70748079e-01 -1.25265464e-01 -1.58962473e-01 7.31908083e-01
6.87149242e-02 1.20606589e+00 -9.34994757e-01 -4.43545878e-01
-4.02410597e-01 9.25810516e-01 -1.48461923e-01 2.34729201e-01
-9.49070871e-01 1.73521176e-01 4.22932804e-02 5.00306845e-01
-1.73700318e-01 5.12227975e-03 -8.34994376e-01 -9.13097620e-01
1.98288262e-01 -7.61882246e-01 3.81313175e-01 -7.17851758e-01
-1.43039834e+00 7.54037082e-01 -2.60860324e-01 -1.46152198e+00
1.12361923e-01 -2.57215649e-01 -7.37891674e-01 9.90870714e-01
-1.52010834e+00 -1.01582003e+00 -5.49068451e-01 5.60365021e-01
6.08697116e-01 1.62858829e-01 3.64590496e-01 7.63378620e-01
-3.65120441e-01 4.19427395e-01 3.01396489e-01 1.58528630e-02
8.12698245e-01 -8.62388194e-01 -5.57299890e-03 9.22457039e-01
-5.34979284e-01 3.45445424e-01 6.07510567e-01 -1.04509234e+00
-1.28518093e+00 -1.24015105e+00 2.71993041e-01 8.18244293e-02
5.08063793e-01 2.41520032e-01 -1.13857710e+00 4.15442467e-01
1.99997500e-01 6.58106446e-01 2.70421088e-01 -6.55898333e-01
2.26040825e-01 -2.12093741e-01 -1.42085207e+00 4.51642215e-01
7.23715246e-01 -1.25687003e-01 -3.77360880e-01 3.38686019e-01
6.55087769e-01 -8.44636381e-01 -7.59065568e-01 7.44562626e-01
3.68947476e-01 -7.87036538e-01 1.09150982e+00 -8.48963335e-02
7.78316975e-01 -2.56022573e-01 1.85934991e-01 -1.43839788e+00
-6.20864093e-01 -1.82967678e-01 2.57025063e-01 5.73696733e-01
2.67095417e-01 -4.58361298e-01 9.99282360e-01 5.31288385e-01
-1.05974877e+00 -8.81637871e-01 -1.29463661e+00 -4.92593527e-01
1.46381736e-01 -4.85389560e-01 2.01825276e-01 1.06053889e+00
-4.13686097e-01 5.22221252e-02 -3.43079358e-01 4.37497020e-01
8.46769392e-01 -2.13225842e-01 6.76364675e-02 -8.08662593e-01
-2.47535393e-01 -3.60155612e-01 -1.29352376e-01 -7.98245966e-01
-4.40445811e-01 -1.04380882e+00 3.55362862e-01 -1.65522075e+00
3.19159716e-01 -4.78440076e-01 -5.65242469e-01 3.18430394e-01
-2.79236138e-01 4.79606479e-01 -7.17993826e-02 3.74497592e-01
-6.01439998e-02 7.38957882e-01 2.03842854e+00 -2.24578947e-01
-2.05291659e-01 1.21682100e-01 -3.67312044e-01 9.85508323e-01
5.19436896e-01 -8.70863020e-01 -1.76722884e-01 -6.12734258e-01
5.90753518e-02 6.05861664e-01 5.08349299e-01 -1.18874478e+00
3.89716983e-01 3.65092009e-02 8.39141130e-01 -7.75145173e-01
2.49893695e-01 -1.13363338e+00 2.93836147e-01 9.10351574e-01
-9.56043899e-02 3.75152715e-02 2.03131735e-01 3.86877537e-01
-3.10163677e-01 -1.59877703e-01 1.44812202e+00 -2.06518933e-01
-6.47578761e-03 4.81914639e-01 -2.41100878e-01 -4.16134745e-01
7.46756434e-01 -8.51079896e-02 -4.42666113e-02 -1.06298402e-02
-8.95283043e-01 -5.27363867e-02 1.10769019e-01 -2.72673797e-02
1.06508064e+00 -1.49523342e+00 -7.07110047e-01 8.42198431e-02
-2.32536286e-01 4.66208577e-01 9.77109611e-01 1.35560715e+00
-8.72233927e-01 2.14332044e-02 -8.30758810e-02 -5.92833042e-01
-8.72743309e-01 2.75235027e-01 6.33226931e-01 -7.29468882e-01
-9.79843915e-01 8.27815175e-01 3.27715576e-01 -1.20425917e-01
-4.65135872e-02 -6.58448398e-01 2.66281277e-01 -5.35971999e-01
4.74342436e-01 3.48502278e-01 6.51700199e-01 -3.81445318e-01
-4.86428380e-01 3.83079708e-01 -3.46700877e-01 7.94253647e-02
1.82732153e+00 9.46261361e-02 -9.92215276e-02 -1.16612658e-01
1.03186107e+00 -1.24093510e-01 -1.13326013e+00 -3.97282660e-01
-3.42980087e-01 -3.42569381e-01 5.98321617e-01 -1.03076839e+00
-1.54857671e+00 6.97832167e-01 1.11323392e+00 -2.64678627e-01
1.38861060e+00 -2.14672253e-01 1.15596747e+00 -1.70496747e-01
6.36816025e-02 -8.81303906e-01 3.99159878e-01 1.17178179e-01
1.10360479e+00 -1.17227268e+00 4.32450444e-01 -5.67728460e-01
-4.26541984e-01 1.26412392e+00 6.29241407e-01 -3.05489480e-01
6.45036578e-01 4.58714545e-01 1.29522994e-01 -4.22284395e-01
6.02282351e-03 6.07631326e-01 2.41089478e-01 3.47481996e-01
5.91855943e-01 -1.26350224e-01 -4.28482562e-01 7.45566010e-01
2.96048075e-01 2.12122142e-01 6.04169786e-01 1.09293103e+00
-2.20384076e-01 -7.84699738e-01 -7.11286962e-01 4.13805991e-01
-6.54258192e-01 -2.43263930e-01 2.85757989e-01 6.60242558e-01
-5.00040166e-02 5.06626248e-01 -4.42187369e-01 5.58108091e-02
4.59679306e-01 -4.34042752e-01 5.60990870e-01 -4.88214672e-01
-8.58925879e-01 5.45850933e-01 -2.15853408e-01 -6.03184104e-01
-3.79759014e-01 -4.74658161e-01 -1.67422104e+00 -1.89685319e-02
-4.12446022e-01 -3.69244576e-01 6.00515306e-01 8.45434070e-01
-6.94974384e-04 1.21130776e+00 5.41401148e-01 -8.93464386e-01
-5.94639182e-01 -1.07285798e+00 -3.97219479e-01 2.99027830e-01
3.44490439e-01 -5.44137716e-01 -1.34836242e-01 -1.69761285e-01] | [13.497515678405762, -2.5445101261138916] |
e1e0953d-49d8-46ca-bfa3-555b3aabdc37 | continuous-ppg-based-blood-pressure | 2011.02231 | null | https://arxiv.org/abs/2011.02231v2 | https://arxiv.org/pdf/2011.02231v2.pdf | Continuous PPG-Based Blood Pressure Monitoring Using Multi-Linear Regression | In this work, we present the Senbiosys blood pressure monitoring algorithm (SB-BPM) that solely requires a photoplethysmography (PPG) signal. The technology is based on pulse wave analysis (PWA) of PPG signals retrieved from different body locations to continuously estimate the systolic blood pressure (SBP) and the diastolic blood pressure (DBP). | ['Antonino Caizzone', 'Assim Boukhayma', 'Serj Haddad'] | 2020-11-04 | null | null | null | null | ['photoplethysmography-ppg'] | ['medical'] | [ 1.11057006e-01 -9.31058824e-02 1.97709143e-01 -3.68828654e-01
1.40973762e-01 -3.52337629e-01 -1.51897579e-01 -4.34767485e-01
-1.50967479e-01 1.14682412e+00 2.18693435e-01 -2.85409153e-01
2.35184714e-01 -6.86529934e-01 1.84334561e-01 -7.53898799e-01
-3.86565059e-01 7.77778402e-02 1.42425656e-01 2.23064974e-01
9.40008909e-02 4.85367835e-01 -6.88727140e-01 -3.60733271e-01
1.10453796e+00 1.38002121e+00 -8.19491327e-01 1.13626826e+00
1.97625265e-01 1.85989827e-01 -7.07842469e-01 -8.20253640e-02
5.39006770e-01 -1.00222874e+00 2.64500882e-02 -4.09656703e-01
8.95766437e-01 -5.02285719e-01 -2.48334840e-01 7.31202245e-01
1.08005297e+00 -2.48614416e-01 1.46775618e-02 -7.67777562e-01
-1.95745826e-01 3.54684204e-01 -3.75640243e-01 5.16618967e-01
1.40340433e-01 3.71634275e-01 -1.29329279e-01 -6.82303250e-01
3.02251545e-03 6.14439726e-01 1.13517904e+00 5.76111853e-01
-1.70220304e+00 -2.97685087e-01 -8.63684416e-01 1.38539951e-02
-1.24747503e+00 -6.36159062e-01 9.37566876e-01 -3.32844257e-01
4.47836727e-01 8.96777391e-01 1.18441606e+00 4.02949244e-01
6.81839466e-01 -3.69432449e-01 1.82235479e+00 -5.40260196e-01
2.54622757e-01 4.05703664e-01 6.49562716e-01 5.35504997e-01
5.89023888e-01 2.44332388e-01 -4.88033146e-01 -6.27595186e-01
9.75141823e-01 -2.43431792e-01 -7.91895390e-01 -2.31428519e-01
-7.07311749e-01 1.79649472e-01 -4.69792753e-01 3.02767903e-01
-8.47623944e-01 -7.62536377e-02 2.12861419e-01 5.23087025e-01
3.26876879e-01 5.25399387e-01 -1.92434743e-01 -6.65111840e-01
-1.23495579e+00 1.06144555e-01 1.38012922e+00 2.21636772e-01
1.87236875e-01 2.72817969e-01 -5.78650296e-01 5.11462033e-01
6.45392001e-01 9.95357513e-01 2.65873700e-01 -1.18055928e+00
-1.26720384e-01 2.22558096e-01 5.81117034e-01 -1.25142062e+00
-3.20803553e-01 8.52639079e-02 -6.75214291e-01 1.09430574e-01
7.81803727e-01 -6.34284854e-01 -6.39075756e-01 1.07232237e+00
4.04457539e-01 4.52353120e-01 -2.45521650e-01 1.22257757e+00
1.06046975e+00 -1.52837727e-02 3.67783695e-01 -7.57554829e-01
1.49849212e+00 -3.09809327e-01 -9.23220396e-01 -3.95043828e-02
-3.55323344e-01 -2.87715524e-01 2.14480922e-01 2.31596142e-01
-1.14783275e+00 -7.80283391e-01 -8.28097701e-01 3.60020190e-01
-8.28383937e-02 -3.02083343e-01 1.27896279e-01 1.39507473e+00
-9.82183874e-01 1.13705158e+00 -6.79198205e-01 -9.61916298e-02
-8.04638714e-02 -1.68761775e-01 -7.29107112e-02 5.14495075e-01
-1.55744994e+00 1.19497907e+00 -2.92114645e-01 5.30811429e-01
3.64950240e-01 -1.47112095e+00 -7.28518486e-01 -1.14265479e-01
-5.20528913e-01 -6.50384426e-01 3.29006582e-01 1.90801814e-01
-2.45938540e+00 1.00174272e+00 -3.72753084e-01 -3.23685557e-01
3.91000867e-01 -3.15137804e-01 -8.13303471e-01 6.62504792e-01
-5.96009731e-01 -4.77288693e-01 8.96420002e-01 -3.64849389e-01
3.84163141e-01 -7.12889493e-01 -1.08286059e+00 -2.77811706e-01
3.42282057e-01 1.31844297e-01 2.85677463e-01 1.21157974e-01
5.33258617e-01 -3.81031811e-01 -2.17517748e-01 -3.76533046e-02
-1.97145358e-01 4.94372725e-01 1.23234652e-01 -1.21841955e+00
1.46928465e+00 -1.94856501e+00 -1.82260260e-01 6.15092099e-01
3.79319102e-01 7.51518130e-01 6.66228771e-01 2.56423891e-01
5.03125675e-02 3.03720050e-02 -1.66820198e-01 1.24139488e-01
7.59502053e-02 -9.67884995e-03 -4.86628026e-01 9.10399199e-01
-4.48202938e-01 1.06620693e+00 -5.73780835e-01 -6.53695345e-01
8.87319684e-01 3.57249916e-01 3.19675058e-01 3.03539902e-01
6.60259068e-01 8.61899018e-01 1.10579394e-01 6.46061003e-01
1.27362359e+00 3.98149371e-01 3.74166161e-01 -6.66649342e-01
-6.17646158e-01 9.61360857e-02 -8.90382290e-01 1.10501480e+00
2.51403511e-01 4.39588755e-01 2.63637334e-01 -5.50254643e-01
1.76298928e+00 8.75842273e-01 6.15071118e-01 -9.69054699e-01
1.38478234e-01 1.89136580e-01 2.56989360e-01 -1.10778046e+00
-3.91370416e-01 -7.63314426e-01 3.83374870e-01 4.02180582e-01
-1.55381501e-01 -1.25097781e-01 -4.50285673e-02 -4.94884044e-01
7.60036469e-01 1.80486724e-01 6.37226105e-01 -9.46478605e-01
9.80041504e-01 -2.88097620e-01 4.39067394e-01 8.44239891e-01
-1.25377131e+00 3.82959157e-01 5.68484187e-01 -1.05865538e+00
-8.47968459e-01 -1.25315452e+00 -1.01533520e+00 -1.40058771e-01
1.31343797e-01 -1.07566707e-01 -2.69524217e-01 5.12776196e-01
7.78301954e-01 3.20707351e-01 -3.38000387e-01 1.32842824e-01
-5.36779821e-01 -7.47450590e-01 5.20929873e-01 2.21528426e-01
5.11915386e-01 -7.91559100e-01 -1.11532533e+00 3.88784409e-01
1.05059899e-01 -7.69222617e-01 7.90770277e-02 -5.24164498e-01
-1.06320798e+00 -9.54453528e-01 -1.03232706e+00 9.66193601e-02
1.49713740e-01 -7.10472345e-01 1.07238817e+00 -6.20667815e-01
-8.17757308e-01 4.88346189e-01 2.20336258e-01 -4.95193720e-01
-1.96049109e-01 -7.53787398e-01 4.46689367e-01 3.63798857e-01
1.18509924e+00 -1.49793780e+00 -1.13973236e+00 1.66773945e-01
5.67202389e-01 -3.47697228e-01 -4.80808225e-03 -3.63182157e-01
6.89051449e-01 -8.66978109e-01 4.02944952e-01 -6.88113391e-01
9.24933791e-01 -1.12500243e-01 -7.78985441e-01 -5.38443625e-02
-7.80153871e-01 -8.50796878e-01 4.44506288e-01 -1.65855527e-01
-6.19742513e-01 -6.05240345e-01 5.53108342e-02 -2.39884272e-01
-2.96330154e-01 6.38910607e-02 3.82005811e-01 -3.64561349e-01
1.01956058e+00 7.85484195e-01 6.21551931e-01 -5.23695171e-01
-2.75292713e-02 6.73552990e-01 1.13642478e+00 -4.50667083e-01
2.69652426e-01 1.73715070e-01 6.42696202e-01 -1.02914476e+00
-2.36649632e-01 -3.14184427e-01 -9.48284566e-01 -7.14627802e-01
6.11590624e-01 -4.17441696e-01 -1.31496656e+00 5.69251716e-01
-4.82728720e-01 8.46086293e-02 -3.65086883e-01 9.38451588e-01
-3.29289824e-01 7.77608335e-01 -7.84080923e-01 -1.26464498e+00
-1.08303416e+00 1.51563764e-01 7.84961432e-02 6.16364658e-01
-6.28467679e-01 -1.01055264e+00 8.27741325e-01 -6.79024905e-02
1.38331509e+00 1.27048707e+00 2.22967252e-01 -9.40957852e-03
2.19780162e-01 -8.21465015e-01 4.11160812e-02 5.42688131e-01
2.83776402e-01 -2.40517497e-01 -8.86388302e-01 2.33749315e-01
6.41899168e-01 4.79396254e-01 1.26869425e-01 1.05438554e+00
5.76902390e-01 -6.21625260e-02 1.43546257e-02 7.08981216e-01
1.54322314e+00 3.39577556e-01 1.38888586e+00 -1.47002071e-01
2.51902878e-01 2.18571916e-01 -1.54732600e-01 9.42977011e-01
-1.45529777e-01 2.53112853e-01 -4.45658475e-01 -1.02250762e-02
1.69112105e-02 1.81236699e-01 9.67655554e-02 3.56114775e-01
-6.23480916e-01 8.53780925e-01 -5.57909548e-01 -1.15392298e-01
-1.29313505e+00 -9.10490334e-01 -4.97029215e-01 2.32024384e+00
1.28989518e+00 -3.16701174e-01 1.34006917e-01 -1.75380588e-01
6.94304526e-01 -5.10420697e-03 -5.10407269e-01 -7.77252078e-01
2.81984538e-01 9.37440693e-01 4.93766040e-01 4.96560842e-01
-8.05068970e-01 -7.78095350e-02 7.72377348e+00 -9.59715784e-01
-1.35178018e+00 -3.18181276e-01 2.28601381e-01 1.37905955e-01
4.11966175e-01 5.80048487e-02 -5.52875340e-01 1.17112505e+00
1.42060244e+00 -8.99069011e-01 9.53682736e-02 7.76118934e-01
4.60283786e-01 -5.06705105e-01 -8.97835374e-01 1.03157330e+00
3.63489315e-02 -1.08163297e+00 -1.03745127e+00 -1.99021339e-01
-2.70472348e-01 -3.18517864e-01 -6.02525711e-01 7.44201019e-02
-8.39333951e-01 -5.22839963e-01 -5.53273320e-01 1.49187577e+00
9.66288030e-01 2.26422817e-01 4.52278197e-01 -2.55624801e-01
-7.06960440e-01 6.25717938e-01 -7.38014042e-01 -4.30728018e-01
4.97320771e-01 1.27267492e+00 -4.38632369e-01 6.92956150e-02
1.44266233e-01 4.20654148e-01 -1.17992379e-01 1.69731843e+00
-3.26018244e-01 8.15901279e-01 -5.35446346e-01 2.33366609e-01
-6.10722482e-01 -5.66700220e-01 7.47497737e-01 7.48149693e-01
2.02034086e-01 7.43496656e-01 -9.14542452e-02 1.76122665e+00
5.32201111e-01 -4.53356542e-02 -3.44730496e-01 3.69381011e-01
5.47486067e-01 1.19278574e+00 2.77366694e-02 -5.48595905e-01
-1.88201234e-01 1.55390844e-01 -9.19020712e-01 4.19284433e-01
-3.95200342e-01 -8.32617104e-01 6.15258276e-01 4.88314271e-01
-4.49827135e-01 -5.39096668e-02 -8.23512316e-01 -1.03218520e+00
2.04249814e-01 1.27386814e-02 3.31028640e-01 -5.61091661e-01
-1.04402208e+00 4.64602560e-02 -7.08833709e-02 -1.06336021e+00
-1.29706919e-01 -2.59454876e-01 -1.09213221e+00 2.31804109e+00
-1.56929243e+00 -1.13445155e-01 -5.53317249e-01 2.53191173e-01
-6.72565997e-01 3.54655117e-01 1.18084157e+00 3.40148032e-01
-4.70402151e-01 2.09775805e-01 -6.03023946e-01 1.35541171e-01
6.97305977e-01 -1.52173007e+00 1.71178296e-01 7.22890258e-01
-1.26688039e+00 7.68429756e-01 9.95152652e-01 -6.48145795e-01
-1.61934423e+00 -4.13860202e-01 1.38380086e+00 -4.55660731e-01
2.90704757e-01 2.42022470e-01 -6.53540194e-01 2.03280449e-01
3.44832577e-02 6.03770137e-01 1.12473154e+00 -3.87997121e-01
7.71062225e-02 -5.55075109e-01 -1.73743236e+00 1.29236251e-01
-9.28193256e-02 -1.99045628e-01 -9.19153690e-01 -2.41000205e-01
1.35080084e-01 -6.92369580e-01 -1.97880578e+00 6.09406233e-01
1.13735485e+00 -7.14834392e-01 9.44652736e-01 -5.97066820e-01
4.82471436e-02 -2.08589360e-01 4.36645210e-01 -9.64602113e-01
-5.27304173e-01 -1.07867289e+00 -5.92203319e-01 1.02430844e+00
-1.88760668e-01 -1.32383573e+00 5.42294979e-01 2.14480186e+00
2.77059805e-02 -2.30782360e-01 -1.09880781e+00 -5.00295222e-01
-3.31604004e-01 4.60155129e-01 3.22044581e-01 7.10059702e-01
1.30771863e+00 -2.54116245e-02 -9.50892806e-01 6.73488155e-02
1.20034385e+00 5.63553274e-01 3.11552823e-01 -1.37242913e+00
-1.13964468e-01 4.01769578e-02 -6.00357473e-01 -2.88139999e-01
-8.07935536e-01 4.14963439e-03 -1.15729004e-01 -1.20888722e+00
-4.19035465e-01 6.46607950e-02 -7.36114323e-01 1.93940818e-01
-1.26594424e-01 3.48810405e-01 -1.50533110e-01 2.00445894e-02
2.23367095e-01 -2.83008784e-01 1.02003765e+00 6.51297390e-01
-9.06664789e-01 2.97127694e-01 -3.15203488e-01 -1.12588480e-02
8.23258758e-01 -2.96335310e-01 4.57885414e-02 7.33979881e-01
-1.99535564e-01 6.40341341e-01 2.69424677e-01 -1.15010691e+00
1.63208634e-01 -2.58952510e-02 8.97808254e-01 -5.17244339e-01
1.99244604e-01 -5.83050609e-01 5.02617002e-01 1.11829221e+00
1.43891528e-01 -5.26254237e-01 2.17221782e-01 -1.40360281e-01
-1.71922088e-01 7.16986787e-03 9.75478649e-01 -6.53785586e-01
-2.83102781e-01 1.61687747e-01 -4.22701478e-01 -2.13462606e-01
7.12063849e-01 -5.57448566e-01 -6.89169288e-01 5.29403687e-02
-1.11411440e+00 1.70855373e-01 2.49725953e-02 -3.57120156e-01
6.88345373e-01 -1.38884032e+00 -8.86411667e-01 8.55373800e-01
-1.63153440e-01 -7.24886239e-01 6.75739408e-01 1.68569267e+00
-8.88437390e-01 5.39817214e-01 -6.15563214e-01 -4.57392782e-01
-1.15871036e+00 1.82897598e-01 1.18598652e+00 3.44245553e-01
-1.26847255e+00 4.82311249e-01 -8.88644397e-01 3.91764075e-01
1.87471136e-01 2.45020673e-01 -4.77698475e-01 -3.50335270e-01
1.26505256e+00 6.09808624e-01 -3.89718413e-01 1.32194951e-01
-4.82181758e-01 5.48723698e-01 7.08659291e-01 2.39138156e-01
1.00748980e+00 -4.31301713e-01 -6.43804848e-01 5.31434059e-01
4.80323046e-01 -1.92958847e-01 -5.83046854e-01 -3.18017393e-01
-2.47271508e-01 -9.23984349e-01 -1.43506676e-01 -1.15035403e+00
-6.99525893e-01 7.01227963e-01 9.43346918e-01 6.73317730e-01
1.06020224e+00 -5.01349270e-01 7.06841707e-01 -2.35966280e-01
1.57461569e-01 -1.00683737e+00 -7.96516538e-01 -3.88339639e-01
7.29285300e-01 -2.84606427e-01 3.76129746e-01 -4.60183889e-01
-2.70078123e-01 1.47195208e+00 3.54162782e-01 -3.86671007e-01
1.16204262e+00 2.89617300e-01 6.26917481e-01 -2.08640695e-01
-2.92120576e-01 2.17789292e-01 2.30873004e-01 6.59823656e-01
4.82276827e-01 3.32441509e-01 -1.27278292e+00 5.51623821e-01
1.10955477e-01 8.37394714e-01 5.31800210e-01 6.87684536e-01
-8.85957301e-01 -6.78643346e-01 -3.23548794e-01 5.67537129e-01
-5.97415507e-01 1.89029856e-03 -8.95422697e-02 2.34662637e-01
-3.22281271e-01 7.57646978e-01 -6.22401107e-03 -4.71903123e-02
6.17895901e-01 9.93781626e-01 7.38842607e-01 -1.96321607e-01
-2.74460465e-01 -1.27847353e-02 -6.99665695e-02 -7.04709351e-01
-7.29971647e-01 -2.07494482e-01 -5.21482944e-01 -3.15660208e-01
3.54101628e-01 1.09631397e-01 7.99851477e-01 4.48105425e-01
4.56543267e-01 -6.77255094e-02 7.46600747e-01 -4.07938629e-01
-8.03160846e-01 -1.25348401e+00 -1.59109414e+00 3.18923086e-01
3.66288245e-01 -2.00838938e-01 -6.50417149e-01 5.87545000e-02] | [14.020322799682617, 2.94793701171875] |
f7ca995a-cc3d-4faf-8f9b-60a221bb78e8 | sequential-optimization-for-efficient-high | 1511.04511 | null | http://arxiv.org/abs/1511.04511v3 | http://arxiv.org/pdf/1511.04511v3.pdf | Sequential Optimization for Efficient High-Quality Object Proposal Generation | We are motivated by the need for a generic object proposal generation
algorithm which achieves good balance between object detection recall, proposal
localization quality and computational efficiency. We propose a novel object
proposal algorithm, BING++, which inherits the virtue of good computational
efficiency of BING but significantly improves its proposal localization
quality. At high level we formulate the problem of object proposal generation
from a novel probabilistic perspective, based on which our BING++ manages to
improve the localization quality by employing edges and segments to estimate
object boundaries and update the proposals sequentially. We propose learning
the parameters efficiently by searching for approximate solutions in a
quantized parameter space for complexity reduction. We demonstrate the
generalization of BING++ with the same fixed parameters across different object
classes and datasets. Empirically our BING++ can run at half speed of BING on
CPU, but significantly improve the localization quality by 18.5% and 16.7% on
both VOC2007 and Microhsoft COCO datasets, respectively. Compared with other
state-of-the-art approaches, BING++ can achieve comparable performance, but run
significantly faster. | ['Philip H. S. Torr', 'Ming-Ming Cheng', 'Ziming Zhang', 'Venkatesh Saligrama', 'Yun Liu', 'Yanjun Zhu', 'Xi Chen'] | 2015-11-14 | null | null | null | null | ['object-proposal-generation'] | ['computer-vision'] | [-1.48407802e-01 1.01080351e-01 -2.34459400e-01 -1.58913895e-01
-1.25612199e+00 -4.06049758e-01 3.90804827e-01 1.88160628e-01
-4.59160358e-01 3.34645182e-01 -2.47934669e-01 1.20852731e-01
-1.96331199e-02 -7.80652940e-01 -7.28856623e-01 -4.72721636e-01
1.80920474e-02 6.93495035e-01 1.15679312e+00 3.34246904e-01
4.81813818e-01 5.93622327e-01 -1.56648040e+00 -4.96085770e-02
7.13299870e-01 1.11735213e+00 4.29244310e-01 5.99432766e-01
-6.22214079e-02 6.58879383e-03 -4.27659780e-01 -3.05808842e-01
2.63729811e-01 1.17406398e-01 -6.39634371e-01 3.21553737e-01
8.04165781e-01 -1.59719780e-01 -1.68805629e-01 9.52123225e-01
6.98820412e-01 -8.78158584e-02 6.42510355e-01 -1.33999813e+00
-4.28150535e-01 7.12375879e-01 -6.68241322e-01 -1.04959710e-02
-2.84280721e-02 1.62654996e-01 9.97493625e-01 -1.33955967e+00
5.92057765e-01 1.28296065e+00 8.73957098e-01 4.20856267e-01
-1.40020859e+00 -5.75330794e-01 3.12718302e-01 1.91685006e-01
-1.98393130e+00 -2.69478321e-01 3.61749053e-01 -2.74705261e-01
7.01036215e-01 1.94654405e-01 5.64296782e-01 5.44789135e-01
-2.39617229e-02 1.06785691e+00 7.85170078e-01 -3.04975033e-01
3.15165460e-01 8.92327875e-02 1.76207438e-01 8.27335358e-01
5.45113683e-01 -7.28855748e-03 -5.54312825e-01 -2.82380611e-01
7.72443414e-01 -1.10531144e-01 -3.28691304e-01 -6.66193604e-01
-1.40945017e+00 7.90577412e-01 6.88943326e-01 -2.67417021e-02
-1.70683742e-01 6.51622474e-01 1.63020939e-01 -3.74486953e-01
1.02478908e-02 3.49814117e-01 -3.89302075e-01 2.06682369e-01
-1.11822355e+00 6.04050875e-01 6.47178888e-01 1.44261777e+00
7.69889891e-01 -2.45592564e-01 -6.05356157e-01 6.67364240e-01
7.14087784e-01 6.44152164e-01 2.91587442e-01 -1.03579044e+00
2.12974429e-01 4.65858787e-01 4.84748095e-01 -8.36351633e-01
-3.38957727e-01 -5.90241134e-01 -2.70692229e-01 -1.60036474e-01
3.20542097e-01 3.38931680e-01 -9.98907447e-01 1.44448638e+00
6.53684616e-01 2.78763354e-01 -3.76279473e-01 9.08523917e-01
7.78554142e-01 8.27720404e-01 3.48183932e-03 -3.78795201e-04
1.50992060e+00 -1.38061619e+00 -4.05354738e-01 -7.28012994e-02
6.17716193e-01 -9.40316975e-01 8.56602907e-01 4.04054761e-01
-1.06756878e+00 -7.94803441e-01 -1.08334374e+00 5.08605689e-02
-3.53657864e-02 5.58890939e-01 6.68353558e-01 7.14397490e-01
-1.03132474e+00 3.96141797e-01 -9.46872294e-01 -4.39192802e-01
6.28938854e-01 4.91273016e-01 -2.49675266e-03 -1.74422130e-01
-2.95634687e-01 6.99027896e-01 8.78105581e-01 -2.09565565e-01
-8.17214131e-01 -8.12295854e-01 -6.17963195e-01 2.61428244e-02
6.10972226e-01 -6.51322186e-01 1.46151721e+00 -2.18694121e-01
-1.41475582e+00 5.81665218e-01 -5.79212792e-02 -5.55388272e-01
4.63438511e-01 -2.65082657e-01 -1.27041742e-01 4.50803638e-02
2.67907381e-01 1.30068874e+00 7.32111990e-01 -1.34501958e+00
-1.05464590e+00 -2.67792702e-01 -3.85678470e-01 -1.95604116e-02
6.52377158e-02 -1.48621425e-01 -1.16219759e+00 -5.99035382e-01
6.28682077e-01 -9.80299234e-01 -4.61575806e-01 5.35294414e-01
-3.39849800e-01 -6.07600391e-01 1.00446820e+00 -2.50886790e-02
1.06168902e+00 -2.06098437e+00 -2.64458746e-01 2.24157423e-01
1.94582522e-01 3.13713461e-01 -2.69387782e-01 1.05262987e-01
5.80083966e-01 -3.70944738e-02 -1.58026874e-01 -5.40646911e-01
2.34024927e-01 1.09085202e-01 -2.85878748e-01 6.53102636e-01
1.75360754e-01 1.07887471e+00 -8.65397930e-01 -7.94306219e-01
2.28589490e-01 3.78156930e-01 -8.80861759e-01 6.57776371e-02
-2.58140981e-01 -1.47197738e-01 -5.35984099e-01 8.85361493e-01
1.00881743e+00 -3.88091892e-01 7.87468180e-02 -4.57100213e-01
-1.81500420e-01 6.83413669e-02 -1.68128061e+00 1.67159307e+00
-1.61230907e-01 4.44793284e-01 -7.29352329e-03 -7.48493493e-01
1.03678024e+00 -7.58263841e-02 2.06529200e-01 -3.80571574e-01
4.69361804e-02 5.22841454e-01 -1.90481111e-01 -5.96635342e-02
9.53218579e-01 2.32621729e-01 -8.05663615e-02 2.08174467e-01
2.56908506e-01 -3.86586517e-01 8.41584504e-02 1.74847841e-01
8.83782387e-01 3.07464272e-01 4.02807087e-01 -4.64462548e-01
2.10859805e-01 2.88793854e-02 5.08801758e-01 1.25526369e+00
-4.43352342e-01 7.13167727e-01 1.50431925e-02 -1.73929617e-01
-9.59398746e-01 -1.42313230e+00 -5.07250667e-01 8.47621977e-01
6.13905966e-01 -5.11469364e-01 -6.47818208e-01 -7.10818529e-01
1.48399055e-01 4.45773721e-01 -3.70490044e-01 1.07527591e-01
-6.01267457e-01 -8.30780447e-01 3.43897700e-01 6.45337701e-01
5.97335875e-01 -8.88325989e-01 -7.52310395e-01 4.58157241e-01
-3.17145661e-02 -1.19267058e+00 -5.22724807e-01 2.41196193e-02
-9.19635355e-01 -7.89503515e-01 -6.63184404e-01 -6.80460811e-01
5.97257733e-01 4.28194255e-01 1.07564998e+00 1.54291600e-01
-5.72821140e-01 3.02534580e-01 -3.41699719e-01 -5.23761034e-01
-7.82518238e-02 4.06986058e-01 1.87857337e-02 -1.38427362e-01
-1.18824393e-02 -1.54728010e-01 -8.05171371e-01 6.61089122e-01
-7.01148808e-01 -7.06939623e-02 7.18856156e-01 6.42758071e-01
1.01635182e+00 -1.73224017e-01 4.96532232e-01 -5.06991625e-01
-1.90377370e-01 -2.13132501e-01 -1.02202761e+00 1.84370518e-01
-5.94082952e-01 4.42399718e-02 5.28879054e-02 -5.03135502e-01
-6.93011165e-01 3.92005682e-01 -2.36867890e-01 -2.43691862e-01
5.10718822e-02 -2.19787121e-01 -1.70424476e-01 -4.30994689e-01
6.31307304e-01 1.85260296e-01 -4.05351430e-01 -5.25450468e-01
6.75308406e-01 5.51744699e-01 7.14686453e-01 -6.29114747e-01
8.74029756e-01 7.37650514e-01 -1.43863156e-01 -6.07804120e-01
-8.62614453e-01 -7.88802087e-01 -4.68891621e-01 4.32098797e-03
5.11303365e-01 -8.41911018e-01 -8.57303202e-01 2.58305252e-01
-1.30618393e+00 -2.27008179e-01 -4.44172293e-01 5.60254931e-01
-7.03466654e-01 3.09226483e-01 -3.20525646e-01 -7.33387470e-01
-2.29471162e-01 -1.11284757e+00 1.42751575e+00 2.87922382e-01
1.57995418e-01 -3.73879045e-01 -1.27092093e-01 1.04413489e-02
2.97811508e-01 1.11142183e-02 2.67275065e-01 -3.58503819e-01
-1.25745821e+00 -2.36283630e-01 -7.34754205e-01 -1.10085554e-01
-2.64013082e-01 -1.34499282e-01 -9.36083853e-01 -4.89296049e-01
-3.42136025e-01 2.41554771e-02 1.03244436e+00 4.63108212e-01
1.28603971e+00 -2.65451342e-01 -8.91342640e-01 6.69525743e-01
1.75384426e+00 1.11532109e-02 5.13479948e-01 2.30870321e-01
4.98853624e-01 2.16265723e-01 1.03247356e+00 4.65170056e-01
4.57225412e-01 1.22313380e+00 5.94827533e-01 1.64941072e-01
-3.95404041e-01 -3.01980734e-01 3.49635594e-02 5.58627963e-01
2.70579278e-01 -2.48465806e-01 -7.78510869e-01 9.07614589e-01
-2.12089562e+00 -5.20179689e-01 -2.93959022e-01 2.04817367e+00
6.20827019e-01 1.69773459e-01 3.83100599e-01 3.21568386e-03
7.41665125e-01 -1.63000077e-01 -2.63274968e-01 2.01615930e-01
1.69138357e-01 1.27221346e-01 8.76291871e-01 3.54380697e-01
-1.19313669e+00 1.09186256e+00 6.61489201e+00 1.33202219e+00
-9.74686801e-01 2.62362361e-01 2.74127722e-01 7.87611902e-02
6.40590191e-02 1.70084551e-01 -1.70396698e+00 4.49931979e-01
6.42427146e-01 -7.95500819e-03 -1.16757229e-01 1.26570737e+00
-1.02680109e-01 -2.65610695e-01 -1.08019161e+00 1.18771839e+00
1.41037151e-01 -1.54781139e+00 7.99428821e-02 -7.19114617e-02
7.91197240e-01 2.41194859e-01 -1.07458662e-02 2.84451783e-01
1.12679387e-02 -6.83826625e-01 1.26352739e+00 3.44145447e-01
3.35884571e-01 -6.16180897e-01 7.16163874e-01 3.31710964e-01
-1.51356590e+00 2.72243544e-02 -6.69261694e-01 3.70493442e-01
4.29977715e-01 6.00283206e-01 -1.19020104e+00 4.24599200e-01
7.82727480e-01 2.54100144e-01 -8.23371172e-01 1.90719974e+00
-1.15822703e-01 5.88134706e-01 -7.24741101e-01 -2.36450776e-01
2.42866755e-01 1.83532953e-01 5.13434052e-01 1.45890141e+00
5.33934891e-01 5.00382669e-02 5.91023624e-01 1.08489382e+00
-2.09423453e-02 1.54948190e-01 -1.23536944e-01 3.78700912e-01
8.99611473e-01 1.27303195e+00 -1.20539033e+00 -5.08063436e-01
-2.54550930e-02 8.34353089e-01 3.10760438e-01 -4.60844375e-02
-1.24869788e+00 -5.09227335e-01 1.88243583e-01 3.57280880e-01
7.88799226e-01 -3.21177185e-01 -1.93373188e-01 -8.03300261e-01
7.04804659e-02 -2.94561148e-01 1.16748333e-01 -6.88028216e-01
-1.02203989e+00 5.20595253e-01 1.25610769e-01 -1.21473312e+00
2.18938932e-01 -6.54035985e-01 -3.03199768e-01 3.35856080e-01
-1.41844976e+00 -1.36561692e+00 -4.79361176e-01 1.66498825e-01
5.29913127e-01 3.32411647e-01 4.59530085e-01 3.57228607e-01
-3.01664293e-01 6.96170986e-01 6.48262352e-02 -2.11207852e-01
5.24820924e-01 -1.19237518e+00 6.80767000e-01 7.60565996e-01
3.82644832e-01 3.98942977e-01 5.00782251e-01 -3.75081778e-01
-1.23714435e+00 -1.43906593e+00 7.76430547e-01 -5.68490565e-01
4.13233519e-01 -4.19548601e-01 -6.99244499e-01 4.81834978e-01
-3.78601789e-01 4.46261525e-01 5.11271954e-02 -2.16599748e-01
-3.67858469e-01 -1.06324278e-01 -1.23862791e+00 6.12317562e-01
1.22769690e+00 -1.28269330e-01 -4.76203382e-01 4.87446427e-01
9.70525265e-01 -5.13137341e-01 -7.48630226e-01 5.28530300e-01
4.32291269e-01 -9.09716487e-01 1.12441063e+00 8.07102323e-02
-4.20948565e-01 -9.08264816e-01 -4.13496703e-01 -7.09508419e-01
-3.63592654e-01 -5.84551990e-01 -2.73376673e-01 1.31367612e+00
2.53526062e-01 -6.10722005e-01 8.49512517e-01 1.50846720e-01
-2.53463924e-01 -8.77466857e-01 -1.03469050e+00 -1.16346920e+00
-4.43802774e-01 -8.43423605e-01 6.33472979e-01 2.47458458e-01
-5.72540700e-01 6.01444021e-03 -2.06435993e-02 4.34117407e-01
9.63470340e-01 3.44754308e-01 9.88529146e-01 -1.00226009e+00
-3.29401374e-01 -4.60234165e-01 -6.28981769e-01 -1.61906123e+00
-3.81751031e-01 -8.10043812e-01 4.02992427e-01 -1.42101145e+00
3.84811908e-01 -8.20412576e-01 -4.41667587e-02 4.23577011e-01
-2.30411857e-01 9.01541889e-01 3.88549507e-01 2.33094171e-01
-1.13080680e+00 5.83713114e-01 1.01098228e+00 -4.87390868e-02
-2.91907579e-01 1.40902221e-01 -4.90473241e-01 9.12020147e-01
3.80000621e-01 -7.01907277e-01 -3.27414982e-02 -2.69043565e-01
-2.28749901e-01 -3.16841036e-01 5.93407512e-01 -1.32056367e+00
2.11751372e-01 -3.99155077e-03 1.77453309e-01 -1.16619492e+00
4.06973600e-01 -6.80746913e-01 1.76825430e-02 6.66539192e-01
9.73537862e-02 -2.41642699e-01 1.84629306e-01 7.61243820e-01
1.00428328e-01 -3.29669356e-01 1.11378455e+00 1.73764750e-01
-1.00960612e+00 5.15056908e-01 -6.70304894e-02 -8.63497928e-02
1.12580955e+00 -3.79170269e-01 -2.74482548e-01 2.16645271e-01
-6.85643852e-01 2.25017652e-01 5.06927192e-01 3.46868455e-01
6.77439749e-01 -1.51637590e+00 -5.38279712e-01 1.12513984e-02
4.68810558e-01 1.59120917e-01 3.97481397e-02 8.69322956e-01
-6.32202506e-01 5.35722256e-01 2.04568848e-01 -1.18777490e+00
-9.98580992e-01 5.98924696e-01 1.82624862e-01 -7.28020146e-02
-6.64201379e-01 1.04490411e+00 3.71284783e-01 -4.54096824e-01
3.93184602e-01 -5.46442807e-01 3.33225667e-01 -2.76201487e-01
4.60508138e-01 5.48458099e-01 -6.77681640e-02 -4.16441232e-01
-5.52543581e-01 8.79723847e-01 -9.28283036e-02 -5.17398119e-02
9.08097506e-01 -8.80445018e-02 3.09909999e-01 -1.23336520e-02
9.49436903e-01 -2.27206275e-02 -1.35077119e+00 -3.15742970e-01
2.97139883e-01 -6.62613451e-01 -1.98592469e-02 -6.16442978e-01
-7.83255100e-01 5.08904159e-01 9.89864349e-01 -1.02534659e-01
6.81947172e-01 5.53515077e-01 7.69816279e-01 4.26607728e-01
8.49750996e-01 -8.75057518e-01 2.72386879e-01 2.12726027e-01
7.55026042e-01 -1.20944989e+00 2.16507375e-01 -8.41612935e-01
-1.37208626e-01 8.10968399e-01 7.10656047e-01 -4.34108824e-01
5.29205263e-01 2.87642837e-01 -3.59756678e-01 -1.45319045e-01
-3.99136484e-01 -3.59238416e-01 5.34531891e-01 6.24688327e-01
4.74278592e-02 1.64972126e-01 -3.65073115e-01 5.51344693e-01
3.96040920e-03 -1.19981721e-01 2.36751229e-01 6.65647268e-01
-9.10029948e-01 -9.56331253e-01 -6.86705589e-01 3.04620475e-01
-2.22281009e-01 1.86414391e-01 1.87701300e-01 1.02000439e+00
4.48316842e-01 5.47661722e-01 1.32252276e-01 2.22383942e-02
3.47514838e-01 -3.10019791e-01 5.91605067e-01 -7.48834610e-01
-1.91836342e-01 2.50903279e-01 -2.06454441e-01 -8.26967478e-01
-3.96302015e-01 -6.01328671e-01 -1.33963048e+00 8.95441771e-02
-8.55807960e-01 1.65133700e-01 8.00155997e-01 7.23319769e-01
4.83113706e-01 3.05477202e-01 3.21703285e-01 -1.21372855e+00
-7.10263252e-01 -8.34553659e-01 -5.58843136e-01 -8.94491225e-02
1.38462260e-02 -8.47864568e-01 -7.12341294e-02 -1.53707772e-01] | [8.757713317871094, -0.22126996517181396] |
6d4159b3-cfe4-4252-9db6-b960c4da43b1 | fusing-heterogeneous-factors-with-triaffine-1 | null | null | https://openreview.net/forum?id=OXXX_dfeH7v | https://openreview.net/pdf?id=OXXX_dfeH7v | Fusing Heterogeneous Factors with Triaffine Mechanism for Nested Named Entity Recognition | Nested entities are observed in many domains due to their compositionality, which cannot be easily recognized by the widely-used sequence labeling framework.
A natural solution is to treat the task as a span classification problem.
To learn better span representation and increase classification performance, it is crucial to effectively integrate heterogeneous factors including inside tokens, boundaries, labels, and related spans which could be contributing to nested entities recognition.
To fuse these heterogeneous factors, we propose a novel triaffine mechanism including triaffine attention and scoring.
Triaffine attention uses boundaries and labels as queries, and uses inside tokens and related spans as keys and values for span representations.
Triaffine scoring interacts with boundaries and span representations for classification.
Experiments show that our proposed method achieves the state-of-the-art $F_1$ scores on four nested NER datasets: ACE2004, ACE2005, GENIA, and KBP2017. | ['Anonymous'] | 2021-11-16 | null | null | null | acl-arr-november-2021-11 | ['nested-named-entity-recognition'] | ['natural-language-processing'] | [-2.55426288e-01 -1.96714908e-01 -2.33707041e-01 -4.38616842e-01
-7.16790557e-01 -8.58812392e-01 3.25021923e-01 5.12464881e-01
-7.50958443e-01 7.02594876e-01 5.08716047e-01 -1.25316799e-01
4.69935574e-02 -7.71424890e-01 -7.00972497e-01 -3.23059320e-01
-1.58211738e-01 2.99400032e-01 3.95194054e-01 -2.43010312e-01
4.13310751e-02 3.71325105e-01 -1.17066252e+00 4.46125537e-01
1.43604434e+00 1.12062025e+00 3.98536362e-02 4.53616798e-01
-6.08848989e-01 8.73217165e-01 -5.89870512e-01 -7.18447030e-01
4.96129617e-02 -1.84992000e-01 -1.09286940e+00 -4.36875880e-01
3.11092168e-01 1.50748156e-02 -2.21063167e-01 9.89512146e-01
3.10908377e-01 2.32744306e-01 6.47120297e-01 -1.07632613e+00
-9.90761876e-01 1.10608470e+00 -4.04922813e-01 3.64770919e-01
3.49797308e-01 -1.34863883e-01 1.55691302e+00 -9.72164869e-01
4.62458611e-01 1.24157870e+00 7.09676921e-01 2.21298426e-01
-1.01645458e+00 -5.59891403e-01 3.92022133e-01 5.02122521e-01
-1.51273453e+00 -5.09089381e-02 6.23253524e-01 -2.90936351e-01
1.11356604e+00 2.59572685e-01 3.28781426e-01 9.61400568e-01
-2.47435585e-01 1.13130832e+00 7.04125762e-01 -4.25589085e-01
-1.05996318e-01 -2.20360402e-02 6.35129809e-01 7.87390292e-01
4.19998728e-02 -2.23659396e-01 -3.84698778e-01 -1.55540258e-01
4.66363728e-01 3.92298587e-02 -1.92258224e-01 1.95043478e-02
-1.21705949e+00 7.09951520e-01 7.55693018e-01 3.79045129e-01
-4.11938667e-01 5.88803142e-02 6.96653724e-01 1.11816451e-01
1.31074205e-01 5.89486957e-01 -6.36824608e-01 -1.75054044e-01
-6.32542789e-01 2.16394484e-01 5.60873270e-01 1.12564600e+00
6.90355182e-01 -1.33658543e-01 -5.48196197e-01 9.79291677e-01
9.69671980e-02 3.93979400e-01 5.44330776e-01 -5.04066050e-01
8.52972388e-01 8.62736702e-01 1.42153591e-01 -7.83529282e-01
-5.13120353e-01 -2.94889987e-01 -7.52187908e-01 -5.96337557e-01
3.27721000e-01 -1.25441596e-01 -1.05748391e+00 1.93586946e+00
2.32771724e-01 1.15248762e-01 -6.35712873e-03 8.40168953e-01
9.45694149e-01 7.90804803e-01 4.52498615e-01 1.27834352e-02
1.76975477e+00 -1.25328720e+00 -7.66683817e-01 -1.48917735e-01
7.94778049e-01 -4.85695899e-01 1.34022999e+00 3.50600295e-02
-9.14435744e-01 -7.04677045e-01 -1.00381327e+00 -5.45358837e-01
-6.47516549e-01 2.72268087e-01 2.92235225e-01 1.91452086e-01
-3.03067327e-01 4.30277586e-01 -6.58819318e-01 -9.01417881e-02
8.25804248e-02 3.56342606e-02 -3.73568386e-01 -8.10692981e-02
-1.97429204e+00 1.01155913e+00 7.41628647e-01 3.48112285e-01
-6.05416358e-01 -7.56294906e-01 -1.19895267e+00 4.43164736e-01
4.03121442e-01 -2.83640265e-01 1.09622049e+00 -3.04178327e-01
-8.44533920e-01 5.06341338e-01 -2.40804508e-01 -5.13285339e-01
2.75295138e-01 -5.33336937e-01 -6.73279524e-01 5.09015620e-02
1.26467004e-01 5.93564332e-01 3.15953523e-01 -7.49032974e-01
-5.60089052e-01 -2.59547323e-01 9.57280844e-02 1.62983090e-01
-6.86740279e-01 -9.72790457e-03 -3.56127173e-01 -6.92751467e-01
-2.79022962e-01 -7.04681158e-01 -2.29615912e-01 -5.00960827e-01
-4.97403860e-01 -8.29484701e-01 6.70259237e-01 -1.01730919e+00
1.76542914e+00 -2.06310797e+00 1.58568189e-01 -2.92574279e-02
1.04785219e-01 4.17622298e-01 -3.68742168e-01 5.84843636e-01
-3.61009297e-04 4.79942322e-01 -2.34494001e-01 -1.17033631e-01
3.42756838e-01 2.74134912e-02 -3.14389884e-01 3.18919346e-02
5.85103571e-01 1.00910425e+00 -8.74194741e-01 -5.99999487e-01
-7.25282580e-02 3.36600721e-01 -2.11628243e-01 3.27848852e-01
-2.82863587e-01 -1.58504307e-01 -4.01505202e-01 6.36353135e-01
4.64663625e-01 -3.11833531e-01 4.20911051e-02 -3.44073564e-01
-6.82862028e-02 5.46634853e-01 -1.11624181e+00 1.56994331e+00
-4.17964071e-01 3.69040459e-01 -2.67565578e-01 -8.94195735e-01
1.00714314e+00 4.49587822e-01 1.51670605e-01 -6.24510884e-01
-2.84363260e-03 2.11940318e-01 -1.27253488e-01 -6.23087227e-01
7.20472932e-01 -9.65291634e-02 -6.79231703e-01 1.67973587e-04
1.63239434e-01 4.61510688e-01 5.67266405e-01 9.19089764e-02
1.31411004e+00 -1.37635469e-01 3.58186871e-01 -1.97671413e-01
7.39773512e-01 -9.86215621e-02 1.05446422e+00 3.92361045e-01
-3.14037889e-01 1.78832203e-01 7.35710204e-01 -5.67625701e-01
-1.00747287e+00 -9.41865921e-01 -7.02023879e-03 1.36773980e+00
9.67277065e-02 -4.73739326e-01 -6.28063917e-01 -1.15354431e+00
1.59484632e-02 4.44925457e-01 -7.42205441e-01 -8.56595114e-02
-8.00525784e-01 -2.80796587e-01 9.97013211e-01 1.08581674e+00
6.32201195e-01 -1.38794041e+00 -2.76620060e-01 4.03744847e-01
-5.05736053e-01 -1.25459516e+00 -9.57804859e-01 4.14574504e-01
-4.63361144e-01 -8.85663450e-01 -7.96452999e-01 -9.69803452e-01
3.37453336e-01 -1.77824479e-02 1.23634636e+00 -5.32311052e-02
-1.59118548e-01 -3.15948129e-01 -6.95937276e-01 -2.01762542e-01
1.31522745e-01 7.27032244e-01 -1.07480101e-01 -2.81708866e-01
4.69550580e-01 -3.20904672e-01 -4.53869760e-01 2.07918718e-01
-1.00396264e+00 -3.59676480e-01 7.57372618e-01 1.07913256e+00
4.63524491e-01 -2.47814655e-01 7.87059963e-01 -1.02966821e+00
7.14867353e-01 -5.17744482e-01 -3.90546948e-01 7.29728639e-01
-5.09293735e-01 2.90966749e-01 1.19559157e+00 -4.58974123e-01
-1.02539361e+00 -4.92735542e-02 -2.43670866e-01 -4.94723976e-01
-1.94067702e-01 8.11143517e-01 -5.35496116e-01 3.41578752e-01
6.13851726e-01 9.63118374e-02 -6.33695960e-01 -5.70432901e-01
5.87207973e-01 6.74587965e-01 4.56871778e-01 -7.52997100e-01
6.24838233e-01 -7.53547475e-02 -4.18708980e-01 -7.07287967e-01
-9.67271388e-01 -5.38580656e-01 -5.46584308e-01 2.97062874e-01
1.04144084e+00 -8.68454695e-01 -7.12104976e-01 3.00260037e-01
-1.31057715e+00 9.08001419e-03 -5.06453663e-02 3.55924875e-01
6.08941652e-02 4.89861399e-01 -9.39102590e-01 -5.53877354e-01
-5.64117014e-01 -9.79838729e-01 8.70020390e-01 6.18325949e-01
-2.18544334e-01 -8.62263203e-01 -2.92142224e-03 2.86795020e-01
3.36874276e-01 2.38961335e-02 1.25634515e+00 -9.23067093e-01
-2.73958623e-01 -1.91388920e-01 -4.16189611e-01 3.05904061e-01
-1.46455392e-01 -9.63434577e-02 -7.09604561e-01 -2.32593268e-02
-6.02103353e-01 -5.10202706e-01 1.03090370e+00 -1.44252241e-01
1.16900158e+00 -4.00602520e-01 -1.93971694e-01 5.08951902e-01
1.32296610e+00 2.87864327e-01 5.08760333e-01 2.46751636e-01
8.44564319e-01 6.11866713e-01 7.31208563e-01 2.65327424e-01
6.98988616e-01 4.07038420e-01 2.13599578e-01 1.38088778e-01
2.26031557e-01 -4.79206294e-01 3.22505891e-01 9.94022191e-01
2.11860910e-01 -2.96942353e-01 -1.06951165e+00 8.09187531e-01
-1.75096166e+00 -1.10741234e+00 5.63410074e-02 1.73448098e+00
8.85078490e-01 1.41655877e-01 1.75997093e-01 -4.83158566e-02
8.80357981e-01 3.03107232e-01 -6.27448976e-01 -3.85405093e-01
-8.75443146e-02 6.85943142e-02 3.14037442e-01 1.41863853e-01
-1.26425076e+00 1.10668874e+00 5.56148005e+00 9.57732320e-01
-1.02344167e+00 -1.57165170e-01 4.88370985e-01 3.29604119e-01
-5.39369047e-01 -2.23661363e-01 -1.09578407e+00 6.73539400e-01
9.44713235e-01 -1.47719085e-01 2.99953580e-01 9.17763829e-01
-2.91473955e-01 3.89544815e-01 -1.04355586e+00 7.50972927e-01
-1.46754578e-01 -1.10550630e+00 -9.73407403e-02 -1.38631970e-01
5.80103159e-01 -7.77632371e-02 -1.31443694e-01 9.67822254e-01
5.52234888e-01 -1.02123928e+00 7.10732758e-01 3.26201946e-01
6.06211901e-01 -8.20186496e-01 9.01823640e-01 2.39772588e-01
-1.72938526e+00 -2.30799645e-01 -3.32820714e-01 2.71920472e-01
2.39653304e-01 5.40320694e-01 -6.96139753e-01 7.17218280e-01
7.03594267e-01 5.03223240e-01 -4.63438839e-01 1.06640267e+00
-4.88333821e-01 6.30542278e-01 -2.20141992e-01 -3.28552037e-01
3.38020355e-01 1.82597861e-02 2.66949594e-01 1.58120954e+00
2.52020210e-01 9.48057398e-02 2.80855924e-01 7.20284224e-01
-6.25533879e-01 2.55656213e-01 -1.92461938e-01 -4.06237692e-01
1.00020170e+00 1.32121658e+00 -5.41846633e-01 -4.16974962e-01
-4.95768487e-01 8.84845853e-01 9.42443609e-01 2.97452956e-01
-1.10086894e+00 -1.07545149e+00 6.17407620e-01 -2.84135878e-01
6.64711297e-01 -1.77053720e-01 -2.24549577e-01 -1.38130176e+00
3.02013367e-01 -7.43673205e-01 8.35018039e-01 -4.74356771e-01
-1.60410774e+00 8.63829017e-01 -1.99877292e-01 -1.06250048e+00
-1.51840001e-02 -4.24590498e-01 -5.90744078e-01 8.22938144e-01
-1.59251738e+00 -1.16713023e+00 -2.49870256e-01 3.64974141e-01
4.55587983e-01 1.52398646e-01 8.12105596e-01 4.59258586e-01
-8.02000463e-01 8.65735710e-01 -9.43135470e-03 8.06385636e-01
7.94990301e-01 -1.50869548e+00 5.17592967e-01 8.75932574e-01
2.45846987e-01 8.77793372e-01 3.34094644e-01 -4.42514390e-01
-1.05350661e+00 -1.09696651e+00 1.16974938e+00 1.24244820e-02
6.69240654e-01 -4.18407321e-01 -1.21592081e+00 6.50556743e-01
2.63530970e-01 3.46263856e-01 7.88918912e-01 4.68615711e-01
-8.52979243e-01 -2.54148126e-01 -9.10144925e-01 4.13195610e-01
1.13815427e+00 -7.98761904e-01 -9.46971059e-01 -6.39838353e-02
1.15434253e+00 -2.98907608e-01 -1.12702823e+00 5.10116935e-01
4.66510534e-01 -3.65675777e-01 8.42526019e-01 -8.44465673e-01
4.39477891e-01 -2.77345479e-01 -1.15002558e-01 -1.45939720e+00
-4.85571653e-01 -2.71314651e-01 -3.03682595e-01 1.68823719e+00
7.67627835e-01 -4.04011309e-01 6.01149380e-01 5.36795974e-01
-2.54882693e-01 -8.93560946e-01 -8.17783237e-01 -9.56046641e-01
1.88315734e-01 -5.48677854e-02 9.20044303e-01 1.18018043e+00
1.96791887e-01 5.85152090e-01 -2.55677551e-01 3.06079090e-01
3.44267398e-01 2.69805074e-01 2.12348491e-01 -1.12879455e+00
-7.09023178e-02 -4.69181687e-01 -2.61508971e-01 -1.17809856e+00
2.40449131e-01 -8.88999879e-01 2.13343263e-01 -1.45177364e+00
9.26686078e-02 -5.68234205e-01 -6.58097386e-01 6.84310079e-01
-5.96022248e-01 -1.39538258e-01 3.12175751e-01 -4.17606831e-02
-8.68022978e-01 7.31374085e-01 8.72485280e-01 -2.12603748e-01
-5.82612678e-02 -5.35335839e-01 -7.04248548e-01 4.40973133e-01
6.11275852e-01 -2.82529652e-01 -5.66017181e-02 -7.91014254e-01
2.73855180e-01 2.37664312e-01 -1.33593857e-01 -8.73271227e-01
3.72249782e-01 -1.17584743e-01 3.79971743e-01 -7.42993891e-01
1.07261412e-01 -8.23900342e-01 -2.03413606e-01 1.87450141e-01
-6.14981890e-01 3.26106042e-01 5.92586137e-02 4.74396616e-01
-5.67453682e-01 -2.74534166e-01 4.16968465e-01 -2.21154615e-01
-8.63045096e-01 3.80502433e-01 -1.07453682e-01 4.28076535e-01
9.50190067e-01 2.70230561e-01 -6.44315302e-01 1.67125002e-01
-7.83040047e-01 8.09465706e-01 1.43553972e-01 5.53172112e-01
4.56361502e-01 -1.56966174e+00 -5.82450211e-01 -4.39311564e-02
2.32102603e-01 1.47614434e-01 5.35843372e-01 5.53133011e-01
-3.99898320e-01 4.00587052e-01 -1.13801137e-01 -1.93338409e-01
-1.19816947e+00 7.79756308e-01 1.11157820e-01 -8.99394512e-01
-1.27127931e-01 9.13341999e-01 2.70483345e-02 -6.54627025e-01
3.51732939e-01 -4.59587276e-01 -4.07040596e-01 2.78902471e-01
6.25949085e-01 4.45289046e-01 5.48162684e-02 -4.55185294e-01
-3.82070392e-01 4.30212855e-01 -4.44973975e-01 1.62231013e-01
1.24627936e+00 -2.45043021e-02 -3.08149550e-02 3.76233131e-01
1.27865160e+00 2.22877283e-02 -7.32885540e-01 -4.70982343e-01
5.35975397e-01 -3.35287862e-02 -4.78767723e-01 -7.74010420e-01
-8.13094735e-01 1.15526986e+00 8.95388946e-02 5.43538630e-01
9.59441662e-01 -1.00690603e-01 1.38272500e+00 3.64158630e-01
1.75009206e-01 -1.03356469e+00 8.10234770e-02 9.02688146e-01
6.33497357e-01 -1.03821945e+00 -5.46239316e-01 -4.52848196e-01
-7.87224948e-01 9.81066465e-01 9.87479448e-01 -3.03048361e-02
4.37519968e-01 1.49718001e-01 -7.10143968e-02 6.45102933e-02
-7.49771297e-01 -3.47177029e-01 5.69451571e-01 1.17837735e-01
6.07858837e-01 6.39294162e-02 -5.25845408e-01 1.09961808e+00
2.23180046e-03 -3.38431418e-01 1.15275383e-02 7.61115611e-01
-4.88857388e-01 -1.17627382e+00 -1.59486279e-01 3.68114263e-01
-6.31268978e-01 -1.91154689e-01 -2.49011666e-01 6.08477831e-01
3.49467456e-01 7.09524155e-01 -8.71265978e-02 -4.32286084e-01
5.56676805e-01 5.26094437e-01 1.75340213e-02 -5.05962431e-01
-8.71903777e-01 -2.91751981e-01 1.86448470e-01 -4.08396602e-01
-8.29799771e-02 -4.58273768e-01 -1.60671854e+00 -7.36296996e-02
-5.34824252e-01 7.00518191e-01 2.70176470e-01 8.68535578e-01
4.05069113e-01 5.85208058e-01 5.68634093e-01 -1.87655166e-01
-7.49714017e-01 -1.24049461e+00 -4.48213309e-01 6.71840727e-01
1.60955310e-01 -5.42682171e-01 -1.48299798e-01 -1.40536323e-01] | [9.549274444580078, 9.31125259399414] |
a0d85b0b-1092-43f2-884f-24f2e22ddd5f | 3d-part-based-sparse-tracker-with-automatic | null | null | http://openaccess.thecvf.com/content_cvpr_2016/html/Bibi_3D_Part-Based_Sparse_CVPR_2016_paper.html | http://openaccess.thecvf.com/content_cvpr_2016/papers/Bibi_3D_Part-Based_Sparse_CVPR_2016_paper.pdf | 3D Part-Based Sparse Tracker With Automatic Synchronization and Registration | In this paper, we present a part-based sparse tracker in a particle filter framework where both the motion and appearance model are formulated in 3D. The motion model is adaptive and directed according to a simple yet powerful occlusion handling paradigm, which is intrinsically fused in the motion model. Also, since 3D trackers are sensitive to synchronization and registration noise in the RGB and depth streams, we propose automated methods to solve these two issues. Extensive experiments are conducted on a popular RGBD tracking benchmark, which demonstrate that our tracker can achieve superior results, outperforming many other recent and state-of-the-art RGBD trackers. | ['Tianzhu Zhang', 'Adel Bibi', 'Bernard Ghanem'] | 2016-06-01 | null | null | null | cvpr-2016-6 | ['occlusion-handling'] | ['computer-vision'] | [-2.50561953e-01 -5.14987946e-01 -1.65706038e-01 -3.16424221e-02
-3.91009480e-01 -3.55955601e-01 6.21718168e-01 -4.16254073e-01
-4.68519062e-01 4.57260907e-01 -9.46838483e-02 1.16990685e-01
2.15932906e-01 -3.20871145e-01 -5.13518333e-01 -9.05993402e-01
1.34985700e-01 2.65956581e-01 7.29435980e-01 2.60478202e-02
-1.37623101e-01 6.94122434e-01 -1.75744295e+00 -6.33179069e-01
6.76351726e-01 1.13925564e+00 1.03854783e-01 5.52839875e-01
7.31640682e-02 4.38412130e-01 -5.90645432e-01 -8.49902108e-02
5.16429424e-01 -1.69593364e-01 -2.88853664e-02 3.09491128e-01
8.17832530e-01 -1.59415409e-01 -6.82193935e-01 1.13162971e+00
7.93580174e-01 4.93599065e-02 3.16409200e-01 -1.27248788e+00
-3.36678594e-01 -3.95265162e-01 -7.11451828e-01 1.29402086e-01
8.15576434e-01 2.61829913e-01 2.72468448e-01 -1.01394188e+00
6.25306904e-01 1.34187949e+00 8.97504330e-01 6.95347369e-01
-1.00358438e+00 -5.52922308e-01 4.30738360e-01 -1.66772187e-01
-1.33030319e+00 -2.47659430e-01 8.41805398e-01 -3.58739227e-01
7.35821307e-01 1.90107957e-01 1.28995371e+00 1.08789802e+00
4.35911298e-01 9.01935577e-01 9.89227891e-01 -1.44590199e-01
2.96394169e-01 -2.85823613e-01 -1.13217963e-03 6.97695374e-01
6.06386960e-01 6.48043513e-01 -9.18040812e-01 -3.34803313e-01
1.11236167e+00 1.08262271e-01 -2.94841707e-01 -9.36976790e-01
-1.44465721e+00 5.39859951e-01 4.47141647e-01 3.75299975e-02
-3.61165732e-01 4.31113362e-01 -3.37155871e-02 -1.79526955e-01
7.68751204e-01 -2.98350096e-01 -2.53994137e-01 -2.90911555e-01
-8.78953278e-01 2.28387445e-01 7.51015604e-01 1.29015172e+00
6.33269072e-01 1.28486037e-01 -2.15009660e-01 1.58666953e-01
1.11962485e+00 1.34832752e+00 3.27471912e-01 -9.44514811e-01
6.96408227e-02 4.53401864e-01 3.64018083e-01 -9.64048684e-01
-5.57142973e-01 -3.92384946e-01 -7.04049766e-01 3.87599379e-01
2.53370196e-01 2.01727957e-01 -1.07077432e+00 1.38193488e+00
1.12871742e+00 8.08824420e-01 -2.31101289e-01 1.20455265e+00
1.10315537e+00 1.18942812e-01 -1.43622264e-01 -3.17117512e-01
1.24035263e+00 -8.82906199e-01 -1.03981793e+00 -2.22892299e-01
1.31527841e-01 -9.46554959e-01 4.47435051e-01 2.03828678e-01
-1.01943827e+00 -9.14545715e-01 -1.03632259e+00 3.34524214e-02
-5.50094247e-02 2.07802393e-02 8.73125732e-01 8.19226861e-01
-1.06980419e+00 3.70043665e-01 -1.51066959e+00 -5.58710635e-01
1.82020292e-01 4.96899217e-01 -1.98075488e-01 -9.00760945e-03
-7.13935435e-01 1.01277208e+00 1.13082506e-01 3.47005844e-01
-6.77411318e-01 -4.70251024e-01 -9.38595116e-01 -7.16484666e-01
1.27463154e-02 -1.13714015e+00 1.07651651e+00 -1.13515161e-01
-1.83744490e+00 8.66505265e-01 -5.88759363e-01 -7.19512403e-02
6.51846886e-01 -3.61738563e-01 -2.63635725e-01 -5.77941276e-02
7.04359040e-02 4.16158795e-01 1.02155924e+00 -1.25823975e+00
-7.58222580e-01 -4.53925490e-01 -2.93956846e-01 1.99179769e-01
-6.39361441e-02 -1.65685430e-01 -1.16476989e+00 -5.83546937e-01
7.54882455e-01 -1.22099674e+00 -3.44686985e-01 5.02816141e-01
-1.43451080e-01 -2.89507121e-01 1.03556180e+00 -2.25639701e-01
9.26646411e-01 -2.05857730e+00 2.94157624e-01 2.69218199e-02
3.23469073e-01 1.53847918e-01 2.05741256e-01 -7.97969252e-02
4.87475574e-01 -5.43439209e-01 6.20803535e-02 -1.06041670e+00
3.18650573e-01 4.11547393e-01 -1.29809305e-01 1.24462414e+00
-1.25979453e-01 8.18810761e-01 -1.13154185e+00 -6.77634418e-01
6.66130304e-01 9.20362711e-01 -6.39159158e-02 3.41050267e-01
-2.93921921e-02 9.01736736e-01 -7.45758176e-01 1.20818377e+00
9.02541041e-01 -3.19145501e-01 -3.17915678e-01 -4.21470761e-01
-3.18402886e-01 1.87603995e-01 -1.42362201e+00 2.40246725e+00
9.12005082e-02 3.29095751e-01 1.30682513e-02 -3.01735461e-01
8.32436562e-01 2.88275838e-01 9.86827433e-01 -6.35878444e-01
2.60155201e-01 2.98833638e-01 -4.72906083e-01 -6.98586255e-02
6.09143317e-01 1.89014956e-01 1.34583443e-01 2.22085342e-01
-3.67012098e-02 -3.24571937e-01 1.34060383e-02 1.33295074e-01
1.18902588e+00 9.90783334e-01 5.84485047e-02 -1.45278484e-01
5.42179227e-01 8.56540427e-02 1.05809081e+00 9.39751804e-01
-6.54398501e-01 6.92090034e-01 -4.21931952e-01 -5.85796833e-01
-4.93443936e-01 -1.17166591e+00 -3.32444608e-01 2.78556347e-01
9.04416800e-01 -4.56989318e-01 -1.65856481e-01 -5.79912782e-01
4.33355182e-01 -1.85813427e-01 -5.54143012e-01 2.48690322e-02
-6.58357859e-01 -7.63753235e-01 1.34240225e-01 6.24803901e-01
4.11116749e-01 -3.68896902e-01 -8.61774862e-01 3.95265281e-01
1.05978265e-01 -1.47692847e+00 -4.58377391e-01 4.19187248e-02
-1.28268695e+00 -1.22152114e+00 -6.80508733e-01 -4.99501675e-01
6.43536091e-01 7.38887489e-01 1.18462265e+00 2.07075730e-01
-2.63048321e-01 1.04343295e+00 -4.29915845e-01 -4.13319379e-01
8.19624886e-02 -3.16857100e-01 4.67460930e-01 -1.77681848e-01
5.08438528e-01 -3.75036538e-01 -7.27629542e-01 5.12291849e-01
-4.80009347e-01 -1.93324968e-01 4.47691202e-01 5.19884765e-01
1.19700825e+00 -4.47834969e-01 -5.01209974e-01 -8.88167247e-02
-1.63809404e-01 -8.87709484e-02 -1.11905062e+00 1.20681070e-01
-4.74552423e-01 -1.44116297e-01 -1.65299162e-01 -6.76620841e-01
-8.33565533e-01 8.45185041e-01 3.77269015e-02 -9.26007569e-01
1.32348478e-01 -1.94998831e-01 -2.15765268e-01 -9.63250935e-01
8.98432732e-02 3.60526830e-01 5.21222427e-02 -5.44471800e-01
2.83025533e-01 3.66879776e-02 7.08146393e-01 -5.29136121e-01
1.54739785e+00 1.15632260e+00 3.93768370e-01 -7.02315271e-01
-7.44259238e-01 -8.26304078e-01 -9.62755084e-01 -4.97836083e-01
7.15086997e-01 -1.26947296e+00 -1.02781689e+00 6.24488711e-01
-1.27816987e+00 1.56760365e-02 -3.84816229e-01 8.61917198e-01
-5.67141175e-01 6.03632390e-01 -4.38233584e-01 -9.00370896e-01
-2.01273486e-01 -1.28556955e+00 1.68970966e+00 5.76167524e-01
1.69558734e-01 -1.09578359e+00 6.08247459e-01 -2.66677737e-01
2.81908602e-01 4.96815652e-01 -4.74452078e-01 4.51534428e-02
-9.63962913e-01 -4.19283181e-01 2.43109856e-02 -1.83417052e-01
3.06492150e-01 7.80483931e-02 -9.34453964e-01 -5.96562684e-01
2.81809032e-01 6.76223412e-02 6.09692812e-01 5.31458914e-01
3.92980456e-01 4.59621876e-01 -8.16160679e-01 1.13334656e+00
1.31952453e+00 9.52620804e-03 4.29826975e-01 5.54400861e-01
9.84921157e-01 -5.11698388e-02 1.05788422e+00 5.97408652e-01
6.04204357e-01 1.09422243e+00 6.59211755e-01 -1.60834536e-01
-4.02068257e-01 -5.50108887e-02 3.74574363e-01 9.37343419e-01
-2.10881442e-01 1.38046637e-01 -4.66244131e-01 1.56728581e-01
-2.21157098e+00 -6.55306816e-01 -5.78014314e-01 2.05360317e+00
6.27307236e-01 -3.23996805e-02 1.07472412e-01 -1.71308339e-01
4.90586638e-01 2.78631002e-01 -6.62508011e-01 6.09067976e-01
-4.14652646e-01 2.51222640e-01 8.64334106e-01 3.29687804e-01
-1.43010366e+00 9.83989716e-01 7.26966524e+00 4.22296524e-01
-7.90670633e-01 2.23970205e-01 -3.92234772e-01 -1.82122365e-01
-3.63115850e-03 1.04100592e-01 -1.42968440e+00 5.77960074e-01
5.18743813e-01 8.94976035e-02 -1.17444329e-01 7.69976079e-01
1.76881254e-01 -3.91273648e-01 -9.42306519e-01 1.32128644e+00
1.32976174e-01 -1.02792513e+00 -5.10996878e-01 3.78829539e-01
7.17601180e-01 1.98899359e-01 -4.13812064e-02 -8.69233906e-02
3.47924799e-01 -4.49353486e-01 8.93659413e-01 7.12674141e-01
1.59963071e-01 -3.13511699e-01 6.02595508e-01 8.00645873e-02
-1.81806040e+00 5.73812842e-01 -6.56512976e-01 1.39654651e-01
4.41924810e-01 6.89669192e-01 3.84281464e-02 7.73188531e-01
1.07323205e+00 1.32752001e+00 -5.27162731e-01 1.60065651e+00
-5.04977703e-01 6.97781295e-02 -7.47383654e-01 7.94952139e-02
-6.21207543e-02 -1.51064217e-01 7.66246736e-01 7.37113118e-01
4.32658851e-01 1.64075390e-01 4.92932826e-01 6.81996405e-01
3.31799537e-01 -4.77531046e-01 -4.55899268e-01 6.00241840e-01
5.26374936e-01 1.27031350e+00 -7.84797847e-01 -2.30623767e-01
-6.82661653e-01 9.45187151e-01 -1.59696192e-01 2.37982228e-01
-1.05980885e+00 1.59362659e-01 1.10629189e+00 -2.68616736e-01
6.39518142e-01 -6.84246778e-01 1.52690280e-02 -1.49377346e+00
3.60257365e-02 -4.14904803e-01 1.47282839e-01 -6.79725528e-01
-1.31128871e+00 4.23135757e-01 -2.49587715e-01 -1.93685818e+00
5.55182211e-02 -6.69434488e-01 -3.24189842e-01 7.01849043e-01
-1.83499837e+00 -1.22610223e+00 -6.83545291e-01 1.05950403e+00
1.27161697e-01 1.76457718e-01 5.84654152e-01 4.89244699e-01
-4.81207848e-01 3.09387475e-01 9.73880813e-02 2.99048908e-02
7.71494389e-01 -1.01308191e+00 5.11661530e-01 1.03032100e+00
3.74291986e-01 5.54005384e-01 7.13531017e-01 -9.29396272e-01
-2.32768536e+00 -7.89871931e-01 4.95988548e-01 -8.82557452e-01
6.33730292e-01 -3.05008769e-01 -5.13743579e-01 6.16606772e-01
-7.81328976e-02 6.94860101e-01 5.11464536e-01 -1.75510496e-01
-5.35850376e-02 -4.22175899e-02 -9.15231884e-01 -1.34264643e-03
1.36818337e+00 -2.58041888e-01 -6.08574450e-01 3.82685035e-01
6.95492506e-01 -1.29520822e+00 -1.12728918e+00 4.92132992e-01
7.04185367e-01 -8.77310574e-01 1.41650152e+00 -3.22089978e-02
-6.77212179e-01 -1.01745689e+00 -2.79198229e-01 -8.31584156e-01
-3.35540593e-01 -7.33301163e-01 -1.05521810e+00 1.00486934e+00
-5.52344799e-01 -4.70319211e-01 1.06372237e+00 5.06911099e-01
-2.01278150e-01 -2.03960270e-01 -1.48273039e+00 -1.19988334e+00
-6.75117433e-01 -4.80791241e-01 4.66556698e-01 5.37861109e-01
-6.70551956e-01 -1.90773427e-01 -5.33354938e-01 5.89896142e-01
1.44888353e+00 3.24360311e-01 1.14119279e+00 -1.38072383e+00
-8.16598907e-02 -1.90843895e-01 -8.06445181e-01 -1.55649424e+00
7.09123537e-02 -3.17466110e-01 2.68084258e-01 -1.40764904e+00
-1.17482565e-01 -3.70198637e-01 -3.29421580e-01 1.37000903e-01
-3.30495507e-01 4.68163073e-01 2.33910441e-01 6.21394336e-01
-1.00785017e+00 9.60458100e-01 1.21807289e+00 -2.04181239e-01
-1.75095960e-01 2.13305473e-01 9.53876134e-03 7.32719481e-01
1.53664062e-02 -7.31016278e-01 8.16898644e-02 -6.21099710e-01
-1.88070551e-01 -2.46824622e-01 5.67326486e-01 -1.23756492e+00
7.92856395e-01 3.30432393e-02 6.77724719e-01 -1.31122196e+00
7.43042946e-01 -1.10498476e+00 3.69544804e-01 8.08405459e-01
6.36992931e-01 1.86215311e-01 2.16755956e-01 8.65055263e-01
-1.67805925e-01 4.08757329e-01 5.82672834e-01 6.61327690e-02
-8.60088646e-01 9.30706501e-01 1.01098455e-01 -1.01080433e-01
1.04778528e+00 -6.66977882e-01 -1.58422634e-01 -9.63258743e-02
-4.89871114e-01 2.66391188e-01 8.95170987e-01 4.94207174e-01
7.02284098e-01 -1.85238683e+00 -3.67677331e-01 2.14640856e-01
1.51111662e-01 1.53696850e-01 -9.21342289e-04 1.17436922e+00
-3.56376410e-01 2.52945453e-01 5.03897406e-02 -1.35215700e+00
-1.11908293e+00 5.07713854e-01 2.70185143e-01 1.70593485e-02
-8.77453208e-01 7.80427754e-01 -9.87597331e-02 -2.02850372e-01
6.31617010e-01 -3.94106537e-01 2.31035545e-01 -1.85825780e-01
5.55451989e-01 3.92714679e-01 -1.62994072e-01 -1.05536556e+00
-9.87574577e-01 1.32245350e+00 5.37242532e-01 1.59589320e-01
1.20108438e+00 -3.96058321e-01 -1.10835694e-02 4.04540241e-01
7.46157944e-01 -7.31664598e-02 -1.79179168e+00 -3.85081708e-01
-3.70926745e-02 -9.65436399e-01 8.91479552e-02 -1.15182310e-01
-1.34234965e+00 4.41063255e-01 9.92040575e-01 6.99475482e-02
9.58595753e-01 2.43388917e-02 8.28591287e-01 7.85234421e-02
8.22947681e-01 -8.91162634e-01 4.30110237e-03 5.73073268e-01
2.88617939e-01 -1.41206110e+00 6.70604706e-01 -4.42459792e-01
1.01632811e-01 8.50202799e-01 5.93309045e-01 -3.17699969e-01
8.13016772e-01 3.90594542e-01 7.50666037e-02 -1.71880454e-01
-3.67864817e-01 -5.54812074e-01 4.50631917e-01 9.60624456e-01
2.50986874e-01 -4.57484573e-01 -2.33492672e-01 2.23625898e-02
2.58678228e-01 -1.56901002e-01 -9.39937085e-02 1.37454128e+00
-2.87409395e-01 -1.28200448e+00 -8.99060369e-01 -2.64597237e-01
2.14969255e-02 3.55451435e-01 -1.54310018e-01 1.16094995e+00
1.13823618e-04 7.92266607e-01 -7.55780786e-02 -3.65851402e-01
5.61901093e-01 -3.98883879e-01 8.69849265e-01 -2.75495291e-01
-6.11834764e-01 5.89589119e-01 -4.80826676e-01 -1.30972016e+00
-1.37654603e+00 -8.76898825e-01 -1.13657928e+00 -2.76289910e-01
-5.39891660e-01 -5.46321347e-02 8.83176327e-01 8.69173169e-01
3.63723516e-01 4.90024954e-01 5.41858673e-01 -1.54448783e+00
-3.85378838e-01 -5.83822548e-01 -7.12360203e-01 2.41591707e-01
4.00641412e-01 -1.26761615e+00 -3.97694737e-01 -1.72417596e-01] | [6.689198970794678, -2.183455228805542] |
7e3f3616-ada6-4e6a-b42e-1184277de6a9 | efficiently-maintaining-next-basket | 2201.13313 | null | https://arxiv.org/abs/2201.13313v1 | https://arxiv.org/pdf/2201.13313v1.pdf | Efficiently Maintaining Next Basket Recommendations under Additions and Deletions of Baskets and Items | Recommender systems play an important role in helping people find information and make decisions in today's increasingly digitalized societies. However, the wide adoption of such machine learning applications also causes concerns in terms of data privacy. These concerns are addressed by the recent "General Data Protection Regulation" (GDPR) in Europe, which requires companies to delete personal user data upon request when users enforce their "right to be forgotten". Many researchers argue that this deletion obligation does not only apply to the data stored in primary data stores such as relational databases but also requires an update of machine learning models whose training set included the personal data to delete. We explore this direction in the context of a sequential recommendation task called Next Basket Recommendation (NBR), where the goal is to recommend a set of items based on a user's purchase history. We design efficient algorithms for incrementally and decrementally updating a state-of-the-art next basket recommendation model in response to additions and deletions of user baskets and items. Furthermore, we discuss an efficient, data-parallel implementation of our method in the Spark Structured Streaming system. We evaluate our implementation on a variety of real-world datasets, where we investigate the impact of our update techniques on several ranking metrics and measure the time to perform model updates. Our results show that our method provides constant update time efficiency with respect to an additional user basket in the incremental case, and linear efficiency in the decremental case where we delete existing baskets. With modest computational resources, we are able to update models with a latency of around 0.2~milliseconds regardless of the history size in the incremental case, and less than one millisecond in the decremental case. | ['Sebastian Schelter', 'Benjamin Longxiang Wang'] | 2022-01-27 | null | null | null | null | ['next-basket-recommendation'] | ['miscellaneous'] | [ 1.23515446e-02 -3.87917429e-01 -2.72718251e-01 -6.07325852e-01
-2.68525064e-01 -6.90595806e-01 2.13547930e-01 8.65856647e-01
-8.38278472e-01 4.57506001e-01 2.45495699e-02 -7.15288103e-01
-3.04555655e-01 -1.13349104e+00 -6.56222224e-01 -2.41843805e-01
-9.20979381e-02 5.70490777e-01 3.14301461e-01 -3.63342822e-01
4.58154976e-01 6.84643269e-01 -1.72404456e+00 4.83119786e-01
5.46522021e-01 1.06666124e+00 -2.21307129e-02 4.07472700e-01
-7.89844319e-02 3.00350487e-01 -4.66848493e-01 -6.94276214e-01
7.23042667e-01 2.66773492e-01 -7.08819389e-01 -3.21086973e-01
8.07343498e-02 -6.04102731e-01 -1.84739783e-01 9.40969944e-01
3.45823646e-01 3.07745039e-01 7.99644887e-02 -1.27294064e+00
-4.35273111e-01 7.57237017e-01 -7.68278360e-01 3.42273235e-01
4.89545554e-01 -3.69264811e-01 9.35767353e-01 -3.21575105e-01
5.66466630e-01 5.54701984e-01 6.17476404e-01 2.32972875e-01
-1.29464138e+00 -7.77072012e-01 5.27210116e-01 2.03091905e-01
-1.31272256e+00 -4.76704150e-01 2.44541764e-01 -2.89069802e-01
1.04014301e+00 8.78142893e-01 3.56745571e-01 3.42016160e-01
1.31140530e-01 5.25482774e-01 5.66790700e-01 -2.77497232e-01
4.96856511e-01 4.31865245e-01 5.95655501e-01 1.00614410e-02
9.50747728e-01 -9.52100605e-02 -3.79883260e-01 -7.96909690e-01
9.49481949e-02 3.19399297e-01 -1.54067669e-02 -3.86553258e-01
-7.10608542e-01 8.46835911e-01 -4.76562344e-02 1.21328548e-01
-4.98633295e-01 -9.19957533e-02 6.00254357e-01 5.95585883e-01
4.85663444e-01 3.10241878e-01 -7.60453224e-01 -4.79465351e-02
-6.84365392e-01 5.06810486e-01 1.19196081e+00 8.74778032e-01
5.47171652e-01 -4.45047528e-01 3.55759233e-01 7.54163027e-01
3.05587113e-01 2.33406082e-01 4.21358615e-01 -6.64148092e-01
4.99794126e-01 4.52407300e-01 5.89650571e-01 -1.02582014e+00
-3.45891654e-01 -1.57937765e-01 -5.87091446e-01 -1.06327944e-01
5.01249731e-01 -8.21439624e-02 -2.83507824e-01 1.50954163e+00
6.76445425e-01 -1.27913415e-01 -1.24733783e-01 5.70679605e-01
4.27140474e-01 4.27565366e-01 -6.61312342e-02 -6.56532407e-01
1.38380814e+00 -2.14099765e-01 -5.39484203e-01 1.76928535e-01
7.74727046e-01 -7.89455414e-01 7.36812294e-01 7.67260134e-01
-9.12032664e-01 -2.22357884e-01 -9.67179298e-01 -6.71693087e-02
-4.98101860e-01 -2.62905836e-01 8.35438728e-01 1.13446665e+00
-6.42759085e-01 6.71860278e-01 -9.19834137e-01 -4.23283160e-01
2.09542766e-01 6.05632663e-01 -1.85674414e-01 -2.56500810e-01
-9.67877328e-01 5.12074590e-01 1.56316325e-01 -2.61374936e-02
-2.88714588e-01 -9.02697623e-01 -2.68718302e-01 1.92342058e-01
5.74339867e-01 -4.26180065e-01 1.22080350e+00 -3.65047991e-01
-9.30959165e-01 4.49953943e-01 -1.27870888e-02 -6.83718860e-01
4.54316795e-01 -3.89317602e-01 -7.53914714e-01 -4.15735364e-01
-1.75325781e-01 -1.72656119e-01 3.21941823e-01 -9.20455158e-01
-1.14408314e+00 -7.54886925e-01 3.08561593e-01 1.85148433e-01
-5.35717845e-01 2.54681975e-01 -4.50738132e-01 -5.92638195e-01
-5.65110380e-03 -1.05342519e+00 -5.01787484e-01 -3.42168123e-01
-7.78310895e-02 -2.21876562e-01 6.83979690e-01 -5.31382740e-01
1.68835640e+00 -2.12415743e+00 -5.89020073e-01 7.41967380e-01
8.67899414e-03 2.29808509e-01 1.46051928e-01 6.32455468e-01
4.67327796e-03 1.75569743e-01 2.27123335e-01 -4.12062183e-02
1.61318723e-02 2.51418084e-01 -6.05792165e-01 4.48332220e-01
-8.82197499e-01 2.24975497e-01 -6.44200206e-01 1.23864532e-01
-1.77996941e-02 1.83830351e-01 -7.67956972e-01 -5.69278710e-02
-2.66674757e-01 -3.60502005e-01 -4.59539711e-01 3.31172854e-01
7.08281934e-01 -1.26670569e-01 7.21819937e-01 -6.51833341e-02
-2.14047894e-01 5.59798360e-01 -1.72727430e+00 1.23245430e+00
-1.57428935e-01 -6.52596429e-02 1.58724278e-01 -9.01250362e-01
8.10186744e-01 1.51072204e-01 7.96173155e-01 -6.95042968e-01
-2.75850564e-01 1.17906943e-01 -1.55936182e-01 -2.18516141e-01
8.90941024e-01 2.79039592e-01 5.90947084e-03 1.06951249e+00
-8.07998061e-01 6.02245688e-01 2.80408174e-01 4.16760147e-01
1.32061815e+00 -3.34731579e-01 3.92359078e-01 -1.88357845e-01
2.22506508e-01 -3.77205200e-02 6.70259237e-01 8.93497050e-01
1.84154138e-01 -2.88462397e-02 1.70634985e-01 -9.34366643e-01
-7.66900361e-01 -5.79097509e-01 -1.21264763e-01 1.54007995e+00
1.61951900e-01 -8.88443589e-01 -3.69935036e-01 -7.44079530e-01
6.01028442e-01 1.04935491e+00 -4.14775699e-01 9.49045867e-02
-4.25171435e-01 -1.02507353e+00 1.10005438e-01 3.40310723e-01
2.12104693e-01 -7.35431969e-01 -9.52466965e-01 4.17318404e-01
7.90969357e-02 -6.19613111e-01 -6.81406736e-01 -4.27288888e-03
-8.75307024e-01 -1.19489968e+00 1.97701722e-01 -8.42025876e-02
4.87235457e-01 5.22422969e-01 9.26482737e-01 3.02199900e-01
-1.56206295e-01 4.80899870e-01 -3.97970587e-01 -6.82725132e-01
-1.19173877e-01 1.59591138e-01 2.64760584e-01 -3.45458873e-02
8.98114800e-01 -5.79210222e-01 -6.34313643e-01 6.00636661e-01
-1.10626757e+00 -2.49098703e-01 -3.63581479e-02 3.15405488e-01
5.82067132e-01 4.64049369e-01 6.67388499e-01 -1.48088253e+00
8.04605663e-01 -6.96274936e-01 -9.73569095e-01 4.74125087e-01
-1.33902621e+00 -1.93133160e-01 5.73468804e-01 -3.23635280e-01
-1.06895757e+00 1.63527757e-01 -1.52839512e-01 1.36403769e-01
-5.09431846e-02 6.87304080e-01 -1.37692794e-01 1.88770637e-01
6.07844114e-01 -8.32212418e-02 -6.37974963e-02 -8.26155126e-01
3.76123726e-01 8.70344758e-01 3.81581485e-01 -7.26418734e-01
3.46798450e-01 5.45661986e-01 -3.27439517e-01 -6.17234409e-01
-3.39122862e-01 -8.23570490e-01 -1.88467711e-01 2.06438780e-01
1.46339655e-01 -6.66810215e-01 -1.26958227e+00 4.62892428e-02
-6.98327601e-01 9.67962444e-02 -2.96513259e-01 3.91778886e-01
-1.85932830e-01 4.61262703e-01 -6.17716432e-01 -7.88714707e-01
-6.00232065e-01 -6.78095162e-01 3.84592801e-01 -2.10796118e-01
-2.10205793e-01 -5.43764174e-01 8.11435729e-02 3.70339006e-01
3.41214538e-01 -7.54452497e-02 8.65805030e-01 -1.18195295e+00
-4.46764797e-01 -5.42634547e-01 -3.37090828e-02 2.29327399e-02
7.47327805e-02 -3.77940893e-01 -3.84660721e-01 -6.91947937e-01
-3.66255790e-02 3.04346919e-01 4.14635628e-01 9.35129002e-02
1.21271968e+00 -6.55537665e-01 -5.03169715e-01 2.92356044e-01
1.45582354e+00 6.51186109e-01 6.10958576e-01 2.79259861e-01
3.39546084e-01 4.93095785e-01 9.04041767e-01 9.16455507e-01
4.43254113e-01 7.85630941e-01 1.60684988e-01 3.69797260e-01
5.42829037e-01 -6.78525940e-02 9.84487757e-02 6.41629517e-01
-1.56633005e-01 -2.04036474e-01 -5.56217968e-01 5.28038919e-01
-2.07809877e+00 -1.01461387e+00 -1.56158179e-01 2.90697479e+00
7.80855119e-01 1.47355244e-01 3.61736536e-01 1.71442375e-01
5.79984069e-01 -4.66305226e-01 -8.31154525e-01 -7.24778414e-01
5.49274206e-01 2.02564299e-01 1.02804625e+00 2.78023273e-01
-1.00082433e+00 4.62341696e-01 6.02113724e+00 4.06679928e-01
-9.45273340e-01 1.15494400e-01 5.11511624e-01 -5.17806232e-01
-4.50004190e-01 -5.85618690e-02 -1.01032865e+00 5.60937703e-01
1.23402405e+00 -6.22453868e-01 5.99995971e-01 1.11044514e+00
1.59725755e-01 -1.83797568e-01 -1.26139843e+00 9.36170816e-01
-2.51955003e-01 -1.35050070e+00 -2.34152317e-01 5.98254263e-01
4.91718471e-01 1.41825870e-01 -5.50914705e-02 1.41466215e-01
6.00987673e-01 -4.54417408e-01 3.59090745e-01 3.64526600e-01
4.48832303e-01 -1.02645957e+00 5.82539082e-01 4.48644698e-01
-1.09527111e+00 -2.78488159e-01 -5.98617613e-01 -2.66429663e-01
2.30181906e-02 8.33216012e-01 -8.00990582e-01 4.47131455e-01
1.11462724e+00 1.72746703e-01 -2.39987969e-01 1.06341124e+00
4.79444593e-01 5.54405391e-01 -9.04498041e-01 -4.48714150e-03
-3.12495589e-01 -2.87398785e-01 1.82660609e-01 9.94434059e-01
3.83245707e-01 5.53631842e-01 4.74642590e-02 5.42691797e-02
-2.33636394e-01 4.61458445e-01 -3.45021874e-01 2.36940458e-01
7.74897277e-01 9.29044604e-01 -6.15012109e-01 -3.84688228e-01
-5.06564677e-01 4.69466984e-01 2.72824373e-02 1.56671017e-01
-4.94855791e-01 -4.88830298e-01 9.37019527e-01 6.62891805e-01
5.85403502e-01 2.43586376e-02 -2.46176854e-01 -1.06175840e+00
1.56708628e-01 -1.08431423e+00 8.88637900e-01 -2.67031174e-02
-1.23852205e+00 2.23367453e-01 6.68295622e-02 -7.19464123e-01
-3.15834343e-01 -1.99700341e-01 -1.03304937e-01 6.78152323e-01
-8.72780323e-01 -5.05016804e-01 3.69471237e-02 6.55859649e-01
7.41382092e-02 -8.66803750e-02 1.00631380e+00 6.90596998e-01
-3.64236146e-01 8.27413857e-01 6.16267741e-01 -3.19976926e-01
7.99158931e-01 -8.94039214e-01 5.92767358e-01 7.80463755e-01
3.24881494e-01 1.14774930e+00 8.62802923e-01 -8.95773292e-01
-1.54956448e+00 -9.82887924e-01 1.04540837e+00 -4.69550580e-01
2.77538508e-01 -3.57192904e-01 -1.00189579e+00 9.07707810e-01
-2.49633253e-01 3.16732563e-02 1.10839856e+00 7.29059637e-01
-6.22681141e-01 -7.28209019e-01 -1.51704586e+00 2.19307438e-01
8.90587628e-01 -2.54635572e-01 -3.66647005e-01 3.92683744e-01
5.87153196e-01 -1.96988210e-01 -1.15081489e+00 1.54608190e-01
9.81325984e-01 -8.36887658e-01 8.59432459e-01 -8.81748021e-01
-2.90257812e-01 -4.58756506e-01 -3.52849305e-01 -7.61075139e-01
-3.05908740e-01 -7.00507045e-01 -2.68846840e-01 1.03952515e+00
4.16683197e-01 -8.38369131e-01 1.00174510e+00 1.46269965e+00
3.25095773e-01 -6.45576477e-01 -6.60005331e-01 -5.90245187e-01
-4.10467207e-01 -6.99382842e-01 1.14594293e+00 1.11760056e+00
1.67736009e-01 -2.51532849e-02 -5.13508320e-01 3.21765155e-01
4.95681673e-01 3.71492773e-01 9.59079862e-01 -1.34355915e+00
-5.07803559e-01 7.34559745e-02 -1.35449573e-01 -8.96424592e-01
-5.21332562e-01 -6.81576133e-01 -5.14573812e-01 -1.36157167e+00
1.10737190e-01 -9.65988457e-01 -7.67706394e-01 6.03302598e-01
2.56202072e-01 -5.68799339e-02 2.46952757e-01 2.52287269e-01
-6.41244948e-01 -2.56827980e-01 2.58847922e-01 1.23476192e-01
-5.65254629e-01 6.07970893e-01 -1.22827184e+00 4.99847680e-01
6.32374644e-01 -6.56585395e-01 -6.07755065e-01 -2.04929233e-01
7.98579335e-01 -1.24511316e-01 -2.97068596e-01 -6.29545748e-01
5.04854858e-01 -2.32334733e-01 2.45852947e-01 -8.46185505e-01
6.75012097e-02 -1.07771432e+00 7.75046825e-01 4.14961249e-01
-4.54610884e-01 2.56467581e-01 1.09779805e-01 6.75222516e-01
2.99796760e-01 -2.47274548e-01 3.66489857e-01 6.34917244e-02
-3.40788454e-01 3.78189474e-01 -1.80166334e-01 -3.44318211e-01
1.13102353e+00 -7.35717192e-02 -3.11014384e-01 -3.25614482e-01
-6.97586954e-01 2.72525847e-01 4.59860235e-01 4.46024358e-01
4.30378467e-01 -1.00905311e+00 -5.26506722e-01 2.70812690e-01
7.35993534e-02 -3.39649439e-01 9.52495635e-02 3.71711731e-01
-3.37881982e-01 3.75580281e-01 7.67402425e-02 -1.85397733e-02
-1.62881100e+00 1.07515442e+00 -3.52674752e-01 -3.70576411e-01
-5.24552524e-01 5.42448282e-01 -1.44620672e-01 -1.92372888e-01
4.17686224e-01 -2.37118110e-01 -3.19390707e-02 7.04462826e-02
1.01323843e+00 6.86930835e-01 5.65687835e-01 -1.38683692e-01
-3.31962883e-01 -2.37469792e-01 -8.97138119e-01 1.97779268e-01
1.49786806e+00 -2.27699429e-01 -1.97695538e-01 9.20162126e-02
8.99538636e-01 3.50128531e-01 -6.25080764e-01 -2.24527165e-01
1.54932588e-01 -7.44323969e-01 -5.00477143e-02 -8.60779405e-01
-1.00909603e+00 1.29654109e-01 6.58167481e-01 5.82305610e-01
1.17636847e+00 -3.49903166e-01 9.63264763e-01 7.77303636e-01
7.69310951e-01 -1.25327635e+00 -7.46055782e-01 -3.43418904e-02
5.53366482e-01 -8.11983585e-01 6.76154554e-01 -2.79055178e-01
-3.97621363e-01 7.33279645e-01 2.86502659e-01 7.26423711e-02
1.11979115e+00 2.01795846e-01 -1.92688271e-01 1.49962306e-01
-1.00242591e+00 3.04313511e-01 -9.06570405e-02 4.38056916e-01
2.40356311e-01 3.88272196e-01 -8.28113437e-01 8.12193215e-01
-1.56781524e-01 2.37859294e-01 4.81306970e-01 1.19140375e+00
-3.14070016e-01 -1.59612608e+00 -3.41638625e-01 1.05363178e+00
-7.51527131e-01 3.70701887e-02 -3.15780900e-02 5.97536922e-01
1.12682246e-01 1.07060778e+00 2.29255170e-01 -4.06939179e-01
5.51716745e-01 -9.59215090e-02 1.33591652e-01 -5.01755774e-01
-9.32282388e-01 7.62445107e-02 2.63488293e-01 -6.33075297e-01
4.90061333e-03 -9.20236290e-01 -1.07694483e+00 -8.19983661e-01
-3.83104503e-01 5.39777398e-01 1.02644050e+00 6.21350348e-01
7.49003887e-01 -5.88173606e-02 5.92604637e-01 -1.04765154e-01
-8.44641745e-01 -4.67344671e-01 -6.70878887e-01 5.54282963e-01
-1.68275774e-01 -2.86841244e-01 2.85359845e-02 2.18557622e-02] | [6.113073825836182, 6.393284797668457] |
3c76d07c-11f3-44b9-bf92-df698ace0896 | learning-a-compressed-sensing-measurement | 1806.10175 | null | https://arxiv.org/abs/1806.10175v4 | https://arxiv.org/pdf/1806.10175v4.pdf | Learning a Compressed Sensing Measurement Matrix via Gradient Unrolling | Linear encoding of sparse vectors is widely popular, but is commonly data-independent -- missing any possible extra (but a priori unknown) structure beyond sparsity. In this paper we present a new method to learn linear encoders that adapt to data, while still performing well with the widely used $\ell_1$ decoder. The convex $\ell_1$ decoder prevents gradient propagation as needed in standard gradient-based training. Our method is based on the insight that unrolling the convex decoder into $T$ projected subgradient steps can address this issue. Our method can be seen as a data-driven way to learn a compressed sensing measurement matrix. We compare the empirical performance of 10 algorithms over 6 sparse datasets (3 synthetic and 3 real). Our experiments show that there is indeed additional structure beyond sparsity in the real datasets; our method is able to discover it and exploit it to create excellent reconstructions with fewer measurements (by a factor of 1.1-3x) compared to the previous state-of-the-art methods. We illustrate an application of our method in learning label embeddings for extreme multi-label classification, and empirically show that our method is able to match or outperform the precision scores of SLEEC, which is one of the state-of-the-art embedding-based approaches. | ['Dmitry Storcheus', 'Daniel Holtmann-Rice', 'Sujay Sanghavi', 'Afshin Rostamizadeh', 'Shanshan Wu', 'Felix X. Yu', 'Alexandros G. Dimakis', 'Sanjiv Kumar'] | 2018-06-26 | null | null | null | null | ['extreme-multi-label-classification'] | ['methodology'] | [ 3.38257700e-01 3.13012302e-01 -4.01264578e-01 -3.66338491e-01
-1.21628153e+00 -3.04865748e-01 3.15829992e-01 1.00610867e-01
-4.01248425e-01 5.75002015e-01 4.34735745e-01 -1.71646342e-01
-2.98633188e-01 -5.42681158e-01 -9.58621323e-01 -8.51948559e-01
-3.68108392e-01 5.40313303e-01 -1.08193576e-01 -1.44486025e-01
2.26967514e-01 2.48885676e-01 -1.34840250e+00 2.34493881e-01
3.44834059e-01 1.00431406e+00 2.58845091e-02 5.91728985e-01
1.84705555e-02 8.80639434e-01 -7.81229362e-02 -3.32983583e-01
4.78110433e-01 -3.35913569e-01 -6.02643251e-01 2.39325821e-01
5.10375381e-01 -1.75200731e-01 -3.01820844e-01 9.99566674e-01
4.67843205e-01 -1.27356976e-01 6.20092869e-01 -8.59222651e-01
-7.46318281e-01 7.12649465e-01 -6.22076094e-01 -9.11791995e-02
2.82835692e-01 -3.42381954e-01 1.25827348e+00 -1.22923863e+00
6.43597841e-01 9.83127236e-01 1.02840424e+00 4.16108310e-01
-1.35941494e+00 -5.66149473e-01 -1.03458166e-01 1.97519571e-01
-1.59930134e+00 -7.81716168e-01 8.46655488e-01 -3.93588632e-01
8.58701587e-01 1.34498298e-01 4.58304167e-01 1.01373136e+00
-2.42813770e-02 6.92275703e-01 1.17365575e+00 -5.98977387e-01
4.15245980e-01 4.69686747e-01 -1.34669412e-02 1.22319436e+00
4.08254623e-01 2.49724850e-01 -6.52771771e-01 -3.44174832e-01
5.69630980e-01 8.31974298e-02 -3.80486280e-01 -7.14651346e-01
-1.26721311e+00 1.22691798e+00 3.91240865e-01 4.93405342e-01
-2.22648337e-01 5.10272145e-01 3.03589463e-01 4.72882599e-01
4.53463048e-01 3.54910463e-01 -4.52512771e-01 -1.53834894e-01
-1.30700505e+00 -1.35282725e-01 9.17163730e-01 7.88347602e-01
9.31688368e-01 3.04059416e-01 1.73667923e-01 8.21773231e-01
3.78210664e-01 4.94779885e-01 6.72617078e-01 -1.06661439e+00
3.23818386e-01 2.53177255e-01 -1.58681050e-01 -1.10029316e+00
-4.09386128e-01 -8.00076365e-01 -9.55349624e-01 3.68630402e-02
2.35262409e-01 3.73486290e-03 -5.36395431e-01 1.66124880e+00
1.31444007e-01 4.72570479e-01 2.08007339e-02 5.97961307e-01
3.26543808e-01 4.73565787e-01 -5.60136616e-01 -3.77139777e-01
1.03518021e+00 -8.88416171e-01 -5.60144663e-01 -4.07386094e-01
9.61760581e-01 -6.57594621e-01 8.77443075e-01 5.41362762e-01
-9.39013183e-01 -3.58076274e-01 -1.22367787e+00 1.10951528e-01
-2.69648768e-02 2.30500564e-01 6.88085914e-01 8.89511347e-01
-1.11726892e+00 7.41972744e-01 -8.08922112e-01 -3.51200342e-01
5.33067465e-01 4.44505841e-01 -6.13862813e-01 -4.62191641e-01
-7.84856439e-01 8.42605352e-01 2.54638016e-01 -3.43835980e-01
-8.38468075e-01 -7.14823067e-01 -9.82047856e-01 1.28451094e-01
4.57028329e-01 -4.55009043e-01 9.05815482e-01 -6.38147116e-01
-1.42014837e+00 8.64558518e-01 -2.26996318e-01 -5.69710732e-01
1.52957797e-01 -5.36477864e-02 -3.09866726e-01 2.44645283e-01
-6.22275732e-02 5.08972406e-01 1.04566121e+00 -1.29386616e+00
-2.00161472e-01 -3.00441682e-01 -2.75749445e-01 -6.54868856e-02
-8.16423357e-01 -3.58898640e-01 -2.34606564e-01 -6.56796634e-01
4.18009073e-01 -1.09529245e+00 -3.09992760e-01 2.13589966e-01
-2.49579459e-01 2.14442179e-01 4.21967924e-01 -6.03762269e-01
1.40605080e+00 -2.19852924e+00 3.41768622e-01 3.69497180e-01
3.49460065e-01 7.16230348e-02 -1.40841395e-01 6.74840569e-01
-2.35453025e-01 -7.22881407e-02 -5.75336337e-01 -6.46210849e-01
7.99519941e-02 4.48784888e-01 -2.58416653e-01 9.85491276e-01
-1.74554989e-01 6.28609359e-01 -9.05239463e-01 -2.94056624e-01
6.78528398e-02 6.05145156e-01 -8.16727281e-01 -7.70130288e-03
8.40423687e-04 1.46538213e-01 -1.35496899e-01 6.69717848e-01
4.77928609e-01 -6.22617364e-01 4.20158386e-01 -3.36029202e-01
2.34729990e-01 5.49581349e-02 -1.51175725e+00 2.09189463e+00
-5.01083195e-01 6.09965682e-01 -1.68856827e-03 -1.60493839e+00
8.00363779e-01 3.13602537e-01 8.24315488e-01 -3.99294138e-01
1.11273088e-01 5.50153077e-01 -2.98090607e-01 -3.81376505e-01
2.41803452e-01 -3.28356624e-01 2.79909112e-02 5.75215280e-01
2.93272823e-01 1.03226252e-01 -1.43561596e-02 1.85085982e-01
1.39109135e+00 -3.33484590e-01 3.79551858e-01 -3.55638325e-01
3.74962091e-01 -3.17967534e-01 4.56750333e-01 6.51580870e-01
8.29977468e-02 7.24878669e-01 3.22146356e-01 -3.97388905e-01
-1.13571155e+00 -8.15338314e-01 -3.18657279e-01 9.57554340e-01
-2.73593336e-01 -7.30528593e-01 -4.58386004e-01 -7.18650579e-01
1.08148992e-01 6.99748278e-01 -7.28019178e-01 -7.95387402e-02
-3.86024863e-01 -9.99472618e-01 5.02143204e-01 4.23615247e-01
2.12819174e-01 -5.36881506e-01 -3.46946329e-01 2.89924920e-01
5.32168522e-03 -1.07810915e+00 -3.08304727e-01 5.12164116e-01
-9.38354373e-01 -1.00219154e+00 -5.40540993e-01 -6.55925453e-01
7.47572660e-01 1.63506389e-01 1.01896882e+00 1.76683869e-02
-1.51579484e-01 5.15810728e-01 -4.22998697e-01 3.90033983e-03
-4.74289924e-01 1.24047443e-01 1.87418059e-01 1.66286901e-01
3.02359045e-01 -8.25363636e-01 -4.35287535e-01 -3.23872305e-02
-1.08714139e+00 -2.99864292e-01 7.26364911e-01 1.15983808e+00
7.27023840e-01 -6.09786659e-02 4.06425029e-01 -1.23063231e+00
4.40721840e-01 -6.69027448e-01 -4.40309227e-01 1.08427308e-01
-1.21174955e+00 6.67004228e-01 6.42338514e-01 -2.72763878e-01
-2.39149004e-01 3.51984143e-01 -4.39564407e-01 -6.25715792e-01
2.96469748e-01 5.98842621e-01 2.20033452e-01 -4.07164335e-01
9.15964007e-01 3.82438064e-01 8.49146172e-02 -6.69195116e-01
5.11690199e-01 7.37826467e-01 2.66007841e-01 -4.26672399e-01
9.33258832e-01 7.11971641e-01 1.91168949e-01 -7.79594481e-01
-1.08119571e+00 -6.34171963e-01 -4.07872945e-01 1.96809500e-01
4.23162460e-01 -1.03806615e+00 -4.41744596e-01 -1.19514249e-01
-7.53567159e-01 -1.26131624e-01 -7.66313314e-01 5.59089720e-01
-7.51711607e-01 4.35633093e-01 -5.53264856e-01 -4.51384395e-01
-2.28356093e-01 -1.12323272e+00 1.04812479e+00 -4.04352397e-01
-4.38704016e-03 -1.15229499e+00 3.96111161e-01 1.88462481e-01
6.18767083e-01 4.60040383e-02 7.42851377e-01 -7.16256380e-01
-4.49094564e-01 -1.54826671e-01 -9.32648182e-02 6.44661546e-01
-1.32413670e-01 -5.46116710e-01 -1.04073441e+00 -8.55914056e-01
2.83187568e-01 -4.55460280e-01 1.14648330e+00 1.53756723e-01
1.17834651e+00 -5.49610317e-01 -2.99459249e-01 8.55017304e-01
1.85158980e+00 -3.99249136e-01 6.41222000e-01 1.30794585e-01
6.66904449e-01 1.02455564e-01 1.17960550e-01 6.27870262e-01
2.82360673e-01 7.02033579e-01 4.89360839e-01 1.02063738e-01
-1.89559534e-01 -2.94349670e-01 5.38651466e-01 1.30744588e+00
2.36326739e-01 8.59457031e-02 -5.49328029e-01 6.18397593e-01
-1.67073846e+00 -9.44601774e-01 4.19368967e-02 2.30596304e+00
8.98638606e-01 -3.15077528e-02 -1.29650027e-01 5.59047699e-01
1.58812657e-01 5.05244434e-01 -4.52576309e-01 -3.43397141e-01
1.06239334e-01 6.66421533e-01 8.90429974e-01 6.60584688e-01
-8.83920074e-01 5.48287392e-01 6.75503826e+00 8.42349946e-01
-1.04014969e+00 4.49941486e-01 3.24518204e-01 -2.74500519e-01
-5.06248891e-01 -1.09014045e-02 -7.71074891e-01 4.31162655e-01
1.19174695e+00 1.41399443e-01 6.60927773e-01 1.02257955e+00
-3.39997381e-01 1.89049542e-01 -1.33373582e+00 1.36144626e+00
5.08587301e-01 -1.66133392e+00 -2.32395291e-01 2.04767510e-01
8.36513758e-01 2.07427293e-01 2.44370878e-01 2.54190773e-01
1.79950371e-01 -1.14364350e+00 4.74157929e-01 3.25049758e-01
1.16368639e+00 -4.07768995e-01 5.53292096e-01 4.20003057e-01
-9.72298980e-01 -2.52472252e-01 -6.07414544e-01 5.04433885e-02
1.40334085e-01 1.25651896e+00 -7.74364531e-01 3.90716046e-01
3.50282401e-01 9.64106262e-01 -4.72740114e-01 7.43024111e-01
-1.02138780e-01 8.20855141e-01 -5.17523408e-01 1.63763255e-01
1.65022030e-01 -2.87782382e-02 5.57070374e-01 1.23074567e+00
7.23225951e-01 -9.69123989e-02 1.49633363e-01 4.79850501e-01
-3.23127776e-01 1.16508678e-01 -7.84739316e-01 -2.99784429e-02
4.16123301e-01 9.44631457e-01 -3.99026662e-01 -2.17754409e-01
-6.13542259e-01 1.12604332e+00 4.95528489e-01 8.08912050e-03
-5.88506043e-01 -7.43568540e-02 4.92238432e-01 2.15246826e-01
8.21322441e-01 -3.30470502e-01 -1.48576990e-01 -1.47975147e+00
1.11282848e-01 -1.05938518e+00 2.98781395e-01 -2.71703809e-01
-1.27171504e+00 3.38324398e-01 -3.81591648e-01 -1.12848961e+00
-3.90014529e-01 -6.31549537e-01 8.45447481e-02 4.02552277e-01
-1.66516650e+00 -8.58845174e-01 -2.43970775e-04 4.63897318e-01
2.86174119e-01 -4.09025043e-01 1.19678414e+00 6.21958911e-01
-2.59851068e-01 8.49247575e-01 8.19959283e-01 -6.36623427e-02
5.94375014e-01 -1.14904773e+00 -1.04298770e-01 6.43091857e-01
9.37241375e-01 4.23396230e-01 6.26182377e-01 -9.16930884e-02
-1.74669135e+00 -9.40817118e-01 9.69451904e-01 -3.08961451e-01
6.85185492e-01 -3.80732477e-01 -7.34262943e-01 9.24163818e-01
5.98169155e-02 3.87669355e-01 1.10417056e+00 2.45519772e-01
-6.33726895e-01 -3.42945367e-01 -1.27810526e+00 4.06515040e-02
1.12472510e+00 -6.02795839e-01 -5.23473501e-01 6.56579316e-01
5.64670384e-01 -3.68231237e-01 -8.74614775e-01 3.08390290e-01
5.81543565e-01 -1.18699515e+00 1.24345112e+00 -2.22086981e-01
4.54521805e-01 -7.72391632e-02 -7.19129860e-01 -1.36274564e+00
-4.08259332e-01 -4.85533625e-01 -5.42229116e-01 6.44702315e-01
3.17097515e-01 -6.58994675e-01 9.90742028e-01 3.54775153e-02
-3.94914858e-02 -1.01051342e+00 -1.11716270e+00 -8.22533965e-01
2.64814179e-02 -5.23159027e-01 2.75612235e-01 1.13935947e+00
-9.45423788e-04 3.16618502e-01 -6.85271621e-01 1.77522957e-01
8.07206690e-01 1.14375772e-02 4.42239374e-01 -1.11915803e+00
-8.81193519e-01 -1.03996135e-01 -6.06829047e-01 -1.33696735e+00
1.36889845e-01 -1.29862571e+00 -2.72255629e-01 -1.29670477e+00
1.84078127e-01 -6.70689285e-01 -5.73822796e-01 4.94438559e-01
2.36774459e-01 4.08596396e-01 2.67570794e-01 1.65792674e-01
-5.87940097e-01 5.11265576e-01 7.58052528e-01 -1.96683034e-01
1.27642527e-01 -3.25564653e-01 -1.02359319e+00 5.56935012e-01
5.28442264e-01 -9.67327356e-01 -3.80121142e-01 -4.87662524e-01
4.96677369e-01 4.71169874e-02 1.85163423e-01 -1.38279772e+00
2.54538357e-01 2.38424465e-01 1.02336325e-01 -1.52843758e-01
4.57039624e-01 -1.03078604e+00 1.77514210e-01 5.35194159e-01
-4.42358851e-01 -1.28107950e-01 -1.12154335e-01 7.50578046e-01
-7.95226023e-02 -5.91824055e-01 7.93299139e-01 -1.76295862e-01
-3.85057420e-01 2.97267765e-01 -6.13235161e-02 2.07549721e-01
7.79170692e-01 1.25938386e-01 2.55170427e-02 -4.00036633e-01
-6.98164403e-01 -2.09873468e-01 4.70508397e-01 -6.05862141e-02
6.23899400e-01 -1.38530278e+00 -7.82494009e-01 5.21830261e-01
1.12170026e-01 -2.73792863e-01 -1.10488988e-01 8.22494030e-01
-3.55380565e-01 3.08953494e-01 2.46003658e-01 -5.57102084e-01
-9.09938276e-01 7.06640184e-01 2.02352360e-01 -5.70157647e-01
-6.19748771e-01 1.01226389e+00 -3.25803488e-01 -5.20379186e-01
2.87186712e-01 -2.51292050e-01 2.12393448e-01 8.97991564e-03
4.99682784e-01 3.60214025e-01 1.66948706e-01 -3.76350850e-01
-2.71487653e-01 7.69116759e-01 1.51616130e-02 -1.09417140e-01
1.69333184e+00 3.82264145e-02 1.73829477e-02 5.95937669e-01
1.87668252e+00 3.14103454e-01 -9.65392590e-01 -5.23555458e-01
-1.11080118e-01 -6.52880907e-01 2.08095044e-01 -4.70869035e-01
-1.24134517e+00 9.38245535e-01 7.07227349e-01 1.65717289e-01
9.99302328e-01 6.28988817e-02 8.97766769e-01 6.01686239e-01
5.73782861e-01 -9.75331843e-01 1.60962865e-01 2.61457324e-01
6.58406734e-01 -1.41036189e+00 3.32901001e-01 -2.74761349e-01
-2.84520835e-01 9.92419124e-01 -2.22897485e-01 -2.75481045e-01
9.06900764e-01 4.15322304e-01 -2.28222921e-01 -3.55865598e-01
-6.29651785e-01 5.61749563e-02 8.15070868e-02 4.25299227e-01
2.65094280e-01 -1.56347908e-03 -2.88810670e-01 3.01718682e-01
-1.72367632e-01 6.89113736e-02 4.07900065e-01 9.23247397e-01
-5.63131452e-01 -1.49927032e+00 -2.70785123e-01 6.90847039e-01
-3.78192753e-01 -1.91152036e-01 1.43156633e-01 5.37196100e-01
1.17767505e-01 7.22784400e-01 -3.11397582e-01 -4.98647690e-01
1.26530454e-01 2.30333656e-01 5.70498586e-01 -6.40297651e-01
-2.23746613e-01 -1.61400706e-01 -7.11784586e-02 -8.23944926e-01
-5.32110214e-01 -8.15808654e-01 -1.09947050e+00 -4.01981950e-01
-2.21446872e-01 1.14617832e-01 6.78887069e-01 6.95093691e-01
4.51432317e-01 2.75237501e-01 8.59842420e-01 -6.73549712e-01
-1.03699386e+00 -7.03413904e-01 -7.58660316e-01 4.29024786e-01
4.32194322e-01 -5.78941643e-01 -6.47362411e-01 7.03057200e-02] | [7.29800271987915, 4.427329063415527] |
87dba07e-fc07-464d-b7b8-ffaaab223e1d | efficient-and-interpretable-infrared-and | 2005.05896 | null | https://arxiv.org/abs/2005.05896v2 | https://arxiv.org/pdf/2005.05896v2.pdf | Efficient and Model-Based Infrared and Visible Image Fusion Via Algorithm Unrolling | Infrared and visible image fusion (IVIF) expects to obtain images that retain thermal radiation information from infrared images and texture details from visible images. In this paper, a model-based convolutional neural network (CNN) model, referred to as Algorithm Unrolling Image Fusion (AUIF), is proposed to overcome the shortcomings of traditional CNN-based IVIF models. The proposed AUIF model starts with the iterative formulas of two traditional optimization models, which are established to accomplish two-scale decomposition, i.e., separating low-frequency base information and high-frequency detail information from source images. Then the algorithm unrolling is implemented where each iteration is mapped to a CNN layer and each optimization model is transformed into a trainable neural network. Compared with the general network architectures, the proposed framework combines the model-based prior information and is designed more reasonably. After the unrolling operation, our model contains two decomposers (encoders) and an additional reconstructor (decoder). In the training phase, this network is trained to reconstruct the input image. While in the test phase, the base (or detail) decomposed feature maps of infrared/visible images are merged respectively by an extra fusion layer, and then the decoder outputs the fusion image. Qualitative and quantitative comparisons demonstrate the superiority of our model, which can robustly generate fusion images containing highlight targets and legible details, exceeding the state-of-the-art methods. Furthermore, our network has fewer weights and faster speed. | ['Junmin Liu', 'Chunxia Zhang', 'Chengyang Liang', 'Shuang Xu', 'Jiangshe Zhang', 'Zixiang Zhao'] | 2020-05-12 | null | null | null | null | ['infrared-and-visible-image-fusion'] | ['computer-vision'] | [ 6.09895647e-01 -1.98262855e-01 1.06641725e-01 -8.73730797e-03
-4.38497812e-01 -6.79659322e-02 4.44525033e-01 -2.83368468e-01
-1.96578607e-01 5.00302255e-01 8.89071375e-02 -3.66230130e-01
-5.54841980e-02 -8.95181060e-01 -6.68855786e-01 -1.01721942e+00
5.14508963e-01 -2.07016766e-01 -6.09224774e-02 -2.61862725e-01
-5.97271556e-03 5.25332928e-01 -1.60375118e+00 4.13404018e-01
1.08594561e+00 1.38841891e+00 4.27833974e-01 7.97778845e-01
-1.49564639e-01 9.10967052e-01 -3.60605627e-01 -9.44986492e-02
4.18373078e-01 -4.90089446e-01 -4.24285531e-01 2.53387123e-01
3.83773714e-01 -5.31431913e-01 -4.62437987e-01 1.13012397e+00
4.81390059e-01 1.24375470e-01 5.21734536e-01 -1.00047946e+00
-9.90116715e-01 1.12225093e-01 -7.93987691e-01 -1.73065960e-02
2.60170251e-01 1.79083094e-01 4.93124485e-01 -7.56406605e-01
2.16126263e-01 1.12859023e+00 7.27318466e-01 3.90866041e-01
-1.05219507e+00 -6.39005899e-01 -5.48525825e-02 1.34497613e-01
-1.08084214e+00 -2.87224501e-01 1.05736673e+00 -2.03495756e-01
7.10723460e-01 3.96468103e-01 7.18218327e-01 8.44365001e-01
5.98635614e-01 5.91886938e-01 9.73229945e-01 -5.42644203e-01
5.35340756e-02 -1.38030127e-02 1.51581824e-01 7.52462804e-01
3.28901291e-01 5.73208332e-01 -3.57711375e-01 1.84518605e-01
8.52946579e-01 3.60500723e-01 -5.49910784e-01 1.22861356e-01
-1.12623036e+00 3.98271680e-01 7.25766003e-01 3.27087313e-01
-6.89072847e-01 8.15055668e-02 -2.56389119e-02 3.91544402e-01
5.69167554e-01 -1.88301384e-01 -1.63531318e-01 5.67213833e-01
-1.02518380e+00 -8.72805566e-02 4.90570158e-01 7.18740761e-01
1.13202775e+00 3.23774934e-01 -1.67699412e-01 6.88616157e-01
4.53374982e-01 3.46762061e-01 4.39784795e-01 -6.58752501e-01
4.88374025e-01 7.71980345e-01 7.09562004e-02 -9.51089501e-01
-3.47645968e-01 -6.04692757e-01 -1.40333188e+00 7.52800822e-01
-2.34361053e-01 -2.76760906e-01 -1.26611793e+00 1.27708888e+00
2.89842576e-01 1.65747762e-01 4.37030703e-01 9.75040674e-01
1.09850621e+00 1.00757611e+00 -1.37849435e-01 -3.26869607e-01
1.42047870e+00 -1.19580841e+00 -8.68970931e-01 -2.33537912e-01
2.17902940e-02 -1.11521471e+00 5.84762752e-01 5.30713797e-01
-1.08759058e+00 -1.27477145e+00 -1.26001310e+00 -2.07652435e-01
-2.88533986e-01 6.00816965e-01 4.89417046e-01 4.81247962e-01
-1.25998759e+00 5.17858207e-01 -5.79976499e-01 -1.74802601e-01
1.91226587e-01 3.42422903e-01 -4.98921901e-01 -6.24779724e-02
-9.71137404e-01 8.46739113e-01 7.13041425e-01 7.99806416e-01
-8.37165833e-01 -5.03141820e-01 -8.23309660e-01 1.42280370e-01
1.76476255e-01 -1.04835451e+00 8.93307745e-01 -1.41808498e+00
-1.80803156e+00 3.98138404e-01 -1.02111831e-01 -9.19373035e-02
2.17778176e-01 -1.94326252e-01 -5.61669350e-01 1.54019654e-01
-3.37967753e-01 4.58689779e-01 1.30924213e+00 -1.66270494e+00
-7.89348722e-01 -1.37432829e-01 1.56188667e-01 3.20332736e-01
-2.97575295e-01 -7.01216236e-02 -4.29484308e-01 -7.19197571e-01
4.37164366e-01 -4.70804662e-01 -1.37595087e-02 2.03843549e-01
-3.33084673e-01 9.39414054e-02 1.16781890e+00 -1.06617498e+00
1.16568542e+00 -2.20332837e+00 2.57116377e-01 2.56404020e-02
3.55576575e-01 4.42874432e-01 -1.19343020e-01 3.91552210e-01
-6.22478247e-01 -1.54728547e-01 -3.06024343e-01 -4.17654157e-01
-3.52443188e-01 -7.84252062e-02 -7.93136656e-02 4.98024732e-01
4.20763753e-02 9.00665164e-01 -4.84310329e-01 -3.65195721e-01
6.92928374e-01 7.13164270e-01 -1.84377000e-01 3.64782512e-01
-7.78572168e-03 6.07226968e-01 -2.67515719e-01 6.57517552e-01
1.21811604e+00 -1.15459666e-01 -2.49133222e-02 -9.30330157e-01
-4.38128173e-01 -3.46156329e-01 -1.33575106e+00 1.57353628e+00
-6.83074832e-01 3.66624326e-01 3.74012440e-01 -1.21424043e+00
1.09625304e+00 5.03250778e-01 3.57012093e-01 -7.59673238e-01
3.23846161e-01 4.95675355e-02 -3.49512815e-01 -6.50552392e-01
3.73791128e-01 -2.24497736e-01 4.26939100e-01 3.21727455e-01
1.46436125e-01 1.45352483e-01 -2.41302192e-01 -2.55957365e-01
5.02627909e-01 4.44209129e-01 1.64009780e-01 2.99828313e-02
9.31636810e-01 -1.77243769e-01 4.62644964e-01 4.59602684e-01
2.66879231e-01 6.83694899e-01 -7.19771013e-02 -8.33807528e-01
-9.54057574e-01 -9.98858690e-01 2.20914036e-01 6.65213585e-01
4.67224836e-01 7.81267881e-02 -6.56179726e-01 -2.63043106e-01
-3.42779368e-01 3.25853139e-01 -5.97295702e-01 -1.80699587e-01
-5.20573854e-01 -6.55979991e-01 7.97332674e-02 3.69980395e-01
1.29199314e+00 -1.03443587e+00 -5.55321693e-01 1.88209608e-01
-3.04994613e-01 -8.28661501e-01 -2.31203213e-01 2.12751076e-01
-9.85365152e-01 -1.05137801e+00 -6.24171853e-01 -8.95257413e-01
7.37562835e-01 7.51731873e-01 7.98883140e-01 3.03358495e-01
-2.13677883e-01 -6.41549975e-02 -3.97625118e-01 -3.12380284e-01
-2.90842146e-01 -2.89889187e-01 -4.42658931e-01 5.78983009e-01
-4.02754825e-03 -5.88220656e-01 -7.53838420e-01 -6.68022484e-02
-1.23458767e+00 6.81075811e-01 1.04117298e+00 9.21264052e-01
4.70097303e-01 4.32534993e-01 1.52377803e-02 -4.03414220e-01
3.38298440e-01 -1.13269366e-01 -6.95428252e-01 3.98566276e-01
-4.51915532e-01 4.92885597e-02 8.69514465e-01 -2.12325931e-01
-1.55645430e+00 2.33778849e-01 -4.06737626e-02 -5.99584877e-01
-2.05800846e-01 5.04890859e-01 -1.47761524e-01 -3.88079464e-01
5.86379051e-01 5.92870295e-01 1.85948610e-02 -6.28099024e-01
1.92901492e-01 7.69713342e-01 7.96363533e-01 -2.26571918e-01
1.17224824e+00 5.69209456e-01 -1.79115176e-01 -6.39294624e-01
-6.81939721e-01 -3.18618119e-01 -5.81387877e-01 -5.27271867e-01
1.19896674e+00 -1.09768629e+00 -7.82364726e-01 6.24934137e-01
-1.38344610e+00 -3.73990051e-02 -6.07009465e-03 6.77494586e-01
-3.85409474e-01 3.43576819e-01 -5.67697108e-01 -8.36983383e-01
-6.06019318e-01 -1.06984341e+00 1.17571473e+00 6.44738197e-01
5.13466895e-01 -9.10011292e-01 -1.40941098e-01 4.40052897e-01
5.54392338e-01 5.23333371e-01 7.50109613e-01 1.54849023e-01
-6.54715538e-01 -1.11004680e-01 -4.11364526e-01 8.35300863e-01
2.30781227e-01 6.56490996e-02 -1.06535983e+00 -2.50490427e-01
5.10118663e-01 -1.90279052e-01 1.17936242e+00 4.31718379e-01
1.16706097e+00 -4.43946451e-01 -2.22795248e-01 1.02938735e+00
1.93084073e+00 3.95360619e-01 9.99896765e-01 4.04574692e-01
7.07030773e-01 3.20891529e-01 2.11255670e-01 8.50683972e-02
1.63097437e-02 4.45603520e-01 5.85614502e-01 -9.78248537e-01
-3.46771330e-01 6.96894228e-02 3.97444427e-01 8.28244328e-01
-5.11948824e-01 -2.60884643e-01 -5.51944554e-01 7.96248466e-02
-1.85472846e+00 -1.05327034e+00 -1.09543890e-01 2.10593534e+00
6.01200819e-01 -8.84769559e-02 -3.82375062e-01 2.91031361e-01
7.86807716e-01 3.33202362e-01 -2.68004477e-01 -2.68836230e-01
-1.74967170e-01 3.67291749e-01 4.77553666e-01 6.80719614e-01
-1.03919542e+00 5.72481811e-01 6.15754366e+00 7.64838815e-01
-1.34076226e+00 9.35590193e-02 5.76169550e-01 3.91239166e-01
-2.88758546e-01 1.53267488e-01 -3.65434587e-01 2.74327785e-01
5.88638127e-01 2.72548020e-01 7.07414567e-01 4.09325272e-01
3.45225215e-01 -1.59624904e-01 -7.57760108e-01 1.06058359e+00
1.93527833e-01 -1.29524970e+00 2.50993073e-01 -2.42589846e-01
7.18490362e-01 -3.26719284e-01 8.71820562e-03 -2.78676730e-02
9.19205546e-02 -1.06946814e+00 7.79785633e-01 9.95672941e-01
8.16593409e-01 -7.93574929e-01 9.55180764e-01 4.65993941e-01
-1.52806056e+00 -3.60498846e-01 -4.12032664e-01 -4.10971195e-02
-2.84845214e-02 6.59240723e-01 -2.93998271e-02 1.46905732e+00
6.81049943e-01 6.32869780e-01 -4.42732871e-01 6.83987975e-01
-1.19394220e-01 2.24094257e-01 -5.19114845e-02 3.74999911e-01
1.85091078e-01 -5.47318399e-01 1.81960821e-01 9.42401528e-01
3.53109390e-01 1.48960128e-01 1.78258017e-01 9.64277506e-01
1.47556990e-01 -3.31443220e-01 -4.50024635e-01 2.97118425e-01
3.62191722e-02 1.62514555e+00 -5.78345358e-01 -5.77907145e-01
-6.23371065e-01 1.20610499e+00 -1.01243788e-02 7.29171276e-01
-9.97775316e-01 -5.17204940e-01 1.17372207e-01 -2.68352538e-01
3.54204446e-01 -2.03690156e-01 -3.45690161e-01 -1.17075133e+00
8.22294652e-02 -8.35485697e-01 7.46715292e-02 -1.28068161e+00
-9.90329027e-01 7.57937372e-01 -1.57728419e-01 -1.51636255e+00
2.67569125e-01 -6.06802583e-01 -7.89453685e-01 1.28741205e+00
-1.95011008e+00 -1.63022006e+00 -1.04235554e+00 7.84090400e-01
5.56823552e-01 -1.19219525e-02 5.89502633e-01 2.89467990e-01
-6.68560326e-01 1.54630706e-01 2.14888766e-01 -2.23595351e-02
3.47718954e-01 -8.25396240e-01 -1.19279049e-01 1.00428593e+00
-2.10662469e-01 5.01603186e-01 4.76871610e-01 -5.48017740e-01
-1.51337218e+00 -1.19052315e+00 5.90711534e-01 2.14100048e-01
9.32278633e-02 -1.23414926e-01 -7.00280845e-01 5.65631747e-01
5.42261004e-01 5.03052659e-02 3.30514491e-01 -5.49288929e-01
-2.24389181e-01 -5.86065412e-01 -1.03280091e+00 3.57371062e-01
6.69073045e-01 -4.95444834e-01 -5.48745930e-01 3.07707071e-01
6.78296328e-01 -4.97206688e-01 -7.58941352e-01 5.07094502e-01
6.61335528e-01 -1.28349626e+00 1.10998261e+00 -4.71395813e-02
6.61233962e-01 -7.88757503e-01 -8.85292049e-03 -1.03934073e+00
-7.20279932e-01 -5.20112216e-01 -4.05803770e-02 1.00265265e+00
2.04135656e-01 -6.37055576e-01 3.63496482e-01 9.63506326e-02
-3.78423423e-01 -6.22130752e-01 -7.36510336e-01 -4.58360136e-01
-4.76051897e-01 7.26201385e-02 5.02368689e-01 7.54994929e-01
-5.89035332e-01 2.66032636e-01 -4.90064651e-01 4.51222122e-01
7.09076226e-01 1.15012497e-01 5.06498814e-01 -1.04318142e+00
-3.33508343e-01 -2.80177861e-01 -9.14945304e-02 -8.53059351e-01
-1.57800362e-01 -6.31967485e-01 4.27996144e-02 -1.87551916e+00
2.40512252e-01 1.91333648e-02 -3.90958875e-01 6.05419695e-01
-3.84564012e-01 3.14295292e-01 1.38451800e-01 3.00141990e-01
-1.35472193e-01 6.38039827e-01 1.42069292e+00 -2.68743753e-01
-1.68946072e-01 -1.82224661e-01 -7.43152678e-01 5.23467779e-01
7.28641331e-01 -8.89118165e-02 -3.90056759e-01 -6.62571728e-01
3.86264995e-02 3.26647252e-01 7.03682661e-01 -1.34912443e+00
5.46264291e-01 -1.27118662e-01 8.45095873e-01 -6.95425332e-01
1.73597664e-01 -1.11951888e+00 7.14800417e-01 6.16799712e-01
-4.33248654e-02 -5.85357994e-02 1.33793652e-01 4.95723605e-01
-3.98445874e-01 -3.42740789e-02 9.11778390e-01 -1.72886238e-01
-5.27264953e-01 3.30269396e-01 -1.95287094e-01 -9.55823243e-01
9.70010817e-01 -6.27230167e-01 -4.65049565e-01 -1.28000602e-01
-6.02775753e-01 -2.01535821e-01 3.17706257e-01 1.80841327e-01
7.98638165e-01 -1.46188116e+00 -7.98124671e-01 4.41307545e-01
-2.19504401e-01 1.96160212e-01 6.67440057e-01 9.00169849e-01
-8.72100353e-01 7.04856738e-02 -2.97729075e-01 -4.58662122e-01
-1.13556659e+00 8.04650843e-01 4.75452274e-01 -2.46015847e-01
-7.51924694e-01 5.25183022e-01 3.32975417e-01 -2.59381533e-01
1.51289374e-01 -4.57736880e-01 -3.55335951e-01 -3.29051405e-01
6.42479062e-01 2.36078724e-01 1.33559316e-01 -5.53459227e-01
4.17520143e-02 8.82340968e-01 1.76547796e-01 5.58511391e-02
1.43995154e+00 -1.56726390e-01 -4.39651340e-01 4.99823466e-02
1.31716692e+00 -3.55076969e-01 -1.27169991e+00 -1.76228061e-01
-6.82006955e-01 -1.82206333e-01 4.77726698e-01 -8.14006090e-01
-1.35856473e+00 8.33138764e-01 8.50735784e-01 7.61129409e-02
1.84953892e+00 -5.29170871e-01 7.90265501e-01 1.97296068e-01
-1.25482202e-01 -1.02449489e+00 1.42518550e-01 1.74877435e-01
9.98384893e-01 -9.29976583e-01 1.68526337e-01 -3.22619945e-01
-1.46782696e-01 1.58704615e+00 6.10915899e-01 -7.10829571e-02
5.81612945e-01 3.40508282e-01 1.45370904e-02 -1.45061374e-01
-5.17679751e-01 -1.34349048e-01 2.85333693e-01 6.13158226e-01
2.39880100e-01 -2.99026996e-01 -1.39371231e-01 3.29319149e-01
1.18371971e-01 3.62569183e-01 1.20370671e-01 9.17511404e-01
-7.33892381e-01 -6.84270799e-01 -8.29274058e-01 2.26582259e-01
-1.67578176e-01 -1.91475123e-01 -3.18429880e-02 4.37871993e-01
5.53747773e-01 1.06587005e+00 1.38823316e-02 -7.93103814e-01
1.43615440e-01 -1.78218633e-01 3.09786767e-01 -1.89920053e-01
-6.12610400e-01 1.99931458e-01 -2.20215276e-01 -5.74490309e-01
-8.41400027e-01 5.60609102e-02 -8.46695483e-01 -4.64362562e-01
-5.89521646e-01 -8.97834226e-02 7.65407026e-01 9.19606209e-01
2.43365496e-01 1.02987838e+00 9.37321723e-01 -1.12159896e+00
-2.56629467e-01 -1.02661335e+00 -4.37823921e-01 9.73956957e-02
6.85358763e-01 -3.51174355e-01 -4.50347692e-01 4.65366036e-01] | [10.547784805297852, -2.003321886062622] |
441c48d5-088c-4c69-9a0e-c79999fa073c | making-images-undiscoverable-from-co-saliency | 2009.09258 | null | https://arxiv.org/abs/2009.09258v5 | https://arxiv.org/pdf/2009.09258v5.pdf | Can You Spot the Chameleon? Adversarially Camouflaging Images from Co-Salient Object Detection | Co-salient object detection (CoSOD) has recently achieved significant progress and played a key role in retrieval-related tasks. However, it inevitably poses an entirely new safety and security issue, i.e., highly personal and sensitive content can potentially be extracting by powerful CoSOD methods. In this paper, we address this problem from the perspective of adversarial attacks and identify a novel task: adversarial co-saliency attack. Specially, given an image selected from a group of images containing some common and salient objects, we aim to generate an adversarial version that can mislead CoSOD methods to predict incorrect co-salient regions. Note that, compared with general white-box adversarial attacks for classification, this new task faces two additional challenges: (1) low success rate due to the diverse appearance of images in the group; (2) low transferability across CoSOD methods due to the considerable difference between CoSOD pipelines. To address these challenges, we propose the very first black-box joint adversarial exposure and noise attack (Jadena), where we jointly and locally tune the exposure and additive perturbations of the image according to a newly designed high-feature-level contrast-sensitive loss function. Our method, without any information on the state-of-the-art CoSOD methods, leads to significant performance degradation on various co-saliency detection datasets and makes the co-salient objects undetectable. This can have strong practical benefits in properly securing the large number of personal photos currently shared on the Internet. Moreover, our method is potential to be utilized as a metric for evaluating the robustness of CoSOD methods. | ['Song Wang', 'Yang Liu', 'Huazhu Fu', 'Felix Juefei-Xu', 'Hongkai Yu', 'Qing Guo', 'Wei Feng', 'Ruijun Gao'] | 2020-09-19 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Gao_Can_You_Spot_the_Chameleon_Adversarially_Camouflaging_Images_From_Co-Salient_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Gao_Can_You_Spot_the_Chameleon_Adversarially_Camouflaging_Images_From_Co-Salient_CVPR_2022_paper.pdf | cvpr-2022-1 | ['co-saliency-detection'] | ['computer-vision'] | [ 5.61984718e-01 -1.96003392e-01 3.12911242e-01 1.62409917e-01
-9.51466084e-01 -8.70246589e-01 7.07871079e-01 9.79480073e-02
-2.53550142e-01 4.59516436e-01 -5.98733611e-02 -3.75941023e-03
1.03838496e-01 -5.54571390e-01 -7.02260196e-01 -6.84133351e-01
1.70783594e-01 -1.29611611e-01 5.24092674e-01 -4.54619169e-01
5.56780584e-02 7.49549210e-01 -1.39169741e+00 1.65845051e-01
9.55621719e-01 1.04443359e+00 1.34172022e-01 4.17454302e-01
4.79990989e-01 5.12074411e-01 -9.15557504e-01 -8.49248350e-01
5.75378239e-01 -4.16295767e-01 -5.56437492e-01 -1.10315293e-01
8.55958700e-01 -2.76225030e-01 -4.88958746e-01 1.58586574e+00
8.72334838e-01 -6.62168115e-02 4.06475782e-01 -1.50151730e+00
-6.72878921e-01 1.91841707e-01 -6.98116124e-01 3.34597707e-01
2.86191076e-01 4.10302311e-01 7.12101996e-01 -8.47973108e-01
4.58988607e-01 1.34919357e+00 4.44258302e-01 6.98678732e-01
-1.30485129e+00 -8.31276119e-01 2.21616909e-01 2.72709578e-01
-1.32015669e+00 -2.87325650e-01 9.32231009e-01 -2.50195891e-01
2.69700527e-01 7.05213904e-01 3.55487615e-01 1.44481695e+00
9.23399478e-02 7.83616900e-01 1.10588109e+00 -2.04971969e-01
2.61398464e-01 2.73348033e-01 -3.97045672e-01 3.71289879e-01
2.42284566e-01 1.99732289e-01 -3.56678247e-01 -3.02691042e-01
4.23363596e-01 1.69000149e-01 -4.38801736e-01 -4.33326691e-01
-1.22420371e+00 6.23195767e-01 8.64471793e-01 8.92234519e-02
-1.39206693e-01 8.75051543e-02 3.41179460e-01 1.96687222e-01
2.70515978e-01 7.43376672e-01 -6.20219223e-02 3.81460279e-01
-8.32363725e-01 4.98888373e-01 3.01952809e-01 6.48228765e-01
4.10145849e-01 1.09765999e-01 -4.69569832e-01 5.64267874e-01
-1.75634623e-01 7.06144452e-01 3.27784568e-01 -4.24092501e-01
3.48360509e-01 3.08152765e-01 1.43796697e-01 -1.52435136e+00
-8.79634842e-02 -5.10530889e-01 -8.03022504e-01 5.50569773e-01
3.94300640e-01 -7.77170733e-02 -8.30603898e-01 1.92941737e+00
4.32691485e-01 2.22838074e-01 -2.99035627e-02 1.26356947e+00
5.69697976e-01 4.06473070e-01 2.23668501e-01 -3.28910351e-02
1.49320185e+00 -9.47216809e-01 -6.81825101e-01 -5.44158578e-01
-1.01549581e-01 -1.00439990e+00 1.15197623e+00 2.54454404e-01
-1.08022678e+00 -5.16411304e-01 -1.28153110e+00 1.15281627e-01
-5.70940971e-01 -1.63471565e-01 4.32065517e-01 6.54869020e-01
-7.53114581e-01 3.75127673e-01 -4.06573445e-01 -1.48820385e-01
6.74434662e-01 2.84943074e-01 -1.88055307e-01 -1.20062090e-01
-1.39819479e+00 8.83288920e-01 1.23975202e-01 -6.10776059e-02
-1.14044213e+00 -7.51271188e-01 -7.00601280e-01 6.78245723e-02
5.99234879e-01 -5.53358197e-01 8.10171008e-01 -1.31685603e+00
-9.90062296e-01 9.34682727e-01 3.40117306e-01 -4.76494402e-01
9.12487268e-01 -4.71355498e-01 -6.33688509e-01 3.81874770e-01
4.61226404e-02 5.61050951e-01 1.59105802e+00 -1.30740273e+00
-3.34922343e-01 -4.29284126e-01 6.12668544e-02 2.31112331e-01
-5.84209621e-01 3.27866465e-01 -6.10539794e-01 -1.25485384e+00
-2.93676347e-01 -1.01734972e+00 -2.13557929e-01 4.55506474e-01
-6.53268993e-01 2.63805598e-01 1.09859967e+00 -5.66240609e-01
1.03987718e+00 -2.53867149e+00 2.28315771e-01 3.01399715e-02
2.48763755e-01 8.05513263e-01 -2.59136319e-01 2.26736769e-01
-1.92234516e-01 2.24682406e-01 -3.14358592e-01 -2.27991790e-01
-1.19737178e-01 -4.37138528e-01 -7.92053938e-01 5.48133850e-01
4.99857932e-01 9.70123470e-01 -1.07464325e+00 -2.63636619e-01
1.91033944e-01 5.96117496e-01 -2.57565200e-01 2.53267467e-01
-2.06098333e-01 2.64295012e-01 -3.50479364e-01 6.28651798e-01
1.05750155e+00 -1.45119391e-02 -2.15496749e-01 -2.80080765e-01
3.23024005e-01 -2.36620665e-01 -1.05215347e+00 1.33171463e+00
-1.07835837e-01 4.88919020e-01 7.27205426e-02 -7.30039120e-01
6.34043574e-01 3.20673175e-02 4.56750132e-02 -6.75725520e-01
-2.00071558e-02 2.15987816e-01 -1.96187869e-01 -2.74846196e-01
4.50914383e-01 -1.73419737e-03 -3.56871784e-01 1.32244304e-01
-4.98574376e-02 -9.69982818e-02 -2.88666189e-01 4.18731600e-01
1.18906748e+00 -2.15452567e-01 1.49216443e-01 -1.78512275e-01
5.74374974e-01 -2.19905749e-01 6.63479567e-01 9.20307934e-01
-4.85347986e-01 1.04256284e+00 4.54809904e-01 -2.07575232e-01
-9.37543035e-01 -1.28166580e+00 -9.60246921e-02 8.91399324e-01
6.93543673e-01 -1.56672239e-01 -8.59786928e-01 -1.06359804e+00
1.87554196e-01 4.25412804e-01 -6.31702840e-01 -6.22926116e-01
-4.86538738e-01 -6.68535173e-01 6.79712296e-01 1.28707439e-01
6.40573561e-01 -1.13905227e+00 -4.01593983e-01 -8.42914730e-02
-9.37251374e-02 -1.34597898e+00 -9.31731164e-01 -2.08401620e-01
-2.04212874e-01 -1.01339781e+00 -9.24424350e-01 -7.35455155e-01
7.42228627e-01 7.30583787e-01 9.69991863e-01 1.86356127e-01
-5.45691490e-01 1.60137758e-01 -3.47061008e-01 -6.63423300e-01
-3.56725216e-01 -2.75929540e-01 7.45266750e-02 4.81350183e-01
1.42928466e-01 -3.86927456e-01 -9.05708373e-01 6.09648168e-01
-1.27751160e+00 -1.48584768e-01 5.01268864e-01 8.43814015e-01
4.47226584e-01 2.46480465e-01 7.35554755e-01 -8.59131157e-01
4.75814223e-01 -4.35707152e-01 -6.40176415e-01 2.44572923e-01
-4.15416390e-01 -3.75777870e-01 8.24697196e-01 -5.83066702e-01
-8.06905568e-01 -8.98155198e-02 1.02272637e-01 -7.57631421e-01
8.35485682e-02 1.07703777e-02 -5.62819898e-01 -3.89762014e-01
8.73096406e-01 1.81128412e-01 3.92392930e-03 -3.09678465e-01
3.47631872e-01 4.28812504e-01 7.93430686e-01 -2.10533112e-01
1.44393694e+00 6.86473787e-01 -2.06065644e-03 -5.98345816e-01
-9.28344011e-01 -2.22000718e-01 -5.25696166e-02 -2.67928075e-02
8.10699761e-01 -9.69290853e-01 -4.29311126e-01 8.65826547e-01
-9.91259754e-01 -4.83355932e-02 -2.43806109e-01 -1.71113983e-02
-2.31388360e-01 6.79952443e-01 -3.85655552e-01 -6.36981070e-01
-5.76989055e-01 -1.22871196e+00 1.04881930e+00 2.18824014e-01
2.67800801e-02 -5.62417150e-01 -4.00292724e-01 4.72598076e-01
4.77039546e-01 5.53565204e-01 6.73867524e-01 -5.88684499e-01
-6.75966740e-01 -4.15179402e-01 -3.25263321e-01 5.60383976e-01
5.14589027e-02 -1.68368593e-01 -1.07826340e+00 -7.50893652e-01
4.56497446e-02 -4.59517181e-01 7.89476931e-01 -3.88351157e-02
1.42669857e+00 -4.25182313e-01 -2.04481453e-01 5.92512608e-01
1.34786010e+00 -1.12076104e-01 7.45032310e-01 3.13757658e-01
6.11423731e-01 5.35652697e-01 8.07444096e-01 1.51629955e-01
-2.46036842e-01 1.00430059e+00 7.71785557e-01 -4.71884578e-01
-3.87988061e-01 -4.11051184e-01 5.19018173e-01 -6.94957096e-03
4.39190745e-01 -4.29013044e-01 -5.19651175e-01 6.02698863e-01
-1.81781173e+00 -1.02308750e+00 1.63814306e-01 2.36990619e+00
7.85691440e-01 2.50583857e-01 1.96048245e-01 1.36693016e-01
1.08692324e+00 4.79618758e-01 -6.95026815e-01 -7.31889009e-02
-4.26555187e-01 1.54667377e-01 4.44143414e-01 1.27474675e-02
-1.42344284e+00 8.98644149e-01 5.41670799e+00 1.26354921e+00
-1.24053967e+00 1.43553287e-01 6.28585577e-01 -2.10725829e-01
-2.92641789e-01 -1.71832412e-01 -6.75633550e-01 8.30763638e-01
3.50375175e-01 -2.43072331e-01 4.92887557e-01 8.09543610e-01
-6.96048290e-02 7.98551589e-02 -8.17488253e-01 1.02384591e+00
3.78898978e-01 -1.08019412e+00 2.13521332e-01 -2.80594174e-02
8.01173329e-01 -2.47248590e-01 6.90685272e-01 3.76862697e-02
1.13156401e-01 -9.47631299e-01 8.30882728e-01 4.06872295e-02
9.73904252e-01 -9.02287543e-01 5.36498010e-01 1.67643979e-01
-7.36818671e-01 -1.71764717e-01 -3.67045850e-01 3.83191466e-01
8.35442618e-02 7.05949605e-01 -4.57641691e-01 3.72036397e-01
7.80906022e-01 3.38275701e-01 -7.82225609e-01 1.00651324e+00
-3.86490464e-01 3.88546675e-01 -1.86684951e-01 1.34758994e-01
1.25214547e-01 1.90302640e-01 1.01263905e+00 1.06288397e+00
4.27692942e-02 -3.05303663e-01 -1.04662636e-02 9.77181733e-01
-3.52282792e-01 5.58720389e-03 -5.75903356e-01 1.39928132e-01
5.43515325e-01 1.35246408e+00 -5.00013649e-01 -4.14824896e-02
-2.73306668e-01 1.47993469e+00 1.65429652e-01 4.06572282e-01
-9.73162174e-01 -3.92703682e-01 8.26839566e-01 2.43738696e-01
4.20229644e-01 1.98527634e-01 -1.29693359e-01 -1.27379906e+00
3.87825668e-01 -1.29615915e+00 4.01982695e-01 -6.67512119e-01
-1.45192003e+00 6.50208890e-01 -2.57100731e-01 -1.45851731e+00
2.75413990e-01 -2.65739113e-01 -6.65083349e-01 9.28137660e-01
-1.61766768e+00 -1.25056100e+00 -4.86316621e-01 6.85852945e-01
4.39719617e-01 -2.76385486e-01 7.02613890e-01 2.63931751e-01
-5.81666887e-01 1.05805576e+00 2.20153555e-02 1.77785903e-01
9.94311690e-01 -1.15994239e+00 6.45969689e-01 1.38688338e+00
1.40073404e-01 4.38095838e-01 7.36271560e-01 -5.29657602e-01
-1.24650824e+00 -1.39850688e+00 4.40248668e-01 -6.08873427e-01
5.98533094e-01 -6.91003382e-01 -9.05550241e-01 3.12988430e-01
1.24865882e-02 4.01359767e-01 4.34206963e-01 -3.93451452e-01
-4.77788270e-01 -2.93691963e-01 -1.43400598e+00 9.58158255e-01
9.41544116e-01 -7.14319885e-01 -3.88491333e-01 3.98192048e-01
8.92770708e-01 -3.43141377e-01 -3.71850014e-01 5.99542320e-01
1.85621962e-01 -9.80415106e-01 1.36144662e+00 -4.68152255e-01
4.36012983e-01 -5.20192623e-01 1.09474147e-02 -1.26425374e+00
-3.07317734e-01 -1.03530777e+00 -1.69287086e-01 1.39946651e+00
6.62464947e-02 -5.29465675e-01 5.72133243e-01 4.13103640e-01
2.11170927e-01 -5.46829045e-01 -9.10026968e-01 -7.97933459e-01
-1.11397140e-01 -7.78503204e-03 5.67828715e-01 8.59111726e-01
-3.61634344e-01 4.71419841e-02 -6.74827874e-01 5.06683886e-01
8.12916279e-01 1.41389975e-02 8.77613187e-01 -7.86730945e-01
-2.83543229e-01 -1.85252130e-01 -7.04181135e-01 -8.38748276e-01
-8.76872241e-02 -6.12834156e-01 -7.90533945e-02 -6.92258775e-01
2.50459164e-01 -4.67420220e-01 -5.82358718e-01 2.95687646e-01
-7.72775412e-01 7.04260170e-01 5.49926519e-01 2.73480088e-01
-5.94217300e-01 5.59137106e-01 1.15706444e+00 -3.56108904e-01
1.15504153e-01 3.28370444e-02 -9.79198873e-01 5.65133154e-01
6.10466480e-01 -5.18146813e-01 -4.06974256e-01 -1.78438395e-01
1.39806658e-01 -3.08194458e-01 8.80647719e-01 -1.07605970e+00
2.01565176e-02 -5.10333106e-02 4.07446831e-01 -3.39762479e-01
3.83387923e-01 -9.31360424e-01 -3.05501763e-02 4.00427163e-01
-2.06889749e-01 -1.09341122e-01 3.26229841e-01 8.66910875e-01
-2.72211879e-01 -1.39439523e-01 1.20473421e+00 9.88165289e-03
-7.53480732e-01 2.28697971e-01 -6.85164034e-02 2.13694617e-01
1.35262644e+00 -4.94416766e-02 -5.44228673e-01 -4.12229955e-01
-3.75351191e-01 5.22051528e-02 7.42819667e-01 6.60010338e-01
6.94719851e-01 -1.33146417e+00 -7.28676200e-01 3.54595184e-01
2.82665253e-01 -1.02091782e-01 4.92073268e-01 4.72831279e-01
-3.44350755e-01 -1.37471139e-01 -1.84650570e-01 -3.59309137e-01
-1.35085201e+00 1.15290034e+00 2.86836326e-01 -1.41785756e-01
-4.90430236e-01 8.73602867e-01 6.22349679e-01 -9.67545377e-04
2.01924562e-01 5.38705826e-01 1.91966388e-02 -6.62812442e-02
7.39151716e-01 1.31194562e-01 1.08369835e-01 -6.73951566e-01
-4.41426277e-01 2.61751622e-01 -4.16782320e-01 2.33179033e-01
1.00728714e+00 -9.79618449e-03 7.00481161e-02 -1.52662054e-01
1.19353044e+00 2.98499435e-01 -1.28846848e+00 -3.27093571e-01
-3.60512048e-01 -8.58113825e-01 5.80440685e-02 -8.81230235e-01
-1.15564263e+00 6.99680150e-01 8.34529042e-01 2.05222249e-01
1.29707086e+00 -6.78530931e-02 1.00239205e+00 -2.11226746e-01
1.91993520e-01 -9.66987967e-01 4.35336947e-01 -1.45632416e-01
1.19740593e+00 -1.31811547e+00 1.62964299e-01 -6.95982456e-01
-6.90440893e-01 5.04139185e-01 6.53097272e-01 -2.01485261e-01
4.28477287e-01 2.10043266e-02 1.31990910e-01 -1.51598573e-01
-3.18949848e-01 -1.88851133e-02 4.40918565e-01 6.98952973e-01
-3.03337038e-01 -4.67577390e-02 -1.54000238e-01 3.39270443e-01
2.27818880e-02 -3.69613409e-01 4.06941563e-01 6.95874453e-01
-2.31765732e-02 -8.84013832e-01 -5.99623561e-01 2.19230056e-01
-9.33317304e-01 -1.83320060e-01 -5.24578452e-01 4.76720691e-01
4.96457517e-02 8.95532310e-01 -1.81561917e-01 -4.42630768e-01
4.08102959e-01 -4.38874364e-01 5.56432493e-02 -3.94818664e-01
-7.39011705e-01 -3.30722816e-02 -3.13627124e-01 -6.49613976e-01
-6.60317019e-02 -6.54021800e-01 -6.41134679e-01 -1.86420843e-01
-3.79238755e-01 -3.99522185e-02 4.90163773e-01 4.94800985e-01
5.36362708e-01 3.58277202e-01 1.01729882e+00 -8.64093542e-01
-9.57454622e-01 -5.44401944e-01 -4.67911422e-01 1.06246495e+00
5.00578701e-01 -5.73739469e-01 -6.38570786e-01 -1.50264382e-01] | [5.63074254989624, 7.8706793785095215] |
854ea9f2-3685-4e1e-8e88-967d90be8f35 | two-sample-testing-in-reinforcement-learning | 2201.08078 | null | https://arxiv.org/abs/2201.08078v2 | https://arxiv.org/pdf/2201.08078v2.pdf | Two-Sample Testing in Reinforcement Learning | Value-based reinforcement-learning algorithms have shown strong performances in games, robotics, and other real-world applications. The most popular sample-based method is $Q$-Learning. It subsequently performs updates by adjusting the current $Q$-estimate towards the observed reward and the maximum of the $Q$-estimates of the next state. The procedure introduces maximization bias with approaches like Double $Q$-Learning. We frame the bias problem statistically and consider it an instance of estimating the maximum expected value (MEV) of a set of random variables. We propose the $T$-Estimator (TE) based on two-sample testing for the mean, that flexibly interpolates between over- and underestimation by adjusting the significance level of the underlying hypothesis tests. A generalization, termed $K$-Estimator (KE), obeys the same bias and variance bounds as the TE while relying on a nearly arbitrary kernel function. We introduce modifications of $Q$-Learning and the Bootstrapped Deep $Q$-Network (BDQN) using the TE and the KE. Furthermore, we propose an adaptive variant of the TE-based BDQN that dynamically adjusts the significance level to minimize the absolute estimation bias. All proposed estimators and algorithms are thoroughly tested and validated on diverse tasks and environments, illustrating the bias control and performance potential of the TE and KE. | ['Ostap Okhrin', 'Martin Waltz'] | 2022-01-20 | null | null | null | null | ['hypothesis-testing', 'hypothesis-testing'] | ['methodology', 'miscellaneous'] | [-2.47477144e-01 2.75541663e-01 -2.65415668e-01 -3.63370389e-01
-7.42401123e-01 -2.43120566e-01 2.94531971e-01 7.65931457e-02
-1.09783828e+00 1.23077619e+00 -6.64677143e-01 -2.55332291e-01
-6.19430482e-01 -1.00179136e+00 -9.75386977e-01 -8.27115893e-01
-5.20876348e-01 4.77058828e-01 3.63587767e-01 -2.34443501e-01
6.58219159e-01 9.84511077e-02 -1.47016931e+00 -5.49308598e-01
1.03620946e+00 1.44483662e+00 1.75038770e-01 4.51144934e-01
5.55316247e-02 4.57121998e-01 -8.40338469e-01 -3.94823074e-01
4.70235050e-01 -4.87044841e-01 -3.61727059e-01 -3.10503930e-01
-1.43240109e-01 -3.46593022e-01 1.08512394e-01 1.20870912e+00
6.70185804e-01 3.78536612e-01 5.33666730e-01 -1.49158621e+00
-1.95516914e-01 6.81158543e-01 -5.90968728e-01 5.54528087e-02
-1.90469176e-02 5.42656362e-01 7.09865570e-01 -5.19896567e-01
3.43153208e-01 1.01551390e+00 7.12004840e-01 5.51151812e-01
-1.08365393e+00 -8.63947988e-01 1.27242550e-01 2.18946055e-01
-1.32900429e+00 -2.11343076e-02 5.86340427e-01 -3.45358640e-01
8.67036939e-01 -3.47104996e-01 9.17289555e-01 7.08321452e-01
4.51507062e-01 6.26536846e-01 1.31737792e+00 -3.43791068e-01
1.00607872e+00 2.05488831e-01 -3.75472814e-01 8.49185467e-01
2.47426152e-01 6.09756351e-01 -4.38048095e-01 -8.80359858e-02
8.00504565e-01 -4.31973219e-01 -9.52041242e-03 -9.34007108e-01
-8.18129063e-01 9.46553826e-01 3.16574484e-01 -1.19031250e-01
-5.80605030e-01 7.21152723e-01 5.02787828e-01 2.60219485e-01
2.96546400e-01 5.11347890e-01 -7.05047846e-01 -3.57713610e-01
-8.59388292e-01 6.71476007e-01 5.97766340e-01 7.03748226e-01
8.94724369e-01 4.58289236e-01 -2.24838763e-01 7.12500215e-01
2.18244910e-01 5.98598063e-01 6.82468712e-01 -1.17363131e+00
1.65098488e-01 3.05955023e-01 6.52351081e-01 -6.42446816e-01
-2.96446562e-01 -5.72951496e-01 -4.28394228e-01 5.84727526e-01
5.11906087e-01 -5.28985322e-01 -5.35142183e-01 2.16742301e+00
4.28318977e-01 1.47208080e-01 -1.82381913e-01 6.81121349e-01
9.43317637e-02 1.87645793e-01 1.70142680e-01 -4.39484835e-01
6.54627085e-01 -3.94558191e-01 -5.54875016e-01 -2.80094147e-01
4.14023280e-01 1.46837130e-01 1.07570159e+00 6.18750870e-01
-1.21867251e+00 -6.23256803e-01 -1.12611675e+00 8.11763167e-01
-3.95078629e-01 -2.53437787e-01 4.20189112e-01 7.78885603e-01
-1.07203269e+00 9.65481281e-01 -6.03592753e-01 -8.85068066e-03
3.09273124e-01 6.27694547e-01 1.06831692e-01 3.94301265e-01
-1.43885624e+00 9.05987322e-01 5.07797837e-01 1.16622411e-01
-1.14658868e+00 -5.26122153e-01 -6.33313894e-01 -8.60976875e-02
6.24943733e-01 -3.25755000e-01 1.28411651e+00 -1.02625561e+00
-2.12780499e+00 5.36370397e-01 3.10150862e-01 -8.44878793e-01
7.13441253e-01 -4.53083143e-02 7.59410113e-02 -9.23084244e-02
-2.58553233e-02 8.12802494e-01 1.15669680e+00 -8.84590030e-01
-6.53517485e-01 -3.66627693e-01 -2.11032783e-03 4.88045365e-02
1.21694252e-01 -4.96044040e-01 2.19255567e-01 -2.11984321e-01
-1.03050418e-01 -7.40986943e-01 -5.04189193e-01 1.84729788e-02
-2.85435691e-02 -3.94280910e-01 2.34862581e-01 -1.79351673e-01
1.12884104e+00 -1.94602072e+00 -8.39220062e-02 4.14744288e-01
-1.08110279e-01 2.06650004e-01 -2.10760236e-01 2.35490918e-01
9.29214731e-02 -2.01386198e-01 -2.58173257e-01 2.15151817e-01
2.27416679e-01 2.16043457e-01 -6.82511404e-02 5.06890714e-01
1.87989727e-01 7.02322423e-01 -1.13869631e+00 -3.42008889e-01
1.75875157e-01 -5.68834804e-02 -8.47563684e-01 4.19833720e-01
-3.75190139e-01 1.81596041e-01 -3.97273928e-01 9.75252464e-02
5.52122891e-01 9.10483375e-02 2.32324824e-01 1.95399866e-01
-2.03073010e-01 -6.62320480e-02 -1.54356933e+00 1.32366407e+00
-3.51022005e-01 2.11699218e-01 1.16251998e-01 -1.56419170e+00
1.26393366e+00 -3.87337655e-02 5.02726912e-01 -8.49010408e-01
3.96275252e-01 3.99079293e-01 1.84205011e-01 -4.29606318e-01
4.95236158e-01 -3.90674710e-01 -2.59374499e-01 2.52479255e-01
3.29509735e-01 -4.94222701e-01 1.76696137e-01 -1.95416793e-01
9.95682478e-01 4.78177160e-01 6.05706394e-01 -4.88931656e-01
5.75275242e-01 -3.72860104e-01 6.45736933e-01 1.12404835e+00
-8.45178068e-01 -2.95235991e-01 8.11912835e-01 -2.54215240e-01
-6.81967556e-01 -1.31750381e+00 -3.35840546e-02 1.12434089e+00
1.41820535e-01 1.86189458e-01 -9.11488831e-01 -9.20037687e-01
3.21356595e-01 8.99770081e-01 -7.00369000e-01 -4.97237206e-01
-2.89021194e-01 -7.16908872e-01 3.05730432e-01 3.82456332e-01
7.20298946e-01 -1.33749580e+00 -1.11792839e+00 3.05822164e-01
7.82328695e-02 -4.43464190e-01 -6.00658432e-02 6.02124810e-01
-7.39947438e-01 -1.07815838e+00 -4.65922743e-01 -3.13706636e-01
2.69040912e-01 -5.02641261e-01 1.05115163e+00 -4.33201849e-01
4.86778058e-02 4.73787189e-01 -1.53119758e-01 -5.67367136e-01
-2.55028695e-01 -2.71927029e-01 3.49730670e-01 -1.61612406e-01
2.51064301e-01 -4.73459482e-01 -8.64187121e-01 3.00887406e-01
-6.05796218e-01 -6.90963686e-01 5.13624012e-01 1.02340484e+00
6.89090490e-01 -3.09524257e-02 1.11802065e+00 -4.92240667e-01
8.59125078e-01 -4.71349150e-01 -1.06490302e+00 -1.26617044e-01
-1.04787540e+00 4.97346967e-01 6.93136871e-01 -6.47057235e-01
-6.50171578e-01 -3.74602377e-01 -1.67713553e-01 -3.50393802e-01
2.06476182e-01 2.89602071e-01 8.33676979e-02 -9.73175466e-02
6.82220936e-01 2.36778647e-01 1.24192119e-01 1.74094170e-01
3.22560310e-01 5.03838956e-01 2.52848297e-01 -7.80755281e-01
3.26004148e-01 7.24401474e-02 7.38514438e-02 -4.06526476e-01
-6.27373219e-01 2.11736951e-02 -1.03495106e-01 -6.69285119e-01
4.33651000e-01 -3.69038016e-01 -1.35447145e+00 4.38812643e-01
-6.06623590e-01 -9.50785697e-01 -8.21109056e-01 4.80739743e-01
-1.23050690e+00 7.26586580e-02 -2.12962776e-01 -1.21411085e+00
-1.95705116e-01 -1.03192580e+00 5.86454630e-01 3.18273097e-01
5.04486002e-02 -6.62682235e-01 3.67018312e-01 -3.24462920e-01
4.08599854e-01 2.10510701e-01 7.33853996e-01 -5.46029508e-01
-2.02260375e-01 -1.58064887e-01 7.77007341e-02 4.32438970e-01
-3.99800241e-01 -1.60297185e-01 -5.69598734e-01 -5.28439939e-01
-1.65912643e-01 -5.76344490e-01 5.61121762e-01 9.63359833e-01
1.30063188e+00 -3.65119465e-02 8.54398012e-02 3.07561755e-01
1.60138535e+00 6.52427852e-01 4.66929704e-01 6.39753759e-01
-2.13330358e-01 4.71793503e-01 8.96058679e-01 9.76942718e-01
1.71435252e-01 4.19063270e-01 8.76763999e-01 5.61874390e-01
6.28090680e-01 -3.11438322e-01 4.57721442e-01 2.82628179e-01
1.08405724e-01 -4.60537709e-02 -6.36467934e-01 4.85687912e-01
-1.95473564e+00 -1.04429042e+00 4.96809959e-01 2.63113379e+00
8.60374451e-01 4.99107599e-01 3.78753275e-01 1.28821999e-01
7.02017367e-01 1.14774145e-01 -1.19754064e+00 -8.02976489e-01
4.87020940e-01 7.18547106e-01 7.41815507e-01 3.65897506e-01
-6.87479973e-01 7.99344599e-01 6.95380592e+00 9.99130070e-01
-9.27445948e-01 2.76449155e-02 5.20312726e-01 -7.93974176e-02
9.29321572e-02 -2.54486293e-01 -6.58682764e-01 6.05246365e-01
1.20404565e+00 -9.06090513e-02 5.94653845e-01 1.17948377e+00
4.73840326e-01 -6.22178376e-01 -8.46651852e-01 7.55133629e-01
-4.24947232e-01 -9.63265061e-01 -4.22245920e-01 -4.27025184e-02
5.05957127e-01 -6.53051538e-03 1.47836939e-01 9.48769629e-01
6.61665797e-01 -6.38596654e-01 9.53273058e-01 5.14782429e-01
7.67196894e-01 -1.04986572e+00 7.99731553e-01 6.07352734e-01
-8.89470398e-01 -5.42213202e-01 -3.66196722e-01 -2.06065819e-01
-2.50592202e-01 4.93911177e-01 -5.10808766e-01 7.63976248e-03
8.09552789e-01 2.77079314e-01 -1.80601567e-01 7.70429611e-01
-2.59461969e-01 4.08348083e-01 -1.82412297e-01 -7.82449603e-01
3.90493482e-01 -4.72035795e-01 3.28554243e-01 9.28608298e-01
3.65152150e-01 -1.84646145e-01 -4.54423465e-02 1.04850161e+00
2.24967465e-01 -4.03604358e-02 -2.97095507e-01 3.14521432e-01
7.46216238e-01 9.39498186e-01 -5.03811717e-01 -3.10357213e-01
1.89356908e-01 5.39180934e-01 5.95897198e-01 1.79718316e-01
-9.57251728e-01 -6.02790952e-01 6.87311649e-01 -1.67213634e-01
6.01908386e-01 -5.70676513e-02 1.24634415e-01 -6.43935800e-01
-1.85385808e-01 -7.15921283e-01 2.54460007e-01 -5.56644499e-01
-1.07035220e+00 2.26633713e-01 2.01420516e-01 -1.23976886e+00
-6.43746555e-01 -7.54550815e-01 -3.19025785e-01 5.71553469e-01
-1.40249670e+00 -1.64194703e-02 1.43860921e-01 5.09124875e-01
3.09243828e-01 -4.00199480e-02 5.27591825e-01 -2.51269490e-01
-6.46161139e-01 6.57803953e-01 3.06387186e-01 -1.35788292e-01
5.64621806e-01 -1.45259547e+00 -2.08674252e-01 4.73024726e-01
-5.53241611e-01 2.37288266e-01 9.57554162e-01 -3.36803705e-01
-1.04751968e+00 -8.28196943e-01 4.86183986e-02 1.30606638e-02
7.52411187e-01 -5.24058342e-02 -4.80854273e-01 3.52616191e-01
-4.63579781e-02 1.24149978e-01 3.16824019e-01 -4.01611663e-02
-9.39812809e-02 -5.95387220e-01 -1.52292407e+00 5.40687680e-01
7.82583356e-01 -9.31305736e-02 -2.86242634e-01 -1.41702324e-01
3.86674732e-01 -2.35413119e-01 -7.43805826e-01 4.02243972e-01
6.88718617e-01 -1.45936751e+00 7.82300055e-01 -4.57571775e-01
2.59791881e-01 2.22136695e-02 -6.30561039e-02 -1.70824170e+00
-7.76705816e-02 -5.59924304e-01 -1.69369519e-01 6.59452021e-01
5.66159077e-02 -7.60322511e-01 1.04137921e+00 3.97053331e-01
1.90290451e-01 -1.00942206e+00 -1.33619845e+00 -9.09075320e-01
2.84054577e-01 -6.01244032e-01 4.19665903e-01 5.07510245e-01
2.38454208e-01 -2.88976878e-01 -1.62194356e-01 -4.07357514e-02
1.01389253e+00 -1.01602189e-01 7.23630011e-01 -7.66371846e-01
-6.04065180e-01 -5.74355066e-01 -3.02558810e-01 -1.01325917e+00
2.76315033e-01 -3.65610242e-01 7.09203959e-01 -8.79576027e-01
-1.34749919e-01 -4.16118860e-01 -7.01954126e-01 4.09634821e-02
-1.83723897e-01 -2.36653492e-01 -6.30516857e-02 -3.67663026e-01
-6.96792543e-01 8.62842798e-01 1.10961628e+00 1.14897430e-01
-4.26120877e-01 1.35896400e-01 -1.69992223e-01 7.71615982e-01
9.92872655e-01 -5.49048066e-01 -4.47452277e-01 2.19276726e-01
4.12567705e-01 4.90420371e-01 2.61016667e-01 -1.06507325e+00
-7.81616569e-03 -4.21794146e-01 2.67184585e-01 -3.71298403e-01
3.43030393e-02 -6.11989260e-01 -3.36379409e-01 9.69541252e-01
-6.26336575e-01 7.02207983e-02 7.58124292e-02 8.24459672e-01
1.18717611e-01 -4.86067325e-01 1.10868740e+00 -2.75899410e-01
-6.90571606e-01 3.19109946e-01 -6.84947789e-01 3.80540937e-01
1.10202038e+00 -2.33304948e-01 5.97021058e-02 -4.92379367e-01
-4.76475954e-01 3.38531107e-01 -1.54595360e-01 -1.21742778e-01
6.18574023e-01 -1.03030610e+00 -2.83931673e-01 2.40522265e-01
-5.22451662e-02 -2.76694119e-01 1.54346660e-01 6.61078751e-01
-6.86310604e-02 4.52557355e-02 -3.45042318e-01 -5.00174224e-01
-4.56020683e-01 4.99735326e-01 9.41598237e-01 -5.41621745e-01
1.19693033e-01 7.51943111e-01 -3.25973809e-01 -6.88423753e-01
3.01898420e-01 -3.81091654e-01 -6.51047304e-02 -1.47552133e-01
2.66304433e-01 6.15187883e-01 -2.93425202e-01 8.95382240e-02
-2.23488048e-01 4.14894521e-01 3.31277519e-01 -5.10537803e-01
1.14994025e+00 -3.83353755e-02 1.65747821e-01 4.96686667e-01
7.58941472e-01 -3.57736707e-01 -1.77355957e+00 -1.81183457e-01
1.31915540e-01 -2.55591989e-01 -2.60001607e-03 -8.08411241e-01
-9.53261256e-01 6.49790525e-01 8.81690204e-01 2.53975362e-01
9.73075509e-01 -3.88864428e-01 2.54594684e-01 4.49970305e-01
8.17342520e-01 -1.62412047e+00 2.45087951e-01 4.27429467e-01
4.12112474e-01 -1.16121233e+00 -6.61959797e-02 4.46713209e-01
-4.89478916e-01 8.92914355e-01 9.26422358e-01 -7.30144620e-01
7.78686523e-01 2.08104894e-01 -1.84445664e-01 1.28205344e-01
-7.47900844e-01 -2.85843402e-01 -3.69583338e-01 7.29200065e-01
-3.77828218e-02 7.66150355e-02 -7.02759981e-01 5.41431725e-01
-6.08248077e-02 2.25338355e-01 3.42472345e-01 7.77019918e-01
-7.20093966e-01 -7.66438901e-01 -2.16298655e-01 4.44279462e-01
-1.41391620e-01 1.91492602e-01 1.80519536e-01 1.08767664e+00
2.27508053e-01 9.22854543e-01 2.12577373e-01 -4.07308191e-01
3.81103694e-01 5.20609543e-02 6.04444444e-01 -1.15447409e-01
-5.31618297e-01 -1.79053068e-01 -2.61762649e-01 -9.73242521e-01
-5.05526364e-01 -5.01946390e-01 -1.17091846e+00 -2.18865439e-01
-3.62266004e-01 4.22165513e-01 6.53333306e-01 1.03219128e+00
1.06253505e-01 3.56897026e-01 8.87056351e-01 -8.64350915e-01
-1.25845587e+00 -9.64813232e-01 -9.78678524e-01 5.47047071e-02
1.52222380e-01 -9.14200842e-01 -6.05562270e-01 -6.00974858e-01] | [4.239161968231201, 2.3050150871276855] |
f97a4687-5fd8-4f8c-9037-30ce0a982579 | rics-a-2d-self-occlusion-map-for-harmonizing | 2205.06975 | null | https://arxiv.org/abs/2205.06975v1 | https://arxiv.org/pdf/2205.06975v1.pdf | RiCS: A 2D Self-Occlusion Map for Harmonizing Volumetric Objects | There have been remarkable successes in computer vision with deep learning. While such breakthroughs show robust performance, there have still been many challenges in learning in-depth knowledge, like occlusion or predicting physical interactions. Although some recent works show the potential of 3D data in serving such context, it is unclear how we efficiently provide 3D input to the 2D models due to the misalignment in dimensionality between 2D and 3D. To leverage the successes of 2D models in predicting self-occlusions, we design Ray-marching in Camera Space (RiCS), a new method to represent the self-occlusions of foreground objects in 3D into a 2D self-occlusion map. We test the effectiveness of our representation on the human image harmonization task by predicting shading that is coherent with a given background image. Our experiments demonstrate that our representation map not only allows us to enhance the image quality but also to model temporally coherent complex shadow effects compared with the simulation-to-real and harmonization methods, both quantitatively and qualitatively. We further show that we can significantly improve the performance of human parts segmentation networks trained on existing synthetic datasets by enhancing the harmonization quality with our method. | ['Honglak Lee', 'Xin Sun', 'Duygu Ceylan', 'Jimei Yang', 'Ruben Villegas', 'Yunseok Jang'] | 2022-05-14 | null | null | null | null | ['image-harmonization'] | ['computer-vision'] | [ 1.51556745e-01 1.39090244e-03 2.81764060e-01 -3.46681237e-01
-5.41850924e-01 -3.24030191e-01 7.10403264e-01 -4.28633213e-01
2.21515417e-01 5.49880803e-01 1.29184514e-01 -3.46620947e-01
2.39457294e-01 -7.12363958e-01 -9.18784916e-01 -7.38282561e-01
-2.70367824e-02 4.80494738e-01 4.86325502e-01 -1.09449357e-01
-1.62000313e-01 8.01636755e-01 -1.62455225e+00 2.55623996e-01
1.06498861e+00 7.32357800e-01 2.18959808e-01 7.76665449e-01
-1.97824776e-01 7.21611798e-01 -5.95669508e-01 -9.16530341e-02
5.43042779e-01 -5.01170456e-01 -5.05281508e-01 3.52612644e-01
8.19625914e-01 -4.71292794e-01 -4.27382410e-01 5.24509609e-01
4.66272175e-01 1.77336135e-03 5.10753274e-01 -1.27013409e+00
-3.66836786e-01 -1.67010069e-01 -9.40376580e-01 -1.55494854e-01
3.02450210e-01 3.57998252e-01 7.43681610e-01 -5.27994752e-01
8.49004269e-01 1.47222435e+00 9.64088261e-01 2.97346324e-01
-1.33371556e+00 -4.72633660e-01 7.54691064e-02 6.42023459e-02
-9.11949754e-01 -2.25990951e-01 1.08458853e+00 -4.30566698e-01
8.02341938e-01 3.08181256e-01 1.13074386e+00 9.76923525e-01
3.32947850e-01 8.62144768e-01 1.30275023e+00 -5.17136395e-01
1.83734328e-01 -5.32639250e-02 5.17605916e-02 7.47393548e-01
-5.36070801e-02 3.50636691e-01 -7.42682338e-01 5.43733016e-02
9.77009833e-01 -3.33010644e-01 -3.73854518e-01 -6.91463947e-01
-1.02694952e+00 6.58852935e-01 5.66392243e-01 -8.02754760e-02
-1.81528449e-01 3.86306256e-01 -2.69103199e-02 -2.61968583e-01
7.85043240e-01 6.01398051e-01 -3.15904230e-01 2.62904406e-01
-1.03328252e+00 5.15919149e-01 7.88958192e-01 7.75569677e-01
7.43599892e-01 2.59062707e-01 -1.28247380e-01 6.38502538e-01
1.33859247e-01 7.13959098e-01 1.01929149e-02 -1.67950940e+00
-2.70280875e-02 4.84558970e-01 3.98992628e-01 -1.01053774e+00
-5.47720790e-01 -4.94014889e-01 -6.24670684e-01 7.15789795e-01
4.86683875e-01 -9.68233589e-03 -1.04142308e+00 1.89949119e+00
5.93664110e-01 5.16341031e-01 6.88462108e-02 1.02238119e+00
7.81917274e-01 8.24378133e-01 -2.31032595e-01 -1.22518770e-01
1.03418612e+00 -1.13504779e+00 -5.87823987e-01 -3.71817410e-01
3.32935840e-01 -9.66771305e-01 1.06207275e+00 1.86415330e-01
-1.31610060e+00 -6.54557407e-01 -1.00306129e+00 -2.07753301e-01
1.05422642e-02 -4.68990117e-01 9.99124706e-01 5.35088003e-01
-1.05685651e+00 5.83164334e-01 -9.87960100e-01 -4.08842206e-01
5.76811433e-01 -9.20738839e-03 -6.83939606e-02 -4.49755937e-02
-9.72461581e-01 9.41109598e-01 -1.72322974e-01 -2.01632887e-01
-6.90461457e-01 -1.17609751e+00 -7.79808760e-01 -8.08168650e-02
-2.93600652e-03 -1.01716304e+00 1.15158153e+00 -8.57134581e-01
-1.29801571e+00 1.02524555e+00 -2.74684995e-01 -5.42953849e-01
6.57100499e-01 -1.51241407e-01 7.82033727e-02 2.87335902e-01
4.53979038e-02 1.08681166e+00 3.86940330e-01 -1.88497031e+00
-4.18571025e-01 -2.06058696e-01 1.56175271e-01 5.17519951e-01
9.03159454e-02 -2.92656690e-01 -6.47260368e-01 -5.59899211e-01
2.12044641e-01 -1.03943205e+00 -3.31830353e-01 4.98929411e-01
-5.27453005e-01 2.90006012e-01 1.15820694e+00 -5.73064864e-01
5.17512619e-01 -1.74985039e+00 -1.11114077e-01 -2.04355389e-01
1.73517823e-01 2.72887647e-01 -1.76616475e-01 2.38067508e-01
6.74882457e-02 1.13346539e-02 -3.76076996e-01 -7.34177351e-01
3.62449735e-02 5.44397593e-01 -7.34670103e-01 4.07775521e-01
1.10877909e-01 1.16061556e+00 -7.95328081e-01 -6.06962621e-01
6.72052920e-01 8.57133806e-01 -6.10588253e-01 4.02458310e-01
-5.36977172e-01 7.36683249e-01 -3.79136443e-01 4.02069271e-01
9.78324831e-01 -3.03095311e-01 -1.09532803e-01 -3.14981073e-01
-1.55354649e-01 1.12122260e-01 -1.07235968e+00 1.64686072e+00
-5.28825998e-01 9.27857339e-01 4.08936292e-02 -7.28855729e-01
5.46925902e-01 6.02912977e-02 7.79345214e-01 -9.51225221e-01
-1.82585016e-01 -6.37110695e-02 -3.20627123e-01 -4.87000495e-01
2.84222931e-01 -1.22434147e-01 3.98726165e-01 5.12233734e-01
-3.97668332e-01 -9.73654628e-01 -1.61686793e-01 9.73393321e-02
8.76886487e-01 6.63309693e-01 -1.25997886e-01 -1.73891112e-01
-7.67629743e-02 3.53390604e-01 5.68741739e-01 8.48297536e-01
-6.86489716e-02 1.05692005e+00 4.01851714e-01 -5.86303353e-01
-1.18788540e+00 -1.18828392e+00 -1.88693121e-01 5.06155968e-01
5.76057374e-01 -1.38893351e-01 -8.77564371e-01 -2.31419116e-01
4.19435725e-02 9.15272534e-01 -5.24044752e-01 4.76761907e-02
-6.83844686e-01 -9.80198920e-01 4.06351954e-01 4.62783247e-01
8.36067498e-01 -8.78272295e-01 -1.19473672e+00 9.51745808e-02
-1.60580412e-01 -1.41249299e+00 -1.47248164e-01 2.52776772e-01
-7.20395982e-01 -1.08262479e+00 -6.11263633e-01 -5.87257326e-01
3.25077623e-01 8.32571149e-01 1.46601117e+00 1.40504017e-01
-5.78388453e-01 5.35652220e-01 -4.42306064e-02 -4.19032961e-01
-5.00734508e-01 -5.30659318e-01 -2.10395083e-01 -3.06184381e-01
-1.08339533e-01 -7.81203330e-01 -9.04214919e-01 5.63104451e-01
-9.70935285e-01 8.66052389e-01 1.21279113e-01 5.70854485e-01
5.57350099e-01 6.53325394e-02 1.41988292e-01 -7.46859550e-01
1.01489142e-01 4.48100194e-02 -6.39596164e-01 2.12134540e-01
-2.96771854e-01 -4.00360972e-02 3.18208277e-01 -2.55153269e-01
-1.40081394e+00 1.78013265e-01 -5.99776842e-02 -5.11068165e-01
-1.63596228e-01 -3.08246184e-02 2.83851475e-02 -3.40821803e-01
5.40831685e-01 -1.75682291e-01 -1.20685667e-01 -2.01338992e-01
5.48967361e-01 9.39414948e-02 6.33385301e-01 -7.81147778e-01
9.41239178e-01 9.60773408e-01 3.75830084e-01 -9.39017951e-01
-1.15094376e+00 -1.65467020e-02 -6.03810072e-01 -3.56874675e-01
1.24535990e+00 -8.31439376e-01 -5.93070686e-01 6.96221113e-01
-1.44111156e+00 -8.57723892e-01 -4.08095926e-01 2.43039817e-01
-7.12597072e-01 4.59683508e-01 -5.82192540e-01 -7.40782559e-01
7.41381794e-02 -1.31047547e+00 1.33073068e+00 3.52898151e-01
-1.44087672e-01 -1.14415097e+00 4.10348400e-02 5.19283295e-01
2.86972046e-01 8.32011700e-01 1.12056315e+00 2.38572195e-01
-9.94553804e-01 2.32356712e-01 -3.30448717e-01 1.64285168e-01
2.78088525e-02 2.36329243e-01 -1.30993426e+00 1.07383877e-01
2.56377667e-01 -4.36764926e-01 7.99700677e-01 8.64306986e-01
1.00329423e+00 1.59971744e-01 -4.27403003e-01 1.04867578e+00
1.27639866e+00 1.60302445e-01 6.04827046e-01 1.72272831e-01
9.36548531e-01 8.88157725e-01 4.63425517e-01 2.81127632e-01
3.29871535e-01 1.16153204e+00 4.63061512e-01 -9.07473922e-01
-8.50035191e-01 -2.35408157e-01 -1.76749259e-01 4.95724082e-01
6.36323169e-02 -4.59252864e-01 -1.00789714e+00 1.96610540e-01
-1.90903306e+00 -7.10204065e-01 -4.03100640e-01 1.96046937e+00
8.39940846e-01 1.99150845e-01 -3.20064127e-01 -1.92490727e-01
4.18916106e-01 3.95241499e-01 -8.49514067e-01 -1.59067333e-01
-6.56133235e-01 1.71641991e-01 3.77783179e-01 9.08980072e-01
-9.60581958e-01 1.16369140e+00 6.64711905e+00 6.22584343e-01
-1.14110601e+00 -1.11560263e-01 9.72965181e-01 -7.87490755e-02
-5.56574523e-01 8.08468834e-02 -5.36612630e-01 1.43209547e-01
3.32941979e-01 1.97215557e-01 2.90981323e-01 5.12915492e-01
6.09810174e-01 -4.95770663e-01 -1.18997312e+00 9.48854744e-01
2.00047269e-01 -1.58021235e+00 1.74819380e-02 -5.91179505e-02
1.28395951e+00 -9.19340700e-02 1.41349882e-01 -6.30491748e-02
3.87168020e-01 -1.03985167e+00 7.46798575e-01 4.21812564e-01
5.20722628e-01 -3.44191104e-01 1.95209697e-01 2.92538315e-01
-1.00112677e+00 4.55265611e-01 -2.40739360e-01 -1.25905676e-02
5.40478706e-01 6.78877413e-01 -7.88891613e-01 3.41032058e-01
5.90400100e-01 6.00707352e-01 -5.22263706e-01 1.08803225e+00
-2.76233017e-01 5.06644309e-01 -4.80037332e-01 4.22551185e-01
2.14052767e-01 -3.92956287e-01 5.28638721e-01 1.06015515e+00
6.22004196e-02 2.21281782e-01 2.29343981e-01 1.12881970e+00
2.39927359e-02 -3.46330225e-01 -6.79441571e-01 3.74310613e-01
1.59476995e-01 9.95598912e-01 -9.08919275e-01 -6.80541471e-02
-3.59147429e-01 9.83835876e-01 5.21935709e-02 8.30902040e-01
-1.14111900e+00 1.19815387e-01 7.14684248e-01 2.48654068e-01
1.33816004e-01 -4.98686284e-01 -7.04476953e-01 -9.18197930e-01
-1.82116441e-02 -5.53562522e-01 -2.28942409e-01 -1.50429213e+00
-1.10809791e+00 4.31912810e-01 1.43003076e-01 -1.00510347e+00
1.16199497e-02 -5.50861180e-01 -6.74288273e-01 8.10542583e-01
-1.60515928e+00 -1.46377742e+00 -6.93471193e-01 3.43888789e-01
5.54901719e-01 2.97915488e-01 7.54482687e-01 -2.24946830e-02
-2.26619571e-01 1.48809236e-02 1.23782389e-01 -3.69994789e-01
7.20949769e-01 -1.28021061e+00 6.46759391e-01 7.91249931e-01
2.66057581e-01 1.04259215e-01 9.48922753e-01 -6.13183141e-01
-1.37411380e+00 -9.26123857e-01 4.12044734e-01 -6.85954034e-01
2.56652921e-01 -4.83683616e-01 -9.34859514e-01 4.84461099e-01
3.54384303e-01 8.06326717e-02 2.41206601e-01 -1.99487031e-01
-4.01172578e-01 -2.72807963e-02 -1.01914167e+00 9.57878411e-01
1.24929059e+00 -3.67030561e-01 -3.71589035e-01 6.02506280e-01
1.05475926e+00 -7.54054904e-01 -4.95274514e-01 6.17329895e-01
4.94356513e-01 -1.42867446e+00 1.43183756e+00 -4.38341081e-01
7.38890350e-01 -3.79209101e-01 -2.05088943e-01 -1.24410295e+00
5.25751039e-02 -6.90358400e-01 -1.62310079e-02 9.56270933e-01
-7.21328184e-02 -3.13426435e-01 1.09391391e+00 9.07178760e-01
-3.55073959e-01 -8.96188676e-01 -8.22002530e-01 -5.66140652e-01
8.41171667e-02 -5.86589992e-01 4.65962172e-01 8.47732186e-01
-7.83371031e-01 1.69934556e-01 -3.44912320e-01 3.19198966e-01
8.93288493e-01 6.81694269e-01 9.98326898e-01 -1.11403310e+00
-5.37398398e-01 -4.11780685e-01 -3.84912267e-02 -1.28032422e+00
2.63253093e-01 -6.27697408e-01 -1.45996269e-02 -1.83879805e+00
2.48733312e-01 -5.93011737e-01 3.18621308e-01 2.54548967e-01
-1.99472666e-01 4.27512020e-01 2.02892810e-01 2.27669820e-01
-5.28623164e-01 6.44819677e-01 1.68270493e+00 -5.10020293e-02
-3.66324544e-01 -2.20303178e-01 -3.21433842e-01 9.63310659e-01
4.93298978e-01 -7.95167014e-02 -5.62279344e-01 -8.22797298e-01
-4.58298586e-02 1.93792492e-01 7.42693186e-01 -1.18468201e+00
6.08723378e-03 -1.34108439e-01 7.09548950e-01 -9.08579826e-01
8.09786737e-01 -6.60984874e-01 3.01904410e-01 3.22264791e-01
-2.59562999e-01 -3.34763885e-01 4.74191517e-01 4.74464923e-01
1.07545488e-01 1.04725480e-01 1.04955363e+00 -1.11653544e-01
-5.87510169e-01 1.99129879e-01 -4.58577909e-02 2.45191231e-01
9.14018035e-01 -3.28423500e-01 -4.37169671e-01 -5.88999331e-01
-3.80694062e-01 1.65430337e-01 9.36633706e-01 1.60280377e-01
3.94360512e-01 -1.12024844e+00 -4.75401312e-01 2.22351879e-01
-1.61071211e-01 2.78625518e-01 2.67819762e-01 5.71689069e-01
-9.64073658e-01 2.37931266e-01 -1.82328150e-01 -1.00688934e+00
-1.11856484e+00 2.45149657e-01 6.04873419e-01 -7.64652938e-02
-8.84645462e-01 8.03940296e-01 9.06731367e-01 -3.31917137e-01
4.99745667e-01 -2.86785483e-01 3.72862548e-01 -4.31566089e-01
1.56948850e-01 1.20126098e-01 -1.99322224e-01 -4.31344181e-01
-8.68719742e-02 1.03456831e+00 2.75767565e-01 -2.34191269e-01
1.42225671e+00 -7.82770216e-02 2.17284009e-01 3.60514700e-01
1.07197952e+00 7.59144649e-02 -2.04530454e+00 -3.78498137e-02
-4.59235907e-01 -5.57666123e-01 1.38578802e-01 -8.22568297e-01
-1.17508745e+00 1.01380241e+00 6.15691364e-01 8.53875056e-02
1.02002978e+00 2.95170564e-02 1.09330583e+00 2.27266014e-01
1.75246134e-01 -7.63740718e-01 2.85337657e-01 3.90062600e-01
7.84768820e-01 -1.27080882e+00 1.72696859e-01 -6.74058437e-01
-5.78668058e-01 9.94660079e-01 6.36829972e-01 -7.73410276e-02
6.20027781e-01 4.72578496e-01 3.27628255e-01 -2.15738341e-01
-4.98272240e-01 -8.37439895e-02 1.49032027e-01 7.86358178e-01
2.71000654e-01 -1.84375346e-01 1.54860467e-01 -5.73547482e-02
-1.62023157e-01 -3.23277861e-01 4.42633629e-01 6.50720835e-01
-2.58316338e-01 -1.05787599e+00 -4.15208310e-01 -2.87465416e-02
1.93517104e-01 1.00985564e-01 -3.40270370e-01 9.74112213e-01
3.02951068e-01 8.36004198e-01 6.89838827e-02 -1.28817141e-01
2.93495625e-01 -2.24203542e-01 7.28403747e-01 -3.70236218e-01
-2.41659656e-01 2.09078565e-01 5.40203340e-02 -7.35998094e-01
-6.92364395e-01 -5.22612333e-01 -1.33218944e+00 -1.04511470e-01
4.84143868e-02 -2.27654666e-01 8.07345092e-01 7.57472873e-01
2.63312638e-01 6.04370475e-01 6.03816509e-01 -1.45803368e+00
7.19571859e-02 -3.04978490e-01 -3.83096427e-01 6.42269433e-01
2.70134211e-01 -7.90596724e-01 -4.81566727e-01 1.84947610e-01] | [9.3689603805542, -3.010524034500122] |
69e3e8b7-ec20-47f9-9d70-cd3b89a5dc3c | matching-distributions-between-model-and-data | null | null | https://aclanthology.org/2021.acl-long.421 | https://aclanthology.org/2021.acl-long.421.pdf | Matching Distributions between Model and Data: Cross-domain Knowledge Distillation for Unsupervised Domain Adaptation | Unsupervised Domain Adaptation (UDA) aims to transfer the knowledge of source domain to the unlabeled target domain. Existing methods typically require to learn to adapt the target model by exploiting the source data and sharing the network architecture across domains. However, this pipeline makes the source data risky and is inflexible for deploying the target model. This paper tackles a novel setting where only a trained source model is available and different network architectures can be adapted for target domain in terms of deployment environments. We propose a generic framework named \textit{Cross-domain Knowledge Distillation} (CdKD) without needing any source data. CdKD matches the joint distributions between a trained source model and a set of target data during distilling the knowledge from the source model to the target domain. As a type of important knowledge in the source domain, for the first time, the gradient information is exploited to boost the transfer performance. Experiments on cross-domain text classification demonstrate that CdKD achieves superior performance, which verifies the effectiveness in this novel setting. | ['Zhoujun Li', 'Lei Cheng', 'Yun Liu', 'XiaoMing Zhang', 'Bo Zhang'] | 2021-08-01 | null | null | null | acl-2021-5 | ['cross-domain-text-classification'] | ['natural-language-processing'] | [ 9.70490053e-02 1.36454195e-01 -3.25456530e-01 -6.33874059e-01
-6.47555411e-01 -8.87043118e-01 6.26816690e-01 -8.90736505e-02
-4.79678601e-01 8.64495635e-01 -8.38454813e-02 -1.86033025e-01
-3.74362655e-02 -7.98891187e-01 -8.52407992e-01 -5.20781755e-01
1.98440418e-01 7.56244898e-01 3.00998509e-01 -1.96029112e-01
-3.15668792e-01 2.65365481e-01 -8.70433092e-01 8.06642324e-02
1.01340330e+00 8.77173424e-01 6.15846157e-01 1.57356918e-01
-5.19220471e-01 5.11616230e-01 -7.77525425e-01 -4.94609654e-01
4.44882989e-01 -5.78421950e-01 -9.50405180e-01 -1.04603067e-01
9.62822884e-02 -3.34521711e-01 -3.65174651e-01 9.85416532e-01
5.06747067e-01 1.14484385e-01 7.48364091e-01 -1.40025222e+00
-8.26102853e-01 8.45303059e-01 -3.87461782e-01 8.27785656e-02
-1.66203380e-01 -9.72146466e-02 5.52817583e-01 -8.26516807e-01
6.78253293e-01 9.90982950e-01 5.44914901e-01 8.60285461e-01
-9.98751521e-01 -1.01233077e+00 6.18236125e-01 3.09677683e-02
-1.25900054e+00 -3.99833798e-01 9.08315599e-01 -3.09724033e-01
4.83989030e-01 -5.13939559e-01 7.97384977e-02 1.52358556e+00
-4.40285742e-01 1.04472733e+00 8.73127043e-01 -4.19202685e-01
4.36640143e-01 6.05948091e-01 8.46093893e-02 2.45778024e-01
1.96860790e-01 3.23841907e-02 -6.33800089e-01 -6.77070245e-02
6.29437268e-01 7.94641599e-02 -1.78864554e-01 -6.95776820e-01
-1.17744684e+00 6.62446380e-01 4.30236667e-01 3.14440161e-01
-3.31007123e-01 -4.72881794e-01 5.36178410e-01 4.72008348e-01
5.27479351e-01 2.69031137e-01 -9.05797005e-01 1.17338477e-02
-6.16154432e-01 3.13817081e-03 9.18551326e-01 1.41774893e+00
9.76045430e-01 1.00392409e-01 2.46307952e-03 1.06304657e+00
2.15332285e-01 7.77545869e-01 7.84623921e-01 -4.93548870e-01
8.25539351e-01 7.14125693e-01 1.00435309e-01 -4.90319461e-01
-6.64995052e-03 -4.81060803e-01 -7.59662211e-01 3.16365110e-03
5.46070993e-01 -5.50211906e-01 -8.84492576e-01 1.96480215e+00
5.34719050e-01 2.95877337e-01 5.93143642e-01 6.54301107e-01
5.31410575e-01 6.45253479e-01 1.14397034e-01 4.22118604e-02
9.29978609e-01 -1.08918214e+00 -3.29791903e-01 -6.54688537e-01
6.29729509e-01 -4.79609489e-01 1.00455260e+00 1.19004257e-01
-7.01578915e-01 -6.60399675e-01 -1.08219826e+00 2.19433755e-01
-5.06698489e-01 2.32008085e-01 1.46375790e-01 2.75877416e-01
-7.41399705e-01 3.83691043e-01 -5.16911983e-01 -5.66415787e-01
5.71232557e-01 2.48040885e-01 -4.00026023e-01 -5.15467644e-01
-1.41673386e+00 7.82999098e-01 9.76812005e-01 -2.89170265e-01
-1.13626218e+00 -7.41203010e-01 -7.62079239e-01 1.21181004e-01
3.72641474e-01 -5.26989102e-01 1.36929905e+00 -1.51183844e+00
-1.71431684e+00 6.48171484e-01 1.20653592e-01 -4.08777833e-01
5.68411887e-01 -1.41484454e-01 -4.82393533e-01 1.81994438e-02
2.61772871e-01 5.44315577e-01 1.10049331e+00 -1.10967910e+00
-8.68303001e-01 -3.24116975e-01 7.26872385e-02 3.97212207e-01
-8.39005470e-01 -4.07463759e-01 -6.38972759e-01 -6.42524481e-01
-3.42944205e-01 -7.49457359e-01 7.06124157e-02 -2.04627991e-01
-2.62066901e-01 -3.32619518e-01 8.95493269e-01 -5.64425051e-01
1.07914281e+00 -2.29294109e+00 3.51974964e-02 2.23119512e-01
6.75958395e-02 6.40580237e-01 -4.07417893e-01 2.26711690e-01
-2.41564140e-01 -1.94258034e-01 -4.03762192e-01 -2.16518506e-01
1.11138500e-01 3.06354493e-01 -3.73004913e-01 1.04325637e-01
2.65015304e-01 6.07220531e-01 -1.04799747e+00 -4.23508793e-01
-1.34120166e-01 3.17686111e-01 -3.19274753e-01 7.16405094e-01
-3.54189366e-01 6.06515229e-01 -8.56657922e-01 2.09709555e-01
8.22202921e-01 -4.09752846e-01 4.74690974e-01 -2.33070962e-02
3.74710590e-01 1.69344395e-01 -1.09603572e+00 2.03086853e+00
-6.55269206e-01 3.13390106e-01 8.29119310e-02 -1.29535353e+00
1.33075345e+00 2.24626198e-01 2.84122109e-01 -7.19897330e-01
3.11829196e-03 2.81090111e-01 -1.22157566e-01 -3.56078148e-01
8.47923011e-02 -1.88863233e-01 -2.83735663e-01 5.30458093e-01
4.83591497e-01 3.78411859e-01 -1.44035861e-01 2.10165113e-01
8.33269596e-01 4.04899150e-01 3.30489665e-01 -1.69419140e-01
3.08269471e-01 2.08547980e-01 6.87833250e-01 6.64027929e-01
-1.99781567e-01 3.12642336e-01 1.44650370e-01 -3.03984374e-01
-1.04129028e+00 -1.08600664e+00 8.84914696e-02 1.42513919e+00
1.39835507e-01 7.92366192e-02 -7.06107676e-01 -1.39273477e+00
2.14675851e-02 7.91667879e-01 -6.99019849e-01 -5.21805882e-01
-4.13385093e-01 -4.09885377e-01 5.64862251e-01 7.29339063e-01
8.85619998e-01 -6.72828078e-01 -1.12068132e-01 1.97585702e-01
-2.10129946e-01 -1.13045192e+00 -6.95221901e-01 3.52774978e-01
-7.86499441e-01 -1.01798666e+00 -1.01975203e+00 -1.00601578e+00
8.86272907e-01 2.56888419e-01 1.03715026e+00 -5.36325693e-01
3.78545910e-01 3.31711113e-01 -4.49224859e-01 -4.56111312e-01
-6.71519637e-01 4.98878002e-01 1.37138382e-01 1.74033448e-01
7.77867258e-01 -6.16326332e-01 -3.86418521e-01 4.93507743e-01
-9.13017809e-01 7.71324243e-03 7.35600650e-01 8.32372904e-01
3.75910074e-01 2.40819126e-01 9.34727430e-01 -1.14819360e+00
6.02199674e-01 -9.69250143e-01 -6.38706744e-01 4.46706504e-01
-6.65551782e-01 1.60002246e-01 1.04219365e+00 -8.28124940e-01
-1.58918023e+00 1.21605635e-01 2.49529853e-01 -6.57913089e-01
-3.91410887e-01 6.04251623e-01 -6.44438148e-01 3.29268962e-01
1.04051530e+00 2.19202280e-01 -5.68664819e-02 -7.25917220e-01
2.93877959e-01 9.01829183e-01 5.43535709e-01 -8.16823125e-01
1.00533676e+00 2.83315271e-01 -5.45483708e-01 -2.28249148e-01
-1.16669285e+00 -4.63363796e-01 -9.57285285e-01 2.29436204e-01
4.50042754e-01 -1.10870981e+00 2.61909943e-02 5.40145755e-01
-1.05278730e+00 -7.75044739e-01 -4.33418989e-01 5.15662074e-01
-2.31201187e-01 1.32722959e-01 -2.16770452e-02 -4.01107281e-01
-1.85959518e-01 -8.46338391e-01 6.95809662e-01 3.18875343e-01
1.72769323e-01 -1.51807725e+00 2.97759026e-02 8.44607688e-03
5.14267325e-01 -6.94808885e-02 7.44588673e-01 -1.43229210e+00
-1.68934762e-01 -1.33442044e-01 -3.11942607e-01 6.39668107e-01
4.82563406e-01 -5.59649885e-01 -1.11256850e+00 -4.72761869e-01
-3.95104736e-02 -5.27438521e-01 5.42165577e-01 -3.44466306e-02
1.02348018e+00 -2.94123322e-01 -4.59253430e-01 4.80105162e-01
1.18616104e+00 1.79824010e-01 1.90955102e-01 4.79951113e-01
7.41262257e-01 5.70526004e-01 7.79165208e-01 3.28465492e-01
5.93507290e-01 5.67875206e-01 1.46845700e-02 -1.93790019e-01
-1.81553528e-01 -6.30305529e-01 5.65931320e-01 7.63171732e-01
5.56517005e-01 -2.60479838e-01 -1.08338344e+00 8.56018066e-01
-1.74660242e+00 -5.48616469e-01 6.64892554e-01 2.30520129e+00
1.19607413e+00 7.74211362e-02 1.35836005e-01 -5.22773445e-01
9.75185633e-01 -1.99852839e-01 -1.26006496e+00 1.33584589e-01
4.30395529e-02 1.42540652e-02 5.16713917e-01 3.36371869e-01
-1.04562926e+00 9.47772682e-01 5.82116604e+00 9.66166914e-01
-1.27644682e+00 3.09531450e-01 3.62222254e-01 1.41001314e-01
-1.64024860e-01 -6.05037324e-02 -1.02004933e+00 6.07728422e-01
1.17308331e+00 -4.96148676e-01 1.43035293e-01 1.26214015e+00
-3.07851732e-01 2.15129361e-01 -1.28239906e+00 7.26289451e-01
-3.06896605e-02 -9.26802635e-01 2.33406946e-01 -2.65654206e-01
6.74365401e-01 2.61209220e-01 -1.46310017e-01 8.15286279e-01
8.66548538e-01 -5.68531156e-01 4.10513997e-01 2.06032991e-01
1.00232375e+00 -7.11859226e-01 7.36848474e-01 6.46494448e-01
-1.00239420e+00 5.83282998e-03 -5.62567294e-01 3.78211856e-01
-1.51540011e-01 4.11024362e-01 -1.34707558e+00 7.30222940e-01
7.88118124e-01 9.24176335e-01 -4.78513032e-01 6.95985436e-01
-2.54480332e-01 7.94798911e-01 -3.78285050e-01 2.80648172e-01
8.08145255e-02 -7.72149190e-02 3.55332792e-01 1.09897614e+00
3.69742870e-01 -2.74094939e-01 4.47527409e-01 7.74207532e-01
-5.01988173e-01 1.05041368e-02 -7.92260170e-01 -7.27522150e-02
8.05938423e-01 1.01555574e+00 -2.97767341e-01 -5.21637857e-01
-4.15451825e-01 1.16687834e+00 6.72769368e-01 6.61835909e-01
-7.15988278e-01 -5.19437432e-01 4.81001288e-01 -1.00307547e-01
3.78538996e-01 9.16839838e-02 5.36476038e-02 -1.37506342e+00
1.23989597e-01 -7.80130148e-01 7.55162716e-01 -5.86789429e-01
-1.80020916e+00 7.02551544e-01 3.24310601e-01 -1.47054112e+00
-2.85380930e-01 -4.63268131e-01 -5.14756322e-01 1.27005613e+00
-2.07672548e+00 -1.28508580e+00 -3.63974571e-01 1.09489024e+00
5.29356897e-01 -5.48769891e-01 8.81979465e-01 4.12659317e-01
-4.65124726e-01 9.69650209e-01 6.00683153e-01 4.73274171e-01
1.25025320e+00 -1.12764645e+00 5.04896700e-01 7.66050279e-01
-2.90145487e-01 5.23133218e-01 2.52662629e-01 -5.39992630e-01
-9.66685295e-01 -1.53695703e+00 5.07121325e-01 -4.24732476e-01
6.43196166e-01 -4.63667303e-01 -1.33306229e+00 9.43967819e-01
1.64157763e-01 7.78231472e-02 8.58053505e-01 8.31708908e-02
-6.83629572e-01 -3.68612021e-01 -1.15470040e+00 3.06914181e-01
8.22857499e-01 -5.52902937e-01 -6.96628392e-01 2.72028744e-01
7.52637506e-01 -3.43735039e-01 -8.86643052e-01 6.79958910e-02
9.57064107e-02 -5.06995440e-01 7.86238194e-01 -8.53690147e-01
1.58566579e-01 -3.02233875e-01 3.92447896e-02 -1.75902879e+00
-1.31309003e-01 -3.29231948e-01 -9.26796719e-02 1.68244755e+00
6.07330680e-01 -8.68187070e-01 6.24167085e-01 6.88191593e-01
-1.15358187e-02 -1.56353071e-01 -7.51763940e-01 -1.15760446e+00
6.45496905e-01 -1.17487028e-01 8.84584308e-01 1.51883459e+00
-1.05685771e-01 6.58630788e-01 -1.66715130e-01 3.33733767e-01
5.33239067e-01 2.40758538e-01 9.59809303e-01 -1.33667254e+00
-2.68756777e-01 -7.43768290e-02 5.43121062e-02 -1.34869468e+00
4.77429926e-01 -1.15234709e+00 7.77951181e-02 -1.33203793e+00
7.73843005e-02 -9.41816330e-01 -5.56113303e-01 8.23632002e-01
-3.11835110e-01 -3.73393178e-01 4.70772907e-02 4.75316942e-01
-6.45604134e-01 7.06246853e-01 1.12695777e+00 -2.15029657e-01
-2.64159888e-01 4.53241132e-02 -1.06005085e+00 5.92806458e-01
9.83345270e-01 -5.98693073e-01 -8.85094643e-01 -7.99481690e-01
-1.56948939e-01 -2.00723737e-01 8.21740180e-02 -8.24563444e-01
3.63733739e-01 -2.19146222e-01 4.48894292e-01 -1.13836132e-01
9.14604031e-03 -1.12073016e+00 -1.80976838e-01 -7.61574656e-02
-3.96276593e-01 -3.95226508e-01 3.96175444e-01 8.20808172e-01
-2.66884625e-01 -3.75491619e-01 1.01668203e+00 2.67201513e-02
-1.03063178e+00 4.40061510e-01 1.62045687e-01 4.51297075e-01
1.07379425e+00 -1.10049255e-01 -3.99162740e-01 -2.34851673e-01
-7.46927321e-01 5.38374543e-01 5.00611603e-01 5.96055686e-01
2.99367309e-01 -1.49462223e+00 -7.30919302e-01 2.01484099e-01
4.75570560e-01 4.34056103e-01 3.10963541e-01 3.64616394e-01
5.27318455e-02 1.08176433e-01 -2.98187286e-01 -5.54664910e-01
-6.82678998e-01 7.57720053e-01 5.01618981e-01 -4.35667723e-01
-5.12498975e-01 7.82854259e-01 5.93577206e-01 -9.78183627e-01
2.35572234e-01 1.81994632e-01 -7.20201433e-02 -2.07543094e-02
5.88273227e-01 1.64590217e-02 -1.64748412e-02 -2.95696735e-01
-1.47659570e-01 2.36960709e-01 -5.35832405e-01 1.88638687e-01
1.27565753e+00 -4.54134673e-01 3.32946062e-01 3.97407204e-01
1.20649660e+00 -2.99238980e-01 -1.80363107e+00 -1.06752920e+00
5.28150573e-02 -1.53891176e-01 -2.20200688e-01 -1.10318506e+00
-1.11734605e+00 1.01582360e+00 5.62534273e-01 -1.89639032e-01
1.26999044e+00 5.17948344e-02 6.82805181e-01 6.14959776e-01
3.75502378e-01 -1.29885662e+00 1.28816739e-01 6.64255202e-01
5.28257906e-01 -1.42857814e+00 -4.26353961e-01 -1.31511390e-01
-9.04971838e-01 9.79429901e-01 1.08362460e+00 1.23972215e-01
7.27406323e-01 3.75781432e-02 3.25504571e-01 6.83318973e-02
-5.23376822e-01 -8.47532526e-02 -6.08852459e-03 1.02106261e+00
1.36288014e-04 -4.49755713e-02 4.91291642e-01 9.74636376e-01
9.80503112e-02 7.93729573e-02 1.85321122e-01 9.73476887e-01
-2.07834080e-01 -1.41949546e+00 -1.95581824e-01 2.19850451e-01
-1.08179584e-01 -3.58769409e-02 -4.26241040e-01 7.85418093e-01
8.76980368e-03 6.93926990e-01 -1.59410417e-01 -1.03117734e-01
5.61566174e-01 3.95215213e-01 6.31494895e-02 -8.09668124e-01
-2.99317360e-01 -1.06832378e-01 -1.34665921e-01 -2.22928226e-01
-4.44467008e-01 -2.95065612e-01 -1.18787134e+00 -6.67194799e-02
-2.68196940e-01 5.29924691e-01 4.77820754e-01 9.84896183e-01
7.48820543e-01 4.76428211e-01 7.13906288e-01 -4.30959642e-01
-7.09041655e-01 -1.07743502e+00 -6.90373182e-01 4.57518935e-01
2.69379050e-01 -7.95539737e-01 -5.14499582e-02 4.95997041e-01] | [10.413710594177246, 3.1192824840545654] |
40aad670-ad52-45eb-9619-9e9555015c79 | ntu-rgbd-a-large-scale-dataset-for-3d-human | 1604.02808 | null | http://arxiv.org/abs/1604.02808v1 | http://arxiv.org/pdf/1604.02808v1.pdf | NTU RGB+D: A Large Scale Dataset for 3D Human Activity Analysis | Recent approaches in depth-based human activity analysis achieved outstanding
performance and proved the effectiveness of 3D representation for
classification of action classes. Currently available depth-based and
RGB+D-based action recognition benchmarks have a number of limitations,
including the lack of training samples, distinct class labels, camera views and
variety of subjects. In this paper we introduce a large-scale dataset for RGB+D
human action recognition with more than 56 thousand video samples and 4 million
frames, collected from 40 distinct subjects. Our dataset contains 60 different
action classes including daily, mutual, and health-related actions. In
addition, we propose a new recurrent neural network structure to model the
long-term temporal correlation of the features for each body part, and utilize
them for better action classification. Experimental results show the advantages
of applying deep learning methods over state-of-the-art hand-crafted features
on the suggested cross-subject and cross-view evaluation criteria for our
dataset. The introduction of this large scale dataset will enable the community
to apply, develop and adapt various data-hungry learning techniques for the
task of depth-based and RGB+D-based human activity analysis. | ['Tian-Tsong Ng', 'Gang Wang', 'Amir Shahroudy', 'Jun Liu'] | 2016-04-11 | ntu-rgbd-a-large-scale-dataset-for-3d-human-1 | http://openaccess.thecvf.com/content_cvpr_2016/html/Shahroudy_NTU_RGBD_A_CVPR_2016_paper.html | http://openaccess.thecvf.com/content_cvpr_2016/papers/Shahroudy_NTU_RGBD_A_CVPR_2016_paper.pdf | cvpr-2016-6 | ['3d-human-action-recognition'] | ['computer-vision'] | [ 2.30408028e-01 -5.09581387e-01 -4.99610245e-01 -4.39205557e-01
-6.26502156e-01 -1.29964843e-01 4.11229759e-01 -2.33458459e-01
-2.95411021e-01 4.45671171e-01 7.45270431e-01 4.03626055e-01
-6.62222952e-02 -5.03754616e-01 -2.09653154e-01 -7.36546099e-01
-4.00990069e-01 6.73898235e-02 3.64692390e-01 -2.42715571e-02
1.18888691e-01 7.18781114e-01 -1.77451563e+00 5.77528000e-01
-1.31142586e-01 1.37806451e+00 -2.47280747e-01 7.16716647e-01
1.73813954e-01 1.31495702e+00 -6.62117302e-01 7.29075670e-02
3.25447351e-01 -7.34530091e-01 -7.65054703e-01 5.15981615e-01
4.28067327e-01 -5.52134871e-01 -7.80203223e-01 3.15192550e-01
9.47455704e-01 3.22525740e-01 4.83675301e-01 -1.35270357e+00
-3.46820444e-01 -1.72077432e-01 -4.37864363e-01 7.68552661e-01
9.18529093e-01 2.15465590e-01 5.59746563e-01 -4.75505173e-01
6.50743961e-01 1.04539084e+00 6.33537233e-01 7.79966354e-01
-5.08206844e-01 -3.87681156e-01 1.32853165e-01 5.89901984e-01
-1.05786681e+00 -3.27778548e-01 1.04118299e+00 -5.53969204e-01
1.40939963e+00 -4.64410670e-02 1.22246814e+00 1.64882505e+00
4.14389879e-01 1.33800006e+00 1.09783101e+00 -2.22304836e-01
1.65014848e-01 -5.56050658e-01 -3.29253711e-02 8.79251897e-01
-1.57024160e-01 1.77487694e-02 -9.49011505e-01 7.25168455e-03
8.92593682e-01 4.45118129e-01 -9.99776199e-02 -7.83949077e-01
-1.42453253e+00 5.51448107e-01 1.42460346e-01 3.99611354e-01
-5.11600971e-01 2.89092422e-01 9.40583050e-01 8.32064524e-02
5.42159200e-01 -3.43458876e-02 -6.14789069e-01 -7.98045874e-01
-5.35159707e-01 2.20741376e-01 3.14237833e-01 5.56870878e-01
1.83502063e-01 8.59198198e-02 -3.28413963e-01 7.06084549e-01
1.73782513e-01 5.18772244e-01 9.95651662e-01 -1.16634297e+00
5.10843694e-01 9.21516657e-01 -9.93652269e-02 -9.84229922e-01
-8.25573564e-01 1.27957389e-01 -8.90138805e-01 3.10281545e-01
4.27922994e-01 1.61219388e-01 -7.85592616e-01 1.37168479e+00
4.07404274e-01 8.18628818e-02 1.54440431e-02 1.01100194e+00
8.45928073e-01 2.12964818e-01 8.81740749e-02 -5.93486913e-02
1.24100363e+00 -9.40169692e-01 -5.98459840e-01 -4.55773361e-02
8.95865500e-01 -1.25888899e-01 9.60411549e-01 4.76215839e-01
-8.76220047e-01 -8.16556633e-01 -9.54682648e-01 1.09118689e-02
-5.11067569e-01 4.35785621e-01 8.37080002e-01 6.93581104e-01
-6.68735564e-01 6.60418808e-01 -1.22688031e+00 -5.05579174e-01
7.00279117e-01 2.68549085e-01 -8.26781750e-01 1.08530775e-01
-1.10276818e+00 7.02320516e-01 1.58774555e-01 9.69312638e-02
-1.08269501e+00 -1.54937729e-01 -9.71696079e-01 -5.96759796e-01
2.89005846e-01 -5.47447443e-01 9.91442144e-01 -9.47605908e-01
-1.69151199e+00 1.24095249e+00 1.51588798e-01 -4.83548373e-01
4.27027404e-01 -4.57323283e-01 -4.34259802e-01 3.49844009e-01
4.68375832e-02 3.14346403e-01 6.35633767e-01 -3.39675516e-01
-6.83033764e-01 -8.76121640e-01 5.03374636e-02 3.09266388e-01
-1.79706857e-01 -1.83271855e-04 -3.83323431e-01 -7.46709704e-01
2.23759994e-01 -9.43685591e-01 -8.24832842e-02 1.43387884e-01
6.52562901e-02 -3.89292866e-01 6.82975113e-01 -5.98469317e-01
9.57560360e-01 -2.14379549e+00 3.63234818e-01 -2.42629156e-01
6.86665624e-02 2.54570305e-01 5.00261858e-02 2.97459304e-01
-9.93488356e-02 -3.72461170e-01 2.43865177e-01 -6.48323223e-02
-1.88651323e-01 3.20399433e-01 2.41314501e-01 7.57271469e-01
-6.99658692e-02 8.51576924e-01 -8.17294478e-01 -5.77710927e-01
6.75677121e-01 5.22015095e-01 -2.37161472e-01 2.36845851e-01
2.13162854e-01 5.67833543e-01 -7.26094842e-01 1.14390767e+00
-9.62741449e-02 -1.43312037e-01 -5.61166890e-02 -2.98919439e-01
3.24904889e-01 8.98184627e-02 -1.13619661e+00 2.22694492e+00
-1.92412093e-01 6.01702213e-01 -5.99222541e-01 -1.17818391e+00
7.46772945e-01 4.22806621e-01 1.28195798e+00 -1.02815783e+00
3.62355858e-01 -1.46319792e-01 -3.96267116e-01 -8.81098151e-01
1.54715804e-02 1.61133185e-01 -2.58783281e-01 5.08050799e-01
3.27388316e-01 4.71634358e-01 2.73692816e-01 -2.79972941e-01
1.45203531e+00 7.09339678e-01 6.01879776e-01 2.58641660e-01
6.59490824e-01 -2.52694160e-01 6.33215964e-01 4.40858155e-01
-9.78885293e-01 5.62390506e-01 4.19365674e-01 -1.02273202e+00
-6.65879965e-01 -7.77372897e-01 2.33618170e-01 1.27884340e+00
5.36553599e-02 -3.55965406e-01 -4.79321182e-01 -9.20940518e-01
-5.34783453e-02 -8.47539082e-02 -9.70786035e-01 -2.91188031e-01
-7.58217394e-01 -7.31904626e-01 7.22962856e-01 1.10863507e+00
9.90668178e-01 -1.26347911e+00 -1.25681424e+00 2.69524217e-01
-2.52563685e-01 -1.16679478e+00 -1.54090345e-01 1.26290038e-01
-1.23270702e+00 -1.46973932e+00 -8.89247239e-01 -4.61823493e-01
2.90206194e-01 2.86404669e-01 1.01140046e+00 -4.92049068e-01
-4.92377222e-01 9.34831321e-01 -7.03320265e-01 -2.06686839e-01
3.79619263e-02 -2.91671634e-01 2.88757563e-01 2.92869434e-02
5.97862720e-01 -4.93365854e-01 -7.27226019e-01 4.55483168e-01
-5.49704194e-01 -3.27728599e-01 5.89555442e-01 6.16743445e-01
8.12927127e-01 -9.26025584e-02 1.64389521e-01 -1.37108073e-01
1.98988214e-01 -1.23735115e-01 -2.17946470e-02 2.66499609e-01
-2.22358048e-01 -7.89642632e-02 2.87580285e-02 -5.43016553e-01
-7.64058292e-01 3.62500727e-01 -2.95849908e-02 -5.40436745e-01
-3.49923521e-01 9.33460444e-02 -2.30082378e-01 2.25284491e-02
6.50485814e-01 3.52195442e-01 9.21262279e-02 -5.37556410e-01
1.00824788e-01 4.25322771e-01 4.65257227e-01 -2.93348402e-01
2.38755774e-02 6.22618020e-01 2.38743290e-01 -6.70704007e-01
-8.27144980e-01 -6.80356264e-01 -1.10449386e+00 -5.98516345e-01
1.25848842e+00 -1.28441060e+00 -7.49758184e-01 1.06676841e+00
-7.21560359e-01 -3.94615412e-01 -4.15157884e-01 7.79814780e-01
-1.07383430e+00 4.39937621e-01 -6.15676045e-01 -7.32929528e-01
-2.32916400e-01 -9.03954566e-01 1.43866241e+00 4.66149598e-02
-2.27196813e-01 -7.35615432e-01 4.56106156e-01 8.18942904e-01
-7.49721949e-04 7.60098696e-01 4.01226282e-01 -3.57976586e-01
-1.85032785e-01 -5.79071641e-01 2.88384527e-01 4.65750128e-01
4.36854184e-01 -2.71570206e-01 -7.99647748e-01 -7.28469789e-02
-5.98305091e-03 -7.68600404e-01 7.23088741e-01 6.04535818e-01
1.30471992e+00 1.90258488e-01 -3.47341329e-01 4.14868742e-01
9.74981427e-01 2.57274598e-01 7.93693006e-01 4.41342711e-01
8.31691980e-01 2.88266629e-01 7.61633694e-01 7.87794411e-01
7.62817860e-02 9.30603564e-01 2.92018682e-01 -5.52921556e-02
-8.51206034e-02 1.97743042e-03 5.83492339e-01 4.98620808e-01
-7.19676554e-01 -1.01686113e-01 -8.42618585e-01 3.00870001e-01
-1.94512570e+00 -1.37226331e+00 2.43290812e-01 1.97445452e+00
4.83353645e-01 1.66330963e-01 7.96960175e-01 5.27847707e-01
2.19129071e-01 3.87063086e-01 -5.69694638e-01 6.84793759e-03
-6.18511215e-02 6.53490648e-02 4.14321214e-01 -2.33634531e-01
-1.64336979e+00 6.19454920e-01 6.46276522e+00 5.71046472e-01
-9.39766884e-01 6.75603449e-02 5.16409218e-01 -5.02446771e-01
7.15118647e-01 -7.49945700e-01 -7.15737462e-01 2.92006731e-01
9.11400855e-01 2.63116419e-01 -1.20255232e-01 1.16395998e+00
3.63848984e-01 -2.56848574e-01 -1.09604645e+00 1.55438769e+00
5.46741128e-01 -1.04618967e+00 -2.06216976e-01 4.25593928e-02
5.06939471e-01 1.45904735e-01 -3.42671931e-01 2.87572265e-01
-1.61442384e-01 -7.23060489e-01 4.64543134e-01 6.82806671e-01
3.68294775e-01 -5.53625762e-01 6.76940441e-01 9.71863568e-02
-1.33734214e+00 -3.29874039e-01 -1.62661031e-01 -2.65781015e-01
1.76622167e-01 2.00521588e-01 -1.09904870e-01 5.09234428e-01
1.11032844e+00 1.52879286e+00 -7.73183703e-01 5.88886976e-01
4.74192314e-02 1.49846956e-01 -4.55635367e-03 -1.15626872e-01
2.43580207e-01 5.20011634e-02 1.09716453e-01 8.69828701e-01
1.95627123e-01 2.98268586e-01 2.47360155e-01 5.31872660e-02
2.93020308e-01 -1.04633063e-01 -8.76777887e-01 -2.38549381e-01
-3.79111230e-01 8.68824303e-01 -7.40580142e-01 -3.19706380e-01
-6.74898326e-01 1.28481865e+00 5.69671951e-02 3.13594341e-02
-9.84390318e-01 -1.32605005e-02 8.76406670e-01 2.99170054e-02
2.68811613e-01 -4.33226556e-01 2.12439641e-01 -1.29386866e+00
1.48716137e-01 -1.09271908e+00 7.41895318e-01 -7.64791012e-01
-1.15569901e+00 3.79108518e-01 1.57821849e-01 -1.77144122e+00
-4.66742009e-01 -9.22418714e-01 -1.72848314e-01 8.76731649e-02
-7.67870188e-01 -1.13270032e+00 -5.79058588e-01 1.02986515e+00
7.37898588e-01 -5.92772663e-01 8.78695667e-01 3.78522903e-01
-6.27602279e-01 1.34796903e-01 -1.01627462e-01 5.86414933e-01
4.96496409e-01 -9.08595204e-01 1.61941350e-01 5.01513302e-01
4.17563707e-01 -5.37480712e-02 1.10428087e-01 -4.92632657e-01
-1.53126633e+00 -8.32890034e-01 5.97524464e-01 -9.97194648e-01
3.07504088e-01 -1.89016208e-01 -2.22282201e-01 6.64347291e-01
-2.78155267e-01 3.97610724e-01 1.03239381e+00 -3.10606253e-03
-1.01751767e-01 -3.35600972e-01 -9.85626757e-01 2.10419059e-01
1.46825767e+00 -5.76836884e-01 -6.26640677e-01 4.78372425e-01
1.56639069e-01 -4.48116094e-01 -1.09027624e+00 6.19095981e-01
9.41940129e-01 -1.29872692e+00 1.29231811e+00 -8.77543271e-01
4.45326000e-01 -1.52560487e-01 -4.32148576e-01 -7.14387715e-01
-3.19397748e-01 -1.39676481e-02 -4.98066813e-01 7.67202914e-01
-2.08205208e-01 -3.40463184e-02 1.16152358e+00 2.97730535e-01
-8.68099928e-02 -1.03068507e+00 -1.08236170e+00 -7.69373894e-01
-4.19574291e-01 -6.58453703e-01 1.10954814e-01 5.82341373e-01
8.80347844e-03 1.51052782e-02 -6.52199328e-01 -3.78798127e-01
4.31035668e-01 7.85972029e-02 9.73105609e-01 -9.66901302e-01
-2.39144593e-01 -2.13440165e-01 -1.32293677e+00 -9.73901093e-01
2.29910538e-02 -3.79304588e-01 -3.20462763e-01 -1.51311529e+00
8.83385390e-02 1.10039033e-01 -4.61589605e-01 6.67187214e-01
2.52363414e-01 5.12739301e-01 -1.16875850e-01 1.70323029e-01
-9.70119119e-01 7.24839509e-01 1.24302793e+00 -3.26754749e-01
-1.43598750e-01 3.16534966e-01 -4.82525304e-02 8.08781326e-01
6.01054192e-01 -2.59565949e-01 -5.67690313e-01 -1.86984673e-01
-9.02622491e-02 5.92482053e-02 5.63109457e-01 -1.59120071e+00
-2.19607100e-01 -2.06790045e-01 9.84381378e-01 -7.23257720e-01
6.92767739e-01 -6.75764561e-01 1.90470610e-02 7.06782877e-01
-3.38049054e-01 -1.28745548e-02 -7.56312981e-02 7.49645710e-01
-3.85451347e-01 4.31561321e-01 5.72398365e-01 -6.54362738e-01
-1.31967568e+00 4.69361722e-01 -4.57316488e-01 1.94623008e-01
1.33761346e+00 -7.45280027e-01 2.04744250e-01 -2.43222579e-01
-9.48819160e-01 -2.43650064e-01 1.06459059e-01 7.97081172e-01
6.42000258e-01 -1.75437605e+00 -4.07080352e-01 2.85252512e-01
4.60838735e-01 -5.49650609e-01 4.52136457e-01 8.73382807e-01
-4.18592989e-01 5.23437142e-01 -9.52452600e-01 -7.47265100e-01
-1.49402618e+00 4.34699237e-01 6.13927841e-01 -3.44884187e-01
-8.57891202e-01 7.06439555e-01 -1.88567862e-01 -1.71394229e-01
5.28104365e-01 -5.50149560e-01 -3.14545244e-01 2.60399580e-01
6.06829286e-01 7.37136245e-01 -7.01633692e-02 -8.06379497e-01
-8.50768745e-01 8.11248064e-01 4.33095872e-01 1.26154661e-01
1.29178572e+00 4.41826647e-03 4.22275424e-01 7.84230113e-01
1.25717688e+00 -7.27370262e-01 -1.39075327e+00 -7.48041570e-02
-2.10297599e-01 -5.81442177e-01 -8.55694041e-02 -6.29098058e-01
-1.31958449e+00 9.08272803e-01 1.23280847e+00 -2.41757393e-01
1.28020632e+00 1.02465332e-01 8.17172706e-01 5.55632830e-01
6.26139939e-01 -1.45677853e+00 7.39680648e-01 2.83582896e-01
7.50071287e-01 -1.52389383e+00 3.57905805e-01 1.05411321e-01
-8.82780135e-01 1.03529942e+00 6.32769227e-01 -1.46904185e-01
6.85174406e-01 -1.88736871e-01 2.03556612e-01 -4.94492412e-01
-4.85988706e-01 -3.26430410e-01 2.45867655e-01 7.74505496e-01
5.51669359e-01 -1.30326897e-01 -7.29670599e-02 3.37886423e-01
2.98721701e-01 5.05818129e-01 9.03507024e-02 1.39841557e+00
-2.51624942e-01 -1.04664278e+00 -1.77007988e-02 3.37238133e-01
-4.78730381e-01 7.07778037e-01 -4.83179808e-01 9.95567441e-01
2.89830565e-01 6.61796510e-01 -1.05137810e-01 -5.64659894e-01
6.59219146e-01 3.23353022e-01 8.87730360e-01 -2.32113779e-01
-4.52386528e-01 -1.73981786e-01 5.63111231e-02 -1.29268825e+00
-1.31757212e+00 -8.90053332e-01 -9.96192813e-01 -8.86048004e-02
2.59952933e-01 -5.36596656e-01 3.71627718e-01 1.10036480e+00
5.23542225e-01 4.58996356e-01 4.88110512e-01 -1.08659256e+00
-2.39185154e-01 -9.52813864e-01 -8.19766939e-01 7.70172238e-01
1.41002953e-01 -1.11122167e+00 -1.19857371e-01 2.38458410e-01] | [7.850456237792969, 0.41867539286613464] |
dcc50ec1-c223-4139-aac7-f3b58edb1a4b | context-aware-attention-for-understanding | 1809.08726 | null | https://arxiv.org/abs/1809.08726v2 | https://arxiv.org/pdf/1809.08726v2.pdf | Context-Aware Attention for Understanding Twitter Abuse | The original goal of any social media platform is to facilitate users to indulge in healthy and meaningful conversations. But more often than not, it has been found that it becomes an avenue for wanton attacks. We want to alleviate this issue and hence we try to provide a detailed analysis of how abusive behavior can be monitored in Twitter. The complexity of the natural language constructs makes this task challenging. We show how applying contextual attention to Long Short Term Memory networks help us give near state of art results on multiple benchmarks abuse detection data sets from Twitter. | ['Kilol Gupta', 'Tuhin Chakrabarty'] | 2018-09-24 | null | null | null | null | ['abuse-detection'] | ['natural-language-processing'] | [-2.06409708e-01 -3.15466784e-02 -7.37963676e-01 -4.37563986e-01
-3.02413523e-01 -5.98188400e-01 6.68578744e-01 4.68347698e-01
-5.22651792e-01 8.01854551e-01 5.65348446e-01 -4.59918588e-01
1.50799036e-01 -6.67414606e-01 -1.75123841e-01 -7.30900317e-02
-3.17631811e-01 8.88994262e-02 -7.39376694e-02 -6.69439852e-01
3.95152658e-01 6.60630882e-01 -8.15300167e-01 3.51946324e-01
4.35743541e-01 3.61404240e-01 -8.66227984e-01 3.74611378e-01
-4.46123749e-01 1.16729987e+00 -6.95366859e-01 -7.82058358e-01
-2.37673029e-01 -1.96860746e-01 -1.10864079e+00 -1.83386356e-01
2.89918482e-01 -5.78649223e-01 -3.39242905e-01 1.23863292e+00
4.43844050e-01 1.05380371e-01 2.63984859e-01 -1.25066245e+00
-5.78490257e-01 7.02625275e-01 -5.50344825e-01 9.31336522e-01
6.10401452e-01 7.84053952e-02 8.78716648e-01 -3.75335485e-01
6.75323367e-01 1.37135279e+00 7.87672043e-01 5.20116150e-01
-1.00637996e+00 -9.60828722e-01 2.83137765e-02 -6.57236800e-02
-1.04507947e+00 -7.25311220e-01 7.52773106e-01 -4.11792427e-01
9.82923985e-01 3.87948304e-01 7.22892225e-01 1.87241209e+00
6.74383193e-02 7.21516311e-01 1.07636571e+00 -3.30086611e-02
-1.65010005e-01 3.00805360e-01 9.17661130e-01 4.60472792e-01
3.45676005e-01 -3.92308950e-01 -4.64165211e-01 -8.47801626e-01
1.18778653e-01 2.47144833e-01 -1.21323848e-02 4.78475720e-01
-3.44181806e-01 1.44001949e+00 4.66550171e-01 1.13110840e+00
-2.09225997e-01 3.10897995e-02 8.58714938e-01 2.89476722e-01
6.89737737e-01 6.07726157e-01 1.72867656e-01 -7.06637561e-01
-8.48460793e-01 3.23322147e-01 1.05799496e+00 1.40884012e-01
1.00570321e-01 -1.91393495e-01 5.01927696e-02 1.00267577e+00
1.48194209e-01 2.96534061e-01 4.24598187e-01 -7.83031583e-01
2.71950990e-01 4.97526884e-01 -8.55499953e-02 -1.68989456e+00
-5.64278305e-01 -3.19173276e-01 -6.43477082e-01 -2.73406684e-01
6.78284347e-01 -2.93211222e-01 -3.31419915e-01 1.85413742e+00
1.12209305e-01 4.74549234e-02 -7.17367589e-01 5.08854687e-01
4.13989753e-01 2.58092850e-01 5.43984473e-01 -3.86420250e-01
1.22647226e+00 -3.85031909e-01 -1.25807488e+00 -5.42078555e-01
8.67149770e-01 -5.34337699e-01 9.47216511e-01 2.89904386e-01
-1.02191603e+00 3.00260216e-01 -8.18594038e-01 -9.73717421e-02
-7.40122736e-01 -1.02012372e+00 7.96760261e-01 1.37348616e+00
-5.90313077e-01 8.51679206e-01 -7.20151246e-01 -6.31488264e-01
8.14616084e-01 -5.94102107e-02 -4.34393644e-01 1.69223428e-01
-1.52136886e+00 1.07503378e+00 -1.66616648e-01 -1.32497117e-01
-3.86360079e-01 -4.01913136e-01 -7.29860485e-01 -3.85691643e-01
3.84524673e-01 -1.01749696e-01 1.12565243e+00 -9.58252072e-01
-1.01119339e+00 1.14018023e+00 -1.48334056e-01 -4.68989789e-01
3.50170076e-01 -4.00698096e-01 -6.73354685e-01 1.15889236e-01
1.99969813e-01 -1.64999850e-02 7.31529236e-01 -5.62782288e-01
1.77552551e-01 -6.69879317e-01 1.92786217e-01 -4.64694649e-01
-1.08645737e+00 1.04676735e+00 6.05305843e-02 -7.58636594e-01
-5.41654229e-01 -1.00876820e+00 -7.57083893e-02 -1.02102481e-01
-8.99280727e-01 -2.92462498e-01 9.73893940e-01 -9.53897774e-01
1.69350147e+00 -2.05091310e+00 -3.13422889e-01 3.28934759e-01
4.33900416e-01 5.97938716e-01 3.61029357e-01 5.39299726e-01
-1.27679020e-01 1.04837680e+00 -1.97139129e-01 -5.67278326e-01
-7.95284510e-02 3.43686342e-01 -4.50094610e-01 8.22336733e-01
-7.94707090e-02 9.06146169e-01 -1.06185269e+00 -3.75200808e-01
-1.38625100e-01 6.78331375e-01 -4.69034672e-01 8.03008303e-02
1.59175500e-01 3.08096319e-01 -5.42129099e-01 6.34664237e-01
5.10293365e-01 -3.09734583e-01 4.25311923e-02 3.67389619e-01
-6.80253059e-02 6.79799080e-01 -4.81995314e-01 1.14869487e+00
-7.06934109e-02 8.86181295e-01 4.19661075e-01 -5.73945880e-01
2.77357846e-01 1.52173012e-01 4.36776161e-01 -6.32830203e-01
4.89878237e-01 -1.96870063e-02 1.59695894e-01 -7.98688769e-01
2.52483249e-01 -5.73058486e-01 -3.25372934e-01 8.26307416e-01
-4.62414742e-01 2.60893732e-01 -9.15365890e-02 4.49322909e-01
1.16119218e+00 -6.30144179e-01 3.88052523e-01 -1.69470668e-01
2.33315155e-01 -2.26314068e-01 2.22665474e-01 7.80417502e-01
-7.49006271e-01 1.15554050e-01 8.80181909e-01 -3.30442429e-01
-6.77048981e-01 -7.14264929e-01 -2.94567108e-01 1.39212894e+00
-4.95970756e-01 -6.23789966e-01 -8.28427017e-01 -8.95406067e-01
5.61840534e-02 8.65466237e-01 -7.65976250e-01 -1.71542436e-01
-6.21308446e-01 -9.74180579e-01 1.03163815e+00 2.33992323e-01
6.36698186e-01 -1.03936291e+00 -2.66770214e-01 1.76900223e-01
-2.62080938e-01 -1.11721730e+00 -4.09578949e-01 -3.14208746e-01
-5.13346851e-01 -1.15726793e+00 -3.53932798e-01 1.06607601e-01
8.70822668e-02 1.84735030e-01 1.09722006e+00 6.17621601e-01
-3.49130571e-01 1.58839971e-01 -2.57859617e-01 -3.79689991e-01
-3.69537026e-01 2.60104477e-01 1.66800708e-01 1.06451906e-01
7.53724456e-01 -1.06309175e+00 -2.78263599e-01 5.77539392e-02
-1.05933845e+00 -7.12209404e-01 -3.03223670e-01 1.52336732e-01
-4.38696772e-01 -4.77047861e-01 7.26207197e-01 -1.48701608e+00
1.08386111e+00 -1.20503175e+00 8.90544280e-02 -4.39892679e-01
-4.20253754e-01 -5.20802021e-01 3.18394452e-01 -3.75635475e-01
-4.56943035e-01 -6.38715088e-01 -7.13857532e-01 -3.58667970e-02
-2.14878887e-01 4.60514277e-01 2.83324152e-01 -2.88093351e-02
7.93630183e-01 -2.40015164e-01 1.15943335e-01 -6.51597381e-01
1.59575969e-01 9.42272305e-01 -2.19765499e-01 -1.10327981e-01
6.26492381e-01 7.66042590e-01 -2.52124310e-01 -1.45773578e+00
-1.30191612e+00 -4.65996206e-01 -2.73655087e-01 -1.10244155e-01
9.71130431e-01 -3.37695748e-01 -9.67080116e-01 3.89090002e-01
-1.22550285e+00 -1.61916494e-01 2.83588499e-01 -1.93386286e-01
7.99821615e-02 5.05050361e-01 -1.10416257e+00 -1.02269661e+00
-3.25010091e-01 -6.58969879e-01 4.20500785e-01 -2.38714606e-01
-1.04695249e+00 -1.26259041e+00 3.64952832e-01 5.36259472e-01
9.73126948e-01 5.97582757e-01 6.01451457e-01 -1.23258018e+00
1.57352477e-01 -4.01619285e-01 -2.24248037e-01 3.36838186e-01
3.29303265e-01 -2.87999492e-02 -1.20471752e+00 -2.30301008e-01
3.23662966e-01 -5.55648148e-01 7.63284504e-01 -9.99082699e-02
1.22840238e+00 -8.66098404e-01 -3.67593914e-01 1.26081750e-01
9.51106608e-01 -3.77590239e-01 7.50772297e-01 2.83702701e-01
6.53783441e-01 8.23181272e-01 -2.10444909e-02 4.25605267e-01
1.03674270e-01 6.49964154e-01 3.44159096e-01 1.94565877e-01
2.14328125e-01 -1.95499972e-01 4.52540725e-01 3.06126535e-01
2.82719553e-01 -4.60568145e-02 -1.11931109e+00 5.25600493e-01
-1.51920295e+00 -1.38140047e+00 -2.79294908e-01 1.76993239e+00
8.15058053e-01 3.67542535e-01 7.58152544e-01 8.10573623e-02
6.04565620e-01 7.03451097e-01 -2.75518954e-01 -9.16460693e-01
5.40565960e-02 9.88074169e-02 3.02675575e-01 8.76203001e-01
-1.14603746e+00 6.81175053e-01 7.34889936e+00 7.59857237e-01
-1.10537124e+00 6.51971400e-01 9.58222866e-01 -5.56538284e-01
-2.13730726e-02 -4.64647532e-01 -7.60844290e-01 6.82168067e-01
1.39116549e+00 -5.07249124e-02 4.30085838e-01 7.49838710e-01
3.89976442e-01 -2.14171913e-02 -7.91385770e-01 9.22578514e-01
2.56108969e-01 -1.02082860e+00 -3.21456015e-01 1.87310010e-01
2.46773824e-01 3.36913168e-01 1.85030624e-01 2.61395603e-01
1.31591573e-01 -1.46133935e+00 -1.14194946e-02 2.55475342e-01
2.12890714e-01 -7.11919785e-01 3.83715689e-01 6.19425237e-01
-3.57371747e-01 -1.60959959e-01 2.50488333e-02 -5.83102584e-01
4.33189929e-01 7.71428645e-01 -5.72244346e-01 -3.07691634e-01
5.49443185e-01 9.26092684e-01 -5.72608411e-01 8.30297947e-01
-2.60614067e-01 9.56346393e-01 -5.03228247e-01 -2.97791898e-01
4.40528214e-01 1.05327792e-01 8.83245289e-01 1.53425133e+00
-2.84705251e-01 -2.48404108e-02 -5.11831976e-02 7.11888254e-01
-4.01646823e-01 2.20464259e-01 -1.27018058e+00 -7.08747566e-01
4.28082161e-02 1.21461177e+00 -4.42744434e-01 -9.59143788e-02
-3.33846509e-01 7.47985303e-01 6.41916156e-01 1.24694174e-02
-7.17112184e-01 4.51860484e-03 1.03036058e+00 7.98609376e-01
-4.11369503e-01 -2.31445953e-01 -1.70480609e-01 -1.09663832e+00
-1.89955115e-01 -1.01890349e+00 6.38150871e-01 -4.03414816e-01
-1.65153897e+00 2.58933842e-01 3.21907667e-03 -3.61944258e-01
-4.73554164e-01 -2.44863823e-01 -7.27793515e-01 5.50670683e-01
-1.34594476e+00 -8.13469768e-01 1.61688894e-01 5.09115100e-01
2.18932524e-01 3.01979035e-01 8.02545488e-01 6.28176272e-01
-7.49173641e-01 4.92866784e-01 -4.22978818e-01 5.99480510e-01
6.20971859e-01 -6.96450710e-01 1.76659092e-01 5.94786942e-01
-5.73693849e-02 1.03861797e+00 9.42795932e-01 -9.58144784e-01
-1.08171189e+00 -9.83134806e-01 1.14005101e+00 -1.05023468e+00
1.18905199e+00 -5.56401134e-01 -1.07155859e+00 1.11592495e+00
4.68267173e-01 -1.90833718e-01 1.10449469e+00 5.13544321e-01
-6.85724437e-01 4.49244678e-01 -1.37517738e+00 7.14952409e-01
1.16940939e+00 -8.15515101e-01 -6.20018303e-01 7.95712173e-01
5.60885370e-01 1.15094624e-01 -4.03578311e-01 -2.00766936e-01
3.92078102e-01 -1.06606114e+00 8.07992518e-01 -1.22329307e+00
4.24406528e-01 5.53411782e-01 2.39891931e-01 -9.87922132e-01
-5.19497469e-02 -1.09857881e+00 -1.17772721e-01 1.42798293e+00
3.00308466e-01 -8.06985557e-01 6.11140370e-01 1.10972095e+00
3.85599762e-01 -4.51318741e-01 -1.05270994e+00 -5.96252441e-01
5.74312627e-01 -7.82714427e-01 6.34551868e-02 1.65985107e+00
6.42979920e-01 4.86508638e-01 -7.44289517e-01 -2.55555958e-01
5.12280345e-01 -3.44259590e-01 5.04005551e-01 -9.23642159e-01
-1.22908488e-01 -5.85220397e-01 -1.24192573e-01 -5.11019111e-01
3.08137655e-01 -8.59027445e-01 -7.38481104e-01 -8.80329430e-01
4.93315786e-01 -1.68397978e-01 -1.27568588e-01 6.55573785e-01
3.13432142e-02 6.68322027e-01 1.63178533e-01 2.91137360e-02
-7.16729939e-01 1.27786130e-01 8.12901378e-01 -6.98262900e-02
-2.56001592e-01 -2.49348190e-02 -1.12254906e+00 8.80056977e-01
9.64626908e-01 -7.50709236e-01 5.64843006e-02 -3.98161337e-02
8.95600140e-01 -1.55114546e-01 4.32028294e-01 -5.92694223e-01
-1.12066053e-01 -1.73781872e-01 -3.56552228e-02 -1.18485741e-01
5.71421504e-01 -5.49464643e-01 -4.10827160e-01 5.63074708e-01
-4.09307539e-01 1.11775972e-01 2.85805970e-01 4.77856100e-01
7.34400526e-02 -3.65472317e-01 8.92365336e-01 -1.43299371e-01
1.27894476e-01 2.41129324e-01 -7.00763762e-01 3.55648369e-01
7.68023908e-01 2.87981242e-01 -5.01281142e-01 -9.50622857e-01
-1.06355608e+00 9.66964215e-02 3.36650312e-01 4.79133755e-01
3.28987598e-01 -1.15149331e+00 -4.77059841e-01 -2.98817605e-02
-2.37278432e-01 -1.06431508e+00 5.48367202e-02 1.10615492e+00
-5.31101460e-03 3.53037536e-01 -1.30474299e-01 -7.72700906e-02
-1.33140123e+00 6.52200222e-01 4.76125211e-01 -2.00504333e-01
-4.06299382e-01 7.00414360e-01 -5.15800834e-01 -1.33220181e-01
8.20058584e-02 1.92669362e-01 -5.29906631e-01 4.96857643e-01
1.20439744e+00 5.48940599e-01 -1.15579389e-01 -9.27582860e-01
-4.98887718e-01 -8.86061862e-02 -3.40049595e-01 -8.66699871e-03
1.54431331e+00 -5.92962094e-02 -6.36969268e-01 5.94186485e-01
1.67980778e+00 5.01091361e-01 -1.99644566e-01 3.82168032e-02
2.34780550e-01 -8.98402989e-01 -1.15004480e-01 -5.07429183e-01
-8.25079083e-01 9.64305580e-01 2.82770783e-01 1.05648637e+00
4.53565538e-01 3.35845537e-02 1.17971992e+00 3.23746830e-01
4.30319682e-02 -8.74941945e-01 2.02386618e-01 5.93981862e-01
1.05215645e+00 -1.28578389e+00 -9.75888316e-03 -5.09580314e-01
-2.94502944e-01 7.95457006e-01 2.11317971e-01 -3.28373373e-01
1.00911880e+00 1.53354883e-01 -1.44395858e-01 -6.04465485e-01
-2.18382701e-01 1.80928096e-01 -2.89755136e-01 3.26105267e-01
7.11674392e-01 -1.47216648e-01 -4.93639410e-01 5.74579716e-01
-1.79466859e-01 -1.07006192e-01 6.39533460e-01 8.47067893e-01
-4.83422488e-01 -1.23385036e+00 -3.43174964e-01 6.08102322e-01
-1.44340098e+00 -1.13788113e-01 -1.02226186e+00 8.20680141e-01
-3.00904095e-01 1.20137405e+00 -9.57402810e-02 -1.23524956e-01
8.15805569e-02 5.21260083e-01 4.55547310e-02 -7.61209607e-01
-1.15906346e+00 -3.39373529e-01 7.75236011e-01 -8.46207023e-01
-2.57532209e-01 -7.21651554e-01 -8.66252780e-01 -9.69565451e-01
8.85896236e-02 3.74736004e-02 5.37015676e-01 1.11188149e+00
3.93345147e-01 -1.90455206e-02 4.16025490e-01 -3.98658931e-01
-5.03758132e-01 -1.05182004e+00 -4.13997173e-01 9.19928253e-01
6.69668317e-01 -5.25752664e-01 -6.95811272e-01 -7.19467998e-01] | [8.711833953857422, 10.490084648132324] |
8e98f969-ba2c-468a-989f-ea7f71ea2f50 | i-see-you-a-vehicle-pedestrian-interaction | 2211.09342 | null | https://arxiv.org/abs/2211.09342v1 | https://arxiv.org/pdf/2211.09342v1.pdf | I see you: A Vehicle-Pedestrian Interaction Dataset from Traffic Surveillance Cameras | The development of autonomous vehicles arises new challenges in urban traffic scenarios where vehicle-pedestrian interactions are frequent e.g. vehicle yields to pedestrians, pedestrian slows down due approaching to the vehicle. Over the last years, several datasets have been developed to model these interactions. However, available datasets do not cover near-accident scenarios that our dataset covers. We introduce I see you, a new vehicle-pedestrian interaction dataset that tackles the lack of trajectory data in near-accident scenarios using YOLOv5 and camera calibration methods. I see you consist of 170 near-accident occurrences in seven intersections in Cusco-Peru. This new dataset and pipeline code are available on Github. | ['Harley Vera', 'Edwin Alvarez', 'Patricia Condori', 'Jorshinno Sumire', 'Hanan Quispe'] | 2022-11-17 | null | null | null | null | ['camera-calibration'] | ['computer-vision'] | [-8.08235228e-01 -2.31672496e-01 -4.09386009e-01 -4.22501564e-01
-5.84757388e-01 -5.18410683e-01 8.15275013e-01 -5.19090295e-02
-5.05671680e-01 8.43285859e-01 2.89816577e-02 -6.43928051e-01
1.40528038e-01 -8.86198223e-01 -8.60791862e-01 -5.68363428e-01
-1.17493533e-02 4.87334758e-01 6.72698617e-01 -4.96786445e-01
-2.26911619e-01 5.24495900e-01 -1.82568252e+00 3.26873213e-02
6.88844860e-01 2.16632739e-01 1.51136220e-01 1.00281549e+00
2.18882963e-01 3.93811733e-01 5.29595092e-02 -4.41006660e-01
3.00141633e-01 1.28656670e-01 -3.32288623e-01 -2.31765717e-01
6.81671679e-01 -5.98436892e-01 -8.82310808e-01 6.46847188e-01
2.73489863e-01 2.35516027e-01 7.69830048e-01 -1.94687855e+00
-1.15268394e-01 1.83960155e-01 -1.57818779e-01 5.87993503e-01
2.80340731e-01 8.85667324e-01 4.55963135e-01 -9.48899508e-01
6.94155335e-01 1.46472931e+00 8.44260991e-01 2.85089940e-01
-1.04653740e+00 -7.91133881e-01 -1.43724069e-01 6.62534595e-01
-1.46888590e+00 -5.55538595e-01 2.68188566e-01 -8.07654858e-01
1.19771743e+00 -2.41058059e-02 5.78512907e-01 1.43713415e+00
1.07101656e-01 7.67890394e-01 3.59440267e-01 1.21421017e-01
-2.99687665e-02 4.16970372e-01 6.42616868e-01 1.36392877e-01
4.19639230e-01 5.80916047e-01 -1.19886838e-01 2.09310710e-01
5.37398234e-02 -1.25855148e-01 5.05603909e-01 -1.49788126e-01
-9.44913149e-01 9.11948681e-01 2.90145397e-01 -1.65042534e-01
-2.63651609e-01 2.74901807e-01 5.48848987e-01 2.41829343e-02
-2.01718416e-03 -5.72804928e-01 -4.29956585e-01 -5.30288637e-01
-6.93606853e-01 9.27542806e-01 5.07533848e-01 1.59957469e+00
1.02617013e+00 -4.41154204e-02 8.80334824e-02 4.70727265e-01
3.76923949e-01 1.00782573e+00 -4.41466838e-01 -1.12006688e+00
8.86139333e-01 3.64176422e-01 2.00896204e-01 -9.58117485e-01
-6.28993034e-01 2.04971179e-01 -6.02032781e-01 1.58863440e-01
6.75759375e-01 -3.19060951e-01 -6.29171312e-01 1.33606052e+00
4.94452655e-01 3.56443048e-01 6.44946769e-02 6.38112545e-01
1.41945660e+00 6.30900621e-01 6.65298581e-01 2.31735319e-01
1.14325571e+00 -9.23116565e-01 -6.60807848e-01 -2.60153294e-01
7.91826069e-01 -4.03271735e-01 5.72677076e-01 -1.91753030e-01
-9.17206049e-01 -8.58652771e-01 -5.03497064e-01 -5.19544072e-02
-1.06367755e+00 -1.36656193e-02 3.92818898e-01 6.83224916e-01
-7.82265306e-01 4.09458093e-02 -6.90928459e-01 -8.85363281e-01
6.30800843e-01 2.19751537e-01 -6.09079063e-01 9.08748135e-02
-1.34148228e+00 1.10261607e+00 2.29975507e-01 -7.49427080e-02
-1.18567371e+00 -9.77679908e-01 -1.04073107e+00 -3.63769591e-01
2.99852014e-01 -6.16609573e-01 1.18961537e+00 -3.10958028e-01
-7.83052683e-01 6.26316011e-01 -4.37886626e-01 -6.58870339e-01
9.14680541e-01 -9.20569971e-02 -7.39148915e-01 -2.55373895e-01
5.10263979e-01 9.58451748e-01 8.75311643e-02 -1.27911973e+00
-1.21725178e+00 -1.47234276e-01 1.08656839e-01 -1.51460692e-01
5.94591677e-01 1.50654763e-01 -5.74422359e-01 -7.03881867e-03
-8.02794993e-01 -1.23720479e+00 -4.36492980e-01 -2.60678560e-01
-4.06491637e-01 -4.25344408e-01 1.31160009e+00 -4.87751484e-01
1.18820143e+00 -2.10157299e+00 -4.90236849e-01 5.18463850e-02
6.16719089e-02 6.69489980e-01 -8.86162966e-02 7.52750218e-01
-3.74143459e-02 1.19972959e-01 -2.06362993e-01 -2.27338195e-01
6.90914094e-02 3.96739662e-01 2.33132485e-02 6.14686728e-01
1.27087414e-01 1.05724621e+00 -1.05274260e+00 -5.03779888e-01
9.75360453e-01 5.79632699e-01 -4.58301038e-01 -1.13219336e-01
2.06030771e-01 5.63258886e-01 -2.69201040e-01 5.64694941e-01
1.25065124e+00 6.48683429e-01 -4.04310465e-01 2.05772996e-01
-8.33475173e-01 4.33816165e-02 -1.03972590e+00 8.85158122e-01
-2.33953953e-01 1.33485687e+00 -1.35642052e-01 -7.13850737e-01
3.63158971e-01 1.06222227e-01 5.05994737e-01 -7.70707190e-01
8.80448073e-02 4.33909781e-02 3.05861328e-02 -9.15377200e-01
5.13288736e-01 4.92179602e-01 -1.49342477e-01 -4.69476938e-01
-1.95634499e-01 -9.48725920e-03 6.30995095e-01 2.27338746e-01
8.91933858e-01 5.36814146e-02 2.14110523e-01 -2.74039239e-01
6.85841084e-01 6.88068092e-01 4.97268409e-01 7.33778179e-01
-7.85000801e-01 4.90668863e-01 3.46911758e-01 -5.60226083e-01
-1.37925494e+00 -1.43277264e+00 -5.97512126e-01 5.73669314e-01
2.29926124e-01 -3.62048805e-01 -7.68546343e-01 -4.47091073e-01
4.05257791e-01 1.12148929e+00 -3.30023736e-01 2.45964918e-02
-7.10066199e-01 -4.91606683e-01 6.58536613e-01 4.43213195e-01
7.18633294e-01 -8.09376001e-01 -3.76517475e-01 2.43451342e-01
-1.46699250e-01 -1.51681936e+00 -1.56140178e-01 -3.94249618e-01
-2.75513798e-01 -1.36112571e+00 -3.35130215e-01 -4.51843262e-01
2.89370596e-01 7.24068642e-01 1.10128105e+00 -5.46570942e-02
-2.77966350e-01 4.33893889e-01 -1.31066710e-01 -5.96375525e-01
-3.84967566e-01 1.65891722e-01 2.21979752e-01 -3.23351979e-01
1.14688981e+00 -1.71124458e-01 -6.00294650e-01 6.66341245e-01
-2.32110471e-01 -1.13478012e-01 6.56633154e-02 2.61648983e-01
2.50258625e-01 5.91810755e-02 5.02005398e-01 -5.31890213e-01
1.09618632e-02 -1.29197335e+00 -8.15134227e-01 -3.30403596e-01
-1.37139082e-01 -6.18476808e-01 4.43747818e-01 -2.90518347e-03
-1.00925052e+00 2.91084617e-01 -5.01508713e-01 -9.23004672e-02
-1.08423269e+00 -1.55122718e-03 -2.99312443e-01 2.89381444e-01
3.17144334e-01 -2.61976928e-01 -1.74350977e-01 -2.17098758e-01
4.44233805e-01 8.53007197e-01 6.47491693e-01 -1.27808794e-01
9.03379023e-01 6.65431261e-01 1.60393029e-01 -1.43919992e+00
-1.27307743e-01 -8.06026757e-01 -1.09997737e+00 -7.19105244e-01
1.35005724e+00 -1.28554547e+00 -1.21413040e+00 5.05931616e-01
-1.17470527e+00 -4.99927938e-01 1.71014462e-02 7.07463145e-01
-5.27235866e-01 1.74348220e-01 -4.82898727e-02 -9.79192734e-01
4.12269711e-01 -1.54539454e+00 8.64234447e-01 4.43443328e-01
-2.05976918e-01 -1.02971864e+00 1.08067617e-01 2.51497984e-01
2.35087171e-01 2.88320720e-01 2.36326531e-01 -1.09321252e-01
-8.10064018e-01 -2.36850575e-01 -4.50920075e-01 -6.31729737e-02
-1.03713110e-01 6.13461614e-01 -9.06358063e-01 1.65172517e-01
-1.03826141e+00 4.47731465e-01 8.58091831e-01 8.23284745e-01
7.14555979e-01 1.26357079e-01 -9.49711561e-01 4.56968218e-01
1.16839778e+00 3.97822320e-01 8.29639554e-01 7.17909753e-01
7.54549444e-01 9.55769956e-01 7.91323423e-01 2.00803220e-01
1.38982058e+00 1.05954051e+00 4.06136453e-01 2.89843641e-02
-4.89733636e-01 -2.52667010e-01 2.42734984e-01 2.15451524e-01
-7.68524855e-02 -2.02949435e-01 -1.43014991e+00 1.03307080e+00
-2.14426875e+00 -1.42262256e+00 -1.46929693e+00 1.98658347e+00
3.76095921e-02 6.93303645e-02 6.89170837e-01 -3.22571158e-01
6.05499029e-01 -3.10417354e-01 -3.06724995e-01 -4.28174406e-01
-1.58307165e-01 -6.11472607e-01 1.13234353e+00 9.17126119e-01
-1.33893466e+00 1.21939158e+00 7.07306242e+00 6.57174468e-01
-4.18047279e-01 2.58376002e-01 4.17238027e-01 4.31710780e-02
-4.65555079e-02 1.58127964e-01 -1.57703292e+00 8.23517680e-01
1.52782440e+00 3.24960500e-02 6.66761249e-02 9.10978913e-01
1.08927858e+00 -5.98548412e-01 -8.67722750e-01 1.07342005e+00
-3.81154567e-01 -1.20121694e+00 -3.07076603e-01 1.92038015e-01
4.64388132e-01 3.45471382e-01 -6.43755868e-02 5.77902138e-01
6.16913855e-01 -8.05484354e-01 7.89964020e-01 8.46144021e-01
5.14973104e-01 -1.07507920e+00 7.32665658e-01 4.53653693e-01
-1.43477440e+00 -1.23007491e-01 -2.74238527e-01 1.10705262e-02
7.24651098e-01 2.10123762e-01 -8.00304651e-01 3.93969387e-01
1.17849302e+00 1.02614594e+00 -8.74586642e-01 1.22136831e+00
8.88361037e-02 5.86278677e-01 -3.09338480e-01 1.60661235e-01
6.01896405e-01 -5.11761785e-01 8.91424358e-01 1.61965311e+00
2.35559702e-01 4.23653051e-02 -7.81340972e-02 6.28916979e-01
5.13639271e-01 -3.25092733e-01 -1.54873574e+00 6.24166608e-01
5.29106081e-01 1.08621895e+00 -1.02917135e-01 -3.01267266e-01
-7.76042700e-01 2.82714754e-01 -4.00749259e-02 4.63349700e-01
-1.37223494e+00 -1.12778977e-01 1.45441902e+00 5.24597526e-01
4.82463464e-02 -6.01880372e-01 -2.39005521e-01 -6.54012561e-01
-4.28006321e-01 -3.48701984e-01 1.37801468e-01 -7.46327996e-01
-8.35161150e-01 1.69156060e-01 9.04116750e-01 -1.36011338e+00
-2.55224794e-01 -6.02542162e-01 -8.37828159e-01 6.26550496e-01
-1.48584151e+00 -1.36199594e+00 -4.59972888e-01 4.78530020e-01
9.60893273e-01 -4.15230900e-01 1.23466969e-01 8.55026007e-01
-9.65860844e-01 4.39836174e-01 1.09295405e-01 8.62297881e-03
4.89825189e-01 -9.41165864e-01 9.11562264e-01 7.52918601e-01
-6.54847383e-01 3.71011585e-01 8.92255425e-01 -7.57791638e-01
-1.15398729e+00 -1.41254652e+00 1.23345125e+00 -1.12705350e+00
6.95177674e-01 -3.65269065e-01 -4.92990136e-01 1.07172024e+00
4.59217548e-01 -9.45575610e-02 3.42407435e-01 -3.32982332e-01
1.80727869e-01 -1.91591695e-01 -1.04092813e+00 7.79678464e-01
1.07883716e+00 -3.27343136e-01 -2.35959604e-01 2.90131211e-01
3.39993924e-01 -2.77216166e-01 -4.90148306e-01 9.41768065e-02
5.26930749e-01 -8.41636717e-01 1.26951492e+00 -6.75297081e-01
-4.04032692e-03 -5.83143294e-01 -1.90787315e-01 -1.13658047e+00
-3.07403386e-01 -2.86801845e-01 1.05991945e-01 1.40526855e+00
5.40227532e-01 -6.63640022e-01 4.55090970e-01 9.77915823e-01
-4.05009866e-01 2.79103555e-02 -1.30053651e+00 -8.41359615e-01
2.84849972e-01 -9.93654966e-01 7.87943423e-01 4.20125097e-01
-4.70526665e-01 -7.16460543e-03 -2.91073143e-01 4.78796989e-01
9.39906061e-01 -8.57798219e-01 1.59927022e+00 -1.17061460e+00
6.53945506e-01 -4.58642453e-01 -6.87758207e-01 -8.83446038e-01
3.51070464e-01 -6.34404600e-01 -9.86885279e-02 -1.59318542e+00
1.52884081e-01 -5.63523531e-01 6.46898031e-01 1.53139427e-01
3.82038616e-02 4.46674079e-02 -4.23990889e-03 -7.21985698e-02
-4.48846817e-01 4.63261545e-01 7.98426449e-01 -3.48545551e-01
-4.13525775e-02 2.77232319e-01 -1.60016730e-01 8.12440217e-01
8.09296548e-01 -4.53026533e-01 -1.44638166e-01 -2.45624736e-01
8.83755740e-03 -8.72486830e-02 8.75172555e-01 -1.13418961e+00
2.80808479e-01 -3.46635461e-01 3.55487950e-02 -1.48796690e+00
2.76750475e-01 -1.18890309e+00 7.14654207e-01 2.49596789e-01
2.56677032e-01 2.90004879e-01 5.70462346e-01 5.66650629e-01
-1.80381425e-02 -1.08581759e-01 7.90677488e-01 1.41103184e-02
-1.13718081e+00 5.83169937e-01 -1.11009216e+00 -2.17063740e-01
1.73295593e+00 -4.89638239e-01 -3.99786919e-01 -2.42203280e-01
-6.43239141e-01 7.84230709e-01 3.53095412e-01 8.25688958e-01
2.66390473e-01 -1.55273747e+00 -8.72976184e-01 1.00793047e-02
2.71524817e-01 -4.63250466e-02 4.94078994e-01 9.50871408e-01
-7.53367901e-01 9.18685555e-01 -3.36831510e-01 -8.25783312e-01
-1.32816994e+00 6.12781525e-01 4.61390108e-01 4.90289807e-01
-7.66920865e-01 2.53233016e-01 1.82842836e-01 -8.05842638e-01
-1.33689836e-01 -1.37201354e-01 -4.93015140e-01 1.67453796e-01
4.83736098e-01 1.21567810e+00 -1.89303383e-01 -1.71119797e+00
-5.71830451e-01 5.17016888e-01 3.80809754e-01 1.69910491e-01
1.02941072e+00 -8.69328082e-01 3.42790484e-01 4.60679412e-01
1.24983358e+00 -2.93273658e-01 -1.34910750e+00 2.94238836e-01
-1.17452912e-01 -6.24019802e-01 -1.22302666e-01 -4.72058505e-02
-8.15663874e-01 8.83857965e-01 7.65979350e-01 1.86004475e-01
4.48944926e-01 1.52268186e-01 9.27324355e-01 2.42000088e-01
3.02589476e-01 -1.24393773e+00 -8.71043921e-01 8.49264264e-01
6.51349366e-01 -1.92797935e+00 -3.00808311e-01 -4.32929456e-01
-8.14137816e-01 9.41709459e-01 9.25620735e-01 2.22020131e-02
1.20864940e+00 7.55182654e-02 -1.21166363e-01 -1.76528201e-01
-5.84161222e-01 -7.66309083e-01 -1.38654385e-03 1.24698091e+00
5.10691553e-02 4.10011888e-01 -1.63224176e-01 3.13513279e-01
-3.10808837e-01 -1.84696183e-01 5.84294796e-01 2.86662877e-01
-2.27700025e-01 -8.82233560e-01 -3.89675736e-01 2.06202030e-01
7.03173354e-02 1.54263020e-01 -3.08776479e-02 1.32296741e+00
4.95918453e-01 1.50261605e+00 3.21242720e-01 -2.60500908e-01
7.65537858e-01 -3.20722073e-01 -2.48780072e-01 -9.11757946e-02
-3.91904712e-01 -5.06778061e-01 4.99441564e-01 -5.45588315e-01
-2.33584508e-01 -1.26338685e+00 -1.29332840e+00 -1.13625896e+00
1.10756546e-01 -1.93309307e-01 6.28248215e-01 8.05628300e-01
3.08885068e-01 3.77112746e-01 6.50624156e-01 -1.31218374e+00
3.64108473e-01 -7.82838106e-01 -2.43186697e-01 3.11849684e-01
6.20174527e-01 -9.93490875e-01 -2.22428024e-01 7.16929436e-02] | [6.103032112121582, 0.86481773853302] |
524fe954-2336-4bf2-a008-149439574cfd | rdfnet-regional-dynamic-fista-net-for | 2302.02519 | null | https://arxiv.org/abs/2302.02519v1 | https://arxiv.org/pdf/2302.02519v1.pdf | RDFNet: Regional Dynamic FISTA-Net for Spectral Snapshot Compressive Imaging | Deep convolutional neural networks have recently shown promising results in compressive spectral reconstruction. Previous methods, however, usually adopt a single mapping function for sparse representation. Considering that different regions have distinct characteristics, it is desirable to apply various mapping functions to adjust different regions' transformations dynamically. With this in mind, we first introduce a regional dynamic way of using Fast Iterative Shrinkage-Thresholding Algorithm (FISTA) to exploit regional characteristics and derive dynamic sparse representations. Then, we propose to unfold the process into a hierarchical dynamic deep network, dubbed RDFNet. The network comprises multiple regional dynamic blocks and corresponding pixel-wise adaptive soft-thresholding modules, respectively in charge of region-based dynamic mapping and pixel-wise soft-thresholding selection. The regional dynamic block guides the network to adjust the transformation domain for different regions. Equipped with the adaptive soft-thresholding, our proposed regional dynamic architecture can also learn appropriate shrinkage scale in a pixel-wise manner. Extensive experiments on both simulated and real data demonstrate that our method outperforms prior state-of-the-arts. | ['Jianan Li', 'Shaocong Dong', 'Tingfa Xu', 'Shiyun Zhou'] | 2023-02-06 | null | null | null | null | ['spectral-reconstruction'] | ['computer-vision'] | [ 4.14154530e-01 -3.64422083e-01 -2.69758612e-01 -4.14117724e-01
-5.57467103e-01 -2.24887982e-01 2.06782252e-01 -3.74401778e-01
-2.60754138e-01 4.31626171e-01 3.43300670e-01 8.43507499e-02
-5.37774712e-02 -6.93991899e-01 -7.40082502e-01 -9.69102800e-01
-6.94384053e-02 -5.45874313e-02 2.98224062e-01 -3.17223877e-01
1.90252766e-01 3.55974525e-01 -1.14375544e+00 3.40099454e-01
8.84390950e-01 1.07940161e+00 4.45076376e-01 2.08552644e-01
9.94594693e-02 6.43711805e-01 -2.53282208e-02 4.24282759e-01
6.78543448e-01 -5.99793553e-01 -2.48925045e-01 2.36937344e-01
3.64927590e-01 -2.83674806e-01 -3.88278782e-01 1.25440562e+00
5.21227241e-01 2.77565122e-01 3.07748020e-01 -5.50600410e-01
-5.98715365e-01 1.06601930e+00 -9.91359890e-01 4.88676727e-01
3.33409198e-02 2.03346312e-01 8.63849699e-01 -1.00069594e+00
5.81902027e-01 9.78779316e-01 6.75244212e-01 2.86283582e-01
-1.23002768e+00 -8.24152470e-01 4.76336867e-01 2.02465132e-01
-1.58536541e+00 -6.95051849e-01 1.44893718e+00 -2.06834853e-01
5.18391669e-01 3.54452357e-02 8.37899387e-01 8.27953160e-01
-3.37103941e-02 5.60328424e-01 1.17538750e+00 -3.39331001e-01
3.14118892e-01 -3.37411612e-01 -2.78461516e-01 7.63280869e-01
3.56637947e-02 -1.37779057e-01 -7.62093365e-01 9.92602855e-02
1.20219243e+00 -6.20276295e-03 -7.26486325e-01 -6.60986781e-01
-1.50102353e+00 6.60924017e-01 7.28200018e-01 5.93316853e-01
-6.85638666e-01 4.79721397e-01 3.75929058e-01 4.46631461e-01
1.14277706e-01 -7.04589859e-02 -2.62093991e-01 3.93805683e-01
-1.43241453e+00 -2.92138979e-02 4.37467724e-01 5.55307388e-01
9.23370063e-01 5.69661498e-01 -1.40477985e-01 1.09458458e+00
4.07323807e-01 2.66370624e-01 7.36632586e-01 -8.29301178e-01
4.85307693e-01 2.14092955e-01 -3.06453016e-02 -1.46780550e+00
-3.80192876e-01 -9.14787352e-01 -1.30431044e+00 1.37655765e-01
3.78565900e-02 -7.71730840e-02 -1.09860396e+00 1.76372850e+00
4.02484715e-01 5.35178542e-01 -1.88307658e-01 1.22983944e+00
4.31198418e-01 8.10321569e-01 -2.54801158e-02 -4.39135462e-01
1.05727291e+00 -8.14748645e-01 -6.35611355e-01 -7.90347308e-02
2.38411576e-01 -5.84268689e-01 7.03189075e-01 3.67312849e-01
-1.26618266e+00 -7.03671813e-01 -1.40995872e+00 1.52728304e-01
1.55691594e-01 1.81592643e-01 3.95333171e-01 2.86979914e-01
-1.25650597e+00 3.53128016e-01 -9.21885967e-01 -8.48394856e-02
4.15887415e-01 2.82978445e-01 9.97123569e-02 2.20688015e-01
-1.34559131e+00 3.81508708e-01 5.43968439e-01 5.01578867e-01
-1.09866726e+00 -7.78649747e-01 -8.28775048e-01 2.83465445e-01
3.73924583e-01 -8.75813842e-01 9.43687379e-01 -1.36585176e+00
-1.60537004e+00 6.39327228e-01 -3.66222560e-02 -5.46878219e-01
5.07805765e-01 8.46670941e-02 -5.37852705e-01 4.11070973e-01
4.19336893e-02 5.23488164e-01 1.48570228e+00 -1.43681669e+00
-5.50306201e-01 -2.12210253e-01 -4.23583180e-01 4.05483186e-01
-6.24646544e-01 -6.97415471e-02 -5.71168542e-01 -1.05187690e+00
7.90256023e-01 -6.38331532e-01 -6.44480288e-01 -4.29580808e-02
-2.13217050e-01 3.15241843e-01 8.65875483e-01 -5.96522748e-01
1.56015325e+00 -2.23674679e+00 4.77907538e-01 6.26458228e-01
4.35543001e-01 -6.52222037e-02 -9.34132654e-03 -1.38209835e-01
-3.88739407e-01 -2.98310101e-01 -5.83521426e-01 -1.05346471e-01
-3.34635913e-01 -1.84236184e-01 -3.21156919e-01 6.98880255e-01
-1.79031119e-01 5.81381679e-01 -7.22943246e-01 -5.67996919e-01
3.56087714e-01 6.58020735e-01 -5.47528744e-01 -2.05103830e-01
-1.09376580e-01 4.34873343e-01 -6.79673254e-01 8.61223161e-01
9.02639270e-01 -2.67095685e-01 4.25938517e-01 -7.62088120e-01
-2.94691324e-01 -1.86346799e-01 -1.55055964e+00 2.02084064e+00
-4.92292196e-01 4.19856459e-01 6.93291485e-01 -1.54260182e+00
1.06771421e+00 4.23522666e-02 8.60025525e-01 -6.54968679e-01
2.40885139e-01 5.17321706e-01 -6.77576959e-02 -1.60163298e-01
6.11505389e-01 -1.76083177e-01 5.14395023e-03 2.63394177e-01
-2.14841977e-01 1.50603786e-01 -4.75844368e-02 3.52354012e-02
9.73516881e-01 9.24858302e-02 2.68436879e-01 -4.10235524e-01
8.77966642e-01 -2.03303963e-01 7.06846595e-01 6.19777918e-01
-3.00975204e-01 7.97714114e-01 -8.60674866e-03 -6.09345257e-01
-8.83002877e-01 -9.12845433e-01 -1.14520736e-01 1.17234826e+00
4.48216051e-01 -4.88656163e-02 -5.86267710e-01 -1.87664360e-01
-1.35618746e-01 9.38749164e-02 -5.88703871e-01 -3.63170318e-02
-1.03594041e+00 -7.48319507e-01 9.44159105e-02 3.67963791e-01
9.45081592e-01 -1.07165420e+00 -7.24959671e-01 5.62205911e-01
-4.98059057e-02 -1.01272213e+00 -5.50580621e-01 4.56585765e-01
-9.61744070e-01 -4.67540234e-01 -9.45022821e-01 -1.12263417e+00
7.16631174e-01 6.32844329e-01 8.31470132e-01 5.29087186e-02
5.95442764e-03 1.02093063e-01 -4.15985942e-01 3.47951621e-01
-1.91579968e-01 2.85960078e-01 -5.26352823e-02 4.54701304e-01
-2.32703462e-01 -1.29896390e+00 -1.01709902e+00 2.20901981e-01
-1.11255610e+00 1.44373447e-01 7.52186596e-01 7.43488431e-01
1.03727841e+00 8.18855092e-02 5.42829871e-01 -8.95812988e-01
5.42915106e-01 -4.55165893e-01 -6.23253465e-01 2.63888240e-01
-4.31369245e-01 3.41781899e-02 9.52789307e-01 -3.91557455e-01
-9.88171637e-01 3.66302550e-01 1.06809519e-01 -8.31894696e-01
2.85487086e-01 6.71404481e-01 -9.99048259e-03 -3.83510947e-01
6.13518357e-01 6.04781568e-01 -2.07518905e-01 -2.23775730e-01
3.27958763e-01 5.03402412e-01 6.77927554e-01 -5.70376337e-01
1.15420294e+00 7.56490767e-01 -1.46872282e-01 -4.96062070e-01
-8.20525527e-01 -2.51362562e-01 -5.54808199e-01 -4.18783665e-01
7.32432961e-01 -1.35333252e+00 -3.03330153e-01 5.81259251e-01
-6.90099716e-01 -4.56070244e-01 -2.77341783e-01 1.25447139e-01
-5.37103713e-01 3.97523880e-01 -4.82486546e-01 -2.44313985e-01
-5.64081192e-01 -1.15403032e+00 8.75942409e-01 3.63661081e-01
2.24358529e-01 -8.19609165e-01 -8.04394949e-03 1.68897673e-01
8.04240167e-01 2.78380275e-01 5.84608257e-01 -9.98707637e-02
-7.43730545e-01 2.26139382e-01 -1.95710793e-01 2.67208010e-01
1.43625468e-01 -1.93586975e-01 -5.92173576e-01 -5.69310367e-01
1.76026106e-01 -1.10363550e-01 1.24388397e+00 8.08928549e-01
1.63319063e+00 -1.42326847e-01 -2.17417464e-01 1.35749114e+00
1.66714168e+00 -3.78756272e-03 6.59806967e-01 6.39849186e-01
8.18226635e-01 8.04597978e-03 4.44911242e-01 6.64564610e-01
2.98129171e-01 5.86651206e-01 4.47286040e-01 -2.72703260e-01
-2.22659141e-01 -1.91012826e-02 4.59475517e-01 1.08035016e+00
-7.28472099e-02 6.29026592e-02 -5.16456783e-01 4.76023227e-01
-1.84125423e+00 -1.01815176e+00 3.31768662e-01 1.96165884e+00
9.56839800e-01 7.48917758e-02 -1.72308013e-01 -1.42694741e-01
9.81055021e-01 9.41232145e-01 -7.42624044e-01 -1.23003721e-01
-2.20987096e-01 2.27995872e-01 5.88374555e-01 2.60419220e-01
-1.19806159e+00 9.27776933e-01 5.52117252e+00 9.36777413e-01
-1.53036070e+00 2.03540415e-01 8.29750478e-01 -1.94630548e-01
-5.62068284e-01 5.53273670e-02 -5.32822251e-01 4.92469400e-01
4.41095173e-01 2.71677095e-02 6.81610405e-01 6.26930773e-01
4.82404679e-01 -1.13832802e-02 -5.01036167e-01 1.04942834e+00
5.79203814e-02 -1.74561274e+00 1.19276278e-01 -3.28770965e-01
1.16150403e+00 -1.01770740e-03 1.98439673e-01 5.64956339e-03
2.01940343e-01 -6.63939714e-01 9.78162110e-01 4.98158365e-01
7.50398934e-01 -8.06736648e-01 1.71824321e-01 5.07772118e-02
-1.62026465e+00 -2.97086567e-01 -4.09741402e-01 4.68662679e-01
1.96980014e-01 9.00013924e-01 -2.23996267e-01 5.22677124e-01
9.68603551e-01 1.08606660e+00 -1.22355975e-01 7.28613615e-01
-7.08108470e-02 6.38318956e-01 -2.78358907e-01 3.61313969e-01
2.65229762e-01 -5.29095590e-01 6.67695642e-01 1.15393448e+00
5.68296492e-01 2.32012510e-01 2.35032216e-01 8.19501400e-01
-2.63515890e-01 1.20487865e-02 -1.16651103e-01 3.01433116e-01
6.37054443e-01 1.33000290e+00 -9.18152213e-01 -2.72940487e-01
-3.58768016e-01 9.75554645e-01 1.31543100e-01 6.15231216e-01
-9.33564603e-01 -1.55713752e-01 4.34368074e-01 2.00940773e-01
8.06237757e-01 -4.74496931e-01 -5.07652879e-01 -1.21683109e+00
4.63295244e-02 -9.98304904e-01 3.77092212e-01 -7.26353526e-01
-8.16246390e-01 4.95655119e-01 -2.45180592e-01 -1.56904244e+00
2.64931291e-01 1.28172729e-02 -6.46571994e-01 6.91634238e-01
-1.78803897e+00 -1.07349956e+00 -6.38065696e-01 8.30000401e-01
7.34319508e-01 -2.35691965e-01 8.04051310e-02 2.61130095e-01
-7.37408161e-01 4.96514201e-01 2.30425745e-01 -1.86380614e-02
6.83686197e-01 -6.94124043e-01 -1.74707294e-01 1.19265890e+00
-7.43083432e-02 6.42138898e-01 6.69106901e-01 -5.11423409e-01
-1.49756122e+00 -1.05908215e+00 1.69895440e-02 6.53675914e-01
8.95989358e-01 -2.49590911e-02 -9.80661333e-01 4.69685227e-01
1.56761765e-01 3.41081828e-01 1.81719616e-01 -2.86997974e-01
-2.34122261e-01 -6.09302878e-01 -1.04739046e+00 5.20508170e-01
1.15329182e+00 -3.35901886e-01 -1.49408564e-01 2.92481393e-01
7.91873753e-01 -6.08359754e-01 -8.17610145e-01 5.55197775e-01
4.37512815e-01 -1.05584574e+00 1.17173624e+00 7.88747966e-02
5.15923619e-01 -7.04058170e-01 -1.89048544e-01 -1.22825396e+00
-8.18012595e-01 -7.32334971e-01 -1.83158547e-01 8.21198523e-01
2.38113388e-01 -6.01170897e-01 8.16517830e-01 -1.63964450e-03
-4.84570771e-01 -8.95066321e-01 -9.84303474e-01 -3.23033601e-01
-3.17413867e-01 -1.94339052e-01 4.40028429e-01 1.17967451e+00
-5.11061072e-01 -2.36028701e-01 -4.95457619e-01 4.77708876e-01
6.37149334e-01 4.63156521e-01 4.08865482e-01 -7.33553290e-01
-3.48508358e-01 -6.35529876e-01 -2.20628470e-01 -1.25007880e+00
5.20599224e-02 -7.89717317e-01 1.69142976e-01 -1.23639619e+00
2.86843330e-01 -5.85268795e-01 -7.97551632e-01 5.53016543e-01
-2.25402266e-02 2.93517917e-01 5.20273335e-02 4.96223748e-01
-5.97218931e-01 7.22373068e-01 1.41618884e+00 -1.95965931e-01
-5.01046002e-01 -3.07146937e-01 -7.79142082e-01 5.60342729e-01
7.65739918e-01 -3.49458694e-01 -3.47380072e-01 -5.39366007e-01
2.52504110e-01 6.54702634e-02 2.70242363e-01 -1.31747448e+00
3.76072228e-01 -3.36934030e-01 5.66576660e-01 -4.93595481e-01
-1.45404143e-02 -9.11592484e-01 2.22264528e-01 4.67790782e-01
-4.39157426e-01 -2.05921516e-01 -3.83151993e-02 6.79863572e-01
-3.65014315e-01 4.81938422e-02 1.15120447e+00 -2.28401616e-01
-1.05656838e+00 4.75135654e-01 -2.52319515e-01 -2.24235967e-01
9.32488918e-01 -5.53656936e-01 1.52937412e-01 -3.51759404e-01
-8.66997957e-01 1.84597373e-01 4.01175141e-01 7.17492253e-02
7.64482975e-01 -1.31068885e+00 -7.11351633e-01 3.18929285e-01
-2.78429270e-01 2.02077270e-01 6.34915709e-01 1.09385324e+00
-6.75084412e-01 -1.19755574e-01 -2.59748429e-01 -7.02334762e-01
-9.37552333e-01 3.36086303e-01 5.86276889e-01 -2.25790650e-01
-7.55485177e-01 9.80026484e-01 1.72350835e-02 1.75149180e-02
5.74562810e-02 -3.16348523e-01 -9.68453586e-02 7.97163025e-02
1.86597034e-01 7.41800293e-02 -7.32275322e-02 -6.00691259e-01
-2.03807920e-01 9.32970822e-01 -9.93028134e-02 2.51461603e-02
1.65492499e+00 -3.50328535e-01 -3.22477341e-01 1.08500794e-01
1.08101463e+00 3.13301869e-02 -1.43111205e+00 -5.21287799e-01
-3.99295300e-01 -4.90452975e-01 4.61911500e-01 -4.24103588e-01
-1.89114249e+00 3.31274331e-01 8.27300131e-01 1.43645585e-01
1.79377258e+00 -2.41145596e-01 1.09040368e+00 1.77749321e-01
3.05306196e-01 -1.39140952e+00 1.52455017e-01 3.40324938e-01
9.09761667e-01 -9.99666214e-01 3.79443586e-01 -3.89852285e-01
-3.01110417e-01 1.16515028e+00 6.06087446e-01 -4.21035767e-01
7.11306393e-01 3.13296825e-01 2.03404389e-02 -1.51128933e-01
-4.02605712e-01 1.04569145e-01 1.81183890e-01 2.69729584e-01
2.51126021e-01 5.14173135e-02 -3.46044153e-01 2.44185835e-01
-3.25576030e-02 2.03672666e-02 3.08400452e-01 7.45089173e-01
-8.37859035e-01 -8.27995420e-01 -5.77137291e-01 4.05663610e-01
-1.62305124e-02 -1.01413377e-01 2.04511732e-01 4.99608576e-01
7.77616575e-02 4.17306304e-01 -2.65950412e-02 -3.86660516e-01
-4.11148146e-02 -2.97071517e-01 3.07211071e-01 -4.22391325e-01
-6.64143384e-01 3.47554326e-01 -3.82565260e-01 -6.54756844e-01
-7.22133100e-01 -7.63114214e-01 -1.36543965e+00 -5.52822463e-03
-1.32942311e-02 -2.81170249e-01 4.40900147e-01 7.03586936e-01
2.14224815e-01 7.13664055e-01 1.09255385e+00 -1.05279160e+00
-5.43277800e-01 -6.77746356e-01 -5.29505253e-01 1.27211645e-01
3.52939516e-01 -4.28278059e-01 -4.86164033e-01 1.77515820e-01] | [11.117714881896973, -2.004115104675293] |
b089badb-02d0-4edb-a154-bf3e6f7ba953 | automated-ischemic-stroke-lesion-segmentation | 2209.09546 | null | https://arxiv.org/abs/2209.09546v2 | https://arxiv.org/pdf/2209.09546v2.pdf | Automated ischemic stroke lesion segmentation from 3D MRI | Ischemic Stroke Lesion Segmentation challenge (ISLES 2022) offers a platform for researchers to compare their solutions to 3D segmentation of ischemic stroke regions from 3D MRIs. In this work, we describe our solution to ISLES 2022 segmentation task. We re-sample all images to a common resolution, use two input MRI modalities (DWI and ADC) and train SegResNet semantic segmentation network from MONAI. The final submission is an ensemble of 15 models (from 3 runs of 5-fold cross validation). Our solution (team name NVAUTO) achieves the top place in terms of Dice metric (0.824), and overall rank 2 (based on the combined metric ranking). | ['Andriy Myronenko', 'Daguang Xu', 'Yufan He', 'Dong Yang', 'Md Mahfuzur Rahman Siddique'] | 2022-09-20 | null | null | null | null | ['ischemic-stroke-lesion-segmentation'] | ['medical'] | [ 1.53806999e-01 2.18999326e-01 -2.92473465e-01 -5.33975303e-01
-1.30421579e+00 -7.18960464e-01 5.65145433e-01 -1.04563542e-01
-7.80847609e-01 7.30590820e-01 5.25236905e-01 -4.86382484e-01
-2.62680471e-01 -2.08074823e-01 -2.29061887e-01 -1.88654736e-01
-3.77724916e-01 8.55751753e-01 7.44179428e-01 1.33684233e-01
3.91150057e-01 7.85350442e-01 -4.74420726e-01 6.64876640e-01
9.11090493e-01 7.93004036e-01 1.93397552e-01 7.71820486e-01
6.41619414e-02 7.08354473e-01 -3.63889813e-01 -1.80262566e-01
4.47334707e-01 -5.70135117e-01 -1.44353449e+00 -4.08803225e-01
6.58562779e-01 -4.70569253e-01 -3.93518001e-01 8.45083654e-01
8.47399890e-01 -3.41229551e-02 9.69265223e-01 -8.70313764e-01
-2.05981731e-01 6.56860530e-01 -3.40052992e-01 1.18013299e+00
-1.82488412e-01 2.11411640e-01 4.33270842e-01 -6.80924416e-01
1.23865044e+00 8.99118066e-01 1.00228071e+00 5.81408381e-01
-1.04462564e+00 -3.91769826e-01 -4.96525802e-02 6.07769489e-01
-1.01374435e+00 -2.77564466e-01 3.65295745e-02 -7.08792150e-01
1.11628747e+00 3.92606854e-01 7.63637602e-01 1.04483414e+00
3.07507098e-01 1.12779462e+00 1.58442330e+00 1.01573125e-01
3.76260310e-01 -5.06962657e-01 6.27379060e-01 1.39844880e-01
2.37844333e-01 -1.28551513e-01 8.55451226e-02 -4.29232530e-02
7.08322406e-01 -2.72596031e-01 -2.84903318e-01 -2.83608437e-01
-1.41992533e+00 5.99615693e-01 7.85014510e-01 3.28474820e-01
-5.33652544e-01 -2.62992624e-02 5.63976049e-01 1.29222110e-01
4.17546749e-01 4.98551518e-01 -3.38731349e-01 -2.00794876e-01
-1.48686731e+00 3.75677049e-01 3.59191418e-01 5.23170769e-01
-3.12620431e-01 -3.49934608e-01 -4.46408689e-01 9.21967924e-01
-1.76328540e-01 2.46061087e-01 7.81828046e-01 -1.19865882e+00
3.02159131e-01 1.75053552e-01 -1.21485539e-01 -4.25898969e-01
-1.16020751e+00 -5.68470895e-01 -6.20382547e-01 4.11669612e-01
6.48636162e-01 -3.94803852e-01 -1.46073449e+00 1.31564546e+00
-2.50521153e-01 1.12274988e-02 -2.17294857e-01 1.19384325e+00
1.01935565e+00 -1.20124385e-01 3.31584513e-01 2.06613809e-01
1.13510072e+00 -1.08998132e+00 -3.20745289e-01 -3.14181536e-01
8.52560222e-01 -3.54071736e-01 7.06123352e-01 3.13141286e-01
-1.26819348e+00 -7.35788420e-02 -8.77951026e-01 8.55151266e-02
-3.81374151e-01 -1.98777504e-02 2.90061086e-01 7.02926397e-01
-1.52768135e+00 6.99524522e-01 -1.27952266e+00 -5.69871783e-01
1.15835714e+00 9.66957286e-02 -3.96412164e-01 -2.38963608e-02
-8.88185918e-01 1.68817461e+00 4.01498288e-01 -7.00844824e-02
-8.94010961e-01 -9.82404828e-01 -2.56277621e-01 -6.32320344e-01
-4.67432998e-02 -6.23250544e-01 1.12885153e+00 -5.13082623e-01
-1.10717869e+00 1.42225540e+00 2.40624938e-02 -8.88708353e-01
1.27487242e+00 -8.28286111e-02 -1.41306400e-01 8.09100211e-01
3.51907015e-01 1.03427339e+00 -9.38514397e-02 -9.29848850e-01
-5.05262673e-01 -6.74180686e-01 -1.68997332e-01 4.12978560e-01
3.77807468e-01 4.49786663e-01 -1.47509769e-01 -6.03209555e-01
4.93896604e-01 -6.89469874e-01 -6.62381113e-01 -2.12521106e-03
-3.38700444e-01 1.04496270e-01 5.45641124e-01 -1.26415586e+00
6.76156163e-01 -1.67840385e+00 1.04433097e-01 2.90155709e-01
7.03577936e-01 1.24098636e-01 -6.12717867e-02 -5.30732691e-01
-4.49653149e-01 4.48671103e-01 -6.98971272e-01 -2.94119984e-01
-3.28411102e-01 -8.85140449e-02 1.48881540e-01 6.39626443e-01
-1.41730383e-01 1.25932789e+00 -8.28703046e-01 -6.35645390e-01
1.92912638e-01 1.74382254e-02 -2.46076092e-01 -1.50754824e-01
5.07993639e-01 5.90159357e-01 -4.31685627e-01 1.75731421e-01
7.39828467e-01 -7.81639665e-02 -5.60521265e-04 -2.88408965e-01
-3.45733799e-02 9.56448093e-02 -7.18172908e-01 1.98453593e+00
7.84097016e-02 6.03070080e-01 -1.65802404e-01 -1.22039664e+00
5.07275045e-01 2.61860937e-01 1.09748423e+00 -8.00631762e-01
3.34562473e-02 4.54268426e-01 1.79447174e-01 -5.04379153e-01
-2.31843397e-01 8.80789533e-02 8.91949609e-02 5.61077297e-01
1.22657433e-01 -3.33377570e-01 3.33916605e-01 4.51539189e-01
1.32101226e+00 7.30990916e-02 6.81095272e-02 -6.78162456e-01
5.44577353e-02 4.30934787e-01 2.14950144e-01 1.13378763e+00
-9.76281583e-01 1.31742179e+00 8.60346913e-01 -5.12671530e-01
-1.22712493e+00 -1.55797088e+00 -5.02321720e-01 3.39043617e-01
-8.83923396e-02 2.56996378e-02 -1.26769483e+00 -9.89653170e-01
-3.98884356e-01 6.92299068e-01 -9.10564661e-01 5.84271885e-02
-9.14213538e-01 -1.13730299e+00 1.03233910e+00 8.91064167e-01
6.39022887e-01 -1.06091106e+00 -1.17258632e+00 3.97166722e-02
-4.35330331e-01 -1.09899688e+00 -3.74653488e-01 -4.21418287e-02
-1.29899073e+00 -1.35340989e+00 -1.41342545e+00 -6.99642658e-01
1.55916765e-01 5.34832384e-03 1.01074314e+00 -7.80850276e-02
-5.49376190e-01 2.05769897e-01 -4.09032792e-01 -2.31246889e-01
-7.31841326e-02 3.29552859e-01 -3.59900534e-01 -6.80915654e-01
1.65271208e-01 -4.57993448e-01 -1.06346607e+00 1.50446936e-01
-7.01652408e-01 2.21255735e-01 2.89061457e-01 4.54998791e-01
7.01394796e-01 -7.43720949e-01 8.07351649e-01 -6.11884713e-01
7.25047886e-01 -4.79547054e-01 -1.95363969e-01 3.92669410e-01
-4.73024249e-01 -5.84793329e-01 1.23203851e-01 5.27227633e-02
-6.43842816e-01 -1.63970247e-01 -1.34384677e-01 -9.14010629e-02
-5.61436951e-01 3.65829557e-01 4.17815536e-01 9.87515412e-03
9.12263632e-01 -1.53608784e-01 2.96605527e-01 -5.85041642e-01
4.98836368e-01 4.59949642e-01 9.10558820e-01 -3.10279429e-01
1.27635049e-02 5.68035483e-01 -4.31057736e-02 -3.30760747e-01
-8.23463261e-01 -2.90395826e-01 -1.15054703e+00 -3.08488220e-01
1.34408343e+00 -5.68353295e-01 8.37971792e-02 7.25041568e-01
-9.97703671e-01 -8.36299837e-01 -2.91152805e-01 8.06176484e-01
-8.09139729e-01 2.88832277e-01 -5.71988583e-01 3.93720344e-02
-6.39052927e-01 -1.66768074e+00 5.36782205e-01 1.57419145e-01
-3.82258266e-01 -8.99375200e-01 1.25769839e-01 4.14246649e-01
7.45531499e-01 7.04528630e-01 7.43567705e-01 -9.76794302e-01
1.45306224e-02 -1.33835420e-01 -7.92050540e-01 1.90947503e-01
-4.96694371e-02 -4.63519275e-01 -5.44795036e-01 3.84370587e-03
-1.54741481e-01 -1.46062806e-01 1.13666415e+00 1.30331469e+00
1.25971723e+00 5.11237442e-01 -2.44899809e-01 5.77398479e-01
1.10688233e+00 2.69875735e-01 7.22753108e-01 8.46177697e-01
5.20140707e-01 3.97613972e-01 -1.13959797e-01 -1.11846194e-01
4.85899985e-01 5.84902227e-01 2.00400680e-01 -3.81912172e-01
-9.06598866e-01 4.68617022e-01 -9.56461579e-02 4.02798027e-01
-1.44209161e-01 2.47609854e-01 -1.44835925e+00 7.28106976e-01
-1.60352409e+00 -8.44778180e-01 -5.90417922e-01 1.84004045e+00
6.35816097e-01 2.29293957e-01 5.39684713e-01 -4.32816386e-01
6.45926476e-01 -4.83677164e-02 -7.05266476e-01 -2.56661415e-01
-4.73326474e-01 4.42652047e-01 8.14000785e-01 5.22128463e-01
-1.29349089e+00 9.49206114e-01 7.81030846e+00 7.76154459e-01
-8.87563288e-01 6.62815392e-01 9.92530942e-01 -2.08574295e-01
1.54697532e-02 -1.56107396e-01 -1.17589556e-01 4.53536510e-01
1.06240261e+00 -1.83746204e-01 3.85090709e-01 3.53437603e-01
2.85027087e-01 -2.11683676e-01 -7.19826460e-01 6.30481064e-01
1.37896212e-02 -1.55213237e+00 -1.98900774e-01 -2.51352698e-01
8.11971903e-01 1.05815279e+00 -1.41575575e-01 -2.17335951e-02
2.40795940e-01 -1.24765873e+00 6.43437922e-01 8.60652566e-01
8.58616531e-01 -3.60151827e-01 9.03025568e-01 -1.37640452e-02
-5.68139791e-01 3.05599362e-01 6.40378101e-03 5.35916626e-01
6.39371395e-01 2.48869225e-01 -6.35630310e-01 5.50157547e-01
1.15917134e+00 8.73206556e-01 -8.13342810e-01 1.77178478e+00
-7.29309171e-02 7.67020226e-01 -3.74507755e-01 5.65413356e-01
6.02411091e-01 -9.88152027e-02 7.23710477e-01 1.44290459e+00
2.42198389e-02 3.93929899e-01 1.25340626e-01 7.41740525e-01
1.81431726e-01 8.95089507e-02 -2.32341900e-01 6.65040553e-01
2.69722473e-02 9.76359963e-01 -1.41756678e+00 -7.76170731e-01
-1.67305283e-02 8.64213526e-01 5.32997819e-03 2.51212239e-01
-5.76959610e-01 -3.30083787e-01 2.11619496e-01 -2.49942448e-02
-1.17104806e-01 -1.40684932e-01 -1.13558650e+00 -9.20311213e-01
-2.04316765e-01 -6.55011952e-01 6.93304181e-01 -1.00464952e+00
-1.23214388e+00 7.10644126e-01 4.46167946e-01 -6.18237615e-01
2.45835483e-02 -6.24292552e-01 -7.48921931e-01 1.10713279e+00
-1.40639603e+00 -5.50209582e-01 -2.73893923e-01 4.09414023e-01
4.71350998e-01 -1.36165440e-01 7.56559789e-01 4.83427912e-01
-4.47744608e-01 4.45438713e-01 7.33387619e-02 3.11207920e-01
6.69983983e-01 -1.43044949e+00 7.86023498e-01 7.90861189e-01
-3.75269532e-01 2.38240913e-01 2.16989294e-01 -8.43207777e-01
-3.23728025e-01 -9.29924309e-01 5.91701984e-01 -7.91906893e-01
7.34474659e-01 4.27392781e-01 -7.44919300e-01 6.49659276e-01
1.94383934e-01 -3.01186163e-02 4.31658357e-01 -5.02297521e-01
8.24256390e-02 5.16118169e-01 -1.47855723e+00 5.67887545e-01
1.26847148e+00 -1.81544870e-01 -9.87174451e-01 6.66760027e-01
3.66385669e-01 -7.54161596e-01 -1.21665323e+00 5.88703394e-01
6.43192232e-01 -7.30308533e-01 1.07244468e+00 -1.20582831e+00
4.42370176e-01 -7.73487911e-02 -6.16924092e-02 -1.32157207e+00
-1.81115150e-01 -7.37901255e-02 4.21814978e-01 1.50279820e-01
5.41163385e-01 -5.98459303e-01 7.61883438e-01 7.32267380e-01
-7.49990165e-01 -8.87167335e-01 -1.20210958e+00 -7.46391833e-01
9.00157452e-01 -5.44271231e-01 3.30575317e-01 7.91844308e-01
-8.48174393e-02 -3.63715321e-01 2.34054565e-01 -1.96498394e-01
9.36116040e-01 -2.23219499e-01 7.88243338e-02 -1.08555472e+00
3.57384354e-01 -1.27641511e+00 -5.27985156e-01 -5.95417678e-01
-1.09954968e-01 -1.71564436e+00 -2.42149666e-01 -2.30261326e+00
5.50503671e-01 -4.15836871e-01 -6.93341196e-01 7.38969684e-01
1.68196410e-02 6.99768186e-01 2.68105716e-01 3.16200376e-01
-5.61023831e-01 -2.32727658e-02 1.30765581e+00 -5.82716838e-02
-9.80297476e-02 -1.21277183e-01 -6.07058167e-01 7.26712883e-01
1.24954724e+00 -4.98678237e-01 -1.12666219e-01 -6.90900862e-01
-5.67100465e-01 -7.02582970e-02 6.52005553e-01 -1.02727187e+00
-3.19292173e-02 3.04816421e-02 5.74047804e-01 -7.18247592e-01
-1.63712353e-01 -3.37473869e-01 -1.38776839e-01 6.86297297e-01
-5.89432418e-01 1.24100670e-01 2.18112007e-01 -3.34600478e-01
2.49797449e-01 -3.93943518e-01 9.81159389e-01 -3.46425056e-01
-5.88146448e-01 4.34072316e-01 -6.25226676e-01 6.24958456e-01
7.75889874e-01 -4.46423352e-01 -4.96731967e-01 7.23165506e-03
-1.50732040e+00 3.89112383e-01 8.27505961e-02 4.17290062e-01
5.63200116e-01 -1.18258858e+00 -1.10094833e+00 -3.83971930e-01
-3.18785101e-01 -3.36401701e-01 4.97231364e-01 1.53949142e+00
-9.85554755e-01 5.23393393e-01 -8.16294014e-01 -6.03944421e-01
-8.71937811e-01 -2.68049002e-01 9.62933898e-01 -4.36486602e-01
-1.36649382e+00 7.41546750e-01 -2.45226085e-01 -5.25464714e-01
1.00071475e-01 -1.96961939e-01 -3.02965462e-01 1.48735508e-01
6.54624522e-01 7.02518463e-01 4.12072808e-01 -5.41622519e-01
-7.34063268e-01 4.59614694e-01 -1.82872400e-01 -5.68468273e-01
1.41588151e+00 6.95356578e-02 -5.91127314e-02 -1.10354565e-01
1.38097095e+00 -8.86250436e-01 -1.11323047e+00 9.13073123e-02
2.89075285e-01 -3.15014683e-02 4.59987789e-01 -1.44520617e+00
-1.33933854e+00 7.79206693e-01 1.34986377e+00 -1.92040280e-01
7.28276312e-01 1.11594945e-01 9.57911849e-01 5.42944781e-02
1.41360715e-01 -1.08593202e+00 -3.83307934e-01 3.98101866e-01
8.59645128e-01 -1.00333452e+00 -5.09110792e-03 1.09676242e-01
-9.05303299e-01 1.06982887e+00 3.14513415e-01 -4.80036795e-01
8.57958436e-01 1.29081517e-01 3.74255955e-01 -3.66563141e-01
-4.98618558e-02 1.01975398e-02 4.90072578e-01 8.54618967e-01
2.15713948e-01 3.57177794e-01 -8.75352979e-01 7.04714239e-01
-2.66874731e-01 1.15453802e-01 5.07254004e-01 6.35078430e-01
-4.33631569e-01 -7.82209754e-01 -1.20078428e-02 9.74368274e-01
-5.90545416e-01 -6.05420880e-02 -3.53763074e-01 6.41412318e-01
-6.28117174e-02 6.42210364e-01 -8.73454809e-02 9.54298228e-02
5.52403510e-01 -3.45553942e-02 5.93050897e-01 -1.81406960e-01
-6.55133963e-01 -8.64157081e-02 8.57086703e-02 -8.78152192e-01
-4.03720349e-01 -1.08604133e+00 -1.55868721e+00 1.32846996e-01
3.73216927e-01 -6.89401105e-02 7.95954466e-01 1.06727779e+00
2.46116459e-01 6.51585519e-01 2.32833654e-01 -8.34725916e-01
-1.94427565e-01 -1.00958943e+00 -3.89561027e-01 3.88356835e-01
-1.18331611e-01 -5.74802458e-01 -1.05047785e-01 -1.89263653e-02] | [14.282951354980469, -2.1158077716827393] |
b43ddd2d-b0cf-48fe-9d42-7db246849226 | differential-privacy-may-have-a-potential | 2306.17370 | null | https://arxiv.org/abs/2306.17370v1 | https://arxiv.org/pdf/2306.17370v1.pdf | Differential Privacy May Have a Potential Optimization Effect on Some Swarm Intelligence Algorithms besides Privacy-preserving | Differential privacy (DP), as a promising privacy-preserving model, has attracted great interest from researchers in recent years. Currently, the study on combination of machine learning and DP is vibrant. In contrast, another widely used artificial intelligence technique, the swarm intelligence (SI) algorithm, has received little attention in the context of DP even though it also triggers privacy concerns. For this reason, this paper attempts to combine DP and SI for the first time, and proposes a general differentially private swarm intelligence algorithm framework (DPSIAF). Based on the exponential mechanism, this framework can easily develop existing SI algorithms into the private versions. As examples, we apply the proposed DPSIAF to four popular SI algorithms, and corresponding analyses demonstrate its effectiveness. More interestingly, the experimental results show that, for our private algorithms, their performance is not strictly affected by the privacy budget, and one of the private algorithms even owns better performance than its non-private version in some cases. These findings are different from the conventional cognition, which indicates the uniqueness of SI with DP. Our study may provide a new perspective on DP, and promote the synergy between metaheuristic optimization community and privacy computing community. | ['Meiyi Xie', 'Hong Zhu', 'Zhiqiang Zhang'] | 2023-06-30 | null | null | null | null | ['metaheuristic-optimization'] | ['methodology'] | [ 9.07507837e-02 -1.73250288e-01 -2.41437882e-01 1.09388418e-01
-1.66751623e-01 -6.20026886e-01 3.84989738e-01 2.65538871e-01
-3.97177219e-01 7.88456142e-01 -4.16086167e-02 -4.34682854e-02
-3.47678512e-01 -9.90969598e-01 -2.06355721e-01 -1.32566464e+00
1.14264395e-02 -7.12136179e-02 -3.22913527e-02 -2.36108437e-01
2.78586030e-01 2.56554753e-01 -1.38563061e+00 -2.64309645e-01
1.32614422e+00 9.89172637e-01 -8.52880925e-02 -1.82281256e-01
-6.03691973e-02 1.56084642e-01 -6.25968397e-01 -8.12591016e-01
7.08615482e-01 -4.63861287e-01 -2.23785862e-01 -3.55866879e-01
-5.66530764e-01 6.47346973e-02 -3.11379075e-01 1.39699697e+00
5.15470386e-01 4.74935956e-02 6.56522736e-02 -1.74728680e+00
-8.39636266e-01 6.20862663e-01 -8.67360473e-01 -8.55310410e-02
2.76719511e-01 5.18801622e-02 9.93124962e-01 -1.51094511e-01
3.81183386e-01 8.18152130e-01 6.64672613e-01 3.23238432e-01
-7.99002469e-01 -9.01475489e-01 1.71775132e-01 1.43961564e-01
-1.42292356e+00 2.49502473e-02 1.03067505e+00 1.58887386e-01
2.47460812e-01 8.57455611e-01 8.90002012e-01 9.29648638e-01
1.94019839e-01 1.18593311e+00 1.41104138e+00 -2.19696552e-01
6.20445132e-01 4.57757920e-01 -4.02551796e-03 3.10762167e-01
8.55087996e-01 3.37273687e-01 -4.24408764e-01 -6.71154797e-01
1.73838943e-01 3.27607423e-01 -6.45178914e-01 -6.29943907e-01
-1.13171136e+00 6.24065459e-01 2.07613677e-01 3.36362213e-01
-2.14304358e-01 -4.27809685e-01 4.15341228e-01 3.13245893e-01
5.68191350e-01 3.95597368e-01 -3.62230718e-01 -3.11511755e-01
-4.53634530e-01 2.07743749e-01 1.04836321e+00 7.53833532e-01
4.86782700e-01 -2.65946537e-01 2.08051037e-02 3.36419255e-01
2.28273362e-01 1.28189385e-01 5.70440352e-01 -6.92113638e-01
3.02137017e-01 8.15161586e-01 9.52798724e-02 -1.59838891e+00
-3.32208201e-02 -6.93864346e-01 -1.13944888e+00 -7.04467669e-02
1.83039010e-01 -2.49866977e-01 2.67384589e-01 1.63634861e+00
5.70559919e-01 3.80931765e-01 3.81451666e-01 8.58869076e-01
4.14971381e-01 5.37278950e-01 -2.98308760e-01 -7.16199100e-01
9.93063211e-01 -8.97689104e-01 -7.48490334e-01 1.26753181e-01
4.67257977e-01 -2.40600049e-01 8.36753964e-01 4.89615560e-01
-7.40905941e-01 8.81298482e-02 -1.13948464e+00 3.19839835e-01
-5.64191699e-01 -9.62801799e-02 1.01394689e+00 1.21452653e+00
-9.59685981e-01 5.29353082e-01 -7.50620306e-01 -5.41348815e-01
6.00476980e-01 4.08603340e-01 -2.80779570e-01 1.29916266e-01
-1.14781845e+00 4.00520802e-01 2.84326464e-01 3.95497456e-02
4.08920385e-02 -4.97886211e-01 -4.60534781e-01 2.04948559e-01
6.11737132e-01 -8.19051266e-01 5.71299553e-01 -1.04006267e+00
-1.64667642e+00 6.96523607e-01 1.42698791e-02 -6.44876778e-01
8.21320474e-01 2.35204190e-01 -4.19921994e-01 -4.48636301e-02
-2.19624117e-01 -8.51701424e-02 5.77983677e-01 -1.33589733e+00
-7.29545534e-01 -7.40895391e-01 1.17151402e-01 2.56713688e-01
-9.40624237e-01 2.21653935e-02 -3.00852209e-01 -8.60952258e-01
1.05617814e-01 -7.95157194e-01 -3.22951108e-01 1.57600582e-01
-3.34133446e-01 -1.90867446e-02 8.94518614e-01 -1.09849714e-01
1.61939216e+00 -2.49369454e+00 2.26454705e-01 3.92733067e-01
3.49615633e-01 3.12367857e-01 2.02325732e-01 5.95776618e-01
3.01691294e-01 2.87712306e-01 -5.45856655e-01 -3.64214212e-01
2.19405219e-01 3.50833595e-01 -2.23210409e-01 7.43034124e-01
-2.97826380e-01 7.27456212e-01 -8.19302022e-01 -2.07946613e-01
-1.79634407e-01 1.54638097e-01 -6.49876177e-01 -3.46396305e-02
3.09997145e-02 3.86042327e-01 -8.91707242e-01 7.67325163e-01
1.14478445e+00 -1.52250245e-01 2.77914852e-01 1.30876780e-01
-3.97007108e-01 -4.32449609e-01 -1.07363701e+00 1.41169393e+00
-6.43842109e-03 1.29086301e-01 2.87248164e-01 -1.07256615e+00
9.28385794e-01 1.29537433e-01 6.80557668e-01 -4.13593829e-01
1.67474538e-01 2.47127935e-01 -1.04492091e-01 -3.37585509e-01
4.25563782e-01 1.11145116e-01 -3.53135541e-02 6.45272970e-01
-8.00029576e-01 3.86788905e-01 -1.11026473e-01 -1.50801629e-01
8.79267335e-01 -1.97417095e-01 5.32527089e-01 -2.54016608e-01
7.03750253e-01 -1.77580312e-01 1.02599251e+00 7.12828696e-01
-5.52865386e-01 3.28124493e-01 5.12164414e-01 -2.20797628e-01
-3.89366955e-01 -5.79019308e-01 -1.33268848e-01 6.41877830e-01
8.41646075e-01 -4.76043433e-01 -7.78975606e-01 -7.67648876e-01
2.99432576e-01 5.25254905e-01 -5.22975028e-01 -3.35170329e-01
-1.83667898e-01 -1.21028364e+00 5.71871340e-01 -5.38563989e-02
1.12432683e+00 -6.81864977e-01 -5.55447519e-01 -4.16788794e-02
5.66550940e-02 -5.83835065e-01 -4.01127815e-01 -2.57572442e-01
-7.31274605e-01 -9.03516293e-01 -6.25020862e-01 -5.50037205e-01
5.54107666e-01 6.65173948e-01 4.41851050e-01 2.50788093e-01
2.09087893e-01 4.53624666e-01 -6.03963435e-01 -5.48852146e-01
-3.74905556e-01 1.79997191e-01 1.77382484e-01 5.46801209e-01
5.89227378e-01 -7.84087777e-01 -6.21819139e-01 3.77728522e-01
-1.08232260e+00 -1.36747748e-01 5.69206595e-01 8.26231241e-01
4.68956739e-01 5.54512322e-01 5.41284800e-01 -7.38231838e-01
8.13917100e-01 -8.17171156e-01 -6.51082575e-01 4.46367860e-01
-9.55282748e-01 -2.35702738e-01 8.42896581e-01 -1.81956172e-01
-1.15728045e+00 -2.34851062e-01 5.13025261e-02 -1.94076374e-01
3.44652496e-02 5.19738257e-01 -7.02215195e-01 -4.47179824e-01
1.46286145e-01 6.38046682e-01 5.50959766e-01 -4.44945395e-01
-3.72979231e-03 8.62552941e-01 1.40280336e-01 -5.11132479e-01
8.57805133e-01 8.80140305e-01 1.33095354e-01 -5.87047100e-01
-1.74206659e-01 -1.51793927e-01 2.06460003e-02 3.02510291e-01
3.66479069e-01 -5.93465388e-01 -1.02372062e+00 6.92737877e-01
-8.33058894e-01 5.44166744e-01 -3.54680747e-01 4.04277474e-01
-3.03590268e-01 8.71339917e-01 -2.12188080e-01 -9.97150064e-01
-3.74086201e-01 -1.22217941e+00 5.65604508e-01 5.05254507e-01
2.49685764e-01 -9.18647289e-01 -1.38689011e-01 2.82350332e-01
6.11177564e-01 5.14357746e-01 6.00015402e-01 -9.00488257e-01
-6.22417390e-01 -3.74854922e-01 -3.83401886e-02 2.02102661e-01
3.19349080e-01 -1.37644410e-01 -7.33006179e-01 -6.34712279e-01
7.50985503e-01 2.87825882e-01 4.86446232e-01 7.66813383e-02
1.25814271e+00 -7.78902113e-01 -5.18713415e-01 9.59127367e-01
1.43974710e+00 5.54840684e-01 4.03457165e-01 8.77254069e-01
1.68799937e-01 3.78089964e-01 7.31118202e-01 1.01041555e+00
4.95449334e-01 5.53631127e-01 5.53701639e-01 -1.40709102e-01
7.89330721e-01 -3.02089304e-01 3.86866361e-01 6.75378025e-01
-6.77273571e-02 -4.35422093e-01 -5.00349164e-01 -4.60670292e-02
-1.92441499e+00 -9.41187978e-01 8.87394771e-02 2.29341197e+00
5.71086705e-01 -2.77821690e-01 1.17600285e-01 2.41950542e-01
7.29452968e-01 3.65121484e-01 -6.82375729e-01 -2.50565916e-01
-5.91850102e-01 -1.68286487e-01 5.72673440e-01 -2.50280917e-01
-9.44568574e-01 4.15826082e-01 5.38656330e+00 1.01740777e+00
-9.26756918e-01 8.83790106e-02 6.69440210e-01 1.71518978e-03
-6.13246500e-01 3.05984557e-01 -3.79768103e-01 9.17934120e-01
3.91317874e-01 -1.07857740e+00 6.15698576e-01 8.67240191e-01
2.37463310e-01 2.51766462e-02 -6.69683218e-01 1.12218845e+00
2.86928080e-02 -1.19251800e+00 -7.32971542e-03 5.01807392e-01
6.29206002e-01 -4.18295771e-01 3.51921171e-01 1.71270017e-02
-8.24792031e-03 -5.68292916e-01 3.07924211e-01 2.46906430e-01
7.78934956e-02 -1.02824020e+00 8.82189095e-01 6.07888401e-01
-8.83649707e-01 -4.49945301e-01 -4.07635719e-01 -2.56948203e-01
-1.23514697e-01 5.21303356e-01 2.36362666e-02 1.25459146e+00
7.68964291e-01 8.00156355e-01 -5.23757756e-01 1.22750664e+00
1.28086165e-01 4.57556367e-01 -4.37805027e-01 -4.55441713e-01
2.37439331e-02 -7.79975176e-01 9.42433178e-01 5.41337907e-01
5.21483362e-01 3.29727679e-01 1.26350731e-01 7.16359556e-01
1.47473803e-02 4.91276056e-01 -7.88756311e-01 -8.00116882e-02
8.33591878e-01 9.60678399e-01 -5.55516720e-01 1.24462083e-01
-4.73460317e-01 1.07230210e+00 -1.12211250e-01 2.24412367e-01
-8.09845686e-01 -5.23723900e-01 8.82256687e-01 -3.03102374e-01
4.35551405e-02 -5.66240922e-02 -5.43857515e-01 -1.41815412e+00
3.66812557e-01 -9.93009329e-01 6.58946514e-01 -1.61816180e-01
-1.42752814e+00 3.37997049e-01 -2.30271906e-01 -1.55668426e+00
3.91433895e-01 -4.66371924e-02 -6.76702082e-01 4.36089486e-01
-1.39670479e+00 -8.16159725e-01 -1.47473156e-01 5.38671374e-01
-1.78545803e-01 -3.70595455e-01 7.61561275e-01 2.40739241e-01
-9.45807397e-01 1.04938591e+00 9.05744374e-01 -3.82892102e-01
3.42373073e-01 -6.18117452e-01 -4.89349430e-03 8.82815063e-01
-8.54965206e-03 8.83103371e-01 6.62124336e-01 -4.06450719e-01
-2.03425908e+00 -7.14483142e-01 6.25196934e-01 -1.91702172e-01
5.74177623e-01 -1.38721958e-01 -8.14886630e-01 3.81942242e-01
2.55395383e-01 -2.44931892e-01 9.46404099e-01 -2.79728740e-01
5.68019897e-02 -3.88553113e-01 -1.71325719e+00 8.65104198e-01
1.14510548e+00 -2.01570801e-02 -2.52081811e-01 2.79080510e-01
8.67082357e-01 -2.37699434e-01 -8.21747422e-01 2.75581956e-01
3.37338448e-01 -1.40804994e+00 7.82892168e-01 -1.62784621e-01
-2.27149427e-02 -4.26812768e-01 -1.73718989e-01 -9.87225115e-01
-4.14816923e-02 -9.75797474e-01 -1.26073107e-01 1.45642984e+00
-6.74481019e-02 -1.55091608e+00 1.06087983e+00 8.88013422e-01
3.83623064e-01 -7.92609453e-01 -1.26126206e+00 -1.10281563e+00
-8.17487761e-03 -6.62778914e-02 1.30500269e+00 1.26426041e+00
1.38400301e-01 -5.33708513e-01 -4.00494754e-01 2.84926593e-01
6.20766282e-01 4.46122110e-01 8.64735305e-01 -1.11698782e+00
-5.65252304e-01 -6.02896810e-01 -4.58306670e-01 -9.23668742e-01
2.06771150e-01 -9.37548280e-01 -5.90496182e-01 -7.85453737e-01
1.75477058e-01 -6.57850862e-01 -4.57986385e-01 3.11576009e-01
-1.62250310e-01 -1.58876017e-01 1.24559633e-01 4.73899126e-01
-3.57208073e-01 1.04557478e+00 1.12946045e+00 -6.52089864e-02
-4.06186610e-01 3.64233881e-01 -1.28544605e+00 4.57801223e-01
1.02537549e+00 -4.13722068e-01 -4.16334599e-01 -1.25404313e-01
1.03196457e-01 -9.95866284e-02 1.55187666e-01 -8.39795470e-01
5.24153888e-01 -5.04578054e-01 -2.80310661e-01 -1.66719392e-01
-3.95958386e-02 -1.21712780e+00 5.42831361e-01 5.69981158e-01
4.86033596e-02 1.32682249e-01 -5.93278222e-02 8.82469893e-01
-3.39708537e-01 -8.18814989e-03 3.54902774e-01 -1.22597508e-01
-4.13569421e-01 5.51553965e-01 -2.24284213e-02 -1.31657973e-01
1.38290489e+00 -7.39790916e-01 -4.04513359e-01 -1.42272845e-01
-1.50582477e-01 3.66924554e-01 1.01592541e+00 2.75796682e-01
4.02909726e-01 -1.11921299e+00 -4.36008245e-01 3.90009731e-01
8.15714747e-02 -2.05989286e-01 2.61249959e-01 1.07872939e+00
-2.91589975e-01 1.77972227e-01 6.04694011e-03 -2.67520219e-01
-1.09024262e+00 8.19199026e-01 -8.01164657e-03 -1.32548600e-01
-6.54519320e-01 6.50660694e-01 9.69140455e-02 -2.52925932e-01
3.23069453e-01 3.53033878e-02 -6.97777048e-03 -9.34645385e-02
3.40019852e-01 5.80922186e-01 -2.43768573e-01 -4.93288100e-01
-3.77854109e-01 3.12510699e-01 -4.69609397e-03 1.35831311e-01
1.32054520e+00 -2.53380835e-01 -5.59392571e-01 -8.22064802e-02
1.18194044e+00 3.93869817e-01 -9.25063014e-01 -1.42469838e-01
-1.02621848e-02 -1.01487470e+00 -1.20970346e-01 -5.03550053e-01
-1.40491986e+00 3.65176737e-01 4.85139281e-01 5.23060620e-01
1.66031659e+00 -5.14300823e-01 9.09446537e-01 1.89718932e-01
1.00585961e+00 -7.75784016e-01 -4.87583041e-01 1.56567041e-02
5.69352925e-01 -1.04320300e+00 7.92107806e-02 -4.89979148e-01
-6.99051321e-01 7.50142038e-01 4.92848843e-01 1.84148267e-01
6.13865197e-01 1.33000508e-01 -2.91901052e-01 1.86144844e-01
-5.28600097e-01 2.77994841e-01 -3.60136390e-01 7.02412724e-01
-1.97964400e-01 5.60266115e-02 -1.04773390e+00 1.17547131e+00
-2.18066886e-01 7.15714991e-02 3.89192879e-01 1.18597889e+00
-6.33183196e-02 -1.48978281e+00 -3.43167365e-01 2.34830499e-01
-4.76070970e-01 3.89026552e-02 -5.56166589e-01 6.45436227e-01
1.80477157e-01 8.76514971e-01 -2.29323864e-01 -5.62903106e-01
1.45995677e-01 -4.91819471e-01 -1.10911615e-01 -1.02586849e-02
-1.00602925e+00 -4.43291306e-01 -3.18425655e-01 -4.08636540e-01
-5.22811353e-01 -8.24337363e-01 -9.29064751e-01 -6.90539241e-01
-4.86380935e-01 6.47138536e-01 3.99160475e-01 5.64923346e-01
8.60181928e-01 -1.87979817e-01 1.10999095e+00 -3.13556552e-01
-7.85396814e-01 -4.24386673e-02 -8.93348455e-01 1.09509997e-01
-4.07902822e-02 -5.28263390e-01 -7.37490237e-01 -8.09595346e-01] | [5.864260196685791, 6.533927917480469] |
9b62a645-0923-4602-86e5-9658d235782b | unified-functional-hashing-in-automatic | 2302.05433 | null | https://arxiv.org/abs/2302.05433v1 | https://arxiv.org/pdf/2302.05433v1.pdf | Unified Functional Hashing in Automatic Machine Learning | The field of Automatic Machine Learning (AutoML) has recently attained impressive results, including the discovery of state-of-the-art machine learning solutions, such as neural image classifiers. This is often done by applying an evolutionary search method, which samples multiple candidate solutions from a large space and evaluates the quality of each candidate through a long training process. As a result, the search tends to be slow. In this paper, we show that large efficiency gains can be obtained by employing a fast unified functional hash, especially through the functional equivalence caching technique, which we also present. The central idea is to detect by hashing when the search method produces equivalent candidates, which occurs very frequently, and this way avoid their costly re-evaluation. Our hash is "functional" in that it identifies equivalent candidates even if they were represented or coded differently, and it is "unified" in that the same algorithm can hash arbitrary representations; e.g. compute graphs, imperative code, or lambda functions. As evidence, we show dramatic improvements on multiple AutoML domains, including neural architecture search and algorithm discovery. Finally, we consider the effect of hash collisions, evaluation noise, and search distribution through empirical analysis. Altogether, we hope this paper may serve as a guide to hashing techniques in AutoML. | ['Esteban Real', 'Quoc V. Le', 'David R. So', 'Chen Liang', 'Jonathan Dungay', 'Connal de Souza', 'Michael Munn', 'Yingjie Miao', 'Stephen Jonany', 'Ryan Gillard'] | 2023-02-10 | null | null | null | null | ['automl'] | ['methodology'] | [-4.60124202e-02 -2.34557912e-01 -2.33471453e-01 -1.73404545e-01
-8.92153740e-01 -6.82119071e-01 3.33754063e-01 6.11004114e-01
-6.84675574e-01 5.37886977e-01 -2.32583001e-01 -3.88206899e-01
1.36576191e-01 -7.45366514e-01 -1.05548310e+00 -7.39682257e-01
-2.20644310e-01 5.99161685e-01 3.92983526e-01 5.27995788e-02
6.08184457e-01 4.89615709e-01 -1.99322391e+00 3.10112059e-01
3.76286119e-01 9.16104913e-01 8.67234468e-02 6.99762046e-01
-2.15897083e-01 6.44984066e-01 -5.74994385e-01 -7.10405767e-01
1.04208097e-01 -2.21165299e-01 -1.00627565e+00 -4.39004183e-01
5.19142151e-01 -1.89779148e-01 -2.41151065e-01 1.11191690e+00
5.01784265e-01 -1.13900840e-01 4.41516131e-01 -1.47867763e+00
-4.86753315e-01 9.58257496e-01 -4.94044036e-01 -8.43413323e-02
1.98105097e-01 1.17168337e-01 1.29594672e+00 -8.72159600e-01
4.83433843e-01 1.28384268e+00 9.20667768e-01 6.08187258e-01
-1.52488232e+00 -5.83933592e-01 -3.04173619e-01 4.28180724e-01
-1.68948472e+00 -4.91297722e-01 3.75476837e-01 -2.62233347e-01
1.21835566e+00 1.66166946e-01 5.55331886e-01 5.22851348e-01
2.57933706e-01 1.05213702e+00 6.46809220e-01 -6.62038505e-01
5.26938260e-01 3.12595209e-03 3.46752733e-01 1.15887737e+00
2.78904766e-01 9.97177735e-02 -5.52613974e-01 -5.68459928e-01
1.77532047e-01 -2.78295130e-01 -1.93945542e-01 -4.41372067e-01
-9.25314307e-01 1.08504725e+00 4.77971494e-01 1.14023961e-01
-3.74784395e-02 5.65506816e-01 6.88343406e-01 5.19881129e-01
-1.21373802e-01 4.85353351e-01 -3.20396185e-01 -6.98465332e-02
-1.03319836e+00 2.32087687e-01 9.81084406e-01 7.81189680e-01
9.94870901e-01 -2.31373295e-01 1.71311840e-01 7.86544859e-01
1.49393812e-01 2.26698399e-01 7.64790654e-01 -9.98144567e-01
-1.44045390e-02 4.54979897e-01 -3.55531842e-01 -7.72497594e-01
-2.64462501e-01 -1.79690599e-01 -6.57236576e-01 2.53533036e-01
2.75685757e-01 1.53416067e-01 -5.23141801e-01 1.99291050e+00
2.27409557e-01 -1.64305512e-03 -1.68808680e-02 4.74488825e-01
3.59891742e-01 6.93565786e-01 -8.81035700e-02 -1.56476364e-01
1.30632555e+00 -9.46951449e-01 -2.25283951e-01 -6.60780519e-02
1.12002873e+00 -7.40569234e-01 9.43948448e-01 5.73735833e-01
-1.36413383e+00 -5.14316678e-01 -1.26076674e+00 -5.71813136e-02
-3.57007295e-01 -1.26856193e-01 7.25598752e-01 7.73087263e-01
-1.41520143e+00 8.00968826e-01 -8.71623456e-01 -2.47625172e-01
2.31386751e-01 5.46830952e-01 -1.27924800e-01 4.67806077e-03
-9.69420791e-01 7.44497716e-01 6.90278947e-01 -4.20745045e-01
-4.30278331e-01 -3.59815896e-01 -6.76439464e-01 2.04158872e-01
2.85256077e-02 -4.11800206e-01 1.45595229e+00 -9.61717486e-01
-1.07560623e+00 8.85209024e-01 -1.22369364e-01 -8.80884051e-01
6.98499335e-03 3.42296250e-02 -1.74334526e-01 2.77818710e-01
-2.67401755e-01 8.76172423e-01 7.92186081e-01 -7.96253204e-01
-6.94701314e-01 -4.33030784e-01 -1.30977482e-01 -7.12393671e-02
-5.39672017e-01 1.17379285e-01 -5.55856287e-01 -4.61020052e-01
-3.98420505e-02 -1.14598179e+00 -1.20606646e-01 2.77600259e-01
8.94997492e-02 -5.26689172e-01 4.87600833e-01 -3.32852304e-02
1.54696393e+00 -2.45182586e+00 1.63470000e-01 4.66764480e-01
4.42678392e-01 2.15198725e-01 -1.08785238e-02 2.75428414e-01
1.27285510e-01 1.23108111e-01 -2.36561626e-01 -1.97107837e-01
1.91274539e-01 1.53203234e-01 -2.95304596e-01 5.33175588e-01
9.95753482e-02 8.79824638e-01 -6.66575611e-01 -6.62929833e-01
-2.78458357e-01 2.75126636e-01 -8.87506127e-01 -1.88228697e-01
-3.16376328e-01 -6.43593490e-01 -1.20147310e-01 5.93920648e-01
3.51405531e-01 -7.24975586e-01 1.77768499e-01 -2.93042570e-01
-9.53924954e-02 2.01058939e-01 -1.16811967e+00 1.61612582e+00
-3.17122877e-01 6.58373058e-01 -1.01192094e-01 -9.74694312e-01
8.69017839e-01 1.87305421e-01 1.07087381e-01 -6.53624117e-01
-4.63601202e-02 5.81769228e-01 -4.99034859e-02 7.77804386e-03
5.08130670e-01 3.48769754e-01 -1.68484941e-01 6.76372945e-01
-1.82805920e-03 4.65396941e-02 2.41711676e-01 2.96236545e-01
1.29995918e+00 -2.59149432e-01 3.46191823e-01 -1.83713526e-01
4.80475247e-01 1.10669270e-01 1.53222293e-01 1.00018167e+00
-1.14947371e-01 3.00258249e-01 3.04311872e-01 -5.75266778e-01
-1.13591993e+00 -8.65505695e-01 -2.45362684e-01 1.23225093e+00
8.83025527e-02 -7.97199667e-01 -9.74691689e-01 -5.21084130e-01
1.32339403e-01 4.12452281e-01 -3.68288904e-01 -6.08135700e-01
-8.20321977e-01 -7.14629889e-01 9.26322162e-01 5.63680768e-01
3.10895473e-01 -1.04390037e+00 -1.13711929e+00 4.05539632e-01
2.60207087e-01 -6.49618089e-01 -4.46863502e-01 6.61619782e-01
-9.17890906e-01 -8.80811572e-01 -3.69778901e-01 -1.03272355e+00
4.21349406e-01 1.30168200e-01 1.27284944e+00 5.57328582e-01
-5.85000932e-01 1.30498350e-01 -1.36208817e-01 1.52505904e-01
-5.79737484e-01 2.73876876e-01 1.99405700e-02 -3.53101730e-01
5.66190362e-01 -4.17935580e-01 -4.63853180e-01 1.98541388e-01
-8.18217099e-01 -1.69984892e-01 7.64826775e-01 1.08786547e+00
5.27030647e-01 1.44914284e-01 3.54292750e-01 -6.81150079e-01
6.02875054e-01 -4.22188193e-01 -8.02601576e-01 4.70034540e-01
-9.78387833e-01 6.59441888e-01 8.56435180e-01 -5.23331881e-01
-2.74854213e-01 2.75345236e-01 -1.33899704e-01 -5.05923867e-01
7.59245828e-03 5.47566175e-01 1.83753639e-01 -3.83776397e-01
9.04944718e-01 3.37880999e-01 2.18956411e-01 -3.79353434e-01
3.15402478e-01 5.96514106e-01 6.81108952e-01 -9.30302799e-01
4.91402417e-01 3.02400831e-02 2.32638530e-02 -5.54002881e-01
-2.13378161e-01 -2.94470370e-01 -1.65091708e-01 1.45053267e-01
4.65317845e-01 -6.35052681e-01 -8.81555974e-01 4.36034262e-01
-1.31566513e+00 -3.04166675e-01 -1.32391647e-01 2.19760716e-01
-5.14436960e-01 3.50200623e-01 -9.88306820e-01 -7.08790898e-01
-5.60557902e-01 -1.43738127e+00 9.03537750e-01 1.16648816e-01
-4.77849096e-01 -6.90663517e-01 6.67018741e-02 -3.27521354e-01
3.59231502e-01 -2.71760404e-01 1.46801639e+00 -7.31088340e-01
-7.06718743e-01 -1.81516558e-01 -2.48534799e-01 2.90945619e-02
-5.16908050e-01 8.70287567e-02 -7.37193108e-01 -5.74042857e-01
-2.73982406e-01 -7.10654318e-01 8.97109151e-01 -5.04304394e-02
1.40018320e+00 -4.84186471e-01 -4.62850183e-01 6.73752546e-01
1.58943141e+00 2.63215393e-01 4.97036368e-01 3.17847729e-01
4.02205110e-01 3.96119356e-01 2.23058388e-01 4.50314850e-01
7.70186037e-02 7.13894367e-01 2.02215090e-01 1.91271588e-01
-1.14204630e-01 -2.02924293e-02 2.94344544e-01 8.80871117e-01
3.96025866e-01 -1.99699655e-01 -1.07214975e+00 4.18751627e-01
-1.74327755e+00 -9.05884087e-01 1.35581270e-01 2.36951709e+00
1.11090577e+00 3.00954551e-01 1.71953604e-01 4.78672475e-01
8.56148064e-01 -3.11303675e-01 -5.49294889e-01 -6.76654816e-01
-6.91468343e-02 6.13120615e-01 6.66325510e-01 3.60089988e-01
-8.68292034e-01 7.96354055e-01 6.29820251e+00 1.18243361e+00
-1.14427996e+00 -1.62294060e-02 6.40543163e-01 -2.79679410e-02
-1.84329942e-01 -1.11373916e-01 -1.08450246e+00 4.85213488e-01
1.20767283e+00 -2.23718926e-01 7.69067705e-01 1.13142931e+00
-7.81450689e-01 3.32177393e-02 -1.35222101e+00 1.26597631e+00
1.42421216e-01 -1.52429998e+00 4.94991951e-02 4.76322733e-02
2.69006997e-01 2.01579720e-01 4.82762046e-03 3.34116906e-01
3.72519374e-01 -8.43542039e-01 9.63576674e-01 -6.52546585e-02
7.77123451e-01 -1.03959620e+00 3.81318718e-01 2.36353636e-01
-1.42539465e+00 -3.04699510e-01 -4.28046674e-01 3.30329388e-01
-2.74796367e-01 3.70970964e-01 -6.75568938e-01 -1.65848672e-01
6.82383597e-01 1.10778131e-01 -7.13060141e-01 1.39470816e+00
1.24837637e-01 4.42893386e-01 -4.97584015e-01 -5.65678298e-01
2.90026695e-01 3.24282914e-01 1.30036891e-01 1.43943572e+00
4.86340612e-01 -1.80352554e-01 1.64975628e-01 8.04647684e-01
-1.99141890e-01 1.65447831e-01 -7.00705707e-01 -1.12469770e-01
9.43975747e-01 9.95414555e-01 -9.10434365e-01 -3.01089495e-01
-3.85128826e-01 9.68709230e-01 5.63604295e-01 -5.84139861e-02
-8.56537819e-01 -6.80899203e-01 2.86669970e-01 -1.79465830e-01
4.31954801e-01 -1.50751129e-01 -1.48481771e-01 -8.69779944e-01
-6.29455224e-02 -1.03677058e+00 5.52925110e-01 -4.46471035e-01
-1.08399594e+00 7.09058642e-01 -1.24235302e-01 -7.92561352e-01
-4.42337871e-01 -6.51930094e-01 -2.49311015e-01 4.31324422e-01
-9.93847489e-01 -4.07941848e-01 7.59842172e-02 6.01095796e-01
4.04655933e-01 -3.37429672e-01 8.69864643e-01 4.46036756e-01
-4.52231050e-01 1.18384945e+00 1.41785234e-01 -1.36248946e-01
5.76485574e-01 -1.18483365e+00 6.09066248e-01 5.40811479e-01
4.85558569e-01 8.53609443e-01 5.33043444e-01 -3.17425698e-01
-1.82762372e+00 -7.37061024e-01 9.16448355e-01 -3.72540858e-03
7.48161376e-01 -2.36849964e-01 -1.12277281e+00 2.74757147e-01
-2.92369872e-02 9.64465588e-02 5.21739781e-01 -2.09295247e-02
-7.88744330e-01 -1.30681172e-01 -1.05488813e+00 6.60392582e-01
8.70357096e-01 -7.15386093e-01 -2.33750850e-01 2.16789365e-01
7.65261114e-01 -2.04764485e-01 -5.42449474e-01 2.93871731e-01
7.38173962e-01 -9.48011518e-01 1.03967297e+00 -3.05494934e-01
2.32155502e-01 -1.84862629e-01 -3.87868702e-01 -8.57949436e-01
-3.82484078e-01 -7.43825495e-01 -6.19785190e-01 8.74306500e-01
5.20616412e-01 -5.38711786e-01 9.40152645e-01 4.55765069e-01
-3.41725014e-02 -9.33363974e-01 -9.29885805e-01 -9.20081675e-01
-5.88561445e-02 -2.47333959e-01 7.75068998e-01 7.30509281e-01
3.81108150e-02 4.61423546e-01 -8.75391960e-02 -1.61264315e-01
7.31109023e-01 2.53200024e-01 4.76838470e-01 -1.27662528e+00
-7.93627143e-01 -9.50862706e-01 -5.87414980e-01 -1.00445712e+00
3.06614965e-01 -1.13727880e+00 1.26108155e-01 -6.31481767e-01
1.56781957e-01 -5.67636192e-01 -4.16433781e-01 7.36390233e-01
1.29902661e-01 3.79466504e-01 1.13137469e-01 4.31269705e-01
-7.21022010e-01 1.56661585e-01 4.95314807e-01 -1.67929828e-01
3.19763739e-03 -2.93232590e-01 -3.77956480e-01 5.67257524e-01
8.42237771e-01 -6.25239372e-01 -2.77249515e-01 -3.41727078e-01
4.44431484e-01 5.33286436e-03 4.40760106e-01 -1.23740172e+00
6.86484337e-01 4.06264365e-01 1.62976786e-01 -4.34631199e-01
1.90522581e-01 -7.67794251e-01 7.87577704e-02 8.10335517e-01
-5.39803207e-01 6.94970548e-01 2.75601894e-01 3.37282717e-01
-7.81145599e-03 -6.37189150e-01 9.52186048e-01 -9.90696996e-02
-7.46762693e-01 7.91461095e-02 -1.93546027e-01 8.74637738e-02
8.03841770e-01 -2.65868306e-01 -4.65608329e-01 -3.92734781e-02
-2.91649222e-01 4.79250848e-02 7.01222897e-01 -1.04757294e-01
7.05563605e-01 -1.27228665e+00 -4.80056435e-01 3.78299564e-01
1.64484531e-01 -2.77629942e-01 -3.66736740e-01 6.56295776e-01
-5.96598566e-01 2.02595025e-01 -8.28753635e-02 -5.87751865e-01
-1.41145813e+00 8.23715746e-01 4.05292436e-02 -2.03095414e-02
-4.81427401e-01 1.04197156e+00 -1.69883937e-01 2.99920863e-03
7.42773950e-01 4.64087501e-02 4.60828066e-01 -2.52849087e-02
6.56580508e-01 2.16047540e-01 3.00194412e-01 -1.77666306e-01
-4.23602909e-01 5.19066095e-01 -3.65682453e-01 -7.90127665e-02
1.01304042e+00 2.85508454e-01 -3.96298766e-01 2.54905820e-01
1.61271799e+00 -2.78539449e-01 -7.36557066e-01 -2.34435022e-01
3.44281405e-01 -8.40659589e-02 5.41345440e-02 -4.75948662e-01
-1.06368160e+00 8.47227335e-01 6.90392911e-01 3.46243352e-01
1.11414433e+00 3.91885489e-02 1.05135417e+00 1.09129083e+00
6.79968774e-01 -9.30439889e-01 2.82474235e-02 6.06132030e-01
3.56036723e-01 -7.94847310e-01 -1.68591499e-01 6.76174164e-02
-4.31687571e-02 1.33668935e+00 5.50076962e-01 -9.50808972e-02
3.42444122e-01 6.21847034e-01 -5.03780961e-01 -1.39946183e-02
-1.06426549e+00 5.27227670e-02 2.10298300e-02 1.04777493e-01
3.85122895e-01 -1.58257455e-01 -2.82027155e-01 2.24408939e-01
-3.26454818e-01 -6.47454932e-02 1.10946849e-01 1.09608996e+00
-7.14546442e-01 -1.26923692e+00 -3.29408497e-01 3.07856143e-01
-3.83022755e-01 -4.31809545e-01 -1.80466771e-01 7.06491172e-01
-7.53060803e-02 4.52073693e-01 1.30223766e-01 -5.15044808e-01
-1.78269997e-01 2.86427051e-01 7.24461973e-01 -3.21280748e-01
-7.18134105e-01 -6.85909092e-02 -1.11276694e-01 -6.37451828e-01
7.52014220e-02 -4.07648861e-01 -1.53870761e+00 -4.19069439e-01
-3.31860244e-01 3.93630624e-01 7.94546425e-01 5.46624362e-01
3.88480455e-01 2.11659130e-02 4.56038117e-01 -6.07847691e-01
-9.91924882e-01 -1.30398884e-01 -2.51244605e-01 1.71267644e-01
2.97478437e-01 -4.35959399e-01 -4.82146502e-01 -1.26629472e-02] | [8.552741050720215, 3.3786227703094482] |
ba9dbb95-25f6-4051-9ef6-592a0f93eab3 | intrusion-detection-with-segmented-federated | null | null | https://ieeexplore.ieee.org/document/9207094 | https://ieeexplore.ieee.org/document/9207094 | Intrusion Detection with Segmented Federated Learning for Large-Scale Multiple LANs | Traditional approaches to cybersecurity issues usually protect users from attacks after the occurrence of specific types of attacks. Besides, patterns of recent cyberattacks tend to be changeable, which add up to unpredictability of them. On the other hand, machine learning, as a new method used to detect intrusion, is attracting more and more attention. Moreover, through the sharing of local training data, the centralized learning approach has proven to improve a model's performance. In this research, a segmented federated learning is proposed, different from a collaborative learning based on single global model in a traditional federated learning model, it keeps multiple global models which allow each segment of participants to conduct collaborative learning separately and rearranges the segmentation of participants dynamically as well. Furthermore, these multiple global models interact with each other for updating parameters, thus being adaptable to various participants' LANs. A dataset covering two months' traffic data from 20 participants' LANs in the LAN-Security Monitoring Project is used. We adopt three types of knowledge-based methods for labeling network events and train a CNN model based on the dataset. At last, we achieve validation accuracies of 0.923, 0.813 and 0.877 individually with these labeling methods. | ['Hiroshi Esaki', 'Hideya Ochiai', 'Yuwei Sun'] | 2020-09-28 | null | null | null | international-joint-conference-on-neural | ['network-intrusion-detection'] | ['miscellaneous'] | [-4.36586171e-01 -4.28627670e-01 -2.88389444e-01 -3.91111821e-01
-2.20353380e-01 -6.25232756e-01 2.47183055e-01 3.93434346e-01
-5.20514011e-01 6.30965889e-01 -3.75408620e-01 -5.31104624e-01
-3.40167493e-01 -1.19563997e+00 -4.45656657e-01 -5.78589678e-01
-2.35596791e-01 2.80247837e-01 6.43748522e-01 -1.81179881e-01
3.13750446e-01 7.22167373e-01 -1.20439112e+00 4.49477762e-01
5.56050003e-01 9.60123360e-01 -5.17996311e-01 3.71561021e-01
-3.43709022e-01 6.54263735e-01 -9.72614884e-01 -3.77046913e-01
1.76582992e-01 -5.31429723e-02 -6.93950772e-01 -3.82607102e-01
-1.27084211e-01 1.23178046e-02 -4.03956145e-01 1.00161982e+00
3.94401968e-01 1.90106660e-01 -7.95138255e-03 -1.41766083e+00
-2.17746317e-01 8.29331279e-01 -3.89557540e-01 3.17903131e-01
7.69859329e-02 1.50263861e-01 5.64066350e-01 -9.93406028e-03
3.58071685e-01 1.01275527e+00 5.52858591e-01 3.24437529e-01
-7.91124105e-01 -1.18656254e+00 7.03364611e-01 5.80614567e-01
-1.22601998e+00 -3.60907085e-04 8.90706837e-01 -9.88509580e-02
6.63287401e-01 2.62315840e-01 3.92113119e-01 1.16502845e+00
4.12997723e-01 3.06573778e-01 8.86196673e-01 -4.03138041e-01
3.59690040e-01 5.40684044e-01 4.65772599e-01 3.83275121e-01
4.85093921e-01 2.06387490e-01 -1.36706546e-01 -4.11608547e-01
3.81625861e-01 4.82099205e-01 2.73487670e-03 -1.27126575e-01
-7.80948579e-01 5.60913324e-01 3.35730761e-01 7.41437376e-01
-7.69910067e-02 -3.50410968e-01 5.85053086e-01 5.81172645e-01
2.63114452e-01 2.17252806e-01 -9.90081191e-01 -1.36493042e-01
-4.06991094e-01 -1.02863736e-01 1.27361333e+00 5.33445716e-01
7.92001128e-01 1.43968761e-01 2.97564805e-01 3.29878867e-01
1.90399140e-01 1.17237017e-01 5.99545598e-01 -2.81099826e-01
3.74014705e-01 1.03573537e+00 -1.25474632e-01 -1.40770245e+00
-5.44099212e-01 -6.89372838e-01 -8.78628254e-01 8.18601251e-02
3.65672201e-01 -4.16968584e-01 -4.62886155e-01 1.66266060e+00
6.00561261e-01 7.18169510e-01 -1.78339645e-01 3.13845992e-01
1.22006536e-01 6.08498037e-01 1.63393021e-01 -3.30766082e-01
9.47312236e-01 -7.36495376e-01 -6.76657796e-01 3.66326213e-01
6.55305028e-01 -6.28304601e-01 4.36826646e-01 9.42082465e-01
-3.31554651e-01 -5.97458005e-01 -9.08024728e-01 8.76873970e-01
-9.73848224e-01 -5.87752938e-01 5.90377808e-01 1.25034273e+00
-5.22331595e-01 5.10486424e-01 -5.61867714e-01 -1.88182697e-01
1.52299494e-01 4.80559826e-01 -5.37000922e-03 1.97548270e-01
-1.40670145e+00 7.51548052e-01 6.84066713e-01 -5.85172251e-02
-7.85055518e-01 -5.57348490e-01 -2.70712763e-01 2.47869119e-01
5.63198924e-01 -3.82819027e-02 1.17110586e+00 -7.73902953e-01
-1.40543616e+00 8.62530470e-02 4.48820919e-01 -3.98879439e-01
3.54578406e-01 1.54224113e-02 -1.29972231e+00 -8.11691135e-02
-3.10197979e-01 -2.96388060e-01 6.91111147e-01 -1.16142881e+00
-8.11639190e-01 -2.92634398e-01 1.75063834e-01 -4.26256359e-01
-8.50989223e-01 3.72636765e-01 -1.63404465e-01 -3.82937819e-01
-2.44303375e-01 -5.32569230e-01 -3.76401037e-01 -5.69436073e-01
-1.06081009e-01 -3.01020175e-01 1.38270855e+00 -3.55786532e-01
1.76949084e+00 -1.94862676e+00 -4.40403819e-01 9.31901097e-01
1.35183573e-01 6.73710167e-01 -1.24580808e-01 5.13878345e-01
-2.33948335e-01 4.39318299e-01 2.01392516e-01 3.93827371e-02
-4.77181599e-02 4.88878414e-02 -1.36081189e-01 2.08861485e-01
-3.96942556e-01 3.74570936e-01 -5.36247492e-01 -5.00690520e-01
3.76376390e-01 1.73044384e-01 -4.17633235e-01 1.83022097e-01
-1.11830018e-01 5.84960639e-01 -7.58516133e-01 6.56037331e-01
6.85833991e-01 -2.16411650e-01 5.01424015e-01 1.14786550e-01
-1.90755770e-01 -1.55997157e-01 -1.51545477e+00 1.18390393e+00
-5.06441653e-01 -1.51483500e-02 1.53106987e-01 -1.31350851e+00
1.09765887e+00 6.83149815e-01 7.53403902e-01 -6.17362618e-01
4.35968637e-01 5.13676517e-02 -1.43633853e-03 -3.78767580e-01
-1.34027034e-01 2.31225610e-01 -3.17287028e-01 9.24914420e-01
-1.15084127e-02 6.96617246e-01 -1.12351239e-01 1.85453862e-01
1.15725803e+00 -5.47456145e-01 1.88632935e-01 1.94995493e-01
9.84533548e-01 -3.20965469e-01 8.04842353e-01 9.01955128e-01
-4.32455003e-01 -2.28602722e-01 3.77713859e-01 -1.01983142e+00
-5.55941820e-01 -8.13347816e-01 1.42654125e-02 9.36320305e-01
2.52077311e-01 -5.07195294e-01 -5.43809772e-01 -1.18962955e+00
-2.18922779e-01 5.58309078e-01 -2.13111967e-01 -4.97762918e-01
-7.13276207e-01 -5.98252475e-01 6.73495591e-01 9.60525796e-02
9.62180018e-01 -1.15606439e+00 -2.97984958e-01 4.25602138e-01
-3.70869935e-02 -9.20584500e-01 -1.50612950e-01 1.00807279e-01
-5.88670611e-01 -1.52171564e+00 2.83672541e-01 -5.63804626e-01
3.83675992e-01 3.23861837e-01 6.80743515e-01 4.85184759e-01
-2.25873858e-01 4.31506395e-01 -5.02424777e-01 -4.79254335e-01
-3.47371846e-01 3.92698973e-01 3.30245316e-01 4.94235277e-01
5.03056943e-01 -7.19464839e-01 -2.84454823e-01 6.05060637e-01
-9.78850245e-01 -6.40488625e-01 3.52371573e-01 5.02519667e-01
-8.60156119e-02 9.06630874e-01 8.31719756e-01 -1.12310731e+00
6.35711730e-01 -7.69227743e-01 -6.74224496e-01 4.75953788e-01
-7.89625645e-01 -4.37716603e-01 1.07108903e+00 -7.16955304e-01
-1.30428350e+00 -4.03900653e-01 2.01462969e-01 -3.52694929e-01
-7.02452838e-01 4.36727285e-01 -4.99715775e-01 -3.51015538e-01
6.94790304e-01 -8.57254490e-02 -9.42273214e-02 -4.12647843e-01
-8.14993586e-03 8.15990508e-01 6.95872977e-02 -6.50133371e-01
1.14478004e+00 7.68783540e-02 -3.09287667e-01 -4.63147104e-01
-5.54187715e-01 -4.25925493e-01 -3.77729267e-01 -3.85545731e-01
4.47341740e-01 -3.92284393e-01 -1.26582015e+00 8.25394869e-01
-1.10765028e+00 7.03299344e-02 1.30463943e-01 5.66427350e-01
1.62716493e-01 4.76678759e-01 -6.67200625e-01 -7.64541328e-01
-2.71287292e-01 -7.78705060e-01 -5.30603714e-02 5.41491807e-01
-3.14680785e-02 -1.20322919e+00 2.08961666e-01 3.28550696e-01
7.89540291e-01 2.23737746e-01 1.01109600e+00 -1.35988450e+00
-6.54203475e-01 -8.38133395e-01 8.39677826e-02 4.72575098e-01
3.38353574e-01 9.78596658e-02 -8.09201300e-01 -5.55083632e-01
2.17296600e-01 -6.64201612e-03 1.56973794e-01 -1.92339644e-01
1.38077915e+00 -5.01157224e-01 -4.42354143e-01 2.53768533e-01
1.24541748e+00 8.37377906e-01 4.25254345e-01 6.39037848e-01
4.99375165e-01 5.75071692e-01 4.42143708e-01 5.38570225e-01
1.81977868e-01 3.82852703e-01 5.73745847e-01 9.79282558e-02
6.40846312e-01 -1.22709200e-01 2.56645024e-01 5.46413422e-01
8.60096440e-02 -2.00851887e-01 -9.35452580e-01 -3.44454311e-02
-1.78849828e+00 -1.16345632e+00 1.77398473e-01 2.26382327e+00
5.02784908e-01 5.09432614e-01 4.70654666e-02 3.38540882e-01
1.02029955e+00 1.64428532e-01 -3.70738745e-01 -5.99897981e-01
2.64431983e-01 4.11386997e-01 4.21763062e-01 2.40469024e-01
-1.14494300e+00 8.39785755e-01 5.27849150e+00 7.89551139e-01
-1.29365814e+00 2.21622854e-01 3.93684089e-01 1.48807704e-01
1.38828829e-01 3.06153744e-01 -8.85262728e-01 7.43332446e-01
1.40605175e+00 -1.15571380e-01 4.70933586e-01 7.99415231e-01
1.75237596e-01 1.21835649e-01 -6.25297844e-01 8.03171992e-01
-2.08301723e-01 -1.25634158e+00 1.47194371e-01 6.24349266e-02
5.73508024e-01 -2.94817239e-01 -1.25798807e-01 5.21791816e-01
4.47182983e-01 -6.55613899e-01 1.56964287e-01 5.95157206e-01
4.21026535e-02 -1.32001293e+00 8.93067002e-01 7.79008925e-01
-1.20354271e+00 -6.03686571e-01 2.01988220e-02 -2.67014559e-02
-4.21389863e-02 5.15088618e-01 -4.90272254e-01 9.48055029e-01
9.09542263e-01 3.35797638e-01 -6.61249578e-01 1.11071467e+00
1.24620250e-03 8.20468962e-01 -2.57697612e-01 -3.44436169e-02
-1.73692615e-03 -4.06009406e-02 3.30712348e-01 8.65373373e-01
-1.83504149e-02 5.61434217e-02 5.76258838e-01 3.96579802e-01
1.09157056e-01 1.37063459e-01 -4.74566042e-01 1.61766320e-01
8.03364694e-01 1.31973851e+00 -5.84561169e-01 -2.16222852e-01
-3.88369888e-01 2.35438347e-01 4.92997766e-02 3.58550310e-01
-9.18877244e-01 -8.59526634e-01 3.85826290e-01 -1.75917298e-02
-1.16425371e-02 -3.43096852e-01 -1.10146038e-01 -1.22151721e+00
-1.96525604e-01 -1.17691982e+00 8.11701119e-01 1.68391969e-02
-1.38595510e+00 5.90487063e-01 -5.20800576e-02 -1.26486135e+00
1.87659457e-01 -4.14477646e-01 -1.13107443e+00 5.80707967e-01
-1.30474758e+00 -1.10906184e+00 -2.90208489e-01 1.18306160e+00
1.39722586e-01 -5.07211506e-01 9.67096627e-01 5.63888431e-01
-1.13000298e+00 7.99372137e-01 -2.42485285e-01 3.95998865e-01
8.26986074e-01 -6.02453768e-01 1.16717191e-02 9.47983921e-01
4.39712591e-02 8.80495429e-01 1.58263564e-01 -7.86868751e-01
-1.04500878e+00 -9.98890042e-01 8.10893416e-01 -2.10122034e-01
8.26238215e-01 -1.51269004e-01 -1.10792494e+00 7.54241407e-01
1.84649289e-01 5.89740090e-03 1.05862761e+00 4.63243127e-01
-6.20876729e-01 -6.16847157e-01 -1.36120617e+00 4.01038706e-01
6.24510407e-01 -3.10810000e-01 -2.59844244e-01 1.99535966e-01
6.78531826e-01 -1.66444838e-01 -8.37815285e-01 2.67703950e-01
3.14172477e-01 -9.77397501e-01 6.85578585e-01 -8.16346169e-01
-6.81578159e-01 -3.05792183e-01 1.25002459e-01 -8.89880836e-01
-3.86776328e-01 -6.34916186e-01 -2.39193544e-01 1.30534565e+00
2.21843272e-01 -1.31108606e+00 8.47011149e-01 6.41243994e-01
7.98738897e-02 -4.30147797e-01 -1.02463651e+00 -7.75187790e-01
-1.28103897e-01 -5.04228294e-01 1.09082782e+00 1.30695128e+00
1.45801559e-01 -7.44712502e-02 -3.30700397e-01 4.67426896e-01
5.97645879e-01 -3.74088176e-02 8.83366704e-01 -1.54874456e+00
-3.14914733e-01 -3.21435541e-01 -3.25544924e-01 -3.76673430e-01
2.35585883e-01 -5.86025715e-01 -7.68151164e-01 -7.29570508e-01
-3.27089399e-01 -8.91484201e-01 -9.79169548e-01 7.68802583e-01
4.86810923e-01 -3.97403419e-01 4.71175350e-02 1.51524052e-01
-7.29683638e-01 3.36363958e-03 7.69305289e-01 -2.47537047e-01
-5.93416870e-01 5.63188314e-01 -5.53348422e-01 7.00725973e-01
1.37012601e+00 -5.15884876e-01 -4.91795212e-01 -8.93235486e-03
-6.95020109e-02 1.24866948e-01 2.36072540e-01 -1.20686412e+00
7.60514379e-01 -5.35662711e-01 4.41708475e-01 -5.69490194e-01
-1.05338119e-01 -1.24196720e+00 3.52692306e-01 6.70666277e-01
-9.00212973e-02 8.61501098e-02 7.49533102e-02 6.88038528e-01
-2.51570731e-01 -8.30054060e-02 7.50246763e-01 -2.04308361e-01
-5.60358346e-01 6.26194119e-01 -4.18695748e-01 -3.47163498e-01
1.43610299e+00 -2.90048808e-01 -2.87788421e-01 -1.42042130e-01
-7.78246462e-01 3.33050102e-01 1.87242782e-04 6.30661964e-01
5.30645967e-01 -9.65305686e-01 -4.18368839e-02 4.90480125e-01
-2.48144418e-01 -1.59373507e-01 6.25128806e-01 6.15914345e-01
-1.68238282e-01 3.78553033e-01 -2.76664019e-01 -2.21892789e-01
-1.20584190e+00 8.06486428e-01 4.02259499e-01 -4.73206401e-01
-4.64247555e-01 1.44012317e-01 -6.44186139e-01 -6.65624738e-01
5.46985745e-01 1.76711738e-01 -4.81537044e-01 1.76779702e-01
4.84836608e-01 4.97569978e-01 3.22862357e-01 -1.31682247e-01
-4.26859617e-01 2.40634099e-01 -4.96699482e-01 2.55914420e-01
1.10014057e+00 8.34763721e-02 -2.97517866e-01 2.43600234e-01
1.11321712e+00 1.77614316e-02 -5.00837445e-01 -4.66370642e-01
2.71549523e-01 -5.49985588e-01 -2.06948906e-01 -8.15710008e-01
-1.27802682e+00 4.02774572e-01 6.05269432e-01 5.25245786e-01
1.15408146e+00 -5.39898574e-01 8.49682450e-01 5.80939293e-01
8.91025722e-01 -1.06215024e+00 1.66529506e-01 5.43868423e-01
-3.54233831e-02 -8.68427217e-01 -2.92457342e-01 -2.15722218e-01
-1.83980063e-01 1.29634333e+00 9.81153548e-01 -3.67225297e-02
1.17431641e+00 1.37373090e-01 4.17264663e-02 9.17316005e-02
-8.36646259e-01 5.26542604e-01 -3.32217008e-01 5.96094668e-01
-1.92478985e-01 7.16151521e-02 -1.92541450e-01 7.49628961e-01
1.16408043e-01 -6.17907010e-02 4.35406476e-01 1.17776513e+00
-5.94917417e-01 -1.64845002e+00 -4.24020857e-01 2.10925415e-01
-4.13171351e-01 4.87284571e-01 -6.01025559e-02 7.21183777e-01
4.61738706e-01 1.28936577e+00 -1.19797818e-01 -8.31346691e-01
4.18116122e-01 1.51870012e-01 -1.91444367e-01 -6.12193465e-01
-1.06838977e+00 -2.61245787e-01 -2.28803113e-01 -5.95822692e-01
-1.44586042e-01 -3.08665901e-01 -1.07392347e+00 -7.10088551e-01
-4.79839981e-01 6.88824356e-01 5.79284012e-01 8.11562657e-01
2.44390205e-01 4.72401142e-01 1.43961251e+00 -1.44062713e-01
-7.90535688e-01 -5.80307961e-01 -6.37214541e-01 8.74455199e-02
-8.38733986e-02 -5.16487062e-01 -6.75929189e-01 -4.55562830e-01] | [5.288843154907227, 7.178753852844238] |
db885865-37a1-44e2-9fca-2aaab2330a25 | deformable-kernel-networks-for-joint-image | 1910.08373 | null | https://arxiv.org/abs/1910.08373v3 | https://arxiv.org/pdf/1910.08373v3.pdf | Deformable Kernel Networks for Joint Image Filtering | Joint image filters are used to transfer structural details from a guidance picture used as a prior to a target image, in tasks such as enhancing spatial resolution and suppressing noise. Previous methods based on convolutional neural networks (CNNs) combine nonlinear activations of spatially-invariant kernels to estimate structural details and regress the filtering result. In this paper, we instead learn explicitly sparse and spatially-variant kernels. We propose a CNN architecture and its efficient implementation, called the deformable kernel network (DKN), that outputs sets of neighbors and the corresponding weights adaptively for each pixel. The filtering result is then computed as a weighted average. We also propose a fast version of DKN that runs about seventeen times faster for an image of size 640 x 480. We demonstrate the effectiveness and flexibility of our models on the tasks of depth map upsampling, saliency map upsampling, cross-modality image restoration, texture removal, and semantic segmentation. In particular, we show that the weighted averaging process with sparsely sampled 3 x 3 kernels outperforms the state of the art by a significant margin in all cases. | ['Bumsub Ham', 'Jean Ponce', 'Beomjun Kim'] | 2019-10-17 | null | null | null | null | ['depth-map-super-resolution'] | ['computer-vision'] | [ 4.98344809e-01 -1.05012648e-01 -1.67526212e-02 -5.49505293e-01
-7.70792544e-01 -5.38364127e-02 4.27462310e-01 -2.15418458e-01
-6.70135856e-01 6.55498207e-01 6.31394684e-01 1.68172151e-01
-6.95939362e-02 -7.43592262e-01 -1.01419008e+00 -6.23016775e-01
1.81012526e-01 -2.63057679e-01 7.17464328e-01 4.49653938e-02
5.09775519e-01 8.42206419e-01 -1.79474461e+00 5.15386641e-01
9.36307728e-01 1.05925679e+00 3.53406489e-01 6.95873857e-01
-5.65299541e-02 8.49889100e-01 -2.27625161e-01 8.32454413e-02
2.69718587e-01 -1.22818135e-01 -8.02208602e-01 2.09756911e-01
1.06659341e+00 -4.72268552e-01 -6.14410937e-01 1.18723464e+00
3.51300091e-01 4.60543156e-01 5.21513283e-01 -6.85797155e-01
-9.55968022e-01 4.47791457e-01 -7.59216309e-01 3.45204234e-01
-3.65906596e-01 6.25088960e-02 5.68742216e-01 -1.16078043e+00
6.20605826e-01 1.31950045e+00 7.75254250e-01 6.90583050e-01
-1.48418915e+00 -4.06993181e-01 3.00574392e-01 1.15138285e-01
-1.12417924e+00 -5.67333996e-01 6.04742348e-01 -2.03467086e-01
9.80407417e-01 2.84715265e-01 5.31302452e-01 5.38919449e-01
1.88305110e-01 9.56026971e-01 1.15638387e+00 -3.86942178e-01
3.02888244e-01 -2.29320213e-01 4.90765385e-02 7.61182189e-01
1.37599811e-01 8.75969306e-02 -7.41361082e-01 3.47368978e-03
1.41737831e+00 1.62245572e-01 -4.44576085e-01 -3.78099293e-01
-1.23341680e+00 6.97080433e-01 7.87273228e-01 2.11660355e-01
-4.79435295e-01 3.85980427e-01 7.71514401e-02 -5.43191619e-02
8.22213709e-01 2.96805859e-01 -4.83405381e-01 4.81297463e-01
-1.24496520e+00 2.50251442e-01 4.06479508e-01 4.87997353e-01
1.02016926e+00 1.92058638e-01 -4.38485533e-01 9.62771475e-01
-1.89156979e-01 1.91261172e-01 4.19312209e-01 -1.36418140e+00
-1.63179234e-01 5.22655725e-01 1.45944461e-01 -8.80750060e-01
-2.29550898e-01 -2.44175196e-01 -9.18177903e-01 7.95698404e-01
4.20637250e-01 7.48942643e-02 -1.39707255e+00 1.65309167e+00
3.89555693e-01 4.76106346e-01 -1.46533147e-01 1.09367990e+00
8.04455638e-01 5.09986818e-01 9.07438993e-02 2.34871507e-01
1.15957987e+00 -1.22517407e+00 -5.72178721e-01 -2.88611650e-01
1.15349656e-02 -8.34612072e-01 1.06263316e+00 1.11922048e-01
-1.33412647e+00 -6.13851428e-01 -8.39104772e-01 -6.57622039e-01
-3.68540853e-01 3.93889040e-01 7.25547552e-01 1.13505110e-01
-1.45599592e+00 9.56206560e-01 -9.54421937e-01 -2.37110525e-01
8.25986087e-01 3.34716380e-01 -3.39676231e-01 -2.09510103e-01
-8.04167151e-01 7.87726820e-01 8.79927799e-02 -7.14600310e-02
-9.40134704e-01 -1.05386198e+00 -9.12233949e-01 1.80079609e-01
1.38159335e-01 -7.18676209e-01 9.09463882e-01 -1.16545296e+00
-1.51224804e+00 9.83165503e-01 -3.94351095e-01 -6.82519138e-01
1.74908206e-01 -4.45659101e-01 -8.66602808e-02 2.98120409e-01
2.67229408e-01 1.02667582e+00 1.30553269e+00 -9.53794301e-01
-1.00473869e+00 -3.50104630e-01 2.75851786e-02 2.16530785e-01
-9.49290022e-02 -4.75400873e-02 -4.97157186e-01 -9.47629571e-01
2.16012985e-01 -4.74635631e-01 -4.80783939e-01 4.71276462e-01
-3.65131944e-01 8.23046044e-02 7.76226759e-01 -8.09475422e-01
9.51196790e-01 -2.28724313e+00 3.33900660e-01 -4.93110269e-02
4.04078901e-01 2.28652328e-01 -2.95800239e-01 -2.51372784e-01
-1.13155432e-01 -3.93789172e-01 -5.07167876e-01 -2.15855643e-01
-4.06104207e-01 -4.79729474e-02 -1.94179252e-01 5.07125318e-01
3.70943010e-01 9.94715273e-01 -6.65623248e-01 -1.09582081e-01
4.64392036e-01 8.96262407e-01 -5.98889530e-01 -1.39657175e-02
3.92815992e-02 2.86949009e-01 -1.60555154e-01 5.16396284e-01
8.62710536e-01 -2.53623694e-01 -3.67629528e-01 -7.40707815e-01
-4.58501071e-01 -3.59425694e-02 -1.11902785e+00 1.91468155e+00
-4.05500770e-01 9.05596137e-01 3.58887911e-01 -8.92563581e-01
6.51592970e-01 -2.47395247e-01 2.69818574e-01 -5.89998305e-01
-7.57969990e-02 1.66725591e-01 -4.11591440e-01 -8.98641497e-02
6.49445415e-01 1.17789708e-01 4.06073600e-01 3.19648385e-01
2.57536650e-01 -1.16427287e-01 -1.07359692e-01 4.84779328e-02
1.04755497e+00 6.17641583e-02 1.90678567e-01 -5.34917772e-01
4.72580314e-01 -1.85913056e-01 3.94087732e-01 7.64975011e-01
-7.83135295e-02 8.92934263e-01 2.10642681e-01 -8.35999429e-01
-1.28843284e+00 -1.16425908e+00 -8.50348398e-02 1.17585659e+00
1.76126152e-01 1.06592871e-01 -1.07690251e+00 -4.36271280e-01
1.28190428e-01 4.64779258e-01 -9.30561602e-01 -3.86758670e-02
-6.57252014e-01 -4.49975997e-01 2.26976290e-01 7.17913330e-01
8.55462432e-01 -1.16826940e+00 -7.01488614e-01 2.53408283e-01
-2.36486667e-03 -9.58909214e-01 -8.07451427e-01 2.77055562e-01
-9.17257845e-01 -1.01348066e+00 -1.00531709e+00 -9.70318079e-01
8.22044075e-01 5.53066015e-01 9.81921852e-01 -1.86572790e-01
-5.89798808e-01 2.59052247e-01 1.79041207e-01 -7.77436122e-02
2.39350140e-01 -9.21410173e-02 -2.54290313e-01 2.80001938e-01
1.01746656e-01 -6.28527045e-01 -9.45905328e-01 8.70251507e-02
-1.23595572e+00 3.27920347e-01 9.13066804e-01 9.32994127e-01
9.00275767e-01 -6.62915036e-02 1.52265817e-01 -9.27213609e-01
4.40576464e-01 2.37664264e-02 -5.43487787e-01 1.51187092e-01
-1.70719847e-01 3.52211028e-01 2.64903694e-01 -5.76464236e-01
-1.22703898e+00 3.68709415e-01 9.58864242e-02 -4.98557985e-01
-2.57333279e-01 -4.29951102e-02 1.17686331e-01 -5.05369604e-01
7.37574518e-01 2.60495991e-01 7.58766383e-02 -5.00977576e-01
6.55089259e-01 1.99328393e-01 8.33172381e-01 -3.22036117e-01
6.71145082e-01 1.12655592e+00 -3.39418203e-02 -8.30268621e-01
-1.21341980e+00 -3.41049761e-01 -5.61153531e-01 -1.17439348e-02
1.06542528e+00 -1.01496899e+00 -3.07268232e-01 6.86680079e-01
-9.92757440e-01 -6.64646089e-01 -6.20068848e-01 3.92190039e-01
-6.57965422e-01 7.31194904e-03 -8.92312825e-01 -3.01206559e-01
-3.98580045e-01 -1.01752377e+00 1.12437344e+00 5.56590497e-01
6.06369823e-02 -8.44308972e-01 -2.42726192e-01 5.01687862e-02
8.99925292e-01 1.18477181e-01 7.78338373e-01 4.31888476e-02
-8.45106900e-01 1.23785175e-01 -8.75094473e-01 6.41597629e-01
1.46003902e-01 -1.92681342e-01 -1.10953343e+00 -1.67638972e-01
-1.33068457e-01 -4.32381555e-02 1.53860533e+00 1.18107557e+00
1.44129205e+00 -1.12793632e-01 -2.19448224e-01 1.03155351e+00
1.55130959e+00 -3.19182813e-01 8.33589911e-01 3.07001442e-01
8.93692851e-01 5.90633869e-01 1.86837211e-01 3.70759182e-02
6.46706298e-02 3.85140866e-01 2.86274880e-01 -6.39095783e-01
-6.45748198e-01 6.94117174e-02 -3.65084335e-02 1.95350945e-01
-1.99232116e-01 3.16997916e-01 -3.22636664e-01 7.82683492e-01
-1.80459845e+00 -8.51351261e-01 2.27397010e-01 2.13259888e+00
8.55608046e-01 -5.99125512e-02 -2.66337305e-01 -2.14047477e-01
8.97360265e-01 2.53113329e-01 -7.13883281e-01 -1.49020836e-01
-3.46484125e-01 6.90481901e-01 8.54490876e-01 8.70034337e-01
-1.34779918e+00 1.20953119e+00 6.57067966e+00 1.11347008e+00
-1.21859574e+00 5.28456308e-02 9.78720129e-01 -1.93371609e-01
-1.76991045e-01 -1.65459186e-01 -7.02788889e-01 2.44686410e-01
4.53623533e-01 1.92622513e-01 6.24285638e-01 7.49026775e-01
6.19680695e-02 -4.45612192e-01 -7.58606672e-01 1.02755094e+00
1.08694673e-01 -1.81313193e+00 2.67396092e-01 -1.77409306e-01
1.13899767e+00 1.03698261e-01 4.18761015e-01 -1.47205189e-01
3.18114102e-01 -1.11016309e+00 6.75688088e-01 8.31633806e-01
8.07607114e-01 -8.72237325e-01 2.74136424e-01 -1.19442925e-01
-1.01717556e+00 -8.23434517e-02 -6.77517414e-01 1.56305894e-01
-4.26569916e-02 8.60438943e-01 -7.82884732e-02 1.04322173e-02
1.05871022e+00 7.87053108e-01 -4.09483820e-01 1.09462225e+00
-9.17707384e-02 2.94653714e-01 -2.55044580e-01 3.90944391e-01
3.80596340e-01 -1.00930566e-02 4.45155829e-01 1.22046137e+00
2.34804139e-01 2.27998897e-01 -2.06149429e-01 1.01687241e+00
-1.37614548e-01 -9.97041091e-02 -2.94553667e-01 4.81402695e-01
1.99136049e-01 1.39223206e+00 -8.60759556e-01 -5.13155043e-01
-5.06693244e-01 1.37521613e+00 6.02566600e-01 6.84665918e-01
-4.38478947e-01 -5.47394514e-01 1.07045650e+00 2.38757417e-01
7.43227661e-01 3.20242532e-02 -4.00929302e-01 -1.16205180e+00
-1.01455033e-01 -5.27742445e-01 -1.66344177e-02 -9.71015573e-01
-1.25993204e+00 5.27373016e-01 -3.49278420e-01 -8.89870524e-01
3.06985348e-01 -6.14103496e-01 -5.77760816e-01 1.10390675e+00
-1.86281371e+00 -1.08299518e+00 -4.15800840e-01 7.23351419e-01
6.50600672e-01 8.24218839e-02 5.66498220e-01 1.49060428e-01
-1.26794651e-01 1.41311467e-01 1.41315728e-01 1.05673134e-01
7.45331705e-01 -1.24774575e+00 5.45984864e-01 9.38435614e-01
-7.51689970e-02 4.55878288e-01 5.76950610e-01 -5.14674485e-01
-8.86315286e-01 -1.31675982e+00 5.60878217e-01 -2.02528492e-01
4.35191154e-01 -1.19047284e-01 -1.18113017e+00 5.56622982e-01
1.21086791e-01 6.11664355e-01 1.14522196e-01 -3.68796885e-01
-3.87481749e-01 -2.18425989e-01 -1.29433107e+00 6.87788069e-01
9.27671075e-01 -5.36287844e-01 -2.42976829e-01 1.24997705e-01
7.15039432e-01 -6.24616504e-01 -5.76067507e-01 2.90645838e-01
6.22195899e-01 -1.22429597e+00 1.31358254e+00 -4.43695992e-01
5.56556463e-01 -3.92790914e-01 -1.78577408e-01 -1.22883618e+00
-7.10089743e-01 -2.97773212e-01 8.13385174e-02 6.94351435e-01
1.85246676e-01 -4.77797210e-01 9.64118183e-01 5.28030455e-01
-1.49167120e-01 -7.13014603e-01 -8.59561861e-01 -2.11644158e-01
-1.18558519e-01 -1.70136631e-01 4.51342911e-01 7.81606138e-01
-7.09863365e-01 -8.05727243e-02 -3.15239996e-01 2.72750616e-01
1.03234160e+00 1.15779318e-01 4.52251464e-01 -1.05282354e+00
-1.35240203e-04 -5.54068625e-01 -4.84312624e-01 -1.20577872e+00
1.30395323e-01 -4.50340480e-01 4.36318479e-02 -1.40202034e+00
2.48713404e-01 -5.97381070e-02 -3.98974627e-01 4.10431981e-01
-3.99624527e-01 6.47693396e-01 -1.28235444e-01 3.49885710e-02
-5.11521518e-01 3.24702054e-01 1.45426941e+00 -1.57341287e-01
-2.57983953e-01 -2.17873082e-01 -7.01304734e-01 9.25687015e-01
6.45019233e-01 -8.93026963e-02 -2.01668084e-01 -6.43252850e-01
-3.29247296e-01 -2.80226886e-01 6.68190241e-01 -1.11983514e+00
3.21627438e-01 -9.52524841e-02 9.08386350e-01 -6.09539866e-01
3.73058647e-01 -5.00535190e-01 -7.05016926e-02 1.58393517e-01
-6.02314949e-01 -4.21814024e-01 3.41836482e-01 5.57698727e-01
-2.97140628e-01 9.85129476e-02 1.08514714e+00 -2.11803168e-01
-1.09016013e+00 5.21318376e-01 -2.98703671e-01 -2.96759307e-02
8.02256465e-01 -2.92604744e-01 -1.53418332e-01 -2.56305814e-01
-9.87132311e-01 -3.01635146e-01 5.50033927e-01 1.73609480e-01
1.00470090e+00 -1.41083181e+00 -6.75076783e-01 3.84471834e-01
-1.71883896e-01 3.43468077e-02 6.17150724e-01 8.18148911e-01
-5.90240836e-01 1.98455572e-01 -5.77431619e-01 -4.04020458e-01
-1.10165584e+00 3.21901888e-01 5.36144733e-01 3.94737311e-02
-8.70528042e-01 1.20266259e+00 6.70641541e-01 -1.29421398e-01
2.76895076e-01 -6.16337538e-01 -1.21301860e-01 -3.52707267e-01
9.60117042e-01 3.79279494e-01 -1.25898356e-02 -4.63149577e-01
-9.38552618e-02 8.53474796e-01 -1.89682096e-01 1.29134068e-02
1.64451194e+00 -2.18299955e-01 -4.11191046e-01 1.45130038e-01
1.09606433e+00 -6.36624396e-02 -1.97061539e+00 -7.37383068e-01
-3.34463805e-01 -6.82816684e-01 4.88995969e-01 -6.42043233e-01
-1.45623469e+00 6.50978088e-01 7.49118209e-01 -1.77315280e-01
1.25773549e+00 4.38060462e-02 7.75151372e-01 -6.77269846e-02
2.45514624e-02 -1.07847261e+00 6.36506230e-02 3.09685320e-01
8.45433652e-01 -1.23919463e+00 4.43516113e-02 -4.67090130e-01
-4.17566597e-01 1.13031042e+00 5.29010057e-01 -7.50730455e-01
7.23717213e-01 4.52647835e-01 -3.41789541e-03 -1.18956022e-01
-5.84668338e-01 -5.46886504e-01 6.02580369e-01 8.14662397e-01
3.36290836e-01 6.09211512e-02 1.02593578e-01 4.81218547e-02
2.62889743e-01 -3.14012952e-02 2.85774797e-01 7.69651949e-01
-8.20060253e-01 -5.62469065e-01 -3.55546385e-01 5.34503639e-01
-4.74164695e-01 -4.29360688e-01 -2.30359677e-02 4.72120643e-01
1.11739412e-01 3.63485664e-01 3.36373419e-01 -1.19744707e-02
2.38560259e-01 -2.46686548e-01 5.98428726e-01 -4.90759462e-01
-3.40036124e-01 2.91712955e-02 -4.14324135e-01 -1.03710604e+00
-6.33582652e-01 -4.30376679e-01 -1.02452815e+00 -2.07624450e-01
4.16875370e-02 -3.28091621e-01 6.12410545e-01 6.57688439e-01
3.75032008e-01 7.25213170e-01 2.04193816e-01 -1.40994036e+00
-1.87366351e-01 -8.37289929e-01 -7.38151908e-01 3.46336246e-01
6.23110354e-01 -5.21350265e-01 -4.16007638e-01 2.00443551e-01] | [10.895106315612793, -1.4665940999984741] |
ee45676d-0416-4e8b-a00d-555e14f5a57e | a-continuum-of-generation-tasks-for | 2210.10817 | null | https://arxiv.org/abs/2210.10817v1 | https://arxiv.org/pdf/2210.10817v1.pdf | A Continuum of Generation Tasks for Investigating Length Bias and Degenerate Repetition | Language models suffer from various degenerate behaviors. These differ between tasks: machine translation (MT) exhibits length bias, while tasks like story generation exhibit excessive repetition. Recent work has attributed the difference to task constrainedness, but evidence for this claim has always involved many confounding variables. To study this question directly, we introduce a new experimental framework that allows us to smoothly vary task constrainedness, from MT at one end to fully open-ended generation at the other, while keeping all other aspects fixed. We find that: (1) repetition decreases smoothly with constrainedness, explaining the difference in repetition across tasks; (2) length bias surprisingly also decreases with constrainedness, suggesting some other cause for the difference in length bias; (3) across the board, these problems affect the mode, not the whole distribution; (4) the differences cannot be attributed to a change in the entropy of the distribution, since another method of changing the entropy, label smoothing, does not produce the same effect. | ['David Chiang', 'Darcey Riley'] | 2022-10-19 | null | null | null | null | ['story-generation'] | ['natural-language-processing'] | [ 2.41415009e-01 2.07540423e-01 -3.65291268e-01 -1.36816740e-01
-7.27841198e-01 -8.59887660e-01 1.00775361e+00 -9.36884061e-03
-4.79590058e-01 9.16814089e-01 7.16149628e-01 -4.63912785e-01
5.68786077e-02 -3.47679138e-01 -7.12460637e-01 -6.94583118e-01
4.76717144e-01 4.15055782e-01 2.46763721e-01 -1.02445558e-01
5.61019719e-01 -7.70357996e-02 -1.09688640e+00 7.50150532e-02
8.50511611e-01 1.80029526e-01 1.78304017e-01 4.68975365e-01
-1.60202637e-01 6.42717540e-01 -6.02528453e-01 -3.85105848e-01
3.10597450e-01 -7.63024211e-01 -9.20713782e-01 4.01526242e-02
3.31649631e-01 4.27195169e-02 -2.74023190e-02 1.01773703e+00
3.83722454e-01 1.30063906e-01 9.89070892e-01 -1.09905112e+00
-6.72008574e-01 8.67017865e-01 -6.92185462e-01 2.45437529e-02
2.99311727e-01 3.93160671e-01 8.93010199e-01 -5.43674588e-01
8.53255510e-01 1.50975490e+00 6.21861815e-01 4.29768503e-01
-1.60709357e+00 -4.45932060e-01 1.00253485e-01 -4.75439548e-01
-1.32105422e+00 -5.92926919e-01 4.35816973e-01 -8.13814998e-01
5.16837478e-01 1.86370105e-01 3.29130560e-01 1.41598725e+00
5.13601840e-01 3.43117751e-02 1.54683435e+00 -4.89839226e-01
-1.87671243e-03 3.46576661e-01 1.61421746e-01 3.75422210e-01
5.45622051e-01 7.01619312e-02 -5.16084909e-01 -3.25427175e-01
6.58910096e-01 -4.16762322e-01 -3.92235100e-01 -5.58945835e-02
-1.58006060e+00 8.24253082e-01 -4.68948558e-02 6.52067780e-01
3.23585048e-02 2.42883191e-01 4.57777411e-01 6.00192189e-01
5.22706568e-01 1.01468062e+00 -5.45168102e-01 -3.70639563e-01
-7.98479557e-01 6.03582740e-01 8.16754222e-01 7.02018857e-01
6.87770307e-01 -1.83223262e-01 -5.03887177e-01 8.72574151e-01
-4.29578573e-02 2.97450453e-01 7.78444707e-01 -9.54870462e-01
6.51251912e-01 3.13611299e-01 3.67705137e-01 -7.12618470e-01
-5.97521663e-01 -2.26369157e-01 -2.77685851e-01 9.78464708e-02
1.29785693e+00 -5.71099579e-01 -7.23399460e-01 2.37802196e+00
-1.98441833e-01 -5.82298577e-01 -4.08640653e-01 1.10100245e+00
2.66119391e-01 4.76893216e-01 2.96916127e-01 -3.05060118e-01
1.44246864e+00 -6.59152091e-01 -7.20368147e-01 -6.20332599e-01
9.00267780e-01 -1.02840269e+00 1.57952619e+00 6.65254146e-02
-1.05184448e+00 -3.47092390e-01 -8.80067885e-01 -2.87262410e-01
-2.14050114e-01 -2.31309667e-01 5.11936367e-01 6.09396636e-01
-9.03270841e-01 7.40478575e-01 -4.76545185e-01 -5.50371349e-01
-2.94926614e-01 8.33494291e-02 -2.45776668e-01 1.23254888e-01
-1.14403498e+00 1.28733921e+00 2.92755216e-01 -5.45902669e-01
-1.81563407e-01 -5.34908712e-01 -6.79244220e-01 9.12893638e-02
3.81845325e-01 -8.85528564e-01 1.26276326e+00 -1.36982584e+00
-1.39044499e+00 8.51710260e-01 -2.86517531e-01 9.70141068e-02
7.66505480e-01 -8.96571428e-02 6.63813204e-02 -6.22737348e-01
3.60243529e-01 6.84843421e-01 8.08920443e-01 -1.21792614e+00
-4.42138761e-02 -4.50250745e-01 -2.28022099e-01 4.53796566e-01
-3.39891911e-02 4.82957363e-02 -8.78247768e-02 -8.81315887e-01
2.25429679e-03 -1.16248775e+00 -1.03678843e-02 -4.11842257e-01
-3.56120616e-01 -2.77390510e-01 2.04294100e-01 -5.64922571e-01
1.15530217e+00 -2.15501571e+00 1.80114880e-01 -2.19770819e-01
1.65951520e-01 -2.97575146e-01 -2.86149681e-01 4.74662751e-01
-1.23778433e-01 5.26922226e-01 -3.56656283e-01 -3.21413130e-01
1.64653957e-01 -4.59901877e-02 -2.11560890e-01 4.47984159e-01
1.82359084e-01 8.12811792e-01 -6.85605466e-01 -3.14567924e-01
-4.23519582e-01 1.90678164e-02 -4.99733210e-01 -2.42896274e-01
-2.21805364e-01 3.95250916e-01 -6.48716167e-02 1.27180398e-01
6.03702247e-01 -5.51866032e-02 2.20416620e-01 3.85294557e-01
-3.57477009e-01 7.28999913e-01 -7.76572287e-01 1.45223796e+00
-2.25655481e-01 1.08300364e+00 -7.34786317e-02 -6.94080830e-01
6.59627020e-01 3.91793728e-01 1.59297034e-03 -6.35098636e-01
-6.31906278e-03 4.75525975e-01 6.15725935e-01 -5.59419930e-01
7.99804389e-01 -5.96529543e-01 -4.85028233e-03 8.55617344e-01
-3.39461386e-01 -3.61865312e-01 1.66038826e-01 -3.52354832e-02
1.05208981e+00 1.47971809e-02 -1.61148347e-02 -4.50466841e-01
-2.37458959e-01 1.34415507e-01 6.56784117e-01 9.07042563e-01
-2.26373821e-01 6.78146899e-01 9.43632483e-01 -9.25664827e-02
-1.42095637e+00 -8.24572206e-01 -3.00800651e-01 1.11289001e+00
-1.29106678e-02 -1.64997697e-01 -9.63846803e-01 -5.08812785e-01
5.27089611e-02 1.02025187e+00 -6.51167810e-01 -1.64851546e-01
-5.10972083e-01 -8.25758338e-01 5.77151895e-01 1.54036328e-01
2.71819681e-01 -9.61078942e-01 -5.80742598e-01 1.73305124e-02
-6.04514658e-01 -1.09984756e+00 -8.00885558e-01 1.40612438e-01
-1.05956602e+00 -5.94941556e-01 -7.32889771e-01 -4.21677023e-01
4.85209942e-01 2.96599180e-01 1.27233207e+00 1.78684413e-01
4.27468151e-01 -6.88683912e-02 -1.29307613e-01 -6.27400219e-01
-6.35364890e-01 5.42197347e-01 -7.63087766e-03 -5.53967416e-01
2.01197103e-01 -4.35636461e-01 -2.18004391e-01 4.08647001e-01
-7.83262491e-01 -5.73948622e-02 5.72326660e-01 6.02115333e-01
1.88542306e-02 -3.23148608e-01 9.04592872e-01 -9.72857237e-01
1.06054473e+00 -5.61127782e-01 -2.33026683e-01 3.89561104e-03
-7.83829510e-01 2.81025499e-01 3.66258174e-01 -8.26043665e-01
-8.90424550e-01 -3.57370257e-01 3.56203169e-01 5.30914627e-02
-1.33859739e-01 3.95890117e-01 1.29035205e-01 3.80727649e-01
8.85008752e-01 3.87054384e-02 2.31912240e-01 -4.18560237e-01
1.01754770e-01 4.71469194e-01 9.57111195e-02 -6.54011548e-01
8.66753995e-01 1.29541785e-01 -2.12132573e-01 -7.24951684e-01
-7.28143275e-01 1.88143268e-01 -3.70242327e-01 1.25328988e-01
1.00284410e+00 -7.92723358e-01 -1.59991801e-01 1.88454688e-01
-1.20066965e+00 -9.06183302e-01 -1.90548614e-01 8.12889814e-01
-5.76372325e-01 1.10135660e-01 -7.35214293e-01 -6.78549170e-01
2.02701967e-02 -1.25629735e+00 8.99285913e-01 -2.46169940e-02
-1.00282252e+00 -9.83625531e-01 2.10702986e-01 3.62011999e-01
5.48532009e-01 1.54741362e-01 1.33767653e+00 -6.88333333e-01
-3.05834770e-01 6.14130013e-02 -2.90098041e-01 6.42746780e-03
2.56649137e-01 1.14395007e-01 -6.79449499e-01 -1.13226444e-01
2.67429024e-01 -2.61873454e-01 7.96993077e-01 5.74404597e-01
5.98663867e-01 -3.06116909e-01 -2.77310938e-01 2.03609213e-01
1.14312911e+00 3.52296163e-03 5.87673783e-01 3.30330551e-01
6.59214675e-01 8.87473583e-01 1.91321477e-01 -5.27939200e-02
3.33417207e-01 9.58816588e-01 -3.07904989e-01 -1.23017274e-01
-1.50581241e-01 -1.72640890e-01 7.85806477e-01 6.18908346e-01
-2.88404729e-02 -1.64677992e-01 -7.87298739e-01 5.04252315e-01
-1.77896714e+00 -9.98892486e-01 -4.11003649e-01 2.38278866e+00
9.82170403e-01 3.63777578e-01 3.02569628e-01 -4.81886566e-02
6.17814124e-01 2.65586436e-01 -3.71050745e-01 -6.83180630e-01
2.56637782e-02 -4.37831312e-01 6.03888810e-01 6.82965159e-01
-3.72557253e-01 7.26095080e-01 7.39077330e+00 7.58578718e-01
-1.21115708e+00 1.28532708e-01 5.77263057e-01 -2.45745882e-01
-6.11431658e-01 2.76461661e-01 -3.10498327e-01 6.99214160e-01
7.89822936e-01 -4.18963641e-01 4.27454501e-01 4.29625571e-01
4.58779186e-01 -4.14394855e-01 -1.26499069e+00 4.09095466e-01
-5.58490083e-02 -7.64121652e-01 -1.64342105e-01 3.39911371e-01
7.51317263e-01 -5.96541204e-02 -5.47620580e-02 3.20833176e-01
2.20483288e-01 -1.01903796e+00 1.15995753e+00 2.91482717e-01
7.00946808e-01 -5.64687073e-01 5.58669925e-01 6.93543255e-01
-4.71252412e-01 2.04979017e-01 -3.09859216e-01 -6.18844807e-01
9.70074236e-02 8.11301589e-01 -7.41767764e-01 5.62639795e-02
1.09583125e-01 -1.03020824e-01 -6.28281891e-01 4.64853674e-01
-1.94382057e-01 7.57114410e-01 1.12713855e-02 9.43197533e-02
5.87494522e-02 -4.04691875e-01 8.23057115e-01 1.20313561e+00
2.65798718e-01 -1.86578006e-01 -1.73689947e-01 1.15768921e+00
-7.99657851e-02 1.00365682e-02 -9.86042798e-01 -2.84626514e-01
5.31980753e-01 8.54251564e-01 -9.93659258e-01 -1.52530417e-01
-4.61709559e-01 9.30979133e-01 2.61192977e-01 4.58129615e-01
-8.25167656e-01 -2.30732560e-01 6.14464819e-01 4.37266737e-01
-1.30212069e-01 -4.16241705e-01 -8.51278484e-01 -1.14549160e+00
9.14325938e-02 -9.52036798e-01 -1.55231923e-01 -6.09785080e-01
-1.18294990e+00 1.53993115e-01 3.29018831e-02 -6.80755556e-01
-2.59658247e-01 -1.70835942e-01 -4.90197718e-01 1.15893996e+00
-1.00207925e+00 -4.64564115e-01 1.28515139e-01 2.61430293e-01
7.09720790e-01 2.44799510e-01 4.77624953e-01 1.21018238e-01
-5.60338616e-01 6.22066379e-01 -2.05865830e-01 -6.37128651e-02
1.28336978e+00 -1.19427502e+00 4.81838465e-01 6.14869833e-01
-2.16048248e-02 9.25480545e-01 9.47376013e-01 -7.40470588e-01
-9.81952250e-01 -7.22306550e-01 1.42924774e+00 -1.09886301e+00
5.10171354e-01 -4.83339161e-01 -9.83754396e-01 7.69673169e-01
3.27186882e-01 -6.40038967e-01 4.07293350e-01 4.25392568e-01
-5.77483416e-01 5.05007505e-01 -9.85609651e-01 9.43546176e-01
1.04918408e+00 -4.84442055e-01 -4.30058807e-01 3.85217637e-01
1.14054239e+00 -2.75951475e-01 -5.76857448e-01 1.42533690e-01
6.60645008e-01 -8.99874747e-01 3.10744673e-01 -5.42248130e-01
8.64124298e-01 -6.48647696e-02 -4.31531668e-02 -1.59279680e+00
-7.02438593e-01 -3.42739075e-01 4.26841021e-01 1.37755811e+00
9.32086706e-01 -8.95945489e-01 2.37963036e-01 7.75974751e-01
-1.03686944e-01 -6.26537621e-01 -4.97483671e-01 -1.02110112e+00
6.96764529e-01 -2.19803199e-01 5.46453476e-01 1.19149256e+00
1.75797537e-01 8.46440554e-01 -1.29436880e-01 -3.86766434e-01
9.38893408e-02 -2.32692778e-01 6.75357282e-01 -1.06918049e+00
-3.91196549e-01 -6.67077124e-01 2.40591958e-01 -8.22673261e-01
5.59631921e-02 -7.59220302e-01 1.81993380e-01 -1.30188906e+00
4.97284621e-01 -5.57979822e-01 3.86620998e-01 2.33300939e-01
-4.99413371e-01 3.25217694e-02 5.25908291e-01 5.37502229e-01
-1.54501181e-02 1.33122787e-01 1.34253287e+00 3.23827207e-01
-3.71853292e-01 -1.55977339e-01 -1.13088739e+00 7.38490880e-01
7.99601853e-01 -5.98156929e-01 -5.05363941e-01 -7.58102179e-01
5.22307217e-01 9.70948413e-02 1.65681347e-01 -7.92800725e-01
-3.30485880e-01 -3.69786412e-01 1.98020950e-01 3.13797623e-01
3.85191920e-03 -4.53845769e-01 1.22832887e-01 4.09368575e-01
-5.86685598e-01 3.97632629e-01 2.10213780e-01 2.86497921e-01
2.40851939e-01 -4.18622196e-01 7.65793204e-01 -9.16309953e-02
2.10880890e-01 -2.81904578e-01 -8.27928603e-01 4.38781351e-01
5.75039685e-01 -8.25262144e-02 -5.35006225e-01 -4.32684124e-01
-3.41158569e-01 -1.52808920e-01 9.59038496e-01 5.45937598e-01
-4.28068429e-01 -1.28144777e+00 -8.99946749e-01 -2.14688480e-01
-8.65871459e-02 -2.73449957e-01 -9.40028504e-02 1.08588183e+00
-2.24547330e-02 3.56960058e-01 1.97712213e-01 -3.59005153e-01
-8.38635445e-01 4.30594891e-01 2.66623676e-01 -4.71700668e-01
-3.17430705e-01 5.64272106e-01 3.26053113e-01 -3.06426287e-01
-9.69740674e-02 -1.90734982e-01 2.00976133e-01 3.48351359e-01
2.46401325e-01 3.76462698e-01 -1.12810880e-01 -3.34426582e-01
-2.28894204e-01 5.67868352e-01 -4.07931283e-02 -6.27899647e-01
8.19178820e-01 -2.39473253e-01 -3.71625572e-02 9.94190753e-01
8.29222560e-01 3.34434241e-01 -1.15958345e+00 -6.11404628e-02
1.36121914e-01 -3.24792176e-01 -3.24018151e-01 -6.61647499e-01
-6.18165314e-01 5.78832746e-01 1.44420624e-01 4.62121546e-01
5.02659917e-01 -5.14279418e-02 5.59289813e-01 5.56908436e-02
3.55645537e-01 -1.26301658e+00 -2.71042101e-02 6.09883845e-01
9.18205082e-01 -1.12102997e+00 6.71603680e-02 -2.73200423e-01
-9.98794079e-01 6.43717527e-01 4.79795635e-01 3.99741828e-02
6.87443540e-02 1.45027384e-01 6.10960871e-02 -1.98092163e-01
-1.05432999e+00 5.43769673e-02 2.09277049e-02 4.32524860e-01
1.07180047e+00 2.41705999e-01 -1.13603926e+00 3.32801163e-01
-6.26943469e-01 -1.04794510e-01 5.70165873e-01 5.14110208e-01
-2.74040610e-01 -1.19104970e+00 -6.15675688e-01 5.00280082e-01
-4.77134883e-01 -1.01985954e-01 -8.06649268e-01 9.30174232e-01
2.04141065e-01 1.00134873e+00 2.62176365e-01 -2.34455913e-01
1.17715232e-01 6.40689075e-01 4.87733901e-01 -8.62625360e-01
-6.69245124e-01 -5.65646216e-02 3.28189880e-01 -1.77779928e-01
7.45662004e-02 -8.35358918e-01 -1.05718303e+00 -5.84699452e-01
-5.37086308e-01 5.95018044e-02 6.36248648e-01 1.07672715e+00
2.79054642e-01 2.92335868e-01 4.17669624e-01 -8.21787596e-01
-7.57929027e-01 -1.33558416e+00 -5.32482982e-01 6.96694434e-01
2.54777849e-01 -6.63228333e-01 -7.01093197e-01 2.08782647e-02] | [11.4000244140625, 9.438535690307617] |
43db8815-8d98-47f0-82f5-f37c39b8d578 | survey-of-hallucination-in-natural-language | 2202.03629 | null | https://arxiv.org/abs/2202.03629v5 | https://arxiv.org/pdf/2202.03629v5.pdf | Survey of Hallucination in Natural Language Generation | Natural Language Generation (NLG) has improved exponentially in recent years thanks to the development of sequence-to-sequence deep learning technologies such as Transformer-based language models. This advancement has led to more fluent and coherent NLG, leading to improved development in downstream tasks such as abstractive summarization, dialogue generation and data-to-text generation. However, it is also apparent that deep learning based generation is prone to hallucinate unintended text, which degrades the system performance and fails to meet user expectations in many real-world scenarios. To address this issue, many studies have been presented in measuring and mitigating hallucinated texts, but these have never been reviewed in a comprehensive manner before. In this survey, we thus provide a broad overview of the research progress and challenges in the hallucination problem in NLG. The survey is organized into two parts: (1) a general overview of metrics, mitigation methods, and future directions; and (2) an overview of task-specific research progress on hallucinations in the following downstream tasks, namely abstractive summarization, dialogue generation, generative question answering, data-to-text generation, machine translation, and visual-language generation. This survey serves to facilitate collaborative efforts among researchers in tackling the challenge of hallucinated texts in NLG. | ['Pascale Fung', 'Andrea Madotto', 'Wenliang Dai', 'Yejin Bang', 'Etsuko Ishii', 'Yan Xu', 'Dan Su', 'Tiezheng Yu', 'Rita Frieske', 'Nayeon Lee', 'Ziwei Ji'] | 2022-02-08 | null | null | null | null | ['generative-question-answering', 'data-to-text-generation'] | ['natural-language-processing', 'natural-language-processing'] | [ 4.45625365e-01 6.28689587e-01 2.06507832e-01 -6.11987114e-02
-1.16261303e+00 -5.15961051e-01 9.36835229e-01 1.56065404e-01
7.57559463e-02 1.09457278e+00 1.21141434e+00 -2.58871578e-02
3.92803669e-01 -6.18739426e-01 -1.99255928e-01 -3.52129668e-01
2.53092945e-01 6.47762418e-01 -4.22024608e-01 -6.71798468e-01
4.38609153e-01 1.90044224e-01 -1.19240761e+00 8.07728112e-01
1.22992706e+00 3.34751487e-01 2.12923110e-01 9.96290505e-01
-2.60198712e-01 1.23767388e+00 -1.42119133e+00 -4.70056355e-01
-3.59877914e-01 -1.26284063e+00 -1.17065430e+00 1.47116661e-01
2.65864104e-01 -4.25601900e-01 -4.76232260e-01 7.16095090e-01
1.27989697e+00 1.60002023e-01 6.03977263e-01 -1.23101616e+00
-9.50149953e-01 6.89855993e-01 1.87296681e-02 8.69257078e-02
9.24721718e-01 4.29532886e-01 8.34483743e-01 -9.73236084e-01
8.11106682e-01 1.48126602e+00 5.47425091e-01 9.68539476e-01
-8.70145798e-01 -3.04875851e-01 -2.69723445e-01 7.55409598e-02
-8.71925592e-01 -7.28054702e-01 4.36542749e-01 -2.61907578e-01
1.70699263e+00 3.67566824e-01 8.13627779e-01 1.57898223e+00
4.48880285e-01 1.16951728e+00 7.09883988e-01 -3.87106776e-01
3.73331197e-02 8.00733492e-02 -3.84980798e-01 4.17071432e-01
2.06510536e-02 -6.07036538e-02 -9.65257049e-01 -2.82025367e-01
3.08937579e-01 -7.84604847e-01 -4.07937765e-01 3.64623964e-01
-1.34865499e+00 1.14025843e+00 6.50562569e-02 3.18473339e-01
-5.02426744e-01 -1.56491935e-01 6.67288840e-01 3.25463772e-01
6.79840088e-01 1.11969471e+00 5.78642860e-02 -3.78375083e-01
-1.14648640e+00 8.36624026e-01 1.23666155e+00 9.51239586e-01
1.06812373e-01 7.94759572e-01 -7.65829623e-01 1.01069224e+00
-6.92135096e-02 5.27387917e-01 9.94416893e-01 -9.26912844e-01
7.54897177e-01 2.57711828e-01 2.04222724e-02 -8.54899108e-01
-4.21340019e-01 -2.36109123e-01 -1.10590005e+00 -2.06686944e-01
-2.09312607e-02 -6.72466993e-01 -7.56145716e-01 1.60366035e+00
-2.97953159e-01 -5.15297055e-01 6.45818532e-01 6.86854184e-01
1.42378974e+00 1.09653819e+00 -1.18282065e-01 -3.99732023e-01
1.13473642e+00 -9.89311695e-01 -1.05214298e+00 -3.98081303e-01
5.55357039e-01 -1.08657730e+00 9.35200572e-01 2.29449525e-01
-1.57487786e+00 -3.72443974e-01 -8.99349570e-01 -4.48470652e-01
-1.98212802e-01 -2.60787457e-01 3.76125395e-01 4.74638700e-01
-1.39984798e+00 4.66196209e-01 -4.31095153e-01 -7.13127494e-01
2.15104654e-01 -3.63287441e-02 -6.51871935e-02 2.19548464e-01
-1.70091534e+00 1.44241011e+00 5.57219148e-01 -8.96211118e-02
-9.20325220e-01 -5.52796423e-01 -9.16234672e-01 -5.88496402e-02
-8.68196338e-02 -1.29581630e+00 1.77923954e+00 -5.07551610e-01
-1.61571085e+00 6.70692682e-01 -2.32368231e-01 -7.29016900e-01
4.93601829e-01 -3.29638630e-01 -3.74045283e-01 7.92192370e-02
1.48497954e-01 9.26906347e-01 4.50271368e-01 -1.05604208e+00
-2.85800725e-01 -2.03334734e-01 -1.99227944e-01 1.04200578e+00
1.22231953e-02 1.63285747e-01 2.36503690e-01 -6.67064667e-01
-4.99437571e-01 -4.82717037e-01 -3.07832092e-01 -7.41351902e-01
-8.71893585e-01 -3.65784913e-01 6.86029911e-01 -9.41386163e-01
1.27744591e+00 -1.57117712e+00 5.31592593e-02 -7.19952643e-01
2.45778546e-01 5.14586806e-01 -1.85084060e-01 1.37618506e+00
7.01935031e-04 9.22663510e-02 -2.74340868e-01 -4.94948834e-01
-4.82541025e-02 -1.21195436e-01 -1.05084288e+00 -1.42743513e-01
3.07063907e-01 1.39083302e+00 -1.22917140e+00 -3.45033556e-01
2.36241698e-01 2.79597610e-01 -1.46925896e-01 4.60160583e-01
-4.29614723e-01 3.74038577e-01 -1.36801571e-01 3.04291338e-01
1.28959969e-01 -6.83038682e-02 -1.35681868e-01 2.70419091e-01
-2.07995936e-01 7.30872929e-01 -2.35726088e-01 1.64334881e+00
-4.73495543e-01 9.28705037e-01 -3.33167136e-01 -7.41438866e-01
9.32754636e-01 9.52509344e-01 -1.70174073e-02 -6.45369351e-01
-1.73811793e-01 3.54573995e-01 -5.35369664e-02 -6.59647286e-01
1.19605029e+00 -4.91046429e-01 -4.09354627e-01 6.78868711e-01
1.36016965e-01 -8.37758720e-01 3.11024755e-01 5.82773983e-01
8.94291759e-01 -3.22710603e-01 7.61284828e-01 1.67186990e-01
1.72797129e-01 3.21835548e-01 -7.09352270e-02 9.78140831e-01
1.39422774e-01 8.20503473e-01 4.65533346e-01 -3.16758007e-02
-1.17213631e+00 -1.07147634e+00 4.78484750e-01 6.42809391e-01
-2.38376588e-01 -4.96492803e-01 -1.07297075e+00 -1.99181274e-01
-3.57230842e-01 1.50436246e+00 -2.61265427e-01 -4.86647993e-01
-4.86577064e-01 -9.58154142e-01 1.15545785e+00 4.17381674e-01
4.77762878e-01 -1.83457279e+00 -5.26981592e-01 4.87675637e-01
-9.37999010e-01 -8.94564092e-01 -3.09181482e-01 -3.47636014e-01
-9.92301464e-01 -5.62118053e-01 -1.01871693e+00 -7.17933774e-01
2.15323210e-01 1.72627866e-01 1.34453642e+00 -2.88444519e-01
-1.64267719e-01 3.03092957e-01 -3.81725550e-01 -6.77299261e-01
-1.12073958e+00 2.06505090e-01 -7.77918324e-02 -5.10424614e-01
1.35127991e-01 -4.55744535e-01 -3.69003057e-01 -4.96170044e-01
-9.78412390e-01 3.75045598e-01 6.70515358e-01 9.56504583e-01
-2.57027801e-02 -5.43532610e-01 1.19754195e+00 -7.50864029e-01
1.91862988e+00 -5.30535817e-01 2.07629636e-01 2.82536745e-01
-4.82089490e-01 -1.92827322e-02 8.79898548e-01 -1.48895323e-01
-1.27737224e+00 -6.19748533e-01 -4.30723011e-01 1.95711464e-01
-1.85098097e-01 7.45313346e-01 -1.14345457e-02 6.94495142e-01
1.08291531e+00 7.55048215e-01 1.16371341e-01 -2.33660460e-01
6.89292789e-01 1.02457607e+00 6.83564425e-01 -4.36503328e-02
5.28614879e-01 -1.04149431e-02 -5.38207352e-01 -1.23387623e+00
-8.06744576e-01 -1.15715556e-01 -1.43383592e-01 -1.65725008e-01
8.16656709e-01 -8.56498659e-01 -6.71184883e-02 7.89465785e-01
-1.71697533e+00 -4.07659531e-01 -5.73064387e-01 4.52229194e-02
-9.61278260e-01 5.40381491e-01 -8.88594866e-01 -8.72526467e-01
-1.15144849e+00 -9.46235418e-01 1.09852231e+00 2.77816623e-01
-1.03214514e+00 -1.10313606e+00 3.53828788e-01 5.17086029e-01
5.53614974e-01 4.12905514e-01 1.01915729e+00 -8.89982343e-01
-2.09180862e-01 -1.75322235e-01 3.11827958e-02 3.45402688e-01
6.65104613e-02 -4.10539329e-01 -8.41118634e-01 -7.68203437e-02
2.19473436e-01 -7.66339719e-01 7.61757255e-01 2.29315832e-01
5.96153498e-01 -7.65386701e-01 -4.47741412e-02 1.41307414e-01
6.76237464e-01 2.77978957e-01 9.43199635e-01 -1.39328778e-01
4.45442528e-01 7.74760127e-01 4.00928229e-01 4.64897722e-01
4.33177322e-01 4.54634994e-01 -4.83321073e-03 -1.58369616e-01
-6.12240493e-01 -6.69391334e-01 6.18859231e-01 9.72542524e-01
1.21599011e-01 -9.82658386e-01 -5.92444897e-01 6.10005379e-01
-1.75386167e+00 -1.42921257e+00 -3.15373391e-01 1.74923670e+00
1.01757145e+00 -4.37960088e-01 -8.15916210e-02 -1.05122112e-01
7.00492859e-01 5.53947449e-01 -5.93146801e-01 -8.91867161e-01
-3.62123013e-01 9.31141153e-02 -3.05920988e-01 7.32192278e-01
-5.63103974e-01 1.18981075e+00 7.21271610e+00 6.49161637e-01
-1.11242402e+00 6.50274232e-02 5.62011003e-01 -1.52188778e-01
-4.48578864e-01 -2.96481520e-01 -4.53912020e-01 2.92150438e-01
1.12182820e+00 -9.20298874e-01 3.36180598e-01 4.71922487e-01
6.02016687e-01 4.40065414e-02 -9.69829440e-01 9.06763494e-01
5.79338431e-01 -1.50845909e+00 5.59184313e-01 -2.34636024e-01
9.51331556e-01 1.06812075e-01 7.07436055e-02 5.44014752e-01
2.97577858e-01 -1.25635111e+00 6.33035004e-01 3.57262939e-01
8.38793218e-01 -6.71209514e-01 6.57631397e-01 5.46577692e-01
-4.37366605e-01 1.56779543e-01 -2.89206415e-01 -2.86522955e-01
7.30412006e-01 5.69346189e-01 -1.47725093e+00 5.17354310e-01
-6.76191505e-03 3.55753809e-01 -2.14899704e-01 8.93354952e-01
-4.89138514e-01 4.97152448e-01 4.17518049e-01 -2.67309546e-01
2.64241874e-01 1.62821531e-01 9.39431727e-01 1.35866201e+00
5.21918416e-01 2.20707893e-01 -1.10952303e-01 9.83487785e-01
-1.17988326e-01 1.58011988e-01 -8.87885928e-01 -4.93481100e-01
3.63644361e-01 9.32279885e-01 -1.28566816e-01 -5.89568615e-01
-5.90088889e-02 1.45111454e+00 6.99699745e-02 4.56867486e-01
-5.11309326e-01 -5.81713378e-01 3.73641700e-01 1.92953292e-02
-5.11174738e-01 -8.24663565e-02 -4.29840267e-01 -1.12948406e+00
-2.61759371e-01 -1.32550097e+00 1.64245203e-01 -1.20113862e+00
-1.23733222e+00 8.97594929e-01 1.12806551e-01 -8.48214865e-01
-1.19939590e+00 9.57910344e-02 -9.50050592e-01 1.20594049e+00
-8.88999104e-01 -8.09927166e-01 -1.73915923e-01 3.43737811e-01
1.09982193e+00 -3.73427391e-01 1.24330711e+00 -2.88554579e-01
-2.74639368e-01 3.49426746e-01 -2.27804333e-02 -1.66594610e-01
8.98459375e-01 -1.16497278e+00 1.05620837e+00 7.90723264e-01
-8.50710049e-02 5.29093385e-01 9.31043923e-01 -8.89131844e-01
-9.26513195e-01 -1.16160166e+00 1.58589840e+00 -5.09202719e-01
3.99470627e-01 -3.24240476e-01 -6.89642012e-01 5.51122069e-01
9.25508797e-01 -1.05242956e+00 6.87853813e-01 -3.15381050e-01
4.66893949e-02 5.13315320e-01 -9.87358093e-01 9.56628919e-01
7.94404626e-01 -5.04498363e-01 -1.00076890e+00 7.69321501e-01
8.80284846e-01 -6.25612259e-01 -4.40603167e-01 8.51180926e-02
2.15548173e-01 -1.00223660e+00 6.51185393e-01 -7.13122070e-01
9.29567993e-01 1.09837130e-01 3.96536916e-01 -1.93431008e+00
-1.20886080e-02 -1.22421730e+00 -2.86820143e-01 1.03259504e+00
5.68479121e-01 -5.57381392e-01 4.53328609e-01 5.09021163e-01
-7.57684708e-01 -6.98438346e-01 -4.47570294e-01 -4.39101070e-01
2.57482469e-01 -4.83861677e-02 3.98846388e-01 8.14489365e-01
5.14136851e-01 1.09714329e+00 -5.97423017e-01 -5.42790890e-01
1.98263273e-01 1.23258270e-02 7.39728749e-01 -8.19089413e-01
-2.65152231e-02 -6.40947998e-01 3.53461839e-02 -1.09807861e+00
1.07159717e-02 -1.05678952e+00 3.02074671e-01 -2.35990000e+00
3.04512262e-01 5.07813156e-01 6.66632533e-01 3.33237618e-01
-3.42781007e-01 1.58695593e-01 3.31066340e-01 1.34404674e-01
-4.15818423e-01 8.47697437e-01 1.44760752e+00 -2.42821574e-02
-3.43521148e-01 4.29418460e-02 -1.31995249e+00 4.85953480e-01
1.06638193e+00 -1.26537934e-01 -6.58648133e-01 -4.45059687e-01
1.64219812e-01 6.80567741e-01 1.63956180e-01 -7.99358487e-01
1.51772231e-01 7.45577663e-02 3.54281813e-01 -8.23803902e-01
5.06158292e-01 3.88431311e-01 -8.01045299e-02 4.81484026e-01
-7.58341789e-01 2.47737631e-01 1.39424697e-01 1.50314018e-01
-4.71482247e-01 -3.36075455e-01 5.58324993e-01 -5.12646079e-01
-4.12946582e-01 -3.15501653e-02 -1.06121743e+00 5.18583715e-01
6.09279871e-01 -1.24076702e-01 -5.58714986e-01 -1.49824071e+00
-5.32692671e-01 2.29916871e-01 1.79342255e-01 6.73202872e-01
9.04955864e-01 -1.13080192e+00 -1.37956333e+00 -2.44732887e-01
2.27500293e-02 -1.48771748e-01 4.17266101e-01 6.46740079e-01
-4.53297943e-01 8.64806175e-01 -1.75337568e-02 -7.51034468e-02
-1.05842519e+00 2.21529126e-01 3.07751447e-01 -5.89657784e-01
-5.33387482e-01 7.95440137e-01 1.67004779e-01 -4.24730778e-01
-6.12734221e-02 5.62010966e-02 -1.90376684e-01 5.85356615e-02
8.24444890e-01 5.15777946e-01 1.46449342e-01 -5.12049437e-01
1.15811318e-01 -3.30987871e-01 -1.70393452e-01 -5.10913968e-01
8.51067007e-01 -1.16116844e-01 -2.04068810e-01 4.11808819e-01
9.11563456e-01 -1.63548470e-01 -5.26056886e-01 2.31741205e-01
-2.28281781e-01 1.94357097e-01 -3.49193156e-01 -1.22616041e+00
-3.87615025e-01 1.04381728e+00 -2.43006676e-01 4.76653546e-01
8.92317593e-01 -1.48676764e-02 1.36781001e+00 4.51487899e-01
3.00585777e-02 -9.54203606e-01 4.55571353e-01 1.07673645e+00
1.64144742e+00 -6.59304619e-01 -2.54448265e-01 5.37144393e-02
-1.18394339e+00 1.06654084e+00 5.50479293e-01 2.40727454e-01
-2.74312168e-01 1.43515870e-01 2.24909931e-01 -1.66076779e-01
-1.03966641e+00 8.93819705e-02 1.43973514e-01 8.55478227e-01
8.65441322e-01 -2.60622110e-02 -3.04547101e-01 3.31821203e-01
-1.08783817e+00 -1.20935626e-01 8.49629998e-01 4.42782611e-01
-4.75528032e-01 -1.13988280e+00 -4.45833176e-01 5.61697960e-01
-2.42894888e-01 -5.31891882e-01 -1.13113678e+00 5.74845254e-01
-4.16120529e-01 1.45362139e+00 -5.82275391e-02 -1.95921436e-01
3.33550155e-01 3.42890263e-01 4.98236924e-01 -1.05472422e+00
-8.25730026e-01 -2.20663339e-01 6.64490819e-01 -3.26651782e-01
1.30603677e-02 -5.87806106e-01 -1.21317434e+00 -3.78150523e-01
-1.55523121e-01 3.63201767e-01 4.15065795e-01 8.96023750e-01
5.52841246e-01 4.94169742e-01 2.09536761e-01 -8.35996807e-01
-8.72212291e-01 -1.41960835e+00 -2.08142951e-01 1.99801356e-01
1.55549422e-01 1.90063179e-01 -2.38400877e-01 -1.02934755e-01] | [12.072305679321289, 9.127997398376465] |
726b0eba-7832-4241-80c6-88f3aa807bfe | deep-set-conditioned-latent-representations-1 | 2212.11030 | null | https://arxiv.org/abs/2212.11030v1 | https://arxiv.org/pdf/2212.11030v1.pdf | Deep set conditioned latent representations for action recognition | In recent years multi-label, multi-class video action recognition has gained significant popularity. While reasoning over temporally connected atomic actions is mundane for intelligent species, standard artificial neural networks (ANN) still struggle to classify them. In the real world, atomic actions often temporally connect to form more complex composite actions. The challenge lies in recognising composite action of varying durations while other distinct composite or atomic actions occur in the background. Drawing upon the success of relational networks, we propose methods that learn to reason over the semantic concept of objects and actions. We empirically show how ANNs benefit from pretraining, relational inductive biases and unordered set-based latent representations. In this paper we propose deep set conditioned I3D (SCI3D), a two stream relational network that employs latent representation of state and visual representation for reasoning over events and actions. They learn to reason about temporally connected actions in order to identify all of them in the video. The proposed method achieves an improvement of around 1.49% mAP in atomic action recognition and 17.57% mAP in composite action recognition, over a I3D-NL baseline, on the CATER dataset. | ['Steven Latre', 'Jose Oramas', 'Peter Hellinckx', 'Kevin Mets', 'Tom De Schepper', 'Akash Singh'] | 2022-12-21 | deep-set-conditioned-latent-representations | https://www.scitepress.org/Link.aspx?doi=10.5220/0010838400003124 | https://orbi.uliege.be/bitstream/2268/290532/1/SinghAl-deepSetConditionedRepresentations_VISAPP_2022.pdf | international-joint-conference-on-computer-1 | ['composite-action-recognition', 'atomic-action-recognition'] | ['computer-vision', 'computer-vision'] | [ 7.51717091e-01 1.33080613e-02 -4.77636576e-01 -4.07380998e-01
-4.38898265e-01 -5.32141268e-01 1.02838755e+00 -2.86605638e-02
-1.33306086e-01 3.87802124e-01 5.34203827e-01 -2.26361817e-03
-2.17979133e-01 -5.50488710e-01 -7.15475857e-01 -7.46228874e-01
-2.96571761e-01 6.23941720e-01 2.10191533e-01 2.16499716e-01
-1.29031181e-01 7.53896713e-01 -1.81757569e+00 8.52072120e-01
-2.81707942e-02 1.29998124e+00 -3.64780486e-01 7.21958876e-01
8.60690027e-02 2.04222417e+00 -6.19469881e-01 -4.92821895e-02
2.29246661e-01 -5.72693110e-01 -1.03856575e+00 4.07236427e-01
6.79463923e-01 -5.97397983e-01 -7.76404440e-01 5.95526755e-01
3.91374603e-02 4.20865119e-01 9.93267834e-01 -1.70719552e+00
-5.13477623e-01 7.95683384e-01 -4.03528899e-01 3.35001349e-01
4.25274104e-01 1.54112741e-01 1.19083524e+00 -3.85052234e-01
5.40407836e-01 1.52295506e+00 4.43133116e-01 6.46097422e-01
-1.03701067e+00 -5.22961676e-01 5.56958139e-01 4.25012529e-01
-1.03484631e+00 -5.18507421e-01 4.31252182e-01 -4.06098068e-01
1.54659712e+00 8.76266435e-02 5.35506070e-01 1.47639179e+00
2.43515477e-01 1.33711231e+00 9.42834437e-01 -1.11292519e-01
1.05853595e-01 -5.85006416e-01 -5.42243049e-02 7.52350807e-01
-2.52112627e-01 -1.29766157e-02 -9.24157381e-01 1.80680603e-01
8.66118491e-01 5.77218056e-01 3.51370305e-01 -2.42794663e-01
-1.59009945e+00 2.53416091e-01 2.70993710e-01 1.53395325e-01
-5.37558794e-01 7.08370626e-01 6.70717418e-01 8.02198946e-02
3.18593293e-01 1.06742062e-01 -4.20685083e-01 -4.05250072e-01
-6.97510898e-01 4.82781827e-02 6.03575826e-01 8.92310858e-01
4.43881601e-01 8.99030492e-02 -5.16907156e-01 4.89421874e-01
1.58997759e-01 5.17681599e-01 2.74757713e-01 -1.49836802e+00
3.46161127e-01 8.51268589e-01 -1.53212354e-01 -9.51327860e-01
-3.63061875e-01 8.61483365e-02 -1.01799464e+00 1.78204730e-01
2.17273101e-01 3.54851454e-01 -1.08828187e+00 1.69110024e+00
9.28341448e-02 5.70150495e-01 3.88585567e-01 7.68623292e-01
6.29047096e-01 8.06065381e-01 4.59083527e-01 -8.71347338e-02
1.20478451e+00 -1.11918390e+00 -7.86343098e-01 -2.90370494e-01
5.96701741e-01 -1.82432517e-01 5.35135388e-01 4.25740093e-01
-9.16941524e-01 -6.11066937e-01 -7.32388437e-01 -3.65942121e-02
-3.54698032e-01 1.76680565e-01 8.65209401e-01 -2.11050771e-02
-8.56142819e-01 4.77711856e-01 -1.10471559e+00 -4.31207210e-01
7.81731367e-01 3.33924860e-01 -5.43361545e-01 -1.09023554e-02
-1.12377918e+00 7.62833714e-01 5.62862039e-01 -3.10153309e-02
-1.62921119e+00 -1.98091656e-01 -8.39440584e-01 -1.82542935e-01
7.81945467e-01 -2.72829711e-01 1.21135259e+00 -1.09333849e+00
-1.34263122e+00 9.79365110e-01 -6.08890364e-03 -8.67384076e-01
9.16854069e-02 -5.28978407e-01 -5.30512929e-01 5.63474596e-01
1.31174877e-01 9.93231833e-01 8.77089620e-01 -8.08994591e-01
-8.78132582e-01 -2.04650775e-01 4.54341352e-01 3.23079109e-01
6.80584982e-02 1.57177836e-01 -4.94440556e-01 -5.71411014e-01
5.62887639e-02 -1.10619962e+00 -1.37451766e-02 8.59622359e-02
-2.85589367e-01 -6.85589552e-01 9.57479894e-01 -1.77520081e-01
7.37834513e-01 -2.16771197e+00 4.38590437e-01 -2.95600325e-01
1.10143796e-01 2.80814152e-02 -1.26146421e-01 3.91657233e-01
-2.54425853e-01 -3.30274552e-01 -1.09330811e-01 -2.30286494e-01
1.20771892e-01 8.40027452e-01 -5.77328384e-01 4.65555668e-01
4.94357497e-01 9.14489329e-01 -1.09680259e+00 -5.63281655e-01
3.24393123e-01 3.05905521e-01 -2.57946014e-01 1.96381792e-01
-5.71734309e-01 3.39935958e-01 -3.85920048e-01 8.87448251e-01
-3.46197486e-02 -4.98127848e-01 3.28300536e-01 -2.03742713e-01
2.05814332e-01 2.19167501e-01 -1.03186238e+00 1.87659085e+00
-4.34668213e-02 6.38388634e-01 -5.28117239e-01 -1.33174324e+00
6.96554482e-01 5.64223647e-01 9.41834271e-01 -7.92213261e-01
3.50781791e-02 -2.34672979e-01 -2.90636748e-01 -5.55347323e-01
1.10248603e-01 -1.21338353e-01 -3.83611917e-01 7.80412316e-01
3.50509346e-01 2.35386401e-01 1.73244998e-01 4.49545741e-01
1.53977847e+00 8.00668716e-01 2.98694909e-01 3.11681002e-01
5.47576606e-01 -7.95502737e-02 5.67509472e-01 7.40580618e-01
-3.96667153e-01 2.36952424e-01 7.36844063e-01 -8.09308112e-01
-6.28648758e-01 -1.14753032e+00 4.04585034e-01 1.47007418e+00
2.11070672e-01 -2.85371870e-01 -3.63091916e-01 -9.46448982e-01
-8.77372026e-02 5.51692069e-01 -6.85582578e-01 -4.87348855e-01
-6.41062319e-01 -1.76476970e-01 9.00117457e-01 9.81882632e-01
6.70039892e-01 -1.48917377e+00 -9.48010504e-01 1.43198162e-01
-2.26682305e-01 -1.49629498e+00 2.05805544e-02 4.91483927e-01
-7.06812024e-01 -1.12339830e+00 -2.51216024e-01 -5.05499959e-01
4.70492035e-01 2.03264341e-01 1.17887497e+00 -2.71940947e-01
-3.32253277e-01 9.61124837e-01 -3.93836230e-01 -1.81857601e-01
-4.07016605e-01 -3.80574316e-01 3.85356992e-01 3.82633954e-01
6.27752006e-01 -4.96435136e-01 -2.88555264e-01 4.38277632e-01
-1.07449698e+00 2.59705216e-01 7.69748151e-01 4.61877912e-01
7.60035515e-01 4.21491772e-01 1.46652192e-01 -6.72455728e-01
-1.49456799e-01 -4.94104058e-01 -2.49449492e-01 4.73814815e-01
-1.06473945e-01 3.05773675e-01 3.68874937e-01 -6.39062047e-01
-1.08908093e+00 4.34176981e-01 3.74445647e-01 -1.01312065e+00
-6.78239703e-01 1.36505678e-01 2.47766636e-02 6.35644376e-01
5.11801183e-01 2.57184625e-01 -1.79853424e-01 -1.13408461e-01
5.53238094e-01 3.20972234e-01 7.31249392e-01 -6.83933616e-01
4.46674198e-01 8.70802104e-01 3.42239052e-01 -4.53014851e-01
-1.38191712e+00 -4.85610813e-01 -1.04644144e+00 -7.19684124e-01
1.49484897e+00 -1.21278071e+00 -9.19427216e-01 6.51637793e-01
-9.84272420e-01 -7.19151974e-01 -4.49388236e-01 3.75369161e-01
-1.00853074e+00 1.35420978e-01 -6.61152959e-01 -9.36683297e-01
1.95611328e-01 -9.20643568e-01 1.37368953e+00 -8.21625739e-02
-4.30388510e-01 -7.68895507e-01 -1.60612524e-01 5.79611957e-01
-8.07765946e-02 4.63167697e-01 7.22419024e-01 -6.93463624e-01
-8.51988971e-01 1.92094911e-02 -2.09841490e-01 3.26888889e-01
3.69760960e-01 -1.82872601e-02 -8.24357867e-01 9.96259898e-02
-2.22028285e-01 -8.74295056e-01 9.67135727e-01 2.23962173e-01
1.28921449e+00 -2.23997846e-01 -3.02106589e-01 1.54152647e-01
1.04205024e+00 5.55310249e-01 5.54802775e-01 -4.56846580e-02
8.42847526e-01 7.13927686e-01 7.76454210e-01 4.45678741e-01
3.23418558e-01 5.41500092e-01 6.51798785e-01 2.28653908e-01
-2.88316280e-01 -2.19072983e-01 7.45769501e-01 4.84634310e-01
-3.06795180e-01 -3.41228008e-01 -1.13617384e+00 4.23791289e-01
-2.10669780e+00 -1.34149909e+00 2.07067952e-01 1.58097613e+00
7.36288309e-01 3.88250083e-01 1.80314511e-01 1.70526952e-01
5.36265612e-01 4.67116207e-01 -8.92691374e-01 -1.07028812e-01
-2.65996724e-01 -8.71771276e-02 3.89945447e-01 1.03349440e-01
-1.57839477e+00 1.02821541e+00 6.26331472e+00 7.59687483e-01
-9.27246630e-01 -6.96831346e-02 6.93483233e-01 -1.76339194e-01
3.09101880e-01 -6.11863695e-02 -8.07877481e-01 -3.97958383e-02
1.19894838e+00 4.30349857e-01 4.51105982e-01 5.75136781e-01
-1.18432976e-01 -2.67584115e-01 -1.64540565e+00 1.13701844e+00
4.36861813e-01 -1.23499060e+00 3.86295229e-01 -1.96124226e-01
7.34822810e-01 -4.56320606e-02 -2.13293284e-02 4.87651110e-01
6.32795513e-01 -1.19592237e+00 8.61426890e-01 1.00862217e+00
6.86578453e-01 -4.79212761e-01 2.82892078e-01 3.70285034e-01
-1.41765296e+00 -3.11971366e-01 -2.83237187e-05 -3.62336457e-01
1.00307114e-01 -2.68860199e-02 -4.34736758e-01 3.82409841e-01
8.35771143e-01 1.59824371e+00 -3.77298713e-01 2.95605898e-01
-2.88050890e-01 5.62934160e-01 -1.11942291e-01 8.43089744e-02
5.82533419e-01 -5.52695021e-02 2.83527642e-01 9.54047680e-01
-1.24291591e-01 4.22167867e-01 2.17242494e-01 5.20119786e-01
3.47915888e-02 -6.84364796e-01 -8.92224371e-01 -6.26021624e-01
-2.30123568e-02 9.54148054e-01 -9.07495618e-01 -6.32711530e-01
-5.01319408e-01 1.11940336e+00 1.99164852e-01 3.63825530e-01
-1.29595077e+00 3.14259261e-01 7.43114293e-01 -3.36369634e-01
3.68672043e-01 -3.54243785e-01 3.07285219e-01 -9.90491927e-01
-2.33378798e-01 -1.00727475e+00 7.09211111e-01 -1.06740761e+00
-1.24877572e+00 4.26467776e-01 2.95636863e-01 -1.53246021e+00
-2.77114868e-01 -7.38417208e-01 -2.12661680e-02 1.53237432e-01
-9.26291168e-01 -1.28310120e+00 -1.62046358e-01 7.57777512e-01
8.59308481e-01 -3.87581527e-01 8.85341942e-01 3.04558694e-01
-4.60610062e-01 6.68406254e-03 -3.27779740e-01 3.67320091e-01
4.68480766e-01 -1.16843414e+00 9.89248753e-02 6.84828460e-01
7.26181686e-01 2.32662708e-01 2.71881372e-01 -6.30359948e-01
-1.49056613e+00 -1.28517365e+00 6.55235052e-01 -7.22695112e-01
8.81569386e-01 -3.70164543e-01 -5.80853701e-01 1.16251516e+00
1.60496533e-01 1.91497028e-01 5.35586178e-01 -9.51381922e-02
-7.01035023e-01 -2.95697272e-01 -6.22391045e-01 6.41954422e-01
1.54112685e+00 -1.02586937e+00 -6.76632941e-01 5.21125019e-01
7.52333760e-01 -2.54325807e-01 -8.66129816e-01 4.66210842e-01
5.64454257e-01 -9.40464795e-01 1.21530533e+00 -1.02968657e+00
6.23458028e-01 -4.36524987e-01 -4.21397805e-01 -5.42823613e-01
-2.15073287e-01 -3.36315006e-01 -6.08616531e-01 9.79572237e-01
4.14276831e-02 -1.73790798e-01 6.69684231e-01 5.01151502e-01
-7.13364035e-02 -6.52185678e-01 -9.40027177e-01 -8.35831940e-01
-5.23868859e-01 -5.96335948e-01 2.30404064e-01 8.29930961e-01
-3.42228979e-01 5.05782485e-01 -7.15912282e-01 1.65529907e-01
5.83508611e-01 1.94039956e-01 5.33177018e-01 -1.12513411e+00
-2.64581531e-01 -2.82581449e-01 -7.50376821e-01 -1.37667954e+00
5.53753376e-01 -7.93306649e-01 9.75371748e-02 -1.63246536e+00
2.31423080e-01 -2.78174996e-01 -6.73799574e-01 1.06794417e+00
3.52547675e-01 2.95009881e-01 1.12558857e-01 3.49454373e-01
-1.51140845e+00 5.00752032e-01 1.13142502e+00 -5.44355690e-01
1.72313333e-01 -5.34089357e-02 -8.25158283e-02 7.36055315e-01
5.28807998e-01 -4.51146871e-01 -7.05329478e-01 -4.35691744e-01
3.80902797e-01 4.38822150e-01 5.81941247e-01 -1.37743163e+00
2.67867655e-01 -3.81020069e-01 3.21759522e-01 -9.73192453e-01
8.18069518e-01 -1.15336931e+00 3.60054225e-01 3.74417394e-01
-8.40107799e-01 -2.28657082e-01 1.45959973e-01 9.10312176e-01
-2.74211258e-01 2.38672540e-01 4.21797454e-01 -2.55550593e-01
-1.14541268e+00 4.54308748e-01 -7.34353364e-01 -9.51823071e-02
1.43383968e+00 -1.25475481e-01 -2.26322487e-01 -4.06533927e-01
-1.04343605e+00 1.01519674e-01 -8.42803344e-02 7.12010443e-01
6.71925128e-01 -1.51484430e+00 -3.74941051e-01 -6.93638995e-02
2.39471123e-01 -6.69653565e-02 3.40866417e-01 8.61919999e-01
-4.12735254e-01 5.31927466e-01 -5.17707765e-01 -8.62255335e-01
-1.29112816e+00 6.30453289e-01 4.12973225e-01 -1.83987930e-01
-7.85395682e-01 9.33900416e-01 2.40484923e-01 -2.09632680e-01
5.84398925e-01 -6.00246191e-01 -3.94045264e-01 2.45290071e-01
3.35272491e-01 2.75127500e-01 -5.32319546e-01 -9.60611820e-01
-5.30918002e-01 3.33167732e-01 1.82539850e-01 5.79963513e-02
1.40413857e+00 7.54589513e-02 -2.69875914e-01 1.10713792e+00
1.06802177e+00 -6.97168767e-01 -1.62966621e+00 -5.19428492e-01
1.13943107e-01 -9.22389850e-02 -2.55312622e-01 -7.66106069e-01
-9.34647560e-01 8.55405569e-01 4.49388236e-01 2.18420640e-01
1.27395380e+00 4.11160946e-01 5.70579410e-01 7.52586961e-01
4.15616274e-01 -9.41683054e-01 6.14453375e-01 7.46152103e-01
7.14983225e-01 -1.14713097e+00 1.49329612e-02 4.07846831e-02
-9.92749393e-01 1.08065760e+00 6.18526995e-01 -6.16239458e-02
4.25747514e-01 1.40474364e-01 2.93879956e-03 -5.80477417e-01
-1.19806063e+00 -2.76428163e-01 2.52889514e-01 3.59767526e-01
5.32265902e-02 1.11086080e-02 6.42728865e-01 -1.13800913e-01
5.92290163e-01 -1.61566675e-01 7.83245042e-02 1.02994382e+00
-1.45393178e-01 -8.14757526e-01 -5.37773073e-02 5.20762622e-01
-2.74277687e-01 1.45573005e-01 -4.02392596e-01 5.07973075e-01
4.00916815e-01 8.00763845e-01 4.79003787e-01 -3.01329464e-01
1.60329506e-01 2.93296248e-01 5.28897583e-01 -6.25209749e-01
-2.04190597e-01 -1.71722263e-01 2.09004670e-01 -9.28523183e-01
-1.34318459e+00 -9.50846553e-01 -1.57118976e+00 1.87328354e-01
2.98691988e-01 -3.38496208e-01 1.10484600e-01 1.17108035e+00
2.03665838e-01 8.41880858e-01 2.97777057e-01 -7.36308515e-01
-1.60525426e-01 -7.19411194e-01 -4.90793258e-01 6.80171072e-01
3.37100416e-01 -9.55600083e-01 -2.40418926e-01 6.72835588e-01] | [8.534077644348145, 0.7215256690979004] |
874f68f7-088d-4870-be99-581f493ee14b | deep-transfer-reinforcement-learning-for-text | 1810.06667 | null | http://arxiv.org/abs/1810.06667v2 | http://arxiv.org/pdf/1810.06667v2.pdf | Deep Transfer Reinforcement Learning for Text Summarization | Deep neural networks are data hungry models and thus face difficulties when
attempting to train on small text datasets. Transfer learning is a potential
solution but their effectiveness in the text domain is not as explored as in
areas such as image analysis. In this paper, we study the problem of transfer
learning for text summarization and discuss why existing state-of-the-art
models fail to generalize well on other (unseen) datasets. We propose a
reinforcement learning framework based on a self-critic policy gradient
approach which achieves good generalization and state-of-the-art results on a
variety of datasets. Through an extensive set of experiments, we also show the
ability of our proposed framework to fine-tune the text summarization model
using only a few training samples. To the best of our knowledge, this is the
first work that studies transfer learning in text summarization and provides a
generic solution that works well on unseen data. | ['Naren Ramakrishnan', 'Chandan K. Reddy', 'Yaser Keneshloo'] | 2018-10-15 | null | null | null | null | ['transfer-reinforcement-learning'] | ['methodology'] | [ 2.80511737e-01 1.93704665e-01 -2.30251908e-01 -4.03822243e-01
-9.68269408e-01 -2.44691238e-01 5.23880363e-01 2.73967505e-01
-5.37422597e-01 1.00635302e+00 3.95967185e-01 -1.97681174e-01
6.74367100e-02 -5.85570574e-01 -8.11441958e-01 -3.73697758e-01
1.41155422e-01 7.24875689e-01 2.52027065e-01 -5.34921646e-01
5.41882098e-01 5.79391234e-02 -1.16611969e+00 3.72192204e-01
9.63347375e-01 5.12188494e-01 1.55263349e-01 8.12673748e-01
-6.75020441e-02 7.29954064e-01 -9.27737176e-01 -2.11913288e-01
2.47316565e-02 -6.78060949e-01 -1.15351987e+00 5.66411130e-02
8.06955636e-01 -6.82307243e-01 -2.85494119e-01 7.91213632e-01
1.02092493e+00 4.19974059e-01 7.79264212e-01 -8.39962602e-01
-1.07792819e+00 5.42105675e-01 -4.93269682e-01 4.01386827e-01
4.04569805e-01 -1.33910000e-01 9.44351137e-01 -5.90804517e-01
6.08870864e-01 1.14509714e+00 6.55993283e-01 9.71106231e-01
-9.51082468e-01 -3.89735162e-01 -3.60388920e-04 1.80882081e-01
-5.87428033e-01 -2.82853782e-01 8.28711748e-01 -5.48001193e-02
1.32124054e+00 2.20190678e-02 2.85681129e-01 1.40521359e+00
3.05489361e-01 1.22486579e+00 1.00492525e+00 -5.30132353e-01
3.82072210e-01 1.34235099e-02 4.41097245e-02 5.37964344e-01
3.14551383e-01 -4.57253397e-01 -7.00313747e-01 4.46970165e-02
4.74488944e-01 -2.53719509e-01 -1.78346694e-01 -3.12137514e-01
-1.09494805e+00 1.10017407e+00 6.22622252e-01 5.11396587e-01
-2.86295623e-01 2.78240889e-01 7.96120942e-01 4.43674803e-01
8.06005001e-01 6.22008801e-01 -5.63905299e-01 -1.91469967e-01
-1.30476892e+00 2.46244162e-01 8.77208233e-01 7.97526181e-01
6.02844417e-01 3.11454087e-01 -2.29143798e-01 1.00486445e+00
-2.05530062e-01 1.28351867e-01 9.59606886e-01 -5.98214447e-01
6.37817264e-01 4.25145388e-01 4.37411852e-02 -6.15540445e-01
-4.52145249e-01 -3.11395288e-01 -8.27687919e-01 8.18408057e-02
2.35308632e-01 -4.60026860e-01 -8.26767027e-01 1.50708354e+00
-6.13973476e-02 1.38782606e-01 4.35348928e-01 6.49661481e-01
1.09391880e+00 7.91349173e-01 -1.14263231e-02 -1.08908460e-01
8.95036817e-01 -1.37858832e+00 -6.03865087e-01 -2.91276723e-01
6.32886827e-01 -6.04558885e-01 1.05590546e+00 3.51678878e-01
-1.21321487e+00 -3.64578158e-01 -9.98220623e-01 -3.28505725e-01
-3.80168289e-01 1.03695780e-01 5.73660791e-01 3.74710977e-01
-1.19956696e+00 8.90093803e-01 -8.32970381e-01 -1.16618383e+00
5.53058803e-01 4.72208202e-01 -2.32973367e-01 3.59394737e-02
-1.05322683e+00 1.23827982e+00 6.31070018e-01 -2.55591989e-01
-7.72526264e-01 -3.04594278e-01 -6.46908760e-01 1.91542149e-01
3.10326576e-01 -1.13734555e+00 1.68923819e+00 -1.18588054e+00
-1.92548859e+00 8.58471930e-01 7.60112004e-03 -8.64440739e-01
6.18989766e-01 -4.58295733e-01 -1.48358420e-01 3.31385672e-01
1.06882155e-01 9.57793534e-01 1.05753779e+00 -1.17061162e+00
-5.08139968e-01 -2.49688387e-01 -9.72104445e-02 3.76967967e-01
-7.30392635e-01 -1.78645745e-01 1.55679360e-01 -7.35354662e-01
-4.96899396e-01 -7.92899847e-01 -1.02933370e-01 -4.25164998e-01
-2.39245951e-01 -5.04614472e-01 9.64139044e-01 -6.33128285e-01
9.85231817e-01 -1.65573323e+00 3.10492724e-01 -5.27721703e-01
-1.25265315e-01 7.14414835e-01 -2.18219966e-01 9.61655796e-01
7.08932132e-02 7.14141726e-02 -2.64837772e-01 -3.76997352e-01
-9.12748128e-02 1.89556777e-01 -3.18061620e-01 2.90875196e-01
3.02253425e-01 1.16101766e+00 -1.05529273e+00 -4.19239879e-01
8.17720369e-02 1.26927927e-01 -3.71358186e-01 2.06330866e-01
-3.91501486e-01 4.47249502e-01 -7.50881076e-01 2.58598238e-01
3.43916833e-01 -1.42160952e-01 -1.48446754e-01 1.21661555e-02
8.70276019e-02 7.50457793e-02 -5.70722699e-01 2.15921617e+00
-3.83195460e-01 7.86013722e-01 -2.07272217e-01 -1.52911043e+00
9.28051293e-01 3.70389968e-01 3.28613609e-01 -5.84341764e-01
2.88079649e-01 2.79337108e-01 -1.13502011e-01 -7.66310632e-01
9.30432558e-01 -2.16280952e-01 9.98363178e-03 7.61105955e-01
4.61236596e-01 -2.79262900e-01 1.81754857e-01 1.22104377e-01
1.02645409e+00 1.84636265e-01 5.97444654e-01 -3.03755105e-01
5.01004755e-01 1.17834970e-01 4.40468313e-03 9.43473458e-01
-2.16803953e-01 5.32602847e-01 3.24240327e-01 -4.38063234e-01
-1.25477338e+00 -6.51115417e-01 -3.42287146e-03 1.45002353e+00
-1.42037600e-01 -2.03201011e-01 -1.19561350e+00 -7.22627759e-01
-1.82767361e-01 9.53792453e-01 -7.77715862e-01 -2.34669104e-01
-6.41039431e-01 -6.14904523e-01 5.83812535e-01 7.00573742e-01
7.91958213e-01 -1.53273976e+00 -6.82676315e-01 2.95833439e-01
-4.07952536e-03 -9.58929956e-01 -3.53159010e-01 5.98360375e-02
-1.31833768e+00 -7.27412879e-01 -9.49850023e-01 -1.17285407e+00
5.53953707e-01 3.13960701e-01 1.05042803e+00 -7.33249113e-02
-5.24035767e-02 6.19940579e-01 -4.60910171e-01 -7.35387504e-01
-6.80141270e-01 6.63690329e-01 -2.24257275e-01 -2.99140513e-01
2.81260490e-01 -5.52818537e-01 -6.80117130e-01 -1.19789010e-02
-1.08999300e+00 -8.72072801e-02 6.99490786e-01 1.14712143e+00
1.25179306e-01 -3.38279068e-01 1.35832894e+00 -9.77310419e-01
1.30259395e+00 -5.15754521e-01 -1.15547642e-01 2.45403334e-01
-5.91694653e-01 1.50896907e-01 8.28768194e-01 -3.95757109e-01
-1.14199495e+00 -1.42039344e-01 1.95230413e-02 -1.17995478e-01
-3.13665748e-01 5.13656080e-01 4.91183877e-01 1.32837012e-01
9.49916542e-01 1.57997996e-01 1.33605860e-03 -4.17806387e-01
4.43070740e-01 8.02325845e-01 4.87552613e-01 -5.07393241e-01
4.53335643e-01 5.41707456e-01 -1.10674605e-01 -1.01471043e+00
-1.06477427e+00 -5.91018498e-01 -5.97908616e-01 1.68666348e-01
9.91835415e-01 -6.60991669e-01 -3.44317198e-01 3.93270582e-01
-1.01194656e+00 -5.52976906e-01 -3.31601292e-01 3.88622165e-01
-1.05172658e+00 5.79220533e-01 -7.05253363e-01 -3.90394479e-01
-1.00792193e+00 -7.85697997e-01 1.26838815e+00 3.29588592e-01
-2.73370147e-01 -1.31775320e+00 4.02215272e-01 3.50637943e-01
6.12923384e-01 1.61291003e-01 6.36960030e-01 -1.01574218e+00
-1.67368025e-01 -1.55835226e-01 7.97071680e-02 4.84490931e-01
2.00519845e-01 -2.48842418e-01 -1.00381875e+00 -6.73738122e-01
6.22094721e-02 -8.83855522e-01 1.32562327e+00 5.89318097e-01
1.04610980e+00 -2.76836395e-01 -1.48893028e-01 2.40236670e-01
1.23191059e+00 -2.65567094e-01 6.90193772e-01 5.90348959e-01
5.22807360e-01 4.68525141e-01 6.35007143e-01 2.33276114e-01
2.20660418e-01 2.58159399e-01 3.85974765e-01 -2.14457363e-01
-1.49783909e-01 -2.57330090e-01 2.46207222e-01 6.88635409e-01
-5.28028980e-03 -6.25483274e-01 -6.61899924e-01 6.87276185e-01
-2.27371860e+00 -1.02454352e+00 1.80235788e-01 1.88801932e+00
8.70480955e-01 2.29381565e-02 2.86793321e-01 -1.66903898e-01
7.76459098e-01 1.93379745e-01 -5.96459985e-01 -8.45108688e-01
1.20642977e-02 3.73851210e-01 2.46991947e-01 3.45109195e-01
-1.16522706e+00 1.40548909e+00 6.52237177e+00 7.60310113e-01
-1.20003569e+00 3.70453969e-02 3.95791858e-01 3.69354375e-02
8.46729502e-02 -2.74352491e-01 -5.59603631e-01 8.89373869e-02
1.02071548e+00 -1.82118818e-01 3.32500160e-01 9.28870082e-01
1.00806288e-01 -1.32155508e-01 -1.30593121e+00 7.81299114e-01
6.07790887e-01 -1.28018022e+00 2.11571813e-01 -2.37890303e-01
1.21483254e+00 2.80946434e-01 5.97179309e-03 4.18925732e-01
1.52525499e-01 -1.13499188e+00 1.40951753e-01 2.33170256e-01
4.28892314e-01 -6.36195719e-01 6.84905469e-01 5.44722021e-01
-3.70148271e-01 -8.40782896e-02 -6.45544589e-01 -6.82559535e-02
8.08429495e-02 4.38779928e-02 -1.32904017e+00 6.23712540e-01
6.00434840e-01 1.01813471e+00 -7.18014121e-01 1.01528430e+00
-1.71136543e-01 8.03321838e-01 4.04806510e-02 -4.76162314e-01
5.89253128e-01 -5.33596054e-03 5.61466694e-01 1.45654631e+00
2.78332919e-01 -2.46345803e-01 1.26846090e-01 4.88043517e-01
-4.43586648e-01 3.81010532e-01 -8.25503469e-01 1.22887120e-01
-1.73124984e-01 1.21165240e+00 -6.09476864e-01 -4.55771446e-01
-4.11761045e-01 1.10313177e+00 7.15974391e-01 4.88553375e-01
-4.15517747e-01 -2.72115886e-01 -1.67072952e-01 -1.00790970e-01
6.73459589e-01 -1.09409198e-01 -1.56900525e-01 -1.33114016e+00
-1.39850989e-01 -8.98004532e-01 4.76553142e-01 -8.58490169e-01
-1.58482671e+00 4.67076957e-01 2.29426026e-01 -1.17264521e+00
-5.76193035e-01 -5.46099246e-01 -8.85335147e-01 4.41840887e-01
-1.67283964e+00 -1.26904368e+00 -2.34433860e-01 6.88062131e-01
9.98501599e-01 -4.73328203e-01 1.02364337e+00 -2.88803607e-01
-3.29254359e-01 4.72696394e-01 5.84466577e-01 -1.94348954e-02
1.16997111e+00 -1.42193043e+00 4.04824942e-01 4.62130636e-01
8.44291318e-03 2.44828373e-01 9.29487765e-01 -5.29343247e-01
-1.14237618e+00 -9.92301702e-01 5.28996229e-01 -2.01891229e-01
6.82122111e-01 -6.51360452e-02 -1.08198035e+00 9.40602064e-01
9.51968610e-01 -2.74221957e-01 6.40292048e-01 2.46519689e-04
5.53681478e-02 6.30323589e-02 -1.11811054e+00 5.67041695e-01
6.33025825e-01 -1.42677307e-01 -1.12335277e+00 6.86031938e-01
3.69403630e-01 -4.90004957e-01 -7.03361094e-01 1.95218831e-01
2.32505754e-01 -8.94892871e-01 6.78360343e-01 -1.03022218e+00
7.11835325e-01 1.78502396e-01 1.61766052e-01 -1.81688440e+00
-1.49293885e-01 -7.07229376e-01 -1.68490916e-01 1.27229929e+00
2.45515659e-01 -6.12818062e-01 8.20023000e-01 -4.39296737e-02
-3.32840562e-01 -6.44786716e-01 -7.98854470e-01 -7.20172346e-01
6.82468116e-01 1.73033640e-01 7.12035596e-02 9.54097211e-01
1.74134076e-01 9.20444965e-01 -5.00069022e-01 -3.82211417e-01
4.13063794e-01 1.10001095e-01 9.73027229e-01 -1.10203862e+00
-2.42104650e-01 -4.42749262e-01 -2.56996959e-01 -8.41574371e-01
6.29309297e-01 -9.02885020e-01 -1.02751315e-01 -1.97335124e+00
3.31763476e-01 2.38943964e-01 -2.23854482e-02 3.79739344e-01
-3.12680185e-01 -2.01603491e-02 1.88250896e-02 9.13027078e-02
-8.65576029e-01 8.00263822e-01 1.38689399e+00 -2.83597499e-01
-2.08871454e-01 -1.65784813e-03 -7.67177641e-01 5.55023253e-01
1.52250409e+00 -5.29448569e-01 -5.22316992e-01 -6.37983084e-01
-4.12772223e-02 1.05407843e-02 1.45480856e-01 -7.77264833e-01
2.20193639e-01 1.24254316e-01 3.96275342e-01 -5.44045091e-01
1.55568600e-01 -3.37160498e-01 -7.57680118e-01 4.24597204e-01
-5.81830323e-01 4.65837158e-02 4.34374690e-01 7.66292274e-01
-2.21032679e-01 -6.82169497e-01 7.89344132e-01 -3.52790833e-01
-5.67436695e-01 -1.35706188e-02 -4.00021017e-01 2.77550459e-01
9.34344411e-01 -2.31368408e-01 -4.94264007e-01 -7.91401267e-01
-4.70917821e-01 3.08827460e-01 4.44448322e-01 5.19270539e-01
4.52437609e-01 -9.47431684e-01 -1.12387133e+00 -3.08173656e-01
-2.97185574e-02 -1.80723108e-02 1.52098671e-01 3.96018803e-01
-5.90342224e-01 5.77700794e-01 -5.02238691e-01 -6.45306706e-01
-1.04145372e+00 7.11697221e-01 2.89882302e-01 -5.09324372e-01
-7.71494627e-01 3.78317714e-01 2.87828762e-02 -3.88776153e-01
1.30934790e-01 -1.68243781e-01 -2.64611751e-01 -8.31988901e-02
4.39752340e-01 5.14706969e-01 2.94954926e-01 -3.31996351e-01
-2.08264720e-02 5.71387172e-01 -4.86139894e-01 -8.85787159e-02
1.60358787e+00 4.33171093e-02 6.48897886e-02 4.29741234e-01
1.28424978e+00 -4.03328598e-01 -1.17382908e+00 -2.28281245e-01
-6.54665455e-02 -3.73164713e-02 -8.49005999e-04 -6.92444444e-01
-5.64137995e-01 1.07368851e+00 3.91807556e-01 3.55296195e-01
1.03941536e+00 -7.44841471e-02 9.15333748e-01 8.77306879e-01
-1.59281939e-01 -1.48101234e+00 7.47005939e-01 5.54789782e-01
1.10664439e+00 -1.49820077e+00 2.25258127e-01 1.10494554e-01
-8.45518887e-01 1.39225471e+00 6.49507582e-01 -6.59270644e-01
1.02753185e-01 -1.62291482e-01 -9.49770063e-02 -1.05299823e-01
-7.18645334e-01 -5.44908382e-02 1.47065178e-01 6.43840194e-01
6.08680725e-01 -2.29097843e-01 -5.01307011e-01 3.21697220e-02
-2.87403345e-01 2.42828041e-01 8.78244817e-01 1.17305791e+00
-7.06954598e-01 -1.14803672e+00 -2.44136229e-01 4.63367820e-01
-6.67665601e-01 2.06613559e-02 -6.33762717e-01 8.82708132e-01
-6.53722346e-01 9.41869438e-01 -1.37936711e-01 1.57071233e-01
3.14321429e-01 6.16616011e-02 7.96331406e-01 -8.07330072e-01
-7.57572949e-01 1.29806176e-02 1.22037619e-01 7.23977983e-02
-8.35713685e-01 -5.20132303e-01 -1.21328175e+00 -2.34651670e-01
-2.23138541e-01 6.88090548e-02 7.08989799e-01 7.76531339e-01
3.50642711e-01 5.64722955e-01 4.37488586e-01 -1.16079485e+00
-1.03205752e+00 -1.35610473e+00 -4.97665823e-01 4.79293913e-01
4.40474302e-01 -2.71903992e-01 -5.42936213e-02 1.52760863e-01] | [12.297266006469727, 9.383249282836914] |
120e23a2-261a-4699-ab2c-0a717b6a3e33 | weighted-model-estimation-for-offline-model | null | null | http://proceedings.neurips.cc/paper/2021/hash/949694a5059302e7283073b502f094d7-Abstract.html | http://proceedings.neurips.cc/paper/2021/file/949694a5059302e7283073b502f094d7-Paper.pdf | Weighted model estimation for offline model-based reinforcement learning | This paper discusses model estimation in offline model-based reinforcement learning (MBRL), which is important for subsequent policy improvement using an estimated model. From the viewpoint of covariate shift, a natural idea is model estimation weighted by the ratio of the state-action distributions of offline data and real future data. However, estimating such a natural weight is one of the main challenges for off-policy evaluation, which is not easy to use. As an artificial alternative, this paper considers weighting with the state-action distribution ratio of offline data and simulated future data, which can be estimated relatively easily by standard density ratio estimation techniques for supervised learning. Based on the artificial weight, this paper defines a loss function for offline MBRL and presents an algorithm to optimize it. Weighting with the artificial weight is justified as evaluating an upper bound of the policy evaluation error. Numerical experiments demonstrate the effectiveness of weighting with the artificial weight. | ['Kei Senda', 'Toru Hishinuma'] | 2021-12-01 | null | https://openreview.net/forum?id=zdC5eXljMPy | https://openreview.net/pdf?id=zdC5eXljMPy | neurips-2021-12 | ['density-ratio-estimation'] | ['methodology'] | [-1.11102872e-01 2.15899125e-01 -7.01671660e-01 -1.02924883e-01
-6.77784860e-01 -1.26075596e-01 3.99436831e-01 4.39593226e-01
-9.45647418e-01 1.22435403e+00 -1.33691907e-01 -5.81531882e-01
-4.86286491e-01 -6.13872647e-01 -8.05599928e-01 -7.88404167e-01
-1.75137743e-01 4.59416240e-01 -1.40410298e-02 -2.82018837e-02
2.28488132e-01 3.55871677e-01 -1.30366480e+00 -5.57963490e-01
1.16302311e+00 1.09473836e+00 3.75434905e-01 4.74318296e-01
-1.76209152e-01 6.80402160e-01 -8.33687603e-01 -1.97638839e-01
4.39032137e-01 -4.38918322e-01 -4.50686365e-01 -1.68158673e-02
-9.77777094e-02 -6.41645908e-01 -1.77801952e-01 1.10674191e+00
5.99045873e-01 4.79716361e-01 5.85765183e-01 -1.45003402e+00
3.72912973e-01 5.05497456e-01 -4.53136116e-01 1.79038271e-01
-6.40438795e-02 3.39695364e-01 4.57245111e-01 -2.30763450e-01
2.07520753e-01 1.12405860e+00 4.86184925e-01 4.93474692e-01
-1.22178161e+00 -6.27765715e-01 5.41295528e-01 2.95748919e-01
-1.18118119e+00 9.49752331e-02 6.89660251e-01 -4.50962871e-01
5.71286023e-01 -1.09011658e-01 9.12434518e-01 7.73081362e-01
3.02527964e-01 8.72760713e-01 1.34903610e+00 -6.35515928e-01
6.51854575e-01 5.30014396e-01 -1.62032038e-01 4.69517082e-01
2.51656175e-01 5.45175552e-01 -1.22266829e-01 -2.57874131e-01
8.34366024e-01 -1.46428972e-01 -3.89735132e-01 -6.01650774e-01
-6.42999291e-01 7.92438209e-01 1.27850482e-02 -1.13865852e-01
-3.97445589e-01 2.45479748e-01 3.79113823e-01 4.03918624e-01
6.22405708e-01 3.85401219e-01 -4.63378012e-01 -5.11505187e-01
-7.68489301e-01 3.49457026e-01 6.68455124e-01 8.27553213e-01
6.09458327e-01 2.18937039e-01 -2.25459754e-01 6.61520302e-01
2.73091733e-01 6.51490152e-01 5.17328560e-01 -8.31342280e-01
6.00816846e-01 3.02385062e-01 7.81708956e-01 -4.93552208e-01
-1.32309139e-01 -4.09929127e-01 -5.01721740e-01 6.26780689e-01
1.03948104e+00 -6.82464719e-01 -5.46710253e-01 1.92618179e+00
6.45838797e-01 1.44103378e-01 8.47621448e-03 4.79779035e-01
-3.86952996e-01 5.74148238e-01 2.86352038e-01 -8.21438789e-01
8.80374968e-01 -6.81688011e-01 -1.00383902e+00 1.01692222e-01
5.13207674e-01 -3.04285020e-01 9.67602313e-01 4.39337343e-01
-1.05639470e+00 -4.97912049e-01 -9.05625403e-01 7.11507618e-01
-2.11493552e-01 7.07363635e-02 2.81650782e-01 5.03428340e-01
-7.16699183e-01 1.01695311e+00 -9.46648717e-01 -2.32241392e-01
2.32370347e-02 3.40099037e-01 3.38545859e-01 5.17043710e-01
-1.34394729e+00 1.22814357e+00 7.42151141e-01 -8.05024207e-02
-8.21706772e-01 -9.12406981e-01 -6.55920863e-01 4.28664386e-02
6.40647352e-01 -3.33751053e-01 1.52229393e+00 -8.73808801e-01
-1.88907731e+00 -6.34327531e-02 1.41039705e-02 -9.78365362e-01
9.89917397e-01 -1.16912223e-01 -3.67675275e-01 3.86188626e-02
-4.40636009e-01 2.45908335e-01 1.33337545e+00 -1.11682427e+00
-9.43608522e-01 -2.28246674e-01 1.09718315e-01 3.09260905e-01
-4.99592960e-01 -4.12058026e-01 2.26348415e-01 -6.02651298e-01
-6.60958409e-01 -6.00774467e-01 -5.70930302e-01 2.16077007e-02
1.27866089e-01 -2.76951045e-01 7.40768611e-01 -8.42470288e-01
1.64215493e+00 -1.92622709e+00 -2.81393021e-01 4.69316423e-01
-1.74022779e-01 4.19883043e-01 1.21345580e-01 5.75650215e-01
-2.69886982e-02 -2.33726293e-01 -2.77113169e-01 -5.08143678e-02
1.46315768e-01 4.59811874e-02 -6.89090669e-01 4.05070037e-01
1.22490518e-01 4.37006056e-01 -1.13657677e+00 -3.29172879e-01
4.25664455e-01 9.20569226e-02 -3.00092727e-01 4.09690857e-01
-4.76754665e-01 4.84259963e-01 -5.81739426e-01 -5.23834787e-02
6.15261316e-01 1.64563090e-01 3.66358042e-01 -4.21769805e-02
-1.95361063e-01 -2.74407655e-01 -1.33785319e+00 1.01112163e+00
-8.01641881e-01 1.41776368e-01 -4.01755683e-02 -1.34948671e+00
1.13139999e+00 2.64481306e-01 6.90506876e-01 -4.75476891e-01
1.22598059e-01 2.07613453e-01 -3.03650219e-02 -3.60578179e-01
2.79361516e-01 -3.80583078e-01 1.22019328e-01 3.87833476e-01
-6.99389130e-02 -3.28899384e-01 3.29802006e-01 -1.55063331e-01
6.72880888e-01 4.65046316e-01 5.98360240e-01 -3.34771663e-01
5.49436271e-01 -1.02919854e-01 5.91857255e-01 7.00368285e-01
-4.96665716e-01 -1.97376624e-01 7.24090099e-01 -2.48969525e-01
-9.04690087e-01 -8.06816161e-01 -2.06723407e-01 4.87514615e-01
5.47662675e-02 -1.78142443e-01 -8.00635219e-01 -9.87561345e-01
3.71039748e-01 1.01632774e+00 -4.87246662e-01 -3.79320920e-01
-4.61028218e-01 -6.13917947e-01 2.78705992e-02 5.27489543e-01
4.19456869e-01 -6.99735045e-01 -7.52241850e-01 5.59202433e-01
1.66787341e-01 -7.10811496e-01 -3.90437841e-01 2.51883894e-01
-9.98359740e-01 -9.81271088e-01 -9.19500053e-01 -1.31485030e-01
6.48353398e-01 -2.38263190e-01 8.01073372e-01 -2.08886430e-01
1.62128508e-01 8.16902637e-01 -4.30435687e-02 -7.51437426e-01
-5.48858583e-01 -2.17979014e-01 3.28891993e-01 1.64689243e-01
5.65192029e-02 -3.33230585e-01 -5.03114760e-01 3.51112783e-01
-7.06364036e-01 -2.65082896e-01 1.70594335e-01 1.12474239e+00
6.58495963e-01 2.97937840e-01 8.01779628e-01 -6.01636887e-01
7.86478877e-01 -2.99453318e-01 -1.40754151e+00 6.17138565e-01
-1.07562125e+00 4.12193626e-01 9.51387405e-01 -7.63217866e-01
-1.26987326e+00 -1.39605612e-01 6.01629354e-02 -5.53165495e-01
1.98835596e-01 3.37954223e-01 -5.36347367e-02 1.42254278e-01
1.62934214e-01 1.24160476e-01 4.89740223e-01 -3.99697095e-01
1.39817432e-01 7.30355740e-01 7.56272823e-02 -8.57040584e-01
5.66577375e-01 1.23386942e-01 2.83044130e-01 -6.46421850e-01
-8.59077215e-01 -3.42212141e-01 -2.69748956e-01 -4.01468545e-01
4.19047594e-01 -5.51304936e-01 -9.85812843e-01 4.12586242e-01
-5.91880322e-01 -8.50127995e-01 -8.02917957e-01 8.38207006e-01
-1.16232312e+00 2.36768365e-01 -1.26606181e-01 -1.52873683e+00
-1.99340627e-01 -1.08037210e+00 5.12948096e-01 3.18971455e-01
2.76920293e-02 -1.37259793e+00 2.36678302e-01 -2.03311145e-01
3.20193350e-01 3.10651839e-01 9.42943215e-01 -3.92362714e-01
-4.62198555e-02 -2.46654138e-01 2.39414737e-01 6.71021700e-01
2.66026147e-02 -7.11442083e-02 -6.31956398e-01 -7.08667696e-01
3.38989645e-01 -2.89091229e-01 5.51917970e-01 5.99574506e-01
1.52173090e+00 -4.80421662e-01 -1.93288967e-01 1.89230889e-01
1.58135903e+00 5.55898964e-01 3.72788817e-01 3.42964560e-01
1.69832274e-01 4.91161883e-01 1.38593519e+00 1.06006193e+00
-7.98217580e-03 6.49838626e-01 3.75253499e-01 2.46411383e-01
3.52222323e-01 -5.34287989e-01 4.96962905e-01 4.96619374e-01
9.62111950e-02 3.44572105e-02 -6.68439567e-01 2.63675451e-01
-1.91726184e+00 -9.16661084e-01 5.05266547e-01 2.87220407e+00
9.66226935e-01 2.79819697e-01 4.65555042e-01 6.81186616e-02
9.39227402e-01 -8.06315839e-02 -1.00741196e+00 -4.40205932e-01
5.15049219e-01 2.24444672e-01 6.85465932e-01 6.09965980e-01
-8.56895208e-01 4.42914397e-01 5.91402340e+00 1.43861568e+00
-1.11642778e+00 -1.83545038e-01 6.79619312e-01 1.58180445e-01
-1.38539877e-02 1.57483369e-01 -1.08199608e+00 8.49999905e-01
1.31459367e+00 -6.27845466e-01 5.74753702e-01 1.03150892e+00
5.83376706e-01 -4.64663416e-01 -9.22083497e-01 6.49816096e-01
-5.65613568e-01 -7.39343345e-01 -1.21338658e-01 1.61301255e-01
6.17769659e-01 -5.14506876e-01 1.15113296e-02 8.17284822e-01
2.29131237e-01 -5.41914225e-01 6.32764280e-01 8.09826374e-01
7.54531026e-01 -1.19920409e+00 7.13252366e-01 6.83425784e-01
-1.07575238e+00 -5.49962223e-01 -4.15322155e-01 -1.62655264e-02
1.88183174e-01 5.03726661e-01 -8.18406820e-01 4.49089974e-01
9.00825337e-02 4.68051791e-01 -9.78949293e-02 1.30302787e+00
-1.91050306e-01 7.12396920e-01 -3.50037396e-01 -2.52009004e-01
4.37375754e-02 -4.94035393e-01 5.29897153e-01 7.68894434e-01
4.10516202e-01 -4.07957554e-01 4.72379446e-01 4.43899095e-01
2.76134461e-01 2.31694654e-02 -4.49543536e-01 -9.88626704e-02
5.61460614e-01 9.57783639e-01 -3.59734952e-01 -4.32071149e-01
-8.44939947e-02 3.02524269e-01 4.21016276e-01 4.25116330e-01
-1.03259277e+00 -4.18123364e-01 4.62945968e-01 4.26006280e-02
2.07535237e-01 -1.57606885e-01 1.75545007e-01 -7.52016485e-01
-7.60779306e-02 -8.78481030e-01 4.15747881e-01 -1.06533490e-01
-1.12229288e+00 1.04648262e-01 6.27932131e-01 -1.75081038e+00
-6.90246582e-01 -5.12221277e-01 -4.58104074e-01 8.48785341e-01
-1.47320545e+00 -3.44689786e-01 1.52911484e-01 3.49300176e-01
4.62601036e-01 -1.07549109e-01 5.36240876e-01 -3.93008552e-02
-7.07925081e-01 6.39944077e-01 6.03412449e-01 -4.03288066e-01
4.47716445e-01 -1.45007110e+00 -3.18504907e-02 4.86033171e-01
-4.15380210e-01 3.01641196e-01 8.80019486e-01 -6.16148293e-01
-1.05171871e+00 -1.01378286e+00 2.30716750e-01 2.07739342e-02
9.06491041e-01 -4.57115378e-03 -8.00702274e-01 2.74194926e-01
-2.88672205e-02 -6.01553544e-02 2.13619858e-01 -3.70133787e-01
2.60195941e-01 -4.44268525e-01 -1.41289616e+00 6.60700679e-01
5.65648496e-01 -1.94019064e-01 -1.70630693e-01 1.48726732e-01
6.16715491e-01 -3.50068241e-01 -1.27811670e+00 6.50885582e-01
5.08658886e-01 -4.04164910e-01 6.52389228e-01 -8.63045573e-01
-7.16241151e-02 -1.12191759e-01 1.47000253e-01 -1.87184918e+00
1.96609408e-01 -8.96486700e-01 -7.43276060e-01 1.18391788e+00
1.62347183e-01 -8.48869264e-01 6.18401289e-01 6.67194903e-01
3.45975846e-01 -1.05681968e+00 -1.05891585e+00 -1.24381459e+00
1.65687472e-01 -2.45257452e-01 6.29790068e-01 5.64569712e-01
1.34755641e-01 -1.33381963e-01 -3.41143131e-01 -1.58394009e-01
8.11306775e-01 -3.99638116e-02 7.91466117e-01 -8.45737696e-01
-6.03253603e-01 -3.66901606e-01 -1.27372116e-01 -1.07714641e+00
2.48356313e-01 -2.29778558e-01 2.66793370e-02 -1.32875383e+00
-2.08394900e-01 -4.61132079e-01 -5.84784269e-01 1.17095776e-01
-2.79819936e-01 -7.12028742e-01 1.69218525e-01 -2.60638922e-01
-2.61163920e-01 1.16859794e+00 1.44726360e+00 -3.80992107e-02
-3.12734306e-01 6.46707594e-01 -6.94756657e-02 8.18537593e-01
1.05158782e+00 -4.70825523e-01 -7.50225186e-01 2.61175424e-01
-1.74150810e-01 5.75467646e-01 3.00049912e-02 -9.40685153e-01
-1.06933251e-01 -6.71545744e-01 1.35385633e-01 -6.44727349e-01
2.90387750e-01 -1.07842267e+00 -3.01343370e-02 7.67009199e-01
-5.10527909e-01 -1.90397352e-02 2.45659843e-01 1.00012732e+00
-7.21106604e-02 -6.47947609e-01 8.64049613e-01 -6.25644326e-02
-5.18861234e-01 1.97467640e-01 -4.25155640e-01 2.15827286e-01
1.06751490e+00 -6.12737834e-02 -2.28808634e-02 -6.52449667e-01
-6.75379395e-01 4.71499681e-01 9.69404727e-02 6.85120700e-04
4.25896227e-01 -1.15421844e+00 -2.33268648e-01 5.26704127e-03
-1.40281796e-01 -4.02019888e-01 9.33551267e-02 7.80398786e-01
-3.71873528e-02 2.43824095e-01 -6.96666166e-02 -2.11182758e-01
-7.85976291e-01 6.91018820e-01 5.15471339e-01 -7.81259418e-01
-2.43382841e-01 1.74636573e-01 -2.66448945e-01 -2.08371356e-01
5.89784086e-01 -3.47844481e-01 -2.49917299e-01 1.27931431e-01
6.96463645e-01 6.15838110e-01 -1.04257643e-01 5.21670245e-02
2.58580267e-01 1.96522936e-01 -2.00913921e-02 -3.47535223e-01
9.98076320e-01 -1.71057865e-01 3.40722293e-01 5.76773942e-01
8.17400753e-01 -4.29648936e-01 -1.95436704e+00 -1.77963302e-01
1.87621981e-01 -5.31988978e-01 1.26327053e-01 -5.72853506e-01
-7.69043446e-01 7.13585556e-01 9.01230216e-01 2.28149891e-01
1.06426489e+00 -7.40720749e-01 5.55405796e-01 3.19775671e-01
7.94637084e-01 -1.67985702e+00 -6.12065122e-02 4.42281872e-01
7.87063837e-01 -1.11017454e+00 6.13714121e-02 1.79876447e-01
-5.76962411e-01 1.01829863e+00 7.68819809e-01 -1.17164277e-01
9.20984149e-01 2.82721013e-01 -1.55433640e-01 5.90815961e-01
-7.91726410e-01 -2.68537402e-01 1.06917337e-01 5.82183957e-01
-6.39703870e-02 1.37635872e-01 -8.31331015e-01 2.88265049e-01
1.87140286e-01 2.75187314e-01 3.55845600e-01 1.03721488e+00
-4.20658141e-01 -1.26502645e+00 -3.22595477e-01 5.08120418e-01
-2.89442599e-01 2.40046322e-01 3.96728814e-01 9.07324970e-01
-3.16910416e-01 7.58797050e-01 2.52260137e-02 -1.91809107e-02
3.65393668e-01 1.17121644e-01 6.03052676e-01 -1.21199608e-01
-3.44776809e-01 -1.72649156e-02 -1.22935390e-02 -3.47949654e-01
-1.50028974e-01 -3.58597428e-01 -1.10122502e+00 3.42919528e-02
-5.51179051e-01 5.09148300e-01 8.85877609e-01 1.04633176e+00
2.82061752e-02 3.47378165e-01 1.11377239e+00 -5.82226157e-01
-1.75485444e+00 -9.84137833e-01 -8.27149808e-01 2.04646841e-01
2.32902899e-01 -1.04814839e+00 -5.99797010e-01 -4.01513904e-01] | [4.246837615966797, 2.471400737762451] |
09f8b157-c1c2-4853-9ed8-183363f7bda5 | next-step-conditioned-deep-convolutional | 1702.03865 | null | http://arxiv.org/abs/1702.03865v1 | http://arxiv.org/pdf/1702.03865v1.pdf | Next-Step Conditioned Deep Convolutional Neural Networks Improve Protein Secondary Structure Prediction | Recently developed deep learning techniques have significantly improved the
accuracy of various speech and image recognition systems. In this paper we show
how to adapt some of these techniques to create a novel chained convolutional
architecture with next-step conditioning for improving performance on protein
sequence prediction problems. We explore its value by demonstrating its ability
to improve performance on eight-class secondary structure prediction. We first
establish a state-of-the-art baseline by adapting recent advances in
convolutional neural networks which were developed for vision tasks. This model
achieves 70.0% per amino acid accuracy on the CB513 benchmark dataset without
use of standard performance-boosting techniques such as ensembling or multitask
learning. We then improve upon this state-of-the-art result using a novel
chained prediction approach which frames the secondary structure prediction as
a next-step prediction problem. This sequential model achieves 70.3% Q8
accuracy on CB513 with a single model; an ensemble of these models produces
71.4% Q8 accuracy on the same test set, improving upon the previous overall
state of the art for the eight-class secondary structure problem. Our models
are implemented using TensorFlow, an open-source machine learning software
library available at TensorFlow.org; we aim to release the code for these
experiments as part of the TensorFlow repository. | ['Navdeep Jaitly', 'Akosua Busia'] | 2017-02-13 | null | null | null | null | ['protein-secondary-structure-prediction'] | ['medical'] | [ 4.99853492e-01 -3.74091044e-02 -7.82189593e-02 -5.18128157e-01
-1.05511355e+00 -5.27875602e-01 4.43032414e-01 -5.46502694e-02
-7.97352374e-01 7.78918803e-01 1.67076383e-02 -8.42535377e-01
3.13523173e-01 -1.64570242e-01 -9.41376448e-01 -8.41533244e-01
-1.34002626e-01 3.94343734e-01 2.00240374e-01 -3.36951911e-01
9.76623893e-02 4.59981471e-01 -1.19284010e+00 1.13664305e+00
5.27214050e-01 1.07088828e+00 3.02541792e-01 1.23805964e+00
4.66416515e-02 1.01335382e+00 -5.71108043e-01 -4.37449634e-01
1.09663755e-01 -2.10151508e-01 -1.29074752e+00 -3.34205300e-01
6.59408152e-01 -3.96777332e-01 -1.52345896e-01 5.72352648e-01
7.15292454e-01 -1.70071080e-01 3.16877276e-01 -7.93739974e-01
-7.52202988e-01 4.55806762e-01 -3.33847433e-01 5.34135163e-01
9.48050767e-02 4.41914797e-01 1.26373148e+00 -1.01244318e+00
5.58080912e-01 1.05822670e+00 1.09035861e+00 5.82444727e-01
-1.59335744e+00 -6.45226181e-01 -1.24320380e-01 5.91698527e-01
-8.82200301e-01 -5.21314561e-01 2.82460839e-01 -4.13167536e-01
1.94873965e+00 -3.71967740e-02 5.43272436e-01 1.27371395e+00
5.65065086e-01 1.03878212e+00 1.12503040e+00 -3.34909320e-01
6.48448020e-02 -4.46289122e-01 2.95986623e-01 9.35536325e-01
-3.26351970e-01 2.34684840e-01 -6.77140236e-01 -3.74401718e-01
3.15436184e-01 -3.05207074e-01 -1.59556076e-01 -2.75222659e-01
-1.46461594e+00 9.19422209e-01 4.74081308e-01 2.21877337e-01
-4.47594881e-01 1.77064762e-01 5.55073440e-01 5.24853587e-01
5.27592719e-01 5.14556229e-01 -1.22835612e+00 -3.40695530e-01
-9.03459251e-01 3.66805911e-01 8.93642545e-01 4.82739627e-01
4.40037370e-01 -9.95772555e-02 -1.76867798e-01 8.28702629e-01
2.79480577e-01 2.11196005e-01 6.17733359e-01 -8.57199311e-01
1.74619839e-01 2.62804508e-01 1.32149942e-02 -2.29565665e-01
-7.46275485e-01 -6.69071972e-01 -5.75309455e-01 2.74557889e-01
4.38341588e-01 -1.52134851e-01 -1.21763802e+00 1.43590426e+00
8.36736858e-02 3.10029015e-02 2.86369056e-01 7.35545218e-01
8.73462677e-01 7.56048024e-01 3.08694392e-01 7.81761948e-04
1.35240066e+00 -1.25221455e+00 -3.94789129e-01 1.90676734e-01
9.77345288e-01 -1.04206383e+00 7.71286726e-01 7.24559784e-01
-7.83639371e-01 -6.84523404e-01 -1.26436257e+00 -1.28599495e-01
-2.32245132e-01 2.02773213e-01 7.19253778e-01 6.43614233e-01
-1.23519456e+00 8.66622090e-01 -1.01479948e+00 -2.36417472e-01
4.57539558e-01 5.23097575e-01 -4.46950138e-01 2.85318881e-01
-1.26664948e+00 1.09702337e+00 4.40688133e-01 -2.20833018e-01
-1.03315008e+00 -9.78449106e-01 -3.92947108e-01 -4.28328253e-02
-2.89619640e-02 -5.86968839e-01 1.91609132e+00 -1.23487687e+00
-1.66978121e+00 9.26873922e-01 -1.79729953e-01 -1.01717508e+00
2.86390215e-01 -3.14427942e-01 -3.12821180e-01 2.63132244e-01
-1.21920362e-01 9.15113151e-01 5.59974968e-01 -6.61402643e-01
-7.33648360e-01 -1.44595608e-01 -1.37572676e-01 -5.67915924e-02
3.73002030e-02 8.82446691e-02 3.13958116e-02 -4.61247057e-01
-4.06982988e-01 -1.04288590e+00 -1.67870477e-01 -9.77774635e-02
-1.09532766e-01 -3.74872535e-01 6.53893232e-01 -7.35695541e-01
6.34579420e-01 -1.74395728e+00 3.55971046e-02 -2.71540940e-01
3.09717715e-01 8.85150433e-01 -5.16477406e-01 5.16933143e-01
-6.19007409e-01 -8.22809935e-02 -2.15130150e-01 -1.80164143e-01
-3.10014993e-01 -3.49940099e-02 -9.01674405e-02 3.94863516e-01
5.34975350e-01 9.97229755e-01 -6.88246667e-01 2.51859780e-02
1.80688873e-01 7.43697047e-01 -6.07122481e-01 1.33904725e-01
-2.96672463e-01 3.82889450e-01 -6.31578267e-02 3.77594203e-01
5.42379022e-01 -7.63498724e-01 3.19404542e-01 -3.53590190e-01
-3.88414562e-02 4.45426852e-01 -4.07548845e-01 1.98379910e+00
-1.90464169e-01 6.68656170e-01 -1.97624892e-01 -1.20015800e+00
8.17104220e-01 3.39596510e-01 7.58409917e-01 -6.28863931e-01
-1.17300004e-01 2.02528000e-01 6.59748673e-01 -3.91820073e-01
1.15489520e-01 -1.97837085e-01 5.74949443e-01 2.52424508e-01
5.68961561e-01 2.50848174e-01 1.00216987e-02 2.48353109e-01
1.55684149e+00 4.95846570e-01 2.73998052e-01 -5.35863042e-01
5.26682496e-01 1.95197031e-01 2.60034710e-01 6.70469642e-01
-5.63700259e-01 4.55184579e-01 3.71274173e-01 -1.00789928e+00
-1.50679219e+00 -8.92660022e-01 -2.82049149e-01 1.51677895e+00
-6.32090747e-01 -4.64536041e-01 -6.91875815e-01 -9.16271746e-01
2.34753755e-03 3.80347908e-01 -5.93560755e-01 -3.52446064e-02
-5.10466993e-01 -1.36368787e+00 7.41540313e-01 5.90377688e-01
2.80534625e-01 -1.18296146e+00 -5.18259168e-01 7.07487762e-01
-1.68052822e-01 -1.01785028e+00 -4.49320003e-02 9.87711728e-01
-6.27711058e-01 -1.20940471e+00 -8.02737832e-01 -8.04968715e-01
-1.79033235e-01 -4.50374447e-02 1.40696251e+00 1.12936169e-01
-2.78905779e-01 -9.05916765e-02 -5.75125873e-01 -3.90568376e-01
-6.88582957e-01 4.53984708e-01 -9.59473625e-02 -1.33297950e-01
7.93515325e-01 -3.44146729e-01 -8.87944162e-01 9.14760083e-02
-7.30853140e-01 3.11923951e-01 8.02221656e-01 1.26484442e+00
5.35011649e-01 -7.13187099e-01 7.28971183e-01 -7.56424963e-01
3.10190648e-01 -3.97795141e-01 -4.71304834e-01 1.77719548e-01
-6.96983159e-01 4.70081151e-01 6.85343325e-01 -1.91813365e-01
-8.01613808e-01 6.32636607e-01 -8.44623744e-01 -9.33953747e-02
-2.07024693e-01 5.06358802e-01 3.16503644e-01 -2.08129764e-01
8.02651525e-01 4.54542935e-01 2.93111801e-01 -5.70856631e-01
3.63029480e-01 8.98594797e-01 2.98919976e-01 -4.18748140e-01
7.93988481e-02 2.91466385e-01 2.84517277e-02 -7.34798014e-01
-9.98042762e-01 -6.32520437e-01 -7.35776424e-01 2.20185816e-01
9.17125702e-01 -9.35286045e-01 -1.07064247e+00 7.80018866e-01
-1.17165720e+00 -6.39056623e-01 3.43752176e-01 2.47143850e-01
-7.60424376e-01 6.21656060e-01 -9.25011396e-01 -2.81019330e-01
-8.49245489e-01 -1.35167789e+00 1.27743173e+00 -3.21931630e-01
-1.09936595e-01 -6.68280721e-01 4.42839086e-01 6.01082623e-01
5.16549945e-01 -1.30579472e-01 9.05219555e-01 -1.07950854e+00
-2.08740517e-01 3.34379554e-01 -2.59479046e-01 4.72602040e-01
-9.32645127e-02 -4.72679511e-02 -1.10778964e+00 -5.22423565e-01
-3.23000133e-01 -6.27879798e-01 1.37445629e+00 3.46525520e-01
1.10561562e+00 1.35632053e-01 -2.22919434e-01 7.27495253e-01
1.21483600e+00 4.57242608e-01 4.98622686e-01 6.45319343e-01
5.11537373e-01 2.68391967e-01 2.93222934e-01 2.27528334e-01
1.62796289e-01 7.77185321e-01 4.95758504e-01 -2.09120959e-01
-1.95548698e-01 1.49172217e-01 3.00223172e-01 5.49522638e-01
-4.41717617e-02 -1.13279186e-01 -1.20647252e+00 3.28540593e-01
-1.82574880e+00 -1.00170445e+00 -2.14684933e-01 1.61373556e+00
1.09381652e+00 1.06206378e-02 3.04955184e-01 2.12908909e-02
3.05411249e-01 -3.86796682e-03 -7.01482892e-01 -5.36062360e-01
-4.40587811e-02 5.97499490e-01 6.00835145e-01 4.11169350e-01
-1.49524319e+00 1.03427672e+00 7.16448450e+00 8.48647416e-01
-1.20301223e+00 7.38500059e-02 9.07581449e-01 -1.20614909e-01
1.88386530e-01 -4.08679813e-01 -9.99918520e-01 3.15426201e-01
1.57272041e+00 3.63849789e-01 2.57912695e-01 8.75727832e-01
9.04975981e-02 1.40825376e-01 -1.18150699e+00 8.03434670e-01
-7.57052656e-03 -1.67043471e+00 -5.87705113e-02 8.41948763e-02
5.47799706e-01 8.22907865e-01 3.92943136e-02 3.75406861e-01
5.16769648e-01 -1.22235560e+00 3.72988433e-01 1.78781807e-01
7.01426089e-01 -8.10015440e-01 8.37331116e-01 9.39945802e-02
-8.35221946e-01 6.66582659e-02 -4.28884029e-01 -1.70846507e-01
-1.30764442e-02 4.93641108e-01 -1.13640475e+00 3.63682568e-01
9.10135508e-01 6.42991722e-01 -5.37655830e-01 8.17801118e-01
1.11669935e-01 9.06133771e-01 -8.88973475e-02 -2.46014535e-01
5.89047551e-01 2.60441482e-01 1.95766419e-01 1.48214722e+00
-3.01103175e-01 -1.72857177e-02 1.13819800e-01 3.15610111e-01
-1.37959868e-01 1.09352343e-01 -2.14524120e-01 3.44467759e-02
-8.56732726e-02 1.23006725e+00 -4.66758549e-01 -6.75285518e-01
-6.70106471e-01 1.00497150e+00 7.34542251e-01 2.37717018e-01
-9.30713654e-01 -5.38246989e-01 9.44931388e-01 -2.90258944e-01
8.41774523e-01 -1.99529976e-01 -1.37213767e-01 -1.09899139e+00
-4.26599860e-01 -1.39354396e+00 2.35464051e-01 -6.92992687e-01
-1.34441149e+00 8.14588845e-01 -5.20506263e-01 -6.57488286e-01
-1.00392066e-01 -1.27148414e+00 2.94065718e-02 8.14054549e-01
-1.55988061e+00 -1.12304986e+00 2.08565310e-01 4.23462152e-01
7.06077397e-01 -5.35161078e-01 1.10541797e+00 1.94495440e-01
-2.96664119e-01 5.57639122e-01 5.69896698e-01 1.50027052e-01
8.05797398e-01 -1.48119271e+00 1.03115916e+00 6.79494262e-01
2.54500031e-01 5.11035323e-01 5.94852448e-01 -3.75583857e-01
-1.29567730e+00 -1.01903093e+00 9.75545108e-01 -7.17608154e-01
8.48553836e-01 -3.20497781e-01 -9.13212240e-01 7.45683074e-01
3.74566972e-01 -3.58343050e-02 1.01212025e+00 3.06963295e-01
-6.09729469e-01 6.68338239e-02 -8.54850292e-01 2.53310531e-01
1.00150216e+00 -5.19069254e-01 -7.33853936e-01 5.18095911e-01
8.48655343e-01 -2.51647264e-01 -1.02144051e+00 5.65359473e-01
8.93423200e-01 -1.08805966e+00 9.84525323e-01 -1.12851858e+00
7.40125060e-01 -1.49407893e-01 -3.04460317e-01 -1.43745017e+00
-6.01564407e-01 -5.09037077e-01 -2.08526939e-01 3.84665519e-01
7.03831792e-01 -4.97814566e-01 1.01027274e+00 2.10701711e-02
-4.92850065e-01 -1.10595953e+00 -9.23344433e-01 -6.76632285e-01
4.50844616e-01 -2.60934383e-01 3.11928511e-01 6.40120864e-01
2.52562761e-02 7.16307461e-01 -3.44033331e-01 -1.84866861e-01
3.45856696e-01 1.63689211e-01 5.96757352e-01 -8.65053177e-01
-6.33684039e-01 -3.09307694e-01 -4.84971076e-01 -1.23116577e+00
2.71841586e-01 -9.96654212e-01 4.69113663e-02 -1.18738866e+00
4.23881710e-01 3.81754413e-02 -6.97691143e-01 8.03635418e-01
-2.59890229e-01 5.41541457e-01 1.71189785e-01 1.28277615e-01
-6.79955542e-01 5.01687050e-01 9.79942143e-01 -3.47446233e-01
1.43056110e-01 -1.40992552e-01 -5.15322208e-01 3.26117724e-01
1.02786779e+00 -2.47645378e-01 -5.38615249e-02 -3.56244922e-01
1.25413025e-02 -1.51567444e-01 3.25610101e-01 -1.14646757e+00
-2.16207206e-01 3.03734213e-01 8.02302063e-01 -6.64893448e-01
3.71705711e-01 -5.03747463e-01 -9.74257290e-02 1.03299975e+00
-6.07580900e-01 1.50144279e-01 3.92796218e-01 4.65343386e-01
2.92639364e-03 2.27731198e-01 9.74504173e-01 -1.56012148e-01
-8.20183575e-01 1.71902075e-01 -7.56920755e-01 -3.10218900e-01
9.03248191e-01 1.59795076e-01 -5.52213728e-01 -1.25980705e-01
-1.06524956e+00 -4.95285504e-02 1.58423334e-01 3.90641659e-01
3.48273665e-01 -9.26403284e-01 -1.00198090e+00 2.35252947e-01
2.18412519e-01 -7.64069855e-01 6.11750036e-02 8.20055366e-01
-8.61361742e-01 9.16693866e-01 -5.35416603e-01 -8.85221779e-01
-1.54708719e+00 6.36153936e-01 7.48171449e-01 -4.06701863e-01
-4.75870639e-01 1.15785253e+00 1.48469508e-01 -5.26215196e-01
8.21280107e-02 -4.03033912e-01 -1.42362922e-01 -2.68607229e-01
6.90154076e-01 7.57308006e-02 6.30002320e-01 -5.44822097e-01
-5.25189638e-01 2.81081975e-01 -5.88410318e-01 1.40313715e-01
1.39381814e+00 2.72152901e-01 1.35770470e-01 2.59056479e-01
1.46141624e+00 -6.30883753e-01 -1.39028394e+00 -1.24361955e-01
7.53728673e-02 8.84037986e-02 8.86603445e-02 -1.43687284e+00
-7.39503384e-01 9.54458952e-01 9.55614150e-01 -1.20128289e-01
8.72805476e-01 -1.68927357e-01 8.48666072e-01 7.49508440e-01
2.33102947e-01 -6.85815632e-01 4.86012138e-02 9.54644203e-01
5.65758407e-01 -1.56856990e+00 -1.79129332e-01 -3.32854912e-02
-5.14965057e-01 1.33201981e+00 3.79327983e-01 -9.99372378e-02
5.73097765e-01 3.29803824e-01 3.55148345e-01 -2.11333439e-01
-1.31258035e+00 -1.38454929e-01 1.60059065e-01 7.30718255e-01
1.12481093e+00 1.57410219e-01 -2.35099420e-01 1.87160924e-01
-1.81368604e-01 7.76334926e-02 2.35269845e-01 8.08472931e-01
-5.90488434e-01 -1.19529617e+00 1.00964261e-02 5.06798327e-01
-1.03112531e+00 -6.43844664e-01 -4.81971055e-01 3.43319893e-01
-1.04620293e-01 9.66090500e-01 -1.16470620e-01 -4.94465202e-01
-3.68742482e-03 4.69862640e-01 5.46011567e-01 -4.60100621e-01
-9.61638272e-01 -4.91922460e-02 3.33112836e-01 -8.10351193e-01
-5.69005549e-01 -4.71576363e-01 -1.10345507e+00 -4.44839776e-01
-1.04272798e-01 8.08640495e-02 7.42807150e-01 9.67475593e-01
7.24532127e-01 5.67571759e-01 1.43546104e-01 -8.86424482e-01
-8.73507261e-01 -1.20343256e+00 -1.05731733e-01 2.86370456e-01
6.23899400e-01 -3.86523336e-01 1.56014130e-01 2.41383031e-01] | [4.710888862609863, 5.659704685211182] |
a3ad7018-06c4-4dd1-a50e-2ff7ee541307 | navigating-explanatory-multiverse-through | 2306.02786 | null | https://arxiv.org/abs/2306.02786v1 | https://arxiv.org/pdf/2306.02786v1.pdf | Navigating Explanatory Multiverse Through Counterfactual Path Geometry | Counterfactual explanations are the de facto standard when tasked with interpreting decisions of (opaque) predictive models. Their generation is often subject to algorithmic and domain-specific constraints -- such as density-based feasibility for the former and attribute (im)mutability or directionality of change for the latter -- that aim to maximise their real-life utility. In addition to desiderata with respect to the counterfactual instance itself, the existence of a viable path connecting it with the factual data point, known as algorithmic recourse, has become an important technical consideration. While both of these requirements ensure that the steps of the journey as well as its destination are admissible, current literature neglects the multiplicity of such counterfactual paths. To address this shortcoming we introduce the novel concept of explanatory multiverse that encompasses all the possible counterfactual journeys and shows how to navigate, reason about and compare the geometry of these paths -- their affinity, branching, divergence and possible future convergence -- with two methods: vector spaces and graphs. Implementing this (interactive) explanatory process grants explainees more agency by allowing them to select counterfactuals based on the properties of the journey leading to them in addition to their absolute differences. | ['Yueqing Xuan', 'Edward Small', 'Kacper Sokol'] | 2023-06-05 | null | null | null | null | ['navigate'] | ['reasoning'] | [ 2.15921924e-01 4.76880938e-01 -4.64538515e-01 -2.68522710e-01
-2.19733685e-01 -8.04073513e-01 1.06194592e+00 3.07943225e-01
-3.36013407e-01 1.08260429e+00 3.76445204e-01 -1.16858888e+00
-9.27164435e-01 -8.54295433e-01 -3.93603176e-01 -5.49483955e-01
-4.42233056e-01 6.30711317e-01 -1.59267947e-01 -1.42281473e-01
5.06698668e-01 5.25687933e-01 -1.57414305e+00 -8.55509341e-02
1.03773928e+00 7.08766460e-01 -2.05874160e-01 5.00762701e-01
-3.83259773e-01 2.66687363e-01 -1.56899795e-01 -8.21984947e-01
1.58404231e-01 -4.53774601e-01 -9.09310937e-01 6.69380426e-02
-2.89038628e-01 -8.06131512e-02 1.67279571e-01 8.89640212e-01
6.16363175e-02 1.69251636e-01 1.03805983e+00 -1.57655203e+00
-4.90760237e-01 6.60798669e-01 -3.21245402e-01 1.62978888e-01
4.21910971e-01 2.63955563e-01 1.07889330e+00 -1.92987815e-01
7.15767026e-01 1.14430690e+00 3.65606904e-01 3.49095792e-01
-1.55658412e+00 -2.31813252e-01 4.14176255e-01 -9.80326384e-02
-9.88325357e-01 -3.26282769e-01 7.84323454e-01 -4.65679497e-01
6.86436951e-01 6.50385082e-01 7.76190221e-01 1.00667930e+00
3.33965957e-01 3.17611814e-01 1.04846144e+00 -4.22008783e-01
6.27403557e-01 2.11007342e-01 -2.64460087e-01 1.55940071e-01
6.02563858e-01 3.87424380e-01 -1.66192308e-01 -3.66330951e-01
5.27054429e-01 -3.54661047e-01 -3.66560549e-01 -7.86283731e-01
-1.14531386e+00 9.24698412e-01 2.67517835e-01 2.91559100e-01
-4.77278084e-01 -5.08941188e-02 3.19949955e-01 1.88470557e-01
2.65855014e-01 7.44951010e-01 -4.40440893e-01 -6.06570356e-02
-7.35698819e-01 5.11361301e-01 8.48875940e-01 6.15764081e-01
5.34082592e-01 -3.31210494e-02 1.99641600e-01 2.34590799e-01
3.51964176e-01 -1.00784436e-01 4.58827019e-01 -6.71223521e-01
6.51591301e-01 7.32847929e-01 4.92290258e-01 -1.00998878e+00
-4.15571004e-01 -5.80061555e-01 -6.74323142e-01 3.88712794e-01
8.75697196e-01 -9.51871499e-02 -4.81862098e-01 1.86307037e+00
6.23472095e-01 6.07998669e-03 -1.83663517e-02 8.96773994e-01
-8.91656335e-03 5.02648890e-01 5.56490183e-01 -6.56686246e-01
9.99232948e-01 -2.29610708e-02 -4.41337645e-01 -4.33172733e-02
9.32911873e-01 -3.21339965e-01 1.02980030e+00 2.20908225e-01
-9.85871553e-01 1.94160808e-02 -9.59400296e-01 4.53483790e-01
-3.31102788e-01 -5.92565119e-01 9.25942838e-01 8.21930110e-01
-7.78628886e-01 9.16124463e-01 -5.05501509e-01 -2.24823236e-01
1.54134035e-01 3.22893292e-01 -2.80009389e-01 3.50613207e-01
-1.15983033e+00 9.67087686e-01 5.70534289e-01 1.12927713e-01
-2.56777048e-01 -6.80573761e-01 -5.50106585e-01 1.03227861e-01
5.46380699e-01 -9.89071429e-01 8.52462590e-01 -1.31427824e+00
-1.07674456e+00 6.99118257e-01 1.31446898e-01 -6.90674961e-01
1.16987455e+00 3.08504164e-01 -5.59576929e-01 -3.44571501e-01
1.69483423e-01 4.12132710e-01 4.31479722e-01 -1.05833709e+00
-6.79039836e-01 -3.63973349e-01 4.65943336e-01 5.83127737e-01
3.07664394e-01 -2.26224065e-01 2.22992733e-01 -3.46583605e-01
2.31941193e-01 -7.43595600e-01 -5.43523431e-01 9.43873376e-02
-8.44829917e-01 -2.68724591e-01 3.45022738e-01 -7.39099309e-02
1.27968371e+00 -1.85796976e+00 8.72426108e-03 3.98013741e-01
6.95553515e-03 -2.17633069e-01 3.07674646e-01 5.13136327e-01
-4.95595038e-01 4.64113146e-01 -4.68343943e-01 9.19234976e-02
2.37707853e-01 8.85490514e-03 -3.23459595e-01 7.90105104e-01
1.04530096e-01 6.12359822e-01 -9.78907347e-01 -3.74015331e-01
3.66035759e-01 1.22688428e-01 -3.47861648e-01 -2.53853589e-01
-4.20245409e-01 2.18984336e-01 -5.45880139e-01 1.02023885e-01
4.65490162e-01 1.03622027e-01 3.82067561e-01 2.67987520e-01
-4.46007818e-01 5.84416986e-01 -1.74214864e+00 1.12010849e+00
-4.36769515e-01 4.16716814e-01 -3.17669004e-01 -8.37783873e-01
7.65973330e-01 1.46100655e-01 2.79728413e-01 -3.49420309e-01
-1.26610875e-01 4.43497449e-01 1.81416631e-01 -2.95695007e-01
5.53946078e-01 -6.63422227e-01 3.82679105e-02 6.27803743e-01
-5.15674055e-01 -3.18391286e-02 1.12514243e-01 9.52848494e-02
5.62296271e-01 4.03638721e-01 8.69888246e-01 -7.04779088e-01
3.92707556e-01 1.27794594e-01 4.71226096e-01 4.38804895e-01
3.68498936e-02 4.95084018e-01 1.04506719e+00 -4.12555486e-01
-9.68095958e-01 -1.11693585e+00 -2.95844287e-01 3.93011659e-01
1.41841903e-01 3.11844517e-03 -4.33577150e-01 -8.43349516e-01
1.88865602e-01 1.56037569e+00 -9.52948868e-01 -7.05707818e-02
-3.24101537e-01 -7.60744572e-01 -9.78781059e-02 1.47314325e-01
1.41688228e-01 -6.35690331e-01 -1.22941303e+00 2.81149149e-01
1.07328789e-02 -1.19872794e-01 -8.57176483e-02 2.31362745e-01
-9.56011653e-01 -1.33723140e+00 -4.40293491e-01 1.96638614e-01
6.46514893e-01 -1.09234035e-01 9.13636923e-01 -1.22303646e-02
-9.75269591e-04 2.38022745e-01 -3.06061767e-02 -3.90088946e-01
-4.92502123e-01 -4.22366232e-01 -1.50320619e-01 5.55762649e-02
5.46054766e-02 -6.72098637e-01 -6.78031206e-01 1.62824050e-01
-8.76399398e-01 1.83027998e-01 2.29271054e-01 6.30795658e-01
5.89455068e-01 2.71330953e-01 6.32695496e-01 -9.53142762e-01
6.75458312e-01 -8.40347052e-01 -6.45767629e-01 3.69007230e-01
-9.62026358e-01 3.71250331e-01 5.53327262e-01 -3.95199627e-01
-1.24601173e+00 -2.45760590e-01 8.08278620e-02 2.25615487e-01
-2.94294804e-01 6.87853992e-01 -5.59213996e-01 5.40427685e-01
6.45568669e-01 -3.18196937e-02 -1.97466582e-01 -1.82827011e-01
8.52890491e-01 3.08993518e-01 3.72574955e-01 -5.18904388e-01
6.01444066e-01 5.57586789e-01 4.06555593e-01 -5.91531634e-01
-2.45613933e-01 3.68551686e-02 -7.28501379e-01 -3.36070925e-01
6.33290291e-01 -2.57593930e-01 -8.26844931e-01 -3.66093546e-01
-9.69236195e-01 2.88027059e-02 -6.92139328e-01 5.32354534e-01
-8.09980392e-01 4.72019799e-02 3.48270684e-01 -1.34333968e+00
3.36889982e-01 -1.11770415e+00 2.92624503e-01 -9.49771851e-02
-8.36920679e-01 -1.27949178e+00 -1.56803206e-01 -2.06716940e-01
2.66646624e-01 7.88433552e-01 1.28046823e+00 -7.92138517e-01
-5.43143988e-01 -2.55838752e-01 3.18865106e-02 -5.15921414e-01
4.26190719e-02 3.06871682e-01 -5.39984524e-01 3.13455649e-02
4.91534248e-02 3.42076689e-01 5.81978679e-01 7.33747780e-01
6.82596266e-01 -9.24003005e-01 -6.79443598e-01 3.19374889e-01
1.53831303e+00 4.99982387e-01 5.29090822e-01 6.17832363e-01
7.08988011e-02 1.21465194e+00 4.73121405e-01 4.34377104e-01
4.19481963e-01 8.69571149e-01 7.82996595e-01 1.21375389e-01
4.23694044e-01 -3.84173393e-01 -1.55124888e-01 -2.28623927e-01
-1.00595161e-01 -3.23991239e-01 -8.18932354e-01 7.12969065e-01
-1.72172654e+00 -1.12412190e+00 -5.55382073e-01 2.69923329e+00
2.55566806e-01 4.66603547e-01 4.68554795e-01 3.55963796e-01
6.71432257e-01 2.01046422e-01 -5.04577100e-01 -8.62765253e-01
-8.45148116e-02 -5.14532387e-01 4.96730417e-01 6.01540685e-01
-7.21989095e-01 1.08488016e-01 5.55163336e+00 6.24296665e-01
-8.27451587e-01 -2.46895269e-01 9.87264216e-01 -9.14837644e-02
-1.13306081e+00 5.26937485e-01 -3.00688058e-01 6.01467550e-01
9.80347216e-01 -5.62035203e-01 9.06739607e-02 7.56293774e-01
4.28011626e-01 -5.04756868e-01 -1.32848835e+00 2.06477985e-01
-6.77138865e-01 -1.46333587e+00 2.08689094e-01 3.65033239e-01
3.91970009e-01 -7.34361231e-01 1.01154797e-01 -2.66613185e-01
4.46469486e-01 -1.07314789e+00 1.30888569e+00 4.41888690e-01
7.52529681e-01 -1.12764812e+00 4.98466700e-01 5.22509754e-01
-9.71937120e-01 -1.45946398e-01 1.51376829e-01 -2.65174717e-01
3.18409234e-01 4.81210768e-01 -6.30161226e-01 8.89697552e-01
9.25098211e-02 8.89943838e-02 -9.89531279e-02 9.44997072e-01
-1.22957215e-01 4.40078437e-01 -3.45033854e-01 -3.06997240e-01
4.43816781e-01 -4.83452022e-01 8.61019075e-01 9.85525966e-01
4.69662040e-01 1.25463426e-01 -6.44078135e-01 1.23248625e+00
4.69043791e-01 5.53233139e-02 -8.91137421e-01 4.15076762e-02
6.49626017e-01 6.27494872e-01 -1.05088902e+00 8.45476836e-02
-2.22777903e-01 5.80828011e-01 -6.89890981e-02 4.39408600e-01
-6.50745869e-01 -6.80267736e-02 8.05327058e-01 5.35856962e-01
-3.53835642e-01 -5.35585023e-02 -7.09068060e-01 -6.97903335e-01
1.52864698e-02 -5.47124267e-01 6.20610893e-01 -4.31006163e-01
-8.34251463e-01 2.56009787e-01 3.14808309e-01 -1.04276538e+00
-4.68613744e-01 -1.68730497e-01 -8.57657671e-01 1.17173445e+00
-1.15521097e+00 -8.86150718e-01 5.62418878e-01 2.25512564e-01
2.71290392e-01 2.24752039e-01 6.20904386e-01 -3.87802631e-01
-4.24649090e-01 1.17505506e-01 8.66764784e-03 -6.51138842e-01
-1.01142339e-01 -1.38251007e+00 3.53974938e-01 8.10792327e-01
1.70899287e-01 7.67296016e-01 1.28359640e+00 -7.61049271e-01
-1.04882932e+00 -6.31516397e-01 1.32794333e+00 -3.35972667e-01
8.49266112e-01 -1.17400028e-02 -7.30610728e-01 6.32890999e-01
-1.47404939e-01 -6.94639862e-01 1.36094049e-01 3.66604298e-01
-1.30731976e-02 9.77900028e-02 -1.04319942e+00 1.15427232e+00
9.39382136e-01 -3.56291160e-02 -5.73536992e-01 2.12591309e-02
5.08688211e-01 7.08488971e-02 -2.97981113e-01 1.86080173e-01
5.71071446e-01 -1.45761561e+00 8.59169245e-01 -7.13340819e-01
5.13535798e-01 -2.49319702e-01 -2.48415917e-02 -1.32726526e+00
-1.22747645e-01 -8.40831280e-01 4.71001387e-01 1.38843572e+00
7.14026809e-01 -8.49686503e-01 6.95203006e-01 1.17365420e+00
1.29070148e-01 -9.92866397e-01 -1.43024421e+00 -7.02622473e-01
1.62706912e-01 -8.17811608e-01 1.11147046e+00 1.04476821e+00
2.34450415e-01 -7.64315110e-03 -1.25458702e-01 -1.11824639e-01
6.14293575e-01 2.58572400e-01 5.51871896e-01 -1.36582708e+00
-1.74364686e-01 -8.00212026e-01 -3.43568236e-01 -6.20008230e-01
-4.69485484e-03 -8.36786091e-01 -4.82355297e-01 -1.62032592e+00
-1.84857517e-01 -7.11160719e-01 9.31308717e-02 -1.89477190e-01
-1.17889047e-02 -7.23465502e-01 8.10966268e-02 1.80178523e-01
-3.71068567e-02 4.39736485e-01 8.91865730e-01 3.07041049e-01
-5.22029221e-01 2.93288857e-01 -8.93611670e-01 7.86614835e-01
6.44734263e-01 -3.42618555e-01 -6.73297584e-01 2.26716921e-01
5.07247865e-01 6.61906242e-01 5.44138014e-01 -4.33396727e-01
6.71119690e-02 -7.54574299e-01 1.90050572e-01 -2.51156658e-01
1.87207729e-01 -8.37573886e-01 9.47546721e-01 5.45884073e-01
-7.14966774e-01 1.44048214e-01 -3.69894318e-02 9.16362643e-01
2.60087579e-01 -3.82701755e-01 5.34559071e-01 -1.67120740e-01
-5.05291522e-01 -1.35596126e-01 -3.58467340e-01 -1.54094845e-01
1.33782160e+00 -7.59442568e-01 -1.76626429e-01 -5.13141155e-01
-5.83791137e-01 2.30452925e-01 6.51395679e-01 1.87095940e-01
3.54912847e-01 -9.86830473e-01 -6.40227199e-01 -1.13163911e-01
-1.00923248e-03 -3.48085582e-01 2.56619126e-01 7.80903578e-01
-1.86937138e-01 5.89277804e-01 -1.38297141e-01 -4.66852225e-02
-8.32264185e-01 8.43858063e-01 5.20429611e-01 -2.25216851e-01
-6.21984422e-01 5.21647155e-01 3.70902300e-01 -1.18173644e-01
-2.10094064e-01 -8.17562118e-02 -2.70854115e-01 3.36340427e-01
8.70815217e-02 6.20795906e-01 -1.95403084e-01 -5.93591511e-01
-4.15970325e-01 1.26048476e-01 5.16401947e-01 -3.02627027e-01
1.16626775e+00 -5.99235237e-01 3.58361304e-02 5.75211823e-01
6.61414444e-01 -4.95385379e-02 -1.44680607e+00 2.39004597e-01
4.84952629e-01 -6.93352878e-01 -3.04464519e-01 -9.17197168e-01
-3.99660975e-01 6.12640858e-01 7.35560656e-02 7.25521505e-01
8.23278844e-01 -4.16944288e-02 -2.49253467e-01 -2.31109530e-01
3.26700181e-01 -8.62899601e-01 -6.66292727e-01 -3.02860409e-01
1.02529562e+00 -8.97523642e-01 1.30383790e-01 -2.92106956e-01
-6.31925583e-01 1.10776031e+00 5.41978516e-02 2.28925556e-01
3.73516768e-01 -1.51587993e-01 -3.02529871e-01 -6.11653849e-02
-8.88294458e-01 -1.10034198e-01 4.29464161e-01 6.33528292e-01
5.12051761e-01 5.43566704e-01 -9.05079484e-01 2.87526906e-01
-4.47754204e-01 -3.38909894e-01 6.21007740e-01 4.52251226e-01
-3.09267454e-02 -8.09979558e-01 -4.19803530e-01 6.94046497e-01
-2.97324508e-01 2.26746276e-02 -2.60145187e-01 1.47003078e+00
1.68663204e-01 1.07846320e+00 2.14565262e-01 1.24100395e-01
3.88416737e-01 3.72357038e-03 2.73975655e-02 -1.74810573e-01
-3.72622997e-01 -4.75283340e-02 6.35755301e-01 -4.04783010e-01
-2.25709185e-01 -9.20569837e-01 -1.00557947e+00 -5.69713712e-01
-5.51909566e-01 4.19215143e-01 7.94843495e-01 1.17038584e+00
7.57682472e-02 2.20976144e-01 3.88750285e-01 -5.55676877e-01
-7.72946715e-01 -3.18045706e-01 -7.87885845e-01 2.93328732e-01
3.37855905e-01 -6.80617690e-01 -5.80398560e-01 -2.73815006e-01] | [8.644810676574707, 5.584219932556152] |
518b4a28-74fc-49f7-9da3-b38bd95f6968 | mcl-iitk-at-semeval-2021-task-2-multilingual | 2104.01567 | null | https://arxiv.org/abs/2104.01567v1 | https://arxiv.org/pdf/2104.01567v1.pdf | MCL@IITK at SemEval-2021 Task 2: Multilingual and Cross-lingual Word-in-Context Disambiguation using Augmented Data, Signals, and Transformers | In this work, we present our approach for solving the SemEval 2021 Task 2: Multilingual and Cross-lingual Word-in-Context Disambiguation (MCL-WiC). The task is a sentence pair classification problem where the goal is to detect whether a given word common to both the sentences evokes the same meaning. We submit systems for both the settings - Multilingual (the pair's sentences belong to the same language) and Cross-Lingual (the pair's sentences belong to different languages). The training data is provided only in English. Consequently, we employ cross-lingual transfer techniques. Our approach employs fine-tuning pre-trained transformer-based language models, like ELECTRA and ALBERT, for the English task and XLM-R for all other tasks. To improve these systems' performance, we propose adding a signal to the word to be disambiguated and augmenting our data by sentence pair reversal. We further augment the dataset provided to us with WiC, XL-WiC and SemCor 3.0. Using ensembles, we achieve strong performance in the Multilingual task, placing first in the EN-EN and FR-FR sub-tasks. For the Cross-Lingual setting, we employed translate-test methods and a zero-shot method, using our multilingual models, with the latter performing slightly better. | ['Ashutosh Modi', 'Deepak Mahajan', 'Jay Mundra', 'Rohan Gupta'] | 2021-04-04 | null | https://aclanthology.org/2021.semeval-1.62 | https://aclanthology.org/2021.semeval-1.62.pdf | semeval-2021 | ['sentence-pair-classification'] | ['natural-language-processing'] | [ 1.62156954e-01 -1.10100217e-01 1.57954305e-01 -3.63231868e-01
-1.29087043e+00 -8.71270716e-01 8.86429131e-01 1.96254551e-01
-8.48219156e-01 9.78666127e-01 2.21124485e-01 -8.32309067e-01
8.20695758e-02 -3.80647451e-01 -6.53030872e-01 -3.09176058e-01
4.23191972e-02 6.40943527e-01 9.41045210e-02 -6.90310359e-01
-8.47990811e-03 -1.24621131e-01 -1.26872194e+00 6.91569686e-01
8.56007874e-01 6.24839962e-01 6.58322155e-01 5.80874920e-01
-3.48323673e-01 6.86677039e-01 -4.76906389e-01 -6.60468221e-01
1.61964774e-01 -4.89847928e-01 -1.10046995e+00 -4.41412389e-01
5.68002522e-01 5.06742954e-01 2.93094397e-01 1.02523243e+00
5.01919091e-01 2.95574255e-02 4.98330295e-01 -9.55194294e-01
-5.09775579e-01 1.02398682e+00 -3.70449007e-01 2.77246386e-01
7.80345798e-01 -3.72270569e-02 1.25041914e+00 -1.14986968e+00
8.80626082e-01 1.56275892e+00 6.05682254e-01 4.72310871e-01
-1.27459538e+00 -4.89010185e-01 2.05387264e-01 4.35110629e-01
-1.29756963e+00 -7.17859328e-01 4.87284899e-01 -3.27862144e-01
1.47991681e+00 2.33456492e-01 1.89191356e-01 1.34480727e+00
1.01927467e-01 6.84731185e-01 1.56891274e+00 -1.02295732e+00
7.73954242e-02 2.31678024e-01 9.83410552e-02 2.89425164e-01
-2.29278564e-01 -4.17312644e-02 -4.88908768e-01 5.56025766e-02
-1.41011938e-01 -6.94687426e-01 -2.36454055e-01 1.91520631e-01
-1.32103026e+00 6.33751214e-01 9.94746163e-02 1.10647058e+00
-2.35982254e-01 -9.48678181e-02 6.36348546e-01 7.77738273e-01
5.84959149e-01 6.59713149e-01 -9.76399064e-01 -1.24350907e-02
-8.86171281e-01 2.63571560e-01 7.21754968e-01 8.26512396e-01
6.27008379e-01 -3.98997784e-01 -2.70469964e-01 1.15516090e+00
2.64522702e-01 5.78109682e-01 8.73522818e-01 -6.38094723e-01
9.19949770e-01 1.29237160e-01 -9.81558934e-02 -5.78857422e-01
-2.21765235e-01 -2.86761373e-01 -4.56981331e-01 -2.58944333e-01
4.18786883e-01 -7.36719593e-02 -6.99596345e-01 2.09098744e+00
6.43819347e-02 5.42733818e-02 6.26847148e-01 4.84606236e-01
6.01607263e-01 5.80182254e-01 3.54774803e-01 -2.22614765e-01
1.72254276e+00 -7.59598374e-01 -7.43938804e-01 -5.19487917e-01
1.04242945e+00 -1.17619741e+00 1.32193589e+00 1.81809708e-01
-7.86364079e-01 -7.29128361e-01 -1.14014316e+00 -2.27831379e-01
-7.96079218e-01 1.81449011e-01 1.21206855e-02 3.05594236e-01
-1.15381038e+00 5.45705974e-01 -2.96958506e-01 -6.34650409e-01
-3.49686652e-01 8.14797636e-03 -4.99792129e-01 -2.66543031e-01
-1.96244478e+00 1.37402928e+00 5.72360039e-01 -2.23099917e-01
-3.07463676e-01 -4.61825669e-01 -1.12318206e+00 -2.60681868e-01
2.02664152e-01 -3.50699812e-01 1.16376734e+00 -8.91695261e-01
-1.10049927e+00 1.42300820e+00 -2.47506425e-01 -5.44855893e-01
2.41326064e-01 -1.20543256e-01 -9.03453469e-01 -3.16630512e-01
7.18381166e-01 7.29187012e-01 4.13215280e-01 -1.19584346e+00
-8.95200789e-01 -4.05649453e-01 -4.13155966e-02 2.18678713e-01
3.17970142e-02 4.02364761e-01 -2.67357856e-01 -7.20741928e-01
-2.11627260e-01 -9.45204318e-01 1.06207192e-01 -8.32175791e-01
-3.08071792e-01 -6.34324074e-01 7.80041635e-01 -9.53067482e-01
1.17168880e+00 -2.11206388e+00 9.54804197e-02 6.88794851e-02
-3.72544646e-01 2.69914031e-01 -6.07713103e-01 5.40895760e-01
-3.77479643e-01 1.26996323e-01 -3.50467563e-01 -8.54006648e-01
-4.83703688e-02 5.54415166e-01 -3.28504235e-01 -1.95424613e-02
3.21914345e-01 8.67282689e-01 -9.97158885e-01 -4.45076138e-01
-3.90989743e-02 2.44601011e-01 -1.81206867e-01 -9.94496004e-05
-1.07765950e-01 4.65717942e-01 1.57932788e-01 1.52796119e-01
4.25044924e-01 2.87998587e-01 6.08301282e-01 -6.28345013e-02
-1.92113534e-01 9.69173074e-01 -1.04615879e+00 1.96407855e+00
-1.18394363e+00 4.57108915e-01 -1.30613307e-02 -9.43814158e-01
8.06909800e-01 5.76951027e-01 -4.49103266e-02 -1.04247081e+00
-7.73153976e-02 7.01016486e-01 1.54675767e-01 -4.66091156e-01
7.30517030e-01 -3.75342816e-01 -4.40568507e-01 3.41364026e-01
4.18232679e-01 -2.64298648e-01 5.99764287e-01 1.78764760e-01
1.00658941e+00 1.56737000e-01 5.32709062e-01 -5.69792807e-01
7.19411194e-01 -2.58536249e-01 3.75463337e-01 5.35386145e-01
6.80310130e-02 2.37057313e-01 3.49495500e-01 9.73510966e-02
-7.89014101e-01 -1.03380787e+00 -2.92679578e-01 1.17735898e+00
-1.36717573e-01 -6.18782282e-01 -6.73721850e-01 -8.54747474e-01
-1.33620963e-01 1.26243722e+00 -3.84120524e-01 1.02576159e-01
-6.78752363e-01 -3.60579550e-01 5.76666832e-01 1.72286123e-01
1.75298676e-01 -1.13547444e+00 -5.94759434e-02 5.03611445e-01
-7.16401160e-01 -1.56700075e+00 -7.11521208e-01 4.75868464e-01
-2.63061464e-01 -1.00035024e+00 -4.16056126e-01 -1.08391809e+00
1.48438469e-01 -1.63480729e-01 1.37218380e+00 -2.11680010e-01
-1.75910387e-02 2.77956247e-01 -5.36693275e-01 -3.57837342e-02
-7.25209355e-01 1.79857045e-01 2.73957700e-01 -5.69593720e-02
5.54354429e-01 -4.01503414e-01 1.48726955e-01 8.01028088e-02
-6.56731069e-01 -2.00888187e-01 3.34835291e-01 8.60844314e-01
5.28229535e-01 -2.57372051e-01 7.55296826e-01 -1.04254818e+00
7.67101169e-01 -4.73498225e-01 -2.64004171e-01 4.60723937e-01
-4.50252324e-01 3.02840739e-01 7.65049458e-01 -3.08611333e-01
-8.09281409e-01 -1.92081761e-02 -4.30866361e-01 -4.05892394e-02
-1.26881868e-01 7.05454171e-01 -6.01144373e-01 4.47986841e-01
5.61206639e-01 7.74335787e-02 -3.37525249e-01 -6.46996379e-01
5.47566295e-01 8.48164320e-01 8.00404191e-01 -8.07364702e-01
3.68921965e-01 -3.48057181e-01 -4.79726821e-01 -6.50320828e-01
-9.67090547e-01 -5.22643030e-01 -6.43538713e-01 -1.02573425e-01
1.07301402e+00 -9.70515847e-01 -1.78058505e-01 3.03376943e-01
-1.61398447e+00 -3.21920425e-01 -2.08992362e-01 4.13083345e-01
-3.14119041e-01 3.12957883e-01 -4.60636139e-01 -5.47525525e-01
-3.34569931e-01 -1.25316238e+00 1.21611476e+00 -3.94454092e-01
-4.85381961e-01 -1.22537291e+00 3.09166640e-01 2.27205127e-01
3.83015603e-01 -1.66360125e-01 1.12837839e+00 -9.71654117e-01
1.16002977e-01 7.24605992e-02 3.85535210e-02 5.36389291e-01
2.08359018e-01 -3.56326371e-01 -1.00737059e+00 -2.60225594e-01
-6.60895333e-02 -4.83109832e-01 7.90951073e-01 -5.96444607e-02
6.26335144e-01 -1.25631824e-01 -1.53960407e-01 9.80074108e-02
1.24475706e+00 1.34856552e-01 5.31422019e-01 2.90817529e-01
3.48879963e-01 8.47920179e-01 5.14306247e-01 -9.77334976e-02
8.02468598e-01 1.01859653e+00 -1.34156272e-01 -9.40992869e-03
-3.37200105e-01 -3.21937412e-01 7.26606905e-01 1.37342727e+00
3.47539604e-01 -3.43940318e-01 -9.24305379e-01 7.83865869e-01
-1.76561844e+00 -8.11323702e-01 -3.85718465e-01 2.11065674e+00
1.23799074e+00 2.38220274e-01 -2.58990973e-01 1.31611615e-01
6.75687730e-01 5.49922176e-02 2.79796962e-02 -4.98206526e-01
-5.54652512e-01 4.78814453e-01 1.07810870e-01 9.87077594e-01
-1.18210554e+00 1.28162050e+00 5.36228609e+00 1.09980094e+00
-1.02997541e+00 4.23987538e-01 5.58694363e-01 2.93191046e-01
-5.03040254e-01 2.43713543e-01 -9.75504875e-01 6.30440593e-01
1.24394393e+00 1.35849109e-02 4.52853471e-01 4.36046690e-01
-1.28305033e-01 -3.56335074e-01 -1.39465392e+00 8.24310660e-01
2.58271366e-01 -8.39133203e-01 -1.99821338e-01 -3.89346391e-01
6.30903840e-01 2.79867977e-01 -2.86364198e-01 7.37833440e-01
4.05529618e-01 -7.10588217e-01 8.30715895e-01 3.31444532e-01
1.04577398e+00 -4.81942058e-01 8.68322253e-01 2.57958144e-01
-1.44519448e+00 5.48185587e-01 -9.21303965e-03 1.34727582e-01
2.64721215e-01 6.34808719e-01 -6.53663635e-01 8.34149659e-01
5.56123793e-01 7.67267108e-01 -7.67885625e-01 3.05705220e-01
-5.40677607e-01 5.16504347e-01 -2.08995566e-01 6.97448999e-02
1.08632594e-01 -6.30880818e-02 7.41817415e-01 1.50529289e+00
3.83628011e-01 -4.31372762e-01 3.97611290e-01 6.10844135e-01
-1.45045146e-01 4.03874606e-01 -7.42018282e-01 2.68348873e-01
5.48094332e-01 1.15852368e+00 -3.48762333e-01 -6.21689618e-01
-4.24572885e-01 1.20199764e+00 6.24265373e-01 1.73469841e-01
-4.16813940e-01 -5.30839860e-01 4.39581543e-01 -1.55257761e-01
1.32273003e-01 -1.60064608e-01 -7.73741305e-02 -1.25432050e+00
1.40668094e-01 -9.63723779e-01 2.69469559e-01 -7.85354972e-01
-1.60953224e+00 9.44713414e-01 -8.67216140e-02 -1.06995368e+00
-7.54430950e-01 -8.18135560e-01 -5.00533581e-01 1.25519073e+00
-1.63685858e+00 -1.14022624e+00 5.24400473e-01 6.27850294e-01
4.51338410e-01 -3.89023274e-01 1.14314663e+00 6.43369496e-01
-1.41985744e-01 7.04088986e-01 -4.74985465e-02 1.02330871e-01
1.20164227e+00 -1.40582132e+00 5.84820330e-01 9.57957625e-01
4.42923188e-01 5.44008434e-01 7.00411260e-01 -5.42551994e-01
-8.69398117e-01 -1.24152994e+00 2.04432797e+00 -4.27780867e-01
1.18922770e+00 -6.33963466e-01 -9.46875811e-01 4.75402504e-01
5.78736305e-01 -1.22066088e-01 6.66039884e-01 4.66241211e-01
-5.03019333e-01 6.48855120e-02 -9.03934419e-01 4.30537194e-01
9.22969699e-01 -1.16771388e+00 -9.99742210e-01 3.76348585e-01
9.44360852e-01 -9.10181701e-02 -7.75571287e-01 1.26447260e-01
1.03523351e-01 -4.48776841e-01 5.95969737e-01 -7.18071401e-01
5.37692249e-01 -3.26868594e-01 -6.62657678e-01 -1.82751858e+00
-5.08147571e-03 -2.12260112e-01 5.62080503e-01 1.61050344e+00
8.25453043e-01 -6.40652895e-01 -3.16209048e-01 -1.84987411e-02
-2.53489792e-01 -3.77342314e-01 -1.28741586e+00 -1.01555216e+00
2.30742395e-01 -8.50460947e-01 4.94559735e-01 1.26678538e+00
1.62888944e-01 9.84417021e-01 -6.29170835e-02 -9.90992188e-02
2.18772769e-01 -2.88800374e-02 2.80929953e-01 -9.83624220e-01
-5.93348742e-01 -2.78628320e-01 -1.05524361e-01 -6.60504103e-01
7.06856132e-01 -1.56715071e+00 3.39245647e-01 -1.18640292e+00
-4.28736657e-02 -3.66466194e-01 -3.00882757e-01 8.96329820e-01
-4.65687960e-01 1.83687851e-01 4.07078564e-01 -8.61712769e-02
-5.65732658e-01 4.59181219e-01 7.59599268e-01 -2.27199331e-01
-3.05528613e-03 -3.48377287e-01 -4.92081344e-01 4.13227081e-01
5.81887186e-01 -4.75257725e-01 4.69994098e-02 -5.32642663e-01
2.22189516e-01 2.17759400e-03 2.19065472e-01 -7.55477428e-01
-5.17645618e-03 4.15252954e-01 -8.13419148e-02 -3.16702604e-01
1.96150187e-02 -6.37795568e-01 -1.64104968e-01 3.39517951e-01
-4.61797953e-01 3.95208776e-01 3.75740528e-01 -3.90713215e-02
-4.78431076e-01 -3.72757226e-01 7.13902056e-01 4.23737755e-03
-7.55860686e-01 -3.44170660e-01 -5.16307890e-01 4.64818209e-01
7.06763804e-01 4.87382025e-01 -2.91157007e-01 -1.06072687e-01
-8.61883044e-01 4.04481292e-01 1.43454596e-01 8.22295368e-01
1.73725754e-01 -1.50586009e+00 -1.01805961e+00 2.60846496e-01
5.11876941e-01 -5.67364097e-01 -5.87641597e-02 7.50499904e-01
2.18419284e-01 4.95452434e-01 1.14212491e-01 -3.64412397e-01
-1.25068617e+00 4.09810394e-01 3.55558783e-01 -7.47334480e-01
-4.81551373e-03 6.42372906e-01 -1.63630217e-01 -1.16446757e+00
-1.17500469e-01 -2.47381479e-01 -2.68244863e-01 2.91360050e-01
3.58622015e-01 1.71518456e-02 4.92290050e-01 -8.98469031e-01
-7.00636029e-01 4.71499532e-01 -9.21577960e-02 -4.23688173e-01
9.70952272e-01 -1.42827898e-01 -4.55654949e-01 8.36015701e-01
1.42760241e+00 2.42092803e-01 -2.93604285e-01 -5.26441216e-01
5.37871718e-01 1.48156121e-01 -1.42735904e-02 -1.21062052e+00
-5.03312528e-01 7.23535717e-01 5.95275223e-01 5.62912524e-02
8.76854420e-01 2.87422001e-01 6.14520729e-01 3.34615827e-01
5.71248114e-01 -1.30553389e+00 -4.48826432e-01 1.20960319e+00
1.10185194e+00 -1.38156331e+00 -5.20164430e-01 -1.42555252e-01
-6.12034321e-01 8.79083693e-01 3.60869765e-01 1.25366628e-01
4.91029352e-01 3.56784821e-01 2.24334970e-01 4.23639379e-02
-9.20902729e-01 -4.25063759e-01 5.11654198e-01 2.78514951e-01
8.74220490e-01 4.34952080e-01 -7.63346374e-01 7.08912730e-01
-5.71717083e-01 -3.05815876e-01 1.28288925e-01 5.99721849e-01
-2.84255110e-02 -1.66767323e+00 -2.95906037e-01 2.12558985e-01
-3.69469196e-01 -7.20809400e-01 -4.40413982e-01 7.74808407e-01
4.40956295e-01 1.16240776e+00 5.40496372e-02 -4.18106347e-01
4.41136837e-01 5.04584253e-01 4.20057535e-01 -6.89355671e-01
-7.03605652e-01 6.64500296e-02 5.32992244e-01 -4.13125098e-01
-4.61825788e-01 -9.26295578e-01 -1.20842123e+00 1.80257156e-01
-3.45344320e-02 3.45782876e-01 7.03689098e-01 1.34287190e+00
1.94040552e-01 5.88295639e-01 6.20440066e-01 -6.09371722e-01
-3.95605832e-01 -1.35580313e+00 -4.34382409e-01 5.18912137e-01
5.53318709e-02 -4.40070629e-01 -3.07549715e-01 7.98299387e-02] | [10.927149772644043, 9.983525276184082] |
86dbdf5b-f3c9-4f4b-b677-edcb8fc902bf | ems-net-efficient-multi-temporal-self | 2303.13753 | null | https://arxiv.org/abs/2303.13753v1 | https://arxiv.org/pdf/2303.13753v1.pdf | EMS-Net: Efficient Multi-Temporal Self-Attention For Hyperspectral Change Detection | Hyperspectral change detection plays an essential role of monitoring the dynamic urban development and detecting precise fine object evolution and alteration. In this paper, we have proposed an original Efficient Multi-temporal Self-attention Network (EMS-Net) for hyperspectral change detection. The designed EMS module cuts redundancy of those similar and containing-no-changes feature maps, computing efficient multi-temporal change information for precise binary change map. Besides, to explore the clustering characteristics of the change detection, a novel supervised contrastive loss is provided to enhance the compactness of the unchanged. Experiments implemented on two hyperspectral change detection datasets manifests the out-standing performance and validity of proposed method. | ['Bo Du', 'Chen Wu', 'Meiqi Hu'] | 2023-03-24 | null | null | null | null | ['change-detection'] | ['computer-vision'] | [ 3.25087845e-01 -6.13697886e-01 2.39662960e-01 -1.10235766e-01
9.05922279e-02 -3.59687716e-01 3.92763406e-01 -8.80832747e-02
-1.87899217e-01 7.20520198e-01 1.30366623e-01 2.06112508e-02
-8.24649155e-01 -1.02384973e+00 -2.24219248e-01 -1.00245440e+00
-1.29873425e-01 -7.99027011e-02 2.02651158e-01 -4.17734057e-01
1.11380599e-01 7.66913593e-01 -1.75248051e+00 -1.21904403e-01
1.03195596e+00 9.67556655e-01 4.80285019e-01 4.94499773e-01
1.31240845e-01 3.84142280e-01 -1.84564859e-01 2.18305424e-01
3.92547399e-01 -2.85145193e-01 -5.52613735e-01 2.65940070e-01
3.89987975e-01 -3.64414752e-01 -6.40193224e-01 1.67530489e+00
6.87561393e-01 3.05094272e-01 7.02738285e-01 -1.05637527e+00
-6.55883968e-01 5.27192354e-01 -9.80316401e-01 1.04718733e+00
-3.05985570e-01 3.37229282e-01 8.39785457e-01 -7.03879237e-01
4.96413499e-01 1.05177212e+00 7.11444438e-01 -2.61915416e-01
-5.58742940e-01 -6.52625084e-01 2.63141632e-01 8.51207078e-01
-1.46237683e+00 -3.99379909e-01 1.06514299e+00 -4.40085828e-01
6.87460363e-01 5.78060985e-01 9.28732991e-01 2.08047464e-01
2.40474548e-02 5.74480772e-01 9.84495223e-01 -9.05385837e-02
-1.65196598e-01 -1.94249019e-01 3.11218023e-01 6.25222266e-01
6.65748537e-01 2.34445587e-01 2.69471169e-01 3.19340348e-01
4.38316166e-01 3.17015201e-01 -6.39635444e-01 1.29573420e-01
-8.88780832e-01 4.74723697e-01 8.90726924e-01 7.27563798e-01
-3.82007927e-01 -1.10452659e-01 3.85964453e-01 2.84809500e-01
7.45405972e-01 1.89318746e-01 -3.87921512e-01 1.96171582e-01
-7.81010807e-01 -3.90384793e-01 -6.37510344e-02 4.89803135e-01
1.02629817e+00 2.07885325e-01 -1.44347653e-01 6.49814904e-01
1.81803644e-01 7.81310558e-01 7.93127179e-01 -4.44091976e-01
3.46121222e-01 8.18314731e-01 -6.76937923e-02 -1.61882889e+00
-7.04520464e-01 -8.86401892e-01 -1.32964003e+00 2.10547224e-01
-5.67984402e-01 -1.06275134e-01 -6.84042394e-01 1.28585815e+00
5.22962093e-01 1.48973003e-01 -2.98768897e-02 9.02920008e-01
9.11477268e-01 8.92786741e-01 -1.68164492e-01 -4.18032616e-01
9.67334569e-01 -8.28001559e-01 -7.58606613e-01 1.26435488e-01
1.91845521e-02 -1.21158488e-01 7.24961221e-01 -1.45950153e-01
-6.44380748e-01 -6.17048025e-01 -1.13846457e+00 2.52154052e-01
-6.47859812e-01 1.29434586e-01 5.23663759e-01 3.44535440e-01
-8.71691525e-01 6.56570196e-01 -5.34167528e-01 -4.62071508e-01
7.28031456e-01 2.14925990e-01 -2.12808311e-01 1.13604374e-01
-1.30436325e+00 5.64729095e-01 8.79594386e-01 5.04714310e-01
-5.73427320e-01 -7.64672041e-01 -5.80705225e-01 2.99410224e-01
1.77567765e-01 -3.41102421e-01 6.09561622e-01 -1.26227665e+00
-1.14884388e+00 5.47890365e-01 7.82480016e-02 -7.29664937e-02
3.56475204e-01 3.37030172e-01 -8.51198256e-01 4.51182395e-01
2.50832010e-02 3.14558357e-01 8.01323354e-01 -1.21689463e+00
-9.76597488e-01 -5.41531920e-01 -3.28112900e-01 7.25163043e-01
-7.34857798e-01 -8.78660753e-02 -2.97063757e-02 -5.07988334e-01
4.23298210e-01 -5.38549304e-01 -3.04472204e-02 -8.44881609e-02
-2.52834648e-01 -5.72195649e-02 1.36897743e+00 -1.03024590e+00
1.22949970e+00 -2.31933594e+00 -1.46852806e-01 3.90369087e-01
3.27046931e-01 5.61038196e-01 -2.14497700e-01 8.26187655e-02
-3.89646769e-01 2.80083627e-01 -6.36193097e-01 2.86055833e-01
-9.81860794e-03 -1.59641892e-01 -6.09466061e-02 7.14364827e-01
1.66861236e-01 7.38652349e-01 -7.80611277e-01 -3.24584484e-01
4.06204462e-01 1.21387452e-01 -8.65445957e-02 -1.59720078e-01
9.33953971e-02 2.98713982e-01 -4.58617866e-01 8.97150755e-01
1.08180702e+00 -5.34822643e-02 -3.69424105e-01 -4.53420490e-01
-5.72361410e-01 -4.65918511e-01 -1.34806013e+00 1.06688929e+00
1.74372748e-01 8.47510457e-01 1.66304410e-01 -1.05911803e+00
6.41757905e-01 -4.33821119e-02 7.73569822e-01 -6.12262845e-01
2.00718775e-01 -1.03205897e-01 -1.86469033e-01 -1.06689274e+00
4.75457609e-01 3.61679904e-02 5.98216474e-01 1.57009616e-01
-5.69847465e-01 -3.33934054e-02 7.66654685e-02 -3.14317852e-01
8.86008620e-01 -3.16042632e-01 4.26909775e-01 -2.89899498e-01
6.50854707e-01 1.77731127e-01 4.15150315e-01 3.39074552e-01
-6.20539010e-01 8.11249763e-02 -2.59369463e-01 -3.53247613e-01
-7.49532282e-01 -5.29221952e-01 -2.55406678e-01 7.60433853e-01
5.22814631e-01 4.48563278e-01 -1.44203335e-01 -5.12377262e-01
7.89535269e-02 4.56778556e-01 -5.89354455e-01 -3.24306577e-01
-1.72841355e-01 -1.36204469e+00 4.36261833e-01 1.98327392e-01
1.26262581e+00 -1.10504532e+00 -4.13893998e-01 2.64903396e-01
-2.61638403e-01 -4.95225132e-01 -2.01960400e-01 1.99419297e-02
-7.75861144e-01 -1.14958489e+00 -2.70836443e-01 -1.06900930e+00
5.42786002e-01 8.12904000e-01 5.42441964e-01 -4.12044115e-02
-5.12178242e-01 9.83814001e-02 -4.13139105e-01 -3.01855177e-01
6.16144203e-02 -1.54602066e-01 -1.03070825e-01 2.13995039e-01
3.86564918e-02 -8.05306494e-01 -6.71860039e-01 1.44038005e-02
-1.10706139e+00 -1.63274541e-01 6.39531255e-01 7.53052950e-01
4.79416579e-01 1.38177812e+00 5.00765204e-01 -3.78482699e-01
2.97315925e-01 -5.69283009e-01 -5.78890264e-01 2.33667582e-01
-8.36388409e-01 -3.30405772e-01 4.82860267e-01 -1.50757834e-01
-1.44384634e+00 -1.99868307e-01 1.47964790e-01 -3.82003486e-01
-3.63976717e-01 7.60511756e-01 -3.00018013e-01 -3.57303798e-01
4.65214610e-01 6.36368632e-01 -2.04539791e-01 -3.29236776e-01
1.71794981e-01 9.77588713e-01 6.99274063e-01 6.77541569e-02
1.11862636e+00 7.03890562e-01 8.35549906e-02 -9.48963702e-01
-4.86707896e-01 -8.15678358e-01 -5.40695131e-01 -3.41885060e-01
5.92076004e-01 -9.71156538e-01 -4.47337806e-01 8.75473619e-01
-8.18507195e-01 -1.26733348e-01 -3.87205780e-01 3.20736170e-01
-1.01354055e-01 6.44970238e-01 -3.77425641e-01 -7.26432800e-01
-6.75795436e-01 -5.12107611e-01 7.35824287e-01 4.23996270e-01
8.29342544e-01 -9.51450229e-01 2.00085595e-01 1.21803284e-01
5.74936569e-01 4.12295014e-01 9.15554166e-01 -1.32269815e-01
-5.49865603e-01 2.87977364e-02 -7.44886875e-01 1.31773680e-01
6.67788804e-01 2.84270078e-01 -8.61677051e-01 -2.67886043e-01
1.38240129e-01 2.76368260e-01 1.25319362e+00 6.03143692e-01
1.27859664e+00 -4.45519716e-01 -4.23814029e-01 9.51200783e-01
1.95914471e+00 5.02307773e-01 7.34045565e-01 5.33740461e-01
9.10261214e-01 4.90590096e-01 3.92507493e-01 5.87531686e-01
2.23129436e-01 1.38649926e-01 7.03202784e-01 -4.89322156e-01
-2.79550105e-01 2.11500108e-01 2.11641088e-01 7.39610195e-01
-2.68092770e-02 -1.42770335e-01 -6.81270003e-01 6.67037904e-01
-1.68388140e+00 -1.79999626e+00 -4.57710534e-01 1.52261937e+00
7.26251841e-01 -3.82438272e-01 -2.77258903e-01 3.77416939e-01
1.47370660e+00 4.31643248e-01 -9.45994556e-01 4.63034898e-01
-7.84285605e-01 1.30597213e-02 5.19933879e-01 2.95328319e-01
-1.48669183e+00 7.84726918e-01 5.65048790e+00 9.69862044e-01
-1.30090952e+00 2.95785576e-01 4.10840839e-01 8.73789713e-02
-3.13028812e-01 -3.40739578e-01 -3.79529387e-01 5.20994723e-01
1.44911125e-01 -2.27775499e-01 4.87204015e-01 4.34694856e-01
7.60789335e-01 9.15105082e-03 -1.39551699e-01 1.13670313e+00
7.14344531e-02 -9.79588985e-01 1.76357701e-01 -2.22910821e-01
1.05704570e+00 2.57725865e-01 3.41670960e-02 2.40086839e-02
1.58993267e-02 -4.02311653e-01 4.89875913e-01 8.33211780e-01
8.70165586e-01 -6.78249180e-01 7.01442063e-01 -1.20224487e-02
-1.61262786e+00 -5.46507180e-01 -5.17879665e-01 2.58069206e-02
-2.43813679e-01 9.54615355e-01 -4.82893139e-01 9.01673734e-01
7.30933785e-01 1.37687707e+00 -7.62203753e-01 1.45926583e+00
1.37621770e-02 5.84251225e-01 -2.44734600e-01 1.43732727e-01
3.03294927e-01 -6.16745353e-01 1.03946054e+00 1.15547407e+00
4.54163313e-01 4.85251814e-01 -2.04114560e-02 7.05171406e-01
1.06940567e-01 1.69316977e-01 -5.36905169e-01 -1.12160936e-01
3.22030306e-01 1.27200520e+00 -7.56792486e-01 -3.19689840e-01
3.83201279e-02 1.00630236e+00 -9.92487594e-02 4.61352229e-01
-8.02004218e-01 -7.92268634e-01 7.08314598e-01 -2.23201513e-01
6.62936985e-01 5.73010854e-02 -3.57699901e-01 -1.12899721e+00
7.06830397e-02 -4.93062496e-01 5.35400450e-01 -1.02929342e+00
-1.24589002e+00 3.36901665e-01 -1.98841974e-01 -1.24725640e+00
7.66260147e-01 -2.87367314e-01 -9.22236145e-01 5.30393481e-01
-2.07100511e+00 -1.03815854e+00 -9.66369271e-01 8.34437549e-01
3.66916776e-01 -7.72433504e-02 1.76178128e-01 4.00113225e-01
-1.24979997e+00 1.95715666e-01 7.12246776e-01 -1.39105648e-01
2.59688258e-01 -9.63006556e-01 -2.45517507e-01 1.41987753e+00
-4.24456477e-01 -2.04972535e-01 4.05730128e-01 -8.10044587e-01
-8.51712108e-01 -1.68333793e+00 2.70879418e-01 5.59756696e-01
7.55790412e-01 7.04621613e-01 -7.48645782e-01 3.41528922e-01
1.70257345e-01 -3.19534183e-01 2.71082997e-01 -4.10968900e-01
-1.98732689e-02 -6.68937564e-01 -1.39908719e+00 5.32528579e-01
1.26678312e+00 -3.16491634e-01 -4.42124784e-01 5.85068285e-01
8.13192308e-01 6.31958544e-02 -6.15755498e-01 6.82705998e-01
9.94960740e-02 -7.75738418e-01 7.44040489e-01 -3.06642115e-01
3.27846766e-01 -6.61176503e-01 -1.31300122e-01 -1.44710660e+00
-1.17884934e+00 -2.41898045e-01 1.94320321e-01 1.14801228e+00
-7.07283765e-02 -8.68496835e-01 4.18818533e-01 -7.95800611e-02
-5.30386567e-01 -1.28839478e-01 -7.64681458e-01 -5.82997859e-01
-4.45945948e-01 -1.28943529e-02 9.26425397e-01 1.38622570e+00
-3.29652071e-01 -1.51170762e-02 3.16467322e-02 7.88867056e-01
4.37733799e-01 2.16793209e-01 -4.13844315e-03 -1.54153943e+00
1.61966383e-01 -1.09713316e+00 -6.87774062e-01 -4.18660134e-01
3.11275870e-01 -9.56044078e-01 -5.88520542e-02 -1.70701766e+00
5.08224905e-01 -1.93838030e-01 -6.50199294e-01 7.14979172e-01
-4.06578183e-01 3.16267371e-01 -2.70018280e-01 4.08009768e-01
-4.12269711e-01 1.09918785e+00 1.30454195e+00 -8.21062088e-01
-3.80931616e-01 -9.70653743e-02 -6.42782748e-01 3.96502316e-01
9.43650246e-01 -3.76512110e-01 -3.84410053e-01 -3.12123954e-01
1.37131125e-01 -4.38922852e-01 4.30409402e-01 -1.40776396e+00
3.47789407e-01 -3.74534518e-01 4.34542090e-01 -8.60133469e-01
-1.06664911e-01 -1.01664364e+00 3.24692190e-01 7.41364777e-01
5.24989031e-02 -2.61975914e-01 3.31971228e-01 6.33074403e-01
-2.06319839e-01 -3.46510321e-01 9.65288997e-01 -7.33703841e-03
-1.33168828e+00 6.97321415e-01 -3.72417778e-01 -3.28940988e-01
1.05367494e+00 -4.23964083e-01 -5.46339452e-01 -2.71544829e-02
-6.13344550e-01 1.37813374e-01 2.20628619e-01 9.14958417e-02
6.52423084e-01 -1.32939529e+00 -8.71412158e-01 2.65124500e-01
1.60908967e-01 -1.36555970e-01 6.08623564e-01 7.18800128e-01
-7.25650191e-01 1.60488356e-02 -4.04984444e-01 -3.55935067e-01
-1.32552969e+00 4.44108367e-01 8.31098855e-01 -6.51332503e-03
-6.87367797e-01 6.47766650e-01 -6.93281814e-02 -3.10483456e-01
-3.89568061e-01 -3.50629032e-01 -6.27785504e-01 4.20173317e-01
4.29114610e-01 8.68861675e-01 1.62906930e-01 -6.72218978e-01
-1.78948745e-01 6.90256834e-01 5.14351547e-01 1.65072680e-01
1.67731428e+00 -6.37247920e-01 -7.08455086e-01 1.89997464e-01
1.14868569e+00 -2.57343262e-01 -1.21827519e+00 -3.23072016e-01
-4.28625762e-01 -3.78208071e-01 5.65710247e-01 -8.20716202e-01
-1.36302352e+00 4.14298713e-01 1.23275506e+00 3.17415148e-01
1.55577302e+00 -5.08163095e-01 4.87430960e-01 7.10984528e-01
6.49127811e-02 -9.91220295e-01 -1.57348171e-01 6.51273489e-01
9.76710558e-01 -1.31242442e+00 6.58313558e-02 -2.30882481e-01
-1.82390600e-01 1.00783706e+00 6.12438738e-01 2.84373611e-01
1.01776600e+00 -1.73653722e-01 -3.49850178e-01 -6.69002473e-01
-7.78571889e-02 -6.59530997e-01 1.87362030e-01 6.69399738e-01
-4.69384640e-01 2.50887454e-01 -1.34834781e-01 1.04011022e-01
-1.75094008e-02 -3.35665584e-01 4.43366766e-01 6.27633989e-01
-1.05170488e+00 -4.30086665e-02 -5.65960467e-01 7.01399922e-01
1.80729583e-01 -2.39230260e-01 -3.34790558e-01 4.86520350e-01
5.41631699e-01 1.15366149e+00 1.51255742e-01 -5.34528494e-01
2.11711317e-01 -1.45040244e-01 1.64005533e-01 -1.92350924e-01
-4.24427211e-01 -2.18210176e-01 -2.27644563e-01 -1.82185441e-01
-7.43126512e-01 -8.45652342e-01 -9.93241251e-01 -2.15406135e-01
-5.89008749e-01 4.51061800e-02 5.74556351e-01 8.99179935e-01
4.32714939e-01 6.83022439e-01 1.32351708e+00 -7.60949790e-01
-4.40822124e-01 -1.30116975e+00 -1.05144346e+00 3.69686723e-01
4.93857086e-01 -5.14790118e-01 -6.82313919e-01 1.07970024e-02] | [9.822577476501465, -1.3733774423599243] |
420e76d2-cca4-4cc0-8bba-2548770dfe15 | dagrid-directed-accumulator-grid | 2306.02589 | null | https://arxiv.org/abs/2306.02589v1 | https://arxiv.org/pdf/2306.02589v1.pdf | DAGrid: Directed Accumulator Grid | Recent research highlights that the Directed Accumulator (DA), through its parametrization of geometric priors into neural networks, has notably improved the performance of medical image recognition, particularly with small and imbalanced datasets. However, DA's potential in pixel-wise dense predictions is unexplored. To bridge this gap, we present the Directed Accumulator Grid (DAGrid), which allows geometric-preserving filtering in neural networks, thus broadening the scope of DA's applications to include pixel-level dense prediction tasks. DAGrid utilizes homogeneous data types in conjunction with designed sampling grids to construct geometrically transformed representations, retaining intricate geometric information and promoting long-range information propagation within the neural networks. Contrary to its symmetric counterpart, grid sampling, which might lose information in the sampling process, DAGrid aggregates all pixels, ensuring a comprehensive representation in the transformed space. The parallelization of DAGrid on modern GPUs is facilitated using CUDA programming, and also back propagation is enabled for deep neural network training. Empirical results show DAGrid-enhanced neural networks excel in supervised skin lesion segmentation and unsupervised cardiac image registration. Specifically, the network incorporating DAGrid has realized a 70.8% reduction in network parameter size and a 96.8% decrease in FLOPs, while concurrently improving the Dice score for skin lesion segmentation by 1.0% compared to state-of-the-art transformers. Furthermore, it has achieved improvements of 4.4% and 8.2% in the average Dice score and Dice score of the left ventricular mass, respectively, indicating an increase in registration accuracy for cardiac images. The source code is available at https://github.com/tinymilky/DeDA. | ['Jiahao Li', 'Jinwei Zhang', 'Rongguang Wang', 'Xiang Chen', 'Renjiu Hu', 'Hang Zhang'] | 2023-06-05 | null | null | null | null | ['image-registration', 'skin-lesion-segmentation', 'lesion-segmentation'] | ['computer-vision', 'medical', 'medical'] | [ 2.93036193e-01 2.26503313e-01 -1.06100127e-01 -2.49875367e-01
-5.18268466e-01 -9.81822237e-02 1.82428688e-01 1.97626814e-01
-5.29742599e-01 6.09273672e-01 -1.80702597e-01 -1.96735650e-01
9.40160230e-02 -9.08391535e-01 -4.69995677e-01 -1.06055009e+00
-7.25577176e-02 2.11292133e-01 1.58180222e-01 2.34307304e-01
-1.47070056e-02 8.34349155e-01 -1.08371043e+00 1.99988395e-01
9.96092141e-01 1.33605814e+00 7.41902664e-02 3.29683721e-01
6.40579239e-02 7.52345383e-01 -4.21297669e-01 -3.61518860e-01
2.95768350e-01 -1.65607959e-01 -5.13687909e-01 -1.31527185e-01
6.26908362e-01 -4.56949651e-01 -4.93173897e-01 8.67672682e-01
7.71335304e-01 -1.26564130e-01 4.92336124e-01 -8.21673512e-01
-2.78253555e-01 4.36878443e-01 -7.66498029e-01 1.67540029e-01
-4.68451887e-01 2.84121335e-01 6.51890993e-01 -7.02590108e-01
5.42996466e-01 6.16513014e-01 1.04919839e+00 5.08551538e-01
-1.15734565e+00 -6.70153916e-01 -3.63648206e-01 -4.47765477e-02
-1.51109481e+00 -3.64656337e-02 9.16004539e-01 -3.35972518e-01
9.10801530e-01 4.56130981e-01 9.09817398e-01 5.38613379e-01
4.57735449e-01 7.22716212e-01 1.00448859e+00 -3.75171691e-01
-1.17459418e-02 -1.25058457e-01 2.24325165e-01 7.78382421e-01
3.49877626e-01 3.11337374e-02 -3.08754355e-01 -6.58074096e-02
1.18538368e+00 4.21632864e-02 -3.43034029e-01 -2.21135795e-01
-1.12988949e+00 5.77296555e-01 8.21562290e-01 2.02132910e-01
-5.27625799e-01 9.05532762e-02 5.26614726e-01 -1.61101729e-01
7.31269240e-01 2.50571936e-01 -2.19927222e-01 4.72512320e-02
-1.21234727e+00 2.26532249e-03 5.75880349e-01 4.84078467e-01
4.64518815e-01 3.26486588e-01 -2.51089662e-01 9.89524603e-01
-5.55189066e-02 4.09185320e-01 3.72388452e-01 -1.02685606e+00
2.27724880e-01 8.51632178e-01 -4.73581135e-01 -1.12814546e+00
-8.55123520e-01 -9.49652255e-01 -1.56607139e+00 3.73374730e-01
6.22581899e-01 -2.95669526e-01 -9.05239224e-01 1.44218898e+00
4.82396036e-01 1.21357761e-01 -2.53455728e-01 9.28877652e-01
7.71297216e-01 4.54975665e-01 1.62122980e-01 -1.93722785e-01
1.26946831e+00 -8.79208267e-01 -4.07032520e-01 -3.81409079e-02
7.92481184e-01 -5.23192823e-01 7.94135451e-01 4.43791479e-01
-1.29016662e+00 -6.00434721e-01 -8.97478044e-01 -9.45830569e-02
-4.01548818e-02 3.75356764e-01 6.86962306e-01 7.89601862e-01
-1.20513463e+00 7.89324164e-01 -1.11609757e+00 7.63241500e-02
1.02542818e+00 5.33933938e-01 -9.04440060e-02 -4.68825512e-02
-8.81801546e-01 5.87662280e-01 2.28615955e-01 3.04270655e-01
-3.23903292e-01 -1.25233746e+00 -6.32798076e-01 -3.42980549e-02
7.34262392e-02 -8.46731186e-01 7.69457579e-01 -9.69446003e-01
-1.27949369e+00 9.31136727e-01 1.81570724e-01 -7.79450178e-01
7.30081975e-01 7.56474063e-02 -9.55737755e-02 3.59336883e-01
-1.90286055e-01 1.09893239e+00 5.27131200e-01 -7.97581315e-01
-2.88582355e-01 -5.75027704e-01 -4.30106729e-01 2.61653930e-01
-4.79888111e-01 -1.10837340e-01 -5.92468321e-01 -9.76376057e-01
2.38150001e-01 -9.29342628e-01 -4.59141642e-01 5.37425160e-01
-2.66184807e-01 1.49446726e-01 7.77190089e-01 -8.94405842e-01
1.10257244e+00 -2.04235411e+00 -3.12865973e-01 4.25520182e-01
6.91374063e-01 4.16743726e-01 2.25161165e-01 -2.57847726e-01
-1.35028988e-01 -1.72087207e-01 -5.37549436e-01 -3.56485695e-01
-4.08480287e-01 8.94649699e-02 1.39726117e-01 6.76233351e-01
1.29630342e-01 1.04162395e+00 -3.86786014e-01 -4.81847078e-01
3.46347690e-01 9.21448290e-01 -6.31828189e-01 -2.13357851e-01
3.04384977e-01 3.59318018e-01 -1.68007493e-01 6.68701530e-01
9.91002262e-01 -4.59663093e-01 2.74901837e-01 -3.33495945e-01
-3.46761383e-02 2.62918975e-02 -9.60082114e-01 1.62344217e+00
-2.25244448e-01 7.00094402e-01 7.22133666e-02 -9.38665509e-01
1.10847998e+00 2.07711145e-01 9.17654157e-01 -9.35768068e-01
2.56584764e-01 1.89996049e-01 5.41158468e-02 -7.11870268e-02
3.11059117e-01 4.54312786e-02 3.11844349e-01 2.89533675e-01
-2.94632226e-01 1.11136243e-01 -3.32940146e-02 -3.44821773e-02
8.42782497e-01 -1.96720138e-02 1.70713346e-02 -5.81521094e-01
2.94599950e-01 2.31603503e-01 6.66658461e-01 6.46274924e-01
-3.00286412e-01 9.22914028e-01 4.45951968e-01 -6.80938601e-01
-1.07397711e+00 -9.46134329e-01 -5.88691294e-01 4.75950211e-01
7.02846199e-02 -1.02550365e-01 -1.18898070e+00 -3.68949682e-01
-9.90754142e-02 3.71130705e-01 -5.71469903e-01 4.23761532e-02
-8.89224887e-01 -1.28300750e+00 7.73951173e-01 7.75670648e-01
8.15961003e-01 -9.62019563e-01 -9.45733905e-01 2.31533647e-01
6.72910288e-02 -8.59579802e-01 -3.35092515e-01 1.28124401e-01
-1.35252571e+00 -9.21437085e-01 -9.42710042e-01 -6.49489939e-01
1.03190863e+00 -5.12380041e-02 1.08181238e+00 1.82973266e-01
-8.16192985e-01 -1.74151659e-01 1.02868915e-01 -2.21776709e-01
-1.43079907e-01 1.57715335e-01 -2.62815714e-01 -2.30986863e-01
1.50200933e-01 -5.62717438e-01 -1.05333865e+00 2.46628717e-01
-8.44783723e-01 5.13001442e-01 4.17933106e-01 1.04955220e+00
7.86889732e-01 -1.86032161e-01 3.71543407e-01 -8.85177374e-01
2.92573482e-01 -1.42128900e-01 -5.20251274e-01 -7.99383894e-02
-7.84278810e-01 -3.67389679e-01 5.08694232e-01 -3.12048227e-01
-9.97309446e-01 1.52923942e-01 -3.54068071e-01 -3.42384666e-01
-1.32133529e-01 3.55975240e-01 2.89703935e-01 -2.82436490e-01
6.97806180e-01 1.57104343e-01 5.46718597e-01 -3.10756356e-01
-1.03220887e-01 3.77083451e-01 4.34247851e-01 -3.90804261e-01
2.14009419e-01 6.27134383e-01 1.90268904e-01 -6.92101538e-01
-4.26957756e-01 -2.17218414e-01 -7.53015876e-01 -3.42401624e-01
8.66020083e-01 -8.51710141e-01 -5.92765212e-01 9.61129069e-01
-9.78121698e-01 -5.51375985e-01 -4.80665714e-01 4.66473162e-01
-2.42408350e-01 2.97812104e-01 -8.32710385e-01 -3.36269319e-01
-9.03596222e-01 -1.35006428e+00 6.12496495e-01 2.84864038e-01
-2.59354889e-01 -1.16610873e+00 -3.21671635e-01 4.69299138e-01
5.63773155e-01 5.50866246e-01 8.32293391e-01 -4.39691365e-01
-4.62891608e-01 -2.95910597e-01 -5.59400499e-01 6.14587843e-01
-7.23638609e-02 -1.31159782e-01 -9.56779361e-01 -2.14123279e-01
5.80849685e-02 9.45737958e-02 7.56394923e-01 9.82816219e-01
1.55602539e+00 -6.58105835e-02 -3.31012428e-01 9.96098161e-01
1.39021420e+00 2.50763595e-01 8.29766095e-01 2.54590660e-01
8.00427556e-01 4.56108719e-01 2.43892178e-01 4.71160799e-01
2.13640109e-01 5.96774399e-01 4.33797240e-01 -8.71358156e-01
-6.54005170e-01 1.78809851e-01 -2.61297047e-01 8.31547737e-01
-2.64805734e-01 2.06365824e-01 -1.17105496e+00 3.87442976e-01
-1.45270610e+00 -5.39951444e-01 -3.80523652e-01 2.07483363e+00
1.03710616e+00 -9.59675200e-03 4.26781438e-02 9.53601897e-02
6.24552906e-01 1.48369148e-01 -6.53389990e-01 -2.33056962e-01
-1.77845716e-01 3.93167436e-01 7.46910572e-01 1.98451355e-01
-1.17348313e+00 6.91140234e-01 5.41358757e+00 1.12618124e+00
-1.46437860e+00 1.10465720e-01 1.50345683e+00 -2.01638326e-01
9.61675048e-02 -4.54534024e-01 -6.91185951e-01 4.40717489e-01
6.61572516e-01 1.36279196e-01 -6.78263837e-03 7.90468931e-01
3.16770375e-01 -1.30588591e-01 -4.95198935e-01 8.75775278e-01
1.83157753e-02 -1.71672416e+00 -6.31034747e-02 1.89097062e-01
8.37422550e-01 3.03504556e-01 7.59358183e-02 -1.93947405e-01
-2.07455933e-01 -1.10835719e+00 3.73998225e-01 3.77546549e-01
1.05944157e+00 -8.11870337e-01 8.27352166e-01 1.19474508e-01
-8.09977949e-01 2.19227895e-01 -3.00250202e-01 -5.84397279e-03
-7.36737531e-03 1.16277564e+00 -8.67115080e-01 3.59230429e-01
7.31886029e-01 5.81014037e-01 -5.21210313e-01 1.13596904e+00
2.80065268e-01 7.89098680e-01 -5.33202708e-01 1.81088418e-01
2.56228775e-01 -4.40074503e-01 4.07815486e-01 1.28077233e+00
1.98078170e-01 1.50949275e-02 -8.70785713e-02 8.98989975e-01
-1.76244289e-01 2.70731598e-01 -1.34498492e-01 6.95754170e-01
2.23036706e-01 1.33211875e+00 -1.03029084e+00 -2.85199255e-01
-1.92464173e-01 7.30528295e-01 1.04437256e-03 1.18335217e-01
-1.03119266e+00 -4.23866063e-01 5.01199603e-01 3.80721599e-01
6.90429583e-02 -9.82038900e-02 -1.18081319e+00 -7.26816297e-01
2.12721065e-01 -6.62250340e-01 3.28164935e-01 -4.34636295e-01
-8.94833624e-01 7.59415090e-01 -3.39124829e-01 -1.09135926e+00
1.50142133e-01 -5.95685303e-01 -5.62967062e-01 9.17733431e-01
-1.49564481e+00 -1.13809538e+00 -5.40244758e-01 3.91431123e-01
3.17235827e-01 2.97904145e-02 7.54515469e-01 5.55582821e-01
-7.88901269e-01 7.78396428e-01 2.06042364e-01 3.62866938e-01
5.58923185e-01 -1.00165606e+00 3.13410670e-01 7.34459937e-01
-1.04986556e-01 5.30807257e-01 1.33051828e-01 -6.06413543e-01
-1.12493742e+00 -1.32879424e+00 5.69840550e-01 9.08686593e-02
4.05639201e-01 -3.40979435e-02 -9.86938059e-01 2.97376871e-01
-2.29690913e-02 4.67177898e-01 6.37202740e-01 -2.88193822e-01
-9.28476155e-02 -3.84796768e-01 -1.33096576e+00 5.92676938e-01
7.30479836e-01 -3.33147421e-02 1.17232077e-01 3.37845504e-01
2.71292150e-01 -8.91139388e-01 -1.21949458e+00 6.04816854e-01
6.71699464e-01 -1.06521571e+00 1.18762052e+00 -6.15414642e-02
5.73391736e-01 -1.31356388e-01 2.56153256e-01 -8.64059806e-01
-2.85486817e-01 -3.54660928e-01 -2.80931103e-03 9.44690168e-01
3.66505355e-01 -7.25036740e-01 1.17331350e+00 7.31215298e-01
-3.46703470e-01 -1.41071689e+00 -1.11397207e+00 -3.10358435e-01
3.30461442e-01 -4.77095902e-01 2.53205240e-01 8.87932777e-01
-3.71777534e-01 -4.14441586e-01 -1.12263009e-01 -3.52792218e-02
7.30727434e-01 -2.47704208e-01 2.65032291e-01 -1.18801546e+00
-9.27207619e-02 -6.43415213e-01 -4.34821337e-01 -9.57166016e-01
-1.34146050e-01 -1.05783963e+00 -4.01578069e-01 -1.35584235e+00
1.68381616e-01 -7.63272703e-01 -2.41364807e-01 6.48565590e-01
-4.17106487e-02 1.03326190e+00 1.61331490e-01 2.94500709e-01
-5.34140021e-02 2.11218908e-01 1.53377712e+00 -1.59281895e-01
-2.48712391e-01 8.99381377e-03 -6.09184623e-01 8.02479744e-01
1.18188834e+00 -3.35490227e-01 -1.32561937e-01 -5.37392318e-01
-2.35225692e-01 -8.97198841e-02 6.26813650e-01 -1.24763548e+00
3.95834416e-01 2.91142225e-01 7.69394338e-01 -5.50594985e-01
2.76046544e-01 -6.96346462e-01 2.37015039e-01 7.40273118e-01
-1.84746802e-01 -9.50573385e-02 5.56449234e-01 -2.49687787e-02
-1.72061890e-01 -3.16269286e-02 1.11939728e+00 7.77156278e-02
-2.79777586e-01 4.32223469e-01 -2.41100177e-01 -1.47152081e-01
1.03553820e+00 -6.37957811e-01 -2.91037619e-01 1.58979952e-01
-6.56843841e-01 -1.62060186e-01 3.04697126e-01 -1.38725698e-01
4.99899328e-01 -1.05259109e+00 -7.16789663e-01 2.91730225e-01
-4.85928118e-01 4.81194019e-01 7.30715215e-01 1.35197735e+00
-1.06712258e+00 2.59679973e-01 -3.66121113e-01 -9.05546308e-01
-1.44503570e+00 -9.01489183e-02 5.65672159e-01 -4.88915890e-01
-1.10054207e+00 9.00031209e-01 1.28091618e-01 -2.55972534e-01
1.68216303e-01 -4.03405517e-01 -1.15941959e-02 -5.12240306e-02
4.41510946e-01 5.15769899e-01 4.15301234e-01 -4.23558235e-01
-2.44533166e-01 5.08199275e-01 -2.36971855e-01 5.46817064e-01
1.44035649e+00 9.62533876e-02 -4.34887797e-01 -1.57994136e-01
1.17969537e+00 -2.19919771e-01 -1.53704536e+00 -2.02074766e-01
-3.71022910e-01 -1.81847364e-01 6.40855134e-01 -1.01821864e+00
-1.76138306e+00 8.98357928e-01 9.31196153e-01 -9.37248543e-02
1.41101372e+00 -3.13702017e-01 1.01309216e+00 -4.95181158e-02
3.07927802e-02 -9.16436076e-01 -2.95099229e-01 2.22428367e-01
6.09920979e-01 -9.45415854e-01 1.88886210e-01 -5.52031934e-01
-6.56587660e-01 1.29800153e+00 4.97963160e-01 -2.06374928e-01
6.14730895e-01 7.44633138e-01 1.58739209e-01 -1.75745994e-01
-2.44437397e-01 3.96265298e-01 3.41475844e-01 4.62082624e-01
5.05578518e-01 1.54381737e-01 -3.00227880e-01 4.57657397e-01
-1.40296027e-01 2.65985057e-02 2.62528211e-01 5.84618568e-01
-1.14124671e-01 -8.41743410e-01 -3.05300623e-01 7.84653068e-01
-6.39016032e-01 -4.11901295e-01 2.20899552e-01 7.53845394e-01
1.89420447e-01 2.42523372e-01 5.05574286e-01 5.49112819e-02
1.06956296e-01 -1.79389820e-01 2.97081113e-01 -2.20864683e-01
-9.49853599e-01 1.82995498e-01 -1.49627119e-01 -5.70534527e-01
-2.36170560e-01 -6.23489141e-01 -1.38103855e+00 -2.96627939e-01
-1.51767552e-01 -3.20219964e-01 8.25341105e-01 6.16958797e-01
5.47247767e-01 8.53236377e-01 3.34179252e-01 -7.74087906e-01
-3.04697365e-01 -6.37635052e-01 -5.85933924e-01 1.50067983e-02
-6.92821592e-02 -3.93894702e-01 -3.83958593e-02 1.19727820e-01] | [14.522920608520508, -2.4902915954589844] |
2a42082e-904f-46c2-b7e6-83b545e4c338 | mlp-singer-towards-rapid-parallel-singing | null | null | https://arxiv.org/abs/2106.07886 | https://arxiv.org/pdf/2106.07886.pdf | MLP Singer: Towards Rapid Parallel Singing Voice Synthesis | Recent developments in deep learning have significantly improved the quality of synthesized singing voice audio. However, prominent neural singing voice synthesis systems suffer from slow inference speed due to their autoregressive design. Inspired by MLP-Mixer, a novel architecture introduced in the vision literature for attention-free image classification, we propose MLP Singer, a parallel Korean singing voice synthesis system. To the best of our knowledge, this is the first work that uses an entirely MLP-based architecture for voice synthesis. Listening tests demonstrate that MLP Singer outperforms a larger autoregressive GAN-based system, both in terms of audio quality and synthesis speed. In particular, MLP Singer achieves a real-time factor of up to 200 and 3400 on CPUs and GPUs respectively, enabling order of magnitude faster generation on both environments. | ['Younggun Lee', 'Hyeongju Kim', 'Jaesung Tae'] | 2021-06-15 | null | null | null | arxiv-2021-6 | ['singing-voice-synthesis'] | ['speech'] | [-5.98624572e-02 3.88929769e-02 2.34426603e-01 2.04501271e-01
-1.02930319e+00 -3.62622291e-01 4.48320955e-01 -5.26316285e-01
-1.76844131e-02 5.26423395e-01 3.57147455e-01 -2.02499762e-01
3.72326493e-01 -4.66313571e-01 -7.39069343e-01 -7.70260692e-01
3.00906122e-01 3.42148304e-01 7.09394109e-04 -8.52947496e-03
-1.75136194e-01 3.04933548e-01 -1.84993446e+00 3.49324256e-01
3.78907353e-01 9.67116296e-01 2.13700905e-01 1.33481646e+00
1.38745308e-01 1.01445353e+00 -8.79110277e-01 5.85081391e-02
3.44130732e-02 -8.74134064e-01 -5.22754431e-01 -1.05414085e-01
7.69980967e-01 -4.71076667e-01 -5.13860881e-01 5.50652146e-01
1.06410611e+00 3.15790415e-01 3.90318841e-01 -8.89234662e-01
-6.99357867e-01 8.33800137e-01 -1.23118691e-01 1.40856728e-01
9.74634960e-02 5.50196230e-01 1.24651349e+00 -9.93612528e-01
1.38761818e-01 1.18116641e+00 5.21049798e-01 9.08449352e-01
-1.02676797e+00 -7.78776109e-01 -4.53788251e-01 8.92997459e-02
-1.09892583e+00 -8.16288173e-01 9.88164663e-01 -4.75519001e-02
1.28365481e+00 4.65896130e-01 6.31951213e-01 1.22306418e+00
-1.80387832e-02 8.27267826e-01 1.14113736e+00 -4.73556697e-01
4.05155122e-01 -6.34212643e-02 -4.49715406e-01 6.67617321e-01
-5.12530804e-01 1.10650852e-01 -1.01114845e+00 1.65971279e-01
8.30514848e-01 -7.51291215e-01 -3.41295928e-01 3.39203387e-01
-1.21180654e+00 7.96483576e-01 3.12828988e-01 2.29298115e-01
-5.65733075e-01 7.64958799e-01 4.63511497e-01 3.82179648e-01
2.36235812e-01 6.39875531e-01 -7.14177340e-02 -5.84698498e-01
-1.33682740e+00 1.28581285e-01 9.61421609e-01 5.39937258e-01
1.72004960e-02 1.31455660e+00 -7.68657476e-02 9.18310821e-01
2.40253255e-01 5.39678276e-01 7.97948122e-01 -1.26606631e+00
-3.93871181e-02 -2.53291339e-01 -3.98182452e-01 -5.31262815e-01
-1.46805346e-01 -8.18206787e-01 -9.36991632e-01 3.81789148e-01
1.88007846e-01 -4.17709976e-01 -7.38441348e-01 1.40437508e+00
1.40766263e-01 5.99331319e-01 2.07232744e-01 1.08572423e+00
1.08414555e+00 9.60029542e-01 -3.31922561e-01 -2.52227038e-01
1.18409944e+00 -1.64709544e+00 -8.01663399e-01 6.68442324e-02
-2.69059539e-01 -9.80520308e-01 1.45693529e+00 8.26936126e-01
-1.31460202e+00 -8.73356342e-01 -1.20551395e+00 -1.29324406e-01
1.00832209e-01 2.43274480e-01 5.67687571e-01 9.24982369e-01
-1.32399154e+00 5.84462523e-01 -8.31091285e-01 -9.31392685e-02
2.93129027e-01 2.96429634e-01 1.47221223e-01 5.66044331e-01
-7.47525215e-01 3.74275655e-01 -1.89822599e-01 -8.49448666e-02
-1.48754430e+00 -8.89039993e-01 -2.29786649e-01 2.84529120e-01
1.89864054e-01 -1.05172384e+00 1.73693168e+00 -9.32428896e-01
-2.47534895e+00 3.97814631e-01 -1.55573830e-01 -8.64375830e-01
3.28243762e-01 -4.58195567e-01 -5.06151736e-01 2.84179211e-01
-4.38376725e-01 6.63771749e-01 1.40960884e+00 -9.44521427e-01
-2.96339422e-01 -1.28222510e-01 -7.86440000e-02 2.65136093e-01
-3.25819284e-01 -8.75117630e-02 -1.89192057e-01 -8.28687370e-01
-4.28984165e-01 -8.34288061e-01 -6.61678240e-02 -1.98491275e-01
-1.16530046e-01 -4.83174697e-02 9.11286533e-01 -7.57948220e-01
9.17796850e-01 -1.85625613e+00 6.15216419e-02 -4.81149912e-01
2.47191012e-01 6.16491973e-01 -1.97731674e-01 3.26656520e-01
1.01394560e-02 -1.97014347e-01 -2.54832804e-01 -5.56974530e-01
-1.77199803e-02 2.25953624e-01 -6.10366762e-01 1.78480268e-01
1.03187867e-01 8.86293232e-01 -6.66216671e-01 -1.28260672e-01
2.74503738e-01 8.63156140e-01 -7.07446694e-01 5.19243777e-01
-1.93834826e-01 5.57425737e-01 1.11549720e-01 6.40065610e-01
1.95250720e-01 3.44097346e-01 -1.36946455e-01 1.04748569e-01
-8.09710026e-02 5.14335513e-01 -8.78887892e-01 1.73904908e+00
-1.02688801e+00 9.34248626e-01 3.02186847e-01 -6.41482413e-01
1.01589668e+00 8.86494517e-01 7.68014938e-02 -3.94476682e-01
1.34940267e-01 4.43779975e-01 1.88344698e-02 -1.64683208e-01
6.48760855e-01 -3.82319272e-01 4.86650825e-01 1.87585071e-01
3.53595763e-01 -4.76121992e-01 -2.89761394e-01 -3.44249219e-01
9.82857823e-01 6.51366562e-02 -2.62887254e-02 -7.06794187e-02
4.42192703e-01 -3.37331653e-01 1.50097907e-01 7.25233495e-01
1.10487081e-02 9.29124653e-01 2.10759580e-01 -1.88106298e-01
-1.15713155e+00 -1.10973275e+00 1.26269057e-01 1.04987335e+00
-7.48442650e-01 -3.41441780e-01 -1.16276598e+00 -1.15230180e-01
-4.86027360e-01 7.67034888e-01 5.51839767e-04 2.45199740e-01
-7.82782257e-01 -3.21215600e-01 1.06041491e+00 6.73084915e-01
4.27136451e-01 -1.62892425e+00 -8.31228673e-01 4.29173678e-01
2.07571164e-01 -1.09074509e+00 -5.92718005e-01 2.43440587e-02
-9.74977672e-01 -4.48510557e-01 -1.12373328e+00 -8.48950863e-01
-6.30480498e-02 -1.09931380e-01 1.15662408e+00 -2.10916132e-01
-2.25997254e-01 3.29519391e-01 -1.19161204e-01 -3.29541236e-01
-7.11244285e-01 7.03183413e-02 3.30504715e-01 -9.55102667e-02
-3.79244417e-01 -9.15429831e-01 -5.17787099e-01 -2.10832298e-01
-6.63376987e-01 -5.28774746e-02 5.94596505e-01 1.07930553e+00
6.24462128e-01 7.72548020e-02 8.88209343e-01 -6.59089446e-01
9.55885231e-01 -2.34718993e-01 -5.09707928e-01 -2.70667225e-01
-5.84159255e-01 -4.99915481e-02 9.81945932e-01 -5.08769155e-01
-9.09290195e-01 7.57380798e-02 -6.88495517e-01 -9.86675501e-01
-2.92769909e-01 3.03837359e-01 1.30571514e-01 1.12934709e-01
6.03870809e-01 5.33226252e-01 -9.27644521e-02 -6.93350255e-01
4.81279492e-01 9.05490100e-01 1.05048883e+00 -2.26317510e-01
5.20516038e-01 1.65457338e-01 6.78743199e-02 -1.49492359e+00
-2.95213044e-01 -2.98414081e-01 -7.73909688e-02 -2.57268190e-01
8.67093682e-01 -1.02646828e+00 -9.92712557e-01 7.96593964e-01
-1.02637362e+00 -5.04640162e-01 -5.27809203e-01 5.05425870e-01
-9.05858576e-01 2.46133417e-01 -8.93681526e-01 -1.11007416e+00
-1.01017356e+00 -1.12503183e+00 9.72442806e-01 1.59744635e-01
-1.11412145e-01 -6.75519109e-01 2.15610862e-01 4.55006421e-01
8.68470013e-01 -2.07846658e-03 5.83784997e-01 -1.45771265e-01
-4.03680831e-01 7.61549920e-02 2.92321056e-01 7.11833477e-01
-4.20345478e-02 -9.05461609e-03 -1.55965078e+00 -3.01540732e-01
3.16888571e-01 -1.48477048e-01 8.78477752e-01 5.33867955e-01
9.45546567e-01 -4.81798023e-01 4.98999596e-01 6.61288679e-01
1.13598502e+00 2.13862538e-01 5.81397116e-01 -3.03441256e-01
8.69204402e-01 1.92945287e-01 1.92338735e-01 2.98445284e-01
-1.04524530e-01 7.57381439e-01 4.90237594e-01 -2.49016091e-01
-9.64298964e-01 -5.41270494e-01 6.31125808e-01 1.45352662e+00
-2.66406268e-01 -3.66818875e-01 -5.47825575e-01 5.57115912e-01
-1.39810526e+00 -8.55924726e-01 1.68961287e-02 2.20120001e+00
6.91199362e-01 -5.81993088e-02 3.00544411e-01 4.27964360e-01
2.78153986e-01 3.83534789e-01 -4.26984042e-01 -8.95386100e-01
-5.12271561e-03 1.02849364e+00 1.67746827e-01 5.56920648e-01
-8.82000089e-01 1.12988627e+00 6.31220675e+00 1.07250619e+00
-1.52189744e+00 2.55157083e-01 3.78775388e-01 -4.40257847e-01
-5.49164638e-02 -2.77859569e-01 -4.96431619e-01 2.45254949e-01
1.51610887e+00 1.82941511e-01 1.11046827e+00 1.03423560e+00
1.80949703e-01 3.99300843e-01 -7.36845136e-01 1.20655513e+00
2.63871700e-01 -1.47482336e+00 1.26788527e-01 7.12268874e-02
6.42449617e-01 2.04302341e-01 3.97228777e-01 2.64209867e-01
-1.40023425e-01 -1.23943186e+00 9.15115356e-01 4.96714711e-01
7.76121616e-01 -1.16091061e+00 5.50133049e-01 2.39709139e-01
-1.09235513e+00 1.39788389e-01 -2.67554045e-01 -1.48521706e-01
1.12173162e-01 2.69861460e-01 -1.17384791e+00 2.82875299e-02
5.80039203e-01 1.72209799e-01 -1.49000332e-01 7.39279687e-01
-2.30149448e-01 1.40803540e+00 -1.78320825e-01 -5.89989014e-02
3.75978589e-01 2.08441317e-01 9.69196081e-01 1.11781883e+00
5.33653736e-01 -5.44740371e-02 -3.04916054e-01 7.96158671e-01
-2.82617658e-01 -8.37822154e-04 -4.30816472e-01 -3.11260879e-01
2.97840059e-01 1.35250139e+00 -3.49929482e-01 -3.12258422e-01
-1.62661061e-01 1.14135051e+00 -2.34755188e-01 1.07194014e-01
-9.75790381e-01 -3.23377609e-01 7.03603923e-01 1.12864666e-01
5.74480832e-01 -5.83744764e-01 -1.72671646e-01 -8.24752510e-01
-2.52207518e-01 -1.23886323e+00 -1.77604437e-01 -8.84334564e-01
-8.47032428e-01 1.07132828e+00 -7.79345334e-01 -1.02057552e+00
-6.29602969e-01 -5.08768678e-01 -7.28936374e-01 7.80879438e-01
-1.41861701e+00 -1.14250720e+00 -1.96374238e-01 4.82445508e-01
1.19137275e+00 -6.61108911e-01 1.23302686e+00 1.86699152e-01
-3.34412336e-01 7.13079572e-01 1.73421457e-01 -1.82093725e-01
4.37648505e-01 -1.35736120e+00 7.53801346e-01 6.39882267e-01
8.56743276e-01 2.80687273e-01 8.30684960e-01 5.29645942e-02
-1.81278944e+00 -8.04055333e-01 6.01841629e-01 -1.07057326e-01
5.62970340e-01 -1.89017892e-01 -7.69077241e-01 8.37554932e-02
8.68147373e-01 -2.12301165e-01 4.80028152e-01 -1.93201825e-01
-2.46412992e-01 -1.76813334e-01 -1.05678654e+00 6.20371580e-01
5.43914974e-01 -7.13417888e-01 -3.44709873e-01 1.73320606e-01
8.02597702e-01 -4.91811901e-01 -8.61968398e-01 2.46135965e-02
5.74111342e-01 -1.08949006e+00 8.56120706e-01 -2.62021840e-01
5.01505673e-01 -2.75770485e-01 -1.57131970e-01 -1.23562944e+00
-6.43871129e-02 -1.14474809e+00 -5.38242102e-01 1.34556854e+00
2.07541943e-01 -3.58181179e-01 7.35196590e-01 -1.73879161e-01
-4.87838924e-01 -7.58246481e-01 -1.12447834e+00 -7.09721565e-01
5.57843149e-02 -5.82985997e-01 5.38515210e-01 4.54112589e-01
-4.27760065e-01 7.10690618e-01 -8.04472268e-01 9.29070115e-02
4.18033332e-01 7.63964579e-02 7.66266704e-01 -7.79894054e-01
-9.96030450e-01 -3.94715816e-01 -1.36911780e-01 -9.97095525e-01
3.90840732e-02 -5.15812278e-01 1.59593090e-01 -1.16957629e+00
-4.10025179e-01 1.08336039e-01 -7.18660504e-02 3.97289366e-01
3.70209932e-01 8.04808259e-01 3.75128657e-01 9.63808745e-02
-1.09801693e-02 6.97293580e-01 1.32517707e+00 4.68371548e-02
-3.23178560e-01 1.12681180e-01 -2.67410517e-01 9.10808563e-01
9.43663418e-01 -3.19968462e-01 -2.79182851e-01 -4.91615355e-01
-2.99629867e-01 4.69105870e-01 3.36369187e-01 -1.62194574e+00
2.41635218e-01 3.82757068e-01 1.41988367e-01 -4.42767411e-01
9.28106546e-01 -4.18001413e-01 2.35096425e-01 5.06817222e-01
-3.83907378e-01 -9.60514247e-02 2.67377377e-01 3.64761978e-01
-5.13233006e-01 1.18266167e-02 8.99999022e-01 -2.30701476e-01
-4.20380682e-01 -7.00175390e-02 -7.37908781e-01 -2.11613148e-01
4.66968805e-01 -7.48597011e-02 -1.69924051e-01 -7.31594205e-01
-6.98021710e-01 -5.59110463e-01 -8.35126936e-02 4.27744657e-01
7.12499201e-01 -1.12445426e+00 -7.65200913e-01 2.60319740e-01
-6.81150079e-01 -2.75357455e-01 3.90409917e-01 8.13296676e-01
-7.28684068e-01 6.15992188e-01 -2.10643813e-01 -4.66757745e-01
-1.44971514e+00 4.68890369e-01 3.34333003e-01 1.03120785e-02
-7.32710004e-01 8.70866477e-01 -4.48368043e-02 -1.65584952e-01
3.64708096e-01 -2.03914762e-01 2.71732011e-03 -1.59609288e-01
4.32929099e-01 7.27866650e-01 2.51484215e-01 -5.53010166e-01
-1.23575628e-01 3.97514343e-01 3.91013533e-01 -3.89339417e-01
1.24184787e+00 3.08167636e-01 5.07592559e-02 5.90130150e-01
9.56753910e-01 3.84262711e-01 -1.21079981e+00 1.63237095e-01
-4.74171996e-01 -1.62959322e-01 6.03237391e-01 -7.98244298e-01
-1.33147430e+00 1.25276315e+00 6.92206860e-01 1.05449051e-01
1.47798347e+00 -2.13792413e-01 1.29094231e+00 2.93421835e-01
-7.38602970e-03 -7.97778964e-01 1.91162258e-01 4.92308289e-01
1.19935822e+00 -6.26538754e-01 -4.87035632e-01 7.39565864e-03
-7.66372502e-01 1.12273407e+00 3.53464276e-01 -5.03304899e-01
2.78969735e-01 5.32597542e-01 2.43853122e-01 1.80707604e-01
-8.84528577e-01 -9.33502540e-02 3.28214586e-01 5.81157625e-01
5.36192834e-01 1.37815714e-01 -5.91287836e-02 5.51006615e-01
-7.88870096e-01 6.67994246e-02 3.91076922e-01 1.56403139e-01
-3.40737849e-01 -8.28843892e-01 -3.03242654e-01 4.00488153e-02
-7.49893665e-01 -4.24217522e-01 -3.81957829e-01 3.87208968e-01
-2.16685891e-01 1.12581205e+00 1.00517869e-01 -3.19223046e-01
1.01196565e-01 2.92138048e-02 5.17180145e-01 -3.95443290e-01
-1.13200843e+00 4.32817250e-01 1.25549480e-01 -2.40998030e-01
-1.28908128e-01 -2.13204846e-01 -1.18516648e+00 -2.08492428e-01
-2.00300917e-01 1.73542216e-01 9.32156563e-01 4.57312435e-01
3.14784467e-01 9.68289375e-01 6.88360393e-01 -8.97736192e-01
-7.00236440e-01 -1.06741238e+00 -6.55053556e-01 -4.03402776e-01
4.36503828e-01 -5.86325675e-02 -3.33882242e-01 1.84555855e-02] | [15.466376304626465, 6.138849258422852] |
e897c417-e987-4ffc-b864-33ccb3b66a77 | federated-variational-inference-towards | 2305.13672 | null | https://arxiv.org/abs/2305.13672v2 | https://arxiv.org/pdf/2305.13672v2.pdf | Federated Variational Inference: Towards Improved Personalization and Generalization | Conventional federated learning algorithms train a single global model by leveraging all participating clients' data. However, due to heterogeneity in client generative distributions and predictive models, these approaches may not appropriately approximate the predictive process, converge to an optimal state, or generalize to new clients. We study personalization and generalization in stateless cross-device federated learning setups assuming heterogeneity in client data distributions and predictive models. We first propose a hierarchical generative model and formalize it using Bayesian Inference. We then approximate this process using Variational Inference to train our model efficiently. We call this algorithm Federated Variational Inference (FedVI). We use PAC-Bayes analysis to provide generalization bounds for FedVI. We evaluate our model on FEMNIST and CIFAR-100 image classification and show that FedVI beats the state-of-the-art on both tasks. | ['Warren Richard Morningstar', 'Arash Afkanpour', 'Karan Singhal', 'Philip Andrew Mansfield', 'Joshua V. Dillon', 'Elahe Vedadi'] | 2023-05-23 | null | null | null | null | ['bayesian-inference', 'generalization-bounds'] | ['methodology', 'methodology'] | [-3.62430394e-01 5.84322810e-02 -4.79141057e-01 -7.02930093e-01
-1.33512616e+00 -6.66655898e-01 7.79703915e-01 -5.92960775e-01
-9.19091702e-02 8.06889713e-01 1.28638208e-01 -2.90970981e-01
-3.47671896e-01 -3.68506372e-01 -1.03231633e+00 -7.85290718e-01
-7.94417039e-02 1.19527447e+00 -1.20490417e-01 7.21525490e-01
-3.81736279e-01 2.81501979e-01 -1.31826770e+00 6.60446286e-01
5.61498880e-01 8.66865933e-01 -6.09077141e-02 9.11567330e-01
-3.50971520e-01 7.97357619e-01 -4.50954705e-01 -7.74302304e-01
1.22359984e-01 -3.48752558e-01 -9.46140826e-01 -8.64030123e-02
3.48927796e-01 -7.96316981e-01 -3.77367020e-01 8.83875668e-01
3.73320997e-01 2.48843674e-02 7.88744926e-01 -1.51328194e+00
-6.16464198e-01 1.13370752e+00 -2.26336703e-01 1.01662435e-01
-1.91847980e-01 1.55466571e-01 9.63408887e-01 -4.10708845e-01
5.35477221e-01 1.31801689e+00 6.12819493e-01 9.75015223e-01
-1.73132825e+00 -5.73519707e-01 4.98804748e-01 -1.03219189e-01
-1.19099033e+00 -5.22154629e-01 3.24244529e-01 -1.77590579e-01
8.58068407e-01 5.35977371e-02 1.87822789e-01 1.64493740e+00
-4.57377322e-02 1.29750013e+00 9.17323887e-01 -3.92518006e-02
8.10730159e-01 1.56107187e-01 1.42906815e-01 4.87587422e-01
2.49115884e-01 2.94909477e-02 -8.12717617e-01 -8.74354541e-01
4.60067302e-01 5.13192296e-01 -4.80127446e-02 -4.69901115e-01
-4.57260579e-01 7.40246415e-01 -1.65381655e-01 -2.38776192e-01
-7.11208344e-01 7.17586517e-01 3.52345049e-01 -1.22704534e-02
5.17189324e-01 -5.08076489e-01 -8.00133646e-01 -5.60336232e-01
-1.03352869e+00 3.39174896e-01 1.32580984e+00 1.08604109e+00
8.06327999e-01 -8.27855989e-02 -3.51411819e-01 7.60262966e-01
3.13321710e-01 6.55940115e-01 3.58707905e-01 -1.57168949e+00
3.01473707e-01 -1.25212064e-02 2.39867359e-01 4.44355011e-01
2.53520608e-01 -4.26892847e-01 -5.11140466e-01 -9.86785963e-02
3.13572854e-01 -3.47305268e-01 -8.64821255e-01 1.91502273e+00
3.20731640e-01 3.41011018e-01 4.00505662e-02 5.58150887e-01
-5.55273443e-02 5.95146120e-01 3.38535011e-01 -2.80128956e-01
8.07155013e-01 -8.55841875e-01 -4.57084477e-01 2.71429539e-01
4.04053360e-01 -1.14727721e-01 1.06465840e+00 8.18207502e-01
-1.11172473e+00 1.22915767e-01 -4.16211009e-01 2.27769271e-01
1.65327072e-01 -3.14511091e-01 8.46155763e-01 1.00369322e+00
-1.11493802e+00 7.97821403e-01 -1.58969438e+00 -3.00622463e-01
1.01677775e+00 3.04400116e-01 9.76559818e-02 -3.19571108e-01
-3.68256480e-01 1.20367475e-01 4.89451066e-02 -6.27355456e-01
-1.60181570e+00 -9.14500952e-01 5.25986888e-02 4.11281049e-01
1.06953785e-01 -1.24007261e+00 2.06853104e+00 -5.59348583e-01
-1.77264440e+00 3.70272219e-01 -3.38107854e-01 -6.44543707e-01
6.08454525e-01 -2.57467896e-01 -7.35822171e-02 -4.78735529e-02
-3.39670002e-01 1.29231825e-01 7.93586373e-01 -1.37294018e+00
-9.87758577e-01 -6.09465718e-01 -1.37171611e-01 -2.51097560e-01
-3.85641485e-01 -2.46393800e-01 -6.29589677e-01 -7.04464503e-03
-1.47438958e-01 -8.89294803e-01 -9.46726883e-04 -1.45418704e-01
-2.38534749e-01 -6.18855655e-01 9.19513524e-01 -2.18568474e-01
9.19641197e-01 -2.05375385e+00 -2.37744614e-01 1.52338535e-01
3.15580994e-01 -2.09332451e-01 1.29139647e-02 4.12764966e-01
6.70471728e-01 1.28317073e-01 2.99740672e-01 -1.19493520e+00
6.35687649e-01 8.22909534e-01 -6.03521109e-01 4.99433398e-01
-6.77823663e-01 7.75741518e-01 -5.37237048e-01 -6.06414378e-01
-1.39067844e-01 7.69896209e-01 -1.12089765e+00 2.84538746e-01
-9.58091497e-01 3.43498349e-01 -6.95133388e-01 5.14147460e-01
7.64094174e-01 -6.36855900e-01 5.93787849e-01 -3.87159325e-02
4.14547980e-01 2.94772536e-01 -1.03721023e+00 1.86247873e+00
-8.57314348e-01 3.66471484e-02 3.82820487e-01 -8.26288819e-01
3.53983313e-01 3.98811251e-01 7.72119224e-01 -1.03397146e-01
5.82289435e-02 2.93606855e-02 -7.54015207e-01 -2.78594166e-01
5.86748235e-02 -7.65840486e-02 8.09893683e-02 1.14737654e+00
6.18917942e-01 5.36828995e-01 -4.03832495e-01 4.83593434e-01
1.13567460e+00 3.47278863e-01 -4.65675443e-01 -8.86908770e-02
-3.27918828e-01 -4.05596763e-01 4.57335055e-01 1.42093325e+00
-1.06463097e-01 3.24807733e-01 4.73404646e-01 -4.18619603e-01
-9.33836043e-01 -1.70433986e+00 -1.01086413e-02 1.65729237e+00
-3.23222995e-01 -5.22287965e-01 -9.29987073e-01 -8.33047748e-01
2.32905194e-01 1.09155166e+00 -4.48552877e-01 2.11014431e-02
-6.94709048e-02 -8.25419605e-01 6.16339326e-01 5.14120877e-01
2.75376707e-01 -3.11211109e-01 -4.89239484e-01 3.11245441e-01
-8.63121524e-02 -8.99155617e-01 -3.76871079e-01 3.66562456e-01
-1.04213548e+00 -7.25336671e-01 -2.94802934e-01 3.82257216e-02
1.04432791e-01 -1.21412233e-01 1.07321775e+00 -5.71919024e-01
-5.61638251e-02 9.60886121e-01 -7.19585689e-03 -2.13888437e-01
-5.93164504e-01 3.30745608e-01 2.46186361e-01 1.14436261e-01
4.29316908e-01 -9.16119933e-01 -6.00938916e-01 -2.85831410e-02
-7.95688391e-01 -2.41099074e-01 2.96609014e-01 8.52440238e-01
7.47569621e-01 -1.85697868e-01 3.13972861e-01 -1.22814298e+00
6.18180156e-01 -7.53657043e-01 -6.78678632e-01 5.96276104e-01
-1.10468400e+00 5.45372367e-01 5.92377484e-01 -6.35506988e-01
-1.57860827e+00 6.09199442e-02 8.90313536e-02 -9.65249717e-01
-2.51280278e-01 1.03085972e-01 -1.99442416e-01 4.93762940e-01
7.53279865e-01 1.94042698e-01 -1.21891070e-02 -1.07385731e+00
6.75985932e-01 9.30987716e-01 7.41895974e-01 -1.42748630e+00
2.47604653e-01 8.04227710e-01 -2.31531233e-01 -7.87705034e-02
-8.33423793e-01 -1.40793249e-01 -7.83346444e-02 -2.97120195e-02
4.03667271e-01 -9.45181370e-01 -1.37523997e+00 3.38877648e-01
-1.07913256e+00 -7.15487123e-01 -5.95466077e-01 4.95998293e-01
-1.01924598e+00 4.90827449e-02 -9.14034486e-01 -1.27815402e+00
-7.77782023e-01 -8.93639863e-01 1.09453917e+00 1.54112443e-01
1.53521284e-01 -1.09427047e+00 5.12921095e-01 3.04445237e-01
7.61802912e-01 -3.04774106e-01 8.31971347e-01 -9.58025098e-01
-8.87105167e-01 -2.79443949e-01 1.15348749e-01 2.01781258e-01
-1.50324181e-01 3.15385349e-02 -1.33464348e+00 -3.38689506e-01
-1.98612794e-01 -4.44589257e-01 7.82257795e-01 4.04251546e-01
1.51637554e+00 -9.07270849e-01 -5.78771591e-01 9.53530967e-01
1.46908844e+00 -2.32727095e-01 1.54974639e-01 -1.63474753e-01
2.60545433e-01 -5.13965972e-02 -2.04782620e-01 1.14857960e+00
5.18721461e-01 4.28008199e-01 4.32646990e-01 6.81274593e-01
7.62786865e-02 -6.84695840e-01 4.81421620e-01 1.05378695e-01
1.19905919e-02 -4.49595600e-01 -7.45556831e-01 4.87849087e-01
-2.04934263e+00 -1.12216961e+00 3.24213177e-01 2.17203450e+00
8.63022029e-01 -2.42631420e-01 4.88205940e-01 -5.01215816e-01
4.87756103e-01 -3.76857191e-01 -1.11305463e+00 -3.33347738e-01
1.14799768e-01 5.08023262e-01 6.65070355e-01 2.82503128e-01
-6.09033644e-01 7.58400500e-01 6.72105455e+00 8.04742396e-01
-8.17345023e-01 9.99157727e-01 7.58311987e-01 -7.02339113e-01
-4.59534585e-01 9.72041413e-02 -9.46341038e-01 5.48033774e-01
1.40576172e+00 -1.71056882e-01 1.12667334e+00 1.39677656e+00
-2.30850503e-01 3.06094825e-01 -1.53610337e+00 1.09448707e+00
-2.88112223e-01 -1.53628552e+00 -3.09956763e-02 5.05835354e-01
8.91817510e-01 8.76674056e-01 2.94140965e-01 5.76152802e-01
1.48125815e+00 -7.29715049e-01 5.92791915e-01 7.38451242e-01
6.51123822e-01 -6.48597956e-01 2.27849290e-01 6.81387365e-01
-5.51968157e-01 -3.11596185e-01 -2.37063408e-01 4.42415714e-01
7.85258561e-02 2.53654271e-01 -5.41779876e-01 1.38424978e-01
1.01631594e+00 1.56514831e-02 -6.15050644e-02 8.91025305e-01
2.40788996e-01 1.20156789e+00 -8.67260754e-01 1.11630075e-01
-1.00652605e-01 6.79845065e-02 1.57498106e-01 9.40811157e-01
2.86839962e-01 -2.39404794e-02 3.98540199e-02 1.13789594e+00
-3.22973251e-01 -3.33606333e-01 1.91085495e-03 2.18606293e-02
6.51929617e-01 9.39665437e-01 -4.84756827e-02 -5.40268660e-01
-2.93300807e-01 1.20278323e+00 5.00616670e-01 7.64050841e-01
-8.95200491e-01 3.67062509e-01 1.11924267e+00 -4.40398008e-02
5.30515254e-01 6.30229563e-02 -6.37763590e-02 -1.47444797e+00
-2.48842061e-01 -6.37603402e-01 7.90797412e-01 -5.31447828e-01
-1.87428105e+00 5.49116433e-01 3.20484251e-01 -3.49516124e-01
-8.02277744e-01 -1.96172357e-01 -6.24828398e-01 8.20663989e-01
-1.00647008e+00 -1.34478545e+00 -4.99495454e-02 9.81784284e-01
2.49851361e-01 -2.01665252e-01 8.64480674e-01 2.29225472e-01
-6.34364128e-01 1.06408763e+00 8.38435888e-01 -2.91583478e-01
4.39731807e-01 -9.68461692e-01 2.23016635e-01 3.95889282e-01
3.22879910e-01 6.72781229e-01 7.38283873e-01 -4.25711334e-01
-1.76271820e+00 -1.21494341e+00 3.45775157e-01 -6.80088997e-01
5.57979941e-01 -3.34188670e-01 -7.53481567e-01 1.16576529e+00
1.57528073e-01 4.75340545e-01 7.58673906e-01 5.02700508e-01
-8.89899313e-01 -3.47409457e-01 -1.36049473e+00 2.35186994e-01
1.18208241e+00 -5.90825021e-01 6.07806817e-02 5.71846604e-01
7.19900727e-01 -1.89218894e-01 -8.92585039e-01 -2.53683925e-01
7.30688095e-01 -8.91935110e-01 6.27147615e-01 -1.24841571e+00
-1.91423833e-01 3.48125577e-01 -6.67061985e-01 -1.04534781e+00
-2.89594889e-01 -1.20991838e+00 -1.05148315e+00 1.46405351e+00
3.12883258e-01 -8.10221493e-01 1.16653621e+00 1.16844058e+00
1.50996983e-01 -5.58534563e-01 -1.20071495e+00 -1.05683875e+00
4.77752954e-01 -8.62816632e-01 1.03492284e+00 3.16052854e-01
4.01200466e-02 8.51377398e-02 -2.05198064e-01 2.32317165e-01
1.35286355e+00 1.99131772e-01 9.34341252e-01 -1.06467450e+00
-1.03313565e+00 -3.49051058e-01 1.73616052e-01 -1.23283267e+00
4.98002440e-01 -7.18804836e-01 -2.22006530e-01 -1.27029908e+00
7.60011017e-01 -7.62852848e-01 -3.17087680e-01 7.26418138e-01
2.76307166e-01 -2.55843371e-01 6.93022609e-02 7.13214099e-01
-1.13642836e+00 5.87060809e-01 2.87119120e-01 1.36729879e-02
-2.11930335e-01 3.94813180e-01 -6.10070050e-01 1.24759145e-01
7.52658069e-01 -4.84000087e-01 -5.65522015e-01 -5.12469530e-01
-1.38083138e-02 1.60985813e-01 5.53723931e-01 -6.40084624e-01
4.76309806e-01 -3.58997732e-01 1.84026793e-01 -4.46938306e-01
3.99505705e-01 -7.33920276e-01 5.40900111e-01 1.25787720e-01
-4.89835292e-01 -4.53777492e-01 -1.51480392e-01 9.18972194e-01
4.82074291e-01 1.94623128e-01 5.76220572e-01 9.85076800e-02
1.57282427e-01 7.73802817e-01 -3.74145210e-01 1.89281836e-01
7.31134355e-01 2.37893820e-01 -4.69758868e-01 -6.94719791e-01
-9.11508799e-01 1.06786229e-01 5.42238533e-01 -2.86699031e-02
2.46410564e-01 -1.04233801e+00 -4.53091830e-01 2.35223584e-02
-8.11370164e-02 -5.63240275e-02 4.33578759e-01 6.75648272e-01
5.89184416e-03 3.42987448e-01 4.66539025e-01 -7.84165144e-01
-7.42204964e-01 7.31504381e-01 5.71808994e-01 -2.40307212e-01
-4.89940464e-01 7.33413219e-01 6.59225062e-02 -3.08904618e-01
5.96138239e-01 -4.43505198e-02 7.84165502e-01 -6.77814364e-01
4.11855727e-01 4.77032423e-01 -5.00869416e-02 4.70306026e-03
-2.67165303e-01 -2.93895900e-01 -1.50853157e-01 -4.78858739e-01
1.37348878e+00 -1.84021175e-01 2.30915755e-01 3.39134157e-01
1.25162971e+00 -3.35974216e-01 -1.91706562e+00 -7.04085171e-01
-2.98993498e-01 -4.07538712e-01 2.05342695e-01 -9.30462420e-01
-1.19505537e+00 7.48395622e-01 8.08754504e-01 2.40927134e-02
8.04416895e-01 6.34644389e-01 7.20834613e-01 4.29062724e-01
8.24243784e-01 -8.39660048e-01 -2.06546232e-01 1.15706928e-01
1.18822850e-01 -8.58433545e-01 -2.72350430e-01 2.54147053e-01
-5.30706465e-01 6.69741213e-01 3.41482848e-01 1.82998255e-01
1.00898206e+00 5.37129939e-01 -2.25726575e-01 1.82153825e-02
-1.58280933e+00 1.77717835e-01 -3.03865165e-01 6.89706445e-01
1.48449495e-01 2.26886258e-01 4.33385819e-01 1.18451047e+00
5.52257709e-02 4.68108684e-01 2.98182853e-02 8.45005453e-01
-1.21887781e-01 -1.20783114e+00 -2.27467120e-01 5.66494763e-01
-7.80927896e-01 1.21001109e-01 1.75348192e-01 1.67009667e-01
-1.78955898e-01 8.86365414e-01 2.56935418e-01 -3.79086524e-01
-2.19881237e-01 5.40090442e-01 7.55790889e-01 -3.95567209e-01
-2.90353566e-01 2.39563003e-01 -2.23005652e-01 -6.91435158e-01
-6.34469017e-02 -1.02685654e+00 -1.10023594e+00 -8.42305899e-01
-1.59833238e-01 2.73110688e-01 1.04534411e+00 8.71129692e-01
9.41533089e-01 1.11075655e-01 6.12420201e-01 -5.99282384e-01
-1.53301704e+00 -7.88527906e-01 -7.10646927e-01 1.93175599e-01
8.20313841e-02 -3.02044809e-01 -4.28019047e-01 2.60364830e-01] | [5.838552474975586, 6.279998302459717] |
e27ffd3e-d2ef-4d59-9829-b2ed0efc90fb | image-augmentation-improves-few-shot | 2208.12613 | null | https://arxiv.org/abs/2208.12613v1 | https://arxiv.org/pdf/2208.12613v1.pdf | Image augmentation improves few-shot classification performance in plant disease recognition | With the world population projected to near 10 billion by 2050, minimizing crop damage and guaranteeing food security has never been more important. Machine learning has been proposed as a solution to quickly and efficiently identify diseases in crops. Convolutional Neural Networks typically require large datasets of annotated data which are not available on demand. Collecting this data is a long and arduous process which involves manually picking, imaging, and annotating each individual leaf. I tackle the problem of plant image data scarcity by exploring the efficacy of various data augmentation techniques when used in conjunction with transfer learning. I evaluate the impact of various data augmentation techniques both individually and combined on the performance of a ResNet. I propose an augmentation scheme utilizing a sequence of different augmentations which consistently improves accuracy through many trials. Using only 10 total seed images, I demonstrate that my augmentation framework can increase model accuracy by upwards of 25\%. | ['Frank Xiao'] | 2022-08-25 | null | null | null | null | ['image-augmentation'] | ['computer-vision'] | [ 4.53342468e-01 5.86658120e-02 -2.52009183e-01 -9.36019495e-02
-2.76148766e-01 -9.43046212e-01 4.71290261e-01 6.00609541e-01
-3.49468708e-01 5.44020176e-01 -1.30591229e-01 -6.01712644e-01
2.68798351e-01 -9.40901756e-01 -6.77616298e-01 -4.12717223e-01
7.81482384e-02 2.08176166e-01 7.77589753e-02 -3.78496975e-01
-6.23819605e-02 6.28833532e-01 -1.39588380e+00 -4.26562205e-02
9.64477956e-01 9.35212433e-01 4.45059031e-01 7.01367259e-01
-2.72245139e-01 4.14512366e-01 -7.14952111e-01 -1.22451298e-01
4.16759223e-01 -1.67058259e-01 -6.76151633e-01 3.08043808e-01
4.19105828e-01 -4.37374532e-01 1.14224829e-01 8.61268461e-01
3.72276127e-01 -3.23663443e-01 2.37244442e-01 -1.11654544e+00
-7.64850795e-01 5.76865315e-01 -1.02026987e+00 -1.05187306e-02
-1.43327475e-01 1.25268772e-01 7.22014368e-01 -5.32158732e-01
3.88833731e-01 6.22542799e-01 9.66285706e-01 1.35166436e-01
-1.48027170e+00 -5.28590381e-01 -4.11442108e-02 -2.89058864e-01
-1.09413576e+00 -2.28681430e-01 3.11398506e-01 -2.80710101e-01
7.65519977e-01 1.48190662e-01 6.72243655e-01 5.32673895e-01
-7.86533505e-02 6.12988114e-01 6.19031608e-01 -4.71619785e-01
-7.14657782e-03 -2.82424479e-03 5.44931227e-03 7.62312829e-01
5.03089011e-01 -1.53740272e-01 -1.51238427e-01 -6.68137595e-02
7.76673198e-01 8.32644850e-02 -7.05966502e-02 -2.89287239e-01
-1.11948848e+00 7.20147729e-01 8.40592384e-01 2.17308730e-01
-5.38475692e-01 -6.40022159e-02 3.60003412e-01 -1.90669037e-02
4.93022382e-01 9.64635015e-01 -9.38076258e-01 2.27258444e-01
-1.02403629e+00 1.57726288e-01 6.62078917e-01 8.36812139e-01
5.45673728e-01 1.75364837e-01 1.81312695e-01 7.14303017e-01
-3.09988379e-01 4.57109183e-01 1.56662688e-01 -8.01892102e-01
2.45024730e-02 9.91040826e-01 3.37471783e-01 -9.11538124e-01
-4.47104186e-01 -4.17385936e-01 -1.03620076e+00 1.46031275e-01
6.71892345e-01 -3.37751091e-01 -1.34074163e+00 1.83313668e+00
2.43794456e-01 7.58431256e-02 -1.55211404e-01 4.01305050e-01
5.72822034e-01 7.12166607e-01 6.78238571e-01 6.54404610e-02
1.22862411e+00 -6.60137534e-01 -5.59817195e-01 -3.01345974e-01
8.41108382e-01 -8.26068640e-01 1.01769376e+00 3.65563452e-01
-9.14746761e-01 -4.79134381e-01 -1.19393659e+00 8.32876414e-02
-7.64447510e-01 5.48254490e-01 9.86411870e-01 6.53087378e-01
-8.75558913e-01 6.40904248e-01 -7.98187196e-01 -5.80756664e-01
7.96877384e-01 4.52776670e-01 -6.84893310e-01 -2.01990113e-01
-5.36079049e-01 9.20798182e-01 5.71659625e-01 1.59041598e-01
-7.55104244e-01 -9.59014058e-01 -6.56142771e-01 2.53825396e-01
4.91440833e-01 -2.42284313e-01 1.07400346e+00 -7.49993563e-01
-8.71757269e-01 7.83668697e-01 3.58838700e-02 -5.38399398e-01
-7.37807974e-02 -5.14598906e-01 -1.71788812e-01 -2.31004506e-01
7.94774815e-02 1.16979992e+00 5.01832426e-01 -1.07438397e+00
-6.38286650e-01 -5.39727092e-01 -7.72330686e-02 5.04144607e-03
-8.93618464e-01 1.00357704e-01 8.07608813e-02 -6.39679790e-01
1.30647421e-01 -1.11347830e+00 -4.78735387e-01 3.47089648e-01
-2.89352477e-01 3.01996052e-01 1.16569233e+00 -1.05979812e+00
7.13584304e-01 -1.82452762e+00 -9.66412723e-02 -1.92014888e-01
9.47561786e-02 8.89497995e-01 -5.16306579e-01 5.44007681e-02
-1.98370293e-01 4.62793171e-01 -4.72771525e-01 1.68345213e-01
-5.13451815e-01 1.60791099e-01 -1.21025868e-01 2.89469566e-02
8.73769104e-01 1.05859745e+00 -8.07600915e-01 -7.28437826e-02
2.80278027e-01 6.86248779e-01 -2.00027049e-01 2.28508443e-01
-3.99158150e-01 4.00318861e-01 -1.56123951e-01 1.01410341e+00
6.91608250e-01 -5.45580566e-01 3.33678246e-01 -2.24769428e-01
-1.96178928e-01 -1.61775708e-01 -7.45837927e-01 1.52405763e+00
-2.84820944e-01 5.31605780e-01 7.49329776e-02 -9.77550566e-01
9.50311005e-01 3.37688833e-01 6.10101819e-01 -3.34467232e-01
1.07594945e-01 9.04089399e-03 2.98152298e-01 -8.00795630e-02
4.56551820e-01 4.64225084e-01 2.31861576e-01 2.57374436e-01
1.82105005e-01 -9.33113918e-02 3.75278771e-01 8.64127353e-02
1.35096073e+00 3.66762698e-01 3.56812716e-01 -4.05343980e-01
8.21129531e-02 5.43174148e-01 5.22230446e-01 3.94096464e-01
-1.58079863e-01 2.66355991e-01 3.40214998e-01 -6.96694195e-01
-1.42023861e+00 -7.50243306e-01 1.07278386e-02 1.17205775e+00
-2.81911075e-01 -9.34368521e-02 -6.76868796e-01 -5.30167580e-01
8.30320343e-02 4.72217768e-01 -6.29872203e-01 -2.45385662e-01
-2.74329454e-01 -1.42911410e+00 6.49692953e-01 7.43566155e-01
8.44914913e-01 -1.08020377e+00 -8.60471070e-01 1.27454653e-01
-8.46409425e-03 -1.37138128e+00 1.53798193e-01 7.17657804e-01
-8.30778003e-01 -8.83233368e-01 -7.92717218e-01 -7.95843840e-01
8.58583391e-01 3.82391483e-01 1.27225709e+00 2.70858943e-01
-8.94279480e-01 -2.52829313e-01 -4.99776453e-01 -8.72867644e-01
-3.34994733e-01 5.93222976e-01 -2.74276048e-01 -6.90341353e-01
3.35122168e-01 -5.10632098e-01 -4.60177600e-01 -8.53397548e-02
-9.73088264e-01 1.13144390e-01 7.67435193e-01 7.88784623e-01
3.36013287e-01 9.97109190e-02 6.23141170e-01 -8.85613680e-01
3.61969829e-01 -4.87288117e-01 -8.02274227e-01 5.40979207e-01
-4.44401175e-01 -2.27980718e-01 4.12367195e-01 -4.68897939e-01
-8.60612929e-01 6.77962780e-01 1.40715867e-01 3.65243964e-02
-3.77513498e-01 6.20226741e-01 -1.23351395e-01 -3.43404859e-01
7.20303416e-01 -3.15652341e-01 1.78617492e-01 -4.01928961e-01
4.32662249e-01 2.90124297e-01 6.41044140e-01 -2.60070205e-01
9.46046948e-01 1.74380168e-01 1.40378341e-01 -1.08061075e+00
-1.01286852e+00 -1.59910202e-01 -8.77765596e-01 -6.92703053e-02
7.08717942e-01 -8.76042366e-01 -5.65050244e-01 6.80279136e-01
-1.02601063e+00 -5.24286449e-01 -3.67510915e-01 2.40069076e-01
1.80483967e-01 7.45323719e-03 -4.51502979e-01 -5.01089871e-01
-4.52826083e-01 -8.37635994e-01 7.46010840e-01 4.50409234e-01
-2.90754326e-02 -6.30495608e-01 -2.88816124e-01 5.12916259e-02
6.13199294e-01 5.45733452e-01 8.13644946e-01 -4.75178689e-01
-2.94305980e-01 -5.77726603e-01 -7.03798711e-01 2.70993233e-01
6.76251531e-01 2.03964919e-01 -1.03570712e+00 -1.24857731e-01
-5.36105514e-01 -6.03257656e-01 7.35142529e-01 5.55126667e-01
1.33285773e+00 2.88930740e-02 -3.03570122e-01 3.89386803e-01
1.50164270e+00 1.17735505e-01 5.80521405e-01 2.49100775e-01
6.95807755e-01 6.56028569e-01 6.53688610e-01 3.35817814e-01
-3.66150588e-02 1.12979934e-01 7.85163522e-01 -5.14979661e-01
1.36136830e-01 1.32399097e-01 -3.36651623e-01 1.69673458e-01
-8.26704502e-03 -5.96097767e-01 -1.22013485e+00 8.49944413e-01
-1.46531022e+00 -5.82340181e-01 2.69759417e-01 2.10101557e+00
7.92874336e-01 1.07721120e-01 -3.94045189e-02 3.79130989e-01
4.88259554e-01 -2.44030803e-01 -7.76041389e-01 -1.79197699e-01
-2.35567093e-01 5.00179946e-01 1.01676869e+00 1.09102219e-01
-1.32266736e+00 1.17194736e+00 6.69595385e+00 1.93087041e-01
-1.15912223e+00 -2.73946732e-01 9.95252073e-01 2.51652896e-01
1.35672912e-01 -1.27990535e-02 -6.04948938e-01 4.17346545e-02
9.34625566e-01 2.28463709e-01 3.12771171e-01 8.11207414e-01
-1.14462590e-02 -3.24981034e-01 -5.38636923e-01 2.45418936e-01
-1.58942178e-01 -1.30458128e+00 -1.38872221e-01 1.37077868e-01
7.32040226e-01 1.96329560e-02 -3.87598611e-02 -1.72143045e-03
8.47628534e-01 -1.15773439e+00 -6.41916180e-03 1.09783746e-01
7.79998183e-01 -7.11858809e-01 6.99781358e-01 2.82156944e-01
-1.23525727e+00 -1.91694126e-01 -1.29524574e-01 -5.66001050e-02
-1.06174953e-01 6.02964878e-01 -1.06558132e+00 2.79526681e-01
8.14839661e-01 5.07961512e-01 -9.20974731e-01 9.38561678e-01
1.22126902e-03 7.24695504e-01 -6.20607972e-01 2.44282424e-01
1.81048840e-01 5.65859750e-02 1.15275579e-02 7.76585162e-01
3.41625631e-01 5.20406254e-02 3.37170631e-01 7.14114189e-01
-1.14272416e-01 -6.30851313e-02 -7.06141829e-01 -5.87805152e-01
5.35335541e-01 1.57326198e+00 -1.12898254e+00 -4.49179746e-02
-2.24449009e-01 9.69028473e-01 5.36408015e-02 -1.61115706e-01
-5.73683262e-01 -3.54970664e-01 5.20378709e-01 3.87156405e-03
2.21415117e-01 -3.83865565e-01 -2.92461008e-01 -5.20664990e-01
-2.41996601e-01 -8.32418680e-01 -7.15944320e-02 -7.71739423e-01
-1.01330304e+00 4.11098212e-01 -2.51331240e-01 -7.09350705e-01
-4.96889427e-02 -7.30391026e-01 -3.40971470e-01 9.98710215e-01
-1.40045369e+00 -1.59665072e+00 -8.56944203e-01 -1.00630738e-01
3.45192224e-01 -1.61281034e-01 1.32088578e+00 4.00213361e-01
-7.09418654e-01 3.24496001e-01 -8.58399048e-02 1.76767260e-01
5.10571003e-01 -1.03459072e+00 8.37332308e-01 8.38489890e-01
5.99687472e-02 2.72052556e-01 5.29368281e-01 -7.32556820e-01
-1.11954415e+00 -1.18742704e+00 6.82437837e-01 -5.76169230e-02
5.68624914e-01 -9.14791524e-02 -1.00496721e+00 6.37911975e-01
2.03768224e-01 9.29138586e-02 5.20444095e-01 2.94197053e-02
-2.90097445e-01 -7.25939721e-02 -1.28583694e+00 3.57656211e-01
6.41807795e-01 -1.72134057e-01 2.66885221e-01 2.96717316e-01
9.33044970e-01 -1.56644568e-01 -9.88470912e-01 9.17496383e-01
4.61572170e-01 -2.78091908e-01 9.65136409e-01 -7.64414132e-01
4.57813263e-01 -1.53638139e-01 -2.36077458e-01 -1.28276396e+00
-3.13767463e-01 -2.89058357e-01 1.22797817e-01 1.38834739e+00
4.51124489e-01 -2.59542882e-01 9.98770118e-01 6.06427610e-01
1.71267480e-01 -5.28848708e-01 3.46755348e-02 -5.78354895e-01
6.51110560e-02 1.67449981e-01 7.39267290e-01 1.16781867e+00
-4.69016433e-01 1.56539276e-01 -2.91676253e-01 2.66484529e-01
3.34234744e-01 -2.08232462e-01 6.52229249e-01 -1.49959207e+00
2.28873938e-01 -1.75827026e-01 -3.02937537e-01 -2.89043635e-01
-1.67576093e-02 -6.50386155e-01 1.51723593e-01 -1.39165425e+00
2.03927606e-01 -7.01559782e-01 -2.41696313e-01 1.05834448e+00
-4.35182989e-01 3.70674402e-01 2.08623722e-01 -9.76561457e-02
2.09928006e-01 -8.28469079e-03 9.07105327e-01 -5.02413064e-02
-1.85997307e-01 -2.30655029e-01 -8.43692601e-01 7.60122001e-01
1.48042047e+00 -1.78554073e-01 -4.52787399e-01 -7.18852758e-01
1.03495404e-01 -3.69940490e-01 5.79854727e-01 -1.18691361e+00
-2.85837263e-01 -9.86564830e-02 8.59142065e-01 -5.26961923e-01
4.06681657e-01 -8.84536922e-01 2.83889249e-02 6.36554599e-01
-3.67789805e-01 2.96191245e-01 8.05280924e-01 1.94919452e-01
1.70484036e-01 -2.75785714e-01 7.32246935e-01 -4.05191302e-01
-8.18414986e-01 2.62800246e-01 -2.27288399e-02 -2.70103842e-01
1.12628174e+00 1.56592175e-01 -3.22475016e-01 -1.00368366e-01
-5.80204070e-01 2.54203320e-01 3.31184506e-01 4.98212218e-01
9.44265798e-02 -7.89688110e-01 -7.36821473e-01 3.20272855e-02
-5.67239746e-02 -1.79953966e-02 -2.37259999e-01 1.99274197e-01
-7.97092795e-01 2.72064954e-01 -8.54869604e-01 -5.19641876e-01
-1.33237350e+00 4.68050033e-01 1.69416871e-02 -3.79170567e-01
-3.24169755e-01 9.90992725e-01 -2.04164267e-01 -4.78858620e-01
3.34106863e-01 -1.81866631e-01 -3.37792486e-01 8.80406201e-02
4.30681556e-01 1.74756050e-01 1.56213969e-01 -2.45988309e-01
-2.54237391e-02 1.79430451e-02 -1.59559891e-01 9.15516764e-02
1.62327886e+00 3.78691435e-01 -1.20427661e-01 1.46900848e-01
5.91238558e-01 -6.39251709e-01 -1.10524154e+00 -2.21081357e-02
3.61721545e-01 -2.89207637e-01 1.84942305e-01 -1.16438699e+00
-1.18978024e+00 7.12695122e-01 9.16277051e-01 4.26459551e-01
1.42316866e+00 -4.36865360e-01 8.07461739e-01 4.94811535e-01
5.23924977e-02 -6.86663687e-01 -1.45005301e-01 1.45424202e-01
5.01601934e-01 -1.69794619e+00 1.97813794e-01 -4.98231709e-01
-3.96323144e-01 7.27771342e-01 8.85212779e-01 -8.06736499e-02
7.56196439e-01 7.52970040e-01 -4.50188434e-03 5.10892123e-02
-5.82962275e-01 -4.13706332e-01 -1.53778315e-01 7.42498636e-01
7.93758988e-01 1.66980460e-01 3.95036861e-02 1.95272733e-02
1.62348345e-01 1.45119458e-01 2.91620135e-01 1.32005167e+00
-5.46160996e-01 -1.33240747e+00 -3.12274247e-01 8.07415605e-01
-5.58250427e-01 -2.00050056e-01 -5.37305593e-01 7.04376996e-01
2.12130293e-01 7.31698751e-01 4.35752384e-02 -2.00129360e-01
1.22470828e-02 9.63359177e-02 3.00391614e-01 -5.88166654e-01
-6.19205117e-01 -1.98966563e-01 -1.50433883e-01 -2.05918297e-01
-5.28908134e-01 -5.51839888e-01 -9.25432146e-01 -4.23538297e-01
-6.02105856e-01 -4.26969260e-01 1.05782318e+00 7.10062385e-01
4.68090355e-01 6.89492285e-01 3.52682710e-01 -8.11563790e-01
-4.52487111e-01 -9.96426046e-01 -1.56106234e-01 1.71275452e-01
1.34605661e-01 -3.88026714e-01 2.81273365e-01 3.81402850e-01] | [9.130762100219727, -1.549551010131836] |
29c912a5-342b-451c-889a-7840ec1f85af | numglue-a-suite-of-fundamental-yet | 2204.05660 | null | https://arxiv.org/abs/2204.05660v1 | https://arxiv.org/pdf/2204.05660v1.pdf | NumGLUE: A Suite of Fundamental yet Challenging Mathematical Reasoning Tasks | Given the ubiquitous nature of numbers in text, reasoning with numbers to perform simple calculations is an important skill of AI systems. While many datasets and models have been developed to this end, state-of-the-art AI systems are brittle; failing to perform the underlying mathematical reasoning when they appear in a slightly different scenario. Drawing inspiration from GLUE that was proposed in the context of natural language understanding, we propose NumGLUE, a multi-task benchmark that evaluates the performance of AI systems on eight different tasks, that at their core require simple arithmetic understanding. We show that this benchmark is far from being solved with neural models including state-of-the-art large-scale language models performing significantly worse than humans (lower by 46.4%). Further, NumGLUE promotes sharing knowledge across tasks, especially those with limited training data as evidenced by the superior performance (average gain of 3.4% on each task) when a model is jointly trained on all the tasks as opposed to task-specific modeling. Finally, we hope that NumGLUE will encourage systems that perform robust and general arithmetic reasoning within language, a first step towards being able to perform more complex mathematical reasoning. | ['Ashwin Kalyan', 'Chitta Baral', 'Peter Clark', 'Bhavdeep Sachdeva', 'Neeraj Varshney', 'Arindam Mitra', 'Swaroop Mishra'] | 2022-04-12 | null | https://aclanthology.org/2022.acl-long.246 | https://aclanthology.org/2022.acl-long.246.pdf | acl-2022-5 | ['mathematical-reasoning', 'arithmetic-reasoning'] | ['natural-language-processing', 'reasoning'] | [-5.30892909e-02 1.63725287e-01 1.97807342e-01 -3.81854683e-01
-4.20032293e-01 -5.48282206e-01 8.72705877e-01 4.27276373e-01
-5.66107273e-01 4.34727579e-01 2.90734887e-01 -5.75112760e-01
-1.94156244e-01 -1.00995040e+00 -7.25171506e-01 -2.58728601e-02
-3.66674960e-02 8.31183851e-01 -2.93300506e-02 -7.90810347e-01
5.67629218e-01 2.38764197e-01 -1.18860900e+00 7.12216079e-01
1.11127496e+00 6.81130528e-01 -1.82423949e-01 8.19663167e-01
-2.87868023e-01 1.64177811e+00 -7.22495675e-01 -8.15903723e-01
1.84279919e-01 -1.78220510e-01 -1.04889679e+00 -7.69709110e-01
7.97625303e-01 -3.83695602e-01 -2.35866919e-01 7.77467847e-01
6.64663091e-02 9.92296487e-02 7.08462238e-01 -1.37134755e+00
-1.09295285e+00 1.01805377e+00 -1.04318947e-01 1.56788737e-01
4.38710093e-01 4.40398633e-01 1.05945671e+00 -6.79801762e-01
1.22774035e-01 1.43343282e+00 8.72111440e-01 4.34958339e-01
-1.13295579e+00 -6.99092984e-01 5.51369525e-02 1.66436195e-01
-1.10922503e+00 -4.13005650e-01 3.45911741e-01 -3.32448691e-01
1.49752235e+00 6.95770010e-02 6.78641796e-01 6.09566689e-01
3.94491434e-01 9.24701631e-01 1.11491203e+00 -3.51535738e-01
1.73461750e-01 -2.83759534e-01 3.72091025e-01 8.23391616e-01
4.14833367e-01 -4.25023258e-01 -6.51621521e-01 1.05805136e-01
7.53564835e-01 -3.17109466e-01 1.48202658e-01 -1.87993690e-01
-1.65684760e+00 5.85598409e-01 5.75250626e-01 5.31225622e-01
-4.14393812e-01 6.12272501e-01 5.47450066e-01 6.42357349e-01
2.19809726e-01 1.09489930e+00 -5.60469389e-01 -3.17896485e-01
-9.50462580e-01 7.67453015e-01 1.27337193e+00 7.34375417e-01
2.33345047e-01 9.41560045e-02 -1.81501299e-01 5.24621725e-01
2.16474067e-02 4.08806920e-01 3.71365696e-01 -1.30964530e+00
7.34066784e-01 8.34709764e-01 -3.58644873e-02 -9.37484443e-01
-6.31341696e-01 -4.28619951e-01 -7.05986202e-01 3.14410001e-01
1.08635640e+00 1.59918074e-03 -6.71256840e-01 1.88235152e+00
-3.57430398e-01 -6.37754202e-02 3.66987765e-01 6.16721869e-01
7.47304678e-01 7.62970746e-01 4.05397892e-01 4.88323808e-01
1.36826873e+00 -1.14592111e+00 -4.68917221e-01 -7.03840852e-01
1.12246418e+00 -6.35118604e-01 1.23629439e+00 8.14927936e-01
-1.52658153e+00 -5.69109380e-01 -1.13479125e+00 -6.82939231e-01
-7.32640147e-01 -7.66022652e-02 1.29215467e+00 6.25303507e-01
-1.34311187e+00 4.41479266e-01 -5.75083256e-01 -2.04127133e-01
4.77616757e-01 2.59361178e-01 -2.33976528e-01 -3.42409879e-01
-1.20436776e+00 1.72316015e+00 4.75002527e-01 -4.69078869e-02
-5.36023498e-01 -1.23293090e+00 -7.50894010e-01 3.24922562e-01
4.54866469e-01 -1.00884080e+00 1.46076047e+00 -6.76762342e-01
-1.30844879e+00 9.37113881e-01 5.51184304e-02 -8.39900494e-01
6.33929610e-01 -5.68276823e-01 -1.24590516e-01 -8.70845020e-02
-7.77732860e-03 7.88451314e-01 2.32870162e-01 -9.11782563e-01
-2.83361733e-01 -2.50516593e-01 5.10048330e-01 2.08598033e-01
-5.82328178e-02 -4.78260554e-02 -1.65288895e-02 -6.07297063e-01
-8.11753944e-02 -6.13730788e-01 1.20551303e-01 2.85968810e-01
1.92172706e-01 -7.83253431e-01 1.91456839e-01 -7.10701108e-01
1.11620843e+00 -1.75625789e+00 2.28919178e-01 -1.33264318e-01
5.44877112e-01 3.75414431e-01 -2.67190516e-01 4.32007700e-01
-1.06103636e-01 2.36995265e-01 -2.21015051e-01 -2.09261537e-01
5.52900791e-01 1.76446401e-02 -4.13521379e-01 -7.56288646e-03
3.45304489e-01 1.42262971e+00 -1.04673564e+00 -2.99703956e-01
2.44935721e-01 6.05004057e-02 -7.00307131e-01 -6.68994486e-02
-5.10809958e-01 -2.57948488e-01 -1.34415224e-01 4.27615792e-01
4.51283962e-01 -4.74616557e-01 3.89762409e-02 7.67084733e-02
1.09654091e-01 5.51686943e-01 -9.81912315e-01 2.02481961e+00
-7.31134593e-01 6.87257648e-01 -2.07402438e-01 -1.09819305e+00
6.36048496e-01 1.57820314e-01 -4.92225140e-02 -8.75006080e-01
3.20782326e-02 4.05883580e-01 8.89156401e-01 -1.64885148e-01
5.22036791e-01 -6.85735494e-02 -2.86874980e-01 7.42829204e-01
-1.30345756e-02 -1.02324629e+00 5.47859848e-01 5.13945758e-01
1.22288597e+00 9.80876610e-02 4.89083380e-01 -4.95614409e-01
5.83369970e-01 1.68336228e-01 -2.43717596e-01 1.03858578e+00
-1.83395132e-01 7.42072612e-02 5.01000702e-01 -8.91841769e-01
-1.27878749e+00 -1.08700323e+00 6.51400462e-02 1.35828066e+00
-4.83191639e-01 -5.82009196e-01 -7.66296744e-01 -1.15003094e-01
2.39187941e-01 1.18916416e+00 -5.78268349e-01 -2.76090950e-01
-8.10224235e-01 -4.63848710e-01 1.04296780e+00 6.63955450e-01
9.01385427e-01 -1.17557180e+00 -8.99695814e-01 1.97811037e-01
-8.95242672e-03 -1.09999561e+00 1.70371890e-01 6.81012273e-02
-7.69543767e-01 -8.97016406e-01 -6.67915821e-01 -5.84354579e-01
3.14700961e-01 2.74507962e-02 1.81246305e+00 5.71307182e-01
-1.18697263e-01 4.02344853e-01 -6.58190995e-03 -7.47391582e-01
-5.64572036e-01 1.03188522e-01 -2.23186031e-01 -8.38330090e-01
6.42815828e-01 -4.05018687e-01 -1.01960890e-01 -2.02884719e-01
-7.37398922e-01 4.88925606e-01 6.98002219e-01 5.75640798e-01
-4.14449126e-01 -7.88634345e-02 6.06299460e-01 -7.38907576e-01
9.53342438e-01 -4.71253514e-01 -2.91883975e-01 4.69153881e-01
-3.50096315e-01 2.15652525e-01 8.20627153e-01 -2.97570586e-01
-7.92972028e-01 -5.99877656e-01 2.84475148e-01 5.70948012e-02
3.90976202e-03 5.85149586e-01 4.14156020e-01 4.62291129e-02
8.16185296e-01 2.64359981e-01 1.11077383e-01 5.60576096e-02
5.76560140e-01 2.07966566e-01 6.34163320e-01 -1.29541063e+00
7.42465496e-01 7.69984797e-02 1.68430015e-01 -6.56972587e-01
-9.62494791e-01 5.72116971e-02 -4.76923168e-01 7.44840056e-02
7.57938802e-01 -1.04604518e+00 -1.23371780e+00 6.78934753e-01
-1.43076718e+00 -9.01902139e-01 -4.14056517e-02 3.38647246e-01
-7.12665200e-01 1.47273913e-01 -7.21530795e-01 -7.84982145e-01
-2.91034669e-01 -1.01457131e+00 7.52929568e-01 2.79349321e-03
-8.11726272e-01 -1.19831693e+00 -2.88484871e-01 6.50551915e-01
8.39696407e-01 -1.00891471e-01 1.57294178e+00 -7.84170985e-01
-6.60099983e-01 -8.34284723e-02 -6.01353705e-01 3.59658867e-01
-2.66649038e-01 -1.65844232e-01 -6.51421070e-01 2.39460453e-01
-1.50622549e-02 -9.35658455e-01 8.33494306e-01 1.24118432e-01
1.21706581e+00 7.93077517e-03 8.67210850e-02 2.54671276e-01
9.88803923e-01 7.59113207e-02 8.25462997e-01 4.09041792e-01
5.30322134e-01 4.65543211e-01 2.37338528e-01 1.23468451e-01
9.85121608e-01 3.75991702e-01 1.21398322e-01 1.37244642e-01
-1.77003890e-01 -9.22491327e-02 2.49096394e-01 7.01016009e-01
-2.82553583e-01 -8.27463791e-02 -1.62586462e+00 4.56742644e-01
-1.84993756e+00 -8.66221547e-01 -6.98101297e-02 1.89603209e+00
1.26877964e+00 5.12212157e-01 -1.03604816e-01 2.62560934e-01
6.90648379e-03 1.53899789e-01 -4.56516296e-01 -9.32343364e-01
8.30962323e-03 8.01321208e-01 1.50453016e-01 4.77411091e-01
-8.52439046e-01 1.25774002e+00 7.23748016e+00 7.09360123e-01
-7.94066608e-01 -3.54117304e-01 5.90863109e-01 8.25240910e-02
-2.30980024e-01 -1.95673421e-01 -3.27529252e-01 1.55945137e-01
1.09721231e+00 -2.76116759e-01 8.34367692e-01 6.39754534e-01
-2.90114611e-01 -3.78656924e-01 -1.61826849e+00 8.81373107e-01
2.25457013e-01 -1.29216301e+00 3.82675499e-01 -4.23761010e-01
6.10153854e-01 -2.34158561e-01 9.31855217e-02 8.62894893e-01
8.12512815e-01 -1.81341672e+00 8.24627817e-01 6.85768604e-01
1.88856527e-01 -5.20617545e-01 5.68331897e-01 6.34708822e-01
-9.20372844e-01 -9.99404639e-02 -2.21158773e-01 -1.09074867e+00
-1.19916372e-01 3.23852569e-01 -5.98075032e-01 2.81557709e-01
3.62974584e-01 6.38440907e-01 -9.13777947e-01 6.59570038e-01
-2.78229445e-01 2.23105788e-01 -3.85788888e-01 -3.67874354e-01
3.23027551e-01 -1.81329343e-02 -8.57322142e-02 1.17510819e+00
1.24450967e-01 3.45524222e-01 5.61860427e-02 1.21792316e+00
-1.65510565e-01 -5.42787649e-02 -6.35868013e-01 -4.45075959e-01
4.63252127e-01 8.60930979e-01 -4.60422695e-01 -7.08635747e-01
-5.79593241e-01 7.33642220e-01 7.39693463e-01 1.06861211e-01
-8.50309849e-01 -2.64157444e-01 5.42442083e-01 -1.66185200e-01
-2.40308613e-01 -7.59043217e-01 -1.06813395e+00 -1.06896293e+00
-3.24307941e-02 -1.21904075e+00 2.25363865e-01 -1.28754139e+00
-1.44749129e+00 9.89295691e-02 1.25490800e-01 -3.45722854e-01
-3.29266697e-01 -1.20530534e+00 -5.97062767e-01 9.95229781e-01
-1.23228133e+00 -1.12341583e+00 -3.73172879e-01 4.02952850e-01
5.83295286e-01 -3.01899195e-01 1.14706743e+00 2.15214416e-02
4.46931943e-02 5.26620686e-01 -4.40564245e-01 4.06567752e-01
6.72780752e-01 -1.42414820e+00 8.28000307e-01 5.83570302e-01
2.07535520e-01 1.13682163e+00 4.88340735e-01 -4.84216660e-01
-1.63140428e+00 -3.77318144e-01 1.16139245e+00 -9.48681474e-01
1.06274009e+00 -3.99182707e-01 -8.20983827e-01 9.24910903e-01
3.51822108e-01 -2.40008309e-01 3.96497667e-01 2.97158688e-01
-7.97984421e-01 1.54154062e-01 -9.10939574e-01 9.22673404e-01
1.14615452e+00 -7.07529545e-01 -1.37414181e+00 4.80493516e-01
7.28781819e-01 -6.51209235e-01 -9.58270192e-01 1.75578147e-01
8.21958005e-01 -9.02523816e-01 1.27604449e+00 -8.67672384e-01
1.26002955e+00 -1.21228300e-01 -2.29588032e-01 -1.29057848e+00
-2.74506360e-01 -3.21514904e-01 -9.19098929e-02 8.07011724e-01
4.38842595e-01 -6.63018465e-01 5.12668431e-01 9.51984286e-01
-1.24109291e-01 -5.90792418e-01 -5.61666310e-01 -6.80876076e-01
8.50906968e-01 -7.39939690e-01 5.93863964e-01 1.04450428e+00
2.81423122e-01 3.86525422e-01 6.89477250e-02 -2.07965285e-01
3.90924126e-01 -1.38335645e-01 1.01388311e+00 -1.29625678e+00
-1.92420512e-01 -9.31195855e-01 -3.14791024e-01 -9.10853744e-01
3.51157337e-01 -1.16240060e+00 -2.84864455e-01 -1.82079220e+00
6.10782988e-02 -4.70445871e-01 -6.45686761e-02 7.08458364e-01
-1.40911087e-01 9.12805051e-02 7.10026920e-01 5.24996854e-02
-5.93631864e-01 1.34096533e-01 1.35048950e+00 -2.18444631e-01
3.15420032e-01 -3.69727015e-01 -1.00188601e+00 8.79467666e-01
7.69526064e-01 2.89344341e-02 -4.52237397e-01 -9.32065547e-01
6.30619049e-01 -1.22627512e-01 6.68525815e-01 -1.54360437e+00
5.04304767e-01 -1.07822604e-01 5.96899688e-01 -2.09159136e-01
3.75500113e-01 -6.16054654e-01 -2.45370001e-01 7.30327666e-01
-5.83401024e-01 4.47153628e-01 5.62419891e-01 -1.70758858e-01
-3.57447676e-02 -1.79775134e-01 4.77103233e-01 -5.98667562e-01
-8.57240498e-01 -3.54162484e-01 -2.13732377e-01 5.06744862e-01
7.65859187e-01 -4.03077453e-02 -8.69844258e-01 -3.94501328e-01
-3.62733841e-01 3.99309129e-01 2.81521887e-01 3.97620589e-01
2.96457767e-01 -9.90889788e-01 -9.69627619e-01 -9.27318037e-02
3.02918684e-02 2.25148454e-01 -5.55548295e-02 6.36437237e-01
-1.01751935e+00 8.94676030e-01 -5.27651250e-01 -2.33744040e-01
-7.48369157e-01 4.43977416e-01 3.33110392e-01 -6.28531039e-01
-3.93497825e-01 8.89491141e-01 6.06130362e-02 -7.67320395e-01
1.01109877e-01 -9.67899680e-01 9.94908214e-02 -1.74411312e-01
5.75116396e-01 3.29755694e-01 3.04773916e-02 5.83941527e-02
-3.54983628e-01 5.08690298e-01 -1.39985055e-01 -3.10209915e-02
1.29754817e+00 5.57734609e-01 -4.17188555e-01 6.57755673e-01
6.44055367e-01 -2.87019342e-01 -8.00561905e-01 -2.94738293e-01
1.03500761e-01 -1.79549292e-01 -2.13739797e-01 -1.38434875e+00
-5.02175093e-01 1.09766603e+00 -1.50840849e-01 1.01171665e-01
7.33914196e-01 -2.84778208e-01 5.69202363e-01 1.05717790e+00
5.05651832e-01 -9.49886680e-01 3.02599132e-01 1.20672846e+00
1.01813960e+00 -1.22180104e+00 3.44897866e-01 -2.30085701e-01
-6.04145885e-01 1.33243036e+00 7.36549914e-01 -3.84575427e-01
1.70355171e-01 4.47553039e-01 -1.70770422e-01 -2.90255785e-01
-1.01794446e+00 5.65022603e-02 2.22432747e-01 4.92964864e-01
9.50924456e-01 8.71808529e-02 -8.29605311e-02 5.75037062e-01
-7.97507107e-01 4.73304808e-01 4.20717299e-01 9.07721162e-01
-2.83709019e-01 -8.23639989e-01 -2.33571440e-01 4.70017374e-01
-2.20314220e-01 -6.45468771e-01 -3.99670035e-01 1.20593941e+00
-6.34284467e-02 6.71541929e-01 2.53873259e-01 1.20211579e-01
1.75712645e-01 5.33555686e-01 9.18320954e-01 -5.90575755e-01
-9.65933502e-01 -1.00883055e+00 2.39407241e-01 -4.59475875e-01
-1.47892937e-01 -3.26012820e-01 -1.43894112e+00 -9.74867821e-01
4.22573805e-01 -1.65551752e-01 4.73649740e-01 1.20234013e+00
-2.33225487e-02 6.90929830e-01 -4.74627852e-01 -6.92041337e-01
-8.64354134e-01 -9.01421964e-01 -1.63572326e-01 3.95150542e-01
-2.20394544e-02 -4.42427397e-01 -1.05454743e-01 -1.64107978e-01] | [9.54235553741455, 7.323666572570801] |
65086b76-36e4-48c5-85fb-dd88f2370536 | iotmalware-android-iot-malware-detection | 2102.13376 | null | https://arxiv.org/abs/2102.13376v2 | https://arxiv.org/pdf/2102.13376v2.pdf | Collective Intelligence: Decentralized Learning for Android Malware Detection in IoT with Blockchain | The widespread significance of Android IoT devices is due to its flexibility and hardware support features which revolutionized the digital world by introducing exciting applications almost in all walks of daily life, such as healthcare, smart cities, smart environments, safety, remote sensing, and many more. Such versatile applicability gives incentive for more malware attacks. In this paper, we propose a framework which continuously aggregates multiple user trained models on non-overlapping data into single model. Specifically for malware detection task, (i) we propose a novel user (local) neural network (LNN) which trains on local distribution and (ii) then to assure the model authenticity and quality, we propose a novel smart contract which enable aggregation process over blokchain platform. The LNN model analyzes various static and dynamic features of both malware and benign whereas the smart contract verifies the malicious applications both for uploading and downloading processes in the network using stored aggregated features of local models. In this way, the proposed model not only improves malware detection accuracy using decentralized model network but also model efficacy with blockchain. We evaluate our approach with three state-of-the-art models and performed deep analyses of extracted features of the relative model. | ['Waqar Ali', 'Ting Yang', 'Zakria', 'Jay Kumar', 'Wenyong Wang', 'Rajesh Kumar'] | 2021-02-26 | null | null | null | null | ['android-malware-detection'] | ['miscellaneous'] | [-4.72819284e-02 -3.37712735e-01 -6.52065039e-01 6.21227324e-02
-2.65266865e-01 -7.14670420e-01 1.02059209e+00 -2.22401381e-01
-1.50548369e-01 5.81383407e-01 -6.61922321e-02 -8.56441498e-01
3.15402895e-02 -7.73454726e-01 -6.61003649e-01 -8.25271845e-01
-7.70275146e-02 4.20471251e-01 4.63470131e-01 -1.06393613e-01
1.21604949e-01 2.28019252e-01 -1.19764197e+00 2.20896557e-01
7.84190476e-01 1.33273375e+00 -1.49703667e-01 4.79296774e-01
4.45520729e-01 1.10491288e+00 -5.61654747e-01 -2.95367360e-01
2.98861057e-01 -9.53634307e-02 -4.37430173e-01 -6.28348053e-01
-3.77039045e-01 -6.98850274e-01 -8.10588375e-02 1.22901392e+00
2.58359373e-01 -5.60720861e-01 4.73525494e-01 -1.75116563e+00
-6.87210739e-01 1.12834930e+00 -6.11834764e-01 2.18075886e-02
-7.82976225e-02 4.83297646e-01 7.26685524e-01 6.37581646e-02
3.71530622e-01 8.82255495e-01 8.96997094e-01 5.12226880e-01
-8.27661395e-01 -1.05475831e+00 -1.93590507e-01 3.19022834e-01
-8.12446058e-01 -1.64019302e-01 7.45317996e-01 -3.30176950e-01
9.26097810e-01 3.58496726e-01 7.03673720e-01 1.82938385e+00
8.17193031e-01 6.90339506e-01 1.30084050e+00 2.79843301e-01
2.61337131e-01 2.31745258e-01 2.58885741e-01 3.48272175e-01
5.16337812e-01 1.77049235e-01 1.03608489e-01 -6.51642144e-01
1.89846113e-01 5.83153546e-01 -2.71330290e-02 -2.02405035e-01
-1.27988613e+00 7.51321733e-01 2.64895797e-01 6.41906619e-01
-4.40979302e-01 5.05141318e-01 8.56948197e-01 2.80707002e-01
2.80768871e-01 -2.30682306e-02 -9.54885662e-01 -4.08263594e-01
-8.28199863e-01 -1.81897029e-01 1.05580366e+00 3.49159420e-01
4.59283650e-01 1.12723168e-02 3.47792953e-01 -5.08584715e-02
5.89024186e-01 8.15462887e-01 1.04963839e+00 -7.34469593e-01
2.60715574e-01 8.56755018e-01 -2.24842638e-01 -1.08623028e+00
-4.04550195e-01 -5.62378943e-01 -1.28828800e+00 -9.87372771e-02
-1.25022486e-01 -9.02221426e-02 -2.53592104e-01 1.53388059e+00
4.81219918e-01 5.85092366e-01 -3.53243984e-02 6.36773705e-02
3.39400291e-01 7.35802770e-01 -5.98568171e-02 -3.01972568e-01
1.39941752e+00 -8.73792470e-01 -3.57629240e-01 3.44288856e-01
7.73724318e-01 -4.25838739e-01 8.23622465e-01 5.45399666e-01
-5.58499455e-01 -2.84070164e-01 -9.06804442e-01 2.96647668e-01
-6.16423905e-01 -3.35086957e-02 4.20909315e-01 9.38076854e-01
-1.08535242e+00 4.91387010e-01 -1.13961017e+00 -3.38541955e-01
5.78229964e-01 8.32188725e-01 -1.58448592e-01 6.25163078e-01
-1.13816464e+00 5.49670994e-01 4.72975373e-01 -2.48144776e-01
-1.38650143e+00 -2.44205102e-01 -4.32532787e-01 2.85519391e-01
2.94335216e-01 -4.53599095e-01 1.17900014e+00 -8.23812246e-01
-1.60812879e+00 3.71843845e-01 4.79352087e-01 -9.37914729e-01
3.51711601e-01 2.80895531e-02 -5.64443886e-01 -1.97312579e-01
-1.21159390e-01 1.10680260e-01 1.06230628e+00 -1.07012904e+00
-5.23919463e-01 -3.21563423e-01 1.11178167e-01 -6.22111917e-01
-6.62405968e-01 2.51439303e-01 2.83864409e-01 -2.33198032e-01
-6.05672359e-01 -1.08699465e+00 1.96136579e-01 -6.58776879e-01
-4.38904017e-01 -4.51577306e-01 1.91991043e+00 -9.54029381e-01
1.27220571e+00 -1.98323381e+00 -2.59036571e-01 4.80837941e-01
5.05387723e-01 6.71938479e-01 8.48832801e-02 3.07314187e-01
3.33118230e-01 3.10627788e-01 -1.01571597e-01 -2.15654314e-01
2.21306980e-01 7.36167803e-02 -5.06666303e-01 5.79293907e-01
-3.99866134e-01 1.18984532e+00 -7.72732556e-01 -3.63921165e-01
1.92277223e-01 5.98701000e-01 -5.06343782e-01 -1.22733325e-01
-3.97542208e-01 7.54583597e-01 -6.31888449e-01 7.97634482e-01
5.97981215e-01 -6.86251462e-01 4.20234263e-01 -2.80281957e-02
2.00479090e-01 2.75700361e-01 -5.04475772e-01 1.02885222e+00
-5.95803261e-01 4.40933734e-01 3.45652431e-01 -9.14854288e-01
5.10910511e-01 2.72283942e-01 7.97197580e-01 -6.65572047e-01
8.12670648e-01 2.60697216e-01 2.59937078e-01 -4.94449914e-01
1.10322692e-01 5.79312563e-01 -8.76541287e-02 1.08811295e+00
-3.00782353e-01 4.77584511e-01 -4.94446367e-01 2.67662019e-01
1.38408339e+00 5.23246676e-02 5.59149981e-01 -3.32455724e-01
8.77740383e-01 -5.75454652e-01 3.07197660e-01 4.81719583e-01
-6.57953978e-01 -5.47056198e-01 7.89097428e-01 -4.74310756e-01
-8.56931329e-01 -7.68601954e-01 2.82126069e-01 1.00587165e+00
7.96363130e-02 -4.73308206e-01 -9.30058897e-01 -1.23207271e+00
2.51562335e-02 3.38655651e-01 -7.41908312e-01 -2.22758219e-01
-8.02238226e-01 -6.53230906e-01 8.43951762e-01 1.93885267e-01
9.85518992e-01 -1.14158762e+00 -6.81151211e-01 -4.50816005e-02
-1.04179226e-01 -1.04910195e+00 -3.52540761e-01 -6.59545213e-02
-6.20424271e-01 -1.27284813e+00 4.09439839e-02 -6.07899547e-01
2.36437455e-01 7.64953252e-03 7.10434616e-01 3.05370748e-01
2.78422236e-01 2.22712353e-01 -2.29256555e-01 -1.62114784e-01
-1.04527354e+00 4.72525805e-01 5.33518195e-01 2.11618915e-01
2.52805412e-01 -8.41051638e-01 -5.88406980e-01 4.19477016e-01
-8.66106570e-01 -3.82705688e-01 7.40566015e-01 5.42528272e-01
7.64681101e-02 1.41239911e-01 7.04337299e-01 -8.16915035e-01
8.39681208e-01 -9.16509330e-01 -8.29466522e-01 1.50473028e-01
-8.86660993e-01 -1.40181825e-01 1.05706024e+00 -7.48094440e-01
-6.32329583e-01 -3.35197240e-01 -1.29122376e-01 -3.82125527e-01
2.21929058e-01 1.19610436e-01 -3.51717204e-01 -1.44301131e-01
4.90306884e-01 4.52225029e-01 2.53563911e-01 -2.27304161e-01
2.51881629e-01 1.22636425e+00 2.75344133e-01 -3.43999684e-01
7.83402264e-01 5.01629710e-01 8.42321217e-02 -4.83493060e-01
1.39944151e-01 -8.62550512e-02 2.48197699e-03 1.35565922e-01
8.07972133e-01 -6.22665465e-01 -1.71084023e+00 9.57536817e-01
-1.28893387e+00 -3.41469616e-01 1.33793414e-01 1.85420454e-01
-2.27104560e-01 6.35698855e-01 -8.12673688e-01 -8.17219734e-01
-8.71667802e-01 -1.54309559e+00 9.63329792e-01 -1.04899757e-01
-1.19288608e-01 -1.04505205e+00 2.26937607e-01 3.44779551e-01
8.89425755e-01 2.52564073e-01 7.91923583e-01 -1.18736398e+00
-6.76504433e-01 -2.95177042e-01 -2.24577561e-01 8.85482192e-01
4.10301715e-01 3.59961763e-02 -9.54449177e-01 -4.91032422e-01
4.90634531e-01 -8.52375776e-02 3.99947137e-01 1.28827378e-01
1.30293512e+00 -1.16383874e+00 -6.02147698e-01 4.31081235e-01
1.06558967e+00 3.27296376e-01 4.95331705e-01 1.97463587e-01
8.29676449e-01 1.34188235e-01 -1.39212802e-01 3.02437186e-01
4.35328543e-01 3.54487777e-01 9.44893479e-01 4.08735007e-01
3.50740284e-01 -1.45974338e-01 8.97007406e-01 1.02721488e+00
-2.87152342e-02 -3.20562571e-02 -8.31764162e-01 2.55334020e-01
-1.76174057e+00 -1.04370320e+00 2.01019906e-02 1.90458834e+00
5.93182564e-01 2.73612678e-01 3.58661383e-01 -6.52530938e-02
6.32329524e-01 2.36514121e-01 -5.97614229e-01 -3.49044263e-01
1.28571793e-01 1.40165821e-01 6.06701434e-01 2.88457036e-01
-1.00524890e+00 5.61603904e-01 5.10228777e+00 1.11097276e+00
-1.50487649e+00 9.70360577e-01 7.42688835e-01 3.66642252e-02
-1.28309965e-01 4.86678723e-03 -6.03274465e-01 1.21359432e+00
1.33040309e+00 5.67397997e-02 8.00228834e-01 1.15233219e+00
2.80174732e-01 3.05343539e-01 -8.89623225e-01 8.29004049e-01
-1.71331331e-01 -1.32233024e+00 -1.02134205e-01 6.55717790e-01
8.57097983e-01 6.29139006e-01 2.89318621e-01 3.02570641e-01
4.61214513e-01 -7.15319157e-01 4.36672688e-01 2.20195875e-01
6.24593914e-01 -6.74508870e-01 9.79366183e-01 6.54188871e-01
-9.34781551e-01 -6.21037662e-01 2.75024533e-01 5.62963411e-02
-2.96235949e-01 4.01803970e-01 -7.97350764e-01 4.48202699e-01
6.99653089e-01 7.80132592e-01 -6.33965433e-01 2.62295842e-01
5.07267490e-02 9.19520319e-01 -4.90642339e-01 -3.51971060e-01
2.40272418e-01 -1.56473532e-01 3.69879693e-01 7.84083009e-01
4.09841210e-01 -6.32058084e-01 2.20300350e-02 6.78505242e-01
-4.23389494e-01 -1.66451573e-01 -8.12716663e-01 -5.37757836e-02
5.97606719e-01 1.38793194e+00 -6.11375868e-01 -5.42203367e-01
-2.19686851e-01 6.08186901e-01 -2.05169931e-01 6.50849473e-03
-1.08705389e+00 -7.61098042e-02 5.65517426e-01 1.87250122e-01
4.40769456e-02 -1.57199457e-01 -2.51904726e-01 -1.23258924e+00
-8.02687034e-02 -1.24956214e+00 2.39407793e-02 -3.76238585e-01
-1.10182214e+00 5.36949337e-01 -2.07513452e-01 -1.25670505e+00
-4.45155025e-01 -6.00580394e-01 -8.41102779e-01 2.55836040e-01
-1.12144959e+00 -1.63755345e+00 1.63726564e-02 8.61023307e-01
3.44985761e-02 -4.40354079e-01 4.55488890e-01 3.77896100e-01
-5.43295324e-01 5.29536605e-01 4.99784142e-01 9.20970812e-02
4.30855811e-01 -6.94181740e-01 5.22681475e-01 8.85177672e-01
-9.00199786e-02 9.41712439e-01 3.31661433e-01 -9.30112302e-01
-1.36157680e+00 -1.16837728e+00 5.04817009e-01 -7.19736397e-01
1.13867688e+00 -4.47793484e-01 -4.08965319e-01 9.99952793e-01
4.79767770e-01 -5.32198809e-02 3.88748735e-01 -5.16430795e-01
-5.14508188e-01 -6.01517558e-01 -1.49568927e+00 1.58735722e-01
5.55217445e-01 -6.16483390e-01 1.09570995e-01 3.02578181e-01
9.97906804e-01 7.67928809e-02 -6.41286135e-01 2.97750920e-01
8.93276334e-01 -1.14416766e+00 5.92872202e-01 -4.14297342e-01
1.12514138e-01 -1.96010053e-01 -2.81415254e-01 -8.03530157e-01
6.35722131e-02 -1.11191094e+00 -1.14866614e+00 1.11881673e+00
8.21023434e-02 -1.32888436e+00 7.09271491e-01 -4.76545580e-02
1.76050141e-01 -9.93492961e-01 -1.23713350e+00 -7.97422588e-01
-1.15649328e-02 -3.54228407e-01 1.22113740e+00 9.52046990e-01
-1.58433944e-01 4.86592166e-02 -5.33761561e-01 -2.90843938e-02
5.63585460e-01 -2.99716502e-01 8.17647338e-01 -1.38580561e+00
-4.91943359e-01 -5.40398419e-01 -4.11090076e-01 -8.15567076e-01
2.17605367e-01 -6.76861823e-01 -6.03525758e-01 -9.71715391e-01
2.87907064e-01 -3.87645572e-01 -5.79469442e-01 6.77751482e-01
5.22785902e-01 3.66035044e-01 9.34715271e-02 6.26904130e-01
-8.27434957e-01 2.07433209e-01 7.69467533e-01 -4.38182533e-01
-2.09011555e-01 2.32473135e-01 -8.16676974e-01 6.00602806e-01
1.04274058e+00 -4.35776293e-01 -3.82932305e-01 -1.39999986e-01
4.41271693e-01 -3.13976884e-01 6.66250348e-01 -1.04030406e+00
1.29157946e-01 4.16116156e-02 2.73276996e-02 -3.89397234e-01
2.25112657e-03 -1.28612459e+00 2.98827201e-01 1.14420128e+00
2.90791750e-01 3.24533522e-01 -4.34329242e-01 5.58064818e-01
3.53805035e-01 3.59633505e-01 7.26930976e-01 3.28829922e-02
-3.61499861e-02 2.53267735e-01 -2.22290710e-01 -6.20298564e-01
1.43914294e+00 1.42827123e-01 -9.72798228e-01 -5.02414584e-01
-3.62597704e-01 8.29585791e-02 6.29592299e-01 4.45302188e-01
7.87188262e-02 -1.08708525e+00 -1.88431159e-01 5.22182763e-01
-3.26013237e-01 -4.14469242e-01 7.10400790e-02 1.40653145e+00
-4.26695585e-01 5.98153412e-01 -2.07793593e-01 -6.65567219e-01
-1.30834138e+00 7.90332615e-01 1.84051737e-01 -9.47427809e-01
-7.73086846e-02 8.16604421e-02 -2.33694345e-01 -6.04332387e-01
1.42996326e-01 -6.34119570e-01 -3.23942788e-02 -2.24068359e-01
4.21679050e-01 7.89928913e-01 -5.91673329e-02 -8.96374464e-01
-5.02596557e-01 3.50814819e-01 -2.71517318e-02 3.29490751e-01
1.13440835e+00 2.37400010e-02 -7.91274548e-01 8.60835314e-02
1.35469782e+00 2.94407755e-01 -7.51834929e-01 5.82820773e-02
-2.61450708e-01 1.00910589e-01 -2.55188197e-01 -8.88913512e-01
-1.08208215e+00 7.05756605e-01 7.93864846e-01 8.70032966e-01
1.01959562e+00 -8.61623138e-03 1.38200688e+00 5.10944664e-01
7.45252430e-01 -5.74284673e-01 -3.92681584e-02 3.51852655e-01
1.31926641e-01 -9.14328098e-01 -2.08784848e-01 2.06607014e-01
-2.77637094e-01 6.70669556e-01 3.30170929e-01 -1.01650231e-01
1.05598187e+00 2.13639945e-01 -2.26559654e-01 7.84094818e-03
-5.10614634e-01 7.30580688e-01 -4.29333538e-01 7.30426967e-01
-3.23952645e-01 2.98750192e-01 -2.52716541e-01 9.14975166e-01
-1.67410210e-01 1.72294840e-01 3.99407595e-01 7.12483525e-01
-2.36557379e-01 -1.32726014e+00 -2.16932520e-01 7.90943980e-01
-7.43162930e-01 6.39380440e-02 -3.78114939e-01 6.16164863e-01
5.05564034e-01 9.72249031e-01 -3.16057920e-01 -1.11472559e+00
-4.53784108e-01 -8.14248696e-02 -1.31769642e-01 5.47810718e-02
-9.80802417e-01 -2.22484484e-01 -2.03565761e-01 -6.84062243e-01
-3.21597219e-01 -3.92422169e-01 -8.18403542e-01 -7.85616696e-01
-3.41706306e-01 1.84219077e-01 9.24758732e-01 9.39685464e-01
7.96795964e-01 2.95756876e-01 1.04527068e+00 -7.07335949e-01
-8.98350120e-01 -1.27450168e+00 -4.86648381e-01 1.32723093e-01
5.77673733e-01 -3.75973523e-01 -4.38522726e-01 -1.25340357e-01] | [14.423099517822266, 9.681137084960938] |
661cfcf7-075e-439d-a3e7-d35946698c2a | superyolo-super-resolution-assisted-object | 2209.13351 | null | https://arxiv.org/abs/2209.13351v2 | https://arxiv.org/pdf/2209.13351v2.pdf | SuperYOLO: Super Resolution Assisted Object Detection in Multimodal Remote Sensing Imagery | Accurately and timely detecting multiscale small objects that contain tens of pixels from remote sensing images (RSI) remains challenging. Most of the existing solutions primarily design complex deep neural networks to learn strong feature representations for objects separated from the background, which often results in a heavy computation burden. In this article, we propose an accurate yet fast object detection method for RSI, named SuperYOLO, which fuses multimodal data and performs high-resolution (HR) object detection on multiscale objects by utilizing the assisted super resolution (SR) learning and considering both the detection accuracy and computation cost. First, we utilize a symmetric compact multimodal fusion (MF) to extract supplementary information from various data for improving small object detection in RSI. Furthermore, we design a simple and flexible SR branch to learn HR feature representations that can discriminate small objects from vast backgrounds with low-resolution (LR) input, thus further improving the detection accuracy. Moreover, to avoid introducing additional computation, the SR branch is discarded in the inference stage, and the computation of the network model is reduced due to the LR input. Experimental results show that, on the widely used VEDAI RS dataset, SuperYOLO achieves an accuracy of 75.09% (in terms of mAP50 ), which is more than 10% higher than the SOTA large models, such as YOLOv5l, YOLOv5x, and RS designed YOLOrs. Meanwhile, the parameter size and GFLOPs of SuperYOLO are about 18 times and 3.8 times less than YOLOv5x. Our proposed model shows a favorable accuracy and speed tradeoff compared to the state-of-the-art models. The code will be open-sourced at https://github.com/icey-zhang/SuperYOLO. | ['Qian Du', 'Yunsong Li', 'Zhenman Fang', 'Weiying Xie', 'Jie Lei', 'Jiaqing Zhang'] | 2022-09-27 | null | null | null | null | ['real-time-object-detection', 'small-object-detection'] | ['computer-vision', 'computer-vision'] | [ 3.07693183e-01 -3.85644168e-01 1.23917192e-01 -7.86288157e-02
-7.63666332e-01 -1.50676131e-01 1.27857149e-01 -2.89842904e-01
-3.69930476e-01 6.62814319e-01 -4.17377830e-01 -1.40579790e-01
-2.34100148e-01 -1.15008223e+00 -5.84045053e-01 -1.13391066e+00
1.27306744e-01 -4.66313101e-02 4.67399001e-01 -8.82152244e-02
-1.55207291e-01 6.36848330e-01 -1.82488799e+00 4.96016294e-02
1.23410940e+00 1.35973763e+00 6.58337235e-01 4.69171792e-01
1.93053395e-01 6.47131801e-01 -3.08490515e-01 2.56968319e-01
2.74271339e-01 -1.50709540e-01 -1.12413064e-01 -1.80435270e-01
5.26674211e-01 -7.01590121e-01 -3.90640557e-01 1.10816395e+00
6.78016245e-01 1.74780115e-01 4.76106405e-01 -7.28926063e-01
-6.86029553e-01 4.52429622e-01 -1.04975462e+00 5.28154433e-01
-3.00015002e-01 3.53090912e-02 6.45071566e-01 -1.18773341e+00
1.08691886e-01 1.25500560e+00 6.50571048e-01 1.73914924e-01
-9.73725617e-01 -1.13392508e+00 6.12201169e-02 1.67762771e-01
-1.95834851e+00 -3.85592043e-01 4.49339807e-01 -2.96866924e-01
5.05381465e-01 3.42236221e-01 2.00453937e-01 4.86617655e-01
-2.14636773e-01 6.28251493e-01 1.10022068e+00 -1.07482150e-01
-4.42582332e-02 4.72857542e-02 1.50235251e-01 6.26217782e-01
5.32197952e-01 2.10268676e-01 -2.34302610e-01 1.15016423e-01
1.11043310e+00 3.14412385e-01 -3.69757116e-01 4.46624517e-01
-9.95774806e-01 8.22835803e-01 7.81319797e-01 2.79674560e-01
-4.68253911e-01 -1.04426779e-01 -6.03078268e-02 -1.62277237e-01
5.51399350e-01 -6.91467598e-02 -2.62965441e-01 5.30076861e-01
-7.65771508e-01 1.72517933e-02 8.64357650e-02 6.30323470e-01
8.99944127e-01 2.62041956e-01 -3.22884321e-02 8.96447480e-01
3.19753379e-01 1.23664236e+00 -6.78419396e-02 -9.11804199e-01
5.25668502e-01 6.74111128e-01 3.08620483e-01 -1.11046386e+00
-4.84394044e-01 -7.14635909e-01 -1.28551519e+00 2.44149134e-01
1.14137888e-01 -8.84186476e-02 -8.30139518e-01 1.36862171e+00
4.29446399e-01 1.51042968e-01 1.60557315e-01 1.08890367e+00
1.31354451e+00 1.16914105e+00 -2.09838375e-02 -2.25340620e-01
1.38699245e+00 -5.11775374e-01 -3.32672447e-01 -3.39123726e-01
2.42694050e-01 -5.65675974e-01 9.68581676e-01 2.93220520e-01
-8.12054157e-01 -8.91654193e-01 -9.94102299e-01 1.05418772e-01
-2.02656984e-01 1.07342672e+00 5.50076306e-01 2.87651241e-01
-6.13759160e-01 2.10819632e-01 -7.46134460e-01 -2.23939985e-01
6.25890136e-01 6.45928159e-02 -8.07858333e-02 -1.74473971e-01
-1.16413474e+00 6.59783959e-01 5.17587423e-01 6.61311865e-01
-9.41164911e-01 -6.39036000e-01 -7.54250407e-01 1.97239190e-01
6.08806491e-01 -2.42472798e-01 6.84046805e-01 -5.60607851e-01
-1.15649664e+00 5.70787430e-01 -1.65892288e-01 -1.68061793e-01
2.32658952e-01 -2.92179853e-01 -5.69598258e-01 5.10396361e-01
1.66896045e-01 4.14275438e-01 8.35037112e-01 -1.22798038e+00
-9.38576579e-01 -5.84903717e-01 1.71266034e-01 1.16645433e-01
-3.05327773e-01 2.16277272e-01 -2.96246082e-01 -4.01500702e-01
4.24577355e-01 -7.06956744e-01 -2.88598478e-01 1.94035575e-01
-1.75622880e-01 -1.73663184e-01 9.34489667e-01 -7.54365265e-01
1.20381808e+00 -2.28769326e+00 -1.47649378e-01 -2.98544951e-02
3.16087633e-01 7.22006142e-01 -9.54155177e-02 -1.91391364e-01
2.10532740e-01 3.34276333e-02 -3.14855993e-01 2.55327880e-01
-5.37705004e-01 -5.41403592e-02 -1.71394438e-01 5.65853179e-01
2.83437937e-01 6.12729490e-01 -6.31554306e-01 -5.56170046e-01
4.60778028e-01 7.08538294e-01 -1.09821260e-02 1.60759151e-01
2.00345963e-01 4.39184070e-01 -7.59481788e-01 9.56047058e-01
1.29934084e+00 -4.51536834e-01 -2.08200186e-01 -6.50104702e-01
-5.73682666e-01 -1.26203507e-01 -1.50294614e+00 9.82926309e-01
-5.98966718e-01 4.30177033e-01 4.42524374e-01 -8.42709184e-01
1.25614357e+00 -2.00691279e-02 1.97433352e-01 -7.03469932e-01
1.19386919e-01 2.56588638e-01 -3.88405085e-01 -4.94339943e-01
3.77934098e-01 3.03135756e-02 2.42725849e-01 4.68555689e-02
-4.16360736e-01 3.05831730e-01 1.35819152e-01 -4.08441527e-03
3.63517910e-01 1.32092059e-01 2.18591511e-01 -6.93594292e-02
8.01064074e-01 -9.00379196e-02 6.82862103e-01 9.60876882e-01
-2.39840467e-02 4.60071743e-01 -1.46825001e-01 -3.78536344e-01
-5.92892349e-01 -1.14165270e+00 -5.00177324e-01 9.98653352e-01
5.56671083e-01 1.31207094e-01 -3.42116505e-01 -9.43743885e-02
-7.27254376e-02 3.39309871e-01 -3.45698208e-01 2.89134718e-02
-5.61499476e-01 -1.37376964e+00 4.80110854e-01 8.06004882e-01
1.06274092e+00 -8.18291605e-01 -7.43887305e-01 3.30724716e-02
-3.02666306e-01 -1.06439447e+00 2.02828944e-01 -1.14962578e-01
-8.92954826e-01 -9.01226819e-01 -7.07931459e-01 -4.69181329e-01
5.01858652e-01 9.89834011e-01 6.63455367e-01 1.14443786e-01
-5.52007377e-01 -2.56885797e-01 -3.98558974e-01 -2.90447623e-01
-3.83796939e-03 -7.75868371e-02 -7.32763112e-02 1.91492867e-02
2.95168370e-01 -2.85439491e-01 -7.58789122e-01 4.36024129e-01
-9.29345906e-01 4.18470576e-02 8.72478902e-01 6.84030592e-01
6.46136463e-01 3.76075923e-01 6.39205933e-01 -3.32006067e-01
-3.13505560e-01 -3.73639137e-01 -1.09279144e+00 2.99875647e-01
-2.92124659e-01 -3.96357179e-01 5.07363677e-01 -3.22617590e-01
-1.47620654e+00 7.87255690e-02 3.52085121e-02 -3.30810726e-01
-2.30789348e-01 5.10888457e-01 -1.71016112e-01 -2.81890392e-01
6.33506656e-01 3.84790033e-01 -1.04960367e-01 -7.67377377e-01
2.11028799e-01 9.66083348e-01 4.63396132e-01 -2.15020046e-01
9.90948975e-01 7.12567627e-01 7.94799812e-03 -1.15521514e+00
-1.07884014e+00 -4.55821931e-01 -4.51606005e-01 -1.74551934e-01
7.40080416e-01 -1.48480177e+00 -8.15119863e-01 7.06003070e-01
-7.96600938e-01 -6.66902363e-02 1.85492411e-01 7.10272491e-01
1.26598552e-01 3.30817252e-01 -7.08965361e-01 -1.22182894e+00
-6.54162705e-01 -8.85395169e-01 1.19701254e+00 8.23385954e-01
7.12299526e-01 -4.04232532e-01 -5.29253900e-01 4.62095231e-01
5.77900350e-01 1.79823741e-01 5.26658297e-01 7.41698742e-02
-8.74516010e-01 -2.86219623e-02 -1.05470395e+00 4.78658885e-01
2.45523434e-02 1.32523149e-01 -1.04988921e+00 -2.91869104e-01
-1.55521512e-01 -2.58321017e-01 1.30415881e+00 6.52865171e-01
1.18913531e+00 -3.11904633e-03 -3.63081008e-01 7.57142305e-01
1.68379188e+00 1.48302630e-01 6.03229761e-01 2.78371632e-01
8.57515693e-01 4.63501394e-01 1.10950720e+00 5.95008731e-01
3.08988810e-01 4.09482181e-01 6.04818225e-01 -4.01192307e-01
-3.42274383e-02 1.22202270e-01 3.39643717e-01 3.61358404e-01
-5.59999406e-01 -4.78852279e-02 -7.06283212e-01 4.05700982e-01
-1.61115110e+00 -1.18621922e+00 -3.33582550e-01 2.18446159e+00
5.30601501e-01 -1.97188392e-01 -1.52766332e-01 3.54874246e-02
8.95666301e-01 3.39186162e-01 -6.36102200e-01 4.96804059e-01
-3.51126611e-01 5.26584052e-02 7.90972233e-01 4.43691105e-01
-1.37290311e+00 8.71220767e-01 4.88271856e+00 1.10841572e+00
-1.34270346e+00 1.66629910e-01 6.39580667e-01 -1.45585418e-01
1.37758180e-01 -3.31475735e-01 -1.31718087e+00 2.83394605e-01
5.58208168e-01 2.55503535e-01 3.73858631e-01 8.76079381e-01
5.18741608e-01 -3.97674888e-01 -2.35395864e-01 1.07293344e+00
-5.81068769e-02 -1.25311613e+00 -8.90398473e-02 -6.86209723e-02
5.46943128e-01 3.10943425e-01 2.01587435e-02 2.73866624e-01
-7.76513591e-02 -9.38359797e-01 4.73107338e-01 5.45063615e-01
9.18332338e-01 -7.87078798e-01 8.61106277e-01 3.95069808e-01
-1.70501089e+00 -3.85319412e-01 -9.23866272e-01 1.10393744e-02
-7.28432909e-02 8.43831241e-01 -2.64265239e-01 7.75166571e-01
1.15782833e+00 8.28718722e-01 -5.84195793e-01 7.92514443e-01
-1.88660681e-01 7.10238934e-01 -6.55616164e-01 2.29238153e-01
1.83828101e-01 -4.08775270e-01 5.54780424e-01 1.02028739e+00
6.67871237e-01 4.59750295e-01 3.80928427e-01 1.03186285e+00
1.91802979e-01 -9.81538817e-02 -4.20290828e-01 1.61289915e-01
7.27600157e-01 1.59258902e+00 -5.44979572e-01 -4.58634824e-01
-3.38625401e-01 5.11086464e-01 -5.46074025e-02 4.42133665e-01
-1.04473579e+00 -5.10870397e-01 4.12017912e-01 -6.64109513e-02
6.45959496e-01 -1.82970703e-01 -2.69790828e-01 -1.23783767e+00
9.74491835e-02 -6.09090090e-01 4.75087285e-01 -9.63108897e-01
-9.42688644e-01 6.65745676e-01 2.83911489e-02 -1.43662941e+00
4.62980509e-01 -7.42884278e-01 -5.34830153e-01 9.31139648e-01
-1.88788712e+00 -1.27933908e+00 -9.76681232e-01 3.44998628e-01
3.53537053e-01 8.31476226e-02 3.52307439e-01 5.32027543e-01
-1.01220405e+00 2.95924574e-01 4.16231364e-01 3.41744535e-02
3.24563026e-01 -6.07830346e-01 -2.01540485e-01 1.11606956e+00
-3.56349856e-01 4.30382878e-01 3.35073173e-01 -4.48016346e-01
-1.29721904e+00 -1.44093525e+00 3.95660818e-01 -5.70042208e-02
4.97844756e-01 -3.55510004e-02 -1.15075409e+00 2.84227788e-01
-4.72528368e-01 1.73248127e-01 3.16980660e-01 -2.72349656e-01
-2.27692261e-01 -6.92979634e-01 -1.08761477e+00 2.57214427e-01
9.02048826e-01 -3.29772562e-01 -2.60808676e-01 2.20351502e-01
7.38843501e-01 -2.02822730e-01 -9.84259665e-01 8.03527832e-01
5.56843877e-01 -9.01706159e-01 1.31406558e+00 1.65276662e-01
2.92003214e-01 -8.99614155e-01 -4.65415508e-01 -6.54248059e-01
-4.89224195e-01 1.63577944e-01 -3.96287479e-02 1.27201045e+00
1.92461446e-01 -7.35907495e-01 3.60942423e-01 7.01386556e-02
1.38909966e-02 -4.97123241e-01 -8.69341552e-01 -8.46376598e-01
-3.74749124e-01 -7.93443024e-02 4.97724056e-01 6.12994254e-01
-8.25166583e-01 2.29747966e-01 -4.21138912e-01 9.14879322e-01
9.13597405e-01 6.26675963e-01 6.30205989e-01 -1.46454668e+00
-3.35610546e-02 -2.26358265e-01 3.89682986e-02 -9.55874622e-01
-3.34935099e-01 -5.86758196e-01 1.19080968e-01 -1.64735043e+00
3.42517018e-01 -5.72733998e-01 -4.55823660e-01 6.06792450e-01
-4.68884766e-01 5.80181003e-01 9.48851332e-02 4.41818863e-01
-5.09090304e-01 7.17077017e-01 1.14980865e+00 -8.27209279e-02
-2.44695067e-01 1.07793631e-02 -6.31108165e-01 8.35464954e-01
8.93844962e-01 -3.88830215e-01 -1.51384756e-01 -4.58299190e-01
1.11760475e-01 7.30595440e-02 7.33042479e-01 -1.14274859e+00
-4.73572277e-02 -3.09276998e-01 7.07132876e-01 -1.14616537e+00
3.56606483e-01 -5.43025434e-01 5.79232462e-02 5.18278182e-01
1.87074393e-01 -6.13197327e-01 2.36119121e-01 4.72447455e-01
-1.86137259e-01 -1.24596745e-01 1.12125885e+00 -8.16485286e-02
-9.05207515e-01 4.25151885e-01 -2.29477227e-01 -3.58920068e-01
7.97305107e-01 -4.75527011e-02 -6.48143113e-01 8.17374792e-03
-4.66577441e-01 2.92589754e-01 2.09291950e-01 1.20525435e-01
7.18288183e-01 -1.04902530e+00 -9.76408422e-01 1.39771953e-01
1.44418150e-01 3.95746142e-01 8.36596966e-01 1.08411205e+00
-4.97024924e-01 2.37096488e-01 -1.10303432e-01 -8.47448766e-01
-1.24052334e+00 3.04738909e-01 3.43158782e-01 -2.20424663e-02
-7.73614109e-01 7.62901962e-01 5.05056918e-01 -3.54041904e-01
-1.25811011e-01 -2.54258335e-01 -3.99040788e-01 1.10277362e-01
9.20926154e-01 8.71368468e-01 -3.56007904e-01 -6.85762286e-01
-5.23349285e-01 8.29935431e-01 1.07255980e-01 5.05409837e-01
1.40532720e+00 -3.86105865e-01 -2.23030791e-01 1.62894219e-01
7.97303498e-01 -1.68374762e-01 -1.33999848e+00 -5.04041493e-01
-5.61102629e-01 -5.69925725e-01 4.95619804e-01 -6.08215511e-01
-1.24855494e+00 9.97721434e-01 1.04749823e+00 -8.69565159e-02
1.38660932e+00 1.96796224e-01 7.44052351e-01 5.65937698e-01
3.01518619e-01 -8.02886188e-01 -1.66064069e-01 3.60316753e-01
8.15200865e-01 -1.49037170e+00 3.10047537e-01 -4.86297697e-01
-5.19216001e-01 9.91431952e-01 8.06522012e-01 -1.20109454e-01
4.17907298e-01 9.00089890e-02 5.00946790e-02 -1.79791287e-01
-2.58997083e-01 -4.50744033e-01 1.74115330e-01 4.31919396e-01
4.18203063e-02 1.82559267e-01 -9.04195309e-02 6.96481884e-01
4.84609842e-01 -3.16460282e-02 3.59033257e-01 5.80991030e-01
-1.00859785e+00 -2.18016282e-01 -7.52280593e-01 4.90077764e-01
-2.83581913e-01 -2.45095715e-01 2.12755844e-01 6.80801630e-01
2.44560033e-01 1.20793283e+00 -6.02068305e-02 -2.62040526e-01
1.12644918e-01 -6.14102185e-01 1.20051153e-01 -3.02968234e-01
7.22755417e-02 2.92469054e-01 -1.24938436e-01 -5.48915327e-01
-6.90712154e-01 -4.00378346e-01 -1.31752467e+00 -1.12073570e-01
-7.04343617e-01 -8.33912715e-02 5.00997901e-01 8.02895308e-01
3.76142293e-01 4.06787246e-01 7.52543151e-01 -1.16107786e+00
-5.90005100e-01 -1.10819447e+00 -8.46476078e-01 -3.01437885e-01
3.69453669e-01 -9.87196207e-01 -5.54191232e-01 -2.28905067e-01] | [9.18464183807373, -0.9795426726341248] |
c63e16a0-ab24-44e8-94e0-65e362b63597 | autonomous-capability-assessment-of-black-box | 2306.04806 | null | https://arxiv.org/abs/2306.04806v1 | https://arxiv.org/pdf/2306.04806v1.pdf | Autonomous Capability Assessment of Black-Box Sequential Decision-Making Systems | It is essential for users to understand what their AI systems can and can't do in order to use them safely. However, the problem of enabling users to assess AI systems with evolving sequential decision making (SDM) capabilities is relatively understudied. This paper presents a new approach for modeling the capabilities of black-box AI systems that can plan and act, along with the possible effects and requirements for executing those capabilities in stochastic settings. We present an active-learning approach that can effectively interact with a black-box SDM system and learn an interpretable probabilistic model describing its capabilities. Theoretical analysis of the approach identifies the conditions under which the learning process is guaranteed to converge to the correct model of the agent; empirical evaluations on different agents and simulated scenarios show that this approach is few-shot generalizable and can effectively describe the capabilities of arbitrary black-box SDM agents in a sample-efficient manner. | ['Siddharth Srivastava', 'Rushang Karia', 'Pulkit Verma'] | 2023-06-07 | null | null | null | null | ['active-learning', 'active-learning'] | ['methodology', 'natural-language-processing'] | [ 8.17400739e-02 3.60509306e-01 -1.77679896e-01 -3.38658571e-01
-3.94661695e-01 -6.77765727e-01 9.32216167e-01 2.08924294e-01
-3.81239176e-01 6.16584182e-01 -1.01688623e-01 -4.61974144e-01
-5.56068599e-01 -6.04833961e-01 -2.55587786e-01 -7.21336722e-01
-6.42972827e-01 1.28296351e+00 2.66594797e-01 -1.80929080e-01
2.53564328e-01 1.66483924e-01 -1.75855625e+00 -1.60317138e-01
9.82392013e-01 6.42990589e-01 1.39783263e-01 1.17801249e+00
5.21984510e-02 1.45244396e+00 -7.13571191e-01 8.13241303e-02
5.38734794e-01 -1.79629311e-01 -5.44946730e-01 3.51339877e-01
-1.26925379e-01 -6.87853336e-01 -2.37900212e-01 9.35686707e-01
4.96728308e-02 4.36944544e-01 7.73949981e-01 -1.74476027e+00
6.43572072e-03 9.18079555e-01 1.33654401e-01 2.01505885e-01
4.76455867e-01 1.09351194e+00 6.55777633e-01 2.68001924e-03
3.59912157e-01 1.38000059e+00 1.45797366e-02 8.20016444e-01
-1.40228367e+00 -4.55925643e-01 3.09493572e-01 1.52308956e-01
-1.21656156e+00 -6.02842271e-01 2.42141336e-01 -4.61815506e-01
1.15543354e+00 2.11070001e-01 9.78349268e-01 8.37167680e-01
6.31865144e-01 9.47654366e-01 1.12787485e+00 -3.42467904e-01
1.21076560e+00 6.75902143e-02 1.88854679e-01 3.61333400e-01
6.23538196e-01 5.95806956e-01 -4.85994637e-01 -5.76486886e-01
4.23097849e-01 3.41221914e-02 5.12159824e-01 -8.22300851e-01
-8.90258551e-01 5.64293683e-01 -3.92143995e-01 -8.83262381e-02
-7.99532175e-01 4.36755300e-01 -8.15992057e-02 4.72992778e-01
8.69149342e-02 8.93902659e-01 -3.33587229e-01 -5.35202801e-01
-7.99965262e-01 6.59648716e-01 1.30545437e+00 1.09816289e+00
2.83545673e-01 3.86627287e-01 -4.22525182e-02 -1.57936156e-01
5.54651320e-01 5.40407002e-01 6.89848512e-02 -1.59563673e+00
-5.75738363e-02 6.66744292e-01 9.42560375e-01 8.06944817e-02
-3.22384864e-01 1.76470444e-01 -1.28991961e-01 1.11335719e+00
3.02484751e-01 -7.53193080e-01 -9.22398865e-01 1.51229727e+00
-1.13769835e-02 2.80632675e-01 3.83118182e-01 5.23889303e-01
-1.55188695e-01 1.05999339e+00 3.47139180e-01 -5.89443803e-01
8.19082439e-01 -4.12281066e-01 -6.87602937e-01 -5.73840559e-01
3.86947066e-01 1.08707719e-01 7.80356050e-01 4.44843709e-01
-1.50497270e+00 -6.47916794e-02 -1.25720775e+00 9.01979029e-01
3.60464081e-02 -7.94723749e-01 6.87467396e-01 6.54564083e-01
-1.16844141e+00 2.07500681e-01 -1.47746348e+00 -3.78098696e-01
3.47276151e-01 5.14130473e-01 4.51706022e-01 7.57684708e-02
-8.26272368e-01 1.41022289e+00 6.48045361e-01 -5.08632243e-01
-2.01336908e+00 -5.65266788e-01 -5.80317140e-01 2.86225021e-01
1.04844952e+00 -8.04334462e-01 2.09583592e+00 -1.03384686e+00
-1.68504322e+00 1.32814020e-01 2.07979083e-01 -8.73040795e-01
6.77608788e-01 1.59761786e-01 -2.62401283e-01 -1.34883285e-01
-2.29377598e-01 4.20537502e-01 8.06525528e-01 -1.28807986e+00
-1.10000777e+00 -1.36988148e-01 5.83875000e-01 6.87377036e-01
-1.16851248e-01 4.27609049e-02 4.18118358e-01 1.33923307e-01
-3.93194199e-01 -1.08151925e+00 -8.24592650e-01 -2.13173598e-01
5.11375442e-02 -3.56711149e-01 3.50331485e-01 5.50172217e-02
9.97238100e-01 -1.83903468e+00 1.02801666e-01 1.76005527e-01
2.12623969e-01 -2.08523460e-02 4.53124605e-02 9.34819281e-01
5.91208518e-01 3.77596021e-02 -9.50858593e-02 -8.57553557e-02
4.16515976e-01 4.58007634e-01 -2.54452676e-01 2.12378219e-01
-3.39105837e-02 5.01287162e-01 -9.30626690e-01 -3.89358759e-01
4.55752701e-01 -2.89536238e-01 -1.80044055e-01 6.25225425e-01
-9.02940810e-01 3.15429688e-01 -7.02229142e-01 4.90347296e-01
1.52967542e-01 -1.63048476e-01 6.48893774e-01 1.03237617e+00
-3.10602129e-01 2.73564048e-02 -1.43306959e+00 9.31956232e-01
-3.31707329e-01 3.53238225e-01 2.26681009e-01 -6.56672120e-01
6.33092344e-01 4.59140867e-01 4.30073947e-01 -4.27730143e-01
1.69730410e-01 7.58899152e-02 5.19251883e-01 -5.03589809e-01
7.24772960e-02 -2.46148333e-01 -4.20188367e-01 9.22028720e-01
-1.24589041e-01 -3.85417104e-01 4.74923551e-01 3.03899378e-01
1.38436723e+00 -3.92662823e-01 7.43519485e-01 -4.37446833e-01
1.75351307e-01 4.18578655e-01 5.74648798e-01 1.30510378e+00
-4.06602532e-01 -5.81968784e-01 4.01097178e-01 -5.79437077e-01
-9.19323444e-01 -1.12302744e+00 3.79412413e-01 1.06729352e+00
5.44093311e-01 -2.63780981e-01 -6.69124186e-01 -2.31409490e-01
5.38812652e-02 1.61291516e+00 -4.04897898e-01 -3.69069755e-01
-1.44308910e-01 -4.33878362e-01 5.95220039e-03 3.96704704e-01
2.90387422e-01 -1.13844526e+00 -1.64135373e+00 3.64210099e-01
5.07445157e-01 -5.85249305e-01 -1.40172184e-01 1.32768452e-01
-8.17872763e-01 -9.81116593e-01 -1.09047763e-01 -2.56574929e-01
6.40769541e-01 6.95665330e-02 8.24121237e-01 -2.27312952e-01
-2.76929978e-02 8.38385224e-01 -1.71712358e-02 -8.37516963e-01
-9.28076684e-01 -3.11448872e-01 4.65077788e-01 -2.45519325e-01
4.81730759e-01 -3.17040712e-01 -3.03057104e-01 1.47533000e-01
-7.33537436e-01 1.55570552e-01 5.13498366e-01 3.77387255e-01
1.64884627e-01 8.73254001e-01 2.74207711e-01 -6.77593708e-01
1.08310187e+00 -4.31776047e-01 -1.18310976e+00 6.69635832e-01
-9.75482345e-01 3.16414535e-01 5.05962729e-01 -6.61600828e-01
-1.56428063e+00 3.63173157e-01 8.44111383e-01 -2.94203032e-02
-3.28430384e-01 3.10271025e-01 -3.59744102e-01 2.11283684e-01
6.35875821e-01 4.08999830e-01 2.73211718e-01 9.75836739e-02
2.85402387e-01 6.12522960e-01 1.80306748e-01 -8.21795166e-01
9.17430401e-01 2.98237562e-01 3.29279415e-02 -5.65897107e-01
-2.87261337e-01 -6.99147210e-02 -1.26235813e-01 -7.04020023e-01
4.61564481e-01 -6.76549554e-01 -1.02338910e+00 5.63701212e-01
-5.67179620e-01 -8.48370850e-01 -5.77999890e-01 2.16429770e-01
-1.20012200e+00 -8.22897404e-02 -2.10209697e-01 -1.70878518e+00
5.09579144e-02 -1.05145216e+00 3.79390627e-01 6.37764573e-01
-4.96399671e-01 -7.63221264e-01 5.69999516e-01 2.13893563e-01
6.30434930e-01 -2.46812344e-01 9.59464669e-01 -1.15303874e+00
-1.03421009e+00 -2.31094941e-01 7.27463365e-01 -9.66297314e-02
-1.93169743e-01 1.43031448e-01 -5.10159731e-01 -4.98300761e-01
1.54782280e-01 -1.81817055e-01 -9.94972535e-04 6.92462027e-01
5.49853921e-01 -7.92246938e-01 -4.19110686e-01 -1.54403538e-01
1.24430454e+00 1.09887564e+00 4.87248182e-01 3.52135271e-01
-2.90148169e-01 6.50870502e-01 1.00505674e+00 9.89810586e-01
3.43545496e-01 3.02269965e-01 5.01750529e-01 6.78547561e-01
6.53369129e-01 -1.81228280e-01 8.98683310e-01 -6.57275692e-02
-1.37500495e-01 -4.32465255e-01 -1.33524728e+00 2.99695611e-01
-2.32126474e+00 -1.29480863e+00 5.46509266e-01 2.49832368e+00
6.31989539e-01 6.93288863e-01 4.55247641e-01 -2.24625841e-01
5.16833127e-01 -2.39088252e-01 -1.28776634e+00 -4.82603580e-01
2.16028795e-01 -1.28137127e-01 4.63908076e-01 5.83900452e-01
-8.83868039e-01 7.41141617e-01 7.69295883e+00 4.13989902e-01
-4.51472819e-01 1.24558680e-01 4.76552963e-01 -3.57953906e-01
-9.73763540e-02 4.99523073e-01 -5.81451595e-01 5.53627610e-01
1.56844449e+00 -1.13745975e+00 7.75248826e-01 1.07583487e+00
5.45910895e-01 -5.72309077e-01 -1.53091586e+00 5.85592806e-01
-3.49758238e-01 -1.19513774e+00 -2.19447300e-01 1.65045410e-01
9.88510311e-01 -2.28660628e-01 -9.12070274e-02 5.02211869e-01
1.40788519e+00 -9.89311695e-01 9.16813254e-01 8.78327072e-01
1.70823514e-01 -9.39307928e-01 6.92669213e-01 1.22135472e+00
-7.34811187e-01 -9.77703273e-01 3.21109663e-04 -6.17983162e-01
2.35868514e-01 -1.88385263e-01 -1.26470757e+00 -2.43924677e-01
2.81673610e-01 7.65443072e-02 -2.17720136e-01 1.10243142e+00
-4.56684977e-02 3.37728232e-01 -4.01750386e-01 -5.87840378e-01
1.76235080e-01 -1.59385607e-01 8.71002018e-01 4.71656233e-01
3.65697145e-01 4.40114945e-01 4.76181358e-01 8.81326616e-01
6.17120445e-01 -5.93998730e-01 -5.96851349e-01 -5.89860260e-01
9.96665478e-01 8.13564479e-01 -6.53376341e-01 -7.12363780e-01
-7.10064620e-02 3.65184546e-01 -2.65498161e-01 3.67884636e-01
-4.68965352e-01 1.59521908e-01 8.16420853e-01 1.72882766e-01
-2.53958195e-01 -4.26907986e-01 -2.96875656e-01 -8.77588868e-01
-4.97632951e-01 -1.10179102e+00 4.74342257e-01 -7.74030685e-01
-1.29508090e+00 1.43703625e-01 4.56245184e-01 -1.07987690e+00
-8.28022420e-01 -2.44958758e-01 -8.02071452e-01 4.60338593e-01
-6.56173229e-01 -6.05464756e-01 -1.59429818e-01 3.63012642e-01
8.34824860e-01 -6.75743520e-01 6.77263319e-01 -6.75297320e-01
-5.30101776e-01 -1.49241358e-01 2.77158231e-01 -7.47660339e-01
6.66546747e-02 -1.35087550e+00 2.85999596e-01 9.59043324e-01
-2.57622153e-01 5.38120449e-01 1.20868742e+00 -7.54307449e-01
-1.84833395e+00 -7.32484818e-01 1.98473871e-01 -8.03733885e-01
6.62860930e-01 -1.39498278e-01 -3.14257056e-01 8.75103652e-01
3.60456049e-01 -7.62963831e-01 3.47251803e-01 -1.35257021e-01
1.34380937e-01 -1.29761472e-01 -1.15871203e+00 9.12128806e-01
8.57520044e-01 -1.55405253e-01 -9.78384912e-01 2.29780614e-01
5.44559300e-01 1.02445580e-01 -2.30306879e-01 9.68589783e-02
2.17039794e-01 -7.58932710e-01 6.68960154e-01 -8.48724008e-01
-2.08329111e-01 -4.32953745e-01 -1.72145721e-02 -1.38418782e+00
-4.10866946e-01 -1.16769886e+00 -7.38571346e-01 6.85389996e-01
1.95814073e-01 -8.36553752e-01 6.00045562e-01 1.42251086e+00
1.25844434e-01 -2.28266597e-01 -1.05195391e+00 -1.15888715e+00
-7.78502971e-02 -2.21017316e-01 6.83903515e-01 3.78979921e-01
4.84085858e-01 4.06198092e-02 -1.51296049e-01 2.94190854e-01
1.13966441e+00 2.10265517e-02 5.56288540e-01 -1.37237263e+00
-3.53148162e-01 -5.05111754e-01 -2.59583235e-01 -6.92729354e-01
1.95645005e-01 -2.78360546e-01 3.94877642e-01 -1.54058886e+00
5.49688101e-01 -4.48900521e-01 -2.61562586e-01 3.38248789e-01
2.63615638e-01 -1.01321256e+00 3.71165097e-01 4.40176636e-01
-1.05101776e+00 2.53466159e-01 6.70279562e-01 -2.16278031e-01
-4.93946582e-01 3.72378260e-01 -6.55378819e-01 7.47463822e-01
7.93199599e-01 -3.01270097e-01 -9.05686557e-01 9.98102278e-02
4.21100497e-01 6.15190685e-01 2.54224271e-01 -1.12459707e+00
7.46438384e-01 -1.15962851e+00 -1.91986308e-01 -2.90689230e-01
2.60934323e-01 -9.52137351e-01 7.42596507e-01 1.04237199e+00
-8.52815866e-01 -3.98375690e-02 -9.88800600e-02 6.66126430e-01
3.42752993e-01 -4.82661635e-01 6.90273762e-01 -2.87731737e-01
-1.21432722e+00 3.36942762e-01 -1.34095418e+00 -1.21941358e-01
1.72979128e+00 2.08590701e-02 -2.66709715e-01 -9.27212775e-01
-7.14796960e-01 8.55804920e-01 6.22593105e-01 1.02208182e-01
5.50298989e-01 -6.74623251e-01 -6.46246076e-01 -6.06568344e-02
4.57770526e-02 -3.12788427e-01 2.49139309e-01 2.40499645e-01
-3.56844723e-01 3.48413289e-01 -4.95938599e-01 -2.44396091e-01
-9.79842961e-01 7.57938147e-01 6.87471390e-01 -6.69405013e-02
-3.73772413e-01 3.47707927e-01 -6.06097002e-03 -3.62384282e-02
4.00654256e-01 -5.36083151e-03 3.39105981e-03 -1.95565313e-01
6.73488617e-01 5.05867779e-01 -5.55195630e-01 1.11881793e-01
-3.85886222e-01 -1.82417914e-01 -6.11135438e-02 -8.02816629e-01
1.22231805e+00 -1.28828064e-01 1.90606505e-01 7.73903549e-01
-1.70333818e-01 -7.72915721e-01 -1.81248736e+00 -1.99808747e-01
2.64713377e-01 -4.88706261e-01 6.23282865e-02 -1.00646055e+00
-2.83672154e-01 4.63513047e-01 5.98091483e-01 6.90739155e-01
8.52206528e-01 2.08503264e-03 3.84095535e-02 8.31576645e-01
1.05241227e+00 -1.48296285e+00 1.64603591e-01 2.44147971e-01
4.97514486e-01 -9.76445913e-01 -9.22299474e-02 2.64229447e-01
-1.07418919e+00 9.08588052e-01 8.74728084e-01 -1.86115354e-01
3.75076711e-01 5.65894961e-01 -2.07622364e-01 -1.75818443e-01
-1.74532115e+00 -9.09775347e-02 -4.97139633e-01 7.76816547e-01
-5.35983443e-01 6.01258218e-01 2.44630113e-01 4.55766141e-01
1.22324444e-01 1.17637739e-01 1.09276617e+00 1.40586543e+00
-9.27323639e-01 -8.36386323e-01 -3.67325068e-01 4.63274181e-01
3.02527189e-01 4.03474987e-01 -4.69148546e-01 5.47057927e-01
-1.51951730e-01 1.19464314e+00 3.73063534e-01 -1.67055041e-01
4.51428257e-02 2.76925802e-01 4.69051123e-01 -8.84419382e-01
-6.97705969e-02 -1.77410617e-01 3.35676402e-01 -5.19898295e-01
-1.38900965e-01 -1.23813200e+00 -1.35244012e+00 -4.54751849e-01
1.97503254e-01 -5.31309359e-02 2.64452904e-01 1.02538311e+00
2.73704529e-01 3.39642555e-01 7.72504032e-01 -5.05104125e-01
-1.41034496e+00 -6.72586501e-01 -5.14669955e-01 2.00765021e-03
3.07943434e-01 -6.97976470e-01 -5.43443859e-01 -5.41342236e-02] | [4.3485026359558105, 2.158092737197876] |
da606d92-1a1f-4d67-8e93-9d2489d8bf6e | unsupervised-multi-object-segmentation-by | 2210.12148 | null | https://arxiv.org/abs/2210.12148v1 | https://arxiv.org/pdf/2210.12148v1.pdf | Unsupervised Multi-object Segmentation by Predicting Probable Motion Patterns | We propose a new approach to learn to segment multiple image objects without manual supervision. The method can extract objects form still images, but uses videos for supervision. While prior works have considered motion for segmentation, a key insight is that, while motion can be used to identify objects, not all objects are necessarily in motion: the absence of motion does not imply the absence of objects. Hence, our model learns to predict image regions that are likely to contain motion patterns characteristic of objects moving rigidly. It does not predict specific motion, which cannot be done unambiguously from a still image, but a distribution of possible motions, which includes the possibility that an object does not move at all. We demonstrate the advantage of this approach over its deterministic counterpart and show state-of-the-art unsupervised object segmentation performance on simulated and real-world benchmarks, surpassing methods that use motion even at test time. As our approach is applicable to variety of network architectures that segment the scenes, we also apply it to existing image reconstruction-based models showing drastic improvement. Project page and code: https://www.robots.ox.ac.uk/~vgg/research/ppmp . | ['Andrea Vedaldi', 'Christian Rupprecht', 'Iro Laina', 'Subhabrata Choudhury', 'Laurynas Karazija'] | 2022-10-21 | null | null | null | null | ['unsupervised-object-segmentation'] | ['computer-vision'] | [ 3.82829607e-01 4.76363488e-02 -3.96452606e-01 -3.24894220e-01
-5.58086336e-01 -9.40068364e-01 4.37519461e-01 -3.51081461e-01
-6.27108455e-01 5.47386646e-01 -1.50731236e-01 -3.11135978e-01
-5.08755594e-02 -6.07180297e-01 -1.11503255e+00 -9.68197465e-01
-7.44770393e-02 7.80606806e-01 7.23392427e-01 1.19972184e-01
7.52845928e-02 5.77690423e-01 -1.31985676e+00 2.90091962e-01
3.62857580e-01 4.57132548e-01 5.06124973e-01 9.92004633e-01
1.99891433e-01 8.22162330e-01 -3.89131606e-01 1.77078456e-01
4.74010438e-01 -5.55688441e-01 -1.52704191e+00 5.49818218e-01
7.12382615e-01 -6.36807203e-01 -2.90573061e-01 1.11119282e+00
-2.84244865e-02 3.20673674e-01 6.46872520e-01 -1.14744258e+00
-3.80645692e-01 5.29322267e-01 -4.96510834e-01 4.35759991e-01
2.32289851e-01 4.76090789e-01 9.58526909e-01 -4.06137645e-01
1.01542604e+00 1.17389500e+00 4.27391469e-01 7.93821216e-01
-1.35211265e+00 -1.79314718e-01 4.21729594e-01 1.51600853e-01
-8.87532234e-01 -4.16295260e-01 6.23611391e-01 -5.95446348e-01
7.83919334e-01 1.94234550e-01 7.65614152e-01 9.71016169e-01
2.49924660e-02 1.30104709e+00 7.27080524e-01 -1.67208254e-01
1.53559700e-01 -2.25249350e-01 1.69440940e-01 6.92675829e-01
2.48383686e-01 -2.45970953e-02 -1.27909899e-01 3.34467113e-01
9.07781422e-01 -5.91549650e-02 -4.17995721e-01 -7.35295355e-01
-1.68671930e+00 4.99557823e-01 3.13267142e-01 2.87515879e-01
-1.88288808e-01 5.29328465e-01 1.13090016e-01 6.91419542e-02
-3.17461267e-02 2.32550517e-01 -5.93106925e-01 -9.72327441e-02
-1.18013334e+00 4.27567959e-01 7.45069563e-01 9.70554650e-01
7.83350110e-01 7.69871240e-03 1.38545573e-01 2.85951585e-01
2.86786824e-01 3.50632548e-01 5.17800331e-01 -1.63646448e+00
1.91758960e-01 1.94181979e-01 4.48956966e-01 -7.22630620e-01
-4.88822222e-01 -3.82867386e-03 -5.50621688e-01 3.05269063e-01
9.41811800e-01 -1.65923119e-01 -1.22246242e+00 1.68114126e+00
2.86704898e-01 3.82035911e-01 1.12917356e-01 1.13968730e+00
6.48977757e-01 6.62266850e-01 -1.31508261e-01 -1.11837000e-01
9.20650542e-01 -1.19959235e+00 -4.65213478e-01 -4.61225122e-01
6.73048973e-01 -5.87499022e-01 7.14393139e-01 4.80171144e-01
-1.20675647e+00 -6.52745187e-01 -9.24424887e-01 6.33004159e-02
-1.24675035e-01 -1.27550215e-01 8.05695117e-01 4.26585883e-01
-1.13889682e+00 9.86888230e-01 -1.21946776e+00 -4.58595723e-01
5.30677080e-01 5.85461199e-01 -3.86948645e-01 -9.76011679e-02
-6.52803957e-01 6.00800991e-01 6.38824999e-01 1.87758371e-01
-1.24755287e+00 -1.29072830e-01 -7.55599260e-01 -3.13861489e-01
4.00940150e-01 -6.06733143e-01 1.53087127e+00 -1.43785119e+00
-1.18711305e+00 8.17063391e-01 -4.79998976e-01 -6.01031184e-01
7.73627043e-01 -3.47655416e-01 -6.30547851e-02 5.51246762e-01
8.25168043e-02 1.34846735e+00 7.02474475e-01 -1.38218296e+00
-7.65417099e-01 3.90977785e-02 1.43358782e-01 2.50008494e-01
1.95603490e-01 -1.25083700e-01 -7.25181639e-01 -4.29524332e-01
2.55446970e-01 -1.33608150e+00 -5.17078996e-01 3.71365845e-02
-6.47260964e-01 -1.10365793e-01 1.15566289e+00 -4.31445003e-01
5.02658725e-01 -1.80700350e+00 1.46983609e-01 -1.04922868e-01
-4.65987511e-02 2.11099580e-01 -2.33224779e-01 6.32637888e-02
1.81515682e-02 2.15311408e-01 -6.59798443e-01 -1.81903228e-01
-1.86260998e-01 4.79490012e-01 -2.93419272e-01 8.41918588e-01
3.01551849e-01 1.10576069e+00 -9.11154389e-01 -4.98555660e-01
3.61076057e-01 1.37507170e-01 -5.15189946e-01 -1.78738520e-01
-6.57087266e-01 7.75807679e-01 -3.74018788e-01 5.31872928e-01
5.60613334e-01 -4.60260808e-01 8.19720104e-02 1.61805734e-01
5.11849001e-02 1.12804458e-01 -1.41112518e+00 1.69288516e+00
2.57262290e-01 9.40537333e-01 6.97743371e-02 -1.27217507e+00
2.61040539e-01 2.85235554e-01 8.19251180e-01 -1.79463968e-01
-5.46114519e-02 6.86542764e-02 2.50577688e-01 -7.49964893e-01
4.95075285e-01 8.02044868e-02 2.56999493e-01 5.30997813e-01
8.43841359e-02 -2.61730075e-01 4.97077912e-01 1.61775798e-01
9.74442601e-01 7.80296743e-01 -1.22966908e-01 -2.63836414e-01
1.91163585e-01 5.01915216e-01 7.82664895e-01 9.63599861e-01
-4.33199525e-01 9.34517443e-01 3.45159441e-01 -3.48779231e-01
-9.54858840e-01 -9.88421082e-01 -1.38086617e-01 7.86640465e-01
8.36475194e-01 9.76397097e-02 -8.02883267e-01 -8.37537169e-01
-2.52113432e-01 4.60085064e-01 -4.66858357e-01 3.07688445e-01
-8.46726894e-01 -7.82593131e-01 2.82721609e-01 6.50277913e-01
3.76051426e-01 -1.35035896e+00 -1.00332201e+00 2.60959446e-01
-4.05756801e-01 -1.42526257e+00 -3.94199401e-01 2.46314526e-01
-1.07611442e+00 -1.26661479e+00 -6.19095445e-01 -1.03345883e+00
6.61789179e-01 4.97510374e-01 1.05553222e+00 7.91568235e-02
-2.50016630e-01 6.40907407e-01 -6.03005961e-02 -1.57497272e-01
-4.13539827e-01 -1.67323887e-01 -9.83401537e-02 -2.28851005e-01
2.89014786e-01 -2.61262178e-01 -9.06471312e-01 5.92831790e-01
-9.83482420e-01 1.22242048e-01 5.10139942e-01 4.72524613e-01
6.94464266e-01 2.73159891e-01 2.59358257e-01 -8.28036368e-01
-1.79729760e-01 -3.90346706e-01 -5.24903178e-01 7.53099052e-03
-6.15811720e-02 -3.19270566e-02 3.19809526e-01 -6.10032201e-01
-8.47574174e-01 7.06164539e-01 -4.50809151e-02 -4.21695918e-01
-8.22595060e-01 5.74567029e-03 1.13392724e-02 9.70307812e-02
3.54363471e-01 1.95363387e-01 -6.97963312e-02 -2.46900827e-01
4.54126358e-01 1.12195231e-01 7.12278605e-01 -3.76433551e-01
6.77284777e-01 1.15924561e+00 -1.17846616e-02 -9.86816645e-01
-6.25561953e-01 -8.74478161e-01 -1.06463635e+00 -2.82732785e-01
1.29016829e+00 -7.91515410e-01 -4.51088607e-01 3.62327725e-01
-1.23158455e+00 -7.51552522e-01 -2.68710911e-01 5.77386022e-01
-8.95825505e-01 5.30670226e-01 -7.23752797e-01 -6.56423569e-01
2.99283653e-01 -1.46253848e+00 9.23458993e-01 2.96232343e-01
-3.96969646e-01 -1.00674272e+00 -2.70602852e-01 4.24750090e-01
-1.07163452e-01 3.82670790e-01 4.15089160e-01 -6.70885026e-01
-1.25911570e+00 -2.67965230e-03 1.83049470e-01 2.54056036e-01
1.27423212e-01 2.93324053e-01 -7.94243157e-01 -3.18827868e-01
6.05517672e-03 -2.52548724e-01 1.05400598e+00 7.96844244e-01
1.05936182e+00 -2.15409577e-01 -5.82083106e-01 5.55509329e-01
1.44607389e+00 3.45547676e-01 6.41551852e-01 4.83562618e-01
1.02222121e+00 7.34999120e-01 5.37108421e-01 -1.05264708e-01
2.02745900e-01 6.01763964e-01 6.60008788e-01 -1.17008030e-01
-9.89272818e-02 8.73781592e-02 4.88872737e-01 2.44290441e-01
-8.62045586e-02 -4.45484489e-01 -9.58966017e-01 9.38095510e-01
-2.04263806e+00 -1.12295353e+00 -5.35589635e-01 1.93832004e+00
4.84212846e-01 2.76138872e-01 4.01572108e-01 -3.19425277e-02
6.47615850e-01 1.40664130e-01 -7.47362852e-01 -3.65538783e-02
-2.27604911e-01 -1.39071345e-01 7.41875291e-01 5.88506877e-01
-1.43121469e+00 1.12199855e+00 6.41879559e+00 3.60666871e-01
-1.03343248e+00 -2.01148182e-01 7.09717989e-01 -1.62267163e-01
-1.80576950e-01 2.77101517e-01 -7.88068831e-01 2.88780183e-01
6.73161328e-01 2.50182599e-01 2.58208275e-01 7.75680602e-01
4.42620188e-01 -3.92571926e-01 -1.33918178e+00 5.60435474e-01
-4.98826019e-02 -1.30419099e+00 -6.64951056e-02 5.58449328e-02
8.83626938e-01 3.29078883e-01 -7.15903267e-02 -2.21619576e-01
4.45550233e-01 -1.10229838e+00 1.06649768e+00 4.23945636e-01
1.74228474e-02 -4.62101340e-01 4.93865818e-01 6.60364509e-01
-9.06024992e-01 9.92263258e-02 -2.29111150e-01 3.90815958e-02
4.23700124e-01 2.53720403e-01 -7.99683392e-01 3.38308275e-01
8.24584723e-01 8.43913674e-01 -5.24031639e-01 1.15835786e+00
-2.33935773e-01 7.56180167e-01 -4.08452630e-01 1.47638023e-01
5.87995648e-01 -2.99870223e-01 7.62279034e-01 1.35110509e+00
1.15334131e-01 1.25308540e-02 4.14401233e-01 7.40061879e-01
1.69469640e-01 -3.49203050e-01 -6.97614074e-01 6.37715030e-03
1.05438381e-01 1.06975222e+00 -1.42817652e+00 -3.66696477e-01
-4.56218123e-01 1.22183979e+00 -8.13416690e-02 6.98137820e-01
-7.50254393e-01 -8.25850964e-02 4.65815663e-01 6.27808571e-02
8.18347931e-01 -5.73572695e-01 -3.48730050e-02 -1.01624322e+00
-8.83539245e-02 -6.89004421e-01 2.03291863e-01 -8.07765424e-01
-7.91870952e-01 1.51882634e-01 1.05487555e-01 -1.19088733e+00
-4.07391012e-01 -7.88188577e-01 -5.87459087e-01 3.45636398e-01
-1.22390079e+00 -8.65445554e-01 -7.98682198e-02 4.12725091e-01
8.97294044e-01 3.49029541e-01 2.73392498e-01 -4.98642698e-02
-4.02695715e-01 -8.01878944e-02 1.16810434e-01 5.02653956e-01
4.50127006e-01 -1.37901556e+00 4.33921069e-01 1.19713926e+00
6.22087061e-01 4.57871884e-01 8.37409139e-01 -5.96162379e-01
-1.39339709e+00 -1.19215417e+00 3.20556134e-01 -7.96738863e-01
4.66236889e-01 -1.72908455e-01 -9.83197391e-01 1.07444608e+00
3.45295548e-01 2.33988523e-01 1.83701232e-01 -4.60617125e-01
1.36566296e-01 4.41717118e-01 -8.56694818e-01 6.44038618e-01
1.08427334e+00 -3.77348699e-02 -5.99741518e-01 5.23927927e-01
7.23238885e-01 -4.58519369e-01 -4.61115688e-01 4.42775846e-01
4.62313026e-01 -1.07516623e+00 1.03272092e+00 -6.62702620e-01
4.75056678e-01 -6.53115273e-01 -6.89314231e-02 -8.30762684e-01
-1.85847074e-01 -6.13158822e-01 -1.53963253e-01 9.56548393e-01
3.68375063e-01 -3.21677119e-01 1.12959862e+00 5.92401505e-01
-3.79730351e-02 -5.89213908e-01 -7.33426452e-01 -9.57269728e-01
2.28375733e-01 -5.97974479e-01 1.02597877e-01 1.02701175e+00
-4.05106544e-01 1.36373609e-01 -2.43035734e-01 2.98281640e-01
6.72796726e-01 3.41238052e-01 7.62048423e-01 -8.47925127e-01
-5.43954134e-01 -5.58332384e-01 -4.91387904e-01 -1.56564176e+00
3.03889155e-01 -6.99296236e-01 5.37431836e-01 -1.89347160e+00
1.02295049e-01 -2.68819183e-01 -7.71226734e-02 4.27084893e-01
-3.15848514e-02 3.71832162e-01 2.06424952e-01 4.84773368e-01
-8.56102824e-01 4.98916544e-02 1.52423608e+00 -2.87501752e-01
-2.29168132e-01 1.94651484e-01 -2.91516870e-01 1.10286975e+00
1.05455792e+00 -5.03392279e-01 -4.60127145e-01 -5.67527294e-01
-2.02635691e-01 -3.98659706e-03 6.70560241e-01 -1.06518316e+00
2.33567998e-01 -4.59905982e-01 5.52004993e-01 -7.03797281e-01
2.77265996e-01 -7.30966628e-01 1.60599113e-01 6.97571814e-01
-2.76813865e-01 -7.22464845e-02 2.14490786e-01 8.03253949e-01
-2.90412083e-02 -5.35224617e-01 8.32326710e-01 -5.50874591e-01
-1.18884504e+00 2.59772331e-01 -7.61995852e-01 -5.33432551e-02
1.09669662e+00 -4.81715232e-01 -2.25841075e-01 -3.21198225e-01
-7.71736860e-01 3.29282284e-01 7.32672811e-01 4.17538494e-01
4.43421632e-01 -9.17759061e-01 -4.56876755e-01 -1.73714951e-01
-2.87741810e-01 5.24397612e-01 2.17891932e-02 8.36882412e-01
-8.21941435e-01 4.27236646e-01 2.25277003e-02 -1.13677275e+00
-1.16221023e+00 6.19844198e-01 4.45849061e-01 3.13051939e-01
-9.66060579e-01 7.85035729e-01 4.22243625e-01 -1.28095031e-01
1.68964073e-01 -5.85165203e-01 -7.22991228e-02 -2.59728312e-01
2.20291778e-01 1.30615517e-01 -2.82692164e-01 -7.95895636e-01
-2.84121215e-01 7.23716617e-01 -8.36438686e-02 -3.69190425e-01
1.26980650e+00 -2.05349624e-01 6.41195700e-02 7.18958914e-01
1.01188159e+00 -1.99526221e-01 -1.85084760e+00 8.50147456e-02
1.66231290e-01 -2.97150224e-01 -2.80963659e-01 -4.70520198e-01
-1.11509073e+00 8.04625273e-01 4.25894380e-01 3.20724785e-01
9.26741362e-01 1.93699494e-01 7.49044478e-01 5.56300819e-01
3.18479478e-01 -1.08949673e+00 2.21511588e-01 4.13491488e-01
4.65079397e-01 -1.42481744e+00 -6.42423779e-02 -3.57200980e-01
-7.90383935e-01 1.07610774e+00 6.16642773e-01 -3.00049216e-01
3.63283038e-01 2.24299446e-01 9.59552601e-02 -4.70128432e-02
-5.43823779e-01 -3.98725152e-01 1.93125635e-01 5.77856302e-01
1.77262753e-01 -2.24258557e-01 1.92895070e-01 -1.25446454e-01
-2.76819989e-02 -9.57464576e-02 8.04009974e-01 1.11180162e+00
-5.24307966e-01 -8.79848659e-01 -3.78432304e-01 2.89314598e-01
-6.34584665e-01 3.06293637e-01 -3.60609591e-01 1.06837380e+00
2.18557671e-01 8.44521105e-01 2.01309681e-01 -2.71015745e-02
-1.90337803e-02 -3.99008952e-02 7.09657967e-01 -5.35235584e-01
-1.53958857e-01 3.12619954e-01 -5.58159836e-02 -6.79859638e-01
-1.01199853e+00 -1.02783942e+00 -1.67141008e+00 1.17428817e-01
-2.53964901e-01 -5.76367490e-02 4.46563512e-01 1.04832542e+00
-4.53834347e-02 4.57114279e-01 1.43644407e-01 -1.26848209e+00
1.21408835e-01 -5.79969347e-01 -3.27576756e-01 5.43523967e-01
7.56160736e-01 -3.43898058e-01 -4.14983839e-01 7.45016873e-01] | [9.00213623046875, -0.4278172552585602] |
f1015411-d0d3-4f2c-895f-f7cee3728321 | vicunaner-zero-few-shot-named-entity | 2305.03253 | null | https://arxiv.org/abs/2305.03253v1 | https://arxiv.org/pdf/2305.03253v1.pdf | VicunaNER: Zero/Few-shot Named Entity Recognition using Vicuna | Large Language Models (LLMs, e.g., ChatGPT) have shown impressive zero- and few-shot capabilities in Named Entity Recognition (NER). However, these models can only be accessed via online APIs, which may cause data leak and non-reproducible problems. In this paper, we propose VicunaNER, a zero/few-shot NER framework based on the newly released open-source LLM -- Vicuna. VicunaNER is a two-phase framework, where each phase leverages multi-turn dialogues with Vicuna to recognize entities from texts. We name the second phase as Re-Recognition, which recognizes those entities not recognized in the first phase (a.k.a. Recognition). Moreover, we set entity correctness check dialogues in each phase to filter out wrong entities. We evaluate VicunaNER's zero-shot capacity on 10 datasets crossing 5 domains and few-shot capacity on Few-NERD. Experimental results demonstrate that VicunaNER achieves superior performance in both shot settings. Additionally, we conduct comprehensive investigations on Vicuna from multiple perspectives. | ['Bin Ji'] | 2023-05-05 | null | null | null | null | ['few-shot-ner', 'named-entity-recognition-ner'] | ['natural-language-processing', 'natural-language-processing'] | [-3.91940773e-01 2.05442861e-01 1.45851495e-02 -2.06021249e-01
-1.18451095e+00 -7.54223824e-01 6.26174867e-01 -1.66964196e-02
-7.94985533e-01 6.43161535e-01 3.54875296e-01 -2.10559145e-01
1.90969422e-01 -6.86220288e-01 -4.26446497e-01 -2.18591988e-01
2.71862775e-01 6.44525528e-01 5.15784979e-01 -3.36243689e-01
1.06953174e-01 2.83833146e-02 -8.47934306e-01 4.01563615e-01
1.13862038e+00 2.69207865e-01 -1.11221805e-01 7.48414159e-01
-7.04889178e-01 1.04569972e+00 -7.30287671e-01 -9.47525322e-01
7.25950301e-03 -1.04156926e-01 -1.32861817e+00 -5.41007578e-01
-1.79335028e-02 -2.03110382e-01 -3.39603752e-01 9.25388932e-01
7.65644789e-01 4.14313525e-01 2.90777206e-01 -1.30231822e+00
-7.41197944e-01 1.05703914e+00 -1.94111571e-01 1.75212339e-01
3.00713301e-01 3.08518201e-01 1.08247185e+00 -1.39001644e+00
8.26178551e-01 1.11662328e+00 1.00223088e+00 8.54392886e-01
-9.21621203e-01 -5.40864646e-01 -2.80033082e-01 -2.83881333e-02
-1.34289384e+00 -8.45243871e-01 1.75657675e-01 -2.85903931e-01
1.34710288e+00 4.20013428e-01 -1.04249910e-01 1.09314680e+00
-1.80809975e-01 9.66671526e-01 8.81118953e-01 -4.09249008e-01
2.38817602e-01 9.11282375e-02 6.81137562e-01 5.28545201e-01
8.24426264e-02 -4.80561048e-01 -5.19909799e-01 -4.26631212e-01
2.78334200e-01 -1.19473755e-01 -1.11627780e-01 8.09416100e-02
-1.09203923e+00 7.08235145e-01 2.23115385e-02 5.02783477e-01
-5.62195122e-01 -3.04673523e-01 6.48478150e-01 3.24145854e-01
4.53201056e-01 7.22895086e-01 -7.44101822e-01 -7.05107689e-01
-8.21371377e-01 -5.39485663e-02 1.40266955e+00 1.23612905e+00
6.29530668e-01 -3.06473196e-01 -7.41386712e-01 1.07774091e+00
1.68742165e-01 3.89090687e-01 5.90514719e-01 -5.01848459e-01
8.30662966e-01 8.86940360e-01 4.29246813e-01 -6.48501277e-01
-4.62416172e-01 1.45706534e-01 -7.00576186e-01 -6.49009168e-01
3.32143068e-01 -5.82701027e-01 -8.15309584e-01 1.56536853e+00
5.02169371e-01 4.03144598e-01 6.42800987e-01 6.16526306e-01
1.51001525e+00 7.82538235e-01 4.95858669e-01 -1.34241357e-02
1.59629345e+00 -1.31854272e+00 -8.57963324e-01 -1.85259745e-01
1.14930773e+00 -6.91672385e-01 1.09083354e+00 -1.07192636e-01
-8.52007866e-01 -3.93050984e-02 -4.40472066e-01 -2.42841512e-01
-6.33987308e-01 3.41405034e-01 3.07433069e-01 6.29952610e-01
-6.93894565e-01 5.81416845e-01 -8.49372029e-01 -5.77790141e-01
4.60877828e-03 -5.24912365e-02 -5.39536774e-01 8.15472230e-02
-1.84595537e+00 1.00096333e+00 4.57639068e-01 2.00587258e-01
-8.18057775e-01 -7.36788750e-01 -9.14037526e-01 2.13113129e-01
7.30209291e-01 -8.36633965e-02 1.79099607e+00 -1.56831801e-01
-1.44488215e+00 8.06797564e-01 -3.06764215e-01 -4.72891271e-01
4.87709135e-01 -5.30610681e-01 -8.02086771e-01 -1.65766597e-01
4.10299689e-01 1.87878385e-01 2.25556530e-02 -8.47815573e-01
-5.38575113e-01 -3.70338187e-02 1.18471749e-01 1.06174417e-01
-2.85062730e-01 4.11190897e-01 -5.88921547e-01 -2.14861825e-01
-4.86084372e-01 -7.16138840e-01 -2.35909283e-01 -7.97391176e-01
-6.70646310e-01 -6.39094830e-01 4.43512201e-01 -8.85998487e-01
1.87021923e+00 -2.00803542e+00 -3.60570550e-01 -1.52273387e-01
4.13244665e-01 9.42300320e-01 -3.68574023e-01 9.91542578e-01
1.75069794e-01 6.25317514e-01 -8.27007964e-02 -4.30782497e-01
2.50509709e-01 1.33689895e-01 -1.85923427e-01 2.04962920e-02
2.01098368e-01 1.21560061e+00 -1.06408334e+00 -5.75726986e-01
-1.77913774e-02 2.32034713e-01 -2.25473121e-01 3.81185085e-01
5.10899611e-02 -1.06494091e-01 -5.90850890e-01 5.38128436e-01
6.54618084e-01 -3.92552763e-01 2.74219722e-01 -3.35806310e-02
-4.09219831e-01 7.18936980e-01 -1.17222524e+00 1.48918128e+00
-4.53510761e-01 3.09167683e-01 -5.95883131e-02 -3.07905674e-01
8.06270063e-01 7.33272612e-01 3.35023552e-02 -4.34333056e-01
9.43401232e-02 3.30333918e-01 -3.82396370e-01 -6.45424426e-01
1.15760577e+00 2.14590088e-01 -4.69046593e-01 4.24337059e-01
5.24266839e-01 7.03496099e-01 3.84050220e-01 6.29258156e-01
1.39305341e+00 -2.44751260e-01 6.15919411e-01 1.31274074e-01
3.81904513e-01 3.14330727e-01 1.00622785e+00 9.86081004e-01
-4.51774299e-01 2.37154678e-01 5.51761985e-01 -5.83777279e-02
-8.58238935e-01 -6.55950904e-01 2.00448647e-01 1.41214299e+00
1.82008147e-01 -8.13854992e-01 -1.03385520e+00 -1.07837772e+00
-3.26855600e-01 1.00701439e+00 -4.91172403e-01 1.31154647e-02
-4.26981866e-01 -7.31082857e-01 1.41405594e+00 3.65160584e-01
7.65061378e-01 -1.25791395e+00 -3.84590805e-01 4.47788924e-01
-5.41357398e-01 -1.08442318e+00 -6.60920501e-01 4.09416296e-02
-2.35856354e-01 -1.14002287e+00 -6.38743341e-01 -6.37469769e-01
2.08919793e-01 3.38503689e-01 1.21220577e+00 -1.46406874e-01
3.86589998e-03 3.84329081e-01 -8.30490232e-01 -1.32865891e-01
-5.95015705e-01 5.62885404e-01 -5.27540669e-02 -1.40762672e-01
9.97853637e-01 -1.54151648e-01 -2.93697208e-01 5.66506386e-01
-6.84004724e-01 -5.41392863e-02 5.18479943e-01 8.89571846e-01
1.73986703e-01 -3.99997562e-01 6.95814550e-01 -1.28417420e+00
8.76428127e-01 -8.55040550e-01 -2.73199409e-01 9.06393409e-01
-5.20945430e-01 -3.37184221e-02 5.68625808e-01 -4.19134051e-01
-1.36726427e+00 -3.37097377e-01 -2.51954496e-01 -2.09353060e-01
-2.27162004e-01 6.55247927e-01 -3.30368727e-01 3.18545431e-01
6.62542582e-01 2.48624623e-01 -5.98962128e-01 -9.62001085e-01
5.93319893e-01 1.20269287e+00 5.37065864e-01 -4.88276631e-01
7.28536785e-01 -1.53550193e-01 -1.04313767e+00 -8.34870279e-01
-9.29014564e-01 -1.00466752e+00 -5.90700567e-01 -1.06024049e-01
9.75146532e-01 -1.01259398e+00 -6.72758222e-01 5.91666818e-01
-1.47474957e+00 -2.82264948e-01 -2.28104860e-01 2.69378334e-01
9.51389372e-02 3.70869339e-01 -9.56740260e-01 -9.02012110e-01
-8.30266953e-01 -6.25088632e-01 8.27336013e-01 7.52633750e-01
-1.83036909e-01 -8.58804226e-01 4.85026419e-01 3.21515173e-01
3.37297440e-01 -2.51304448e-01 4.00191247e-01 -1.71772361e+00
-4.79050390e-02 -8.23427588e-02 -1.36168733e-01 -3.87998782e-02
-1.33075744e-01 -1.28309548e-01 -9.97490644e-01 -9.05015916e-02
-4.53286469e-01 -4.79310513e-01 5.12795091e-01 -3.67687613e-01
3.74825656e-01 -5.70888460e-01 -3.66660982e-01 2.80145258e-01
1.08892167e+00 2.16430292e-01 7.68810451e-01 4.84534651e-01
6.50046110e-01 4.00714248e-01 7.53857434e-01 3.61538202e-01
7.12465167e-01 3.80445421e-01 -1.43178746e-01 -8.65626261e-02
1.98473021e-01 -5.81899464e-01 4.56294298e-01 1.24609387e+00
6.54997155e-02 -4.43807453e-01 -1.31786549e+00 7.74619281e-01
-1.95684731e+00 -1.15863264e+00 -2.98310548e-01 1.84674239e+00
1.06255734e+00 -1.04846075e-01 -4.51750606e-02 -5.70589304e-01
1.12690222e+00 1.82062015e-01 -5.36996245e-01 -3.50141883e-01
-1.29052162e-01 8.51851702e-02 4.12930071e-01 2.03643143e-01
-1.12248445e+00 1.39849877e+00 5.50286770e+00 9.85883117e-01
-6.65073693e-01 5.46202004e-01 1.81868687e-01 3.26193094e-01
-1.27993941e-01 2.92392701e-01 -1.27354741e+00 5.04664123e-01
1.32048321e+00 -6.36618018e-01 9.68198851e-02 1.09963334e+00
-1.16767921e-01 2.19536796e-01 -6.79987133e-01 7.34703660e-01
-6.23967461e-02 -1.19006228e+00 -1.47065297e-01 -1.73614800e-01
4.85172719e-01 5.32051980e-01 -7.36047029e-01 1.15490890e+00
1.04032946e+00 -7.35735178e-01 3.62987459e-01 5.39399862e-01
7.32149780e-01 -6.45182490e-01 9.27254498e-01 7.11836994e-01
-1.29838848e+00 3.03676814e-01 -4.45821762e-01 4.74098921e-01
4.60580826e-01 4.06777591e-01 -1.02135122e+00 8.64051700e-01
6.40281320e-01 3.26798886e-01 -4.80169535e-01 1.08819902e+00
-3.19420367e-01 8.54501843e-01 -2.12386847e-01 -2.05974057e-01
3.41520429e-01 7.04726651e-02 6.73250198e-01 1.69546795e+00
1.01031318e-01 4.85221773e-01 3.22751999e-01 6.86714411e-01
-6.45906031e-01 2.97082067e-01 -3.67344469e-01 -3.91355425e-01
1.16046154e+00 1.71363604e+00 -3.82245898e-01 -6.42629802e-01
-5.72380126e-01 1.19646263e+00 6.41855180e-01 1.58863783e-01
-7.94175744e-01 -1.02714014e+00 5.31487048e-01 -3.69933844e-01
1.36242926e-01 -4.84320037e-02 2.91875988e-01 -1.70435452e+00
-2.00137407e-01 -9.00882185e-01 6.53898001e-01 -5.80428720e-01
-1.51845300e+00 8.85725975e-01 -3.12744975e-01 -9.90174830e-01
-6.31016269e-02 -2.26441890e-01 -8.25850964e-01 7.01672316e-01
-1.37156415e+00 -1.08493543e+00 -1.73276544e-01 4.50409740e-01
6.74723685e-01 1.50599657e-02 1.07160640e+00 4.75326329e-01
-1.10801446e+00 8.04287672e-01 2.52951384e-01 9.95718718e-01
1.02834201e+00 -1.34264648e+00 1.03353667e+00 1.01291561e+00
1.73743188e-01 9.35918331e-01 2.82536149e-01 -8.61671984e-01
-1.16034043e+00 -1.27797055e+00 1.43162560e+00 -7.03099549e-01
8.20880413e-01 -4.81145620e-01 -1.33135438e+00 8.80808532e-01
2.60692924e-01 -9.51138958e-02 9.62738872e-01 3.22621942e-01
-4.27254915e-01 4.45699662e-01 -1.09684908e+00 5.96673727e-01
8.13807845e-01 -6.24647379e-01 -1.23368251e+00 3.88517380e-02
9.97961283e-01 -5.24538457e-01 -1.12425160e+00 1.07705072e-01
2.12346479e-01 -5.52223802e-01 4.38552678e-01 -1.10250616e+00
9.73889828e-02 -1.70275509e-01 4.70209755e-02 -1.32763100e+00
-7.86120519e-02 -1.05589175e+00 -1.90536231e-01 2.14540505e+00
7.68359065e-01 -6.92930639e-01 2.75068104e-01 1.01049376e+00
-6.63683861e-02 -1.64443433e-01 -9.37725306e-01 -9.56772447e-01
-8.92970264e-02 -2.07835734e-01 8.10520291e-01 1.46711230e+00
2.86977410e-01 7.96455085e-01 -6.49355352e-01 1.96936056e-01
2.33100235e-01 -1.39058426e-01 9.20377076e-01 -1.03657961e+00
-2.61253715e-02 1.50311053e-01 1.28307879e-01 -1.04734886e+00
6.57298863e-02 -9.22049224e-01 3.42206895e-01 -1.48438668e+00
4.66844410e-01 -4.18652058e-01 -2.82419324e-01 1.02412772e+00
-4.62084830e-01 -3.10321599e-01 2.29224712e-01 4.74074394e-01
-1.29754961e+00 6.98001027e-01 4.94219869e-01 2.04521045e-01
-4.90444005e-01 -3.08262676e-01 -5.98177791e-01 6.75646305e-01
7.49651670e-01 -8.07423949e-01 1.67013869e-01 -4.69624907e-01
1.78557158e-01 6.76607639e-02 1.04427468e-02 -6.85428858e-01
5.75220048e-01 -1.61570609e-01 -2.33240664e-01 -5.02922237e-01
-3.92452292e-02 -2.92935461e-01 -4.46049571e-02 -8.64446349e-03
-4.54005927e-01 1.06338076e-01 5.55776991e-04 4.57129985e-01
-1.85485482e-01 -4.09718633e-01 4.50581789e-01 -1.51113927e-01
-1.20121527e+00 2.18005165e-01 -3.35098118e-01 7.82206357e-01
7.66804755e-01 2.09551409e-01 -7.63689280e-01 -3.87921482e-02
-7.28755832e-01 7.38982499e-01 1.65461540e-01 6.02364361e-01
1.97774604e-01 -1.25914013e+00 -7.09477127e-01 -3.28664333e-01
3.91988665e-01 -1.34844974e-01 4.97438550e-01 9.87507582e-01
-1.57570645e-01 3.13572049e-01 2.52491862e-01 -1.79763764e-01
-1.35919559e+00 3.23301196e-01 3.32427710e-01 -8.42022479e-01
-6.77006543e-01 6.90083683e-01 -1.54997170e-01 -1.15387654e+00
-1.14133507e-02 5.87365963e-02 -4.77465957e-01 1.50530413e-01
8.80286098e-01 6.55567825e-01 -2.11895108e-02 -6.63172901e-01
-5.94828069e-01 -8.87778252e-02 -3.99627268e-01 -1.35755047e-01
1.39552522e+00 -3.55907798e-01 -6.94600567e-02 6.29444838e-01
7.83400297e-01 2.78134078e-01 -7.71800697e-01 -5.71455538e-01
6.52970135e-01 -7.55457580e-02 -2.42952541e-01 -1.12643743e+00
-5.90728343e-01 5.94268858e-01 1.05835751e-01 2.12841302e-01
5.83145559e-01 -1.13475487e-01 1.30508196e+00 7.65412152e-01
4.32215840e-01 -1.25750768e+00 -4.30375695e-01 1.16648841e+00
3.46545964e-01 -1.30869401e+00 -5.83762467e-01 -3.12460691e-01
-1.14040387e+00 8.32260728e-01 9.25559819e-01 3.92735928e-01
3.11819255e-01 1.01615004e-01 2.19198942e-01 -2.96056718e-01
-1.09052539e+00 -3.26959342e-01 1.13234669e-02 1.94740981e-01
4.32605207e-01 1.19675428e-01 -4.85005975e-01 1.45774257e+00
1.57713711e-01 9.17299911e-02 5.89951038e-01 1.03015137e+00
-5.18781245e-01 -9.72267568e-01 -7.35373273e-02 1.86732009e-01
-7.05070436e-01 -5.18169224e-01 -7.29409039e-01 8.72997761e-01
-1.66496992e-01 1.20607471e+00 -2.73115695e-01 -5.42523801e-01
7.73475051e-01 5.48598111e-01 -4.07340229e-01 -7.99108326e-01
-1.11033511e+00 -2.57838368e-01 7.01443613e-01 -4.84147936e-01
-1.25305533e-01 -3.32974792e-01 -1.41736007e+00 -4.96457726e-01
-6.32650971e-01 9.41890001e-01 3.19307119e-01 7.79515386e-01
7.61137903e-01 9.77187827e-02 4.80266452e-01 -8.09746236e-02
-7.50134349e-01 -1.28276861e+00 -4.68194187e-01 4.17651713e-01
-1.20060354e-01 -4.16223407e-01 -3.32394898e-01 -2.27199033e-01] | [9.664621353149414, 9.376605987548828] |
9e16d8a8-1a9c-45ca-a9f6-8a306a25a093 | relation-graph-network-for-3d-object | 1912.00202 | null | https://arxiv.org/abs/1912.00202v1 | https://arxiv.org/pdf/1912.00202v1.pdf | Relation Graph Network for 3D Object Detection in Point Clouds | Convolutional Neural Networks (CNNs) have emerged as a powerful strategy for most object detection tasks on 2D images. However, their power has not been fully realised for detecting 3D objects in point clouds directly without converting them to regular grids. Existing state-of-art 3D object detection methods aim to recognize 3D objects individually without exploiting their relationships during learning or inference. In this paper, we first propose a strategy that associates the predictions of direction vectors and pseudo geometric centers together leading to a win-win solution for 3D bounding box candidates regression. Secondly, we propose point attention pooling to extract uniform appearance features for each 3D object proposal, benefiting from the learned direction features, semantic features and spatial coordinates of the object points. Finally, the appearance features are used together with the position features to build 3D object-object relationship graphs for all proposals to model their co-existence. We explore the effect of relation graphs on proposals' appearance features enhancement under supervised and unsupervised settings. The proposed relation graph network consists of a 3D object proposal generation module and a 3D relation module, makes it an end-to-end trainable network for detecting 3D object in point clouds. Experiments on challenging benchmarks ( SunRGB-Dand ScanNet datasets ) of 3D point clouds show that our algorithm can perform better than the existing state-of-the-art methods. | ['Ajmal Mian', 'Syed Zulqarnain Gilani', 'Mingtao Feng', 'Yaonan Wang', 'Liang Zhang'] | 2019-11-30 | null | null | null | null | ['object-proposal-generation'] | ['computer-vision'] | [-1.37676641e-01 1.06950186e-01 -1.07328542e-01 -4.75418687e-01
-3.32791239e-01 -3.18652272e-01 8.49401891e-01 2.06896365e-01
-2.75549620e-01 -2.57478595e-01 -3.91668707e-01 -3.22027534e-01
5.25620580e-02 -7.90901899e-01 -1.04152060e+00 -3.72172982e-01
-3.08077931e-01 8.44981968e-01 7.44109929e-01 -1.01527229e-01
3.48213434e-01 1.33215976e+00 -1.62207079e+00 -2.66054738e-02
5.59604168e-01 1.37692034e+00 3.54406744e-01 2.60981143e-01
-4.44703221e-01 2.86887109e-01 -3.72873396e-01 -8.45021755e-02
6.67826593e-01 2.88473696e-01 -2.25866467e-01 3.10177833e-01
7.10874617e-01 -4.68228042e-01 -3.17138672e-01 8.84227395e-01
3.60248387e-01 -4.97795455e-02 1.00186908e+00 -1.30173528e+00
-5.93033075e-01 1.35085732e-02 -1.01787484e+00 2.92448878e-01
3.09933536e-02 1.98989719e-01 9.78038430e-01 -1.29881799e+00
4.65407729e-01 1.49737811e+00 5.77761173e-01 3.02843839e-01
-1.08245301e+00 -7.71707118e-01 2.87916303e-01 -1.68110617e-02
-1.46294236e+00 2.19538505e-03 1.16260350e+00 -5.44413507e-01
1.46201444e+00 6.83621317e-02 8.12467217e-01 5.83542943e-01
-8.81422982e-02 9.29721713e-01 6.66138232e-01 -3.47837716e-01
-5.73392697e-02 1.21278554e-01 8.47584754e-02 8.51064861e-01
3.78723919e-01 8.24254006e-02 -2.52966672e-01 -1.30627736e-01
1.09559870e+00 3.23284388e-01 8.00654888e-02 -8.95332634e-01
-9.93171692e-01 7.81004906e-01 1.23178065e+00 6.06979765e-02
-5.62384009e-01 1.05800644e-01 1.95812288e-04 -5.68394326e-02
8.02371025e-01 2.17425659e-01 -4.42393124e-01 5.93221307e-01
-5.02878845e-01 5.52884817e-01 2.05731586e-01 1.24073911e+00
1.02214885e+00 -2.78520912e-01 -1.62064359e-01 6.75520599e-01
6.04584932e-01 5.86780012e-01 -4.03067507e-02 -3.49902660e-01
5.23538589e-01 1.35471690e+00 1.41677931e-01 -1.08980811e+00
-5.42273939e-01 -4.94790465e-01 -6.25180364e-01 5.63392460e-01
6.29818290e-02 4.97158617e-01 -1.25858796e+00 1.25554872e+00
8.25054109e-01 4.16584224e-01 -3.66538078e-01 1.16894031e+00
1.18130803e+00 6.18566573e-01 -3.02507281e-01 4.18849498e-01
1.35993302e+00 -6.45753980e-01 1.45561188e-01 -3.21910560e-01
3.61444175e-01 -6.82155430e-01 7.39228547e-01 -2.08934277e-01
-1.17648613e+00 -7.85736799e-01 -8.75043631e-01 -1.36458009e-01
-4.34518337e-01 2.60158509e-01 6.41630352e-01 2.86590129e-01
-6.33807361e-01 3.22370589e-01 -8.62486482e-01 -2.20006287e-01
1.13970542e+00 5.89297593e-01 -3.11933756e-01 4.85723984e-04
-3.89902502e-01 9.12420630e-01 5.05619705e-01 1.36454538e-01
-8.71591568e-01 -7.65337527e-01 -8.74066472e-01 9.64144319e-02
2.60529548e-01 -8.25443029e-01 1.07772660e+00 -3.96551609e-01
-1.02953339e+00 1.43520105e+00 4.05379348e-02 -2.82043755e-01
4.35387164e-01 -1.37600735e-01 9.69044268e-02 -1.03925891e-01
2.19074756e-01 1.00119567e+00 9.93950367e-01 -1.30223811e+00
-1.09398770e+00 -7.14481533e-01 1.53006790e-02 3.29090357e-01
9.71191823e-02 -8.88219774e-02 -7.63167500e-01 -1.52665138e-01
8.03088963e-01 -8.20516765e-01 -2.45089382e-01 3.39603335e-01
-6.44510686e-01 -9.61395383e-01 1.07174456e+00 1.15327261e-01
3.32445890e-01 -2.27207875e+00 -1.64843425e-01 3.05975795e-01
4.15820181e-01 1.74019098e-01 -1.43946007e-01 -1.21626936e-01
5.02463942e-03 2.64612250e-02 1.25068232e-01 -4.61825013e-01
1.85283855e-01 -1.09674418e-02 -2.98030317e-01 6.99076653e-01
8.89187455e-01 1.17179620e+00 -7.25208104e-01 -4.05746907e-01
6.39694393e-01 5.68521082e-01 -4.92360950e-01 4.90151227e-01
-3.84082675e-01 1.04694515e-01 -7.55572140e-01 8.43078613e-01
1.07695603e+00 -3.81713361e-01 -5.52657604e-01 -2.02572539e-01
-1.13509625e-01 3.32990825e-01 -1.04263163e+00 1.43822753e+00
-2.77738363e-01 4.71208572e-01 -2.61037916e-01 -8.48172545e-01
1.47818375e+00 -9.77510735e-02 3.45260680e-01 -4.50825721e-01
8.87778252e-02 2.33125299e-01 -1.48032889e-01 -3.01649421e-01
2.15854868e-01 8.00555050e-02 1.18034907e-01 2.00198382e-01
1.19338259e-01 -7.23334730e-01 -2.88575590e-01 4.28352412e-03
9.13828790e-01 3.26369219e-02 1.75781831e-01 1.13767304e-01
3.16725224e-01 8.64598975e-02 2.46917337e-01 9.11005974e-01
-1.07575394e-03 7.59760201e-01 3.84508580e-01 -8.39591384e-01
-1.19035697e+00 -1.07249975e+00 -3.33620340e-01 6.43897653e-01
5.00742197e-01 -5.24117239e-02 -4.38954495e-02 -1.00281954e+00
4.72334504e-01 4.25745189e-01 -6.47459447e-01 2.09220853e-02
-6.46184742e-01 -5.02687275e-01 9.42119434e-02 5.05426526e-01
3.52234960e-01 -9.71803486e-01 -8.95076334e-01 4.21440639e-02
5.01676798e-01 -1.19336486e+00 -2.16987208e-01 5.23569465e-01
-1.05150938e+00 -1.03732359e+00 -4.63391393e-01 -8.52143943e-01
1.04413164e+00 7.57851422e-01 1.15265691e+00 2.05828547e-01
-4.34585661e-01 2.49272823e-01 -2.38842383e-01 -8.95749211e-01
3.96242030e-02 5.80600724e-02 -3.18300575e-02 -6.91682324e-02
6.94906116e-01 -4.97731477e-01 -7.84135163e-01 4.46019888e-01
-5.93918860e-01 1.13384880e-01 9.49272931e-01 6.29037440e-01
8.27235401e-01 -1.71219230e-01 9.61945485e-03 -6.23561442e-01
1.21674508e-01 -3.33766311e-01 -8.57863426e-01 -1.09982036e-01
-2.14948505e-01 -6.19133301e-02 4.60418314e-02 -3.94961238e-01
-6.22921228e-01 4.80758220e-01 5.54274507e-02 -1.02342677e+00
-5.47217309e-01 1.05689075e-02 -8.77466723e-02 -2.65927464e-01
5.32574773e-01 1.99095085e-01 -2.50536919e-01 -5.51018894e-01
3.66137534e-01 4.93056059e-01 2.81241804e-01 -3.52428287e-01
1.28544664e+00 6.47390664e-01 1.91358402e-01 -8.57254922e-01
-8.75603795e-01 -8.41055632e-01 -9.39952910e-01 -1.91816300e-01
9.32003558e-01 -1.08066845e+00 -6.71364665e-01 1.49107128e-01
-1.57575011e+00 6.54826611e-02 -2.44053721e-01 4.38331068e-01
-3.36185306e-01 -3.30335572e-02 -1.88548505e-01 -8.94245505e-01
-2.76611596e-01 -1.11954725e+00 1.68269145e+00 2.87198722e-01
2.53146857e-01 -4.24171716e-01 -1.90170988e-01 1.48750931e-01
1.15351096e-01 2.37236649e-01 1.03810167e+00 -7.25213289e-01
-1.16897142e+00 -6.10825658e-01 -7.46094644e-01 8.29917118e-02
-4.91927527e-02 -4.41891328e-02 -9.68196750e-01 -5.14860637e-02
-1.10558055e-01 -1.02126800e-01 9.48744833e-01 5.30148208e-01
1.20330310e+00 2.09404994e-03 -7.20434129e-01 6.72572434e-01
1.31054974e+00 -8.50709230e-02 2.22436622e-01 1.44234926e-01
7.94974446e-01 4.46022660e-01 5.64744294e-01 4.03506398e-01
2.03289673e-01 7.39111841e-01 1.09043419e+00 -2.67161906e-01
2.78987605e-02 -3.27585071e-01 -1.30502015e-01 3.40747744e-01
-2.93329209e-01 -9.73632485e-02 -1.04373550e+00 5.07425845e-01
-1.64269400e+00 -5.19404829e-01 -3.79104644e-01 1.96205330e+00
1.69853866e-01 6.59699619e-01 7.00936913e-02 -1.84464395e-01
6.93700731e-01 5.64417839e-02 -6.36502862e-01 1.64799914e-01
-5.38661778e-02 2.50068426e-01 4.66734797e-01 -9.57265273e-02
-1.31790817e+00 9.93755519e-01 4.63971567e+00 7.05601156e-01
-1.06398463e+00 -1.17290102e-01 4.67407346e-01 -1.68204363e-02
-2.62990617e-03 -8.11510757e-02 -1.20392811e+00 2.40597073e-02
-1.17261007e-01 3.33778203e-01 -2.91797996e-01 1.23534119e+00
2.22740676e-02 2.76721776e-01 -1.18622243e+00 1.04637396e+00
-1.13246618e-02 -1.25544834e+00 2.54121512e-01 3.83762084e-02
6.18571639e-01 4.49739546e-01 -5.17546432e-03 2.37759277e-01
1.03278615e-01 -8.46733212e-01 9.57911491e-01 2.31641263e-01
3.63600194e-01 -6.92131639e-01 7.30797827e-01 5.20885766e-01
-1.27658129e+00 3.24395194e-04 -8.27662230e-01 1.21504925e-02
8.60071462e-03 5.82367420e-01 -1.51870203e+00 3.67451191e-01
9.31798756e-01 8.39652717e-01 -6.97172165e-01 1.46381247e+00
-4.31180865e-01 2.96920300e-01 -7.61753380e-01 -3.37286651e-01
5.75011432e-01 1.28025766e-02 7.84149468e-01 1.02749312e+00
4.12883461e-01 1.55028835e-01 2.17832908e-01 1.44228578e+00
-1.30932033e-01 2.59700194e-02 -8.43155801e-01 2.61877596e-01
5.21287799e-01 1.30619752e+00 -1.14188242e+00 -1.21937275e-01
-4.84311849e-01 5.54005504e-01 5.21510601e-01 -2.54785400e-02
-6.22174442e-01 -2.70287573e-01 7.54024804e-01 3.71502280e-01
8.69104028e-01 -3.48640591e-01 -1.93468645e-01 -8.00723374e-01
1.18172795e-01 -2.86423922e-01 1.12595014e-01 -9.93978143e-01
-1.55159283e+00 5.39234936e-01 5.64896055e-02 -1.39639950e+00
2.59127170e-01 -8.77708375e-01 -7.92013407e-01 7.67379165e-01
-1.71872723e+00 -1.48756540e+00 -5.82848370e-01 3.62302572e-01
5.34377754e-01 -1.16177052e-01 3.26842278e-01 9.81107703e-04
-3.10283870e-01 1.31568015e-01 -4.16506559e-01 2.04901576e-01
2.46314973e-01 -1.23590612e+00 8.57910633e-01 4.27896470e-01
5.49724817e-01 3.10205579e-01 1.81032583e-01 -6.39925838e-01
-1.39263856e+00 -1.28426790e+00 6.20216370e-01 -8.47979605e-01
2.60315806e-01 -7.88328528e-01 -9.55279350e-01 4.56694037e-01
-3.61094952e-01 4.39646155e-01 -1.12900874e-02 8.38855058e-02
-4.00655985e-01 -1.55464020e-02 -7.68634260e-01 3.36771250e-01
1.26776791e+00 -2.37623468e-01 -6.22118056e-01 5.09362936e-01
8.31484318e-01 -7.59224057e-01 -4.46398169e-01 7.47954726e-01
1.73123658e-01 -9.43844378e-01 1.53342223e+00 -6.35796726e-01
3.76572400e-01 -6.11542583e-01 5.69071025e-02 -8.33580613e-01
-4.28846061e-01 -1.11889899e-01 -2.90285088e-02 7.64183342e-01
3.35520834e-01 -2.33485505e-01 9.63325799e-01 2.57515937e-01
-4.87812191e-01 -8.24462354e-01 -1.16370821e+00 -6.49789929e-01
-3.31444740e-01 -7.24566340e-01 8.39771509e-01 7.55319476e-01
-7.57870257e-01 4.62181181e-01 1.43078357e-01 6.58132493e-01
7.41421163e-01 5.84778130e-01 1.22320974e+00 -1.56562662e+00
4.42746654e-02 -9.34616148e-01 -9.84546900e-01 -1.47950029e+00
3.81620228e-02 -1.12121272e+00 3.25027108e-02 -1.41217566e+00
9.20329913e-02 -9.26825881e-01 -2.57718354e-01 4.84467804e-01
-1.72684833e-01 3.89810950e-01 2.41955802e-01 2.95422435e-01
-6.55728102e-01 6.57195091e-01 1.26158142e+00 -3.66041780e-01
-4.37653363e-01 3.48094463e-01 -1.39824733e-01 7.07027912e-01
4.06722546e-01 -6.43652558e-01 -2.56123021e-02 -5.50292730e-01
1.36201292e-01 -3.53496939e-01 9.49283898e-01 -9.34478223e-01
2.40164131e-01 7.63669088e-02 7.29534864e-01 -1.44488525e+00
5.23897409e-01 -1.00175631e+00 -2.32340261e-01 2.40688011e-01
-5.75517165e-03 -2.62373358e-01 2.31822848e-01 6.25461280e-01
-4.26386669e-03 -2.57952750e-01 7.57083416e-01 -2.65849710e-01
-6.75748348e-01 6.79382205e-01 2.13620901e-01 -3.20756137e-01
1.14195859e+00 -4.00248855e-01 4.30858210e-02 1.45621523e-01
-3.90551805e-01 2.93462396e-01 3.84317100e-01 5.96836746e-01
1.02203798e+00 -1.28734219e+00 -8.15867007e-01 5.54497361e-01
3.86654496e-01 9.10379529e-01 7.25401491e-02 5.32847106e-01
-4.77672964e-01 4.05897468e-01 -2.78988890e-02 -1.42279053e+00
-1.22303033e+00 5.41679502e-01 3.69131446e-01 1.64609939e-01
-9.54651117e-01 1.23675990e+00 5.42843521e-01 -6.60932183e-01
3.74457538e-01 -8.04512978e-01 -6.98318854e-02 -1.64590910e-01
1.12904780e-01 -2.01905712e-01 1.83697164e-01 -4.53397185e-01
-4.90662694e-01 9.14546847e-01 -1.57724515e-01 5.47597587e-01
1.77519965e+00 2.72661537e-01 7.78028443e-02 2.68327773e-01
9.86657977e-01 -4.16742504e-01 -1.27025521e+00 -5.70249438e-01
4.11512156e-04 -7.60087192e-01 1.97849527e-01 -3.31639647e-01
-1.04409814e+00 1.01067483e+00 7.69911528e-01 2.71869361e-01
6.62221670e-01 7.03983963e-01 3.27982575e-01 5.08595407e-01
2.42159426e-01 -3.56547356e-01 2.88691670e-01 5.16841233e-01
8.23965251e-01 -1.65864861e+00 1.26277745e-01 -4.87241536e-01
-1.21803254e-01 1.22769713e+00 7.95428753e-01 -4.86397028e-01
7.49256730e-01 -1.06048405e-01 -1.84262723e-01 -7.40653813e-01
-4.53227639e-01 -4.89430666e-01 8.77771318e-01 7.06625700e-01
-4.35872339e-02 -4.29576449e-02 4.15418893e-01 4.99039412e-01
-3.80387093e-04 -4.87220794e-01 -2.82823760e-02 7.52833903e-01
-6.24465644e-01 -6.04187965e-01 -5.53782940e-01 5.52175581e-01
8.95378962e-02 2.81348705e-01 -5.21837473e-01 1.06343055e+00
3.64355505e-01 2.36497685e-01 6.13011837e-01 -2.36718386e-01
7.31300950e-01 -2.06423178e-01 4.77133811e-01 -8.15049648e-01
-2.86932707e-01 9.26307514e-02 -3.51826638e-01 -4.12145317e-01
-4.92047131e-01 -5.10924220e-01 -1.13817811e+00 9.27141458e-02
-7.13007092e-01 -2.24699870e-01 9.05191004e-01 6.72520518e-01
4.86569107e-01 2.69465983e-01 7.43373215e-01 -1.58285654e+00
-3.59005570e-01 -8.15049887e-01 -3.15607458e-01 4.23458606e-01
3.15639555e-01 -9.45450544e-01 -2.68987894e-01 -3.87808412e-01] | [7.735595226287842, -2.8771121501922607] |
027d4605-3a74-44ee-aff4-5efd154bf89f | earthnet2021-a-novel-large-scale-dataset-and | 2012.06246 | null | https://arxiv.org/abs/2012.06246v1 | https://arxiv.org/pdf/2012.06246v1.pdf | EarthNet2021: A novel large-scale dataset and challenge for forecasting localized climate impacts | Climate change is global, yet its concrete impacts can strongly vary between different locations in the same region. Seasonal weather forecasts currently operate at the mesoscale (> 1 km). For more targeted mitigation and adaptation, modelling impacts to < 100 m is needed. Yet, the relationship between driving variables and Earth's surface at such local scales remains unresolved by current physical models. Large Earth observation datasets now enable us to create machine learning models capable of translating coarse weather information into high-resolution Earth surface forecasts. Here, we define high-resolution Earth surface forecasting as video prediction of satellite imagery conditional on mesoscale weather forecasts. Video prediction has been tackled with deep learning models. Developing such models requires analysis-ready datasets. We introduce EarthNet2021, a new, curated dataset containing target spatio-temporal Sentinel 2 satellite imagery at 20 m resolution, matched with high-resolution topography and mesoscale (1.28 km) weather variables. With over 32000 samples it is suitable for training deep neural networks. Comparing multiple Earth surface forecasts is not trivial. Hence, we define the EarthNetScore, a novel ranking criterion for models forecasting Earth surface reflectance. For model intercomparison we frame EarthNet2021 as a challenge with four tracks based on different test sets. These allow evaluation of model validity and robustness as well as model applicability to extreme events and the complete annual vegetation cycle. In addition to forecasting directly observable weather impacts through satellite-derived vegetation indices, capable Earth surface models will enable downstream applications such as crop yield prediction, forest health assessments, coastline management, or biodiversity monitoring. Find data, code, and how to participate at www.earthnet.tech . | ['Markus Reichstein', 'Jakob Runge', 'Joachim Denzler', 'Vitus Benson', 'Christian Requena-Mesa'] | 2020-12-11 | null | null | null | null | ['crop-yield-prediction', 'crop-yield-prediction', 'earth-surface-forecasting'] | ['computer-vision', 'miscellaneous', 'time-series'] | [ 2.60898113e-01 -3.18323761e-01 -1.19748555e-01 -2.44859144e-01
-2.83001155e-01 -8.84478211e-01 8.84035766e-01 2.53685564e-01
-2.51718640e-01 8.01232576e-01 2.50444412e-01 -7.74223268e-01
2.12794971e-02 -1.33782125e+00 -7.45887220e-01 -6.58923984e-01
-7.12253034e-01 9.22685582e-03 1.15948394e-01 -6.49131536e-01
-1.59505039e-01 7.77885258e-01 -1.85547161e+00 2.75081515e-01
7.30122387e-01 9.96429741e-01 5.98370314e-01 1.07818305e+00
-6.90672249e-02 1.39275372e-01 -2.13597462e-01 3.21153402e-01
3.86932969e-01 -2.18697335e-03 -4.00486082e-01 -3.82383496e-01
6.46748126e-01 -3.53644490e-01 1.38846666e-01 9.43804562e-01
5.41296482e-01 -3.37972827e-02 4.71267015e-01 -9.10304010e-01
-3.13471794e-01 1.03185795e-01 -3.29202563e-01 3.41752291e-01
-4.28125486e-02 3.77850771e-01 6.14301682e-01 -6.60661161e-01
5.42349637e-01 1.04585135e+00 9.80271995e-01 3.66476215e-02
-1.29925370e+00 -6.17540956e-01 3.01248968e-01 4.08178456e-02
-1.28912687e+00 -4.51423347e-01 1.30731940e-01 -9.82744873e-01
1.15324354e+00 6.70498073e-01 7.07797050e-01 8.22695851e-01
6.51572585e-01 -5.30153736e-02 1.19599533e+00 -1.91790704e-02
2.52332926e-01 -2.26423994e-01 -3.86970758e-01 2.34618276e-01
1.92662314e-01 5.68586826e-01 -3.10009539e-01 1.42430365e-01
4.85760003e-01 9.43825319e-02 -4.62784678e-01 2.84154356e-01
-1.34072566e+00 6.14609480e-01 8.55587184e-01 1.92038387e-01
-8.30723882e-01 1.61672115e-01 3.19914579e-01 4.71970201e-01
1.07661867e+00 1.75589219e-01 -9.49118435e-01 1.56370923e-01
-1.33520663e+00 3.06219935e-01 4.98148829e-01 3.86755019e-01
7.92047620e-01 4.36596870e-01 3.39769959e-01 5.83738744e-01
2.76488066e-01 1.30320168e+00 2.65014116e-02 -8.07041228e-01
1.30860582e-01 3.80225807e-01 3.93424153e-01 -1.38126183e+00
-7.44587004e-01 -3.22069377e-01 -1.40441537e+00 6.13857210e-01
1.17507698e-02 -5.95754504e-01 -8.94826651e-01 1.52141774e+00
4.60593700e-01 3.34643573e-01 1.68724701e-01 1.18350577e+00
7.36678004e-01 1.22786069e+00 3.78799021e-01 -5.87877110e-02
1.06682992e+00 -3.30452085e-01 -4.91682827e-01 -2.52124876e-01
8.69246662e-01 -4.39994782e-01 6.45435691e-01 -3.97992246e-02
-3.58842224e-01 -6.19618297e-01 -7.26870596e-01 6.14631474e-01
-1.16887200e+00 1.98874548e-01 7.23805964e-01 3.14252555e-01
-1.48388052e+00 8.04262996e-01 -7.45620430e-01 -6.80148065e-01
2.33422834e-02 4.88193110e-02 -4.24192041e-01 3.46871883e-01
-1.68815291e+00 1.07148480e+00 3.52860898e-01 5.96235275e-01
-9.35014963e-01 -9.83712852e-01 -9.26834345e-01 1.86027437e-01
-3.80402505e-01 -4.59928602e-01 6.23152256e-01 -1.15632808e+00
-1.27499890e+00 9.03704643e-01 8.72446746e-02 -6.49483263e-01
3.63988191e-01 -4.00169455e-02 -9.79010880e-01 -1.43546864e-01
2.11434662e-01 9.51543450e-01 5.96131086e-01 -9.51606989e-01
-9.21520114e-01 -3.70870262e-01 -1.14738466e-02 1.98508203e-01
-1.05074003e-01 2.46162906e-01 2.63740808e-01 -6.11364961e-01
-1.55333847e-01 -1.07518744e+00 -4.12553102e-01 2.40489557e-01
1.27897575e-01 4.40182000e-01 7.14066029e-01 -1.02273166e+00
8.39605212e-01 -1.98107696e+00 -7.11799711e-02 -7.33935535e-02
-1.04198843e-01 4.71604705e-01 -5.07749021e-01 4.75643754e-01
-2.75442988e-01 5.41472733e-01 -4.39133734e-01 3.26282412e-01
-2.60913789e-01 5.54962941e-02 -7.28247166e-01 6.00466728e-01
2.49872044e-01 6.13861263e-01 -6.89306617e-01 2.02545986e-01
6.00735009e-01 5.94963849e-01 -1.61344737e-01 8.71515125e-02
-3.71128321e-01 6.06792390e-01 -6.46449029e-02 5.39380014e-01
1.21478665e+00 9.08769220e-02 1.80413008e-01 1.18899979e-01
-8.22264254e-01 9.00611728e-02 -1.10824847e+00 1.07427180e+00
-6.79006159e-01 1.02273583e+00 4.30877954e-01 -7.48515129e-01
1.14930820e+00 3.28564674e-01 2.00071797e-01 -7.72779167e-01
-4.42083627e-01 1.21321514e-01 -1.12267919e-01 -4.32990611e-01
5.63842952e-01 -3.20843697e-01 2.47525290e-01 -4.10884582e-02
-3.98795575e-01 -1.19827539e-01 -1.55200422e-01 -2.81846315e-01
5.84725320e-01 3.30263793e-01 9.39265490e-02 -7.27379143e-01
2.97028571e-01 4.01819199e-01 4.74522471e-01 4.82422590e-01
-1.71642125e-01 4.81449217e-01 2.24109367e-01 -1.20319855e+00
-9.51138377e-01 -6.25563562e-01 -5.33881962e-01 1.23938656e+00
-1.73706159e-01 -8.46743435e-02 -2.42623016e-01 -5.68388887e-02
3.44994247e-01 6.48715496e-01 -7.63668597e-01 2.21423998e-01
-3.33713777e-02 -1.31072414e+00 6.75667107e-01 2.72009939e-01
5.90650856e-01 -9.68956828e-01 -7.48038530e-01 2.89006352e-01
-1.53704494e-01 -9.12675738e-01 5.35211921e-01 1.96499497e-01
-9.78747547e-01 -8.80008876e-01 -8.64005446e-01 -5.65442443e-02
8.41697380e-02 3.02230746e-01 1.26175022e+00 -1.69290036e-01
-4.33269367e-02 1.09842092e-01 -1.71467245e-01 -3.96049559e-01
-3.33718330e-01 1.79501444e-01 3.66231263e-01 -1.16735026e-01
9.38855186e-02 -6.65954411e-01 -7.44611800e-01 4.14579570e-01
-9.09913898e-01 2.79034257e-01 3.18548918e-01 2.51937985e-01
4.31807250e-01 -2.16789410e-01 6.71396494e-01 -3.43490183e-01
2.52072364e-02 -1.02541077e+00 -1.06685126e+00 2.29137957e-01
-2.98760027e-01 -4.59121704e-01 5.30102670e-01 -4.55205031e-02
-1.10444975e+00 1.95691660e-01 -1.22839704e-01 -4.60236706e-02
-8.97177875e-01 1.22378135e+00 2.33740672e-01 -6.48968741e-02
8.35657835e-01 3.35656494e-01 -3.60693991e-01 -4.03743982e-01
2.74635822e-01 7.09374189e-01 3.93898070e-01 -1.25542298e-01
6.48397207e-01 6.14269614e-01 1.39745429e-01 -1.43192470e+00
-5.88393986e-01 -3.21124882e-01 -6.55840695e-01 -6.96531653e-01
8.31736565e-01 -1.53705251e+00 -4.11662757e-01 6.29182994e-01
-1.17621517e+00 -7.51667678e-01 2.50654995e-01 5.50295115e-01
-9.78035629e-02 -1.01090953e-01 2.22098883e-02 -8.30449700e-01
-4.48975593e-01 -4.71533746e-01 9.02992606e-01 8.55558459e-03
-8.99918750e-02 -1.24115682e+00 7.36395121e-01 -2.00251937e-01
1.07026899e+00 8.14307690e-01 3.24630708e-01 9.55354944e-02
-4.75708276e-01 -1.23544961e-01 -4.41680372e-01 1.88916817e-01
2.68270940e-01 4.80382621e-01 -1.11805129e+00 -3.60957354e-01
-4.42783475e-01 -1.44366443e-01 1.36294532e+00 7.57544816e-01
8.12356949e-01 -2.73633718e-01 -2.57116526e-01 1.20068538e+00
1.55388117e+00 -1.20545521e-01 6.59892976e-01 6.83850825e-01
4.56534714e-01 8.54220688e-01 4.63616252e-01 4.82493132e-01
3.47343832e-01 2.36394137e-01 9.18697536e-01 -4.31239784e-01
1.19613446e-01 3.22335988e-01 5.84612906e-01 3.03833365e-01
-3.59492868e-01 -7.20482320e-02 -1.67134070e+00 8.09324086e-01
-1.53873253e+00 -1.39625406e+00 -5.71607888e-01 2.25437403e+00
3.00040215e-01 -1.34802774e-01 -3.33371848e-01 -4.56098616e-01
7.61316121e-01 8.06254506e-01 -7.10654855e-01 -3.38787347e-01
-5.47518671e-01 -1.94220975e-01 1.19278932e+00 5.61899364e-01
-1.43916583e+00 1.05397785e+00 6.16599607e+00 2.38042086e-01
-1.88301897e+00 8.27489421e-02 6.36791468e-01 -3.24670151e-02
-3.69237214e-01 5.84724657e-02 -1.03519821e+00 3.06285411e-01
1.59973121e+00 4.24953876e-03 3.30017120e-01 5.86886823e-01
9.54449296e-01 -2.47588143e-01 -3.06665957e-01 3.91618669e-01
-6.02317810e-01 -1.84640670e+00 9.82476771e-02 3.81748527e-02
9.76862967e-01 1.00828493e+00 -9.27108154e-02 3.04505974e-01
5.23957908e-01 -1.00796127e+00 3.77274305e-01 8.76773655e-01
1.25527799e+00 -4.68228221e-01 5.12656450e-01 4.03766751e-01
-1.66020644e+00 1.16433628e-01 -6.34244084e-01 -5.07448733e-01
-1.04027018e-01 8.53865325e-01 -3.39938343e-01 6.61704004e-01
1.10171521e+00 1.19945025e+00 -3.28071654e-01 7.61060238e-01
8.83533657e-02 7.94260204e-01 -7.65377879e-01 4.24131751e-01
4.22111899e-01 -9.30302590e-02 5.68305492e-01 1.30181038e+00
7.14755356e-01 3.26028943e-01 8.04442838e-02 5.38340151e-01
2.26800263e-01 1.02926463e-01 -1.08734035e+00 1.55297294e-01
2.91506857e-01 1.36932683e+00 -4.62662458e-01 -2.91298985e-01
-3.26354653e-01 5.96355319e-01 -9.61373672e-02 5.08077621e-01
-6.68687940e-01 -1.06919140e-01 1.30240822e+00 2.58573711e-01
1.42954975e-01 -4.55832511e-01 -5.64717352e-02 -1.30081785e+00
-3.03698123e-01 -5.98498225e-01 7.57670105e-02 -1.04557538e+00
-6.76173091e-01 4.95147437e-01 -1.03667714e-01 -1.46761262e+00
-7.72249550e-02 -6.37724638e-01 -1.07952642e+00 1.34585047e+00
-2.35926747e+00 -1.29661858e+00 -5.56399465e-01 1.82471141e-01
1.44720748e-01 7.81425610e-02 1.11569512e+00 -2.08696928e-02
-4.82877016e-01 -4.60370898e-01 7.98505127e-01 -3.36171806e-01
7.62206674e-01 -1.05297041e+00 8.97763729e-01 7.69276857e-01
-5.96973419e-01 3.83240432e-02 6.57399178e-01 -9.04425561e-01
-1.12505627e+00 -1.90582144e+00 1.04398274e+00 9.16467831e-02
8.62396896e-01 9.89654735e-02 -1.05683410e+00 5.75351357e-01
4.29931935e-03 2.67932206e-01 5.78224421e-01 1.71765700e-01
-6.14202023e-02 -5.05566835e-01 -9.09110487e-01 3.54818642e-01
5.31658709e-01 -5.63705742e-01 -5.24695069e-02 6.03346527e-01
4.38323170e-01 -7.63643235e-02 -1.15961838e+00 6.95337415e-01
6.84625268e-01 -9.42706823e-01 6.86979353e-01 -8.49269688e-01
7.84884751e-01 -3.79822969e-01 -6.11534894e-01 -1.69228995e+00
-6.57865167e-01 -1.71073601e-01 3.14540148e-01 9.29746747e-01
3.91484112e-01 -5.86743653e-01 4.06966716e-01 5.31462908e-01
-4.55183648e-02 6.01771213e-02 -9.10936952e-01 -8.91068518e-01
5.82568765e-01 -4.45563465e-01 6.75092280e-01 1.33243334e+00
-4.40086216e-01 -2.72165567e-01 -4.16727483e-01 1.11036265e+00
3.19856942e-01 3.83933961e-01 8.43109429e-01 -1.62005305e+00
3.50779355e-01 -4.40288156e-01 -2.71628380e-01 -4.65518951e-01
9.46983248e-02 -6.84499204e-01 -2.02008978e-01 -1.45158148e+00
-3.20300877e-01 -2.87245333e-01 -2.31451601e-01 7.21162975e-01
-3.87402182e-03 1.47259682e-01 3.10043171e-02 2.27403253e-01
1.76479578e-01 7.79111326e-01 7.46703982e-01 -2.40165070e-01
-6.46687821e-02 -2.97135621e-01 1.52867481e-01 7.52245426e-01
1.15239179e+00 -4.71295446e-01 3.00568342e-01 -6.90217078e-01
6.07410848e-01 3.23830307e-01 7.38391161e-01 -1.36693609e+00
-1.55947670e-01 -7.79706717e-01 5.70656061e-01 -7.37136960e-01
-5.95767237e-02 -8.10063839e-01 7.11158991e-01 5.31092525e-01
-1.76609442e-01 3.47725935e-02 8.03117335e-01 3.13181430e-01
-2.58536965e-01 2.99725175e-01 7.73492038e-01 -1.35961592e-01
-1.16672969e+00 5.20698726e-01 -9.07032609e-01 -3.98386031e-01
8.92105937e-01 9.03660804e-02 -4.53004330e-01 -2.77255148e-01
-9.26612914e-01 3.50101829e-01 6.00028932e-01 5.76740742e-01
1.62442133e-01 -1.00178301e+00 -1.23112464e+00 3.96203637e-01
1.37078732e-01 -4.15044576e-01 6.39516532e-01 5.68539381e-01
-7.46516168e-01 5.46681643e-01 -5.50710261e-01 -6.36738598e-01
-1.09771192e+00 2.11652577e-01 8.20493340e-01 -4.49527707e-03
-4.18768644e-01 4.90323573e-01 1.31318480e-01 -1.03744841e+00
-4.81006563e-01 -5.46249092e-01 -4.23579365e-01 5.06269693e-01
7.32036352e-01 8.10148790e-02 6.25535771e-02 -8.18341613e-01
-4.86870199e-01 4.80405360e-01 7.36648977e-01 2.07227655e-02
1.70423377e+00 -3.40612590e-01 -1.19984202e-01 7.23586440e-01
9.07426357e-01 -3.91073793e-01 -1.50441587e+00 -2.15802472e-02
-6.81466833e-02 -4.10464674e-01 8.16392243e-01 -9.15603220e-01
-1.20520866e+00 1.16894615e+00 1.09189880e+00 3.58925611e-01
1.09105790e+00 -6.45945132e-01 3.73727053e-01 6.50715530e-01
-8.77561122e-02 -8.14330876e-01 -1.01007736e+00 7.44884074e-01
1.15897822e+00 -1.62583315e+00 -2.88316514e-02 2.99144179e-01
-2.51273423e-01 1.24335384e+00 4.01939869e-01 1.92654684e-01
9.99783933e-01 1.74514845e-01 2.46023118e-01 -4.74070981e-02
-9.40080166e-01 -4.12957937e-01 1.04352638e-01 6.23337924e-01
4.52935547e-01 6.26396954e-01 2.93101877e-01 7.24066570e-02
7.94057995e-02 3.05819064e-01 5.02031982e-01 5.34252942e-01
-8.18547547e-01 -3.61774921e-01 -7.79532671e-01 4.86866206e-01
-1.64068863e-01 -5.26659429e-01 3.02230902e-02 3.53424698e-01
-1.27470434e-01 6.84783399e-01 3.17441761e-01 -2.32479155e-01
7.20403194e-02 -5.27814254e-02 -4.55775261e-01 -2.39693329e-01
-4.17665154e-01 -4.13214624e-01 1.84728444e-01 -6.46954298e-01
-6.53600931e-01 -7.59023488e-01 -6.38378084e-01 -9.64401186e-01
1.56223029e-03 -1.68765515e-01 1.16707098e+00 7.55688846e-01
7.69685090e-01 3.25486630e-01 6.83496237e-01 -1.74107838e+00
-1.17154777e-01 -1.07719076e+00 -7.17505217e-01 -1.88433602e-01
7.26414144e-01 -4.34113950e-01 -5.22155046e-01 1.35579005e-01] | [9.523691177368164, -1.5625641345977783] |
7cf82dfb-2789-4036-86b4-aaba8f399d3d | all-information-is-necessary-integrating | 2304.13439 | null | https://arxiv.org/abs/2304.13439v1 | https://arxiv.org/pdf/2304.13439v1.pdf | All Information is Necessary: Integrating Speech Positive and Negative Information by Contrastive Learning for Speech Enhancement | Monaural speech enhancement (SE) is an ill-posed problem due to the irreversible degradation process. Recent methods to achieve SE tasks rely solely on positive information, e.g., ground-truth speech and speech-relevant features. Different from the above, we observe that the negative information, such as original speech mixture and speech-irrelevant features, are valuable to guide the SE model training procedure. In this study, we propose a SE model that integrates both speech positive and negative information for improving SE performance by adopting contrastive learning, in which two innovations have consisted. (1) We design a collaboration module (CM), which contains two parts, contrastive attention for separating relevant and irrelevant features via contrastive learning and interactive attention for establishing the correlation between both speech features in a learnable and self-adaptive manner. (2) We propose a contrastive regularization (CR) built upon contrastive learning to ensure that the estimated speech is pulled closer to the clean speech and pushed far away from the noisy speech in the representation space by integrating self-supervised models. We term the proposed SE network with CM and CR as CMCR-Net. Experimental results demonstrate that our CMCR-Net achieves comparable and superior performance to recent approaches. | ['Yuhong Yang', 'Chang Han', 'Weiping tu', 'Xinmeng Xu'] | 2023-04-26 | null | null | null | null | ['speech-enhancement'] | ['speech'] | [ 2.83144236e-01 -9.68073756e-02 2.78901666e-01 -3.47531855e-01
-9.94860530e-01 1.87194981e-02 3.71747702e-01 -3.78607124e-01
-4.14827436e-01 5.57009101e-01 6.03636622e-01 2.15087086e-02
-1.76612943e-01 -3.48597437e-01 -7.36414731e-01 -9.24228311e-01
1.12188168e-01 -4.09790397e-01 2.83295274e-01 -5.53659499e-01
4.35140654e-02 2.08491266e-01 -1.70188332e+00 2.34046713e-01
1.17994523e+00 1.08621573e+00 6.73578382e-01 6.11575961e-01
-3.93855236e-02 9.08008754e-01 -5.25528848e-01 8.29321146e-02
2.25828439e-02 -8.17912102e-01 -3.74129832e-01 2.32501358e-01
-5.13460375e-02 9.66611207e-02 -3.92305642e-01 1.30582356e+00
9.37492847e-01 4.97838974e-01 4.74785298e-01 -1.10699856e+00
-8.74987006e-01 4.89906400e-01 -5.15039325e-01 4.54062849e-01
1.22139901e-01 2.43047565e-01 7.95922637e-01 -1.16179347e+00
8.94280523e-02 1.19569063e+00 6.73694730e-01 8.64399314e-01
-8.89497161e-01 -7.41804659e-01 3.84364694e-01 3.73800218e-01
-1.18015838e+00 -1.09408391e+00 1.25240076e+00 -1.74775422e-01
6.50946200e-01 5.56227624e-01 5.86061180e-01 1.12831163e+00
-6.75686374e-02 8.64092052e-01 1.14805138e+00 -6.93885386e-01
1.81359544e-01 3.71102393e-01 1.28185123e-01 1.86726883e-01
-4.55924869e-01 5.26340604e-01 -4.92499053e-01 4.35736850e-02
5.13814032e-01 -2.99369216e-01 -7.59094238e-01 -1.20421350e-01
-7.94027030e-01 4.11318421e-01 4.73626643e-01 5.87142408e-01
-4.42048430e-01 -2.69467592e-01 2.24425346e-01 3.69093537e-01
7.55387247e-01 2.35156268e-01 -4.25088674e-01 -2.88827475e-02
-6.76973820e-01 -2.29228780e-01 4.28617120e-01 6.54599011e-01
6.73371792e-01 4.76172358e-01 -3.37910384e-01 1.42747748e+00
5.58832407e-01 4.75651175e-01 8.81271720e-01 -6.83171034e-01
6.18980467e-01 1.28718257e-01 -5.08437268e-02 -8.81538808e-01
2.43092445e-03 -9.63149250e-01 -9.74477351e-01 1.61406770e-01
-3.58115405e-01 -3.39277804e-01 -8.64163876e-01 1.99348390e+00
1.98924169e-01 5.85932195e-01 3.45620841e-01 1.06690025e+00
1.05836618e+00 9.13716614e-01 1.47870675e-01 -6.68533862e-01
8.58219802e-01 -1.24857116e+00 -1.29107130e+00 -4.06023562e-01
7.18868226e-02 -8.40550184e-01 1.11103368e+00 2.85897136e-01
-1.24525738e+00 -8.57171297e-01 -1.17408502e+00 1.55756220e-01
-2.24161804e-01 2.10332245e-01 -1.94485690e-02 4.70350653e-01
-1.21104980e+00 5.91689706e-01 -5.27570128e-01 4.19213548e-02
2.37667367e-01 2.19310820e-01 -1.25605941e-01 3.70687455e-01
-1.48173034e+00 7.17169225e-01 1.70839041e-01 5.36604166e-01
-8.35766435e-01 -4.09597814e-01 -9.86818135e-01 1.81693301e-01
4.04573530e-01 -5.45451760e-01 1.30421031e+00 -1.48570752e+00
-1.96054888e+00 3.54474008e-01 -2.48461232e-01 -2.41855815e-01
3.87035966e-01 -3.59868318e-01 -8.76034439e-01 9.77496430e-02
-2.59402096e-01 3.05223733e-01 1.34135973e+00 -1.67808354e+00
-5.15300691e-01 -2.57415593e-01 -4.15366858e-01 5.71514428e-01
-6.37127995e-01 1.59677759e-01 -2.87543118e-01 -9.64577198e-01
5.03630817e-01 -5.14189303e-01 -8.27211812e-02 -7.92763829e-02
-1.73704267e-01 -1.48510203e-01 8.87787282e-01 -9.44891512e-01
1.46596789e+00 -2.32235336e+00 3.63053828e-02 9.46294069e-02
2.46980116e-01 7.09293127e-01 -3.06144327e-01 1.82890534e-01
-5.20946145e-01 -8.76592919e-02 -3.36333632e-01 -4.65503216e-01
-9.66698229e-02 -3.79934255e-03 -4.17261302e-01 3.09479058e-01
4.77068424e-01 6.20900691e-01 -1.06925058e+00 -4.13870931e-01
2.82799989e-01 7.68113971e-01 -4.21569169e-01 7.15234339e-01
2.07617506e-01 5.37381947e-01 -2.40935043e-01 1.85789332e-01
8.73511791e-01 7.55828619e-02 -2.65373439e-01 -3.20524544e-01
-3.22671235e-01 4.63153422e-01 -1.28333640e+00 1.30714464e+00
-5.00201941e-01 4.63435382e-01 6.07510030e-01 -1.05264199e+00
1.11514437e+00 4.90534723e-01 1.50002956e-01 -8.28798831e-01
1.14452943e-01 1.02643825e-01 1.63474027e-03 -8.19227874e-01
2.63461232e-01 -3.62070829e-01 5.87454677e-01 1.12911649e-01
2.53695726e-01 -2.25217622e-02 -3.83239359e-01 7.63773695e-02
7.45230973e-01 -2.35946737e-02 1.46300450e-01 -8.52507949e-02
9.93930817e-01 -7.59189785e-01 9.44444776e-01 5.18284142e-01
-5.85968554e-01 6.47279859e-01 -2.41007167e-03 3.04841131e-01
-6.03922725e-01 -9.94339347e-01 9.46392026e-03 1.12993121e+00
3.51342499e-01 -2.69770026e-01 -7.39952326e-01 -5.50614893e-01
-6.04230046e-01 5.55277407e-01 -4.10003364e-01 -7.74320900e-01
-6.13735855e-01 -6.19501710e-01 5.01118414e-02 5.62182844e-01
6.52697325e-01 -1.25016272e+00 2.99938589e-01 2.21086338e-01
-4.11248326e-01 -6.82936609e-01 -7.37883449e-01 3.31492931e-01
-6.26644731e-01 -5.44879258e-01 -7.60927737e-01 -1.06069613e+00
6.48783088e-01 7.09094346e-01 6.67312801e-01 -1.03910834e-01
2.77142435e-01 4.47630197e-01 -4.33601409e-01 -2.85430312e-01
-5.21010220e-01 -3.68107110e-01 7.50657171e-02 4.64050263e-01
1.19596742e-01 -7.62141228e-01 -6.16201222e-01 2.99994290e-01
-6.79766715e-01 -9.35474560e-02 7.00313568e-01 1.13446033e+00
5.79621315e-01 2.10384443e-01 1.12343729e+00 -4.48317081e-01
8.82072926e-01 -5.22244155e-01 -6.82878196e-02 8.08599368e-02
-5.11670709e-01 -1.77280650e-01 5.82395256e-01 -7.06942737e-01
-1.54804814e+00 -1.95613384e-01 -6.66760981e-01 -6.22826338e-01
-1.03303954e-01 2.15044037e-01 -5.02634823e-01 -3.75226177e-02
5.39973378e-01 7.47938931e-01 1.22526355e-01 -5.43318808e-01
2.24278271e-01 1.15444827e+00 5.99131823e-01 -3.94051433e-01
7.29857922e-01 2.81320568e-02 -7.08055675e-01 -9.23505664e-01
-9.20097053e-01 -6.44064784e-01 -2.97007143e-01 -4.39985663e-01
6.46657109e-01 -1.05754113e+00 -4.68096197e-01 6.71673119e-01
-1.00758028e+00 -1.78266212e-01 -4.18303490e-01 7.57013381e-01
-6.22373760e-01 5.43601513e-01 -6.13247871e-01 -1.32714093e+00
-4.44763809e-01 -1.01282024e+00 8.42567146e-01 3.81840557e-01
1.92248940e-01 -7.24859893e-01 1.88300043e-01 3.86307299e-01
5.71704984e-01 -4.00720060e-01 4.47515786e-01 -6.01710081e-01
-1.46000385e-01 1.75772578e-01 -1.19234495e-01 1.11873305e+00
3.23250204e-01 -2.16586143e-01 -1.50642300e+00 -3.66150498e-01
6.05324805e-01 -1.87545866e-01 9.20986235e-01 4.15915489e-01
1.24759960e+00 -4.27103370e-01 -3.87642793e-02 4.61552739e-01
1.00951576e+00 5.12456000e-01 8.52467299e-01 2.05920041e-01
4.30113077e-01 7.23503530e-01 5.77441633e-01 1.56108513e-01
2.43640646e-01 3.76775861e-01 2.89779902e-01 -4.02203798e-01
-4.31906402e-01 -3.85069340e-01 5.08748531e-01 1.69796574e+00
1.13807343e-01 -1.31240517e-01 -3.69809210e-01 5.96790612e-01
-1.86399591e+00 -9.95184898e-01 9.32574049e-02 2.15616703e+00
1.02236521e+00 2.45731056e-01 -1.30648851e-01 4.43905592e-01
1.13377154e+00 2.84768641e-01 -4.44525719e-01 -4.10849787e-02
-2.78959781e-01 1.01319604e-01 -3.01863581e-01 5.10548592e-01
-1.17518342e+00 6.04952395e-01 5.77844238e+00 1.09594405e+00
-1.30246735e+00 2.31924459e-01 4.55177039e-01 7.34142140e-02
-3.08453709e-01 -3.33575696e-01 -3.61643136e-01 6.51154995e-01
7.92747438e-01 -4.72514555e-02 4.22250211e-01 6.39994025e-01
5.87853253e-01 3.54844779e-01 -5.62093437e-01 1.07170403e+00
2.00826049e-01 -8.25178385e-01 -2.76668340e-01 -3.45826894e-01
5.87876856e-01 -2.06194550e-01 2.14491740e-01 6.45639956e-01
8.76125097e-02 -5.14888704e-01 8.72435868e-01 7.39684105e-01
4.24097419e-01 -6.61392629e-01 6.00804687e-01 5.28196871e-01
-1.28917754e+00 -3.03920358e-01 -2.46607259e-01 2.06456363e-01
6.08038194e-02 4.56754893e-01 -3.22188884e-01 6.44441307e-01
8.89386237e-01 8.61879766e-01 -1.54570445e-01 1.10511935e+00
-5.27039170e-01 8.77887309e-01 1.83158249e-01 1.80282310e-01
-2.59442683e-02 -3.01721871e-01 9.28936243e-01 1.15920329e+00
2.36369193e-01 2.34379679e-01 -8.82935077e-02 8.49144101e-01
4.55807103e-03 2.37682760e-01 -3.31653059e-01 2.41042942e-01
5.82679808e-01 1.17709291e+00 -1.91869922e-02 -2.13172361e-01
-3.65076780e-01 8.30016732e-01 3.86084110e-01 6.00782454e-01
-6.96901560e-01 -6.11945987e-01 4.44334328e-01 -1.22592069e-01
4.14271295e-01 1.30980849e-01 1.57429595e-02 -1.11581266e+00
3.55698854e-01 -8.06315839e-01 8.99937227e-02 -1.04964876e+00
-1.58443499e+00 9.59396064e-01 -6.21975780e-01 -1.62094223e+00
4.86378483e-02 -1.62827909e-01 -8.00390363e-01 1.07893956e+00
-2.02262759e+00 -8.87202859e-01 -1.94744721e-01 6.38928473e-01
6.76817119e-01 -2.06828877e-01 4.31524068e-01 4.77758348e-01
-8.59532654e-01 7.09954560e-01 1.06264442e-01 -1.43938646e-01
8.14013720e-01 -1.12270427e+00 -9.56225023e-02 9.17249978e-01
-1.13715000e-01 5.10856330e-01 5.94452262e-01 -4.42765802e-01
-9.44006741e-01 -1.16196585e+00 7.84687340e-01 8.24192017e-02
6.15472853e-01 -3.36585164e-01 -1.25876451e+00 3.59432459e-01
2.55348235e-01 1.74723312e-01 5.88120043e-01 -9.67773199e-02
-1.63270503e-01 -5.04065514e-01 -9.28993762e-01 5.32066941e-01
1.06475091e+00 -7.32882977e-01 -9.82963145e-01 5.16157523e-02
1.12274706e+00 -2.68703282e-01 -6.06292546e-01 5.81855834e-01
1.33256957e-01 -7.90557027e-01 9.59781706e-01 -4.51499432e-01
2.16033965e-01 -4.48847562e-01 -1.07465208e-01 -1.68780577e+00
-4.02820230e-01 -8.12703609e-01 -2.73839772e-01 1.68506420e+00
2.30326757e-01 -6.40939474e-01 1.88724354e-01 1.27183557e-01
-7.18669295e-01 -7.12756634e-01 -8.72693062e-01 -8.62187088e-01
-1.67946532e-01 -5.26026249e-01 4.57757354e-01 9.10695732e-01
-2.62591373e-02 5.98942518e-01 -5.56100905e-01 4.71540183e-01
4.44128156e-01 -2.48876154e-01 2.99456865e-01 -1.02618825e+00
-3.04541945e-01 -3.80699664e-01 -2.14883704e-02 -1.32211840e+00
3.30753356e-01 -6.43535674e-01 6.00127757e-01 -1.34769237e+00
1.34580478e-01 -3.35462719e-01 -7.20292926e-01 1.06855243e-01
-5.67090333e-01 -1.57074600e-01 2.00232156e-02 1.81879714e-01
-6.59613371e-01 1.34497488e+00 1.43794942e+00 -1.01027921e-01
-4.94996488e-01 2.87318021e-01 -8.79517555e-01 7.45391667e-01
5.64918816e-01 -3.80945474e-01 -4.96895522e-01 -2.10281849e-01
-2.57849097e-01 2.33027920e-01 2.89373577e-01 -1.02560771e+00
4.74035859e-01 5.23358025e-02 2.30930567e-01 -5.30361593e-01
5.41045308e-01 -8.01950395e-01 -4.32788014e-01 1.73787698e-01
-6.39178216e-01 -4.78917956e-01 1.61019102e-01 8.19862068e-01
-6.45182669e-01 -1.95887446e-01 9.67346907e-01 -1.65624656e-02
-5.80953062e-01 1.79712638e-01 -3.31621230e-01 2.55905986e-02
6.60049856e-01 -9.77616012e-02 -3.02994013e-01 -6.07725024e-01
-1.12829268e+00 2.42029309e-01 -3.86195838e-01 5.15571952e-01
8.36894870e-01 -1.52871644e+00 -7.15777218e-01 4.06162381e-01
-3.90945189e-03 -2.80180126e-01 6.99330032e-01 8.06131303e-01
3.34778547e-01 -1.81757268e-02 1.64988756e-01 -5.52051842e-01
-1.23181498e+00 5.24617732e-01 5.87401628e-01 1.04689352e-01
-4.77511346e-01 9.99324977e-01 4.33679760e-01 -6.34831429e-01
5.71453452e-01 -4.10755165e-02 -6.33223355e-01 -9.67570096e-02
7.63526857e-01 3.91409338e-01 2.31418267e-01 -6.63234115e-01
-1.11372001e-01 4.36458826e-01 -5.82353696e-02 -1.38331860e-01
1.58979762e+00 -6.12457752e-01 -1.38029888e-01 5.07624328e-01
1.08313549e+00 8.05564597e-02 -1.56872034e+00 -5.68168581e-01
-3.03516507e-01 -3.16887230e-01 4.65328723e-01 -9.47132468e-01
-1.21640396e+00 9.61502314e-01 1.00654733e+00 3.58053863e-01
1.58718979e+00 -1.44531690e-02 7.28506923e-01 5.97506911e-02
-1.44500032e-01 -1.27976596e+00 4.73499686e-01 3.40092391e-01
1.21588767e+00 -1.32806373e+00 -5.17276227e-01 -5.42003036e-01
-6.28031075e-01 8.69502842e-01 8.29588890e-01 4.60555814e-02
8.55518341e-01 2.33503252e-01 2.78518409e-01 8.80015567e-02
-7.63058722e-01 -4.62962806e-01 5.02583265e-01 8.14937174e-01
2.89760917e-01 -2.42991924e-01 -2.23824009e-01 1.30283415e+00
-3.95567752e-02 -2.28785738e-01 1.24257825e-01 8.72791827e-01
-7.53210247e-01 -8.06655228e-01 -4.05603707e-01 1.55754715e-01
-2.32355326e-01 -2.32817233e-01 -1.96357206e-01 2.96145469e-01
1.54599786e-01 1.38446867e+00 -6.60373643e-02 -6.21021986e-01
6.15824580e-01 -1.04384892e-01 9.01759192e-02 -5.12182236e-01
-6.68484867e-01 5.17124593e-01 -8.13572407e-02 -2.23407611e-01
-4.82492745e-01 -4.97222006e-01 -9.21818376e-01 3.70638251e-01
-7.31097102e-01 2.59851873e-01 5.40697634e-01 9.87740934e-01
3.86348963e-01 1.03869176e+00 1.27183282e+00 -7.25366771e-01
-8.02193105e-01 -1.15491891e+00 -5.95869958e-01 2.49363154e-01
7.90376663e-01 -5.45791328e-01 -9.01713192e-01 9.36605781e-02] | [14.810855865478516, 5.935832977294922] |
08e5c278-6f83-4718-8c91-174643f24d54 | location-free-spectrum-cartography | 1812.11539 | null | https://arxiv.org/abs/1812.11539v2 | https://arxiv.org/pdf/1812.11539v2.pdf | Location-free Spectrum Cartography | Spectrum cartography constructs maps of metrics such as channel gain or received signal power across a geographic area of interest using spatially distributed sensor measurements. Applications of these maps include network planning, interference coordination, power control, localization, and cognitive radios to name a few. Since existing spectrum cartography techniques require accurate estimates of the sensor locations, their performance is drastically impaired by multipath affecting the positioning pilot signals, as occurs in indoor or dense urban scenarios. To overcome such a limitation, this paper introduces a novel paradigm for spectrum cartography, where estimation of spectral maps relies on features of these positioning signals rather than on location estimates. Specific learning algorithms are built upon this approach and offer a markedly improved estimation performance than existing approaches relying on localization, as demonstrated by simulation studies in indoor scenarios. | ['Baltasar Beferull-Lozano', 'Luis Miguel Lopez Ramos', 'Daniel Romero', 'Yves Teganya'] | 2018-12-30 | null | null | null | null | ['spectrum-cartography'] | ['computer-vision'] | [ 5.28831542e-01 -5.47663420e-02 -1.42436504e-01 -8.32730308e-02
-6.32696807e-01 -5.78434706e-01 3.90648276e-01 1.82588294e-01
-3.09092760e-01 1.18896699e+00 2.51925200e-01 -4.92736876e-01
-8.21668506e-01 -9.53766763e-01 -1.47545680e-01 -7.31256843e-01
-6.09537065e-01 -2.85866950e-02 -9.10626724e-02 -2.47773398e-02
2.89183825e-01 4.20082837e-01 -1.09282041e+00 -6.32699609e-01
7.41137624e-01 1.18493700e+00 2.08532229e-01 7.76356339e-01
2.52167974e-02 4.52340603e-01 -1.15373516e+00 1.77331537e-01
1.33002415e-01 -3.65586072e-01 1.09710217e-01 -3.79759073e-01
-2.40836605e-01 -5.25761768e-02 -1.82436377e-01 9.19660509e-01
7.19954669e-01 -2.75329798e-01 4.20031846e-01 -1.22563767e+00
5.07451706e-02 1.08500731e+00 -4.15105999e-01 3.36506218e-02
6.42873049e-01 -7.38708913e-01 6.64203465e-01 -3.76725733e-01
1.02770768e-01 6.79393947e-01 1.13179612e+00 -2.06251308e-01
-1.30943859e+00 -9.98247623e-01 1.94303375e-02 1.55349821e-02
-1.91331017e+00 -3.20082217e-01 8.26708198e-01 -1.36639953e-01
3.56725752e-01 4.72616881e-01 5.74323177e-01 7.62942433e-01
5.47011554e-01 1.18859544e-01 9.84236121e-01 -7.23774076e-01
5.63805580e-01 -1.82670370e-01 -5.32093465e-01 4.55664366e-01
3.87269944e-01 1.35721147e-01 -6.77784264e-01 -4.73764837e-01
5.56656420e-01 -1.07673362e-01 -4.79484916e-01 -7.47989655e-01
-1.34703302e+00 5.32419384e-01 5.73094726e-01 6.03534222e-01
-4.18868512e-01 3.54017943e-01 -3.32620591e-01 3.62426221e-01
1.92229241e-01 6.63643360e-01 -4.21224147e-01 -1.23153195e-01
-1.02657151e+00 -2.12228373e-01 9.05034244e-01 1.17549443e+00
8.79349649e-01 2.73174554e-01 4.59917724e-01 6.33652925e-01
4.31084573e-01 1.18345165e+00 1.52337894e-01 -6.44938886e-01
5.07422686e-01 -1.13238893e-01 4.01470542e-01 -1.18288600e+00
-1.25525773e+00 -1.03463078e+00 -9.33612049e-01 1.82223264e-02
4.08360690e-01 -9.49176788e-01 -7.28663802e-01 1.76438498e+00
-1.88392829e-02 3.21777351e-02 8.04070905e-02 3.16754311e-01
-3.42976183e-01 4.87931818e-01 -1.90342709e-01 -3.47611129e-01
8.13037515e-01 -4.64860685e-02 -9.00500000e-01 -1.43389553e-01
2.32168227e-01 -7.12722659e-01 1.30093530e-01 6.05789065e-01
-5.46283782e-01 -2.78663188e-01 -1.58098769e+00 1.07642746e+00
-6.74448609e-01 -2.26414785e-01 5.92427135e-01 1.62739313e+00
-1.19086969e+00 9.85975750e-03 -5.44920146e-01 -4.26970154e-01
1.62261933e-01 3.98465276e-01 5.24882138e-01 7.29969740e-02
-1.26855671e+00 8.68905246e-01 2.17485189e-01 -6.55416772e-02
4.55145575e-02 -7.51725078e-01 -7.37701297e-01 5.85316569e-02
1.79516718e-01 -4.06298667e-01 1.15862572e+00 -3.78310204e-01
-1.65806770e+00 -3.58247846e-01 9.05589014e-03 -6.16545737e-01
3.28018129e-01 4.66481745e-02 -1.19602191e+00 -1.18433304e-01
2.62460202e-01 3.01899761e-03 5.97900808e-01 -1.04395759e+00
-1.17405844e+00 -4.72546369e-02 9.18653086e-02 1.38999686e-01
1.01605743e-01 -5.95847249e-01 1.15767494e-01 -4.77256387e-01
6.91051066e-01 -9.24982667e-01 -6.24180734e-01 -4.33224976e-01
-6.02019906e-01 6.12023950e-01 6.71073973e-01 -3.00707817e-01
1.25990248e+00 -1.88878453e+00 -5.38319528e-01 1.26451170e+00
-4.94729318e-02 -4.86691982e-01 3.66250068e-01 1.05617404e+00
2.30333433e-01 1.23651639e-01 1.98483661e-01 1.82673097e-01
1.08562097e-01 -3.47121924e-01 -1.71340123e-01 5.85719228e-01
-5.77467144e-01 2.17886627e-01 -8.86444509e-01 4.29792926e-02
4.22396988e-01 2.41625577e-01 -7.98795670e-02 -5.76794207e-01
3.87199193e-01 6.76709056e-01 -7.36684382e-01 6.27894819e-01
6.01540864e-01 -1.49808034e-01 6.22049332e-01 -6.88420609e-02
-5.44252753e-01 1.33169815e-01 -1.27785885e+00 1.68510759e+00
-1.04481757e+00 7.95719147e-01 2.30998531e-01 -7.84441292e-01
9.29511905e-01 3.15003186e-01 1.06785834e+00 -1.00594521e+00
7.40052806e-03 3.28956664e-01 1.08797736e-02 1.56061977e-01
1.66138440e-01 1.67720526e-01 -5.75353801e-01 5.90066612e-01
-3.72877359e-01 1.02662101e-01 -1.52875900e-01 1.69530585e-02
1.61286366e+00 -3.55842590e-01 9.37615693e-01 -3.70867729e-01
4.64910209e-01 -1.13293529e-01 3.32654119e-01 1.21581841e+00
1.75778225e-01 -6.45005843e-03 -1.20766096e-01 -1.57551486e-02
-8.15833032e-01 -1.29379928e+00 -4.59539980e-01 9.79491472e-01
5.55988789e-01 -3.68203402e-01 -3.93122256e-01 -3.22201289e-02
1.93439022e-01 7.53430426e-01 -3.65489990e-01 1.91275761e-01
-3.78353536e-01 -8.02341998e-01 5.53439260e-01 1.30899116e-01
5.48901439e-01 -1.81970954e-01 -7.03826070e-01 6.35934770e-01
8.27417672e-02 -1.11524117e+00 -1.33276343e-01 5.61158001e-01
-1.33232147e-01 -9.95884478e-01 -5.88064790e-01 -4.07440692e-01
5.07852674e-01 5.06431639e-01 6.58444226e-01 -1.82979196e-01
-9.25663672e-03 8.65125716e-01 -3.81890833e-01 -8.49895358e-01
4.36022505e-02 1.71933591e-01 3.96667480e-01 1.01934373e-01
-2.02248439e-01 -1.15510011e+00 -5.67235708e-01 7.27813601e-01
-2.53006909e-02 -1.77207798e-01 8.91509712e-01 4.30727065e-01
1.64183199e-01 5.81470191e-01 1.01684415e+00 -6.41896486e-01
6.30642891e-01 -5.88840425e-01 -9.50730085e-01 1.37799144e-01
-6.28427148e-01 -1.04700409e-01 6.63583696e-01 1.21054754e-01
-9.02728260e-01 2.77791530e-01 1.43189847e-01 5.64151764e-01
1.74793482e-01 6.86596930e-01 -4.13024694e-01 -6.84252501e-01
7.87952721e-01 2.13616207e-01 -5.79808474e-01 -5.55590494e-03
2.87314475e-01 6.10218406e-01 4.89530116e-01 -7.00374991e-02
1.39278734e+00 4.92891967e-01 3.52425009e-01 -1.04845595e+00
-4.08036470e-01 -5.41222632e-01 -5.32031834e-01 -1.82647690e-01
2.78927177e-01 -9.29124415e-01 -6.96583390e-01 1.78515527e-03
-7.45953739e-01 -5.72845107e-03 2.82839149e-01 7.70714939e-01
-6.41470313e-01 -2.58122981e-01 3.25383127e-01 -1.17067885e+00
8.95397589e-02 -5.03896236e-01 6.88905001e-01 3.91292095e-01
-3.40515286e-01 -1.25047052e+00 -1.18617704e-02 -3.47383380e-01
1.04597974e+00 5.43112874e-01 7.93950558e-01 -4.17690337e-01
-9.78448272e-01 -5.45980871e-01 6.80597350e-02 -5.42747080e-01
5.38587213e-01 -8.34386051e-01 -8.08653653e-01 -2.42955312e-01
-2.13602468e-01 4.40877914e-01 -4.99764979e-02 7.98930943e-01
8.94008100e-01 -6.70639127e-02 -1.12025762e+00 6.43828869e-01
1.56017303e+00 7.81675577e-01 4.29581970e-01 2.85411179e-01
1.74064413e-01 -1.52112946e-01 5.88478982e-01 5.83448112e-01
1.00597687e-01 8.29799533e-01 3.38218987e-01 -2.24606413e-02
2.63769150e-01 -2.28351071e-01 -1.81499660e-01 2.98672140e-01
1.82343066e-01 -5.18212974e-01 -8.57429564e-01 1.93204343e-01
-1.66611195e+00 -9.47948933e-01 1.93135053e-01 2.26491237e+00
1.51139677e-01 1.93822831e-01 -1.86280429e-01 3.66591334e-01
9.22429919e-01 3.41407567e-01 -5.81835389e-01 7.19352216e-02
-6.17402280e-03 1.98424906e-01 1.53751993e+00 7.31191933e-01
-9.89963233e-01 1.27396017e-01 7.11795950e+00 4.40124512e-01
-1.02438462e+00 1.09740056e-01 -1.15989268e-01 2.05030844e-01
-2.55946487e-01 -2.38000482e-01 -4.17107970e-01 5.54581463e-01
9.27094102e-01 -3.99275839e-01 5.04645169e-01 6.70493662e-01
4.74302709e-01 -5.44204950e-01 -7.76842177e-01 1.25258446e+00
-1.17869049e-01 -1.49628234e+00 -8.28348398e-01 4.10104066e-01
6.86711967e-01 6.87653199e-02 5.41774556e-02 -5.71513623e-02
5.54669440e-01 -7.28323400e-01 6.15673840e-01 5.95464349e-01
5.69317460e-01 -1.00220811e+00 5.34468174e-01 3.23529989e-01
-1.59351289e+00 -4.90619540e-01 2.14951206e-02 -1.41026229e-01
4.52043116e-01 9.01736915e-01 -1.37553227e+00 8.32923949e-01
2.30201751e-01 1.42857075e-01 -2.03029215e-01 1.63739800e+00
-4.12029117e-01 6.57095551e-01 -7.15588570e-01 -4.68354553e-01
1.12426564e-01 1.52979568e-01 4.53685641e-01 1.03224945e+00
9.29713905e-01 -1.51874587e-01 1.94404230e-01 2.73348808e-01
1.79313377e-01 -1.67437479e-01 -6.82819068e-01 6.86626792e-01
1.44134438e+00 1.15109980e+00 -8.31115007e-01 2.76992589e-01
-5.09213507e-01 3.23175818e-01 -6.89874172e-01 7.01598048e-01
-7.45495498e-01 -8.63756359e-01 4.25968498e-01 1.65226340e-01
3.43216538e-01 -6.69423580e-01 -3.51458490e-01 -3.61026943e-01
-3.18986416e-01 -2.90311724e-01 -2.14465827e-01 -4.98631328e-01
-8.09498489e-01 1.29124105e-01 -7.10141510e-02 -1.54194713e+00
-6.17177248e-01 -2.24650204e-01 -4.55750495e-01 8.55292559e-01
-1.28396475e+00 -7.89939702e-01 -4.22901779e-01 5.90365708e-01
2.53465883e-02 -4.12623346e-01 1.03880751e+00 2.82084614e-01
1.35688320e-01 3.89931023e-01 9.01327491e-01 -5.08900322e-02
2.48676375e-01 -1.42728269e+00 3.88926178e-01 7.97877133e-01
4.41815674e-01 3.66799533e-01 9.38392580e-01 -4.73000705e-01
-1.22191370e+00 -8.45126927e-01 3.77137303e-01 -1.81715548e-01
9.32147980e-01 -6.23514652e-01 1.81992516e-01 2.44905576e-01
1.57137990e-01 -3.17136437e-01 9.95954335e-01 5.61144948e-01
8.52529928e-02 -3.67538452e-01 -9.55244720e-01 5.68249643e-01
1.09829628e+00 -3.79617512e-01 1.44800007e-01 3.20004910e-01
9.29661021e-02 -3.06905985e-01 -7.60345578e-01 8.12328234e-02
8.56132090e-01 -8.62202287e-01 9.08223033e-01 5.94606161e-01
-9.50558960e-01 -5.56843638e-01 -4.61922050e-01 -1.71164668e+00
-4.72924322e-01 -1.14960301e+00 1.79685980e-01 9.10183787e-01
8.18615496e-01 -9.53720987e-01 8.83439302e-01 -1.52856067e-01
6.73237368e-02 -1.33799568e-01 -1.22094905e+00 -8.39890242e-01
-5.97768545e-01 -5.86480558e-01 9.63504612e-01 6.65250480e-01
2.27117434e-01 4.36568469e-01 -4.80151653e-01 9.03366387e-01
9.29589033e-01 -2.68478185e-01 7.98287213e-01 -1.47752690e+00
-3.97470035e-02 -1.44534990e-01 -7.92950273e-01 -1.18839598e+00
-3.37890416e-01 -2.61170924e-01 1.91256642e-01 -1.49242628e+00
-6.51001453e-01 -1.11956644e+00 -4.58105356e-01 2.00081185e-01
4.52074617e-01 2.27887303e-01 -1.58175647e-01 -1.65631175e-01
-5.46395659e-01 5.37104569e-02 5.22595227e-01 -1.43133134e-01
-2.50035733e-01 9.01459217e-01 -6.94962800e-01 6.71173036e-01
1.18602240e+00 -2.84775287e-01 -9.03278947e-01 -1.51326597e-01
4.31615710e-01 3.41127038e-01 8.77350271e-02 -2.00492191e+00
7.31042445e-01 -1.05657049e-01 7.37106979e-01 -6.37772441e-01
5.80657482e-01 -1.33718121e+00 4.76841182e-01 5.73145330e-01
7.90409185e-03 -1.38662779e-03 -7.19367042e-02 1.09538174e+00
1.49479091e-01 1.54055059e-01 4.89879578e-01 3.01012337e-01
-7.71017611e-01 -1.71531901e-01 -7.57480145e-01 -5.90637267e-01
9.86159086e-01 -4.63269621e-01 -2.90952533e-01 -1.18278480e+00
-5.63773274e-01 9.50581953e-02 1.35981113e-01 2.03501210e-01
2.41841674e-01 -1.22528899e+00 -1.41076669e-01 9.14491192e-02
-7.29063824e-02 -6.61092997e-01 -1.69152990e-02 8.33979249e-01
-3.18826199e-01 1.07709372e+00 -2.06208825e-01 -5.63188136e-01
-9.14318860e-01 1.79257337e-02 6.47549212e-01 2.45759878e-02
-3.13506573e-01 4.03541356e-01 -2.81129658e-01 -1.41507640e-01
3.04858953e-01 -2.43579745e-01 -1.86099298e-02 -1.33881867e-01
6.25212729e-01 2.94632047e-01 1.53604046e-01 -3.57833117e-01
-6.78185821e-01 8.03229332e-01 4.04839903e-01 -4.46812034e-01
8.35511923e-01 -8.19914937e-01 4.75311637e-01 2.98228055e-01
7.24166393e-01 8.17096472e-01 -1.03899944e+00 -4.25741881e-01
4.37688231e-01 -3.16864878e-01 -7.34511670e-03 -1.16129947e+00
-5.77343822e-01 3.65837081e-03 9.58426297e-01 9.09629166e-01
1.27394295e+00 -2.07539067e-01 1.90223232e-01 8.41789305e-01
1.42443848e+00 -1.22894347e+00 -3.45736146e-01 3.39146882e-01
2.66427279e-01 -6.95114315e-01 2.43816555e-01 -4.29970533e-01
1.97585195e-01 9.57278252e-01 -2.17005163e-01 1.85904324e-01
1.13729179e+00 5.38478017e-01 4.25025016e-01 1.16290212e-01
-9.70739964e-03 -3.86052281e-01 -1.73413858e-01 1.30488586e+00
9.91550907e-02 1.23818271e-01 1.75633430e-02 1.62033051e-01
-7.76270390e-01 -4.59734768e-01 5.55140734e-01 1.17149758e+00
-1.00537288e+00 -1.16002238e+00 -7.75932252e-01 5.05734921e-01
-3.05095464e-01 -9.70081799e-03 1.81195065e-02 1.06319952e+00
1.00506604e-01 1.32688367e+00 -1.08918637e-01 -5.98876595e-01
3.76148552e-01 -2.00391427e-01 1.96053818e-01 -2.63453245e-01
3.25989544e-01 2.76718475e-02 4.36212957e-01 -5.17770410e-01
-5.04111290e-01 -6.59476817e-01 -1.09782553e+00 -1.78000093e-01
-3.02504271e-01 4.32579875e-01 1.23980284e+00 9.69441772e-01
1.75837517e-01 9.66788232e-01 1.10460341e+00 -9.90755737e-01
-7.09930509e-02 -5.81009686e-01 -1.04908800e+00 -6.51690125e-01
4.45962042e-01 -9.15314138e-01 -2.98395101e-02 -3.92886937e-01] | [6.309925556182861, 1.158268928527832] |
735324cf-1001-4982-9136-98695ca2aabb | dynamic-graph-modules-for-modeling-higher | 1812.05637 | null | https://arxiv.org/abs/1812.05637v3 | https://arxiv.org/pdf/1812.05637v3.pdf | Dynamic Graph Modules for Modeling Object-Object Interactions in Activity Recognition | Video action recognition, a critical problem in video understanding, has been gaining increasing attention. To identify actions induced by complex object-object interactions, we need to consider not only spatial relations among objects in a single frame, but also temporal relations among different or the same objects across multiple frames. However, existing approaches that model video representations and non-local features are either incapable of explicitly modeling relations at the object-object level or unable to handle streaming videos. In this paper, we propose a novel dynamic hidden graph module to model complex object-object interactions in videos, of which two instantiations are considered: a visual graph that captures appearance/motion changes among objects and a location graph that captures relative spatiotemporal position changes among objects. Additionally, the proposed graph module allows us to process streaming videos, setting it apart from existing methods. Experimental results on benchmark datasets, Something-Something and ActivityNet, show the competitive performance of our method. | ['Wei zhang', 'Chenliang Xu', 'Luowei Zhou', 'Hao Huang', 'Jason J. Corso'] | 2018-12-13 | null | null | null | null | ['3d-human-action-recognition'] | ['computer-vision'] | [ 2.59858966e-01 -3.10070395e-01 -3.14891458e-01 -1.30391568e-01
2.69425251e-02 -4.53940660e-01 6.99020565e-01 3.68713528e-01
-1.52897209e-01 3.98053318e-01 4.02842194e-01 1.60987288e-01
-3.63468647e-01 -5.85040390e-01 -7.95280099e-01 -7.48841763e-01
-4.98956352e-01 8.74839649e-02 8.70254755e-01 2.78926790e-01
1.28714606e-01 4.84919786e-01 -1.63558686e+00 5.01629531e-01
1.97392464e-01 9.26135182e-01 1.79343820e-01 7.76479304e-01
8.13165158e-02 1.62820482e+00 -2.86309391e-01 3.04809446e-03
2.24293172e-02 -4.76872593e-01 -7.52773285e-01 6.21542394e-01
5.12243211e-01 -4.55597252e-01 -7.38454819e-01 9.06581759e-01
1.66357514e-02 3.84986162e-01 2.22082630e-01 -1.61662829e+00
-5.12371480e-01 5.86941242e-01 -5.18605471e-01 8.42589855e-01
6.92811668e-01 1.89118072e-01 9.77616429e-01 -6.41034901e-01
5.91240346e-01 1.29892850e+00 2.48847440e-01 2.92392254e-01
-9.27226722e-01 -2.47671381e-01 6.78410470e-01 9.84632969e-01
-1.19097984e+00 -4.71401304e-01 7.74752319e-01 -6.63059771e-01
9.83502746e-01 2.55182207e-01 9.83505189e-01 1.05937910e+00
1.96747959e-01 1.03643322e+00 5.99426687e-01 -7.94934034e-02
2.00478122e-01 -4.22066748e-01 2.38829896e-01 6.34531021e-01
2.16341555e-01 -3.02322745e-01 -7.63165534e-01 -1.17251806e-01
9.54134166e-01 6.46603465e-01 -5.27579427e-01 -7.30084479e-01
-1.45253563e+00 4.87586528e-01 3.40931892e-01 5.26454151e-01
-5.93008459e-01 5.29837251e-01 3.02285075e-01 6.55675679e-02
1.96178019e-01 -1.62389934e-01 -1.83784589e-01 -3.94241452e-01
-5.34259021e-01 1.03952959e-01 5.48836470e-01 1.01246428e+00
6.77556098e-01 -1.86668858e-01 -2.50298530e-01 3.38357687e-01
2.13941708e-01 4.16124724e-02 4.70874697e-01 -8.57772112e-01
5.31443536e-01 9.97385681e-01 2.41825767e-02 -1.52228272e+00
-3.06980371e-01 1.09890521e-01 -7.48228371e-01 -1.95152387e-01
2.73719907e-01 2.68667221e-01 -6.87371373e-01 1.60399199e+00
4.28363085e-01 9.16877449e-01 -1.51809275e-01 8.19083631e-01
7.95957863e-01 6.70868814e-01 2.35730693e-01 -4.82865274e-01
1.35594058e+00 -1.05292678e+00 -9.55040097e-01 -8.52858797e-02
6.44840240e-01 -2.84939498e-01 5.58884740e-01 1.32624090e-01
-1.26934183e+00 -6.47926927e-01 -6.36622012e-01 9.50831249e-02
-3.74043614e-01 -1.75425962e-01 5.49499571e-01 3.58501710e-02
-9.22442555e-01 4.26548362e-01 -1.23061216e+00 -5.00499964e-01
3.73257667e-01 4.08666223e-01 -5.41271091e-01 -4.33130115e-02
-8.98078382e-01 4.25769150e-01 6.18468285e-01 2.11349145e-01
-1.04152811e+00 -3.14493239e-01 -9.99482989e-01 2.56194592e-01
9.38003778e-01 -4.89535332e-01 9.05323505e-01 -1.28966928e+00
-1.09535825e+00 3.81696433e-01 -3.01586300e-01 -3.70134473e-01
3.51445913e-01 -2.75198430e-01 -4.92707103e-01 5.68977714e-01
-2.01120898e-01 3.15748006e-01 8.83930504e-01 -1.08941054e+00
-8.06160092e-01 -4.87218618e-01 4.65983570e-01 2.84833461e-01
-4.90036905e-01 1.03292711e-01 -7.80828834e-01 -5.89842200e-01
1.30970985e-01 -9.42986131e-01 -9.65602621e-02 3.66192721e-02
-2.26907432e-01 -3.87490332e-01 1.36638725e+00 -5.69530010e-01
1.42624760e+00 -2.27878094e+00 4.96419996e-01 1.20568007e-01
3.48223805e-01 3.37059021e-01 -9.01450515e-02 5.65241337e-01
-1.60399377e-01 -7.33110756e-02 1.25383697e-02 -7.42756352e-02
-2.19167396e-01 4.42464679e-01 -7.83561915e-02 5.40163696e-01
1.57627776e-01 9.77021396e-01 -1.16668487e+00 -6.47697389e-01
3.55116576e-01 5.70192933e-01 -5.64680874e-01 1.69609129e-01
-1.78917691e-01 4.68702048e-01 -7.10940063e-01 6.31897390e-01
3.03422034e-01 -5.55979192e-01 3.44639689e-01 -2.74137318e-01
-2.91808192e-02 -6.58282861e-02 -1.45402086e+00 1.59893250e+00
3.50272730e-02 4.97776419e-01 -2.09822148e-01 -1.07801521e+00
2.86388546e-01 6.16152108e-01 9.57345128e-01 -4.60676402e-01
-5.35865165e-02 -3.31016809e-01 1.37018524e-02 -1.05807209e+00
2.61813015e-01 3.71079952e-01 2.48551250e-01 1.64889991e-01
1.06593743e-01 6.49777472e-01 6.19545162e-01 3.50046217e-01
1.43208170e+00 2.16958672e-01 4.84247029e-01 1.76624924e-01
6.91958129e-01 -3.57136279e-01 7.34094858e-01 5.77971995e-01
-3.54580522e-01 5.36267996e-01 6.14761293e-01 -7.09339142e-01
-4.47446644e-01 -9.72512722e-01 4.41776246e-01 1.03061080e+00
4.85719830e-01 -7.70611882e-01 -4.76529241e-01 -7.90031254e-01
-1.03059679e-01 8.81099924e-02 -7.78027594e-01 -1.45579964e-01
-8.61263752e-01 -3.90603274e-01 -2.92813347e-04 8.20976138e-01
5.35570979e-01 -1.25926197e+00 -9.67454672e-01 2.87171304e-01
-4.41621244e-01 -1.55667806e+00 -6.49788618e-01 -3.04288298e-01
-8.47670794e-01 -1.52255464e+00 -3.21916163e-01 -7.69857585e-01
6.49141490e-01 5.51199913e-01 9.16942954e-01 4.02932137e-01
-3.68023992e-01 9.39437330e-01 -6.21039093e-01 3.30848880e-02
-2.13596411e-02 -4.08389211e-01 6.85777441e-02 8.58417332e-01
1.97501928e-01 -6.07647419e-01 -6.85866117e-01 4.55061346e-01
-1.24961996e+00 1.16901033e-01 4.89779264e-01 3.84346128e-01
5.39010525e-01 1.52218252e-01 1.92003325e-01 -5.83808959e-01
-2.01464012e-01 -5.73270917e-01 -3.13155532e-01 5.73732674e-01
2.22890645e-01 -2.68635690e-01 3.68368238e-01 -7.63230443e-01
-9.16430056e-01 3.92499208e-01 5.71388304e-01 -7.66420782e-01
-2.98773944e-01 5.01340747e-01 -5.14872074e-01 1.50076017e-01
-5.59273623e-02 3.92624795e-01 -1.41728714e-01 -2.77050197e-01
2.20651716e-01 1.23756804e-01 3.34438860e-01 -1.28377825e-01
6.42947376e-01 8.08899105e-01 2.51896344e-02 -9.89858747e-01
-5.67186952e-01 -7.37124801e-01 -1.00835514e+00 -5.26710093e-01
1.24907386e+00 -8.57707024e-01 -8.28021169e-01 4.76949424e-01
-9.61660564e-01 -2.89932102e-01 -2.45741025e-01 7.24052310e-01
-7.77415216e-01 5.63615322e-01 -5.64031720e-01 -5.75011790e-01
3.38032842e-01 -1.00963998e+00 1.02715755e+00 1.93296194e-01
-1.23528756e-01 -1.24923491e+00 -2.69633345e-02 4.03948069e-01
3.15377861e-02 5.83633602e-01 7.35861897e-01 -3.76766264e-01
-1.08717835e+00 -2.20965669e-01 2.36121472e-03 -1.29161729e-02
5.39434671e-01 1.73730671e-01 -4.70121771e-01 -3.29981685e-01
-6.28978983e-02 9.56542231e-03 7.24151194e-01 3.95860940e-01
1.36412704e+00 -6.13968670e-01 -6.07538760e-01 4.60593253e-01
1.13235331e+00 3.43427151e-01 6.57904744e-01 6.42131343e-02
1.06338942e+00 5.06909311e-01 5.47090590e-01 5.96524954e-01
4.72434789e-01 9.76862609e-01 6.22140944e-01 1.57953843e-01
-8.70102197e-02 -1.57449216e-01 5.29393375e-01 8.07315171e-01
-5.93554497e-01 -4.15939331e-01 -7.84268200e-01 5.00908077e-01
-2.41592240e+00 -1.47270012e+00 -2.79688627e-01 1.94689929e+00
2.85671711e-01 -1.03588834e-01 2.81921595e-01 7.78543428e-02
7.05166936e-01 3.94629091e-01 -4.67582822e-01 2.91318089e-01
-3.21382843e-02 -4.03886855e-01 1.08488739e-01 1.70551509e-01
-1.27085543e+00 8.89754832e-01 5.49736738e+00 2.72977084e-01
-8.44743848e-01 6.05607964e-02 2.26188719e-01 -3.07308227e-01
2.41511554e-01 7.18600601e-02 -4.57626343e-01 4.47841436e-01
5.41536927e-01 2.10754462e-02 3.08544338e-01 5.44989765e-01
3.44461918e-01 -1.53247386e-01 -1.47710443e+00 1.07770216e+00
3.19785416e-01 -1.20699334e+00 4.48359787e-01 9.28164497e-02
4.46398973e-01 -4.47221935e-01 -1.67480528e-01 8.07141736e-02
2.96680797e-02 -7.90969908e-01 7.18720138e-01 7.75587916e-01
5.47260344e-02 -3.99061888e-01 3.38811368e-01 4.39339280e-01
-1.76870263e+00 -3.51238281e-01 -3.78160365e-02 -2.62642056e-01
3.64706039e-01 -2.78613809e-03 -3.64371717e-01 5.79572797e-01
9.41942811e-01 1.36440408e+00 -7.18808472e-01 1.10828698e+00
-6.09029643e-02 4.84120280e-01 4.46782857e-02 1.56721130e-01
2.78785229e-01 -1.76425979e-01 5.34329712e-01 1.08048201e+00
1.82289064e-01 4.94821489e-01 5.95792353e-01 4.24332380e-01
2.06228524e-01 -4.09977436e-02 -7.73684978e-01 -1.19003288e-01
4.47625481e-03 1.17050004e+00 -9.50562358e-01 -5.81735373e-01
-8.71168494e-01 9.84625697e-01 2.68273532e-01 5.63312113e-01
-1.14536774e+00 2.24321321e-01 8.69609237e-01 2.51289278e-01
5.87362885e-01 -5.37337899e-01 7.11265922e-01 -1.49137640e+00
2.68693596e-01 -8.96551728e-01 7.22121119e-01 -7.60877013e-01
-8.72001708e-01 2.21684530e-01 3.54848146e-01 -1.46906352e+00
-1.12565443e-01 -4.51967001e-01 -5.05220592e-01 -4.20144610e-02
-1.27824211e+00 -1.21561718e+00 -6.08407617e-01 1.10010171e+00
8.28521192e-01 1.77598402e-01 4.00388449e-01 3.51397395e-01
-6.86633468e-01 -1.41040822e-02 -3.65017295e-01 3.79227668e-01
1.21363238e-01 -1.00367093e+00 4.25244644e-02 8.67133975e-01
7.15698063e-01 5.23711979e-01 3.52335125e-01 -7.13269889e-01
-1.80395448e+00 -1.18132699e+00 6.37894034e-01 -5.80984175e-01
7.68746972e-01 -3.97688717e-01 -1.10966957e+00 1.18152916e+00
1.31782681e-01 5.24169087e-01 6.42647326e-01 -2.89895356e-01
-1.94585085e-01 5.22224186e-03 -5.79837024e-01 5.46740174e-01
1.49852693e+00 -4.94561613e-01 -5.17227888e-01 3.73856962e-01
4.61667269e-01 -3.04190487e-01 -7.50356197e-01 4.93816197e-01
4.19170558e-01 -9.71152544e-01 1.12876105e+00 -9.22492445e-01
1.30814180e-01 -5.86621463e-01 -1.02941036e-01 -9.04126227e-01
-6.28936589e-01 -4.80261832e-01 -7.59814203e-01 1.16122484e+00
-2.87777781e-01 -3.06007415e-01 6.16248786e-01 5.33561170e-01
3.01948078e-02 -4.37451392e-01 -8.93467188e-01 -8.89886618e-01
-7.59715557e-01 -2.96693623e-01 3.47573340e-01 1.14871132e+00
1.15028806e-01 1.19226933e-01 -5.24260223e-01 5.84082842e-01
3.84809017e-01 1.56564459e-01 8.18887115e-01 -1.03337979e+00
-5.43557584e-01 -3.69781137e-01 -1.19179583e+00 -1.12673283e+00
1.85232088e-01 -5.07541299e-01 -8.51925761e-02 -1.67577434e+00
5.45440137e-01 1.44759849e-01 -4.89869952e-01 5.54597497e-01
-2.16505140e-01 2.34198689e-01 3.37974995e-01 2.84069002e-01
-1.24177945e+00 5.87062359e-01 1.00703764e+00 -2.11639076e-01
-1.71010599e-01 -1.42639711e-01 -4.70068194e-02 1.02653205e+00
3.90275091e-01 -2.64213502e-01 -6.77199304e-01 -5.20034730e-01
-2.57181423e-03 2.20528290e-01 7.57605731e-01 -1.28526998e+00
3.79601538e-01 -4.33007687e-01 2.60407478e-01 -4.15538698e-01
4.87893879e-01 -1.15556264e+00 4.14241433e-01 4.48626488e-01
-2.44891450e-01 3.01982939e-01 -1.51810542e-01 1.01329267e+00
-4.69928712e-01 8.04683268e-02 3.74633700e-01 -2.00156644e-01
-1.23269427e+00 7.98624575e-01 -5.08234441e-01 -9.88838747e-02
1.62155485e+00 -3.10425729e-01 -6.35460615e-02 -4.07801241e-01
-1.01699817e+00 2.25455880e-01 3.00354421e-01 9.63173270e-01
7.72441804e-01 -1.44273579e+00 -3.75513792e-01 6.48000687e-02
3.26720983e-01 -1.99549526e-01 5.22043169e-01 1.14061618e+00
-4.02668536e-01 2.72387981e-01 -1.40049428e-01 -8.09706569e-01
-1.80461359e+00 8.93549919e-01 1.96574882e-01 -1.95657518e-02
-8.67982626e-01 5.81895709e-01 5.43177068e-01 3.00547272e-01
4.28766757e-01 -6.36693716e-01 -5.84003866e-01 8.87145475e-02
7.30639338e-01 4.30543005e-01 -3.64003122e-01 -1.18404138e+00
-4.00962144e-01 7.59549022e-01 1.41030967e-01 3.24416846e-01
1.24166250e+00 -2.37989172e-01 -1.28523543e-01 9.03348386e-01
1.04165411e+00 -3.26919705e-01 -1.59208071e+00 -4.73782480e-01
2.88929433e-01 -7.83663332e-01 -1.42573178e-01 -1.52445316e-01
-1.29411077e+00 7.45920002e-01 4.00254011e-01 6.40502870e-01
1.16204393e+00 2.36682370e-01 6.52338922e-01 3.77366006e-01
4.41428870e-01 -7.83857882e-01 7.17844665e-01 3.30659062e-01
8.90037715e-01 -1.03060579e+00 3.57871167e-02 -5.26267469e-01
-6.08547151e-01 1.09671259e+00 7.51285851e-01 -2.17976496e-02
7.90129304e-01 -1.91126801e-02 -2.48167753e-01 -3.22093189e-01
-9.46486473e-01 -3.07042241e-01 3.74722600e-01 3.19842398e-01
1.74176097e-01 -2.09720820e-01 5.49412630e-02 4.44666632e-02
5.92466474e-01 6.49611875e-02 4.23419356e-01 1.39756238e+00
-2.82600671e-01 -8.49827170e-01 -1.03567861e-01 1.74826041e-01
-3.94138426e-01 3.05816323e-01 -3.77076179e-01 8.14541101e-01
2.75974929e-01 8.88628602e-01 4.98724878e-01 -2.70386130e-01
3.31151843e-01 2.04857551e-02 5.27200043e-01 -6.16475224e-01
-3.27133954e-01 -1.53346518e-02 -2.95420974e-01 -1.08463037e+00
-1.10990977e+00 -1.07236576e+00 -1.34356797e+00 1.42783433e-01
-1.40801251e-01 -8.31901878e-02 1.69311449e-01 1.01317525e+00
3.99136871e-01 5.43485343e-01 3.97540301e-01 -8.42369914e-01
2.75965221e-02 -7.09732890e-01 -7.15358019e-01 8.91386330e-01
5.51478982e-01 -8.61940920e-01 -1.73789218e-01 6.53907359e-01] | [8.488816261291504, 0.6490561366081238] |
9cbec039-dc54-46cd-aef9-ef3056086806 | pseudo-label-correction-and-learning-for-semi | 2303.02998 | null | https://arxiv.org/abs/2303.02998v1 | https://arxiv.org/pdf/2303.02998v1.pdf | Pseudo-label Correction and Learning For Semi-Supervised Object Detection | Pseudo-Labeling has emerged as a simple yet effective technique for semi-supervised object detection (SSOD). However, the inevitable noise problem in pseudo-labels significantly degrades the performance of SSOD methods. Recent advances effectively alleviate the classification noise in SSOD, while the localization noise which is a non-negligible part of SSOD is not well-addressed. In this paper, we analyse the localization noise from the generation and learning phases, and propose two strategies, namely pseudo-label correction and noise-unaware learning. For pseudo-label correction, we introduce a multi-round refining method and a multi-vote weighting method. The former iteratively refines the pseudo boxes to improve the stability of predictions, while the latter smoothly self-corrects pseudo boxes by weighing the scores of surrounding jittered boxes. For noise-unaware learning, we introduce a loss weight function that is negatively correlated with the Intersection over Union (IoU) in the regression task, which pulls the predicted boxes closer to the object and improves localization accuracy. Our proposed method, Pseudo-label Correction and Learning (PCL), is extensively evaluated on the MS COCO and PASCAL VOC benchmarks. On MS COCO, PCL outperforms the supervised baseline by 12.16, 12.11, and 9.57 mAP and the recent SOTA (SoftTeacher) by 3.90, 2.54, and 2.43 mAP under 1\%, 5\%, and 10\% labeling ratios, respectively. On PASCAL VOC, PCL improves the supervised baseline by 5.64 mAP and the recent SOTA (Unbiased Teacherv2) by 1.04 mAP on AP$^{50}$. | ['Yulan Guo', 'Zhengfa Liang', 'Yusong Tan', 'Ke Liang', 'Wei Chen', 'Yulin He'] | 2023-03-06 | null | null | null | null | ['semi-supervised-object-detection'] | ['computer-vision'] | [ 1.45277530e-01 -1.06592335e-01 1.96612123e-02 -4.59372103e-01
-1.17161191e+00 -4.80899483e-01 4.70563203e-01 3.14813405e-01
-8.32278788e-01 8.67875278e-01 -2.57149011e-01 8.04372281e-02
1.64480671e-01 -4.59457040e-01 -8.43388438e-01 -9.89092350e-01
6.52672470e-01 3.31353486e-01 7.76843786e-01 2.13750958e-01
2.97021791e-02 2.26840824e-01 -1.47235501e+00 1.40964270e-01
1.05816376e+00 1.24694812e+00 2.56244361e-01 4.88848418e-01
-1.67064965e-01 6.70922458e-01 -7.70680368e-01 -4.10092354e-01
1.90476090e-01 -2.20930845e-01 -4.95231241e-01 -1.02537058e-01
8.95026863e-01 1.17759369e-01 3.81782115e-03 1.45840788e+00
6.21771812e-01 1.13215812e-01 7.18674660e-01 -1.12424326e+00
-2.53420979e-01 4.35014188e-01 -1.00183010e+00 -2.35232562e-02
-2.55736679e-01 -6.35662526e-02 8.60240221e-01 -1.16136372e+00
4.66403455e-01 9.92252648e-01 9.87822235e-01 6.05755091e-01
-1.37954938e+00 -1.04780555e+00 6.69512078e-02 -1.67534314e-02
-1.60099423e+00 -2.73338705e-01 4.11634982e-01 -5.53463399e-01
4.61324096e-01 1.87626734e-01 1.79169118e-01 7.67391801e-01
5.93184382e-02 6.65899098e-01 1.38497698e+00 -4.59799469e-01
2.19663918e-01 4.48154926e-01 3.64961475e-01 6.78386867e-01
3.71857584e-01 5.57183288e-02 -4.19033825e-01 5.72905177e-04
3.47220808e-01 -1.08067796e-01 -1.40744165e-01 -4.07464802e-01
-8.69100988e-01 5.24710178e-01 6.45186305e-01 -5.91803640e-02
8.05983096e-02 1.95708200e-01 2.95838118e-01 -2.45542333e-01
6.76301718e-01 3.78354847e-01 -4.06886667e-01 2.41583046e-02
-1.11392713e+00 2.22502321e-01 3.86725008e-01 9.29520071e-01
9.12050486e-01 -4.34630588e-02 -4.81808364e-01 1.14319241e+00
2.87067205e-01 8.11962545e-01 2.80048609e-01 -8.69663000e-01
3.08106720e-01 6.34824514e-01 3.87484998e-01 -7.43766546e-01
-4.53134805e-01 -8.74112606e-01 -6.76510394e-01 2.47573987e-01
7.15339601e-01 -8.71719122e-02 -1.13277638e+00 1.73140109e+00
5.30893624e-01 1.94381446e-01 -3.84047985e-01 9.47319865e-01
9.05506670e-01 4.85042542e-01 3.61946911e-01 -2.13483080e-01
1.17293870e+00 -1.35224748e+00 -6.73203588e-01 -4.29438233e-01
6.62563860e-01 -9.33915675e-01 9.28490877e-01 3.45994473e-01
-7.20852792e-01 -8.00330818e-01 -9.79255199e-01 1.38717413e-01
-1.82501271e-01 5.82364380e-01 9.70563293e-02 5.65033019e-01
-7.15696275e-01 3.76639307e-01 -7.27871776e-01 -1.31124273e-01
6.97332442e-01 2.66309917e-01 -2.07606316e-01 -1.24085322e-01
-8.66809666e-01 7.89882183e-01 2.72129238e-01 -9.24660563e-02
-8.87416482e-01 -6.45360649e-01 -5.75123787e-01 -2.13782579e-01
6.24629378e-01 -1.06795974e-01 1.25163996e+00 -8.09121728e-01
-1.20755780e+00 9.19264078e-01 -1.48476467e-01 -5.03761649e-01
6.78424776e-01 -4.02219713e-01 -4.00108725e-01 -1.38870478e-01
4.42723721e-01 1.03004110e+00 8.98349702e-01 -1.41710365e+00
-1.03934479e+00 -3.15538168e-01 -3.12571079e-01 2.14267880e-01
-6.32384196e-02 -1.09014958e-01 -6.41029000e-01 -6.75155878e-01
2.80424118e-01 -1.16536534e+00 -2.10771799e-01 1.40345588e-01
-4.27378476e-01 -1.86424002e-01 7.61383593e-01 -4.41820234e-01
1.17119706e+00 -2.35801244e+00 -3.11001688e-01 1.12544876e-02
2.46148273e-01 5.94099879e-01 -5.27870096e-02 -3.48271161e-01
4.62251529e-02 -8.66940245e-02 -2.51616061e-01 -7.40521789e-01
-1.96716860e-01 6.64587989e-02 -8.16248879e-02 4.74457979e-01
-6.94092829e-03 6.56689525e-01 -1.02753544e+00 -4.50952530e-01
1.63970202e-01 4.41450566e-01 -3.43427658e-01 1.53711975e-01
-1.50317162e-01 3.06835204e-01 -5.00244722e-02 5.28441250e-01
9.25167799e-01 -2.67308742e-01 -2.23056599e-01 -2.65223414e-01
-2.28763521e-01 1.18950114e-01 -1.49323833e+00 1.47041917e+00
-3.03365350e-01 5.41754365e-01 1.48762064e-02 -4.76536572e-01
1.05899775e+00 -3.82226855e-02 1.50375262e-01 -5.82639635e-01
1.34590656e-01 4.47145522e-01 -2.43294030e-01 -6.41386434e-02
5.40199041e-01 3.55223268e-02 3.79573852e-02 2.26350710e-01
9.12761167e-02 -4.89228219e-02 2.12597311e-01 9.08469483e-02
8.61646056e-01 2.66329616e-01 2.24511445e-01 -2.90594906e-01
5.62567115e-01 1.44001059e-02 8.71661365e-01 8.30144346e-01
-5.95654607e-01 8.53801906e-01 2.73532838e-01 -2.19883859e-01
-7.82883644e-01 -1.04733825e+00 -4.15633440e-01 1.23797786e+00
5.10464013e-01 -2.88907349e-01 -9.95163739e-01 -1.10804343e+00
2.31918007e-01 6.63749158e-01 -5.38702548e-01 -1.81359544e-01
-3.99266332e-01 -9.15672243e-01 5.34013093e-01 6.42036617e-01
6.58953965e-01 -7.83052146e-01 -6.64213598e-02 1.74851775e-01
-1.71226084e-01 -1.15409505e+00 -7.60365427e-01 5.40867746e-01
-4.81132388e-01 -9.31545138e-01 -6.87488675e-01 -7.47712433e-01
8.50422263e-01 5.28047919e-01 7.06285596e-01 -1.74987987e-01
-1.14712454e-01 -1.93752706e-01 -2.35854730e-01 -3.67034793e-01
-2.49498695e-01 6.93231970e-02 2.38312721e-01 1.63554460e-01
4.31091040e-01 8.01349990e-03 -5.19502938e-01 8.65585864e-01
-4.82419103e-01 -8.77843238e-04 5.09633958e-01 1.11787879e+00
1.08700132e+00 -1.74982250e-01 3.71423006e-01 -1.16577613e+00
1.73677877e-02 -2.50569582e-01 -9.28111494e-01 1.01455033e-01
-1.04761124e+00 2.58280188e-02 4.02688086e-01 -5.54298818e-01
-1.18184376e+00 3.71970206e-01 -5.55090532e-02 -5.32107890e-01
-2.42984921e-01 -1.88772932e-01 -5.34260720e-02 -3.83222044e-01
9.76054251e-01 -6.09836727e-02 -4.86379176e-01 -6.69585228e-01
2.78159797e-01 7.35207736e-01 6.50004506e-01 -2.75487810e-01
9.27730799e-01 4.91367757e-01 -2.44835079e-01 -4.55475301e-01
-1.45196545e+00 -6.83088958e-01 -4.51823235e-01 -1.78877562e-01
7.94369340e-01 -1.01575565e+00 -1.54336274e-01 5.44974685e-01
-1.04335320e+00 -2.93109298e-01 -4.45003092e-01 4.10636246e-01
-8.08347091e-02 2.14776069e-01 -3.92909706e-01 -7.63072133e-01
-3.72729242e-01 -1.25834918e+00 1.08356977e+00 5.29994965e-01
-1.15453884e-01 -4.19486076e-01 -6.09380044e-02 5.57888091e-01
3.06764245e-01 5.88063933e-02 2.74721414e-01 -6.37921631e-01
-4.42127705e-01 -2.67410010e-01 -7.49340296e-01 6.89497828e-01
-4.57762629e-02 -1.92202210e-01 -1.33004320e+00 -3.25171322e-01
-2.83779800e-01 -3.96813512e-01 9.43210185e-01 3.61036122e-01
1.18353641e+00 1.28360972e-01 -4.33728039e-01 6.23421371e-01
1.25860822e+00 1.35346219e-01 2.48246044e-01 2.13890389e-01
7.58696973e-01 3.95567656e-01 1.00024426e+00 2.29550138e-01
2.24794134e-01 8.13732684e-01 4.61500704e-01 -9.03996639e-03
-5.86064756e-01 -4.21774179e-01 1.46437258e-01 6.25480115e-01
1.81723997e-01 -9.17617902e-02 -8.65415335e-01 4.58067149e-01
-1.83520687e+00 -4.79846984e-01 -5.40306866e-01 2.13838530e+00
1.01839209e+00 5.48720121e-01 -1.31710961e-01 1.36071280e-01
9.65814233e-01 5.52182458e-02 -4.27844048e-01 1.17772341e-01
-8.73411298e-02 1.72591522e-01 8.52680206e-01 4.44188654e-01
-1.47656214e+00 1.03906226e+00 4.71930981e+00 1.34639680e+00
-1.02320671e+00 5.50048053e-01 6.78353131e-01 6.18846528e-02
3.68453175e-01 -7.00734779e-02 -1.41943729e+00 7.18883276e-01
6.26862049e-01 3.93981755e-01 2.30515480e-01 1.29023695e+00
2.95471735e-02 -4.79287356e-01 -8.63868296e-01 1.14817464e+00
1.45256713e-01 -1.06006467e+00 -3.10735375e-01 -3.35419685e-01
1.19431984e+00 1.81779876e-01 9.39355940e-02 3.74202311e-01
1.54992014e-01 -7.32026815e-01 1.06868422e+00 2.85804540e-01
1.00089121e+00 -6.16655648e-01 1.11265755e+00 4.52587366e-01
-1.20793641e+00 1.16577623e-02 -3.32312793e-01 2.18732014e-01
-5.93413077e-02 8.16445887e-01 -6.61434412e-01 1.04270488e-01
9.33493972e-01 3.31349581e-01 -8.61477315e-01 1.29763746e+00
-7.14518845e-01 7.38061130e-01 -3.46863389e-01 -3.09596621e-02
1.40611142e-01 2.46853139e-02 5.66192567e-01 1.32380271e+00
-6.81850538e-02 -1.70888692e-01 1.01619974e-01 7.94539094e-01
-3.07536244e-01 -4.00563423e-03 7.70298168e-02 5.69555104e-01
6.46426916e-01 1.47732997e+00 -7.50618458e-01 -2.06386670e-01
-6.02358431e-02 7.70514071e-01 3.53845090e-01 2.63580561e-01
-1.06557345e+00 -5.94493687e-01 3.15325499e-01 1.87871233e-01
2.35943556e-01 7.05095455e-02 -6.19814754e-01 -8.16836059e-01
3.36792134e-02 -5.95635116e-01 2.61624426e-01 -6.69795871e-01
-1.03509533e+00 6.60743475e-01 -1.61338568e-01 -1.10897946e+00
3.45747381e-01 -4.25364345e-01 -3.45809191e-01 7.83835948e-01
-1.45870543e+00 -1.02794576e+00 -7.42192030e-01 3.51726674e-02
5.15085161e-01 7.93977678e-02 4.45274532e-01 6.35478377e-01
-8.00020993e-01 9.38978672e-01 2.38299236e-01 4.03021052e-02
1.11002088e+00 -1.32898593e+00 2.02043191e-01 8.22089911e-01
2.07053468e-01 2.17340365e-01 5.82106888e-01 -5.71258485e-01
-4.94262159e-01 -1.63129532e+00 9.82146680e-01 -5.40466309e-01
2.81676531e-01 -5.92536211e-01 -9.23274338e-01 3.01882863e-01
-4.07769799e-01 5.28264344e-01 2.04122216e-01 -2.24887609e-01
-4.54145640e-01 -5.24596453e-01 -1.26724303e+00 3.77697468e-01
9.53019261e-01 -3.64498705e-01 -2.46403560e-01 3.25562626e-01
7.16373503e-01 -5.98256052e-01 -5.17381072e-01 5.75781941e-01
3.75039607e-01 -7.78953075e-01 7.52933264e-01 3.78673971e-02
-4.61870171e-02 -8.25862169e-01 -1.22946523e-01 -1.11704397e+00
-4.10591871e-01 -3.51361454e-01 5.82747757e-02 1.42180395e+00
5.41080177e-01 -3.75202477e-01 9.48754907e-01 2.07482770e-01
-1.89374149e-01 -7.90846825e-01 -1.06569695e+00 -7.38817334e-01
-2.69307643e-01 -4.34298873e-01 2.68609822e-01 7.22905636e-01
-6.46956086e-01 2.59952933e-01 -2.67364889e-01 1.88419610e-01
7.97805190e-01 -2.11782664e-01 7.88913906e-01 -1.08242917e+00
-1.92246795e-01 -2.46589810e-01 -7.31506795e-02 -1.18970418e+00
-1.55506119e-01 -7.17502177e-01 4.41119641e-01 -1.15231407e+00
2.95488268e-01 -8.40509236e-01 -4.20104504e-01 6.01201713e-01
-4.37072009e-01 8.53092313e-01 6.31023869e-02 2.77189136e-01
-9.28623736e-01 5.33961177e-01 9.72326100e-01 -1.05924375e-01
-8.94947425e-02 1.45490304e-01 -3.92801106e-01 9.69181716e-01
6.69310749e-01 -1.00878179e+00 2.93558277e-02 -3.11978847e-01
-3.26374248e-02 -5.50164640e-01 2.61282086e-01 -1.18638015e+00
2.88053870e-01 7.50791430e-02 3.62690449e-01 -7.38341153e-01
2.24078313e-01 -7.35838711e-01 -2.76000630e-02 5.29126287e-01
-2.59713382e-01 -3.91479015e-01 1.65238485e-01 7.40525484e-01
-1.39284834e-01 -3.55804890e-01 1.44185102e+00 1.99600846e-01
-6.99248791e-01 8.24986771e-02 1.10429645e-01 3.29779655e-01
1.11829448e+00 -1.87562749e-01 -3.23771298e-01 -7.79178832e-03
-5.42869508e-01 3.53798360e-01 3.12531769e-01 3.36223125e-01
3.07128519e-01 -1.28702068e+00 -4.96521980e-01 2.08064169e-01
2.57126659e-01 3.50732625e-01 1.15954727e-01 9.21905160e-01
-4.38389242e-01 1.58362702e-01 2.59381354e-01 -8.68363500e-01
-1.30563307e+00 9.86165628e-02 4.05079097e-01 -2.72295773e-01
-1.24011740e-01 1.34969771e+00 1.74702436e-01 -5.62394500e-01
7.52788186e-01 -2.53317058e-01 -9.24896300e-02 7.54097477e-02
4.56399918e-01 7.21702516e-01 2.11015910e-01 -7.07597136e-01
-3.84856343e-01 6.82147264e-01 -2.43254095e-01 2.06428841e-01
9.93972540e-01 -6.68990240e-02 9.49749500e-02 3.78633738e-01
8.94365609e-01 1.15182355e-01 -1.50453484e+00 -4.93723541e-01
6.62110522e-02 -3.73150647e-01 1.57628208e-01 -1.03227186e+00
-9.03751671e-01 8.30011368e-01 9.21761513e-01 -1.78107038e-01
8.72158289e-01 -6.81510344e-02 7.59578764e-01 5.91166554e-06
3.48934263e-01 -1.32351649e+00 1.69645339e-01 5.46566188e-01
5.64530373e-01 -1.35612929e+00 7.04152510e-02 -6.08345330e-01
-6.13316059e-01 6.54551923e-01 1.10973036e+00 5.03923185e-02
5.28319240e-01 2.86792696e-01 1.02599204e-01 2.31702656e-01
-3.49130422e-01 -9.74147096e-02 4.69692826e-01 3.23568970e-01
8.76969099e-02 1.01701260e-01 -3.55573982e-01 7.74008751e-01
1.57870620e-01 -2.45986223e-01 9.29113552e-02 6.94911301e-01
-7.04072356e-01 -8.29115152e-01 -6.67301536e-01 4.84461188e-01
-3.65955800e-01 -5.55210235e-03 -1.91470996e-01 5.33083737e-01
8.72591317e-01 9.07873750e-01 5.51824532e-02 -4.11737353e-01
4.57760543e-01 4.17923331e-02 4.67931479e-03 -7.52372503e-01
-5.90219438e-01 2.97759444e-01 -5.46692684e-02 -3.99639428e-01
-3.12593341e-01 -5.55619061e-01 -1.27484751e+00 -8.91363025e-02
-1.05172503e+00 2.19636351e-01 7.56226659e-01 6.81200981e-01
2.05378130e-01 5.10540545e-01 4.91996884e-01 -6.11615360e-01
-7.21594810e-01 -1.07841349e+00 -5.36948800e-01 2.81230927e-01
1.06917195e-01 -8.96547258e-01 -5.73084891e-01 -1.59821898e-01] | [9.153697967529297, 1.2654545307159424] |
3a694aa4-56a4-40fe-94c3-6db4ab428292 | expr-at-semeval-2018-task-9-a-combined | null | null | https://aclanthology.org/S18-1150 | https://aclanthology.org/S18-1150.pdf | EXPR at SemEval-2018 Task 9: A Combined Approach for Hypernym Discovery | In this paper, we present our proposed system (EXPR) to participate in the hypernym discovery task of SemEval 2018. The task addresses the challenge of discovering hypernym relations from a text corpus. Our proposal is a combined approach of path-based technique and distributional technique. We use dependency parser on a corpus to extract candidate hypernyms and represent their dependency paths as a feature vector. The feature vector is concatenated with a feature vector obtained using Wikipedia pre-trained term embedding model. The concatenated feature vector fits a supervised machine learning method to learn a classifier model. This model is able to classify new candidate hypernyms as hypernym or not. Our system performs well to discover new hypernyms not defined in gold hypernyms. | ["Nicolas B{\\'e}chet", 'Ahmad Issa Alaa Aldine', 'Mounira Harzallah', 'Giuseppe Berio', 'Ahmad Faour'] | 2018-06-01 | null | null | null | semeval-2018-6 | ['hypernym-discovery'] | ['natural-language-processing'] | [ 2.85162460e-02 5.63315809e-01 -3.09276432e-01 -2.71696359e-01
-2.18612626e-02 -5.86050928e-01 8.65962565e-01 6.85019016e-01
-8.27508450e-01 8.57585371e-01 3.16370368e-01 -4.59591269e-01
-4.49506015e-01 -1.25795603e+00 -2.54654318e-01 -5.55994101e-02
-1.64520800e-01 1.02707314e+00 2.82447547e-01 -6.79781497e-01
1.33522570e-01 3.45635831e-01 -1.73496997e+00 2.26764649e-01
8.56337786e-01 1.75511658e-01 1.42970294e-01 4.73609477e-01
-6.74078047e-01 3.37780178e-01 -3.92071277e-01 -6.85953975e-01
4.57305670e-01 2.81215236e-02 -1.41921270e+00 -6.92949295e-01
3.96869361e-01 4.07456726e-01 -3.84585053e-01 1.16423011e+00
1.52738020e-01 2.42690742e-01 7.81916082e-01 -1.36650848e+00
-3.25855941e-01 9.91912067e-01 -1.10927865e-01 2.73687452e-01
7.01435626e-01 -4.62860107e-01 1.47865331e+00 -6.16580188e-01
1.25548899e+00 1.06984198e+00 3.73168141e-01 4.34489787e-01
-8.00374568e-01 -6.75697863e-01 -4.16907430e-01 6.15966260e-01
-1.18375516e+00 1.35410681e-01 2.43580475e-01 -1.76165655e-01
1.91008341e+00 2.72267580e-01 8.36735666e-01 5.44378877e-01
-1.57359585e-01 8.89999121e-02 8.00831735e-01 -1.06643760e+00
-1.13097385e-01 2.33339965e-01 8.22306275e-01 7.67650604e-01
6.68592572e-01 -1.09538145e-01 -3.04893553e-01 -4.66594577e-01
2.28153452e-01 -2.54351705e-01 5.48925158e-03 2.39927545e-02
-1.01678920e+00 8.53502333e-01 1.29384175e-01 7.66431868e-01
-3.84317666e-01 -7.29421079e-02 5.16109765e-01 5.73906064e-01
1.87003221e-02 8.75192761e-01 -6.73722684e-01 8.35931767e-03
-5.71445644e-01 3.14638793e-01 1.21305203e+00 1.00455701e+00
9.02266562e-01 -6.49208009e-01 7.80794546e-02 1.02135396e+00
4.44851488e-01 1.87614053e-01 9.98040557e-01 -2.70172507e-01
5.02026379e-01 1.19489527e+00 -1.45363197e-01 -6.13348246e-01
-6.09488130e-01 1.35210946e-01 1.09691143e-01 5.60895866e-03
1.34312753e-02 -2.30084211e-02 -8.89092803e-01 1.35990882e+00
7.37603664e-01 2.88192689e-01 5.08126020e-01 3.43492955e-01
1.07610178e+00 6.44255877e-01 4.71052200e-01 -1.05761707e-01
1.70032680e+00 -6.64133847e-01 -6.20772839e-01 2.99726158e-01
1.16721725e+00 -8.49409819e-01 7.10202932e-01 2.19826087e-01
-4.84888792e-01 -2.83543430e-02 -1.20833302e+00 1.71611998e-02
-1.27261269e+00 -1.80755794e-01 7.16653824e-01 5.45785606e-01
-5.39369941e-01 7.84021616e-01 -3.08053702e-01 -1.08131146e+00
-8.85094702e-03 6.63414001e-01 -8.96496952e-01 8.61049455e-04
-1.60989821e+00 1.46364236e+00 1.47927701e+00 -4.91320461e-01
-2.01571971e-01 -5.17185450e-01 -1.09588659e+00 -4.18036208e-02
2.83176899e-01 -4.89183187e-01 8.58283341e-01 -4.58701849e-01
-8.50437701e-01 1.17593277e+00 -2.80974396e-02 -7.68155217e-01
-3.43485713e-01 -5.54947299e-04 -1.20779276e+00 -4.38712947e-02
2.36668050e-01 2.64969170e-01 3.35322469e-01 -9.51954901e-01
-1.00589502e+00 -2.63446450e-01 -1.06936870e-02 1.64201722e-01
-6.81236506e-01 2.18787119e-01 -3.57718542e-02 -2.56757021e-01
-4.41939868e-02 -7.78424859e-01 -1.12250276e-01 -9.90871012e-01
-4.00920331e-01 -6.87318861e-01 8.27800810e-01 -4.82111543e-01
1.62530184e+00 -1.44465983e+00 -3.15817446e-01 6.20429635e-01
3.20391685e-01 7.53929019e-01 -2.36167416e-01 9.96174872e-01
-6.18779957e-01 2.96430528e-01 -2.74325460e-01 4.94783670e-01
8.04959564e-04 6.53854012e-01 -1.74897268e-01 2.92802323e-03
7.98978657e-02 8.37139666e-01 -1.06645513e+00 -6.56987190e-01
2.21750081e-01 -2.72252727e-02 -5.98267950e-02 2.47568488e-01
-3.33997369e-01 -2.64762044e-01 -2.96335340e-01 3.96509171e-01
6.10155046e-01 3.32903922e-01 6.21292949e-01 -3.48317593e-01
-9.45411250e-02 4.93369520e-01 -1.25480759e+00 1.56185079e+00
-5.23255885e-01 3.93053949e-01 -9.19266343e-01 -1.04925454e+00
1.10338235e+00 5.55347860e-01 3.80672395e-01 -4.90320623e-01
2.16112837e-01 5.65886557e-01 3.30367535e-01 -1.27675188e+00
6.20358348e-01 -1.21302672e-01 2.80751139e-02 4.55415308e-01
8.41254950e-01 4.31241184e-01 7.10935354e-01 3.57493758e-01
1.56315088e+00 1.89245746e-01 9.53163803e-01 -2.48835996e-01
6.53829515e-01 4.32397008e-01 4.35864985e-01 3.33731204e-01
2.93909132e-01 -2.49200881e-01 4.43292439e-01 -6.72159553e-01
-1.27490079e+00 -9.72140789e-01 -4.62671578e-01 9.02144134e-01
-7.60802254e-02 -1.11329401e+00 -3.50400090e-01 -1.27798748e+00
4.99015413e-02 8.11513543e-01 -5.19960642e-01 7.51873292e-03
-7.37352073e-01 -6.65285051e-01 9.02963877e-01 -3.14585455e-02
-6.71986267e-02 -1.50175369e+00 -6.11950219e-01 3.07849318e-01
1.60377756e-01 -1.02841842e+00 2.80530334e-01 5.15031397e-01
-5.06259203e-01 -1.57541072e+00 1.28437161e-01 -1.10229707e+00
5.09427428e-01 -4.37660575e-01 1.17675054e+00 1.35029063e-01
-8.83810937e-01 2.67051980e-02 -9.14526582e-01 -4.77513760e-01
-4.98449862e-01 3.69984299e-01 1.08601481e-01 -5.22416234e-01
1.17301714e+00 -7.56742001e-01 -6.25518188e-02 -3.64099085e-01
-1.09543025e+00 -5.53895235e-01 3.58079642e-01 6.98169231e-01
1.79451704e-01 4.37083021e-02 4.89531875e-01 -1.65721512e+00
8.95748436e-01 -9.21746850e-01 -3.54787678e-01 6.80384755e-01
-9.96088266e-01 6.38810575e-01 4.88808423e-01 -3.45636934e-01
-9.91527438e-01 2.78567195e-01 -3.50800931e-01 2.34625980e-01
-3.61906052e-01 7.73235381e-01 -2.03930020e-01 4.02671061e-02
7.83722579e-01 -1.96262315e-01 -5.58251560e-01 -6.04591429e-01
8.97915721e-01 8.19820940e-01 6.11026943e-01 -6.82697117e-01
9.29339826e-01 -6.28367290e-02 2.92875618e-01 -7.36817420e-01
-3.63447249e-01 -1.03064358e+00 -1.19025767e+00 1.68572411e-01
7.39995658e-01 -5.31115353e-01 -4.59500134e-01 -3.78095031e-01
-1.40389991e+00 3.85046393e-01 -3.30190301e-01 6.83753312e-01
-8.49041119e-02 4.61915493e-01 -1.29119113e-01 -6.26875341e-01
-7.32326031e-01 -1.52194023e-01 4.87092048e-01 1.94270000e-01
-5.67386687e-01 -1.09870887e+00 8.69138062e-01 1.85609348e-02
-1.30041808e-01 2.86775827e-01 1.53836751e+00 -1.82850850e+00
1.23249277e-01 -6.17541194e-01 -1.23011684e-02 -8.14691260e-02
7.59759992e-02 1.61454126e-01 -6.97009861e-01 2.55141109e-01
-5.58138430e-01 -1.84233710e-01 8.48876536e-01 -3.65499824e-01
4.46683228e-01 -5.52299559e-01 -7.23813832e-01 4.70621109e-01
1.91248035e+00 4.01673242e-02 6.96176648e-01 7.01526284e-01
6.76172733e-01 7.92231977e-01 8.21303368e-01 2.59526581e-01
3.04420829e-01 4.52556789e-01 3.59519452e-01 5.32341123e-01
3.76141779e-02 -5.07884800e-01 -7.34918565e-02 8.44731688e-01
2.26384159e-02 -1.01462804e-01 -1.26336646e+00 6.62873566e-01
-2.00644183e+00 -9.87045109e-01 -6.48471832e-01 2.05606937e+00
9.79547322e-01 -9.43862498e-02 3.55153158e-02 1.65033743e-01
6.82634115e-01 -2.65263230e-01 2.17238784e-01 -9.37661946e-01
-1.31209418e-02 1.09450936e+00 6.66054189e-01 7.00461626e-01
-8.91653121e-01 1.50773764e+00 5.00288105e+00 7.19838440e-01
-5.92137814e-01 1.85964614e-01 -6.85298085e-01 2.53028512e-01
-5.93449712e-01 6.66360974e-01 -1.05018079e+00 7.09880665e-02
1.10091114e+00 -3.85414779e-01 1.80218309e-01 6.75331771e-01
-5.27477562e-01 1.08395681e-01 -9.43837821e-01 7.66054511e-01
3.15948367e-01 -1.19025016e+00 4.00826156e-01 -2.34928969e-02
3.67502123e-01 1.61564291e-01 -6.26633584e-01 5.42330146e-01
5.46844065e-01 -1.24708354e+00 -1.51096269e-01 2.57723093e-01
6.09170854e-01 -7.06618607e-01 7.57759392e-01 1.83997810e-01
-1.33720708e+00 -1.31131858e-01 -7.81799734e-01 -1.03034034e-01
1.63582370e-01 5.14582574e-01 -1.39924657e+00 8.21007907e-01
3.85253429e-01 6.94124162e-01 -5.79805851e-01 1.31149828e+00
-6.98330224e-01 4.01054621e-01 -4.88021761e-01 -1.48454949e-01
1.29236475e-01 -8.36110339e-02 7.11314023e-01 1.54643476e+00
2.73186326e-01 1.00446098e-01 -1.77032322e-01 5.50526798e-01
-1.37299046e-01 8.12370420e-01 -1.01637745e+00 -1.66133270e-01
8.52194011e-01 1.75273454e+00 -4.84874517e-01 -5.31906247e-01
-3.49298030e-01 8.42356026e-01 6.20589554e-01 1.65458377e-02
-4.04839844e-01 -1.00080884e+00 4.77140814e-01 -6.59774393e-02
1.27811125e-03 1.33646622e-01 5.87113798e-02 -1.18589342e+00
-4.24269326e-02 -4.58291143e-01 1.02695584e+00 -3.82191628e-01
-1.29699349e+00 9.49534953e-01 2.28838325e-01 -1.05642498e+00
-3.46456468e-01 -8.42662513e-01 -9.14819360e-01 8.00489068e-01
-1.37676585e+00 -1.49181223e+00 -6.81187883e-02 4.77813870e-01
1.69143483e-01 -5.07865787e-01 1.40640008e+00 3.58246207e-01
-2.27196082e-01 4.31406587e-01 -2.51926959e-01 2.66799182e-01
8.85289550e-01 -1.40546870e+00 5.05299628e-01 6.91681683e-01
8.04860473e-01 9.50164914e-01 6.54857576e-01 -1.07426584e+00
-9.36492920e-01 -1.00574028e+00 1.81482720e+00 -5.50276160e-01
8.12061906e-01 -7.04351887e-02 -8.94602060e-01 6.58316255e-01
5.60465276e-01 8.05415958e-02 1.08876896e+00 4.23278511e-01
-9.93141532e-01 2.00083271e-01 -1.18395364e+00 2.30206981e-01
9.10849094e-01 -2.65194803e-01 -1.39094126e+00 3.25824410e-01
8.38322997e-01 1.62723079e-01 -1.19146740e+00 5.15923083e-01
5.54946423e-01 -2.02659011e-01 8.72023463e-01 -1.41880989e+00
2.19406232e-01 -2.79688716e-01 -1.01407848e-01 -1.30783594e+00
1.51208773e-01 -4.82190937e-01 -2.96341538e-01 1.41527224e+00
9.23471153e-01 -5.40117621e-01 6.01294637e-01 3.65691066e-01
2.44817480e-01 -4.58777755e-01 -9.31208074e-01 -8.31975698e-01
2.71884520e-02 -3.33256543e-01 8.10598016e-01 1.52498543e+00
9.41911638e-01 7.70289838e-01 -1.09857894e-01 5.00365458e-02
5.12980640e-01 -7.55011290e-02 3.24542135e-01 -1.68860674e+00
-3.14801969e-02 -1.89235419e-01 -7.12111235e-01 6.93799928e-02
6.24980152e-01 -1.58672011e+00 -4.01544034e-01 -1.68788946e+00
1.09701775e-01 -5.13292551e-01 -3.07016492e-01 7.44875729e-01
3.15741897e-02 -1.57975644e-01 3.79264588e-03 -1.99401453e-02
-4.65397686e-01 1.11657053e-01 4.23276454e-01 5.44883683e-02
-2.46551201e-01 -2.81533480e-01 -1.83453336e-01 6.67870700e-01
9.33875978e-01 -1.06688690e+00 -2.76124358e-01 -2.47573748e-01
7.18621016e-01 -4.75178629e-01 7.34235048e-02 -6.99815333e-01
3.11544299e-01 -5.98748848e-02 -7.04371929e-02 -2.22735077e-01
-4.33242843e-02 -1.12149572e+00 9.32210777e-03 5.07511675e-01
-4.17734444e-01 9.15364176e-02 -1.45568894e-02 3.06453347e-01
-2.26738766e-01 -1.00727987e+00 3.28017175e-01 -1.87544972e-01
-1.06579435e+00 2.27121949e-01 -1.10766955e-01 1.73519447e-01
1.09082937e+00 -4.89660800e-02 -3.23221028e-01 4.20150816e-01
-8.03657949e-01 2.85739988e-01 1.18507102e-01 7.20332742e-01
4.09511358e-01 -1.35101175e+00 -5.98355711e-01 -2.52503213e-02
7.23211527e-01 -5.05333662e-01 -2.99829006e-01 1.61538944e-01
-8.09705377e-01 3.54858935e-01 -4.82215643e-01 4.12740827e-01
-1.52735317e+00 6.22853220e-01 2.62500513e-02 -4.29995835e-01
-5.72428942e-01 5.69761217e-01 -7.72785902e-01 -9.31165814e-01
8.78660101e-03 -5.23606129e-02 -1.01574695e+00 9.63200107e-02
3.77375364e-01 4.16944355e-01 5.79563081e-02 -7.49208689e-01
-4.37003851e-01 5.83862305e-01 3.86701040e-02 -2.39408284e-01
1.56557798e+00 2.50853956e-01 -7.51232982e-01 2.25483999e-01
1.47176981e+00 1.32281333e-02 2.60633111e-01 -3.85350436e-01
7.75757551e-01 -3.03805113e-01 -5.82923234e-01 -8.56624663e-01
-4.09412026e-01 5.69676280e-01 5.18088222e-01 9.70144570e-02
5.62180400e-01 2.06604987e-01 1.04152966e+00 8.25836957e-01
1.11704230e-01 -1.21921015e+00 -6.41597927e-01 9.85929251e-01
3.65076154e-01 -1.16116393e+00 3.41121620e-03 -5.73071539e-01
-5.00263989e-01 1.54413354e+00 7.50088096e-01 -4.01674211e-01
1.10885370e+00 2.17058957e-01 -2.53250021e-02 -7.85394013e-01
-8.24815929e-01 -9.02723968e-01 3.95720869e-01 9.41835046e-01
5.55114508e-01 9.08938646e-02 -1.32622588e+00 7.86472261e-01
-4.61749554e-01 -2.63258755e-01 3.52115154e-01 5.53602636e-01
-7.69978702e-01 -1.94337749e+00 2.10817352e-01 6.84040666e-01
-3.42786133e-01 -6.64933801e-01 -4.03136760e-01 7.75380671e-01
7.54132986e-01 6.19485795e-01 -1.61840305e-01 -6.20739996e-01
3.33755195e-01 5.33370852e-01 4.69433308e-01 -1.09093392e+00
-7.70730317e-01 -6.41013145e-01 6.25188112e-01 -2.58030415e-01
-2.61434793e-01 -2.72935033e-01 -1.69217432e+00 1.64154232e-01
-5.34060121e-01 5.44667780e-01 7.80853987e-01 1.12477338e+00
-1.86344713e-01 1.30001217e-01 2.55232006e-01 3.29794019e-01
-3.25785249e-01 -1.01182938e+00 -5.07192016e-01 8.29693496e-01
-4.00848597e-01 -4.19656575e-01 2.24661186e-01 -2.57077795e-02] | [9.816080093383789, 8.711825370788574] |
b5632080-a274-4476-85d1-faf03942995e | reducing-audio-membership-inference-attack | 1911.01888 | null | https://arxiv.org/abs/1911.01888v1 | https://arxiv.org/pdf/1911.01888v1.pdf | Reducing audio membership inference attack accuracy to chance: 4 defenses | It is critical to understand the privacy and robustness vulnerabilities of machine learning models, as their implementation expands in scope. In membership inference attacks, adversaries can determine whether a particular set of data was used in training, putting the privacy of the data at risk. Existing work has mostly focused on image related tasks; we generalize this type of attack to speaker identification on audio samples. We demonstrate attack precision of 85.9\% and recall of 90.8\% for LibriSpeech, and 78.3\% precision and 90.7\% recall for VOiCES (Voices Obscured in Complex Environmental Settings). We find that implementing defenses such as prediction obfuscation, defensive distillation or adversarial training, can reduce attack accuracy to chance. | ['Zigfried Hampel-Arias', 'Nina Lopatina', 'Felipe A. Mejia', 'Paul Gamble', 'Maria Alejandra Barrios', 'Michael Lomnitz', 'Lucas Tindall'] | 2019-10-31 | null | null | null | null | ['membership-inference-attack'] | ['computer-vision'] | [ 6.49898112e-01 2.68402398e-01 4.94255796e-02 -1.41690969e-01
-8.80335510e-01 -1.44208157e+00 5.20205677e-01 -4.57046255e-02
-3.45515132e-01 4.77652192e-01 -1.42792553e-01 -8.12483370e-01
5.88405058e-02 -7.06244826e-01 -8.46209228e-01 -6.26494169e-01
-3.46066922e-01 -1.20643955e-02 -1.50814414e-01 2.43401974e-01
1.84157565e-01 7.37878978e-01 -9.37429845e-01 6.04159832e-01
1.64995641e-01 9.62126613e-01 -7.81799436e-01 1.01868677e+00
5.75070381e-01 6.42599642e-01 -1.28729868e+00 -9.88355577e-01
6.87951446e-01 1.21151596e-01 -5.62805176e-01 -4.59704459e-01
8.03208590e-01 -7.13583410e-01 -6.12343729e-01 1.25787568e+00
4.42549497e-01 -1.91810533e-01 6.47182882e-01 -1.59828341e+00
-4.56797421e-01 6.59812689e-01 -5.11321425e-01 3.49865288e-01
2.22503662e-01 4.88337308e-01 8.19714069e-01 -2.35560492e-01
2.46397406e-01 1.21195889e+00 7.54051030e-01 7.30604112e-01
-1.44359231e+00 -1.42806101e+00 -1.63480714e-01 -1.14049621e-01
-1.50196350e+00 -8.32370639e-01 4.08432752e-01 -5.12965202e-01
6.48172498e-01 7.93668985e-01 -1.63019006e-03 1.44119287e+00
1.92569390e-01 4.34163600e-01 1.14209223e+00 -1.70582309e-01
9.74903256e-02 5.67128301e-01 -4.29801904e-02 3.90617907e-01
5.38554847e-01 5.81178963e-01 -3.54871988e-01 -8.37644517e-01
3.67307335e-01 -2.30502769e-01 -3.67815971e-01 1.21093787e-01
-6.87939584e-01 9.39911842e-01 2.85699785e-01 -2.22449720e-01
8.05551931e-02 2.39603877e-01 5.24940491e-01 4.38374341e-01
1.14738442e-01 8.92306328e-01 -4.48945552e-01 2.40117610e-01
-8.48631144e-01 2.21188501e-01 9.56353724e-01 7.70722508e-01
3.40672135e-01 2.92970747e-01 -1.04012132e-01 3.41796488e-01
1.51682481e-01 6.91720605e-01 1.35045052e-01 -1.01602387e+00
4.90343392e-01 -9.64948013e-02 5.53223155e-02 -1.30389798e+00
2.97016297e-02 -3.12805474e-01 -5.08885920e-01 3.86753291e-01
5.87070882e-01 -6.57233596e-01 -7.44188249e-01 1.82773864e+00
2.18039170e-01 1.12328850e-01 3.45868051e-01 5.49871325e-01
3.04500252e-01 5.57888985e-01 2.26989850e-01 1.18911475e-01
1.35126495e+00 -1.76440105e-01 -4.07955766e-01 -4.02444899e-01
3.35217267e-01 -7.72064447e-01 7.71002650e-01 5.96683741e-01
-6.81098521e-01 -1.61162063e-01 -1.17345238e+00 3.85343254e-01
-4.98553783e-01 -2.91302562e-01 6.18488491e-01 1.60273111e+00
-5.47860980e-01 4.42284107e-01 -5.53306043e-01 4.77885157e-02
8.25106978e-01 5.39818406e-01 -4.69593197e-01 8.06615278e-02
-1.25763452e+00 6.40078306e-01 3.35809961e-02 -2.37994343e-01
-1.24551296e+00 -8.70000958e-01 -5.59044302e-01 9.27032307e-02
2.56798357e-01 -1.64383143e-01 1.16903448e+00 -7.78006852e-01
-8.99556875e-01 8.55071604e-01 3.11402738e-01 -9.80984151e-01
5.31090438e-01 -2.37973437e-01 -8.21776271e-01 3.02459389e-01
-2.28790894e-01 3.97367954e-01 1.31863868e+00 -1.12943697e+00
-3.52437645e-01 -4.87411290e-01 1.54638484e-01 -2.32021645e-01
-4.92483437e-01 4.11067903e-01 2.97413737e-01 -7.86459386e-01
-2.79326469e-01 -1.07635581e+00 -6.71245232e-02 5.60150743e-02
-8.69584680e-01 6.87063754e-01 1.29068267e+00 -8.72950435e-01
1.03912401e+00 -2.29793406e+00 -5.43131351e-01 4.85435575e-01
1.21089570e-01 7.62905419e-01 1.36596456e-01 3.61661673e-01
-1.71337575e-01 7.55088687e-01 -7.91550651e-02 8.82294923e-02
1.13115847e-01 -1.35035321e-01 -1.04421496e+00 6.33933425e-01
7.15437382e-02 3.87990713e-01 -2.96543419e-01 8.20434168e-02
-6.16510101e-02 3.52198422e-01 -7.85579383e-01 1.26282275e-01
-6.98005222e-03 1.82666048e-01 -2.26914823e-01 8.30949664e-01
7.49148548e-01 2.19392627e-01 1.09285720e-01 4.74005677e-02
3.24189931e-01 2.83537120e-01 -9.58939791e-01 6.61101222e-01
-1.44529060e-01 9.69022214e-01 4.75757599e-01 -5.94933450e-01
6.87892258e-01 4.87522185e-01 -5.47670154e-03 -1.93825755e-02
2.05863833e-01 -2.51889259e-01 4.16372836e-01 -3.51409286e-01
2.39198923e-01 -4.70422097e-02 -3.86722624e-01 6.11414790e-01
-1.76373512e-01 -1.35011703e-01 -7.25650847e-01 3.10548604e-01
1.28070688e+00 -6.97832942e-01 5.49954586e-02 -6.84663653e-02
2.04748362e-01 -1.24797590e-01 4.30341780e-01 1.03817630e+00
-4.52535689e-01 2.98226833e-01 5.27780473e-01 -2.11399153e-01
-1.02537572e+00 -1.24648082e+00 -2.84638196e-01 9.86543000e-01
-3.34072322e-01 -5.20124853e-01 -1.02069998e+00 -1.01629889e+00
4.30174500e-01 8.05959046e-01 -3.68447363e-01 -4.93003547e-01
-2.57565379e-01 -6.01201415e-01 1.56633103e+00 2.82600909e-01
4.76710290e-01 -5.82040668e-01 -4.11048681e-01 -2.53671139e-01
1.56738043e-01 -1.26386261e+00 -3.60871077e-01 3.08399145e-02
-3.16273898e-01 -1.23614776e+00 4.51764613e-02 -1.79600060e-01
4.74558800e-01 8.61662328e-02 7.87402511e-01 1.16334207e-01
-4.12940502e-01 6.24970675e-01 -3.51356193e-02 -9.86443639e-01
-8.18913698e-01 -8.78229514e-02 3.64740908e-01 2.51025725e-02
3.88856947e-01 -6.51173890e-01 -3.85601252e-01 4.46304530e-01
-8.57092917e-01 -7.62232006e-01 3.65190625e-01 5.89644432e-01
1.27375573e-01 2.89223552e-01 5.32452345e-01 -1.08069360e+00
5.00452816e-01 -4.99216706e-01 -5.79235315e-01 1.17090531e-01
-3.92856359e-01 -4.07162011e-01 6.62629068e-01 -9.49277222e-01
-5.55159569e-01 1.24168478e-01 -1.58949390e-01 -6.46289289e-01
-3.91171247e-01 -9.19634476e-02 -3.65153521e-01 -5.28829694e-01
9.80396569e-01 7.03116804e-02 1.06678180e-01 -1.45775080e-01
1.45861521e-01 7.96815634e-01 4.57917690e-01 -5.32640994e-01
1.19995701e+00 5.50010979e-01 -6.55099154e-02 -8.50156724e-01
-5.27078688e-01 1.06988944e-01 -1.70486003e-01 -3.38449888e-02
6.62604809e-01 -8.91030252e-01 -1.02345550e+00 4.02408391e-01
-8.12398255e-01 -1.51082531e-01 2.66187340e-02 3.37022960e-01
-2.16711357e-01 4.45649624e-01 -4.27036881e-01 -1.07061529e+00
-4.30912763e-01 -1.05218172e+00 5.76909482e-01 -1.30864590e-01
-5.07555842e-01 -4.74907845e-01 -2.67060488e-01 7.17281222e-01
2.71268040e-01 5.01708031e-01 6.04268909e-01 -1.29533660e+00
-5.49658537e-01 -6.82524264e-01 -3.04821823e-02 5.57143748e-01
-3.15451473e-02 1.41254842e-01 -1.49280310e+00 -5.13862789e-01
8.62885192e-02 -4.56247270e-01 5.78324258e-01 -1.99111216e-02
1.28517914e+00 -1.23596013e+00 -3.15395564e-01 6.95804238e-01
9.79601324e-01 4.13829714e-01 6.35040581e-01 1.81626137e-02
4.28671807e-01 6.29502356e-01 2.64645904e-01 3.46027702e-01
-2.86180437e-01 5.05845726e-01 5.72778165e-01 2.55620211e-01
3.94094318e-01 -4.64926600e-01 6.82227969e-01 -3.37704867e-01
4.69100624e-01 -3.13144445e-01 -9.27258015e-01 1.71297193e-01
-1.16097450e+00 -1.39457941e+00 1.61461741e-01 2.31174731e+00
7.73207486e-01 3.14528644e-01 1.73808619e-01 3.18868309e-01
8.33633304e-01 1.74584210e-01 -5.70027113e-01 -6.69903755e-01
5.64027876e-02 2.36124188e-01 9.63695645e-01 4.24675375e-01
-1.44819403e+00 9.20279503e-01 7.32661724e+00 6.74932063e-01
-1.06509042e+00 -1.08668134e-01 8.08938742e-01 -4.34786916e-01
4.20486368e-02 -7.54958913e-02 -7.66023457e-01 4.93465811e-01
1.16263402e+00 -1.43020645e-01 5.77521026e-01 9.40288723e-01
-2.67240494e-01 2.61651933e-01 -1.02298164e+00 7.23084152e-01
8.58769789e-02 -1.16862917e+00 -1.34389758e-01 4.72550601e-01
2.67392367e-01 -2.71234483e-01 7.45306969e-01 2.10891187e-01
7.11740136e-01 -1.39039695e+00 5.95363319e-01 6.21437244e-02
7.48985767e-01 -1.06497705e+00 2.57787555e-01 4.35347736e-01
-6.32068217e-01 -3.16031575e-01 -3.39985907e-01 -2.85436586e-02
-2.39400059e-01 2.28147805e-01 -1.32108819e+00 -2.82163247e-02
7.17292190e-01 -1.11674882e-01 -5.86567998e-01 4.36384261e-01
-2.86477655e-01 1.08874846e+00 -5.09729385e-01 1.85834721e-01
1.06701903e-01 3.12447071e-01 8.90183985e-01 1.27934718e+00
2.71650683e-02 1.64970867e-02 6.52674511e-02 7.56841183e-01
-1.88429803e-01 -3.73350620e-01 -1.04281211e+00 -3.44554275e-01
9.77593839e-01 1.06682479e+00 -1.50208429e-01 9.37637240e-02
1.03212848e-01 9.06180918e-01 -1.20630093e-01 1.26145601e-01
-8.97220910e-01 -4.17443752e-01 1.27600408e+00 1.08461976e-01
2.20634565e-01 -1.23559766e-01 -2.57478893e-01 -9.39579785e-01
-1.41371116e-01 -1.35504484e+00 6.94527030e-01 -4.31643456e-01
-1.26052558e+00 2.87556589e-01 -2.45350331e-01 -1.01104879e+00
-3.33081037e-01 -5.15810430e-01 -7.65380025e-01 7.66382158e-01
-7.70447910e-01 -1.14238334e+00 3.53786677e-01 6.98640287e-01
-1.55200288e-01 -5.71063519e-01 1.03587210e+00 5.43759540e-02
-7.58979142e-01 1.30856252e+00 -1.24780051e-02 7.47229338e-01
7.24884152e-01 -8.80487978e-01 6.13118351e-01 1.06221581e+00
3.48707408e-01 1.01077950e+00 8.09366643e-01 -4.47596222e-01
-1.47238529e+00 -1.17562890e+00 3.46112669e-01 -7.92961001e-01
7.01062918e-01 -6.70725048e-01 -9.23548698e-01 1.01822746e+00
4.23849225e-02 -2.47462496e-01 1.24811149e+00 8.39503333e-02
-1.16221976e+00 -3.48388217e-02 -1.84049380e+00 7.08411813e-01
5.98241687e-01 -1.05788946e+00 -1.77107438e-01 3.67796838e-01
8.63884568e-01 -2.61791766e-01 -9.28560317e-01 1.78567439e-01
8.67702007e-01 -7.46768832e-01 1.34272063e+00 -1.07724023e+00
-9.89567935e-02 -7.14993253e-02 -5.24882615e-01 -8.04687023e-01
-9.11562517e-02 -9.55693364e-01 -2.62728065e-01 1.58791840e+00
5.38153231e-01 -9.17066693e-01 8.68287802e-01 9.60697651e-01
4.06239390e-01 -1.81526691e-02 -9.30843830e-01 -7.67155290e-01
2.90744573e-01 -6.43881917e-01 6.14413023e-01 1.25336373e+00
-9.30007175e-02 -1.03188537e-01 -5.73802054e-01 1.05971134e+00
7.75500178e-01 -3.62265885e-01 1.03944457e+00 -8.42531919e-01
-4.22500134e-01 -6.68530315e-02 -6.42022908e-01 -3.37935328e-01
2.15579942e-01 -7.41834700e-01 -5.34027219e-01 -3.46554697e-01
-1.54878423e-01 -3.76936704e-01 -1.42911717e-01 6.13164544e-01
7.32704327e-02 5.03137350e-01 4.20189589e-01 1.24293230e-01
4.95593473e-02 -1.38085410e-01 4.82088238e-01 -4.24089640e-01
2.20874399e-01 4.33301628e-01 -1.15612090e+00 7.36434281e-01
1.27925122e+00 -7.98622549e-01 -3.55410933e-01 -1.11198448e-01
-1.16251588e-01 -1.58234909e-01 8.25095177e-01 -1.12381947e+00
7.10934550e-02 -1.93915308e-01 6.36823416e-01 -9.55745485e-03
5.62331259e-01 -1.01481962e+00 2.40348160e-01 6.60361648e-01
-5.31892598e-01 -1.41942486e-01 5.03771782e-01 6.15137041e-01
3.17994237e-01 -2.81071186e-01 9.11876976e-01 -1.83369108e-02
-1.22969262e-01 2.66929299e-01 -5.22639215e-01 5.19589148e-02
1.15115941e+00 -1.40320733e-01 -6.57006264e-01 -5.22095144e-01
-7.98455656e-01 -9.47028771e-02 5.05538285e-01 1.91980049e-01
5.04904032e-01 -9.48921621e-01 -5.36108553e-01 4.23998177e-01
-1.18267961e-01 -6.74952209e-01 2.01849043e-01 2.15095639e-01
-1.80962250e-01 9.04306546e-02 -1.69461563e-01 -2.64170170e-01
-1.92118251e+00 8.06183040e-01 4.32450712e-01 2.19773605e-01
-2.90947706e-01 9.12218451e-01 6.74959049e-02 -2.12020993e-01
4.76094574e-01 1.27145767e-01 3.58066857e-01 -1.68538824e-01
9.14794147e-01 3.89681607e-01 -1.25687838e-01 -5.13322413e-01
-4.08340186e-01 -8.54959264e-02 -3.65382224e-01 -3.17485273e-01
7.69315362e-01 3.81786019e-01 2.27038696e-01 -1.73110247e-01
1.24343789e+00 5.76002777e-01 -1.07630754e+00 -1.95818138e-03
-3.16370636e-01 -8.40056181e-01 -8.43856260e-02 -1.00616992e+00
-8.70883465e-01 9.11767602e-01 5.34474671e-01 5.61002135e-01
9.11222339e-01 -1.88691825e-01 4.07660663e-01 5.89277685e-01
2.04581067e-01 -6.50741100e-01 -2.69984543e-01 1.75294235e-01
7.66793549e-01 -8.81854355e-01 2.25432977e-01 -3.94925684e-01
-7.00343609e-01 6.99748516e-01 5.29406846e-01 -4.53441255e-02
6.74594700e-01 6.76530898e-01 -2.90715508e-03 3.04428376e-02
-7.94134796e-01 5.16272128e-01 7.58428425e-02 1.04373968e+00
-1.42916754e-01 2.06420526e-01 3.60281527e-01 8.44371915e-01
-6.81501269e-01 -7.01157570e-01 5.25267363e-01 8.79760921e-01
-2.76691675e-01 -9.26263094e-01 -8.13108563e-01 4.55074698e-01
-1.12043464e+00 -1.97125003e-01 -9.27646756e-01 6.82626605e-01
1.36333570e-01 1.16268826e+00 9.15028974e-02 -8.44254792e-01
1.56820957e-02 1.75268337e-01 1.99467734e-01 -5.06325841e-01
-1.03720069e+00 -3.30919117e-01 4.07721043e-01 -3.50885510e-01
2.14143068e-01 -9.28208590e-01 -8.05218637e-01 -8.03031266e-01
-2.95045108e-01 7.91302323e-02 6.79898560e-01 4.26411271e-01
6.39090240e-01 9.06608254e-02 8.33614349e-01 -2.91009843e-01
-1.04032743e+00 -6.86896801e-01 -5.21252632e-01 1.00701787e-01
5.34643888e-01 -2.32068405e-01 -6.57190740e-01 -3.03954463e-02] | [5.804698944091797, 7.5753254890441895] |
c7e589c0-6409-4db1-89e8-4b6d148492c7 | interference-cancellation-based-channel | 2006.14508 | null | https://arxiv.org/abs/2006.14508v2 | https://arxiv.org/pdf/2006.14508v2.pdf | Interference Cancellation Based Channel Estimation for Massive MIMO Systems with Time Shifted Pilots | In massive multiple-input multiple-output (MIMO) systems with time shifted pilot (TSP) schemes, the inter-group interference caused by the pilot contamination can be eliminated when the number of base station (BS) antennas M approaches infinity. However, M is finite in practice and the effectiveness of the TSP is limited by channel estimation errors. In this paper, it is analytically shown that the mean square channel estimation error (MSCEE) of the TSP is dominated by the intergroup data interference. To reduce the MSCEE in the finite antenna massive MIMO systems, an interference cancellation based channel estimation for the TSP (IC-TSP) is proposed, where the dominant inter-group data interference is canceled based on BS cooperation. To show the advantage of the IC-TSP, the additional overhead of IC-TSP is evaluated by considering different M and the coherence time of BS-BS channels. Furthermore, the impact of sectorization and compressed sensing based BS-BS channel estimation are also discussed. We show that when 128 < M < 2048, with the inter-group data interference from the nearest two cell layers being canceled, the IC-TSP achieves a spectral efficiency gain of more than 1.2 bps/Hz over the TSP. | ['Jinglin Shi', 'Jinhong Yuan', 'Yiqing Zhou', 'Bule Sun'] | 2020-06-25 | null | null | null | null | ['2048'] | ['playing-games'] | [ 4.72655833e-01 5.57641387e-01 1.37186065e-01 7.10074306e-01
-4.87677604e-01 -2.28314519e-01 -8.39077011e-02 8.28066245e-02
-3.05977315e-01 1.24969578e+00 -1.77382380e-01 -8.50397885e-01
-3.03757399e-01 -4.93698448e-01 -4.82770085e-01 -1.39372492e+00
-6.58928216e-01 -3.81842822e-01 1.86133817e-01 -3.99496645e-01
1.00863785e-01 2.25281313e-01 -8.18450272e-01 1.55816460e-02
1.05144191e+00 1.49128664e+00 3.77577186e-01 6.89269900e-01
7.78597072e-02 6.86289549e-01 -1.17605484e+00 5.99407069e-02
5.71709454e-01 -5.48650444e-01 6.52041137e-02 2.47143224e-01
-4.96390432e-01 -2.48354658e-01 -6.87769413e-01 6.81220233e-01
8.73843968e-01 -4.29547817e-01 3.83730650e-01 -1.41883075e+00
1.10822119e-01 5.32946348e-01 -8.35781157e-01 -5.70986904e-02
2.93811671e-02 -4.47951436e-01 3.75184536e-01 -9.23726022e-01
1.10058494e-01 1.03356647e+00 6.74737334e-01 -3.11172068e-01
-7.98449814e-01 -1.04585028e+00 -4.71967161e-01 1.08354956e-01
-2.20525455e+00 -4.84303474e-01 6.73137233e-02 -2.05325738e-01
4.04626101e-01 7.56333768e-01 7.13096678e-01 -1.75460950e-01
3.96915168e-01 2.68533677e-01 7.29414523e-01 -1.09803569e+00
3.37930292e-01 2.46645629e-01 -3.46838057e-01 4.70719725e-01
9.26515758e-01 -1.02536716e-01 -2.81182647e-01 -4.61019784e-01
9.88570750e-01 -3.32809120e-01 -5.86609006e-01 -2.24104822e-01
-1.24965775e+00 5.39928079e-01 2.16132388e-01 6.49606943e-01
-4.80625927e-01 1.78998038e-01 -1.57771245e-01 5.84357440e-01
7.79794157e-02 2.04208896e-01 -2.67279670e-02 4.95229274e-01
-1.07427323e+00 -1.89103082e-01 8.20496321e-01 1.68329370e+00
5.18107057e-01 1.71407074e-01 -5.86267292e-01 5.80431640e-01
1.11520231e-01 1.08074903e+00 -1.99978605e-01 -1.03034163e+00
9.01794136e-01 -6.69081733e-02 4.04418200e-01 -8.76283169e-01
-5.90359032e-01 -1.59668744e+00 -1.45418310e+00 -3.94903809e-01
5.54284930e-01 -8.85095954e-01 -4.56034839e-02 1.36038375e+00
-2.68304795e-02 3.08961235e-02 4.53776807e-01 6.36992693e-01
-2.81938408e-02 4.70445603e-01 -6.80228949e-01 -1.01093411e+00
9.40217733e-01 -6.15230322e-01 -1.07297862e+00 -2.55680770e-01
9.50887322e-01 -7.66258001e-01 -1.63480222e-01 5.17091274e-01
-1.13340354e+00 -2.06745848e-01 -1.36597419e+00 8.03897202e-01
4.32957470e-01 6.35852575e-01 1.79176360e-01 1.09642589e+00
-9.00756955e-01 1.84411334e-03 -2.99366206e-01 -3.20441842e-01
3.88622403e-01 6.31244481e-01 1.16536934e-02 -9.26937833e-02
-1.20453775e+00 3.27678829e-01 3.34889114e-01 1.06121838e-01
2.29879841e-01 -5.27507424e-01 -3.73226434e-01 2.26512909e-01
3.15170348e-01 -5.22498548e-01 8.77658784e-01 -7.60851562e-01
-1.02218425e+00 -1.57862738e-01 -2.75199085e-01 -5.58433235e-01
2.45612442e-01 2.53340214e-01 -5.62858403e-01 4.67281014e-01
1.37240961e-01 -2.63594538e-01 2.61770368e-01 -1.19159913e+00
-9.44410145e-01 -3.93042445e-01 4.74663600e-02 1.76594883e-01
-4.34158146e-01 -2.80324906e-01 -5.56322634e-01 -7.64695704e-01
5.13685703e-01 -1.34988856e+00 -6.67824090e-01 -3.64636540e-01
-6.33355498e-01 8.71141851e-01 4.57096428e-01 -6.43526733e-01
2.02944851e+00 -2.58868623e+00 -1.69847682e-01 6.05363309e-01
1.12955868e-01 2.82155395e-01 3.54663044e-01 9.31691468e-01
3.99462074e-01 -3.84318866e-02 3.48374993e-02 7.99964592e-02
-5.87700546e-01 -9.03402716e-02 1.07276104e-01 6.07687235e-01
-4.75534678e-01 2.40099490e-01 -5.66451550e-01 -2.52975553e-01
-1.12532610e-02 7.29448199e-02 -5.17873764e-01 -3.25957090e-01
7.63668239e-01 7.55190790e-01 -5.74545145e-01 4.16481376e-01
1.18508613e+00 -5.05128264e-01 4.96408165e-01 -2.23011047e-01
-6.11545384e-01 -2.74853140e-01 -1.35898781e+00 1.12484932e+00
-4.40074265e-01 5.70990741e-01 6.27449811e-01 -9.92484510e-01
4.14175779e-01 8.00978899e-01 4.85272348e-01 -1.01996052e+00
2.03130260e-01 6.88398719e-01 4.16690737e-01 2.88038757e-02
2.60565192e-01 -3.16753909e-02 -1.31623104e-01 1.60004362e-01
-1.89560473e-01 5.49965262e-01 3.38238478e-01 4.54866201e-01
1.21644568e+00 -8.76246691e-01 9.32591498e-01 -5.68135440e-01
7.35503197e-01 -4.02833402e-01 7.42618680e-01 8.66499662e-01
9.50914547e-02 1.44676507e-01 2.78922975e-01 5.40416241e-01
-8.44828308e-01 -6.03926718e-01 -2.72495061e-01 3.87659460e-01
7.50574946e-01 -3.23119164e-01 -5.52879751e-01 2.48153165e-01
1.80497006e-01 8.08039248e-01 -1.61312953e-01 -7.16245398e-02
-1.50406823e-01 -1.04545927e+00 5.39267063e-01 -5.74904978e-02
8.11020017e-01 2.91668713e-01 -3.69855642e-01 6.48200333e-01
-3.98309290e-01 -1.41929114e+00 -1.19703569e-01 2.44764239e-01
-5.23695529e-01 -8.09961081e-01 -1.06702638e+00 -2.36745417e-01
6.29043937e-01 1.08224642e+00 2.10232437e-01 5.22392653e-02
2.59828538e-01 2.27854520e-01 -4.56705749e-01 -5.82982779e-01
-2.26666600e-01 -5.16918786e-02 2.96491310e-02 2.24917501e-01
-3.00484776e-01 -3.85865778e-01 -6.58375025e-01 6.51042819e-01
-2.49046609e-01 3.45941693e-01 8.12886357e-01 5.66695809e-01
1.51520064e-02 3.53561431e-01 7.90203989e-01 -1.83547452e-01
8.31093863e-02 -5.26249528e-01 -4.39467847e-01 1.84615731e-01
-1.21686876e-01 -2.00304449e-01 5.90121090e-01 -1.28829824e-02
-8.84270489e-01 -1.72326609e-01 -1.29071577e-02 2.61411726e-01
6.40230775e-01 4.75267768e-01 -2.55323678e-01 -6.97589815e-01
6.02855206e-01 1.40851587e-01 -2.41215214e-01 8.20645317e-02
-1.27989799e-01 1.14959836e+00 -2.09593460e-01 2.01812014e-01
8.92666578e-01 6.59299076e-01 6.77484691e-01 -1.59711266e+00
-5.37812471e-01 -4.83640939e-01 -4.06293690e-01 -5.55490740e-02
5.29527403e-02 -1.44678366e+00 -6.02075934e-01 1.60985082e-01
-1.12349474e+00 -2.48407200e-01 2.43456066e-01 1.08963072e+00
-2.98995614e-01 6.06859267e-01 -3.98961604e-01 -1.36963367e+00
-2.54930288e-01 -8.37178349e-01 6.38021469e-01 -1.21744625e-01
4.48135883e-02 -4.98454601e-01 -8.06099057e-01 1.28404479e-02
7.55522490e-01 3.21426243e-01 6.16071045e-01 -1.35775998e-01
-7.18874991e-01 -1.26312330e-01 -1.86903805e-01 -6.50889352e-02
1.12349458e-01 -7.90969372e-01 -5.16008496e-01 -7.46888697e-01
-1.23430692e-01 6.33683145e-01 1.78601101e-01 7.69113839e-01
7.73178101e-01 -3.88174266e-01 -9.76816535e-01 2.90988207e-01
1.44008493e+00 4.91489470e-01 9.36878324e-01 2.68702190e-02
2.83447087e-01 1.54713746e-02 1.13559175e+00 1.06027448e+00
1.31499276e-01 9.78121161e-01 2.22178057e-01 -1.88802913e-01
8.07993039e-02 5.42033076e-01 7.67057911e-02 9.12285089e-01
-7.49792485e-03 -7.59347439e-01 -4.04484391e-01 1.84057519e-01
-1.69678819e+00 -6.85695291e-01 -9.75962758e-01 2.86092615e+00
3.84504467e-01 8.60071704e-02 -1.06994614e-01 7.50390291e-01
8.58688712e-01 -6.07163310e-01 -1.29042655e-01 1.04016274e-01
-5.91343462e-01 -1.70167416e-01 1.61973190e+00 4.20635760e-01
-7.42561340e-01 1.23701461e-01 4.65970659e+00 1.36636949e+00
-9.12119567e-01 2.76414931e-01 3.69643569e-01 1.68334395e-02
7.89387375e-02 -2.78378781e-02 -8.34224880e-01 6.09783113e-01
9.37585711e-01 -4.33675289e-01 1.22526996e-01 1.20184518e-01
6.84743762e-01 -7.89414525e-01 -4.55957681e-01 1.14877343e+00
-2.97242761e-01 -1.21438265e+00 -2.09535748e-01 8.23233008e-01
6.42953753e-01 -3.46815765e-01 -3.19086730e-01 1.65565640e-01
-5.26797414e-01 -3.47580492e-01 6.65863216e-01 3.39327246e-01
1.07153177e+00 -7.88858533e-01 8.09103489e-01 6.40686512e-01
-1.47146535e+00 -1.00435317e+00 -1.03228666e-01 -3.82397026e-01
3.22566986e-01 1.40405905e+00 -7.38800526e-01 1.21519876e+00
-7.55666848e-03 -1.24856003e-01 -1.53191298e-01 1.48447466e+00
1.68949008e-01 6.19648933e-01 -5.46748281e-01 1.54322833e-01
9.68004018e-02 -3.06809068e-01 7.03834951e-01 9.31005299e-01
1.45393443e+00 4.52809840e-01 3.35655324e-02 3.27925086e-02
2.40526319e-01 2.90723473e-01 -2.81872511e-01 3.84494901e-01
1.09461355e+00 9.14665043e-01 -4.75075454e-01 -3.52573037e-01
-6.98864996e-01 8.79055917e-01 -5.21649361e-01 6.80719733e-01
-6.02556527e-01 -9.39297497e-01 3.20002943e-01 4.81278449e-01
3.71062487e-01 -6.06214762e-01 -5.97137094e-01 -6.49097800e-01
-8.91623124e-02 -6.21327639e-01 -1.71248898e-01 -2.80807555e-01
-2.14620203e-01 -4.21381742e-02 -4.66343939e-01 -1.70805776e+00
-1.03574619e-01 -1.28753856e-01 -1.17688566e-01 8.51017237e-01
-1.16754079e+00 -4.99365062e-01 -1.57978237e-01 3.48303139e-01
-2.24221498e-01 -2.61003405e-01 6.65904760e-01 1.02419686e+00
-3.37460995e-01 8.49094093e-01 9.29320812e-01 -1.76243350e-01
3.09329689e-01 -5.51694036e-01 -3.49024236e-01 7.59679317e-01
-7.29691625e-01 4.37527269e-01 7.24287987e-01 -5.69730282e-01
-1.46737862e+00 -1.14830303e+00 8.71008515e-01 6.54327929e-01
2.86749274e-01 -5.43024540e-01 -2.40220547e-01 1.87194392e-01
1.79865524e-01 -1.46877497e-01 1.26736236e+00 -5.60523748e-01
6.49098873e-01 -2.97005344e-02 -9.81822371e-01 8.15192401e-01
9.35108721e-01 7.43594393e-02 4.65002030e-01 3.33159059e-01
3.48807245e-01 -2.07760558e-01 -8.92422915e-01 4.43952262e-01
5.27894676e-01 -8.62572372e-01 7.64232576e-01 5.09742022e-01
-5.33182681e-01 -5.48904896e-01 -6.12235963e-01 -1.22095704e+00
-6.09378159e-01 -9.67369437e-01 -1.42474249e-01 7.21849501e-01
3.37619841e-01 -4.81044918e-01 5.23906231e-01 -1.13979690e-01
-9.42430273e-02 -3.93135279e-01 -1.24114871e+00 -1.06928146e+00
-2.52840489e-01 -8.29350725e-02 2.44580299e-01 5.47906756e-01
3.34982187e-01 5.76833546e-01 -5.09731948e-01 7.27286220e-01
7.38833427e-01 -3.72041881e-01 1.00011659e+00 -1.32270968e+00
-5.31308472e-01 1.59890488e-01 -5.63574582e-02 -1.53811049e+00
-6.14033878e-01 -5.18781781e-01 -4.87034231e-01 -1.49298048e+00
-3.72704089e-01 -8.30922008e-01 -3.24437022e-02 -1.68918952e-01
1.13898650e-01 2.27853969e-01 3.28376085e-01 9.61522534e-02
-5.41804910e-01 4.83989805e-01 1.37609959e+00 4.58700210e-01
-2.41507322e-01 5.13572872e-01 -6.61888421e-01 4.54187572e-01
7.96398222e-01 -1.55919403e-01 -3.06603014e-01 1.94889009e-02
1.46459356e-01 9.61167216e-01 3.83665636e-02 -1.54151630e+00
2.05818266e-01 1.77358404e-01 1.77242219e-01 -6.32335961e-01
5.77630103e-01 -1.12082684e+00 4.05990124e-01 9.54697609e-01
2.37533867e-01 -1.04143560e+00 2.63850361e-01 8.56080294e-01
5.59538370e-04 4.27688360e-02 7.10497200e-01 3.01425576e-01
5.89544587e-02 6.49295794e-03 -1.16257107e+00 -4.09755439e-01
1.34452295e+00 -4.45643246e-01 -5.84662855e-02 -7.69372940e-01
-5.60291588e-01 4.62269485e-01 -1.53972924e-01 -4.19256508e-01
4.71133040e-03 -1.39662182e+00 -6.28161371e-01 -1.08038997e-02
1.11274809e-01 -7.63142526e-01 5.33285022e-01 1.76556921e+00
-8.64941105e-02 1.26564014e+00 7.31095523e-02 -6.56869709e-01
-1.34016645e+00 9.22070444e-02 1.38103336e-01 -1.41779393e-01
8.34256224e-03 4.95129645e-01 4.16834176e-01 5.28870940e-01
3.60109061e-02 2.31171489e-01 4.58530821e-02 -2.65824348e-01
5.44211745e-01 7.93538809e-01 2.94637650e-01 -7.05724001e-01
-3.60199548e-02 4.08831716e-01 4.65019643e-01 -2.03541785e-01
5.63906610e-01 -9.90504265e-01 -1.67624708e-02 -2.59966888e-02
9.30384636e-01 6.53427839e-01 -6.46581590e-01 -4.28350329e-01
-1.85314029e-01 -6.48853898e-01 1.42828882e-01 -3.48797411e-01
-7.72443831e-01 5.80088615e-01 6.71502113e-01 3.47445905e-01
1.11944985e+00 -5.06479025e-01 6.91533148e-01 5.24049878e-01
1.36886406e+00 -1.02303755e+00 -2.80919015e-01 5.26936293e-01
8.30570340e-01 -4.20772523e-01 1.55271497e-02 -1.09908473e+00
-2.28318945e-01 9.34887588e-01 -7.21131116e-02 3.96374762e-01
7.70923436e-01 3.90208870e-01 -4.94651556e-01 2.59575009e-01
-1.25306159e-01 -4.15934592e-01 -2.62259841e-01 4.08583999e-01
5.07186890e-01 1.48331106e-01 -9.35175180e-01 7.19924092e-01
-1.75669685e-01 3.24542336e-02 8.77446651e-01 9.81260240e-01
-9.82197762e-01 -1.15214169e+00 -1.02071619e+00 7.88329482e-01
-6.14310086e-01 -3.28884274e-01 2.61358351e-01 6.67900503e-01
1.83956146e-01 1.65878367e+00 -1.44706205e-01 -2.92756766e-01
3.93840134e-01 -5.54693758e-01 4.74200189e-01 -4.48061407e-01
1.36024594e-01 5.36273360e-01 5.04271030e-01 -2.10876837e-01
4.89621758e-02 -4.01514292e-01 -1.41175103e+00 -3.33036363e-01
-5.95040202e-01 6.43680036e-01 3.65212351e-01 9.63702023e-01
6.76620901e-01 7.52222359e-01 1.00250673e+00 -4.62957233e-01
-1.77328274e-01 -6.51374817e-01 -1.36375785e+00 -4.90560442e-01
3.79939228e-01 -7.43968368e-01 -5.04384220e-01 -4.88564193e-01] | [6.1847381591796875, 1.4240915775299072] |
8ea83600-a7ca-4fea-a2d4-5a3107bea16d | vibration-based-damage-detection-in-wind | 1804.00558 | null | http://arxiv.org/abs/1804.00558v1 | http://arxiv.org/pdf/1804.00558v1.pdf | Vibration-Based Damage Detection in Wind Turbine Blades using Phase-Based Motion Estimation and Motion Magnification | Vibration-based Structural Health Monitoring (SHM) techniques are among the
most common approaches for structural damage identification. The presence of
damage in structures may be identified by monitoring the changes in dynamic
behavior subject to external loading, and is typically performed by using
experimental modal analysis (EMA) or operational modal analysis (OMA). These
tools for SHM normally require a limited number of physically attached
transducers (e.g. accelerometers) in order to record the response of the
structure for further analysis. Signal conditioners, wires, wireless receivers
and a data acquisition system (DAQ) are also typical components of traditional
sensing systems used in vibration-based SHM. However, instrumentation of
lightweight structures with contact sensors such as accelerometers may induce
mass-loading effects, and for large-scale structures, the instrumentation is
labor intensive and time consuming. Achieving high spatial measurement
resolution for a large-scale structure is not always feasible while working
with traditional contact sensors, and there is also the potential for a lack of
reliability associated with fixed contact sensors in outliving the life-span of
the host structure. Among the state-of-the-art non-contact measurements,
digital video cameras are able to rapidly collect high-density spatial
information from structures remotely. In this paper, the subtle motions from
recorded video (i.e. a sequence of images) are extracted by means of
Phase-based Motion Estimation (PME) and the extracted information is used to
conduct damage identification on a 2.3-meter long Skystream wind turbine blade
(WTB). The PME and phased-based motion magnification approach estimates the
structural motion from the captured sequence of images for both a baseline and
damaged test cases on a wind turbine blade. | ['Christopher Niezrecki', 'Zhu Mao', 'Peyman Poozesh', 'Aral Sarrafi'] | 2018-03-30 | null | null | null | null | ['motion-magnification'] | ['computer-vision'] | [ 1.68895945e-01 -6.09945297e-01 4.72795188e-01 5.22090256e-01
-5.97195148e-01 -6.27353370e-01 -1.40473396e-01 1.58834770e-01
-1.75571311e-02 3.43505442e-01 -7.97965750e-02 -7.80265480e-02
-4.39833492e-01 -7.23542511e-01 -2.18954027e-01 -1.04843211e+00
-3.70235771e-01 -8.54335502e-02 5.81938624e-01 -2.01432273e-01
2.58909345e-01 8.00408423e-01 -1.74265718e+00 -2.66403973e-01
3.06918830e-01 1.10512006e+00 4.73349452e-01 7.28815734e-01
5.33628285e-01 2.87282497e-01 -9.96659517e-01 5.51965177e-01
-1.99484602e-01 -3.97008955e-01 -7.17779398e-01 3.50026786e-01
-8.55655745e-02 -7.84891844e-01 2.84938037e-01 6.10210121e-01
6.15786076e-01 2.66888559e-01 5.64396620e-01 -6.70312285e-01
2.45765179e-01 -4.15884145e-02 -8.30168903e-01 3.05653661e-01
6.43889010e-01 3.54954042e-02 5.87931514e-01 -1.01677227e+00
3.25149834e-01 8.31957996e-01 8.52983117e-01 1.87215641e-01
-1.10004723e+00 -1.57674357e-01 -6.49865866e-01 2.73897231e-01
-1.09606016e+00 -2.43161291e-01 1.26251376e+00 -8.64405394e-01
8.03189933e-01 4.23086792e-01 7.83592880e-01 7.91549563e-01
3.77310604e-01 -1.34012103e-01 7.82314420e-01 -3.61520320e-01
4.95603472e-01 -5.89104235e-01 -1.17572643e-01 4.52400833e-01
4.56628978e-01 3.16612601e-01 -6.23628914e-01 -2.32769072e-01
8.13969314e-01 -2.83799432e-02 -7.92164087e-01 -1.52844125e-02
-9.17071223e-01 3.22343677e-01 1.95280448e-01 5.16347885e-01
-5.27893484e-01 9.12201256e-02 4.60634261e-01 2.22412810e-01
3.30837786e-01 3.60041022e-01 -1.62384480e-01 -6.08773410e-01
-1.10067701e+00 -1.59205973e-01 5.20492077e-01 -2.51682103e-01
4.68303353e-01 3.98116410e-01 5.23596823e-01 5.90960205e-01
4.28413332e-01 5.74802101e-01 6.31778896e-01 -7.97826231e-01
1.31137863e-01 2.79611498e-01 2.21657068e-01 -1.25766945e+00
-5.29322147e-01 -1.35500520e-01 -7.99202919e-01 6.56965077e-01
2.74499983e-01 -4.56264287e-01 -2.72499382e-01 1.16081750e+00
6.47271872e-01 1.31066412e-01 -1.57450631e-01 1.13758910e+00
3.21120292e-01 6.49378777e-01 -4.53764528e-01 -6.00010097e-01
1.33083665e+00 -8.38710889e-02 -7.08665729e-01 -2.02042893e-01
7.18979418e-01 -8.16501200e-01 1.10876167e+00 5.20752370e-01
-7.78256655e-01 -4.55398083e-01 -1.35988915e+00 7.00898945e-01
1.06430165e-01 7.80828074e-02 -4.09140021e-01 4.95347172e-01
-6.85829937e-01 8.42466414e-01 -1.40038228e+00 -1.54521197e-01
-3.03582758e-01 9.90659185e-03 -4.48736846e-01 1.45789951e-01
-8.14835072e-01 8.54183078e-01 -2.15550944e-01 4.71516520e-01
-9.82576668e-01 -5.40298223e-01 -6.52774751e-01 1.91497933e-02
3.70680481e-01 -2.73752093e-01 9.51937020e-01 2.86894199e-03
-1.67364240e+00 2.14534417e-01 9.82431043e-03 4.66149598e-02
2.40018237e-02 -5.40803075e-01 -3.30835849e-01 7.39613414e-01
-8.87221396e-02 -6.77736461e-01 1.30233884e+00 -1.38629138e+00
-2.27892876e-01 -2.68465102e-01 -3.62626165e-01 -1.54797986e-01
-5.60281157e-01 1.74518749e-01 6.97429657e-01 -6.80043399e-01
3.39475393e-01 -6.99266076e-01 5.54518178e-02 -4.21228111e-01
-4.31858957e-01 2.61033356e-01 1.71560991e+00 -1.09163702e+00
1.32768106e+00 -2.24856043e+00 2.44817644e-01 2.56023973e-01
-2.60797739e-02 2.74999201e-01 4.23563391e-01 1.04562843e+00
-1.31465709e-02 -2.31764704e-01 -5.58090627e-01 -1.44399151e-01
-5.24734437e-01 1.75550833e-01 -1.07180454e-01 7.74999559e-01
-6.94587156e-02 1.59436651e-02 -5.77494800e-01 -1.01836175e-01
3.02774817e-01 2.13360429e-01 -1.62604928e-01 3.59963328e-01
4.42800581e-01 6.70641065e-01 -2.85978615e-01 7.70050764e-01
3.62901956e-01 3.62719238e-01 2.16218457e-03 -3.39272529e-01
-3.69645357e-01 5.55698352e-04 -1.49975550e+00 1.19108021e+00
-5.69486439e-01 3.49952191e-01 7.99173176e-01 -1.31385243e+00
1.02466214e+00 9.61040139e-01 6.52520299e-01 -1.71319414e-02
3.01957000e-02 4.71694648e-01 -2.66658545e-01 -1.08934546e+00
2.41738617e-01 -1.96214214e-01 -1.20701835e-01 3.84061545e-01
-3.00575972e-01 -2.15764523e-01 -3.59802023e-02 -3.80460709e-01
1.47102702e+00 -3.14156488e-02 1.08512506e-01 -1.96232229e-01
6.18435502e-01 3.69223952e-02 7.47865975e-01 -1.50817394e-01
7.98717067e-02 3.63271356e-01 7.16479421e-02 -1.05221614e-01
-7.56386578e-01 -8.21741045e-01 -6.43000007e-02 3.10541749e-01
1.09785020e-01 -3.06628287e-01 -7.47483194e-01 7.96685740e-02
-4.60341908e-02 1.29622385e-01 9.83826723e-03 -4.30860102e-01
-8.39664996e-01 -6.55895889e-01 2.34117299e-01 7.40094304e-01
2.28546098e-01 -7.67171681e-01 -1.38367033e+00 6.37898922e-01
-3.46882232e-02 -9.03556764e-01 -7.79228136e-02 -4.21356112e-02
-1.23139083e+00 -1.19554567e+00 -3.48387897e-01 -6.73375130e-01
6.25596523e-01 3.19487751e-01 6.33995354e-01 4.16239738e-01
-4.36203182e-01 5.44079304e-01 -5.67079008e-01 5.19943796e-03
-2.89263368e-01 -3.95725995e-01 5.30371964e-01 1.14508577e-01
-4.79423016e-01 -8.27656925e-01 -8.67026806e-01 7.40972757e-01
-8.52625370e-01 -5.06829023e-01 2.25587144e-01 1.01201999e+00
4.36975211e-01 7.92950571e-01 6.51952207e-01 -1.39188766e-01
6.92757547e-01 -2.95370609e-01 -6.89300954e-01 -3.35712999e-01
-2.96594203e-01 -6.56176984e-01 7.62991071e-01 -4.73395020e-01
-9.42107916e-01 -1.75610393e-01 -2.81289011e-01 -4.39528108e-01
-3.90028805e-01 1.29314506e+00 -2.02988937e-01 -3.34757171e-03
6.14410400e-01 -1.30852014e-01 3.00256580e-01 -8.33242536e-01
-3.68597120e-01 6.81379974e-01 8.24428976e-01 -5.15771806e-01
1.16695142e+00 4.37228888e-01 4.89752561e-01 -1.56354094e+00
-2.04407737e-01 -6.58961058e-01 -5.31114697e-01 -7.85407186e-01
6.74092174e-01 -5.45816898e-01 -7.82977581e-01 9.27014530e-01
-8.59807670e-01 -2.84676135e-01 -3.68266195e-01 8.41773272e-01
-1.00361317e-01 8.27664673e-01 -8.84651601e-01 -1.11370707e+00
-4.26315695e-01 -1.11440587e+00 1.18303144e+00 -4.32080775e-02
-5.54628491e-01 -9.34143424e-01 5.48945218e-02 6.08524203e-01
4.15861815e-01 9.98694897e-01 7.74744749e-01 4.64349508e-01
-1.16207898e-01 -7.46421397e-01 7.68416166e-01 4.31859195e-01
4.70508128e-01 2.25407839e-01 -8.06159496e-01 -7.54464924e-01
8.93556535e-01 -1.02366917e-01 2.17236072e-01 5.64267933e-01
6.12243116e-01 -2.25980058e-01 -3.89932156e-01 2.28346333e-01
1.59631836e+00 1.42895058e-01 2.57213354e-01 1.61835045e-01
7.18827724e-01 8.96333098e-01 9.76331651e-01 6.79951072e-01
-4.73813474e-01 8.29976678e-01 7.89933264e-01 1.32332236e-01
1.69821605e-01 2.05143362e-01 7.40558624e-01 1.29004395e+00
-4.78968054e-01 -1.04447611e-01 -8.71037424e-01 7.51357317e-01
-1.17909610e+00 -9.72891629e-01 -6.63613081e-01 2.47512174e+00
5.36890447e-01 -8.79776478e-02 5.84262088e-02 1.19742262e+00
7.26296246e-01 -3.34418863e-02 -2.63489515e-01 -2.55312651e-01
2.02899963e-01 1.67295203e-01 3.17514539e-01 4.31224406e-01
-5.82020521e-01 -1.87411293e-01 5.57433558e+00 5.76415598e-01
-1.23185205e+00 8.66641700e-02 -2.89416224e-01 -4.87703597e-03
1.28542721e-01 1.66725561e-01 -3.49300057e-01 5.92343092e-01
1.03570735e+00 2.41860762e-01 1.45788774e-01 5.05298078e-01
6.84481025e-01 -7.12548018e-01 -5.30424118e-01 7.38169909e-01
-3.49553257e-01 -9.22917783e-01 -7.36802518e-01 3.43478978e-01
2.93500543e-01 -4.04182404e-01 -3.90352875e-01 -6.90226972e-01
-5.54637074e-01 -4.90030646e-01 7.45602131e-01 4.12316084e-01
8.40101779e-01 -6.11480653e-01 9.55204308e-01 5.96020997e-01
-1.60455954e+00 -2.43580207e-01 -7.92165473e-02 -3.89875263e-01
9.76906478e-01 1.25005794e+00 -4.17538673e-01 7.40166545e-01
1.11288643e+00 5.81613421e-01 -3.30137312e-02 7.90989816e-01
-2.07885727e-01 1.29060936e+00 -6.27242923e-01 3.31563830e-01
-1.32102817e-01 -4.37065631e-01 1.18292022e+00 3.25758457e-01
9.03117299e-01 -7.36226812e-02 7.12677017e-02 3.49112600e-01
4.32976395e-01 -3.40885758e-01 -6.15911245e-01 1.86843917e-01
5.99969149e-01 1.37678921e+00 -6.17479324e-01 1.19975992e-01
-2.53733188e-01 5.16368330e-01 -5.44693291e-01 8.18232223e-02
-3.37822944e-01 -3.30441713e-01 6.58304930e-01 7.57928967e-01
2.04839408e-01 -8.76420975e-01 -4.60453629e-02 -6.06744826e-01
3.80800694e-01 -4.09675837e-01 2.59860575e-01 -8.01104069e-01
-1.30001605e+00 4.27261330e-02 1.90943003e-01 -1.70806623e+00
-3.38427901e-01 -5.42887390e-01 -1.16931593e+00 8.19890678e-01
-7.76422799e-01 -6.11627162e-01 -5.07824063e-01 3.77134234e-01
3.19776505e-01 2.02453047e-01 7.46620536e-01 2.75272369e-01
-7.14066088e-01 -1.98525473e-01 2.84726471e-01 9.07599833e-03
3.93411994e-01 -1.04203618e+00 -1.17563374e-01 1.07341003e+00
-5.13774931e-01 1.91847965e-01 1.01343966e+00 -8.51834595e-01
-1.78556061e+00 -5.54681301e-01 3.05508882e-01 6.61766604e-02
7.09558249e-01 2.88625192e-02 -1.18706179e+00 2.80127041e-02
1.57774594e-02 -1.81073118e-02 6.53249383e-01 -4.84498203e-01
6.48421466e-01 -6.11683764e-02 -1.19926977e+00 1.74826950e-01
6.98940694e-01 -5.93447506e-01 -6.84400141e-01 1.04949540e-02
7.71493837e-02 -3.77451688e-01 -1.43099546e+00 5.17789841e-01
3.86615843e-01 -7.85801411e-01 1.06219399e+00 1.92920178e-01
3.04912060e-01 -7.67472863e-01 7.20230192e-02 -1.33380663e+00
-2.28681371e-01 -5.78758776e-01 -3.50334048e-01 1.64630771e+00
-2.80649304e-01 -8.41628015e-01 5.12192249e-01 4.46411557e-02
-4.04421568e-01 -7.33740866e-01 -1.22087312e+00 -1.06212735e+00
-3.47649425e-01 -2.13358998e-01 2.80682296e-01 7.80299962e-01
3.19868594e-01 1.86002672e-01 -1.48729861e-01 7.54548848e-01
5.66733837e-01 -3.35082784e-02 3.92909348e-01 -1.44514847e+00
-3.10302556e-01 5.05166575e-02 -7.31665432e-01 -4.23808306e-01
-2.40733236e-01 1.81468017e-02 1.33866549e-01 -1.46571183e+00
-7.13782132e-01 1.36726111e-01 3.20994347e-01 3.33301513e-03
1.54463677e-02 1.71657994e-01 -3.67775261e-01 6.40464187e-01
6.70516074e-01 4.38078403e-01 1.04263628e+00 5.69621101e-02
-7.55692869e-02 1.85325101e-01 2.88988024e-01 7.19913363e-01
6.60804212e-01 -4.46687281e-01 -3.45511973e-01 -2.11874589e-01
2.35452384e-01 5.99293530e-01 7.07889676e-01 -1.46849656e+00
2.11352617e-01 -5.31693082e-03 1.73106343e-01 -8.22227955e-01
2.82114685e-01 -1.05524135e+00 6.12473011e-01 8.85954142e-01
4.81289923e-01 4.18770522e-01 2.96259741e-03 4.87307400e-01
-7.23553479e-01 -4.40525621e-01 7.19990969e-01 1.33489877e-01
-4.74082500e-01 -2.89540231e-01 -1.06762242e+00 -4.00130689e-01
8.65440011e-01 -5.83642721e-01 -2.63104767e-01 -1.27933890e-01
-5.73180079e-01 -2.59190351e-01 5.53049326e-01 -2.86161482e-01
8.03463280e-01 -1.03614581e+00 -3.05591673e-01 3.27133328e-01
-2.96527863e-01 3.76691282e-01 6.24943614e-01 1.22010565e+00
-8.89851987e-01 -2.33301371e-01 -5.04368544e-02 -8.29329073e-01
-1.41777253e+00 2.51375288e-01 3.40250671e-01 -1.24868318e-01
-8.33340824e-01 5.45284808e-01 -4.41911221e-01 5.42570293e-01
-4.61896449e-01 -3.89884442e-01 -2.88200706e-01 2.95234740e-01
3.58153433e-01 1.01300049e+00 6.29237831e-01 -8.12416911e-01
-3.08640152e-01 1.16068375e+00 8.82303119e-01 -2.26461500e-01
1.26440775e+00 -3.94275159e-01 -2.62308091e-01 5.94529510e-01
9.39229429e-01 1.51879907e-01 -1.31019151e+00 2.83327788e-01
7.34471297e-03 -5.03659248e-01 3.70763749e-01 -1.48459792e-01
-1.17405260e+00 8.50263953e-01 4.90743041e-01 6.44285142e-01
1.53855038e+00 -1.57091558e-01 1.14105916e+00 -1.80834215e-02
6.10369205e-01 -1.24244845e+00 1.89903975e-01 1.00468777e-01
9.13088739e-01 -5.21792591e-01 3.15154456e-02 -4.29948509e-01
-5.61099015e-02 1.27201355e+00 2.19162092e-01 -9.81923863e-02
8.17587793e-01 6.39721632e-01 1.84635639e-01 -4.88585114e-01
-4.04580623e-01 1.18464515e-01 -4.99082953e-02 5.53478837e-01
2.82914340e-01 -1.14288896e-01 -2.94522554e-01 2.17836246e-01
-6.28803745e-02 -4.84187603e-01 4.75926787e-01 1.72555101e+00
-6.86443448e-01 -9.05891120e-01 -1.27062321e+00 3.81969631e-01
-6.30790353e-01 5.90856910e-01 6.53932765e-02 5.05727053e-01
-2.09175155e-01 1.30801594e+00 -1.21272571e-01 -5.64090490e-01
7.13333547e-01 2.94699712e-04 4.96580303e-02 -3.83530408e-01
-5.53286910e-01 1.97898850e-01 1.88435137e-01 -4.98567939e-01
-5.53855419e-01 -9.36869383e-01 -1.26477885e+00 -2.35346094e-01
-7.49387085e-01 1.62032321e-01 6.94651365e-01 1.00693154e+00
1.76023200e-01 7.99135029e-01 9.21340823e-01 -1.22324836e+00
-4.79628831e-01 -1.17773175e+00 -1.27695286e+00 4.69261289e-01
4.59247231e-01 -1.17138863e+00 -8.47508669e-01 3.49759459e-01] | [6.545657157897949, 2.4896340370178223] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.