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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
a3469b37-960a-4df9-9ec9-22a778c22781 | from-the-string-landscape-to-the-mathematical | 2202.06086 | null | https://arxiv.org/abs/2202.06086v1 | https://arxiv.org/pdf/2202.06086v1.pdf | From the String Landscape to the Mathematical Landscape: a Machine-Learning Outlook | We review the recent programme of using machine-learning to explore the landscape of mathematical problems. With this paradigm as a model for human intuition - complementary to and in contrast with the more formalistic approach of automated theorem proving - we highlight some experiments on how AI helps with conjecture formulation, pattern recognition and computation. | ['Yang-Hui He'] | 2022-02-12 | null | null | null | null | ['automated-theorem-proving', 'automated-theorem-proving'] | ['miscellaneous', 'reasoning'] | [ 1.59710541e-01 7.76852846e-01 -2.38496989e-01 -2.75652528e-01
-1.21920988e-01 -4.83220607e-01 7.62536883e-01 4.72110033e-01
-1.01808898e-01 8.51778865e-01 -1.53409421e-01 -1.39482403e+00
-5.48884451e-01 -9.36195433e-01 -5.57753086e-01 -1.05450712e-02
-3.58248502e-01 6.83487594e-01 2.00851649e-01 -4.74748671e-01
7.93215752e-01 1.55103564e-01 -1.19723630e+00 6.39412045e-01
6.25294030e-01 8.03270161e-01 -4.92769718e-01 9.83914256e-01
-1.02453925e-01 1.69352651e+00 -2.67943978e-01 -7.95050740e-01
-2.79768091e-02 -5.05060077e-01 -1.42568707e+00 -1.08499572e-01
1.87741607e-01 1.31786466e-01 -2.72485912e-01 1.28740358e+00
-4.65781689e-01 -6.25388026e-01 5.19088507e-01 -1.80462885e+00
-5.16720831e-01 6.73302889e-01 -1.17824018e-01 2.15278283e-01
1.11890006e+00 1.45961449e-01 1.27246976e+00 -1.79001927e-01
4.47189569e-01 1.47245669e+00 8.44653904e-01 4.79453146e-01
-1.30542815e+00 -3.79851758e-01 -1.98735043e-01 8.90887260e-01
-1.31011796e+00 -4.17484939e-01 9.53326523e-01 -6.62682295e-01
8.45459580e-01 5.98150790e-01 9.47077036e-01 3.56697768e-01
5.40712416e-01 9.90811825e-01 1.48043656e+00 -1.21819603e+00
1.21524125e-01 5.43914855e-01 6.05606318e-01 9.97669995e-01
3.83730024e-01 3.11059117e-01 -4.90730166e-01 -4.32380497e-01
7.55860627e-01 -4.79528755e-01 -6.32300973e-02 -5.45292079e-01
-1.30205560e+00 1.18059933e+00 3.74565907e-02 4.47598726e-01
-1.67541206e-02 3.04598331e-01 5.71491420e-01 9.04430091e-01
-2.08438747e-02 1.08540285e+00 -8.48911822e-01 -1.36241645e-01
-9.44782913e-01 7.38739967e-01 1.39334059e+00 4.73323673e-01
4.69652027e-01 -7.48997554e-02 5.69860101e-01 -4.78091031e-01
4.99059409e-01 1.03970151e-02 -1.11985967e-01 -1.22495317e+00
2.00054020e-01 7.43550539e-01 1.66559130e-01 -1.16374791e+00
-1.63771585e-01 -1.28495380e-01 -4.77514297e-01 4.63362545e-01
5.76437056e-01 1.02282673e-01 -3.76056358e-02 1.10581315e+00
-3.98219377e-02 1.53484359e-01 4.30434421e-02 4.45652246e-01
4.08765256e-01 3.79242986e-01 -3.10114384e-01 -7.13732779e-01
9.54885960e-01 -6.00198030e-01 -7.45389402e-01 3.12939793e-01
9.16385829e-01 -4.16299582e-01 6.61418855e-01 1.07147443e+00
-1.27028680e+00 -1.69005901e-01 -1.06210542e+00 1.18344046e-01
-4.10539925e-01 -4.75247324e-01 1.54008067e+00 6.77405655e-01
-1.09555995e+00 6.07723057e-01 -3.11149687e-01 8.19094628e-02
6.15649343e-01 4.49632168e-01 -4.09529284e-02 -1.39417946e-01
-1.17053652e+00 1.33381939e+00 6.73962355e-01 -7.81027973e-02
-2.72508025e-01 -7.11907804e-01 -9.51965451e-01 -2.72612780e-01
8.03904831e-01 -6.12595677e-01 1.51640570e+00 -6.98000133e-01
-1.30998433e+00 1.05743551e+00 -2.43270487e-01 -9.44146037e-01
4.41619992e-01 -4.39132117e-02 -5.32135785e-01 -1.66346159e-04
-2.10510001e-01 3.53751659e-01 5.89018226e-01 -1.13069713e+00
-3.25831801e-01 -3.49068731e-01 4.94552553e-01 -5.43148220e-01
5.60424209e-01 3.84629071e-01 6.33721411e-01 -1.17693856e-01
2.36305922e-01 -4.84632820e-01 -2.71301240e-01 6.83645755e-02
-2.10294485e-01 -9.31676567e-01 5.67504644e-01 -1.15726925e-01
1.34226477e+00 -1.46606517e+00 3.32766891e-01 4.87785280e-01
7.48086512e-01 1.49888188e-01 7.30654955e-01 5.20164907e-01
-6.43757284e-01 3.80818695e-01 2.12870896e-01 4.70390081e-01
4.33787078e-01 3.56723428e-01 -4.40357745e-01 2.81024784e-01
4.18477535e-01 1.38453710e+00 -1.10870790e+00 -9.03127015e-01
3.76852244e-01 -4.82538819e-01 -4.93385911e-01 -1.31991550e-01
-7.08525240e-01 -1.43350273e-01 -5.97429037e-01 7.49594808e-01
3.72825444e-01 -5.43566704e-01 4.32197928e-01 3.00935924e-01
-1.30781323e-01 7.17927217e-01 -1.30591917e+00 1.41701615e+00
1.52279109e-01 9.89323497e-01 1.16746100e-02 -1.57869196e+00
6.53864264e-01 4.10579175e-01 -1.06961392e-01 -4.52049673e-01
1.77325025e-01 2.24837616e-01 5.69357514e-01 -7.52971828e-01
-9.03402194e-02 -5.83480597e-01 -9.30520296e-02 9.80611742e-01
-2.96101987e-01 -8.45211685e-01 1.15176894e-01 2.69007772e-01
8.12284827e-01 9.65997428e-02 7.91393995e-01 -6.14759922e-01
8.89938712e-01 6.69303060e-01 3.04604322e-01 7.49839902e-01
-6.18640967e-02 -4.93228465e-01 8.52562547e-01 -1.32547200e+00
-9.18036163e-01 -5.01611948e-01 -3.20031077e-01 6.63696527e-01
-1.82565361e-01 -9.10848379e-01 -7.22793996e-01 -5.67699075e-01
1.03329025e-01 9.23912764e-01 -5.29377878e-01 3.96606280e-03
-5.52601337e-01 -2.87303805e-01 4.89607960e-01 3.03376466e-01
2.22441807e-01 -1.21222711e+00 -8.52012098e-01 -2.98072342e-02
2.47287720e-01 -5.93018353e-01 7.60811090e-01 4.44516391e-01
-9.72554266e-01 -1.72543943e+00 3.39788973e-01 -7.81220615e-01
3.21044117e-01 -2.59171128e-01 1.47848117e+00 6.65136039e-01
-4.05826062e-01 2.70844489e-01 -1.62488446e-01 -8.81349504e-01
-9.99486029e-01 -5.46157420e-01 1.34400025e-01 -8.99629414e-01
8.06364119e-01 -5.59154570e-01 3.93725753e-01 -1.38320118e-01
-8.00760865e-01 2.70648420e-01 2.51497298e-01 6.54002130e-01
-4.00239080e-01 5.39649725e-01 1.27208412e-01 -8.27923954e-01
8.30030560e-01 -9.55943465e-02 -7.07945645e-01 5.66615999e-01
-7.45004296e-01 2.85405606e-01 5.86592972e-01 -1.74302131e-01
-1.65381461e-01 -5.98635316e-01 5.43065310e-01 4.72902507e-02
-2.76769638e-01 5.59851170e-01 -1.26659214e-01 -4.27570701e-01
6.67746663e-01 1.42732739e-01 1.68903768e-01 -5.46121858e-02
2.60296017e-01 3.57288539e-01 4.67029184e-01 -1.09878588e+00
1.24030709e+00 -1.09039567e-01 5.52966535e-01 -5.50552070e-01
-8.33870173e-01 -9.96472873e-03 -8.08602631e-01 7.99191669e-02
1.22346416e-01 -3.54816228e-01 -1.28251183e+00 -1.14276811e-01
-1.43908501e+00 -2.33221665e-01 -3.26389462e-01 5.21199346e-01
-9.36595976e-01 4.61769015e-01 -2.91779995e-01 -1.46494186e+00
-6.60487339e-02 -7.60700703e-01 4.72139686e-01 -4.24348302e-02
-7.30974913e-01 -1.34161246e+00 1.18286237e-01 4.05054301e-01
1.36538997e-01 3.14990848e-01 1.49647284e+00 -7.17651129e-01
-3.46971363e-01 -2.73407221e-01 -3.59212935e-01 2.22422987e-01
-3.01098168e-01 -7.37350956e-02 -7.49197841e-01 2.62329400e-01
3.57875109e-01 -7.29319632e-01 2.89884329e-01 -6.44413382e-03
1.19394004e+00 -6.18835092e-01 -1.99888840e-01 -3.42681408e-02
1.29093957e+00 -1.39176995e-01 6.50586486e-01 6.36648595e-01
-1.79733172e-01 4.54436541e-01 5.98414063e-01 2.73011029e-01
2.81308532e-01 4.24234360e-01 2.17083860e-02 1.42573401e-01
3.96700144e-01 -4.04829308e-02 -9.44858566e-02 4.23295677e-01
-4.79147851e-01 2.71002799e-01 -1.33640635e+00 2.06563234e-01
-1.99614716e+00 -1.36345828e+00 -5.05591214e-01 1.57281375e+00
1.26191151e+00 7.80343413e-01 4.82023776e-01 9.42319334e-01
3.96295995e-01 -2.10896477e-01 1.24670312e-01 -1.04053581e+00
7.67261386e-02 6.30702317e-01 -2.53198836e-02 7.44215965e-01
-7.68231332e-01 8.45844448e-01 8.52079391e+00 6.15849376e-01
-3.76142949e-01 -1.62807792e-01 2.03144595e-01 5.57770252e-01
-2.92847961e-01 4.41655427e-01 -1.90334126e-01 -2.76144724e-02
7.55553126e-01 -1.84188977e-01 7.66534626e-01 7.25626886e-01
-2.69282665e-02 -3.44493717e-01 -1.76754868e+00 7.95572817e-01
2.53533155e-01 -1.89401567e+00 2.48202868e-02 -7.87200332e-02
3.79876822e-01 -5.75840354e-01 -3.12839657e-01 1.25863016e-01
2.98706859e-01 -1.37591672e+00 9.48961616e-01 5.81878662e-01
2.26357162e-01 -7.17451334e-01 8.84400547e-01 6.89020276e-01
-5.81614494e-01 -2.03618586e-01 -6.45036697e-02 -1.32753336e+00
-4.99280661e-01 2.94110239e-01 -8.48062396e-01 5.67826152e-01
2.05549046e-01 3.75051022e-01 -5.99243641e-01 1.02048063e+00
-3.49926680e-01 3.93815666e-01 -2.84454674e-01 -5.72013974e-01
1.95763230e-01 -6.24246821e-02 5.73172808e-01 1.41143131e+00
-7.21375048e-01 5.71031332e-01 1.34937331e-01 1.42433012e+00
7.28113413e-01 -5.32833755e-01 -6.05358660e-01 -1.26362175e-01
9.97624025e-02 6.94728434e-01 -4.95684236e-01 -5.46403944e-01
-2.08975047e-01 1.24956146e-01 -5.76288700e-02 -1.72338679e-01
-6.80669487e-01 -3.75679553e-01 -3.71054150e-02 3.91600043e-01
-5.20634539e-02 -4.84435737e-01 -8.51690233e-01 -1.17804658e+00
-8.39566404e-04 -1.53652668e+00 1.45856038e-01 -8.44101191e-01
-9.65411603e-01 -1.87823959e-02 3.85154605e-01 -4.86534357e-01
-5.17810643e-01 -1.07276070e+00 -1.14227557e+00 6.59866214e-01
-1.39384317e+00 -1.29987073e+00 4.52758133e-01 5.12431681e-01
1.77838892e-01 -9.19898301e-02 1.24895823e+00 -6.91746652e-01
4.88693081e-02 2.82303035e-01 -4.77257937e-01 8.84126425e-02
-1.04064889e-01 -1.43349469e+00 2.68553168e-01 4.37167823e-01
3.86786640e-01 7.61323452e-01 1.14788687e+00 -2.31597692e-01
-1.88516963e+00 -1.70031134e-02 1.30263913e+00 -7.94163167e-01
1.37340772e+00 -2.09752589e-01 -6.80803895e-01 5.22200108e-01
2.72393614e-01 -3.20915610e-01 5.69965005e-01 6.68474078e-01
-4.60774660e-01 2.30893239e-01 -1.02675176e+00 4.38768446e-01
6.78747118e-01 -7.67956078e-01 -1.69186211e+00 6.41812980e-01
5.39461732e-01 -2.35766061e-02 -8.26304138e-01 2.31322706e-01
7.16771662e-01 -8.02424788e-01 7.57026315e-01 -1.33101642e+00
5.24550200e-01 -4.00304884e-01 -1.21528260e-01 -5.19648373e-01
-3.17450762e-01 -1.41418636e+00 -4.66517776e-01 6.63073242e-01
3.32828045e-01 -4.06303406e-01 8.72606874e-01 8.24508786e-01
3.57798666e-01 -8.56822431e-01 -9.28082466e-01 -5.52642822e-01
5.06974697e-01 -1.05556810e+00 2.49072790e-01 1.16089761e+00
1.40707171e+00 4.37634826e-01 2.35343009e-01 -3.95950019e-01
6.77316844e-01 5.29621780e-01 8.07439089e-01 -1.61422157e+00
-4.93938953e-01 -8.17337036e-01 -8.76176417e-01 -5.73127687e-01
4.35535014e-01 -8.84593189e-01 -2.35958755e-01 -1.22753263e+00
2.85120070e-01 -1.80548698e-01 6.93834969e-04 6.04588449e-01
3.43695045e-01 -1.40592948e-01 1.06721791e-02 -4.94837537e-02
-1.20206380e+00 -2.97624230e-01 1.15282369e+00 -7.86141399e-03
8.20334852e-02 8.00546780e-02 -9.46524680e-01 1.18115520e+00
7.11518884e-01 -2.74928957e-01 1.15980702e-02 -9.54774674e-03
1.05217469e+00 2.30880573e-01 7.54746318e-01 -8.23442996e-01
4.10881460e-01 -7.51400590e-01 7.96509385e-02 -4.57048178e-01
-3.01616788e-01 -7.65474498e-01 -1.81240886e-01 8.73611808e-01
-6.01321280e-01 3.08632374e-01 3.43419760e-01 1.76880464e-01
-3.62540819e-02 -4.45702702e-01 4.10055310e-01 -4.55940098e-01
-7.19168067e-01 -5.19345820e-01 -4.98540789e-01 7.52262548e-02
8.06824267e-01 -1.43341705e-01 -3.30873758e-01 -1.54491514e-01
-5.64597905e-01 4.46422756e-01 4.97452319e-01 -8.83160159e-02
6.10437512e-01 -7.31889486e-01 -6.87213242e-01 9.08295363e-02
-1.62201807e-01 -1.28425598e-01 -7.11115837e-01 1.10305297e+00
-7.36790955e-01 1.08437216e+00 -5.71151003e-02 -3.58999223e-01
-1.11386597e+00 9.42019820e-01 5.49834788e-01 -3.97662342e-01
-2.86706030e-01 7.48375833e-01 -1.74516588e-01 -2.25239486e-01
2.48464555e-01 -5.08337498e-01 9.91134346e-02 -8.09924424e-01
6.32217288e-01 2.69748896e-01 -3.61052930e-01 2.05868229e-01
-7.19142795e-01 3.15899372e-01 -4.86626998e-02 5.51111214e-02
1.50661600e+00 3.76931697e-01 -8.73569965e-01 8.64592135e-01
5.95826864e-01 -1.95401847e-01 -1.38929859e-01 -2.53780872e-01
6.38188601e-01 -1.64795280e-01 9.48005840e-02 -1.12861240e+00
1.91683486e-01 8.32186282e-01 -1.71953291e-01 7.96626508e-01
5.75734973e-01 2.81901538e-01 2.64369696e-01 1.00146794e+00
5.33850372e-01 -9.97239828e-01 -1.61661670e-01 5.60425699e-01
6.06990337e-01 -1.35905921e+00 8.10824811e-01 -3.30355108e-01
-2.60231376e-01 1.71879065e+00 1.97611541e-01 -3.93905580e-01
5.69222093e-01 5.30227304e-01 -5.32615364e-01 -6.57988906e-01
-9.54725862e-01 2.68690944e-01 2.99205542e-01 5.02782464e-01
5.97426832e-01 2.12013066e-01 -9.80059728e-02 6.48476958e-01
-7.31694877e-01 7.19835579e-01 4.49918002e-01 1.34298074e+00
-7.20158994e-01 -1.25937200e+00 -8.03310573e-01 1.92321360e-01
-3.63776326e-01 -1.98494256e-01 -8.64312828e-01 1.19968748e+00
5.21613955e-01 1.17957878e+00 -2.90612310e-01 -3.68509859e-01
-3.36201549e-01 3.17108095e-01 1.26142907e+00 -6.04126394e-01
-3.90826672e-01 -2.87818909e-01 4.10746783e-01 -3.89427185e-01
-8.03843915e-01 -6.86086893e-01 -9.07696664e-01 -9.25666809e-01
-7.73834229e-01 8.30045402e-01 2.64396191e-01 1.45394492e+00
-6.36074066e-01 1.32307440e-01 3.14653158e-01 -2.61747748e-01
-5.79020023e-01 -6.02653682e-01 -5.01116574e-01 -3.67819458e-01
4.87629920e-01 -4.12364751e-01 -5.53740978e-01 7.42416233e-02] | [8.890679359436035, 6.963185787200928] |
edc526ef-d5c3-44ef-8794-38add614c3bc | universal-dependencies-and-morphology-for | null | null | https://aclanthology.org/E17-1034 | https://aclanthology.org/E17-1034.pdf | Universal Dependencies and Morphology for Hungarian - and on the Price of Universality | In this paper, we present how the principles of universal dependencies and morphology have been adapted to Hungarian. We report the most challenging grammatical phenomena and our solutions to those. On the basis of the adapted guidelines, we have converted and manually corrected 1,800 sentences from the Szeged Treebank to universal dependency format. We also introduce experiments on this manually annotated corpus for evaluating automatic conversion and the added value of language-specific, i.e. non-universal, annotations. Our results reveal that converting to universal dependencies is not necessarily trivial, moreover, using language-specific morphological features may have an impact on overall performance. | ["Rich{\\'a}rd Farkas", "Zsolt Sz{\\'a}nt{\\'o}", "Katalin Simk{\\'o}", 'Veronika Vincze'] | 2017-04-01 | null | null | null | eacl-2017-4 | ['morphological-tagging'] | ['natural-language-processing'] | [-3.12429309e-01 4.57449496e-01 2.45570630e-01 -8.53653550e-01
-8.39055955e-01 -6.60995841e-01 1.24534383e-01 4.98957783e-01
-7.08273411e-01 1.32801986e+00 4.95870829e-01 -5.80808043e-01
-5.16514480e-02 -7.07489252e-01 -4.07303572e-01 -2.31028154e-01
9.57786664e-02 6.46421015e-01 4.15414512e-01 -7.37815738e-01
5.98183908e-02 3.50972116e-01 -1.09504116e+00 2.29937613e-01
1.17474365e+00 -2.53083766e-03 2.60186851e-01 5.94923198e-01
-1.41327843e-01 3.35303634e-01 -9.07638431e-01 -1.20305240e+00
2.43840739e-02 -4.77133363e-01 -1.21294320e+00 -3.61597955e-01
3.25650007e-01 -5.81570864e-02 1.30628884e-01 1.05296171e+00
4.23869997e-01 -1.75908078e-02 5.68130136e-01 -3.91890138e-01
-8.04430664e-01 1.19982159e+00 8.35976973e-02 4.82476592e-01
4.63237077e-01 -2.68648863e-01 1.12480462e+00 -4.94104534e-01
9.77449656e-01 1.26549006e+00 5.58643639e-01 4.81081814e-01
-1.09753370e+00 -3.39455575e-01 -2.52122264e-02 1.44283980e-01
-1.28680205e+00 -6.62564874e-01 3.28949720e-01 -2.95680165e-01
1.56291759e+00 6.88076913e-02 4.69801605e-01 6.14863813e-01
4.07695442e-01 1.42182978e-02 1.12583625e+00 -1.11411953e+00
-3.00139576e-01 2.99101412e-01 2.54722357e-01 5.71457922e-01
7.35431850e-01 -2.29452625e-01 -8.93334672e-02 2.15172321e-01
5.46236157e-01 -1.11987710e+00 -1.45908415e-01 2.64841020e-01
-8.80436659e-01 7.43861377e-01 -6.91910414e-03 8.49227369e-01
5.82902282e-02 -2.07607955e-01 6.15047812e-01 5.65401316e-01
2.54559457e-01 4.47627217e-01 -8.23496163e-01 -9.53939259e-02
-6.75550640e-01 2.80538127e-02 9.31348324e-01 1.13468921e+00
7.13141382e-01 6.65688440e-02 1.81323737e-01 9.11935925e-01
4.40536477e-02 3.77957970e-01 4.22597557e-01 -8.03823829e-01
6.57584131e-01 3.57935339e-01 1.51634030e-02 -6.54174805e-01
-5.48261285e-01 -1.07589990e-01 1.91286355e-02 -1.69022352e-01
7.56996393e-01 -2.39229277e-01 -8.46201658e-01 1.73540902e+00
1.92222133e-01 -9.33233917e-01 4.30276543e-01 5.51700890e-01
8.85066569e-01 4.37427074e-01 4.21665996e-01 -4.87823844e-01
1.41192257e+00 -6.18270993e-01 -1.17740858e+00 -1.33051500e-01
1.18207848e+00 -1.28487527e+00 1.15722835e+00 3.07612419e-01
-1.34966707e+00 -3.31548184e-01 -9.28152382e-01 -1.99934170e-01
-5.60378432e-01 1.65225208e-01 7.13647783e-01 1.11266196e+00
-1.03319466e+00 7.28165627e-01 -6.65431619e-01 -8.32777202e-01
-3.65002811e-01 3.55712771e-01 -7.53708601e-01 1.84237510e-01
-1.40212035e+00 1.42925036e+00 7.66550064e-01 -1.89399183e-01
3.51812273e-01 -6.70201927e-02 -1.03196728e+00 -1.65455952e-01
1.53735861e-01 -2.95838654e-01 1.46541488e+00 -8.17489624e-01
-1.36998665e+00 1.14295304e+00 -1.59549713e-01 -3.12476099e-01
1.73418984e-01 -2.32603505e-01 -7.86510170e-01 -8.70295539e-02
2.96503514e-01 2.78871924e-01 3.57740410e-02 -6.66012883e-01
-8.30630541e-01 -2.58447349e-01 1.21728279e-01 -5.90145439e-02
-2.75557697e-01 7.41046309e-01 -2.64178514e-01 -8.66388321e-01
1.04304545e-01 -7.85441220e-01 1.15937091e-01 -1.08299863e+00
1.35308653e-01 -5.18728554e-01 1.60083652e-01 -1.44798839e+00
1.54700851e+00 -1.79873860e+00 4.18075398e-02 -2.94422992e-02
-5.09327352e-01 2.55269349e-01 1.63055360e-01 5.50281405e-01
-9.75374728e-02 4.32891905e-01 -4.35283929e-01 -1.36332870e-01
2.66306894e-03 9.32033300e-01 2.37331063e-01 3.20920914e-01
2.64016926e-01 8.55808854e-01 -9.05346155e-01 -7.19075441e-01
1.25576869e-01 2.07931340e-01 -5.24082482e-01 -8.34082142e-02
1.38596222e-01 3.38833600e-01 1.05141468e-01 7.74191082e-01
6.36657834e-01 8.33082914e-01 5.95699012e-01 -1.32082775e-01
-4.37189549e-01 1.05740476e+00 -8.84425223e-01 1.67729795e+00
-6.31611407e-01 2.72283226e-01 -6.75325468e-02 -7.83231497e-01
9.93013620e-01 4.85130548e-01 -4.59757119e-01 -7.49364793e-01
3.01514626e-01 7.78329074e-01 5.67139745e-01 -4.23963845e-01
7.98554063e-01 -3.87566596e-01 -4.95159864e-01 1.79804698e-01
6.33137882e-01 -4.35724437e-01 9.40873265e-01 -6.30556196e-02
6.52881861e-01 3.66730869e-01 8.31398904e-01 -7.32659876e-01
6.07597172e-01 2.29580671e-01 8.88026416e-01 2.83344060e-01
-1.33817166e-01 6.14588082e-01 4.93499815e-01 -2.86482632e-01
-1.01790118e+00 -8.96624088e-01 -4.39893693e-01 1.10696638e+00
-6.28667772e-01 -6.69646144e-01 -9.00654018e-01 -7.84963906e-01
-4.64329869e-01 1.32437956e+00 -4.11929101e-01 7.00470060e-02
-1.32712805e+00 -7.29621828e-01 7.46011555e-01 5.18523574e-01
9.54790227e-03 -1.22687781e+00 -7.17898011e-01 6.82180226e-01
-2.62260497e-01 -1.33598673e+00 -1.77054182e-01 4.83059764e-01
-7.51203835e-01 -8.46864104e-01 -2.24468082e-01 -1.02611661e+00
5.12851834e-01 -4.50844049e-01 1.26650393e+00 2.05494732e-01
-1.87176857e-02 9.42821242e-03 -9.32236016e-01 -4.70200419e-01
-9.51651394e-01 3.24442774e-01 -5.82860224e-02 -8.22689295e-01
4.76472169e-01 -4.47452486e-01 3.25390935e-01 -1.57299131e-01
-7.07226932e-01 -5.86847484e-01 4.47478116e-01 5.82682967e-01
1.75332695e-01 -2.38881543e-01 6.09433234e-01 -1.49102139e+00
6.88629687e-01 2.00433135e-02 -4.23834622e-01 2.49785677e-01
-4.19066966e-01 5.91152646e-02 6.17599785e-01 1.70049518e-01
-1.39278710e+00 -9.08270571e-03 -7.95985997e-01 6.57281697e-01
-2.66655862e-01 6.02441728e-01 -5.84686875e-01 -5.73345535e-02
6.77969515e-01 -3.34313095e-01 -5.99262834e-01 -6.75748706e-01
5.16288102e-01 5.12625694e-01 7.45962918e-01 -8.83106411e-01
4.69534308e-01 -1.88664928e-01 -1.55583262e-01 -8.34497750e-01
-4.92168635e-01 -1.69454545e-01 -1.40749252e+00 1.89636379e-01
8.43697608e-01 -5.75382113e-01 3.37921202e-01 6.41552284e-02
-1.55393469e+00 -3.90457451e-01 -4.10333037e-01 5.57903647e-01
-4.32984710e-01 6.71037078e-01 -7.91468024e-01 -3.86549175e-01
-3.39474410e-01 -7.97340751e-01 3.62837195e-01 3.08138356e-02
-7.94565737e-01 -1.25881767e+00 2.95148969e-01 -1.48314238e-03
1.23456739e-01 1.53439254e-01 1.38086987e+00 -6.85578346e-01
3.49593401e-01 -7.76061565e-02 2.16624495e-02 5.93608201e-01
4.09953207e-01 3.40481907e-01 -5.00070095e-01 6.12782408e-03
-1.31989762e-01 -1.47424176e-01 4.58842069e-01 1.23310745e-01
2.83772703e-02 -2.51032948e-01 1.47377178e-01 3.45363706e-01
1.32322884e+00 2.50879198e-01 6.20527864e-01 5.40898979e-01
4.34513271e-01 9.81254697e-01 9.38598454e-01 1.13241196e-01
7.09439278e-01 5.11907637e-01 3.32569629e-02 1.78155109e-01
-3.77826065e-01 1.70853753e-02 6.26071990e-01 1.45018792e+00
-3.19685400e-01 -5.26345931e-02 -9.76641595e-01 9.77855563e-01
-1.48859668e+00 -3.68227065e-01 -6.53791666e-01 2.14538860e+00
1.08912218e+00 3.51897478e-01 6.08916767e-02 1.99635670e-01
9.18022096e-01 -4.10898656e-01 4.70051199e-01 -1.39638758e+00
-5.30011535e-01 8.55677903e-01 4.74451661e-01 8.34121346e-01
-8.25963438e-01 1.70014751e+00 6.87102365e+00 3.48965794e-01
-8.01625788e-01 4.31518525e-01 -1.04760520e-01 2.46599764e-01
-2.84643829e-01 2.13769153e-01 -8.84116352e-01 9.89716966e-03
1.35689580e+00 -2.08405763e-01 5.94767705e-02 3.91317904e-01
1.50618553e-02 -1.60392106e-01 -9.23076332e-01 3.58923107e-01
9.03519318e-02 -7.85259485e-01 1.28836870e-01 -1.02758169e-01
3.39896530e-01 -6.80695772e-02 -5.69347978e-01 3.32350135e-01
4.86994386e-01 -7.27414668e-01 1.09609365e+00 -1.37776896e-01
7.22671211e-01 -9.71455812e-01 1.25312793e+00 -1.82454959e-01
-1.15981209e+00 2.87341923e-01 -8.40696931e-01 -4.24307168e-01
4.50682610e-01 4.22023624e-01 -5.50068080e-01 9.76796329e-01
7.48504639e-01 2.51975775e-01 -8.26217473e-01 7.50458658e-01
-9.93258774e-01 7.13442624e-01 -3.90424341e-01 6.93493783e-02
1.15448661e-01 -3.55184108e-01 4.08344299e-01 1.82578135e+00
4.78679627e-01 3.03935587e-01 -4.59861279e-01 2.81419158e-01
3.48375469e-01 9.00635660e-01 -3.36214215e-01 2.21196964e-01
5.08705497e-01 1.32451308e+00 -1.02933383e+00 -3.18874151e-01
-6.88856363e-01 9.31638479e-01 7.67703712e-01 -3.87987383e-02
-5.58222413e-01 -6.60583258e-01 3.81229818e-01 2.08360031e-01
3.95960242e-01 -3.24096650e-01 -4.09984052e-01 -1.01643145e+00
7.23119602e-02 -6.29050136e-01 7.90421784e-01 -3.89932513e-01
-8.85875940e-01 8.86805236e-01 1.71033084e-01 -4.63010222e-01
-1.82689711e-01 -7.80704618e-01 -5.16372919e-01 1.03125787e+00
-1.38656950e+00 -1.17946696e+00 3.04240584e-01 3.29505086e-01
3.20363104e-01 1.75799429e-01 1.08518732e+00 4.81743574e-01
-4.79925036e-01 9.63261306e-01 -2.21804649e-01 1.77646667e-01
1.07190931e+00 -1.43419814e+00 6.70804262e-01 1.20069301e+00
5.81864938e-02 6.69613719e-01 8.57404590e-01 -7.26464570e-01
-5.95026553e-01 -9.33458269e-01 1.92033756e+00 -4.65454817e-01
7.97492206e-01 -2.26026773e-01 -1.05126500e+00 1.01111567e+00
5.61381519e-01 -4.37769294e-01 8.09774518e-01 1.10003121e-01
-1.48930654e-01 9.45983976e-02 -1.22750497e+00 5.24022877e-01
1.14142299e+00 -2.28289291e-01 -1.34920967e+00 6.64348528e-02
7.44101882e-01 -3.08780402e-01 -1.21438599e+00 3.08040798e-01
3.91862243e-01 -8.74338150e-01 2.76971906e-01 -6.11972570e-01
2.71607041e-01 -1.37708718e-02 -2.25368202e-01 -1.60061467e+00
-5.42565286e-01 -5.60912430e-01 7.62717485e-01 1.78287899e+00
7.96593130e-01 -8.57325256e-01 3.12488645e-01 6.30447030e-01
-7.35189140e-01 -3.64808552e-02 -1.20160460e+00 -8.51005137e-01
7.26258993e-01 -2.13065788e-01 6.64474070e-01 1.02122283e+00
6.58953547e-01 4.97454733e-01 -6.01705499e-02 3.12536992e-02
1.45264342e-01 -1.77447066e-01 5.66407204e-01 -1.30378950e+00
-3.74901026e-01 -1.59115970e-01 -3.06452841e-01 -2.48141423e-01
3.07146162e-01 -1.04384673e+00 1.78306364e-02 -1.59571528e+00
-3.35712880e-01 -4.60584313e-01 1.17197866e-02 6.29985273e-01
-1.79460332e-01 2.47740954e-01 2.19636396e-01 -1.34541348e-01
-5.34628145e-03 -4.66185343e-03 8.23985457e-01 3.27358425e-01
-3.51746887e-01 -4.54498172e-01 -7.83971727e-01 7.74173915e-01
1.14774787e+00 -7.37537086e-01 2.24120617e-01 -7.88498044e-01
3.71283412e-01 -1.41291678e-01 -4.71778512e-01 -8.38355482e-01
-1.32260129e-01 -9.40607041e-02 -2.99799163e-02 -2.51940578e-01
-1.63819268e-01 -4.91322994e-01 6.81525990e-02 5.75962126e-01
2.85837144e-01 4.26239997e-01 5.29384196e-01 -3.36436898e-01
-2.35293284e-01 -9.88260567e-01 7.15603590e-01 -3.87897938e-01
-6.66531563e-01 -4.85157937e-01 -6.90211773e-01 2.28089154e-01
8.24533463e-01 -2.35003173e-01 -2.32061088e-01 -1.26659404e-03
-7.81366527e-01 -1.55774862e-01 6.06286347e-01 1.85854271e-01
6.21106178e-02 -8.85244429e-01 -8.98954928e-01 3.66175212e-02
1.09530501e-01 -3.81930649e-01 -1.48195460e-01 5.66773057e-01
-9.90297914e-01 6.85241401e-01 -5.53335071e-01 5.10684699e-02
-1.40662634e+00 2.97760993e-01 -4.88690771e-02 -2.67731369e-01
-4.69135165e-01 6.78544879e-01 -1.48219675e-01 -8.23677301e-01
-3.31229270e-01 -5.30928791e-01 -5.37376225e-01 1.78420603e-01
9.27971676e-02 2.77774006e-01 7.09541976e-01 -1.17971647e+00
-5.17681062e-01 5.44141769e-01 -2.00516433e-01 -3.15063894e-01
1.14984751e+00 -2.62717605e-01 -5.98554373e-01 4.24541950e-01
5.96322834e-01 7.15617836e-01 -2.58544862e-01 1.12667494e-01
4.77000982e-01 -9.19109732e-02 -2.97122687e-01 -7.22161472e-01
-6.73022687e-01 7.16875136e-01 -7.66715268e-03 1.34718701e-01
8.98680866e-01 1.22663818e-01 7.84599841e-01 4.68632936e-01
4.16654378e-01 -1.42256606e+00 -8.71657133e-01 1.04480767e+00
6.45364523e-01 -6.88557327e-01 -6.27915934e-02 -1.11715567e+00
-3.89865249e-01 1.39312220e+00 4.98095006e-01 -8.96027014e-02
4.09906089e-01 3.71293306e-01 4.33268100e-01 1.34183750e-01
-4.45695758e-01 -4.00132596e-01 -2.62075961e-01 9.63109672e-01
1.04282117e+00 4.06926364e-01 -1.51636684e+00 8.06237757e-01
-9.80732858e-01 -3.53026211e-01 9.99408185e-01 8.93908501e-01
-3.67346764e-01 -1.90900719e+00 -4.69106317e-01 1.68419078e-01
-8.49384010e-01 -4.48424846e-01 -4.77968067e-01 1.37576413e+00
2.30883077e-01 9.40710247e-01 2.30093463e-03 -1.34635851e-01
6.56688035e-01 3.78965825e-01 1.00374329e+00 -1.12356329e+00
-8.57454538e-01 1.46402016e-01 8.42832208e-01 1.84077723e-03
-3.89406264e-01 -9.02772963e-01 -1.46659112e+00 -3.37757587e-01
-4.41907257e-01 4.71675545e-01 4.30076659e-01 1.02212453e+00
-2.79692113e-01 4.07662570e-01 -1.63480937e-01 -4.95767504e-01
-3.18681538e-01 -1.20042074e+00 -5.73883116e-01 2.11130500e-01
-3.16877902e-01 -4.96492803e-01 -3.86244029e-01 2.49521688e-01] | [10.388339042663574, 10.112144470214844] |
47b307ce-b573-454f-8d92-d4919620df27 | bev-mae-bird-s-eye-view-masked-autoencoders | 2212.05758 | null | https://arxiv.org/abs/2212.05758v1 | https://arxiv.org/pdf/2212.05758v1.pdf | BEV-MAE: Bird's Eye View Masked Autoencoders for Outdoor Point Cloud Pre-training | Current outdoor LiDAR-based 3D object detection methods mainly adopt the training-from-scratch paradigm. Unfortunately, this paradigm heavily relies on large-scale labeled data, whose collection can be expensive and time-consuming. Self-supervised pre-training is an effective and desirable way to alleviate this dependence on extensive annotated data. Recently, masked modeling has become a successful self-supervised learning approach for point clouds. However, current works mainly focus on synthetic or indoor datasets. When applied to large-scale and sparse outdoor point clouds, they fail to yield satisfactory results. In this work, we present BEV-MAE, a simple masked autoencoder pre-training framework for 3D object detection on outdoor point clouds. Specifically, we first propose a bird's eye view (BEV) guided masking strategy to guide the 3D encoder learning feature representation in a BEV perspective and avoid complex decoder design during pre-training. Besides, we introduce a learnable point token to maintain a consistent receptive field size of the 3D encoder with fine-tuning for masked point cloud inputs. Finally, based on the property of outdoor point clouds, i.e., the point clouds of distant objects are more sparse, we propose point density prediction to enable the 3D encoder to learn location information, which is essential for object detection. Experimental results show that BEV-MAE achieves new state-of-the-art self-supervised results on both Waymo and nuScenes with diverse 3D object detectors. Furthermore, with only 20% data and 7% training cost during pre-training, BEV-MAE achieves comparable performance with the state-of-the-art method ProposalContrast. The source code and pre-trained models will be made publicly available. | ['Yongtao Wang', 'Zhiwei Lin'] | 2022-12-12 | null | null | null | null | ['point-cloud-pre-training'] | ['computer-vision'] | [-1.21614024e-01 -1.95247069e-01 -2.39920646e-01 -4.77157176e-01
-7.65949130e-01 -3.16271305e-01 3.57428491e-01 2.88880393e-02
-3.39276105e-01 3.02824438e-01 -3.39195549e-01 -2.84117848e-01
3.48292083e-01 -9.62682366e-01 -1.16768122e+00 -5.98556817e-01
1.28421396e-01 4.16933626e-01 5.79527855e-01 -3.88926230e-02
-5.01197055e-02 7.13136911e-01 -1.83590984e+00 9.25507098e-02
8.49107087e-01 9.99593616e-01 8.59976828e-01 2.68301815e-01
-3.40992093e-01 3.36857438e-01 -3.74731630e-01 -1.76411465e-01
4.00561422e-01 3.44865210e-02 -1.16753921e-01 3.55606407e-01
6.76456511e-01 -5.91309965e-01 -3.59782577e-01 1.07281232e+00
5.74301064e-01 -1.46530062e-01 7.32911468e-01 -1.14645183e+00
-6.48245633e-01 1.54153004e-01 -5.49576461e-01 9.67466608e-02
-1.40907288e-01 2.42586404e-01 7.66722441e-01 -1.50532353e+00
2.56272793e-01 1.14363551e+00 7.81828642e-01 6.44222558e-01
-1.12982810e+00 -9.85475600e-01 3.67679805e-01 -1.33972047e-02
-1.92004740e+00 -2.67381340e-01 9.57688808e-01 -4.82873648e-01
1.15678775e+00 -1.09219849e-01 7.87000656e-01 9.16457772e-01
-1.33094683e-01 8.72764707e-01 9.08377171e-01 -3.57928872e-01
3.48881692e-01 4.70873654e-01 -3.44186813e-01 8.55792224e-01
3.75444949e-01 4.29059744e-01 -4.75378543e-01 1.42923251e-01
9.23755646e-01 3.16408157e-01 -1.02252670e-01 -7.03935146e-01
-9.86418068e-01 8.29204202e-01 9.77532744e-01 7.42736682e-02
-2.42289588e-01 4.18367162e-02 -3.91895324e-02 2.82512465e-03
6.80954218e-01 -1.77871376e-01 -3.58377010e-01 3.27161551e-01
-9.58217323e-01 7.95841888e-02 2.41455406e-01 1.41523194e+00
1.18277645e+00 2.39299223e-01 1.74541160e-01 6.62274897e-01
7.97420144e-01 1.00214028e+00 2.85028201e-02 -5.26350975e-01
5.59173226e-01 8.02731454e-01 1.40035659e-01 -7.46553421e-01
-3.68306525e-02 -5.75788915e-01 -7.51313865e-01 5.97012222e-01
-3.18374000e-02 2.49448165e-01 -1.20451045e+00 1.40739906e+00
5.65955222e-01 3.22345883e-01 4.24012132e-02 1.03445542e+00
1.12486386e+00 7.44071186e-01 -1.19016431e-01 7.97355920e-02
1.10491467e+00 -6.41281843e-01 -2.20975518e-01 -5.52023172e-01
5.26333272e-01 -5.80658257e-01 1.03719008e+00 -1.17087081e-01
-8.49578440e-01 -9.18837428e-01 -1.10844684e+00 -8.17250386e-02
-3.01461130e-01 4.09618676e-01 5.48821390e-01 6.46944225e-01
-7.15916038e-01 2.06534758e-01 -1.02977347e+00 -2.96295822e-01
8.89205217e-01 3.84778053e-01 -3.01160485e-01 -2.64607489e-01
-6.68206453e-01 7.66265452e-01 3.01863343e-01 7.72913918e-03
-1.25161254e+00 -6.96301520e-01 -9.64861929e-01 1.37238139e-02
2.32126042e-01 -8.20769489e-01 1.07794225e+00 -3.50129277e-01
-1.15348887e+00 1.18362498e+00 -3.92307043e-01 -3.81820649e-01
2.89350420e-01 -2.45138302e-01 -2.34156862e-01 -7.62112588e-02
3.45796525e-01 1.13770151e+00 1.02056956e+00 -1.63792515e+00
-7.81957626e-01 -5.44732332e-01 -5.14864288e-02 1.83898732e-01
-1.90400928e-01 -2.87695050e-01 -5.66960573e-01 -4.25505906e-01
5.04755855e-01 -8.35153759e-01 -2.33103558e-01 3.93719941e-01
-2.36208424e-01 -3.97337914e-01 8.83527458e-01 6.25644624e-02
6.44399285e-01 -2.37731910e+00 -2.32503697e-01 -1.21209279e-01
2.26732031e-01 2.22774372e-01 1.64447911e-02 1.50124282e-01
1.14035301e-01 -8.09492543e-02 -2.51348823e-01 -7.08933055e-01
-6.59926161e-02 3.01119834e-01 -4.63269413e-01 5.65894008e-01
5.62956512e-01 9.24517035e-01 -8.37841034e-01 -7.27379143e-01
6.60058737e-01 5.13030171e-01 -6.84244156e-01 4.31911230e-01
-3.88971627e-01 4.62153405e-01 -4.42158163e-01 1.05653548e+00
1.08391511e+00 -3.67095232e-01 -5.57969034e-01 -1.72249287e-01
-3.64984065e-01 2.55769908e-01 -1.10000634e+00 1.95667064e+00
-5.81279993e-01 5.07510722e-01 3.73378582e-02 -6.91597998e-01
1.25014865e+00 -5.85265607e-02 2.31434777e-01 -3.23871762e-01
1.75504148e-01 3.76551598e-01 -3.76625538e-01 -2.03560710e-01
3.03116560e-01 -2.09854633e-01 2.26337671e-01 -2.96223667e-02
3.13785642e-01 -6.33044660e-01 -2.53189266e-01 6.88722879e-02
7.72659242e-01 2.34561220e-01 1.55203253e-01 1.09465644e-01
4.06067014e-01 1.59993827e-01 5.39664745e-01 7.66294479e-01
-1.02849498e-01 7.79645920e-01 -3.31603676e-01 -2.58476615e-01
-9.65147018e-01 -1.29661596e+00 -5.13924420e-01 5.27394116e-01
4.50093120e-01 -2.81304061e-01 -4.26730216e-01 -6.92836404e-01
2.15676889e-01 6.14644647e-01 -2.70370662e-01 -9.05217081e-02
-3.25159013e-01 -4.94449466e-01 2.29409739e-01 6.38004839e-01
6.65984273e-01 -7.26365626e-01 -5.33331752e-01 7.59904012e-02
2.29865387e-01 -1.35444343e+00 -9.71746221e-02 4.85081017e-01
-1.16615891e+00 -7.31949389e-01 -6.98973775e-01 -1.07215858e+00
8.57294858e-01 9.28383708e-01 9.76548553e-01 -3.44646350e-02
-1.95840061e-01 2.56084502e-01 -3.34003031e-01 -8.61891568e-01
-7.40990639e-02 6.64431900e-02 2.00672865e-01 -1.89518645e-01
9.46426272e-01 -7.24570215e-01 -5.61754405e-01 3.99082422e-01
-5.64619064e-01 -2.94559682e-03 7.69010484e-01 8.41908276e-01
1.00300097e+00 -8.05731267e-02 1.55440181e-01 -4.00282651e-01
-1.62716120e-01 -3.78086209e-01 -9.44474578e-01 -2.06741884e-01
-4.28124785e-01 -3.04082453e-01 3.04194599e-01 -3.30229431e-01
-7.53126562e-01 6.05063200e-01 -3.71686399e-01 -1.14707625e+00
-5.00101149e-01 -6.39734790e-02 -3.37478757e-01 -3.84666413e-01
7.22405612e-01 5.41956544e-01 -2.70586550e-01 -5.60469151e-01
3.63776028e-01 8.76697540e-01 3.78362626e-01 -2.68852860e-01
1.34767652e+00 8.20683241e-01 -1.55657455e-01 -9.52500522e-01
-9.00071681e-01 -6.97736800e-01 -8.16485822e-01 -2.17784479e-01
8.50274920e-01 -1.67297876e+00 -5.56287944e-01 2.72433400e-01
-1.33524907e+00 -1.72879413e-01 -4.32399213e-01 6.82126284e-01
-4.75826621e-01 2.25971982e-01 -2.23958656e-01 -1.06816459e+00
-1.01400971e-01 -9.89274025e-01 1.65030825e+00 2.00913504e-01
3.93740535e-01 -4.57437545e-01 -8.39774236e-02 2.58784741e-01
5.70267960e-02 -1.13893867e-01 6.59372747e-01 -2.89586306e-01
-1.33550215e+00 -2.90896147e-01 -4.70734417e-01 4.91287529e-01
1.13895439e-01 -3.59199077e-01 -1.27558887e+00 -2.53292531e-01
1.50008127e-01 -4.05815661e-01 9.54490364e-01 3.10967207e-01
1.19961011e+00 1.42727450e-01 -5.89810014e-01 8.05070639e-01
1.45722783e+00 -5.45677394e-02 3.03126872e-01 5.90348616e-02
8.64721060e-01 3.64175409e-01 7.72115827e-01 4.35290605e-01
5.71127951e-01 5.12529373e-01 8.20742130e-01 -1.40629530e-01
-1.70823708e-01 -7.41257250e-01 1.79405317e-01 8.69268835e-01
-3.26543041e-02 -1.49544731e-01 -8.24681222e-01 6.58256590e-01
-1.60275412e+00 -8.18367779e-01 -1.45682245e-01 2.20922947e+00
5.64317048e-01 3.82706076e-01 -9.42572057e-02 1.17516913e-01
5.55618584e-01 1.50018871e-01 -6.04711175e-01 3.05845499e-01
-1.90344006e-01 1.98649809e-01 6.07305706e-01 2.87108302e-01
-1.27845883e+00 1.22713315e+00 5.06448984e+00 8.49407315e-01
-1.17013156e+00 3.16595286e-01 7.50663802e-02 -7.22231045e-02
-2.25771025e-01 -7.00852349e-02 -1.30530405e+00 4.56736624e-01
4.23331201e-01 2.66073197e-01 -1.50728896e-02 1.32378376e+00
1.08679734e-01 7.47938827e-02 -1.06271040e+00 1.38647389e+00
-5.22999167e-02 -1.34945130e+00 -1.23957060e-01 8.25692937e-02
6.17596090e-01 6.25431120e-01 4.90619130e-02 3.47822189e-01
8.40253234e-02 -7.15460658e-01 8.48212898e-01 1.13659792e-01
8.69812369e-01 -4.86036271e-01 5.76428175e-01 7.64719546e-01
-1.35437381e+00 2.63396613e-02 -9.23951566e-01 8.36608093e-03
1.96347728e-01 8.56685519e-01 -8.33837330e-01 3.18176240e-01
1.02959371e+00 8.66335392e-01 -4.83272225e-01 1.21903729e+00
-4.68043685e-01 4.97098863e-01 -6.40990973e-01 -2.88590968e-01
2.26759195e-01 2.05050129e-02 6.46481931e-01 1.02411950e+00
5.18598855e-01 1.10455401e-01 4.67875004e-01 1.20661318e+00
-4.62031215e-02 -3.34255435e-02 -1.06012428e+00 2.55329221e-01
7.29548097e-01 9.63397563e-01 -5.16928613e-01 -1.30086184e-01
-6.51955545e-01 8.18946838e-01 3.38707626e-01 1.44468293e-01
-7.78343916e-01 -2.11658165e-01 7.03947425e-01 2.76572049e-01
8.09774518e-01 -5.44771016e-01 -2.14611068e-01 -1.21248174e+00
1.41990989e-01 -2.99838036e-01 -1.39197379e-01 -8.75613689e-01
-1.38053203e+00 6.28189325e-01 -6.00325838e-02 -1.76160800e+00
6.82617649e-02 -6.79900765e-01 -3.98750365e-01 7.72898436e-01
-1.83107829e+00 -1.39772606e+00 -6.61968172e-01 5.54234028e-01
5.81408918e-01 -1.44940689e-01 7.68217564e-01 4.90461498e-01
-2.77410090e-01 3.33159298e-01 -5.64581864e-02 7.51678050e-02
4.47579861e-01 -1.01207578e+00 6.03920460e-01 7.70561755e-01
5.24284184e-01 4.02047753e-01 2.28062183e-01 -7.38645911e-01
-1.45745468e+00 -1.45732582e+00 6.36664391e-01 -6.36500299e-01
2.94379115e-01 -8.91570449e-01 -9.69336927e-01 5.04494905e-01
-3.11068892e-01 4.39751506e-01 5.22553563e-01 -7.68294632e-02
-2.77142853e-01 -2.40294173e-01 -9.98609543e-01 3.71112376e-01
1.49389219e+00 -6.81222022e-01 -6.52376890e-01 4.89467680e-01
9.45217073e-01 -6.26323581e-01 -5.14457405e-01 7.34390378e-01
-1.41154579e-03 -1.04053140e+00 1.19752836e+00 -2.12054491e-01
2.32446715e-01 -7.99527287e-01 -5.06805658e-01 -9.61096346e-01
-4.03001547e-01 -5.07406890e-02 -3.72990280e-01 1.14350414e+00
2.82337099e-01 -3.19035769e-01 1.31331313e+00 -4.51993532e-02
-4.48506147e-01 -6.95744634e-01 -1.09247887e+00 -9.64142978e-01
-2.37065822e-01 -7.50923753e-01 5.26701570e-01 6.87581658e-01
-7.69672811e-01 4.39038962e-01 -2.53879249e-01 6.99574471e-01
8.82327318e-01 4.38242167e-01 1.19744468e+00 -1.23713958e+00
-1.22271456e-01 -6.31785195e-04 -6.89080775e-01 -1.68586075e+00
7.99880922e-02 -9.87265944e-01 3.23296726e-01 -1.51863074e+00
-1.16616376e-01 -7.51336217e-01 -4.71316949e-02 3.62121016e-01
-9.50957462e-02 3.54941696e-01 2.32412294e-01 4.28025782e-01
-5.14199674e-01 1.00822043e+00 9.71407413e-01 -3.55350554e-01
-2.92887807e-01 2.69193381e-01 -2.50293225e-01 7.14772165e-01
5.12516797e-01 -7.44697392e-01 -3.51946592e-01 -7.29114354e-01
-1.82858363e-01 -3.99511337e-01 7.05032468e-01 -1.32045960e+00
3.20172042e-01 3.50710265e-02 6.30966604e-01 -1.31194234e+00
7.33864129e-01 -1.13522565e+00 -1.61536396e-01 4.31197464e-01
3.63734603e-01 -2.83740550e-01 2.45792821e-01 8.44583988e-01
-2.74000376e-01 -2.34205619e-01 7.66586363e-01 -2.37131879e-01
-8.70641708e-01 6.65404677e-01 -1.94335654e-02 -2.68270582e-01
1.02614009e+00 -4.06990021e-01 2.34597712e-03 -7.94713050e-02
-3.46295625e-01 2.53957808e-01 6.45129263e-01 3.71563733e-01
1.06055820e+00 -1.32979107e+00 -6.37591302e-01 5.03647804e-01
5.55747747e-01 8.17675650e-01 3.02120715e-01 3.90442312e-01
-3.97497445e-01 3.87436211e-01 6.17160164e-02 -1.32647777e+00
-1.03400552e+00 7.55991817e-01 1.97828233e-01 4.25585061e-01
-7.79721916e-01 1.11342812e+00 4.18698519e-01 -6.13885880e-01
4.29709464e-01 -4.88861084e-01 2.14953855e-01 -2.47619286e-01
2.85539836e-01 -1.32622793e-01 1.05563872e-01 -5.87704003e-01
-5.59885502e-01 8.33188355e-01 1.19565576e-01 2.65280038e-01
1.33503795e+00 -6.92121238e-02 3.24368685e-01 5.26376426e-01
9.91738856e-01 5.39801978e-02 -1.39791071e+00 -5.42774796e-01
-4.12404209e-01 -8.10760677e-01 3.44777048e-01 -2.01232612e-01
-8.84716392e-01 1.29721129e+00 9.44851577e-01 -1.46969870e-01
8.68059754e-01 3.52644771e-01 5.94218850e-01 5.25755525e-01
7.88067043e-01 -6.14924908e-01 -1.15902228e-02 4.15215939e-01
6.51497722e-01 -1.73740339e+00 1.76030695e-02 -7.08162129e-01
-3.25829864e-01 7.11343288e-01 9.59535658e-01 -1.77688941e-01
8.96588147e-01 1.07416362e-01 7.00718835e-02 -1.63068205e-01
-4.33872759e-01 -6.77222192e-01 2.36587569e-01 8.32257926e-01
-7.85562210e-03 -6.66503608e-02 4.52751338e-01 4.59058493e-01
-1.18566252e-01 -4.07182537e-02 -1.13311306e-01 9.34968352e-01
-6.75660431e-01 -8.29846144e-01 -3.70031416e-01 2.78813511e-01
1.25511453e-01 -1.75027281e-01 -1.43110186e-01 7.81773150e-01
4.68958706e-01 6.94025517e-01 1.83207661e-01 -5.62618256e-01
4.61877823e-01 -1.96799949e-01 5.95365107e-01 -1.01986635e+00
-9.68541130e-02 1.11386038e-01 -4.00591344e-01 -2.79279739e-01
-5.23055732e-01 -3.60967934e-01 -1.09750700e+00 -5.62292263e-02
-8.03977191e-01 -1.04317203e-01 8.07262778e-01 6.67479038e-01
5.98870397e-01 2.69298851e-01 7.44787395e-01 -1.27318239e+00
-4.82368827e-01 -8.08804035e-01 -4.93409067e-01 9.98254586e-03
4.24303889e-01 -9.79630470e-01 -2.89319217e-01 -1.63118735e-01] | [7.816186904907227, -2.9499926567077637] |
118a7576-e4e5-441b-b2a1-d0e9035a1ec1 | pic-a-phrase-in-context-dataset-for-phrase | 2207.09068 | null | https://arxiv.org/abs/2207.09068v5 | https://arxiv.org/pdf/2207.09068v5.pdf | PiC: A Phrase-in-Context Dataset for Phrase Understanding and Semantic Search | While contextualized word embeddings have been a de-facto standard, learning contextualized phrase embeddings is less explored and being hindered by the lack of a human-annotated benchmark that tests machine understanding of phrase semantics given a context sentence or paragraph (instead of phrases alone). To fill this gap, we propose PiC -- a dataset of ~28K of noun phrases accompanied by their contextual Wikipedia pages and a suite of three tasks for training and evaluating phrase embeddings. Training on PiC improves ranking models' accuracy and remarkably pushes span-selection (SS) models (i.e., predicting the start and end index of the target phrase) near-human accuracy, which is 95% Exact Match (EM) on semantic search given a query phrase and a passage. Interestingly, we find evidence that such impressive performance is because the SS models learn to better capture the common meaning of a phrase regardless of its actual context. SotA models perform poorly in distinguishing two senses of the same phrase in two contexts (~60% EM) and in estimating the similarity between two different phrases in the same context (~70% EM). | ['Anh Nguyen', 'Trung Bui', 'Seunghyun Yoon', 'Thang M. Pham'] | 2022-07-19 | null | null | null | null | ['semantic-retrieval'] | ['natural-language-processing'] | [ 1.15994811e-01 -9.33803096e-02 -5.53150952e-01 -1.97463989e-01
-1.21774173e+00 -9.18350399e-01 6.81289315e-01 7.62940407e-01
-9.95830536e-01 5.34028172e-01 9.12262261e-01 -5.26685894e-01
-1.07523553e-01 -7.32538581e-01 -6.44826233e-01 -5.22844017e-01
1.12350546e-01 4.80623156e-01 3.14229935e-01 -2.43920743e-01
6.03788912e-01 -1.85106158e-01 -1.32767570e+00 3.66505146e-01
4.58239764e-01 7.54448891e-01 4.41615969e-01 7.38636076e-01
-5.49851716e-01 7.88983256e-02 -4.66306150e-01 -3.72660637e-01
-8.64626095e-03 -1.87702283e-01 -1.04018259e+00 -7.64722109e-01
6.95950389e-01 1.83902085e-01 -2.02150464e-01 9.02137935e-01
4.15521353e-01 2.27215171e-01 5.93593419e-01 -6.79817796e-01
-1.08233118e+00 7.36014962e-01 -2.42646873e-01 7.81625569e-01
7.51883745e-01 -1.26667023e-01 2.10971355e+00 -1.20876145e+00
6.69187725e-01 9.99126256e-01 8.64496291e-01 5.83726525e-01
-1.28608418e+00 -2.75290161e-01 4.49402109e-02 1.64795056e-01
-1.11341321e+00 -1.03918031e-01 1.68473437e-01 -2.09971294e-01
1.58187103e+00 1.61928415e-01 5.40233850e-01 1.47496402e+00
1.91869676e-01 3.72323900e-01 7.60714114e-01 -5.99698603e-01
1.00472115e-01 4.42045256e-02 8.04223299e-01 3.19763184e-01
4.35335070e-01 -1.15280025e-01 -8.23568761e-01 -5.29644072e-01
2.54093241e-02 -2.21177951e-01 -4.05357659e-01 -3.11021861e-02
-1.31528473e+00 8.12146366e-01 2.98408151e-01 5.57801545e-01
-1.77820355e-01 3.87977362e-01 6.60914779e-01 2.18453884e-01
2.58747876e-01 9.95589852e-01 -8.91134501e-01 -6.91925585e-01
-6.67702138e-01 7.55048215e-01 9.59950745e-01 8.36531162e-01
5.75515270e-01 -6.62321508e-01 -1.91711158e-01 8.78410757e-01
5.78464717e-02 4.68580812e-01 7.33029425e-01 -5.30678928e-01
6.34836316e-01 3.85402590e-01 2.15966985e-01 -8.75266731e-01
-3.74145299e-01 -2.85398185e-01 6.92904964e-02 -5.42543769e-01
5.39365411e-01 1.01728380e-01 -5.26114285e-01 1.98197770e+00
-1.24848723e-01 2.30374306e-01 9.93613452e-02 7.02967107e-01
7.62182534e-01 8.69716048e-01 5.32796621e-01 -1.53452875e-02
1.66412270e+00 -8.35404575e-01 -3.10727626e-01 -6.46021008e-01
7.88091958e-01 -7.34931111e-01 1.56262493e+00 -2.02401429e-01
-8.89699996e-01 -3.84477615e-01 -1.09159815e+00 -4.59409744e-01
-7.54233479e-01 -6.28477037e-01 3.49064559e-01 2.94618309e-01
-8.00010383e-01 6.29544675e-01 -4.90314811e-01 -6.69338167e-01
1.07670568e-01 -4.47175205e-02 -4.31286067e-01 -1.43913403e-01
-1.47454178e+00 1.11765170e+00 4.45520014e-01 -5.94520867e-01
-3.13438475e-01 -1.32274318e+00 -8.33157182e-01 2.89488196e-01
1.16244614e-01 -6.75016701e-01 1.06604743e+00 -3.93098414e-01
-6.26507759e-01 1.11957824e+00 -4.08728927e-01 -5.52114904e-01
-1.99341193e-01 -6.23523593e-01 -5.31751871e-01 1.56736344e-01
4.97584015e-01 5.23400962e-01 6.14245057e-01 -8.31519842e-01
-8.22303295e-01 -2.51149803e-01 1.86364487e-01 3.65102828e-01
-6.25739098e-01 3.57639045e-02 -3.94996196e-01 -6.02127969e-01
1.05777524e-01 -8.62549186e-01 9.26843286e-02 -2.51656801e-01
-1.37889504e-01 -6.70462549e-01 5.98879755e-01 -5.46426356e-01
1.69313645e+00 -2.35444713e+00 -2.47205701e-02 -2.09621247e-02
-1.52232805e-02 1.47610128e-01 -3.41477722e-01 9.19011474e-01
-6.05269037e-02 6.27208412e-01 -9.82651561e-02 -2.16480494e-01
3.90237898e-01 4.14063185e-01 -7.98480332e-01 7.99348205e-02
1.98559582e-01 9.60620105e-01 -1.34292316e+00 -3.93279791e-01
-2.54294246e-01 1.88855276e-01 -6.45105422e-01 -1.32364482e-01
-3.89519721e-01 -9.06113982e-02 -4.10337359e-01 3.58526289e-01
8.10439140e-02 -1.61296532e-01 2.53776789e-01 1.40880281e-02
9.93610770e-02 9.86962140e-01 -7.55458891e-01 1.78618932e+00
-6.84569240e-01 8.01248550e-01 -6.14457726e-01 -5.35539150e-01
5.80577970e-01 5.09766579e-01 2.37886906e-01 -9.90072370e-01
-3.41837883e-01 4.76094514e-01 -1.41544387e-01 -6.11863077e-01
9.88566399e-01 -1.75818503e-01 -5.36102951e-01 6.72667265e-01
1.10789604e-01 2.22308002e-02 3.46600980e-01 1.63476542e-01
1.47395170e+00 3.43750231e-02 3.03432792e-01 -5.88893294e-01
2.67143637e-01 1.67196840e-01 5.86946726e-01 9.78684306e-01
-2.87097752e-01 7.47705698e-01 5.56166351e-01 -5.27655602e-01
-1.46301079e+00 -1.30490971e+00 -3.92272353e-01 1.62297964e+00
3.07589144e-01 -8.79292667e-01 -4.23224449e-01 -8.00445914e-01
3.60299468e-01 1.05300307e+00 -7.03360021e-01 -1.15747191e-01
-7.88950205e-01 -6.97578192e-01 5.93766153e-01 8.48088324e-01
-8.84941742e-02 -1.03609097e+00 -4.37400639e-01 3.01525444e-01
-3.54817361e-01 -1.17206764e+00 -7.25129783e-01 3.81959766e-01
-6.25972271e-01 -1.13307166e+00 -3.14430773e-01 -1.08878815e+00
2.36735374e-01 1.71277761e-01 1.71186626e+00 1.35362580e-01
-5.89682870e-02 4.83769476e-01 -6.64306283e-01 -2.71017551e-01
-8.20599943e-02 1.89203843e-01 2.51473546e-01 -7.88789332e-01
1.11957085e+00 -5.57767808e-01 -7.20949411e-01 1.24690548e-01
-8.30290496e-01 -4.55011368e-01 3.66483390e-01 1.07483101e+00
6.24850929e-01 -3.30758512e-01 6.68153465e-01 -6.77043617e-01
9.16328311e-01 -7.63036668e-01 7.32441545e-02 3.69096667e-01
-7.93751478e-01 1.96053565e-01 4.35546190e-01 -3.71557146e-01
-5.34510851e-01 -7.01210439e-01 -3.08122337e-01 1.23113118e-01
2.31237933e-02 6.09856129e-01 5.87261856e-01 5.78898013e-01
8.49995494e-01 2.13328749e-01 -4.85135257e-01 -5.12266517e-01
4.03450787e-01 5.11867225e-01 4.85381246e-01 -9.49313521e-01
5.99195421e-01 3.80894318e-02 -3.40168089e-01 -7.35111117e-01
-1.31461978e+00 -9.35963929e-01 -4.42830473e-01 4.72883046e-01
1.04351628e+00 -8.17285419e-01 -1.57216653e-01 -4.62287098e-01
-1.09336722e+00 9.33701098e-02 -2.50297576e-01 4.42456216e-01
-3.46079290e-01 2.60895818e-01 -4.32612747e-01 -2.29288861e-01
-4.72983837e-01 -7.55524516e-01 1.04444587e+00 9.63759869e-02
-7.70074904e-01 -1.09185851e+00 4.58197922e-01 1.21962793e-01
3.70336324e-01 -1.94657922e-01 1.52996588e+00 -1.21496415e+00
-1.26453817e-01 -2.71808952e-01 -2.64743388e-01 2.16382965e-01
-1.08156003e-01 -1.55744180e-01 -9.54138279e-01 -2.06398845e-01
-3.75831246e-01 -2.32744247e-01 7.80336082e-01 -1.43413357e-02
7.94739962e-01 -1.65526271e-01 -4.84036207e-01 1.81964219e-01
1.79330969e+00 -5.97632416e-02 3.69308531e-01 7.56819546e-01
2.41270572e-01 2.30267122e-01 5.21755815e-01 7.75201768e-02
1.69186056e-01 7.58824766e-01 5.02585620e-02 8.60034466e-01
-2.30245963e-01 -7.05164909e-01 2.89482951e-01 9.14138794e-01
3.67579430e-01 -3.64674032e-01 -1.15005410e+00 1.07054567e+00
-1.57234740e+00 -1.01864159e+00 8.97195712e-02 2.09688640e+00
9.87731218e-01 5.36505461e-01 -1.91351742e-01 -1.29436567e-01
4.58879501e-01 4.33591545e-01 -1.71881631e-01 -8.28115523e-01
3.88102233e-02 6.68193102e-01 4.23959702e-01 5.53158283e-01
-9.71378624e-01 8.45803440e-01 6.54636431e+00 9.02438402e-01
-6.34652555e-01 1.35741353e-01 -8.74027051e-03 4.93760183e-02
-6.76303387e-01 1.87837377e-01 -1.03157902e+00 6.77924633e-01
1.12609112e+00 -3.15279752e-01 1.87342301e-01 8.67996097e-01
-1.13522537e-01 5.37282042e-02 -1.49852800e+00 8.01477373e-01
1.07182525e-01 -1.45862484e+00 2.72119612e-01 -1.55166700e-01
5.95653772e-01 2.71724135e-01 1.44232124e-01 7.57690907e-01
-1.36293471e-01 -1.01664829e+00 5.10580838e-01 2.51388878e-01
5.38434863e-01 -5.08656442e-01 8.03244233e-01 2.06361994e-01
-1.09216070e+00 -1.19312525e-01 -5.90808690e-01 -1.11405827e-01
1.20601736e-01 2.22815514e-01 -6.05086029e-01 1.96841478e-01
8.27685595e-01 5.70939124e-01 -5.55690587e-01 7.97136605e-01
-4.00123626e-01 7.85934925e-01 -1.43239945e-01 -4.82653230e-01
6.98177814e-01 1.22994967e-01 8.77158999e-01 1.62825191e+00
4.96092848e-02 -1.19387418e-01 4.68532965e-02 5.82634687e-01
-1.29704550e-01 2.19174340e-01 -6.59744084e-01 -3.60745072e-01
7.71471322e-01 8.74504030e-01 -2.23046899e-01 -1.59295708e-01
-7.86764801e-01 7.57273734e-01 5.76599658e-01 2.28076816e-01
-6.96116686e-01 -4.94888306e-01 1.20261741e+00 -1.22795753e-01
3.86672378e-01 -2.89782226e-01 -4.19601709e-01 -1.09517598e+00
2.98222482e-01 -4.12138999e-01 4.62794632e-01 -6.76885605e-01
-1.59023356e+00 4.08747613e-01 -1.39089048e-01 -9.07350361e-01
-1.20382912e-01 -7.77733207e-01 -8.19422603e-01 9.88330305e-01
-1.58237350e+00 -7.28450954e-01 1.25846744e-01 -4.87102047e-02
7.22414851e-01 1.61189921e-02 1.30723608e+00 1.76410168e-01
-1.77011907e-01 6.91714466e-01 2.50226766e-01 2.78271943e-01
8.66702020e-01 -1.53396595e+00 6.62415683e-01 5.59995055e-01
6.57473266e-01 1.10112381e+00 9.20598567e-01 -3.60490292e-01
-1.28276002e+00 -7.38936186e-01 1.71727002e+00 -1.05914485e+00
1.20874131e+00 -2.26095796e-01 -1.05007291e+00 7.28607893e-01
3.31659466e-01 -1.19470641e-01 1.04372120e+00 7.47509956e-01
-9.20520544e-01 1.59222677e-01 -7.98936367e-01 8.22142541e-01
1.05701184e+00 -1.00616753e+00 -1.52003157e+00 2.66751409e-01
1.31857634e+00 -2.03390718e-01 -8.93843651e-01 2.96161562e-01
6.97879076e-01 -6.25195742e-01 1.12317550e+00 -9.77528930e-01
7.74206638e-01 -1.40400559e-01 -6.56015515e-01 -1.28523171e+00
-2.73640543e-01 -7.91459158e-02 -9.83509570e-02 1.04117584e+00
8.41148973e-01 -3.93507153e-01 7.45943427e-01 5.42691946e-01
-1.61145017e-01 -1.12997055e+00 -9.25113201e-01 -8.28487754e-01
3.78991425e-01 -7.67739713e-01 4.74243611e-01 8.30478370e-01
4.36343014e-01 4.68991250e-01 3.56393814e-01 2.35492155e-01
2.08275095e-01 2.16739878e-01 2.28601545e-01 -8.03865254e-01
-1.87800676e-01 -5.05420089e-01 -5.01233459e-01 -1.07627058e+00
2.86027610e-01 -1.06630337e+00 -1.61330923e-02 -1.45452774e+00
3.75545859e-01 -4.77012634e-01 -6.65444255e-01 2.11852014e-01
-5.81317782e-01 2.34291792e-01 -4.38410528e-02 8.08275416e-02
-6.04586244e-01 2.48448953e-01 6.56458795e-01 -1.23883620e-01
1.04810841e-01 -3.96749377e-01 -7.86310673e-01 5.52454472e-01
6.69670820e-01 -6.47039890e-01 -2.69674659e-01 -8.24176192e-01
7.82730460e-01 -3.14783901e-01 2.84732401e-01 -8.98577988e-01
1.23679154e-01 -7.44678918e-03 3.78443524e-02 -3.74464005e-01
1.25301510e-01 -6.38373673e-01 -5.13613641e-01 2.43944421e-01
-8.50163996e-01 5.66540599e-01 2.39183828e-01 9.92523491e-01
-3.07369828e-01 -7.23168373e-01 3.42061967e-01 -1.46119222e-01
-1.21173286e+00 4.97913398e-02 -1.00126289e-01 7.35373199e-01
6.81061387e-01 -1.66923180e-01 -5.06358862e-01 -3.92406657e-02
-7.23080754e-01 1.56110585e-01 3.26497644e-01 7.96477735e-01
4.83276337e-01 -1.23141503e+00 -5.23960710e-01 -6.31332025e-02
7.12962389e-01 -3.38449985e-01 -6.46577850e-02 3.82836640e-01
-4.68031526e-01 6.62365496e-01 2.46303946e-01 -3.51895601e-01
-1.08656538e+00 4.76446241e-01 -4.98474836e-02 -2.79525578e-01
-7.41544604e-01 1.02267969e+00 -2.01583520e-01 -4.56533611e-01
2.55594939e-01 -3.79190564e-01 -8.55130404e-02 7.44762942e-02
6.98193192e-01 1.29588917e-01 -9.01907310e-02 -3.83848190e-01
-3.82235616e-01 7.02090681e-01 -2.47862890e-01 -1.37332842e-01
1.17485905e+00 -3.95670049e-02 -1.32170215e-01 7.02650189e-01
1.59811389e+00 2.38747045e-01 -6.64785504e-01 -5.01553595e-01
6.23652518e-01 -4.50887471e-01 -2.65660167e-01 -7.46814728e-01
-1.09076738e-01 6.81733787e-01 2.21030012e-01 3.12612385e-01
4.04876620e-01 1.32542208e-01 1.25676191e+00 5.83026350e-01
4.60173398e-01 -1.21259916e+00 2.00519741e-01 9.16797638e-01
4.59778070e-01 -9.97301936e-01 -5.16289435e-02 4.74518016e-02
-5.40348828e-01 1.04721212e+00 4.80375081e-01 -2.63778180e-01
7.47759104e-01 -2.61437058e-01 -1.62885949e-01 -2.77814120e-01
-1.00079441e+00 -7.28562996e-02 2.39699915e-01 3.62089038e-01
5.46018660e-01 3.05500231e-03 -7.00715065e-01 6.02512658e-01
-4.40787435e-01 -5.28565109e-01 1.64797276e-01 1.02217031e+00
-7.51028001e-01 -1.01736581e+00 1.47233099e-01 5.42432189e-01
-8.20819080e-01 -6.79786205e-01 -1.83090027e-02 6.86692297e-01
-3.46908569e-02 7.92307973e-01 3.36695611e-01 -3.18688422e-01
2.58785218e-01 6.58660710e-01 1.01056784e-01 -1.05798316e+00
-5.50475478e-01 -6.68706834e-01 2.30181053e-01 -4.86821443e-01
-1.42639801e-01 -6.01369619e-01 -1.21259284e+00 -2.05946296e-01
-1.33391112e-01 5.73063493e-01 6.67285383e-01 1.06611598e+00
3.08038235e-01 1.14463635e-01 3.50372344e-01 -2.43257344e-01
-7.32044220e-01 -8.56373191e-01 -4.46171165e-01 7.85868764e-01
2.70788103e-01 -3.88656944e-01 -5.84333897e-01 -2.75874555e-01] | [10.538680076599121, 8.74164867401123] |
8d193ba5-c800-4001-9fa4-c823939cecb7 | fourier-rnns-for-modelling-noisy-physics-data | 2302.06534 | null | https://arxiv.org/abs/2302.06534v1 | https://arxiv.org/pdf/2302.06534v1.pdf | Fourier-RNNs for Modelling Noisy Physics Data | Classical sequential models employed in time-series prediction rely on learning the mappings from the past to the future instances by way of a hidden state. The Hidden states characterise the historical information and encode the required temporal dependencies. However, most existing sequential models operate within finite-dimensional Euclidean spaces which offer limited functionality when employed in modelling physics relevant data. Alternatively recent work with neural operator learning within the Fourier space has shown efficient strategies for parameterising Partial Differential Equations (PDE). In this work, we propose a novel sequential model, built to handle Physics relevant data by way of amalgamating the conventional RNN architecture with that of the Fourier Neural Operators (FNO). The Fourier-RNN allows for learning the mappings from the input to the output as well as to the hidden state within the Fourier space associated with the temporal data. While the Fourier-RNN performs identical to the FNO when handling PDE data, it outperforms the FNO and the conventional RNN when deployed in modelling noisy, non-Markovian data. | ['Lorenzo Zanisi', 'Stanislas Pamela', 'Vignesh Gopakumar'] | 2023-02-13 | null | null | null | null | ['time-series-prediction'] | ['time-series'] | [ 2.90590942e-01 -2.22201601e-01 2.38738030e-01 -2.01299116e-01
-7.01386994e-03 -2.77838081e-01 9.44930613e-01 -1.36954159e-01
-5.55593371e-01 6.14650309e-01 2.37664014e-01 -5.19594312e-01
-6.35757506e-01 -7.88836420e-01 -6.26986623e-01 -1.02089095e+00
-7.26179123e-01 3.72722656e-01 1.22752741e-01 -5.96693382e-02
-4.74111736e-03 7.20840394e-01 -1.68860555e+00 1.28359973e-01
4.08959568e-01 1.07126665e+00 2.33292073e-01 1.06905615e+00
-2.66570061e-01 1.18442142e+00 -4.21005845e-01 2.90925711e-01
3.16097319e-01 -5.13430357e-01 -7.77660906e-01 -3.96112800e-01
-1.55859768e-01 -8.78080949e-02 -6.54336214e-01 4.67876166e-01
1.52830377e-01 7.86877513e-01 5.48839748e-01 -1.12769485e+00
-5.56545854e-01 3.17301154e-01 5.08792818e-01 5.54736555e-01
2.55424798e-01 4.96422499e-02 7.97516108e-01 -5.74995995e-01
5.52398384e-01 1.06387568e+00 8.58815730e-01 5.23551404e-01
-1.41034901e+00 -1.66920841e-01 -6.77564996e-04 3.03324908e-01
-1.02859235e+00 -1.95767909e-01 6.20406508e-01 -5.39927065e-01
1.42879891e+00 3.94129306e-01 7.93950379e-01 1.45483983e+00
5.98773301e-01 5.16404212e-01 9.37828183e-01 -4.33795303e-01
4.97849166e-01 -2.64818996e-01 2.22234856e-02 3.17344815e-01
-6.96267009e-01 7.29996920e-01 -7.35190094e-01 -2.04065546e-01
9.86215055e-01 1.87514320e-01 -3.71047914e-01 -2.28802100e-01
-1.23109114e+00 6.52705550e-01 1.73515782e-01 5.61155736e-01
-7.50274539e-01 4.24436510e-01 6.76574290e-01 7.78410912e-01
5.51272929e-01 4.33312744e-01 -7.74394095e-01 -3.36808652e-01
-1.10605836e+00 2.11566374e-01 9.97248888e-01 5.69628000e-01
5.90132415e-01 2.77625203e-01 -1.98408455e-01 1.47629529e-01
2.26137131e-01 1.65769175e-01 6.53916657e-01 -1.18107677e+00
7.93278217e-02 2.09092841e-01 2.39716738e-01 -6.03357792e-01
-5.19529104e-01 -2.16407776e-01 -1.09545219e+00 1.77894503e-01
3.06913048e-01 -1.31154045e-01 -9.26256061e-01 1.66949451e+00
2.34885722e-01 6.83291376e-01 2.66603649e-01 3.67902726e-01
2.23703861e-01 1.09099305e+00 1.16966657e-01 -4.95091319e-01
7.41922677e-01 -4.93736923e-01 -8.41744661e-01 4.01496530e-01
7.32235074e-01 -4.44143742e-01 5.95614135e-01 4.37291205e-01
-9.69043851e-01 -8.82177055e-01 -7.29375541e-01 -2.77291127e-02
-6.81266665e-01 -1.35785952e-01 5.87266743e-01 8.59849304e-02
-1.47494185e+00 1.24585545e+00 -1.28753364e+00 -1.99999407e-01
-3.51350576e-01 5.67799330e-01 -1.85353681e-01 5.32502055e-01
-1.66787040e+00 1.16376960e+00 7.16707110e-01 4.25730914e-01
-7.69585252e-01 -9.06781197e-01 -7.01087832e-01 -1.26735076e-01
-1.49393141e-01 -7.54801691e-01 1.26822925e+00 -8.35455835e-01
-1.97943902e+00 1.64455593e-01 -9.24222097e-02 -1.02869511e+00
6.32828057e-01 -1.30815223e-01 -6.05542481e-01 -1.38812419e-02
-2.79979914e-01 4.55241084e-01 9.29129720e-01 -4.58289415e-01
-3.16945583e-01 1.52983457e-01 1.27318770e-01 -5.02027124e-02
-1.71766914e-02 -2.83550262e-01 2.24897757e-01 -7.39182711e-01
1.04980871e-01 -8.91440809e-01 -4.69176829e-01 -1.28401995e-01
4.89955954e-02 -4.62216526e-01 1.11302733e+00 -5.69455266e-01
1.35167396e+00 -2.26231098e+00 7.26214051e-01 1.94116175e-01
-1.12296470e-01 2.27997050e-01 3.63319993e-01 7.71284521e-01
-5.45949876e-01 -2.60564297e-01 -3.63426536e-01 -4.68421727e-01
6.47310540e-02 8.77433479e-01 -6.28614068e-01 5.62493861e-01
2.14169249e-01 7.27177620e-01 -8.14202249e-01 -1.48859009e-01
4.32804406e-01 7.67243147e-01 -1.88251823e-01 1.50464535e-01
-4.59180534e-01 9.20004904e-01 -3.05738240e-01 -2.74708956e-01
1.84137374e-01 -8.27476829e-02 -2.01022014e-01 1.65113539e-01
-6.14917874e-01 4.31884825e-01 -1.17107224e+00 1.79085183e+00
-7.62141526e-01 6.04751825e-01 -1.18097048e-02 -1.23730612e+00
8.94790232e-01 9.15777445e-01 8.92320216e-01 -6.65908813e-01
-5.35188522e-03 2.71589130e-01 -3.22959870e-02 -5.99200487e-01
2.56485641e-01 -5.71046948e-01 9.34048966e-02 7.05418050e-01
3.70587200e-01 1.04118675e-01 1.29728585e-01 -1.07967928e-01
1.29654479e+00 6.40456736e-01 -8.35504681e-02 -1.99077979e-01
8.13981712e-01 -3.39742243e-01 4.17550296e-01 5.64693511e-01
5.91237582e-02 2.01418951e-01 4.01284397e-01 -9.23843086e-01
-1.04182994e+00 -9.84833121e-01 -1.97784394e-01 8.83636355e-01
-5.93066335e-01 -4.17457610e-01 -2.85965592e-01 -1.59609690e-01
-1.75416023e-01 1.02167690e+00 -8.35522413e-01 -3.70882839e-01
-7.65297830e-01 -3.97719741e-01 4.13125217e-01 5.33292353e-01
3.37886810e-01 -1.52886701e+00 -8.95954847e-01 5.00295639e-01
1.56322479e-01 -8.58679533e-01 -1.39885888e-01 6.80839956e-01
-9.97466743e-01 -5.90963304e-01 -5.07318318e-01 -3.07890356e-01
1.23875655e-01 -5.30570090e-01 7.87217736e-01 -3.52369368e-01
-1.95553198e-01 6.16342008e-01 -2.39674240e-01 -3.52631211e-01
-7.17268109e-01 -3.63608450e-02 2.39060000e-01 2.65518725e-01
1.07326038e-01 -1.05239022e+00 -1.53168336e-01 -1.40717044e-01
-1.20248902e+00 -5.27192131e-02 2.93027818e-01 5.74903429e-01
3.51469159e-01 3.27004105e-01 1.42968088e-01 -5.81114173e-01
3.97173673e-01 -5.30915737e-01 -5.52755713e-01 4.42392267e-02
-2.15381071e-01 6.65132761e-01 8.92844141e-01 -6.98904157e-01
-1.04645336e+00 2.24338919e-01 -1.59449205e-01 -6.62879288e-01
-6.19016960e-02 4.64092702e-01 4.69069898e-01 1.30979672e-01
6.35532677e-01 3.00997764e-01 -5.50070032e-02 -5.32928109e-01
1.85842186e-01 2.49155119e-01 4.51166332e-01 -5.69044709e-01
5.16523600e-01 5.31101525e-01 5.62671721e-01 -7.59880126e-01
-5.81072867e-01 -6.04025841e-01 -1.19997001e+00 -2.08213836e-01
9.68485475e-01 -3.42671752e-01 -6.15272105e-01 5.72912872e-01
-1.24927115e+00 -3.73503268e-01 -9.92251337e-01 8.07907939e-01
-1.03278983e+00 2.12324008e-01 -7.00614929e-01 -1.15118456e+00
-1.13326021e-01 -7.46934831e-01 8.77559662e-01 -2.75598943e-01
-4.46257442e-01 -1.59923005e+00 3.55135083e-01 -5.30557215e-01
6.11712098e-01 4.54596102e-01 9.87674475e-01 -4.16163176e-01
-2.65098393e-01 -2.96719193e-01 4.78229463e-01 5.23267806e-01
-9.45314988e-02 -3.76007194e-03 -7.44385958e-01 2.32678335e-02
7.27808595e-01 3.93373147e-02 9.13150907e-01 4.50525343e-01
9.91676033e-01 -8.70154053e-02 3.20252627e-02 6.20411217e-01
1.25117636e+00 2.31640548e-01 4.70419705e-01 1.22200795e-01
4.89339769e-01 5.15799820e-01 2.89015900e-02 3.62001061e-01
2.64593847e-02 3.03903371e-01 2.00269252e-01 3.06871533e-01
2.54608780e-01 -7.08678961e-02 6.54225588e-01 8.83724093e-01
-3.14176857e-01 -5.22008389e-02 -1.07619882e+00 2.87864804e-01
-2.04864860e+00 -1.03288114e+00 -2.59658962e-01 2.11605310e+00
6.60825491e-01 2.17245325e-01 -2.39382312e-02 3.77838582e-01
4.01646435e-01 2.08048418e-01 -5.24638236e-01 -7.00133383e-01
1.72207784e-02 6.90642715e-01 4.60155308e-01 5.50621152e-01
-1.00723279e+00 3.73401433e-01 6.81428576e+00 4.01375949e-01
-1.16286421e+00 1.89986840e-01 -3.68644260e-02 -2.44882077e-01
-1.04740411e-01 7.68320560e-02 -4.80467021e-01 4.13183242e-01
2.06200647e+00 3.15506279e-01 5.06152213e-01 2.47631371e-01
5.76433063e-01 -3.03588714e-02 -1.35414267e+00 7.88823247e-01
-5.00287652e-01 -1.23983300e+00 -5.26382821e-03 6.18976913e-02
5.01785636e-01 5.96043877e-02 -5.76365627e-02 3.59506011e-01
2.29049653e-01 -1.14790452e+00 7.52125025e-01 1.44643104e+00
5.15666842e-01 -5.78255594e-01 5.46435416e-01 7.92899847e-01
-1.12843323e+00 -3.37676167e-01 -7.49272108e-02 -5.55569470e-01
4.55166668e-01 3.74098897e-01 -4.08356398e-01 5.54510057e-01
7.50287175e-01 9.55202520e-01 -4.41093743e-03 5.66494942e-01
1.08404599e-01 5.99069774e-01 -6.12613320e-01 2.07186624e-01
6.09100997e-01 -3.83092225e-01 6.27822220e-01 1.03291965e+00
4.73479152e-01 -1.88503966e-01 1.21693507e-01 8.37049842e-01
5.20093501e-01 -3.32091510e-01 -5.90825915e-01 -2.79592484e-01
-7.19870627e-02 7.20916867e-01 -6.11691892e-01 -2.68641591e-01
-3.06213111e-01 1.04837477e+00 1.53352082e-01 5.52043557e-01
-7.82070756e-01 -6.09232448e-02 6.11295342e-01 2.63871439e-02
4.37077165e-01 -7.14360476e-01 3.00213724e-01 -8.76194894e-01
6.39494583e-02 -2.26712570e-01 4.76289451e-01 -6.77210748e-01
-1.25387311e+00 5.52489579e-01 2.74275988e-01 -1.01530802e+00
-6.64102972e-01 -8.83349657e-01 -6.11272156e-01 1.22335362e+00
-1.06038713e+00 -9.14856315e-01 3.79663020e-01 7.08306849e-01
2.59495109e-01 1.25234455e-01 1.21047044e+00 2.03903988e-02
-2.72095799e-01 -3.08811158e-01 3.82586837e-01 -1.70679942e-01
8.68090317e-02 -1.40464580e+00 5.07517993e-01 6.14908159e-01
9.17477906e-02 6.69261515e-01 9.08413827e-01 -5.63842177e-01
-1.18752849e+00 -9.89069045e-01 1.08780801e+00 -4.04218376e-01
1.01436496e+00 -4.34449553e-01 -1.04183316e+00 7.68364012e-01
1.12953670e-01 3.15772861e-01 5.35051763e-01 -1.55326113e-01
2.06761360e-02 1.70821011e-01 -7.67453253e-01 2.42311940e-01
8.80185068e-01 -1.02985978e+00 -4.49777812e-01 3.97170514e-01
7.42066026e-01 -4.80984330e-01 -1.13943303e+00 2.33605936e-01
3.03366005e-01 -8.57583880e-01 9.08652842e-01 -8.16136420e-01
4.69457246e-02 -2.89107531e-01 -1.55469447e-01 -1.11613536e+00
-4.65527982e-01 -1.16287100e+00 -7.46404409e-01 6.66556120e-01
1.21545129e-01 -6.57575905e-01 5.18449843e-01 5.21908224e-01
-5.50638251e-02 -7.70058930e-01 -1.47473145e+00 -8.01540256e-01
1.03729010e-01 -9.36553061e-01 4.28901821e-01 6.60610914e-01
-5.18799424e-01 1.30239233e-01 -4.75445718e-01 1.97211713e-01
1.17160082e-01 -2.55070120e-01 4.94225100e-02 -1.40160251e+00
-6.44749701e-01 -3.46800774e-01 -5.70280969e-01 -8.20836306e-01
3.89614582e-01 -8.42429817e-01 4.18297723e-02 -1.29220021e+00
-5.27957082e-01 -1.25450045e-01 -5.25041759e-01 8.28010887e-02
4.22716469e-01 -3.82580817e-01 -5.69405369e-02 3.84922534e-01
-2.07001448e-01 5.19236267e-01 9.45213675e-01 3.27034146e-01
-2.75594771e-01 4.18949485e-01 4.31118518e-01 5.54657161e-01
6.56977534e-01 -4.00966704e-01 -5.62074900e-01 -1.74110994e-01
4.83982831e-01 1.54058993e-01 7.37637579e-01 -1.21053541e+00
5.37309289e-01 2.08713231e-03 3.90039802e-01 -6.51940286e-01
4.66059476e-01 -9.63862121e-01 5.71627736e-01 5.05296409e-01
-4.77147400e-01 4.23883706e-01 1.75681621e-01 8.78404915e-01
-7.95109197e-02 -2.73937285e-01 4.37259942e-01 -5.01095951e-01
-6.50913358e-01 2.86840498e-01 -7.95008719e-01 -2.71358848e-01
9.70490396e-01 -2.27652773e-01 5.55832982e-01 -4.63646472e-01
-1.38505793e+00 4.15534340e-02 -4.25416902e-02 1.78555593e-01
3.07379514e-01 -1.07676828e+00 -1.96135551e-01 4.76151317e-01
-5.69070578e-01 3.44531946e-02 3.52997452e-01 9.91872907e-01
-4.31325912e-01 7.69695818e-01 -1.63406640e-01 -6.15260243e-01
-7.24399149e-01 6.45230532e-01 7.13946700e-01 -4.80574429e-01
-7.82955170e-01 8.19109261e-01 -2.57362336e-01 -6.63265884e-01
2.76610225e-01 -8.39361072e-01 -1.09003291e-01 2.27198198e-01
1.90112531e-01 6.06790662e-01 1.16691425e-01 -7.81984448e-01
3.96066457e-02 1.20694995e-01 2.80242562e-01 -5.07773817e-01
1.58783340e+00 6.36212751e-02 -4.07730699e-01 1.39273024e+00
1.47253847e+00 -7.04113007e-01 -1.42202401e+00 -3.74873906e-01
2.32393607e-01 3.94409388e-01 2.75978357e-01 -3.91439378e-01
-5.27846992e-01 9.93618011e-01 4.33959693e-01 7.54613042e-01
1.11089849e+00 -2.72605717e-01 8.18203688e-01 3.50055248e-01
2.40583032e-01 -1.16009510e+00 -2.16735542e-01 1.01496756e+00
7.81696022e-01 -4.59617555e-01 -2.28053421e-01 2.51707435e-01
-2.01400653e-01 1.64836109e+00 -2.72682216e-02 -2.72765547e-01
1.28645301e+00 2.52223432e-01 -1.76402092e-01 -2.89982617e-01
-1.02408719e+00 -3.83141004e-02 4.52039272e-01 3.77719223e-01
3.41521472e-01 -1.07252605e-01 2.48978764e-01 6.62629083e-02
-4.47536618e-01 5.89586124e-02 2.17569649e-01 1.14476955e+00
-1.10825365e-02 -1.17329919e+00 -1.93748578e-01 3.38264108e-01
-3.70081693e-01 7.05781728e-02 -6.62008822e-02 7.82510161e-01
3.84207666e-01 4.47705418e-01 4.08594906e-01 -1.70470893e-01
2.20996633e-01 7.73565173e-01 4.44328666e-01 -5.81780791e-01
-7.38209665e-01 -2.24860370e-01 -2.19882593e-01 -7.56768763e-01
-7.68351555e-01 -1.06986833e+00 -1.37705564e+00 -1.54887751e-01
-1.07240714e-01 1.87341854e-01 6.12181842e-01 1.44160867e+00
1.94365025e-01 8.99140000e-01 2.86504954e-01 -1.06857669e+00
-6.35488272e-01 -9.66109574e-01 -5.44094801e-01 1.48590297e-01
9.73809481e-01 -5.04679203e-01 -5.23715436e-01 -4.00022883e-03] | [6.947344779968262, 3.350409984588623] |
2f22a726-2f47-475d-9d92-da9e02da167c | higher-order-implicit-fairing-networks-for-3d | 2111.00950 | null | https://arxiv.org/abs/2111.00950v1 | https://arxiv.org/pdf/2111.00950v1.pdf | Higher-Order Implicit Fairing Networks for 3D Human Pose Estimation | Estimating a 3D human pose has proven to be a challenging task, primarily because of the complexity of the human body joints, occlusions, and variability in lighting conditions. In this paper, we introduce a higher-order graph convolutional framework with initial residual connections for 2D-to-3D pose estimation. Using multi-hop neighborhoods for node feature aggregation, our model is able to capture the long-range dependencies between body joints. Moreover, our approach leverages residual connections, which are integrated by design in our network architecture, ensuring that the learned feature representations retain important information from the initial features of the input layer as the network depth increases. Experiments and ablations studies conducted on two standard benchmarks demonstrate the effectiveness of our model, achieving superior performance over strong baseline methods for 3D human pose estimation. | ['A. Ben Hamza', 'Jianning Quan'] | 2021-11-01 | null | null | null | null | ['3d-pose-estimation'] | ['computer-vision'] | [-1.81386992e-01 3.31369489e-01 -1.86821222e-01 -3.17815661e-01
-1.16453923e-01 -3.48431736e-01 3.02103758e-01 -1.01265706e-01
-4.33605999e-01 4.06634480e-01 6.20203972e-01 4.01946276e-01
1.51336387e-01 -4.76337850e-01 -7.56948233e-01 -8.32811818e-02
-5.83467543e-01 4.87697989e-01 2.17742503e-01 -3.71095985e-01
-4.00829315e-01 6.29383385e-01 -1.24346733e+00 -1.66321740e-01
2.87832052e-01 9.57104325e-01 -4.24545854e-01 6.91231251e-01
4.89552051e-01 5.05556822e-01 -5.86779952e-01 -2.85057962e-01
4.99617010e-01 -1.18522421e-01 -6.75892472e-01 1.09445505e-01
7.80790806e-01 -6.43304408e-01 -8.19388390e-01 5.32163382e-01
6.37296319e-01 2.78034419e-01 5.55336177e-01 -1.22453570e+00
-1.46952361e-01 1.74463034e-01 -4.99432743e-01 -1.15290016e-01
4.61936563e-01 3.12051058e-01 1.23605597e+00 -7.71105707e-01
8.09253097e-01 1.49103820e+00 1.04121470e+00 6.20099127e-01
-1.28704000e+00 -4.36799824e-01 4.15905327e-01 -1.27077460e-01
-1.28919947e+00 -3.14098924e-01 7.85054505e-01 -3.45480710e-01
1.15264213e+00 -1.87444717e-01 1.08917332e+00 1.23227203e+00
4.54359680e-01 8.00380290e-01 4.43439633e-01 -4.72473316e-02
-1.52233660e-01 -4.48344618e-01 1.53945789e-01 1.02645683e+00
4.63047355e-01 1.28896385e-02 -7.35922515e-01 3.15935612e-02
1.05431247e+00 2.62435451e-02 -1.58537552e-01 -9.73222733e-01
-9.81085122e-01 6.63316309e-01 1.05051970e+00 -1.03192315e-01
-4.57847208e-01 7.60896385e-01 3.81742686e-01 8.04393888e-02
5.18740714e-01 3.88279140e-01 -6.26827180e-01 -6.44633770e-02
-6.34436309e-01 5.67592084e-01 8.60714555e-01 8.66683424e-01
4.66280431e-01 -1.41963929e-01 -3.21827352e-01 4.60219860e-01
4.70790058e-01 3.08987707e-01 -1.08538426e-01 -9.61717010e-01
5.06410241e-01 8.60807180e-01 -1.68868899e-02 -1.17265272e+00
-1.02640009e+00 -8.09013486e-01 -7.38535464e-01 2.89032370e-01
4.28231686e-01 -2.66279489e-01 -1.21008158e+00 2.05226350e+00
5.18638968e-01 -1.06249377e-01 -4.36224669e-01 1.07522511e+00
8.54805291e-01 8.35055783e-02 1.19109295e-01 5.25372088e-01
1.17265892e+00 -1.02042603e+00 -2.99697697e-01 -6.71555221e-01
3.46478015e-01 -3.15377176e-01 7.74320543e-01 -3.22431661e-02
-1.15357292e+00 -7.04895020e-01 -1.12485456e+00 -4.08407032e-01
-3.42192985e-02 8.64073783e-02 7.56138802e-01 4.71051335e-01
-1.03514123e+00 7.09546268e-01 -1.24555779e+00 -3.68239284e-01
5.39828539e-01 7.43172228e-01 -6.57761455e-01 -1.33457318e-01
-1.17457259e+00 9.02579427e-01 5.94107658e-02 4.83600199e-01
-7.84722984e-01 -5.90203643e-01 -1.17992568e+00 7.66311139e-02
3.91999990e-01 -1.19945025e+00 8.59414756e-01 -1.84990451e-01
-1.34522390e+00 7.05171585e-01 -4.75544818e-02 -4.16084558e-01
8.59355032e-01 -8.28734398e-01 2.46055737e-01 2.63599366e-01
-2.37515584e-01 9.40783560e-01 8.71962190e-01 -8.06923032e-01
-9.10619423e-02 -6.42264009e-01 2.31382236e-01 4.13021922e-01
-2.84805357e-01 -5.67005038e-01 -9.50320601e-01 -5.56475341e-01
2.40137711e-01 -1.28957260e+00 -5.23543179e-01 4.12912011e-01
-4.23773974e-01 -1.86933488e-01 4.61567074e-01 -7.52971947e-01
9.97110069e-01 -1.99855340e+00 6.33578718e-01 5.14324188e-01
7.02465475e-01 -1.66768342e-01 -2.44232655e-01 2.02112556e-01
2.25781918e-01 -1.56100571e-01 9.90234315e-02 -6.77061319e-01
1.38146743e-01 1.84764162e-01 3.64450961e-01 6.95480645e-01
4.48003888e-01 1.38868344e+00 -5.95175207e-01 -4.22343820e-01
2.85360515e-01 9.45141435e-01 -8.44959199e-01 1.19651556e-01
-5.62176406e-02 5.15109599e-01 -4.06153917e-01 4.82613087e-01
2.33381063e-01 -4.20354396e-01 1.56452671e-01 -4.87906277e-01
6.41254961e-01 4.00477469e-01 -1.05893707e+00 2.21817994e+00
-4.17360097e-01 3.04291189e-01 4.65568490e-02 -4.31337625e-01
7.09453702e-01 8.81668702e-02 6.84841931e-01 -3.65729660e-01
3.31828982e-01 -6.60505444e-02 5.91154024e-02 -2.36120187e-02
2.60673523e-01 4.01007384e-01 -2.07740426e-01 2.43699715e-01
1.46556243e-01 -9.22851115e-02 1.13773890e-01 2.70316988e-01
1.27697909e+00 4.97678876e-01 1.24615163e-01 -1.13877624e-01
1.32383376e-01 -2.50616521e-01 5.46103120e-01 5.06831944e-01
-3.67824376e-01 7.02898741e-01 6.60197556e-01 -7.30948150e-01
-1.00456595e+00 -1.15534842e+00 2.26958290e-01 9.68467534e-01
1.56483203e-02 -6.29313052e-01 -6.10141516e-01 -8.46277952e-01
3.34820122e-01 -1.47911340e-01 -9.45303380e-01 -4.70294118e-01
-9.63448584e-01 -3.70464504e-01 5.65166354e-01 8.80101323e-01
3.83601844e-01 -6.86083615e-01 -9.05227661e-01 2.87035763e-01
2.28110868e-02 -1.34187567e+00 -5.09460330e-01 2.50089914e-01
-8.72835636e-01 -1.18986833e+00 -6.83315337e-01 -5.97838402e-01
7.22897589e-01 -1.06009826e-01 1.40996456e+00 2.57320076e-01
-5.29642820e-01 4.21469331e-01 -3.62758413e-02 -6.50095567e-02
1.96600437e-01 5.89868605e-01 5.00304028e-02 -4.91259724e-01
3.26975793e-01 -6.25081003e-01 -8.09463620e-01 1.96640432e-01
-4.09629107e-01 -1.67053789e-01 5.93871415e-01 8.41041446e-01
3.73371959e-01 -2.04764336e-01 1.28655612e-01 -6.05930209e-01
4.48416352e-01 -7.16498122e-02 -3.11661154e-01 -5.26482612e-02
-1.94555312e-01 2.47465447e-01 2.28065565e-01 -2.69856811e-01
-5.19108474e-01 3.71892720e-01 -1.25489056e-01 -4.52090591e-01
-5.72827309e-02 3.17178994e-01 -1.17752597e-01 -2.78177887e-01
6.14588082e-01 -4.13781822e-01 3.01146954e-01 -4.44524169e-01
3.52857053e-01 -7.41898194e-02 6.28713071e-01 -5.20105064e-01
9.91325378e-01 3.56979132e-01 4.64757949e-01 -5.26699483e-01
-9.09280300e-01 -2.41723359e-01 -1.07665694e+00 -2.38323182e-01
9.21617866e-01 -1.22692323e+00 -8.64416599e-01 4.43211317e-01
-1.12477207e+00 -2.56822973e-01 -1.67261213e-01 5.23076773e-01
-3.58145475e-01 1.76456168e-01 -9.26250339e-01 -5.67624211e-01
-4.12899435e-01 -1.11332142e+00 1.35427988e+00 -9.23612043e-02
-7.86305964e-01 -9.59461153e-01 -1.85560323e-02 1.27353072e-01
2.47851998e-01 6.90011084e-01 7.64752448e-01 -3.12592924e-01
-3.70584011e-01 -4.94643986e-01 -8.67617950e-02 2.02797458e-01
1.73223093e-02 -3.25362951e-01 -7.86679387e-01 -6.59824491e-01
-6.06309831e-01 -5.75787365e-01 1.00687337e+00 4.76847619e-01
7.63564169e-01 -5.24783414e-03 -3.10122818e-01 8.25841844e-01
8.13131809e-01 -8.31958711e-01 2.83943862e-01 3.06213927e-03
1.23722732e+00 6.44257963e-01 2.22661734e-01 3.67034644e-01
5.32983482e-01 6.47531748e-01 5.84167123e-01 -3.01883638e-01
-4.40368474e-01 -4.55416441e-01 1.33495450e-01 4.29525882e-01
-3.31617177e-01 1.35517448e-01 -8.40631187e-01 3.30366403e-01
-1.75252783e+00 -4.69417959e-01 8.72840509e-02 2.08224154e+00
5.64121544e-01 5.74676633e-01 4.45667773e-01 -2.57700346e-02
3.69865149e-01 4.20715719e-01 -8.43243062e-01 6.97327927e-02
1.34371370e-01 3.57316136e-01 5.64732373e-01 4.09378678e-01
-1.23795521e+00 1.01425362e+00 6.88628721e+00 4.01560925e-02
-8.41849983e-01 -3.24434251e-01 3.01893473e-01 -5.24031401e-01
-4.18577120e-02 -3.40756744e-01 -8.38003099e-01 -6.32255673e-02
3.66579413e-01 3.80399883e-01 2.51553327e-01 5.96533000e-01
-1.34548053e-01 2.42474422e-01 -1.35420692e+00 8.85098696e-01
1.50777763e-02 -7.56921470e-01 -1.80299029e-01 2.62565225e-01
6.30515099e-01 2.49841601e-01 5.64772524e-02 1.95468619e-01
4.64941233e-01 -1.24714231e+00 5.92912436e-01 2.32940152e-01
5.91760933e-01 -9.43318307e-01 6.40371442e-01 2.43312512e-02
-1.36350751e+00 1.10370494e-01 -2.72475511e-01 -2.98998982e-01
1.48299694e-01 6.22035205e-01 -8.05172563e-01 2.71606892e-01
7.70469129e-01 9.64071929e-01 -8.12364578e-01 8.15018237e-01
-5.82137406e-01 2.20331401e-01 -7.25423813e-01 4.98213293e-03
1.84322372e-01 3.36983353e-01 4.42956716e-01 8.44298303e-01
-1.60388246e-01 -2.74058789e-01 5.11367440e-01 7.05396056e-01
-3.20470512e-01 -1.13849968e-01 -5.75817585e-01 5.60740121e-02
1.98475286e-01 1.22399116e+00 -5.87253392e-01 5.74089214e-02
-2.48904243e-01 9.89294171e-01 6.60144925e-01 3.68198723e-01
-7.72844374e-01 -2.54986197e-01 9.74616647e-01 2.44723544e-01
4.13317412e-01 -7.66303182e-01 -2.61895865e-01 -1.05682206e+00
3.96040082e-01 -7.98565149e-01 4.23652917e-01 -2.18804821e-01
-1.14745998e+00 5.35266936e-01 -7.97467586e-03 -7.81641066e-01
-4.86942053e-01 -6.00542784e-01 -2.29787663e-01 7.21697748e-01
-1.27233565e+00 -1.47422278e+00 -4.95801628e-01 6.59571826e-01
1.09109171e-01 3.23557943e-01 6.08694732e-01 2.15893492e-01
-4.66477096e-01 7.41931379e-01 -7.37202764e-01 5.22555470e-01
5.89012623e-01 -1.02960730e+00 1.11529982e+00 7.55937338e-01
1.40600458e-01 7.85607517e-01 5.43890059e-01 -7.93017805e-01
-1.45331681e+00 -8.38192999e-01 7.34485388e-01 -5.87231100e-01
4.53813732e-01 -8.58191490e-01 -6.06598437e-01 8.16865683e-01
-2.88052648e-01 3.84591162e-01 4.31294441e-01 6.30854070e-01
-6.40315175e-01 8.05926770e-02 -8.19537878e-01 5.87490201e-01
1.62260401e+00 -4.57611352e-01 -4.68333811e-01 1.61067069e-01
7.64247179e-01 -8.14469516e-01 -9.58142519e-01 6.58232093e-01
1.10125089e+00 -6.08509004e-01 1.29343915e+00 -7.65222132e-01
3.71318012e-01 -1.07475258e-01 9.07810256e-02 -1.10883212e+00
-5.39046884e-01 -5.51605403e-01 -4.62842822e-01 5.07229567e-01
3.59285742e-01 -2.27757543e-01 1.22404456e+00 8.20838928e-01
1.89542025e-01 -8.97245824e-01 -7.77503908e-01 -5.80036640e-01
-1.61210634e-02 -2.04542935e-01 4.23713028e-01 3.38371545e-01
-2.88292110e-01 5.96205413e-01 -7.16222703e-01 1.63165957e-01
7.84778714e-01 -1.15146801e-01 1.14548373e+00 -1.50543654e+00
-5.02000213e-01 -2.12745056e-01 -8.32760632e-01 -1.38027012e+00
2.67553359e-01 -6.10508323e-01 1.45248786e-01 -1.46997690e+00
-5.63734733e-02 -2.02516302e-01 -3.08269024e-01 6.81149065e-01
-3.37719023e-01 6.12778366e-01 3.08575362e-01 -6.45903312e-03
-6.00146770e-01 5.17478943e-01 1.36429155e+00 -9.59615558e-02
-3.49342406e-01 8.18912219e-03 -3.07812661e-01 7.65001476e-01
5.25748432e-01 -3.43159348e-01 -3.54299992e-01 -7.94013023e-01
3.22993457e-01 -3.46786320e-01 7.26702869e-01 -1.26530647e+00
1.82111904e-01 3.03255081e-01 1.01734698e+00 -6.24857545e-01
7.31581628e-01 -8.03100944e-01 -5.50254434e-02 8.21472943e-01
-3.92491996e-01 2.23496303e-01 8.83631036e-02 5.83916306e-01
6.12866953e-02 4.50554132e-01 5.88896811e-01 -1.80859327e-01
-6.21220410e-01 6.61696017e-01 2.28073001e-01 2.60869712e-01
6.43349409e-01 -1.11620665e-01 1.88542664e-01 -4.70823556e-01
-8.22558165e-01 3.95871609e-01 5.44532061e-01 6.46475852e-01
5.42155325e-01 -1.34974742e+00 -5.66414595e-01 1.91976771e-01
2.02686992e-02 2.89874613e-01 4.91730198e-02 8.36217165e-01
-6.20841920e-01 1.99673742e-01 -3.04891229e-01 -8.04503560e-01
-1.19880879e+00 1.17047183e-01 3.12888771e-01 -5.43699205e-01
-9.46482539e-01 9.39862728e-01 7.17477798e-02 -5.81123114e-01
6.47152662e-01 -3.54366928e-01 3.20458412e-02 -1.70333892e-01
1.34445950e-01 2.51566052e-01 -9.17305611e-03 -7.04970062e-01
-6.58463538e-01 7.67244279e-01 -7.26273432e-02 1.18394308e-01
1.42775404e+00 -3.33801433e-02 1.54522792e-01 1.65336177e-01
1.50535691e+00 -1.27167180e-01 -1.58466852e+00 -1.68249771e-01
-2.11790681e-01 -2.95242369e-01 -8.77101943e-02 -5.50878048e-01
-1.25777972e+00 8.26498806e-01 3.57944995e-01 -4.52618450e-01
7.70017326e-01 1.18185664e-02 9.96772110e-01 6.16291165e-01
3.83403480e-01 -1.04978871e+00 2.35990688e-01 5.87483227e-01
7.60092258e-01 -1.13414395e+00 3.64605904e-01 -4.76327777e-01
-2.80493200e-01 9.67884183e-01 8.34346056e-01 -6.00560308e-01
6.45746946e-01 1.58301666e-01 4.79643941e-02 -3.50460887e-01
-6.34753644e-01 -3.22250009e-01 7.31311679e-01 5.40523112e-01
5.39659023e-01 -2.11967379e-01 -1.26467310e-02 2.72468418e-01
-3.69085938e-01 -1.81900531e-01 -1.54990166e-01 1.00413287e+00
-2.73577452e-01 -9.54665959e-01 -5.21970429e-02 2.69185394e-01
-3.45279127e-01 1.62406519e-01 -7.44742155e-01 1.05856609e+00
-1.03831276e-01 5.62525272e-01 2.87089078e-03 -5.20566702e-01
6.81905746e-01 -4.78813145e-03 7.80223727e-01 -5.86813569e-01
-6.53689921e-01 1.15178406e-01 1.81111977e-01 -1.15323007e+00
-3.60914409e-01 -3.79106253e-01 -1.29391253e+00 -1.52248308e-01
-7.20248148e-02 -2.43218973e-01 5.38027823e-01 8.92102659e-01
5.49750507e-01 7.55508184e-01 2.19934702e-01 -1.13055348e+00
-5.88138461e-01 -8.57888877e-01 -3.57318699e-01 6.15845919e-01
4.44847196e-01 -1.00576532e+00 -1.74108416e-01 -4.13378626e-01] | [6.99744987487793, -0.7798519730567932] |
c571dff5-1900-49ba-9ca3-37f4fad3f42f | better-low-resource-entity-recognition | 2305.13582 | null | https://arxiv.org/abs/2305.13582v2 | https://arxiv.org/pdf/2305.13582v2.pdf | Better Low-Resource Entity Recognition Through Translation and Annotation Fusion | Pre-trained multilingual language models have enabled significant advancements in cross-lingual transfer. However, these models often exhibit a performance disparity when transferring from high-resource languages to low-resource languages, especially for languages that are underrepresented or not in the pre-training data. Motivated by the superior performance of these models on high-resource languages compared to low-resource languages, we introduce a Translation-and-fusion framework, which translates low-resource language text into a high-resource language for annotation using fully supervised models before fusing the annotations back into the low-resource language. Based on this framework, we present TransFusion, a model trained to fuse predictions from a high-resource language to make robust predictions on low-resource languages. We evaluate our methods on two low-resource named entity recognition (NER) datasets, MasakhaNER2.0 and LORELEI NER, covering 25 languages, and show consistent improvement up to +16 F$_1$ over English fine-tuning systems, achieving state-of-the-art performance compared to Translate-train systems. Our analysis depicts the unique advantages of the TransFusion method which is robust to translation errors and source language prediction errors, and complimentary to adapted multilingual language models. | ['Alan Ritter', 'Vedaant Shah', 'Yang Chen'] | 2023-05-23 | null | null | null | null | ['named-entity-recognition-ner', 'low-resource-named-entity-recognition', 'cross-lingual-transfer'] | ['natural-language-processing', 'natural-language-processing', 'natural-language-processing'] | [-1.74488306e-01 -2.86185473e-01 -5.19083798e-01 -5.46088338e-01
-1.57360828e+00 -8.49460483e-01 4.72968459e-01 -5.76122999e-02
-1.05423617e+00 1.13413632e+00 3.93548280e-01 -5.32274246e-01
5.32783687e-01 -5.60723066e-01 -9.28003371e-01 4.15900238e-02
3.97458076e-01 8.54464054e-01 -1.78074449e-01 -5.05229235e-01
-4.82081085e-01 1.32872164e-01 -9.05196130e-01 5.39115608e-01
1.20629907e+00 2.94876337e-01 2.44339958e-01 2.28691399e-01
-3.99060339e-01 5.55318773e-01 -3.81978810e-01 -8.25189173e-01
2.31668070e-01 -2.11654052e-01 -9.60217476e-01 -4.88252908e-01
3.17387372e-01 1.74111575e-01 7.40440041e-02 9.09430861e-01
5.97029924e-01 -1.50212914e-01 5.45489192e-01 -7.00573921e-01
-1.09458947e+00 1.04839146e+00 -2.82553017e-01 1.24755614e-01
3.99213731e-01 -1.14980102e-01 1.01807475e+00 -1.24651015e+00
1.02870166e+00 1.16394508e+00 1.05228508e+00 5.60719609e-01
-1.18318403e+00 -1.01009500e+00 1.33438967e-02 -2.14770913e-01
-1.82899344e+00 -8.94936144e-01 1.30452484e-01 -3.47334862e-01
1.68042064e+00 -1.54812649e-01 2.83558909e-02 9.96924281e-01
1.68200478e-01 6.51567340e-01 1.31647527e+00 -7.52383113e-01
-3.90325278e-01 5.58677614e-01 -3.44384134e-01 6.05278075e-01
7.51530528e-02 1.87909100e-02 -6.63296103e-01 1.14579931e-01
2.48805463e-01 -3.59954506e-01 -8.63894597e-02 3.99533808e-01
-1.48271835e+00 6.50020540e-01 2.83110142e-01 5.81325114e-01
-3.29569161e-01 -2.78407663e-01 5.96310258e-01 4.83717322e-01
1.01734221e+00 5.43005228e-01 -1.22174299e+00 -7.89916664e-02
-9.85722363e-01 -2.72590995e-01 8.64655674e-01 1.31752801e+00
1.05567288e+00 1.54679462e-01 -4.39363606e-02 1.31484687e+00
1.26226038e-01 8.75326335e-01 6.64146006e-01 -2.17788458e-01
8.99176359e-01 5.27840376e-01 -4.72720563e-02 -1.56710014e-01
-2.83954889e-01 -4.11677003e-01 -6.64265752e-01 -4.17527139e-01
1.94145650e-01 -4.44146276e-01 -8.08655322e-01 1.92479646e+00
2.55121365e-02 -2.63790727e-01 6.95273459e-01 4.32659000e-01
8.25530112e-01 6.83478653e-01 5.97802579e-01 2.22028028e-02
1.39111459e+00 -1.03170681e+00 -4.99159396e-01 -5.08541226e-01
1.21732843e+00 -1.12968206e+00 1.27602732e+00 -2.01793343e-01
-8.84586692e-01 -6.15501344e-01 -6.76466227e-01 -3.96567523e-01
-9.65109050e-01 7.16472149e-01 4.21626419e-01 5.93840241e-01
-1.21290004e+00 2.54150301e-01 -6.16880298e-01 -7.53621578e-01
-5.01502771e-03 2.78780788e-01 -8.25577259e-01 -2.52869666e-01
-1.53789163e+00 1.36360157e+00 7.19538271e-01 -1.49124116e-01
-7.03797698e-01 -1.01413786e+00 -1.03196108e+00 -2.98649490e-01
5.90420440e-02 -3.54276150e-01 1.08314419e+00 -8.94164383e-01
-1.30742490e+00 1.31069696e+00 -2.75854200e-01 -2.72237688e-01
4.08096701e-01 -3.95746857e-01 -9.70728576e-01 -5.49014747e-01
5.85243225e-01 6.96517885e-01 1.79776214e-02 -8.25659394e-01
-9.00086641e-01 -6.66269287e-02 -3.71099353e-01 3.93915385e-01
-3.53068352e-01 7.45386600e-01 -5.55590391e-01 -6.04923904e-01
-5.66183209e-01 -1.01311064e+00 -1.87507316e-01 -7.52444446e-01
-2.56836444e-01 -2.61741698e-01 3.62450004e-01 -1.07495892e+00
1.21124053e+00 -1.85921025e+00 -1.37644291e-01 -1.83543935e-01
-4.97787476e-01 4.59379435e-01 -4.11437631e-01 5.70873976e-01
-8.56679901e-02 3.96035910e-01 -7.46279433e-02 -4.56478328e-01
-1.15634143e-01 2.99523592e-01 -2.43005291e-01 2.74700612e-01
5.00018954e-01 9.57129180e-01 -9.77001011e-01 -3.98377091e-01
1.42930904e-02 6.63301945e-01 -4.39940035e-01 2.50659078e-01
9.84060243e-02 6.64705038e-01 -1.63652673e-01 7.54947603e-01
2.79926270e-01 1.46595418e-01 2.22441912e-01 -2.05146462e-01
-5.26309669e-01 7.94051290e-01 -5.54426670e-01 1.98541844e+00
-1.18440509e+00 2.48529926e-01 -2.68007405e-02 -4.71902370e-01
9.78658915e-01 5.66260278e-01 2.72929639e-01 -7.42777050e-01
-1.18435822e-01 1.02181768e+00 -2.09970936e-01 -1.56861976e-01
7.58592486e-01 -4.02694643e-01 -5.11833072e-01 3.78045619e-01
6.25362635e-01 -1.70779347e-01 3.62872034e-01 -1.51250198e-01
7.30592430e-01 5.80207407e-01 4.84960943e-01 -4.99984950e-01
6.18602574e-01 3.78152460e-01 7.57446349e-01 3.18835318e-01
-2.17335775e-01 2.37896144e-01 -3.63940150e-01 -2.23687351e-01
-1.08894777e+00 -8.08767259e-01 -3.39726776e-01 1.74294686e+00
-4.50720400e-01 -4.98374969e-01 -5.56716919e-01 -9.59562480e-01
-9.52308103e-02 8.36413741e-01 -3.93053770e-01 1.01155572e-01
-7.51163661e-01 -8.64626348e-01 1.17448199e+00 5.55389583e-01
4.20850992e-01 -1.11240852e+00 2.56287158e-01 4.69881862e-01
-5.27644217e-01 -1.48092484e+00 -7.34256387e-01 3.77852142e-01
-3.35436672e-01 -5.12733102e-01 -7.96701074e-01 -1.09319735e+00
4.58515614e-01 -1.38044462e-01 1.66254091e+00 -3.80051941e-01
9.75042358e-02 1.62107453e-01 -2.60798454e-01 -5.39698362e-01
-8.33482027e-01 6.99844956e-01 4.53188598e-01 -3.42999309e-01
8.80998611e-01 2.41824277e-02 5.77807687e-02 4.39242274e-01
-5.31613767e-01 -9.02456697e-03 6.78805232e-01 8.23631406e-01
9.45477247e-01 -3.66496861e-01 7.98008084e-01 -1.02547646e+00
2.85638303e-01 -6.33892655e-01 -4.63409305e-01 6.98052824e-01
-5.12445271e-01 1.25952169e-01 9.18240428e-01 -2.66496986e-01
-1.34059370e+00 1.04320690e-01 -4.16333944e-01 -8.17632526e-02
-1.14242256e-01 8.27130079e-01 -3.18019331e-01 1.03850149e-01
7.18500376e-01 4.37821113e-02 -5.91190875e-01 -8.18785906e-01
6.32372856e-01 1.02888834e+00 5.71173728e-01 -8.67730796e-01
6.21713579e-01 -7.28493929e-02 -6.89251065e-01 -6.73676968e-01
-1.06536365e+00 -6.52435422e-01 -1.11457551e+00 2.49477357e-01
9.86953139e-01 -1.59340930e+00 1.11814559e-01 1.56628713e-01
-1.07195675e+00 -3.94130796e-01 -2.77490199e-01 7.10431695e-01
-2.46543124e-01 -2.45999068e-01 -9.35086370e-01 -4.30231839e-01
-6.04318678e-01 -1.19851315e+00 1.14429045e+00 -1.02819383e-01
-2.61236340e-01 -1.46013546e+00 5.02604842e-01 2.79819399e-01
5.34847081e-01 -2.57521451e-01 8.01597536e-01 -1.14069343e+00
9.30005461e-02 -9.64985937e-02 -1.70901522e-01 4.16409731e-01
3.00263762e-01 -2.24612743e-01 -7.96478927e-01 -4.83234376e-01
-5.64344764e-01 -6.99553370e-01 4.48980242e-01 -2.15932161e-01
3.77217323e-01 -1.42050356e-01 -3.34891617e-01 7.37227261e-01
1.55245399e+00 -2.30167210e-01 1.79351985e-01 4.71464932e-01
8.29759359e-01 5.60596645e-01 5.36513090e-01 -1.80820882e-01
9.16711390e-01 6.61223054e-01 -4.82759833e-01 -4.99671221e-01
-4.36457247e-01 -6.52110517e-01 8.19415510e-01 1.73834801e+00
-1.72964577e-02 -4.96261045e-02 -1.38598847e+00 8.10657740e-01
-1.41286516e+00 -4.65897918e-01 8.22978839e-02 2.19541836e+00
1.49230683e+00 -2.76493192e-01 -1.14135720e-01 -8.91383767e-01
7.04838872e-01 -1.92973360e-01 -2.68671721e-01 -5.59368491e-01
-6.14357233e-01 4.28925395e-01 8.67215395e-01 5.00084579e-01
-1.08038199e+00 1.77228343e+00 7.00034189e+00 9.56106603e-01
-1.21872640e+00 5.43979883e-01 4.47461814e-01 3.03386033e-01
-3.51861209e-01 3.74500342e-02 -1.43638861e+00 2.31873006e-01
1.54313397e+00 -4.46431518e-01 3.70597214e-01 7.59169161e-01
-8.84494334e-02 5.26002049e-01 -1.13891983e+00 5.91953635e-01
1.67692378e-01 -1.03202736e+00 2.14225322e-01 -1.63455263e-01
9.91994202e-01 9.56562698e-01 -1.78517312e-01 1.01768553e+00
9.84078765e-01 -1.08827698e+00 6.56527162e-01 1.81611225e-01
1.55302536e+00 -7.72855759e-01 9.95437741e-01 3.63354117e-01
-1.39830124e+00 5.34790874e-01 -4.98279512e-01 2.56356239e-01
3.08693230e-01 2.66030014e-01 -9.85549033e-01 9.28566515e-01
7.97175109e-01 7.92915881e-01 -4.98815477e-01 3.59450936e-01
-3.21414202e-01 6.77024305e-01 -3.38099062e-01 3.25516790e-01
2.81218082e-01 -1.30995184e-01 1.14541210e-01 1.95371854e+00
4.86140996e-01 -3.31826836e-01 5.18478572e-01 5.81507444e-01
-6.72502160e-01 9.25862968e-01 -7.05917060e-01 -2.52783597e-01
4.54880595e-01 1.24542141e+00 -1.28586441e-01 -5.08323491e-01
-9.28756237e-01 9.77243483e-01 8.07428122e-01 2.72550285e-01
-5.34565747e-01 -3.59321326e-01 5.81128120e-01 -8.90827924e-02
-2.62285732e-02 -1.65360808e-01 -4.94282693e-02 -1.60665643e+00
-2.50949949e-01 -1.03650355e+00 5.57230294e-01 -5.06894171e-01
-1.67546988e+00 1.11396861e+00 -2.54672378e-01 -1.24258280e+00
-4.38366741e-01 -7.02591956e-01 -2.87449849e-03 1.54407704e+00
-1.90867472e+00 -1.88179207e+00 4.26807135e-01 6.83143914e-01
5.39912999e-01 -5.48392951e-01 1.39335239e+00 7.45788693e-01
-4.54001158e-01 9.88009334e-01 2.29943410e-01 5.62943339e-01
1.40884566e+00 -1.08797836e+00 5.15494287e-01 9.63276327e-01
3.23110163e-01 7.99251914e-01 1.76653519e-01 -8.44940782e-01
-1.13496554e+00 -1.57099259e+00 1.77998734e+00 -8.61925721e-01
9.88503695e-01 -5.01004219e-01 -9.64746714e-01 1.20950174e+00
4.71402973e-01 6.20360598e-02 1.29177809e+00 5.65938771e-01
-7.06524611e-01 1.39212072e-01 -1.07200122e+00 5.42964876e-01
8.89823973e-01 -1.05013967e+00 -7.41092384e-01 3.24137300e-01
7.56379008e-01 -3.75303626e-01 -1.31979787e+00 4.72785830e-01
3.92314643e-01 -1.05599232e-01 6.77988946e-01 -8.49975824e-01
2.13730693e-01 -2.25083053e-01 -4.63038117e-01 -1.75896275e+00
-2.48308495e-01 -4.47505265e-01 6.23795569e-01 1.64942765e+00
1.16845226e+00 -6.78380847e-01 6.71472251e-02 3.32468927e-01
-3.56971920e-01 -1.55637026e-01 -1.00605619e+00 -9.07782972e-01
7.30078638e-01 -4.51979160e-01 5.63709140e-01 1.46767521e+00
1.37127023e-02 6.78942502e-01 -3.54708582e-01 4.00299579e-02
2.67170042e-01 -2.65183747e-01 6.37627065e-01 -7.96746612e-01
-1.88146979e-01 -1.34794161e-01 3.57153290e-03 -7.27880716e-01
8.08074355e-01 -1.59471357e+00 3.25631738e-01 -1.27569151e+00
2.53205508e-01 -9.58568811e-01 -5.23414314e-01 1.09514630e+00
-3.83725017e-01 6.78525686e-01 5.65805957e-02 4.71629769e-01
-4.62960750e-01 4.60121185e-01 6.49866760e-01 2.93514635e-02
-1.62310138e-01 -3.54271054e-01 -6.75670564e-01 7.03545630e-01
5.05451679e-01 -5.43792665e-01 2.14775264e-01 -1.04241538e+00
2.23474681e-01 -2.47259229e-01 -5.43710291e-01 -6.93171859e-01
-5.66076953e-03 -2.24436279e-02 2.98860699e-01 -3.06942254e-01
-9.02903825e-02 -6.53697133e-01 9.94111300e-02 1.54288188e-01
-3.33688319e-01 4.79407907e-01 5.83202779e-01 -7.82795325e-02
-4.49607044e-01 1.61282551e-02 6.81287766e-01 -3.07209194e-01
-9.26120162e-01 3.18442345e-01 -2.82924116e-01 4.51294452e-01
5.11586130e-01 2.80300975e-01 -3.87722820e-01 1.41485864e-02
-5.79141200e-01 1.12143204e-01 5.92516601e-01 8.54238153e-01
-2.18859494e-01 -1.40186679e+00 -1.30241120e+00 2.17056721e-01
5.87786138e-01 -4.41379666e-01 -1.00690134e-01 6.75883353e-01
-3.60399872e-01 7.01108932e-01 -2.20158920e-01 -4.13406610e-01
-8.44191909e-01 3.55414689e-01 3.01689714e-01 -6.54957294e-01
-1.77703276e-01 7.41563797e-01 1.07394829e-02 -1.47434103e+00
-3.86102587e-01 1.01584308e-02 -6.25594333e-02 -5.42171299e-02
2.94113129e-01 7.02139065e-02 4.27126199e-01 -1.44619918e+00
-6.79023266e-01 5.63570142e-01 -1.79952636e-01 -2.93370873e-01
1.21079516e+00 -3.04180592e-01 -2.43296862e-01 5.94150603e-01
1.23409259e+00 6.34894788e-01 -5.57460904e-01 -7.13255048e-01
3.06684434e-01 -3.94186452e-02 -1.39310881e-01 -1.22268558e+00
-7.38301218e-01 6.89465106e-01 3.15847129e-01 -6.16356671e-01
8.53555083e-01 7.19768330e-02 7.47996747e-01 3.12312841e-01
7.76199937e-01 -1.30117154e+00 -6.58711195e-01 1.13335872e+00
6.13800168e-01 -1.47392368e+00 -2.63267249e-01 -2.38290757e-01
-8.03354442e-01 7.19761729e-01 6.04911864e-01 1.93114787e-01
6.69234872e-01 5.06591558e-01 6.90615833e-01 2.48554438e-01
-6.74627662e-01 -3.58061254e-01 5.25602102e-01 6.00548267e-01
1.25182545e+00 3.69479090e-01 -2.51488477e-01 7.17070222e-01
-5.04710257e-01 1.21478485e-02 6.19434975e-02 7.46443510e-01
4.25053686e-02 -1.47306550e+00 -1.59603447e-01 9.07618850e-02
-1.00254703e+00 -7.87840486e-01 -2.02378199e-01 8.95509005e-01
3.33273888e-01 8.27664554e-01 -3.42695005e-02 -1.72175825e-01
5.23852825e-01 4.74599630e-01 1.09683506e-01 -1.02533484e+00
-9.67778742e-01 6.47683069e-02 5.19160211e-01 -3.46084416e-01
-4.30949658e-01 -5.96245766e-01 -1.04846966e+00 -1.46049857e-01
-2.56402671e-01 3.92009169e-01 7.52067327e-01 9.90166068e-01
5.01132607e-01 1.97752669e-01 2.81143576e-01 -5.37230730e-01
-3.10334414e-01 -1.12200749e+00 -2.23517537e-01 4.35902208e-01
-1.58587262e-01 -2.55976170e-01 -1.43863216e-01 2.01421604e-01] | [10.708081245422363, 9.934703826904297] |
2dc7e91c-f88b-4f5b-82b7-3e04223dea9c | robust-multi-agent-coordination-via | 2305.05909 | null | https://arxiv.org/abs/2305.05909v1 | https://arxiv.org/pdf/2305.05909v1.pdf | Robust multi-agent coordination via evolutionary generation of auxiliary adversarial attackers | Cooperative multi-agent reinforcement learning (CMARL) has shown to be promising for many real-world applications. Previous works mainly focus on improving coordination ability via solving MARL-specific challenges (e.g., non-stationarity, credit assignment, scalability), but ignore the policy perturbation issue when testing in a different environment. This issue hasn't been considered in problem formulation or efficient algorithm design. To address this issue, we firstly model the problem as a limited policy adversary Dec-POMDP (LPA-Dec-POMDP), where some coordinators from a team might accidentally and unpredictably encounter a limited number of malicious action attacks, but the regular coordinators still strive for the intended goal. Then, we propose Robust Multi-Agent Coordination via Evolutionary Generation of Auxiliary Adversarial Attackers (ROMANCE), which enables the trained policy to encounter diversified and strong auxiliary adversarial attacks during training, thus achieving high robustness under various policy perturbations. Concretely, to avoid the ego-system overfitting to a specific attacker, we maintain a set of attackers, which is optimized to guarantee the attackers high attacking quality and behavior diversity. The goal of quality is to minimize the ego-system coordination effect, and a novel diversity regularizer based on sparse action is applied to diversify the behaviors among attackers. The ego-system is then paired with a population of attackers selected from the maintained attacker set, and alternately trained against the constantly evolving attackers. Extensive experiments on multiple scenarios from SMAC indicate our ROMANCE provides comparable or better robustness and generalization ability than other baselines. | ['Yang Yu', 'Chao Qian', 'Li-He Li', 'Cong Guan', 'Feng Chen', 'Hao Yin', 'Ke Xue', 'Zi-Qian Zhang', 'Lei Yuan'] | 2023-05-10 | null | null | null | null | ['multi-agent-reinforcement-learning', 'smac-1', 'smac'] | ['methodology', 'playing-games', 'playing-games'] | [-1.23754285e-01 -1.19088024e-01 -8.26016143e-02 3.45695078e-01
-6.26777828e-01 -7.43195713e-01 5.46948791e-01 6.68896213e-02
-5.92748225e-01 8.54008198e-01 -1.00068383e-01 -1.04743935e-01
-3.49439532e-01 -8.47288370e-01 -5.09205580e-01 -1.28604937e+00
-5.96495569e-01 7.94767261e-01 2.60001302e-01 -5.00640988e-01
6.50738552e-02 3.40022981e-01 -1.08121181e+00 -2.18102887e-01
9.95287240e-01 5.30695617e-01 -1.05283791e-02 4.78293091e-01
4.83320773e-01 8.73830795e-01 -1.24755406e+00 -2.07379073e-01
5.62118053e-01 -4.20110047e-01 -3.73856366e-01 3.86835188e-02
-4.51914549e-01 -4.48259145e-01 -3.06293547e-01 1.16325891e+00
8.83464754e-01 4.24387634e-01 4.58867043e-01 -1.79069984e+00
-1.20155446e-01 9.80546176e-01 -8.31963122e-01 1.30587220e-01
4.27445382e-01 6.51595592e-01 7.77066231e-01 1.82231650e-01
2.63884336e-01 1.54908347e+00 2.61320502e-01 1.03106070e+00
-1.17596030e+00 -9.45825756e-01 5.28883159e-01 6.72922730e-02
-1.05773783e+00 3.48100066e-02 7.47371078e-01 1.02184109e-01
4.00963277e-01 2.32251614e-01 4.74599034e-01 1.72918165e+00
3.39196652e-01 7.56037354e-01 1.04010105e+00 1.39929261e-02
7.94532061e-01 -7.29022995e-02 -4.55260992e-01 3.20554972e-01
4.80283856e-01 4.75555152e-01 -4.30429913e-02 -7.95787573e-01
6.67346537e-01 -1.84959974e-02 -9.50289369e-02 -3.05066854e-01
-1.18383694e+00 9.30669665e-01 1.29826665e-01 -2.05168687e-02
-8.47245455e-01 -4.22445417e-04 5.26121140e-01 6.48731709e-01
-1.56385422e-01 8.40614498e-01 -5.23235440e-01 -2.17387557e-01
-1.86438367e-01 5.86244583e-01 8.34742010e-01 5.49829841e-01
3.22874904e-01 6.92097485e-01 -1.01647966e-01 4.52799380e-01
1.84325293e-01 5.89596272e-01 5.37002861e-01 -1.18291903e+00
6.98658586e-01 3.68655801e-01 2.68253446e-01 -9.41992342e-01
-4.48111624e-01 -4.56511170e-01 -7.94152200e-01 6.88091636e-01
2.65683800e-01 -8.82943571e-01 -4.75678861e-01 2.13296437e+00
6.55152261e-01 3.35791588e-01 4.92067814e-01 8.14809620e-01
5.89366592e-02 6.85037792e-01 5.16008064e-02 -6.29715085e-01
1.03347731e+00 -6.06259048e-01 -3.94224972e-01 -1.63115352e-01
3.42680544e-01 -2.35292003e-01 7.97479093e-01 4.83099520e-01
-9.15844262e-01 -3.22353415e-04 -9.69804347e-01 1.40423012e+00
9.64777842e-02 -6.54047430e-01 4.24868107e-01 7.65946805e-01
-4.30517077e-01 3.74598891e-01 -1.00874841e+00 1.38791064e-02
3.40494424e-01 6.70513034e-01 -9.91671458e-02 1.76319152e-01
-1.35588610e+00 6.20900750e-01 5.06712735e-01 -4.92550939e-01
-1.77610612e+00 -4.62334633e-01 -6.16656899e-01 -6.57191314e-03
1.00995040e+00 -6.55669034e-01 1.09748280e+00 -1.05450177e+00
-1.88563120e+00 -1.64018512e-01 7.85585344e-01 -5.49864829e-01
6.49560988e-01 2.90227741e-01 -3.46839219e-01 1.68955877e-01
1.12667844e-01 1.39457032e-01 1.04118288e+00 -1.62965393e+00
-7.71870077e-01 -1.86629370e-01 6.35391235e-01 7.49740720e-01
-5.55990398e-01 1.62039876e-01 1.63408533e-01 -8.35674345e-01
-3.43090922e-01 -1.14797711e+00 -7.03837395e-01 -8.56886089e-01
-3.20760965e-01 -2.50621475e-02 8.52783382e-01 -1.46521673e-01
1.04519999e+00 -2.00234509e+00 4.50627804e-01 4.69028324e-01
6.77838251e-02 3.74917001e-01 -5.19164503e-01 7.49878407e-01
3.12778242e-02 7.09440932e-02 -1.98524803e-01 -1.50045723e-01
1.08093150e-01 5.49761236e-01 -3.54792386e-01 4.84264970e-01
-8.77231807e-02 3.89629126e-01 -1.05886269e+00 -1.86321780e-01
-1.37166739e-01 -1.04104258e-01 -7.42079139e-01 5.18753409e-01
-4.27600920e-01 8.98186564e-01 -1.03006530e+00 7.59460747e-01
5.16062617e-01 2.37087116e-01 3.99888545e-01 5.42496204e-01
1.44277558e-01 -2.51519501e-01 -1.59838128e+00 9.61654186e-01
-2.79949367e-01 -4.12920564e-01 4.07199442e-01 -1.04884541e+00
6.84698701e-01 4.44484860e-01 9.00470793e-01 -5.86943269e-01
3.34500939e-01 -7.43487254e-02 6.13104045e-01 -2.41693065e-01
6.35648146e-02 -3.09749506e-02 -5.03854573e-01 8.84066582e-01
-2.72445917e-01 -2.80821212e-02 4.67236750e-02 1.85926318e-01
1.43727040e+00 -2.01042429e-01 3.46466660e-01 1.00495592e-01
6.51847363e-01 7.13165104e-02 1.09883618e+00 1.14583731e+00
-5.66877067e-01 1.72005109e-02 8.07676315e-01 -2.92582273e-01
-8.33166659e-01 -9.02704418e-01 5.22330105e-01 1.24877572e+00
4.07393128e-01 -1.83905274e-01 -6.64056063e-01 -1.09731472e+00
1.30209759e-01 6.93866491e-01 -6.30621314e-01 -4.88232344e-01
-7.61792839e-01 -1.13157856e+00 8.99628401e-01 -3.82310003e-02
6.00742996e-01 -1.35841107e+00 -6.92480326e-01 4.56489384e-01
8.50148574e-02 -7.68229365e-01 -4.20588762e-01 1.71797678e-01
-3.44174892e-01 -1.28766847e+00 -3.51967722e-01 -3.93573523e-01
5.71060359e-01 1.86912641e-01 5.75299561e-01 -1.29799629e-02
1.29792824e-01 6.70252860e-01 -5.83713114e-01 -3.75067353e-01
-8.53986800e-01 -8.78118277e-02 8.58127654e-01 7.16427118e-02
-1.06411003e-01 -8.04458797e-01 -4.18047071e-01 5.07308602e-01
-1.14956868e+00 -6.95138156e-01 4.86488432e-01 1.15257037e+00
1.44492030e-01 6.91079915e-01 8.83338094e-01 -4.98752087e-01
1.19232273e+00 -7.02808261e-01 -7.64915586e-01 1.61617383e-01
-3.14832151e-01 1.93054229e-02 1.27650106e+00 -1.18079567e+00
-7.71443725e-01 -1.81748062e-01 7.76631609e-02 -5.72264850e-01
-1.20314792e-01 1.68986037e-01 -5.01727939e-01 -1.87923685e-01
7.13177860e-01 3.75936866e-01 2.73094922e-01 9.88163650e-02
1.17163351e-02 3.26722711e-01 7.91699961e-02 -1.16508722e+00
1.34055209e+00 2.53347605e-01 6.11835495e-02 -3.73350531e-01
-2.09923863e-01 9.12168026e-02 1.50071576e-01 -1.72955543e-01
3.97689253e-01 -8.64510238e-01 -1.22801006e+00 6.11187875e-01
-8.06494832e-01 -4.60900545e-01 -3.31810415e-01 4.74513769e-01
-5.84949195e-01 4.48583573e-01 -4.42472607e-01 -9.61539924e-01
-2.03377321e-01 -1.39016902e+00 3.96514744e-01 4.38651592e-01
2.58474231e-01 -8.84297550e-01 4.91550118e-01 1.16932727e-01
3.98791492e-01 5.46235144e-01 6.84434712e-01 -1.10979438e+00
-2.71684617e-01 -9.08880830e-02 6.98940873e-01 2.72360593e-01
8.25796947e-02 -1.94115698e-01 -3.44401926e-01 -8.73575509e-01
3.02775085e-01 -5.53023875e-01 8.03685263e-02 7.78756738e-02
9.26814079e-01 -8.61223340e-01 -1.90307066e-01 4.33650106e-01
1.06311858e+00 7.44969308e-01 2.64210761e-01 8.08982551e-01
3.64148438e-01 3.79492015e-01 8.21322918e-01 9.92516935e-01
3.58787358e-01 5.16066551e-01 1.18834066e+00 3.43380839e-01
6.90383434e-01 -1.83443353e-01 9.40336764e-01 2.12133288e-01
-1.26069948e-01 -1.96991265e-01 -6.10437691e-01 7.14732930e-02
-2.02309155e+00 -1.29436195e+00 6.38987243e-01 2.29674459e+00
7.69299328e-01 3.14056963e-01 6.62941337e-01 -3.96317542e-02
7.54415572e-01 1.65415227e-01 -9.16303515e-01 -3.38528037e-01
-1.72804877e-01 -1.19585544e-01 5.61198056e-01 2.78000474e-01
-9.97202218e-01 9.39693928e-01 5.18549156e+00 9.77769613e-01
-1.00054669e+00 1.43103823e-01 3.86595398e-01 -2.30369240e-01
4.35084477e-02 1.51847720e-01 -6.85920477e-01 5.68087697e-01
7.28045523e-01 -3.75241727e-01 9.79916334e-01 8.47541809e-01
4.24129337e-01 1.82549596e-01 -6.37882352e-01 5.49002409e-01
-1.29026905e-01 -7.32200503e-01 -8.16033259e-02 1.85554847e-01
9.01539981e-01 -1.69403479e-01 1.58403292e-01 7.96111405e-01
1.12814593e+00 -7.43327379e-01 4.62435991e-01 -7.67397881e-03
2.57836372e-01 -1.32523251e+00 6.71347916e-01 8.50383878e-01
-9.21898603e-01 -9.11475122e-01 -2.20925480e-01 -3.74017097e-02
-4.62894216e-02 -2.64876753e-01 -5.36210060e-01 6.35166287e-01
4.05042797e-01 6.46342933e-02 -2.04137206e-01 7.46366382e-01
-3.47851217e-01 4.88756478e-01 -3.81231546e-01 -6.99756593e-02
6.02468848e-01 -1.79595053e-01 1.15631938e+00 3.82956684e-01
5.88571578e-02 1.70393020e-01 9.43134367e-01 3.81268710e-01
1.76929355e-01 -7.40730241e-02 -6.90672755e-01 1.39469057e-01
8.81498516e-01 1.32178092e+00 -3.81186575e-01 -9.38391834e-02
-6.18345067e-02 7.68527627e-01 2.37542942e-01 4.57995027e-01
-1.23009264e+00 -2.32781187e-01 9.88765597e-01 -2.78769463e-01
6.29673004e-02 -1.31487414e-01 1.93172321e-01 -1.02533829e+00
-2.56446242e-01 -1.82869232e+00 6.48725986e-01 6.78263605e-02
-1.49566710e+00 5.49811780e-01 -1.26894832e-01 -1.46246016e+00
-4.44295913e-01 -8.37830678e-02 -1.13789749e+00 3.81176531e-01
-1.06080270e+00 -9.73618090e-01 2.08751455e-01 1.04624391e+00
3.94562364e-01 -8.48096728e-01 6.93648458e-01 3.48764546e-02
-1.13034880e+00 8.59538257e-01 2.06442401e-01 -1.09619372e-01
5.76422393e-01 -1.15022862e+00 -7.43772760e-02 9.39059138e-01
-3.47502202e-01 5.37783921e-01 9.03614283e-01 -7.06990123e-01
-1.68469107e+00 -1.20892799e+00 -3.78214598e-01 -3.09086204e-01
1.02645659e+00 -2.29653791e-01 -5.44604778e-01 4.27259475e-01
6.55651987e-02 -1.43869579e-01 3.96982759e-01 -3.12477559e-01
-1.51820675e-01 -1.06531076e-01 -1.32421088e+00 1.10807264e+00
7.24946558e-01 7.92911947e-02 -4.54103082e-01 4.30142313e-01
8.03537190e-01 -3.90460610e-01 -6.03905499e-01 2.95485348e-01
-5.99144446e-03 -6.93124592e-01 9.43150520e-01 -1.02351511e+00
-1.02825530e-01 -3.70678306e-01 -1.11990318e-01 -1.72438824e+00
-2.82591671e-01 -1.28353906e+00 -2.03854203e-01 1.07975495e+00
-2.76968628e-02 -8.82800758e-01 8.61933589e-01 2.76717991e-01
5.21205813e-02 -4.97679502e-01 -1.28477538e+00 -1.13464844e+00
2.47294366e-01 3.47695425e-02 7.77044296e-01 1.06372142e+00
1.37215421e-01 1.46631664e-02 -7.54366934e-01 7.81013310e-01
8.06889176e-01 -2.00315267e-01 1.17270589e+00 -7.52699196e-01
-8.20817351e-01 -4.88501042e-01 -6.82168081e-02 -4.56111223e-01
4.59543079e-01 -4.00994897e-01 -3.08057610e-02 -8.61194491e-01
-7.68950135e-02 -6.49855375e-01 -3.56915116e-01 6.11528575e-01
-3.18641424e-01 -3.88146192e-01 3.91693473e-01 2.78987646e-01
-9.40502286e-01 8.43611777e-01 1.26761281e+00 -2.15705067e-01
-7.47359574e-01 5.30323446e-01 -8.20230126e-01 6.08956039e-01
1.03317332e+00 -7.16344178e-01 -5.06800473e-01 1.25073008e-02
-1.20264925e-01 5.98815441e-01 9.63835567e-02 -1.15709066e+00
2.78117120e-01 -8.63562703e-01 -1.30033419e-01 8.52035061e-02
1.54311776e-01 -9.81825650e-01 1.32801458e-01 1.07281518e+00
-1.32279769e-01 4.49743420e-01 9.46924239e-02 7.71029890e-01
6.03919439e-02 -2.56729573e-01 8.39127302e-01 -1.21652626e-01
-2.61668414e-01 5.98722816e-01 -6.95990980e-01 2.12775156e-01
1.61813426e+00 2.26849705e-01 -4.05289710e-01 -6.02333367e-01
-4.34966475e-01 9.30240154e-01 4.86309230e-01 3.03880334e-01
4.14795667e-01 -1.16158700e+00 -8.79188538e-01 1.99531429e-02
-2.71794379e-01 -1.27105638e-01 2.62871981e-01 5.67159891e-01
-1.03192799e-01 -2.96995968e-01 -5.17335474e-01 -1.12503901e-01
-1.16163599e+00 8.18451107e-01 5.15685022e-01 -6.78753674e-01
-4.31423336e-01 4.31921899e-01 -8.56667534e-02 -5.27110755e-01
3.87668192e-01 3.84630293e-01 -4.42705512e-01 -4.18006144e-02
4.91120398e-01 4.78805602e-01 -4.90248889e-01 -4.16515619e-01
-3.82695198e-01 1.43628120e-01 -1.79646507e-01 -2.42550194e-01
1.35767746e+00 4.39528152e-02 1.21031545e-01 -2.22994223e-01
3.41562510e-01 1.94748685e-01 -1.55450034e+00 -1.76484138e-01
-2.40108520e-01 -3.08907300e-01 -4.47726190e-01 -5.31231046e-01
-1.09380436e+00 1.98565200e-01 4.03678983e-01 5.74122608e-01
1.10153544e+00 -5.09012640e-01 5.71087122e-01 5.96779943e-01
6.75434589e-01 -1.16169059e+00 5.29216468e-01 7.28093505e-01
5.49620092e-01 -1.06641126e+00 -4.81989309e-02 2.47056633e-01
-1.20639980e+00 8.55451107e-01 1.02354956e+00 -4.29184884e-01
1.56323299e-01 3.86924416e-01 -2.06263345e-02 1.26074422e-02
-9.25018430e-01 -5.65368198e-02 -6.30462646e-01 1.13010550e+00
-7.63161421e-01 1.21267229e-01 -9.51215923e-02 7.29260027e-01
-1.79099590e-02 -7.79677033e-01 9.05783772e-01 1.08045578e+00
-5.03981292e-01 -1.38277352e+00 -8.31426680e-01 1.63469493e-01
-4.64024365e-01 4.02076721e-01 -2.18933284e-01 8.07405293e-01
1.73091888e-01 1.07156610e+00 -3.35300148e-01 -5.12907684e-01
3.41802925e-01 -4.35223043e-01 -7.00006448e-03 -4.66742367e-01
-1.00543153e+00 1.29182234e-01 -1.88426629e-01 -5.23288012e-01
-1.18334956e-01 -7.59695709e-01 -1.21955967e+00 -4.32527006e-01
-1.00317277e-01 4.65221584e-01 1.68960959e-01 8.01664710e-01
1.09729454e-01 7.06171572e-01 1.45952559e+00 -6.38155162e-01
-1.46571422e+00 -7.82444894e-01 -6.59776092e-01 3.14006865e-01
1.72633186e-01 -9.12448943e-01 -7.25937545e-01 -8.12589884e-01] | [3.7970683574676514, 2.2970290184020996] |
b82f9024-1b9c-4749-b94d-818b84177aa8 | fast-classification-of-small-x-ray-1 | null | null | https://www.nature.com/articles/s41524-019-0196-x | https://www.nature.com/articles/s41524-019-0196-x.pdf | Fast classification of small X-ray diffraction datasets using data augmentation and deep neural networks | X-ray diffraction (XRD) data acquisition and analysis is among the most time-consuming steps in the development cycle of novel thin-film materials. We propose a machine learning-enabled approach to predict crystallographic dimensionality and space group from a limited number of thin-film XRD patterns. We overcome the scarce data problem intrinsic to novel materials development by coupling a supervised machine learning approach with a model-agnostic, physics-informed data augmentation strategy using simulated data from the Inorganic Crystal Structure Database (ICSD) and experimental data. As a test case, 115 thin-film metal-halides spanning three dimensionalities and seven space groups are synthesized and classified. After testing various algorithms, we develop and implement an all convolutional neural network, with cross-validated accuracies for dimensionality and space group classification of 93 and 89%, respectively. We propose average class activation maps, computed from a global average pooling layer, to allow high model interpretability by human experimentalists, elucidating the root causes of misclassification. Finally, we systematically evaluate the maximum XRD pattern step size (data acquisition rate) before loss of predictive accuracy occurs, and determine it to be 0.16° 2θ, which enables an XRD pattern to be obtained and classified in 5.5 min or less. | ['Giuseppe Romano', 'Siyu I. P. Tian', 'Aaron Gilad Kusne', 'Tonio Buonassisi', 'Noor Titan Putri Hartono', 'Felipe Oviedo', 'Charles Settens', 'Brian L. DeCost', 'Zhe Liu', 'Savitha Ramasamy', 'Zekun Ren', 'Shijing Sun'] | 2019-05-17 | null | null | null | npj-computational-materials-2019-5 | ['material-classification', 'material-recognition', 'x-ray-diffraction'] | ['computer-vision', 'computer-vision', 'miscellaneous'] | [ 4.95353431e-01 1.32785007e-01 -1.26090124e-01 -4.88364339e-01
-5.82836986e-01 -2.12966949e-01 6.21502519e-01 2.30228156e-01
-4.00314689e-01 7.87468970e-01 -5.68649828e-01 -4.33656365e-01
-3.80173504e-01 -7.68567562e-01 -7.03129292e-01 -9.44380701e-01
4.93111322e-03 8.37909877e-01 3.25864069e-02 2.93396026e-01
4.87398177e-01 1.09495366e+00 -1.78457654e+00 6.62872493e-01
9.03028667e-01 1.33627987e+00 5.67663610e-01 4.06883776e-01
-8.90722200e-02 4.51915413e-01 -5.57958841e-01 2.29633555e-01
-1.77173577e-02 -1.88086197e-01 -7.51665890e-01 2.00831562e-01
3.85841876e-01 -1.12736754e-01 4.60641049e-02 6.97637916e-01
3.11271787e-01 -2.70336628e-01 9.09880936e-01 -7.06135631e-01
-7.54210472e-01 4.07127500e-01 -1.83369398e-01 2.16332167e-01
2.17960745e-01 5.09419262e-01 7.53486037e-01 -1.00019157e+00
9.98942196e-01 6.91050351e-01 3.65179926e-01 5.43812335e-01
-1.42448950e+00 -6.35431528e-01 -1.07239328e-01 4.14087236e-01
-1.33015287e+00 -2.63087034e-01 8.05344641e-01 -6.68271065e-01
1.41199207e+00 2.31944665e-01 6.24711692e-01 1.19463408e+00
4.28689629e-01 2.42537379e-01 1.12639964e+00 -4.72337842e-01
8.01999092e-01 2.79558152e-02 1.44892231e-01 3.22280049e-01
4.78052646e-01 1.45149752e-01 -2.79911667e-01 1.05305307e-01
6.17935836e-01 -2.17667460e-01 -1.48289457e-01 -2.37644121e-01
-8.27464819e-01 3.87802064e-01 4.02036607e-01 3.72439831e-01
-5.14035225e-01 -3.50136131e-01 -4.76301163e-02 4.79273170e-01
3.39361280e-01 1.21833980e+00 -5.42819977e-01 2.38296092e-01
-9.81177807e-01 3.89949173e-01 4.72267240e-01 7.94936419e-01
5.78944683e-01 -5.57910204e-02 2.59920180e-01 6.28858030e-01
-6.56326127e-04 5.10936081e-01 4.52104181e-01 -5.07914126e-01
2.18953490e-01 8.73287320e-01 -2.16814294e-03 -3.61832380e-01
-8.43871832e-01 -2.65155971e-01 -7.07559109e-01 4.17226642e-01
2.36681297e-01 1.26809001e-01 -1.01445735e+00 1.09735632e+00
2.00532287e-01 -5.61669469e-01 -6.89117461e-02 9.35212374e-01
5.61736166e-01 5.76160252e-01 1.49133746e-02 -5.01225948e-01
1.17710173e+00 -5.33351898e-01 -3.39007556e-01 4.77743931e-02
7.86337197e-01 -4.54875082e-01 1.23814535e+00 1.08460784e+00
-9.81137395e-01 -6.06055975e-01 -1.78969049e+00 5.24624325e-02
-6.23119056e-01 2.29092509e-01 6.67905927e-01 5.62766194e-01
-2.56201237e-01 1.18625927e+00 -9.71864581e-01 -2.61409264e-02
7.12605774e-01 7.87403643e-01 -5.03186226e-01 -6.05192445e-02
-6.58033073e-01 8.15964162e-01 3.63528907e-01 -7.48780891e-02
-4.32493001e-01 -1.14117026e+00 -2.87559062e-01 -2.61397570e-01
1.43042684e-01 -2.49613240e-01 1.11414170e+00 -6.71317458e-01
-1.43580830e+00 6.91587031e-01 8.36098045e-02 -3.65210146e-01
2.44694784e-01 7.47613087e-02 -7.31604218e-01 9.08365324e-02
9.09219608e-02 2.99193144e-01 6.12623274e-01 -1.14255059e+00
-4.40879822e-01 -4.53366220e-01 -2.96784341e-01 -4.05579507e-01
-3.24687570e-01 -2.57256716e-01 9.41705555e-02 -1.80112258e-01
3.31598014e-01 -6.29606009e-01 -4.70236018e-02 -1.24778673e-01
-4.22750235e-01 -1.67022496e-01 8.37242067e-01 -4.24411178e-01
1.05422473e+00 -1.74285066e+00 2.21938461e-01 4.51845735e-01
3.98665547e-01 1.75000116e-01 1.97742015e-01 2.55298048e-01
-4.66780990e-01 -6.77251220e-02 -3.55600834e-01 9.37849507e-02
-1.37447387e-01 -6.56719059e-02 4.33971249e-02 4.68685895e-01
6.61987185e-01 6.64670348e-01 -4.31002319e-01 2.09543139e-01
3.21599752e-01 1.35577574e-01 -5.73636830e-01 2.38977462e-01
-7.48112023e-01 3.68105352e-01 -4.99529749e-01 9.15317535e-01
7.73937702e-01 -5.84227204e-01 3.70751113e-01 -4.72347945e-01
-4.56949979e-01 5.99455178e-01 -9.05794144e-01 1.44159603e+00
-3.69619846e-01 3.66352767e-01 -3.20844531e-01 -8.27812493e-01
1.16819155e+00 4.61990125e-02 5.39269924e-01 -1.52189195e+00
9.89542678e-02 4.04983282e-01 3.98752153e-01 -6.18562818e-01
1.21567145e-01 -2.28871927e-01 2.12631315e-01 6.44222498e-01
1.00230031e-01 -1.02050304e-01 1.95951164e-01 -5.32647073e-01
1.17338920e+00 -1.09789066e-01 1.23381978e-02 -6.29358292e-01
5.96981168e-01 2.08313018e-01 9.82600674e-02 5.68331838e-01
4.76771593e-01 7.54981160e-01 4.18608546e-01 -9.48101461e-01
-1.74739683e+00 -9.03614402e-01 -5.43738842e-01 5.10638058e-01
-3.03080827e-01 -3.53205889e-01 -8.35598052e-01 -3.09292972e-01
8.70863497e-02 7.11515188e-01 -6.15394294e-01 -2.79720366e-01
-6.07884407e-01 -1.16819489e+00 3.60592082e-02 5.92666268e-01
1.01203203e-01 -1.09731340e+00 -6.02789044e-01 3.79687756e-01
8.04074943e-01 -9.26667929e-01 3.40572268e-01 7.49651253e-01
-6.59372687e-01 -1.41370320e+00 -2.79032320e-01 -4.36563700e-01
8.36220741e-01 -2.23508373e-01 8.25577796e-01 -4.14688662e-02
-9.04170513e-01 -8.57687518e-02 -1.47965252e-01 -2.73814768e-01
-6.04319215e-01 3.94784868e-01 4.01819855e-01 -3.36071104e-01
6.23078883e-01 -6.64686859e-01 -5.38316548e-01 2.55009949e-01
-7.41480231e-01 2.44249940e-01 7.15279520e-01 5.70736170e-01
8.76514971e-01 1.02415264e-01 5.57731688e-01 -9.24678087e-01
4.12845433e-01 -3.10993940e-01 -8.95139813e-01 1.94571763e-01
-1.02596700e+00 2.50962645e-01 1.07986045e+00 -3.82787943e-01
-7.13791132e-01 6.35899156e-02 -3.29430372e-01 -2.90267795e-01
-4.59569603e-01 3.67184937e-01 -4.51709032e-01 -1.43913403e-01
9.31446493e-01 9.08102617e-02 -4.47731242e-02 -5.95246196e-01
-6.80262893e-02 8.37108433e-01 2.52922893e-01 -7.63242364e-01
5.48811257e-01 1.48990512e-01 1.96211085e-01 -1.13690186e+00
-5.81257761e-01 -6.43121079e-03 -1.02146780e+00 -1.16330095e-01
8.81194234e-01 -7.21144319e-01 -7.59271085e-01 3.70031267e-01
-9.30018485e-01 -2.38513395e-01 -1.37086198e-01 4.91345555e-01
-5.02092659e-01 -6.23499192e-02 -4.62835491e-01 -4.58548665e-01
-4.20302361e-01 -1.27471161e+00 8.08271110e-01 -2.37600431e-02
-5.78274250e-01 -5.84735572e-01 -3.70280713e-01 2.58479506e-01
2.78250664e-01 2.78184563e-01 1.55027628e+00 -7.33131468e-01
-7.47209609e-01 -2.48531982e-01 -2.65248474e-02 2.97714531e-01
3.97826046e-01 6.70410097e-02 -9.94895637e-01 -2.75981218e-01
8.25276077e-02 -1.89805239e-01 5.47766268e-01 2.81331271e-01
1.41325665e+00 6.87367143e-03 -3.67160708e-01 5.68625510e-01
1.35068953e+00 7.21490800e-01 6.56626403e-01 6.23188138e-01
6.89374804e-01 3.71145546e-01 3.89125913e-01 4.60888267e-01
-1.80098057e-01 7.08007455e-01 1.54649228e-01 6.65249899e-02
1.31701808e-02 8.31480920e-02 -1.16835132e-01 6.10234618e-01
-2.84484506e-01 -1.78906262e-01 -1.09294510e+00 -8.32460895e-02
-1.05023122e+00 -5.40634930e-01 -2.51703590e-01 2.13562107e+00
7.65613258e-01 6.46448553e-01 1.32758453e-01 5.13950169e-01
2.91112840e-01 -2.30712384e-01 -9.65874732e-01 -3.12385261e-01
-3.20012301e-01 7.25671768e-01 4.37120497e-01 1.49393156e-01
-9.08626378e-01 5.46875477e-01 6.38600683e+00 4.11020666e-01
-1.55855644e+00 -4.45491582e-01 6.12138450e-01 -3.19865078e-01
-1.92722753e-01 -3.99264216e-01 -7.64752090e-01 5.28977334e-01
1.16075802e+00 3.76665406e-02 4.02084947e-01 8.87109399e-01
9.86048505e-02 4.54574823e-02 -1.64147949e+00 8.25743914e-01
-2.24353433e-01 -1.98350883e+00 1.69964448e-01 2.92663485e-01
5.00369370e-01 -5.59991300e-02 3.01588140e-02 -2.43829802e-01
-2.80841976e-01 -1.10376549e+00 6.34588599e-01 5.52995145e-01
9.49604571e-01 -8.36749077e-01 3.26008856e-01 2.89972164e-02
-6.81676090e-01 -2.56252348e-01 -2.97186285e-01 -5.70268258e-02
-2.08964691e-01 6.25322402e-01 -8.71752441e-01 4.71469462e-01
7.21847951e-01 6.03017092e-01 -5.68720102e-01 5.84997177e-01
1.79427966e-01 3.33570629e-01 -5.10886014e-01 -2.55667061e-01
1.45596758e-01 -1.60624370e-01 1.25800133e-01 7.52339303e-01
3.68724495e-01 1.87494650e-01 -4.18139756e-01 1.39912570e+00
1.54284537e-02 -2.50517607e-01 -2.35994801e-01 -2.38568425e-01
4.35541689e-01 8.84596169e-01 -7.61773825e-01 -3.69984168e-03
-3.80753338e-01 6.21157289e-01 6.17763363e-02 8.65378454e-02
-5.12191951e-01 -2.05233037e-01 4.96531099e-01 7.28168786e-01
5.26376903e-01 -2.18475848e-01 -5.36464810e-01 -7.25686252e-01
3.43940526e-01 -7.58949757e-01 -1.09585173e-01 -7.50285983e-01
-1.15853477e+00 5.85829675e-01 -1.24146223e-01 -9.97729957e-01
-2.09385589e-01 -1.35974884e+00 -4.58237052e-01 6.59439385e-01
-1.19680464e+00 -6.90746129e-01 -1.78529769e-01 -8.30583796e-02
3.88715476e-01 -4.76162195e-01 9.30462301e-01 4.97670412e-01
-6.90358937e-01 6.39310002e-01 4.05559391e-01 -4.77994978e-01
1.72573164e-01 -1.20229352e+00 4.91555929e-01 2.66011447e-01
-3.79183143e-01 3.80087405e-01 8.27449560e-01 -5.72633922e-01
-1.85778332e+00 -1.11725450e+00 6.03487909e-01 -3.43040347e-01
6.64926171e-01 -6.93537712e-01 -1.23558033e+00 2.10917711e-01
-3.48517954e-01 -2.37647325e-01 6.53003156e-01 6.53061410e-03
-2.04673484e-01 -1.31707609e-01 -1.08048189e+00 4.58961397e-01
9.96602952e-01 -4.66570556e-01 -2.55162895e-01 4.92969781e-01
5.09798706e-01 -1.46469742e-01 -1.44186974e+00 4.88216698e-01
5.00722349e-01 -8.46966028e-01 8.10042739e-01 -6.85302496e-01
7.58494854e-01 -2.75015205e-01 -1.89692378e-01 -7.69408226e-01
-3.30955356e-01 -1.68498799e-01 6.96482956e-02 8.23309541e-01
8.00309598e-01 -3.71328771e-01 1.19128346e+00 6.88351154e-01
-3.80176812e-01 -9.91523385e-01 -1.05837548e+00 -8.17831874e-01
3.15798163e-01 -3.98696721e-01 8.40445817e-01 8.22239339e-01
-8.42653364e-02 3.59818697e-01 4.63302024e-02 2.26066753e-01
4.23979700e-01 2.76110172e-01 4.16708440e-01 -1.56219172e+00
-1.05959766e-01 -3.97556573e-01 -4.33556348e-01 -4.59709734e-01
4.11063619e-02 -9.11917031e-01 -4.09717470e-01 -1.14504111e+00
-2.09848315e-01 -6.47033095e-01 -1.58578247e-01 3.18388700e-01
6.58307374e-01 -1.76308379e-01 -2.62795627e-01 1.79096952e-01
-3.06825370e-01 4.62139279e-01 1.14660978e+00 -2.75844127e-01
-2.00565889e-01 -2.07802996e-01 -3.29690874e-01 2.91636705e-01
6.47634983e-01 -3.37337971e-01 -1.77765056e-01 -2.10620418e-01
1.62727565e-01 -1.28449947e-01 2.00413331e-01 -1.53298867e+00
1.50419310e-01 -4.89375181e-02 9.15596783e-01 -6.73356295e-01
2.12481990e-01 -9.10933316e-01 4.95847821e-01 5.57963729e-01
-2.32141599e-01 -2.88198322e-01 2.67451227e-01 4.11206633e-02
3.67966443e-02 -1.66381031e-01 7.42993176e-01 -3.06055862e-02
-4.38934147e-01 3.49446058e-01 -4.86769736e-01 -7.68523455e-01
9.81062174e-01 -4.14774209e-01 -2.07867667e-01 6.31250799e-01
-8.47616911e-01 -7.06894994e-02 6.40999258e-01 4.32541758e-01
5.43156981e-01 -1.18023825e+00 -3.29003125e-01 8.58332157e-01
4.38636452e-01 1.25702888e-01 3.56582075e-01 3.67627800e-01
-7.82414556e-01 6.16527259e-01 -4.31846648e-01 -7.53976882e-01
-8.87036741e-01 6.32614911e-01 4.40849870e-01 6.89877048e-02
-7.52362847e-01 4.57886338e-01 -1.71880499e-01 -3.33220840e-01
-1.23541750e-01 -5.64039052e-01 3.51839401e-02 -1.27823323e-01
5.30542612e-01 1.45692378e-01 8.19076717e-01 -1.23974197e-01
-2.66999245e-01 5.04823089e-01 -5.90631604e-01 5.34530222e-01
1.83135796e+00 3.84709418e-01 8.65204111e-02 4.34441894e-01
1.28775430e+00 -7.17875123e-01 -1.30924213e+00 -1.74927711e-02
3.87571752e-01 4.78791259e-02 6.57968223e-03 -8.74394119e-01
-7.16128826e-01 6.63782358e-01 8.99056435e-01 3.23817879e-01
1.02465034e+00 1.64868400e-01 4.17317778e-01 7.89816141e-01
1.87370583e-01 -1.45908785e+00 -3.26749355e-01 1.85394108e-01
8.72214019e-01 -1.05908203e+00 3.70570868e-01 -3.79183143e-01
-2.73567811e-02 1.47078001e+00 7.45192826e-01 -1.08220175e-01
4.58957523e-01 5.08285642e-01 -2.48591363e-01 -5.61078906e-01
-1.03953159e+00 5.09950995e-01 2.18340442e-01 6.87295139e-01
3.55470598e-01 1.17704680e-03 -5.25160227e-03 9.09783185e-01
-4.24571246e-01 1.19781025e-01 3.29308033e-01 1.03132915e+00
-5.18395782e-01 -1.19450009e+00 -1.55569553e-01 7.05042005e-01
-3.23797986e-02 8.48561525e-03 -4.44207400e-01 9.65577900e-01
1.05471671e-01 3.84414941e-01 6.08624756e-01 -5.33808708e-01
5.77683806e-01 4.19216007e-02 8.01849246e-01 -4.04906869e-01
-2.28302211e-01 -2.52113163e-01 4.48149182e-02 -3.28892797e-01
-3.42885524e-01 -5.99707544e-01 -1.41750562e+00 -4.43900198e-01
-3.13606501e-01 -3.95571254e-02 7.79877305e-01 1.10368395e+00
3.85569185e-01 7.84907579e-01 7.23308623e-01 -9.19867218e-01
-8.55140612e-02 -8.28630030e-01 -7.35547185e-01 2.05482334e-01
1.82160631e-01 -7.15584815e-01 -2.10937619e-01 -1.01348154e-01] | [5.21409797668457, 5.335574626922607] |
cd8629eb-2ac0-4a46-a2ba-a9c995051be1 | task-prior-conditional-variational-auto | 2205.15014 | null | https://arxiv.org/abs/2205.15014v1 | https://arxiv.org/pdf/2205.15014v1.pdf | Task-Prior Conditional Variational Auto-Encoder for Few-Shot Image Classification | Transductive methods always outperform inductive methods in few-shot image classification scenarios. However, the existing few-shot methods contain a latent condition: the number of samples in each class is the same, which may be unrealistic. To cope with those cases where the query shots of each class are nonuniform (i.e. nonuniform few-shot learning), we propose a Task-Prior Conditional Variational Auto-Encoder model named TP-VAE, conditioned on support shots and constrained by a task-level prior regularization. Our method obtains high performance in the more challenging nonuniform few-shot scenarios. Moreover, our method outperforms the state-of-the-art in a wide range of standard few-shot image classification scenarios. Among them, the accuracy of 1-shot increased by about 3\%. | ['Zaiyun Yang'] | 2022-05-30 | null | null | null | null | ['few-shot-image-classification'] | ['computer-vision'] | [ 1.40980184e-01 -2.95990586e-01 -5.61501265e-01 -3.26246768e-01
-1.03579760e+00 1.85581118e-01 4.72691506e-01 -3.09752285e-01
-3.43292505e-01 8.95953059e-01 5.63635081e-02 5.47073603e-01
2.14134261e-01 -7.82513022e-01 -5.93828619e-01 -8.14437568e-01
6.18091524e-01 4.39674437e-01 3.42179984e-01 -1.01828612e-01
7.20913149e-03 -3.44892353e-01 -1.86339450e+00 7.75842369e-02
7.63164103e-01 1.07044375e+00 4.27244186e-01 4.20127690e-01
-1.45060390e-01 1.01586092e+00 -4.63744551e-01 -3.25014234e-01
-4.51973863e-02 -7.40614295e-01 -4.79153961e-01 5.20473480e-01
2.12080270e-01 -5.02544165e-01 -4.50961530e-01 1.35320222e+00
5.64593434e-01 5.27408838e-01 1.17941260e+00 -1.42798781e+00
-9.96482611e-01 3.92495960e-01 -7.16754496e-01 2.15333372e-01
-8.86297598e-02 1.68261215e-01 1.03250504e+00 -1.27999520e+00
7.79670358e-01 9.98784363e-01 3.76077056e-01 7.67383397e-01
-1.20259416e+00 -5.60033441e-01 -1.58081250e-03 4.13250655e-01
-1.57526910e+00 -6.20354176e-01 7.48363376e-01 -4.79578495e-01
5.62515497e-01 -3.03168148e-01 4.36121285e-01 1.43066382e+00
2.63131559e-02 1.00608552e+00 8.41489375e-01 -3.62998784e-01
6.95592046e-01 3.24950457e-01 2.51914889e-01 5.92732966e-01
1.68608636e-01 -2.80303299e-01 -6.92343831e-01 -1.25390843e-01
5.79689085e-01 5.75591564e-01 -1.10669725e-01 -3.88254344e-01
-8.58693480e-01 1.27428186e+00 1.94812313e-01 3.81273359e-01
-3.80454183e-01 1.58935472e-01 5.82529843e-01 1.55179590e-01
8.44277620e-01 4.08866443e-02 3.58048007e-02 -4.69620712e-02
-1.11269689e+00 5.39691672e-02 6.59266949e-01 1.39349926e+00
9.96792078e-01 4.04588908e-01 -7.97885060e-01 1.09433544e+00
7.55898505e-02 3.48904669e-01 5.30977190e-01 -8.96973491e-01
8.38562548e-02 -3.52427177e-02 2.84655422e-01 -4.82282579e-01
3.24826956e-01 -3.67204458e-01 -8.13069940e-01 -2.79372841e-01
-1.33536637e-01 -1.73556894e-01 -1.27873468e+00 1.71980095e+00
1.26912192e-01 7.13582039e-01 5.78410067e-02 9.07646239e-01
1.00304759e+00 8.16292405e-01 2.36595735e-01 -7.94621825e-01
1.16389382e+00 -1.17037046e+00 -1.07921946e+00 -2.09002212e-01
4.68925655e-01 -4.68236059e-01 1.40206933e+00 2.57427879e-02
-7.70527840e-01 -5.58584213e-01 -1.14470816e+00 5.67942001e-02
-2.05256462e-01 -1.10111304e-01 5.58463931e-01 5.07018566e-01
-5.32310069e-01 2.44323641e-01 -4.09127086e-01 -4.45553362e-01
8.02505016e-01 -4.26757425e-01 5.13243955e-03 -6.86664701e-01
-1.26188302e+00 5.40767074e-01 3.34874392e-01 -6.76783919e-01
-1.53314137e+00 -9.04935658e-01 -1.05434418e+00 2.22599640e-01
7.84650981e-01 -6.97909653e-01 1.39732862e+00 -5.34359455e-01
-1.47533619e+00 8.31773520e-01 -3.48140687e-01 -2.27800101e-01
4.35808241e-01 -1.68537006e-01 -2.38721013e-01 3.17231357e-01
5.15535414e-01 6.00651026e-01 1.14548326e+00 -1.31310093e+00
-2.69272357e-01 -1.09595664e-01 -4.59302589e-03 1.63459703e-01
-6.26871765e-01 -1.52792647e-01 -6.53902948e-01 -5.47248065e-01
-4.06217903e-01 -5.92464030e-01 -1.98665217e-01 7.59561434e-02
-1.97538897e-01 -3.97488594e-01 1.03508389e+00 9.08010453e-02
1.00810182e+00 -2.30800366e+00 4.54660952e-02 -6.99301839e-01
1.32805228e-01 2.65670508e-01 -6.32551238e-02 3.32546294e-01
2.20629454e-01 -7.04348981e-02 -2.98638165e-01 -7.28694022e-01
1.64554298e-01 4.51770723e-01 -4.31619436e-01 4.31676209e-01
1.01726696e-01 1.02340209e+00 -1.14713418e+00 -9.90455091e-01
4.01671320e-01 4.48506504e-01 -4.38571572e-01 4.66508925e-01
-3.90521020e-01 1.68271393e-01 -4.45772320e-01 7.53876269e-01
3.88350427e-01 -6.88872457e-01 -3.73039365e-01 9.20126140e-02
2.69826382e-01 -4.51868623e-01 -8.53930891e-01 1.95966029e+00
-3.69878769e-01 4.41671252e-01 -4.07976270e-01 -1.06521964e+00
7.27993488e-01 6.13430619e-01 5.92239499e-01 -3.60421598e-01
3.31191748e-01 -9.51122046e-02 -4.06526893e-01 -4.87767965e-01
3.79169732e-01 -8.04279268e-01 -2.18909636e-01 3.33141685e-01
7.36315846e-01 -2.65913785e-01 2.89363176e-01 4.92001384e-01
7.77617693e-01 -9.60304867e-03 5.27052879e-01 -2.33211014e-02
-1.37041286e-01 -9.71512720e-02 8.65352690e-01 1.08213067e+00
-5.66455364e-01 9.58217680e-01 3.94847393e-01 -1.49701852e-02
-1.13279903e+00 -1.04983282e+00 -3.05275559e-01 1.26038945e+00
3.90214086e-01 -3.71374816e-01 -7.15682030e-01 -4.17932332e-01
-3.15091550e-01 1.13087749e+00 -8.03611517e-01 -4.73766476e-01
3.75748247e-01 -8.86233628e-01 1.74672991e-01 4.63201702e-01
6.25979662e-01 -8.91530037e-01 -6.19399369e-01 2.75932163e-01
-3.13064307e-01 -1.43513310e+00 -4.19534832e-01 1.33027807e-01
-5.82450092e-01 -7.52792656e-01 -1.27303112e+00 -6.94853067e-01
3.44718874e-01 6.40308559e-01 9.74097610e-01 -4.74032968e-01
-4.97722596e-01 4.77927357e-01 -7.43734658e-01 -4.80398357e-01
1.04357921e-01 -2.10744619e-01 7.53622577e-02 2.70881593e-01
6.97305381e-01 -3.87977213e-01 -5.37075102e-01 2.44619131e-01
-8.95768642e-01 -1.72177792e-01 2.05551401e-01 1.27301395e+00
8.34328473e-01 1.76064540e-02 8.44308317e-01 -1.11618721e+00
3.79604787e-01 -9.19690192e-01 -1.83664281e-02 3.40593398e-01
-5.15789449e-01 -3.96671236e-01 6.29015684e-01 -7.54032373e-01
-1.27073061e+00 -2.38488883e-01 2.67811507e-01 -1.27588570e+00
-1.37269080e-01 4.41099375e-01 -3.67254787e-03 2.15774462e-01
8.14472318e-01 4.00294989e-01 -2.45831266e-01 -2.38934830e-01
4.40630108e-01 6.74095571e-01 2.17695966e-01 -2.71559209e-01
5.90600312e-01 4.93213534e-01 -2.46115819e-01 -1.23764193e+00
-1.41536069e+00 -8.38663101e-01 -4.64319885e-01 -3.72559965e-01
1.01299465e+00 -1.27105343e+00 -9.44898054e-02 4.16521221e-01
-9.92007017e-01 -2.16947153e-01 -6.31401539e-01 5.64125359e-01
-9.17681932e-01 2.28356391e-01 -7.95659184e-01 -1.01983583e+00
-1.46141008e-01 -1.08315015e+00 1.11067390e+00 3.55838627e-01
8.89529362e-02 -7.94141352e-01 2.67696520e-03 2.74796754e-01
2.62933612e-01 5.49168773e-02 8.06203663e-01 -5.81710160e-01
-3.71310920e-01 -3.29312742e-01 -2.58295059e-01 3.23021680e-01
6.21980848e-03 -3.10606480e-01 -1.21951234e+00 -2.43865594e-01
1.01054303e-01 -1.11398542e+00 1.18522167e+00 5.79834223e-01
1.22619760e+00 7.37829730e-02 -2.55025893e-01 3.34260166e-01
1.70409822e+00 2.60977373e-02 6.84775531e-01 -4.82167870e-01
7.10769355e-01 2.09403217e-01 9.49153423e-01 8.31879795e-01
9.99366790e-02 5.25953889e-01 2.55011111e-01 2.36336380e-01
-5.17005995e-02 -2.69707918e-01 1.75353199e-01 8.93799484e-01
2.41009500e-02 -4.15806293e-01 -4.47892457e-01 8.26142848e-01
-2.03705525e+00 -1.27055073e+00 3.08971137e-01 2.02200174e+00
7.73788035e-01 9.13244635e-02 7.07790107e-02 -1.01633713e-01
8.37621748e-01 6.74115062e-01 -8.67682636e-01 3.61731499e-02
2.94174384e-02 -4.08547781e-02 1.62192449e-01 1.05466366e-01
-1.16383684e+00 1.11954415e+00 6.71310616e+00 1.28148758e+00
-7.38721490e-01 6.77924156e-01 6.37900412e-01 -3.40123892e-01
-1.51086465e-01 -9.21828449e-02 -8.46400082e-01 6.82833076e-01
8.29728484e-01 -4.37628627e-01 1.44156009e-01 1.19602096e+00
-1.70592859e-01 -1.70458764e-01 -9.17082608e-01 1.25181925e+00
6.47211671e-01 -1.32906878e+00 6.79264069e-02 -1.40387908e-01
1.12173462e+00 2.52882838e-02 1.58403456e-01 8.75407279e-01
3.58507127e-01 -7.79670894e-01 3.61730158e-01 6.66484892e-01
1.26337326e+00 -7.82856226e-01 6.35346889e-01 6.14694059e-01
-9.69511271e-01 1.59215461e-02 -8.72350574e-01 -1.89468376e-02
2.51162708e-01 6.75752699e-01 -3.52949679e-01 2.99332082e-01
6.35948479e-01 1.01028466e+00 1.05146291e-02 9.12370741e-01
4.79262508e-03 8.46782088e-01 1.57323524e-01 -2.07084358e-01
4.20996368e-01 -1.01177767e-01 5.17252266e-01 1.00870430e+00
2.41049573e-01 5.18669009e-01 3.70244086e-01 1.01752412e+00
-2.98252165e-01 1.44536182e-01 -8.77584815e-01 -1.70657098e-01
3.78601998e-01 1.03086734e+00 -6.10734165e-01 -7.63131797e-01
-6.88204885e-01 1.20705009e+00 3.06735754e-01 5.92928171e-01
-9.09646690e-01 -6.35060668e-01 4.65531558e-01 -2.16528416e-01
6.06756449e-01 1.81128100e-01 2.95479484e-02 -1.66587186e+00
-2.94747323e-01 -3.29248130e-01 3.47864687e-01 -9.74964082e-01
-1.69228017e+00 3.37201416e-01 1.74493313e-01 -1.50480545e+00
-3.21900129e-01 -3.54172230e-01 -8.16363931e-01 5.45452297e-01
-1.34401834e+00 -1.17291188e+00 -2.84778535e-01 7.17383325e-01
1.14175534e+00 -3.63749921e-01 9.73291516e-01 3.84782046e-01
-5.38094938e-01 5.46292305e-01 2.17287078e-01 1.04061179e-02
7.83624351e-01 -9.33876753e-01 -2.36190841e-01 6.75587237e-01
1.70775041e-01 3.19128215e-01 7.72968531e-01 -5.23160100e-01
-1.04025686e+00 -1.26620066e+00 5.28871059e-01 -1.02356181e-01
5.20648301e-01 -3.84083718e-01 -1.01788390e+00 6.43967450e-01
1.65304378e-01 4.98928785e-01 1.16745591e+00 1.95516478e-02
-4.90264386e-01 1.37767851e-01 -1.11434078e+00 6.31065011e-01
9.64738548e-01 -8.26576293e-01 -6.94112897e-01 5.41931748e-01
1.05055809e+00 1.61596403e-01 -7.22893953e-01 1.96064651e-01
2.40052059e-01 -6.99413657e-01 8.09037209e-01 -8.42673838e-01
7.32965410e-01 -1.95920374e-02 -7.55529046e-01 -1.48407352e+00
-4.76083547e-01 -1.37828499e-01 -4.01555300e-01 9.71140563e-01
-2.23864838e-02 -2.50694871e-01 6.82610989e-01 2.86177605e-01
-7.83980340e-02 -6.71378791e-01 -7.93930590e-01 -8.63641083e-01
-1.49495527e-01 -2.78953344e-01 5.76279163e-02 1.10280979e+00
1.29469484e-02 7.15088725e-01 -1.04702520e+00 -3.31671387e-01
1.11231315e+00 1.30945012e-01 5.48043311e-01 -1.29976964e+00
-3.35632175e-01 -5.66244917e-03 -4.82914448e-01 -9.04923916e-01
3.78205150e-01 -7.08047390e-01 3.58251989e-01 -1.41430271e+00
9.15870965e-01 6.94456398e-02 -5.81858516e-01 2.80742496e-01
-4.18613166e-01 4.90238786e-01 1.05897561e-01 3.85535449e-01
-1.07805514e+00 1.19127369e+00 1.20901549e+00 -2.99500138e-01
1.59836873e-01 -2.00822070e-01 -5.17055094e-01 6.94014072e-01
4.85711515e-01 -6.04525924e-01 -7.96083391e-01 5.11148944e-02
-3.15649480e-01 2.59960502e-01 2.00807914e-01 -1.00670230e+00
3.03556204e-01 -4.17536438e-01 3.01631987e-01 -6.44607842e-01
8.25036347e-01 -5.09537995e-01 -1.97484627e-01 1.71602026e-01
-3.78346086e-01 -1.02364826e+00 -3.53077710e-01 1.14853847e+00
-3.56015533e-01 -5.49382091e-01 1.10009098e+00 -3.79726648e-01
-1.01776600e+00 9.65488553e-01 -2.70795822e-01 4.16168034e-01
1.35283828e+00 -2.02142909e-01 -2.44177982e-01 -5.21189034e-01
-6.94896042e-01 1.29629597e-01 3.47194046e-01 2.90725440e-01
8.65048110e-01 -1.62299085e+00 -7.16727316e-01 -1.29277453e-01
9.14577901e-01 -9.89899337e-02 6.18660569e-01 7.37543464e-01
1.58085182e-01 6.24792017e-02 -2.38029197e-01 -6.17531061e-01
-7.95505166e-01 8.83186758e-01 3.86438631e-02 1.92046627e-01
-6.91574931e-01 8.49580169e-01 3.95855725e-01 7.51429871e-02
2.19358787e-01 4.99139458e-01 -2.20769033e-01 3.36305022e-01
7.68543601e-01 3.47428560e-01 -5.45256615e-01 -5.37413359e-01
-1.45001570e-02 3.99162471e-01 -2.07238317e-01 -1.33312508e-01
1.40987682e+00 -6.86623007e-02 2.95799911e-01 1.42284274e+00
1.39353442e+00 -6.93665743e-01 -1.48351252e+00 -7.41238952e-01
-6.21823430e-01 -6.99496150e-01 2.44262561e-01 -2.68509209e-01
-8.23606908e-01 1.03848147e+00 4.02574986e-01 -4.11750041e-02
7.49847114e-01 2.13785857e-01 7.57380009e-01 4.30311322e-01
7.20719993e-01 -1.34534383e+00 5.18334627e-01 4.13781047e-01
4.67907727e-01 -1.75002432e+00 -6.07550219e-02 -4.05265212e-01
-1.04614890e+00 5.11850297e-01 6.46855593e-01 -1.98515892e-01
9.77684617e-01 -9.88001451e-02 -3.20872694e-01 -2.31827438e-01
-1.00497591e+00 -4.84571129e-01 1.79641128e-01 5.32411098e-01
1.72805712e-01 2.04932645e-01 -2.10819870e-01 6.86140776e-01
5.24242282e-01 4.46338952e-01 3.87674451e-01 9.75514770e-01
-7.06623375e-01 -2.92144507e-01 9.06853154e-02 6.94367111e-01
-3.63201439e-01 -8.21069330e-02 9.02999490e-02 4.11361873e-01
-6.17148876e-02 1.04538405e+00 7.34725520e-02 -3.48951548e-01
-6.23068474e-02 3.14179033e-01 4.03177083e-01 -1.11705065e+00
2.31571361e-01 2.65749425e-01 -2.26136982e-01 -3.09700638e-01
-5.12816548e-01 -5.42405128e-01 -8.09505463e-01 -1.51917160e-01
-5.31475723e-01 4.13261130e-02 2.16673110e-02 1.02402365e+00
9.68336388e-02 5.67049563e-01 6.92932665e-01 -8.03743720e-01
-7.40977645e-01 -1.17481029e+00 -1.18766081e+00 5.86952806e-01
1.16526678e-01 -1.12834096e+00 -5.34038365e-01 2.39690676e-01] | [10.006508827209473, 2.9332520961761475] |
886bff8f-5861-446c-9bcf-59647e008ea8 | invariant-slot-attention-object-discovery | 2302.04973 | null | https://arxiv.org/abs/2302.04973v1 | https://arxiv.org/pdf/2302.04973v1.pdf | Invariant Slot Attention: Object Discovery with Slot-Centric Reference Frames | Automatically discovering composable abstractions from raw perceptual data is a long-standing challenge in machine learning. Recent slot-based neural networks that learn about objects in a self-supervised manner have made exciting progress in this direction. However, they typically fall short at adequately capturing spatial symmetries present in the visual world, which leads to sample inefficiency, such as when entangling object appearance and pose. In this paper, we present a simple yet highly effective method for incorporating spatial symmetries via slot-centric reference frames. We incorporate equivariance to per-object pose transformations into the attention and generation mechanism of Slot Attention by translating, scaling, and rotating position encodings. These changes result in little computational overhead, are easy to implement, and can result in large gains in terms of data efficiency and overall improvements to object discovery. We evaluate our method on a wide range of synthetic object discovery benchmarks namely CLEVR, Tetrominoes, CLEVRTex, Objects Room and MultiShapeNet, and show promising improvements on the challenging real-world Waymo Open dataset. | ['Thomas Kipf', 'Aravindh Mahendran', 'Gamaleldin F. Elsayed', 'Mehdi S. M. Sajjadi', 'Sjoerd van Steenkiste', 'Ondrej Biza'] | 2023-02-09 | null | null | null | null | ['object-discovery'] | ['computer-vision'] | [-7.87091479e-02 -5.33673055e-02 -2.05742881e-01 -4.10482109e-01
-6.06221259e-01 -7.22091794e-01 8.62944484e-01 1.74875423e-01
-1.44478902e-01 5.55240333e-01 2.64128894e-01 -5.97766265e-02
-1.90586478e-01 -6.31454587e-01 -1.12815332e+00 -4.87267077e-01
5.90627901e-02 7.94414043e-01 5.27295709e-01 -1.66349962e-01
1.23722501e-01 5.34468651e-01 -1.87879825e+00 3.32388490e-01
4.24977034e-01 9.99425471e-01 1.38738483e-01 4.49553788e-01
-1.04971521e-01 7.97984242e-01 -2.26185828e-01 -3.15031826e-01
5.38260043e-01 -2.06484944e-01 -1.10247648e+00 2.41555050e-01
9.93746817e-01 -2.73723453e-01 -4.71955299e-01 7.62825727e-01
4.15873438e-01 2.06792682e-01 4.38895881e-01 -1.42329824e+00
-8.61595988e-01 7.05872416e-01 -5.84659696e-01 2.73332130e-02
2.98935641e-02 6.76126033e-02 1.42790294e+00 -8.09957564e-01
8.19084048e-01 1.46065712e+00 5.58197379e-01 4.22721297e-01
-1.51616752e+00 -3.85831237e-01 3.62887651e-01 5.02346575e-01
-1.20588696e+00 -6.34079456e-01 7.43584752e-01 -3.26041460e-01
1.09117115e+00 4.54260290e-01 7.06100881e-01 9.73536670e-01
-1.64263651e-01 9.09206092e-01 8.44301879e-01 -3.06151450e-01
3.19933444e-01 -1.95075154e-01 -1.45175550e-02 8.06685090e-01
2.01169044e-01 -6.04206063e-02 -5.61721385e-01 -7.67047405e-02
1.10163355e+00 7.12065920e-02 -3.56518961e-02 -1.24383783e+00
-1.62836432e+00 6.71942234e-01 9.43752170e-01 2.73756566e-03
9.34133120e-03 5.91002107e-01 4.74815160e-01 2.12774903e-01
2.15577319e-01 8.45387697e-01 -4.79891270e-01 -1.09210804e-01
-4.19056177e-01 5.70937276e-01 5.11154175e-01 1.17125797e+00
5.25769472e-01 1.89085696e-02 -2.33859524e-01 8.68628144e-01
-1.40088229e-02 4.37219024e-01 3.31932932e-01 -1.22888339e+00
3.13549936e-01 4.45306867e-01 2.45840877e-01 -9.26081300e-01
-5.59339643e-01 -4.34658229e-01 -4.89485770e-01 2.24067241e-01
3.81432980e-01 4.40488547e-01 -1.28164756e+00 1.72921634e+00
5.05704880e-01 5.87271340e-02 -1.66478202e-01 1.11610115e+00
9.87501264e-01 5.13090134e-01 2.55255662e-02 3.82003009e-01
1.46473598e+00 -1.28595340e+00 -3.24323863e-01 -1.74668700e-01
4.59220529e-01 -7.13027298e-01 1.13845885e+00 3.40229779e-01
-1.27748644e+00 -4.96326685e-01 -8.63036394e-01 -5.57940423e-01
-3.97862196e-01 2.82902252e-02 1.08348465e+00 2.40296409e-01
-7.54143417e-01 5.27155459e-01 -1.06431890e+00 -2.98086405e-01
7.07584858e-01 4.28047478e-01 -4.56094176e-01 -4.09917459e-02
-5.48125207e-01 6.14263654e-01 3.33193570e-01 -1.27888575e-01
-8.91805649e-01 -9.53782856e-01 -9.06818151e-01 1.57810539e-01
6.59172416e-01 -7.76310086e-01 1.40581357e+00 -7.61678934e-01
-1.36007988e+00 7.72595704e-01 1.11190667e-02 -4.86274689e-01
4.50217992e-01 -2.67055124e-01 -2.00769249e-02 -9.41866934e-02
2.24347919e-01 1.26454687e+00 7.24122226e-01 -1.29217017e+00
-6.54210746e-01 -3.04811597e-01 5.02740145e-01 2.85453826e-01
-2.44053692e-01 -2.49139130e-01 -5.25938988e-01 -7.40878046e-01
4.63564396e-01 -1.17746818e+00 -1.24750458e-01 6.15541995e-01
-3.68923426e-01 -3.90713066e-01 8.50333869e-01 -2.42457420e-01
3.58985126e-01 -2.14542818e+00 4.16859001e-01 -1.72136024e-01
4.33020324e-01 1.31366238e-01 -2.34637827e-01 9.81417447e-02
-2.31693089e-02 -3.14299092e-02 2.48605516e-02 -4.43626583e-01
2.12484732e-01 4.82675076e-01 -5.78530550e-01 3.72877032e-01
2.81473458e-01 1.28492832e+00 -9.77776408e-01 -2.26969391e-01
4.08931762e-01 4.21664804e-01 -9.77102995e-01 1.64296422e-02
-7.65322626e-01 3.29279602e-01 -2.10965380e-01 6.42907202e-01
4.13849860e-01 -5.84492743e-01 -8.31540897e-02 -4.11344230e-01
3.26233394e-02 5.54340541e-01 -1.22805250e+00 2.01372695e+00
-5.49795568e-01 7.50332713e-01 -2.44823813e-01 -8.93091142e-01
6.50852263e-01 -5.29875793e-02 4.23293203e-01 -8.65629256e-01
1.54686356e-02 -7.10670080e-04 6.04075305e-02 -2.02116802e-01
6.16865635e-01 2.49438167e-01 1.69417150e-02 4.21899021e-01
1.48398265e-01 -2.72957593e-01 2.60024101e-01 7.92208984e-02
9.51289117e-01 4.13396448e-01 1.95670515e-01 -2.22608581e-01
-9.27760750e-02 1.57550141e-01 4.67937887e-01 6.28532350e-01
1.81989655e-01 8.65491807e-01 4.05254424e-01 -9.75033998e-01
-1.32751846e+00 -1.38277602e+00 -1.41156361e-01 1.19520772e+00
5.05373180e-01 -4.65483516e-01 -2.87385643e-01 -5.34742415e-01
1.86001897e-01 3.23649526e-01 -6.30267978e-01 -5.01923598e-02
-6.12932444e-01 -4.30994242e-01 3.14290404e-01 9.09742713e-01
4.78247017e-01 -1.18172681e+00 -8.86880159e-01 1.70071632e-01
-1.06028050e-01 -1.38928354e+00 -3.67663175e-01 2.11820304e-01
-6.90748155e-01 -9.76290047e-01 -6.33319616e-01 -8.16869020e-01
6.62447929e-01 4.77205575e-01 1.32766271e+00 -1.87604129e-02
-7.45291591e-01 2.37909853e-01 -2.10373074e-01 -3.70801777e-01
1.57732710e-01 2.39594072e-01 -1.73027720e-02 -1.69701248e-01
9.43291858e-02 -4.89454240e-01 -4.44942504e-01 5.37878036e-01
-6.72510862e-01 2.53111660e-01 3.38483691e-01 8.88746321e-01
6.77357018e-01 -2.43479773e-01 2.52878815e-01 -6.53024852e-01
-1.17003471e-02 -1.97483405e-01 -9.10884321e-01 1.03040531e-01
-2.04008922e-01 4.00005639e-01 5.33666015e-01 -4.94835466e-01
-8.61024678e-01 2.26600721e-01 1.49447769e-01 -6.07087910e-01
-2.14768037e-01 2.32941424e-03 3.35737318e-02 -2.44775146e-01
6.57777250e-01 -1.67133939e-02 -2.37034768e-01 -5.92368305e-01
6.36864066e-01 1.86135277e-01 7.00667202e-01 -8.51536572e-01
8.99339736e-01 8.73701274e-01 9.55536589e-02 -6.81553304e-01
-9.43880856e-01 -3.82343501e-01 -5.16558766e-01 3.18369538e-01
8.05501044e-01 -9.02277350e-01 -8.87524188e-01 3.73876095e-01
-1.19063234e+00 -3.61913413e-01 -6.22633815e-01 2.27084920e-01
-8.06960702e-01 -2.08579786e-02 -2.69160986e-01 -2.97615707e-01
-4.06147353e-02 -1.20084155e+00 1.28019285e+00 4.92259823e-02
-2.55069137e-01 -6.80120468e-01 -1.45928368e-01 2.66861141e-01
5.25509715e-01 3.51120025e-01 1.12259042e+00 -8.36397186e-02
-1.30530500e+00 2.91430354e-01 -5.07349730e-01 -1.05084233e-01
1.25211343e-01 -1.96742013e-01 -7.50908673e-01 -2.73299545e-01
-6.16923809e-01 -7.71364033e-01 8.08243454e-01 4.97963168e-02
1.62948680e+00 -2.47200906e-01 -3.05438399e-01 9.02734935e-01
1.25609183e+00 -1.25540018e-01 3.57613474e-01 3.49976718e-01
9.96893167e-01 3.08117807e-01 3.39332253e-01 3.13291401e-01
5.83894610e-01 1.11833000e+00 7.40451992e-01 -1.15922593e-01
-6.17646277e-01 -3.03858221e-01 -2.13341549e-01 4.23758745e-01
-8.36335421e-02 -1.10234492e-01 -9.66091812e-01 8.84613454e-01
-1.95104527e+00 -7.96291411e-01 2.11128339e-01 1.92304981e+00
6.92045689e-01 9.25467461e-02 1.28037274e-01 1.15921656e-02
2.58768857e-01 1.39092863e-01 -7.26481497e-01 -1.64568365e-01
-1.17297545e-01 2.90198475e-01 4.95329469e-01 2.46777371e-01
-1.08365059e+00 1.04001474e+00 5.94729948e+00 4.98483688e-01
-1.29247630e+00 3.97223271e-02 4.36018109e-01 -3.14252198e-01
-3.51251751e-01 -8.29192698e-02 -6.52328491e-01 5.38041107e-02
3.12508613e-01 1.63335428e-01 6.56468213e-01 1.10125268e+00
-4.01348591e-01 1.72164202e-01 -1.55201924e+00 1.23233128e+00
1.20576523e-01 -1.81454849e+00 9.70402136e-02 -1.32462531e-01
8.57596755e-01 2.71207303e-01 1.93444476e-01 1.16669692e-01
3.31118196e-01 -1.02452505e+00 1.17811048e+00 1.07174963e-01
8.46065819e-01 -5.33333063e-01 1.90184385e-01 1.63283087e-02
-1.25857401e+00 -1.40716344e-01 -5.49413145e-01 -7.18713105e-02
3.55207734e-02 9.55816060e-02 -8.00033271e-01 3.25380206e-01
1.10602582e+00 7.45730698e-01 -6.26362324e-01 1.17400980e+00
-5.73796071e-02 2.35099822e-01 -4.65882123e-01 -5.01647852e-02
2.88188100e-01 2.99565971e-01 5.35568833e-01 6.63465679e-01
-3.87970395e-02 -1.33703977e-01 1.49794787e-01 1.11091447e+00
-3.44656050e-01 -2.42342874e-01 -4.35915202e-01 -4.04794794e-03
3.72714460e-01 1.12317717e+00 -1.05573940e+00 -2.32325122e-01
-3.44608366e-01 7.45310426e-01 5.17754138e-01 1.25173554e-01
-9.45946932e-01 -2.41476640e-01 8.74830604e-01 2.10312828e-01
7.82611728e-01 -3.84014636e-01 -4.55191851e-01 -1.12129116e+00
2.10075021e-01 -9.08791244e-01 7.58276379e-04 -8.72189105e-01
-1.02833867e+00 5.47399700e-01 2.33235225e-01 -1.06181180e+00
2.90114735e-03 -8.04731071e-01 -2.53817439e-01 3.31272781e-01
-1.32608771e+00 -1.28888059e+00 -5.05256295e-01 4.44308400e-01
8.37644637e-01 -1.07560627e-01 7.43286133e-01 3.58514577e-01
-1.40627146e-01 6.44567311e-01 -1.10967100e-01 3.51157114e-02
4.83110040e-01 -1.27376330e+00 9.25457537e-01 4.06627655e-01
8.04793954e-01 5.80826283e-01 6.32162809e-01 -1.70498103e-01
-1.55164874e+00 -1.12343132e+00 4.88095015e-01 -8.93485665e-01
5.15653253e-01 -9.00964737e-01 -6.78776383e-01 7.91518509e-01
-5.82668073e-02 6.46998048e-01 7.34185353e-02 3.37084085e-01
-8.58055115e-01 -7.76699185e-02 -9.10071254e-01 8.52006972e-01
1.71915913e+00 -3.83955777e-01 -6.01934910e-01 5.51598370e-01
9.12523508e-01 -9.06746149e-01 -7.90859282e-01 4.48437691e-01
6.94820285e-01 -7.18858004e-01 1.37184024e+00 -9.16736364e-01
5.45279860e-01 -4.37353104e-01 -1.50905609e-01 -1.17285419e+00
-4.97502714e-01 -6.18411899e-01 -2.02786595e-01 8.99131358e-01
2.79061794e-01 -4.29881364e-01 9.75666642e-01 4.69562352e-01
-1.86573967e-01 -9.35193419e-01 -9.06239569e-01 -8.83383036e-01
-1.41530871e-01 -2.22050056e-01 9.99296546e-01 8.27943861e-01
-4.54164535e-01 3.31932545e-01 -3.16307008e-01 3.57628986e-03
7.35987425e-01 4.60693866e-01 1.03543842e+00 -1.18992388e+00
-4.75400776e-01 -4.49282438e-01 -7.61047125e-01 -1.29109478e+00
1.30955845e-01 -8.43046904e-01 -2.68409122e-02 -1.41046071e+00
1.81334242e-01 -8.35271180e-01 -1.04145922e-01 8.37436438e-01
1.74728915e-01 6.37182057e-01 4.11424607e-01 9.59556773e-02
-9.24665809e-01 6.88004553e-01 1.31075478e+00 -2.27454290e-01
-3.74341086e-02 -4.16157931e-01 -5.87684393e-01 7.41769254e-01
5.89613318e-01 -3.54786396e-01 -5.09522498e-01 -1.05859065e+00
2.65931517e-01 -3.04453969e-01 8.64161968e-01 -1.06542575e+00
-5.67631274e-02 -2.04034895e-01 3.90194416e-01 -5.56669056e-01
6.99441850e-01 -8.01293433e-01 7.16801658e-02 1.78380698e-01
-5.17614901e-01 2.08400011e-01 2.43872955e-01 4.60941911e-01
5.65044805e-02 3.06203395e-01 7.70117760e-01 -1.62378162e-01
-9.01618004e-01 5.13112307e-01 4.01013404e-01 2.65528113e-01
1.02610350e+00 -1.27293050e-01 -5.64609587e-01 -1.56606331e-01
-5.64888120e-01 1.41345918e-01 6.42464459e-01 1.08049786e+00
3.91447544e-01 -1.58087611e+00 -3.65693659e-01 2.46488720e-01
4.62798893e-01 3.71374190e-01 1.01688355e-01 5.23755193e-01
-5.17093956e-01 5.48013926e-01 -5.59061468e-01 -1.01042235e+00
-1.08311725e+00 6.23092175e-01 2.57494837e-01 2.30070025e-01
-9.38945174e-01 1.01353908e+00 3.95576715e-01 -5.39206684e-01
4.58632112e-01 -6.53044224e-01 2.61361569e-01 -4.12111521e-01
2.45793551e-01 2.69442827e-01 1.59995928e-01 -7.01215208e-01
-3.10683787e-01 7.53809988e-01 -1.94350302e-01 8.00300315e-02
1.38328671e+00 1.18678883e-01 -8.86223540e-02 3.80453020e-01
1.19994438e+00 -2.24957138e-01 -1.44837320e+00 -4.77051914e-01
-2.01648504e-01 -6.28428936e-01 -9.87717882e-03 -5.81131220e-01
-1.10434926e+00 7.44788766e-01 4.73437935e-01 1.58577889e-01
6.28406107e-01 4.46077138e-01 8.02210033e-01 7.91750908e-01
5.80096483e-01 -8.94149005e-01 3.57388139e-01 4.16018873e-01
1.07957697e+00 -1.39433014e+00 -1.10181989e-02 -3.88242632e-01
-4.32102025e-01 8.87237489e-01 8.82134616e-01 -3.99557263e-01
5.03367186e-01 3.83566283e-02 -2.18722045e-01 -2.06394270e-01
-7.78378844e-01 -2.01028243e-01 4.47970062e-01 5.39989412e-01
2.31801361e-01 8.13135877e-02 1.09186277e-01 8.04395601e-02
-5.17930627e-01 -3.81253242e-01 3.25623214e-01 8.76446426e-01
-2.94294775e-01 -1.03391874e+00 -2.26402491e-01 3.81977797e-01
-1.45434618e-01 1.33163720e-01 -1.62822232e-01 8.25337946e-01
2.24711701e-01 3.70907515e-01 4.01633739e-01 1.62366442e-02
3.44700724e-01 -3.41943771e-01 7.87209272e-01 -7.34181464e-01
-3.34362015e-02 -2.60016888e-01 1.76102929e-02 -8.03476751e-01
-3.66222918e-01 -6.20075405e-01 -1.27136683e+00 -7.71634579e-02
-1.92403987e-01 -1.50843382e-01 6.01061583e-01 9.33409214e-01
6.21899784e-01 6.74645782e-01 2.72082061e-01 -1.11565161e+00
-5.95021665e-01 -4.47976798e-01 2.71234196e-03 6.41961336e-01
6.21128738e-01 -1.16151333e+00 4.07674015e-02 9.24600810e-02] | [7.795662879943848, -2.7174532413482666] |
98de3c0b-96c6-4aa9-a336-9c4270d59448 | speech-emotion-diarization-which-emotion | 2306.12991 | null | https://arxiv.org/abs/2306.12991v1 | https://arxiv.org/pdf/2306.12991v1.pdf | Speech Emotion Diarization: Which Emotion Appears When? | Speech Emotion Recognition (SER) typically relies on utterance-level solutions. However, emotions conveyed through speech should be considered as discrete speech events with definite temporal boundaries, rather than attributes of the entire utterance. To reflect the fine-grained nature of speech emotions, we propose a new task: Speech Emotion Diarization (SED). Just as Speaker Diarization answers the question of "Who speaks when?", Speech Emotion Diarization answers the question of "Which emotion appears when?". To facilitate the evaluation of the performance and establish a common benchmark for researchers, we introduce the Zaion Emotion Dataset (ZED), an openly accessible speech emotion dataset that includes non-acted emotions recorded in real-life conditions, along with manually-annotated boundaries of emotion segments within the utterance. We provide competitive baselines and open-source the code and the pre-trained models. | ['Alya Yacoubi', 'Alaa Nfissi', 'Mirco Ravanelli', 'Yingzhi Wang'] | 2023-06-22 | null | null | null | null | ['speaker-diarization', 'speech-emotion-recognition'] | ['speech', 'speech'] | [ 1.48292318e-01 7.64070451e-02 8.79762024e-02 -9.95168209e-01
-8.04371178e-01 -5.50176799e-01 3.97068620e-01 -1.21124014e-01
-3.23430479e-01 3.55919331e-01 6.21867836e-01 -1.60201192e-01
5.86301804e-01 -1.18345447e-01 -2.54489809e-01 -6.06431067e-01
-1.61783963e-01 1.25638142e-01 -4.50047642e-01 -4.33436960e-01
-1.36099324e-01 1.99219748e-01 -1.64210856e+00 5.62583685e-01
5.84650397e-01 1.62993133e+00 -1.51191086e-01 8.20348680e-01
-4.46999460e-01 9.76776540e-01 -1.16283941e+00 -4.62677866e-01
-4.86278027e-01 -5.42113841e-01 -1.06615984e+00 3.63173366e-01
-1.23701297e-01 9.09863859e-02 -1.51687220e-01 8.44623685e-01
5.16234159e-01 5.32801628e-01 6.07324064e-01 -1.51458657e+00
-5.85167050e-01 5.41363180e-01 1.12866424e-01 3.38495314e-01
6.83663547e-01 -1.94466978e-01 1.15875494e+00 -8.59818578e-01
5.97847998e-01 1.29940498e+00 3.15111458e-01 9.47963536e-01
-8.22836459e-01 -6.04597151e-01 4.27971601e-01 1.82386905e-01
-1.23644090e+00 -1.02334094e+00 9.94513154e-01 -3.54743958e-01
1.25341284e+00 6.77239716e-01 2.86815107e-01 1.72191143e+00
-1.86222091e-01 1.11831260e+00 8.72841060e-01 -1.40765950e-01
4.04511362e-01 9.52846929e-02 3.26629937e-01 1.02422880e-02
-1.16843534e+00 -3.68607044e-02 -7.40506709e-01 -7.52613395e-02
4.86489013e-02 -2.90608048e-01 -4.37346876e-01 4.89794463e-01
-1.15929282e+00 6.33134544e-01 -1.53875276e-01 4.63241994e-01
-6.47032380e-01 -2.23948538e-01 9.92520273e-01 7.38134980e-01
8.32131326e-01 1.64028183e-01 -8.07828665e-01 -1.03452361e+00
-6.64565146e-01 -9.76234898e-02 8.97236228e-01 9.28397596e-01
4.55995440e-01 4.27309036e-01 -1.10917285e-01 1.25602043e+00
-9.09814704e-03 2.56141156e-01 5.99685490e-01 -7.66391933e-01
2.89422512e-01 3.78370941e-01 1.13942601e-01 -7.15189636e-01
-3.10524285e-01 1.02881178e-01 -7.84243882e-01 -3.15652788e-01
6.65454939e-02 -6.69321954e-01 -7.75386453e-01 1.92639256e+00
3.08181465e-01 1.00221291e-01 6.54782236e-01 8.35119426e-01
1.30173790e+00 1.12996125e+00 2.93392330e-01 -5.84982812e-01
1.62774897e+00 -9.03395832e-01 -1.40782046e+00 -2.36867309e-01
4.61498171e-01 -6.90646172e-01 1.35658824e+00 3.65496486e-01
-8.56002748e-01 -3.33062589e-01 -7.05803037e-01 -1.83344912e-02
-6.47931397e-01 9.25201699e-02 4.62510794e-01 5.73755443e-01
-8.78833890e-01 -9.19937566e-02 -4.84577954e-01 -8.11232254e-02
-9.29356902e-04 -6.85751736e-02 -4.62972164e-01 7.98474193e-01
-1.60341918e+00 6.82547867e-01 9.26133431e-03 1.10522471e-01
-8.84944618e-01 -4.95989770e-01 -1.08527637e+00 1.72438830e-01
4.00341511e-01 1.94200560e-01 1.83711433e+00 -1.51713276e+00
-2.18108726e+00 1.08428133e+00 -7.15022683e-01 -2.31810436e-01
-7.49529451e-02 -1.62857324e-02 -1.15033305e+00 2.58337766e-01
-3.30012083e-01 4.24903363e-01 9.11256194e-01 -9.42916095e-01
-6.47808313e-01 -2.07665861e-01 -2.22520873e-01 1.45132661e-01
-2.89759547e-01 9.37065721e-01 -1.87044397e-01 -7.88858056e-01
-1.00231066e-01 -7.79749393e-01 1.74032167e-01 -6.59613371e-01
-4.13809031e-01 -7.29557395e-01 9.98437583e-01 -7.33758450e-01
1.47503090e+00 -2.80448771e+00 7.70458803e-02 -2.81845510e-01
-4.72829901e-02 -1.02751173e-01 -1.28022015e-01 3.92668247e-01
-6.08281374e-01 3.24247301e-01 -1.39141828e-01 -5.77447712e-01
5.49496830e-01 2.65019119e-01 -6.01025462e-01 9.60751325e-02
4.03435290e-01 6.47035778e-01 -6.45204723e-01 -3.82716835e-01
-2.29760315e-02 3.98166180e-01 -7.53581151e-02 6.71882987e-01
-1.35694280e-01 2.99201757e-01 -2.99310446e-01 6.88282251e-01
2.93267667e-01 2.89362848e-01 -2.70196497e-01 -4.11190614e-02
-9.07174274e-02 7.18826532e-01 -9.94076312e-01 1.41407132e+00
-5.90153039e-01 7.69995451e-01 5.89833260e-01 -8.81576359e-01
1.17038059e+00 1.15276217e+00 4.52126682e-01 -6.00232303e-01
4.67943192e-01 -1.10927001e-01 -2.58636087e-01 -7.18838036e-01
5.88205278e-01 -4.70929593e-01 -7.14027941e-01 3.64721984e-01
1.96829796e-01 -4.33443904e-01 -9.40381885e-02 6.38225973e-02
1.06314862e+00 -5.15022457e-01 3.11934590e-01 4.11487818e-02
5.26085496e-01 -3.42479795e-01 7.49331295e-01 2.02265218e-01
-8.78842235e-01 4.54113096e-01 6.59119070e-01 -3.12079936e-01
-4.11785275e-01 -9.94740367e-01 -2.02489048e-01 1.51297748e+00
-5.06888628e-02 -5.69447696e-01 -8.13643575e-01 -9.07539845e-01
-5.39483190e-01 9.74571705e-01 -7.66131401e-01 -4.93544489e-02
-1.73220992e-01 -2.59096444e-01 8.47693264e-01 2.84182191e-01
2.69382507e-01 -1.73835051e+00 -3.75170261e-01 1.61551967e-01
-8.03905189e-01 -1.32935059e+00 -8.72119009e-01 6.39624536e-01
1.25134671e-02 -5.67081332e-01 -3.97126466e-01 -8.51142287e-01
9.17686373e-02 -1.55337170e-01 1.29921865e+00 -3.67618501e-01
4.51701321e-02 4.67041612e-01 -9.35490072e-01 -5.74553967e-01
-6.40561283e-01 -2.12792531e-01 1.06530963e-02 4.07888055e-01
7.03514397e-01 -3.81088763e-01 -3.35881934e-02 4.14176017e-01
-7.71492541e-01 -2.91596264e-01 -9.38758403e-02 6.34327948e-01
4.42890257e-01 2.19637677e-01 1.04210794e+00 -4.38898176e-01
1.15310335e+00 -5.62521935e-01 7.08119646e-02 3.18393737e-01
1.99722409e-01 -1.67626262e-01 6.25903606e-01 -7.04144895e-01
-1.41400111e+00 -1.42683074e-01 -7.00710595e-01 -3.26129138e-01
-8.37791920e-01 4.13571715e-01 -5.38913190e-01 6.32817566e-01
3.28792363e-01 2.77160317e-01 -1.70049846e-01 -2.02497348e-01
4.65021014e-01 1.35717130e+00 6.69355631e-01 -6.88310742e-01
-9.07877758e-02 4.04999517e-02 -9.57539856e-01 -1.17634451e+00
-6.42051876e-01 -6.16239011e-01 -1.08911127e-01 -3.76872718e-01
1.10409391e+00 -8.41370285e-01 -1.05825496e+00 3.39019209e-01
-1.45279217e+00 -3.54204565e-01 -2.37151563e-01 3.27330083e-01
-7.73126960e-01 1.45437002e-01 -9.07439530e-01 -1.32533324e+00
-4.38011616e-01 -1.16787493e+00 1.31549191e+00 7.85938203e-02
-8.34056020e-01 -7.70136774e-01 -2.37604320e-01 1.91088364e-01
3.33871603e-01 1.61152139e-01 6.56414151e-01 -1.01273286e+00
5.27211785e-01 -1.42322471e-02 1.64132863e-01 6.22736454e-01
2.86763132e-01 2.47292906e-01 -1.20004308e+00 3.16546619e-01
3.40224445e-01 -8.02107513e-01 2.87664890e-01 7.54457265e-02
1.18402100e+00 -4.01431799e-01 2.26025760e-01 2.92866558e-01
4.63972241e-01 7.89128184e-01 5.17223656e-01 -3.06281894e-02
-1.73173212e-02 1.08289409e+00 6.50337517e-01 6.65873349e-01
6.20726764e-01 5.57281733e-01 2.45068118e-01 -1.46678369e-02
2.51325637e-01 1.04548596e-01 6.46880865e-01 1.28061235e+00
4.48901534e-01 -5.47918439e-01 -7.60796070e-01 5.33326387e-01
-1.52704561e+00 -1.18653524e+00 1.04952633e-01 1.50326288e+00
1.15131438e+00 -1.94726929e-01 1.85249835e-01 9.73377153e-02
7.06259668e-01 5.19374549e-01 -4.47151989e-01 -1.13174641e+00
-3.39780122e-01 1.15613751e-01 -6.08627498e-01 4.93419051e-01
-1.29093540e+00 1.15275860e+00 6.62033653e+00 8.87454093e-01
-1.43625033e+00 1.58291608e-01 8.37277949e-01 -1.24368779e-01
-1.93091348e-01 -4.04703766e-01 -5.19718349e-01 5.70569992e-01
1.43448496e+00 -2.77158856e-01 4.55684066e-01 9.55295146e-01
4.52555716e-01 1.31924912e-01 -1.14973962e+00 1.31981504e+00
1.53298661e-01 -8.23776186e-01 -4.30037439e-01 -4.78798330e-01
3.66531461e-01 -1.06552884e-01 -1.28191769e-01 7.82424510e-01
2.09219307e-01 -9.48668003e-01 9.90400493e-01 2.73769230e-01
7.97258973e-01 -7.38334119e-01 4.79360700e-01 3.55944782e-01
-1.26305175e+00 1.73517182e-01 1.64435834e-01 -1.53549701e-01
4.62569326e-01 3.29663843e-01 -6.79297209e-01 3.73027086e-01
9.61550832e-01 5.09411931e-01 -3.72178317e-03 2.03808710e-01
-3.07541132e-01 9.44212675e-01 -1.77280009e-01 -2.55810380e-01
2.34811813e-01 -1.23019278e-01 5.10343313e-01 1.72526097e+00
8.16731453e-02 4.09168422e-01 -1.40567288e-01 5.70621490e-01
-2.62408733e-01 2.84324169e-01 -3.62428010e-01 -6.38677299e-01
5.88488698e-01 1.35300303e+00 -3.42156500e-01 -4.37782049e-01
-4.29262310e-01 1.22473288e+00 1.62717104e-01 7.58218884e-01
-9.28505480e-01 -6.08196676e-01 1.38690364e+00 -5.79908431e-01
1.33005947e-01 2.35782005e-02 1.39469907e-01 -1.15702033e+00
7.81861320e-02 -1.30395424e+00 3.53694409e-01 -9.47563946e-01
-1.37404728e+00 1.33126318e+00 -4.92737204e-01 -9.26704228e-01
-5.20673573e-01 -2.93468654e-01 -8.57071340e-01 6.70413136e-01
-1.21283805e+00 -6.04545891e-01 -1.77844152e-01 6.16255820e-01
9.70917106e-01 8.30381934e-04 1.23331165e+00 3.41960907e-01
-7.72851467e-01 6.24401748e-01 -3.81760985e-01 2.87172109e-01
8.39546978e-01 -1.17338598e+00 2.41921708e-01 5.68880916e-01
6.17444376e-03 2.91467428e-01 9.21573818e-01 -1.96578965e-01
-1.10772109e+00 -7.76029468e-01 1.29703808e+00 -4.18981493e-01
9.05831158e-01 -6.98997796e-01 -1.04875278e+00 7.26384878e-01
5.65324366e-01 -1.39290780e-01 1.02528048e+00 4.52183366e-01
-3.29793543e-01 1.08647019e-01 -9.56785262e-01 5.12456000e-01
7.72881985e-01 -1.01359606e+00 -9.45792139e-01 -1.73094943e-02
1.21651852e+00 -2.90990353e-01 -9.30828631e-01 3.79497170e-01
2.33639792e-01 -7.85222650e-01 6.19381607e-01 -9.11930919e-01
2.75520980e-01 7.15772957e-02 -4.70319450e-01 -1.62270546e+00
2.93148488e-01 -9.46531773e-01 1.02273658e-01 1.75974298e+00
4.92523313e-01 -4.75765377e-01 4.19676512e-01 7.95573413e-01
-5.44339418e-01 -7.34135807e-01 -1.26575994e+00 -6.17833376e-01
-1.35870829e-01 -9.85567033e-01 9.45435226e-01 1.20630813e+00
6.51297390e-01 4.76455271e-01 -2.84484178e-01 1.30175263e-01
-2.56758481e-01 -1.24439094e-02 5.20792782e-01 -8.97805154e-01
1.94182977e-01 -6.70937181e-01 -3.46627146e-01 -1.04391897e+00
9.63653564e-01 -5.33629179e-01 5.96885324e-01 -8.99121583e-01
-4.36345249e-01 -1.77449193e-02 -2.33249608e-02 5.38419783e-01
-1.55089960e-01 -3.89315337e-01 -1.44286796e-01 -3.33578408e-01
-8.22084785e-01 1.09908545e+00 9.01843786e-01 -9.92456451e-02
-1.69379577e-01 -1.27530456e-01 -5.08109689e-01 7.05091357e-01
7.25455701e-01 -2.66483456e-01 -3.53984863e-01 -2.78380681e-02
2.08247732e-02 6.03943706e-01 -6.37933835e-02 -4.91919339e-01
8.81834999e-02 -4.16735381e-01 -3.25873941e-01 -5.38902342e-01
7.15966582e-01 -7.42425442e-01 -1.24487102e-01 -4.30397242e-01
-6.46567941e-01 9.46413353e-02 2.45866388e-01 1.32036000e-01
-9.93360162e-01 -9.20178965e-02 5.68958998e-01 2.29149282e-01
-8.30424964e-01 3.10775846e-01 -9.08709764e-01 3.14630270e-01
9.60303962e-01 -4.64887582e-02 -3.21389228e-01 -8.55550289e-01
-1.12266600e+00 3.07917923e-01 -5.34559786e-02 8.57272148e-01
6.51152492e-01 -1.24520588e+00 -4.96759266e-01 1.96187153e-01
4.50850368e-01 -3.42456907e-01 3.64452273e-01 5.75751305e-01
2.71538615e-01 1.94100410e-01 2.67321140e-01 -2.11035326e-01
-1.35810614e+00 5.00073850e-01 5.10953367e-01 2.78926969e-01
-2.90995628e-01 1.07708609e+00 3.07003945e-01 -6.18782103e-01
7.50077367e-01 -4.97134775e-01 -1.61711827e-01 3.28370064e-01
7.67248631e-01 -7.97820091e-02 5.20418733e-02 -1.06109416e+00
-5.13089061e-01 -1.14947118e-01 2.62671351e-01 -3.68638426e-01
1.12066245e+00 -5.67851901e-01 -7.44564831e-02 1.05931938e+00
1.36965907e+00 -1.78724620e-02 -1.15621436e+00 -1.64252251e-01
1.72940582e-01 6.86000511e-02 7.97283947e-02 -9.70870793e-01
-7.29095876e-01 1.00408947e+00 2.56036222e-01 7.39477456e-01
1.30574012e+00 2.98719525e-01 9.48666155e-01 2.21028820e-01
3.20741013e-02 -1.39816773e+00 1.83945134e-01 9.65669870e-01
1.28434420e+00 -1.29913449e+00 -1.04598486e+00 -3.55012983e-01
-1.31353688e+00 8.98128629e-01 5.85316539e-01 4.10357714e-01
7.82017887e-01 4.79535341e-01 7.09200978e-01 -1.93748683e-01
-1.19561732e+00 -2.67333925e-01 2.07392856e-01 3.66226435e-01
7.45134890e-01 3.71985018e-01 6.26432570e-03 1.28759396e+00
-5.99051058e-01 -3.81348997e-01 9.63264555e-02 6.16111636e-01
-2.83415258e-01 -8.68882835e-01 -2.29895920e-01 1.07586659e-01
-4.93190676e-01 -1.21149294e-01 -9.07760024e-01 3.78224283e-01
-1.13951802e-01 1.70106649e+00 2.31122836e-01 -5.05824685e-01
5.41152775e-01 6.75167203e-01 -2.33962134e-01 -4.99557585e-01
-8.01108778e-01 8.05692077e-02 6.47221208e-01 -5.89598119e-01
-2.96834260e-01 -6.61597311e-01 -1.60317814e+00 -4.56975922e-02
-7.06578791e-02 6.69179559e-01 8.96083117e-01 1.11996865e+00
4.68311399e-01 7.17732310e-01 9.97358084e-01 -5.57054877e-01
-3.68757308e-01 -1.16532326e+00 -7.56647587e-01 7.81176984e-01
5.56050599e-01 -4.32994574e-01 -8.35871816e-01 2.62616843e-01] | [13.40550422668457, 5.9009504318237305] |
fe3c4732-92a4-4656-b8c0-ef2befc5cfd0 | dense-cnn-learning-with-equivalent-mappings | 1605.07251 | null | http://arxiv.org/abs/1605.07251v1 | http://arxiv.org/pdf/1605.07251v1.pdf | Dense CNN Learning with Equivalent Mappings | Large receptive field and dense prediction are both important for achieving
high accuracy in pixel labeling tasks such as semantic segmentation. These two
properties, however, contradict with each other. A pooling layer (with stride
2) quadruples the receptive field size but reduces the number of predictions to
25\%. Some existing methods lead to dense predictions using computations that
are not equivalent to the original model. In this paper, we propose the
equivalent convolution (eConv) and equivalent pooling (ePool) layers, leading
to predictions that are both dense and equivalent to the baseline CNN model.
Dense prediction models learned using eConv and ePool can transfer the baseline
CNN's parameters as a starting point, and can inverse transfer the learned
parameters in a dense model back to the original one, which has both fast
testing speed and high accuracy. The proposed eConv and ePool layers have
achieved higher accuracy than baseline CNN in various tasks, including semantic
segmentation, object localization, image categorization and apparent age
estimation, not only in those tasks requiring dense pixel labeling. | ['Jian-Hao Luo', 'Chen-Wei Xie', 'Jianxin Wu'] | 2016-05-24 | null | null | null | null | ['image-categorization'] | ['computer-vision'] | [ 2.64586389e-01 5.39017737e-01 -7.20395222e-02 -7.83491373e-01
6.23727925e-02 -1.47677764e-01 3.30404878e-01 -1.21207185e-01
-6.34800673e-01 7.30664849e-01 -2.77738981e-02 -1.11847922e-01
2.51660854e-01 -1.07519424e+00 -9.72893953e-01 -6.16170943e-01
2.54066378e-01 3.06606740e-01 8.68526936e-01 2.20539913e-01
3.55118185e-01 3.39705914e-01 -1.78242540e+00 3.37279111e-01
9.80648816e-01 1.55623519e+00 4.97707248e-01 5.62546134e-01
-3.07496637e-01 8.54455888e-01 -2.82192498e-01 -5.41236520e-01
2.18914270e-01 1.36963800e-01 -1.12534165e+00 -1.81780621e-01
4.57239211e-01 -3.83584976e-01 -1.43637016e-01 1.03571224e+00
3.55520815e-01 2.21150974e-03 8.09747696e-01 -1.25384402e+00
-9.11332130e-01 6.89250588e-01 -6.16873622e-01 -1.32277504e-01
-1.55966237e-01 -3.20138991e-01 6.11253381e-01 -7.17998624e-01
2.66206086e-01 1.37762439e+00 1.12059379e+00 7.85734057e-01
-8.66613746e-01 -8.95523429e-01 3.70410621e-01 9.39285457e-02
-1.27710223e+00 -8.05349424e-02 2.47646406e-01 -4.65103030e-01
1.06512022e+00 -4.06729691e-02 6.32853806e-01 8.85912597e-01
1.17739566e-01 7.46486545e-01 1.19267488e+00 -1.86326638e-01
1.24746129e-01 2.62885362e-01 3.22954386e-01 6.92755461e-01
1.93029344e-01 -7.32703805e-02 -3.42519671e-01 3.37557554e-01
9.37068641e-01 3.23498368e-01 1.98784005e-02 7.87945613e-02
-1.09520602e+00 6.91177249e-01 7.90353358e-01 1.35004684e-01
-3.80299211e-01 3.12049061e-01 1.82856470e-01 -6.68695793e-02
4.07310694e-01 4.28222537e-01 -5.95765054e-01 2.51799077e-01
-7.27672279e-01 3.22194040e-01 5.66873491e-01 1.04200792e+00
1.19318533e+00 -9.40135494e-02 -5.49201548e-01 9.94710088e-01
1.02311447e-01 2.96753228e-01 8.70061576e-01 -1.07510948e+00
1.64548934e-01 8.58969808e-01 2.19288394e-02 -7.31042206e-01
-5.40048838e-01 -1.73365623e-01 -1.02257907e+00 1.64682567e-01
4.68594879e-01 -8.13541040e-02 -1.37371647e+00 1.77994347e+00
1.18554413e-01 3.82396907e-01 -4.34757210e-02 7.13894546e-01
9.94681597e-01 7.65043557e-01 5.44846117e-01 2.61431247e-01
1.30259860e+00 -1.26006293e+00 -2.69693613e-01 -6.24952972e-01
5.82080066e-01 -6.99153483e-01 1.10291553e+00 2.33352661e-01
-9.08858478e-01 -1.06094635e+00 -9.38196003e-01 -3.83329779e-01
-6.39995933e-01 2.86569715e-01 1.00582027e+00 6.06413722e-01
-1.32618332e+00 8.86981428e-01 -6.75440431e-01 -4.75056201e-01
8.07365894e-01 8.57423902e-01 -4.26507115e-01 -8.76016542e-02
-9.97536719e-01 7.18965650e-01 8.39361191e-01 1.57044008e-01
-4.50926751e-01 -6.42997801e-01 -7.62884557e-01 1.57366976e-01
-1.04242094e-01 -5.64666927e-01 1.13909757e+00 -1.03098428e+00
-1.41398311e+00 9.16555464e-01 -1.54899016e-01 -5.89935005e-01
3.16917270e-01 -2.37766623e-01 -2.01512640e-03 -1.84875995e-01
-1.39510110e-02 1.77033508e+00 7.49357462e-01 -7.38129258e-01
-7.28387475e-01 -3.21639359e-01 1.02875277e-01 -2.58151647e-02
-6.07235968e-01 -1.29696429e-01 -4.61900294e-01 -4.13726956e-01
3.73224288e-01 -6.21352494e-01 -3.61928463e-01 1.26583040e-01
-3.13852370e-01 -6.07810855e-01 6.54374242e-01 -7.33708262e-01
8.96521449e-01 -2.03011727e+00 -2.18834385e-01 9.04544517e-02
1.94259137e-01 3.56771648e-01 -8.77624601e-02 -3.87039691e-01
-7.97545724e-03 1.72998205e-01 -1.15048394e-01 -4.04735178e-01
-7.56001770e-02 2.43596569e-01 -1.97310537e-01 1.45352170e-01
4.02444899e-01 1.05201161e+00 -4.89199728e-01 -6.91428006e-01
7.98972994e-02 3.23414564e-01 -6.84671223e-01 3.04631799e-01
-1.63510025e-01 3.05670649e-01 -3.23124439e-01 5.39687574e-01
9.32893455e-01 -3.28424603e-01 -1.82745934e-01 -2.56309867e-01
-5.14570065e-02 -7.50301108e-02 -9.20991540e-01 1.44572151e+00
-4.55750227e-01 4.62228715e-01 -2.79793590e-01 -1.12669253e+00
1.40800905e+00 3.34095396e-02 1.06127523e-02 -9.33369100e-01
2.96794415e-01 3.50644499e-01 -7.69877732e-02 -2.59010345e-01
5.20012259e-01 4.63942923e-02 7.88605884e-02 1.46951407e-01
3.84485126e-01 4.79668155e-02 1.09913550e-01 -1.73883319e-01
7.88782001e-01 3.48872870e-01 -3.48345749e-02 -4.32521224e-01
5.81038535e-01 -3.03583562e-01 1.00499070e+00 7.94101357e-01
-2.02186272e-01 8.21664751e-01 6.67235255e-01 -6.77590430e-01
-1.29338932e+00 -8.24530721e-01 -2.20156908e-01 9.93615866e-01
3.57226789e-01 -1.81176603e-01 -1.13408518e+00 -6.81163013e-01
-7.50969425e-02 2.40446240e-01 -9.00364101e-01 -8.92965868e-02
-4.80121642e-01 -7.70221770e-01 4.84421939e-01 1.13207352e+00
1.13517904e+00 -1.38042843e+00 -5.66020846e-01 1.88399944e-02
-1.95796601e-02 -1.36671424e+00 -5.12507223e-02 2.75410920e-01
-1.01826596e+00 -9.24613237e-01 -8.83351445e-01 -1.24494076e+00
1.07875347e+00 -2.79659957e-01 9.21090305e-01 2.34324515e-01
-2.61250675e-01 -3.39336127e-01 -2.40701735e-01 -6.76171124e-01
-1.02206804e-01 3.69494677e-01 -6.34263605e-02 -9.95599329e-02
6.26683056e-01 -4.22498196e-01 -8.05392921e-01 4.66249824e-01
-6.20459139e-01 5.45818925e-01 7.22556233e-01 9.57060158e-01
7.57793725e-01 -2.30115369e-01 6.17517471e-01 -1.12759030e+00
1.15166530e-01 -1.65033460e-01 -6.01972044e-01 1.95536822e-01
-7.90404081e-01 3.79358381e-02 6.27196670e-01 -5.34227490e-01
-1.17126787e+00 2.35565767e-01 -6.53397381e-01 -1.61453262e-01
-3.17819476e-01 -1.24032475e-01 8.98456573e-02 -1.97134539e-01
4.77548122e-01 -1.63278356e-01 -7.72203729e-02 -5.59536755e-01
-1.08666606e-02 7.23460138e-01 3.79734099e-01 -2.29947105e-01
5.04468918e-01 1.59734130e-01 -2.14833021e-01 -3.90680134e-01
-1.02397072e+00 -1.24546714e-01 -8.68385434e-01 1.70010075e-01
1.04670787e+00 -9.04925048e-01 -9.58692849e-01 1.08054543e+00
-1.19260287e+00 -6.12260580e-01 -3.80938470e-01 5.03214300e-01
-5.38232088e-01 9.04226974e-02 -6.83346570e-01 -4.16454077e-01
-4.54502732e-01 -1.08986175e+00 9.66602981e-01 7.52627313e-01
-8.14396441e-02 -8.32192302e-01 -5.33902347e-01 2.02430367e-01
5.61209798e-01 -1.66577831e-01 9.69332874e-01 -4.52079296e-01
-3.03907245e-01 -2.73238838e-01 -8.80625010e-01 6.58184826e-01
-4.31357384e-01 -1.99610740e-01 -1.39062297e+00 1.05067855e-02
-3.34171355e-01 -5.25319159e-01 1.28744948e+00 7.66367912e-01
1.98036861e+00 -5.96675426e-02 -5.38174629e-01 9.04029071e-01
1.29625249e+00 8.53741989e-02 9.53899562e-01 1.65315360e-01
6.63476706e-01 6.55296803e-01 5.74301541e-01 1.28658876e-01
2.11801261e-01 3.11395854e-01 4.13128793e-01 -4.70266491e-01
-3.09827656e-01 -3.74327660e-01 -5.04224701e-03 5.34262180e-01
-2.95758843e-01 8.28864574e-02 -8.00378263e-01 5.86723089e-01
-1.81274235e+00 -5.89423954e-01 -2.59361062e-02 2.00447035e+00
7.34097302e-01 4.69205856e-01 -1.97444677e-01 -1.85673814e-02
6.98646009e-01 -4.61673364e-02 -6.30358219e-01 -6.83675945e-01
2.82029565e-02 4.57964957e-01 8.82501841e-01 2.69608259e-01
-1.18109858e+00 1.31901968e+00 6.98645210e+00 8.75090957e-01
-1.18924320e+00 3.51894975e-01 1.21015310e+00 5.05102158e-01
-4.28484902e-02 -1.78851083e-01 -1.37380707e+00 3.33300382e-01
7.90173113e-01 3.37418199e-01 1.18313730e-01 1.09052885e+00
-3.48123878e-01 -3.01089823e-01 -1.05851161e+00 9.25234735e-01
-4.15620767e-02 -1.40220010e+00 -3.59640606e-02 -1.78721979e-01
8.13188970e-01 -4.26523462e-02 1.26660213e-01 5.59759617e-01
2.47456878e-01 -1.33331430e+00 6.54835105e-01 4.58375126e-01
7.59863913e-01 -4.76964265e-01 1.27455509e+00 2.97873765e-01
-1.05347002e+00 -3.94226134e-01 -9.87642944e-01 -3.01702678e-01
-1.51345626e-01 4.46729302e-01 -7.94503987e-01 -8.14643502e-02
1.20255256e+00 6.63303256e-01 -9.36574399e-01 1.04742420e+00
-2.07449108e-01 5.08332491e-01 -2.91408509e-01 -2.27204740e-01
2.56398082e-01 -2.42618285e-03 -3.82821262e-01 9.88646388e-01
4.65739757e-01 -2.07276598e-01 -3.10368873e-02 1.10990393e+00
-2.84088284e-01 1.47528395e-01 -3.35968733e-01 9.80818197e-02
3.05059075e-01 1.28123772e+00 -1.05958939e+00 -4.30662006e-01
-5.43361545e-01 9.51859176e-01 6.05792046e-01 2.89624333e-01
-7.27904558e-01 -5.09534419e-01 5.83489478e-01 1.24148212e-01
3.88674855e-01 2.00046971e-01 -5.64075470e-01 -7.87562013e-01
-1.71219453e-01 -1.54923841e-01 -9.70844086e-03 -7.33314455e-01
-1.11820865e+00 7.85742164e-01 -1.23259969e-01 -9.15896893e-01
-3.33064012e-02 -1.20023394e+00 -7.14245200e-01 1.12924123e+00
-1.69627666e+00 -1.39582407e+00 -6.09609485e-01 5.92135727e-01
4.60889488e-01 -4.96305116e-02 9.51048613e-01 3.72877061e-01
-6.56481862e-01 6.90582871e-01 -3.22206914e-01 2.54980057e-01
6.07259274e-01 -1.26684129e+00 4.49449807e-01 7.06624150e-01
-2.77814984e-01 3.41871083e-01 2.07993254e-01 -4.23075795e-01
-5.77533901e-01 -1.36807680e+00 9.86012518e-01 -1.52544171e-01
1.68501854e-01 -5.38403213e-01 -8.82623434e-01 7.56214440e-01
-5.95564209e-02 2.28286654e-01 4.82051551e-01 2.62669712e-01
-2.23070204e-01 -3.95598769e-01 -1.08947408e+00 3.62113804e-01
1.01785743e+00 -1.55066118e-01 -4.29568797e-01 3.02032351e-01
9.34251547e-01 -3.30406517e-01 -8.65489662e-01 5.33439040e-01
6.84519887e-01 -1.02874541e+00 1.03813219e+00 -4.07069743e-01
6.24897182e-01 -5.41836880e-02 1.08696800e-02 -7.90226281e-01
-5.17385662e-01 1.62861407e-01 2.34828591e-01 1.16324246e+00
5.50828934e-01 -7.15875804e-01 1.07707655e+00 5.83507180e-01
-2.45223492e-01 -9.80175912e-01 -4.16234642e-01 -5.29841423e-01
2.70486355e-01 -4.27701741e-01 7.73760498e-01 4.18066442e-01
-5.04591823e-01 6.81162626e-02 -2.82004714e-01 4.69491668e-02
2.64109999e-01 7.41104037e-02 4.77364361e-01 -1.58189225e+00
-9.50616151e-02 -3.28471422e-01 -6.23851418e-01 -1.17121696e+00
3.17244530e-01 -7.10864067e-01 2.67506242e-01 -1.72260630e+00
3.43019992e-01 -7.21868396e-01 -4.18385416e-01 8.95612478e-01
-1.98181495e-01 8.49301696e-01 -8.02223906e-02 1.01995468e-03
-3.94278675e-01 3.39769095e-01 1.24749780e+00 -2.00701296e-01
-1.43662002e-02 2.98300944e-02 -6.85533762e-01 1.15610218e+00
7.93739736e-01 -1.70208633e-01 -2.44824380e-01 -4.97302383e-01
6.62657917e-02 -5.84172904e-01 2.92387187e-01 -1.29546809e+00
1.72838435e-01 -1.16704563e-02 1.01222360e+00 -5.07114112e-01
3.62956733e-01 -5.95131099e-01 -1.54216453e-01 5.99977434e-01
-2.71295041e-01 -2.93956578e-01 1.69111237e-01 1.67578787e-01
-3.95635128e-01 -5.66912949e-01 9.48796153e-01 -2.78948456e-01
-1.43087375e+00 6.85091734e-01 8.95954203e-04 -3.29649806e-01
1.10797679e+00 -7.29872465e-01 -3.91248047e-01 3.24316062e-02
-8.12207103e-01 2.90969402e-01 2.42747068e-01 5.12698710e-01
5.84471822e-01 -1.22792280e+00 -3.86116236e-01 4.35357958e-01
-2.94833817e-03 2.48837516e-01 3.90391082e-01 8.61960590e-01
-8.37203205e-01 2.98772812e-01 -7.37360179e-01 -5.74080110e-01
-9.63862956e-01 5.38482785e-01 3.03363353e-01 -1.95893943e-01
-4.91094410e-01 1.37242985e+00 6.42796397e-01 -8.54754567e-01
4.15134668e-01 -4.16006595e-01 -5.75873435e-01 -1.54117495e-01
5.97923279e-01 9.74302739e-02 -1.94793269e-01 -3.62447172e-01
-1.33370981e-01 9.17327583e-01 -1.19212084e-01 5.28710842e-01
1.26086116e+00 -7.48050809e-02 -1.66687667e-01 1.23622879e-01
1.05259490e+00 -5.68574011e-01 -1.52430022e+00 -9.17039365e-02
-1.74104214e-01 -3.29490215e-01 -3.46590169e-02 -6.51135683e-01
-1.10469687e+00 1.39893639e+00 8.83351743e-01 -2.51697656e-02
1.15890706e+00 5.63682839e-02 7.92115867e-01 2.38255903e-01
2.29199529e-01 -1.38638830e+00 5.62578402e-02 5.83160937e-01
5.33038557e-01 -1.33288169e+00 -2.64480203e-01 -8.11397433e-01
-5.24216294e-01 1.16411626e+00 1.36816072e+00 -2.11804658e-01
7.84454584e-01 2.53840208e-01 -1.05707183e-01 2.19519079e-01
-4.16088402e-01 -2.00780258e-01 1.09759748e-01 7.37406433e-01
4.67794687e-01 1.25347778e-01 -2.76983619e-01 8.83024216e-01
-2.70711422e-01 2.29408503e-01 1.35486960e-01 4.84299004e-01
-6.51232719e-01 -1.01596463e+00 -1.14960201e-01 7.45551705e-01
-3.98420930e-01 -4.06537317e-02 6.08191602e-02 4.55105871e-01
8.09291303e-01 4.25519794e-01 6.91265881e-01 -5.23599088e-01
1.67592749e-01 1.60353586e-01 4.13055271e-01 -5.47499835e-01
-4.77530718e-01 -3.13928038e-01 -2.81804055e-01 -6.36977077e-01
-3.09963316e-01 -1.12443477e-01 -1.36191881e+00 -3.52911830e-01
-3.34860265e-01 -4.54194918e-02 6.24317586e-01 8.98724496e-01
1.21347271e-01 6.31558955e-01 2.68183410e-01 -8.36900532e-01
-3.52392018e-01 -1.32775533e+00 -8.26808631e-01 3.63922060e-01
-2.61744291e-01 -6.99403763e-01 4.46530804e-02 1.46699771e-01] | [9.49726676940918, 0.31758517026901245] |
425e45e7-a35f-49cd-aab2-37d401df4003 | network-cooperation-with-progressive | 2002.11919 | null | https://arxiv.org/abs/2002.11919v1 | https://arxiv.org/pdf/2002.11919v1.pdf | Network Cooperation with Progressive Disambiguation for Partial Label Learning | Partial Label Learning (PLL) aims to train a classifier when each training instance is associated with a set of candidate labels, among which only one is correct but is not accessible during the training phase. The common strategy dealing with such ambiguous labeling information is to disambiguate the candidate label sets. Nonetheless, existing methods ignore the disambiguation difficulty of instances and adopt the single-trend training mechanism. The former would lead to the vulnerability of models to the false positive labels and the latter may arouse error accumulation problem. To remedy these two drawbacks, this paper proposes a novel approach termed "Network Cooperation with Progressive Disambiguation" (NCPD) for PLL. Specifically, we devise a progressive disambiguation strategy of which the disambiguation operations are performed on simple instances firstly and then gradually on more complicated ones. Therefore, the negative impacts brought by the false positive labels of complicated instances can be effectively mitigated as the disambiguation ability of the model has been strengthened via learning from the simple instances. Moreover, by employing artificial neural networks as the backbone, we utilize a network cooperation mechanism which trains two networks collaboratively by letting them interact with each other. As two networks have different disambiguation ability, such interaction is beneficial for both networks to reduce their respective disambiguation errors, and thus is much better than the existing algorithms with single-trend training process. Extensive experimental results on various benchmark and practical datasets demonstrate the superiority of our NCPD to other state-of-the-art PLL methods. | ['Chen gong', 'Jian Yang', 'Jiehui Deng', 'Yao Yao'] | 2020-02-22 | null | null | null | null | ['partial-label-learning'] | ['methodology'] | [ 3.02952558e-01 2.28187531e-01 -2.85467327e-01 -1.81527957e-01
-8.70516002e-02 -2.89832413e-01 4.10710663e-01 1.46150559e-01
-5.44107378e-01 9.92735386e-01 -4.52682287e-01 -1.86418548e-01
-4.26079661e-01 -8.79492164e-01 -3.33971739e-01 -1.07030594e+00
1.56240970e-01 4.60213929e-01 4.05861169e-01 -1.15555078e-01
9.38952416e-02 1.97130635e-01 -1.62834680e+00 -4.72387820e-02
1.34464192e+00 1.01184905e+00 2.20409527e-01 -3.05045038e-01
-4.55551505e-01 8.56332242e-01 -6.97143972e-01 -3.33677024e-01
1.88450649e-01 -4.02248539e-02 -6.68257296e-01 1.04782350e-01
-1.85553003e-02 1.03709407e-01 1.13820679e-01 1.31062913e+00
4.35878247e-01 1.33883938e-01 2.63413101e-01 -1.40113068e+00
-4.52780366e-01 7.34729171e-01 -6.88654006e-01 1.12156697e-01
-3.89040355e-03 3.38199064e-02 9.53980744e-01 -7.84386277e-01
3.19145292e-01 1.02365267e+00 7.48208165e-01 5.13426304e-01
-9.56945956e-01 -1.02341783e+00 7.32975006e-01 1.07383661e-01
-1.58315611e+00 -1.10955492e-01 1.08792889e+00 -2.42793486e-01
3.67103696e-01 5.06525338e-02 1.94797933e-01 7.52815604e-01
-1.73392519e-01 6.97824061e-01 1.31274819e+00 -3.05980355e-01
1.12221636e-01 4.55605447e-01 4.41834927e-01 7.05580056e-01
3.40588748e-01 -6.96526607e-04 -1.62544578e-01 2.27396656e-03
4.34180409e-01 1.56692073e-01 -3.53424430e-01 -2.21953005e-01
-9.78743732e-01 4.84309971e-01 6.33310199e-01 5.83196759e-01
-4.33605760e-01 -5.91622174e-01 3.21401626e-01 3.00467074e-01
4.38316375e-01 4.73611802e-01 -6.51140273e-01 4.63927388e-01
-6.14920616e-01 -2.15811834e-01 6.09489560e-01 7.22399473e-01
1.07966888e+00 -9.67672467e-02 -8.39647129e-02 9.19305146e-01
3.49010825e-01 2.05835421e-02 6.88253641e-01 -4.61475521e-01
4.83265787e-01 1.22338724e+00 1.84072759e-02 -1.22965991e+00
-4.39858854e-01 -1.05137312e+00 -1.04444146e+00 -5.34057580e-02
3.16493481e-01 -3.24747115e-01 -7.61046827e-01 1.90997601e+00
5.20126879e-01 4.30500776e-01 1.67420179e-01 8.48882079e-01
8.38743627e-01 6.39762521e-01 4.17106926e-01 -6.50973499e-01
1.03860140e+00 -9.31940436e-01 -6.96113527e-01 -2.23973513e-01
7.48492897e-01 -6.72230005e-01 7.85100639e-01 5.41542113e-01
-6.37267351e-01 -7.85672009e-01 -1.07530963e+00 6.45886004e-01
-2.80306071e-01 1.41554162e-01 6.67465270e-01 5.24962425e-01
-7.86030650e-01 5.35522878e-01 -2.20523149e-01 -9.10984352e-03
3.69264096e-01 5.40951192e-01 -1.36796117e-01 -3.08706462e-02
-1.53769052e+00 7.72420049e-01 7.85977066e-01 3.97524238e-01
-3.50406319e-01 -4.66103613e-01 -3.80058169e-01 5.22065014e-02
9.58427846e-01 -2.68300265e-01 9.90746081e-01 -1.13599861e+00
-1.00917137e+00 4.66565669e-01 1.47092417e-01 -1.05630934e-01
4.86237586e-01 -4.97140456e-04 -7.00208545e-01 -2.06341296e-01
2.29867563e-01 6.15103304e-01 6.72543347e-01 -1.56945610e+00
-1.05184698e+00 -8.10655355e-02 1.41229317e-01 4.82016206e-01
-5.87661564e-01 -3.43089759e-01 -3.61326426e-01 -4.56813008e-01
4.68695581e-01 -6.99040353e-01 -1.98437154e-01 -2.90265054e-01
-4.53870356e-01 -6.97281778e-01 1.01732695e+00 -1.09004691e-01
1.46052194e+00 -2.05289268e+00 -1.66026905e-01 4.52183485e-01
2.94928849e-01 7.32683122e-01 -6.42031133e-02 5.04310206e-02
-1.97404757e-01 2.59550780e-01 -7.51512423e-02 -9.30411890e-02
-2.45867744e-01 5.51282227e-01 -5.84957749e-02 -1.89667419e-02
3.13071847e-01 3.12508404e-01 -1.06490755e+00 -7.06384242e-01
1.26922699e-02 9.09784213e-02 -2.19947681e-01 1.52786151e-01
-1.65627703e-01 5.53018510e-01 -7.32401550e-01 6.34628177e-01
7.77001798e-01 -3.57918799e-01 3.58697236e-01 -3.51155758e-01
-3.84625420e-02 1.16966687e-01 -1.53028345e+00 9.65309620e-01
-3.13252568e-01 -2.52020329e-01 -9.15572420e-03 -1.27988350e+00
1.10438812e+00 3.94571215e-01 4.91997093e-01 -5.86246431e-01
2.33072102e-01 3.43836159e-01 1.31950989e-01 -5.15365243e-01
9.99483094e-02 -2.74276257e-01 1.91512808e-01 3.47409785e-01
-7.85986483e-02 5.01503885e-01 3.49853724e-01 4.17108312e-02
7.74956405e-01 -2.83765662e-02 1.63124964e-01 -1.27299562e-01
9.78012741e-01 -1.54527515e-01 1.15682888e+00 7.74667680e-01
-3.81195694e-01 -1.26404718e-01 3.48092794e-01 -5.37883103e-01
-3.51656377e-01 -8.12293768e-01 -1.53006896e-01 1.02943850e+00
5.76874912e-01 -1.12949438e-01 -4.84861940e-01 -1.27322567e+00
-1.56870112e-01 4.76195365e-01 -3.86835545e-01 -3.43777984e-01
-5.04299164e-01 -1.01098478e+00 3.64109427e-01 3.91477495e-01
1.07762980e+00 -1.10246694e+00 -6.44011870e-02 3.35644692e-01
-2.12667286e-01 -7.75909305e-01 -1.08352251e-01 3.14329922e-01
-7.90088892e-01 -1.21440005e+00 -2.62334675e-01 -1.09980249e+00
9.43607807e-01 4.57314759e-01 8.10301781e-01 6.54110014e-01
3.79592270e-01 -1.14466518e-01 -4.49879766e-01 -3.09895068e-01
-3.83103609e-01 1.49149418e-01 2.69674480e-01 2.25294575e-01
4.45414782e-01 -6.67485893e-01 -3.43327016e-01 6.50018990e-01
-9.19687927e-01 8.42525363e-02 9.28890228e-01 1.01587355e+00
5.18674254e-01 8.09030473e-01 1.13943660e+00 -1.02429831e+00
6.16740644e-01 -6.57405496e-01 -4.94829535e-01 4.15283591e-01
-1.12346387e+00 -3.66687961e-02 8.90470743e-01 -7.07204103e-01
-1.31970549e+00 -1.05263460e-02 4.94976640e-02 -2.76473582e-01
-1.85632482e-01 7.75817692e-01 -3.42169017e-01 -2.10598379e-01
4.58662450e-01 1.39094964e-01 -7.01692924e-02 -3.91316801e-01
-2.91080084e-02 4.89857882e-01 3.41535449e-01 -6.68714881e-01
8.53661597e-01 -7.97539484e-03 -1.89991115e-04 -1.73120797e-01
-1.15940452e+00 -5.15839040e-01 -5.57069898e-01 -4.44772840e-01
4.05937642e-01 -7.48925924e-01 -6.14734709e-01 6.11903250e-01
-9.25782800e-01 7.15397149e-02 9.87553503e-03 3.25457036e-01
3.01133394e-01 3.33295047e-01 -4.59374934e-01 -6.86937511e-01
-2.08378628e-01 -1.11826563e+00 4.79594827e-01 6.80906177e-01
1.91072188e-02 -1.03105330e+00 -3.33243072e-01 2.83437252e-01
2.25239888e-01 6.18187226e-02 9.64702249e-01 -1.09419250e+00
-4.47707474e-01 -1.74663141e-01 -3.82429242e-01 4.34719294e-01
4.12597179e-01 -1.07056759e-01 -9.28697467e-01 -2.46343642e-01
9.16319713e-02 -2.78873563e-01 5.97622931e-01 -4.26402241e-02
9.44397092e-01 -3.40505779e-01 -5.71569622e-01 -3.61735187e-02
1.37731254e+00 4.97283995e-01 2.76037991e-01 3.72191459e-01
8.20200801e-01 5.82535684e-01 8.96049917e-01 1.96577847e-01
3.54306966e-01 3.43881518e-01 3.56619388e-01 -2.50627100e-01
2.18601778e-01 -1.12350874e-01 -7.33372048e-02 1.00660694e+00
-7.09186494e-02 -2.76451647e-01 -8.14150572e-01 3.41832638e-01
-1.83308733e+00 -5.63246131e-01 -1.87209591e-01 2.06656718e+00
1.01721764e+00 5.36790490e-01 -2.04471827e-01 5.44290781e-01
1.16210151e+00 2.82387529e-02 -4.33145851e-01 1.70925468e-01
-1.74611688e-01 -3.82010415e-02 7.49696270e-02 1.73539326e-01
-1.17889833e+00 7.82917857e-01 4.62784767e+00 1.11212265e+00
-1.29877627e+00 1.82845742e-01 8.26135576e-01 3.54423672e-01
3.12540606e-02 4.43302579e-02 -8.76266301e-01 6.30086601e-01
3.69592726e-01 1.75358146e-01 6.54792488e-02 8.39082479e-01
3.32678210e-05 -2.70415813e-01 -8.43253911e-01 5.11741042e-01
-2.12495923e-01 -8.96751106e-01 8.79113153e-02 -2.76023835e-01
8.81234646e-01 -3.32140356e-01 -1.25817806e-01 5.80156028e-01
3.05072993e-01 -5.66032529e-01 3.74669075e-01 5.43635607e-01
3.57375950e-01 -8.52148890e-01 1.09940219e+00 8.04089606e-01
-1.28914976e+00 -2.99101770e-01 -2.52754003e-01 -7.95457661e-02
-1.23444192e-01 7.90492952e-01 -8.17339778e-01 9.35454190e-01
4.79482532e-01 5.46441495e-01 -5.68241715e-01 9.84616697e-01
-4.94372636e-01 7.15012133e-01 -3.54446203e-01 -5.91355152e-02
2.89180219e-01 -1.23755276e-01 4.31413800e-01 6.72768116e-01
2.26326302e-01 2.20021904e-01 6.80970192e-01 6.78126335e-01
-1.49511039e-01 2.73592532e-01 -3.48457813e-01 3.64380836e-01
9.56889272e-01 1.45613372e+00 -8.97109210e-01 -4.82350051e-01
-3.85277659e-01 3.95461321e-01 4.57853168e-01 3.25248241e-01
-6.59169734e-01 -2.74400502e-01 -8.89806822e-02 -1.42634079e-01
5.52129291e-04 3.25417399e-01 -2.89871156e-01 -8.46930981e-01
1.51010588e-01 -7.07195818e-01 6.17538512e-01 -5.89687288e-01
-1.54081082e+00 5.34128606e-01 -1.28817901e-01 -1.36998796e+00
2.01618373e-01 -3.22831213e-01 -7.98504651e-01 7.90577590e-01
-1.72216105e+00 -9.73474264e-01 -3.40186059e-01 4.37552184e-01
3.26259732e-01 -8.17310736e-02 6.01149142e-01 5.98125696e-01
-9.17942643e-01 6.81831360e-01 -2.25964636e-01 6.28622202e-03
7.32450724e-01 -1.01092339e+00 -5.26642561e-01 7.94540942e-01
-1.84397846e-01 6.78246796e-01 3.68484020e-01 -7.66774714e-01
-6.40160978e-01 -1.25964749e+00 1.00483334e+00 6.86055869e-02
5.08967519e-01 1.40769780e-01 -1.28953207e+00 4.08481389e-01
-1.17344670e-01 -4.18048687e-02 5.82666576e-01 2.17401683e-01
-2.17722371e-01 -4.39279169e-01 -1.05630589e+00 5.63429117e-01
9.08915520e-01 -3.04546673e-02 -6.85832858e-01 3.23426634e-01
5.93405545e-01 -1.46255493e-01 -8.39399278e-01 8.44849885e-01
1.14800781e-01 -8.39062333e-01 7.69949794e-01 -2.06116155e-01
1.78854361e-01 -6.94978356e-01 3.53789926e-01 -1.33117533e+00
-3.78131032e-01 -2.56657571e-01 1.10522211e-01 1.60361838e+00
4.07054603e-01 -9.86717761e-01 5.93200266e-01 5.27526796e-01
-2.32947588e-01 -9.01838720e-01 -8.62573802e-01 -8.02365720e-01
-2.29778156e-01 -4.17790897e-02 7.85867512e-01 1.58176219e+00
5.83369657e-02 5.22337735e-01 -2.77805716e-01 3.95040780e-01
4.26047444e-01 8.91797766e-02 2.97867596e-01 -1.72253013e+00
-8.09179768e-02 -4.67550516e-01 -3.74966003e-02 -8.21100652e-01
1.61091775e-01 -8.31281900e-01 2.21449137e-01 -1.10006750e+00
3.90238017e-02 -1.24104345e+00 -8.17106426e-01 9.62334514e-01
-7.02748537e-01 3.56457755e-02 -1.84747472e-03 5.91675460e-01
-9.61219251e-01 3.83430719e-01 1.41363561e+00 -3.67861353e-02
-3.86264026e-01 1.15978770e-01 -7.60551512e-01 8.47928464e-01
8.30959320e-01 -6.69381678e-01 -6.62481248e-01 -1.50563166e-01
2.37317279e-01 -4.61250357e-02 1.88102841e-01 -1.15093172e+00
5.60852885e-01 -1.57229275e-01 1.80889949e-01 -5.02951026e-01
-1.37272328e-01 -1.00349283e+00 1.30230948e-01 5.97052395e-01
-2.81930774e-01 -2.46149421e-01 9.58459917e-03 6.05769634e-01
-1.74465328e-01 -4.63137329e-01 6.33892000e-01 -1.16249740e-01
-7.89112926e-01 2.34819427e-01 -2.39544287e-01 -1.40734479e-01
1.02509332e+00 -1.79390743e-01 -5.26587784e-01 6.88922107e-02
-8.18636477e-01 6.97799623e-01 1.05269648e-01 2.92567134e-01
3.29468399e-01 -1.36436200e+00 -3.80577534e-01 1.45752311e-01
-2.61371881e-02 3.98869365e-01 2.55112529e-01 9.86567736e-01
1.67390719e-01 1.91165671e-01 -9.91567299e-02 -5.65798342e-01
-1.11691296e+00 7.44435847e-01 2.98943371e-01 -6.54823124e-01
-1.14286356e-01 8.03997934e-01 2.63574421e-01 -5.82694948e-01
3.48481774e-01 -9.70987231e-03 -8.02704453e-01 2.28696883e-01
1.54786512e-01 1.87456965e-01 1.36124164e-01 -5.60935259e-01
-2.47374773e-01 3.76347095e-01 -5.11494517e-01 5.74770451e-01
1.00096977e+00 -2.34808847e-01 -2.20146790e-01 2.90690511e-01
5.95770538e-01 -1.52325556e-01 -1.11086857e+00 -7.79683709e-01
1.69853419e-01 -4.78915349e-02 -1.28514376e-02 -9.98286188e-01
-1.33415306e+00 4.77793694e-01 4.99801278e-01 6.01580262e-01
1.42342722e+00 -3.15467715e-01 5.67253947e-01 5.22378743e-01
4.79577690e-01 -1.19921601e+00 1.83147222e-01 4.64341640e-01
3.39788407e-01 -1.23689282e+00 -1.29096895e-01 -8.25630844e-01
-4.55833763e-01 9.06129956e-01 1.16973984e+00 6.61766455e-02
4.35723186e-01 -6.60186633e-02 3.81993540e-02 -1.01073064e-01
-6.79440796e-01 -2.24496275e-02 2.28051692e-01 3.25388223e-01
1.50546581e-01 -8.24807510e-02 -6.88726723e-01 9.38849568e-01
3.24217528e-01 -8.83944482e-02 1.57019854e-01 9.63896990e-01
-5.76036096e-01 -1.19786060e+00 -4.35301185e-01 4.52189177e-01
-2.56209850e-01 -8.24723616e-02 -2.44449243e-01 8.76753867e-01
8.49965334e-01 1.10673535e+00 -2.21400097e-01 -6.19719863e-01
2.85086900e-01 2.02539623e-01 -6.34307191e-02 -5.61365426e-01
-6.68015659e-01 1.62158161e-01 1.35789946e-01 -1.27166256e-01
-8.65967393e-01 -4.57319766e-01 -1.37689507e+00 -4.70544063e-02
-7.68386900e-01 4.09473628e-01 8.31670687e-02 1.35503304e+00
1.96020693e-01 8.12486947e-01 8.69466782e-01 -4.40579206e-01
-5.79770267e-01 -9.81944084e-01 -4.48462129e-01 2.83703417e-01
2.83223074e-02 -1.09574950e+00 -4.48778361e-01 -1.42226472e-01] | [9.459600448608398, 3.9557602405548096] |
086576ba-0a3f-46f0-9ff2-15c7907a4b42 | valerian-invariant-feature-learning-for-imu | 2303.06048 | null | https://arxiv.org/abs/2303.06048v1 | https://arxiv.org/pdf/2303.06048v1.pdf | VALERIAN: Invariant Feature Learning for IMU Sensor-based Human Activity Recognition in the Wild | Deep neural network models for IMU sensor-based human activity recognition (HAR) that are trained from controlled, well-curated datasets suffer from poor generalizability in practical deployments. However, data collected from naturalistic settings often contains significant label noise. In this work, we examine two in-the-wild HAR datasets and DivideMix, a state-of-the-art learning with noise labels (LNL) method to understand the extent and impacts of noisy labels in training data. Our empirical analysis reveals that the substantial domain gaps among diverse subjects cause LNL methods to violate a key underlying assumption, namely, neural networks tend to fit simpler (and thus clean) data in early training epochs. Motivated by the insights, we design VALERIAN, an invariant feature learning method for in-the-wild wearable sensor-based HAR. By training a multi-task model with separate task-specific layers for each subject, VALERIAN allows noisy labels to be dealt with individually while benefiting from shared feature representation across subjects. We evaluated VALERIAN on four datasets, two collected in a controlled environment and two in the wild. | ['Rong Zheng', 'Boyu Wang', 'Yujiao Hao'] | 2023-03-03 | null | null | null | null | ['human-activity-recognition', 'human-activity-recognition'] | ['computer-vision', 'time-series'] | [ 2.92007834e-01 -1.21907875e-01 -1.05389692e-01 -5.25375605e-01
-9.40012932e-01 -3.09256852e-01 1.26737043e-01 -1.77450359e-01
-5.13458431e-01 7.76444435e-01 5.33650398e-01 3.91822219e-01
-1.60956442e-01 -2.31401205e-01 -8.69242549e-01 -4.67960536e-01
-1.42721951e-01 1.41876072e-01 -4.64717090e-01 2.32014641e-01
-7.28228450e-01 -1.77691747e-02 -1.45547295e+00 8.39709789e-02
4.74033594e-01 1.15810537e+00 -2.82635033e-01 2.69635439e-01
7.07788825e-01 9.01899517e-01 -7.91161597e-01 -3.07490602e-02
3.15056920e-01 -4.05671149e-01 -5.01995623e-01 8.59521106e-02
8.18948746e-01 -2.00409129e-01 -3.99908632e-01 7.07992196e-01
9.99694228e-01 3.69119853e-01 4.69029844e-01 -1.42422211e+00
-6.47873342e-01 4.26542133e-01 -5.68949729e-02 1.32256836e-01
3.36666882e-01 3.49292755e-01 7.58507013e-01 -7.46187568e-01
4.68212396e-01 9.44036484e-01 1.42543864e+00 7.41705298e-01
-1.63474417e+00 -7.57558942e-01 2.50141881e-02 -6.25780821e-02
-1.49051809e+00 -6.30674899e-01 8.29201758e-01 -7.20278561e-01
7.86395371e-01 1.70232460e-01 5.78304946e-01 2.27611446e+00
-6.17656717e-03 7.51928031e-01 1.22930622e+00 1.18926622e-01
5.48402309e-01 -1.81655556e-01 3.61413866e-01 2.97359079e-01
5.81169009e-01 3.42582911e-02 -9.74613070e-01 -3.80589753e-01
5.39026082e-01 2.91593313e-01 -1.56435221e-01 -2.91927606e-01
-1.39650857e+00 4.75605369e-01 4.32712525e-01 2.61642992e-01
-7.21307397e-01 4.00518358e-01 6.16492450e-01 1.34459600e-01
5.38210094e-01 6.21017814e-01 -6.47046685e-01 -4.47011381e-01
-9.98238087e-01 4.85321641e-01 7.56284416e-01 7.74295449e-01
5.28733253e-01 1.54059395e-01 -4.79033619e-01 9.07338917e-01
-1.08104749e-02 6.06510997e-01 6.05241179e-01 -9.43901420e-01
2.83726990e-01 3.63172114e-01 1.97313592e-01 -8.21391881e-01
-1.02469099e+00 -8.57553482e-01 -9.51089501e-01 -1.93267062e-01
4.56385076e-01 -7.70864725e-01 -8.33731413e-01 2.37385726e+00
1.25692636e-01 2.31449068e-01 -2.19274655e-01 1.03520012e+00
7.07092166e-01 5.97687811e-02 5.29928803e-01 2.18776539e-02
1.26006150e+00 -8.38617682e-01 -7.93257236e-01 -6.70255840e-01
7.40778744e-01 -1.00530289e-01 1.19593370e+00 5.94779372e-01
-5.84339499e-01 -7.14008391e-01 -1.14609671e+00 -7.17206076e-02
-2.53060967e-01 2.85080016e-01 6.79862261e-01 5.95364928e-01
-7.83706605e-01 7.51472592e-01 -1.00208259e+00 -6.53056562e-01
7.50778317e-01 2.10012183e-01 -5.93405545e-01 4.42663627e-03
-9.62485671e-01 6.28951788e-01 1.54170662e-01 2.86912680e-01
-1.28588629e+00 -6.09183133e-01 -8.49739850e-01 -1.88105524e-01
4.72980231e-01 -7.49935448e-01 1.39041531e+00 -9.64353085e-01
-1.03044939e+00 6.92879438e-01 9.46169067e-03 -4.60701197e-01
5.38115978e-01 -5.79315782e-01 -5.89073420e-01 -3.07846874e-01
3.02813977e-01 4.20075744e-01 7.23995864e-01 -1.01363361e+00
-1.78908736e-01 -5.55601120e-01 -4.89153564e-01 -1.51438296e-01
-3.25899750e-01 -2.95837313e-01 9.14245173e-02 -9.33899879e-01
-5.08038104e-02 -1.00127268e+00 -1.08913360e-02 -1.25026241e-01
-3.76211137e-01 -2.89270222e-01 3.83711666e-01 -7.79946923e-01
1.11533535e+00 -2.21997261e+00 -3.40561569e-02 2.14096308e-01
6.62893414e-01 1.31331310e-01 -1.61970779e-01 1.09808572e-01
-4.24943641e-02 -1.40372608e-02 2.85571143e-02 -6.42735481e-01
1.76090419e-01 3.85640711e-01 2.89415449e-01 7.56287158e-01
-6.53070351e-03 1.07247448e+00 -8.96577656e-01 -2.04402134e-01
-1.32422715e-01 5.22056103e-01 -4.41868454e-01 4.61080670e-01
-5.83250728e-03 7.06675649e-01 -3.90062898e-01 8.78353298e-01
2.48212829e-01 -2.89604127e-01 5.52712157e-02 -5.54149747e-01
3.03763092e-01 2.18790367e-01 -1.12748039e+00 2.02541661e+00
-2.21446171e-01 6.70734167e-01 2.92871322e-04 -1.06830192e+00
7.94804335e-01 4.41152900e-01 9.16904628e-01 -7.80622780e-01
4.67232674e-01 1.50486559e-01 -1.18099347e-01 -7.00006247e-01
1.38521895e-01 1.72990829e-01 -4.41715926e-01 4.19387609e-01
4.21285808e-01 8.24257076e-01 -2.51048338e-02 -3.14121306e-01
1.56673980e+00 2.32037947e-01 2.65149087e-01 -2.40006521e-01
-3.12674820e-01 -3.06647718e-01 1.11271822e+00 1.20198047e+00
-8.56961310e-01 6.19394958e-01 2.61979215e-02 -7.11100578e-01
-9.40317571e-01 -1.15185702e+00 -2.73883492e-01 1.42484951e+00
-2.99244225e-01 -5.34353375e-01 -9.89571214e-01 -5.93895614e-01
7.47545511e-02 2.42586881e-01 -1.05078506e+00 -3.00779492e-01
-4.41409737e-01 -7.81794190e-01 9.36904192e-01 7.27101803e-01
7.52488792e-01 -1.14344275e+00 -9.19715762e-01 3.97722483e-01
-4.40860450e-01 -1.28173637e+00 -4.09567654e-01 6.35736883e-01
-5.61398208e-01 -1.04783535e+00 -6.03056312e-01 -4.34426427e-01
2.84673154e-01 1.18835054e-01 1.26612747e+00 -3.70650053e-01
-2.10771203e-01 6.05365872e-01 -4.15898740e-01 -4.71672624e-01
1.72970071e-01 2.97874331e-01 6.05863392e-01 1.36161521e-01
8.70785177e-01 -1.01087344e+00 -5.35773158e-01 3.78681719e-01
-5.76674044e-01 -3.31332594e-01 4.95468020e-01 8.33949387e-01
6.20637417e-01 -4.35540974e-01 8.20630252e-01 -5.80476701e-01
6.71076775e-01 -8.45743239e-01 2.02789932e-01 1.81933073e-03
-5.42149961e-01 4.28109020e-02 4.29166973e-01 -8.22334647e-01
-6.35340810e-01 -1.79322734e-02 -1.02389678e-01 -3.27604562e-01
-5.47109902e-01 4.07432884e-01 -3.22803766e-01 9.95237380e-02
1.44243753e+00 -1.28514901e-01 -2.37825409e-01 -8.15683186e-01
8.35159719e-02 8.05638790e-01 9.62061107e-01 -8.94365966e-01
5.04879296e-01 1.56541109e-01 -1.97124645e-01 -8.35361898e-01
-1.28198934e+00 -4.39477652e-01 -4.89326239e-01 -1.66160703e-01
9.55611825e-01 -1.45999348e+00 -6.38089001e-01 6.37993693e-01
-7.39786148e-01 -7.10271835e-01 -5.61715007e-01 4.64151442e-01
-5.50067842e-01 -2.07321748e-01 -4.87420201e-01 -7.57691860e-01
-4.44031417e-01 -7.73129284e-01 1.11115682e+00 2.07269534e-01
-8.07125747e-01 -6.75035179e-01 3.10858279e-01 4.94218558e-01
4.75211293e-01 8.66212785e-01 2.12825030e-01 -1.03975964e+00
8.62315521e-02 -2.47298002e-01 1.22216202e-01 5.97384751e-01
1.60567939e-01 -9.35034573e-01 -1.60177803e+00 -3.80562901e-01
-1.89546011e-02 -9.17890787e-01 5.31348765e-01 2.87256390e-01
1.24510396e+00 -3.44560832e-01 -1.99527800e-01 7.72623658e-01
8.87469351e-01 -2.96362251e-01 4.82220590e-01 2.59041339e-01
8.84176075e-01 1.53006062e-01 1.32761568e-01 5.84337711e-01
4.19095457e-01 6.82610393e-01 -6.67463895e-03 -1.82855815e-01
-1.77633867e-01 -5.00251770e-01 3.89047086e-01 5.49725473e-01
9.70624853e-03 4.06070463e-02 -6.41624629e-01 6.08756185e-01
-2.10666561e+00 -7.40718305e-01 2.00685710e-01 2.09479046e+00
8.91261101e-01 -8.49758014e-02 5.87842703e-01 4.53742072e-02
2.09592044e-01 1.40164390e-01 -7.66215861e-01 1.75987422e-01
-2.38338470e-01 9.55182314e-02 5.27323902e-01 -2.14256838e-01
-1.44365287e+00 4.74679500e-01 5.99806023e+00 5.20255983e-01
-8.75266194e-01 6.66772723e-01 4.97850627e-01 -4.40021902e-01
2.92268485e-01 -4.73207831e-01 -5.81663430e-01 6.64348960e-01
1.28306782e+00 3.41940433e-01 5.92040598e-01 1.13429463e+00
2.44782791e-01 -6.81289881e-02 -1.55424714e+00 1.61558807e+00
1.76196292e-01 -8.29227865e-01 -5.84920585e-01 9.24227908e-02
7.22342968e-01 4.31635141e-01 -1.60109907e-01 5.51645219e-01
1.27074808e-01 -1.05324578e+00 8.95275056e-01 7.11265564e-01
7.02484906e-01 -3.14313740e-01 6.62129581e-01 3.40046853e-01
-9.38537836e-01 -2.33893022e-01 -1.55777901e-01 -2.96113789e-01
-8.02517757e-02 6.66945219e-01 -3.70200902e-01 7.56425187e-02
1.14115393e+00 8.19431722e-01 -7.93749332e-01 8.18785310e-01
-3.11296489e-02 8.77475858e-01 -3.84163260e-01 1.25431225e-01
-9.32953805e-02 2.13650048e-01 3.56099576e-01 1.27808440e+00
2.16967672e-01 -1.69451073e-01 5.62318087e-01 9.39758420e-01
-4.07172680e-01 -2.90598094e-01 -7.63440192e-01 -2.75574382e-02
4.25979733e-01 1.18994355e+00 -1.03206694e-01 -8.85774046e-02
-2.98615426e-01 1.04362357e+00 5.41199327e-01 5.10777593e-01
-8.31308901e-01 2.15560999e-02 8.56378317e-01 1.93649903e-01
-1.71566576e-01 -3.78086001e-01 -4.91079867e-01 -1.28235078e+00
4.23268527e-01 -1.02148604e+00 2.76614100e-01 -5.89833617e-01
-1.65976989e+00 4.36217010e-01 -1.68766323e-02 -1.13709700e+00
-1.82951614e-01 -5.39713800e-01 -3.54437739e-01 6.32097185e-01
-9.88555372e-01 -1.27707255e+00 -7.40647137e-01 8.11478853e-01
3.12165827e-01 -8.42056237e-03 9.94523823e-01 7.20704734e-01
-1.05193746e+00 1.04246509e+00 1.51916295e-02 4.54111636e-01
9.01877224e-01 -9.64845896e-01 3.11001867e-01 8.54554534e-01
1.35614946e-01 7.49583662e-01 5.71027696e-01 -6.79955542e-01
-1.23582327e+00 -1.35027587e+00 6.98531449e-01 -8.34550321e-01
3.61090630e-01 -8.81498516e-01 -8.20194781e-01 1.02927816e+00
-1.94021210e-01 4.06520784e-01 1.02579093e+00 5.52892685e-01
-4.38259602e-01 -2.63931304e-01 -1.15236938e+00 3.38579565e-01
1.62225389e+00 -6.18003309e-01 -4.66801971e-01 4.50825840e-01
4.42976683e-01 -1.80119753e-01 -1.18016195e+00 3.38809609e-01
8.09990585e-01 -5.00746667e-01 8.40444744e-01 -8.72110426e-01
-1.09757252e-01 -1.00122057e-01 -4.21937078e-01 -1.52141273e+00
-5.84099889e-01 -8.39299083e-01 -3.99874091e-01 1.09301805e+00
2.50710607e-01 -4.43959951e-01 6.94693685e-01 8.80096614e-01
-2.07760394e-01 -3.15724492e-01 -1.00683475e+00 -9.76956308e-01
-2.20372632e-01 -7.55817235e-01 3.72225434e-01 9.68807518e-01
-4.23804820e-02 5.90184569e-01 -9.12909627e-01 -1.35642558e-01
8.61709535e-01 -7.21225619e-01 9.32982385e-01 -1.70377660e+00
-2.64528900e-01 2.77090490e-01 -3.88322979e-01 -7.42550552e-01
8.83771628e-02 -6.27165973e-01 3.18697661e-01 -1.16067600e+00
2.00342387e-01 -4.71235633e-01 -5.44943154e-01 9.15676296e-01
-7.00730830e-02 4.13185716e-01 1.28592581e-01 2.60913253e-01
-9.96008813e-01 5.88122189e-01 5.32497704e-01 -1.69036135e-01
-2.02269286e-01 -1.96486443e-01 -8.46182525e-01 7.59292841e-01
8.48209262e-01 -6.08742177e-01 -3.60524684e-01 -3.80857050e-01
2.77403384e-01 -5.26196122e-01 6.93422914e-01 -1.67441130e+00
-2.97011789e-02 -1.13132773e-02 7.82309532e-01 1.31209064e-02
3.63962859e-01 -9.82245624e-01 3.05033684e-01 -3.49707380e-02
-4.25078422e-01 -1.71381339e-01 4.21560481e-02 5.56896269e-01
3.47702116e-01 3.20720315e-01 5.39031565e-01 -8.29908028e-02
-6.53975666e-01 3.44141245e-01 -2.24333286e-01 4.89865065e-01
5.96135080e-01 -1.69253513e-01 -4.16190028e-01 -2.15151891e-01
-9.72585022e-01 2.87753884e-02 8.05899650e-02 6.32400274e-01
1.62098482e-01 -1.62482357e+00 -5.33673882e-01 4.35811162e-01
5.80780208e-01 -1.02520771e-01 3.00272822e-01 9.48948860e-01
2.20027074e-01 1.35409564e-01 -2.95482755e-01 -5.89385867e-01
-5.54653823e-01 3.61027598e-01 6.33559227e-01 -1.62652936e-02
-5.92858493e-01 6.06749177e-01 -1.11165345e-01 -6.25412941e-01
6.62492096e-01 -5.33738852e-01 2.21019641e-01 2.28106547e-02
6.06414020e-01 6.65426254e-01 2.69445717e-01 -5.49839258e-01
-4.64199573e-01 1.62043408e-01 2.46450901e-01 1.72692984e-01
1.24103069e+00 -3.93335149e-02 5.10261297e-01 7.08240211e-01
1.30732167e+00 -4.79665875e-01 -1.63976073e+00 -5.60810506e-01
-2.38693152e-02 -1.05908319e-01 5.13650067e-02 -1.06229305e+00
-7.85541832e-01 5.16153991e-01 1.23182750e+00 -1.71561912e-01
1.01778483e+00 -7.51117617e-02 9.85991001e-01 7.57555366e-01
4.80396807e-01 -1.64233708e+00 2.13536426e-01 3.23147714e-01
6.37579322e-01 -1.38416588e+00 -1.63810372e-01 1.55775920e-01
-5.66202700e-01 4.73341823e-01 6.30189836e-01 -1.02365106e-01
6.09570980e-01 1.40404612e-01 3.72632407e-02 -2.70239681e-01
-3.59093666e-01 -2.35991299e-01 2.90298074e-01 1.02380049e+00
3.42111915e-01 3.43718559e-01 -1.71723869e-02 1.35415220e+00
-2.02028617e-01 4.18115646e-01 1.89325064e-01 1.10547018e+00
-1.66918486e-01 -6.46132112e-01 -3.68679255e-01 5.35747349e-01
-4.68555689e-01 1.86593726e-01 -3.05694848e-01 5.78250527e-01
6.86702430e-01 1.12865174e+00 -2.29813918e-01 -7.96957493e-01
7.74074376e-01 4.67489749e-01 2.77787298e-01 -3.99048150e-01
-7.81679571e-01 -4.54080589e-02 3.54668289e-01 -9.31580365e-01
-4.73408967e-01 -8.06474388e-01 -7.60297537e-01 -1.83213890e-01
2.16522813e-01 -1.79584369e-01 5.39884150e-01 1.11141682e+00
8.14114511e-01 7.98362970e-01 3.70736361e-01 -9.25301671e-01
-7.00846970e-01 -1.39442611e+00 -8.08904767e-01 8.76066983e-01
5.53102493e-01 -9.65487421e-01 -2.94971704e-01 7.08649978e-02] | [7.758265018463135, 0.9508281350135803] |
99e414be-1be7-463d-82c7-4e2d91e9cbb0 | neural-network-fast-classifies-biological | 2012.10331 | null | https://arxiv.org/abs/2012.10331v1 | https://arxiv.org/pdf/2012.10331v1.pdf | Neural network fast-classifies biological images using features selected after their random-forests-importance to power smart microscopy | Artificial intelligence is nowadays used for cell detection and classification in optical microscopy, during post-acquisition analysis. The microscopes are now fully automated and next expected to be smart, to make acquisition decisions based on the images. It calls for analysing them on the fly. Biology further imposes training on a reduced dataset due to cost and time to prepare the samples and have the datasets annotated by experts. We propose here a real-time image processing, compliant with these specifications by balancing accurate detection and execution performance. We characterised the images using a generic, high-dimensional feature extractor. We then classified the images using machine learning for the sake of understanding the contribution of each feature in decision and execution time. We found that the non-linear-classifier random forests outperformed Fisher's linear discriminant. More importantly, the most discriminant and time-consuming features could be excluded without any significant loss in accuracy, offering a substantial gain in execution time. It suggests a feature-group redundancy likely related to the biology of the observed cells. We offer a method to select fast and discriminant features. In our assay, a 79.6 $\pm$ 2.4 % accurate classification of a cell took 68.7 $\pm$ 3.5 ms (mean $\pm$ SD, 5-fold cross-validation nested in 10 bootstrap repeats), corresponding to 14 cells per second, dispatched into 8 phases of the cell cycle using 12 feature-groups and operating a consumer market ARM-based embedded system. Interestingly, a simple neural network offered similar performances paving the way to faster training and classification, using parallel execution on a general-purpose graphic processing unit. Finally, this strategy is also usable for deep neural networks paving the way to optimising these algorithms for smart microscopy. | ['Jacques Pécréaux', 'Marc Tramier', 'Otmane Bouchareb', 'Baptiste Giroux', 'Jérémy Pont', 'Thomas Walter', 'Youssef El Habouz', 'Florian Sizaire', 'Maël Balluet'] | 2020-12-18 | null | null | null | null | ['cell-detection'] | ['computer-vision'] | [ 5.17273664e-01 -7.75709748e-02 4.22818989e-01 -3.28443259e-01
-4.25028622e-01 -4.57825124e-01 4.27253366e-01 3.57527196e-01
-1.09150589e+00 8.70648026e-01 -8.42786968e-01 -3.15979779e-01
-3.81779999e-01 -7.04538524e-01 -3.89477640e-01 -1.21532226e+00
-2.90623516e-01 7.93734193e-01 2.13165238e-01 1.04756802e-01
5.22107661e-01 1.00064766e+00 -1.95631468e+00 2.79037803e-01
4.65803504e-01 1.12297273e+00 5.22175729e-01 1.04683566e+00
1.00424074e-01 3.45996410e-01 -6.09572113e-01 -8.01110268e-02
6.89492673e-02 -1.55169964e-01 -6.51431322e-01 7.24875107e-02
-4.37417664e-02 -1.48865134e-01 4.78460789e-01 6.08973980e-01
6.80643141e-01 -3.41642499e-01 8.48932683e-01 -9.18321431e-01
2.59104282e-01 -6.97941631e-02 -1.33642420e-01 2.15453297e-01
1.81058049e-01 3.44071388e-01 5.02167284e-01 -6.21741712e-01
7.13329852e-01 7.23782599e-01 5.03734708e-01 3.57139766e-01
-1.48543584e+00 -4.37900454e-01 -3.96632940e-01 2.05167577e-01
-1.21947598e+00 -4.94711667e-01 2.36561432e-01 -7.22927392e-01
1.01085556e+00 4.18887794e-01 8.22122514e-01 6.86298013e-01
4.80574548e-01 1.84246466e-01 1.49619806e+00 -5.10081887e-01
4.64308411e-01 3.67392033e-01 1.05026715e-01 6.17543399e-01
2.50356257e-01 -3.55127011e-03 -8.77741352e-02 4.87003811e-02
5.03902316e-01 7.72429407e-02 -1.47768468e-01 7.64830336e-02
-1.06200004e+00 5.45288026e-01 -4.73012365e-02 5.06909788e-01
-2.47580603e-01 -1.81469962e-01 4.92333770e-01 2.32992753e-01
3.54806371e-02 5.50591767e-01 -7.83537686e-01 -1.11659490e-01
-7.38366544e-01 -1.35633454e-01 8.51592064e-01 3.17284167e-01
7.72488832e-01 -1.83063447e-01 1.64192259e-01 4.29881305e-01
1.58361867e-02 3.48860532e-01 7.47584581e-01 -9.45639491e-01
-4.24334496e-01 1.01150239e+00 -1.10871226e-01 -8.53393555e-01
-9.70548570e-01 -3.41890723e-01 -1.04063356e+00 6.49580002e-01
9.68101263e-01 9.70521793e-02 -6.63900197e-01 1.18868136e+00
3.10317904e-01 -4.55152929e-01 -2.13249147e-01 6.83618605e-01
1.89847007e-01 3.34183306e-01 -2.18966335e-01 -5.04149914e-01
1.63840961e+00 -1.17805757e-01 -4.19728607e-01 2.94185400e-01
6.77988350e-01 -6.12836540e-01 9.87374008e-01 9.13618088e-01
-7.00552464e-01 -6.65456116e-01 -1.10427809e+00 1.40881285e-01
-7.40155876e-01 4.06569451e-01 6.18726790e-01 7.65646458e-01
-9.80177402e-01 9.18115973e-01 -7.98638523e-01 -3.91410083e-01
4.60245758e-01 9.37562585e-01 -6.05908036e-01 2.80312806e-01
-5.29584229e-01 6.63064420e-01 3.55745047e-01 1.70240596e-01
-3.94506514e-01 -3.83826494e-01 -4.15737122e-01 1.89233907e-02
-5.66530749e-02 -5.89031994e-01 5.21636546e-01 -8.13117862e-01
-1.67580640e+00 1.42575336e+00 -2.25791425e-01 -5.04775941e-01
5.39197922e-01 3.19707602e-01 -1.41381606e-01 2.75154948e-01
-8.76552388e-02 6.38670146e-01 6.64341450e-01 -8.83322835e-01
-7.41802156e-01 -6.91689610e-01 -4.11606044e-01 -5.19861460e-01
-1.51210561e-01 -8.58098418e-02 -1.04263924e-01 -3.00054345e-02
-1.70698855e-02 -9.38915074e-01 -2.24686891e-01 -6.92660436e-02
-8.54929611e-02 -3.34666595e-02 7.22887814e-01 -4.38932449e-01
7.54893601e-01 -1.99723613e+00 -1.36477366e-01 2.81143308e-01
2.01171100e-01 4.21442360e-01 2.02940077e-01 2.64476478e-01
-6.63849786e-02 -5.15459403e-02 -1.27229407e-01 -1.37915671e-01
-2.15721413e-01 4.93119078e-05 6.55703992e-02 8.30786467e-01
2.40096048e-01 4.59023535e-01 -3.68291676e-01 -5.08724153e-01
3.79089773e-01 6.46387100e-01 -2.41711274e-01 9.49000940e-02
1.27979487e-01 4.54958051e-01 -6.47039968e-04 7.09538460e-01
5.05883396e-01 -2.65610337e-01 3.20150286e-01 -1.47968918e-01
-3.30246538e-01 -1.50333807e-01 -1.09271669e+00 1.27844787e+00
-2.70379752e-01 8.14284384e-01 1.33388355e-01 -1.26553571e+00
1.15688860e+00 -2.57122833e-02 4.06900287e-01 -7.13847101e-01
5.34293473e-01 4.79406357e-01 1.52045175e-01 -4.77110833e-01
6.74512312e-02 -4.08323891e-02 2.93687969e-01 2.75290608e-01
2.28170350e-01 -3.45592164e-02 4.72001880e-01 -2.29927897e-01
1.12336695e+00 8.23427588e-02 4.21445608e-01 -6.04261279e-01
8.73584151e-01 3.85013013e-03 2.49863893e-01 5.02794623e-01
-3.42193037e-01 4.85864073e-01 5.89144945e-01 -8.14130366e-01
-8.03504944e-01 -4.54279542e-01 -4.07132179e-01 8.04180264e-01
-1.81578130e-01 4.73130047e-02 -9.89430308e-01 -3.50886941e-01
-2.30423674e-01 2.51558334e-01 -5.40100276e-01 1.88017279e-01
-3.40045035e-01 -9.77622926e-01 3.74811679e-01 -8.02814960e-02
3.12520936e-02 -1.06864655e+00 -1.39008343e+00 3.15139800e-01
3.95967543e-01 -8.09396327e-01 4.75136042e-01 8.08647633e-01
-9.89028573e-01 -1.25356328e+00 -3.63368601e-01 -4.90928918e-01
7.82972932e-01 -9.46132988e-02 7.47199655e-01 3.15751344e-01
-9.28072691e-01 9.15879309e-02 -7.31678307e-02 -5.21201015e-01
-2.27143407e-01 5.39528988e-02 1.39114037e-01 -1.76080674e-01
6.10174775e-01 -6.67560995e-01 -7.77219892e-01 3.75061810e-01
-8.11903417e-01 -2.28192106e-01 7.94870734e-01 1.06032348e+00
8.63666117e-01 1.54381379e-01 3.99568439e-01 -7.91596055e-01
8.37066770e-02 -3.72392163e-02 -8.19020152e-01 -7.81353042e-02
-7.53073812e-01 -1.49049321e-02 1.08902228e+00 -2.06207067e-01
-4.29028541e-01 4.06146646e-01 -2.23233089e-01 9.18360725e-02
-6.91419125e-01 1.60738125e-01 -1.47214666e-01 -3.32853913e-01
7.38621473e-01 2.46280193e-01 5.50689161e-01 -1.52224585e-01
-2.88571000e-01 6.89909935e-01 1.99704468e-01 -2.13454455e-01
3.34494978e-01 8.37625444e-01 5.54751933e-01 -9.53465879e-01
-5.45140021e-02 -4.91327703e-01 -7.14736402e-01 -4.05062646e-01
6.91265821e-01 -5.03547728e-01 -1.42280722e+00 6.07686937e-01
-9.96985197e-01 -2.98527986e-01 -1.54227242e-01 4.56551760e-01
-6.52809322e-01 3.13892126e-01 -4.77431834e-01 -1.00221574e+00
-5.27749717e-01 -1.24974489e+00 8.80248010e-01 2.79605895e-01
-3.45845938e-01 -7.04262793e-01 -7.33921379e-02 3.89502227e-01
2.84922957e-01 2.28697434e-01 8.35106254e-01 -8.64438593e-01
-4.24345762e-01 -6.32918060e-01 -2.93068081e-01 1.35321572e-01
-1.10612316e-02 4.90919143e-01 -1.28605294e+00 -3.62059087e-01
2.32368141e-01 -1.03664726e-01 7.10239410e-01 5.15578866e-01
1.27498782e+00 -4.60467637e-02 -5.62264979e-01 5.86685419e-01
1.64213192e+00 4.37510073e-01 7.24816263e-01 5.38061678e-01
1.48925766e-01 1.01747227e+00 5.24833381e-01 4.11387891e-01
-1.73189610e-01 5.99780321e-01 4.73369688e-01 -2.16747373e-01
3.19104195e-01 5.59050202e-01 1.81554437e-01 4.54832673e-01
-2.41640955e-01 -1.06003299e-01 -9.03691590e-01 2.40039051e-01
-1.32320046e+00 -7.73552120e-01 -4.75310892e-01 2.45592022e+00
5.40200770e-01 3.79175425e-01 3.26063126e-01 6.73926175e-01
4.71631348e-01 -7.03960419e-01 -2.40132421e-01 -4.98767674e-01
-4.91629466e-02 4.00883526e-01 6.72871590e-01 3.31931800e-01
-8.86283875e-01 2.98957705e-01 5.27494240e+00 9.68412519e-01
-1.53931975e+00 -4.63822305e-01 9.68869090e-01 -1.26594886e-01
3.70332211e-01 -8.32944065e-02 -1.00280178e+00 7.91229010e-01
1.26817632e+00 2.56541789e-01 2.51686901e-01 7.41536200e-01
2.55637407e-01 -7.45263934e-01 -1.07305384e+00 9.93888974e-01
-1.94367349e-01 -1.28815842e+00 -2.31208250e-01 5.39820254e-01
7.80955143e-03 -1.47620529e-01 -1.48523048e-01 -2.40769405e-02
-3.60801250e-01 -1.10308301e+00 4.16771382e-01 6.69606388e-01
8.96332979e-01 -7.47289479e-01 1.02800643e+00 5.59833229e-01
-6.36279225e-01 -7.02775195e-02 -2.79521108e-01 -3.98671687e-01
-1.44432962e-01 8.41393292e-01 -9.82689679e-01 3.40722859e-01
6.76794350e-01 8.90209377e-02 -6.26628697e-01 6.71481669e-01
5.51127672e-01 2.92955816e-01 -5.39050937e-01 -2.86670297e-01
-1.00178446e-03 -3.36909205e-01 3.09406281e-01 1.42182553e+00
4.06615138e-01 -1.70573115e-01 -3.66252661e-01 5.28951526e-01
5.52802205e-01 1.22520469e-01 -3.74715865e-01 3.60788740e-02
1.61673293e-01 1.72087312e+00 -1.54422462e+00 -1.65683925e-01
5.89445271e-02 7.55359888e-01 2.00852737e-01 -2.67977983e-01
-4.64582950e-01 -7.50419617e-01 3.46824974e-01 3.75290036e-01
2.93717623e-01 -7.54846707e-02 -3.81959617e-01 -6.10926926e-01
-1.10814847e-01 -6.30795538e-01 1.69082820e-01 -3.88147086e-01
-9.76560175e-01 5.46741068e-01 -4.02804255e-01 -1.21013570e+00
-2.74312615e-01 -1.41056621e+00 -2.25885496e-01 6.95824504e-01
-1.33804905e+00 -7.70085871e-01 -2.27600396e-01 2.62669832e-01
1.93962008e-01 -1.54957294e-01 1.23133433e+00 2.06194922e-01
-6.37675047e-01 1.78093106e-01 2.03917027e-01 -1.27950668e-01
5.67955554e-01 -1.37950814e+00 -2.69853652e-01 4.61395591e-01
-1.37169724e-02 6.60768509e-01 7.30228603e-01 -1.42452672e-01
-1.42051518e+00 -5.53305209e-01 9.36039090e-01 -2.72509962e-01
4.98770922e-01 -3.26394528e-01 -8.76106143e-01 2.03114189e-02
5.34546515e-03 3.11333919e-03 1.07744217e+00 -2.01289207e-01
2.00472325e-01 -3.69274318e-01 -1.33007228e+00 3.08274299e-01
5.05362213e-01 -2.50522852e-01 5.53115085e-02 2.55686045e-01
-5.18665463e-02 -6.31095916e-02 -8.37800682e-01 3.09851140e-01
8.44621420e-01 -1.51070356e+00 6.85825884e-01 -1.38511196e-01
2.28934839e-01 -4.64524209e-01 1.43787876e-01 -5.79965353e-01
-1.96094334e-01 -5.00830591e-01 2.98682272e-01 9.71435368e-01
2.97776550e-01 -8.56927752e-01 1.05475414e+00 5.03803194e-01
1.22794829e-01 -9.86306250e-01 -1.19463325e+00 -6.78704500e-01
-1.99710578e-01 -2.84328312e-01 8.77541378e-02 6.03426337e-01
3.00362110e-02 2.74011821e-01 1.33023500e-01 -1.92932725e-01
4.75879580e-01 -1.88303867e-03 8.50994349e-01 -1.48787963e+00
-3.62830877e-01 -7.03930676e-01 -9.30095613e-01 -5.23436666e-01
-5.11084497e-02 -5.54334521e-01 -1.86110973e-01 -9.26535726e-01
4.39750031e-02 -3.17605019e-01 -1.93758354e-01 1.85204044e-01
6.49888292e-02 3.56514871e-01 -7.60370567e-02 2.68951297e-01
-3.23363572e-01 -2.48450398e-01 7.27980316e-01 7.08908588e-02
-8.97656605e-02 3.34359035e-02 -1.94675326e-01 6.61620677e-01
8.96015346e-01 -4.41998690e-01 -5.26182875e-02 1.06983826e-01
2.92342663e-01 -2.90392697e-01 4.58987623e-01 -1.36842382e+00
4.11139578e-01 1.42975152e-01 7.77869999e-01 -5.68613112e-01
4.32518482e-01 -1.13089716e+00 4.88343537e-01 9.65458393e-01
9.89577100e-02 -3.17998797e-01 2.00407863e-01 3.16800386e-01
-1.33495748e-01 -4.76004988e-01 9.45103228e-01 -2.67253369e-01
-5.10895789e-01 -1.69075787e-01 -8.21583986e-01 -5.84622443e-01
1.33914363e+00 -8.87023747e-01 -2.54135728e-01 1.94088638e-01
-6.93305552e-01 -2.46091723e-01 7.01836348e-01 -5.00422180e-01
2.98342407e-01 -5.20017087e-01 -3.62364739e-01 5.09989679e-01
-5.24745462e-03 -9.07320753e-02 5.98120272e-01 1.18252969e+00
-8.67329597e-01 5.80748379e-01 -5.66195905e-01 -9.29807246e-01
-1.54001296e+00 5.53972840e-01 3.21350455e-01 -4.19512242e-01
-1.75657626e-02 7.08607733e-01 -2.22995400e-01 -4.06818651e-02
-1.93226621e-01 -2.64606178e-01 -4.95648712e-01 4.53780442e-01
5.84845006e-01 5.94220221e-01 4.43444163e-01 -3.65538001e-01
-4.59076792e-01 7.17604518e-01 -7.60436058e-02 1.22015707e-01
1.20639992e+00 -1.37290619e-02 -3.67904007e-01 5.35547137e-01
1.17220426e+00 -1.05705611e-01 -1.16039550e+00 4.56179976e-01
1.37307063e-01 -9.22076032e-02 -9.75320861e-02 -7.28872955e-01
-7.26274192e-01 1.00921464e+00 8.08231413e-01 6.17391229e-01
1.27871776e+00 -3.91247869e-01 1.78215384e-01 4.05923158e-01
6.21861696e-01 -1.12708247e+00 -4.19400334e-01 3.10154259e-01
4.38567907e-01 -1.28635323e+00 2.94639438e-01 -3.42299342e-01
-2.11546272e-01 1.58025897e+00 2.98909724e-01 -1.56059563e-02
3.56577843e-01 4.75473613e-01 1.43273547e-01 -2.69286990e-01
-8.02040100e-01 -8.95759463e-02 -1.30297124e-01 8.41515958e-01
4.12255347e-01 7.35755116e-02 -4.55639422e-01 5.28887570e-01
-5.87766320e-02 2.88544059e-01 4.51384723e-01 7.74300635e-01
-5.34241915e-01 -9.92670417e-01 -2.74057865e-01 6.37202263e-01
-5.83099782e-01 2.49616995e-01 -2.73679137e-01 8.22261155e-01
3.24493706e-01 7.33905673e-01 1.94173321e-01 -3.77521247e-01
1.44657537e-01 5.26439287e-02 4.36464876e-01 -2.45188162e-01
-6.62300110e-01 4.25581522e-02 -2.81240314e-01 -6.47292972e-01
-4.32456434e-01 -5.86925328e-01 -1.28328776e+00 -3.59105140e-01
-3.77559543e-01 8.95377696e-02 1.02606094e+00 8.30721974e-01
5.27751863e-01 2.86262929e-01 6.02608562e-01 -1.19888937e+00
-1.51561782e-01 -8.39722335e-01 -7.53625512e-01 1.01211062e-02
1.40669286e-01 -4.85333860e-01 -6.74432576e-01 4.27366018e-01] | [14.710262298583984, -3.09609055519104] |
ba5b7d9b-d720-497a-b881-9dc02c09eb70 | quantum-tomography-benchmarking | 2012.15656 | null | https://arxiv.org/abs/2012.15656v2 | https://arxiv.org/pdf/2012.15656v2.pdf | Quantum tomography benchmarking | Recent advances in quantum computers and simulators are steadily leading us towards full-scale quantum computing devices. Due to the fact that debugging is necessary to create any computing device, quantum tomography (QT) is a critical milestone on this path. In practice, the choice between different QT methods faces the lack of comparison methodology. Modern research provides a wide range of QT methods, which differ in their application areas, as well as experimental and computational complexity. Testing such methods is also being made under different conditions, and various efficiency measures are being applied. Moreover, many methods have complex programming implementations; thus, comparison becomes extremely difficult. In this study, we have developed a general methodology for comparing quantum state tomography methods. The methodology is based on an estimate of the resources needed to achieve the required accuracy. We have developed a software library (in MATLAB and Python) that makes it easy to analyze any QT method implementation through a series of numerical experiments. The conditions for such a simulation are set by the number of tests corresponding to real physical experiments. As a validation of the proposed methodology and software, we analyzed and compared a set of QT methods. The analysis revealed some method-specific features and provided estimates of the relative efficiency of the methods. | ['Yu. I. Bogdanov', 'A. Yu. Chernyavskiy', 'B. I. Bantysh'] | 2020-12-31 | null | null | null | null | ['quantum-state-tomography'] | ['medical'] | [ 1.29502043e-01 -5.60926318e-01 1.21902868e-01 -1.93892837e-01
-5.73282242e-01 -5.03472626e-01 5.93445420e-01 4.06214520e-02
-5.34009516e-01 9.29139078e-01 -6.22933626e-01 -5.31137645e-01
-9.02197957e-02 -1.12893224e+00 -3.15794080e-01 -7.95115411e-01
-3.31744030e-02 4.57854062e-01 3.80904168e-01 -3.47947598e-01
7.09168136e-01 5.49822330e-01 -1.48572075e+00 -3.12139660e-01
1.05276728e+00 7.57077277e-01 2.08783180e-01 5.10162175e-01
-2.85464767e-02 2.16761664e-01 -5.70505381e-01 -2.44074345e-01
1.56743959e-01 -7.50527203e-01 -8.02930832e-01 -3.37104529e-01
-3.84812206e-01 -1.23674810e-01 -2.53785133e-01 1.45033097e+00
6.94945693e-01 1.59966379e-01 6.09191775e-01 -1.16388679e+00
-8.66990834e-02 5.41368067e-01 -7.00229499e-03 2.06300646e-01
5.76976299e-01 2.81957865e-01 5.99402845e-01 -4.15703923e-01
4.08331156e-01 7.02891827e-01 3.73005837e-01 3.44253778e-01
-1.25495780e+00 -7.74592817e-01 -9.78986561e-01 4.21907246e-01
-1.70895374e+00 -2.71705776e-01 6.92176104e-01 -2.04417244e-01
9.39482272e-01 1.50343835e-01 7.87742138e-01 6.75968528e-01
5.74458838e-01 -1.19258896e-01 1.70793343e+00 -8.01115096e-01
5.95324874e-01 6.24945045e-01 2.33975217e-01 5.80481052e-01
4.49180335e-01 3.40447128e-01 -1.52786076e-01 -2.30961844e-01
7.83620715e-01 -3.43002498e-01 -2.87049282e-02 -3.20445865e-01
-1.06352699e+00 6.09338582e-01 1.08730674e-01 7.73956180e-01
-1.21726632e-01 5.56818694e-02 3.95999312e-01 2.40441725e-01
-1.39332682e-01 5.32483280e-01 7.01786205e-02 -4.54385549e-01
-8.82787287e-01 3.14860821e-01 9.46689725e-01 7.18336701e-01
8.90132427e-01 -9.84389409e-02 2.43730456e-01 2.27454215e-01
-1.86784361e-02 4.69571561e-01 5.25155067e-01 -5.74818134e-01
1.67346105e-01 3.73197377e-01 3.58295560e-01 -8.46503377e-01
-3.84975672e-01 -2.37228021e-01 -8.33925962e-01 3.01092207e-01
4.76769090e-01 -2.03285608e-02 -4.15748596e-01 1.43147147e+00
3.49853456e-01 -3.40091772e-02 1.67512879e-01 7.53988028e-01
5.01120746e-01 8.49783480e-01 -2.23301873e-01 -4.95053589e-01
1.13859820e+00 -3.16600502e-01 -8.78546715e-01 6.08926594e-01
8.88391554e-01 -9.68629301e-01 9.81821835e-01 5.27437985e-01
-9.20879662e-01 -5.00552535e-01 -1.20514452e+00 2.41898179e-01
-5.43050528e-01 -1.93278834e-01 7.94488132e-01 1.22804785e+00
-9.56316233e-01 1.18009031e+00 -7.62878180e-01 -4.54330802e-01
-3.15593034e-01 5.95644414e-01 -1.07363112e-01 3.45711082e-01
-1.30886924e+00 1.17166674e+00 6.37339830e-01 2.03571215e-01
-5.85018516e-01 -8.79793987e-02 -3.24313909e-01 2.78198496e-02
9.47351158e-02 -5.26579320e-01 1.13470209e+00 -5.45896888e-01
-2.01448607e+00 5.66238344e-01 -1.13952737e-02 -3.52810323e-01
4.68704581e-01 3.62382352e-01 -7.39900768e-01 4.90164571e-02
-5.82703529e-03 -1.01528093e-01 4.73963976e-01 -7.18629301e-01
-1.71248734e-01 -1.77131936e-01 1.79020509e-01 -5.70294522e-02
7.63372343e-04 2.88578957e-01 2.91885901e-02 1.64485738e-01
2.42290631e-01 -8.97982955e-01 -2.65367746e-01 -5.01591206e-01
-2.50954658e-01 -2.91451328e-02 5.69389224e-01 1.89162157e-02
1.16961145e+00 -1.96064222e+00 -5.96194416e-02 4.25447255e-01
-3.88561115e-02 2.44396821e-01 4.56403375e-01 8.50438118e-01
9.73375291e-02 5.30237034e-02 -2.24791363e-01 7.08142892e-02
1.02537431e-01 4.73771896e-03 9.21519846e-03 6.05704129e-01
-1.74152851e-01 4.33973908e-01 -7.56819010e-01 -6.01818025e-01
5.55465162e-01 3.68288279e-01 -4.71538246e-01 4.32905778e-02
7.70836174e-02 7.76442885e-01 -6.70194924e-01 5.19806683e-01
7.76498020e-01 -2.25933373e-01 2.34566286e-01 -3.05950344e-01
-7.04560339e-01 5.37074208e-01 -1.45157492e+00 1.40589666e+00
-4.29428190e-01 2.90877461e-01 -2.39763111e-01 -1.15755010e+00
8.79121840e-01 4.28790003e-01 3.32136631e-01 -8.60351443e-01
6.41387224e-01 7.20593393e-01 4.35176820e-01 -5.20460248e-01
6.51293218e-01 -4.89837289e-01 -1.16028331e-01 6.15687430e-01
1.13844007e-01 -7.92596459e-01 5.14650941e-01 -1.65450554e-02
8.49341750e-01 1.43521041e-01 6.39130771e-01 -5.42210400e-01
7.23842978e-01 1.32437080e-01 2.88867176e-01 6.59024417e-01
-4.70915258e-01 3.81875634e-02 4.35871273e-01 -3.84306103e-01
-1.29639006e+00 -8.45687568e-01 -5.80608666e-01 3.75476629e-02
5.89258611e-01 -4.42076743e-01 -8.06907713e-01 1.57772735e-01
-5.72846949e-01 6.62631035e-01 -8.11451003e-02 -3.84487100e-02
-3.98700267e-01 -1.04199767e+00 4.64320689e-01 -2.43381783e-02
6.91088855e-01 -1.01541245e+00 -8.58757198e-01 3.04601610e-01
8.70151967e-02 -1.15090036e+00 4.12614554e-01 2.07029521e-01
-1.15587962e+00 -9.75982189e-01 -2.80744851e-01 -3.11878592e-01
5.09248674e-01 2.70205170e-01 8.13624322e-01 1.06871836e-01
-2.53525048e-01 -1.04228787e-01 -4.48013186e-01 -7.36122802e-02
-7.01837420e-01 -5.17441370e-02 3.50236923e-01 -3.34980100e-01
1.99321240e-01 -8.26466084e-01 -4.70741540e-01 1.05811551e-01
-9.11718488e-01 -5.71537316e-02 6.30979598e-01 7.93571234e-01
3.46402735e-01 4.01406288e-01 1.33365959e-01 -6.85278535e-01
6.58657789e-01 -2.64533073e-01 -1.20143723e+00 1.43229708e-01
-7.31304109e-01 2.48453379e-01 9.39654768e-01 3.60098295e-02
-8.18145275e-01 -2.39087358e-01 -3.36209804e-01 3.00337702e-01
-1.14573181e-01 4.61668044e-01 -4.58308049e-02 -6.51359677e-01
6.70423508e-01 3.74222726e-01 -2.35488653e-01 -1.20040059e-01
3.82380299e-02 6.98671162e-01 1.18988827e-01 -7.03531325e-01
8.61649156e-01 2.52129436e-01 4.59936082e-01 -1.01203859e+00
-3.27603102e-01 -1.11995764e-01 -5.76280773e-01 -2.72080511e-01
6.62987947e-01 -2.69021034e-01 -9.70804393e-01 5.15183687e-01
-9.81419921e-01 1.22122392e-01 1.01354823e-01 9.30803657e-01
-4.25721347e-01 6.14353657e-01 -4.18457210e-01 -1.01039088e+00
-2.88121581e-01 -1.62962997e+00 4.69506145e-01 5.06895781e-01
3.35308202e-02 -8.97420645e-01 2.37426907e-01 8.23859051e-02
5.27394831e-01 1.53761685e-01 8.53101969e-01 -2.51930714e-01
-8.76308560e-01 -2.85057783e-01 -1.15327664e-01 1.40894771e-01
-2.01007053e-01 2.60675281e-01 -9.04664099e-01 -3.45110953e-01
2.60572672e-01 -2.25147173e-01 1.20321773e-01 2.05183610e-01
1.01024938e+00 2.08309725e-01 -2.26758271e-01 5.01376152e-01
1.82614374e+00 4.94325161e-01 8.67809117e-01 3.17027897e-01
1.76043645e-01 1.61554560e-01 6.75449371e-01 3.19916248e-01
-1.13963872e-01 8.68768334e-01 3.25883985e-01 2.75992095e-01
5.80883205e-01 4.17807251e-02 2.86448926e-01 1.12910151e+00
-5.38629353e-01 9.24711451e-02 -7.46004283e-01 1.02675140e-01
-1.34416318e+00 -1.09607768e+00 -4.57055658e-01 2.69932413e+00
5.24742067e-01 3.17878276e-01 6.37312755e-02 5.43593407e-01
6.50193214e-01 -2.42319942e-01 -7.78077990e-02 -7.41499305e-01
3.17410111e-01 6.99967623e-01 4.38067615e-01 2.58287996e-01
-6.85456216e-01 7.83494234e-01 6.68706274e+00 9.39728200e-01
-1.41075766e+00 2.20574379e-01 -4.05349433e-02 4.72019404e-01
-1.01965360e-01 5.94199836e-01 -6.13344610e-01 4.84641135e-01
1.30068088e+00 -5.23578048e-01 4.16046232e-01 6.58357620e-01
2.45714366e-01 -7.22459972e-01 -8.99687588e-01 1.10396659e+00
-3.73872638e-01 -1.22173727e+00 -2.08992228e-01 6.18549064e-02
6.60213113e-01 -1.51841015e-01 -2.32844546e-01 2.21953005e-01
-3.19168180e-01 -7.95817256e-01 5.09904981e-01 3.60367060e-01
7.08723485e-01 -8.26464236e-01 1.09563243e+00 4.77231830e-01
-9.82322812e-01 2.02087313e-01 -4.72069979e-01 -5.38535476e-01
2.19515949e-01 5.78211606e-01 -7.69011438e-01 9.17028189e-01
3.64966005e-01 1.33363098e-01 -1.52611092e-01 1.22438240e+00
-1.17984347e-01 4.42240864e-01 -3.42260510e-01 -6.89419627e-01
1.27379000e-01 -8.77021670e-01 2.46604443e-01 8.11144769e-01
7.14587927e-01 1.23977572e-01 -3.43479395e-01 1.30204964e+00
3.84637743e-01 2.68000633e-01 -7.42354155e-01 -3.95152718e-01
5.67116678e-01 1.21940780e+00 -1.00861871e+00 -3.14669073e-01
-3.24668109e-01 6.83187306e-01 -1.33920148e-01 -7.65632093e-02
-1.08785903e+00 -4.97109920e-01 3.15352827e-01 -7.85328448e-02
-1.73327729e-01 -6.25348628e-01 -3.00304651e-01 -1.19078302e+00
-2.50844896e-01 -5.82441688e-01 -2.29833573e-01 -5.13069332e-01
-7.66189039e-01 6.12028301e-01 9.02549773e-02 -1.39720571e+00
-1.23589300e-01 -6.50382936e-01 -6.06492281e-01 1.00927258e+00
-1.21473897e+00 -3.74364555e-01 -2.21576095e-01 4.02691752e-01
-1.59953654e-01 1.90866321e-01 1.09806955e+00 5.62262654e-01
-6.49818659e-01 3.85447025e-01 3.66895258e-01 -2.09938675e-01
2.56647170e-01 -8.45048666e-01 1.25526249e-01 8.65685165e-01
-5.69405034e-02 9.10668135e-01 1.15810001e+00 -3.22288305e-01
-1.65507412e+00 -1.97984487e-01 7.96651125e-01 -1.40951455e-01
8.47595572e-01 -4.22982037e-01 -6.55718386e-01 2.41515443e-01
-4.45022956e-02 -1.43061623e-01 4.49183851e-01 1.53636448e-02
1.96896300e-01 -9.01627988e-02 -1.20772004e+00 3.94184917e-01
5.51231921e-01 -8.39473069e-01 -4.37392384e-01 1.55323938e-01
1.63366467e-01 -3.57706904e-01 -9.03855681e-01 3.33063066e-01
6.35433555e-01 -1.55775332e+00 5.03869951e-01 1.42276615e-01
5.40613867e-02 -5.14475286e-01 2.65457854e-02 -9.90826488e-01
1.85004428e-01 -7.34001040e-01 3.76929045e-01 8.91765118e-01
3.33454192e-01 -1.01985669e+00 5.44039905e-01 6.23366892e-01
-3.94599373e-03 -4.14985716e-01 -1.03284860e+00 -9.84856308e-01
-4.20132615e-02 -5.24351358e-01 5.40696084e-01 8.10110390e-01
3.45288038e-01 1.76528931e-01 -1.37149513e-01 2.22660884e-01
6.18273795e-01 3.44007313e-01 8.92471850e-01 -1.01230109e+00
-3.47999930e-01 -3.44460070e-01 -8.54290605e-01 -6.29123569e-01
-2.81972975e-01 -6.76262259e-01 -3.02315474e-01 -1.09636438e+00
4.22516912e-01 -6.35140240e-01 -2.37717927e-01 -1.79580301e-01
2.25613162e-01 8.69407207e-02 -4.50432263e-02 2.96636492e-01
-1.27178863e-01 5.74187756e-01 1.24704635e+00 3.44288498e-01
-1.68816969e-01 1.88345775e-01 -8.09031259e-03 4.97974813e-01
9.04348254e-01 -6.97660148e-01 -2.19709516e-01 2.43784577e-01
4.91285980e-01 2.89126992e-01 3.77867013e-01 -1.51961696e+00
2.10226610e-01 4.66652215e-03 -1.67776987e-01 -7.15290010e-01
5.01436234e-01 -7.08452761e-01 4.58419681e-01 8.33115220e-01
3.84149522e-01 2.74730604e-02 -1.94166645e-01 3.50712389e-02
-2.55097955e-01 -6.99464381e-01 9.10696924e-01 -2.20590740e-01
-7.30468869e-01 -1.88568812e-02 -2.91498661e-01 -3.08367759e-01
1.16119719e+00 -4.98107851e-01 -1.32223740e-01 -1.59606412e-01
-4.33539659e-01 -2.83396721e-01 8.07093799e-01 -4.51072544e-01
2.86606461e-01 -9.76062357e-01 -3.90992612e-02 1.76238388e-01
-3.12713198e-02 -4.98451442e-01 3.03728849e-01 1.19331181e+00
-1.09053683e+00 6.31097257e-01 -7.36307979e-01 -5.62815785e-01
-1.01357818e+00 5.12271166e-01 4.04924273e-01 -2.11303756e-01
-4.21386629e-01 2.73988336e-01 -3.49595964e-01 -3.46105158e-01
-3.92302603e-01 -5.33005536e-01 -4.14617453e-03 -4.24277186e-01
3.39430213e-01 4.43258137e-01 1.13781333e-01 -5.55506408e-01
-1.74638480e-01 6.75382853e-01 4.16913956e-01 -3.76234084e-01
1.13533616e+00 9.02357846e-02 -4.60761189e-01 5.63547015e-01
9.61269379e-01 5.93680963e-02 -4.36656862e-01 1.35699496e-01
-1.01628073e-01 -3.47101122e-01 -5.10663502e-02 -2.64045238e-01
-3.66869688e-01 1.05120206e+00 7.38496482e-01 5.83913326e-01
9.48436797e-01 -4.59289581e-01 6.99284971e-01 5.25742173e-01
1.26330841e+00 -1.22619081e+00 -4.71703023e-01 4.21930969e-01
7.62512162e-02 -1.11720133e+00 2.09669635e-01 -5.22925198e-01
-7.57897422e-02 1.27278912e+00 1.64882332e-01 -1.82570949e-01
5.71118593e-01 1.48715660e-01 -2.22459361e-01 -3.09384048e-01
-2.59241581e-01 -1.08755089e-01 -1.80348650e-01 1.65453106e-01
6.47095442e-01 3.88673931e-01 -1.01737952e+00 1.14775904e-01
-5.02697349e-01 2.81338602e-01 8.82496595e-01 1.07558990e+00
-4.43662405e-01 -1.53216052e+00 -5.98881423e-01 -2.75451923e-03
-3.59395385e-01 1.97477695e-02 1.77693963e-02 1.07643163e+00
1.03130005e-01 1.01511514e+00 -4.42122519e-01 -5.58863521e-01
3.70041765e-02 -2.72040498e-02 9.30526912e-01 -3.62716436e-01
-3.95258904e-01 -3.83216470e-01 2.08174288e-02 -5.26336432e-01
-6.52449369e-01 -4.53407615e-01 -1.55044377e+00 -8.14874232e-01
-7.76784360e-01 6.95803642e-01 1.12081802e+00 1.21931505e+00
-2.01553609e-02 3.45593303e-01 5.90263307e-01 -8.14113915e-01
-7.47763991e-01 -8.68998349e-01 -8.86923969e-01 8.65493566e-02
-2.35837385e-01 -8.13634872e-01 -3.19516331e-01 -6.00127220e-01] | [5.583437919616699, 4.958820343017578] |
950d50ba-c489-4ef2-a341-2b1eed737479 | discrepancy-optimal-meta-learning-for-domain | null | null | https://openreview.net/forum?id=eJyt4hJzOLk | https://openreview.net/pdf?id=eJyt4hJzOLk | Discrepancy-Optimal Meta-Learning for Domain Generalization | This work attempts to tackle the problem of domain generalization (DG) via learning to reduce domain shift with an episodic training procedure. In particular, we measure the domain shift with $\mathcal{Y}$-discrepancy and learn to optimize $\mathcal{Y}$-discrepancy between the unseen target domain and source domains only using source-domain samples. Theoretically, we give a PAC-style generalization bound for discrepancy-optimal meta-learning and further make comparisons with other DG bounds including ERM and domain-invariant learning. The theoretical analyses show that there is a tradeoff between classification performance and computational complexity for discrepancy-optimal meta-learning. The theoretical results also shed light on a bilevel optimization algorithm for DG. Empirically, we evaluate the algorithm with DomainBed and achieves state-of-the-art results on two DG benchmarks. | ['Chen Jia'] | 2021-09-29 | null | null | null | null | ['style-generalization'] | ['computer-vision'] | [ 1.48830220e-01 1.94846198e-01 -6.32285535e-01 -5.31703413e-01
-1.22066104e+00 -5.90635955e-01 5.71388453e-02 4.24207337e-02
-3.52582663e-01 1.26191604e+00 -2.95417368e-01 -1.74344793e-01
-5.35334408e-01 -7.77620375e-01 -8.68253410e-01 -8.24487984e-01
-2.70403773e-01 7.70568609e-01 6.95540309e-02 -3.49341512e-01
2.17516407e-01 1.93175808e-01 -1.50423074e+00 2.37746790e-01
1.32904589e+00 1.32705152e+00 -1.23508364e-01 2.61589170e-01
-1.73021466e-01 2.89933115e-01 -8.43426466e-01 -3.45635206e-01
6.47898912e-01 -4.92902428e-01 -8.85469496e-01 1.42534450e-01
6.27932787e-01 -6.89385980e-02 -2.13205233e-01 1.17668104e+00
4.06233132e-01 3.94828588e-01 8.67672741e-01 -1.30160224e+00
-8.26854706e-01 4.52538103e-01 -4.78375137e-01 5.27202785e-01
-7.26041496e-02 -1.91081405e-01 8.04860055e-01 -6.39524698e-01
5.65663457e-01 1.16563141e+00 6.14138901e-01 7.93619454e-01
-1.50389266e+00 -8.17270577e-01 4.62437004e-01 8.05096328e-02
-1.16423798e+00 1.41402185e-02 7.80145407e-01 -4.13800657e-01
8.54482651e-01 -9.23015550e-02 1.62557602e-01 1.11091900e+00
4.75151092e-02 7.56864667e-01 1.20422947e+00 -4.70525414e-01
7.35076845e-01 3.67353171e-01 1.96409568e-01 3.73827696e-01
4.72657353e-01 3.20169568e-01 -4.68128055e-01 -1.81946278e-01
6.59957647e-01 -2.65196025e-01 -1.06897265e-01 -8.64507139e-01
-7.26713300e-01 9.64215934e-01 2.45262906e-01 7.04442859e-02
2.70734169e-02 -1.41881630e-01 5.29087424e-01 8.46248329e-01
8.65764618e-01 6.17677987e-01 -9.65641499e-01 -3.88441756e-02
-8.73117030e-01 5.15209436e-01 7.88757384e-01 1.32410336e+00
8.63055646e-01 1.69341967e-01 9.32076797e-02 1.12009788e+00
-2.47234836e-01 5.12655854e-01 7.01273203e-01 -1.22792113e+00
8.97903025e-01 7.23032355e-01 1.08876511e-01 -4.27772641e-01
-2.69270301e-01 -6.78917825e-01 -6.69965029e-01 2.18436345e-01
7.24760413e-01 -1.86472774e-01 -6.44791961e-01 1.95620465e+00
3.44503760e-01 1.70142412e-01 2.80682832e-01 6.43527269e-01
3.71858031e-01 4.01376277e-01 5.81995137e-02 -4.05399531e-01
7.94166803e-01 -9.15914357e-01 -2.91069329e-01 -4.14086848e-01
1.04676139e+00 -1.80479690e-01 1.34471309e+00 4.99389201e-01
-1.06562865e+00 -6.15974307e-01 -1.31144810e+00 2.07684770e-01
-4.86739755e-01 -2.02592030e-01 4.52524185e-01 1.00262189e+00
-5.81631303e-01 9.69610929e-01 -5.11096954e-01 -2.19762161e-01
5.84497690e-01 4.21298504e-01 -9.54642966e-02 -1.59219671e-02
-1.10503495e+00 5.67860544e-01 8.20321083e-01 -7.91882813e-01
-8.53120506e-01 -1.19396329e+00 -7.40664959e-01 -2.21669406e-01
3.35207909e-01 -7.00067818e-01 1.23431242e+00 -9.80986476e-01
-1.49826777e+00 1.35532522e+00 -1.05541319e-01 -7.71843910e-01
5.47812939e-01 -1.10113308e-01 -6.72026575e-01 -1.17693022e-01
1.63712904e-01 5.72541177e-01 8.33043694e-01 -1.12269568e+00
-9.96826172e-01 -6.64078891e-01 4.92637940e-02 1.34416685e-01
-6.93468750e-01 -5.10491073e-01 1.05896980e-01 -8.87596250e-01
3.02814335e-01 -8.39843690e-01 2.93680355e-02 -1.86161906e-01
1.98703900e-01 -4.13492084e-01 7.86600590e-01 -4.55508053e-01
1.29663074e+00 -2.10649991e+00 1.24686055e-01 7.46584088e-02
9.04082954e-02 2.34358892e-01 -1.33089811e-01 -7.30384663e-02
-1.97250381e-01 -3.72386687e-02 -5.11853397e-01 -1.46187112e-01
1.03032731e-01 2.94167787e-01 -7.01562703e-01 5.25525749e-01
-7.75574818e-02 7.03094125e-01 -8.75541210e-01 -1.99006736e-01
-2.19138250e-01 -1.47752389e-01 -8.74113381e-01 5.31612057e-03
-5.90306342e-01 3.32728475e-01 -2.25586534e-01 7.59848714e-01
1.08823752e+00 -4.75344390e-01 3.02975535e-01 3.54759097e-01
3.30548346e-01 3.70922953e-01 -1.26454997e+00 1.94777000e+00
-4.67936993e-01 4.17921513e-01 -9.42430124e-02 -1.58279896e+00
1.25469530e+00 -1.80106476e-01 5.02639592e-01 -1.04294610e+00
8.11218843e-03 6.46987379e-01 -4.10094917e-01 -2.07439829e-02
4.04140234e-01 -3.01316530e-01 -3.58465075e-01 3.04717898e-01
2.51484603e-01 -7.41085410e-03 7.83360004e-02 -4.23927188e-01
8.59633446e-01 2.93266773e-01 1.64532781e-01 -6.72106087e-01
6.56336010e-01 2.81478763e-01 7.08906770e-01 7.98444629e-01
-2.73616135e-01 2.27946386e-01 5.61099648e-01 -4.63424236e-01
-1.23454058e+00 -1.52090669e+00 -4.18767601e-01 1.55066741e+00
5.39992273e-01 1.02815308e-01 -9.06844854e-01 -1.01677239e+00
4.37409729e-01 8.93997014e-01 -5.21308959e-01 -5.60580254e-01
-8.09932768e-01 -7.67893672e-01 4.30437148e-01 7.74768949e-01
7.98097491e-01 -4.34280038e-01 -2.95896351e-01 2.02417746e-01
-5.19878082e-02 -9.66988027e-01 -2.94652820e-01 5.46249926e-01
-1.55660355e+00 -7.50791132e-01 -8.60715091e-01 -1.01739681e+00
4.45694327e-01 4.25835550e-02 1.29984212e+00 -5.53194463e-01
1.90168954e-02 2.67165542e-01 -1.55259550e-01 -3.55869591e-01
-3.62504900e-01 5.20299375e-01 2.07492322e-01 -6.06196642e-01
7.86952078e-01 -8.16861153e-01 -5.47393799e-01 7.24891901e-01
-6.19362712e-01 -6.78059161e-01 2.54466087e-01 1.07702065e+00
8.68332863e-01 1.45666331e-01 1.02053106e+00 -7.86603630e-01
6.70662224e-01 -5.58919966e-01 -9.56647992e-01 3.03073108e-01
-1.05542564e+00 4.00674105e-01 8.53163004e-01 -8.10166001e-01
-1.04500306e+00 -2.73299545e-01 2.74077713e-01 -4.77870107e-01
-1.38642192e-01 -3.37078683e-02 -2.88496822e-01 -1.99881550e-02
1.27271807e+00 3.99740517e-01 -2.05679193e-01 -5.84879577e-01
3.63920152e-01 7.34979391e-01 5.96984029e-01 -1.29926753e+00
6.11495912e-01 5.59207559e-01 -4.10363488e-02 -5.83979845e-01
-1.29622829e+00 -3.67612481e-01 -6.40581548e-01 4.25689310e-01
2.15971917e-01 -9.91804898e-01 -4.63422894e-01 5.45721985e-02
-6.82958424e-01 -6.77678585e-01 -6.66318953e-01 3.53815824e-01
-1.12480462e+00 3.06000501e-01 -1.37748614e-01 -6.12268448e-01
-1.76377237e-01 -6.85945809e-01 9.23801005e-01 4.83598597e-02
-1.05176181e-01 -1.29184759e+00 1.18247025e-01 2.25042060e-01
1.22817591e-01 3.06384295e-01 1.09523141e+00 -8.46786737e-01
-3.58379096e-01 2.74725348e-01 -1.33536339e-01 7.08339930e-01
-1.04783960e-02 -1.04514420e+00 -9.77996230e-01 -6.10835552e-01
1.64553538e-01 -4.25247937e-01 8.38988543e-01 5.50329626e-01
1.49658227e+00 -3.17596614e-01 -5.06486297e-01 1.11018169e+00
1.47510386e+00 3.93712491e-01 5.07688403e-01 8.07666421e-01
8.53942111e-02 3.55965406e-01 1.07176828e+00 6.69876158e-01
1.66665763e-01 5.12356579e-01 1.53617233e-01 2.98407406e-01
-2.21845917e-02 -3.14702928e-01 1.86203390e-01 5.42390831e-02
3.41144592e-01 -1.16390154e-01 -7.47540236e-01 6.83851719e-01
-1.71590459e+00 -7.91816592e-01 3.31180602e-01 2.40825057e+00
1.02837265e+00 3.30954820e-01 4.18292433e-01 9.55417454e-02
7.75494337e-01 -7.55492449e-02 -1.15102899e+00 -4.95955348e-01
-4.53983158e-01 5.57636738e-01 7.96136677e-01 3.61318588e-01
-1.22587204e+00 1.01276016e+00 6.97115278e+00 1.05329740e+00
-9.68375564e-01 1.15797698e-01 6.27576888e-01 -2.81404316e-01
-1.32804543e-01 -3.21407437e-01 -1.24307799e+00 5.27801454e-01
1.03796685e+00 -6.13146603e-01 5.91968000e-01 1.51424408e+00
-4.32301491e-01 2.21537426e-01 -1.56510985e+00 1.16513824e+00
-1.99084282e-01 -1.40924764e+00 -2.35560820e-01 1.67864725e-01
1.19448090e+00 -1.77204907e-01 4.21875298e-01 7.90159166e-01
3.41762513e-01 -8.88694167e-01 5.15442789e-01 -1.02743357e-01
1.00599575e+00 -8.91211331e-01 3.82271469e-01 5.11524737e-01
-8.90655160e-01 -6.49286926e-01 -7.55901337e-01 -1.63313106e-03
-2.94662088e-01 6.40874386e-01 -7.55000591e-01 4.21713412e-01
8.59099984e-01 7.36672759e-01 -2.80164957e-01 7.82042205e-01
4.73367237e-02 3.79013330e-01 -2.70871162e-01 9.23368558e-02
2.70357151e-02 -1.10943876e-01 6.14406884e-01 1.11641788e+00
5.22014260e-01 1.68921482e-02 1.21609848e-02 1.06887770e+00
-3.88031632e-01 -5.22878282e-02 -5.39337516e-01 9.51864198e-02
6.54238522e-01 3.66153240e-01 -2.85066992e-01 -3.30071539e-01
-4.87446599e-02 1.06400287e+00 3.70185196e-01 2.70693421e-01
-8.72213960e-01 -4.47875172e-01 1.10488021e+00 3.16072524e-01
5.05476058e-01 -2.03723401e-01 -8.36317658e-01 -1.16385269e+00
3.00536424e-01 -8.98636103e-01 9.26009536e-01 -1.21486850e-01
-1.55417168e+00 2.47187883e-01 3.24998558e-01 -1.49424613e+00
-3.00422758e-01 -9.01756704e-01 2.83478461e-02 8.28073263e-01
-1.80220294e+00 -5.44595718e-01 -1.99695960e-01 7.18157411e-01
6.19582951e-01 -5.77307105e-01 7.05552816e-01 3.07433099e-01
-1.96607351e-01 1.26360154e+00 7.41065562e-01 -1.12048656e-01
8.85695159e-01 -1.30033350e+00 3.26614231e-01 3.17091882e-01
-3.58120441e-01 4.31790054e-01 5.91494083e-01 -5.72260678e-01
-9.97117519e-01 -1.19916892e+00 5.36950111e-01 -5.06435215e-01
4.19453681e-01 -2.94384658e-01 -9.91648912e-01 7.10325003e-01
-2.23899290e-01 -1.89376026e-02 6.97086573e-01 3.13761920e-01
-7.68459380e-01 -4.62765217e-01 -1.73720849e+00 3.36324871e-01
1.61088896e+00 -3.02886575e-01 -9.05759037e-01 3.25678140e-01
9.36082602e-01 -6.83953166e-01 -1.13239408e+00 4.88512754e-01
2.88424700e-01 -9.79958951e-01 1.25183785e+00 -8.49207103e-01
2.32960865e-01 1.33465677e-01 -4.18937892e-01 -1.43234754e+00
-2.18045145e-01 -4.05525327e-01 -3.88836116e-01 9.15526628e-01
3.68890941e-01 -9.83815134e-01 1.19267046e+00 4.04509932e-01
-4.60223369e-02 -6.18845463e-01 -1.14160049e+00 -1.79687965e+00
1.10427392e+00 -2.61387169e-01 7.58503497e-01 9.52107549e-01
2.09824238e-02 -2.30734758e-02 -4.99312468e-02 1.27855599e-01
9.09519494e-01 3.38025421e-01 6.31156504e-01 -1.48353088e+00
-3.84523481e-01 -4.82818782e-01 -2.74466634e-01 -1.53816497e+00
5.14168203e-01 -1.10594404e+00 -4.90341276e-01 -9.11509633e-01
-1.49833709e-01 -5.45881510e-01 -4.21534568e-01 1.94541752e-01
1.38317063e-01 -3.87102991e-01 -2.05550388e-01 7.34168962e-02
-6.92378879e-01 4.99467999e-01 1.18722892e+00 -3.06552052e-01
-6.18762255e-01 -4.81461436e-02 -9.28876042e-01 4.42768961e-01
9.91008043e-01 -6.86249018e-01 -6.52979553e-01 -3.38181823e-01
-2.81271134e-02 -5.39722890e-02 1.13760784e-01 -1.05822682e+00
-9.59754065e-02 -4.90362287e-01 4.52269435e-01 -5.54853499e-01
2.60135114e-01 -6.62290871e-01 -4.50249642e-01 4.48075682e-01
-4.63169068e-01 -3.15882891e-01 4.52377677e-01 6.11278415e-01
-7.20607117e-02 -1.25823781e-01 1.33308852e+00 -4.08382267e-02
-8.14079404e-01 2.08537862e-01 1.84318408e-01 8.61945808e-01
1.24308038e+00 -5.29291630e-01 -5.28873444e-01 2.91689690e-02
-7.55362689e-01 2.56798804e-01 4.48258966e-01 3.78289193e-01
3.35631311e-01 -1.44955850e+00 -6.13442779e-01 2.63145566e-01
3.26472580e-01 9.37865749e-02 1.30018830e-01 3.00583065e-01
-1.48545787e-01 4.67219293e-01 -2.24124536e-01 -7.47782528e-01
-8.12145770e-01 7.63964653e-01 5.69666564e-01 -4.20649469e-01
-4.00996387e-01 9.35401857e-01 1.88491315e-01 -7.26434648e-01
5.44738770e-01 -3.48961532e-01 4.69818681e-01 -8.49974900e-02
4.27107602e-01 8.89252782e-01 7.78015330e-02 1.29837140e-01
-3.67437184e-01 5.23779929e-01 -9.35843140e-02 -3.68555938e-03
1.28126013e+00 -7.09385872e-02 2.54538774e-01 3.01674783e-01
1.60692716e+00 -5.18068910e-01 -1.58665729e+00 -5.83619893e-01
4.03494358e-01 -7.20394731e-01 -3.30606014e-01 -9.01810288e-01
-6.06277943e-01 8.27467740e-01 1.02681351e+00 9.87844989e-02
1.21019602e+00 1.08292878e-01 8.52515578e-01 6.55069470e-01
4.74986017e-01 -1.91316962e+00 6.39992952e-01 6.81762576e-01
7.13017285e-01 -1.29209840e+00 -1.07361712e-01 -1.81228369e-01
-4.96954501e-01 9.57143128e-01 8.29503000e-01 -3.25596005e-01
6.90850973e-01 1.52641922e-01 -2.28041708e-01 3.13188255e-01
-6.05351686e-01 2.57328711e-02 1.57728568e-01 1.07616889e+00
1.00067906e-01 2.56010387e-02 -2.10955501e-01 8.66923749e-01
-4.58047330e-01 1.53718770e-01 -5.37656434e-02 7.98683584e-01
-8.30303729e-01 -1.26889944e+00 -4.39682633e-01 1.11670524e-01
-2.28960499e-01 2.55869359e-01 -1.50299340e-01 9.36383724e-01
2.30225995e-01 7.91640997e-01 7.80130699e-02 -2.36079797e-01
5.56924343e-01 3.67763996e-01 8.47090185e-01 -4.44645941e-01
-1.55680090e-01 -1.59956276e-01 -2.72667378e-01 -3.48711103e-01
-6.02048114e-02 -5.73193014e-01 -1.36065114e+00 -3.86207968e-01
2.08587706e-01 6.01603761e-02 4.09971535e-01 8.04899335e-01
4.86503124e-01 2.86352962e-01 8.11229229e-01 -3.18144232e-01
-1.30113196e+00 -5.10538697e-01 -8.28632534e-01 5.59571564e-01
3.55622262e-01 -8.40876520e-01 -5.85784674e-01 -8.85859579e-02] | [10.393521308898926, 3.014510154724121] |
26d5ec09-0810-4626-8ff7-e8ed9ad17bdb | continuous-cost-aggregation-for-dual-pixel | 2306.07921 | null | https://arxiv.org/abs/2306.07921v1 | https://arxiv.org/pdf/2306.07921v1.pdf | Continuous Cost Aggregation for Dual-Pixel Disparity Extraction | Recent works have shown that depth information can be obtained from Dual-Pixel (DP) sensors. A DP arrangement provides two views in a single shot, thus resembling a stereo image pair with a tiny baseline. However, the different point spread function (PSF) per view, as well as the small disparity range, makes the use of typical stereo matching algorithms problematic. To address the above shortcomings, we propose a Continuous Cost Aggregation (CCA) scheme within a semi-global matching framework that is able to provide accurate continuous disparities from DP images. The proposed algorithm fits parabolas to matching costs and aggregates parabola coefficients along image paths. The aggregation step is performed subject to a quadratic constraint that not only enforces the disparity smoothness but also maintains the quadratic form of the total costs. This gives rise to an inherently efficient disparity propagation scheme with a pixel-wise minimization in closed-form. Furthermore, the continuous form allows for a robust multi-scale aggregation that better compensates for the varying PSF. Experiments on DP data from both DSLR and phone cameras show that the proposed scheme attains state-of-the-art performance in DP disparity estimation. | ['Georgios Evangelidis', 'Sagi Katz', 'Sagi Monin'] | 2023-06-13 | null | null | null | null | ['disparity-estimation', 'stereo-matching-1'] | ['computer-vision', 'computer-vision'] | [ 4.32926536e-01 -1.88939050e-01 1.51640266e-01 -4.02581125e-01
-7.50484705e-01 -2.74184644e-01 3.55730325e-01 -5.10579348e-02
-4.46515799e-01 6.61722124e-01 -5.08772843e-02 -4.54904288e-02
1.86281037e-02 -9.18444395e-01 -7.69859970e-01 -8.42031538e-01
5.03500164e-01 1.10332109e-02 6.09064162e-01 2.79959687e-03
5.43783486e-01 5.74746072e-01 -1.98581874e+00 1.86733678e-02
1.27925062e+00 1.17856014e+00 5.45949101e-01 6.07349217e-01
1.37095088e-02 4.29452300e-01 -3.15933883e-01 -5.93042850e-01
6.79588854e-01 -1.87011361e-01 -2.06977487e-01 3.02980036e-01
1.09847856e+00 -7.98939764e-01 -1.70615420e-01 1.12297797e+00
5.21386862e-01 -8.12493786e-02 4.34290290e-01 -8.97960544e-01
-1.08622558e-01 -5.34675062e-01 -9.72572565e-01 5.26326895e-03
7.35827565e-01 6.31714091e-02 8.22531104e-01 -7.03431308e-01
4.40897763e-01 1.07277596e+00 6.24962568e-01 1.52845025e-01
-1.18006146e+00 -2.97876805e-01 -1.72105566e-01 4.51526307e-02
-1.29126585e+00 -4.47793365e-01 8.70412350e-01 -2.86184311e-01
8.04099023e-01 2.65734553e-01 7.29468226e-01 3.87755573e-01
3.24228078e-01 2.95262277e-01 1.27547479e+00 -5.84127486e-01
1.69702336e-01 1.71446040e-01 -7.63184279e-02 5.07112503e-01
2.54772246e-01 1.69429943e-01 -5.39865971e-01 -7.58819878e-02
1.11122549e+00 -6.03152886e-02 -5.90027869e-01 -7.05701768e-01
-1.01224303e+00 4.89537060e-01 4.29732084e-01 1.48352265e-01
-3.02921265e-01 -3.88204344e-02 -6.67814687e-02 5.77928275e-02
3.91864985e-01 2.48917341e-02 9.72583366e-04 -1.02075361e-01
-7.86382020e-01 2.35616714e-01 4.23287213e-01 8.20307255e-01
1.11057889e+00 -1.33251995e-01 2.53392249e-01 9.64934945e-01
2.54169762e-01 6.81667984e-01 1.43296510e-01 -1.38383484e+00
7.42402375e-01 5.10604978e-01 2.69955724e-01 -1.21457386e+00
-1.24101885e-01 -2.72435069e-01 -6.23505831e-01 7.58112371e-01
6.84877574e-01 1.44421011e-01 -4.06567931e-01 1.54264665e+00
5.08196592e-01 2.39896234e-02 -1.30282164e-01 9.58377481e-01
3.88774902e-01 6.58132911e-01 -5.64943075e-01 -3.69346559e-01
1.17980146e+00 -5.85458398e-01 -6.34348094e-01 -1.57066151e-01
7.99179077e-02 -8.64320695e-01 8.48048091e-01 5.65093398e-01
-1.40177941e+00 -5.74611723e-01 -1.10281932e+00 -3.92508626e-01
1.28751814e-01 -2.80295573e-02 3.28311324e-01 7.35217571e-01
-1.28002715e+00 3.57976794e-01 -5.76581895e-01 -3.23060274e-01
-1.20600447e-01 4.02278900e-01 -2.19948038e-01 -1.64434463e-01
-6.91087306e-01 8.32006276e-01 -3.68327163e-02 2.92493310e-02
-6.23078123e-02 -6.53267562e-01 -8.35032642e-01 5.59224971e-02
8.12326968e-02 -9.09482241e-01 7.41591752e-01 -8.61926198e-01
-1.72019613e+00 1.11719501e+00 -5.30172050e-01 -2.49856055e-01
7.74025083e-01 -1.93965614e-01 -1.81336571e-02 5.18492401e-01
6.31917492e-02 6.53554142e-01 8.23231161e-01 -1.33892703e+00
-8.08012187e-01 -6.94929302e-01 2.46812329e-01 6.69219434e-01
-2.23924443e-01 -1.04815893e-01 -5.81803143e-01 -2.34792218e-01
4.20355380e-01 -5.93679309e-01 -6.40064403e-02 3.98832858e-01
-4.75327447e-02 2.73319989e-01 6.11311555e-01 -4.95072991e-01
1.15988326e+00 -2.21338582e+00 1.11763492e-01 -1.86915677e-02
4.42375429e-02 -6.13991469e-02 2.09495187e-01 2.90535092e-01
2.10274503e-01 -6.92052484e-01 -3.71344656e-01 -6.09948337e-01
-4.31027651e-01 4.27268110e-02 5.50899021e-02 6.50862455e-01
-1.52837619e-01 3.45378250e-01 -6.72808707e-01 -5.17867684e-01
6.61135077e-01 5.59922457e-01 -5.42172492e-01 2.31448799e-01
1.21240005e-01 4.22987759e-01 -1.37415037e-01 5.93871951e-01
1.27878368e+00 1.98116750e-01 -2.38885712e-02 -3.82565200e-01
-6.12906694e-01 -8.17924589e-02 -1.55541599e+00 1.68337071e+00
-5.16199768e-01 6.91496313e-01 2.96253920e-01 -7.04809010e-01
1.10875595e+00 -4.71867695e-02 5.72277486e-01 -8.23323011e-01
-1.27654016e-01 7.45738745e-01 -4.14962888e-01 -2.02148557e-01
6.49661362e-01 -1.61864653e-01 3.57863396e-01 -2.66922731e-02
-4.09020871e-01 -4.84849393e-01 -7.52422288e-02 -1.88474163e-01
6.64103329e-01 1.16085000e-01 4.17465329e-01 -3.12163711e-01
9.38994288e-01 -2.53936589e-01 6.24763072e-01 3.74017656e-01
-1.26165643e-01 1.14061344e+00 1.69500381e-01 -2.14646116e-01
-1.06485128e+00 -9.75005209e-01 -3.92677248e-01 9.64038596e-02
6.77949667e-01 6.21403120e-02 -7.56442785e-01 -3.53237614e-02
8.40977803e-02 2.55179942e-01 -1.11099198e-01 4.23553765e-01
-7.07533479e-01 -3.33718717e-01 -1.78454127e-02 3.08693975e-01
9.10101771e-01 -4.04516280e-01 -1.06145406e+00 8.04421678e-02
-3.35383385e-01 -1.16764557e+00 -5.95216036e-01 -1.86059609e-01
-1.08692682e+00 -1.08464217e+00 -1.10711169e+00 -7.40879714e-01
6.77377522e-01 8.48231912e-01 8.58307123e-01 -2.84467191e-01
-1.93548188e-01 4.13356960e-01 4.90242653e-02 1.31322324e-01
-4.42294814e-02 -5.75709939e-01 -7.36486986e-02 2.46094421e-01
2.88269579e-01 -7.11120605e-01 -9.37709093e-01 5.56633472e-01
-7.99152613e-01 1.69707179e-01 2.51315653e-01 6.82820678e-01
7.36857295e-01 -5.22265732e-02 1.45305410e-01 -3.00607651e-01
2.20270649e-01 7.55487308e-02 -1.17343533e+00 1.59885623e-02
-6.54734850e-01 -3.30544114e-01 4.25396889e-01 8.48471969e-02
-1.46020329e+00 1.10452510e-01 -1.12656765e-02 -4.19101715e-01
-5.74272834e-02 -1.99310467e-01 -2.11294800e-01 -4.17551041e-01
5.76372206e-01 2.41073310e-01 2.22534537e-01 -2.67955482e-01
1.98584586e-01 7.59860992e-01 7.45355070e-01 -3.60920161e-01
6.05410635e-01 1.05923939e+00 3.28782976e-01 -1.10529196e+00
-2.96676099e-01 -7.40255177e-01 -6.21167421e-01 -4.17836010e-01
9.52950060e-01 -9.95678604e-01 -8.00652385e-01 8.05862010e-01
-1.09964466e+00 7.31912777e-02 -2.00987250e-01 5.46989679e-01
-7.78951168e-01 8.59659374e-01 -6.21531904e-01 -8.91265988e-01
-3.88746485e-02 -1.20352566e+00 1.20269513e+00 3.84376734e-01
1.65345341e-01 -1.00576222e+00 7.21558407e-02 6.00523710e-01
3.47920448e-01 2.31907278e-01 4.99302447e-01 6.73504114e-01
-1.10571456e+00 3.48084792e-02 -3.77634704e-01 5.73392749e-01
1.64796829e-01 1.07590601e-01 -1.27878833e+00 -1.39496535e-01
5.97296953e-01 6.61104172e-02 6.53303325e-01 9.36840475e-01
6.77105367e-01 1.16187155e-01 -1.20103441e-01 9.68154788e-01
1.96980238e+00 3.32563609e-01 9.35376048e-01 6.98274672e-01
6.06962800e-01 9.08477247e-01 8.77923846e-01 3.62524658e-01
5.35238266e-01 1.20623517e+00 5.76522827e-01 -2.77759254e-01
-2.19644189e-01 1.79800559e-02 2.07562909e-01 4.72041309e-01
-1.53342724e-01 -4.87033166e-02 -6.46260321e-01 3.67735952e-01
-1.65860224e+00 -8.48986924e-01 -6.35448694e-01 2.76149464e+00
5.91708899e-01 -1.22023836e-01 -2.43834332e-02 2.41628379e-01
7.45404124e-01 1.85401276e-01 -3.31579983e-01 -5.01737297e-01
-5.23798406e-01 -1.05462745e-01 7.53183126e-01 8.90559375e-01
-7.77588785e-01 3.76318663e-01 6.00257254e+00 8.06309164e-01
-1.08739018e+00 -1.01257667e-01 4.46985036e-01 1.16092570e-01
-5.57645857e-01 3.91167887e-02 -8.40142369e-01 5.66197276e-01
3.83051038e-01 -1.83881536e-01 2.37333715e-01 6.15759373e-01
4.08327103e-01 -8.52989078e-01 -9.35324848e-01 1.42801797e+00
2.09128410e-01 -9.82887268e-01 -2.03757852e-01 3.67232084e-01
8.54534626e-01 -9.81032923e-02 1.80810571e-01 -5.79878688e-01
-5.22571385e-01 -4.76119578e-01 6.49205148e-01 3.29774171e-01
8.30393255e-01 -7.64825761e-01 4.88544464e-01 4.45018709e-01
-1.15469372e+00 -3.99202630e-02 -5.62293410e-01 -8.23647752e-02
4.56734955e-01 8.40344489e-01 -3.53073418e-01 7.21427500e-01
5.63631117e-01 7.70100534e-01 -2.20544428e-01 1.22657692e+00
1.01705797e-01 -4.13199306e-01 -4.69944835e-01 4.12025839e-01
-8.23377743e-02 -9.18109477e-01 7.90417075e-01 9.24573839e-01
6.02367401e-01 7.80132338e-02 -2.72925168e-01 7.78324366e-01
3.19269896e-01 9.92546156e-02 -7.03814149e-01 6.96950197e-01
3.52112055e-01 9.36750352e-01 -4.87287968e-01 -4.51941974e-02
-7.39643514e-01 1.07834482e+00 7.88355768e-02 3.33628565e-01
-5.39026141e-01 -2.72768050e-01 7.25381792e-01 3.88896286e-01
1.25835463e-01 -2.65870005e-01 -5.51521838e-01 -1.24559665e+00
5.96970201e-01 -6.57391369e-01 1.11121066e-01 -7.75513351e-01
-9.61111605e-01 4.38538343e-01 -2.97558885e-02 -1.74891365e+00
-2.02066764e-01 -4.62891996e-01 -4.73713845e-01 1.08681965e+00
-2.04491687e+00 -8.04941177e-01 -5.98263204e-01 6.96354747e-01
5.59401274e-01 2.46659413e-01 5.42387843e-01 5.48312724e-01
-2.17931554e-01 3.69351357e-01 2.68184602e-01 -5.86430490e-01
6.62473083e-01 -1.02221334e+00 -1.98114174e-03 9.94426191e-01
-4.54978824e-01 4.67723817e-01 6.40378296e-01 -3.71054620e-01
-1.23200119e+00 -3.80723119e-01 8.55599105e-01 -2.32387260e-01
1.06002361e-01 -2.17904616e-02 -8.59175920e-01 1.68155417e-01
1.29812598e-01 -2.41182342e-01 1.93189278e-01 -4.49661702e-01
-1.21603601e-01 -6.09392166e-01 -1.30281901e+00 1.97570831e-01
9.15608346e-01 -6.19980395e-01 -3.64236385e-01 6.70254556e-03
3.77561122e-01 -6.70397162e-01 -5.90762556e-01 3.68836373e-01
8.48148525e-01 -1.94039798e+00 1.10901368e+00 6.24578595e-01
4.41914469e-01 -6.10518217e-01 -3.75968575e-01 -8.80762696e-01
1.41308561e-01 -5.58946252e-01 4.25556675e-02 1.15471804e+00
6.64272485e-03 -1.05627108e+00 6.92538142e-01 7.73559690e-01
-6.23069294e-02 -6.63471699e-01 -1.15859115e+00 -8.31387222e-01
-4.44838077e-01 -1.58532992e-01 1.71533614e-01 6.59035385e-01
-2.05015123e-01 -7.24485815e-02 -2.83714890e-01 2.00931892e-01
1.18005717e+00 1.56177908e-01 8.65050197e-01 -1.10316503e+00
-5.12238801e-01 -3.49774629e-01 -6.95585072e-01 -1.64079809e+00
-2.56294429e-01 -2.20036238e-01 -1.04469597e-01 -1.45252311e+00
1.90794140e-01 -6.28148258e-01 2.84460068e-01 -5.19519806e-01
-1.00398049e-01 1.57314241e-01 8.18006545e-02 4.89182472e-01
6.58667311e-02 3.26360971e-01 1.17931294e+00 3.45947653e-01
-3.51461947e-01 1.67030588e-01 -3.07951838e-01 8.50137770e-01
3.84152323e-01 -1.20565087e-01 -5.56590915e-01 -8.42370272e-01
1.58122554e-01 3.38094264e-01 2.45734796e-01 -1.07726049e+00
4.57752824e-01 2.61804014e-02 1.29481718e-01 -7.36787856e-01
8.73476267e-01 -1.02110648e+00 3.59402865e-01 3.23282063e-01
2.05030113e-01 -6.39749542e-02 -8.76290724e-02 5.69073379e-01
-6.20714664e-01 -2.09405765e-01 1.18448055e+00 -1.24261323e-02
-4.96782839e-01 4.96197902e-02 1.88454688e-01 -3.17277700e-01
1.14177251e+00 -1.14400148e+00 -3.31582844e-01 -5.27799845e-01
-1.07691914e-01 5.67834787e-02 1.17618835e+00 -5.23256287e-02
7.20987380e-01 -1.17404032e+00 -5.00045955e-01 3.52949589e-01
-7.39883631e-02 1.22440830e-01 4.45204645e-01 8.96928370e-01
-1.02767467e+00 4.96697128e-01 -2.41527706e-01 -1.07135570e+00
-1.41297042e+00 1.23672850e-01 5.26779234e-01 7.47819915e-02
-6.73340380e-01 8.85211945e-01 4.68192518e-01 2.65933424e-02
2.30330005e-01 -3.91194582e-01 -7.46516064e-02 -3.90502438e-02
4.75788981e-01 6.45753503e-01 2.75978874e-02 -7.18688428e-01
-2.88054287e-01 1.45923746e+00 3.53841692e-01 -3.07134509e-01
1.04688895e+00 -7.53641009e-01 -1.08250588e-01 3.08905035e-01
1.25109315e+00 4.85026449e-01 -1.59876406e+00 -8.30921307e-02
-4.99557287e-01 -1.26465499e+00 6.95094466e-02 -1.76190704e-01
-9.19339359e-01 9.31882024e-01 6.46911383e-01 1.42635569e-01
1.46467209e+00 -4.42733496e-01 8.33789587e-01 -1.39743865e-01
6.91863537e-01 -1.01769078e+00 -1.21326581e-01 1.12834647e-01
5.83649278e-01 -1.25922048e+00 4.61149924e-02 -9.42070901e-01
-4.15109366e-01 1.29699445e+00 4.71552163e-01 -1.04007654e-01
3.19749266e-01 2.01363310e-01 1.23090714e-01 1.43896848e-01
-2.27896065e-01 -1.88116178e-01 1.45166829e-01 6.11686170e-01
2.60338187e-01 -1.89261630e-01 -4.98402029e-01 -4.22027528e-01
6.03482574e-02 -3.72042730e-02 6.46944702e-01 6.79982960e-01
-6.33699596e-01 -1.08842194e+00 -6.24307573e-01 -8.32434744e-02
-2.12764442e-01 3.09397113e-02 1.53402075e-01 5.83984554e-01
1.39656529e-01 1.01322687e+00 1.78733006e-01 9.02424678e-02
6.53721631e-01 -3.89057994e-01 6.65400028e-01 -3.93251240e-01
-2.19114453e-01 2.91371137e-01 -4.69411276e-02 -8.32678437e-01
-7.10161686e-01 -6.87046707e-01 -1.02980947e+00 -2.59362906e-01
-3.25786710e-01 -2.48874471e-01 8.39505792e-01 4.24923211e-01
1.88235253e-01 -5.37280217e-02 9.50722158e-01 -8.51441860e-01
-2.98340410e-01 -5.11111975e-01 -7.96712160e-01 4.28030342e-01
5.40385425e-01 -5.85077226e-01 -7.68070936e-01 -8.53196681e-02] | [9.206759452819824, -2.507826089859009] |
62d75100-d30a-4b57-92cf-c6e6db0e820e | ireen-iterative-reverse-engineering-of-black | 2006.10720 | null | https://arxiv.org/abs/2006.10720v2 | https://arxiv.org/pdf/2006.10720v2.pdf | IReEn: Reverse-Engineering of Black-Box Functions via Iterative Neural Program Synthesis | In this work, we investigate the problem of revealing the functionality of a black-box agent. Notably, we are interested in the interpretable and formal description of the behavior of such an agent. Ideally, this description would take the form of a program written in a high-level language. This task is also known as reverse engineering and plays a pivotal role in software engineering, computer security, but also most recently in interpretability. In contrast to prior work, we do not rely on privileged information on the black box, but rather investigate the problem under a weaker assumption of having only access to inputs and outputs of the program. We approach this problem by iteratively refining a candidate set using a generative neural program synthesis approach until we arrive at a functionally equivalent program. We assess the performance of our approach on the Karel dataset. Our results show that the proposed approach outperforms the state-of-the-art on this challenge by finding an approximately functional equivalent program in 78% of cases -- even exceeding prior work that had privileged information on the black-box. | ['Mateusz Malinowski', 'Mario Fritz', 'Hossein Hajipour'] | 2020-06-18 | null | https://openreview.net/forum?id=HyLGEHFzqzM | https://openreview.net/pdf?id=HyLGEHFzqzM | neurips-workshop-cap-2020-12 | ['computer-security'] | ['miscellaneous'] | [ 6.12622499e-01 4.98420596e-01 -1.46206683e-02 -8.00216049e-02
-4.45130885e-01 -9.01596725e-01 8.44843030e-01 2.18102440e-01
-9.32962969e-02 4.03121352e-01 -1.35710984e-01 -8.66302907e-01
2.47201733e-02 -8.38352203e-01 -9.92527723e-01 -3.95357519e-01
5.93216158e-02 3.99185061e-01 2.52720416e-01 -2.94248223e-01
2.45954365e-01 2.70483106e-01 -1.64823806e+00 1.28518835e-01
6.47517145e-01 3.88703078e-01 -8.89505073e-02 7.73830473e-01
3.74947041e-01 8.06960583e-01 -7.02669501e-01 -4.53344882e-01
1.82032973e-01 -5.20057261e-01 -9.16661322e-01 1.67549521e-01
1.23181097e-01 -2.84479916e-01 1.61677629e-01 1.36939073e+00
-2.11861789e-01 -3.68133217e-01 3.30436587e-01 -1.57702458e+00
-3.58871579e-01 8.43262434e-01 -1.89562380e-01 -4.25460517e-01
5.53536475e-01 3.87560219e-01 1.14494264e+00 -3.18061143e-01
6.65245533e-01 1.09237492e+00 2.85568595e-01 6.96271837e-01
-1.79342532e+00 -2.76627213e-01 1.37094362e-02 -2.54215151e-01
-1.23146379e+00 -3.80447537e-01 6.19508684e-01 -8.08889985e-01
8.38716328e-01 4.52586472e-01 4.21870679e-01 1.01116037e+00
1.57747686e-01 4.07243460e-01 1.31436110e+00 -6.70889199e-01
5.92293978e-01 4.18913543e-01 5.21433711e-01 8.70219529e-01
5.26902497e-01 3.98462296e-01 -1.73795566e-01 -6.50820613e-01
3.22854817e-01 -1.65813521e-01 -1.75474763e-01 -7.02752709e-01
-1.18561244e+00 8.24845552e-01 -6.09029271e-02 2.04553932e-01
-6.16640672e-02 4.58580583e-01 2.68785626e-01 5.49379885e-01
-7.86773860e-02 8.16312730e-01 -4.00128096e-01 -1.59551829e-01
-6.93603873e-01 4.10083920e-01 1.42246234e+00 8.00175071e-01
6.54087365e-01 2.00825572e-01 2.69837707e-01 -3.16554010e-01
2.99950957e-01 1.12441644e-01 1.00810438e-01 -8.16551864e-01
3.23873609e-01 8.05063248e-01 1.88265622e-01 -5.98308742e-01
-1.49683490e-01 -2.99648702e-01 -4.22822088e-01 8.72688055e-01
6.06707931e-01 -2.29357451e-01 -6.64974153e-01 1.84048748e+00
1.25083834e-01 -8.41064937e-03 2.16030642e-01 6.74347401e-01
2.39405960e-01 5.84457457e-01 -3.36882740e-01 -1.45229269e-02
1.51703894e+00 -7.68547893e-01 -2.57800698e-01 -4.11136508e-01
5.09795308e-01 -5.39579093e-02 1.09201109e+00 5.43901026e-01
-7.47490942e-01 -3.48334163e-01 -1.53099477e+00 3.67990106e-01
-2.20266759e-01 9.66219977e-02 7.08401084e-01 1.18064415e+00
-1.31902027e+00 3.55070710e-01 -8.09463739e-01 -2.45812684e-01
2.22033501e-01 5.79669595e-01 -4.17759180e-01 2.10337073e-01
-7.48639703e-01 7.00912356e-01 4.40039992e-01 -1.98786885e-01
-1.33297348e+00 -5.77676415e-01 -1.11213851e+00 2.07715765e-01
9.14615691e-01 -7.62068093e-01 1.49471569e+00 -1.16718817e+00
-1.57259250e+00 7.91143060e-01 -1.29946962e-01 -8.03858042e-01
7.44419396e-01 9.37992483e-02 -1.73160106e-01 -3.06237161e-01
-3.17692965e-01 3.06161702e-01 9.44542348e-01 -1.42298901e+00
-4.26825106e-01 -3.69457871e-01 1.01868689e+00 -5.60434222e-01
1.33910701e-01 1.24197036e-01 2.51585953e-02 -2.79759616e-01
-3.79418671e-01 -1.31816781e+00 -2.96665072e-01 -3.19260582e-02
-5.16840219e-01 9.72360820e-02 7.76718199e-01 -6.29681110e-01
1.13129425e+00 -2.18113637e+00 3.10961634e-01 4.00264829e-01
4.43568707e-01 5.34503199e-02 2.91022837e-01 4.64778870e-01
-3.05343002e-01 4.80529219e-01 -5.46972096e-01 -7.09187463e-02
5.04448593e-01 9.89304334e-02 -6.44299626e-01 6.77994013e-01
5.12712836e-01 8.85180593e-01 -7.09368825e-01 -4.26299945e-02
1.35085909e-02 1.45800769e-01 -8.54596257e-01 2.40315080e-01
-8.66498768e-01 5.48843980e-01 -5.74187815e-01 3.69926423e-01
4.29576069e-01 -2.68418849e-01 5.81196785e-01 3.37118506e-01
-9.23837498e-02 4.71082032e-02 -1.17138207e+00 1.58666694e+00
-6.45650446e-01 5.24664581e-01 -4.62234505e-02 -8.51114035e-01
8.57694805e-01 3.21253002e-01 -1.73552126e-01 -1.86117768e-01
4.34136130e-02 1.79933593e-01 4.06657666e-01 -2.82493025e-01
4.38306868e-01 -2.10766926e-01 -5.62065601e-01 8.98414731e-01
-4.01543468e-01 -1.33366078e-01 1.07792400e-01 2.35013515e-02
1.31760633e+00 5.48015416e-01 8.25035036e-01 -3.73348206e-01
6.98891282e-01 5.10262921e-02 2.49731869e-01 1.01375067e+00
3.61848772e-01 2.21756428e-01 1.37012470e+00 -3.89918834e-01
-1.16456795e+00 -6.60950005e-01 3.65821451e-01 7.69795179e-01
-2.45442763e-01 -7.64541328e-01 -1.08829355e+00 -8.71429741e-01
-1.49613827e-01 1.04137027e+00 -7.46638298e-01 -2.88672626e-01
-6.36211634e-01 -3.31872195e-01 6.70794129e-01 2.12322041e-01
1.28660142e-01 -1.04043972e+00 -1.38531506e+00 -7.67681226e-02
2.24048540e-01 -9.32267010e-01 -2.08502263e-01 2.40509287e-01
-5.42976677e-01 -1.20400250e+00 -1.11531600e-01 -5.07462561e-01
1.03167963e+00 -5.42589068e-01 1.17644048e+00 4.24623847e-01
-1.60588279e-01 1.84334695e-01 -1.68108419e-01 -1.63642973e-01
-1.20167851e+00 1.66060254e-01 -3.16833079e-01 -1.15837539e-02
-5.24851354e-03 -4.36549485e-01 -1.60344746e-02 6.23055249e-02
-1.34506810e+00 9.19290558e-02 4.96284246e-01 7.24242449e-01
5.00244275e-02 5.03641009e-01 2.01491043e-01 -1.42918360e+00
5.98088026e-01 -2.89226860e-01 -1.41982853e+00 3.16407382e-01
-6.06030107e-01 8.33351254e-01 1.16304994e+00 -3.76102775e-01
-8.33823562e-01 4.98468310e-01 1.46049157e-01 8.12493414e-02
-3.37567508e-01 5.27304053e-01 -7.10363746e-01 -7.52037764e-02
5.94846725e-01 2.94588715e-01 -6.05989173e-02 -1.38488486e-01
1.58088401e-01 3.82156938e-01 5.39827764e-01 -9.95111704e-01
1.24387264e+00 3.25738877e-01 5.15991867e-01 -3.49613130e-01
2.41044816e-02 2.17548162e-01 -4.01514947e-01 1.97028250e-01
4.02913570e-01 -3.55830997e-01 -1.25724876e+00 2.60920793e-01
-1.33515620e+00 -2.10473850e-01 -2.79267609e-01 -3.32453609e-01
-7.74889231e-01 1.88207895e-01 3.13801691e-02 -1.07429588e+00
4.08586450e-02 -1.82603645e+00 1.06033766e+00 -9.77780074e-02
-5.29061615e-01 -7.02331662e-01 2.97604799e-01 9.03882310e-02
2.74987936e-01 4.67340082e-01 1.44068658e+00 -1.07903254e+00
-7.98742414e-01 -3.32934856e-01 2.52274007e-01 9.69680101e-02
1.38458200e-02 -6.57297075e-02 -1.09528720e+00 -2.20729291e-01
7.58105963e-02 -7.92516686e-04 2.30335489e-01 -2.35790238e-01
9.55494821e-01 -6.25827610e-01 -1.40482098e-01 3.49207580e-01
1.38837445e+00 2.48576909e-01 6.37097538e-01 2.68201470e-01
2.58627057e-01 7.14226961e-01 2.55674720e-01 3.80901456e-01
2.45682254e-01 8.71137321e-01 7.05096602e-01 1.44014910e-01
4.29506302e-01 -4.56688672e-01 5.03187299e-01 -3.21972340e-01
1.95165187e-01 -6.66739866e-02 -9.99240398e-01 3.83717537e-01
-1.80556130e+00 -8.84883106e-01 1.84933245e-01 2.37983394e+00
6.85077190e-01 2.58031398e-01 3.12040448e-01 2.45629266e-01
6.55810416e-01 -4.12587356e-03 -4.77131844e-01 -5.25239050e-01
3.54238361e-01 3.31532121e-01 2.60670245e-01 6.77008212e-01
-1.03225970e+00 4.03995126e-01 5.42224169e+00 2.71593750e-01
-9.72823739e-01 2.72337869e-02 4.16320294e-01 4.24184620e-01
-5.09862840e-01 6.81016505e-01 -5.94093204e-01 3.65967304e-01
1.16426456e+00 -4.37812418e-01 8.10875356e-01 9.33432341e-01
-4.39969311e-03 -2.37979308e-01 -1.82441592e+00 2.43343443e-01
1.56997666e-02 -9.85620141e-01 -2.18993917e-01 3.16284716e-01
4.12454009e-01 -7.30309784e-01 2.08560646e-01 2.83609390e-01
5.27070463e-01 -1.23576093e+00 1.24327934e+00 3.66245568e-01
5.15972376e-01 -7.80517876e-01 6.04894400e-01 5.98136425e-01
-9.68947709e-01 -7.77438059e-02 4.22658533e-01 -1.87431812e-01
-3.23242307e-01 7.35462382e-02 -1.01630664e+00 5.73790669e-01
1.04308106e-01 2.58907914e-01 -5.96896768e-01 7.82605529e-01
-4.97694790e-01 3.66133660e-01 -2.00391263e-01 -1.63359389e-01
-1.56677458e-02 -3.02704656e-03 6.06228292e-01 9.16955054e-01
2.93135643e-01 -8.44496787e-02 6.03788942e-02 1.48157573e+00
-1.99105591e-02 -5.30347109e-01 -7.56465673e-01 -3.66072744e-01
-6.00299388e-02 1.04178655e+00 -6.11881852e-01 -1.77614421e-01
-3.49793643e-01 7.53881991e-01 5.42306416e-02 2.69629568e-01
-8.86262953e-01 -2.40707889e-01 3.45193207e-01 1.82085618e-01
3.29018801e-01 -2.62247801e-01 -3.57614681e-02 -1.07404327e+00
3.95336390e-01 -1.40681183e+00 2.07244992e-01 -6.55737221e-01
-6.04175508e-01 7.48940527e-01 1.92377061e-01 -1.11388695e+00
-7.93703020e-01 -6.05098903e-01 -5.98905325e-01 9.12282109e-01
-1.12432075e+00 -1.27040493e+00 1.17206335e-01 1.62245572e-01
2.54057705e-01 -3.16667110e-02 9.89991486e-01 -2.09120318e-01
-3.68334651e-01 4.07264262e-01 -3.21480900e-01 -2.45335370e-01
5.37886359e-02 -1.32291722e+00 6.30778313e-01 1.31127679e+00
4.71761934e-02 1.10430634e+00 1.13784623e+00 -6.21228993e-01
-2.09074569e+00 -9.68471229e-01 6.36342824e-01 -5.89966476e-01
9.85632122e-01 -7.74213791e-01 -6.23657703e-01 1.10491896e+00
2.35718414e-01 -1.09121479e-01 2.55419374e-01 -2.93032378e-01
-4.70219612e-01 1.26539871e-01 -1.11473441e+00 6.50406182e-01
7.10752785e-01 -7.35921204e-01 -6.17132306e-01 -5.40209748e-02
7.80888021e-01 -3.33362043e-01 -4.05042201e-01 -1.15505811e-02
6.35445416e-01 -8.67994487e-01 6.27129793e-01 -7.92923748e-01
5.39554775e-01 -8.50058496e-01 -1.29314726e-02 -1.14997387e+00
4.58452672e-01 -1.06593263e+00 -1.59090430e-01 1.12448037e+00
6.29861653e-01 -7.74715960e-01 8.08699131e-01 8.91273916e-01
9.16652530e-02 -2.62962192e-01 -7.23363817e-01 -6.30200028e-01
-1.98032603e-01 -4.19278234e-01 8.70721519e-01 5.54925680e-01
3.41230482e-01 1.74852788e-01 -3.44569653e-01 5.90553582e-01
5.95549047e-01 5.29887557e-01 9.07860041e-01 -1.13374388e+00
-9.73116696e-01 -4.23803985e-01 -4.19540972e-01 -6.53292894e-01
7.54566431e-01 -8.95151198e-01 1.40917808e-01 -9.22294438e-01
3.65749031e-01 7.70769641e-02 2.74234354e-01 5.63686609e-01
1.72066599e-01 -1.83901533e-01 2.37077788e-01 -5.44176027e-02
-3.75290871e-01 1.21758036e-01 5.54643810e-01 -3.86437356e-01
-5.45709841e-02 3.35944325e-01 -9.34694946e-01 6.05679095e-01
5.95020771e-01 -6.01706803e-01 -4.01290536e-01 -1.26669496e-01
6.88876152e-01 4.38745528e-01 9.42953706e-01 -8.16138625e-01
2.80170381e-01 -3.17816496e-01 -4.99674678e-01 5.45476153e-02
2.06286058e-01 -1.34584701e+00 8.17775548e-01 8.27536523e-01
-4.04017746e-01 2.50907749e-01 1.22435950e-01 3.86089057e-01
-3.99075225e-02 -7.29331553e-01 3.64858806e-01 -9.86927003e-02
-7.24573791e-01 -2.18224213e-01 -6.37280643e-01 -3.12014341e-01
1.11181164e+00 5.12389354e-02 -6.10716045e-01 -4.05524045e-01
-5.74487388e-01 -1.28435656e-01 1.06418812e+00 2.25671783e-01
3.96356136e-01 -6.93134546e-01 -3.78065467e-01 5.46301782e-01
3.79700899e-01 -2.77218401e-01 -3.01913798e-01 5.92118084e-01
-5.96072912e-01 5.05869806e-01 -1.67300582e-01 -2.64629126e-01
-1.23126101e+00 6.76145911e-01 3.82359833e-01 -4.20448840e-01
-3.16561073e-01 2.81466544e-03 4.77036715e-01 -5.29826581e-01
-1.52927548e-01 -4.93193537e-01 -1.22657595e-02 -5.56838870e-01
4.45312858e-01 -7.24290162e-02 5.51734120e-02 -3.74780238e-01
-5.35591602e-01 3.12463224e-01 1.68967079e-02 -3.55896384e-01
1.35245216e+00 4.58569258e-01 -5.16512752e-01 2.42353857e-01
8.44068110e-01 1.23605810e-01 -8.33518803e-01 -9.40107089e-03
2.04398856e-01 -3.53858769e-01 -1.71660706e-01 -7.20800996e-01
-7.62236416e-01 7.50915885e-01 2.37142548e-01 5.22946537e-01
9.81381714e-01 1.48716927e-01 6.81667253e-02 6.64336026e-01
7.74792194e-01 -2.60910183e-01 -1.83194816e-01 3.40372115e-01
6.57710493e-01 -9.51949656e-01 2.99707297e-02 -3.33238810e-01
-5.13438821e-01 1.11890888e+00 5.02792060e-01 -1.17755622e-01
2.28353605e-01 7.37532198e-01 -4.30410475e-01 -6.32959530e-02
-7.65966952e-01 1.68536589e-01 1.06375106e-01 5.35879850e-01
9.03362632e-02 2.86632199e-02 8.22922215e-02 8.59263599e-01
-3.73863548e-01 1.61449134e-01 9.90889370e-01 1.37787104e+00
-1.94829016e-04 -1.40704715e+00 -5.19190431e-01 6.63787425e-02
-4.50800776e-01 -1.45572394e-01 -6.49884939e-01 9.59895909e-01
-1.97620481e-01 9.27245796e-01 -3.55360508e-01 -4.04224545e-01
3.24173838e-01 1.03180803e-01 4.35656518e-01 -7.48524785e-01
-8.43799412e-01 -2.36454606e-01 3.18300247e-01 -4.00936574e-01
-2.51915574e-01 -5.55722594e-01 -1.00490963e+00 -3.16602826e-01
-1.41092375e-01 1.80287167e-01 5.93301952e-01 9.15076673e-01
1.23075359e-01 5.42275071e-01 4.96827215e-01 -5.85853815e-01
-7.50683784e-01 -3.18094283e-01 -4.29459512e-01 1.46982253e-01
7.07903147e-01 -2.72069484e-01 -2.97124147e-01 3.73635143e-01] | [8.22209358215332, 7.345916271209717] |
ed942d58-4399-45ed-8ad8-57351b247952 | bridging-natural-language-processing-and | 2304.09616 | null | https://arxiv.org/abs/2304.09616v2 | https://arxiv.org/pdf/2304.09616v2.pdf | Bridging Natural Language Processing and Psycholinguistics: computationally grounded semantic similarity datasets for Basque and Spanish | We present a computationally-grounded word similarity dataset based on two well-known Natural Language Processing resources; text corpora and knowledge bases. This dataset aims to fulfil a gap in psycholinguistic research by providing a variety of quantifications of semantic similarity in an extensive set of noun pairs controlled by variables that play a significant role in lexical processing. The dataset creation has consisted in three steps, 1) computing four key psycholinguistic features for each noun; concreteness, frequency, semantic and phonological neighbourhood density; 2) pairing nouns across these four variables; 3) for each noun pair, assigning three types of word similarity measurements, computed out of text, Wordnet and hybrid embeddings. The present dataset includes noun pairs' information in Basque and European Spanish, but further work intends to extend it to more languages. | ['I. San Martin', 'M. Arantzeta', 'J. Goikoetxea'] | 2023-04-19 | null | null | null | null | ['word-similarity', 'semantic-textual-similarity', 'semantic-similarity'] | ['natural-language-processing', 'natural-language-processing', 'natural-language-processing'] | [-2.44024098e-01 -8.61573592e-02 -3.98736186e-02 -4.36890572e-01
-2.63680786e-01 -6.19185686e-01 9.01883543e-01 1.12738323e+00
-1.12861753e+00 4.57351685e-01 7.22006559e-01 -1.00099981e-01
-3.21558148e-01 -1.06696081e+00 -1.86625645e-01 -2.77875096e-01
3.55713554e-02 7.09947169e-01 2.51237065e-01 -6.74254477e-01
7.78709888e-01 4.98146474e-01 -1.74374688e+00 3.36258896e-02
7.30474472e-01 7.14991808e-01 5.01387477e-01 1.78617731e-01
-2.25994408e-01 3.22740763e-01 -1.95969507e-01 -6.79732442e-01
-3.39997187e-02 -3.73186439e-01 -1.05294228e+00 -3.66814375e-01
1.97369456e-01 3.18621427e-01 1.30279586e-01 1.24880791e+00
5.32700121e-01 6.88999474e-01 8.91741693e-01 -8.29273880e-01
-1.21736753e+00 9.57577586e-01 1.02691710e-01 6.75269127e-01
8.18303943e-01 -2.20218766e-02 1.46891344e+00 -1.10120428e+00
7.23112643e-01 1.65817189e+00 5.14915347e-01 2.29195297e-01
-1.16014564e+00 -4.11009640e-01 -3.35311800e-01 4.36133832e-01
-1.34945059e+00 -1.70942634e-01 6.15771770e-01 -4.21219528e-01
1.27836859e+00 -7.56026730e-02 7.85877228e-01 9.38442886e-01
4.82473075e-02 -1.85618494e-02 1.37476456e+00 -9.26413119e-01
4.35612202e-01 3.81678224e-01 5.53997874e-01 3.52959871e-01
3.55525203e-02 2.28611603e-01 -5.86767316e-01 -5.58644980e-02
6.32387519e-01 -3.04396242e-01 6.48420900e-02 -1.01318825e-02
-1.54674983e+00 1.17807400e+00 3.23330075e-01 8.68481636e-01
-5.57641685e-01 -2.68299192e-01 7.62287438e-01 4.67815757e-01
3.99455369e-01 7.39712894e-01 -6.86009884e-01 -2.44510368e-01
-4.01584893e-01 6.24829710e-01 8.62504184e-01 7.57210910e-01
1.03992963e+00 -3.13907713e-01 7.75607005e-02 1.19509637e+00
3.94768119e-01 3.96833688e-01 1.15686572e+00 -4.89605308e-01
3.61682773e-01 5.98262966e-01 -2.98022717e-01 -1.33652258e+00
-5.53593695e-01 4.91504818e-01 -1.92665353e-01 -3.01782846e-01
2.98823565e-01 2.70003438e-01 -4.44435209e-01 1.93755317e+00
3.89017791e-01 -3.50858003e-01 1.31015167e-01 7.90769160e-01
1.26997077e+00 4.66591865e-01 5.72438478e-01 -1.33898601e-01
2.03751731e+00 -3.86953622e-01 -7.35959232e-01 -1.45971760e-01
6.24557197e-01 -8.77884030e-01 1.58161139e+00 -2.32775509e-02
-1.17500734e+00 -8.16438735e-01 -8.00556898e-01 -4.29442525e-01
-1.18769062e+00 -3.89108926e-01 7.29710579e-01 6.96058869e-01
-1.06307197e+00 4.31950361e-01 -3.61919343e-01 -9.19623613e-01
-2.10852325e-01 1.32971942e-01 -6.66088283e-01 7.10034147e-02
-1.70448828e+00 1.52896142e+00 1.06687200e+00 -3.76582175e-01
-3.24767500e-01 -7.10719883e-01 -1.36346686e+00 -6.04377948e-02
1.58999950e-01 -5.01863718e-01 8.05849373e-01 -7.16357052e-01
-1.22400260e+00 1.61125946e+00 -3.47210504e-02 -1.98873803e-01
-2.29228362e-01 7.53455162e-02 -6.60656512e-01 -1.70126036e-01
5.65094531e-01 7.05268741e-01 3.07187229e-01 -8.62412810e-01
-4.05507565e-01 -7.93428838e-01 4.06684019e-02 4.26406443e-01
-6.05705202e-01 6.82578564e-01 -1.14266593e-02 -1.07136285e+00
6.55020848e-02 -6.69794440e-01 -4.53348495e-02 -3.33091915e-01
5.77304177e-02 -8.29513550e-01 -7.36471564e-02 -6.23693585e-01
1.05387414e+00 -2.13484883e+00 2.11843312e-01 1.21578962e-01
1.71523556e-01 5.78910895e-02 -2.99065560e-01 7.11509645e-01
-1.11202702e-01 1.06309026e-01 -8.64794329e-02 -3.20104845e-02
3.74444276e-01 3.83462191e-01 1.45065516e-01 3.40837717e-01
1.21865183e-01 9.99103665e-01 -1.11606061e+00 -7.47230411e-01
4.45518583e-01 1.22331038e-01 -6.76135540e-01 -7.99116194e-02
1.14568405e-01 -2.56008297e-01 -1.80346534e-01 3.17530364e-01
4.24364775e-01 7.00354040e-01 1.81911513e-01 -1.16816059e-01
-3.13862860e-01 7.45654523e-01 -1.13114059e+00 1.76970351e+00
-5.33605516e-01 3.23046237e-01 -3.87073457e-01 -1.02905500e+00
1.09910965e+00 5.66929132e-02 6.94113225e-02 -9.29989219e-01
6.09689593e-01 2.81706780e-01 2.11315244e-01 -5.49613833e-01
9.18682098e-01 -3.82110476e-01 -4.56198066e-01 3.36046815e-01
3.98633897e-01 -5.81618547e-01 6.24181986e-01 -2.21489504e-01
7.91580021e-01 -2.26570681e-01 8.44891906e-01 -9.80374753e-01
6.72568679e-01 -4.41468395e-02 4.41247344e-01 2.73281515e-01
-3.71594220e-01 2.19312012e-01 2.82554805e-01 -4.30035859e-01
-1.01678264e+00 -1.39448178e+00 -7.30407417e-01 1.47557127e+00
9.75326747e-02 -4.30200160e-01 -6.83506846e-01 -1.05148733e-01
7.56497011e-02 9.51126277e-01 -8.35189819e-01 -2.24141791e-01
-2.56460756e-01 -5.50631285e-01 5.55318475e-01 5.88277280e-01
1.95007980e-01 -1.66605914e+00 -4.46732581e-01 1.68672219e-01
-2.17975155e-02 -1.04787385e+00 -3.14923078e-01 3.32943588e-01
-6.81081057e-01 -9.23427880e-01 -9.57974494e-02 -1.24849176e+00
6.18853755e-02 5.00208475e-02 1.59361267e+00 8.65020882e-03
-4.54938561e-01 1.59267798e-01 -7.96822727e-01 -6.98943734e-01
-5.28814495e-01 -1.38521999e-01 4.55283284e-01 -5.62981188e-01
1.23845530e+00 -5.37293434e-01 5.38955070e-02 5.75037338e-02
-9.69056487e-01 -6.24020815e-01 2.23677084e-01 7.43776679e-01
4.51577812e-01 -1.31141603e-01 5.16780734e-01 -6.15645945e-01
1.09160542e+00 -8.25842798e-01 -2.59868115e-01 -1.25453874e-01
-3.89140189e-01 -2.80761626e-02 4.67788517e-01 -2.65904903e-01
-8.28604162e-01 -4.48104501e-01 -2.55965769e-01 3.29453588e-01
-5.42256773e-01 6.30127668e-01 -3.36007655e-01 3.63209099e-01
8.32632780e-01 5.24009727e-02 -2.58423761e-02 -3.44683915e-01
7.95253634e-01 6.11122310e-01 6.44182622e-01 -8.55152428e-01
5.53163230e-01 7.06761703e-03 -2.79972047e-01 -1.20145822e+00
-4.95817244e-01 -6.85408771e-01 -1.14542508e+00 1.80870876e-01
1.14307153e+00 -6.52402878e-01 -7.97694504e-01 1.77260935e-01
-1.21278608e+00 3.42839181e-01 -5.25439501e-01 8.47812533e-01
-5.32580614e-01 2.99458832e-01 -5.61316669e-01 -2.78050721e-01
-1.78253040e-01 -8.00779462e-01 8.44278276e-01 -7.29145017e-03
-7.32264578e-01 -1.49070144e+00 4.95268226e-01 1.44811094e-01
1.66951954e-01 -1.36892532e-03 1.33021939e+00 -1.20311451e+00
3.83867085e-01 2.56893761e-03 -2.15126380e-01 4.35503006e-01
6.64672256e-02 -1.57836944e-01 -6.87995255e-01 7.87906498e-02
2.23289251e-01 -5.87720096e-01 6.47619188e-01 2.19018057e-01
4.39348817e-01 1.31712466e-01 2.57363785e-02 2.33041734e-01
1.53829980e+00 1.51084527e-01 5.53077400e-01 3.96214277e-01
4.34652090e-01 8.25444281e-01 5.48469424e-01 3.05444747e-01
6.11179054e-01 5.39701045e-01 5.64744771e-02 3.46747369e-01
1.45872682e-02 -1.82545170e-01 3.11421931e-01 1.27092779e+00
-1.63071677e-02 -2.85382327e-02 -1.00370896e+00 9.46214736e-01
-1.51418507e+00 -8.41175795e-01 -1.91107616e-01 2.06963539e+00
9.60222840e-01 -1.23009183e-01 1.71217144e-01 4.22795981e-01
8.10713768e-01 5.89247867e-02 7.59911090e-02 -1.06881642e+00
-2.49259412e-01 6.13475144e-01 -1.87367518e-02 3.91786635e-01
-8.32823396e-01 1.44443333e+00 6.21785593e+00 7.90933073e-01
-5.37429273e-01 1.59631014e-01 -1.50220469e-01 3.38538915e-01
-5.46169937e-01 -1.84384175e-02 -6.35690153e-01 4.05011773e-01
1.14755821e+00 -3.39319050e-01 3.22050899e-01 5.96769571e-01
1.74116209e-01 -3.26767772e-01 -1.19421947e+00 8.48476946e-01
3.56445014e-01 -1.02528286e+00 1.91445574e-01 -2.47024998e-01
2.12897703e-01 -4.93592769e-02 -3.28988522e-01 3.37303072e-01
3.13366115e-01 -9.45522428e-01 8.36320460e-01 2.95009971e-01
5.10627747e-01 -9.47488248e-01 6.54823422e-01 1.37515917e-01
-1.12863016e+00 1.27718121e-01 -8.02422404e-01 -4.90365088e-01
4.43023778e-02 4.59103853e-01 -4.06668961e-01 2.69632995e-01
8.34559560e-01 7.84295082e-01 -6.82600141e-01 5.17903209e-01
-4.03862804e-01 2.94801712e-01 -1.57841280e-01 -6.24793112e-01
3.24581563e-01 -4.51809108e-01 4.88191247e-01 1.39376259e+00
-1.46278098e-01 8.50374624e-02 7.64629766e-02 1.23312652e+00
1.42346308e-01 9.79920566e-01 -7.55382836e-01 -1.39365062e-01
9.99115109e-01 1.17592943e+00 -1.01096535e+00 -7.34691992e-02
-4.30884242e-01 9.53236818e-01 6.49256587e-01 -2.84485042e-01
-6.84541821e-01 -6.88408434e-01 9.13502395e-01 -1.50752902e-01
-1.82410330e-01 -2.79532403e-01 -3.71334285e-01 -8.67124498e-01
-8.22637379e-02 -5.21302283e-01 2.67930001e-01 -5.96614301e-01
-1.80534136e+00 5.31534612e-01 1.82750002e-01 -5.20123541e-01
-2.47928068e-01 -9.92139876e-01 -2.63735294e-01 1.24144840e+00
-9.71196592e-01 -1.05579901e+00 -8.82990938e-03 7.26247251e-01
4.67793912e-01 -2.91067749e-01 1.29521048e+00 2.67694741e-01
-2.42140800e-01 3.72356147e-01 -1.83939651e-01 2.16537759e-01
6.82761371e-01 -1.35448360e+00 6.73499525e-01 3.07671607e-01
4.08120751e-01 8.46425056e-01 5.18215120e-01 -4.83012676e-01
-1.04628408e+00 -5.38056731e-01 1.79855537e+00 -7.40252137e-01
1.08531916e+00 -4.36655223e-01 -8.23862374e-01 5.33145368e-01
3.32261056e-01 -2.12042689e-01 8.92242968e-01 3.74132901e-01
-3.52721810e-01 3.25993985e-01 -1.27837038e+00 6.17010713e-01
1.14603221e+00 -9.44680572e-01 -1.58578753e+00 3.18616927e-01
8.55476201e-01 3.33641954e-02 -1.19960928e+00 -2.24821940e-02
2.81693071e-01 -8.37166190e-01 1.06557357e+00 -7.75943041e-01
5.63888490e-01 1.57313764e-01 -4.24486250e-01 -1.68175638e+00
-5.58343291e-01 -1.20413691e-01 7.97370076e-01 1.43364418e+00
3.28098208e-01 -6.65844083e-01 1.85527712e-01 4.45738941e-01
-2.12971315e-01 -2.90640771e-01 -8.88898551e-01 -7.73369133e-01
5.53228617e-01 -8.80081177e-01 6.40431285e-01 1.22928023e+00
4.52107131e-01 6.20926321e-01 5.48303187e-01 -2.75997639e-01
1.67570010e-01 -3.02160949e-01 1.48983538e-01 -1.53949964e+00
1.75516903e-01 -7.64143050e-01 -8.63955438e-01 -3.13158035e-01
7.32966661e-01 -1.32473731e+00 -1.18441932e-01 -1.20949125e+00
2.62805820e-01 -3.81580651e-01 1.37079461e-02 2.32584700e-01
-1.40633225e-01 2.51042098e-01 9.52166840e-02 -2.34069154e-01
-3.26310135e-02 6.01714551e-01 9.65252757e-01 2.94481218e-01
-6.63295239e-02 -6.28973305e-01 -6.08579338e-01 8.60472918e-01
6.59502625e-01 -4.08065796e-01 -3.21450949e-01 -3.71170312e-01
3.66765916e-01 -3.45648348e-01 3.39744657e-01 -7.23095000e-01
5.02431914e-02 -3.13324302e-01 1.11843571e-01 -2.04822212e-01
2.57982910e-01 -6.74803674e-01 -4.05406773e-01 2.27314651e-01
-3.81097913e-01 8.07968676e-01 2.90791094e-01 1.72849983e-01
-4.74885792e-01 -7.22473204e-01 9.74421799e-01 -3.12221766e-01
-1.05531871e+00 -3.16948071e-02 -5.61135173e-01 6.70526087e-01
1.10178471e+00 -4.10327077e-01 1.03452519e-01 2.47853950e-01
-6.73674941e-01 -2.23673373e-01 4.72005099e-01 8.19324374e-01
5.26198387e-01 -1.58943319e+00 -7.87295103e-01 2.33709365e-01
4.62231696e-01 -7.35683799e-01 -2.47093022e-01 3.84512037e-01
-5.61550021e-01 3.20911676e-01 -5.63899219e-01 -1.14198476e-01
-1.00840235e+00 8.73534501e-01 -1.08748935e-02 1.62316322e-01
-3.15385878e-01 8.20989668e-01 -3.04247469e-01 -9.37567174e-01
-1.87215582e-01 -4.17545676e-01 -7.13228643e-01 6.95188701e-01
3.29276294e-01 5.11649251e-01 2.46444866e-01 -1.20776379e+00
-4.24741894e-01 6.65042758e-01 1.62256852e-01 -4.33939487e-01
1.25789320e+00 -2.22096369e-01 -7.07441866e-01 8.78666937e-01
1.37562943e+00 -1.21712089e-01 -9.86961368e-03 -3.87088418e-01
4.58358675e-01 -4.13501501e-01 -2.87146103e-02 -4.79454279e-01
-5.01853347e-01 6.50473416e-01 3.70570838e-01 3.52001995e-01
6.99800253e-01 1.68476716e-01 7.19064295e-01 3.51132333e-01
3.03303957e-01 -1.48257864e+00 -1.27175033e-01 1.20967805e+00
7.36659586e-01 -1.11728740e+00 -3.55436832e-01 -3.18444788e-01
-6.53286815e-01 9.70117688e-01 3.66380423e-01 -2.85199046e-01
1.01092625e+00 -9.35339183e-03 -1.76749285e-02 -5.94505012e-01
-3.86119634e-01 -6.03142858e-01 4.12080824e-01 7.84844935e-01
9.35701907e-01 2.23167747e-01 -1.24804652e+00 8.09470415e-01
-8.99680138e-01 -5.16221642e-01 3.62136006e-01 6.58464372e-01
-5.43461323e-01 -1.05479109e+00 -3.26328516e-01 3.66898924e-01
-1.80976540e-01 -7.04402983e-01 -4.96286482e-01 1.05825269e+00
4.26859885e-01 1.00318205e+00 4.55596596e-01 -8.95404443e-02
5.38014114e-01 4.50293943e-02 6.17532670e-01 -1.02407575e+00
-9.29745853e-01 -6.42081738e-01 1.15074575e-01 -5.08136034e-01
-4.75121289e-01 -9.01235342e-01 -1.42701101e+00 -4.26092654e-01
-1.45025492e-01 2.37141237e-01 7.63304889e-01 1.22530103e+00
-2.50149965e-01 2.45215535e-01 2.02864096e-01 -9.80940640e-01
-4.46895123e-01 -1.24272895e+00 -1.03099859e+00 9.70499992e-01
-5.02913117e-01 -8.51642132e-01 -4.78066236e-01 -3.40718217e-02] | [10.445927619934082, 9.061966896057129] |
936b2441-bf1e-402c-b73f-b11a145cef5b | progen2-exploring-the-boundaries-of-protein | 2206.13517 | null | https://arxiv.org/abs/2206.13517v1 | https://arxiv.org/pdf/2206.13517v1.pdf | ProGen2: Exploring the Boundaries of Protein Language Models | Attention-based models trained on protein sequences have demonstrated incredible success at classification and generation tasks relevant for artificial intelligence-driven protein design. However, we lack a sufficient understanding of how very large-scale models and data play a role in effective protein model development. We introduce a suite of protein language models, named ProGen2, that are scaled up to 6.4B parameters and trained on different sequence datasets drawn from over a billion proteins from genomic, metagenomic, and immune repertoire databases. ProGen2 models show state-of-the-art performance in capturing the distribution of observed evolutionary sequences, generating novel viable sequences, and predicting protein fitness without additional finetuning. As large model sizes and raw numbers of protein sequences continue to become more widely accessible, our results suggest that a growing emphasis needs to be placed on the data distribution provided to a protein sequence model. We release the ProGen2 models and code at https://github.com/salesforce/progen. | ['Ali Madani', 'Nikhil Naik', 'Eli N. Weinstein', 'Jeffrey Ruffolo', 'Erik Nijkamp'] | 2022-06-27 | null | null | null | null | ['protein-design'] | ['medical'] | [ 4.44765985e-01 6.25010058e-02 -1.98386118e-01 -4.51822609e-01
-6.23299956e-01 -7.44532347e-01 2.46454984e-01 2.97379136e-01
-3.84179413e-01 1.00740278e+00 4.27001193e-02 -5.60408533e-01
1.68238476e-01 -4.14300382e-01 -1.00738275e+00 -7.52758443e-01
2.36880891e-02 9.62888658e-01 1.11127272e-01 -3.55497360e-01
1.19615003e-01 4.02717620e-01 -1.35969353e+00 4.78469461e-01
8.80232334e-01 3.91491413e-01 6.14250779e-01 9.17188823e-01
-1.87992305e-01 2.96867937e-01 -5.02439916e-01 -2.84566075e-01
4.59629372e-02 -5.66087425e-01 -7.23679245e-01 -2.79640734e-01
-7.94983357e-02 1.27734214e-01 -1.10460825e-01 7.12534130e-01
7.47656047e-01 2.57892683e-02 6.18470311e-01 -8.08711410e-01
-1.16289961e+00 4.23920423e-01 -3.91943365e-01 3.86612177e-01
1.51243344e-01 7.94537306e-01 1.00182855e+00 -7.27355599e-01
1.01415753e+00 1.07590652e+00 6.43135846e-01 8.49474788e-01
-1.69461429e+00 -4.01753545e-01 9.29398346e-04 1.69153869e-01
-9.56656456e-01 -2.76509494e-01 3.07352364e-01 -6.38089836e-01
1.87248111e+00 1.36112258e-01 6.00547552e-01 1.22950935e+00
3.87578189e-01 5.66652238e-01 6.22429788e-01 -2.80364037e-01
1.89284265e-01 -1.89071774e-01 2.28632122e-01 5.41323125e-01
3.83776017e-02 -1.65457111e-02 -6.13172293e-01 -8.49729955e-01
4.07872736e-01 2.05269717e-02 -2.65729755e-01 -3.45088989e-01
-1.02552998e+00 9.34076130e-01 1.23406075e-01 2.04696711e-02
-7.53190696e-01 5.22459038e-02 3.46801877e-01 1.53895393e-01
4.55457330e-01 8.63505721e-01 -1.02629268e+00 -3.48691583e-01
-5.14376760e-01 7.23319411e-01 5.47344983e-01 8.17676127e-01
5.22020698e-01 -1.94462150e-01 1.69989556e-01 1.01016164e+00
-9.62447301e-02 4.07087654e-01 6.76249862e-01 -6.78699315e-01
-2.68770643e-02 4.78723377e-01 3.37757021e-01 -4.68321830e-01
-4.38403845e-01 -2.05835879e-01 -1.88927591e-01 -1.37689024e-01
3.15707624e-01 3.78839299e-02 -9.72419381e-01 1.96526277e+00
2.71367460e-01 -1.95755079e-01 1.36823699e-01 6.40553772e-01
3.57814431e-01 8.27252150e-01 5.18703043e-01 -3.14529389e-02
1.36089993e+00 -7.88436055e-01 -2.56696939e-01 -3.57810020e-01
5.56583226e-01 -6.43534541e-01 1.05027807e+00 3.72627527e-01
-1.12072563e+00 -5.01968324e-01 -1.08532989e+00 -1.59829974e-01
-4.99959916e-01 -3.57084513e-01 7.70319760e-01 3.03260773e-01
-7.96461821e-01 7.44711339e-01 -9.66538668e-01 -6.49534643e-01
3.92917454e-01 4.85184461e-01 -3.46934319e-01 -7.44914562e-02
-1.05912089e+00 1.17277718e+00 6.66066468e-01 -2.04833016e-01
-8.40986311e-01 -8.00754428e-01 -4.73846287e-01 1.07464142e-01
-9.12866741e-03 -8.30563366e-01 1.19549322e+00 -8.44384313e-01
-1.03388155e+00 9.08649027e-01 -3.18377346e-01 -7.41176903e-01
-1.01643018e-01 -5.42838052e-02 -1.36899948e-01 -9.49674025e-02
-2.88114280e-01 1.02371860e+00 1.57060862e-01 -7.93963969e-01
-4.09071892e-01 -4.12396252e-01 -2.79184639e-01 -9.19210818e-03
2.12336220e-02 2.82506377e-01 -8.99656955e-03 -5.62854469e-01
-4.12806660e-01 -9.97635245e-01 -5.52931726e-01 -1.35106325e-01
2.99834739e-02 -2.13758007e-01 2.12084562e-01 -7.80503154e-01
9.78786647e-01 -1.67535305e+00 5.44206142e-01 -1.24961399e-01
1.90525904e-01 6.74833715e-01 -6.66062236e-01 6.33243799e-01
-2.27594629e-01 1.89424768e-01 -3.55170548e-01 3.53863001e-01
-1.03658952e-01 1.24570979e-02 -3.22537720e-02 1.45234406e-01
7.16636479e-01 1.13347542e+00 -8.37060928e-01 9.26121771e-02
4.51596901e-02 6.50108695e-01 -8.35337460e-01 3.47607315e-01
-9.04422045e-01 4.09866840e-01 -5.79631209e-01 5.88903487e-01
5.25174737e-01 -7.29353845e-01 6.50294602e-01 1.55745000e-01
2.57506937e-01 3.17193896e-01 -1.09899014e-01 1.54720855e+00
2.30988823e-02 3.12979847e-01 -1.46972030e-01 -8.23367894e-01
1.02157462e+00 2.45014094e-02 6.15887821e-01 -4.08045918e-01
-3.87876551e-03 1.07702442e-01 6.42912805e-01 -4.25098240e-01
2.79070050e-01 -2.62127608e-01 1.53808713e-01 4.39388067e-01
1.94389075e-01 2.12864950e-01 4.46743965e-01 -1.59612503e-02
1.29762125e+00 5.19470811e-01 4.08212125e-01 -2.22678259e-01
3.36206704e-01 6.55087173e-01 8.20246458e-01 5.38184881e-01
-1.23967983e-01 3.64684671e-01 4.50660735e-01 -7.75663435e-01
-1.75671828e+00 -6.89087451e-01 -1.03147499e-01 1.49813735e+00
-4.30898219e-01 -5.89812934e-01 -8.77190828e-01 -3.16230059e-01
3.26392621e-01 6.23898566e-01 -5.37259400e-01 -4.51412499e-01
-6.14610255e-01 -1.47108448e+00 6.12367988e-01 3.68872941e-01
-4.16278362e-01 -1.43183827e+00 -5.44734299e-01 6.94687128e-01
-1.56931385e-01 -6.03105187e-01 -5.55540740e-01 5.64304531e-01
-7.07034230e-01 -1.13550520e+00 -7.98840582e-01 -7.42919087e-01
5.51737010e-01 -1.27027286e-02 1.34152269e+00 5.55655546e-02
-8.13525677e-01 -2.11882174e-01 -1.74751192e-01 -6.50103033e-01
-8.49307835e-01 2.20948666e-01 5.74528985e-02 -5.92233002e-01
1.06811488e+00 -5.56076884e-01 -5.76233745e-01 3.00936341e-01
-8.87243330e-01 2.90483505e-01 5.59666038e-01 1.25827789e+00
7.36463845e-01 -5.77810824e-01 9.33486402e-01 -1.01166010e+00
8.05822968e-01 -7.14251995e-01 -5.79657197e-01 5.61379075e-01
-7.80984700e-01 4.77159679e-01 6.81672871e-01 -6.04089737e-01
-9.20595229e-01 6.78787082e-02 -4.93167669e-01 -9.89589021e-02
-1.64392382e-01 5.20802259e-01 1.00110523e-01 1.67573988e-01
8.78797829e-01 4.52705264e-01 4.61213559e-01 -6.18416309e-01
4.20389891e-01 7.18866229e-01 2.33285636e-01 -7.37138331e-01
8.23415667e-02 -7.43536577e-02 -3.34620833e-01 -5.76382637e-01
-3.65458637e-01 -3.37252438e-01 -4.03345883e-01 3.77819955e-01
6.89011276e-01 -6.64427340e-01 -9.93487298e-01 3.78814608e-01
-9.19374406e-01 -6.10958099e-01 3.56026255e-02 4.04572114e-02
-7.96402216e-01 3.63093555e-01 -9.51120019e-01 -5.84968925e-01
-7.16614068e-01 -1.36513042e+00 9.20967519e-01 1.01964436e-01
-5.16215622e-01 -7.65028894e-01 3.10366303e-01 5.91065526e-01
5.27847052e-01 3.63787711e-02 1.54679227e+00 -8.97347450e-01
-4.07206148e-01 6.25292212e-02 2.42904928e-02 7.75291100e-02
4.63697724e-02 4.68835868e-02 -5.50173044e-01 -2.68693954e-01
-3.16354066e-01 -6.64772928e-01 7.68051684e-01 4.25264120e-01
1.00367856e+00 -9.80151519e-02 -4.99088645e-01 5.63966572e-01
1.18563020e+00 4.44961637e-01 3.38298619e-01 2.87565231e-01
3.04989368e-01 7.23339736e-01 6.29768610e-01 2.13392064e-01
5.16036227e-02 7.13362992e-01 8.30069408e-02 2.93939877e-02
1.15621597e-01 -4.19339299e-01 2.27954194e-01 4.94458407e-01
9.68725756e-02 -5.84340274e-01 -1.14640164e+00 4.03211951e-01
-1.97013175e+00 -1.01733303e+00 1.33047342e-01 2.12008929e+00
1.17755854e+00 6.71185320e-03 2.25794628e-01 -8.01573336e-01
6.83518052e-01 -2.43237466e-01 -1.45598924e+00 -4.79973912e-01
-4.13346916e-01 2.43821442e-01 4.61639851e-01 4.26668733e-01
-7.10455418e-01 9.75772560e-01 7.34176397e+00 8.15867245e-01
-9.96815979e-01 -1.80236563e-01 1.09280026e+00 -2.91316062e-01
-3.87706369e-01 -1.18604720e-01 -9.60735023e-01 6.79495513e-01
1.50660396e+00 -4.52116847e-01 5.35511851e-01 8.76793385e-01
2.39916146e-01 2.80735135e-01 -9.52258408e-01 5.81162512e-01
-1.53933182e-01 -1.73507392e+00 4.59492914e-02 2.75039077e-01
4.96408522e-01 6.19625747e-01 -9.90355015e-02 2.77130306e-01
6.08678341e-01 -1.20617807e+00 1.68546915e-01 4.30452764e-01
6.50271297e-01 -7.67342627e-01 6.44318104e-01 3.98258924e-01
-5.19462764e-01 1.29933953e-01 -7.90196836e-01 2.27192700e-01
3.65934610e-01 2.77364582e-01 -1.07814133e+00 -2.94473656e-02
4.01223183e-01 4.71875399e-01 -4.93303508e-01 5.66867650e-01
2.56273031e-01 5.52635014e-01 -1.97558030e-01 -2.60428876e-01
5.93332984e-02 -1.95589781e-01 2.20294207e-01 1.13733089e+00
4.43356559e-02 1.24743268e-01 2.84297056e-02 1.02209878e+00
-1.37951458e-02 5.17911762e-02 -2.63325185e-01 -7.25778818e-01
4.27936047e-01 1.01426458e+00 -3.89361113e-01 -3.57705265e-01
-3.07101756e-01 8.64926696e-01 7.08166838e-01 2.73784727e-01
-9.26727116e-01 -1.57482550e-01 1.25209606e+00 8.06381628e-02
2.33775631e-01 -1.01905465e-02 2.15662643e-01 -9.11974549e-01
-3.97938579e-01 -1.51712024e+00 3.46373588e-01 -1.02702963e+00
-1.56361628e+00 5.15107512e-01 -2.87774473e-01 -2.79317915e-01
-4.57477927e-01 -1.13866520e+00 -4.79041412e-03 1.22992873e+00
-1.01170123e+00 -8.75132918e-01 2.90506840e-01 -1.69180498e-01
8.03093314e-01 -1.99573651e-01 1.07456446e+00 7.77733102e-02
-5.96149623e-01 3.57899457e-01 5.43304145e-01 -4.68667328e-01
7.84611344e-01 -9.42074120e-01 1.38573039e+00 3.23250771e-01
-2.61068761e-01 1.09183502e+00 9.77238357e-01 -8.49582970e-01
-1.44983006e+00 -8.62208247e-01 9.27117348e-01 -8.55237067e-01
4.97405738e-01 -5.67683578e-01 -1.25854218e+00 7.38105416e-01
3.00554074e-02 -5.56400359e-01 1.08158731e+00 1.03050895e-01
-2.95663625e-01 4.42446053e-01 -1.21589625e+00 5.94435990e-01
1.23086333e+00 -3.29805434e-01 -3.41126353e-01 5.30020237e-01
1.01136112e+00 -1.54457539e-01 -9.64470804e-01 4.28510606e-01
6.44813061e-01 -6.97917700e-01 1.18539143e+00 -1.45857871e+00
4.33244884e-01 -1.06348753e-01 4.71559018e-02 -1.18057108e+00
-8.92069697e-01 -5.02313137e-01 -4.94533367e-02 7.47975767e-01
9.13363159e-01 -7.97573090e-01 1.00584316e+00 5.89619339e-01
-8.89329761e-02 -1.04919112e+00 -3.46821189e-01 -4.79161114e-01
3.40684205e-01 1.51082069e-01 7.90067613e-01 7.24733174e-01
2.36141771e-01 4.98916864e-01 -3.69164497e-01 -3.85119528e-01
2.24390969e-01 1.03306003e-01 6.21113122e-01 -9.99654651e-01
-7.83952534e-01 -5.03207505e-01 -2.19389677e-01 -9.96739924e-01
4.67841476e-02 -8.75350475e-01 -5.55765210e-03 -1.19236612e+00
6.65834188e-01 -1.64998487e-01 -5.34087062e-01 4.52393711e-01
-5.85682511e-01 -3.76298018e-02 -1.15791738e-01 1.64204344e-01
-2.91617095e-01 4.86528933e-01 8.35029244e-01 -5.82289547e-02
9.01097357e-02 -3.50442320e-01 -9.54348445e-01 3.53060871e-01
1.19111955e+00 -3.65754873e-01 -3.51011366e-01 -3.13595325e-01
1.81876972e-01 -1.65264055e-01 -1.46740168e-01 -5.47704697e-01
-1.91887662e-01 -6.24398232e-01 4.34551209e-01 -3.79020721e-01
5.26768208e-01 -1.77586272e-01 6.77056968e-01 7.53541887e-01
-6.82133853e-01 4.04284149e-01 3.42782021e-01 5.78819513e-01
2.19933599e-01 -2.86804307e-02 8.01681280e-01 -4.93573517e-01
-4.57116187e-01 3.20753455e-01 -5.05508721e-01 -6.09502941e-02
8.63519907e-01 -1.02320097e-01 -5.10624468e-01 -3.81859429e-02
-7.61369407e-01 1.02580950e-01 9.66794074e-01 6.57422900e-01
3.20869178e-01 -8.13263416e-01 -7.67530024e-01 2.43364066e-01
5.03691554e-01 -6.15130663e-01 2.48891920e-01 3.19812983e-01
-8.08470786e-01 9.28970158e-01 -3.93176258e-01 -3.36633950e-01
-1.41838062e+00 8.80147815e-01 3.29725266e-01 -2.14769423e-01
-3.11884493e-01 9.97729719e-01 3.11672568e-01 -7.99522340e-01
-3.20078015e-01 1.09827220e-01 2.31062576e-01 -4.69617009e-01
6.00904047e-01 8.75767022e-02 -9.35743898e-02 -5.13041198e-01
-2.75186747e-01 1.65081859e-01 -3.88663292e-01 2.93698847e-01
1.53018260e+00 2.07831010e-01 -1.17562391e-01 1.19299516e-01
8.29531133e-01 -5.34147263e-01 -1.40211356e+00 -6.78090081e-02
1.81484491e-01 -1.83205977e-01 -5.98943353e-01 -1.10754395e+00
-3.04238290e-01 7.72197723e-01 5.65343618e-01 -3.28553617e-01
7.65113056e-01 5.87171242e-02 9.80744779e-01 5.16517699e-01
5.73865235e-01 -8.41695130e-01 -1.20237172e-01 5.57328045e-01
5.42433798e-01 -1.11780083e+00 -1.80008709e-01 -5.51094525e-02
-5.83592415e-01 7.37148166e-01 6.72609150e-01 2.48118773e-01
7.90376663e-02 2.22050741e-01 2.96302997e-02 -1.56525344e-01
-1.51874447e+00 6.67254999e-02 7.37759471e-03 9.62702632e-01
9.30662870e-01 1.39884919e-01 -4.72716004e-01 5.77925563e-01
1.26241464e-02 3.00297793e-02 1.68682083e-01 8.90674412e-01
-8.37525189e-01 -1.63971555e+00 -6.83681965e-02 4.94440824e-01
-6.84973598e-01 -4.30095613e-01 -7.37085283e-01 2.02280343e-01
-9.67068002e-02 7.07219422e-01 -1.95903823e-01 -6.05936907e-02
7.49559924e-02 7.19796836e-01 5.57523727e-01 -7.26583481e-01
-6.07781529e-01 -7.11703449e-02 2.62864083e-01 -2.52156019e-01
1.00371249e-01 -5.31409323e-01 -1.28187084e+00 -4.51874375e-01
-3.06834519e-01 3.50458026e-01 5.90422094e-01 3.15799773e-01
1.23319077e+00 3.98991078e-01 -1.05458826e-01 -6.11257970e-01
-7.68289745e-01 -1.13461339e+00 -1.29380092e-01 4.60224509e-01
-4.82500270e-02 -3.29670459e-01 1.55728966e-01 3.40409815e-01] | [4.684179782867432, 5.609221458435059] |
2b16cab3-2bdf-4835-8d18-1b158b0d9f01 | quantization-of-generative-adversarial | 2108.13996 | null | https://arxiv.org/abs/2108.13996v1 | https://arxiv.org/pdf/2108.13996v1.pdf | Quantization of Generative Adversarial Networks for Efficient Inference: a Methodological Study | Generative adversarial networks (GANs) have an enormous potential impact on digital content creation, e.g., photo-realistic digital avatars, semantic content editing, and quality enhancement of speech and images. However, the performance of modern GANs comes together with massive amounts of computations performed during the inference and high energy consumption. That complicates, or even makes impossible, their deployment on edge devices. The problem can be reduced with quantization -- a neural network compression technique that facilitates hardware-friendly inference by replacing floating-point computations with low-bit integer ones. While quantization is well established for discriminative models, the performance of modern quantization techniques in application to GANs remains unclear. GANs generate content of a more complex structure than discriminative models, and thus quantization of GANs is significantly more challenging. To tackle this problem, we perform an extensive experimental study of state-of-art quantization techniques on three diverse GAN architectures, namely StyleGAN, Self-Attention GAN, and CycleGAN. As a result, we discovered practical recipes that allowed us to successfully quantize these models for inference with 4/8-bit weights and 8-bit activations while preserving the quality of the original full-precision models. | ['Dmitry Vetrov', 'Alexander Fritzler', 'Pavel Andreev'] | 2021-08-31 | null | null | null | null | ['neural-network-compression', 'neural-network-compression'] | ['methodology', 'miscellaneous'] | [ 5.12550712e-01 6.29647449e-02 -7.10928142e-02 -1.38555586e-01
-7.81925142e-01 -5.10487080e-01 5.68891168e-01 -3.02942067e-01
-1.82335436e-01 8.11079025e-01 1.17073707e-01 -4.15198833e-01
3.60800475e-01 -1.16834712e+00 -8.46270204e-01 -8.21140528e-01
2.58890986e-01 3.62406611e-01 -8.73951763e-02 -3.72964203e-01
-1.07472397e-01 4.31355655e-01 -1.49071431e+00 2.75121838e-01
7.88026571e-01 1.23008668e+00 1.21757418e-01 8.67705822e-01
-2.37939179e-01 8.45495582e-01 -8.83590519e-01 -1.05696452e+00
3.96348625e-01 -7.30734468e-01 -3.11896503e-01 -2.95857042e-01
5.17435670e-01 -6.69018090e-01 -7.21795201e-01 1.26698112e+00
5.66883266e-01 -2.09138229e-01 5.13483822e-01 -1.37980044e+00
-8.77628446e-01 5.97199619e-01 -2.23724201e-01 -2.46808261e-01
-8.02347660e-02 4.51477259e-01 8.47791076e-01 -3.88318211e-01
4.64198470e-01 1.24708128e+00 6.53721809e-01 8.47072780e-01
-1.00416160e+00 -7.01762795e-01 -4.52461064e-01 3.06180686e-01
-1.38457334e+00 -6.51510715e-01 7.71507740e-01 2.72904802e-02
7.90787399e-01 3.97999257e-01 9.68922377e-01 1.34977639e+00
1.78893149e-01 6.13340259e-01 8.46946001e-01 -2.47849867e-01
5.60007155e-01 -5.34581803e-02 -7.72099435e-01 6.45281196e-01
1.98895842e-01 -3.76473740e-03 -4.88248199e-01 3.99732329e-02
9.87899125e-01 1.95402466e-02 -2.08618045e-01 1.35443315e-01
-9.34395254e-01 8.88111293e-01 5.14477432e-01 1.41223282e-01
-3.87568712e-01 9.14732635e-01 4.77667272e-01 1.99495330e-01
2.25957230e-01 2.66801775e-01 -3.58905904e-02 -5.86160243e-01
-1.14196157e+00 1.97680831e-01 5.17015815e-01 1.12269127e+00
7.08328009e-01 6.64479971e-01 -2.51811177e-01 5.03274381e-01
-6.22250475e-02 7.25949109e-01 5.40129602e-01 -1.00948179e+00
3.47503006e-01 3.52967441e-01 -3.30194175e-01 -9.07583117e-01
1.04036674e-01 -3.22562784e-01 -1.62206435e+00 2.85382390e-01
3.03116411e-01 -6.01444393e-02 -9.81784523e-01 1.65329301e+00
-1.00677395e-02 5.74608743e-02 -2.20774878e-02 8.59721303e-01
5.19610226e-01 8.98614347e-01 1.90271586e-02 -1.23092323e-03
1.27645075e+00 -1.00552893e+00 -8.66539896e-01 -2.39685401e-01
2.85545558e-01 -5.23848355e-01 1.08734679e+00 4.62745935e-01
-1.33832908e+00 -6.42996788e-01 -1.29961216e+00 -4.47032124e-01
-2.74781436e-01 -2.96283104e-02 8.91157746e-01 1.00394630e+00
-1.30054522e+00 6.97708189e-01 -6.96900129e-01 2.82644540e-01
6.05058193e-01 2.27114320e-01 -2.22994879e-01 -1.26564264e-01
-1.20199919e+00 6.01271391e-01 2.37644121e-01 3.35478373e-02
-8.22442055e-01 -7.73135543e-01 -7.14577854e-01 3.14699322e-01
1.63191885e-01 -1.04947114e+00 1.18817663e+00 -1.29683912e+00
-1.95661390e+00 5.11280537e-01 -1.15714774e-01 -8.54598045e-01
5.72406590e-01 2.59366948e-02 -4.91298795e-01 1.61725312e-01
-3.57655287e-01 9.02372420e-01 1.38037968e+00 -7.49414027e-01
-3.28845769e-01 -2.30691016e-01 7.11453408e-02 -8.90183821e-02
-4.36765760e-01 -4.76392627e-01 -5.05858064e-01 -1.01837754e+00
-3.43123734e-01 -8.74058425e-01 -2.44431317e-01 3.28050256e-01
-2.72158086e-01 1.61772102e-01 8.24544907e-01 -7.86590040e-01
1.04922235e+00 -2.23734117e+00 1.52778387e-01 -7.64815286e-02
3.34817111e-01 5.58873594e-01 2.84839887e-02 1.97217748e-01
2.87864089e-01 1.82519615e-01 -2.73072094e-01 -5.26217222e-01
2.00913712e-01 3.67949069e-01 -4.50657457e-01 1.60662666e-01
2.22932249e-01 1.29858148e+00 -7.63990998e-01 -3.65045458e-01
2.14035466e-01 8.73360336e-01 -8.05466235e-01 2.62314081e-01
-4.27188516e-01 3.04155350e-01 -3.05038951e-02 7.15615273e-01
6.23938620e-01 -1.60283029e-01 1.84579074e-01 -4.73202795e-01
3.18992794e-01 3.12129617e-01 -7.19874263e-01 1.93149948e+00
-7.09652483e-01 8.56235743e-01 7.55216032e-02 -7.53904879e-01
7.73815870e-01 1.38008028e-01 -9.20689013e-03 -1.05980575e+00
1.45928636e-01 1.80776745e-01 -2.25947678e-01 -2.41061859e-02
8.89117539e-01 -1.91729069e-01 -1.19850270e-01 9.15684924e-02
3.60040925e-02 -5.46377242e-01 -3.01604029e-02 2.14008614e-01
1.00561023e+00 -2.37702131e-02 8.75841081e-02 5.69394752e-02
1.22803710e-01 -3.35694104e-01 3.85224551e-01 5.12089372e-01
1.60696357e-01 7.58225024e-01 5.06741583e-01 -3.61677587e-01
-1.50405788e+00 -9.90770340e-01 1.90824613e-01 7.14716434e-01
-3.59035693e-02 -7.20229149e-01 -1.11338854e+00 -3.12871695e-01
-3.45879585e-01 7.91292608e-01 -3.48255306e-01 -4.60066587e-01
-5.65488696e-01 -5.56799650e-01 9.44355845e-01 6.32547975e-01
8.43965888e-01 -7.92556167e-01 -7.26628959e-01 1.43993139e-01
-1.11708920e-02 -1.25128496e+00 -5.76421261e-01 4.40169163e-02
-8.30413103e-01 -4.37243253e-01 -8.41791034e-01 -4.23632175e-01
4.79843110e-01 -8.74626040e-02 1.35625577e+00 5.99359581e-03
-2.04960719e-01 -5.16984649e-02 -2.37563834e-01 -3.39178741e-01
-8.37486625e-01 -2.61595962e-03 -1.59622222e-01 -7.89579451e-02
-1.67345002e-01 -9.38207448e-01 -7.24911451e-01 -6.56812117e-02
-1.17074013e+00 3.67273748e-01 8.69656205e-01 1.04993761e+00
6.72709465e-01 1.63022637e-01 4.60066706e-01 -7.71072507e-01
3.65115553e-01 -9.26078334e-02 -7.49853969e-01 1.28261060e-01
-4.90828961e-01 2.60839641e-01 1.17404282e+00 -3.83370548e-01
-7.26264298e-01 -1.63821667e-01 -6.75596356e-01 -6.92074001e-01
2.05213934e-01 1.53350690e-02 -4.81271774e-01 -1.36281401e-01
5.22732615e-01 3.91320229e-01 -3.43463570e-02 -2.00876057e-01
7.34064460e-01 6.51531756e-01 9.91324365e-01 -3.34151357e-01
8.00187588e-01 3.74279767e-01 2.52669662e-01 -8.67003202e-01
-2.47950643e-01 4.11070645e-01 -4.98864986e-03 2.65210960e-02
8.27792287e-01 -1.04002869e+00 -6.43582463e-01 5.58000684e-01
-1.19357169e+00 -5.47635972e-01 -6.49015486e-01 -2.44845878e-02
-4.57457244e-01 3.42030853e-01 -5.99341214e-01 -6.64160609e-01
-7.08060801e-01 -1.26433218e+00 1.08245730e+00 1.51335627e-01
-5.58652095e-02 -8.08736563e-01 -3.48146826e-01 2.60827512e-01
8.77288997e-01 2.43067294e-01 9.45495248e-01 1.63228616e-01
-8.92903566e-01 -1.27468467e-01 -2.45031312e-01 6.16372645e-01
-3.02522955e-03 -1.81507409e-01 -1.14660740e+00 -2.14119971e-01
-5.20436792e-03 -2.52238393e-01 7.59196401e-01 1.05964243e-01
1.65893662e+00 -6.31433785e-01 1.70172200e-01 1.15224111e+00
1.31276834e+00 1.52751297e-01 1.27657950e+00 -2.53726155e-01
8.31355572e-01 -1.97583258e-01 9.25710704e-03 4.35729265e-01
1.83279872e-01 8.58166635e-01 7.41376817e-01 -4.70158271e-02
-5.57898521e-01 -6.17963612e-01 3.24125648e-01 9.08947766e-01
-1.67235062e-01 -4.83136386e-01 -3.79661024e-01 3.17894459e-01
-1.29762280e+00 -8.36934090e-01 2.14905307e-01 2.10864758e+00
9.75446284e-01 2.07402095e-01 -1.05956793e-01 4.08465803e-01
5.02526283e-01 2.10727587e-01 -6.92991734e-01 -5.41048586e-01
-2.73560941e-01 7.41226137e-01 7.92588472e-01 1.19800806e-01
-7.12271810e-01 7.62309790e-01 6.22245121e+00 1.42046678e+00
-1.31130648e+00 3.82930189e-01 9.77942467e-01 -1.52837992e-01
-6.58712566e-01 -2.71575361e-01 -6.45175040e-01 8.09446275e-01
1.34038544e+00 -2.86576003e-02 9.19546664e-01 1.12310135e+00
-2.55025119e-01 1.64230108e-01 -9.99299109e-01 1.49835980e+00
5.78021223e-04 -1.76578629e+00 4.21034962e-01 1.40207976e-01
9.03627515e-01 -2.79034257e-01 3.38310510e-01 2.93157429e-01
1.45876274e-01 -1.33602321e+00 8.52244973e-01 2.14820549e-01
1.60979283e+00 -8.30756247e-01 5.70642292e-01 1.24617077e-01
-9.56827283e-01 1.32850036e-01 -5.46669185e-01 -1.40477379e-03
2.79719412e-01 8.01079869e-01 -6.55550838e-01 3.18834484e-01
4.97111082e-01 2.89972842e-01 -3.50754082e-01 4.83133674e-01
-2.90283114e-01 5.83998680e-01 -2.43661851e-01 -7.75134265e-02
2.90992886e-01 -2.69049287e-01 2.37129509e-01 1.07940972e+00
6.41630590e-01 -4.63231988e-02 -5.54603636e-01 1.00252604e+00
-6.40943646e-01 -3.28639448e-01 -5.64896405e-01 -2.95437992e-01
4.32133704e-01 1.15922534e+00 -5.08971274e-01 -4.74808931e-01
-1.35131985e-01 1.42073619e+00 9.58550200e-02 2.18916573e-02
-1.23002577e+00 -4.26040202e-01 8.74418736e-01 1.74414203e-01
3.16509515e-01 -3.17689538e-01 -4.60624546e-01 -1.12970066e+00
9.01467279e-02 -1.09903049e+00 -1.83830515e-01 -7.93986320e-01
-8.99970055e-01 6.41881704e-01 -5.03027141e-01 -9.60670531e-01
-4.99987632e-01 -4.19668913e-01 -4.22948569e-01 8.14310908e-01
-1.40372753e+00 -1.11285889e+00 -5.21366596e-01 5.71960986e-01
5.36428511e-01 -1.55398771e-01 9.34790254e-01 6.78573489e-01
-2.90038168e-01 1.10506701e+00 2.57439882e-01 3.03693768e-02
3.19147825e-01 -1.02633107e+00 9.75870907e-01 8.58341455e-01
3.05274874e-01 1.50358528e-01 4.89452422e-01 -3.39095712e-01
-2.17922401e+00 -1.33338702e+00 3.34811270e-01 4.10080180e-02
3.76625180e-01 -4.58696902e-01 -6.57634139e-01 4.73287523e-01
2.79551834e-01 2.12833565e-02 3.20641577e-01 -6.84980750e-01
-3.63974124e-01 -2.23785520e-01 -1.32703578e+00 8.56037438e-01
1.16824138e+00 -7.13728845e-01 9.46792588e-02 2.52877951e-01
9.28227305e-01 -5.83841681e-01 -7.45123565e-01 1.17162928e-01
5.46426237e-01 -1.00960100e+00 1.09773016e+00 -1.22179255e-01
8.01870167e-01 -1.57441482e-01 -2.95571268e-01 -1.29716754e+00
-7.39920139e-02 -9.87572968e-01 -4.66581762e-01 1.08122289e+00
8.70430842e-02 -4.33917880e-01 9.06758606e-01 4.75216001e-01
-2.34073102e-01 -6.52356684e-01 -1.15489936e+00 -7.00329304e-01
-6.61304072e-02 -4.54847872e-01 9.74219918e-01 5.67462027e-01
-4.24589396e-01 1.86026335e-01 -7.47934341e-01 -1.59864560e-01
6.15152478e-01 -1.80878088e-01 9.56574738e-01 -5.89145780e-01
-6.49548233e-01 -5.80936253e-01 -8.46819460e-01 -1.33617067e+00
-1.67479515e-02 -6.76949024e-01 -2.10774288e-01 -1.14635253e+00
-3.13241422e-01 -4.35235888e-01 3.42292100e-01 2.68993318e-01
6.82530105e-02 7.48507082e-01 2.98820406e-01 -6.81691617e-03
-2.66219646e-01 7.90081501e-01 1.25932896e+00 -4.64419633e-01
2.12023959e-01 -2.95773029e-01 -6.14227295e-01 4.92875010e-01
7.86656082e-01 -1.87242031e-01 -4.83542770e-01 -6.54076993e-01
5.19641578e-01 1.70047924e-01 5.78335345e-01 -1.25922740e+00
2.12424517e-01 1.61581963e-01 3.56975794e-01 -3.10267240e-01
7.72754252e-01 -7.96181381e-01 5.59540331e-01 5.04751503e-01
5.12691997e-02 7.07136169e-02 1.64660007e-01 4.16071475e-01
-2.61572927e-01 -2.04419881e-01 8.68428886e-01 -1.31276160e-01
-6.40591741e-01 2.79850334e-01 -8.97544622e-02 1.27465934e-01
7.27608025e-01 -9.89244729e-02 -3.23415667e-01 -7.32222855e-01
-2.88825035e-01 -4.33368325e-01 6.65042877e-01 1.78849682e-01
6.78215444e-01 -1.54650414e+00 -5.82884729e-01 4.78573084e-01
-4.20817703e-01 3.26797932e-01 4.04642224e-01 4.04248744e-01
-9.73151743e-01 2.73373902e-01 -4.30485040e-01 -4.05576527e-01
-8.52476537e-01 5.64818323e-01 1.04198165e-01 -1.84291676e-01
-6.08723879e-01 7.46723831e-01 4.20117788e-02 3.46959889e-01
1.82069197e-01 -4.39842999e-01 5.21255195e-01 -2.96111435e-01
6.76637232e-01 4.40645844e-01 1.69778332e-01 -2.76116312e-01
5.99113591e-02 3.44480872e-01 3.04426670e-01 7.91755319e-02
1.08770108e+00 2.27186874e-01 -8.97725672e-02 7.95780495e-02
1.27046001e+00 -5.70999682e-02 -1.37459373e+00 9.57261547e-02
-7.91017950e-01 -4.37882096e-01 1.55682907e-01 -4.99612451e-01
-1.56820762e+00 1.12474418e+00 5.12171388e-01 2.58887142e-01
1.40085185e+00 -3.31501275e-01 1.39160776e+00 2.94953138e-01
5.62014639e-01 -8.81426930e-01 -4.22864147e-02 1.77512124e-01
7.66107261e-01 -9.10794914e-01 -1.53199568e-01 -2.33522043e-01
-4.38729584e-01 1.01464653e+00 2.86414832e-01 1.20957727e-02
2.98883557e-01 6.09753728e-01 -3.79997104e-01 2.59010673e-01
-5.30752003e-01 2.61095434e-01 4.49519232e-02 6.10946476e-01
6.32243454e-02 2.66291022e-01 3.78422104e-02 5.35281062e-01
-6.34592593e-01 1.04263119e-01 4.11715567e-01 5.82646489e-01
6.01009093e-02 -9.76817012e-01 -2.68458128e-01 3.63360852e-01
-4.43993360e-01 -4.12732095e-01 -4.15268028e-03 4.23396856e-01
1.57205313e-01 5.91597080e-01 2.30798930e-01 -6.14736080e-01
-4.18547019e-02 1.28228348e-02 5.03301322e-01 7.81298801e-02
-4.78217512e-01 -1.81555226e-01 2.24930961e-02 -5.84576428e-01
-6.82481304e-02 -1.45525098e-01 -8.85259807e-01 -9.61201668e-01
-3.85629199e-02 -1.31389663e-01 1.07867873e+00 5.78103244e-01
6.29230320e-01 7.62105584e-01 4.64023143e-01 -1.00548732e+00
-6.99963748e-01 -8.43210459e-01 -4.46964949e-01 3.15982223e-01
5.15811853e-02 -1.78265557e-01 -2.17604592e-01 1.92632467e-01] | [11.58951473236084, -0.48503121733665466] |
45293def-e5fc-459f-aa36-a904e211f43c | context-aware-embeddings-for-automatic-art | 1904.04985 | null | http://arxiv.org/abs/1904.04985v1 | http://arxiv.org/pdf/1904.04985v1.pdf | Context-Aware Embeddings for Automatic Art Analysis | Automatic art analysis aims to classify and retrieve artistic representations
from a collection of images by using computer vision and machine learning
techniques. In this work, we propose to enhance visual representations from
neural networks with contextual artistic information. Whereas visual
representations are able to capture information about the content and the style
of an artwork, our proposed context-aware embeddings additionally encode
relationships between different artistic attributes, such as author, school, or
historical period. We design two different approaches for using context in
automatic art analysis. In the first one, contextual data is obtained through a
multi-task learning model, in which several attributes are trained together to
find visual relationships between elements. In the second approach, context is
obtained through an art-specific knowledge graph, which encodes relationships
between artistic attributes. An exhaustive evaluation of both of our models in
several art analysis problems, such as author identification, type
classification, or cross-modal retrieval, show that performance is improved by
up to 7.3% in art classification and 37.24% in retrieval when context-aware
embeddings are used. | ['Yuta Nakashima', 'Noa Garcia', 'Benjamin Renoust'] | 2019-04-10 | null | null | null | null | ['art-analysis'] | ['computer-vision'] | [ 2.85850257e-01 -4.97829169e-01 -1.33175433e-01 -2.45135412e-01
-4.65243787e-01 -7.98169434e-01 1.01881325e+00 7.47161746e-01
-3.65485907e-01 4.00114447e-01 4.60288823e-01 1.71279460e-01
-3.66596729e-01 -1.02024865e+00 -5.96751869e-01 -4.74314809e-01
3.44715059e-01 5.46925545e-01 -2.19039783e-01 -1.01415917e-01
6.48623526e-01 5.99736094e-01 -1.82853639e+00 3.51567537e-01
4.18477565e-01 1.36896455e+00 -5.22525162e-02 6.00550413e-01
-8.93912554e-01 7.15531945e-01 -6.88887417e-01 -5.86028218e-01
-7.59610608e-02 -3.63272637e-01 -7.47062266e-01 1.05874352e-01
7.36730635e-01 1.63830817e-01 -1.09335177e-01 9.35655534e-01
2.61018306e-01 4.48792055e-02 1.19319212e+00 -1.18227696e+00
-1.27934062e+00 5.80609143e-01 -4.95221317e-01 2.29402274e-01
3.31319094e-01 -2.11934343e-01 1.30202198e+00 -9.61768687e-01
6.15686178e-01 1.42484808e+00 3.84897172e-01 1.34881333e-01
-1.28817749e+00 -5.03000498e-01 1.32509097e-01 4.24571186e-01
-1.60717487e+00 -2.30595171e-01 1.30464506e+00 -8.71797323e-01
3.45090032e-01 1.97604433e-01 8.25707197e-01 9.96401191e-01
-7.75236711e-02 7.14798510e-01 1.01808250e+00 -7.01259673e-01
2.24282235e-01 2.04375550e-01 2.61444688e-01 6.83341682e-01
1.19033493e-01 -4.55020219e-01 -4.98195499e-01 -4.82066646e-02
9.55909193e-01 2.91328758e-01 1.88096717e-01 -3.26080710e-01
-1.20939612e+00 9.42108572e-01 7.07566857e-01 6.45937622e-01
-5.02449572e-01 2.44148135e-01 5.97617745e-01 1.12979479e-01
4.73255783e-01 6.82858050e-01 8.75128284e-02 1.94592580e-01
-1.00497997e+00 2.79037878e-02 3.63273382e-01 6.76711321e-01
1.13192070e+00 1.97283253e-01 -4.37386334e-01 1.07913733e+00
3.14401656e-01 5.15235782e-01 6.32902801e-01 -6.32399082e-01
2.72767425e-01 9.68527555e-01 -2.49443904e-01 -1.38108432e+00
1.91322342e-02 -4.25652623e-01 -7.34923303e-01 1.46957114e-01
1.84039712e-01 4.02295113e-01 -1.06227922e+00 1.45197606e+00
-8.13934729e-02 1.83933288e-01 5.59680909e-02 7.11136281e-01
1.15477812e+00 5.27294993e-01 2.35443756e-01 1.36633255e-02
1.81075597e+00 -8.03749740e-01 -9.33705389e-01 -2.90873289e-01
9.70670208e-02 -8.92473459e-01 9.50877070e-01 1.41694576e-01
-9.03514445e-01 -7.46425331e-01 -1.03229511e+00 -1.83014601e-01
-1.11261797e+00 3.85920733e-01 5.40304720e-01 3.57118607e-01
-8.52246463e-01 3.69195163e-01 -2.33623475e-01 -3.34823877e-01
5.04436255e-01 6.65746555e-02 -2.60416955e-01 -3.33888829e-02
-9.39592838e-01 6.72870517e-01 3.41398925e-01 -1.70134366e-01
-4.10095990e-01 -5.45470893e-01 -8.84543180e-01 2.06778511e-01
1.89651474e-01 -6.73757136e-01 4.57122564e-01 -1.26532006e+00
-1.16928399e+00 1.05339766e+00 -5.92181161e-02 -2.36993149e-01
-8.44582543e-03 -2.73381859e-01 -5.63380122e-01 3.87142211e-01
2.83655189e-02 5.01838565e-01 1.05835092e+00 -1.44685197e+00
-2.84211904e-01 -5.53151429e-01 6.69089481e-02 2.51507722e-02
-8.93588185e-01 -1.00198381e-01 -7.65836835e-01 -1.04969382e+00
-4.71643843e-02 -7.08543718e-01 2.25081697e-01 3.41155566e-02
-1.94358751e-01 -4.01245266e-01 8.08739960e-01 -4.75654334e-01
1.31449461e+00 -2.04071283e+00 5.58700860e-01 3.43200266e-01
1.48526311e-01 9.20045674e-02 -1.66573882e-01 2.34573379e-01
1.82532683e-01 2.28279814e-01 -5.64590655e-02 -3.89110684e-01
4.10663113e-02 2.48037204e-01 -1.97763517e-01 -1.24363214e-01
3.89482707e-01 1.02194142e+00 -7.95408487e-01 -9.07205582e-01
4.61226642e-01 8.11536014e-01 -1.51565298e-01 -1.43749490e-02
-2.04971239e-01 2.25662384e-02 -6.46962047e-01 9.95764434e-01
2.11843088e-01 -3.29922229e-01 1.58848688e-01 -3.62193555e-01
1.39592022e-01 -2.51872957e-01 -1.09501708e+00 1.80945218e+00
-8.44768226e-01 9.35173571e-01 -2.66076893e-01 -7.77250528e-01
1.22240198e+00 1.74206838e-01 2.94434667e-01 -7.67071187e-01
1.18996322e-01 -3.47371846e-02 -5.74265420e-01 -3.91527742e-01
8.01997066e-01 3.60245518e-02 -1.80126607e-01 2.86535382e-01
9.16141048e-02 5.00809364e-02 1.85478225e-01 1.11269467e-01
7.74368405e-01 3.14565748e-02 3.09944093e-01 -4.38868217e-02
8.49784315e-01 -1.65606707e-01 2.55703539e-01 5.45279086e-01
1.61050662e-01 5.62490523e-01 3.23403925e-01 -8.20374250e-01
-9.74017799e-01 -1.09404278e+00 -7.31044784e-02 1.13676775e+00
-8.10301751e-02 -5.69395423e-01 -2.62557179e-01 -5.55032969e-01
2.42636293e-01 4.59047318e-01 -9.18712735e-01 -4.26241942e-03
-5.01940191e-01 -2.83556134e-01 2.83952862e-01 6.61296546e-01
3.52017790e-01 -1.14292395e+00 -3.50071222e-01 -2.04973258e-02
-6.00385889e-02 -1.00238407e+00 -3.05361718e-01 -1.31536856e-01
-8.22112501e-01 -9.56713736e-01 -7.92935431e-01 -8.64408076e-01
4.62681234e-01 -1.21359237e-01 1.31603265e+00 -1.07037937e-02
-5.97579479e-01 6.70990527e-01 -4.97974277e-01 -2.76273608e-01
-2.80623510e-02 -4.92681228e-02 -2.46721670e-01 5.22048771e-01
5.37468016e-01 -6.48406148e-01 -5.07951617e-01 -1.37285873e-01
-9.67413187e-01 -3.94792050e-01 8.35180998e-01 6.74418032e-01
8.99173260e-01 -1.93565622e-01 2.39255235e-01 -5.75301230e-01
6.12001479e-01 -4.09028947e-01 -2.33810708e-01 5.27650476e-01
-6.06384575e-01 4.22718883e-01 4.87311155e-01 -7.05471277e-01
-6.76250994e-01 1.50214404e-01 2.65389442e-01 -1.07355046e+00
-2.18881667e-01 4.21770513e-01 -4.51038852e-02 7.88281038e-02
4.66638446e-01 3.49999100e-01 -3.08814377e-01 -6.60549700e-01
4.81169701e-01 5.81060529e-01 5.87956846e-01 -6.20698273e-01
6.15985334e-01 2.48435095e-01 3.64409059e-01 -8.06807220e-01
-9.25591350e-01 -4.72459465e-01 -7.05749393e-01 -4.25048202e-01
1.14611673e+00 -7.85941541e-01 -6.77925825e-01 4.02955301e-02
-1.03133452e+00 3.02678704e-01 -2.70936698e-01 4.61277276e-01
-4.46701109e-01 3.02484244e-01 -3.16924065e-01 -8.33812177e-01
-4.60580617e-01 -1.04238236e+00 1.33438885e+00 1.21893577e-01
-1.75681993e-01 -1.20680892e+00 9.25190002e-02 4.24225211e-01
2.69459307e-01 5.51679671e-01 1.31966603e+00 -6.78548932e-01
-4.54758465e-01 -1.60255551e-01 -3.96615028e-01 8.15086886e-02
3.11549455e-01 3.50486897e-02 -9.66935456e-01 5.60798645e-02
-6.16808414e-01 -2.97160268e-01 1.10823882e+00 2.28598490e-01
1.54962218e+00 -3.25613797e-01 -2.24738047e-01 4.45508748e-01
1.79226387e+00 1.78495049e-01 6.16407871e-01 5.02002001e-01
1.04878175e+00 5.41684031e-01 2.88924336e-01 4.46926266e-01
8.02008659e-02 8.97410333e-01 4.49925750e-01 2.95136213e-01
-2.48916939e-01 -1.67623535e-01 -1.96012575e-02 6.01770103e-01
-4.08087939e-01 -2.17160434e-01 -9.31181848e-01 6.15829170e-01
-1.72586954e+00 -1.00834239e+00 -2.95217172e-03 2.07441044e+00
6.02587223e-01 7.26114437e-02 1.76489085e-01 1.61925063e-01
8.27310085e-01 3.11320364e-01 -3.11563760e-01 -5.19002557e-01
-2.02741876e-01 2.80788451e-01 1.80254325e-01 3.15950513e-01
-1.18064439e+00 1.04183805e+00 5.98514700e+00 8.62524807e-01
-8.76386881e-01 -5.06736897e-02 3.12598616e-01 -1.07500762e-01
-3.88804317e-01 -1.76611587e-01 -4.00652647e-01 5.05582452e-01
5.49415112e-01 1.04403876e-01 2.98117191e-01 7.22175598e-01
-4.20455664e-01 2.96007574e-01 -1.07569265e+00 1.38242805e+00
7.11674452e-01 -1.61350465e+00 6.74353302e-01 1.22966044e-01
4.65211660e-01 -8.10157239e-01 2.26273641e-01 2.76137799e-01
6.00023456e-02 -9.81642842e-01 7.50443578e-01 1.08748496e+00
5.85069001e-01 -9.63880897e-01 5.73563755e-01 -3.75310063e-01
-1.52225292e+00 -2.51115225e-02 -1.17559291e-01 1.08617030e-01
-1.18201241e-01 3.57336760e-01 -3.33503127e-01 6.61625981e-01
6.43681109e-01 1.00346959e+00 -1.05456364e+00 7.88011909e-01
-1.32609040e-01 1.64184600e-01 1.98232651e-01 -2.14039341e-01
1.94573775e-01 -3.08175623e-01 4.76700753e-01 1.28319299e+00
2.37828851e-01 -2.40109608e-01 8.00892636e-02 8.82388711e-01
-3.22022080e-01 3.31911027e-01 -7.56373107e-01 -3.79992723e-01
3.27126861e-01 1.45007277e+00 -8.32405925e-01 -5.38184464e-01
-1.77575648e-01 1.06439495e+00 2.89235562e-01 2.72763193e-01
-4.59637254e-01 -5.39746046e-01 6.80416942e-01 -9.80204716e-02
4.72194344e-01 -2.57903814e-01 -3.30477268e-01 -7.90318370e-01
-6.20729141e-02 -2.08072975e-01 6.02931857e-01 -1.18711483e+00
-1.38813043e+00 7.55655825e-01 -1.69844449e-01 -1.21755838e+00
-2.33521491e-01 -8.53436351e-01 -5.38161278e-01 7.08067298e-01
-1.31632841e+00 -1.52961504e+00 -3.46267343e-01 5.67049086e-01
7.70740628e-01 -6.11390293e-01 9.03369606e-01 1.73571452e-01
-3.34509730e-01 4.72057968e-01 1.58258796e-01 4.36571926e-01
7.19603956e-01 -1.38572669e+00 -2.78407365e-01 4.25819427e-01
7.53021598e-01 4.47484344e-01 5.61682880e-01 -5.72836459e-01
-1.37939739e+00 -9.10984337e-01 8.75075758e-01 -5.13010442e-01
7.60496616e-01 -2.53832102e-01 -8.30994904e-01 4.56266373e-01
3.39096785e-01 -8.73464644e-02 9.46694553e-01 4.50221032e-01
-8.66648436e-01 -1.93467915e-01 -8.53300095e-01 5.10572910e-01
8.12750697e-01 -7.77835965e-01 -8.67945790e-01 1.96075469e-01
5.47124445e-01 2.03030750e-01 -1.31903100e+00 -3.24431248e-02
7.23501444e-01 -3.22735250e-01 1.41146564e+00 -5.27671158e-01
8.66186380e-01 -2.21174181e-01 -3.82160455e-01 -1.22231698e+00
-4.56138998e-01 9.51643884e-02 -2.42103577e-01 1.59155607e+00
1.98420271e-01 2.77816653e-02 6.05738342e-01 6.21290281e-02
1.17288500e-01 -3.95907581e-01 -6.40642524e-01 -7.12974131e-01
-3.30103226e-02 -2.39834011e-01 7.59578228e-01 1.17724001e+00
-5.32172263e-01 7.02180624e-01 -3.49656075e-01 -1.94057748e-02
5.21005213e-01 6.20561659e-01 3.93894553e-01 -1.74911869e+00
-6.26886217e-03 -8.36336255e-01 -8.80267799e-01 -1.77440464e-01
4.47220117e-01 -1.01843786e+00 -6.06396556e-01 -1.81552970e+00
2.49600098e-01 -2.29684070e-01 -5.62210917e-01 4.99585271e-01
-1.53469071e-02 5.29982626e-01 4.21593845e-01 3.24667960e-01
-6.95189178e-01 5.74553192e-01 1.04103494e+00 -7.18631089e-01
-2.27126867e-01 -3.13139468e-01 -4.17663991e-01 6.60239041e-01
7.33787000e-01 -2.21052170e-01 -1.68181390e-01 -5.42858064e-01
4.68994558e-01 -2.31674328e-01 6.62638068e-01 -1.05077910e+00
8.77623558e-02 -1.35216594e-01 8.02501261e-01 -6.27161145e-01
7.49643266e-01 -9.93240058e-01 -2.42067990e-03 1.42719582e-01
-4.45396692e-01 3.09415847e-01 1.89008966e-01 8.64459872e-01
-5.15580833e-01 -3.51897597e-01 3.16493928e-01 -1.98979214e-01
-9.78160560e-01 9.84192789e-02 -2.52880126e-01 -7.15761334e-02
7.32092977e-01 -2.53626406e-01 -1.55102089e-01 -1.67058662e-01
-8.79486620e-01 -8.96194503e-02 3.19792360e-01 6.89567029e-01
7.88083017e-01 -1.79035330e+00 -4.55399126e-01 -8.15125778e-02
6.91144526e-01 -5.51135719e-01 1.48445725e-01 1.23711303e-01
-2.97139108e-01 2.73785353e-01 -5.58887601e-01 -6.04324102e-01
-1.56679785e+00 9.68340576e-01 -1.97834939e-01 -9.45550427e-02
-3.60596478e-01 5.07532835e-01 -8.42502937e-02 2.02636704e-01
2.58873582e-01 8.42346111e-04 -7.95220852e-01 7.37769544e-01
4.98095661e-01 2.66683191e-01 -8.73964950e-02 -9.77499247e-01
-4.37662631e-01 1.24328828e+00 9.28340405e-02 -1.68418124e-01
1.32024837e+00 7.69030228e-02 -1.08011790e-01 8.19544256e-01
1.25147676e+00 3.15404311e-02 -6.69420481e-01 -6.23417735e-01
4.66798581e-02 -5.32435238e-01 3.68692458e-01 -8.20740223e-01
-1.21853089e+00 1.11728108e+00 8.30982983e-01 2.62885541e-01
1.24335814e+00 2.51262009e-01 1.97901085e-01 4.04759526e-01
-1.12384697e-02 -1.14983821e+00 6.11467004e-01 3.13434243e-01
1.08591759e+00 -1.29906261e+00 7.68630803e-02 -8.51323530e-02
-7.22584903e-01 1.32603252e+00 4.14220303e-01 -1.45185832e-02
6.01785898e-01 -3.13653886e-01 -2.33890325e-01 -5.24372697e-01
-3.08977693e-01 -6.17622614e-01 7.91513205e-01 3.13420534e-01
3.21848691e-01 1.99684784e-01 -5.28438091e-02 3.06804657e-01
-1.27487451e-01 -3.85619849e-01 1.77448139e-01 8.36854875e-01
-3.84614527e-01 -1.16336036e+00 -6.23313487e-01 3.74046266e-01
-3.21597129e-01 -8.02527368e-02 -8.54791701e-01 6.31811500e-01
2.04533488e-01 6.64775610e-01 7.37873554e-01 -3.37616801e-01
2.43670464e-01 4.98747855e-01 4.67249840e-01 -4.26030278e-01
-6.86586738e-01 -8.30010995e-02 -2.33198300e-01 -1.38511077e-01
-7.04435706e-01 -5.09038508e-01 -6.76842272e-01 -1.00121021e-01
6.89781979e-02 -2.34126276e-03 9.73140836e-01 8.79186749e-01
2.81780034e-01 8.71355176e-01 4.37632710e-01 -6.99731469e-01
3.74580353e-01 -8.44091058e-01 -4.94776368e-01 6.35134101e-01
2.48599142e-01 -9.55646634e-01 -2.42566377e-01 2.06298649e-01] | [11.227643013000488, 0.4652967154979706] |
6c6e4a81-0c70-468a-a3e2-886c332ad99f | towards-crafting-text-adversarial-samples | 1707.02812 | null | http://arxiv.org/abs/1707.02812v1 | http://arxiv.org/pdf/1707.02812v1.pdf | Towards Crafting Text Adversarial Samples | Adversarial samples are strategically modified samples, which are crafted
with the purpose of fooling a classifier at hand. An attacker introduces
specially crafted adversarial samples to a deployed classifier, which are being
mis-classified by the classifier. However, the samples are perceived to be
drawn from entirely different classes and thus it becomes hard to detect the
adversarial samples. Most of the prior works have been focused on synthesizing
adversarial samples in the image domain. In this paper, we propose a new method
of crafting adversarial text samples by modification of the original samples.
Modifications of the original text samples are done by deleting or replacing
the important or salient words in the text or by introducing new words in the
text sample. Our algorithm works best for the datasets which have
sub-categories within each of the classes of examples. While crafting
adversarial samples, one of the key constraint is to generate meaningful
sentences which can at pass off as legitimate from language (English)
viewpoint. Experimental results on IMDB movie review dataset for sentiment
analysis and Twitter dataset for gender detection show the efficiency of our
proposed method. | ['Suranjana Samanta', 'Sameep Mehta'] | 2017-07-10 | null | null | null | null | ['adversarial-text'] | ['adversarial'] | [ 8.10674727e-01 4.41912591e-01 7.25904182e-02 -4.12904829e-01
-6.53131008e-01 -1.08764219e+00 8.29875052e-01 -4.14876975e-02
-3.39560717e-01 5.38924158e-01 2.02457890e-01 -1.43984407e-01
3.58612925e-01 -8.69643092e-01 -7.77775705e-01 -7.88558960e-01
3.64483923e-01 1.65802240e-01 4.10413649e-03 -3.80044490e-01
2.53278285e-01 4.65686679e-01 -1.15363669e+00 5.66263795e-01
5.65022767e-01 7.70337701e-01 -7.82771185e-02 9.23493743e-01
-1.07973274e-02 7.95432627e-01 -1.37711215e+00 -1.34531760e+00
6.93832934e-01 -7.08739340e-01 -4.15811092e-01 3.51670742e-01
6.24803424e-01 -4.06501979e-01 -4.07493919e-01 1.68129611e+00
5.74278295e-01 -1.52008295e-01 7.33295619e-01 -1.37990582e+00
-9.84231234e-01 8.38585973e-01 -4.93503243e-01 1.40723482e-01
2.59436756e-01 -1.85694508e-02 6.16500080e-01 -7.73491681e-01
4.77240115e-01 1.67317045e+00 2.19212398e-01 1.04644918e+00
-7.25971460e-01 -9.91510630e-01 3.12808275e-01 -7.13647977e-02
-1.04259121e+00 -4.87490416e-01 1.05944753e+00 -3.02709281e-01
3.63283828e-02 8.15925717e-01 3.10332417e-01 1.61539912e+00
4.19076830e-01 7.72290170e-01 9.58151162e-01 -3.88121486e-01
2.83055425e-01 7.37073779e-01 -2.64354825e-01 2.60031044e-01
2.81970769e-01 8.67808796e-03 -1.26564443e-01 -4.09640998e-01
1.30245358e-01 3.36270124e-01 -1.37566984e-01 -6.73950687e-02
-8.91987443e-01 1.12400043e+00 1.72996700e-01 -1.19825847e-01
-1.35557711e-01 -1.06811702e-01 7.53535569e-01 3.72336835e-01
5.85310221e-01 5.25673509e-01 -1.54147431e-01 3.55997711e-01
-6.11039996e-01 4.22078729e-01 4.35504109e-01 1.03651118e+00
1.97733611e-01 2.38641709e-01 -2.65970647e-01 6.58934832e-01
8.94153491e-02 1.02312183e+00 7.21122086e-01 -4.11238968e-01
5.43877423e-01 4.63402510e-01 1.07370876e-01 -1.50743687e+00
3.81290495e-01 -3.90458286e-01 -6.51318729e-01 2.37585500e-01
3.36442351e-01 -3.73473257e-01 -9.85973716e-01 1.49039316e+00
3.61161083e-01 -1.57055125e-01 4.51881409e-01 7.90285468e-01
7.38763154e-01 5.67839801e-01 -6.64134324e-02 -5.00575639e-03
1.32498300e+00 -1.07848859e+00 -8.18338871e-01 -6.34800732e-01
2.50578046e-01 -1.07590246e+00 1.15163112e+00 5.03041923e-01
-6.84196055e-01 -5.24122536e-01 -1.38966072e+00 4.87476885e-01
-6.03420675e-01 1.45632857e-02 2.75248617e-01 1.26026380e+00
-2.14744434e-01 2.36817479e-01 -1.38296694e-01 4.39993851e-02
4.99675214e-01 2.31061012e-01 -3.79644215e-01 -2.10212067e-01
-1.38631535e+00 5.26218653e-01 9.97894444e-03 -2.84536574e-02
-1.07917118e+00 -2.27991641e-01 -8.09574842e-01 -3.56790513e-01
1.98171303e-01 -1.46535680e-01 1.16654277e+00 -2.06819606e+00
-1.14046061e+00 9.88454878e-01 1.59560427e-01 -4.12102073e-01
9.94239032e-01 -1.72124896e-02 -8.68120432e-01 1.28834814e-01
2.55139321e-01 3.38511020e-01 1.75057983e+00 -1.35712183e+00
-6.63657069e-01 -4.80816215e-01 2.58818001e-01 1.64736867e-01
-8.25813711e-01 1.24419883e-01 -1.73364356e-01 -1.37904310e+00
-4.02284235e-01 -8.88205826e-01 -2.75961578e-01 -2.23523796e-01
-7.56003022e-01 3.56719524e-01 1.25505650e+00 -5.77781200e-01
1.10983431e+00 -2.38946152e+00 -1.01696856e-01 1.89497918e-01
3.09252560e-01 5.38878977e-01 5.56047522e-02 3.05961847e-01
-3.69998723e-01 3.87400836e-01 -9.96643901e-02 -2.88061917e-01
-1.06987245e-01 1.78828761e-01 -8.92779827e-01 5.02118409e-01
1.27062246e-01 5.41052699e-01 -9.95342731e-01 -3.08049649e-01
-4.86885086e-02 1.14393175e-01 -4.03886229e-01 1.98784381e-01
-2.94612758e-02 1.84958875e-01 -7.10689604e-01 5.16898930e-01
9.54938233e-01 5.52743614e-01 -1.95190489e-01 -5.15310541e-02
6.81481421e-01 -3.48633111e-01 -1.11164403e+00 7.06304133e-01
-1.95304543e-01 4.74604428e-01 -5.22671752e-02 -9.11892295e-01
1.00950217e+00 1.55591369e-01 -1.16717622e-01 -2.65425801e-01
3.33184540e-01 -1.01412265e-02 1.13613099e-01 -5.80209434e-01
7.03540742e-01 -3.29040915e-01 -6.08008146e-01 2.75419444e-01
-2.17354923e-01 -2.62982428e-01 -1.27079859e-01 3.65603596e-01
9.96742129e-01 -5.85417092e-01 1.29197404e-01 4.97248843e-02
7.10521042e-01 -1.74190119e-01 4.26100045e-01 9.70832407e-01
-4.76414055e-01 5.89023173e-01 5.29796779e-01 -3.10887188e-01
-1.12502515e+00 -1.06498027e+00 3.73552479e-02 1.07805562e+00
5.55512495e-02 -2.89167225e-01 -1.00093603e+00 -1.44325554e+00
6.21464550e-02 7.30129063e-01 -9.63712335e-01 -6.25393450e-01
-3.39565486e-01 -5.36409795e-01 1.00787342e+00 2.30606332e-01
5.38734794e-01 -1.06478477e+00 -2.71298230e-01 2.40691956e-02
1.23805016e-01 -9.99508619e-01 -7.66987383e-01 -5.33788763e-02
-3.52415144e-01 -9.50203717e-01 -7.30002403e-01 -9.22425926e-01
1.24565446e+00 1.61693305e-01 6.90042555e-01 -6.20833114e-02
-1.63062587e-01 9.91676748e-02 -7.06878781e-01 -1.04734087e+00
-1.03405309e+00 -2.52726525e-01 8.64168331e-02 6.07930779e-01
3.88941705e-01 -3.24952751e-02 -3.84867340e-01 4.07212734e-01
-1.30056775e+00 -2.02683359e-01 3.41164112e-01 9.57041442e-01
1.74353436e-01 5.90001404e-01 8.62686157e-01 -1.53417051e+00
7.82084167e-01 -4.09649730e-01 1.46744968e-02 -4.92659509e-02
-2.01628953e-01 -3.42632920e-01 1.35414529e+00 -1.02240944e+00
-9.39953566e-01 -3.84685770e-02 -2.24827919e-02 -5.42813003e-01
-1.20305032e-01 2.12698784e-02 -6.15300894e-01 1.07343927e-01
8.37725282e-01 1.59296364e-01 -3.10296961e-03 -5.16993627e-02
3.91012698e-01 9.94437933e-01 4.96524513e-01 -3.00821662e-01
1.20749438e+00 5.91543436e-01 -4.37840313e-01 -6.24300897e-01
-6.84527397e-01 7.68626928e-02 -2.19464734e-01 -3.81080240e-01
5.70381403e-01 -5.48717022e-01 -8.04249421e-02 7.84459770e-01
-7.11085916e-01 3.13186646e-01 -3.06895882e-01 -2.88593452e-02
6.47618249e-03 5.11171579e-01 -3.70466769e-01 -8.07687879e-01
-2.72922814e-01 -1.31041503e+00 9.96753097e-01 1.86365068e-01
-2.39483997e-01 -8.46995831e-01 -4.61258024e-01 4.53915238e-01
1.19219489e-01 5.87925434e-01 7.82266140e-01 -1.33662832e+00
1.41974837e-01 -1.07628536e+00 2.92350322e-01 8.55257630e-01
5.03852367e-01 3.15261453e-01 -1.11802411e+00 -4.11803395e-01
3.38480353e-01 -2.92232782e-01 5.13960719e-01 -1.29251912e-01
1.27545810e+00 -8.34404409e-01 -2.67797709e-01 3.13142151e-01
9.35491920e-01 4.74707633e-01 5.53127885e-01 1.42733261e-01
7.64594078e-01 7.07904339e-01 8.44037652e-01 4.49977815e-01
-2.12670892e-01 2.40833446e-01 6.29196823e-01 -1.43416390e-01
4.03063059e-01 -4.67881322e-01 5.97777665e-01 3.99246216e-01
5.53678989e-01 -5.39706111e-01 -5.23638964e-01 4.15603250e-01
-1.56003892e+00 -1.09895158e+00 3.01167239e-02 1.97534931e+00
9.19903696e-01 7.54709601e-01 -1.68615028e-01 4.57914144e-01
1.13854170e+00 3.61949742e-01 -6.79498732e-01 -8.39725852e-01
-1.59944683e-01 2.45112866e-01 6.91221833e-01 4.13150668e-01
-1.37970018e+00 9.72173870e-01 5.21605921e+00 1.30912006e+00
-1.29624307e+00 -9.11226794e-02 9.29346323e-01 -1.34639859e-01
-3.74492705e-01 -2.94814259e-01 -8.07289004e-01 7.07924962e-01
5.44104218e-01 -4.36756849e-01 2.79658675e-01 1.22325861e+00
7.64215887e-02 1.50127247e-01 -9.82654452e-01 7.32751131e-01
5.00322938e-01 -9.02059674e-01 4.83395427e-01 -1.46698162e-01
7.25175560e-01 -7.08595514e-01 6.49052680e-01 3.70042801e-01
3.17614377e-01 -1.11956763e+00 1.08103955e+00 1.99726105e-01
6.81702971e-01 -1.22868788e+00 8.77846003e-01 4.29916263e-01
-6.04851127e-01 -1.32469073e-01 -5.23845077e-01 -3.59080285e-02
-5.01341641e-01 5.52639663e-01 -8.90798211e-01 1.50077015e-01
5.14876366e-01 5.50054312e-01 -1.01853383e+00 1.94309518e-01
-2.76632160e-01 6.65339887e-01 7.37825334e-02 -1.99332133e-01
1.74207985e-01 1.01610191e-01 6.71018362e-01 1.21703899e+00
-7.82076120e-02 -3.18007678e-01 5.43259941e-02 4.82222229e-01
-4.68520582e-01 1.76755711e-01 -1.01124430e+00 -3.70293975e-01
4.99837101e-01 1.43623996e+00 -6.68900132e-01 -5.29314756e-01
-1.69184878e-01 1.23871064e+00 -2.62926340e-01 2.03853503e-01
-1.05523634e+00 -8.82332325e-01 5.36270499e-01 -2.64497958e-02
1.90335661e-01 4.59610909e-01 -2.73216575e-01 -9.99806941e-01
1.24256447e-01 -1.41684055e+00 3.41039330e-01 -5.17116904e-01
-1.70312488e+00 7.68779397e-01 -3.44327122e-01 -1.39118946e+00
-2.01486096e-01 -5.41393340e-01 -7.38417804e-01 9.20968711e-01
-7.69087017e-01 -1.16430318e+00 -9.63972136e-02 7.47864306e-01
8.18606615e-01 -6.74744308e-01 8.91352713e-01 1.00458495e-01
-5.14688134e-01 1.15632892e+00 1.47895843e-01 6.07250929e-01
1.00355399e+00 -1.19540858e+00 4.84628052e-01 1.22035825e+00
-2.02846035e-01 6.11808002e-01 1.04236269e+00 -8.27799618e-01
-1.07388759e+00 -1.45886981e+00 5.49093068e-01 -5.68987548e-01
5.68318069e-01 -9.16476071e-01 -4.97745842e-01 7.54354000e-01
1.89357787e-01 4.85913754e-02 8.00059021e-01 -5.79122722e-01
-4.23410773e-01 -1.42328158e-01 -1.73770225e+00 8.99478853e-01
6.42140150e-01 -5.40413916e-01 -6.31035626e-01 4.96542186e-01
6.67482793e-01 -4.24398988e-01 -3.80439013e-01 1.98936194e-01
4.98538792e-01 -6.51374876e-01 1.04644907e+00 -9.78084445e-01
9.30331171e-01 -2.64053911e-01 -3.15337241e-01 -1.41997969e+00
3.69017161e-02 -5.81818938e-01 1.23637721e-01 1.52481163e+00
2.35731661e-01 -5.37130773e-01 7.07647920e-01 4.40647632e-01
1.55068815e-01 -4.65471715e-01 -5.81112921e-01 -5.41921854e-01
2.05173731e-01 -3.90581369e-01 6.14209473e-01 1.12723768e+00
-2.58997947e-01 2.58507669e-01 -6.46221578e-01 2.63989359e-01
5.33207834e-01 -3.53292704e-01 1.07607973e+00 -4.68350261e-01
-2.94166893e-01 -1.81656852e-01 -5.73370755e-01 -5.83676636e-01
1.78385153e-01 -6.78156793e-01 -8.46685544e-02 -7.13241577e-01
-1.26287371e-01 -1.23802349e-01 4.12616245e-02 1.82391211e-01
-5.56798220e-01 4.85840678e-01 2.05934823e-01 2.06796769e-02
-1.18782334e-01 2.87065804e-01 1.31586576e+00 -4.63158578e-01
3.48024040e-01 4.44666833e-01 -1.17679238e+00 1.02910316e+00
8.04492891e-01 -7.52456427e-01 -6.74345970e-01 -9.50543433e-02
1.65574789e-01 -2.32330456e-01 2.89873093e-01 -5.88429093e-01
-2.24293828e-01 -1.97736740e-01 8.13382864e-01 -4.14039344e-01
7.45487288e-02 -9.80138898e-01 -1.64121255e-01 4.81112450e-01
-5.59404314e-01 1.30974203e-01 -3.53390165e-02 7.88895249e-01
-1.92729890e-01 -6.35167718e-01 1.10372806e+00 -2.34134033e-01
-3.81994069e-01 1.14676178e-01 -5.96481264e-01 2.70627767e-01
1.38974237e+00 -1.11998275e-01 -3.92412543e-01 -5.81407964e-01
-7.23518193e-01 -8.79595205e-02 4.29155320e-01 7.74017751e-01
7.22813129e-01 -1.28077376e+00 -6.72878444e-01 4.22821313e-01
2.84459770e-01 -3.48234534e-01 1.11598819e-01 -9.88113433e-02
-4.89099950e-01 -1.37642264e-01 -1.37665540e-01 -7.30780363e-02
-1.53123629e+00 1.04946017e+00 3.87091786e-01 -4.76078168e-02
-1.36100486e-01 1.03205287e+00 5.03150046e-01 -3.97372961e-01
2.43982911e-01 2.06115041e-02 -3.33395720e-01 2.22992614e-01
6.68362916e-01 4.63160512e-04 -4.69299965e-02 -7.28494227e-01
-2.45825246e-01 4.77701165e-02 -4.52906549e-01 5.73998280e-02
8.26644301e-01 -3.25354785e-02 2.72002891e-02 2.69998461e-01
1.28073847e+00 8.67236078e-01 -8.07690799e-01 -1.72697037e-01
-4.37402546e-01 -8.83254945e-01 -3.66410434e-01 -7.59679914e-01
-1.22387350e+00 6.57326519e-01 4.83878702e-01 4.35409606e-01
1.15736032e+00 -2.92808264e-01 7.03491628e-01 1.60125598e-01
1.18577674e-01 -9.89971757e-01 2.77343661e-01 1.84292093e-01
9.75042224e-01 -1.12646329e+00 -1.06756091e-01 -3.86475444e-01
-9.98406649e-01 1.01108813e+00 8.02345455e-01 -4.29754704e-01
5.44963777e-01 2.38579229e-01 2.77930081e-01 1.23619214e-01
-3.22999507e-01 4.82363939e-01 1.09680712e-01 7.77269483e-01
-1.22339107e-01 2.15967745e-01 -2.09173709e-01 1.18609142e+00
-7.40333796e-01 -7.24211991e-01 8.25946808e-01 9.90729868e-01
-2.08497167e-01 -1.13127506e+00 -7.49709845e-01 6.28050983e-01
-1.02873850e+00 -1.60980880e-01 -8.95735383e-01 5.65464199e-01
4.72578436e-01 1.09886479e+00 -1.67324513e-01 -7.71023929e-01
3.78798902e-01 -8.12628418e-02 1.19786628e-01 -7.10974574e-01
-8.98664415e-01 -1.15839012e-01 1.73850477e-01 1.38324603e-01
-3.90756428e-02 -6.11357987e-01 -7.82555163e-01 -4.07274812e-01
-2.03141332e-01 1.54888853e-01 3.58114809e-01 8.53074253e-01
-6.68452308e-02 6.18999839e-01 1.25193417e+00 -4.07381356e-01
-7.64335394e-01 -8.93276691e-01 -6.80383205e-01 8.82957220e-01
3.78819406e-01 -3.31777275e-01 -6.41621590e-01 3.76732796e-01] | [5.911858558654785, 8.017765045166016] |
cfb7ca54-a086-4f88-b73a-8779eae2ebfc | multi-view-learning-for-web-spam-detection | 1305.3814 | null | http://arxiv.org/abs/1305.3814v2 | http://arxiv.org/pdf/1305.3814v2.pdf | Multi-View Learning for Web Spam Detection | Spam pages are designed to maliciously appear among the top search results by
excessive usage of popular terms. Therefore, spam pages should be removed using
an effective and efficient spam detection system. Previous methods for web spam
classification used several features from various information sources (page
contents, web graph, access logs, etc.) to detect web spam. In this paper, we
follow page-level classification approach to build fast and scalable spam
filters. We show that each web page can be classified with satisfiable accuracy
using only its own HTML content. In order to design a multi-view classification
system, we used state-of-the-art spam classification methods with distinct
feature sets (views) as the base classifiers. Then, a fusion model is learned
to combine the output of the base classifiers and make final prediction.
Results show that multi-view learning significantly improves the classification
performance, namely AUC by 22%, while providing linear speedup for parallel
execution. | ['Behrouz Minaei-Bidgoli', 'Ali Hadian'] | 2013-05-16 | null | null | null | null | ['spam-detection'] | ['natural-language-processing'] | [-2.72417337e-01 -5.59127867e-01 -3.92871171e-01 -3.20069015e-01
-8.59761715e-01 -7.83302963e-01 6.24330878e-01 6.08174391e-02
1.09251089e-01 4.14902061e-01 -1.15874372e-01 -4.69903737e-01
2.06600912e-02 -1.07025278e+00 -2.78202623e-01 -6.23658180e-01
-3.10066808e-02 5.19861042e-01 1.14105916e+00 -3.29072148e-01
5.62710583e-01 3.35275620e-01 -1.66115558e+00 1.03695893e+00
1.10046601e+00 1.06325793e+00 -1.41396686e-01 5.84974408e-01
-5.34457266e-01 6.26814127e-01 -5.71628869e-01 -7.54655242e-01
3.71029109e-01 -3.68690401e-01 -6.54669642e-01 6.73617795e-02
7.28648305e-01 -4.26912904e-01 -3.54712516e-01 1.22274959e+00
5.70121370e-02 -1.78329825e-01 5.94274223e-01 -1.32427871e+00
-8.24796036e-02 2.61328757e-01 -6.07143342e-01 4.00937162e-02
3.18066955e-01 -1.85699850e-01 1.30806029e+00 -6.96125686e-01
6.72371686e-01 1.25529516e+00 5.71048975e-01 3.21350873e-01
-1.16374385e+00 -4.11798269e-01 4.73672152e-01 6.18122458e-01
-5.86625040e-01 1.53615728e-01 5.39569438e-01 1.18717197e-02
7.24773824e-01 8.34988058e-01 7.38675237e-01 1.19387662e+00
4.65613306e-01 1.04569113e+00 1.49074221e+00 -1.40067711e-01
1.82210818e-01 4.66777593e-01 7.73953795e-01 8.68311942e-01
8.36011708e-01 -3.07357639e-01 -4.65434194e-01 -1.00517333e+00
-5.12511544e-02 2.71327585e-01 -7.86855295e-02 -7.09363461e-01
-3.62177163e-01 1.11226761e+00 4.26372617e-01 -3.92081179e-02
2.82280669e-02 -2.78617322e-01 6.23130023e-01 7.23253608e-01
5.61937392e-01 9.65238139e-02 -7.33778596e-01 2.05755457e-01
-7.36037433e-01 2.23403275e-01 1.53437865e+00 7.18425870e-01
6.20188117e-01 -4.36947584e-01 1.74358413e-01 8.88548493e-01
4.16748732e-01 6.24791324e-01 7.58352101e-01 -5.37778378e-01
4.35295343e-01 1.13095236e+00 -9.83222574e-03 -8.34571183e-01
-9.03901532e-02 -3.14444125e-01 -3.58750641e-01 2.03709677e-01
4.50959980e-01 4.52712119e-01 -9.30386662e-01 8.20193052e-01
2.73601383e-01 -3.33737314e-01 -2.83257812e-01 7.85884023e-01
3.46900851e-01 3.03388625e-01 2.47953422e-02 -9.82836261e-02
1.57681561e+00 -1.22503722e+00 -2.29090214e-01 -1.93998754e-01
8.08676898e-01 -8.15510392e-01 8.59185040e-01 7.94350207e-01
-6.13724470e-01 3.99825945e-02 -1.03619385e+00 4.10348386e-01
-7.68433809e-01 -9.15154293e-02 6.79322362e-01 7.97483325e-01
-5.92089713e-01 6.45299911e-01 -8.26886117e-01 -4.96304870e-01
4.13159132e-01 2.51695544e-01 -2.65652418e-01 -3.74358371e-02
-1.05244744e+00 8.58015895e-01 1.02275610e-01 -6.26314402e-01
-5.77693760e-01 -1.87821701e-01 -8.29397980e-03 9.69837755e-02
5.04698694e-01 -4.34521556e-01 1.21253085e+00 -1.24519074e+00
-8.76385987e-01 7.22467661e-01 -8.68924931e-02 -5.58723629e-01
6.44964337e-01 -2.79842764e-01 -6.30194485e-01 3.89609963e-01
1.52755201e-01 -4.62642580e-01 1.27465141e+00 -1.61402750e+00
-1.07869601e+00 -7.68004715e-01 -2.17401162e-01 -1.81894839e-01
-7.52196491e-01 -1.70229375e-01 -4.53823775e-01 -2.86280274e-01
5.88900566e-01 -1.06801462e+00 -6.19332753e-02 -3.28417093e-01
-2.28497118e-01 -2.48243332e-01 1.24638212e+00 -8.93154800e-01
1.47853518e+00 -1.76204920e+00 -2.12327614e-01 6.77105248e-01
5.69008529e-01 3.92559767e-01 2.30861425e-01 5.98961651e-01
1.70651600e-01 2.03288764e-01 2.41929218e-01 3.74431878e-01
-1.98705688e-01 -1.25131905e-02 -5.20651221e-01 5.40732741e-01
-4.14570212e-01 4.88733470e-01 -7.69986629e-01 -6.05753362e-01
7.26755857e-02 -2.61605680e-01 -3.61290127e-01 1.91996962e-01
-1.89471871e-01 -6.79412603e-01 -8.26813936e-01 9.20745671e-01
7.50236928e-01 -6.12902880e-01 6.78308666e-01 -5.68044372e-02
3.39712769e-01 4.80141222e-01 -9.20379639e-01 7.37883866e-01
-5.29780626e-01 2.42984131e-01 3.51354152e-01 -9.45202172e-01
6.79745615e-01 -1.06274225e-02 5.87281644e-01 -4.07515317e-01
-5.51226474e-02 6.18521512e-01 -1.17558958e-02 -3.06700081e-01
-1.03518434e-01 3.79768282e-01 7.57461116e-02 6.99708343e-01
7.99953789e-02 3.53996933e-01 5.03196061e-01 5.72719693e-01
1.47625184e+00 -1.18583195e-01 2.28480637e-01 -2.52778769e-01
9.43382978e-01 4.46952432e-01 1.69308126e-01 9.12263155e-01
-2.23081231e-01 2.52152503e-01 9.67589200e-01 -7.60243237e-01
-8.05560648e-01 -1.25446618e+00 -1.15272878e-02 1.49597645e+00
3.86331022e-01 -8.47593129e-01 -9.62044477e-01 -1.74853563e+00
4.85635936e-01 4.83192980e-01 -2.42315277e-01 -1.19736940e-01
-6.64209366e-01 -1.05273247e+00 1.22455299e-01 1.11811124e-01
3.63937616e-01 -4.73683417e-01 -1.24376960e-01 2.27173552e-01
-1.72375645e-02 -6.87706053e-01 -3.20923448e-01 3.75902683e-01
-1.21920788e+00 -1.70434928e+00 -6.62425458e-02 -7.90049255e-01
7.07756102e-01 9.87949789e-01 1.10430896e+00 5.31406045e-01
-1.30960256e-01 1.19491749e-01 -6.41565919e-01 7.69836158e-02
-6.03990436e-01 3.15864801e-01 5.01386635e-02 -1.69914201e-01
7.10008621e-01 -5.42009771e-01 -5.50133467e-01 7.43570089e-01
-7.63048828e-01 -2.71067023e-01 6.74188375e-01 1.20757902e+00
3.51446904e-02 8.95219371e-02 4.74411011e-01 -1.24562347e+00
7.44785726e-01 -4.44484919e-01 -7.04686403e-01 5.75704336e-01
-1.13383877e+00 6.98190257e-02 1.04630542e+00 -3.59488428e-01
-1.06458127e+00 1.55177554e-02 1.38139203e-01 1.51572645e-01
5.05840033e-02 3.54901589e-02 -1.37668431e-01 -2.98080951e-01
9.18014228e-01 4.95061547e-01 3.61985624e-01 -6.73601389e-01
4.49710757e-01 1.35060036e+00 -3.53784114e-01 -2.10050479e-01
7.77772367e-01 5.63905239e-01 -1.67473897e-01 -6.07966483e-01
-9.09232378e-01 -9.19025004e-01 -4.50528353e-01 -1.77438155e-01
2.80775968e-02 -7.01058328e-01 -8.20143700e-01 3.74861717e-01
-7.02896297e-01 5.37157059e-01 5.03004372e-01 3.45915486e-03
-3.98015648e-01 7.86310613e-01 -8.57203126e-01 -5.81300616e-01
-6.32481515e-01 -8.75158191e-01 9.00025070e-01 -1.51486710e-01
-9.70745385e-02 -7.25210309e-01 -1.59700066e-01 7.99313784e-01
2.40851998e-01 -5.96002698e-01 1.04408896e+00 -1.48467267e+00
-7.37765670e-01 -8.27927589e-01 -3.38318676e-01 4.11104620e-01
1.45794719e-01 -3.00926447e-01 -6.58978164e-01 -4.53332484e-01
2.09825173e-01 -4.50974666e-02 1.22514200e+00 -2.57644325e-01
1.25539982e+00 -3.91710609e-01 -8.21927845e-01 3.18552762e-01
1.47260165e+00 -5.90246916e-02 3.02693725e-01 8.39746892e-01
3.47360790e-01 5.27176678e-01 4.58174855e-01 2.62352079e-01
-2.13931859e-01 3.90857786e-01 3.71700555e-01 5.58191299e-01
4.26352136e-02 -1.80924952e-01 6.60954833e-01 8.09251964e-01
3.63812804e-01 -2.75196940e-01 -6.45961702e-01 8.09559003e-02
-2.12149882e+00 -1.18300712e+00 -6.80452347e-01 2.16308904e+00
4.94396985e-01 4.29351956e-01 2.65736163e-01 -1.29797399e-01
7.13196754e-01 -3.58032733e-02 -4.04255122e-01 -2.68862784e-01
1.58365726e-01 -2.29417741e-01 9.37484205e-01 2.11560622e-01
-1.54893601e+00 9.76708531e-01 6.47309828e+00 1.02599084e+00
-8.17382812e-01 1.77043658e-02 3.95138383e-01 -8.91360268e-02
-2.07690865e-01 1.73578784e-01 -1.13837981e+00 7.39904284e-01
1.10329354e+00 -9.21371132e-02 4.51544017e-01 1.42569554e+00
-1.14331014e-01 -2.47361615e-01 -4.45261180e-01 4.81150985e-01
4.31125879e-01 -9.95497227e-01 2.95338005e-01 1.28526866e-01
6.85328841e-01 1.24268048e-01 -3.30687284e-01 1.54946536e-01
4.84162867e-01 -2.78098941e-01 -1.26911048e-02 2.77722865e-01
-1.67941689e-01 -5.15933752e-01 7.86808372e-01 5.12555480e-01
-9.00797486e-01 -4.75110292e-01 -1.70613825e-01 4.69692707e-01
-9.32787061e-02 6.27904296e-01 -6.92824006e-01 5.57870984e-01
8.59858632e-01 4.36559826e-01 -1.13685322e+00 1.05436850e+00
-5.93461804e-02 6.96325421e-01 -2.78674960e-01 -5.65617144e-01
1.73661187e-01 -4.94727999e-01 4.51563269e-01 9.04754639e-01
7.46148229e-02 -4.07601535e-01 4.13762540e-01 4.14909348e-02
1.12759821e-01 3.46241534e-01 -4.73263860e-01 9.05237272e-02
3.81520391e-03 1.63432038e+00 -8.25891197e-01 -6.90442801e-01
-9.46673691e-01 1.02905738e+00 3.59900057e-01 -6.47035753e-03
-6.16414189e-01 -3.48954201e-01 7.38420010e-01 5.98701477e-01
2.58446187e-01 -3.98789942e-02 -4.57897335e-01 -1.50198495e+00
6.27437308e-02 -1.17060697e+00 5.11578500e-01 -2.53968179e-01
-1.96288848e+00 4.65695202e-01 -3.92632723e-01 -1.50796294e+00
-2.12077409e-01 -1.21924996e+00 -2.66417891e-01 5.98262608e-01
-1.01879537e+00 -1.17877233e+00 -2.86331207e-01 2.81645954e-01
6.20159090e-01 -7.28156984e-01 7.17238247e-01 2.03523207e-02
-1.56145081e-01 2.08078459e-01 7.22177982e-01 4.46947105e-02
1.05956113e+00 -1.43736374e+00 4.00137544e-01 4.33890164e-01
8.21112469e-02 6.81563914e-01 4.82644290e-01 -1.18904316e+00
-1.65809166e+00 -7.60298014e-01 7.48941541e-01 -6.74990356e-01
1.20503473e+00 -3.26192260e-01 -9.78308201e-01 3.24147522e-01
-6.02390096e-02 -2.19453543e-01 6.47348046e-01 2.82922626e-01
-7.55440295e-01 -3.61683607e-01 -9.99839246e-01 6.49207890e-01
8.07625890e-01 -2.32419208e-01 -4.97767597e-01 7.42821157e-01
8.45593885e-02 6.59978837e-02 -5.06433845e-01 8.15239176e-02
8.33959937e-01 -1.19578516e+00 1.02456033e+00 -1.07600546e+00
3.20001930e-01 2.13652961e-02 -7.28128701e-02 -1.15953147e+00
-4.32242960e-01 7.89366588e-02 -6.02947891e-01 6.52767479e-01
3.91890079e-01 -7.65845954e-01 1.23464906e+00 1.64817989e-01
3.88247401e-01 -7.31150270e-01 -6.62123024e-01 -7.40091741e-01
-1.75984159e-01 2.17299461e-02 1.40151560e-01 4.15387034e-01
2.09352836e-01 4.10826623e-01 -3.25321048e-01 -1.43709630e-01
9.42696691e-01 7.03455687e-01 6.73443079e-01 -1.57705426e+00
-2.60867141e-02 -4.70052689e-01 -3.85245949e-01 -8.36105049e-01
1.70034304e-01 -1.11156774e+00 -3.32634598e-01 -1.41524708e+00
5.29998600e-01 -4.61034924e-02 -2.36036345e-01 4.89047691e-02
-2.92418540e-01 -9.86936837e-02 4.29208167e-02 5.69455385e-01
-8.59885454e-01 -1.97410751e-02 1.02771211e+00 -2.20636651e-01
-8.89321882e-03 6.61581993e-01 -6.79429889e-01 9.44031656e-01
9.79585350e-01 -4.61733460e-01 -6.50781915e-02 2.94756830e-01
1.11105002e-01 -1.40159994e-01 1.96129486e-01 -6.82924390e-01
2.89941609e-01 -1.51254937e-01 6.44078612e-01 -6.55717194e-01
1.55100405e-01 -1.00462401e+00 -4.07579631e-01 9.06083941e-01
-1.32487610e-01 -1.06887937e-01 -4.59089220e-01 1.03535450e+00
-1.74743995e-01 -5.98019183e-01 7.96623468e-01 -5.32946348e-01
-6.38768017e-01 8.17143619e-02 -6.62414849e-01 -1.47195339e-01
1.02874601e+00 1.11580379e-01 -8.10459673e-01 -3.59076172e-01
-5.85304439e-01 2.74599373e-01 5.77629626e-01 6.35534346e-01
3.60948384e-01 -1.16410506e+00 -3.98987770e-01 3.97954434e-01
2.14418575e-01 -1.18779325e+00 -2.52568424e-01 5.57503283e-01
-8.14173758e-01 4.03949738e-01 1.90801974e-02 -3.98543835e-01
-1.71825862e+00 4.56089497e-01 1.97383150e-01 -3.57368797e-01
-4.75666761e-01 2.61237383e-01 -3.40205759e-01 -4.79545146e-01
1.25458509e-01 6.38308048e-01 -1.15103669e-01 3.09105515e-01
5.31466186e-01 7.17012644e-01 5.27449429e-01 -2.42903054e-01
-3.54825467e-01 -1.08551178e-02 -5.56077957e-01 8.74101818e-02
1.25063217e+00 1.66132957e-01 -5.97863376e-01 -2.04889197e-02
1.26250303e+00 5.43600470e-02 -5.36432862e-01 -6.77657872e-02
5.13310313e-01 -8.62185359e-01 -5.00221252e-02 -9.57500756e-01
-7.35538661e-01 4.61532325e-01 6.78542614e-01 9.39261198e-01
8.24319124e-01 -7.52691701e-02 1.05107188e+00 6.43508017e-01
5.88099778e-01 -1.59452963e+00 1.54156253e-01 5.24999380e-01
3.81802917e-01 -1.41353667e+00 3.20200741e-01 -8.92905116e-01
-5.29731452e-01 1.36755276e+00 1.04112995e+00 -4.00785029e-01
8.75226319e-01 8.44069943e-02 -7.53980428e-02 -1.02256089e-01
-1.08335161e+00 -1.99882701e-01 2.64043957e-01 2.23364696e-01
3.77316922e-01 7.68692642e-02 -9.95619118e-01 5.76190054e-01
1.95914581e-01 -6.78366065e-01 2.61110842e-01 1.02014661e+00
-1.25303328e+00 -1.65266514e+00 -4.89502281e-01 1.60277104e+00
-5.14275014e-01 1.82978231e-02 -7.99818635e-01 5.53724766e-01
-3.57001334e-01 8.96572530e-01 -2.45892078e-01 -5.87473750e-01
2.55828589e-01 4.73772466e-01 2.00769693e-01 -4.39593345e-01
-8.49842429e-01 2.10254744e-01 2.16344818e-01 -7.27933407e-01
1.78572953e-01 -6.62627876e-01 -7.32928634e-01 -3.91389072e-01
-7.70195723e-01 3.74863952e-01 6.95384800e-01 5.87634861e-01
2.71407425e-01 1.46237046e-01 1.09148157e+00 -1.02722608e-01
-1.51656973e+00 -7.98050106e-01 -7.48026133e-01 7.03053772e-01
-1.81759462e-01 -4.32721585e-01 -8.17787409e-01 -2.22636521e-01] | [7.839234352111816, 10.023884773254395] |
29436169-b874-4573-a735-a9a021f73556 | a-feedback-neural-network-for-small-target | 1805.00342 | null | http://arxiv.org/abs/1805.00342v2 | http://arxiv.org/pdf/1805.00342v2.pdf | A Feedback Neural Network for Small Target Motion Detection in Cluttered Backgrounds | Small target motion detection is critical for insects to search for and track
mates or prey which always appear as small dim speckles in the visual field. A
class of specific neurons, called small target motion detectors (STMDs), has
been characterized by exquisite sensitivity for small target motion.
Understanding and analyzing visual pathway of STMD neurons are beneficial to
design artificial visual systems for small target motion detection. Feedback
loops have been widely identified in visual neural circuits and play an
important role in target detection. However, if there exists a feedback loop in
the STMD visual pathway or if a feedback loop could significantly improve the
detection performance of STMD neurons, is unclear. In this paper, we propose a
feedback neural network for small target motion detection against naturally
cluttered backgrounds. In order to form a feedback loop, model output is
temporally delayed and relayed to previous neural layer as feedback signal.
Extensive experiments showed that the significant improvement of the proposed
feedback neural network over the existing STMD-based models for small target
motion detection. | ['Shigang Yue', 'Hongxin Wang', 'Jigen Peng'] | 2018-05-01 | null | null | null | null | ['motion-detection'] | ['computer-vision'] | [ 4.10304070e-01 -6.51345193e-01 -1.89327329e-01 -3.84233706e-02
4.67886686e-01 -7.55452752e-01 3.63340288e-01 -2.49345481e-01
-4.90818739e-01 4.24899369e-01 -1.76143661e-01 -2.87999988e-01
5.45520365e-01 -5.82934082e-01 -5.60278594e-01 -8.70279491e-01
-8.99156183e-02 -4.63373244e-01 1.15118790e+00 -6.50893971e-02
4.66066509e-01 6.89581096e-01 -1.51579213e+00 5.41413605e-01
2.84546077e-01 9.08528626e-01 7.70159900e-01 9.32889760e-01
6.07335288e-03 6.88940227e-01 -7.28577673e-01 4.51228678e-01
4.35443610e-01 -8.00452173e-01 -1.19092315e-01 -2.37920821e-01
2.50585109e-01 -4.34091985e-01 -1.03160046e-01 1.17867076e+00
5.94576418e-01 -1.30253389e-01 8.12895715e-01 -9.99202907e-01
-6.54487789e-01 2.27086022e-01 -1.11054826e+00 1.01467896e+00
8.15083608e-02 5.83976388e-01 3.82887900e-01 -7.50802755e-01
5.28767884e-01 1.56886244e+00 2.19505340e-01 1.12209380e+00
-1.27302206e+00 -8.65479410e-01 1.11062020e-01 2.87347615e-01
-9.62654650e-01 -3.75384569e-01 6.55733407e-01 -5.37241757e-01
1.03603733e+00 2.85198241e-01 7.47259319e-01 1.16176772e+00
5.47579288e-01 7.50425637e-01 9.40300822e-01 -3.61202210e-01
5.80757670e-02 -1.19348131e-01 2.81632632e-01 3.60267431e-01
5.88093102e-01 7.19036579e-01 -5.09995222e-01 2.24523380e-01
1.00281239e+00 -2.49353692e-01 -4.00743663e-01 -1.05199046e-01
-9.98024762e-01 5.41589975e-01 9.57023084e-01 2.92157918e-01
6.03510402e-02 2.06693247e-01 1.46261692e-01 1.58646122e-01
-2.03804314e-01 -1.04820132e-02 2.00571530e-02 1.93698868e-01
-3.46655071e-01 -6.57029003e-02 4.56625670e-01 2.86551178e-01
3.42572570e-01 4.40855980e-01 -3.92212957e-01 5.56488097e-01
6.33949339e-01 9.61334646e-01 7.08009660e-01 -7.10040033e-01
3.61057222e-01 8.50436866e-01 2.11752102e-01 -1.29974043e+00
-6.01900578e-01 -1.72363698e-01 -7.80295908e-01 8.84793758e-01
5.76991498e-01 6.38640001e-02 -8.87885034e-01 1.64615381e+00
4.25643951e-01 1.49108484e-01 1.55211136e-01 1.41063154e+00
9.63423431e-01 1.03202724e+00 -1.05192326e-01 -4.18136805e-01
1.19075298e+00 -5.68692982e-01 -4.84986156e-01 -6.81254566e-01
2.21126303e-01 -7.25119054e-01 1.00546873e+00 1.11612990e-01
-4.62374687e-01 -7.36498535e-01 -1.12333012e+00 2.81164050e-01
-2.34101534e-01 3.77623767e-01 3.76328588e-01 6.09318078e-01
-6.98636115e-01 1.31982431e-01 -8.83367956e-01 -5.70328593e-01
2.95256495e-01 4.21142280e-01 -7.37412320e-03 1.94396347e-01
-8.02494228e-01 7.40161538e-01 3.96787465e-01 5.43125629e-01
-1.27204609e+00 1.46077394e-01 -5.60910225e-01 -2.28496760e-01
7.56639661e-03 -5.13762712e-01 1.28171468e+00 -1.09367383e+00
-1.30969250e+00 8.22875261e-01 -3.52722347e-01 -5.38950503e-01
1.12140656e-01 7.25945160e-02 -3.15869540e-01 8.98208767e-02
-7.06193075e-02 9.71487761e-01 1.20264673e+00 -1.18357325e+00
-7.43920684e-01 -4.74913597e-01 -1.06651083e-01 2.85610288e-01
-2.85764933e-01 4.04836088e-01 8.40860978e-02 -6.46341205e-01
-9.01103020e-02 -8.05916309e-01 -1.90758228e-01 5.51637709e-01
-1.13954142e-01 -2.00396448e-01 1.43349600e+00 1.27764437e-02
1.15043581e+00 -2.15935826e+00 -1.70533448e-01 -4.08838719e-01
1.42683210e-02 8.38023603e-01 -2.30226859e-01 8.44134763e-03
2.27261305e-01 -2.91988969e-01 -4.60904315e-02 6.06206000e-01
-5.54759324e-01 1.69292390e-02 -4.57412422e-01 3.94130647e-01
3.52261931e-01 7.16719329e-01 -6.37188733e-01 -2.63442874e-01
1.67465612e-01 2.26957411e-01 -1.27946839e-01 3.49668503e-01
-3.05916816e-01 5.31588376e-01 -4.67698485e-01 8.00877213e-01
7.64877617e-01 -6.81296438e-02 -3.19110960e-01 -2.52962261e-01
-5.29120862e-01 -1.78271204e-01 -9.76243556e-01 1.01486778e+00
2.09102973e-01 1.08828747e+00 1.98770538e-01 -8.49291503e-01
1.22332525e+00 -1.51239887e-01 -9.36796665e-02 -8.86796296e-01
4.80765462e-01 1.93874642e-01 5.96933186e-01 -6.28819406e-01
3.20120603e-02 1.29338438e-02 2.68326581e-01 2.80027390e-01
-4.11574543e-01 3.46307874e-01 9.15209800e-02 -1.44767374e-01
8.12466025e-01 1.03986382e-01 2.76268989e-01 -2.72558369e-02
3.85438710e-01 2.38896057e-01 7.44535446e-01 5.77562213e-01
-5.47987282e-01 4.98813838e-01 1.06784545e-01 -4.77609396e-01
-5.11515796e-01 -8.95161808e-01 4.70635556e-02 9.04748976e-01
6.67938888e-01 3.11091930e-01 -4.55196440e-01 -1.90595463e-01
-1.87893897e-01 1.91601634e-01 -6.52912021e-01 -4.78817523e-01
-6.23693943e-01 -9.08645272e-01 5.06788194e-01 6.92328572e-01
6.78858757e-01 -1.44301200e+00 -1.72062528e+00 1.49463788e-01
1.97312549e-01 -6.36685669e-01 -4.69965696e-01 4.47284430e-01
-7.99068272e-01 -1.10386586e+00 -9.37701106e-01 -1.01737440e+00
4.90687102e-01 9.63742256e-01 4.13264334e-01 -3.93858612e-01
-5.69645345e-01 -1.70376033e-01 -2.33056188e-01 -8.07442069e-01
-7.63423815e-02 -5.15311182e-01 1.60922930e-01 -1.86830480e-02
6.18021429e-01 -1.96054041e-01 -8.18234205e-01 5.93200564e-01
-8.02350223e-01 -2.10796855e-02 7.20207036e-01 8.01612735e-01
2.25352332e-01 -3.24863166e-01 2.66705841e-01 -2.37842485e-01
5.91832042e-01 -1.26524553e-01 -1.13333809e+00 5.85711375e-02
3.43280770e-02 1.05671331e-01 5.57673931e-01 -9.01105881e-01
-1.07706237e+00 2.23450109e-01 3.06472898e-01 -5.00083625e-01
-2.98344016e-01 3.86666469e-02 2.52665728e-02 -4.54688549e-01
1.01810253e+00 2.50852823e-01 -1.47362053e-01 -2.85644531e-01
-9.03163385e-03 6.86665714e-01 4.75034267e-01 2.16955215e-01
6.88985765e-01 5.85165322e-01 2.41773874e-01 -1.01530099e+00
-2.90900826e-01 -4.35930192e-01 -3.88596088e-01 -4.94076103e-01
9.70015109e-01 -6.00188017e-01 -1.13268030e+00 7.90909052e-01
-1.41225314e+00 -2.41340190e-01 2.16969728e-01 7.22303450e-01
-1.22767776e-01 1.67641461e-01 -4.95323241e-01 -1.04137075e+00
-3.75558168e-01 -9.57519233e-01 6.33907735e-01 1.05511403e+00
-1.23582901e-02 -5.56870639e-01 8.07359591e-02 -3.21825415e-01
4.24687713e-01 1.34907454e-01 5.17505288e-01 -1.84983596e-01
-6.73592389e-01 -5.59112541e-02 -5.23414254e-01 -7.57912099e-02
3.38363856e-01 2.17282832e-01 -1.12689662e+00 -2.20114484e-01
1.86456904e-01 -3.44845563e-01 1.36742592e+00 1.05470264e+00
5.01060009e-01 -1.14352293e-01 -7.22923815e-01 6.41087472e-01
1.36778963e+00 1.08327043e+00 2.49028087e-01 3.32632154e-01
4.30247068e-01 6.13908768e-01 7.39247441e-01 3.54978591e-01
-4.74188209e-01 5.00172555e-01 7.03591585e-01 8.22447892e-03
-3.05246055e-01 -2.76673045e-02 6.41105771e-01 2.82737136e-01
2.86680460e-02 -4.13302004e-01 -5.70223212e-01 3.78618658e-01
-1.71119976e+00 -1.07792115e+00 -4.47787166e-01 2.22327399e+00
3.68572503e-01 4.45350140e-01 2.37969816e-01 -7.95843303e-02
8.27012956e-01 4.65615944e-04 -1.11354387e+00 -3.26449782e-01
-6.16311312e-01 -2.71492332e-01 5.14378250e-01 2.80394435e-01
-1.02421772e+00 8.40252578e-01 6.73081923e+00 4.50406790e-01
-1.66889369e+00 -4.06593502e-01 2.52524227e-01 -7.00092986e-02
1.91341817e-01 -7.28630945e-02 -1.40853453e+00 3.58633876e-01
4.79921430e-01 -5.85843250e-02 1.07490666e-01 6.29707992e-01
5.94900012e-01 -4.71650302e-01 -8.43488932e-01 1.00675857e+00
-1.03251018e-01 -1.16254497e+00 1.52258113e-01 -2.67394241e-02
3.95513922e-01 -1.76227093e-01 2.12392792e-01 -2.91700840e-01
2.19379216e-01 -8.79046500e-01 3.74651343e-01 2.46809676e-01
2.68511802e-01 -3.86280268e-01 5.66187143e-01 6.21883214e-01
-1.38112175e+00 -3.89308035e-01 -1.08269036e+00 -3.95031214e-01
-3.61032709e-02 -2.65782252e-02 -9.88890469e-01 -2.28954732e-01
8.82085681e-01 8.21714222e-01 -6.97149754e-01 1.45558321e+00
3.76255624e-02 6.55616045e-01 -4.13058251e-01 -6.09688580e-01
2.82950848e-01 -8.38540271e-02 9.29927647e-01 1.07781398e+00
2.46837124e-01 3.63082178e-02 -1.31477535e-01 1.05103314e+00
4.06587154e-01 -1.35639042e-01 -1.04496658e+00 -3.69288206e-01
3.47865224e-01 1.05547225e+00 -9.31175053e-01 3.78181525e-02
-2.97479302e-01 1.01203823e+00 -1.30513951e-01 3.65049064e-01
-6.18837953e-01 -5.92897892e-01 4.67363745e-01 -7.56795378e-03
5.33951283e-01 -2.07476303e-01 -6.03527278e-02 -8.15300226e-01
-1.33486971e-01 -3.32456976e-01 4.22549188e-01 -8.71879518e-01
-9.83868062e-01 4.81134534e-01 -2.01128811e-01 -1.62467062e+00
-1.00803100e-01 -1.14174831e+00 -8.72343063e-01 9.07284915e-01
-1.04745340e+00 -1.08088875e+00 -4.24645841e-01 6.92455769e-01
6.34136021e-01 -1.84798643e-01 5.35695910e-01 -1.44785196e-01
-6.44732773e-01 3.83371532e-01 8.62406865e-02 5.74529804e-02
7.03092813e-01 -7.56022573e-01 1.71729177e-01 1.27878535e+00
1.00945443e-01 3.66293192e-01 7.38248229e-01 -7.28803992e-01
-1.36378312e+00 -8.63038421e-01 7.35453293e-02 -7.12476000e-02
3.46346140e-01 -2.94867456e-01 -8.50833297e-01 2.86767989e-01
1.74468592e-01 9.40615311e-02 3.47030967e-01 -7.85118222e-01
-1.07575633e-01 -4.17049617e-01 -8.66962552e-01 6.06390536e-01
9.25975680e-01 1.48487106e-01 -6.00039840e-01 -2.76067734e-01
6.71928644e-01 -7.48703107e-02 2.69787878e-01 3.23585927e-01
8.21004510e-01 -9.19536352e-01 9.22594130e-01 -5.48905492e-01
-6.53931797e-02 -9.73178864e-01 -1.72286302e-01 -1.15243471e+00
-4.43386197e-01 -2.23726466e-01 1.01922259e-01 1.04824102e+00
9.92876887e-02 -4.59155083e-01 7.79649675e-01 -1.03849649e-01
5.13394661e-02 -1.35508040e-02 -9.00709748e-01 -8.53882253e-01
-2.40600586e-01 -1.34440251e-02 -2.77431041e-01 4.14960176e-01
-2.07454756e-01 6.90028787e-01 -5.42644143e-01 2.80811369e-01
7.48978913e-01 3.75978380e-01 4.18808341e-01 -1.07770956e+00
-2.11604610e-01 -5.11046052e-01 -6.88709259e-01 -1.31398749e+00
-4.49750900e-01 -5.18842518e-01 1.57113373e-01 -1.64871097e+00
-4.77157719e-03 3.05663109e-01 -3.22843552e-01 3.97520274e-01
-1.38930902e-01 4.04370874e-01 1.05019681e-01 2.95835257e-01
-2.36145467e-01 2.23692462e-01 1.41990614e+00 -3.07080835e-01
-5.52208602e-01 3.03800732e-01 -3.61032397e-01 6.30730391e-01
7.00745523e-01 -4.43544477e-01 -4.82568651e-01 -6.34709477e-01
-2.06763342e-01 -7.85844773e-02 4.02413100e-01 -1.15325844e+00
5.35302520e-01 -4.96011704e-01 8.63915861e-01 -8.38786483e-01
3.94629151e-01 -7.85309851e-01 -1.25045970e-01 1.25338852e+00
-1.28846899e-01 -1.66055396e-01 5.90222955e-01 6.80872381e-01
-1.24285810e-01 -2.01314434e-01 1.12152278e+00 -2.39766881e-01
-1.26896834e+00 -1.05960019e-01 -1.10918057e+00 -2.65465826e-01
1.31849289e+00 -7.44947076e-01 -6.45427048e-01 -4.86364104e-02
-3.72451723e-01 6.96333721e-02 1.31369352e-01 5.24102747e-01
1.05251408e+00 -1.28599691e+00 -3.67120445e-01 5.82025588e-01
4.23268750e-02 -3.08927715e-01 -7.11928261e-03 6.50305510e-01
-5.24021924e-01 6.62785351e-01 -8.25693011e-01 -8.76519024e-01
-1.70678318e+00 7.56676137e-01 4.46666032e-01 5.31039834e-01
-1.11080907e-01 1.31534886e+00 7.58920252e-01 3.11011136e-01
2.69066989e-01 -6.81405246e-01 -5.61409652e-01 -2.71463782e-01
9.37453806e-01 1.91835701e-01 -5.81965327e-01 -4.95221823e-01
-5.49228668e-01 9.81402874e-01 1.59780562e-01 6.07109778e-02
9.40814793e-01 -2.28856117e-01 5.36977239e-02 8.32161844e-01
7.12060750e-01 -2.98103660e-01 -1.41566229e+00 1.52338937e-01
-1.85367867e-01 -6.95859313e-01 -3.34988028e-01 -6.69886053e-01
-7.55521178e-01 1.51503563e+00 1.29645252e+00 8.75118673e-02
1.34293950e+00 -9.68293697e-02 4.03635353e-01 4.68934059e-01
2.67758399e-01 -5.23316085e-01 5.76081872e-01 3.28531891e-01
7.97379375e-01 -1.28407383e+00 -5.78536868e-01 -2.43316412e-01
-5.23744166e-01 1.38325548e+00 1.31670022e+00 -2.96352923e-01
3.32131863e-01 4.43155766e-01 3.08712840e-01 1.62509859e-01
-1.00048780e+00 -2.85160154e-01 3.26560229e-01 1.10926414e+00
3.46942067e-01 -1.79191440e-01 -2.24836022e-01 3.02762687e-01
3.27923656e-01 6.74753711e-02 2.98876405e-01 9.61495161e-01
-1.31655610e+00 -7.12895811e-01 -7.51399159e-01 6.67300001e-02
-9.23263654e-02 1.91206157e-01 -9.47070658e-01 4.07445848e-01
1.43235266e-01 1.01144898e+00 1.96686640e-01 -6.03948772e-01
2.12374523e-01 -5.24649084e-01 6.02051377e-01 -5.03199041e-01
-1.37035087e-01 3.42755169e-01 -4.77776498e-01 -3.19461882e-01
-4.82350379e-01 -1.62305504e-01 -1.35703528e+00 2.07187414e-01
-3.82429838e-01 -1.87883750e-01 4.22608674e-01 4.29969519e-01
7.90341049e-02 2.41243854e-01 4.12895650e-01 -7.05458939e-01
-3.52923274e-01 -1.03617430e+00 -4.95980382e-01 3.30816507e-02
4.70582306e-01 -7.89479434e-01 -3.01158577e-01 1.61823437e-01] | [8.405130386352539, -0.8393992781639099] |
8ec32017-46f8-4f39-9f15-8ec4d1d20bd0 | end-to-end-table-question-answering-via | 2203.16714 | null | https://arxiv.org/abs/2203.16714v1 | https://arxiv.org/pdf/2203.16714v1.pdf | End-to-End Table Question Answering via Retrieval-Augmented Generation | Most existing end-to-end Table Question Answering (Table QA) models consist of a two-stage framework with a retriever to select relevant table candidates from a corpus and a reader to locate the correct answers from table candidates. Even though the accuracy of the reader models is significantly improved with the recent transformer-based approaches, the overall performance of such frameworks still suffers from the poor accuracy of using traditional information retrieval techniques as retrievers. To alleviate this problem, we introduce T-RAG, an end-to-end Table QA model, where a non-parametric dense vector index is fine-tuned jointly with BART, a parametric sequence-to-sequence model to generate answer tokens. Given any natural language question, T-RAG utilizes a unified pipeline to automatically search through a table corpus to directly locate the correct answer from the table cells. We apply T-RAG to recent open-domain Table QA benchmarks and demonstrate that the fine-tuned T-RAG model is able to achieve state-of-the-art performance in both the end-to-end Table QA and the table retrieval tasks. | ['James Hendler', 'Alfio Gliozzo', 'Michael Glass', 'Mustafa Canim', 'Feifei Pan'] | 2022-03-30 | null | null | null | null | ['table-retrieval'] | ['natural-language-processing'] | [ 3.24106932e-01 1.98889539e-01 -2.37135291e-02 -4.45591629e-01
-2.57937288e+00 -1.03181422e+00 6.20894551e-01 5.04333138e-01
-2.25183353e-01 4.44198251e-01 3.75377506e-01 -3.98797125e-01
-2.29904696e-01 -9.42817509e-01 -8.48169744e-01 -7.83759132e-02
5.05370498e-01 1.60360324e+00 6.34407818e-01 -7.27430761e-01
4.74334121e-01 -1.45670339e-01 -1.49283993e+00 1.05661929e+00
1.06294274e+00 1.23175836e+00 -7.32172206e-02 9.03620124e-01
-4.83846515e-01 1.17078650e+00 -6.50496066e-01 -8.61456811e-01
2.57897884e-01 -4.77810502e-01 -1.27861309e+00 -6.43568039e-01
7.01874375e-01 -5.09957254e-01 -3.05338383e-01 5.59600711e-01
5.93945742e-01 1.18899211e-01 6.51746690e-01 -7.32925773e-01
-7.25732505e-01 7.11003065e-01 -3.75295430e-02 1.77131414e-01
8.65803540e-01 1.00795612e-01 1.75147533e+00 -9.86299038e-01
8.76666486e-01 1.49394703e+00 3.35394710e-01 2.08072290e-01
-1.12190175e+00 -1.58458680e-01 -2.29000404e-01 3.63359272e-01
-1.27028215e+00 -5.64278185e-01 3.64279747e-01 -4.74812351e-02
1.43613696e+00 3.89789164e-01 9.22964048e-03 4.88657832e-01
3.21269691e-01 1.19528592e+00 6.04483128e-01 -5.05964935e-01
2.22709835e-01 -3.33908349e-02 2.61938244e-01 6.23571396e-01
-2.79325396e-01 -4.55953389e-01 -8.05459797e-01 -4.38949317e-01
3.22833583e-02 -5.85988522e-01 -7.71363080e-02 -5.22179842e-01
-1.29057765e+00 9.30810213e-01 5.31155765e-01 -8.43263119e-02
-5.18050075e-01 -1.04805389e-02 6.37479424e-01 4.57745045e-01
2.22884268e-01 8.02157402e-01 -7.07314730e-01 -2.13813096e-01
-9.16929305e-01 9.16090250e-01 1.12017429e+00 1.08288944e+00
7.53214538e-01 -6.79829657e-01 -1.17092979e+00 7.88130879e-01
1.25412658e-01 5.62439322e-01 1.98217466e-01 -9.27630305e-01
1.01439273e+00 9.10558879e-01 2.81475544e-01 -6.82391703e-01
2.09019892e-02 -3.53233516e-01 -1.41964644e-01 -5.12287498e-01
5.85673273e-01 4.62065339e-01 -8.98171902e-01 1.37147427e+00
4.43664134e-01 -7.98570931e-01 2.80370206e-01 8.26606750e-01
1.01928067e+00 1.12845743e+00 -9.23121944e-02 2.04286143e-01
1.76347721e+00 -1.28601813e+00 -7.41060019e-01 -4.67712015e-01
8.81600320e-01 -8.59629273e-01 1.54195786e+00 3.04730982e-01
-1.44492424e+00 -4.35329139e-01 -7.12620974e-01 -9.31579888e-01
-3.57303023e-01 1.17089227e-01 2.47193888e-01 3.21947664e-01
-9.54919875e-01 -1.27717242e-01 -3.84919852e-01 -1.29486546e-01
1.38208359e-01 2.08981588e-01 -2.15920016e-01 -5.98661840e-01
-1.37090611e+00 1.02691007e+00 2.25788385e-01 -2.29284391e-02
-8.08294058e-01 -1.03081918e+00 -8.50218356e-01 3.47712547e-01
7.25008011e-01 -1.16295886e+00 1.92243087e+00 -7.39018852e-03
-1.33592796e+00 8.46993268e-01 -5.91408789e-01 -6.32594228e-01
2.61055589e-01 -5.14977455e-01 -6.86350316e-02 3.92280787e-01
4.77556050e-01 6.29860640e-01 4.17244852e-01 -9.88859832e-01
-5.49529552e-01 -3.53352457e-01 7.48509467e-02 3.73284757e-01
3.50141346e-01 8.37368816e-02 -1.05308890e+00 -3.13064367e-01
2.50850059e-03 -5.89429438e-01 -3.31037715e-02 -1.67944357e-01
-4.44942713e-01 -7.52773523e-01 3.77336353e-01 -1.04566801e+00
1.52900100e+00 -1.69202769e+00 6.09151348e-02 1.14741549e-01
1.51292592e-01 1.45715907e-01 -3.04086030e-01 8.89751256e-01
4.97330397e-01 -2.40502611e-01 -4.26797271e-01 -1.22854508e-01
4.41211432e-01 -9.50277373e-02 -6.72481596e-01 -1.22923136e-01
4.41347241e-01 1.35398114e+00 -8.72017741e-01 -4.99704659e-01
-2.18313515e-01 2.35458449e-01 -7.40847945e-01 5.98168194e-01
-9.91638660e-01 -1.13226980e-01 -6.89038277e-01 7.18506932e-01
3.57117981e-01 -4.92909789e-01 -2.69102231e-02 -8.31281468e-02
4.06732440e-01 1.28948855e+00 -8.28064382e-01 1.95486641e+00
-3.98006648e-01 2.61172801e-01 -9.74554867e-02 -4.57973450e-01
8.50292444e-01 8.51452425e-02 5.46570942e-02 -1.40208316e+00
-6.28160238e-02 6.41832173e-01 -1.44026950e-01 -3.65934044e-01
9.29103792e-01 3.16067040e-01 -5.42714477e-01 6.67739034e-01
-6.83879107e-02 -5.25896132e-01 5.09095669e-01 7.68792927e-01
1.45894945e+00 1.81767404e-01 1.32988319e-01 -9.44892094e-02
8.56559575e-01 7.16661334e-01 2.99651157e-02 1.04042065e+00
2.73287743e-01 6.83147669e-01 4.47979778e-01 -2.17556804e-01
-8.62903476e-01 -1.28079104e+00 2.08345324e-01 1.44240820e+00
-1.79620460e-02 -6.45630300e-01 -8.86846602e-01 -8.32765937e-01
1.66767001e-01 9.25380290e-01 -4.88773584e-01 -2.58147925e-01
-8.96191001e-01 -9.41254050e-02 6.10378981e-01 3.33995104e-01
4.31322217e-01 -1.11214471e+00 -4.82773751e-01 5.46887875e-01
-8.32857728e-01 -1.00334072e+00 -6.69769883e-01 2.29551002e-01
-4.68053669e-01 -9.31201816e-01 -4.71622854e-01 -7.40149021e-01
2.09873438e-01 -2.00910065e-02 2.04849339e+00 6.35824874e-02
3.71787213e-02 3.67120206e-01 -3.33515465e-01 -2.53054917e-01
-6.90512419e-01 6.03626966e-01 -6.46903217e-01 -2.66802639e-01
5.86928785e-01 2.63910502e-01 -7.24953055e-01 3.12447011e-01
-9.10136461e-01 -2.53450006e-01 5.65369189e-01 8.70005131e-01
9.37454760e-01 -5.10099947e-01 7.00901091e-01 -9.60273921e-01
9.84337807e-01 -3.50701720e-01 -7.89729357e-01 7.92209089e-01
-6.53143525e-01 6.53567195e-01 5.76587260e-01 2.77498178e-03
-8.58682156e-01 -2.32528508e-01 -4.86554533e-01 3.47481258e-02
3.79282922e-01 7.10714400e-01 -2.16470748e-01 4.29664403e-01
8.57618332e-01 4.23931301e-01 -8.71203840e-02 -5.04664779e-01
7.77380824e-01 6.88454509e-01 7.06268609e-01 -7.40408123e-01
9.05898631e-01 2.62170807e-02 -4.01807070e-01 1.18281886e-01
-1.21931756e+00 -7.77667582e-01 -4.63862181e-01 1.06869034e-01
7.22056985e-01 -1.01411545e+00 -5.96341670e-01 -1.64878264e-04
-1.06585681e+00 -3.57455015e-01 -3.34452301e-01 -1.88655108e-01
-7.17059135e-01 5.84396534e-02 -6.55214310e-01 -5.82842231e-01
-1.02228570e+00 -1.33250058e+00 1.62969148e+00 -9.45566744e-02
-3.42100680e-01 -6.44870102e-01 1.86396286e-01 1.02211595e+00
6.08753264e-01 -3.81710976e-01 1.49230409e+00 -1.06410408e+00
-1.22343016e+00 -3.35583001e-01 -2.62772888e-01 -1.08180262e-01
-2.88007200e-01 -4.97373730e-01 -8.34493577e-01 -1.68212101e-01
-2.94249952e-01 -8.97810936e-01 9.27204907e-01 -1.07924230e-01
7.32288539e-01 -4.30763781e-01 -2.19323263e-01 1.19220972e-01
1.26928234e+00 1.76132843e-01 8.48058343e-01 6.03997946e-01
3.24058741e-01 7.22593248e-01 1.10336852e+00 1.41428784e-02
8.99150968e-01 8.61117244e-01 1.77675948e-01 3.73314172e-01
-1.82982236e-01 -7.09037840e-01 3.26123118e-01 7.38263249e-01
1.02259123e+00 -6.47269428e-01 -1.05968094e+00 8.25522780e-01
-1.64921558e+00 -6.30181372e-01 1.31722679e-02 2.15930247e+00
1.18699682e+00 1.58104956e-01 4.27045897e-02 1.96360514e-01
1.29841655e-01 1.30248955e-02 -6.36154294e-01 -4.79948670e-01
-1.48957167e-02 5.98330617e-01 1.02982990e-01 5.74484766e-01
-7.16057003e-01 1.16669524e+00 6.26893616e+00 1.12980080e+00
-5.41676223e-01 -1.06154196e-01 6.58085585e-01 -1.77776553e-02
-7.12621748e-01 1.22034602e-01 -1.06696975e+00 -9.55400392e-02
1.29138923e+00 -2.52234191e-01 5.02386808e-01 6.62273943e-01
-5.08477449e-01 -2.33183175e-01 -1.27413607e+00 7.54970372e-01
1.67563036e-01 -1.54760718e+00 4.77259159e-01 -4.39136207e-01
2.80806094e-01 -1.25849783e-01 1.47057742e-01 9.03519630e-01
3.40953380e-01 -1.08565855e+00 7.94173002e-01 4.28337097e-01
8.43857646e-01 -7.22824395e-01 6.25437677e-01 3.28021765e-01
-1.18052375e+00 -8.57827887e-02 -2.83580959e-01 3.52072835e-01
1.72721684e-01 2.54952550e-01 -1.16729808e+00 6.58594906e-01
6.53244197e-01 4.73929644e-02 -8.47257137e-01 8.65753651e-01
-2.44601682e-01 5.13677001e-01 -4.06685770e-01 -2.96042860e-01
4.19860929e-01 1.52065963e-01 2.82096148e-01 9.72893059e-01
1.36013106e-01 -1.55981593e-02 -1.55512556e-01 9.01451886e-01
-4.11369294e-01 1.85929313e-01 -1.79637253e-01 1.10975256e-04
6.65605009e-01 9.85871434e-01 -3.21233600e-01 -5.14212310e-01
-1.45091996e-01 8.97113979e-01 5.38129151e-01 1.70348912e-01
-4.45942581e-01 -7.19215751e-01 3.15705240e-01 3.97205390e-02
5.89807749e-01 1.72709301e-01 -3.45432937e-01 -9.00638461e-01
5.02021790e-01 -1.69884813e+00 9.44090784e-01 -1.01103151e+00
-1.03732312e+00 7.13546216e-01 -1.12866843e-02 -9.35385168e-01
-8.01516533e-01 -3.09467822e-01 -1.90929286e-02 1.16893995e+00
-1.54142368e+00 -9.55759048e-01 6.18930086e-02 6.69478655e-01
7.35783458e-01 -6.86914399e-02 8.26833725e-01 2.14460626e-01
-5.77278920e-02 1.02634776e+00 2.31033742e-01 2.17583790e-01
8.86302233e-01 -1.46151650e+00 9.52111363e-01 7.62580097e-01
3.10815841e-01 9.27761197e-01 6.09405577e-01 -4.89460349e-01
-1.98419082e+00 -8.38565290e-01 1.48272979e+00 -1.20236933e+00
5.41500688e-01 -8.18053901e-01 -1.16376781e+00 5.08930087e-01
3.56232345e-01 -2.47539416e-01 2.71707267e-01 1.16909683e-01
-7.43948162e-01 -3.89118642e-01 -1.00176835e+00 5.67518353e-01
4.81223106e-01 -9.78463590e-01 -9.63077843e-01 5.25153577e-01
1.02116334e+00 -7.92319417e-01 -9.77927685e-01 3.45148444e-01
5.95343351e-01 -7.64511645e-01 1.17495179e+00 -3.38791788e-01
5.81469476e-01 -4.54903483e-01 -2.86758065e-01 -8.81003022e-01
-9.98472329e-03 -7.10649252e-01 -3.69157672e-01 1.13079870e+00
9.26032841e-01 -2.44295821e-01 6.61573946e-01 7.11411238e-01
-6.85430542e-02 -9.03963387e-01 -1.11682415e+00 -4.10246253e-01
3.14968348e-01 -2.28018403e-01 8.23359191e-01 1.08959384e-01
-2.34318018e-01 9.70922172e-01 9.53555405e-02 6.48946986e-02
2.73653895e-01 3.68600190e-01 1.07292593e+00 -9.44066942e-01
-3.67368400e-01 -2.03258932e-01 2.52091110e-01 -1.72577143e+00
-1.58388615e-01 -9.49330926e-01 4.98630077e-01 -2.15461063e+00
1.40825018e-01 -3.31248641e-01 -1.21207781e-01 2.05800161e-01
-5.28617859e-01 -5.80463670e-02 9.04405117e-02 2.86496252e-01
-1.05228662e+00 6.68846190e-01 1.16420329e+00 -3.84133220e-01
-5.04759736e-02 -2.02333942e-01 -9.52485502e-01 -2.77557760e-01
3.09046894e-01 -7.55888700e-01 -6.86768174e-01 -7.51792848e-01
6.15536809e-01 7.76670754e-01 -6.73567876e-02 -8.06851208e-01
5.03427088e-01 3.69327247e-01 2.59097703e-02 -1.18232906e+00
4.12864715e-01 -4.78924185e-01 -4.46882367e-01 2.03159019e-01
-8.30224454e-01 5.52357554e-01 2.49691188e-01 4.48785841e-01
-4.94580239e-01 -4.96123917e-02 3.72447044e-01 -7.11419359e-02
-3.16737622e-01 1.24962784e-01 -4.48483437e-01 7.27983415e-01
2.55130768e-01 1.22989580e-01 -7.40778327e-01 -5.21804869e-01
1.54324815e-01 5.16347885e-01 1.19002439e-01 4.10656482e-01
5.61962008e-01 -9.57400322e-01 -8.48321080e-01 5.92073724e-02
6.43217146e-01 2.88583100e-01 8.43380168e-02 4.60056484e-01
-5.42639077e-01 1.03399062e+00 3.39887589e-01 -6.57127202e-01
-9.52750206e-01 5.34244895e-01 3.31204325e-01 -1.12219286e+00
-4.33038622e-01 1.00444460e+00 -1.57336611e-02 -9.41264987e-01
2.58751184e-01 -4.22414571e-01 -1.14978977e-01 -8.59887227e-02
7.24386692e-01 -3.94060090e-02 6.99865103e-01 -2.56770134e-01
-3.86249065e-01 3.38372231e-01 -5.56477666e-01 -4.21368152e-01
7.93957949e-01 -2.02647084e-03 -2.22207040e-01 2.42975116e-01
1.13158131e+00 1.04770422e-01 -6.10561430e-01 -6.04030550e-01
3.31200600e-01 -1.85593635e-01 -1.51230782e-01 -1.52687037e+00
-4.60552633e-01 7.99804747e-01 1.43658325e-01 -1.77860469e-01
1.20121145e+00 5.76293655e-02 1.33639061e+00 1.07993102e+00
4.33896184e-01 -7.72303700e-01 1.22590706e-01 9.72006977e-01
9.98426855e-01 -1.19163489e+00 -2.22899914e-01 -2.92963803e-01
-6.37254059e-01 8.99558127e-01 4.97618765e-01 1.19823851e-01
1.12722054e-01 -5.74631579e-02 4.15310413e-01 -3.75700593e-01
-1.31452847e+00 -3.36944520e-01 7.77987242e-01 3.61575127e-01
4.53575402e-01 -4.63931322e-01 -8.37955922e-02 5.10613024e-01
-5.39630473e-01 -2.06294835e-01 3.65604497e-02 1.04005599e+00
-3.89585823e-01 -1.18751001e+00 -2.78952628e-01 6.10705316e-01
-5.49195647e-01 -4.89361912e-01 -5.40006697e-01 7.21366525e-01
-7.13830352e-01 1.31845987e+00 -4.87960726e-02 -2.54976571e-01
6.60140157e-01 4.21485275e-01 2.98114479e-01 -5.76500952e-01
-1.35799873e+00 -2.04452589e-01 4.87239987e-01 -7.20695555e-01
2.42759123e-01 -3.97592902e-01 -1.11363053e+00 -1.18534654e-01
-9.07146931e-02 6.59024596e-01 2.41504326e-01 8.36227715e-01
6.39320314e-01 3.38226557e-01 2.24827394e-01 1.57930598e-01
-9.82414722e-01 -1.04097605e+00 4.67875265e-02 2.77806789e-01
3.57162535e-01 -2.47872025e-01 -4.92707603e-02 -2.59725481e-01] | [10.495390892028809, 7.879754543304443] |
8b3155d0-f312-4808-80bc-a71c411841b9 | unsupervised-speech-segmentation-and-variable | 2110.02345 | null | https://arxiv.org/abs/2110.02345v2 | https://arxiv.org/pdf/2110.02345v2.pdf | Unsupervised Speech Segmentation and Variable Rate Representation Learning using Segmental Contrastive Predictive Coding | Typically, unsupervised segmentation of speech into the phone and word-like units are treated as separate tasks and are often done via different methods which do not fully leverage the inter-dependence of the two tasks. Here, we unify them and propose a technique that can jointly perform both, showing that these two tasks indeed benefit from each other. Recent attempts employ self-supervised learning, such as contrastive predictive coding (CPC), where the next frame is predicted given past context. However, CPC only looks at the audio signal's frame-level structure. We overcome this limitation with a segmental contrastive predictive coding (SCPC) framework to model the signal structure at a higher level, e.g., phone level. A convolutional neural network learns frame-level representation from the raw waveform via noise-contrastive estimation (NCE). A differentiable boundary detector finds variable-length segments, which are then used to optimize a segment encoder via NCE to learn segment representations. The differentiable boundary detector allows us to train frame-level and segment-level encoders jointly. Experiments show that our single model outperforms existing phone and word segmentation methods on TIMIT and Buckeye datasets. We discover that phone class impacts the boundary detection performance, and the boundaries between successive vowels or semivowels are the most difficult to identify. Finally, we use SCPC to extract speech features at the segment level rather than at uniformly spaced frame level (e.g., 10 ms) and produce variable rate representations that change according to the contents of the utterance. We can lower the feature extraction rate from the typical 100 Hz to as low as 14.5 Hz on average while still outperforming the MFCC features on the linear phone classification task. | ['Najim Dehak', 'Laureano Moro-Velazquez', 'Piotr Żelasko', 'Jesús Villalba', 'Saurabhchand Bhati'] | 2021-10-05 | null | null | null | null | ['boundary-detection'] | ['computer-vision'] | [ 5.48610747e-01 -4.22949009e-02 -3.70193869e-01 -2.35486060e-01
-1.19933259e+00 -8.53852093e-01 3.44476312e-01 1.47816360e-01
-2.94034183e-01 2.26557106e-01 2.29579002e-01 -5.10601938e-01
3.04116130e-01 -3.87648493e-01 -7.87389338e-01 -6.61209643e-01
-1.41894981e-01 -6.57321587e-02 2.86452323e-01 1.03425264e-01
2.04448313e-01 2.42033720e-01 -1.68990207e+00 4.89068955e-01
9.01374996e-01 1.21104574e+00 3.75445426e-01 1.13058567e+00
-1.13869950e-01 4.55977261e-01 -5.97000420e-01 1.42325133e-01
2.05531716e-01 -6.39343262e-01 -7.54550934e-01 1.94086358e-01
2.17709929e-01 -1.27534971e-01 -9.59406048e-02 9.35930371e-01
4.08230782e-01 1.16827592e-01 5.07381320e-01 -7.26639330e-01
-2.35822603e-01 7.15117693e-01 -2.88495004e-01 4.50714380e-01
2.30077952e-01 4.72342782e-02 1.28680992e+00 -8.65141690e-01
2.30799466e-01 1.19217622e+00 7.89122939e-01 2.12773204e-01
-1.33546412e+00 -5.26526749e-01 3.01020175e-01 1.89090505e-01
-1.14017045e+00 -7.40190208e-01 9.45727050e-01 -5.09925961e-01
1.14663005e+00 2.00784057e-01 4.12122160e-01 9.55844700e-01
-2.52073891e-02 8.30843687e-01 9.15406644e-01 -7.21106648e-01
2.16092497e-01 -3.75709087e-01 1.34334192e-01 5.16345322e-01
-5.24358273e-01 2.59487182e-01 -6.02751613e-01 1.84288353e-01
6.37697458e-01 -3.21242929e-01 -4.91548538e-01 2.87426233e-01
-9.46925819e-01 8.58039021e-01 1.53224960e-01 3.76838177e-01
-2.80649602e-01 2.76997566e-01 4.14895356e-01 2.54309297e-01
5.84190667e-01 3.49745512e-01 -7.36687660e-01 -5.19518614e-01
-1.34621799e+00 -1.08988717e-01 6.71267271e-01 6.81784749e-01
7.13916004e-01 3.30626875e-01 -1.93614826e-01 1.03848279e+00
3.25937241e-01 2.45992988e-01 5.84849119e-01 -1.18514884e+00
3.36317033e-01 -1.67503551e-01 -1.30807340e-01 -4.56312358e-01
-2.47322619e-01 -5.37271023e-01 -3.28447253e-01 -1.25428677e-01
4.58205104e-01 -4.19863045e-01 -1.12344301e+00 1.81081009e+00
3.53063159e-02 5.80962598e-01 -1.28840744e-01 6.37411177e-01
3.14799130e-01 1.14005589e+00 1.06927618e-01 -5.27636468e-01
1.42394602e+00 -9.69835937e-01 -7.52888978e-01 -4.21294004e-01
5.58761358e-01 -9.18690562e-01 1.02450478e+00 4.88922805e-01
-1.32861030e+00 -9.32696283e-01 -1.15769422e+00 4.54056896e-02
-1.30063817e-01 2.62476504e-01 2.75739193e-01 6.86961710e-01
-8.49824309e-01 8.46968889e-01 -1.11694777e+00 1.32869884e-01
2.24064842e-01 2.35182047e-01 1.26290068e-01 3.89357358e-01
-1.25898862e+00 5.11899948e-01 2.96715885e-01 -2.77916826e-02
-6.10269129e-01 -7.81459033e-01 -1.06830287e+00 2.62254715e-01
1.87694147e-01 -1.01061679e-01 1.51770103e+00 -1.19734669e+00
-1.75418758e+00 5.90829074e-01 -5.76861441e-01 -8.34569871e-01
-1.15896396e-01 -2.32826412e-01 -5.33777416e-01 4.63734627e-01
-1.04028314e-01 8.01690340e-01 1.16608262e+00 -1.06482315e+00
-8.14925075e-01 1.92140058e-01 -2.65098691e-01 2.13069901e-01
-5.40244058e-02 1.66358367e-01 -5.43766856e-01 -8.42686832e-01
3.23690265e-01 -8.68183553e-01 4.44838926e-02 -4.75441694e-01
-2.92020887e-01 -2.74097294e-01 7.52344131e-01 -8.84882629e-01
1.42129302e+00 -2.38591409e+00 6.31154031e-02 -4.97741699e-02
-2.65925974e-01 3.79118472e-01 -1.09520927e-01 1.63320825e-01
-2.26596966e-01 2.92510629e-01 -4.30103362e-01 -6.08873427e-01
-8.72292072e-02 2.47688815e-01 -4.77838159e-01 3.48356277e-01
5.46278000e-01 8.36771488e-01 -7.49340236e-01 -3.80586833e-01
1.10400468e-01 5.78951657e-01 -7.37144947e-01 1.25079721e-01
-1.70592591e-01 4.64844137e-01 2.12743193e-01 4.54893768e-01
4.04628366e-01 2.24277332e-01 1.57741867e-02 -2.65413553e-01
-2.77025908e-01 1.00708508e+00 -9.43329632e-01 1.57748163e+00
-5.82522333e-01 9.39291418e-01 1.90251410e-01 -1.26025820e+00
8.27823222e-01 6.26964748e-01 4.21503037e-01 -4.98843729e-01
-3.93199474e-02 3.78853828e-01 2.37557247e-01 -3.33505780e-01
3.33326042e-01 -3.27782393e-01 -3.14271599e-02 1.25347540e-01
2.99245447e-01 -3.02294523e-01 3.56641933e-02 -3.20847243e-01
9.96669948e-01 1.65948883e-01 1.12313934e-01 -2.23071463e-02
3.82283539e-01 -2.93131322e-01 8.78825724e-01 6.78470433e-01
-4.35056329e-01 9.25776303e-01 5.16319692e-01 1.02668978e-01
-8.20693851e-01 -1.21353376e+00 -2.89439201e-01 1.37955272e+00
-1.22207880e-01 -4.49257016e-01 -9.47814405e-01 -4.89066541e-01
-2.37949997e-01 7.37047493e-01 -1.67876258e-01 -1.02378093e-01
-1.00250483e+00 -3.87737393e-01 6.54420435e-01 6.13896906e-01
8.78569484e-02 -1.04877913e+00 -4.32731420e-01 5.02066910e-01
-3.22833568e-01 -1.37459254e+00 -7.57160962e-01 7.38806725e-01
-7.93469131e-01 -5.39645135e-01 -5.18694043e-01 -9.81215954e-01
1.07978515e-01 2.21301895e-03 9.91373718e-01 -1.49131238e-01
4.12156805e-02 3.39454055e-01 -6.45196736e-01 -2.97481149e-01
-5.98418474e-01 2.00718001e-01 -4.95248027e-02 1.55945018e-01
2.19393566e-01 -6.96689367e-01 -5.50605178e-01 1.48609325e-01
-7.13969767e-01 -1.56811431e-01 4.27966595e-01 7.83448935e-01
7.81622648e-01 1.61308691e-01 8.58200908e-01 -7.16186166e-01
4.32460606e-01 -3.83502752e-01 -4.25452948e-01 -1.51714191e-01
-3.65508080e-01 -1.28461033e-01 6.73021436e-01 -6.57481670e-01
-9.46649373e-01 2.23263010e-01 -6.50642931e-01 -5.62734127e-01
-3.38907540e-01 5.31310081e-01 -3.89324352e-02 4.07142311e-01
3.79281938e-01 2.85973579e-01 -1.01541840e-01 -6.02209449e-01
3.83638561e-01 8.28937232e-01 8.50903094e-01 -3.57883036e-01
6.62122786e-01 1.24777228e-01 -6.36029959e-01 -9.26848590e-01
-9.08362031e-01 -5.88871837e-01 -7.09498763e-01 -3.71028669e-02
9.83604312e-01 -9.23949301e-01 -3.38217586e-01 3.06959659e-01
-1.28354180e+00 -4.69613791e-01 -4.62611914e-01 6.83998823e-01
-7.12928295e-01 4.71374482e-01 -1.04823518e+00 -9.33633566e-01
-1.86855823e-01 -1.17686331e+00 1.21541679e+00 2.16746628e-01
-2.86046833e-01 -8.58668149e-01 -1.56699553e-01 2.62467951e-01
2.41688266e-01 -4.78042550e-02 8.84022593e-01 -6.70150995e-01
-3.12562019e-01 4.87185158e-02 2.30411917e-01 6.27556682e-01
2.28534579e-01 2.63338387e-01 -1.36558592e+00 -1.42734751e-01
3.08934718e-01 5.36325909e-02 1.09833407e+00 7.02066720e-01
1.29574597e+00 -2.61379957e-01 -1.77382350e-01 7.20602989e-01
9.84250128e-01 5.66208422e-01 6.03274465e-01 -1.12260610e-01
4.97474432e-01 5.89540601e-01 5.14626324e-01 1.51836127e-01
3.10468376e-02 5.96857846e-01 1.03118353e-01 -5.19118086e-02
-4.45203215e-01 -2.26815030e-01 7.04406261e-01 1.38370967e+00
2.06391081e-01 -1.01853773e-01 -7.87864864e-01 6.24715149e-01
-1.34303856e+00 -9.44500327e-01 -8.47046226e-02 2.09431171e+00
1.27147186e+00 5.79352200e-01 7.30206817e-02 5.90764403e-01
7.96942830e-01 2.89897889e-01 -4.67321664e-01 -8.09024215e-01
4.01050458e-03 6.11063242e-01 4.24956888e-01 7.86620319e-01
-1.22858155e+00 1.03002703e+00 6.39993715e+00 1.07242632e+00
-1.26266038e+00 1.56977504e-01 7.38338113e-01 4.10807841e-02
-2.82749802e-01 1.01746865e-01 -8.27132165e-01 6.13654673e-01
1.44444919e+00 2.89459586e-01 5.48672259e-01 6.85066223e-01
4.15475011e-01 -7.24618584e-02 -1.14619660e+00 8.72723639e-01
-1.84903368e-01 -1.19599831e+00 -5.02905011e-01 -9.89092048e-04
5.66047013e-01 4.26703095e-02 2.19452292e-01 4.88203943e-01
-1.67088836e-01 -1.08045280e+00 9.94963825e-01 2.21631169e-01
8.47802401e-01 -6.83378220e-01 3.24173093e-01 3.50061297e-01
-1.54014552e+00 -1.01289168e-01 -1.42180547e-01 -6.44718483e-02
2.83959478e-01 5.63091397e-01 -8.74855697e-01 2.05417901e-01
4.61796463e-01 5.66524208e-01 -1.76859230e-01 8.32859457e-01
-2.14290917e-01 1.50136054e+00 -3.31567794e-01 4.41647172e-01
3.08026493e-01 5.29360585e-03 5.18857479e-01 1.66573060e+00
2.49349326e-01 -3.77603769e-02 1.82958931e-01 7.71347165e-01
4.81033102e-02 -1.54491782e-01 -7.50156119e-02 -1.74091607e-01
7.91602671e-01 9.80427384e-01 -7.74133921e-01 -1.65450647e-01
-5.48319995e-01 9.91700292e-01 9.89960134e-02 5.18007219e-01
-9.48161244e-01 -4.96316403e-01 8.35189342e-01 -1.33284301e-01
8.29552114e-01 -5.14925659e-01 -4.19764251e-01 -8.66711557e-01
-1.32889211e-01 -8.61862302e-01 6.29763082e-02 -4.04728293e-01
-1.03337967e+00 4.95656729e-01 -2.66095668e-01 -1.21649885e+00
-6.55723751e-01 -4.92777854e-01 -8.19549739e-01 1.00292969e+00
-1.80960584e+00 -6.75658047e-01 3.01788628e-01 2.78265744e-01
1.06264114e+00 3.26363593e-01 7.38765538e-01 1.99852690e-01
-5.87253928e-01 6.37004912e-01 4.97830287e-02 3.81862491e-01
4.92732167e-01 -1.31269360e+00 6.60479188e-01 9.87520456e-01
4.56618249e-01 3.40403318e-01 5.33071756e-01 -4.04006392e-01
-9.41852272e-01 -9.70887899e-01 1.06988120e+00 -2.12589607e-01
5.60066521e-01 -5.44703960e-01 -1.16898644e+00 5.29195786e-01
1.06364660e-01 1.72607839e-01 7.18520463e-01 2.56603416e-02
-3.11125845e-01 -5.61002567e-02 -6.06266797e-01 4.37804490e-01
7.86711514e-01 -1.05395114e+00 -8.09046924e-01 -2.57692747e-02
1.24949944e+00 -3.57097089e-01 -7.46104181e-01 3.40882838e-01
3.76976818e-01 -1.01190937e+00 9.26253557e-01 -2.55745858e-01
2.16145769e-01 -1.96369246e-01 -2.98341751e-01 -1.26015687e+00
-1.34813190e-01 -9.58468437e-01 -3.08550328e-01 1.47986782e+00
6.25625312e-01 -4.01761562e-01 5.11092544e-01 9.94444713e-02
-6.13438070e-01 -7.91798890e-01 -1.05793607e+00 -9.47084188e-01
1.92685142e-01 -9.59192634e-01 2.58337051e-01 5.75215816e-01
1.02120107e-02 3.83526832e-01 -7.98682496e-02 2.83981293e-01
1.33529916e-01 1.30738482e-01 -1.20128021e-02 -1.01541686e+00
-5.65110683e-01 -5.58087289e-01 -9.17892009e-02 -1.54207397e+00
2.63019800e-01 -7.96111345e-01 6.54426217e-01 -1.08057833e+00
-5.23510158e-01 -3.42380285e-01 -3.98301959e-01 2.58439451e-01
-3.11336905e-01 1.59872353e-01 2.68817961e-01 2.43423358e-01
-1.80992514e-01 2.98422843e-01 9.37848926e-01 -7.21869096e-02
-5.68982184e-01 2.00408176e-01 -3.59347165e-01 9.89997804e-01
8.49986196e-01 -3.46250862e-01 -3.31037879e-01 -2.88108319e-01
-3.21359187e-01 2.57579863e-01 4.19655107e-02 -1.12583613e+00
1.47063419e-01 9.42606479e-02 2.36131668e-01 -6.94928467e-01
5.25486052e-01 -4.06344473e-01 -3.35315734e-01 3.30516785e-01
-5.62760651e-01 -3.08041036e-01 1.60749048e-01 5.28235734e-01
-4.01445448e-01 -4.92909998e-01 8.94544482e-01 1.00178838e-01
-5.47629356e-01 2.08272152e-02 -7.64717102e-01 1.50934830e-01
4.91475254e-01 -3.79385918e-01 1.27094626e-01 -3.72435480e-01
-8.19987774e-01 -1.39990821e-01 -6.19227365e-02 3.10080856e-01
3.93076360e-01 -1.11160493e+00 -4.88510340e-01 3.97338599e-01
-3.26196730e-01 -8.26933980e-02 5.33319712e-02 8.80456448e-01
-1.32811606e-01 4.16611463e-01 3.32943052e-01 -9.08577740e-01
-9.33988631e-01 3.94517869e-01 4.13796157e-01 -1.42247498e-01
-4.16360199e-01 1.14234614e+00 1.83912814e-01 8.01200233e-03
4.83109355e-01 -7.63504922e-01 -2.26832837e-01 3.34680229e-01
3.81544203e-01 7.68648908e-02 1.21766314e-01 -8.30630064e-01
-1.95249990e-01 6.00013912e-01 7.46350214e-02 -2.64018178e-01
1.09824944e+00 -4.38688487e-01 3.24420184e-01 8.28158915e-01
1.55055594e+00 1.86925247e-01 -1.82683563e+00 -2.00155020e-01
1.97024882e-01 -1.05141342e-01 2.59913266e-01 -5.05213380e-01
-7.60345399e-01 1.15754580e+00 4.72622514e-01 5.64171851e-01
1.33587348e+00 2.10167444e-03 1.20561790e+00 -1.48256570e-01
-2.03017533e-01 -1.31297243e+00 1.05369762e-01 7.99053311e-01
4.58611786e-01 -9.52700615e-01 -4.68645066e-01 -5.75590551e-01
-4.99214113e-01 1.22849751e+00 1.62274837e-01 -1.46160424e-01
6.84691489e-01 5.31970918e-01 -9.54529718e-02 3.03830594e-01
-7.38541901e-01 -6.39478624e-01 4.86245960e-01 5.69305480e-01
6.94779158e-01 1.10822968e-01 -2.46255826e-02 7.36696422e-01
-4.06819880e-01 -5.03821015e-01 2.58863211e-01 7.56372333e-01
-8.44626427e-01 -1.14671028e+00 -5.36275387e-01 3.47319633e-01
-7.39733219e-01 -4.06632245e-01 -8.84844288e-02 2.58479446e-01
2.59320170e-01 1.40005744e+00 4.61038530e-01 -4.63569582e-01
8.57540444e-02 5.49625695e-01 2.15303943e-01 -7.96174765e-01
-6.30604208e-01 8.01689208e-01 6.79864660e-02 -4.22752708e-01
-3.56385052e-01 -8.26064408e-01 -1.48443937e+00 2.12045893e-01
-5.18775463e-01 1.72123969e-01 6.49521410e-01 1.11273408e+00
1.08586542e-01 6.88995063e-01 8.95534098e-01 -1.02172101e+00
-4.04799372e-01 -8.71846974e-01 -4.55870450e-01 1.96973622e-01
6.79824412e-01 -3.01248521e-01 -5.61569333e-01 5.08468747e-01] | [14.5855712890625, 6.6071624755859375] |
d2b09dfb-f78a-487c-8761-a86a788de4d5 | componerf-text-guided-multi-object | 2303.13843 | null | https://arxiv.org/abs/2303.13843v1 | https://arxiv.org/pdf/2303.13843v1.pdf | CompoNeRF: Text-guided Multi-object Compositional NeRF with Editable 3D Scene Layout | Recent research endeavors have shown that combining neural radiance fields (NeRFs) with pre-trained diffusion models holds great potential for text-to-3D generation.However, a hurdle is that they often encounter guidance collapse when rendering complex scenes from multi-object texts. Because the text-to-image diffusion models are inherently unconstrained, making them less competent to accurately associate object semantics with specific 3D structures. To address this issue, we propose a novel framework, dubbed CompoNeRF, that explicitly incorporates an editable 3D scene layout to provide effective guidance at the single object (i.e., local) and whole scene (i.e., global) levels. Firstly, we interpret the multi-object text as an editable 3D scene layout containing multiple local NeRFs associated with the object-specific 3D box coordinates and text prompt, which can be easily collected from users. Then, we introduce a global MLP to calibrate the compositional latent features from local NeRFs, which surprisingly improves the view consistency across different local NeRFs. Lastly, we apply the text guidance on global and local levels through their corresponding views to avoid guidance ambiguity. This way, our CompoNeRF allows for flexible scene editing and re-composition of trained local NeRFs into a new scene by manipulating the 3D layout or text prompt. Leveraging the open-source Stable Diffusion model, our CompoNeRF can generate faithful and editable text-to-3D results while opening a potential direction for text-guided multi-object composition via the editable 3D scene layout. | ['Lin Wang', 'Hui Xiong', 'Xiaodong Lin', 'Haonan Lu', 'Sijia Li', 'Haotian Bai', 'Yiqi Lin'] | 2023-03-24 | null | null | null | null | ['text-to-3d'] | ['computer-vision'] | [ 3.28915060e-01 -9.66369063e-02 3.36571246e-01 -4.92970824e-01
-6.03191495e-01 -9.37378705e-01 8.35001707e-01 -1.19654834e-01
1.35166243e-01 1.22270115e-01 3.80710036e-01 -2.21005172e-01
-1.20902494e-01 -1.17325699e+00 -8.85501802e-01 -5.24395764e-01
4.98456240e-01 2.77353108e-01 1.92259654e-01 -3.11268359e-01
3.32853258e-01 9.76226985e-01 -1.54062402e+00 3.68084043e-01
1.07521963e+00 7.74082065e-01 6.28154516e-01 8.72111022e-01
-5.16085386e-01 6.02886260e-01 -6.85350716e-01 -2.22407088e-01
4.59020346e-01 -2.09742412e-01 -4.13570642e-01 2.73084819e-01
1.12869227e+00 -7.99936831e-01 -1.37380719e-01 8.48457456e-01
4.87110227e-01 3.70675445e-01 6.98798120e-01 -1.01727939e+00
-1.15153611e+00 2.93674469e-02 -5.27468920e-01 -3.57767165e-01
4.47974503e-01 4.40003723e-01 7.95922935e-01 -8.85636806e-01
8.04476440e-01 1.56597745e+00 4.04474705e-01 4.15107548e-01
-1.44132602e+00 -3.51397067e-01 5.37822068e-01 -4.33857590e-01
-1.07085395e+00 -2.24411219e-01 1.05463970e+00 -6.44492924e-01
1.05038011e+00 5.86971998e-01 6.26126230e-01 1.16743279e+00
3.33143234e-01 4.17798936e-01 1.02979374e+00 -2.57007360e-01
2.19277442e-01 1.50248315e-02 -2.54778028e-01 7.62069404e-01
-2.12262660e-01 1.07832827e-01 -5.98167300e-01 -2.38440167e-02
1.44043946e+00 1.03066459e-01 -2.60247707e-01 -5.75450361e-01
-1.31395090e+00 6.93879068e-01 6.24530852e-01 -2.90657654e-02
-2.27278292e-01 3.19463611e-01 8.38965625e-02 -2.48029493e-02
7.59980500e-01 5.50911427e-01 -5.05799353e-01 6.32288754e-02
-5.70246816e-01 3.89489621e-01 4.24675465e-01 1.22186124e+00
8.61353397e-01 1.68731034e-01 -3.60298008e-01 9.48314130e-01
3.50300223e-01 8.14932346e-01 -5.66840693e-02 -1.30398715e+00
6.69279754e-01 6.55423880e-01 6.10438175e-02 -1.31767201e+00
-3.33967716e-01 -1.19233415e-01 -8.46867502e-01 6.58033133e-01
1.05118878e-01 -7.89307710e-03 -1.05008006e+00 1.62009633e+00
5.53108335e-01 -3.10453504e-01 -2.08485186e-01 8.57572794e-01
9.19188023e-01 9.07232881e-01 -1.41408116e-01 3.30970615e-01
1.16758478e+00 -1.07699788e+00 -5.36710083e-01 -2.54139245e-01
6.25371397e-01 -8.08017612e-01 1.54163599e+00 3.98187153e-02
-9.62399900e-01 -6.96898162e-01 -7.86676168e-01 -5.80001295e-01
-4.05101836e-01 5.13425395e-02 6.54814720e-01 2.49283433e-01
-1.09358501e+00 3.75087112e-01 -6.01177633e-01 -3.07876587e-01
2.17679486e-01 -6.21194132e-02 -2.61410832e-01 -2.54879951e-01
-7.60105193e-01 6.58905983e-01 1.27610877e-01 4.91854362e-02
-7.79372513e-01 -8.92119050e-01 -1.01082325e+00 -1.59410954e-01
4.77019042e-01 -1.14895463e+00 1.11813581e+00 -7.16749787e-01
-1.64844918e+00 6.47222817e-01 -1.74061075e-01 3.68459433e-01
5.32982171e-01 -2.97052026e-01 -2.48422921e-02 3.51014845e-02
2.86421686e-01 1.07829058e+00 9.78098154e-01 -1.70712483e+00
-2.48119593e-01 -1.70277953e-01 5.47279298e-01 5.73219121e-01
-1.32634580e-01 -2.75063962e-01 -4.75142598e-01 -1.08868992e+00
3.24014455e-01 -5.50759614e-01 -2.51114875e-01 6.88318789e-01
-5.31836689e-01 -2.12776400e-02 1.10703683e+00 -4.61595803e-01
8.80192816e-01 -2.33142424e+00 2.59883821e-01 1.15123712e-01
2.38932863e-01 -3.03556383e-01 -4.11375761e-01 3.44222158e-01
-1.35461539e-01 3.74313205e-01 -1.93447933e-01 -4.02886957e-01
-1.53415352e-02 1.37729391e-01 -3.96503955e-01 -4.41699959e-02
3.06308568e-01 1.08403349e+00 -9.70546603e-01 -7.50167370e-02
6.09341800e-01 6.15941703e-01 -9.01836991e-01 2.69781649e-01
-6.99210048e-01 6.31475687e-01 -6.40909672e-01 4.37106341e-01
8.45563233e-01 -3.08498114e-01 -2.71241695e-01 -5.82504869e-01
-2.56018877e-01 2.34560981e-01 -1.20220387e+00 2.03620553e+00
-7.58608997e-01 6.44063473e-01 -2.74563041e-02 -2.77553290e-01
9.99972939e-01 -1.98931754e-01 2.76071012e-01 -7.37199783e-01
-2.79555291e-01 -1.56541467e-02 -5.73811710e-01 -4.00963396e-01
1.02983177e+00 7.35559911e-02 -4.38166149e-02 6.63516521e-01
-2.10329875e-01 -9.53559577e-01 -2.26052299e-01 4.92798477e-01
8.41807604e-01 6.54194295e-01 -1.23512290e-01 -1.37947202e-01
7.90242627e-02 -1.14613689e-01 4.63007763e-02 8.59309614e-01
4.18788582e-01 1.10533619e+00 2.63092697e-01 -3.35707396e-01
-1.34260309e+00 -1.31105018e+00 -2.70971246e-02 9.56920505e-01
2.94236511e-01 -4.98557210e-01 -5.87520838e-01 -7.58494675e-01
-7.97803178e-02 9.77633238e-01 -5.40771365e-01 1.52350396e-01
-3.83520395e-01 -2.99501956e-01 2.12311074e-01 4.80642915e-01
5.96439302e-01 -8.63981664e-01 -5.77785432e-01 1.77802145e-01
-1.65145025e-01 -9.77081478e-01 -9.10423398e-01 1.15289226e-01
-9.14629817e-01 -5.46299994e-01 -8.35743070e-01 -4.62794632e-01
1.07992256e+00 6.30118072e-01 1.02593160e+00 -2.91411635e-02
-2.33345062e-01 6.34920001e-01 -3.23124349e-01 1.21551473e-02
-6.04936659e-01 -1.09208375e-01 -2.60046512e-01 -1.51337028e-01
-3.57676834e-01 -5.72717369e-01 -6.28866613e-01 3.89055014e-01
-1.29283726e+00 8.91777694e-01 1.90701008e-01 4.32243079e-01
6.61095560e-01 -5.09291538e-04 3.57863642e-02 -7.75396109e-01
4.37779188e-01 -6.23059869e-02 -6.26306832e-01 1.74078465e-01
-1.46002293e-01 2.55573958e-01 6.35747015e-01 -4.64842796e-01
-1.46278656e+00 -6.04710095e-02 -1.43339247e-01 -3.05347919e-01
-3.65979284e-01 2.43497476e-01 -3.22184443e-01 -2.67311446e-02
8.47276032e-01 4.69554812e-02 -4.09335196e-01 -4.86546636e-01
1.00186360e+00 3.92439306e-01 4.17127162e-01 -8.26288462e-01
1.02650356e+00 5.80151141e-01 -1.56953812e-01 -7.26504266e-01
-9.48405802e-01 -7.23382235e-02 -7.05993116e-01 -3.36645156e-01
1.13040698e+00 -7.71576047e-01 -2.35813469e-01 7.63440371e-01
-1.58691895e+00 -8.54203403e-01 -4.59294081e-01 -2.74770577e-02
-5.46177447e-01 3.01760197e-01 -5.67290008e-01 -4.50485140e-01
2.23262906e-02 -1.17206264e+00 1.65033972e+00 1.96760774e-01
-3.19760501e-01 -1.07977104e+00 -2.49820754e-01 2.37350345e-01
4.34182018e-01 3.45122635e-01 1.31222689e+00 2.20494136e-01
-1.24056935e+00 2.22683653e-01 -5.00135839e-01 3.60043228e-01
4.49370503e-01 3.43100309e-01 -9.82529640e-01 5.42134279e-03
-8.05991516e-02 -1.61833912e-01 3.84883195e-01 3.30692768e-01
1.27890527e+00 -2.43104681e-01 -1.51971444e-01 8.61117542e-01
1.30325770e+00 9.00133029e-02 5.52060723e-01 3.47400725e-01
1.26124060e+00 7.28059888e-01 3.30738813e-01 3.35339218e-01
5.37759960e-01 8.22283030e-01 3.36598814e-01 -3.60417068e-01
-4.91732180e-01 -7.16197431e-01 3.03038269e-01 8.98358643e-01
1.11392692e-01 -5.92121243e-01 -5.65440178e-01 1.18673041e-01
-1.65428102e+00 -7.84771025e-01 -2.74258912e-01 1.95721591e+00
7.44163156e-01 -2.56561842e-02 -5.18395662e-01 -4.72799599e-01
4.21582818e-01 5.22106886e-01 -6.84744596e-01 -4.35309172e-01
-3.40086371e-01 -2.21699536e-01 3.54702324e-01 8.01064014e-01
-6.61180019e-01 1.00371754e+00 5.60534811e+00 8.47334683e-01
-1.21381950e+00 -2.89447568e-02 6.63027525e-01 -1.35753274e-01
-1.16980052e+00 2.52645966e-02 -8.26980591e-01 3.62230092e-02
2.05523074e-01 1.02434568e-01 6.59113407e-01 7.02505112e-01
5.81756890e-01 -2.84126431e-01 -1.10893357e+00 9.86327112e-01
1.38695762e-01 -1.47616947e+00 5.83093405e-01 1.64807796e-01
1.11963570e+00 -1.46090373e-01 1.09234095e-01 -8.09467137e-02
4.03231859e-01 -8.08074772e-01 1.16124058e+00 7.99216568e-01
1.24459004e+00 -3.89406830e-01 -1.01146542e-01 4.08723056e-01
-1.15652871e+00 2.67035931e-01 -2.26618171e-01 2.25820497e-01
5.07346869e-01 7.38247514e-01 -4.56662863e-01 6.25345588e-01
8.35006714e-01 7.33847439e-01 -6.43627405e-01 6.19390965e-01
-3.48993093e-01 -5.43179438e-02 -4.53052193e-01 1.61542758e-01
1.53611362e-01 -3.26807886e-01 6.56002581e-01 9.26288545e-01
6.01852298e-01 -7.27729872e-03 1.71430819e-02 1.49412286e+00
4.23594676e-02 -1.96718853e-02 -9.39378917e-01 8.77316669e-02
3.39425921e-01 1.38034093e+00 -7.65953064e-01 -2.87833184e-01
-2.35192895e-01 1.44856822e+00 3.02156180e-01 8.38695765e-01
-8.34405005e-01 -3.83443296e-01 4.78968173e-01 1.84633672e-01
2.52071291e-01 -6.40725791e-01 -5.79975903e-01 -1.22134483e+00
1.43428892e-01 -7.30148315e-01 -2.87802309e-01 -1.67967200e+00
-1.45067835e+00 5.93251467e-01 2.27790415e-01 -1.04434097e+00
2.20164452e-02 -6.22493088e-01 -4.65191782e-01 1.12645292e+00
-1.32438755e+00 -1.29787719e+00 -5.68539023e-01 4.44998562e-01
6.92908406e-01 3.47816288e-01 6.85641885e-01 8.30358565e-02
-1.58103988e-01 2.68790632e-01 -4.85036196e-03 -1.72827080e-01
6.98351324e-01 -1.23414946e+00 8.48519802e-01 7.14715660e-01
1.95646420e-01 6.53788567e-01 2.95259446e-01 -8.57014060e-01
-1.19759703e+00 -1.21099985e+00 5.62918603e-01 -9.57089841e-01
2.85964489e-01 -7.68993258e-01 -9.33866501e-01 5.09187222e-01
-5.77579311e-04 -2.86834031e-01 7.66225308e-02 -1.52599245e-01
-2.78979391e-01 -2.02016085e-02 -9.57172155e-01 1.20347035e+00
1.51144075e+00 -9.22141492e-01 -2.09097102e-01 3.38847458e-01
1.24435031e+00 -8.77442718e-01 -7.73288190e-01 3.15211415e-01
2.21139103e-01 -1.08480251e+00 1.03774416e+00 -6.36409596e-02
7.23252833e-01 -6.33050680e-01 -2.52765715e-01 -1.42480624e+00
-4.01558757e-01 -5.83924770e-01 8.13055187e-02 1.30076408e+00
3.40850770e-01 -4.57094640e-01 4.57134336e-01 8.36291492e-01
-3.99308294e-01 -5.60004234e-01 -5.44691265e-01 -3.58146757e-01
-5.49176782e-02 -7.48339057e-01 6.47873521e-01 1.01288533e+00
-7.56667733e-01 3.67311746e-01 -2.22872153e-01 1.61545560e-01
3.89945924e-01 1.63928956e-01 1.09028029e+00 -8.86530578e-01
-3.22535843e-01 -4.45993245e-01 7.11821094e-02 -1.91123021e+00
-2.22646102e-01 -9.65676010e-01 2.36966565e-01 -1.82486355e+00
-2.75989366e-03 -7.90377796e-01 4.50671911e-01 4.38600600e-01
-1.57858595e-01 6.81875050e-02 3.30864221e-01 1.13622569e-01
-1.91425934e-01 8.01843703e-01 1.92485011e+00 -1.33905545e-01
-3.84744018e-01 -4.16422814e-01 -7.26256788e-01 6.92529976e-01
5.64339161e-01 -3.34626466e-01 -5.59470475e-01 -9.96743500e-01
4.32776600e-01 -1.74638018e-01 7.15139866e-01 -6.21260345e-01
-1.17972400e-02 -5.03068507e-01 4.89716619e-01 -8.62133563e-01
3.71733010e-01 -7.29418814e-01 2.53653556e-01 -3.99804056e-01
-3.28138083e-01 1.31389827e-01 3.32563013e-01 5.65232635e-01
2.71967560e-01 -1.69647917e-01 4.80198145e-01 -2.25840256e-01
-6.88593984e-01 4.76684421e-01 -1.79057032e-01 -1.36236355e-01
7.97348201e-01 -4.21539307e-01 -5.50146759e-01 -4.87714589e-01
-4.79944587e-01 1.62217915e-02 9.39804733e-01 7.11399972e-01
6.79795921e-01 -1.49345112e+00 -3.54338616e-01 5.49973071e-01
2.35830337e-01 8.37480605e-01 6.42212033e-01 3.24703991e-01
-6.27035618e-01 1.34069070e-01 1.06986597e-01 -1.00013530e+00
-1.01602209e+00 2.60718644e-01 4.33005244e-01 9.13169086e-02
-9.57388520e-01 8.75489950e-01 8.69952381e-01 -1.04403114e+00
-4.14960943e-02 -7.12353647e-01 4.22345579e-01 -3.12480092e-01
2.89874613e-01 -5.28850667e-02 -1.51277706e-01 -4.65151668e-01
1.89546674e-01 1.01275218e+00 7.58545846e-02 -3.91528904e-01
1.24365819e+00 -4.23406810e-01 -1.09505981e-01 4.61749285e-01
1.19552982e+00 5.50071716e-01 -1.93932354e+00 -8.54993165e-02
-5.90728223e-01 -7.39998221e-01 9.58677083e-02 -9.59678769e-01
-8.37106466e-01 9.84138012e-01 3.54196280e-01 8.80980045e-02
8.42311502e-01 1.85234603e-02 6.01207137e-01 3.07073414e-01
2.86822051e-01 -9.00183380e-01 5.01845062e-01 6.81978643e-01
1.31927419e+00 -9.63829875e-01 -9.81701836e-02 -5.54868460e-01
-6.09493971e-01 1.28060985e+00 8.22742522e-01 2.52426982e-01
3.66922110e-01 2.68924087e-01 3.60796511e-01 -3.59821230e-01
-5.50168633e-01 2.61426091e-01 6.24099851e-01 8.32786441e-01
2.91968584e-01 -1.10803053e-01 5.37991405e-01 4.29153629e-02
-2.78979570e-01 -3.12844932e-01 4.23515052e-01 7.27226734e-01
-2.02750832e-01 -1.10420215e+00 -3.86837959e-01 2.77027637e-01
2.87559330e-01 -3.48906547e-01 -2.30920270e-01 4.60922807e-01
1.55330002e-01 6.29282475e-01 1.42332643e-01 -2.61433631e-01
5.23459613e-01 -2.32812151e-01 5.79725206e-01 -8.84396195e-01
-3.15313548e-01 1.28259182e-01 -7.11458847e-02 -6.92713678e-01
-3.40908855e-01 -2.81746656e-01 -1.20261168e+00 -3.00668091e-01
-2.17558444e-01 -4.87963736e-01 6.80831492e-01 7.79798329e-01
6.04577661e-01 7.42137790e-01 6.28096104e-01 -1.30556107e+00
-1.15602821e-01 -6.30744040e-01 -4.17354256e-01 5.28320909e-01
3.00693661e-01 -5.35837054e-01 -3.12405676e-01 2.36876577e-01] | [9.331156730651855, -3.1705856323242188] |
21a95ee8-aabc-439f-a183-5644b1dcf6cf | on-the-benefits-of-self-taught-learning-for | 2209.10099 | null | https://arxiv.org/abs/2209.10099v4 | https://arxiv.org/pdf/2209.10099v4.pdf | On the benefits of self-taught learning for brain decoding | Context. We study the benefits of using a large public neuroimaging database composed of fMRI statistic maps, in a self-taught learning framework, for improving brain decoding on new tasks. First, we leverage the NeuroVault database to train, on a selection of relevant statistic maps, a convolutional autoencoder to reconstruct these maps. Then, we use this trained encoder to initialize a supervised convolutional neural network to classify tasks or cognitive processes of unseen statistic maps from large collections of the NeuroVault database. Results. We show that such a self-taught learning process always improves the performance of the classifiers but the magnitude of the benefits strongly depends on the number of samples available both for pre-training and finetuning the models and on the complexity of the targeted downstream task. Conclusion. The pre-trained model improves the classification performance and displays more generalizable features, less sensitive to individual differences. | ['Camille Maumet', 'Elisa Fromont', 'Elodie Germani'] | 2022-09-19 | null | null | null | null | ['brain-decoding', 'brain-decoding'] | ['medical', 'miscellaneous'] | [ 3.63512158e-01 1.61446184e-01 1.05140492e-01 -7.74880707e-01
-5.29751778e-01 -6.16451561e-01 6.34936333e-01 3.11054438e-01
-7.57908642e-01 6.12764716e-01 4.52346712e-01 -1.13099232e-01
-4.43744659e-01 -4.62453127e-01 -1.15660059e+00 -5.36815584e-01
-2.54711121e-01 7.10872054e-01 1.36499390e-01 2.08794907e-01
8.22466984e-02 3.91394794e-01 -1.38183391e+00 4.84377325e-01
9.25946712e-01 9.13124502e-01 3.84950221e-01 3.93999368e-01
3.60530555e-01 5.05143166e-01 -2.68239886e-01 -4.24762189e-01
1.80354521e-01 -3.64975601e-01 -6.70444250e-01 -2.39532560e-01
5.85911870e-01 -4.01899636e-01 -5.52684128e-01 8.83572459e-01
5.08977473e-01 2.15725079e-01 9.27053750e-01 -6.21471465e-01
-4.30845886e-01 1.02249837e+00 4.66389805e-02 1.02951860e+00
-3.60125214e-01 6.38860345e-01 8.95019889e-01 -4.31237280e-01
5.67607462e-01 9.12662566e-01 6.90171123e-01 4.17951167e-01
-1.53290880e+00 -8.10503542e-01 -7.33354762e-02 7.18158185e-02
-8.84525061e-01 -7.04061329e-01 2.14424774e-01 -8.52837682e-01
7.26080060e-01 -1.08758882e-01 6.34256899e-01 1.48528409e+00
4.39276367e-01 1.75705597e-01 1.43877578e+00 1.98564872e-01
4.63617295e-01 1.18289076e-01 5.47255218e-01 7.20385969e-01
3.15270364e-01 2.76796877e-01 -4.83268201e-01 -8.48127529e-02
7.65009224e-01 -1.76961198e-01 -1.35139599e-01 -2.94554979e-01
-1.33020937e+00 5.46334326e-01 7.24891424e-01 3.44369531e-01
-6.80951715e-01 2.03249440e-01 6.40517414e-01 1.60349593e-01
6.15359843e-01 9.42481220e-01 -7.77946830e-01 7.97731727e-02
-1.06552076e+00 8.80572423e-02 4.41778362e-01 7.90829360e-02
8.65345061e-01 8.30434561e-02 -5.62263846e-01 6.92230403e-01
-1.93084970e-01 3.96333903e-01 7.71402419e-01 -9.06403780e-01
2.26714090e-01 3.70047033e-01 -5.79311192e-01 -5.50884604e-01
-7.18987525e-01 -7.40831852e-01 -4.81610835e-01 -1.20498590e-01
3.65487933e-01 -4.85681415e-01 -8.59990954e-01 1.84316289e+00
-2.06818596e-01 -2.59545147e-02 -2.43660584e-01 7.39570975e-01
7.43077934e-01 2.41907418e-01 4.48242605e-01 2.08682626e-01
1.23581731e+00 -5.54574668e-01 -2.29957491e-01 -5.33684313e-01
5.61746895e-01 1.47732884e-01 9.78037179e-01 3.22400957e-01
-8.84332418e-01 -6.82608902e-01 -8.60169709e-01 -6.80963546e-02
-4.61864084e-01 1.84751764e-01 9.12793279e-01 4.49555218e-01
-1.24949121e+00 8.40831995e-01 -9.53536808e-01 -1.45724565e-01
1.15882659e+00 6.06623054e-01 -4.96991038e-01 -1.76290050e-01
-8.45677853e-01 1.01860178e+00 7.03614354e-01 -1.97281271e-01
-1.46037626e+00 -1.03342235e+00 -6.28768802e-01 6.74997270e-01
1.03331059e-01 -7.31766403e-01 9.64403093e-01 -1.20233166e+00
-1.15459168e+00 8.80878687e-01 8.86183325e-03 -5.47768831e-01
2.41230920e-01 1.32180423e-01 -1.86591610e-01 1.26413658e-01
2.05574453e-01 9.27598476e-01 7.64907479e-01 -7.36995041e-01
-1.64243206e-01 -6.65086210e-01 -2.51578540e-01 -3.52284685e-02
-2.25971773e-01 -8.61272886e-02 -9.31133479e-02 -3.49114358e-01
-2.19642848e-01 -6.13827825e-01 -1.11187659e-01 -1.07671887e-01
-1.36587441e-01 -1.81064650e-01 2.01714203e-01 -8.06497812e-01
4.80420768e-01 -2.24190068e+00 2.91894853e-01 1.97909161e-01
5.00124931e-01 -3.01670730e-02 -4.72246587e-01 -1.30887046e-01
-4.16972220e-01 -7.25170672e-02 -3.90760690e-01 -1.23418681e-01
-1.79339156e-01 2.10904568e-01 -1.85969751e-02 5.61919093e-01
4.51288104e-01 1.11121345e+00 -8.99273336e-01 5.10548577e-02
-2.38980390e-02 2.31236801e-01 -8.35529566e-01 2.26678312e-01
-2.47028753e-01 7.76328385e-01 -2.28501379e-01 -2.29916330e-02
4.61859286e-01 -2.21721753e-01 2.42590755e-01 -3.50963891e-01
9.35628172e-03 4.68822062e-01 -3.88163209e-01 1.64679205e+00
-4.61099654e-01 8.35235238e-01 -3.04255843e-01 -1.43587291e+00
6.99934542e-01 1.58383399e-02 4.86911744e-01 -8.44050348e-01
3.44138801e-01 8.64375532e-02 6.82803869e-01 -5.31506479e-01
-3.10926110e-01 -6.79388121e-02 1.77381486e-01 6.91320121e-01
9.63573039e-01 3.09682548e-01 1.14812173e-01 7.24667311e-03
1.48069477e+00 -9.02193934e-02 6.74820840e-02 -6.42656267e-01
1.31158561e-01 -2.25077838e-01 2.99091905e-01 8.29297900e-01
-1.63589209e-01 3.28480452e-01 8.79838884e-01 -4.37452823e-01
-1.02627039e+00 -1.11193335e+00 -4.85706061e-01 1.37875438e+00
-7.79049277e-01 -1.72174662e-01 -8.18500459e-01 -7.74783134e-01
2.82832414e-01 7.67376721e-01 -1.03932059e+00 -8.62973690e-01
-3.61221492e-01 -9.32319701e-01 5.12474418e-01 5.51195383e-01
3.40408355e-01 -1.10330570e+00 -6.75656617e-01 -2.33875066e-02
4.14764673e-01 -1.21623313e+00 1.63220223e-02 6.97392344e-01
-1.07797611e+00 -1.01576018e+00 -3.91963929e-01 -5.88699341e-01
6.99403405e-01 -2.32998654e-01 1.09544075e+00 -1.51710823e-01
-2.77116507e-01 2.12239102e-01 1.84840560e-02 -3.30594093e-01
-2.73256510e-01 3.69754523e-01 5.83724976e-02 1.86804473e-01
1.91474065e-01 -7.71654844e-01 -6.12260103e-01 -9.97446701e-02
-7.36096680e-01 -1.19837716e-01 9.26859438e-01 8.35141003e-01
3.69976968e-01 5.21194981e-03 6.73037112e-01 -1.08045030e+00
4.67997283e-01 -7.89063871e-01 -5.39437950e-01 -6.71127159e-03
-5.46218872e-01 6.34894907e-01 5.15914798e-01 -3.79200757e-01
-8.59727383e-01 4.73614112e-02 -9.30731669e-02 -2.38066524e-01
-3.17935586e-01 6.24134898e-01 7.74120837e-02 -1.86304096e-02
7.91476965e-01 1.59086570e-01 1.31784856e-01 -5.40140867e-01
3.61001164e-01 2.20171243e-01 7.57494092e-01 -6.16393209e-01
3.87987763e-01 2.61606991e-01 -1.35438323e-01 -3.04512501e-01
-9.19486165e-01 -6.54600039e-02 -9.65251684e-01 -1.03965633e-01
1.13600755e+00 -8.83292377e-01 -7.48304069e-01 1.37252346e-01
-7.77132869e-01 -9.36319113e-01 -3.59472811e-01 6.51949584e-01
-4.26440656e-01 -3.29465181e-01 -4.00218129e-01 -8.91203806e-02
-3.18391204e-01 -1.35266757e+00 9.03159022e-01 -2.55334992e-02
-1.39214201e-02 -1.08647084e+00 1.08119860e-01 2.33915791e-01
6.27793729e-01 -2.31626481e-02 1.22804153e+00 -1.09178817e+00
-3.09068680e-01 -1.16062418e-01 -4.73209471e-01 3.85360241e-01
1.97787210e-02 -4.74204689e-01 -1.35048628e+00 2.86304932e-02
-1.51863188e-01 -5.02444506e-01 1.42578673e+00 5.24260044e-01
1.85819316e+00 -2.54037023e-01 -1.80795312e-01 7.92997897e-01
1.00745142e+00 -4.12827253e-01 4.88380015e-01 2.51195822e-02
3.89484882e-01 5.26071787e-01 -5.60949862e-01 2.25508451e-01
4.07057434e-01 1.74002722e-01 -3.03231366e-02 9.46869925e-02
-1.64289087e-01 -1.14479639e-01 4.08168197e-01 5.02731144e-01
-4.63685244e-02 3.85689199e-01 -1.12629175e+00 3.27465385e-01
-1.47072530e+00 -8.57303798e-01 3.85447055e-01 2.05914974e+00
1.06202948e+00 8.37621167e-02 3.36439237e-02 -3.74120742e-01
2.90729016e-01 3.11517771e-02 -6.97145820e-01 -1.77931488e-01
-9.14935023e-02 7.63580680e-01 6.49739444e-01 1.38412073e-01
-9.36302364e-01 9.32139516e-01 7.09132624e+00 4.35872495e-01
-1.36271894e+00 4.82599288e-01 9.20156240e-01 -2.04753965e-01
-2.18068168e-01 -2.31036991e-01 -3.15664560e-01 4.36107367e-01
1.65404236e+00 -1.38928518e-01 9.36630428e-01 5.00712395e-01
1.18951276e-01 -1.25664100e-01 -1.44737518e+00 7.26260364e-01
1.51928708e-01 -1.40237975e+00 -9.05201137e-02 1.70970499e-03
5.93240499e-01 6.13989234e-01 2.99554974e-01 6.58903241e-01
3.80118698e-01 -1.18676281e+00 4.67678964e-01 7.97378004e-01
8.03061426e-01 -4.45429236e-01 6.89808547e-01 2.80817658e-01
-1.70706362e-01 -1.78775445e-01 -5.23973525e-01 -8.07305425e-03
-4.85355854e-01 8.75283360e-01 -8.39437366e-01 -6.76695481e-02
6.69726074e-01 7.28832781e-01 -1.22407568e+00 9.90400434e-01
-2.30726302e-02 8.87393296e-01 -1.10207880e-02 2.19141901e-01
1.91586390e-02 8.88124257e-02 -2.49184556e-02 1.14741123e+00
4.34383675e-02 1.72900364e-01 -5.73009700e-02 1.22270441e+00
-3.41974109e-01 -5.41705601e-02 -4.30708915e-01 -1.96218550e-01
1.46599161e-02 1.40585518e+00 -6.11301303e-01 -4.41853076e-01
-1.32031798e-01 8.39729548e-01 8.02981377e-01 3.42114389e-01
-4.94455546e-01 1.82138056e-01 4.01234180e-01 3.62184681e-02
3.39896888e-01 -2.01374069e-01 -5.60623407e-01 -1.17329502e+00
-4.55146968e-01 -9.62588847e-01 3.55544627e-01 -8.70268703e-01
-1.30203140e+00 4.82004076e-01 6.47631884e-02 -1.15064539e-01
-1.90302938e-01 -8.38285387e-01 -6.78102791e-01 9.46058631e-01
-1.24293649e+00 -6.64117575e-01 -2.17398956e-01 4.22420412e-01
2.11476266e-01 -3.80613238e-01 7.93536603e-01 1.82953984e-01
-8.27264845e-01 3.81471336e-01 2.33309433e-01 3.08201939e-01
8.13069999e-01 -1.09051704e+00 1.35947973e-01 5.63589096e-01
1.23292014e-01 6.95194781e-01 2.22569510e-01 -6.33801758e-01
-1.24694920e+00 -1.06769753e+00 5.13759017e-01 -6.71166837e-01
6.25140548e-01 -7.71741927e-01 -9.63137150e-01 9.46888983e-01
1.02361180e-01 3.18065345e-01 6.17556751e-01 5.28734922e-01
-5.91729879e-01 -3.87286961e-01 -1.00404358e+00 2.44502217e-01
1.08861780e+00 -8.39863479e-01 -7.14431584e-01 4.83334571e-01
4.15208519e-01 -1.16778620e-01 -9.50385511e-01 1.47357479e-01
5.36565363e-01 -6.43804193e-01 8.70129824e-01 -1.04442978e+00
5.74628413e-01 4.73886698e-01 -4.24346067e-02 -1.69830334e+00
-7.13436782e-01 9.05724093e-02 4.93448824e-02 7.79763043e-01
6.17484987e-01 -5.31739295e-01 4.17429894e-01 6.45836055e-01
-2.78551280e-01 -5.92193246e-01 -8.49552214e-01 -4.34890211e-01
3.34148884e-01 -3.16452235e-01 4.06779259e-01 8.38935852e-01
-1.97270647e-01 4.67670739e-01 1.53637961e-01 5.32940924e-02
5.89725077e-01 -3.47385734e-01 3.23775262e-01 -1.55056798e+00
-4.44253385e-01 -4.54300612e-01 -3.31741661e-01 -2.77190357e-01
5.66688120e-01 -1.48174167e+00 7.57827237e-02 -1.15362787e+00
5.66251993e-01 -3.49749744e-01 -7.35509574e-01 7.62706220e-01
-3.11320454e-01 3.57399583e-02 -2.95524988e-02 1.99071039e-03
-5.05599499e-01 3.53766292e-01 9.83554602e-01 9.34058987e-03
4.12059464e-02 -1.51611239e-01 -7.78292954e-01 4.02975500e-01
7.68244743e-01 -5.91222823e-01 -2.54798561e-01 -6.90125585e-01
2.97053549e-02 -1.70870632e-01 6.08392894e-01 -1.17243719e+00
-9.44220871e-02 2.49585316e-01 1.12146330e+00 5.11798821e-02
-7.41178310e-03 -4.35612321e-01 -3.40488791e-01 5.18200576e-01
-8.56556475e-01 -1.69716980e-02 4.13178623e-01 3.60435605e-01
3.20551366e-01 -1.30307361e-01 8.16721082e-01 -2.24655539e-01
-4.98536348e-01 5.69096863e-01 -3.59423369e-01 1.87861085e-01
3.40084195e-01 3.20282668e-01 -3.46125841e-01 -1.31059393e-01
-8.00521612e-01 -3.65130380e-02 -1.65011182e-01 1.62579045e-01
2.81884879e-01 -8.74275565e-01 -7.71257758e-01 4.63350892e-01
-1.80306882e-01 -4.09646183e-01 2.29056925e-01 1.03532398e+00
-1.81112707e-01 4.16189194e-01 -7.88507402e-01 -4.65646207e-01
-4.06513304e-01 4.09493327e-01 5.44467926e-01 -7.92223737e-02
-6.71742141e-01 7.13100374e-01 5.19590199e-01 -4.72008169e-01
9.17671546e-02 -5.35643995e-01 -2.11107001e-01 1.79092243e-01
3.87652755e-01 -2.12089848e-02 3.17170590e-01 -2.03493163e-01
-2.30446264e-01 -1.40791804e-01 -1.92134947e-01 -1.35040745e-01
1.84314525e+00 3.24098796e-01 -1.38400942e-01 4.12551045e-01
1.22818792e+00 -3.08384359e-01 -1.40617895e+00 -3.35031748e-01
4.95272838e-02 -7.42750466e-02 6.42377555e-01 -1.06271529e+00
-1.23225129e+00 8.23489785e-01 7.34580457e-01 -5.66628337e-01
8.48342061e-01 1.83924779e-01 2.64883339e-01 6.27566040e-01
7.13142678e-02 -8.75467062e-01 1.32543832e-01 6.17206395e-01
7.43476033e-01 -1.19534707e+00 2.73626111e-02 1.11556888e-01
-6.46755874e-01 1.08583963e+00 6.35838807e-01 -5.36746323e-01
7.25233197e-01 1.30823636e-02 -4.13720876e-01 -5.91322660e-01
-7.30707288e-01 -2.42040634e-01 6.05305552e-01 6.06943846e-01
1.93174005e-01 -6.08307049e-02 1.02027757e-02 1.11480951e+00
-3.80896598e-01 8.85199849e-03 2.55501986e-01 2.07636595e-01
-4.47194427e-01 -4.27737862e-01 1.32829010e-01 1.20447671e+00
-2.72223413e-01 -4.01724458e-01 -1.73063830e-01 4.51570541e-01
3.95161748e-01 2.06308410e-01 4.79518801e-01 -1.86691329e-01
8.72945189e-02 5.13587534e-01 5.80993056e-01 -8.78416121e-01
-7.88199306e-01 -3.76150012e-01 4.29268032e-02 -6.61260962e-01
3.49834887e-03 -8.38693261e-01 -9.50293481e-01 -1.50212739e-02
1.37189806e-01 -1.63086116e-01 6.24606311e-01 1.20779097e+00
5.57945192e-01 8.18311751e-01 4.00510907e-01 -7.95515776e-01
-7.16129124e-01 -1.07210124e+00 -4.05940592e-01 4.43723708e-01
2.83048719e-01 -8.43442261e-01 -8.26996341e-02 -2.95987315e-02] | [12.64585018157959, 3.2165427207946777] |
320d9f6d-06b0-4132-9703-0393c1afc99a | total-energy-shaping-control-for-mechanical | 2301.03746 | null | https://arxiv.org/abs/2301.03746v2 | https://arxiv.org/pdf/2301.03746v2.pdf | Total energy-shaping control for mechanical systems via Control-by-Interconnection | Application of IDA-PBC to mechanical systems has received much attention in recent decades, but its application is still limited by the solvability of the so-called matching conditions. In this work, it is shown that total energy-shaping control of under-actuated mechanical systems has a control-by-interconnection interpretation. Using this interpretation, alternate matching conditions are formulated that defines constraints on the added energy, rather then the total closed-loop energy. It is additionally shown that, for systems that are under-actuated degree one with the mass matrix depending on a single coordinate, the kinetic energy matching conditions resolve to ODEs which can be evaluated numerically. Using this approach controllers are proposed for the benchmark cart-pole and acrobot systems. | ['Joel Ferguson'] | 2023-01-10 | null | null | null | null | ['total-energy', 'acrobot'] | ['miscellaneous', 'playing-games'] | [ 3.38704675e-01 8.36079478e-01 -2.84260750e-01 4.62194800e-01
2.08562553e-01 -8.60475063e-01 5.04164755e-01 1.71158120e-01
-2.86334932e-01 9.30156112e-01 -6.25226378e-01 -3.95951331e-01
-6.99166656e-01 -3.43211561e-01 -4.49643582e-01 -8.45725834e-01
1.18861087e-01 2.54050612e-01 -1.50249064e-01 -5.95342875e-01
-1.27632812e-01 6.29778385e-01 -1.20981777e+00 -7.24283934e-01
8.78781796e-01 8.41817558e-01 1.98171988e-01 5.61712325e-01
7.25960195e-01 3.04662317e-01 -3.35898161e-01 2.76195526e-01
3.18249792e-01 -5.42425632e-01 -6.04724884e-01 1.87258527e-01
-1.15916826e-01 3.00327353e-02 -1.02654226e-01 1.13660192e+00
3.62039596e-01 5.95719159e-01 8.07055712e-01 -1.41300559e+00
-1.00744545e-01 3.81768644e-01 -7.30758905e-02 -2.84373432e-01
3.49691153e-01 -6.36402145e-02 7.97957838e-01 -6.69298947e-01
5.58586776e-01 1.18722808e+00 4.85585213e-01 5.54638207e-01
-1.68316042e+00 -1.43736914e-01 -1.09077357e-01 -2.62216199e-02
-1.31534898e+00 4.71693054e-02 6.06858850e-01 -6.42467678e-01
6.67263031e-01 5.54053128e-01 8.97302508e-01 4.98298854e-01
7.62313426e-01 2.19912231e-01 9.28604901e-01 -5.30653417e-01
3.91804099e-01 1.29130797e-03 3.35753918e-01 5.47781706e-01
7.74751663e-01 1.60824046e-01 3.10113907e-01 -2.21747294e-01
6.02566063e-01 -1.95483595e-01 -4.16702002e-01 -7.95310795e-01
-9.94538546e-01 8.11954975e-01 2.08049595e-01 3.01790178e-01
-3.24241847e-01 8.86146501e-02 2.09757775e-01 1.60339490e-01
7.36883655e-02 7.88379014e-01 -2.43308321e-01 1.95045080e-02
-1.55769750e-01 6.02945507e-01 1.00763929e+00 9.90385652e-01
2.78051764e-01 3.83726180e-01 3.01304668e-01 3.36166382e-01
1.67467222e-01 7.09874332e-01 3.15795578e-02 -9.28288519e-01
1.09129652e-01 7.05637157e-01 5.53246975e-01 -8.96384478e-01
-7.24725902e-01 -1.68689221e-01 -1.08272779e+00 5.21509051e-01
4.44185317e-01 -5.33183873e-01 -3.14346015e-01 1.82909131e+00
3.67195159e-01 -6.64583206e-01 1.13655590e-01 9.67044055e-01
-3.19879800e-01 6.03883564e-01 -4.03934836e-01 -5.13596714e-01
1.25035930e+00 -3.17003220e-01 -1.18730497e+00 1.16703764e-01
2.01210141e-01 -3.89129400e-01 6.03299797e-01 4.37016994e-01
-1.36212480e+00 -5.47927678e-01 -1.40696466e+00 2.55300969e-01
-2.07130283e-01 3.70104343e-01 -3.31929564e-01 3.80757570e-01
-7.77280152e-01 1.05158031e+00 -8.33055794e-01 -2.15225399e-01
-7.56780148e-01 5.03639460e-01 -4.79685701e-02 1.03549862e+00
-1.29587603e+00 1.50061536e+00 6.41708910e-01 4.75686520e-01
-2.08463520e-01 -7.27137089e-01 -5.14816940e-01 -1.64258063e-01
6.19482398e-01 -8.05033147e-01 1.06139481e+00 -4.82512981e-01
-2.11364079e+00 1.77759171e-01 4.19856906e-01 -2.79846579e-01
6.76843226e-01 -6.14522323e-02 -1.24751695e-01 3.53225857e-01
-1.97418138e-01 -6.14975505e-02 9.34095502e-01 -8.87427509e-01
-1.60937533e-01 -6.01481833e-02 1.59531876e-01 1.16231970e-01
-5.62692806e-02 -5.27021289e-01 3.62343311e-01 -5.39501488e-01
8.51744264e-02 -1.35276937e+00 -3.63020986e-01 -9.36350450e-02
-5.12548625e-01 -1.91150263e-01 8.50712538e-01 -2.64693886e-01
1.23544371e+00 -2.13198853e+00 1.04663229e+00 3.69307905e-01
-2.30178237e-01 1.94463477e-01 6.09505117e-01 8.97253215e-01
-2.24110782e-01 5.88655956e-02 -1.24923393e-01 4.19393808e-01
3.12344790e-01 8.75073820e-02 -4.28576529e-01 5.52216470e-01
4.49235409e-01 1.40081480e-01 -6.67899609e-01 -2.59868383e-01
6.38477504e-01 3.98149073e-01 -3.36782098e-01 -1.32491946e-01
-1.18294664e-01 4.70659554e-01 -7.43492007e-01 -3.41264606e-02
4.27784979e-01 1.49266273e-01 4.67698038e-01 -3.43323052e-01
-6.17504299e-01 -3.69772464e-01 -1.42393720e+00 9.78198826e-01
-4.54239249e-01 3.47609431e-01 8.49760532e-01 -1.16703236e+00
7.87764490e-01 4.06241238e-01 7.39407361e-01 -5.38806431e-02
5.75164080e-01 4.91147041e-01 2.77884483e-01 -1.11016370e-01
3.15769106e-01 -2.75905132e-01 -1.59752205e-01 -1.24238588e-01
-1.45518497e-01 -6.98547721e-01 4.67385441e-01 1.62107293e-02
8.58053684e-01 4.75070365e-02 5.54287732e-01 -9.00487006e-01
1.19138861e+00 3.42573076e-01 4.30568606e-01 -1.73745826e-01
1.54483601e-01 -1.83852986e-01 6.82025135e-01 2.97371268e-01
-1.28321624e+00 -5.73315918e-01 -5.77855289e-01 1.87127054e-01
1.65983319e-01 -1.99432731e-01 -9.46286976e-01 2.87066251e-01
4.42081839e-01 7.20533848e-01 -3.31792384e-01 -5.35893977e-01
-5.31563580e-01 -1.97088301e-01 2.70746313e-02 1.25446126e-01
1.00241922e-01 -3.08560431e-01 -1.26549625e+00 5.43837547e-01
1.85032323e-01 -9.66626406e-01 -1.16030119e-01 3.29492718e-01
-9.13486004e-01 -1.36311400e+00 -8.19098711e-01 -3.05600792e-01
7.27725387e-01 -2.58396089e-01 4.38531488e-01 -2.54435480e-01
-3.73198867e-01 6.25750721e-01 1.31458730e-01 -2.82926232e-01
-5.83654702e-01 -4.13472801e-02 5.18488705e-01 -2.35802885e-02
-6.43579960e-01 -1.30708173e-01 -3.35785955e-01 7.79794693e-01
-6.78869069e-01 1.88851431e-01 2.57837683e-01 8.85912657e-01
6.99108541e-01 4.26809788e-02 8.61874402e-01 -1.76269352e-01
4.92822438e-01 5.29506765e-02 -1.37728465e+00 -1.43904686e-01
-6.69930637e-01 1.28065825e-01 1.09758556e+00 -5.27958810e-01
-9.92114902e-01 6.48907900e-01 2.54213095e-01 -4.16295767e-01
3.72214049e-01 3.11722934e-01 -3.49398047e-01 -3.13997418e-01
1.21910989e-01 -1.05584398e-01 4.94497806e-01 -5.56281865e-01
3.20123821e-01 2.18713239e-01 5.30963600e-01 -5.69154739e-01
9.10531223e-01 3.47989053e-02 1.04873288e+00 -9.03596103e-01
-1.59770995e-02 -1.76384017e-01 -5.57204127e-01 -4.65433478e-01
6.97191954e-01 -3.11946869e-01 -1.30726576e+00 2.62826562e-01
-1.02171683e+00 -1.61398306e-01 -5.74723065e-01 5.93503773e-01
-1.07665849e+00 2.39827558e-01 -5.02700031e-01 -1.55355811e+00
-1.13008641e-01 -1.00637233e+00 7.01992512e-01 5.17951995e-02
-3.94331694e-01 -9.61051226e-01 2.37914339e-01 -4.20582980e-01
2.11754695e-01 7.77069688e-01 1.00975835e+00 2.43994910e-02
-2.89518863e-01 -3.69727045e-01 5.44218779e-01 2.87877530e-01
-1.37343958e-01 8.24064612e-02 -3.68114591e-01 -7.24956453e-01
2.53864735e-01 9.00783911e-02 3.40423360e-02 4.06758636e-01
3.31135005e-01 -4.42166716e-01 -6.02325082e-01 -8.05881917e-02
1.88610399e+00 3.87014180e-01 -1.43982723e-01 2.74724215e-02
3.15270811e-01 7.88380086e-01 6.75907433e-01 2.88806677e-01
-3.13483447e-01 1.04707611e+00 5.31997442e-01 2.39902452e-01
3.99291962e-01 2.29128078e-01 4.16927606e-01 6.07944071e-01
-2.48625427e-01 5.09543940e-02 -5.20351589e-01 4.46899414e-01
-1.95036399e+00 -4.88212794e-01 -7.56980717e-01 2.41282487e+00
7.89803743e-01 -1.52524561e-01 2.76322603e-01 6.83849990e-01
1.04095161e+00 -4.93843466e-01 -7.05104709e-01 -6.03501499e-01
-2.21228544e-02 1.42678339e-02 8.60184312e-01 8.93628180e-01
-8.51053953e-01 -2.38155082e-01 6.22271299e+00 5.98384380e-01
-8.54080021e-01 -1.51826963e-01 -1.09324403e-01 1.69231921e-01
3.12310934e-01 1.43000051e-01 -7.79561102e-01 6.32144928e-01
9.85027134e-01 -9.56887066e-01 3.80438298e-01 1.08523405e+00
5.77559114e-01 -4.07421082e-01 -9.71877635e-01 3.96618307e-01
-3.27163935e-01 -5.99451721e-01 -5.80023229e-01 2.88498878e-01
6.56107426e-01 -8.61426294e-01 -3.45225818e-02 -1.00592285e-01
-2.43929103e-01 -3.62153143e-01 7.03987241e-01 5.40931106e-01
4.49012041e-01 -8.43733728e-01 4.67581183e-01 6.19594455e-01
-1.31589961e+00 -3.35408568e-01 -1.69074208e-01 -3.91598552e-01
6.15531921e-01 5.88331044e-01 -4.16048080e-01 1.02262640e+00
-2.83088177e-01 5.12983143e-01 -2.34837413e-01 7.63126671e-01
1.62695125e-02 3.43318135e-02 -6.60634995e-01 -5.98661244e-01
1.81175936e-02 -8.74815166e-01 9.03747737e-01 6.15689754e-01
3.11160326e-01 1.60474271e-01 6.70395885e-03 1.10629582e+00
7.67292500e-01 5.40986247e-02 -5.88228703e-01 1.73220158e-01
-1.41361758e-01 1.34996796e+00 -5.57129860e-01 -1.94969594e-01
1.90716505e-01 5.19776285e-01 -5.57907999e-01 2.25258932e-01
-8.51153612e-01 -9.02078509e-01 6.84116185e-01 6.60507083e-02
-1.19684532e-01 -4.18047994e-01 -4.51448793e-03 -7.09194362e-01
1.22258641e-01 -2.83369124e-01 -2.37802584e-02 -5.16347826e-01
-8.86203885e-01 1.48776755e-01 5.72141469e-01 -1.56821859e+00
-9.46361840e-01 -8.86113524e-01 -2.95749933e-01 1.03529787e+00
-8.81463885e-01 -1.82517007e-01 2.77856261e-01 2.38826603e-01
4.54585850e-02 3.64263237e-01 9.36234653e-01 3.00645590e-01
-9.20089126e-01 -2.73206472e-01 6.54062033e-01 -4.88275826e-01
1.64441839e-01 -1.35180283e+00 -3.23594153e-01 7.11030483e-01
-9.84507024e-01 5.13514757e-01 1.26535833e+00 -4.10016268e-01
-2.03247762e+00 -8.85815442e-01 7.02309549e-01 1.25874236e-01
1.03279340e+00 -7.98769891e-02 -7.92040408e-01 2.78196573e-01
4.76434469e-01 -2.00207010e-01 -2.41570696e-01 -8.92592490e-01
7.39788592e-01 -1.01057440e-01 -9.88925219e-01 4.31550235e-01
5.05684018e-01 -2.45688245e-01 -5.84445953e-01 2.85343587e-01
5.40550768e-01 -4.47927773e-01 -1.08315682e+00 4.48103756e-01
4.96795714e-01 -1.13309972e-01 8.08962345e-01 -3.52961212e-01
1.20050590e-02 -8.19925368e-01 2.23311901e-01 -1.47855306e+00
-3.98120247e-02 -1.15578675e+00 -3.03016275e-01 9.67689216e-01
1.55959561e-01 -9.15499449e-01 1.93557858e-01 4.12552267e-01
-1.05320431e-01 -6.91966295e-01 -1.19095004e+00 -1.25531089e+00
2.08558038e-01 2.79742151e-01 -1.02557920e-01 6.00916743e-01
8.81426036e-01 5.79673290e-01 -1.73452705e-01 3.85751486e-01
4.56090540e-01 -1.43635273e-01 4.64236200e-01 -1.39716291e+00
-3.44486833e-01 -5.42334020e-01 -2.45774746e-01 -6.49605989e-01
2.21181348e-01 -5.69126308e-01 1.62661180e-01 -1.44460273e+00
-5.69691122e-01 -1.13925658e-01 1.24899589e-01 1.52051702e-01
2.24964589e-01 -2.40152534e-02 2.73038268e-01 -1.99560057e-02
2.60547727e-01 5.73359072e-01 1.17145407e+00 7.93923810e-03
-3.01876068e-01 4.08711344e-01 3.18304718e-01 6.32615924e-01
8.25245380e-01 1.38249258e-02 -7.17333615e-01 4.38106179e-01
3.84702891e-01 5.45193851e-01 5.85612714e-01 -1.34044993e+00
4.14698124e-02 -2.34747216e-01 -3.40679556e-01 -8.48981142e-01
4.27476019e-01 -1.23382485e+00 9.23637986e-01 1.23216426e+00
-2.93717712e-01 1.51133701e-01 1.61003783e-01 5.89243591e-01
-1.75113663e-01 -5.52922308e-01 9.07684505e-01 4.52646255e-01
-5.61252125e-02 -4.87074614e-01 -9.08718884e-01 -3.99532706e-01
1.29047525e+00 7.35469088e-02 -4.37994190e-02 -1.98005319e-01
-9.13230300e-01 1.54922128e-01 3.09033424e-01 5.45820128e-03
-1.23069301e-01 -1.42591596e+00 -1.68342665e-01 -3.13303284e-02
-4.00078118e-01 -3.35940659e-01 3.13260347e-01 1.22648442e+00
-3.06438595e-01 8.65474105e-01 -4.75891501e-01 -6.14828289e-01
-1.12628806e+00 7.25720227e-01 4.05787259e-01 5.69955595e-02
-4.69078869e-01 -2.85102367e-01 -1.62724540e-01 -5.92740327e-02
-1.29565537e-01 -8.97255838e-01 1.76050067e-02 -4.83368486e-02
-1.41964316e-01 6.28048360e-01 9.72878337e-02 -5.14943898e-01
-3.28860074e-01 1.07985711e+00 8.36041451e-01 -3.95914942e-01
8.99042010e-01 -1.10949025e-01 3.31395343e-02 5.37848294e-01
8.37934136e-01 -3.31105590e-01 -1.12949228e+00 4.17643696e-01
-1.67849839e-01 -4.89367777e-03 -1.92903906e-01 -4.56908971e-01
-7.34703481e-01 6.10693455e-01 4.29720312e-01 9.28292036e-01
1.03324831e+00 -6.38511539e-01 7.93805346e-02 5.80702245e-01
4.08164680e-01 -1.45210505e+00 -3.25529337e-01 5.80259979e-01
1.07217443e+00 -2.20953718e-01 1.08174175e-01 -9.63958621e-01
-1.29490972e-01 1.38597965e+00 4.06150371e-01 -6.14868283e-01
6.19926870e-01 4.62162912e-01 -6.90761387e-01 5.94429016e-01
-5.37271976e-01 -4.10102010e-02 3.63834649e-01 2.88657755e-01
4.60776180e-01 1.26781493e-01 -1.19139504e+00 4.50245857e-01
2.17849225e-01 6.63031116e-02 7.05546618e-01 9.62998450e-01
-4.41251457e-01 -1.08229911e+00 -7.11439431e-01 -3.08583319e-01
-1.77927688e-01 6.32407725e-01 -4.18208003e-01 1.34879971e+00
-1.98258951e-01 9.58893597e-01 -1.54419038e-02 -5.53338416e-02
9.87031460e-01 1.62396938e-01 5.08289814e-01 -3.20933104e-01
-4.69365835e-01 2.27471352e-01 1.82849213e-01 -4.01550502e-01
-3.93721342e-01 -4.69078451e-01 -1.44140863e+00 -9.67977419e-02
-8.71006489e-01 4.45457608e-01 7.86962032e-01 5.32330275e-01
2.61221230e-01 8.10467184e-01 5.63746691e-01 -1.01232100e+00
-1.20851815e+00 -9.79259670e-01 -7.23795116e-01 -1.10046566e-01
4.25920546e-01 -1.09355319e+00 -5.24797618e-01 -1.56082705e-01] | [5.427215099334717, 2.5769765377044678] |
34ec2bbc-b618-4c34-8958-e925b57b6668 | care-open-knowledge-graph-embeddings | null | null | https://aclanthology.org/D19-1036 | https://aclanthology.org/D19-1036.pdf | CaRe: Open Knowledge Graph Embeddings | Open Information Extraction (OpenIE) methods are effective at extracting (noun phrase, relation phrase, noun phrase) triples from text, e.g., (Barack Obama, took birth in, Honolulu). Organization of such triples in the form of a graph with noun phrases (NPs) as nodes and relation phrases (RPs) as edges results in the construction of Open Knowledge Graphs (OpenKGs). In order to use such OpenKGs in downstream tasks, it is often desirable to learn embeddings of the NPs and RPs present in the graph. Even though several Knowledge Graph (KG) embedding methods have been recently proposed, all of those methods have targeted Ontological KGs, as opposed to OpenKGs. Straightforward application of existing Ontological KG embedding methods to OpenKGs is challenging, as unlike Ontological KGs, OpenKGs are not canonicalized, i.e., a real-world entity may be represented using multiple nodes in the OpenKG, with each node corresponding to a different NP referring to the entity. For example, nodes with labels Barack Obama, Obama, and President Obama may refer to the same real-world entity Barack Obama. Even though canonicalization of OpenKGs has received some attention lately, output of such methods has not been used to improve OpenKG embed- dings. We fill this gap in the paper and propose Canonicalization-infused Representations (CaRe) for OpenKGs. Through extensive experiments, we observe that CaRe enables existing models to adapt to the challenges in OpenKGs and achieve substantial improvements for the link prediction task. | ['Swapnil Gupta', 'Sreyash Kenkre', 'Partha Talukdar'] | 2019-11-01 | null | null | null | ijcnlp-2019-11 | ['open-information-extraction'] | ['natural-language-processing'] | [-2.20784158e-01 1.12410367e+00 -4.14113939e-01 -1.13985479e-01
-3.50011528e-01 -7.33626127e-01 4.46681738e-01 6.94686830e-01
-4.50802036e-02 8.51008177e-01 4.29147780e-01 -5.02856910e-01
-3.75238121e-01 -1.36918139e+00 -8.38523328e-01 -1.81601539e-01
-3.77618194e-01 5.76434374e-01 2.37257496e-01 -2.46443242e-01
-2.90532678e-01 3.01562756e-01 -1.20394254e+00 8.11108202e-02
6.93989396e-01 7.15597570e-01 -3.03714331e-02 2.54728496e-01
-3.66397709e-01 6.21720731e-01 -3.46948028e-01 -1.00506091e+00
1.79443359e-01 1.90143347e-01 -1.29392397e+00 -4.76608217e-01
4.27775025e-01 1.90624699e-01 -6.00313902e-01 9.64433432e-01
1.13838740e-01 1.02062315e-01 7.79379666e-01 -1.57286108e+00
-1.06496716e+00 1.22996962e+00 -1.89835876e-01 8.40626657e-03
4.51309711e-01 -6.57484293e-01 1.88156593e+00 -8.80475402e-01
8.83187473e-01 1.23009658e+00 7.19149828e-01 2.77215362e-01
-1.07824564e+00 -4.53328490e-01 1.14601120e-01 4.14581865e-01
-1.46782064e+00 -1.09809525e-01 5.18248737e-01 -1.76914960e-01
1.20256567e+00 1.51811868e-01 7.72874892e-01 7.47540057e-01
9.31820571e-02 6.80076599e-01 5.62487543e-01 -6.97732627e-01
-2.75804219e-03 8.58745202e-02 6.49856091e-01 8.45877290e-01
9.10631895e-01 -2.69208968e-01 -4.11783814e-01 -2.85210639e-01
3.24234992e-01 -2.98441023e-01 -2.31551006e-01 -5.90046465e-01
-1.03174663e+00 7.86453426e-01 8.90811563e-01 2.35492334e-01
-1.78820536e-01 2.09926307e-01 3.86637926e-01 5.54498807e-02
3.10882270e-01 5.17980337e-01 -6.94104373e-01 1.22233748e-01
-2.56989837e-01 4.37351793e-01 1.20500124e+00 1.27526188e+00
1.06551981e+00 -3.19555849e-01 3.31034005e-01 6.89210773e-01
3.23884100e-01 5.66954836e-02 9.88083109e-02 -6.98473752e-01
7.52971411e-01 1.09081757e+00 -4.74797599e-02 -1.20728636e+00
-5.04374087e-01 -1.38864681e-01 -3.30258906e-01 -4.68739957e-01
2.33583957e-01 5.71143813e-02 -7.88903236e-01 1.64756584e+00
4.95716453e-01 2.29494885e-01 5.32882988e-01 5.71549535e-01
1.40737510e+00 6.23570204e-01 3.24365973e-01 6.87351227e-02
1.79738522e+00 -7.77782202e-01 -7.98807025e-01 -2.26776361e-01
8.92093956e-01 -6.33752719e-02 6.87267363e-01 -2.37532318e-01
-7.05027103e-01 3.47422920e-02 -1.08914030e+00 -5.04102647e-01
-1.06473899e+00 -3.27451915e-01 9.28941965e-01 2.66981184e-01
-7.04787076e-01 5.01449704e-01 -7.38282621e-01 -5.63408017e-01
3.52340102e-01 4.01692867e-01 -8.64808619e-01 -1.72611967e-01
-1.65222442e+00 1.06987381e+00 1.14210606e+00 -5.61788902e-02
-1.83844060e-01 -8.36135685e-01 -1.46594679e+00 3.75547528e-01
8.24194074e-01 -7.96175301e-01 7.88534045e-01 -3.30914319e-01
-6.45984888e-01 8.89898002e-01 2.26228274e-02 -4.97371525e-01
-3.45595151e-01 -1.71660613e-02 -8.12766850e-01 3.46413888e-02
3.37633282e-01 7.00368822e-01 4.39690739e-01 -1.27474713e+00
-7.39565432e-01 -4.63840961e-01 8.29745948e-01 3.56490687e-02
-7.41695464e-01 1.08237430e-01 -2.23784745e-01 -3.04544717e-01
4.99846071e-01 -8.75476956e-01 6.58792108e-02 -5.45437813e-01
-7.85597265e-01 -6.60485625e-01 8.20176303e-01 -5.71907878e-01
1.55594993e+00 -1.93228376e+00 2.45019585e-01 2.36769602e-01
6.35780334e-01 9.49370936e-02 1.84829384e-01 8.12955320e-01
-2.71373063e-01 5.49072981e-01 -1.61108598e-01 1.65283889e-01
4.35965270e-01 7.92756796e-01 -2.77261317e-01 6.08260520e-02
1.96409315e-01 1.25066125e+00 -1.08374512e+00 -7.11143196e-01
-2.70021081e-01 1.39686495e-01 -5.17493784e-01 -2.60326773e-01
-3.05375785e-01 -3.28070343e-01 -4.54266608e-01 8.27673376e-01
3.65714878e-01 -2.68224359e-01 6.21608138e-01 -6.63129926e-01
3.17391474e-03 6.44517362e-01 -1.11802506e+00 1.14711833e+00
-3.22272360e-01 4.46633488e-01 -3.08460057e-01 -9.72027540e-01
6.68688059e-01 4.89650160e-01 5.04807472e-01 -1.41775310e-01
-1.28199652e-01 2.07779050e-01 5.02707660e-02 -3.57722640e-01
9.80677545e-01 -6.63131475e-02 -4.49454606e-01 2.13338777e-01
3.71277392e-01 3.78104374e-02 6.02381110e-01 6.18570745e-01
1.32578933e+00 2.76948273e-01 8.03854942e-01 -1.02740720e-01
1.99276477e-01 2.92259753e-01 8.26681197e-01 3.66683483e-01
2.44727861e-02 1.80546671e-01 7.29706883e-01 -5.09119570e-01
-9.36713874e-01 -1.29648006e+00 -3.82144392e-01 8.00627053e-01
1.97430030e-01 -1.14829445e+00 -4.07373667e-01 -1.12394524e+00
2.92468756e-01 6.82195067e-01 -4.46119636e-01 -4.78674434e-02
-6.31426156e-01 -6.09752476e-01 6.36622608e-01 6.57881320e-01
1.66809529e-01 -8.73851478e-01 -1.75756082e-01 4.03711230e-01
-3.99433076e-01 -1.49691772e+00 3.44299264e-02 2.92757750e-01
-5.26532829e-01 -1.40370846e+00 1.62760183e-01 -8.98297727e-01
7.35124767e-01 -8.00741687e-02 1.29765260e+00 -4.97789159e-02
-2.97172423e-02 4.31631088e-01 -6.62492454e-01 -4.80751216e-01
-1.95823640e-01 3.06793958e-01 2.34826624e-01 -2.45779917e-01
6.57659173e-01 -6.02860689e-01 -9.91421118e-02 5.89243360e-02
-9.46119905e-01 -2.15584673e-02 2.68712074e-01 7.48638809e-01
4.82626289e-01 2.67897218e-01 8.30969393e-01 -1.12594962e+00
4.92827445e-01 -8.17619145e-01 -4.68353838e-01 4.74862456e-01
-6.03909552e-01 -1.00667756e-02 4.49529320e-01 -1.71178311e-01
-6.20367289e-01 -2.35825136e-01 -1.56595167e-02 -1.54893205e-01
1.28257632e-01 1.37034202e+00 -3.27723771e-01 2.12126642e-01
4.60862011e-01 -2.85174996e-01 -5.07049203e-01 -3.11863452e-01
5.99949062e-01 6.26834333e-01 4.59196448e-01 -8.58370364e-01
1.07131732e+00 3.10866624e-01 2.75721192e-01 -6.02080762e-01
-1.08814728e+00 -2.96084106e-01 -8.30961287e-01 1.99470147e-01
8.86373937e-01 -1.03543961e+00 -5.73652148e-01 -3.92540991e-01
-1.06435800e+00 1.83378190e-01 -4.33223069e-01 3.54921490e-01
-1.58948764e-01 3.26243013e-01 -6.49468899e-01 -4.94803905e-01
-1.23797372e-01 -5.85301399e-01 7.64447987e-01 2.57092923e-01
-4.31658596e-01 -1.30792058e+00 -1.93096668e-01 3.34347546e-01
-1.66311756e-01 2.54432321e-01 1.55449259e+00 -9.66653109e-01
-6.35639727e-01 -3.43356073e-01 -3.54072481e-01 9.51727405e-02
3.58319223e-01 1.59756579e-02 -6.20869637e-01 -6.26616776e-02
-7.25616097e-01 -1.58378348e-01 4.63509828e-01 -4.25975509e-02
5.34345925e-01 -5.17059743e-01 -6.99594736e-01 3.42211455e-01
1.46523416e+00 -1.32328004e-01 6.43046200e-01 5.91518939e-01
9.33270872e-01 6.30374193e-01 6.16579115e-01 6.82491660e-02
1.13161874e+00 6.55209959e-01 4.29637760e-01 4.02118444e-01
-3.59522551e-02 -6.44262612e-01 4.25630420e-01 8.68814707e-01
-4.92785312e-02 -2.94621944e-01 -9.75701511e-01 7.86629200e-01
-1.85337210e+00 -8.15560520e-01 -4.62349027e-01 1.72255468e+00
9.00805771e-01 -1.08696356e-01 -3.07435930e-01 3.60116363e-02
4.98459429e-01 1.22219175e-01 -2.24257603e-01 -2.49158621e-01
-1.02375470e-01 1.37442991e-01 4.90571976e-01 3.96901011e-01
-1.02110755e+00 1.19526875e+00 5.08028507e+00 5.06290972e-01
-4.36329991e-01 1.77774936e-01 -1.42192066e-01 4.53116149e-01
-5.70677102e-01 7.71486461e-01 -1.19147456e+00 2.48437136e-01
7.94488490e-01 -4.88564551e-01 2.32018068e-01 7.18128383e-01
-3.95252287e-01 5.78196682e-02 -1.31925595e+00 6.48161888e-01
-1.31094217e-01 -1.43886971e+00 1.24497928e-01 2.67991036e-01
5.63922286e-01 -1.73055008e-01 -5.57107091e-01 7.92439222e-01
7.70058513e-01 -8.25678408e-01 5.98180652e-01 1.49802282e-01
5.51253915e-01 -6.03572547e-01 7.06664324e-01 2.60023326e-01
-1.64405990e+00 -2.65793711e-01 -5.83559453e-01 -7.51070725e-03
2.92178750e-01 5.59565723e-01 -9.78730381e-01 1.32692528e+00
7.08314598e-01 6.51858747e-01 -4.78132755e-01 4.25701559e-01
-6.77282870e-01 5.09531319e-01 -3.51168662e-01 1.66954517e-01
2.30508298e-01 -2.68229872e-01 6.00048959e-01 8.88968229e-01
2.67453581e-01 3.55798900e-01 3.45653474e-01 9.49213147e-01
-4.88676488e-01 2.17232525e-01 -9.43830371e-01 -4.42984790e-01
7.04176903e-01 1.28913653e+00 -4.15390283e-01 -3.61515552e-01
-7.79441655e-01 3.27233106e-01 7.98552752e-01 3.17844450e-01
-7.93653190e-01 -4.88351464e-01 7.31760204e-01 3.43466341e-01
2.41080731e-01 -1.78345546e-01 5.76703139e-02 -1.19462359e+00
1.07181408e-01 -4.20327872e-01 8.83763552e-01 -8.96034777e-01
-1.30480587e+00 4.76579159e-01 2.63167113e-01 -9.25372899e-01
-2.00399727e-01 -5.68811178e-01 -3.06564987e-01 6.50382757e-01
-1.50845230e+00 -1.52205467e+00 1.42721727e-01 2.63729364e-01
-1.43392580e-02 1.39234122e-02 1.06997776e+00 2.43306488e-01
-5.39368629e-01 4.20654058e-01 -2.80772746e-01 4.70037192e-01
4.30017829e-01 -1.48813200e+00 2.77863801e-01 7.77560771e-01
5.28862119e-01 9.34304595e-01 5.17146587e-01 -1.03487027e+00
-1.48978066e+00 -1.20060480e+00 1.43449867e+00 -6.30367577e-01
1.13625371e+00 -4.74524051e-01 -9.98230100e-01 1.39868438e+00
1.28967941e-01 4.99927521e-01 7.99009025e-01 7.39456892e-01
-5.75161219e-01 2.49895416e-02 -9.84227240e-01 7.46440589e-01
1.35615003e+00 -5.05733192e-01 -1.12559843e+00 2.97488481e-01
1.14513636e+00 -5.00141025e-01 -1.43424141e+00 5.52102506e-01
3.25790972e-01 -2.87723213e-01 1.01869833e+00 -1.17927957e+00
3.51191461e-01 -4.24663067e-01 -3.26767772e-01 -1.47277915e+00
-3.16891879e-01 -1.42560527e-01 -7.27557302e-01 1.49134624e+00
6.70297563e-01 -8.31279457e-01 5.78590691e-01 8.35396171e-01
-4.46965814e-01 -6.82135463e-01 -1.01883829e+00 -9.45366859e-01
2.83678044e-02 -4.10483241e-01 8.91646862e-01 1.29566598e+00
7.59253323e-01 5.25673389e-01 -3.15568633e-02 6.70711696e-01
4.67168212e-01 3.60072345e-01 5.67121625e-01 -1.61083162e+00
9.40899178e-02 1.66070074e-01 -8.19964170e-01 -6.67588294e-01
6.09261215e-01 -1.54070008e+00 -3.41555119e-01 -2.16030025e+00
8.77968669e-02 -9.16833341e-01 -1.86466992e-01 1.07992852e+00
-6.28893673e-02 4.24893871e-02 2.27587834e-01 -9.03768763e-02
-3.68947983e-01 4.15825993e-01 9.67633903e-01 -3.19976121e-01
-2.72903237e-02 -4.55102682e-01 -1.02463472e+00 7.17984617e-01
4.96309459e-01 -6.03892803e-01 -3.79432410e-01 -3.08724940e-01
8.38357687e-01 7.46692568e-02 4.83718872e-01 -6.23126805e-01
3.09940070e-01 1.84005145e-02 -1.54586509e-01 -3.60508889e-01
3.06331307e-01 -1.02471960e+00 4.08917159e-01 -1.08844399e-01
1.24510273e-01 -3.36282440e-02 1.29509158e-03 6.47971690e-01
-4.53154862e-01 -6.27052486e-01 6.15257621e-02 -1.33596763e-01
-9.82624888e-01 3.27515543e-01 7.68123120e-02 1.26014695e-01
1.01388419e+00 -2.58001477e-01 -7.32245445e-01 1.05621433e-02
-1.07781053e+00 4.34750706e-01 3.02969038e-01 4.93435264e-01
5.49444556e-01 -1.55406380e+00 -4.13622379e-01 -1.20714210e-01
4.90894794e-01 3.25825602e-01 -1.34072199e-01 7.16107786e-01
-2.28094980e-01 3.07832032e-01 1.61810309e-01 -1.99774262e-02
-1.11424255e+00 6.16409481e-01 -3.83064412e-02 -4.98182535e-01
-8.59238625e-01 5.35213053e-01 6.37109429e-02 -7.75490940e-01
-3.88028547e-02 -4.90778118e-01 -4.13539529e-01 3.44834387e-01
1.11659169e-01 1.72839656e-01 1.19366564e-01 -7.26901174e-01
-3.24454933e-01 2.31449410e-01 -1.11782715e-01 2.42976770e-01
1.43487525e+00 -4.70048673e-02 -5.75988650e-01 3.16176534e-01
1.06131220e+00 2.52968371e-01 -4.54782814e-01 -2.54686803e-01
4.66703743e-01 -3.24178845e-01 -2.08893225e-01 -5.32781839e-01
-8.28226984e-01 2.28787929e-01 -1.64107397e-01 5.09193540e-01
5.16126513e-01 4.79184031e-01 9.48069870e-01 5.30808985e-01
8.77659976e-01 -9.46632743e-01 -5.49073219e-01 7.61375546e-01
7.05480039e-01 -9.68438685e-01 1.64801776e-01 -1.02865613e+00
-4.50565279e-01 1.13893569e+00 5.29681325e-01 1.51588365e-01
8.07436705e-01 1.35486349e-01 -2.96890408e-01 -6.78548872e-01
-7.34052420e-01 -4.63404208e-01 1.60332635e-01 5.98831654e-01
1.69285119e-01 3.67145270e-01 -3.97934973e-01 1.01193893e+00
-3.83608252e-01 -1.95565641e-01 6.93387091e-01 9.95121300e-01
-2.27661341e-01 -1.33321917e+00 -1.54538155e-01 7.16214836e-01
-3.43782783e-01 -3.68109763e-01 -2.21981302e-01 9.88442600e-01
2.73531079e-01 7.58912444e-01 6.60662800e-02 -3.10191363e-01
4.19424295e-01 3.93060625e-01 3.63566458e-01 -9.92151976e-01
-3.84350896e-01 -6.88500822e-01 6.99631393e-01 -1.52443230e-01
-3.74952376e-01 -6.43470109e-01 -1.88132536e+00 -2.94194877e-01
-6.48332953e-01 3.95054340e-01 4.44801062e-01 9.71759140e-01
3.56646836e-01 4.60810959e-01 1.25842184e-01 -3.07360142e-01
-7.71046355e-02 -7.60655165e-01 -8.33800137e-01 3.74129236e-01
-1.81633398e-01 -1.04010308e+00 -1.72288328e-01 1.29863694e-02] | [9.037094116210938, 8.150567054748535] |
be4b08c6-e7c7-4a7b-b1c7-fe8caad69c81 | a-tool-for-extracting-conversational | null | null | https://aclanthology.org/L12-1039 | https://aclanthology.org/L12-1039.pdf | A Tool for Extracting Conversational Implicatures | Explicitly conveyed knowledge represents only a portion of the information communicated by a text snippet. Automated mechanisms for deriving explicit information exist; however, the implicit assumptions and default inferences that capture our intuitions about a normal interpretation of a communication remain hidden for automated systems, despite the communication participants' ease of grasping the complete meaning of the communication. In this paper, we describe a reasoning framework for the automatic identification of conversational implicatures conveyed by real-world English and Arabic conversations carried via twitter.com. Our system transforms given utterances into deep semantic logical forms. It produces a variety of axioms that identify lexical connections between concepts, define rules of combining semantic relations, capture common-sense world knowledge, and encode Grice's Conversational Maxims. By exploiting this rich body of knowledge and reasoning within the context of the conversation, our system produces entailments and implicatures conveyed by analyzed utterances with an F-measure of 70.42{\%} for English conversations. | ['Marta Tatu', 'Dan Moldovan'] | 2012-05-01 | null | null | null | lrec-2012-5 | ['implicatures'] | ['natural-language-processing'] | [ 1.49628460e-01 1.02877843e+00 -1.16798215e-01 -7.50803709e-01
-3.42307746e-01 -7.42658317e-01 1.10064912e+00 3.22285891e-01
5.29210977e-02 8.86383832e-01 9.82226670e-01 -5.63280880e-01
-2.43472546e-01 -9.52320158e-01 -4.04171079e-01 -1.02935284e-02
2.51102060e-01 6.89833403e-01 -8.63714516e-03 -8.56357276e-01
6.63598031e-02 -8.75324011e-02 -1.40552640e+00 1.14250064e+00
4.74990696e-01 9.04969633e-01 -1.50215412e-02 5.41891277e-01
-9.29370463e-01 1.76710093e+00 -7.86053419e-01 -1.11087334e+00
-4.85319287e-01 -4.87226963e-01 -1.63795710e+00 1.33337274e-01
-1.61762774e-01 -3.36839855e-01 -9.20035690e-02 1.01740742e+00
-3.07000995e-01 -1.52378067e-01 5.61702371e-01 -1.63730741e+00
-5.23136199e-01 1.66122508e+00 3.71002287e-01 -5.01140282e-02
1.03256428e+00 -8.60802606e-02 1.43581927e+00 -4.28591818e-01
6.31356895e-01 1.69865763e+00 7.19645202e-01 6.66697323e-01
-8.75959992e-01 -4.44817722e-01 9.74036828e-02 2.30013683e-01
-1.33155274e+00 -6.56183064e-01 5.62167704e-01 -3.92651469e-01
1.31455278e+00 6.71847165e-01 5.53786695e-01 1.28152478e+00
-1.57298133e-01 7.93103516e-01 8.48910272e-01 -4.41636145e-01
1.56856865e-01 6.55460775e-01 3.93469959e-01 6.01783156e-01
1.22078948e-01 -5.65150261e-01 -7.26321638e-01 -5.06192088e-01
1.87127605e-01 -3.33328903e-01 -1.33904532e-01 4.77743894e-01
-9.72281098e-01 9.39588130e-01 -2.44664606e-02 4.25984859e-01
-3.39695901e-01 1.20538436e-01 4.05715346e-01 5.49995363e-01
2.13913798e-01 5.25562584e-01 -6.21304393e-01 -6.38571382e-01
-2.75987744e-01 3.80992532e-01 1.69746935e+00 1.24357367e+00
6.42630935e-01 -5.81130981e-01 3.22212189e-01 6.09245837e-01
4.91174877e-01 6.17679119e-01 3.14851642e-01 -1.68546474e+00
2.59710521e-01 9.45395529e-01 4.04904038e-01 -1.35176444e+00
-3.90491188e-01 -9.65069085e-02 -1.83700711e-01 -6.85376048e-01
5.62439203e-01 -4.96281475e-01 3.17340881e-01 1.76204371e+00
1.61872119e-01 -1.68867886e-01 7.52843916e-01 5.26053786e-01
1.03473341e+00 5.23485482e-01 5.66874035e-02 -2.44939789e-01
1.59703910e+00 -3.21676821e-01 -1.17208672e+00 -3.55332583e-01
9.60507751e-01 -7.32162714e-01 1.08295441e+00 1.34750724e-01
-9.64891493e-01 2.59564787e-01 -9.36170757e-01 -1.04413353e-01
-4.29598421e-01 -3.95512760e-01 9.78013396e-01 5.37838101e-01
-7.49141276e-01 3.34404632e-02 -2.21390337e-01 -3.77921104e-01
7.44507909e-02 3.69652407e-03 2.28192154e-02 4.87931743e-02
-1.75995815e+00 9.03050601e-01 5.03873825e-01 -8.50211754e-02
-2.82762527e-01 -4.17092085e-01 -1.00810730e+00 1.80541530e-01
7.33771324e-01 -7.27376938e-01 1.85529137e+00 -1.29283357e+00
-1.68748939e+00 9.70426798e-01 -4.21596408e-01 -6.34304225e-01
3.74146610e-01 -1.46813005e-01 -6.45795822e-01 3.16990137e-01
2.72525758e-01 4.97409701e-01 2.48111442e-01 -1.18255889e+00
-6.45632982e-01 -6.32397681e-02 9.62219059e-01 3.38662058e-01
-1.85861260e-01 2.80465066e-01 2.84162425e-02 -1.37498394e-01
1.12981349e-01 -7.17691004e-01 3.40465188e-01 -2.71639377e-01
-5.49624622e-01 -3.64645511e-01 6.86999440e-01 -4.49707836e-01
1.23215985e+00 -1.97612858e+00 -3.47865134e-01 3.61691594e-01
2.92052448e-01 -3.26749802e-01 3.84472936e-01 9.12819088e-01
3.08889866e-01 3.73522490e-01 5.23390621e-02 -7.86719844e-03
6.62888706e-01 7.26400733e-01 -7.96520472e-01 -2.79980246e-02
-2.08323933e-02 8.25857878e-01 -1.10165608e+00 -5.16056180e-01
2.30537295e-01 2.24473104e-01 -5.27199745e-01 1.41954988e-01
-8.90428841e-01 -1.29889011e-01 -5.41036189e-01 3.33043225e-02
2.67202705e-01 -4.11585182e-01 7.36257136e-01 -1.10063637e-02
2.28148937e-01 1.12185872e+00 -1.00888324e+00 1.29758680e+00
-6.29278183e-01 7.87435055e-01 2.81468511e-01 -5.22173762e-01
7.91177630e-01 5.06190121e-01 -1.14789367e-01 -9.91307572e-02
3.77590954e-01 -3.92521210e-02 -1.31192759e-01 -9.24574196e-01
3.57037455e-01 -6.73581123e-01 -4.64111328e-01 9.79996860e-01
-3.22065860e-01 -4.36575890e-01 1.55364657e-02 7.59881616e-01
8.67945910e-01 -3.63992155e-01 5.75022399e-01 -3.52791488e-01
6.95523262e-01 3.40665787e-01 7.69868344e-02 7.51054823e-01
1.42380521e-01 -2.13987127e-01 8.98259342e-01 -5.36980748e-01
-4.65374470e-01 -9.15327847e-01 -3.33327963e-03 1.08965194e+00
3.04974884e-01 -9.39143836e-01 -9.17572200e-01 -2.99203783e-01
-1.63487673e-01 1.60574520e+00 -2.58147538e-01 5.85271337e-04
-2.33977661e-01 -2.77618587e-01 1.06821048e+00 2.82788604e-01
6.31342351e-01 -8.98652315e-01 -5.82901180e-01 1.99874938e-01
-1.11660028e+00 -1.81180584e+00 1.51051790e-01 -2.99569041e-01
-2.30970740e-01 -1.35956812e+00 6.73131168e-01 -3.25104505e-01
4.86321777e-01 -1.54368952e-01 1.22652364e+00 3.52937043e-01
2.47357383e-01 5.69195211e-01 -4.62290764e-01 -5.55030584e-01
-1.03046942e+00 -4.60191250e-01 -1.81682944e-01 1.18438520e-01
9.57048714e-01 -3.20516407e-01 -1.03574827e-01 3.27699721e-01
-6.69054806e-01 2.68083155e-01 -1.28130680e-02 4.01238889e-01
-3.63095880e-01 2.80931801e-01 3.53288710e-01 -1.14041841e+00
1.20970798e+00 -7.41204858e-01 1.01435117e-01 2.11819738e-01
-2.53974386e-02 2.06963226e-01 5.49651444e-01 2.13428326e-02
-1.73211098e+00 -5.69180191e-01 -1.79823115e-01 5.69237351e-01
-4.94228452e-01 7.44505167e-01 -1.85051754e-01 6.43601298e-01
6.74096465e-01 7.93487504e-02 1.65726200e-01 9.32624936e-02
7.39005625e-01 1.30725718e+00 6.44396245e-01 -1.25496793e+00
2.80617833e-01 7.21277654e-01 -4.97549415e-01 -1.12255645e+00
-1.44203031e+00 -3.75228643e-01 -1.81454524e-01 -2.21844539e-01
6.69360280e-01 -8.50906849e-01 -1.40292287e+00 -1.03195198e-02
-1.51481426e+00 -3.21450979e-01 -2.39765719e-01 2.49898389e-01
-7.07936883e-01 3.95411789e-01 -7.02163637e-01 -9.83712852e-01
-1.26650229e-01 -7.79572606e-01 6.94708228e-01 -3.77649218e-01
-1.56112039e+00 -1.23581314e+00 -4.84081030e-01 7.51190782e-01
3.45339775e-01 -3.46409008e-02 1.16719007e+00 -1.08643687e+00
-8.62407982e-02 -4.00217362e-02 -2.05281153e-01 1.75999865e-01
3.01529020e-01 -7.97549263e-03 -9.97288406e-01 5.11394799e-01
1.18843704e-01 -4.56455976e-01 -1.19802922e-01 -1.15381859e-01
7.16925025e-01 -1.17504776e+00 -1.39987111e-01 -2.06809655e-01
8.14780235e-01 4.41234075e-02 4.85462934e-01 1.22388624e-01
-3.34141374e-01 1.09738338e+00 1.28101021e-01 6.09561920e-01
9.68322158e-01 2.61575222e-01 1.39282420e-01 6.44621730e-01
1.18600354e-01 -4.99988377e-01 3.69312912e-01 5.90134680e-01
1.80632710e-01 -8.21991935e-02 -1.05618942e+00 3.83409858e-01
-1.72196054e+00 -1.38418591e+00 -2.21064135e-01 1.33568048e+00
1.35745668e+00 4.38038677e-01 -2.87949592e-01 5.93236871e-02
6.92678928e-01 -8.37592483e-02 -1.26524931e-02 -7.00415611e-01
-2.06338078e-01 -3.20273697e-01 -1.37584105e-01 1.20700669e+00
-5.02214670e-01 9.67594743e-01 6.60387945e+00 3.15626889e-01
-6.72233999e-01 -1.08829318e-02 2.38956317e-01 1.39220014e-01
-7.56222010e-01 1.35276511e-01 -6.41310811e-01 2.03833029e-01
1.15997458e+00 -5.23897469e-01 4.96542811e-01 5.72759390e-01
1.41627103e-01 -7.88496211e-02 -1.44393206e+00 7.44309902e-01
7.42268264e-02 -1.71052206e+00 2.55219966e-01 -4.21329379e-01
3.09301525e-01 -3.14482927e-01 -4.28292334e-01 2.45878115e-01
8.17221880e-01 -9.45184350e-01 9.36980188e-01 4.03667629e-01
3.65295976e-01 -5.26269615e-01 9.98398066e-01 7.96412230e-01
-6.46606445e-01 -2.17923462e-01 3.72379534e-02 -9.60266709e-01
4.14302424e-02 4.83911365e-01 -1.14694655e+00 1.09559327e-01
3.10439736e-01 4.71598119e-01 1.72708612e-02 -1.20188043e-01
-5.94660342e-01 5.42261600e-01 -5.54264605e-01 -6.53194726e-01
3.76985013e-01 3.07580251e-02 5.82089305e-01 1.60777211e+00
-1.89635605e-01 6.39923871e-01 -4.45567258e-02 1.11050344e+00
-1.81525081e-01 -1.86779529e-01 -6.30821168e-01 -2.64071405e-01
9.21733022e-01 7.49886453e-01 -5.23650885e-01 -6.85353518e-01
-3.81110609e-01 5.18604040e-01 5.18828258e-02 4.09394383e-01
-5.17126977e-01 -2.05595672e-01 9.02055860e-01 3.62644903e-02
-2.63339370e-01 -8.44845735e-03 -1.16447181e-01 -1.11106610e+00
1.75986681e-02 -9.62295771e-01 4.33887005e-01 -1.17037904e+00
-1.37783206e+00 3.13803136e-01 4.10488516e-01 -3.16855103e-01
-6.64916158e-01 -6.21540189e-01 -6.60551608e-01 4.77631450e-01
-1.13285279e+00 -8.41007829e-01 -2.59708166e-01 5.38631678e-01
4.65840518e-01 1.56731680e-01 1.34327447e+00 -2.87821919e-01
5.82756326e-02 1.88321739e-01 -6.47771299e-01 4.77697521e-01
3.33171338e-01 -1.07291853e+00 4.44761626e-02 1.72853380e-01
-6.63870480e-03 1.28325868e+00 1.34712815e+00 -3.63609642e-01
-1.34240663e+00 -5.82701147e-01 1.47377980e+00 -7.14577973e-01
1.26126993e+00 -1.06818169e-01 -6.87165499e-01 1.37738442e+00
6.24986172e-01 -8.53872597e-01 1.30805230e+00 3.51557195e-01
-7.59404600e-01 2.08736971e-01 -1.29908061e+00 7.99741745e-01
1.03776789e+00 -9.30906355e-01 -1.64890790e+00 5.11754513e-01
9.57181394e-01 -3.17154884e-01 -6.24340236e-01 -1.81774750e-01
4.67959285e-01 -8.01561475e-01 7.12991297e-01 -8.74195337e-01
5.97157896e-01 -1.26137525e-01 -6.02484047e-01 -1.07447994e+00
5.08743584e-01 -1.02213275e+00 -9.21704620e-02 1.20247626e+00
6.58161819e-01 -7.34506488e-01 3.42940599e-01 1.22554302e+00
-7.07676867e-03 -2.24363372e-01 -6.65122867e-01 -2.42901057e-01
-5.17165773e-02 -1.20646775e+00 8.60449493e-01 1.28136218e+00
1.22695315e+00 7.47900188e-01 3.64594519e-01 1.12289123e-01
6.29924238e-01 1.77701548e-01 5.44925272e-01 -1.12421024e+00
-1.67911336e-01 -4.56976891e-01 -2.81419486e-01 -1.20243120e+00
7.23039329e-01 -1.05965400e+00 -1.71471864e-01 -1.54650831e+00
1.30780801e-01 -1.77989662e-01 6.15845084e-01 6.50249898e-01
5.40593266e-01 -2.33373001e-01 -1.33544713e-01 -6.93918169e-02
-6.91108823e-01 2.27462426e-01 8.63710761e-01 -2.25569934e-01
7.87991434e-02 -2.98618972e-01 -1.29443645e+00 1.36972177e+00
7.91428804e-01 -1.29263669e-01 -7.23852575e-01 -4.15441602e-01
1.01967990e+00 1.21908695e-01 5.32814443e-01 -3.79558504e-01
4.42997843e-01 -6.13788843e-01 -3.40929240e-01 -1.51088923e-01
3.98485243e-01 -8.80170166e-01 6.61040619e-02 2.34419763e-01
-1.05135679e+00 -2.21940607e-01 7.18076751e-02 2.47901380e-01
-3.90798897e-01 -3.84659976e-01 2.07377344e-01 -3.16140443e-01
-7.36317754e-01 -7.80708373e-01 -8.94374549e-01 5.59021592e-01
7.95441449e-01 -3.12343445e-02 -5.25732994e-01 -1.08647501e+00
-9.80352283e-01 2.80355334e-01 9.20123328e-03 3.25936466e-01
7.59663463e-01 -7.82244921e-01 -7.18379080e-01 3.57738808e-02
2.14186788e-01 -2.50694096e-01 -1.31648272e-01 5.32724261e-01
-5.86100161e-01 4.65813786e-01 3.32847536e-02 -2.48950109e-01
-1.18998349e+00 -2.06000470e-02 4.57136363e-01 4.60241526e-01
-7.20542252e-01 7.40530193e-01 -9.67776179e-02 -6.16176903e-01
2.41090924e-01 -5.90307295e-01 -9.34186503e-02 -5.00514805e-02
9.43799913e-01 1.01032391e-01 -1.66843399e-01 -6.67279840e-01
-3.45605373e-01 -1.50750965e-01 1.68127075e-01 -1.80647701e-01
1.00208867e+00 -6.18142247e-01 -7.01179326e-01 5.58650911e-01
1.06264317e+00 6.01128712e-02 -4.43966717e-01 -4.17494982e-01
2.46470317e-01 -1.58636317e-01 -3.47639412e-01 -1.00739181e+00
-1.04594246e-01 4.93071645e-01 -8.22471857e-01 6.86150372e-01
3.96105707e-01 5.94760001e-01 8.27176094e-01 1.18665361e+00
3.57895821e-01 -1.05909574e+00 -6.66340142e-02 8.22082996e-01
1.02066469e+00 -8.83569658e-01 -1.90626234e-01 -9.41655695e-01
-8.94272685e-01 1.36511397e+00 3.09637010e-01 5.69500625e-01
5.65399230e-01 7.38041341e-01 3.98287445e-01 -6.69818342e-01
-1.21552908e+00 5.96223138e-02 -4.47097540e-01 3.82448971e-01
5.47974706e-01 4.55667377e-01 -1.66646361e-01 1.04303622e+00
-1.07983470e+00 -1.87382564e-01 1.00871146e+00 7.43800700e-01
-6.27197027e-01 -4.08755660e-01 -1.25299498e-01 4.92946617e-02
-4.40296412e-01 -2.17158392e-01 -1.02789009e+00 7.69196212e-01
-1.77607059e-01 1.66597414e+00 7.62220174e-02 -2.28433177e-01
5.44728562e-02 4.96179253e-01 2.73728639e-01 -6.43589020e-01
-4.65662867e-01 -6.26776636e-01 1.09239054e+00 -4.02675658e-01
-7.21464574e-01 -4.82448608e-01 -2.00799537e+00 -8.88262093e-01
9.19704884e-02 7.22770870e-01 3.84988308e-01 1.72078085e+00
6.57641441e-02 1.21894933e-01 9.79295447e-02 1.80277407e-01
-7.08086431e-01 -7.42489696e-01 -8.64179656e-02 8.22997570e-01
2.01684445e-01 -2.86840320e-01 -6.22851908e-01 3.25353146e-01] | [12.345514297485352, 8.011598587036133] |
e87391de-f084-489d-8ba2-1b88a756d090 | conformer-llms-convolution-augmented-large | 2307.00461 | null | https://arxiv.org/abs/2307.00461v1 | https://arxiv.org/pdf/2307.00461v1.pdf | Conformer LLMs -- Convolution Augmented Large Language Models | This work builds together two popular blocks of neural architecture, namely convolutional layers and Transformers, for large language models (LLMs). Non-causal conformers are used ubiquitously in automatic speech recognition. This work aims to adapt these architectures in a causal setup for training LLMs. Transformers decoders effectively capture long-range dependencies over several modalities and form a core backbone of modern advancements in machine learning. Convolutional architectures have been popular in extracting features in domains such as raw 1-D signals, speech, and images, to name a few. In this paper, by combining local and global dependencies over latent representations using causal convolutional filters and Transformer, we achieve significant gains in performance. This work showcases a robust speech architecture that can be integrated and adapted in a causal setup beyond speech applications for large-scale language modeling. | ['Prateek Verma'] | 2023-07-02 | null | null | null | null | ['speech-recognition'] | ['speech'] | [ 8.68578106e-02 4.10859197e-01 -4.51383948e-01 -5.95222175e-01
-6.49056137e-01 -3.11300159e-01 1.34905064e+00 -2.55789250e-01
6.38915747e-02 5.15906990e-01 1.00511467e+00 -3.96719515e-01
-1.98983312e-01 -7.31465101e-01 -9.22775447e-01 -5.04413068e-01
-2.75085688e-01 3.14396948e-01 7.93392137e-02 -5.21626808e-02
-2.46331096e-01 6.65301085e-01 -1.22647154e+00 7.43107617e-01
9.86500457e-02 8.56667399e-01 4.89709228e-02 8.73545110e-01
-2.37621441e-01 1.57428694e+00 -3.25168490e-01 -2.91646570e-01
-5.41897535e-01 -2.81108916e-01 -8.47812891e-01 -5.00667170e-02
4.03422058e-01 -3.46783608e-01 -9.21730340e-01 3.73765796e-01
3.78974855e-01 -1.88256711e-01 6.76683426e-01 -8.08278620e-01
-6.27525687e-01 1.13926244e+00 -5.41607440e-02 3.55074346e-01
3.10749799e-01 -9.26403701e-02 1.20388758e+00 -8.63314509e-01
3.38172942e-01 1.89994025e+00 6.27701700e-01 5.55371046e-01
-1.23563254e+00 -4.73354667e-01 3.35462034e-01 3.43087733e-01
-7.45421886e-01 -9.79260564e-01 8.33148062e-01 -4.75291789e-01
1.65336406e+00 2.58725230e-03 2.66576618e-01 1.94879770e+00
2.81021297e-01 1.14490283e+00 9.23017561e-01 -4.84839022e-01
-4.43652570e-02 -1.74478620e-01 -9.00731608e-02 5.51502168e-01
-5.51146984e-01 4.05952275e-01 -1.28207338e+00 -2.17783432e-02
1.04337013e+00 -3.87615144e-01 1.85670480e-01 -1.01960905e-01
-1.38598871e+00 7.80173779e-01 4.46613073e-01 5.97522020e-01
-3.52184027e-01 8.74357641e-01 3.63590389e-01 3.42893481e-01
5.77765465e-01 -4.83441763e-02 -5.61880887e-01 -3.18014354e-01
-8.41115654e-01 -1.65737391e-01 7.95562208e-01 6.77254975e-01
1.44097269e-01 3.46615285e-01 -7.49345198e-02 9.62605655e-01
5.77879369e-01 5.44755757e-01 4.42970514e-01 -5.09183288e-01
3.78734767e-01 3.13248426e-01 -4.47329432e-01 -5.02349377e-01
-4.88954931e-01 -5.74480295e-01 -1.04351151e+00 -2.49809653e-01
-4.37107719e-02 1.33574560e-01 -1.04988265e+00 1.90374982e+00
-5.71424514e-02 6.81573093e-01 5.87166250e-02 5.40427446e-01
6.87015831e-01 6.47506535e-01 3.32538366e-01 -2.30137601e-01
1.18198335e+00 -5.60846210e-01 -8.18481088e-01 -3.73033375e-01
2.48078242e-01 -8.31330538e-01 7.13611603e-01 4.89528328e-01
-1.20711815e+00 -4.61942285e-01 -1.03157830e+00 -2.56136268e-01
-3.73246789e-01 6.14389405e-02 7.96759903e-01 3.45992237e-01
-1.27245677e+00 7.33049214e-01 -1.18105352e+00 -3.93048674e-01
3.31355602e-01 4.01127458e-01 -3.65241170e-01 2.08181173e-01
-1.61890161e+00 8.63754809e-01 -2.29772925e-02 -1.63481180e-02
-1.49160814e+00 -6.27719820e-01 -7.01078773e-01 -5.64854518e-02
7.12023601e-02 -7.63567805e-01 1.60982049e+00 -8.04921687e-01
-1.96842408e+00 6.66613936e-01 -3.78195286e-01 -9.38776016e-01
2.77490556e-01 -4.36838001e-01 -4.87692863e-01 -1.23580381e-01
-2.67148882e-01 2.97830641e-01 1.31251872e+00 -8.12871337e-01
-2.32929602e-01 -2.60358036e-01 -2.99238041e-02 -2.18130782e-01
-4.93277311e-01 3.05602849e-01 -2.81574249e-01 -7.64372289e-01
-6.74731061e-02 -6.72015667e-01 -7.79646412e-02 -5.55216789e-01
-3.94855857e-01 -4.94252115e-01 9.71599162e-01 -4.25689638e-01
1.07054067e+00 -1.84971285e+00 5.53518474e-01 -1.09198801e-01
1.95514679e-01 1.41454741e-01 -1.99951306e-01 5.96951544e-01
-3.74341547e-01 1.37886494e-01 3.73740196e-02 -8.06505859e-01
1.44357368e-01 5.10159433e-01 -9.28346932e-01 2.66403615e-01
5.51042557e-01 1.20314169e+00 -7.70748019e-01 -4.00451981e-02
5.25429070e-01 7.99693167e-01 -2.82357395e-01 4.20993030e-01
-6.24055922e-01 5.52807748e-01 -3.40123773e-01 1.17921300e-01
1.21403880e-01 -3.53368938e-01 3.22782576e-01 -6.49568290e-02
-1.47015497e-01 1.15399718e+00 -7.52844334e-01 1.99840248e+00
-1.18879473e+00 9.14338648e-01 2.83781737e-02 -1.21815872e+00
7.34410942e-01 9.22220170e-01 4.53965247e-01 -8.77389550e-01
-9.54832956e-02 1.63652182e-01 -2.17692867e-01 -4.32502240e-01
1.04242668e-01 -3.97128016e-01 -1.21735767e-01 3.84558886e-01
6.46490872e-01 -2.04415172e-01 -1.61078095e-01 3.63303244e-01
1.24465036e+00 1.04104765e-01 1.50131345e-01 -9.46477354e-02
5.01959741e-01 -4.99440402e-01 -6.94458466e-03 6.03618562e-01
2.66517431e-01 4.92840171e-01 7.21333265e-01 -5.14077246e-01
-1.14549839e+00 -1.24848759e+00 -1.53736070e-01 1.19881582e+00
-5.80409288e-01 -5.59839904e-01 -3.31475765e-01 -5.71378708e-01
-2.18299180e-01 7.69582510e-01 -6.03030801e-01 -1.47574201e-01
-7.11478889e-01 -6.19418204e-01 9.24989641e-01 8.34328175e-01
2.43206397e-01 -1.00053382e+00 2.62270141e-02 4.10084277e-01
-4.55489010e-02 -1.43788218e+00 5.06881066e-02 4.87476349e-01
-8.01509082e-01 -6.79231524e-01 -4.58478540e-01 -3.83711487e-01
-1.68216661e-01 -1.97442710e-01 1.22675693e+00 -4.37051147e-01
-6.59684762e-02 5.86487532e-01 -5.79097122e-02 -3.45685303e-01
-7.70170569e-01 1.65775254e-01 3.84980261e-01 1.54881939e-01
2.74358857e-02 -1.11474431e+00 -1.73764408e-01 1.96833685e-01
-6.44600034e-01 8.88583437e-02 7.19629228e-01 8.36005509e-01
8.56328383e-02 -4.11453158e-01 6.35624886e-01 -5.13340473e-01
6.87780440e-01 -6.27742827e-01 -2.76443243e-01 5.43511100e-02
-2.13568255e-01 6.15987301e-01 7.13498950e-01 -3.27773124e-01
-1.38833892e+00 -1.31720662e-01 -4.82290417e-01 -5.23735344e-01
-4.13970023e-01 7.26828694e-01 -3.39903720e-02 6.07631207e-01
5.94106257e-01 3.77091318e-02 -2.40271047e-01 -6.86108112e-01
7.60672212e-01 6.96804404e-01 5.35533607e-01 -7.03925192e-01
4.29103523e-01 5.49912810e-01 2.51930237e-01 -9.45130587e-01
-8.51436734e-01 -4.01884019e-01 -8.12572002e-01 4.49341498e-02
8.61994922e-01 -1.19396138e+00 -5.12345195e-01 5.35222948e-01
-1.44740629e+00 -7.50937045e-01 -1.15996897e-01 6.12906456e-01
-6.91195309e-01 -1.55207649e-01 -8.01530778e-01 -8.00145328e-01
1.75090030e-01 -8.54141235e-01 1.34068429e+00 -9.41229239e-02
-1.36847526e-01 -1.37667596e+00 1.62506163e-01 1.59862906e-01
5.06419480e-01 -2.20071822e-01 7.66519725e-01 -4.18284327e-01
-8.18683922e-01 8.47546011e-02 -3.25558782e-02 4.40626025e-01
2.67567523e-02 6.89219832e-02 -1.48314452e+00 9.87102315e-02
-2.54608303e-01 -4.83485132e-01 1.25089467e+00 4.59484249e-01
9.11759794e-01 -2.23999187e-01 -3.91883671e-01 5.09474993e-01
7.73879111e-01 -6.23609014e-02 5.51279664e-01 -2.45037839e-01
6.88985646e-01 5.18303812e-01 -1.81734711e-01 3.39390039e-01
2.75095224e-01 8.50637794e-01 5.04945159e-01 -1.02890082e-01
-6.69129014e-01 -5.70985973e-01 7.48287916e-01 1.31081533e+00
-3.46148983e-02 -3.91771436e-01 -1.11665857e+00 7.02214181e-01
-1.99651325e+00 -1.04484761e+00 -2.32319802e-01 1.75620091e+00
8.21447730e-01 9.65768024e-02 -2.99028665e-01 -1.66120842e-01
1.40727282e-01 4.09610927e-01 -3.15393418e-01 -4.96163338e-01
-3.08733702e-01 5.45831263e-01 2.13361099e-01 5.90989172e-01
-1.15595233e+00 9.80009854e-01 6.53709745e+00 7.98794091e-01
-1.42591989e+00 3.48982215e-01 5.52364588e-01 -5.86535558e-02
-3.88050854e-01 3.07699572e-02 -7.66510904e-01 -8.00125673e-02
1.70506346e+00 1.85240477e-01 3.67079020e-01 4.73177701e-01
5.49660206e-01 3.74775529e-01 -1.40375566e+00 7.33463943e-01
-2.19182491e-01 -1.74029171e+00 1.11962631e-01 2.92645134e-02
6.87461019e-01 5.48360884e-01 2.81592578e-01 1.72043800e-01
7.89076447e-01 -1.56320488e+00 8.27638745e-01 5.82469344e-01
9.38658178e-01 -4.41843987e-01 3.38819057e-01 3.24884176e-01
-1.25736392e+00 -1.56747445e-01 -6.26131007e-03 -2.50682384e-01
3.87468904e-01 6.53135121e-01 -9.65982854e-01 5.57472289e-01
7.02039540e-01 1.10588694e+00 -1.07150443e-01 2.46293038e-01
-7.47671723e-01 1.07323098e+00 -4.66957778e-01 1.02670841e-01
3.45245540e-01 5.43487906e-01 4.97469157e-01 1.48254645e+00
8.35955888e-02 -2.46733099e-01 -3.29933345e-01 8.40271354e-01
-2.67531574e-01 -2.89443910e-01 -7.83425212e-01 -2.94481367e-01
1.95049062e-01 8.09094369e-01 -8.01737905e-02 -3.48839641e-01
-5.78129649e-01 7.95131624e-01 3.87518585e-01 3.12033951e-01
-8.99094641e-01 4.57585394e-01 9.28939104e-01 9.73780528e-02
1.94240600e-01 -6.78541780e-01 -2.57536083e-01 -1.27946508e+00
-2.87873000e-01 -6.03478193e-01 2.78353274e-01 -8.29014540e-01
-1.39045489e+00 4.74295288e-01 1.51374266e-01 -7.94956625e-01
-7.65068769e-01 -8.21386039e-01 -5.12525558e-01 9.38648343e-01
-1.51266694e+00 -1.69468117e+00 8.50074813e-02 9.01608586e-01
8.87951851e-01 -2.56566137e-01 9.66932535e-01 2.77036697e-01
-4.18468714e-01 1.81311712e-01 -1.00740045e-01 1.20621599e-01
4.60845768e-01 -1.21759415e+00 8.42466056e-01 9.66037512e-01
7.61973858e-01 8.91052604e-01 4.62971300e-01 -4.42086011e-01
-1.41839492e+00 -9.37306404e-01 1.16773891e+00 -4.11099344e-01
1.14826179e+00 -7.48215258e-01 -5.54331303e-01 9.22394335e-01
5.12230098e-01 -2.91220304e-02 2.22283125e-01 5.72245657e-01
-7.18139231e-01 -5.35773067e-03 -1.72637165e-01 5.80325603e-01
1.16426218e+00 -1.25441551e+00 -5.06843984e-01 2.93312848e-01
8.25829744e-01 -2.53818929e-01 -8.70819986e-01 3.60785514e-01
4.71417636e-01 -1.05885875e+00 1.32801640e+00 -6.57150030e-01
6.49880350e-01 6.11176267e-02 -2.15532824e-01 -1.40279782e+00
-1.47341296e-01 -1.07358742e+00 -4.10802811e-01 1.17303097e+00
3.87915909e-01 -5.77647567e-01 3.78906876e-01 1.58726409e-01
-3.11154753e-01 -4.34988767e-01 -1.32502782e+00 -5.20440638e-01
4.11388665e-01 -1.17536914e+00 4.15366888e-01 7.30985165e-01
-5.14570437e-02 8.61155748e-01 -5.28131783e-01 3.92564148e-01
6.19699776e-01 -2.05295146e-01 5.73843181e-01 -1.13547170e+00
-3.74384493e-01 -6.35536909e-01 -2.70327568e-01 -1.55775428e+00
4.50148255e-01 -8.57436001e-01 -2.35356659e-01 -1.53419793e+00
-3.57440747e-02 -5.05685270e-01 -3.06723595e-01 5.79274714e-01
4.60323095e-01 -6.12525083e-02 -6.50148988e-02 1.96092576e-01
-1.01763785e-01 6.81388438e-01 7.73637712e-01 -2.55243093e-01
1.06451325e-01 -2.31964793e-02 -1.01678766e-01 7.60937393e-01
7.34737515e-01 -2.64812708e-01 -5.67346394e-01 -5.73903561e-01
4.16946262e-01 2.82161534e-01 8.22532475e-01 -7.80884087e-01
4.43991423e-01 -1.04986235e-01 1.70678616e-01 -3.38007510e-01
5.31261146e-01 -5.97220898e-01 1.09387182e-01 -7.90639082e-04
-6.66745365e-01 -2.85622001e-01 2.35509530e-01 5.84973097e-01
-5.80112755e-01 4.41165000e-01 5.34017742e-01 -2.83127904e-01
-7.57274508e-01 2.04446107e-01 -3.77531528e-01 -3.62206429e-01
5.23175657e-01 4.22259778e-01 -4.06909496e-01 -5.48277318e-01
-9.40676510e-01 -2.58301139e-01 -3.23656946e-01 8.78097713e-01
5.78412890e-01 -1.19825459e+00 -6.54931605e-01 2.89874077e-01
-9.03787240e-02 -2.47802183e-01 1.34904534e-01 1.12546289e+00
1.72033712e-01 1.07866180e+00 2.56385356e-01 -8.93128157e-01
-1.02520311e+00 3.47664922e-01 5.16901135e-01 -4.90747154e-01
-4.95085299e-01 1.20042121e+00 4.60702926e-01 -2.94113904e-01
2.38516062e-01 -6.08105600e-01 -4.46082046e-03 1.41581208e-01
4.06630009e-01 -1.89939871e-01 1.76004484e-01 -5.39764702e-01
-3.77631336e-01 1.67368621e-01 1.43067598e-01 -5.85380614e-01
1.58169186e+00 -1.62953660e-01 -1.03750974e-01 9.49900806e-01
1.17440176e+00 6.95259497e-02 -1.52155519e+00 -4.09938246e-01
6.63375109e-02 1.97642758e-01 3.07411343e-01 -9.14997876e-01
-8.36240470e-01 1.46589112e+00 2.67617375e-01 3.98034394e-01
8.07382405e-01 6.12162828e-01 5.26720762e-01 8.87151062e-02
2.78776199e-01 -7.07519412e-01 4.46380049e-01 9.11220610e-01
1.19659209e+00 -9.34619009e-01 -3.72577101e-01 7.11254627e-02
-2.67851889e-01 1.48466313e+00 -4.80485000e-02 -6.59643114e-02
7.71079659e-01 8.48594368e-01 -2.89120674e-02 -2.75907099e-01
-1.50967646e+00 -3.01806241e-01 3.42254996e-01 4.64406759e-01
8.70943904e-01 2.61975288e-01 2.75863647e-01 3.64851683e-01
5.43381274e-02 -4.53856401e-02 1.70918629e-01 4.42997664e-01
4.94718663e-02 -1.39239526e+00 -2.24368840e-01 6.35838211e-02
-3.37017745e-01 -5.10256946e-01 -3.67422640e-01 4.85153139e-01
-4.69944440e-02 1.10106850e+00 1.63346231e-01 -4.40200359e-01
1.81344375e-01 2.69519150e-01 5.50418079e-01 -3.41063619e-01
-3.77852142e-01 1.79155782e-01 5.56474626e-01 -7.12839663e-01
-7.42907703e-01 -6.30837142e-01 -1.14272940e+00 -2.34649077e-01
6.88382909e-02 -3.45714331e-01 8.69919121e-01 1.10527265e+00
3.48029405e-01 9.71552849e-01 4.20074970e-01 -7.49745905e-01
-3.62189740e-01 -1.14846337e+00 -8.96519199e-02 1.34269506e-01
6.79759741e-01 -7.00090468e-01 -1.64383233e-01 4.91022050e-01] | [10.995808601379395, 6.609679222106934] |
8728faeb-3e0f-417a-809e-2b6e24d3e58b | facial-surface-analysis-using-iso-geodesic | 1608.08878 | null | http://arxiv.org/abs/1608.08878v1 | http://arxiv.org/pdf/1608.08878v1.pdf | Facial Surface Analysis using Iso-Geodesic Curves in Three Dimensional Face Recognition System | In this paper, we present an automatic 3D face recognition system. This
system is based on the representation of human faces surfaces as collections of
Iso-Geodesic Curves (IGC) using 3D Fast Marching algorithm. To compare two
facial surfaces, we compute a geodesic distance between a pair of facial curves
using a Riemannian geometry. In the classifying step, we use: Neural Networks
(NN), K-Nearest Neighbor (KNN) and Support Vector Machines (SVM). To test this
method and evaluate its performance, a simulation series of experiments were
performed on 3D Shape REtrieval Contest 2008 database (SHREC2008). | ['Bouzid Manaut', 'El Mahdi Barrah', 'Said Safi', 'Rachid Ahdid'] | 2016-08-31 | null | null | null | null | ['3d-shape-retrieval'] | ['computer-vision'] | [-1.63879946e-01 -4.50383015e-02 1.37295336e-01 -7.60738611e-01
-5.05984873e-02 -4.02793735e-01 8.40645850e-01 -4.52728420e-01
-1.12861186e-01 1.97067261e-01 -3.25581878e-01 -4.60926056e-01
-1.27737179e-01 -7.33314335e-01 -3.15172732e-01 -3.43421906e-01
-5.30763745e-01 4.17829365e-01 8.93482566e-02 -2.65674561e-01
4.33600247e-01 1.42052448e+00 -1.68377960e+00 1.33191511e-01
1.46813720e-01 1.10889113e+00 -7.82397628e-01 5.11673212e-01
-1.01198159e-01 -2.84446895e-01 -1.74539387e-01 -5.93438327e-01
6.20316267e-01 -2.11022168e-01 -8.19631755e-01 -8.20031092e-02
5.17633498e-01 -1.07530773e-01 4.89304550e-02 1.03619564e+00
2.60906935e-01 8.28470662e-02 1.43235183e+00 -1.45291233e+00
-7.47476995e-01 -4.25900936e-01 -3.87894690e-01 -4.40785289e-02
6.85439885e-01 -4.75732356e-01 3.29221457e-01 -1.61412907e+00
7.31220126e-01 1.78020120e+00 8.78879547e-01 8.48911285e-01
-9.49303508e-01 -6.14839494e-01 -5.92345178e-01 3.69062394e-01
-1.84502113e+00 -4.33507890e-01 1.01338100e+00 -5.05924106e-01
8.80633175e-01 3.79283220e-01 7.23488033e-01 5.67085207e-01
2.55238444e-01 3.61133188e-01 1.16334045e+00 -6.35672867e-01
3.45725596e-01 8.63451585e-02 2.09674582e-01 1.17919731e+00
2.51145549e-02 1.99124575e-01 -1.11175105e-01 -6.04965091e-01
8.85309637e-01 -2.78494000e-01 1.82550669e-01 -4.39796239e-01
-2.13476241e-01 7.32980549e-01 1.43217936e-01 4.67276841e-01
-2.39141479e-01 -2.47481912e-01 -3.15122567e-02 4.80306178e-01
5.32046199e-01 -3.72298300e-01 5.24854735e-02 2.93029040e-01
-6.98711872e-01 1.85340434e-01 1.14075136e+00 5.89927137e-01
7.91460693e-01 1.56994385e-03 3.43618572e-01 7.48972476e-01
8.29291463e-01 6.47474289e-01 3.79407793e-01 -1.04105961e+00
-1.11628555e-01 8.64237607e-01 -1.63844913e-01 -1.50786209e+00
-2.30275720e-01 5.49030542e-01 -8.89777362e-01 8.13680351e-01
3.64292651e-01 1.18160218e-01 -7.72539854e-01 1.19319189e+00
7.18966246e-01 5.09641528e-01 -6.32023886e-02 8.91902506e-01
7.88852990e-01 4.09138680e-01 -2.36280248e-01 -5.46332113e-02
9.20591950e-01 -2.69457430e-01 -4.59349871e-01 9.97178197e-01
7.49617696e-01 -9.26614583e-01 4.15450931e-01 2.37017334e-01
-9.85182703e-01 -4.92920399e-01 -1.17891705e+00 1.75640434e-01
-8.04738581e-01 1.76284716e-01 1.47217780e-01 1.21521962e+00
-1.42704046e+00 1.14744198e+00 -7.72847831e-01 -3.70249838e-01
4.39649671e-01 6.73389435e-01 -6.72959507e-01 3.20898950e-01
-9.09003854e-01 9.00210977e-01 -1.98381931e-01 1.06420703e-01
-4.24853176e-01 -2.40573913e-01 -8.48785043e-01 -6.04966819e-01
-4.97972697e-01 6.41921535e-03 7.64112890e-01 -8.89071167e-01
-1.87211215e+00 1.57599306e+00 -2.84194469e-01 -3.71730216e-02
2.55669415e-01 1.13330901e-01 -6.84706569e-01 1.78841472e-01
-5.58438122e-01 4.97157484e-01 1.00223625e+00 -9.95014191e-01
1.84789486e-02 -9.68300641e-01 -3.10092717e-01 4.48344611e-02
-5.67714311e-02 2.57204056e-01 -1.46081835e-01 -3.30253601e-01
5.29730082e-01 -1.14820611e+00 4.11525816e-01 4.10163730e-01
-2.62345850e-01 -1.03490555e+00 1.34235823e+00 -7.29427874e-01
7.30592489e-01 -2.21042013e+00 5.56786619e-02 8.84152710e-01
-2.23696068e-01 4.69626546e-01 -2.08574280e-01 1.22446679e-01
-3.20909202e-01 1.67835057e-01 -3.14384043e-01 -3.35451663e-01
-1.23577498e-01 1.13913022e-01 5.39101511e-02 9.37053025e-01
2.19482481e-01 5.32400310e-01 -1.48655251e-01 -6.11992419e-01
-6.81676492e-02 7.01528966e-01 -1.96555063e-01 2.54256967e-02
2.47090176e-01 -7.20047802e-02 -4.39905822e-01 9.10969377e-01
1.04938519e+00 4.95030701e-01 -1.89816386e-01 -1.43603966e-01
9.48693752e-02 -4.56739545e-01 -1.47704816e+00 1.21829355e+00
-9.20493379e-02 4.48892891e-01 6.63609877e-02 -1.05793786e+00
1.55635512e+00 3.59002948e-01 4.00116146e-01 -2.18250483e-01
3.79584491e-01 4.44885373e-01 -4.35889959e-01 -4.10658747e-01
-1.57940924e-01 2.06887677e-01 5.97191274e-01 6.26795232e-01
5.05614206e-02 -1.18684486e-01 -2.18430147e-01 -4.94386196e-01
5.03664017e-01 2.60876477e-01 -1.06007226e-01 -6.97385848e-01
1.23170197e+00 -3.72711897e-01 8.68006125e-02 -1.42059594e-01
-3.40364665e-01 5.95610023e-01 4.32526052e-01 -8.80322933e-01
-8.91763508e-01 -1.21782112e+00 -5.29736757e-01 4.20350283e-01
-1.83720011e-02 4.26913947e-02 -1.35360289e+00 -6.55642271e-01
3.69406223e-01 5.69978893e-01 -7.30014026e-01 -7.64484555e-02
-6.94199800e-01 -3.52348417e-01 6.41296566e-01 1.00782856e-01
3.86207044e-01 -7.94701874e-01 -3.47710788e-01 -3.22469890e-01
7.21064806e-01 -7.12880790e-01 -6.35210514e-01 -6.32295609e-01
-1.07318306e+00 -1.38604641e+00 -8.94644201e-01 -1.13095021e+00
8.95591199e-01 1.79126412e-01 5.64556837e-01 1.57269493e-01
-5.87449193e-01 7.77876139e-01 -1.67321339e-02 -3.19745570e-01
-3.95829260e-01 -3.75136018e-01 7.09432542e-01 4.12486613e-01
9.72537518e-01 -5.45981824e-01 -3.65356863e-01 7.48333573e-01
-5.32560825e-01 -6.28288686e-01 1.35370031e-01 3.97987545e-01
6.51475310e-01 -2.04621255e-01 2.95618951e-01 -2.53562272e-01
8.12924385e-01 -2.20980644e-01 -9.60543871e-01 3.99802238e-01
-5.31141222e-01 -2.82022636e-02 1.84455454e-01 -3.31068099e-01
-7.56677091e-01 1.42578185e-01 -1.24121703e-01 -6.77879870e-01
-2.63717204e-01 -3.53327021e-02 -1.38293937e-01 -8.63405585e-01
9.13563669e-01 1.34846032e-01 5.71866512e-01 -6.01365805e-01
2.03776866e-01 1.05792737e+00 1.61049142e-01 -4.37713385e-01
8.36601138e-01 5.33237100e-01 5.94116092e-01 -1.28191471e+00
-1.01067774e-01 -1.43237382e-01 -1.17737603e+00 -6.06142402e-01
9.18512106e-01 -2.14752078e-01 -1.06913435e+00 5.24387896e-01
-1.39202213e+00 1.11050770e-01 2.79414266e-01 4.76710439e-01
-5.13374507e-01 3.29502672e-01 -3.69869381e-01 -1.02087259e+00
-4.18540329e-01 -1.00944448e+00 9.93023276e-01 2.47470006e-01
-1.24280766e-01 -9.14361000e-01 3.47314924e-01 -6.65593073e-02
2.25176886e-01 3.78374189e-01 9.64616656e-01 -8.02631915e-01
-1.16119675e-01 -5.66527605e-01 -1.26091957e-01 5.56557298e-01
3.34326923e-01 3.82467926e-01 -9.77422297e-01 -1.35331884e-01
7.44797736e-02 1.33134192e-02 2.06143618e-01 5.45936450e-02
1.17385221e+00 -2.59169310e-01 -3.36893737e-01 5.54643869e-01
1.17557919e+00 6.12470388e-01 3.72486919e-01 -1.58941790e-01
3.22058499e-01 1.05300498e+00 1.72667170e-03 4.15234156e-02
8.46455023e-02 7.27911055e-01 1.44708827e-01 3.63524884e-01
-1.22382715e-01 -4.83799316e-02 2.27421284e-01 7.11232424e-01
-6.20403647e-01 5.58453798e-01 -1.01650465e+00 -2.66071260e-02
-1.18842864e+00 -7.83607364e-01 -8.51026326e-02 2.40120530e+00
5.44348478e-01 -1.42719761e-01 1.92454338e-01 3.20846915e-01
8.32246959e-01 -4.69436526e-01 -2.50190020e-01 -9.09926891e-01
-7.27783218e-02 4.07306701e-01 -5.41051328e-02 6.23111367e-01
-1.00740016e+00 7.98219681e-01 6.66959620e+00 6.96539581e-01
-1.30384743e+00 -1.18102551e-01 7.87230730e-01 2.70543665e-01
8.36346671e-02 -5.66228986e-01 -9.37720954e-01 2.90107802e-02
9.33789015e-01 -2.21023589e-01 4.76175040e-01 9.58692253e-01
-2.04863980e-01 1.55836985e-01 -1.10938191e+00 1.36958361e+00
3.90206873e-01 -1.20480156e+00 2.65466899e-01 -6.93268999e-02
6.65374160e-01 -2.94793218e-01 1.39945656e-01 -5.92726022e-02
-2.21945971e-01 -1.36734712e+00 2.41625682e-01 7.41550922e-01
8.21917474e-01 -7.06131279e-01 4.54835713e-01 1.85384527e-02
-1.12436080e+00 5.24218798e-01 -5.92155635e-01 2.36572489e-01
-4.18497920e-01 7.06664920e-02 -7.39738464e-01 2.87237138e-01
6.64498389e-01 2.67817378e-01 -4.02380139e-01 9.56327796e-01
1.04413241e-01 1.60228401e-01 -6.58985019e-01 -6.04935288e-01
1.60909116e-01 -8.44459474e-01 6.58528149e-01 1.03094614e+00
4.14007187e-01 5.32180250e-01 -3.44428390e-01 8.69249821e-01
-2.48177443e-02 6.27755046e-01 -1.11763608e+00 2.58952111e-01
2.51358867e-01 1.22945273e+00 -6.63270891e-01 -4.09492068e-02
-3.01371306e-01 9.21358764e-01 -1.23460311e-02 2.41230458e-01
-2.18673319e-01 -5.20824850e-01 8.08354020e-01 2.36407593e-01
-7.45933205e-02 -5.84811568e-01 -1.55260906e-01 -5.29598832e-01
1.07962992e-02 -3.95471245e-01 1.42440572e-01 -6.21379197e-01
-1.18114698e+00 6.90276146e-01 1.06038295e-01 -1.00131655e+00
-2.67007351e-02 -1.13882327e+00 -7.00383604e-01 1.11053371e+00
-1.05249310e+00 -8.31330538e-01 3.99240404e-02 9.87077594e-01
2.60422863e-02 -6.60128951e-01 1.08716536e+00 1.77181274e-01
-1.12536073e-01 7.15559602e-01 1.01037346e-01 3.37729514e-01
3.02808285e-01 -9.32025790e-01 3.60192060e-01 -8.15494284e-02
1.87253490e-01 4.53914434e-01 5.66772781e-02 -6.05015099e-01
-1.73608649e+00 -1.02649021e+00 1.00137472e+00 -4.30168450e-01
2.42604494e-01 -2.40558028e-01 -9.75218177e-01 4.01808321e-01
-2.30555251e-01 1.68835834e-01 7.13676274e-01 -2.62900382e-01
-3.57742906e-01 -1.57804683e-01 -1.67122924e+00 6.14716172e-01
1.12912285e+00 -5.50697029e-01 -6.34364486e-01 4.38682526e-01
-4.71936241e-02 -1.27824798e-01 -1.00198281e+00 4.56622958e-01
7.58540869e-01 -8.90706003e-01 8.30545127e-01 -7.04112768e-01
-2.16261789e-01 -2.40808517e-01 -4.18421000e-01 -9.05246735e-01
1.14462271e-01 -6.65905178e-01 1.25224605e-01 7.36538827e-01
1.99025199e-01 -6.25367939e-01 1.14819860e+00 6.01956964e-01
2.90492028e-01 -9.48764801e-01 -1.42947507e+00 -1.08816636e+00
3.10881078e-01 -3.20308626e-01 7.27453768e-01 9.21969235e-01
-1.15778469e-01 -2.06056863e-01 2.24984869e-01 9.82180163e-02
1.01075435e+00 -1.32854432e-01 4.36490387e-01 -1.70180833e+00
6.04042590e-01 -7.78966606e-01 -1.06647694e+00 -3.09634835e-01
5.80990076e-01 -1.23879564e+00 -5.52909195e-01 -7.56706297e-01
-4.33163613e-01 -1.84418723e-01 1.29564613e-01 3.32090229e-01
8.19580436e-01 4.82683271e-01 -1.79402024e-01 6.52078679e-03
2.49741580e-02 6.33857667e-01 1.10970294e+00 -3.92679423e-02
-2.12009147e-01 2.64534175e-01 2.17750192e-01 1.02212906e+00
7.56482601e-01 -2.03677624e-01 -5.12088872e-02 -2.59700753e-02
-4.89831924e-01 1.06771857e-01 3.28140020e-01 -9.08777893e-01
1.88391432e-01 1.33754909e-01 6.52061284e-01 -6.01665854e-01
5.76191306e-01 -7.67682314e-01 1.43663555e-01 6.44800365e-01
7.69316927e-02 1.71690524e-01 1.38591871e-01 1.66671857e-01
-1.16098285e-01 -3.30475450e-01 1.01982868e+00 1.84442818e-01
-2.27967650e-01 6.94277763e-01 -2.75358051e-01 -5.73780596e-01
1.25673199e+00 -6.04616106e-01 4.16416943e-01 -8.89184251e-02
-9.38890338e-01 -3.40322495e-01 3.27110648e-01 4.70028341e-01
1.01432598e+00 -2.02622175e+00 -6.49164498e-01 8.51692021e-01
-1.75226390e-01 -7.02620685e-01 -2.16810867e-01 3.96966785e-01
-9.16376173e-01 5.79834521e-01 -4.38353896e-01 -5.63738227e-01
-1.68282795e+00 2.19062388e-01 8.55534136e-01 7.46287525e-01
-2.20380828e-01 8.20661187e-01 -4.25972700e-01 -7.72551298e-01
2.80596524e-01 1.53597325e-01 -6.62587941e-01 -4.80726734e-02
6.33958995e-01 7.08782971e-01 2.70017803e-01 -1.10236728e+00
-4.65058804e-01 1.37249017e+00 3.30212086e-01 -2.35829845e-01
1.09237754e+00 4.80776489e-01 -5.21953046e-01 2.51119852e-01
1.94946134e+00 -5.95356405e-01 -6.55971110e-01 -1.25576496e-01
2.86871761e-01 -4.31550950e-01 -1.30344167e-01 -2.40869746e-01
-9.55532074e-01 8.17316353e-01 1.21684372e+00 2.33856976e-01
8.33189487e-01 -2.36167654e-01 3.90044242e-01 8.19321990e-01
3.92362148e-01 -1.11299777e+00 -2.01617762e-01 4.84156817e-01
1.25968325e+00 -8.32825899e-01 -2.34175682e-01 -6.07467055e-01
-4.13458161e-02 1.88455415e+00 2.35454857e-01 -5.77953160e-01
1.85425985e+00 -5.73357791e-02 1.97603386e-02 -2.15144932e-01
-2.85664022e-01 2.12200806e-01 8.25148106e-01 6.12536907e-01
3.51947427e-01 -2.49441005e-02 -5.39516032e-01 1.40033960e-01
-3.47792119e-01 8.96774456e-02 1.35724217e-01 7.48287618e-01
-4.65507627e-01 -1.01310992e+00 -5.80029845e-01 2.18798772e-01
-1.93881795e-01 5.65037787e-01 -6.55001879e-01 7.30024338e-01
-1.24495029e-01 7.18110442e-01 1.63721547e-01 -5.25394320e-01
4.46620017e-01 4.05045986e-01 8.77089083e-01 -8.49565193e-02
-1.95621513e-04 -3.73279482e-01 -1.81797504e-01 -7.02062309e-01
-3.02816004e-01 -9.93803024e-01 -1.23318076e+00 -4.33063120e-01
1.22313350e-02 1.02086425e-01 1.28160846e+00 8.20441604e-01
3.71951908e-01 -6.47300780e-01 1.06179857e+00 -9.38746631e-01
-7.09915996e-01 -7.16376841e-01 -9.57762241e-01 5.02350807e-01
-1.69380277e-01 -1.03611839e+00 -4.14088368e-01 1.14738733e-01] | [13.257428169250488, 0.6770864129066467] |
96fb2e67-16be-4809-8163-1068ab462b59 | optimization-based-deep-learning-methods-for | 2303.01515 | null | https://arxiv.org/abs/2303.01515v1 | https://arxiv.org/pdf/2303.01515v1.pdf | Optimization-Based Deep learning methods for Magnetic Resonance Imaging Reconstruction and Synthesis | This dissertation is devoted to provide advanced nonconvex nonsmooth variational models of (Magnetic Resonance Image) MRI reconstruction, efficient learnable image reconstruction algorithms and parameter training algorithms that improve the accuracy and robustness of the optimization-based deep learning methods for compressed sensing MRI reconstruction and synthesis. The first part introduces a novel optimization based deep neural network whose architecture is inspired by proximal gradient descent for solving a variational model. The second part is a substantial extension of the preliminary work in the first part by solving the calibration-free fast pMRI reconstruction problem in a discrete-time optimal control framework. The third part aims at developing a generalizable Magnetic Resonance Imaging (MRI) reconstruction method in the meta-learning framework. The last part aims to synthesize target modality of MRI by using partially scanned k-space data from source modalities instead of fully scanned data that is used in the state-of-the-art multimodal synthesis. | ['Wanyu Bian'] | 2023-03-02 | null | null | null | null | ['mri-reconstruction'] | ['computer-vision'] | [ 5.91608822e-01 4.52350587e-01 -1.98482856e-01 -4.40018386e-01
-1.41902566e+00 1.64951473e-01 -3.27166319e-02 -4.86194998e-01
-1.70110703e-01 7.46712089e-01 6.20688975e-01 -3.65244895e-02
-4.60555345e-01 -3.32261264e-01 -9.46012199e-01 -9.23435748e-01
-2.74897903e-01 6.64724052e-01 -5.34965634e-01 -2.60742217e-01
3.03404350e-02 4.99719441e-01 -8.20747316e-01 4.95614469e-01
5.13276041e-01 8.55153024e-01 6.50724113e-01 6.47418618e-01
4.01413143e-01 9.62676823e-01 1.15772590e-01 2.00850725e-01
3.97982746e-01 -6.37053907e-01 -9.34693754e-01 2.72786945e-01
5.17639577e-01 -6.37431622e-01 -6.80515647e-01 9.50467467e-01
1.07246149e+00 3.60824615e-01 7.33843625e-01 -6.93525612e-01
-8.97630036e-01 8.64404202e-01 -4.36149567e-01 3.22272301e-01
7.63670132e-02 -1.33360893e-01 5.29348254e-01 -9.97190356e-01
9.32311654e-01 9.48136389e-01 8.81120265e-01 8.65636408e-01
-9.65729237e-01 -2.57517308e-01 -4.59101409e-01 1.61916479e-01
-1.07909656e+00 -4.83839095e-01 9.91386592e-01 -3.49717766e-01
8.21412742e-01 2.66895175e-01 5.09274900e-01 1.06816936e+00
6.08659863e-01 9.84261930e-01 1.28324485e+00 -3.82809311e-01
3.82217206e-02 -2.78546482e-01 -1.94736078e-01 8.80708158e-01
-3.85260522e-01 2.29885295e-01 -2.56894916e-01 -9.75711420e-02
1.22175753e+00 -1.17088780e-01 -3.47041279e-01 -4.60079223e-01
-1.52685142e+00 1.11157584e+00 5.92049122e-01 7.06524670e-01
-9.24058616e-01 4.32516903e-01 4.48426813e-01 1.34097084e-01
4.15960670e-01 4.06620413e-01 -2.38492027e-01 2.82535583e-01
-1.42593908e+00 2.02257559e-01 5.04327595e-01 4.56239462e-01
3.42310399e-01 8.46232653e-01 -4.67424423e-01 1.04811084e+00
5.06523907e-01 5.81759691e-01 7.82157838e-01 -1.57224488e+00
2.87293941e-01 -2.75271922e-01 -1.09522685e-01 -9.90276277e-01
-6.84486091e-01 -8.22585285e-01 -9.95780289e-01 -3.45037654e-02
-1.97416455e-01 -3.96545082e-01 -9.64696169e-01 1.58330464e+00
3.50713760e-01 3.67354423e-01 -4.55382317e-02 1.27908158e+00
1.24487221e+00 5.55389822e-01 -2.35949084e-01 -7.49322057e-01
9.57846522e-01 -1.11024177e+00 -1.04032826e+00 2.00956762e-01
2.85192221e-01 -5.13148010e-01 7.20048487e-01 3.47259820e-01
-1.62016249e+00 -3.06390077e-01 -1.13160419e+00 -6.72415495e-02
3.26564461e-01 6.70108721e-02 5.65364957e-01 4.53235745e-01
-1.27829111e+00 8.43332648e-01 -1.19497466e+00 3.24804991e-01
5.18104911e-01 4.57880974e-01 -4.01032716e-01 -1.97664827e-01
-1.15471101e+00 1.18163812e+00 1.91853032e-01 1.91678241e-01
-1.50412595e+00 -1.19819796e+00 -9.35553670e-01 -4.23113316e-01
1.00685127e-01 -8.37479889e-01 9.43435669e-01 -1.06234574e+00
-1.75401664e+00 1.10259652e+00 -2.96877921e-02 -3.68504465e-01
6.86713099e-01 8.74019861e-02 -1.15908697e-01 4.78202641e-01
4.38748598e-02 7.25246489e-01 1.14758384e+00 -1.45658791e+00
1.65570483e-01 -4.60403770e-01 -4.14136440e-01 3.51892531e-01
2.44355425e-01 -1.49798557e-01 2.14902535e-02 -8.69931638e-01
3.55264127e-01 -9.32868063e-01 -4.92970735e-01 -3.18222791e-01
-3.73824388e-01 2.81932205e-01 3.47132087e-01 -1.26654446e+00
6.52588129e-01 -1.58534276e+00 7.22245753e-01 5.75819053e-02
2.25857094e-01 -1.37931064e-01 -4.16745991e-01 1.72060117e-01
-6.19977713e-01 -3.90169084e-01 -6.37345850e-01 -4.66583997e-01
-2.23967105e-01 2.64279842e-01 -1.81633711e-01 8.82315576e-01
-3.77358735e-01 1.19819939e+00 -7.95651019e-01 -5.06146967e-01
3.69454443e-01 6.54913425e-01 -5.24990141e-01 5.01658879e-02
8.56283903e-02 1.26755559e+00 -3.58970106e-01 6.67323530e-01
6.76942110e-01 -9.67791006e-02 3.62652801e-02 -7.28689492e-01
1.03621662e-01 -3.55673999e-01 -1.13057899e+00 2.45078111e+00
-5.38420141e-01 4.51409966e-01 7.26173818e-01 -1.84723270e+00
4.35104609e-01 7.06168890e-01 1.47430718e+00 -7.73534417e-01
1.60835087e-01 4.80180860e-01 -3.84261250e-01 -1.07719421e+00
3.35764050e-01 -7.68162489e-01 3.98432344e-01 5.28002024e-01
1.55256376e-01 -3.15262735e-01 -3.13457012e-01 -1.17991038e-01
7.66443431e-01 1.33317292e-01 -1.47054568e-01 -4.79043514e-01
6.20752096e-01 4.52118777e-02 4.00152773e-01 7.37816930e-01
-1.45577610e-01 9.18735743e-01 -2.46901169e-01 -5.66223621e-01
-1.43733501e+00 -1.01101768e+00 -5.35921693e-01 9.56812263e-01
-3.15456718e-01 5.24758995e-01 -7.80297160e-01 -1.08411349e-01
-3.12025875e-01 6.21317029e-01 -6.47478700e-01 2.60443747e-01
-1.11146617e+00 -1.10480988e+00 4.30344790e-01 3.38066548e-01
3.77022177e-01 -1.18063736e+00 -4.14295644e-01 4.51324642e-01
-7.78348505e-01 -9.72998381e-01 -4.31745499e-01 -1.00075535e-01
-1.31725013e+00 -8.01877558e-01 -1.42819417e+00 -9.82702076e-01
5.71215808e-01 -1.36506587e-01 1.02262712e+00 -7.38963298e-03
-5.38680911e-01 8.30985129e-01 -1.25701502e-02 -3.98901552e-02
-5.87044954e-01 -1.61111668e-01 1.63702980e-01 -5.39977401e-02
-5.05773008e-01 -6.99954152e-01 -7.61424661e-01 -7.66511038e-02
-1.11261570e+00 -3.66517762e-03 5.57340026e-01 1.21169686e+00
1.30476665e+00 -2.88446158e-01 6.22703552e-01 -7.75901377e-01
7.09340334e-01 -6.89123392e-01 -1.48174107e-01 2.53420889e-01
-3.76316100e-01 8.01866278e-02 2.19489083e-01 -4.88362402e-01
-8.75807345e-01 2.21904978e-01 -4.62418616e-01 -8.15397978e-01
1.93723425e-01 7.79591918e-01 5.45704484e-01 -4.22677189e-01
9.30351675e-01 6.47915542e-01 3.95665854e-01 -2.69984871e-01
3.00904989e-01 4.40612078e-01 7.74341226e-01 -4.90253359e-01
2.29832917e-01 7.97198474e-01 3.10608357e-01 -8.62730443e-01
-7.47394085e-01 -1.39766529e-01 -5.70222318e-01 -3.64403367e-01
1.01304984e+00 -8.28825235e-01 -4.29471284e-01 2.51562655e-01
-8.70557904e-01 -6.30913198e-01 -4.06099975e-01 8.18406820e-01
-1.22348773e+00 5.74069619e-01 -1.06801987e+00 -4.53516781e-01
-8.84799004e-01 -1.71830797e+00 1.14678717e+00 -2.02858180e-01
1.85108647e-01 -1.27765799e+00 4.20707375e-01 8.92466009e-01
7.15784669e-01 7.18122482e-01 7.32045114e-01 -2.46447697e-01
-4.56714243e-01 1.67955663e-02 3.22469682e-01 4.58827287e-01
-1.00862354e-01 -7.41524100e-01 -6.98866308e-01 -7.69404173e-01
7.16485858e-01 -4.85598624e-01 6.94334507e-01 1.38428009e+00
1.15335822e+00 -3.36672693e-01 -6.02271408e-02 1.08188760e+00
1.65120566e+00 -1.36440635e-01 7.67527461e-01 1.34490415e-01
7.58105755e-01 5.53210557e-01 8.19866359e-02 1.77901968e-01
2.42179081e-01 5.20609677e-01 5.31022787e-01 -1.05369627e-01
-4.56169516e-01 1.80394560e-01 1.53561831e-01 1.13289678e+00
-1.07289262e-01 4.42857414e-01 -6.73690557e-01 6.88448131e-01
-1.80394435e+00 -1.04207230e+00 -7.39889592e-02 1.76892543e+00
8.79487693e-01 -9.15976703e-01 -1.76068604e-01 -9.50067565e-02
4.71183449e-01 4.19994704e-02 -7.39090204e-01 -1.88576266e-01
1.40522756e-02 3.44682813e-01 5.86069942e-01 7.91985989e-01
-1.03853190e+00 5.37126660e-01 7.48427153e+00 9.29668546e-01
-1.40255713e+00 7.00241804e-01 5.79281747e-01 -8.75705108e-02
-4.62445557e-01 -6.05457246e-01 -5.52656886e-04 1.49987355e-01
9.39556241e-01 4.05110836e-01 1.05382609e+00 5.79168141e-01
5.09241998e-01 1.47171944e-01 -9.58462238e-01 1.37901676e+00
1.93250999e-01 -1.83529258e+00 -2.05908045e-01 -2.05201339e-02
1.23682058e+00 4.46315795e-01 4.24608052e-01 -1.12978229e-02
-6.77732229e-02 -1.54841411e+00 6.00908041e-01 7.61891782e-01
9.77057338e-01 -4.52551126e-01 6.08997166e-01 4.55094755e-01
-4.72824514e-01 -7.39434212e-02 -4.09112543e-01 5.86677253e-01
5.19325912e-01 6.37793005e-01 -4.84870821e-01 6.56869709e-01
4.90961611e-01 6.44353628e-01 3.45031828e-01 8.16387117e-01
1.83106139e-01 3.79340529e-01 -4.71550748e-02 5.48876941e-01
3.20144773e-01 -2.37817407e-01 1.10702360e+00 1.11208153e+00
2.02072352e-01 3.97758722e-01 2.22307771e-01 1.19105291e+00
-1.12877838e-01 1.72296539e-01 -5.36905885e-01 2.48873591e-01
-1.49222568e-01 1.32094359e+00 -3.84457529e-01 -1.99473321e-01
1.34969959e-02 8.01834762e-01 6.01448007e-02 3.53034139e-01
-6.17393017e-01 3.04918498e-01 -1.58974558e-01 1.75932541e-01
1.27289966e-01 -2.50336081e-01 -5.08356929e-01 -1.25790620e+00
-3.01882476e-01 -1.01478934e+00 2.84624636e-01 -9.00279939e-01
-1.33218992e+00 5.45075893e-01 3.47437650e-01 -8.07231784e-01
-4.53764409e-01 -3.74987096e-01 -3.26050907e-01 8.13080430e-01
-1.33938956e+00 -1.32702565e+00 -1.12740256e-01 1.01510978e+00
5.67113876e-01 -3.47720057e-01 8.30981731e-01 4.34457541e-01
-1.65906847e-01 4.63040859e-01 4.07460839e-01 5.25616221e-02
2.08811700e-01 -9.56462562e-01 -5.63332379e-01 5.82820952e-01
-2.29510248e-01 4.83815461e-01 8.66656601e-01 -6.54983878e-01
-1.76043570e+00 -7.85371959e-01 2.77948350e-01 3.72685380e-02
4.35392976e-01 4.07766074e-01 -6.77263200e-01 7.52599716e-01
3.27305645e-01 2.53507257e-01 5.34143865e-01 -7.27019072e-01
5.59334338e-01 6.97419941e-02 -1.84040713e+00 2.78430194e-01
6.20871544e-01 -5.27791679e-01 -5.82117975e-01 8.66744339e-01
3.69562507e-01 -9.89922166e-01 -1.39768994e+00 4.93803501e-01
4.36333627e-01 -3.50339860e-01 1.25931454e+00 -5.48936129e-01
7.71526098e-01 1.49197146e-01 -4.58707213e-01 -1.31246114e+00
-4.21682507e-01 -6.63931668e-01 -1.29845455e-01 1.67263329e-01
8.80669057e-02 -3.94782484e-01 9.20851469e-01 4.84040648e-01
-5.50027251e-01 -1.01681387e+00 -1.39397871e+00 -2.83852875e-01
5.36210120e-01 -4.24159050e-01 1.06963150e-01 1.15992486e+00
-4.19510484e-01 -1.32361680e-01 -8.15589130e-01 -7.18649402e-02
1.22564352e+00 -6.09312952e-03 2.44162809e-02 -5.72640300e-01
-3.80614638e-01 -9.55075696e-02 2.96761021e-02 -7.41097093e-01
2.40850642e-01 -1.32921016e+00 1.09680310e-01 -1.78074729e+00
3.42791647e-01 -4.03682500e-01 -3.71781081e-01 3.56685340e-01
3.34074408e-01 5.10179877e-01 1.18921675e-01 5.15909374e-01
-1.42793104e-01 5.46059310e-01 1.97882497e+00 -3.52350712e-01
-7.80972689e-02 -9.29324627e-02 -4.38087285e-01 4.60762739e-01
4.96047795e-01 -5.28403878e-01 -5.13211131e-01 -7.63394475e-01
7.44370073e-02 1.00702620e+00 3.73999804e-01 -6.52529538e-01
1.29090920e-01 -4.21703383e-02 3.92477423e-01 -5.06242275e-01
4.51194584e-01 -7.89346814e-01 4.93305922e-01 6.32893503e-01
-7.00946748e-01 3.41994986e-02 -8.40552747e-02 3.04675162e-01
2.62780450e-02 -5.40921569e-01 1.19794011e+00 -7.73410141e-01
-3.23072881e-01 5.35961509e-01 -2.46239796e-01 8.90601948e-02
6.55799627e-01 -1.07824564e-01 4.47597414e-01 -4.89839822e-01
-1.42222297e+00 -1.72629938e-01 -1.06361240e-01 1.70754474e-02
1.07383454e+00 -1.46664357e+00 -1.25982332e+00 -5.89306839e-02
-7.12504685e-01 -2.71198690e-01 6.83145285e-01 1.44516647e+00
-8.66826236e-01 3.52217883e-01 -5.66934288e-01 -8.03780377e-01
-8.79549325e-01 4.33806539e-01 1.05407584e+00 -5.96124649e-01
-8.24180424e-01 8.86107802e-01 -2.97103465e-01 -9.04546380e-01
5.28061204e-02 1.18537925e-01 -2.41553038e-01 -4.52855080e-01
3.66994590e-01 5.20834804e-01 2.11255163e-01 -8.67841184e-01
-2.68856645e-01 5.60122371e-01 2.21201092e-01 -5.11188567e-01
1.89776850e+00 -6.22770377e-02 -4.30707127e-01 2.32051060e-01
1.55479181e+00 -3.59986246e-01 -9.62459922e-01 -2.38234892e-01
-5.28291702e-01 -1.67021640e-02 6.18731081e-01 -7.75512874e-01
-1.72082114e+00 7.21556842e-01 1.09788442e+00 -4.84926492e-01
9.72226560e-01 -2.27284402e-01 1.10232341e+00 1.43779516e-01
3.41912568e-01 -1.21613550e+00 2.34534144e-01 2.40841687e-01
1.40849924e+00 -1.33439326e+00 4.14934546e-01 2.03243315e-01
-7.84056604e-01 1.21403718e+00 -1.97926745e-01 -5.58531523e-01
8.72384191e-01 1.76163733e-01 -1.01769932e-01 -4.90177244e-01
-8.55514333e-02 6.69632554e-01 6.36595428e-01 6.78906143e-01
4.35759395e-01 1.89979047e-01 -5.29284358e-01 3.11548352e-01
-6.05775900e-02 2.56967038e-01 3.39938521e-01 7.68682480e-01
-1.35696232e-01 -8.01737785e-01 -5.12455881e-01 3.72754037e-01
-9.68589008e-01 -1.19887434e-01 3.76339555e-01 3.41600984e-01
-5.43816537e-02 6.80314302e-01 -4.33566958e-01 -9.41984579e-02
-1.04841284e-01 -2.40412578e-01 1.06734049e+00 -2.75530517e-01
-6.36993170e-01 2.20996499e-01 -1.93411604e-01 -7.00866044e-01
-7.16954350e-01 -6.75401986e-01 -1.39445257e+00 -9.89986677e-03
4.45235781e-02 -2.37931088e-01 1.03266561e+00 1.02427745e+00
8.66481364e-02 7.05133736e-01 5.53908467e-01 -1.43722546e+00
-8.37384164e-01 -9.96896863e-01 -6.37731433e-01 5.15552759e-01
5.42396188e-01 -5.36438525e-01 1.33611888e-01 -6.66456521e-02] | [13.489744186401367, -2.4159626960754395] |
01efa5b9-f9f7-464c-ada2-848a386833f2 | triple-stream-deep-metric-learning-of-great | 2301.02642 | null | https://arxiv.org/abs/2301.02642v1 | https://arxiv.org/pdf/2301.02642v1.pdf | Triple-stream Deep Metric Learning of Great Ape Behavioural Actions | We propose the first metric learning system for the recognition of great ape behavioural actions. Our proposed triple stream embedding architecture works on camera trap videos taken directly in the wild and demonstrates that the utilisation of an explicit DensePose-C chimpanzee body part segmentation stream effectively complements traditional RGB appearance and optical flow streams. We evaluate system variants with different feature fusion techniques and long-tail recognition approaches. Results and ablations show performance improvements of ~12% in top-1 accuracy over previous results achieved on the PanAf-500 dataset containing 180,000 manually annotated frames across nine behavioural actions. Furthermore, we provide a qualitative analysis of our findings and augment the metric learning system with long-tail recognition techniques showing that average per class accuracy -- critical in the domain -- can be improved by ~23% compared to the literature on that dataset. Finally, since our embedding spaces are constructed as metric, we provide first data-driven visualisations of the great ape behavioural action spaces revealing emerging geometry and topology. We hope that the work sparks further interest in this vital application area of computer vision for the benefit of endangered great apes. | ['Tilo Burghardt', 'Hjalmar Kühl', 'Majid Mirmehdi', 'Otto Brookes'] | 2023-01-06 | null | null | null | null | ['metric-learning', 'metric-learning'] | ['computer-vision', 'methodology'] | [ 3.21276218e-01 1.14242872e-02 3.12776566e-01 -3.22410941e-01
-4.33331847e-01 -5.85679531e-01 7.16559887e-01 -1.68056577e-01
-1.27960908e+00 4.75982308e-01 1.87390715e-01 2.85584539e-01
-2.48888284e-01 -4.09850508e-01 -4.18641776e-01 -5.99251509e-01
-1.02475786e+00 4.28334862e-01 4.29619223e-01 -4.22372550e-01
3.46024096e-01 5.79782844e-01 -1.56579685e+00 1.22733973e-01
-7.51479790e-02 7.59066343e-01 -3.81414264e-01 1.18953705e+00
3.56948912e-01 6.47974432e-01 -6.29808843e-01 -8.11998248e-01
7.18580663e-01 -3.69583130e-01 -6.19141400e-01 -3.48385990e-01
1.04043996e+00 -7.47203350e-01 -5.19241750e-01 5.76348484e-01
6.34764433e-01 1.61484387e-02 6.76215112e-01 -1.34135139e+00
-5.40845096e-02 1.19703822e-01 -4.59722668e-01 7.71097958e-01
5.59748530e-01 7.91841567e-01 1.07857704e+00 -3.21727097e-01
7.93499827e-01 1.23197424e+00 1.25495744e+00 7.79705882e-01
-1.42662203e+00 -5.71735382e-01 -4.07001615e-01 4.75426227e-01
-8.74690354e-01 -4.39734131e-01 4.25802022e-01 -7.64788091e-01
1.22441041e+00 8.35933611e-02 1.26996815e+00 1.24021149e+00
9.90454387e-03 7.88437903e-01 1.13015962e+00 1.70601159e-01
1.46973431e-01 -6.65326297e-01 -1.19311042e-01 7.74009287e-01
1.15015902e-01 4.21958625e-01 -8.35365057e-01 -1.72205657e-01
7.97330439e-01 -5.07020727e-02 -1.32500961e-01 -4.03286934e-01
-1.53744948e+00 8.55988503e-01 5.32594264e-01 -4.55514994e-03
-2.14720830e-01 7.54225314e-01 6.71799958e-01 4.25830930e-01
3.41948897e-01 5.34658015e-01 -1.36248544e-01 -1.19858515e+00
-1.30274391e+00 4.87964183e-01 8.03816259e-01 6.01576388e-01
3.28948319e-01 -2.31422111e-02 2.81439036e-01 7.56821811e-01
4.08262849e-01 7.76823521e-01 1.82888806e-01 -1.36984551e+00
4.15327787e-01 4.72934186e-01 -9.79385525e-02 -9.75662291e-01
-6.65314794e-01 1.95291117e-01 -2.81757116e-01 6.34344339e-01
7.11543620e-01 -1.32667094e-01 -7.50071824e-01 1.51478326e+00
3.94513726e-01 1.71498418e-01 -2.74014413e-01 9.91063237e-01
5.19903839e-01 3.02009046e-01 1.64220631e-01 3.87270778e-01
1.27795017e+00 -4.87291306e-01 -1.68816820e-01 -2.08133653e-01
5.87607265e-01 -2.52577305e-01 7.31320322e-01 4.42290604e-01
-7.65298426e-01 -1.97065324e-01 -1.21896958e+00 -8.77441186e-03
-4.80005890e-01 -2.22196698e-01 7.75362909e-01 8.61097932e-01
-1.26056767e+00 7.59385645e-01 -1.15818810e+00 -8.25029194e-01
8.89382303e-01 5.10999858e-01 -8.68538797e-01 2.68266171e-01
-7.90033519e-01 9.02322173e-01 -2.00701863e-01 3.55089843e-01
-1.06442034e+00 -7.19344258e-01 -8.35073590e-01 -5.64163685e-01
-1.28307864e-01 -2.85192221e-01 1.10984087e+00 -4.31939036e-01
-1.18496215e+00 1.12781525e+00 6.49788857e-01 -7.82811284e-01
1.05048215e+00 -3.80447030e-01 -3.15492228e-02 8.03685725e-01
1.04715139e-01 1.19248819e+00 6.45232320e-01 -5.41060805e-01
-9.94198203e-01 -4.99493629e-01 2.40812257e-01 7.05141500e-02
8.55742674e-03 5.56835830e-02 4.68933098e-02 -6.03287339e-01
-3.22461665e-01 -9.02562380e-01 -1.10194482e-01 9.08261538e-01
5.21994054e-01 1.65554821e-01 7.77343154e-01 -7.84732401e-01
9.08456206e-01 -2.04385448e+00 3.70980650e-01 -1.01270296e-01
4.42103595e-01 2.75837481e-01 -7.42174834e-02 5.58027744e-01
2.64789104e-01 -1.22659780e-01 -8.24787080e-01 -1.74938872e-01
2.31278673e-01 4.74109679e-01 -1.23089794e-02 1.29676354e+00
2.33576983e-01 7.23888040e-01 -1.17968035e+00 -6.37544572e-01
5.97378671e-01 6.78245366e-01 -7.62427807e-01 1.94333404e-01
2.69862890e-01 2.72349924e-01 -2.66635250e-02 5.55746615e-01
4.09867883e-01 6.54167354e-01 -2.81987309e-01 1.42752573e-01
-1.49777114e-01 -1.94080621e-01 -7.76843250e-01 1.83963478e+00
-1.81627065e-01 1.18380964e+00 1.99857324e-01 -7.39546657e-01
5.37396014e-01 -1.13606445e-01 8.17276061e-01 -7.27913320e-01
8.46951529e-02 1.71484768e-01 2.88201213e-01 -6.05549693e-01
3.89606953e-01 -2.29312271e-01 -7.16612712e-02 4.60796326e-01
4.86452430e-01 -7.30863810e-02 4.18801099e-01 1.92990825e-01
1.49625230e+00 6.73414886e-01 5.33584580e-02 -3.35211188e-01
3.60702634e-01 1.23122886e-01 1.13631696e-01 5.58391333e-01
-8.63036156e-01 7.75278091e-01 6.85075343e-01 -9.52986062e-01
-1.08476043e+00 -1.15350389e+00 -3.59009832e-01 1.28298604e+00
4.72444482e-02 -5.10867953e-01 -9.69202876e-01 -6.81438744e-01
1.84034318e-01 3.33454809e-03 -1.34862494e+00 3.66452858e-02
-1.09463000e+00 -7.72627115e-01 1.31419683e+00 6.99417710e-01
6.20424688e-01 -1.23702514e+00 -1.65486741e+00 3.00480336e-01
1.81141838e-01 -9.57731426e-01 -1.16018534e-01 -4.78877723e-02
-7.88361967e-01 -1.40322983e+00 -8.87294292e-01 -3.52050871e-01
1.67499110e-01 -1.30236208e-01 6.52765870e-01 1.26594221e-02
-7.31846511e-01 9.60059881e-01 -5.60862362e-01 2.37440154e-01
-9.10704285e-02 -3.72517824e-01 1.62601024e-01 -2.15463042e-01
5.29548824e-01 -7.90041327e-01 -1.06768167e+00 3.90286654e-01
-7.27852404e-01 -3.83740336e-01 1.80057317e-01 6.26927614e-01
6.93508461e-02 -9.79932904e-01 -7.08905309e-02 -2.22916886e-01
3.08525831e-01 -4.26597476e-01 -5.14628351e-01 -1.49163827e-01
8.56002718e-02 3.26369256e-02 3.69618267e-01 -2.54952073e-01
-5.06098628e-01 5.04828244e-02 -2.69382685e-01 2.15056241e-02
-2.67311603e-01 -2.62625486e-01 4.28393871e-01 -3.00835341e-01
8.72120023e-01 5.33406064e-02 1.36081502e-01 -4.42596108e-01
5.84598541e-01 6.48888350e-01 9.26658809e-01 -4.73365396e-01
4.61695671e-01 9.31633115e-01 3.38408768e-01 -1.37881374e+00
2.91436583e-01 -6.50806785e-01 -1.19732106e+00 -7.98162162e-01
1.27055323e+00 -5.74564159e-01 -9.21034455e-01 6.21397614e-01
-7.92064190e-01 -5.32083869e-01 -3.09357017e-01 4.58902180e-01
-9.53368068e-01 6.35524869e-01 -4.40223068e-01 -8.96955132e-01
-1.77262813e-01 -1.01414537e+00 1.49730992e+00 -7.68500343e-02
-4.06700313e-01 -8.75334501e-01 7.36617863e-01 3.70998144e-01
4.30058211e-01 8.61914873e-01 -2.07760688e-02 -5.55027246e-01
-1.15561068e-01 -2.35439509e-01 -3.20454091e-02 1.16201170e-01
-1.70301914e-01 1.93016574e-01 -9.68895912e-01 -3.35948557e-01
-2.66690642e-01 -3.37984115e-01 9.01119828e-01 1.38352001e-02
3.12511295e-01 -1.30615845e-01 -1.59711428e-02 8.35754514e-01
1.02846038e+00 -4.51596826e-02 4.69209939e-01 7.25592732e-01
5.59457362e-01 8.17556858e-01 3.89489174e-01 6.30257726e-01
3.59346867e-01 7.17668355e-01 4.95052189e-01 1.37731463e-01
-3.78566653e-01 -1.61611974e-01 6.82882845e-01 1.94992632e-01
-4.71560061e-01 1.96239501e-01 -1.07514501e+00 7.90415466e-01
-1.69482470e+00 -1.23036540e+00 -1.16605967e-01 2.14012384e+00
3.18736762e-01 -1.16729960e-01 1.01573455e+00 1.66852981e-01
4.00651991e-01 2.65112847e-01 -2.01900005e-01 -5.37863493e-01
1.24686100e-01 3.90000880e-01 7.72297621e-01 4.78163749e-01
-1.29142857e+00 7.52690196e-01 6.82797861e+00 2.61147857e-01
-8.09878290e-01 -7.08752275e-02 -1.04525927e-02 -6.04708970e-01
3.53622854e-01 -3.10065001e-01 -3.25414956e-01 5.35753906e-01
1.07059681e+00 2.62461036e-01 5.10988057e-01 6.42708778e-01
-6.42480850e-02 -3.03856134e-01 -1.35212243e+00 1.31473505e+00
2.42563292e-01 -9.26358581e-01 -4.57095057e-01 4.36153233e-01
1.54324314e-02 5.32476664e-01 -1.99975774e-01 5.58444113e-02
2.07582518e-01 -9.52973008e-01 8.63291502e-01 5.19821107e-01
7.85672545e-01 -6.35812104e-01 5.86435556e-01 5.30359633e-02
-1.22984254e+00 -3.72679383e-02 -2.43940875e-01 -2.20964044e-01
3.29188496e-01 -5.28681993e-01 -5.60034692e-01 -1.11026436e-01
1.02524400e+00 1.05014110e+00 -1.13350320e+00 1.45731282e+00
-9.41686332e-03 6.39498055e-01 -1.03218424e+00 -1.72710270e-01
7.09360361e-01 -3.26148897e-01 6.93153679e-01 1.89447153e+00
-1.06598079e-01 1.37118474e-01 -5.73523819e-01 5.08401096e-01
3.68876874e-01 -1.95384771e-01 -7.80472517e-01 -1.95330083e-02
-2.00159356e-01 1.17292869e+00 -8.63698721e-01 -1.10869855e-01
-4.11656767e-01 1.02069056e+00 2.15138331e-01 -1.24316007e-01
-9.17831361e-01 -4.44209844e-01 1.00231445e+00 2.19765082e-02
4.94387209e-01 -3.76984507e-01 6.00583665e-02 -9.76345062e-01
6.06186613e-02 -4.60562229e-01 4.82466400e-01 -3.30864519e-01
-1.12026751e+00 4.53653961e-01 5.09579659e-01 -1.47572315e+00
-4.27576691e-01 -8.92496228e-01 -3.10841888e-01 1.11214541e-01
-8.45455229e-01 -1.16617107e+00 -1.57561302e-02 3.26248974e-01
3.91544402e-01 -1.04235098e-01 5.65318584e-01 3.85228783e-01
-8.55033025e-02 4.95302945e-01 1.29244402e-01 3.41585368e-01
2.61541933e-01 -1.54833090e+00 7.65572309e-01 9.07556832e-01
4.58366334e-01 5.09536803e-01 6.16834819e-01 -5.33375740e-01
-1.36904442e+00 -5.06563306e-01 2.30864123e-01 -9.80931878e-01
9.28990722e-01 -7.29266942e-01 -5.40661156e-01 5.68917692e-01
-1.12905465e-02 1.93424433e-01 8.22819948e-01 -4.49773878e-01
-2.70897716e-01 -6.94631264e-02 -1.48479056e+00 4.14557487e-01
1.35704339e+00 -3.56187642e-01 -1.11184371e+00 2.32605822e-02
-2.62977183e-02 -8.94373581e-02 -9.74564850e-01 1.02109611e-01
1.30405009e+00 -1.21567941e+00 1.05396616e+00 -6.17233992e-01
3.96144629e-01 -5.19120634e-01 -3.52911085e-01 -1.01680708e+00
1.84365168e-01 -5.27840137e-01 9.10190865e-02 8.15038800e-01
1.31912772e-02 -3.71226400e-01 7.87663162e-01 4.36929554e-01
9.97119248e-02 -3.52072150e-01 -1.72383940e+00 -6.68428123e-01
1.99083239e-01 -7.74855316e-01 1.28554314e-01 4.45522696e-01
5.67468293e-02 -2.04703480e-01 -5.95966816e-01 -2.89060056e-01
1.04002833e+00 -4.38603997e-01 9.86622810e-01 -1.03543496e+00
-1.81206152e-01 -5.84662557e-01 -1.60427868e+00 -8.26771855e-01
-2.61754394e-01 -7.01296628e-01 7.99912959e-02 -1.24187863e+00
-1.68362930e-01 2.25288793e-01 5.58346659e-02 5.09038210e-01
2.55038798e-01 9.10155714e-01 3.37115169e-01 4.95571829e-02
-7.64514685e-01 4.57367450e-01 7.53583491e-01 -3.92302163e-02
2.85132378e-01 -4.47229832e-01 -4.80651371e-02 6.40919745e-01
3.64976823e-01 -5.83310664e-01 -5.22515178e-02 -4.85807538e-01
1.49858356e-01 -2.62737989e-01 7.37782121e-01 -1.52769077e+00
1.77509427e-01 2.80954391e-01 2.09910467e-01 -2.69841850e-01
4.91312772e-01 -9.65907753e-01 6.20364882e-02 7.70688295e-01
-2.44742423e-01 8.44573230e-02 2.91880369e-01 7.77452409e-01
1.28576085e-01 1.07902810e-01 8.16369116e-01 -1.78697318e-01
-1.01324391e+00 9.28842798e-02 -6.05473399e-01 3.32047552e-01
1.03503978e+00 -7.32944846e-01 -2.35804334e-01 -1.51170403e-01
-5.50255418e-01 3.26335132e-01 6.01672530e-01 4.00597841e-01
6.06110454e-01 -1.03598380e+00 -8.54265511e-01 3.34238082e-01
2.20943436e-01 -4.41378236e-01 1.03432231e-01 9.31727350e-01
-1.63800812e+00 1.03871739e-02 -9.32734609e-01 -9.24247563e-01
-1.29625201e+00 4.98957746e-03 4.71147597e-01 5.23087204e-01
-1.00277019e+00 9.23629820e-01 -2.58050114e-01 -4.34338808e-01
1.73808962e-01 -3.68976206e-01 -1.15725495e-01 3.30577761e-01
6.45795822e-01 1.00196791e+00 -3.42275888e-01 -1.33393753e+00
-7.19591439e-01 8.86334836e-01 3.79755795e-01 -5.86696863e-01
1.80335343e+00 8.12761188e-02 1.77462623e-01 2.76200622e-01
1.39082205e+00 -4.30337757e-01 -1.88729167e+00 4.23825622e-01
3.51636559e-02 -7.30962038e-01 -2.58664608e-01 -5.76012671e-01
-8.74885023e-01 1.40516007e+00 1.19680464e+00 -9.81348101e-03
6.64309263e-01 1.36283608e-02 6.68338418e-01 4.45993602e-01
4.94698137e-01 -1.10541368e+00 -2.19652086e-01 1.98305354e-01
9.70005751e-01 -1.09222627e+00 2.47264907e-01 5.44631422e-01
-6.69883966e-01 1.30721927e+00 4.86228526e-01 -4.77866143e-01
2.62260526e-01 2.21508756e-01 1.80614516e-01 -5.49418092e-01
-3.51341546e-01 -4.31201220e-01 -4.73932177e-02 1.00771654e+00
1.27457142e-01 -1.54403900e-03 -4.10168380e-01 5.97011577e-03
-7.34537601e-01 -1.86501026e-01 4.48303401e-01 1.12818718e+00
-3.47357452e-01 -9.61051643e-01 -2.75559396e-01 4.00300920e-01
-3.83448809e-01 1.60243243e-01 -7.73403823e-01 1.24702811e+00
2.10047290e-01 6.70140564e-01 5.41093312e-02 -5.33660710e-01
4.68200624e-01 -2.54670717e-02 8.13805759e-01 -2.38353685e-01
-7.95716286e-01 -2.35104054e-01 1.66376352e-01 -9.62160289e-01
-8.51197779e-01 -1.08195019e+00 -9.41223919e-01 -3.79021585e-01
2.49809459e-01 -1.97722495e-01 6.09832942e-01 6.57958150e-01
-2.20764399e-01 -1.58236057e-01 2.00789735e-01 -1.26333380e+00
-9.61221661e-03 -9.93996203e-01 -5.53717911e-01 5.69271743e-01
6.04621112e-01 -8.92760217e-01 -6.82417691e-01 1.96750134e-01] | [7.642454624176025, -0.5209493637084961] |
69e040ce-b520-4a28-89f2-05f0d45754d3 | mobileinst-video-instance-segmentation-on-the | 2303.17594 | null | https://arxiv.org/abs/2303.17594v1 | https://arxiv.org/pdf/2303.17594v1.pdf | MobileInst: Video Instance Segmentation on the Mobile | Although recent approaches aiming for video instance segmentation have achieved promising results, it is still difficult to employ those approaches for real-world applications on mobile devices, which mainly suffer from (1) heavy computation and memory cost and (2) complicated heuristics for tracking objects. To address those issues, we present MobileInst, a lightweight and mobile-friendly framework for video instance segmentation on mobile devices. Firstly, MobileInst adopts a mobile vision transformer to extract multi-level semantic features and presents an efficient query-based dual-transformer instance decoder for mask kernels and a semantic-enhanced mask decoder to generate instance segmentation per frame. Secondly, MobileInst exploits simple yet effective kernel reuse and kernel association to track objects for video instance segmentation. Further, we propose temporal query passing to enhance the tracking ability for kernels. We conduct experiments on COCO and YouTube-VIS datasets to demonstrate the superiority of MobileInst and evaluate the inference latency on a mobile CPU core of Qualcomm Snapdragon-778G, without other methods of acceleration. On the COCO dataset, MobileInst achieves 30.5 mask AP and 176 ms on the mobile CPU, which reduces the latency by 50% compared to the previous SOTA. For video instance segmentation, MobileInst achieves 35.0 AP on YouTube-VIS 2019 and 30.1 AP on YouTube-VIS 2021. Code will be available to facilitate real-world applications and future research. | ['Xinggang Wang', 'Wenyu Liu', 'Dashan Gao', 'Xiaowen Ying', 'Xin Li', 'Jiancheng Lyu', 'Shuai Zhang', 'Haoyi Jiang', 'Shusheng Yang', 'Tianheng Cheng', 'Renhong Zhang'] | 2023-03-30 | null | null | null | null | ['video-instance-segmentation'] | ['computer-vision'] | [ 9.60724056e-02 -3.14049602e-01 -5.56614161e-01 1.11693600e-02
-8.54665995e-01 -5.33843279e-01 6.98721930e-02 -2.07820311e-01
-5.56745946e-01 3.25091362e-01 -3.72003198e-01 -6.55180097e-01
5.44006750e-02 -8.16369295e-01 -7.59583056e-01 -3.55195850e-01
-2.98279729e-02 1.93121076e-01 1.04933596e+00 3.10636312e-01
1.79394588e-01 2.15626374e-01 -1.67669439e+00 4.98475909e-01
7.13766515e-01 1.33500302e+00 5.38113832e-01 1.11283243e+00
-3.23580839e-02 8.21909368e-01 -5.05037308e-01 -4.94657636e-01
3.37531328e-01 -1.83672890e-01 -6.83462918e-01 2.48778880e-01
3.25955391e-01 -4.34626132e-01 -4.87383157e-01 9.85250175e-01
6.27695978e-01 -1.44617245e-01 1.93377063e-01 -1.40756226e+00
6.21320084e-02 5.11061013e-01 -7.48011291e-01 6.91560447e-01
3.15213412e-01 3.20929587e-01 7.15265274e-01 -9.17424619e-01
4.20958310e-01 9.26022410e-01 6.45299911e-01 1.46404028e-01
-5.95868349e-01 -5.20054996e-01 1.02495596e-01 5.32304287e-01
-1.45028722e+00 -6.25748754e-01 1.59479648e-01 -2.17938032e-02
9.42043304e-01 6.13627791e-01 7.92034864e-01 5.20033240e-01
-5.65600656e-02 1.20916128e+00 6.20929718e-01 -7.69761428e-02
2.07277507e-01 -8.35054815e-02 1.98274389e-01 9.90229189e-01
3.95438522e-02 -2.22492769e-01 -5.65091789e-01 -6.81187212e-02
7.56393313e-01 3.07874773e-02 -2.73070693e-01 -1.34934723e-01
-1.36431885e+00 4.13002282e-01 2.05983669e-01 -8.81142095e-02
-3.29848140e-01 4.56579834e-01 6.88509285e-01 1.19717970e-01
2.98978895e-01 -1.75954461e-01 -4.92266029e-01 -7.06458390e-01
-1.37185991e+00 1.40222989e-03 7.71596551e-01 1.46396458e+00
6.83301985e-01 -4.26880978e-02 -2.23152801e-01 5.83638906e-01
3.26157212e-01 8.59188318e-01 3.11787009e-01 -1.17296684e+00
6.91038132e-01 4.52647477e-01 -1.52889684e-01 -1.08578110e+00
-2.37725556e-01 -2.73187637e-01 -2.54975617e-01 -3.36194456e-01
3.70940417e-01 -2.61235058e-01 -7.04247355e-01 1.01695430e+00
4.66852784e-01 9.50885355e-01 -1.54262677e-01 1.03116882e+00
9.75078285e-01 7.21485138e-01 -1.93342101e-02 -1.47478506e-01
1.72924185e+00 -1.51953101e+00 -3.34705472e-01 -3.99137549e-02
9.03553426e-01 -8.83190155e-01 1.22778964e+00 3.81390929e-01
-1.36171329e+00 -6.61128700e-01 -9.61461902e-01 -1.58622228e-02
-8.42689276e-02 4.78456438e-01 6.15240097e-01 1.01681924e+00
-9.72064734e-01 9.47650746e-02 -1.13121617e+00 -2.17537880e-01
4.57546711e-01 6.41255677e-01 4.00248945e-01 9.98217687e-02
-8.45370114e-01 -4.02176753e-02 3.55997145e-01 -6.57794764e-03
-7.61078954e-01 -8.57197285e-01 -6.10523939e-01 1.82275087e-01
9.76521611e-01 -5.26217461e-01 1.24997532e+00 -8.74052584e-01
-1.43960083e+00 5.75113714e-01 -2.70528913e-01 -5.91448963e-01
5.29488087e-01 -2.97636569e-01 -4.35585678e-01 5.88998854e-01
1.93287402e-01 6.62336230e-01 9.57948446e-01 -7.23607898e-01
-1.23867071e+00 -2.52212614e-01 4.76047933e-01 4.69835401e-01
-3.76275212e-01 1.85952373e-02 -1.61499059e+00 -4.88019198e-01
-4.23079319e-02 -1.07221270e+00 -1.08623996e-01 -1.45059809e-01
-3.43323231e-01 1.61634028e-01 1.18726099e+00 -5.38473129e-01
1.59289336e+00 -2.32006407e+00 -3.45977902e-01 2.68376946e-01
1.98998794e-01 6.40449047e-01 6.17063642e-02 -2.32722938e-01
5.53930044e-01 -1.29464522e-01 1.30181983e-01 -1.81511372e-01
-1.55425757e-01 5.41514345e-02 -1.25725687e-01 2.80617595e-01
-5.03604472e-01 1.02093267e+00 -7.64594793e-01 -7.86386788e-01
3.93106818e-01 3.42184901e-01 -7.91567028e-01 -1.66505441e-01
-1.50600836e-01 -1.34964868e-01 -4.73824501e-01 1.07990479e+00
7.70708025e-01 -3.88864666e-01 1.32868230e-01 -5.30975401e-01
-1.35389328e-01 3.61589552e-03 -1.26020920e+00 1.53367209e+00
-3.49649549e-01 5.92208028e-01 1.78429902e-01 -7.83753693e-01
2.55387932e-01 1.38368919e-01 5.32190025e-01 -9.00908172e-01
1.69341221e-01 3.12605053e-01 -4.74493861e-01 -6.12394571e-01
9.00004268e-01 7.64373004e-01 3.68555598e-02 3.03727090e-01
-4.39257890e-01 2.71588802e-01 3.92005116e-01 3.68114710e-01
1.27824569e+00 2.44692937e-01 -1.04670867e-01 -2.07469285e-01
6.06583416e-01 1.36807993e-01 4.71264213e-01 7.90993392e-01
-2.74368227e-01 3.72084260e-01 2.34584063e-01 -3.13130915e-01
-5.59512019e-01 -1.10153747e+00 -1.59683302e-02 1.16189599e+00
7.47684717e-01 -8.71180236e-01 -1.12555170e+00 -7.96935916e-01
-3.11831325e-01 2.74641961e-01 1.25877067e-01 2.97101915e-01
-6.46255791e-01 -8.18940997e-01 6.95349395e-01 6.38148308e-01
9.49876785e-01 -8.22091877e-01 -1.00933874e+00 2.56582588e-01
-3.22632104e-01 -1.61730409e+00 -8.22622061e-01 -4.07706201e-01
-9.46965516e-01 -1.20071900e+00 -6.47352695e-01 -6.46722376e-01
3.42578351e-01 9.61358190e-01 9.45443690e-01 3.87653798e-01
-3.34521711e-01 6.49714589e-01 -3.33930552e-01 8.79473090e-02
1.94712296e-01 3.64114374e-01 -2.63458610e-01 -1.44041553e-01
3.83555353e-01 -1.01202399e-01 -9.13352668e-01 7.62433231e-01
-8.09084535e-01 7.35357478e-02 1.75945535e-01 4.69732702e-01
7.45358765e-01 8.62579271e-02 2.14688435e-01 -8.40170443e-01
1.50814846e-01 -4.99222219e-01 -8.15376043e-01 1.62903398e-01
-5.35413980e-01 -5.83468139e-01 4.46238577e-01 -2.60193050e-01
-8.58829200e-01 7.52302632e-02 -2.31987774e-01 -3.98698658e-01
2.33711168e-01 2.65652299e-01 -1.99420139e-01 -3.61641459e-02
2.90630192e-01 3.64968032e-01 -1.45401508e-01 -7.36548007e-02
2.87886500e-01 9.17852640e-01 4.94843423e-01 -5.42722583e-01
3.64182323e-01 6.01844907e-01 -2.90841848e-01 -1.06714296e+00
-5.24305582e-01 -6.08321667e-01 -1.30955160e-01 -2.87729144e-01
8.44434500e-01 -1.29481256e+00 -9.82361674e-01 4.83867913e-01
-7.94243872e-01 -6.65925682e-01 4.24469933e-02 4.48425919e-01
-7.55904913e-01 5.10675848e-01 -8.40174675e-01 -3.84451032e-01
-4.34736371e-01 -1.62116826e+00 1.26789987e+00 3.78627628e-01
9.33510959e-02 -7.58309603e-01 -5.81237316e-01 6.98939562e-01
1.83854148e-01 -4.02477056e-01 3.26540828e-01 -1.93188295e-01
-1.22946417e+00 9.20763910e-02 -7.52509236e-01 7.95992911e-02
-1.67242706e-01 1.69535577e-01 -6.78563118e-01 -2.99518526e-01
-2.47894600e-01 2.05818415e-01 5.77930629e-01 4.79260981e-01
1.56125438e+00 -5.96638992e-02 -6.92378461e-01 9.12850082e-01
1.44653368e+00 5.06477356e-01 7.71454275e-01 5.12122750e-01
8.69294465e-01 -9.31692645e-02 1.19778061e+00 5.16386092e-01
6.69672847e-01 9.82918322e-01 3.66857976e-01 -1.47905171e-01
-1.34605199e-01 1.85751751e-01 5.21234572e-01 8.18219304e-01
-2.47548655e-01 -3.86313051e-01 -6.86859667e-01 4.87094343e-01
-2.05086803e+00 -9.15246964e-01 -5.56796312e-01 2.13706636e+00
3.48297387e-01 3.14296007e-01 4.82538581e-01 1.50944337e-01
6.76428676e-01 5.72460219e-02 -4.28919405e-01 -2.99885988e-01
2.61756957e-01 1.89143687e-01 7.33596385e-01 3.66422027e-01
-1.13921332e+00 1.26255286e+00 5.37428522e+00 1.31152666e+00
-1.06400537e+00 3.17431718e-01 6.86665416e-01 -2.84857959e-01
9.36677232e-02 -1.02952175e-01 -1.09805238e+00 9.08810318e-01
1.06811762e+00 4.19430584e-02 3.29604298e-01 1.03865242e+00
2.98522234e-01 -4.29065377e-01 -6.74662292e-01 1.29574156e+00
-4.60933186e-02 -1.51981211e+00 -2.77178377e-01 1.33929506e-01
4.00261045e-01 4.82570678e-02 3.70676517e-02 3.80921245e-01
-3.78929317e-01 -4.04027134e-01 7.68098414e-01 2.76838660e-01
8.26833487e-01 -9.66051221e-01 6.36797488e-01 8.22740123e-02
-1.63716364e+00 5.04376814e-02 -3.42051148e-01 1.52827442e-01
3.24668765e-01 5.05417407e-01 -6.60254419e-01 4.13049489e-01
9.74050283e-01 6.19971395e-01 -6.06108546e-01 1.14877534e+00
2.77678818e-01 7.25288153e-01 -4.81441289e-01 -7.73950573e-03
3.96334648e-01 -7.40031302e-02 4.25405502e-01 1.39474714e+00
4.99873579e-01 6.36983570e-03 3.88715893e-01 1.84473470e-01
-5.57716489e-02 7.95218274e-02 -1.32995054e-01 1.76846251e-01
5.84819257e-01 1.29947174e+00 -1.32699656e+00 -7.76912391e-01
-7.40502596e-01 1.36966157e+00 -2.95075834e-01 2.91543633e-01
-1.77377355e+00 -4.45307851e-01 7.24424124e-01 1.85700640e-01
6.66575015e-01 -4.37326163e-01 -1.53499439e-01 -1.17504168e+00
1.56452134e-01 -9.57442403e-01 2.92183548e-01 -5.99520981e-01
-4.50299323e-01 6.05274498e-01 -1.25460774e-01 -1.15150702e+00
5.21313697e-02 -6.13999069e-01 -2.35549271e-01 1.71582773e-01
-1.24017000e+00 -8.76768827e-01 -5.76159358e-01 8.88422728e-01
9.15985763e-01 -8.80989879e-02 2.45856345e-01 8.51990998e-01
-8.28321815e-01 8.29163432e-01 -5.08001857e-02 -4.07998413e-02
3.42793345e-01 -8.56973767e-01 5.49929738e-01 7.98147559e-01
3.54600996e-01 4.31628644e-01 2.57938713e-01 -6.37603521e-01
-1.76359332e+00 -1.23812246e+00 4.26550716e-01 -3.12925726e-01
4.57576066e-01 -9.72384065e-02 -4.67738837e-01 5.22719681e-01
7.25899125e-03 3.55011970e-01 5.82813680e-01 -3.46072078e-01
1.26925915e-01 -2.45902225e-01 -1.00069702e+00 1.02861679e+00
1.39583826e+00 -4.37647671e-01 5.12287170e-02 4.42053884e-01
7.40177870e-01 -8.36405337e-01 -7.40219951e-01 2.99812794e-01
5.93794107e-01 -1.17943060e+00 1.09506965e+00 -2.83618644e-02
1.87912909e-03 -6.70737863e-01 -2.49642640e-01 -4.09089208e-01
1.10858411e-01 -6.87972784e-01 -5.30287325e-01 1.01655638e+00
3.17869544e-01 -3.59392226e-01 1.22819519e+00 4.25754547e-01
-2.22751588e-01 -1.09178925e+00 -6.73602879e-01 -7.74101496e-01
-8.74593496e-01 -9.77561116e-01 6.77420080e-01 3.90540600e-01
-3.08830053e-01 8.46788660e-02 -1.62726805e-01 2.66757339e-01
5.58270752e-01 1.88835204e-01 1.06438088e+00 -4.12731975e-01
-5.27672589e-01 -1.30293578e-01 -5.55773973e-01 -1.76090705e+00
-1.69930637e-01 -5.86439133e-01 -3.03926647e-01 -1.22616160e+00
1.17675386e-01 -6.16205156e-01 1.56993363e-02 1.74959823e-01
-1.88759357e-01 8.34297359e-01 4.58680570e-01 1.92223251e-01
-1.38016236e+00 1.02008559e-01 9.54274833e-01 3.84398340e-03
-3.25482994e-01 2.93697953e-01 -4.34496731e-01 9.01782870e-01
6.15557969e-01 -2.63278574e-01 -6.77431047e-01 -5.07204473e-01
1.34725273e-01 2.68258423e-01 3.95106792e-01 -1.33790851e+00
3.47897887e-01 1.64550290e-01 9.24585164e-02 -7.90961385e-01
3.69400293e-01 -8.53148878e-01 1.24162093e-01 6.42264605e-01
3.33856672e-01 2.62057453e-01 3.10143679e-01 5.52036762e-01
-1.05956353e-01 -2.15241358e-01 3.59238416e-01 -9.54750521e-06
-1.23913765e+00 5.93924582e-01 -6.31682634e-01 1.03902914e-01
1.40848112e+00 -8.32533538e-01 -1.97584391e-01 -8.33282620e-02
-6.77230597e-01 3.58176440e-01 5.28587699e-01 3.11294347e-01
6.57325506e-01 -1.14712656e+00 -1.93241090e-01 1.22171134e-01
-1.76787138e-01 -2.61125088e-01 3.85249317e-01 1.19333589e+00
-1.09363604e+00 4.44161743e-01 1.54006064e-01 -1.01539004e+00
-1.71761334e+00 5.16543806e-01 5.45132011e-02 -5.74759319e-02
-7.19265699e-01 7.30783701e-01 1.10196315e-01 1.11547589e-01
1.68334723e-01 -3.32071453e-01 2.28794470e-01 8.21686611e-02
6.94324017e-01 9.89782035e-01 1.60333395e-01 -5.89128077e-01
-5.08244395e-01 6.55714691e-01 3.26487496e-02 -8.48055293e-04
7.17122793e-01 -4.62697238e-01 1.72377393e-01 -1.57211170e-01
1.00822937e+00 2.21923262e-01 -1.31024039e+00 1.10271741e-02
1.24668680e-01 -6.41379416e-01 -1.22329332e-02 -2.39844382e-01
-1.47344506e+00 5.97934186e-01 7.34147489e-01 1.92786038e-01
1.33311129e+00 -2.14540720e-01 1.42618990e+00 2.23065272e-01
8.20991278e-01 -1.14842963e+00 1.72568858e-03 3.71353269e-01
1.23303927e-01 -9.45597410e-01 2.31184140e-02 -8.90998006e-01
-4.59433019e-01 1.08099616e+00 6.95928931e-01 -4.87522855e-02
6.47275507e-01 5.84595799e-01 -2.33813915e-02 -1.47880152e-01
-3.97389174e-01 -2.83371300e-01 1.34918138e-01 3.58752042e-01
2.77403414e-01 2.69386340e-02 -4.52780098e-01 3.62506777e-01
8.43007937e-02 2.35292509e-01 3.67477298e-01 8.77491295e-01
-4.09066945e-01 -9.46309626e-01 -3.40315968e-01 5.53266764e-01
-7.75887489e-01 -2.62117028e-01 3.24027717e-01 7.74619102e-01
2.22561911e-01 1.03183699e+00 1.95866615e-01 -3.55397314e-01
1.79397374e-01 -3.99370611e-01 3.92820507e-01 -4.65268463e-01
-6.58874452e-01 3.12029243e-01 2.02254593e-01 -1.17812514e+00
-4.64274198e-01 -5.54808378e-01 -1.49760103e+00 -4.62727875e-01
-3.22750837e-01 7.11199492e-02 6.30412102e-01 7.23670304e-01
7.56140947e-01 6.12812042e-01 2.54999578e-01 -7.59238482e-01
-5.44886515e-02 -2.76967943e-01 -3.84596646e-01 -8.80463943e-02
-2.34574616e-01 -5.32880664e-01 1.93771557e-03 1.18250966e-01] | [9.068193435668945, -0.12996792793273926] |
7698d7aa-043c-4194-9d20-f480b89f0f7a | style-transfer-based-image-synthesis-as-an | 1905.10974 | null | https://arxiv.org/abs/1905.10974v1 | https://arxiv.org/pdf/1905.10974v1.pdf | Style transfer-based image synthesis as an efficient regularization technique in deep learning | These days deep learning is the fastest-growing area in the field of Machine Learning. Convolutional Neural Networks are currently the main tool used for image analysis and classification purposes. Although great achievements and perspectives, deep neural networks and accompanying learning algorithms have some relevant challenges to tackle. In this paper, we have focused on the most frequently mentioned problem in the field of machine learning, that is relatively poor generalization abilities. Partial remedies for this are regularization techniques e.g. dropout, batch normalization, weight decay, transfer learning, early stopping and data augmentation. In this paper, we have focused on data augmentation. We propose to use a method based on a neural style transfer, which allows generating new unlabeled images of a high perceptual quality that combine the content of a base image with the appearance of another one. In a proposed approach, the newly created images are described with pseudo-labels, and then used as a training dataset. Real, labeled images are divided into the validation and test set. We validated the proposed method on a challenging skin lesion classification case study. Four representative neural architectures are examined. Obtained results show the strong potential of the proposed approach. | ['Michał Grochowski', 'Agnieszka Mikołajczyk'] | 2019-05-27 | null | null | null | null | ['skin-lesion-classification'] | ['medical'] | [ 7.11993217e-01 1.85699821e-01 -5.67728244e-02 -3.82872462e-01
-1.63573012e-01 2.30337493e-02 4.59126979e-01 2.97335595e-01
-8.57686996e-01 8.92190456e-01 -2.56385595e-01 8.14270675e-02
-3.25750895e-02 -6.45641685e-01 -5.43370962e-01 -8.27919185e-01
2.23700479e-01 2.00414568e-01 1.68116763e-01 -5.57541773e-02
1.33849606e-01 6.15025401e-01 -1.65979826e+00 1.42004550e-01
1.01037157e+00 8.78763080e-01 1.46651819e-01 3.19296122e-01
-2.81618178e-01 7.79447556e-01 -5.58196425e-01 -1.57775819e-01
1.83775984e-02 -5.09486437e-01 -6.67368770e-01 5.04843235e-01
2.12801158e-01 -1.72692612e-01 1.47047536e-02 1.10651791e+00
5.25572538e-01 1.54387832e-01 6.74387395e-01 -1.00625110e+00
-3.91706735e-01 2.34498426e-01 -4.86274183e-01 1.19040899e-01
-3.53664815e-01 -6.49048164e-02 2.92083651e-01 -7.31039703e-01
6.02814376e-01 9.04503703e-01 4.83243078e-01 9.40700233e-01
-1.22730243e+00 -4.50027078e-01 8.02530944e-02 4.76223022e-01
-1.16089654e+00 -2.40754902e-01 8.94455791e-01 -4.72086698e-01
2.66063720e-01 1.83513105e-01 4.50581878e-01 1.09383714e+00
-4.10103286e-03 7.51309812e-01 1.43326092e+00 -6.99807048e-01
3.93211842e-01 7.27802932e-01 3.09450418e-01 3.85486543e-01
2.52616674e-01 -1.95772499e-01 -5.83630847e-03 1.09361149e-01
4.89617914e-01 -5.90990186e-02 -2.67668694e-01 -3.10684592e-01
-5.71217477e-01 7.77830184e-01 5.70840657e-01 6.96496606e-01
-5.83051085e-01 -1.92565039e-01 3.49668413e-01 1.30960960e-02
7.00241446e-01 1.65513784e-01 -2.55759269e-01 2.83695787e-01
-7.79512227e-01 -1.75364334e-02 4.32231516e-01 3.53347778e-01
6.47239864e-01 8.37367848e-02 -1.92264274e-01 9.80033219e-01
3.55408005e-02 7.42044449e-02 6.77219272e-01 -1.92266077e-01
6.61588237e-02 7.80992329e-01 -1.03493430e-01 -8.89476240e-01
-5.20977736e-01 -8.14231694e-01 -1.07885075e+00 7.33411908e-01
4.13360894e-01 -2.72430182e-01 -1.05351818e+00 1.66533637e+00
3.71086836e-01 8.32314417e-02 1.94586784e-01 9.71692026e-01
7.98100054e-01 6.06636107e-01 2.93937176e-01 -2.76449025e-01
1.01328707e+00 -1.01733124e+00 -7.96170354e-01 -3.70358452e-02
4.21973795e-01 -6.78251445e-01 8.34209383e-01 7.09469676e-01
-9.38245535e-01 -8.08074355e-01 -9.79587197e-01 1.83472991e-01
-4.55999196e-01 6.65484250e-01 2.98194110e-01 5.78943133e-01
-8.86710286e-01 7.13854849e-01 -6.63977623e-01 -5.56615889e-01
4.64358926e-01 3.39183867e-01 -4.71448720e-01 -1.22949533e-01
-8.02850187e-01 7.87947774e-01 7.50654280e-01 4.12939936e-01
-6.89876199e-01 -2.23440647e-01 -3.98551673e-01 8.70140493e-02
2.51756430e-01 -1.72372073e-01 9.77855563e-01 -1.60067618e+00
-1.45633674e+00 9.38234687e-01 3.70103449e-01 -4.26058173e-01
5.57881296e-01 -1.18515380e-01 -5.31865120e-01 3.05686216e-03
-2.81729579e-01 4.62772965e-01 9.26813066e-01 -1.36346829e+00
-5.93336403e-01 -3.59309644e-01 -2.71906376e-01 -1.27993152e-02
-8.21445942e-01 -6.74825832e-02 -2.25623250e-01 -6.14206731e-01
3.01596038e-02 -8.47897887e-01 -2.07865596e-01 -4.78258282e-02
-3.41793358e-01 -1.08296633e-01 8.79662812e-01 -6.73911572e-01
9.54614162e-01 -2.21584988e+00 1.09395385e-01 3.16936076e-01
-6.34463429e-02 9.49448824e-01 -2.21949980e-01 1.95058569e-01
-3.15617532e-01 -9.50885415e-02 -5.36685884e-01 -3.06033552e-01
-4.69767332e-01 1.71770111e-01 2.35070899e-01 3.57558370e-01
5.60206950e-01 2.97385931e-01 -5.98500252e-01 -3.64080817e-01
3.43433052e-01 7.48820543e-01 -1.80869117e-01 3.31122220e-01
-1.59759656e-01 7.63218760e-01 -2.39048526e-01 2.90734321e-01
8.54144812e-01 -2.24719271e-02 1.88818220e-02 -3.70710224e-01
-1.19121838e-02 -4.47920352e-01 -1.19975793e+00 1.40471518e+00
-4.65138406e-01 5.85629404e-01 2.64355958e-01 -1.31333220e+00
1.19367385e+00 2.66118973e-01 2.66588926e-01 -7.02366114e-01
5.55981338e-01 3.98343295e-01 3.04281145e-01 -8.51018071e-01
1.85635075e-01 -3.78398038e-02 5.65319955e-01 1.90261528e-01
2.23253921e-01 1.90668881e-01 4.27974582e-01 -2.74383724e-01
6.70154989e-01 1.27723053e-01 1.65764287e-01 -5.69073930e-02
1.04935801e+00 3.60303558e-02 4.79146391e-01 2.39228755e-01
-8.58616531e-02 5.93394458e-01 4.85434204e-01 -4.65598822e-01
-1.30555153e+00 -4.81070995e-01 -2.83673435e-01 7.81737328e-01
-2.17647016e-01 2.05101326e-01 -1.10916388e+00 -7.10009098e-01
-2.74036109e-01 4.43240315e-01 -7.95434952e-01 -2.05721676e-01
-6.12514079e-01 -9.24535751e-01 3.39934707e-01 2.58247614e-01
6.70661807e-01 -1.34022772e+00 -4.33647811e-01 2.87354380e-01
2.77321517e-01 -9.14222360e-01 2.45367885e-01 2.32987165e-01
-9.50009704e-01 -1.04234767e+00 -1.14137197e+00 -1.04960537e+00
1.03602147e+00 -4.28030342e-02 6.61100745e-01 1.26038790e-01
-6.60131693e-01 -7.88800418e-02 -4.55450892e-01 -2.71126479e-01
-6.64912760e-01 1.82306871e-01 -1.47890925e-01 6.45054758e-01
2.49586720e-02 -4.58882868e-01 -5.03749669e-01 1.20268114e-01
-1.24901843e+00 6.60864860e-02 1.13867319e+00 1.02735031e+00
4.82863128e-01 4.84615229e-02 7.34092832e-01 -1.23127711e+00
7.25607574e-01 -3.54933560e-01 -5.24189413e-01 1.63220018e-01
-5.27040899e-01 -8.31863144e-05 7.78405309e-01 -4.87316728e-01
-1.27290225e+00 1.64214075e-01 -4.01775450e-01 -2.55734563e-01
-6.20071590e-01 6.34110153e-01 -8.89938846e-02 -1.96864113e-01
8.35138977e-01 1.72513753e-01 3.20452124e-01 -6.85072184e-01
1.23721711e-01 6.98027492e-01 3.54861736e-01 -2.02426210e-01
5.09246051e-01 3.06795865e-01 1.21896453e-01 -9.87518072e-01
-5.22867918e-01 -3.11953783e-01 -5.78484297e-01 -3.68802488e-01
8.60519826e-01 -4.00355190e-01 -3.37972373e-01 7.95632184e-01
-1.00200355e+00 -1.95169762e-01 -3.04707080e-01 6.11737192e-01
-1.33388758e-01 2.04804495e-01 -4.60244864e-01 -9.32161212e-01
-4.06757534e-01 -9.68936563e-01 3.82238626e-01 5.76180696e-01
1.74377918e-01 -9.49304819e-01 7.31678903e-02 2.60872841e-01
6.21572614e-01 4.46840137e-01 9.36356008e-01 -8.33361864e-01
-6.31517544e-02 -3.94893140e-01 -3.37267905e-01 1.15542996e+00
3.25186551e-01 8.25396031e-02 -1.14836335e+00 -3.40309322e-01
3.77390198e-02 -5.03298223e-01 9.11372244e-01 2.36939982e-01
1.32887149e+00 -7.86793455e-02 -1.40852764e-01 3.37009013e-01
1.53786397e+00 3.33460897e-01 6.55714750e-01 3.13761085e-01
4.48017389e-01 9.00113881e-01 5.36893487e-01 2.34360874e-01
-3.34462047e-01 4.03605431e-01 6.38686776e-01 -6.40065253e-01
-3.16792160e-01 2.00424835e-01 2.99784038e-02 7.33951688e-01
-3.04297060e-01 -1.46816745e-01 -7.69581914e-01 3.96504074e-01
-1.76213443e+00 -6.75435364e-01 -3.12334180e-01 2.17229319e+00
4.99763012e-01 2.01281190e-01 2.90320888e-02 5.75067222e-01
9.97595012e-01 -1.89737469e-01 -3.49249452e-01 -3.14928085e-01
-2.01447234e-01 5.59741199e-01 2.13498920e-01 1.76035047e-01
-1.31328022e+00 5.35680890e-01 4.67764091e+00 8.90859962e-01
-1.51391268e+00 2.23547518e-02 7.69619226e-01 3.09650898e-01
1.64833516e-01 -4.82101381e-01 -4.32853311e-01 3.79817784e-01
6.95501983e-01 2.85257071e-01 9.50955525e-02 7.55411625e-01
2.29502499e-01 -2.23133117e-01 -7.14881241e-01 8.54309440e-01
2.99816400e-01 -9.55928385e-01 3.59540768e-02 -2.05628365e-01
7.79429018e-01 -4.02441800e-01 2.91614950e-01 1.23545766e-01
-5.35935581e-01 -1.00794065e+00 2.09225103e-01 4.59355175e-01
6.51529908e-01 -7.86434531e-01 1.16931987e+00 2.31891870e-01
-6.20145798e-01 -1.87428117e-01 -2.25529909e-01 3.13015990e-02
-1.46237388e-01 5.61200738e-01 -9.56922829e-01 6.37238085e-01
4.14621174e-01 4.78621662e-01 -6.45865738e-01 1.42554748e+00
-1.78856090e-01 6.37290597e-01 -1.35264605e-01 -2.75365382e-01
3.70784104e-01 -3.64803344e-01 1.36954620e-01 1.05875444e+00
2.12886989e-01 -1.68354899e-01 -7.31612667e-02 7.97226071e-01
2.42596511e-02 5.84563732e-01 -4.84652251e-01 -1.30248502e-01
-2.33440340e-01 1.59110689e+00 -9.48695600e-01 -3.25936556e-01
-2.68569887e-01 8.98129046e-01 1.94094241e-01 2.40899891e-01
-5.23160338e-01 -3.72939020e-01 -4.45223600e-02 1.88143954e-01
-1.96294370e-03 1.69755489e-01 -7.07835257e-02 -8.88611794e-01
7.56104439e-02 -7.84857988e-01 2.37217948e-01 -5.18192470e-01
-1.21881354e+00 1.00151849e+00 -2.17622846e-01 -1.23731101e+00
-7.44246915e-02 -8.41639161e-01 -6.51032865e-01 8.44681203e-01
-1.54073155e+00 -1.00757539e+00 -8.71366203e-01 6.11866653e-01
5.71690619e-01 -4.02778655e-01 8.19102526e-01 7.44737625e-01
-7.54468977e-01 3.18877548e-01 3.24597299e-01 1.18344836e-01
6.39971197e-01 -1.07160723e+00 -1.83725804e-01 8.18877041e-01
-5.63585088e-02 1.51560083e-02 5.90705931e-01 -4.55962330e-01
-6.45766258e-01 -1.09667075e+00 4.93324995e-01 4.70859200e-01
3.35306317e-01 -2.24557862e-01 -1.12978995e+00 1.99720666e-01
2.28365079e-01 1.95471108e-01 5.57289660e-01 -2.49988198e-01
1.50695071e-01 -3.07318538e-01 -1.31240892e+00 3.42349976e-01
2.71600962e-01 -9.51880962e-03 -2.99561262e-01 2.43998244e-01
8.00958648e-02 -1.01390809e-01 -4.42280769e-01 4.67066020e-01
2.31628746e-01 -9.95569289e-01 4.98115212e-01 -6.47933304e-01
4.43688691e-01 -2.06983477e-01 2.90350169e-01 -1.36904895e+00
-6.02577291e-02 -1.56995460e-01 3.36160451e-01 1.35850358e+00
2.97701597e-01 -4.48860258e-01 1.12478161e+00 3.35710198e-01
-3.62509191e-02 -8.55717778e-01 -6.40723765e-01 -5.10976851e-01
3.87182645e-03 9.88902152e-02 -2.70297006e-02 9.28386986e-01
-4.89144236e-01 3.24998975e-01 -4.25769508e-01 -1.63709253e-01
6.92536592e-01 -3.78692836e-01 5.79488695e-01 -1.32274854e+00
-1.21669568e-01 -2.65486687e-01 -5.41843295e-01 -1.68559834e-01
-7.70475541e-04 -6.97572649e-01 -5.40902354e-02 -1.49574208e+00
1.57209605e-01 -6.08021200e-01 -5.14167070e-01 3.57543290e-01
-7.85901919e-02 4.43803221e-01 8.09428021e-02 -1.21667154e-01
-5.74545749e-02 5.18646836e-01 1.19142413e+00 -2.47693032e-01
-2.32073724e-01 2.49563858e-01 -9.03818980e-02 7.38705099e-01
1.21601987e+00 -2.81790555e-01 -3.88251781e-01 -2.82752275e-01
-3.27136576e-01 -2.54244149e-01 3.19722265e-01 -1.30305099e+00
1.52849406e-01 2.03017697e-01 4.86633718e-01 -1.95014104e-01
4.48894203e-01 -1.06041706e+00 1.23618715e-01 7.64397025e-01
-4.33236510e-01 -3.59569460e-01 2.68511623e-01 3.08199733e-01
-4.68116254e-01 -6.95479691e-01 1.06340969e+00 -5.22096977e-02
-8.35934460e-01 1.93487689e-01 -3.27287108e-01 -3.98734808e-01
1.27854323e+00 -1.38400570e-01 -1.28729045e-01 -4.11568768e-02
-1.21214461e+00 -2.04874039e-01 1.07764684e-01 3.43494058e-01
6.69470906e-01 -1.10786891e+00 -8.82713377e-01 2.10655615e-01
1.33050710e-01 -1.71200156e-01 5.34794807e-01 8.86207342e-01
-6.00842297e-01 6.52666315e-02 -8.81577253e-01 -4.54540640e-01
-1.47957754e+00 6.27325058e-01 3.57070595e-01 -8.51399601e-02
-4.11298394e-01 6.35873139e-01 -8.23042914e-03 -2.26193234e-01
4.66568649e-01 -1.85533732e-01 -7.18729675e-01 -1.47571694e-02
4.73425031e-01 3.60691309e-01 3.56486648e-01 -4.67081130e-01
-5.96670285e-02 4.48125750e-01 -2.35091135e-01 1.13058209e-01
1.40543294e+00 3.14764053e-01 -2.34571606e-01 3.39707673e-01
1.08253741e+00 -2.69204915e-01 -9.03997540e-01 -1.81215078e-01
7.00990856e-02 -2.92316079e-01 1.13752581e-01 -9.52229798e-01
-1.25862038e+00 1.09812653e+00 1.21452451e+00 3.83608282e-01
1.30187082e+00 -3.47772568e-01 3.94152343e-01 2.97130048e-01
3.80853983e-03 -1.11089957e+00 8.08663443e-02 6.66777864e-02
8.12626898e-01 -1.31919515e+00 -2.50019491e-01 -4.35144246e-01
-4.12267655e-01 1.41117847e+00 7.56623507e-01 -3.45026404e-01
4.86435920e-01 3.77150625e-02 1.50769740e-01 1.42514244e-01
-4.03105199e-01 -4.00742441e-01 1.57521918e-01 6.30615771e-01
5.80219328e-01 -2.32146725e-01 -7.84068823e-01 4.43316281e-01
5.27952969e-01 3.75977695e-01 5.18420637e-01 8.40572119e-01
-3.97417277e-01 -1.35058725e+00 -5.19730628e-01 4.04216796e-01
-6.29034221e-01 2.78259933e-01 -2.91277736e-01 8.94177794e-01
5.33629894e-01 6.95143878e-01 -1.52793080e-02 -2.53917903e-01
3.48781765e-01 -6.20567352e-02 5.51923811e-01 -4.58453655e-01
-6.74127221e-01 -9.60944444e-02 -5.64439176e-03 -1.84578225e-02
-5.84213018e-01 -2.68500119e-01 -8.93515825e-01 1.53576374e-01
-3.09759974e-01 3.03456515e-01 1.11711371e+00 7.09686220e-01
2.96769850e-02 7.75021493e-01 5.74289322e-01 -7.52751946e-01
-5.38416386e-01 -1.12816024e+00 -5.32648385e-01 7.40907490e-01
2.21369088e-01 -6.18988812e-01 -2.53445357e-01 1.76270887e-01] | [14.99181842803955, -2.552314281463623] |
236b6a09-27b8-439e-a001-10e289bb6acd | let-s-see-clearly-contaminant-artifact | 2104.08852 | null | https://arxiv.org/abs/2104.08852v1 | https://arxiv.org/pdf/2104.08852v1.pdf | Let's See Clearly: Contaminant Artifact Removal for Moving Cameras | Contaminants such as dust, dirt and moisture adhering to the camera lens can greatly affect the quality and clarity of the resulting image or video. In this paper, we propose a video restoration method to automatically remove these contaminants and produce a clean video. Our approach first seeks to detect attention maps that indicate the regions that need to be restored. In order to leverage the corresponding clean pixels from adjacent frames, we propose a flow completion module to hallucinate the flow of the background scene to the attention regions degraded by the contaminants. Guided by the attention maps and completed flows, we propose a recurrent technique to restore the input frame by fetching clean pixels from adjacent frames. Finally, a multi-frame processing stage is used to further process the entire video sequence in order to enforce temporal consistency. The entire network is trained on a synthetic dataset that approximates the physical lighting properties of contaminant artifacts. This new dataset and our novel framework lead to our method that is able to address different contaminants and outperforms competitive restoration approaches both qualitatively and quantitatively. | ['Pedro V. Sander', 'Jing Liao', 'Bo Zhang', 'Xiaoyu Li'] | 2021-04-18 | null | http://openaccess.thecvf.com//content/ICCV2021/html/Li_Lets_See_Clearly_Contaminant_Artifact_Removal_for_Moving_Cameras_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Li_Lets_See_Clearly_Contaminant_Artifact_Removal_for_Moving_Cameras_ICCV_2021_paper.pdf | iccv-2021-1 | ['video-restoration'] | ['computer-vision'] | [ 5.35004377e-01 -2.95388967e-01 5.92408359e-01 5.88717014e-02
-3.40658337e-01 -4.22710329e-01 5.37127435e-01 -1.03364035e-01
-3.72502841e-02 6.67968333e-01 4.66211706e-01 2.23672688e-02
2.02156216e-01 -6.17619812e-01 -8.39758873e-01 -7.81947136e-01
1.82607740e-01 -5.87254286e-01 2.17044294e-01 2.78639253e-02
5.72108030e-01 6.88885748e-01 -1.72345650e+00 4.30292577e-01
1.00075448e+00 7.36380696e-01 5.30641496e-01 8.13073158e-01
2.74930820e-02 1.10444081e+00 -5.45826375e-01 6.56098276e-02
4.31915849e-01 -6.35708690e-01 -5.47798991e-01 6.48424864e-01
8.26713324e-01 -5.42595387e-01 -5.92404068e-01 1.13843632e+00
1.66798994e-01 5.99742413e-01 3.75589967e-01 -8.17952871e-01
-5.32535374e-01 9.70442593e-02 -7.17405498e-01 3.23599279e-01
5.19304454e-01 4.20633376e-01 3.63649756e-01 -1.05900860e+00
6.99957490e-01 1.27077878e+00 3.78841311e-01 5.02185583e-01
-1.13209450e+00 -2.93066114e-01 5.02719402e-01 3.34607452e-01
-9.69403923e-01 -7.37479389e-01 8.36275876e-01 -4.48955953e-01
7.09400773e-01 3.67819130e-01 6.92962527e-01 1.10108757e+00
3.31593573e-01 3.49664718e-01 8.75787437e-01 -4.19885635e-01
3.30344230e-01 -2.02982783e-01 -8.70772377e-02 4.07942176e-01
1.85129002e-01 -4.43369709e-02 -6.25242114e-01 1.27184153e-01
5.76293230e-01 3.42630595e-01 -9.29954648e-01 -1.29585356e-01
-9.46510434e-01 1.04584537e-01 3.59111279e-01 6.40369430e-02
-7.29647219e-01 1.86563209e-01 2.37244457e-01 5.06708547e-02
6.97154701e-01 4.55208182e-01 -1.42333731e-02 2.21271977e-01
-1.02882349e+00 6.68914616e-02 6.10618174e-01 5.40327370e-01
9.50898170e-01 1.42659053e-01 -4.19610232e-01 8.03469419e-01
2.80171394e-01 3.96834075e-01 -1.16009660e-01 -1.33469689e+00
3.43055308e-01 3.34460825e-01 5.23260713e-01 -1.19883752e+00
2.16759205e-01 -2.08461851e-01 -9.18492317e-01 4.98838574e-01
1.29919916e-01 4.41268682e-02 -9.86193240e-01 1.30103517e+00
3.36473107e-01 6.51903093e-01 7.88029656e-03 1.17168939e+00
4.26880598e-01 9.78615224e-01 -7.32067674e-02 -4.01273876e-01
1.03571081e+00 -1.06809306e+00 -7.96824753e-01 -3.80024493e-01
-1.68967500e-01 -9.53370452e-01 8.99447918e-01 5.92988610e-01
-1.31046116e+00 -6.01266921e-01 -1.08122540e+00 -5.89820035e-02
6.47201091e-02 -3.83647978e-02 2.20747385e-02 1.66748971e-01
-1.18951285e+00 9.22195971e-01 -8.44855487e-01 -3.94412279e-01
4.08574849e-01 -9.69504938e-02 -2.72050351e-01 -6.49546623e-01
-7.64553785e-01 8.36864293e-01 4.22973298e-02 6.57011807e-01
-1.30121505e+00 -6.96862698e-01 -1.05101943e+00 1.99635595e-01
2.38435343e-01 -6.49673223e-01 6.06095493e-01 -1.17593765e+00
-1.38083780e+00 3.77317101e-01 -5.64985394e-01 -1.73845395e-01
5.65447628e-01 -5.54516256e-01 -1.32761732e-01 4.30754423e-01
1.73220992e-01 4.71330374e-01 1.22250092e+00 -1.62283874e+00
-7.17408776e-01 1.38209946e-02 -7.32414648e-02 1.65303424e-01
-2.62083769e-01 4.37901244e-02 -8.80024672e-01 -7.56048143e-01
-7.69591182e-02 -6.26315534e-01 -2.88860410e-01 -5.14098722e-03
-5.89899123e-01 4.98299360e-01 8.50217998e-01 -9.89646733e-01
1.04312015e+00 -2.24452496e+00 3.77583176e-01 6.66970238e-02
1.61137402e-01 2.75947064e-01 -4.14734542e-01 3.71161371e-01
-1.43016785e-01 -2.11105645e-01 -5.16241014e-01 -6.71537519e-01
-6.37667894e-01 9.16186720e-02 -4.88908082e-01 7.63870776e-01
4.50400382e-01 4.16602015e-01 -9.85665500e-01 -9.61291343e-02
6.13115072e-01 9.44895804e-01 -4.89854395e-01 5.08719146e-01
-9.25503969e-02 7.45881975e-01 -5.04131056e-02 4.85038966e-01
1.07231903e+00 1.59374088e-01 -6.85018227e-02 -3.67886305e-01
-3.89349669e-01 1.26496498e-02 -1.10392427e+00 1.59084678e+00
-2.98057884e-01 8.68032217e-01 3.96834671e-01 -6.36503339e-01
7.70128429e-01 1.01203017e-01 4.85601425e-01 -6.60472870e-01
-2.08581910e-02 7.63254613e-02 -5.21810472e-01 -8.49239171e-01
6.35785937e-01 1.96451619e-01 6.07129931e-01 1.61551654e-01
-3.05933595e-01 1.01978764e-01 4.25111830e-01 1.83216795e-01
1.19932246e+00 4.33350205e-01 -2.86818862e-01 -2.76959479e-01
8.37407768e-01 -2.91464597e-01 7.12855399e-01 6.14387214e-01
5.97543968e-03 1.00062060e+00 2.76333898e-01 -6.49642110e-01
-1.10498893e+00 -8.91872048e-01 3.16215038e-01 6.53467417e-01
5.35779595e-01 -2.20739707e-01 -8.65149617e-01 -2.98245937e-01
-3.55132997e-01 5.70672810e-01 -6.75044656e-01 -1.99556783e-01
-7.36767709e-01 -5.40492117e-01 -1.40447438e-01 1.60541624e-01
3.90558094e-01 -1.29176486e+00 -6.52300715e-01 3.08545321e-01
-4.40699667e-01 -1.13869512e+00 -6.61076188e-01 1.97407063e-02
-6.32467628e-01 -1.30440581e+00 -8.05145144e-01 -6.14716113e-01
9.77648020e-01 6.81170523e-01 8.39879036e-01 3.47002059e-01
-4.18583155e-01 1.92184702e-01 -3.77496570e-01 8.69571865e-02
-4.81680810e-01 -5.78476012e-01 -2.31111810e-01 6.54328406e-01
-2.52539605e-01 -5.89865625e-01 -9.33421791e-01 3.51088680e-02
-1.18074954e+00 2.92551368e-01 2.91677415e-01 5.65267742e-01
4.39633340e-01 6.79742023e-02 6.99239001e-02 -5.84726095e-01
4.87085551e-01 -4.44936693e-01 -5.58530569e-01 1.66085958e-01
-1.29334047e-01 -5.95503077e-02 7.74046957e-01 -2.81385541e-01
-1.35114467e+00 1.73553899e-01 2.14169353e-01 -8.03814948e-01
-3.24436247e-01 2.43096370e-02 -2.55514055e-01 8.85795802e-02
5.04662037e-01 2.43029892e-01 -2.01794431e-01 -6.05593920e-01
2.49997169e-01 3.87968272e-01 9.16591763e-01 -2.15548515e-01
8.67148221e-01 8.41694355e-01 -1.06789008e-01 -8.33490789e-01
-8.57507348e-01 -4.28302556e-01 -5.64114153e-01 -5.97668588e-01
9.82305288e-01 -7.90545642e-01 -4.23236370e-01 6.43952727e-01
-1.54316163e+00 -3.57926160e-01 -1.22639462e-01 3.05049509e-01
-2.93892324e-01 6.80360079e-01 -8.07206213e-01 -9.11709666e-01
-3.53836089e-01 -1.17383170e+00 9.60660100e-01 3.62018943e-01
7.17461184e-02 -7.29384184e-01 -3.14724166e-04 6.52666390e-02
4.88806754e-01 3.21056783e-01 6.27434313e-01 3.11456323e-01
-9.30710196e-01 3.13274413e-01 -2.91876853e-01 6.58812761e-01
5.86189210e-01 3.09819430e-01 -1.18624234e+00 -3.49291831e-01
2.54345119e-01 4.03316289e-01 1.21833599e+00 5.31305075e-01
9.71129537e-01 -2.73641914e-01 -1.99391454e-01 8.82331789e-01
1.47467637e+00 3.01556736e-01 1.03857219e+00 5.20116806e-01
8.57341290e-01 8.46994400e-01 5.17408967e-01 2.92436570e-01
-9.27945748e-02 2.53290266e-01 6.84065521e-01 -4.17151779e-01
-6.24574482e-01 2.44272333e-02 6.69923902e-01 4.72568870e-01
-1.26503604e-02 -5.36190867e-01 -5.69060743e-01 7.47904837e-01
-1.81295395e+00 -9.67610598e-01 -2.71692187e-01 2.19020247e+00
4.73302841e-01 -1.20821230e-01 -4.95034784e-01 4.19745855e-02
1.01426375e+00 2.39945948e-01 -4.32643116e-01 -1.59517869e-01
-1.56698182e-01 7.60403499e-02 2.44171873e-01 8.92158628e-01
-9.93527472e-01 8.88359129e-01 6.24427032e+00 2.86222905e-01
-1.15962422e+00 -2.48998657e-01 6.55795813e-01 -1.84061602e-01
-3.71479154e-01 5.24099879e-02 -2.24478632e-01 5.41179955e-01
6.46029830e-01 2.32587114e-01 7.12305725e-01 1.95226356e-01
8.43340397e-01 -3.97107780e-01 -9.94136035e-01 8.48267734e-01
3.57882828e-01 -1.24323773e+00 1.65725917e-01 -1.42959610e-01
8.47716630e-01 -1.54863045e-01 -1.54676601e-01 -4.35310096e-01
4.62776385e-02 -8.49720478e-01 7.16982067e-01 9.94694352e-01
5.21045148e-01 -6.39246225e-01 4.34459597e-01 -4.98360135e-02
-9.32207406e-01 -7.15170503e-02 -3.22685212e-01 -6.10705018e-02
4.40254837e-01 8.92254651e-01 -4.40691888e-01 5.55310726e-01
8.75549197e-01 1.18486047e+00 -4.70604748e-01 1.28914523e+00
-4.20910597e-01 3.71164083e-01 1.08824082e-01 6.65878057e-01
5.33462912e-02 -5.49781144e-01 8.13807368e-01 1.37342680e+00
5.09593904e-01 -1.15148783e-01 4.90165548e-03 9.99456465e-01
-9.69779212e-04 -1.84578314e-01 -6.12823069e-01 3.98324758e-01
1.20536342e-01 1.27269685e+00 -8.17150712e-01 -4.17683750e-01
-4.61379528e-01 1.47388375e+00 3.45963956e-04 8.50093246e-01
-6.67898953e-01 -2.22994357e-01 9.19981182e-01 1.69851318e-01
3.08293015e-01 -5.85952848e-02 -3.08193290e-03 -1.34816420e+00
3.04584056e-01 -7.90672839e-01 -1.06236666e-01 -1.14319074e+00
-1.14324212e+00 6.82033122e-01 -4.88525271e-01 -1.27136457e+00
1.89794362e-01 -2.62831479e-01 -8.60955954e-01 1.09328055e+00
-2.04039645e+00 -6.33791089e-01 -9.02528048e-01 4.85001117e-01
8.10230315e-01 1.53038576e-01 3.25475335e-01 4.56204832e-01
-7.86998451e-01 -2.97769547e-01 2.79741473e-02 -8.22040439e-02
9.00482595e-01 -9.45331156e-01 5.27365327e-01 1.77083218e+00
-2.54540741e-01 5.23662031e-01 8.72695088e-01 -8.25698137e-01
-1.28036630e+00 -1.36358535e+00 5.14682174e-01 -1.62625775e-01
2.61331558e-01 -2.43744522e-01 -1.18186188e+00 3.63914073e-01
5.03967285e-01 6.92562386e-02 4.66163494e-02 -6.17403865e-01
-2.16223255e-01 -2.87196487e-01 -8.84663045e-01 6.13924026e-01
8.27496111e-01 -4.42074060e-01 -3.74319822e-01 1.24561675e-01
5.46006739e-01 -2.35912398e-01 -3.69676203e-01 2.24566385e-01
2.54604369e-01 -1.19430268e+00 9.37259138e-01 -2.28652894e-01
8.23379159e-01 -7.59988189e-01 -1.23349037e-02 -1.46718502e+00
-2.86590189e-01 -8.26582313e-01 -2.86878943e-02 1.28781676e+00
1.49551183e-02 -3.68158489e-01 4.60743338e-01 6.50266945e-01
-4.80915010e-01 -1.00621760e-01 -5.06608546e-01 -3.27094585e-01
-4.64965045e-01 -2.34758392e-01 2.86730051e-01 7.54415274e-01
-4.21810091e-01 -4.40195296e-03 -7.82362103e-01 4.89921600e-01
8.51505756e-01 1.07476056e-01 6.11814976e-01 -9.92849886e-01
-1.20653711e-01 -1.56968385e-01 -5.31396270e-02 -9.49994028e-01
1.09587193e-01 -4.71041411e-01 4.87807423e-01 -1.70001721e+00
3.38603914e-01 1.09625049e-02 -2.93108195e-01 2.32523695e-01
-4.92399782e-01 4.28302586e-01 2.56803513e-01 2.45612502e-01
-5.27344704e-01 4.98622388e-01 1.20879030e+00 -2.32872710e-01
-3.66794199e-01 -4.31466073e-01 -7.10367918e-01 6.25229120e-01
5.15031457e-01 -3.54986995e-01 -2.18988672e-01 -8.24455976e-01
-2.61561722e-02 7.07689673e-02 5.99762917e-01 -1.01895082e+00
3.03506196e-01 -2.63396114e-01 6.95909619e-01 -4.27952528e-01
3.39077026e-01 -8.56406331e-01 2.65062809e-01 4.28478211e-01
-1.49004087e-01 -9.06118825e-02 1.50854617e-01 6.75393820e-01
-2.93346941e-01 -2.15082377e-01 9.68972027e-01 -1.39641032e-01
-5.69712460e-01 1.73670962e-01 -6.90401316e-01 -2.86225170e-01
9.10280406e-01 -2.34029666e-01 -3.41962278e-01 -3.24536502e-01
-6.39909267e-01 1.98023021e-01 8.98903310e-01 5.12337208e-01
8.65018547e-01 -9.09797192e-01 -8.27475429e-01 3.68812531e-01
-1.70680687e-01 -2.27315295e-02 5.08826554e-01 6.83023870e-01
-8.66732538e-01 -9.58086774e-02 -3.18726838e-01 -4.58132982e-01
-1.16123009e+00 7.22649097e-01 4.77198452e-01 2.77737647e-01
-1.05347991e+00 5.29454350e-01 4.16184515e-01 3.50880831e-01
2.86428928e-01 -5.01985073e-01 -2.26438925e-01 -7.98180923e-02
9.44088697e-01 6.38895273e-01 9.75914001e-02 -5.40059805e-01
-1.72378018e-01 6.60660148e-01 1.23980306e-01 8.79084244e-02
1.58567524e+00 -5.64487040e-01 -4.01093453e-01 4.59424779e-02
1.10585940e+00 2.59635210e-01 -1.91477048e+00 1.28963485e-01
-2.42578328e-01 -9.66807425e-01 6.42190725e-02 -5.22591949e-01
-1.50475311e+00 7.66128600e-01 5.00103116e-01 1.12466395e-01
1.44459522e+00 -4.47351158e-01 7.46545553e-01 -1.36614770e-01
-3.45310085e-02 -8.58484268e-01 1.10861525e-01 3.25667679e-01
8.32229853e-01 -8.67904544e-01 -4.12809514e-02 -5.70044518e-01
-2.71570057e-01 1.17889535e+00 4.70883697e-01 -2.32828543e-01
1.29721135e-01 3.59094232e-01 2.10117504e-01 -6.17240332e-02
-7.22968876e-01 -1.17275737e-01 1.10257462e-01 6.38736844e-01
2.20081508e-01 -6.05145395e-01 5.53829335e-02 -1.90342918e-01
3.87857199e-01 2.73029450e-02 1.03483093e+00 6.26795471e-01
-5.78927696e-01 -7.95871794e-01 -7.15351760e-01 3.46541889e-02
-3.22200119e-01 -1.45827487e-01 -3.31937015e-01 1.18957222e-01
1.27254009e-01 1.37756264e+00 7.92128891e-02 2.99954880e-03
4.26017463e-01 -1.98958382e-01 2.83511370e-01 -4.24066395e-01
-5.61603844e-01 4.70748425e-01 -1.47545695e-01 -1.01546288e+00
-6.29431129e-01 -5.69892228e-01 -1.02099371e+00 -1.27254426e-01
-1.22440783e-02 -3.47650535e-02 4.71603572e-01 6.48539901e-01
3.81022155e-01 8.60788643e-01 7.28688300e-01 -1.29357731e+00
3.47282678e-01 -8.44868422e-01 -5.17165422e-01 7.21213341e-01
9.49270844e-01 -3.65801603e-01 -6.96164548e-01 5.23681641e-01] | [11.238571166992188, -2.0509748458862305] |
25efc804-73c6-4639-8573-3fbe504cc4fb | imsimcse-improving-contrastive-learning-for | 2305.13192 | null | https://arxiv.org/abs/2305.13192v1 | https://arxiv.org/pdf/2305.13192v1.pdf | ImSimCSE: Improving Contrastive Learning for Sentence Embeddings from Two Perspectives | This paper aims to improve contrastive learning for sentence embeddings from two perspectives: handling dropout noise and addressing feature corruption. Specifically, for the first perspective, we identify that the dropout noise from negative pairs affects the model's performance. Therefore, we propose a simple yet effective method to deal with such type of noise. Secondly, we pinpoint the rank bottleneck of current solutions to feature corruption and propose a dimension-wise contrastive learning objective to address this issue. Both proposed methods are generic and can be applied to any contrastive learning based models for sentence embeddings. Experimental results on standard benchmarks demonstrate that combining both proposed methods leads to a gain of 1.8 points compared to the strong baseline SimCSE configured with BERT base. Furthermore, applying the proposed method to DiffCSE, another strong contrastive learning based baseline, results in a gain of 1.4 points. | ['Lemao Liu', 'Lihui Chen', 'Wei Shao', 'Jiahao Xu'] | 2023-05-22 | null | null | null | null | ['sentence-embeddings', 'sentence-embeddings'] | ['methodology', 'natural-language-processing'] | [ 1.33464992e-01 -1.14857368e-01 -3.93551588e-02 -2.68197656e-01
-1.09856498e+00 -4.11550820e-01 6.06553435e-01 4.03044671e-01
-8.24642181e-01 6.79458559e-01 2.70442575e-01 -3.01431537e-01
6.12068810e-02 -4.42041963e-01 -7.05128133e-01 -6.73228085e-01
-1.68406144e-02 -1.45420432e-02 1.83320031e-01 -4.69192982e-01
2.90758342e-01 2.58178234e-01 -1.27894187e+00 2.46878922e-01
8.59339178e-01 8.31808269e-01 3.49827185e-02 5.97987235e-01
-1.85900941e-01 6.09813690e-01 -6.35030746e-01 -5.31626761e-01
2.51614779e-01 -2.04857960e-01 -8.66285861e-01 -2.37329662e-01
8.67596686e-01 -2.63280511e-01 -2.38709748e-01 9.89056766e-01
7.95244396e-01 1.92348301e-01 4.64703947e-01 -1.11102223e+00
-6.03142679e-01 7.48034835e-01 -5.39181948e-01 4.64707017e-01
7.96926916e-02 2.68087059e-01 1.34625053e+00 -1.41531789e+00
3.04326177e-01 1.34860086e+00 7.19541788e-01 7.59699762e-01
-1.21087801e+00 -4.61952895e-01 4.86346245e-01 1.88981622e-01
-1.00002086e+00 -6.72000170e-01 8.26236308e-01 -1.85390338e-01
1.23559713e+00 2.11456448e-01 1.84017599e-01 1.33397710e+00
-3.91426645e-02 1.01885176e+00 1.11391056e+00 -4.48744774e-01
1.39402881e-01 4.06035125e-01 7.46812820e-01 4.65430081e-01
2.92784423e-01 -8.83582383e-02 -5.62488139e-01 -6.47013113e-02
1.42008126e-01 -2.20066801e-01 -2.00589791e-01 -8.02918524e-02
-7.80039191e-01 7.12635517e-01 3.93165588e-01 3.51734608e-01
1.63334347e-02 2.69462556e-01 9.08282101e-01 5.03136218e-01
6.49458945e-01 5.51581919e-01 -5.22594392e-01 -3.71354312e-01
-7.78710783e-01 2.56404757e-01 5.71499586e-01 7.33962834e-01
4.56790745e-01 2.49969706e-01 -5.36612928e-01 8.11430871e-01
2.04849660e-01 2.32708514e-01 4.01491433e-01 -4.70486730e-01
8.74462187e-01 3.72460961e-01 1.20253444e-01 -8.02515030e-01
-3.99872869e-01 -7.10913956e-01 -8.05880368e-01 3.88839021e-02
2.63467193e-01 -1.94246486e-01 -7.00028896e-01 1.85574913e+00
-3.30794901e-02 3.60504925e-01 1.50962397e-01 7.48380482e-01
9.15247977e-01 5.45238197e-01 1.93132032e-02 -6.47145435e-02
1.29939425e+00 -1.06889236e+00 -8.71385753e-01 -2.50704110e-01
8.22421312e-01 -7.09273100e-01 1.58259761e+00 3.92440975e-01
-1.30628276e+00 -5.51629245e-01 -1.33575892e+00 -1.31974980e-01
-2.95753390e-01 1.13420054e-01 2.98639417e-01 8.39967906e-01
-1.13114762e+00 7.57445753e-01 -7.67414510e-01 -1.98997304e-01
3.91600400e-01 6.05752051e-01 -1.15169264e-01 3.30917574e-02
-1.23113263e+00 9.33372080e-01 3.30201328e-01 2.54112750e-01
-5.13207734e-01 -8.94965708e-01 -7.65357435e-01 3.17587286e-01
1.22655861e-01 -6.37483537e-01 1.02709317e+00 -6.62631094e-01
-1.36535609e+00 5.86242735e-01 -3.99764717e-01 -6.69393122e-01
6.21511519e-01 -7.87602484e-01 -2.01545045e-01 -8.55143741e-02
-1.71513543e-01 4.39381033e-01 6.68764532e-01 -1.27355969e+00
-4.61350173e-01 -8.71587470e-02 1.30406186e-01 1.67306006e-01
-1.12852013e+00 4.17058133e-02 -2.55667120e-01 -6.08918607e-01
-3.41734529e-01 -5.47084749e-01 -2.02513784e-01 -2.39989802e-01
-3.45738053e-01 -4.46818620e-01 8.35987091e-01 -3.80697697e-01
1.40943646e+00 -2.09256840e+00 9.56783816e-02 -5.42548634e-02
2.10437730e-01 7.26210773e-01 -5.29049575e-01 4.83315438e-01
1.61865354e-02 3.11445802e-01 -2.89409995e-01 -9.71513569e-01
8.16713721e-02 5.29926270e-02 -2.51893520e-01 3.71041000e-01
8.84350359e-01 8.27191591e-01 -1.02332807e+00 -2.00188339e-01
2.53385544e-01 5.70430696e-01 -8.35632086e-01 2.37780005e-01
1.62066877e-01 5.60863577e-02 3.14747752e-03 1.69647217e-01
1.08414364e+00 1.72237754e-01 2.15082392e-01 -1.69560134e-01
-9.27680805e-02 6.58429384e-01 -1.19330072e+00 1.56604075e+00
-5.99115312e-01 4.53089535e-01 -1.00156017e-01 -1.05242527e+00
9.86592054e-01 1.59446225e-01 8.36981833e-02 -6.84516191e-01
9.33359563e-02 3.31867218e-01 5.31083392e-03 -3.97800952e-01
8.78412843e-01 -2.07331091e-01 -8.16383138e-02 1.30386025e-01
4.03088659e-01 2.38962457e-01 1.36940524e-01 1.38370559e-01
1.22109354e+00 -1.16437241e-01 -3.10521666e-02 -4.36243147e-01
7.24084318e-01 -5.82880020e-01 6.20958149e-01 9.17589009e-01
-4.71282363e-01 6.80851340e-01 6.75148427e-01 -1.87908188e-01
-9.31614816e-01 -1.19607306e+00 -1.84782092e-02 9.53137457e-01
7.97822420e-03 -5.59728324e-01 -6.45013094e-01 -8.87083054e-01
-1.89807594e-01 6.48456931e-01 -5.18351734e-01 -4.06960398e-01
-8.70934069e-01 -1.02111697e+00 5.25376916e-01 7.02365100e-01
4.77215379e-01 -7.76473820e-01 -2.13871092e-01 2.06989512e-01
-2.73388252e-02 -1.03540254e+00 -5.03388464e-01 3.74027073e-01
-9.54812467e-01 -7.66396940e-01 -3.89738530e-01 -9.85508025e-01
3.74142289e-01 4.04377759e-01 1.18612838e+00 2.68321574e-01
7.93341547e-03 1.42368749e-01 -4.60957497e-01 -7.01133981e-02
-2.25954711e-01 4.53218848e-01 2.97389299e-01 4.43039201e-02
4.54540133e-01 -5.21557748e-01 -6.73320532e-01 -2.21529141e-01
-9.42748487e-01 -5.04464567e-01 4.79584724e-01 1.19140220e+00
9.86948982e-02 -2.67288476e-01 1.01899302e+00 -9.84488428e-01
9.63513196e-01 -5.82740366e-01 -3.21377724e-01 8.65715444e-02
-7.00012982e-01 2.70573080e-01 9.52026904e-01 -2.95391053e-01
-9.97913539e-01 -1.96642101e-01 -4.44390744e-01 -1.44565761e-01
1.52650392e-02 2.65384197e-01 -2.56461412e-01 1.60182416e-01
5.48317254e-01 8.57284740e-02 -1.85401574e-01 -6.24005377e-01
5.78678191e-01 6.63106024e-01 3.16945344e-01 -5.94653189e-01
9.12193239e-01 2.18871772e-01 -2.04126790e-01 -8.82123053e-01
-6.41927004e-01 -5.94293892e-01 -5.63518286e-01 2.37304270e-01
5.73117852e-01 -9.41627145e-01 -6.77912354e-01 4.79300857e-01
-1.31894982e+00 -9.29264799e-02 -2.20865741e-01 3.36818397e-01
-3.18034768e-01 4.63165104e-01 -8.31019640e-01 -8.51503551e-01
-5.81747115e-01 -1.19334579e+00 7.78349698e-01 1.04985215e-01
-8.40904936e-02 -1.21419978e+00 1.27139986e-01 2.17555702e-01
6.51331842e-01 2.22621858e-02 8.95648360e-01 -7.82559156e-01
-2.73961872e-01 5.22830151e-02 -2.48600468e-01 8.78925741e-01
1.20076329e-01 -1.09416120e-01 -1.28123271e+00 -6.23628199e-01
9.27608684e-02 -4.22479898e-01 1.18318510e+00 1.00415774e-01
8.56526613e-01 -3.13606411e-01 7.95569643e-02 4.16961312e-01
1.64348650e+00 -2.08783090e-01 5.80169618e-01 3.54742408e-01
8.04290831e-01 3.66339952e-01 4.70815569e-01 3.86019975e-01
1.99214518e-01 7.72671878e-01 4.44296747e-01 -2.94404775e-01
-4.07035083e-01 -2.53632635e-01 6.26531065e-01 1.41183043e+00
3.12603205e-01 -1.60245016e-01 -5.05697489e-01 7.74836957e-01
-1.83309674e+00 -8.71607840e-01 -3.97611260e-01 2.29760456e+00
8.34459841e-01 4.75714028e-01 9.85497311e-02 3.31980437e-01
4.80347037e-01 2.73031294e-01 -2.56576836e-01 -9.11736488e-01
-3.11973244e-01 5.73832452e-01 3.65135372e-01 6.14229143e-01
-1.21455681e+00 1.02835011e+00 6.21524668e+00 9.17753577e-01
-1.07383144e+00 2.73815900e-01 5.05034089e-01 -4.31403935e-01
-4.50978458e-01 -1.53544933e-01 -8.90870512e-01 4.53504920e-01
1.09678841e+00 1.57997827e-03 1.18261687e-01 4.45382118e-01
4.20316905e-01 2.79923648e-01 -1.19434798e+00 9.17587459e-01
8.05230066e-02 -1.23761702e+00 2.59439964e-02 -2.80843735e-01
7.93862939e-01 -7.84068853e-02 4.87078547e-01 5.75040042e-01
3.28945406e-02 -1.00901794e+00 6.07030928e-01 2.01083615e-01
4.38130736e-01 -1.10204351e+00 1.09361947e+00 8.87285993e-02
-9.82365668e-01 -8.73441696e-02 -4.00797278e-01 -3.20123732e-01
8.12185928e-02 6.07332647e-01 -7.92125702e-01 5.44698536e-01
5.58129251e-01 6.86442137e-01 -7.73632586e-01 1.08001482e+00
-1.91041380e-01 9.43206072e-01 2.21980233e-02 -1.29378185e-01
4.18297678e-01 -3.06408424e-02 5.17549753e-01 1.62005544e+00
7.32949078e-02 -5.51768482e-01 -9.88049805e-02 7.60261178e-01
-3.68171692e-01 2.23680109e-01 -6.09020531e-01 2.96536416e-01
5.91534555e-01 1.20435536e+00 -2.76591510e-01 -1.46296546e-01
-6.47816479e-01 1.13847208e+00 8.05443347e-01 2.88106352e-01
-7.29450166e-01 -6.89549088e-01 8.99209380e-01 -1.14801645e-01
4.58126158e-01 -1.24208637e-01 -4.47507769e-01 -1.14147949e+00
4.22911048e-01 -9.17295277e-01 7.94434026e-02 -9.00971368e-02
-1.53291905e+00 6.49166644e-01 -1.77975029e-01 -1.03674960e+00
9.19646323e-02 -6.43831372e-01 -7.84785032e-01 9.28020716e-01
-1.88203478e+00 -9.77767944e-01 1.82280187e-02 2.09668383e-01
6.00254953e-01 2.88850479e-02 8.26616704e-01 5.88098466e-01
-7.03657210e-01 1.27786136e+00 3.71920377e-01 1.01241425e-01
7.69228876e-01 -1.55052006e+00 7.55849779e-01 1.11742318e+00
6.59060478e-02 8.27391446e-01 6.75287306e-01 -2.01063618e-01
-1.53072429e+00 -1.20627713e+00 1.04853094e+00 -5.71262777e-01
6.88519180e-01 -8.15603614e-01 -1.09788179e+00 3.78996104e-01
5.45312643e-01 -2.29833424e-02 5.61045587e-01 3.73346031e-01
-6.39317274e-01 -3.28464478e-01 -1.07598865e+00 6.64005637e-01
1.06350601e+00 -5.53180456e-01 -7.10033536e-01 2.27579072e-01
9.91992354e-01 -1.99681416e-01 -7.57918179e-01 2.37660497e-01
9.42237154e-02 -9.90709066e-01 9.44851875e-01 -8.88848901e-01
4.59259897e-01 -7.41999149e-02 -1.70804113e-01 -1.54257619e+00
-2.45930925e-01 -7.62403548e-01 -4.54318635e-02 1.65918148e+00
4.46170330e-01 -7.35124946e-01 5.87360859e-01 2.52844542e-01
-3.51095408e-01 -8.52315843e-01 -1.15926600e+00 -1.09581470e+00
7.22641230e-01 -3.40528458e-01 3.12759489e-01 9.07782316e-01
-6.81252219e-03 6.70171797e-01 -6.14545524e-01 -4.49366122e-02
5.07405639e-01 -5.13154030e-01 5.92679679e-01 -9.26742554e-01
-2.72230595e-01 -4.96250659e-01 -3.53611112e-01 -1.18697119e+00
2.74169356e-01 -1.02552605e+00 -5.11577986e-02 -1.41190004e+00
3.25972944e-01 -4.11850452e-01 -6.89382553e-01 1.58562243e-01
-6.80082321e-01 2.60044128e-01 4.31336492e-01 -2.71694422e-01
-5.80954373e-01 7.46044517e-01 6.88026786e-01 -1.23461246e-01
-1.49824470e-01 -2.01332659e-01 -8.48011196e-01 3.49715918e-01
9.25627947e-01 -4.86177564e-01 -3.46537083e-01 -6.89899147e-01
2.00771526e-01 -6.24963224e-01 3.62620473e-01 -9.37465608e-01
6.46464750e-02 3.90210152e-01 -9.62756202e-02 -4.12082493e-01
3.90953153e-01 -5.44800222e-01 -6.81981862e-01 5.37996709e-01
-4.77759361e-01 1.71182975e-01 4.48099315e-01 5.18307328e-01
-3.15648109e-01 -5.24130344e-01 6.81621075e-01 1.46173954e-01
-4.65638429e-01 -7.92109892e-02 -2.56185085e-01 5.45942843e-01
4.71747309e-01 1.53192123e-02 -5.48820794e-01 -1.57797754e-01
-3.58656883e-01 2.90407270e-01 1.47395760e-01 5.79432607e-01
6.99443460e-01 -1.50532460e+00 -9.72082138e-01 7.90706649e-02
1.31503358e-01 -3.47554564e-01 2.14446872e-01 9.25492346e-01
-1.25477687e-01 3.55000526e-01 8.21481794e-02 -4.41143483e-01
-1.33351719e+00 4.50777054e-01 2.02229157e-01 -4.42094594e-01
-4.84816551e-01 9.70705390e-01 -8.64425823e-02 -4.47312862e-01
4.13114280e-01 -5.96518099e-01 -2.13895455e-01 1.35215387e-01
4.94825542e-01 4.92895007e-01 4.55405056e-01 -4.51306641e-01
-5.25068521e-01 4.72765714e-01 -3.98783594e-01 7.57814792e-04
1.48852539e+00 -2.71686073e-02 6.03169091e-02 5.43572009e-01
1.76254642e+00 -2.32800525e-02 -1.14190722e+00 -3.64211768e-01
2.70326138e-01 -3.14433128e-01 2.15760358e-02 -4.86482114e-01
-9.55197096e-01 1.20389354e+00 7.55628407e-01 1.05504446e-01
1.10137272e+00 -3.68377298e-01 8.49747241e-01 3.15118432e-01
4.14598957e-02 -1.09348130e+00 2.27397859e-01 6.98126733e-01
6.18056834e-01 -1.46129906e+00 -2.15256795e-01 -2.17261970e-01
-5.27670145e-01 9.94968593e-01 6.56356990e-01 -4.06106174e-01
4.67226684e-01 2.79872268e-01 -2.65357215e-02 3.19343843e-02
-1.04132915e+00 -1.70383826e-01 1.30676642e-01 4.82140720e-01
6.37051821e-01 -9.97859761e-02 -6.53031409e-01 6.42597020e-01
1.03543237e-01 -2.52357125e-01 5.97961605e-01 7.36878455e-01
-3.77177924e-01 -1.35937202e+00 -1.14924356e-01 2.72914916e-01
-6.08837545e-01 -4.77726072e-01 -1.66944355e-01 7.37143517e-01
-6.73721582e-02 1.12413049e+00 1.19604975e-01 -4.29852933e-01
7.27220714e-01 1.07318990e-01 4.18429852e-01 -5.12545526e-01
-1.12488186e+00 -1.20409943e-01 1.93392232e-01 -3.53441417e-01
-2.70644188e-01 -4.57585901e-01 -1.05035770e+00 -5.87435327e-02
-3.10521394e-01 1.27178922e-01 4.24202442e-01 7.68481195e-01
3.91441792e-01 5.97055495e-01 8.19096327e-01 -5.16027391e-01
-1.13573408e+00 -1.02970707e+00 -2.81492114e-01 6.97119176e-01
5.98175824e-01 -4.90537554e-01 -7.28969395e-01 -4.47826475e-01] | [10.891511917114258, 8.5986328125] |
5a3b1a3c-a069-46d1-a321-fbd8f985217e | ji-yu-bertde-duan-dao-duan-zhong-wen-pian | null | null | https://aclanthology.org/2020.ccl-1.36 | https://aclanthology.org/2020.ccl-1.36.pdf | 基于BERT的端到端中文篇章事件抽取(A BERT-based End-to-End Model for Chinese Document-level Event Extraction) | 篇章级事件抽取研究从整篇文档中检测事件,识别出事件包含的元素并赋予每个元素特定的角色。本文针对限定领域的中文文档提出了基于BERT的端到端模型,在模型的元素和角色识别中依次引入前序层输出的事件类型以及实体嵌入表示,增强文本的事件、元素和角色关联表示,提高篇章中各事件所属元素的识别精度。在此基础上利用标题信息和事件五元组的嵌入式表示,实现主从事件的划分及元素融合。实验证明本文的方法与现有工作相比具有明显的提升。 | ['Bo Xu', 'Shuyi Wang', 'Hui Song', 'Hongkuan Zhang'] | null | null | null | null | ccl-2020-10 | ['document-level-event-extraction'] | ['natural-language-processing'] | [-4.76836711e-01 -8.37558568e-01 4.01335031e-01 2.01676682e-01
7.72331715e-01 -1.22301042e+00 3.33543211e-01 8.35921943e-01
5.52914515e-02 1.65449154e+00 2.12119535e-01 -4.39120471e-01
-9.34749171e-02 -1.11419272e+00 1.97119102e-01 -1.30142069e+00
-4.87259865e-01 1.71132779e+00 5.89283705e-01 -4.92211610e-01
6.83460951e-01 9.55160618e-01 -1.38240266e+00 4.69503999e-01
1.03027713e+00 1.14785862e+00 8.50579143e-01 5.29335439e-01
2.47819617e-01 3.76092374e-01 -1.36114419e+00 -2.61630982e-01
-3.30318779e-01 2.15874761e-01 -4.01453495e-01 9.50025678e-01
8.72509778e-01 -8.61984789e-01 -1.33705723e+00 1.19021189e+00
8.62355173e-01 3.45552474e-01 9.52490509e-01 -7.78002322e-01
-7.07478166e-01 4.21501189e-01 3.87007564e-01 4.24427688e-01
-1.81306064e-01 4.50798236e-02 6.63241506e-01 -8.15531790e-01
2.66975284e-01 1.81332314e+00 1.41460109e+00 3.45603168e-01
-1.63871479e+00 -1.04241192e+00 4.15779054e-01 -3.24779190e-02
-1.76739299e+00 1.18341021e-01 4.10755165e-02 -1.08824268e-01
1.58046758e+00 1.15453291e+00 1.65343499e+00 1.94052720e+00
7.99780250e-01 1.06581616e+00 1.83532226e+00 5.64161003e-01
5.31178415e-01 1.81403303e+00 -4.24907714e-01 2.99778372e-01
1.35252595e+00 -2.02757344e-01 -1.09657511e-01 2.10990697e-01
3.54867965e-01 6.06913030e-01 4.65869643e-02 3.85716736e-01
-9.04310703e-01 9.57511485e-01 3.01562041e-01 1.19524026e+00
-5.50115630e-02 -8.97028208e-01 -8.24881434e-01 6.86577082e-01
2.16474548e-01 -3.45604360e-01 6.28342569e-01 1.00308180e+00
-1.89103596e-02 -3.73869389e-01 1.09079921e+00 1.62041759e+00
1.03701377e+00 5.76875329e-01 -2.89866894e-01 7.14390099e-01
1.66753912e+00 8.88565719e-01 1.36274445e+00 -4.30434972e-01
-5.34184694e-01 8.11424375e-01 -9.50895995e-02 -1.73515534e+00
-8.03437769e-01 -1.42566693e+00 -6.60862029e-01 -9.68142390e-01
-1.45297810e-01 -6.33872151e-01 -1.23210895e+00 1.05067647e+00
4.94769126e-01 -3.17754447e-01 7.59810150e-01 1.01920533e+00
7.78266132e-01 1.62336183e+00 -1.34823069e-01 -4.59795237e-01
1.39977896e+00 -8.13727081e-01 -6.72395587e-01 -3.66623372e-01
-9.18285489e-01 -1.73924065e+00 7.39634275e-01 -3.12451988e-01
-5.24124503e-01 -1.68410107e-01 -2.06875110e+00 3.57239425e-01
-7.47627318e-01 -4.81819332e-01 1.59750736e+00 1.37663853e+00
-1.55839753e+00 4.36949283e-01 -8.69032264e-01 -9.59894776e-01
4.76511747e-01 6.50732994e-01 6.88364357e-02 -1.35018870e-01
-1.09177470e+00 9.67672050e-01 -4.58559126e-01 4.20319647e-01
-1.87311745e+00 -5.58442295e-01 -2.57194728e-01 -2.35550225e-01
8.87731314e-01 -7.01763630e-01 1.32097077e+00 -7.19420493e-01
-9.91663337e-01 9.77226973e-01 -1.03174853e+00 7.43317232e-02
7.13302851e-01 1.10702658e+00 -5.46689093e-01 4.41057235e-02
-2.31867850e-01 1.66462854e-01 9.44298565e-01 -2.76626396e+00
-9.21575606e-01 -1.79011524e+00 1.93082824e-01 6.73751011e-02
3.67550015e-01 5.97216547e-01 2.82178521e-01 -7.71901235e-02
2.84941554e-01 -1.73587513e+00 -6.37455434e-02 -1.36772907e+00
-7.35653818e-01 2.81312257e-01 1.67551661e+00 1.86316922e-01
1.42817307e+00 -2.13527465e+00 -8.16282034e-01 1.09167862e+00
6.54353499e-01 1.06767392e+00 4.46970344e-01 2.06040096e+00
4.55203384e-01 1.02052486e+00 -1.06209743e+00 3.86702567e-01
-2.74773866e-01 4.14589524e-01 9.18592632e-01 3.67047936e-01
-3.69522750e-01 -5.12339324e-02 -2.88578421e-01 2.22416759e-01
6.85365915e-01 9.29748893e-01 5.85296214e-01 3.18732649e-01
8.72572541e-01 2.32001275e-01 -2.73848683e-01 1.34338057e+00
1.11228752e+00 -6.76951170e-01 8.15163493e-01 4.89109993e-01
-2.65485764e-01 -4.29978311e-01 -2.61469913e+00 4.90511626e-01
7.97867402e-02 6.01511121e-01 7.12021112e-01 -9.53428268e-01
1.25100362e+00 1.17684400e+00 6.26226783e-01 1.94376782e-01
4.28959399e-01 8.97552192e-01 3.25752050e-01 -5.50020456e-01
7.12791800e-01 -1.03652728e+00 9.90854800e-01 3.15700948e-01
-7.38736510e-01 -7.83434212e-02 5.61782241e-01 -4.55073446e-01
1.04334593e+00 -1.49867094e+00 2.56704926e-01 -9.35337126e-01
1.13620484e+00 -1.64900362e-01 1.19430192e-01 3.89406443e-01
-8.96599174e-01 2.37163037e-01 -8.79364550e-01 2.46585310e-01
-1.13831602e-01 -1.64864945e+00 -4.14300531e-01 1.11076653e+00
-1.69068456e-01 -1.21635365e+00 -5.34258008e-01 -1.64502308e-01
-7.81928897e-02 5.81921518e-01 3.25836629e-01 3.45360301e-02
-6.32859707e-01 -1.43520939e+00 -8.45784307e-01 -3.84621233e-01
1.05248857e+00 -1.57432830e+00 -1.70948580e-01 -1.70486607e-02
-2.74704278e-01 -6.33556843e-01 -5.10809183e-01 1.24821208e-01
-1.89597118e+00 -9.36344862e-01 1.36239439e-01 -1.57930100e+00
1.67942262e+00 4.25514579e-01 7.53388643e-01 7.76287377e-01
-6.89857006e-01 9.46655273e-01 3.16983044e-01 -1.23850632e+00
2.66433775e-01 -2.35081449e-01 3.46556008e-01 1.26271033e+00
9.65532303e-01 -6.97890878e-01 -1.16601110e+00 8.09710681e-01
-1.37982535e+00 -6.94395661e-01 2.56616265e-01 8.52159619e-01
8.10335457e-01 4.65222210e-01 -5.59090495e-01 -7.51229942e-01
1.19638598e+00 6.80327773e-01 -1.10817766e+00 5.26456296e-01
-1.41707289e+00 -7.29669631e-01 7.70792961e-01 -8.82703960e-01
-9.45928216e-01 -6.72860324e-01 -4.12055224e-01 -5.53181442e-03
1.71278313e-01 -1.08656757e-01 -1.44071028e-01 -6.25936314e-02
-2.77382225e-01 -4.64887284e-02 6.82861686e-01 -5.83800018e-01
-2.22362801e-01 1.61845446e+00 5.60502037e-02 -1.03853166e+00
5.75073540e-01 6.82701468e-01 9.88560915e-02 -7.22479463e-01
2.25083247e-01 4.54692123e-03 -1.19175291e+00 5.86795986e-01
8.58549714e-01 -1.99163985e+00 -2.93584585e-01 7.53516674e-01
-7.24265277e-02 8.45051825e-01 -7.40113914e-01 1.87992454e+00
-1.54126197e-01 -2.82950550e-01 -3.43177378e-01 -9.06659544e-01
-1.42756447e-01 -2.80831337e+00 6.41875029e-01 1.98779416e+00
-4.54151481e-02 -9.50997174e-01 -1.00795478e-01 6.65217578e-01
6.98352039e-01 -2.61793226e-01 1.20296896e+00 -1.42331910e+00
-1.03541958e+00 -5.22836387e-01 4.41239774e-01 1.24288905e+00
7.13314176e-01 7.14295745e-01 -1.69822037e-01 -4.61722732e-01
3.28299068e-02 6.59295976e-01 7.22113132e-01 6.06354058e-01
1.12572126e-02 -7.82688498e-01 -5.54652095e-01 5.23115516e-01
2.50106478e+00 9.00680780e-01 8.85104835e-01 1.88113570e+00
-4.24090952e-01 1.79706067e-01 9.59117338e-02 4.34645146e-01
-1.66778937e-01 1.32171249e+00 -8.65850151e-02 1.25350106e+00
-6.01892829e-01 3.19075473e-02 1.11248910e+00 1.30734849e+00
-2.91733384e-01 -2.31281132e-01 5.97109973e-01 4.57829237e-01
-2.49517822e+00 -1.92257488e+00 3.86345088e-02 2.73947525e+00
2.13222817e-01 -5.69371879e-01 -8.53446946e-02 5.31802893e-01
1.56473923e+00 -3.39710444e-01 -4.89277914e-02 -8.88637185e-01
3.77276093e-01 1.03874898e+00 5.16999483e-01 3.98620099e-01
-1.01534808e+00 4.94374275e-01 8.04057360e-01 1.27723598e+00
-1.83432782e+00 -1.02685772e-01 -2.36178022e-02 -8.83890569e-01
-7.41984665e-01 3.33541065e-01 -1.27044559e+00 1.28274882e+00
1.08700764e+00 -7.85584211e-01 7.92308375e-02 6.33965015e-01
6.47716820e-01 8.21752399e-02 -1.04783320e+00 1.46823490e+00
4.23981160e-01 -1.35220385e+00 7.10625872e-02 5.60667813e-01
1.09079516e+00 -1.62744358e-01 9.23156589e-02 5.25754750e-01
1.39958084e-01 -8.90779316e-01 5.16501725e-01 -7.34985948e-01
-2.00973880e-02 -1.35836208e+00 1.50970900e+00 -3.86538170e-02
-1.66118348e+00 -8.11532497e-01 -3.15972984e-01 -5.69977283e-01
5.13136208e-01 2.35487595e-01 -9.02920604e-01 1.53286770e-01
2.91867316e-01 2.89223373e-01 -1.32783043e+00 1.95984173e+00
-8.51271987e-01 -3.32185239e-01 -8.63465011e-01 -1.97281766e+00
8.53824794e-01 -8.45631123e-01 8.98214042e-01 1.23096585e+00
1.99073702e-01 7.77632415e-01 5.09956539e-01 2.78026491e-01
-2.92501181e-01 2.81227916e-01 -6.59393251e-01 4.56816167e-01
4.82919544e-01 2.12765884e+00 -1.86820853e+00 -4.93669152e-01
-2.38742739e-01 -9.10480380e-01 -2.66839594e-01 -3.23594511e-01
-2.13986173e-01 -3.46376210e-01 8.12356353e-01 1.63846701e-01
2.88290856e-03 -5.36533475e-01 5.70844829e-01 -5.93163848e-01
-1.61712483e-01 -1.14830303e+00 -2.99127728e-01 -4.71478462e-01
-1.53234625e+00 5.18700957e-01 -1.38336718e-01 -1.75049806e+00
1.05544519e+00 -1.42621684e+00 -3.70933235e-01 8.31497833e-02
-3.44972201e-02 -7.26645172e-01 1.84777483e-01 1.08546519e+00
1.00379622e+00 -1.88266143e-01 7.09879935e-01 1.11195993e+00
-7.75466383e-01 7.00487375e-01 1.62603521e+00 7.88934082e-02
3.68913829e-01 -1.56869793e+00 -8.64409387e-01 1.67317063e-01
7.53455341e-01 3.87913167e-01 -1.53010920e-01 -4.29180175e-01
-1.90144777e+00 -9.28727150e-01 7.81995654e-01 -4.45362210e-01
5.52238345e-01 -4.48398054e-01 -1.67308003e-01 1.63679421e+00
2.82465488e-01 2.33213473e-02 7.98714757e-01 -4.07343090e-01
7.90715039e-01 -1.00207686e+00 -2.08007860e+00 1.50193870e+00
1.07588875e+00 3.45731415e-02 -7.54551291e-01 6.80680752e-01
-1.66307867e-01 -7.40351081e-01 -1.43077874e+00 6.10598028e-01
1.73087919e+00 -7.84202456e-01 9.75062907e-01 -1.18567514e+00
-7.57346988e-01 -9.33742449e-02 -1.00122082e+00 -1.36402404e+00
2.16769011e-04 -6.77806139e-01 7.78824866e-01 1.33458817e+00
1.02568519e+00 -1.27727962e+00 4.56031382e-01 1.43101645e+00
-7.06230521e-01 -7.57065654e-01 -2.13146925e+00 -1.25481284e+00
-5.31145096e-01 2.21657082e-02 1.70594901e-01 -9.87533778e-02
-2.45894775e-01 7.40000069e-01 1.26275063e-01 2.67394871e-01
2.43723348e-01 -4.20573533e-01 2.37501755e-01 -1.23568630e+00
1.10697687e+00 -7.61049509e-01 -6.05737090e-01 -1.17984846e-01
-8.69578660e-01 -2.69893527e-01 -1.63800585e+00 -1.94938433e+00
1.22395612e-01 -3.01429302e-01 -2.48570710e-01 -8.37872386e-01
1.34529972e+00 5.00808358e-01 5.92063844e-01 1.05545843e+00
1.20412037e-01 -5.83376229e-01 1.65146637e+00 -1.00017142e+00
-6.26005232e-01 7.88824022e-01 -4.08069432e-01 9.72690105e-01
3.45272541e-01 -3.36458147e-01 -5.39386511e-01 -2.63501406e-01
-9.80901495e-02 -1.16922247e+00 -1.42292166e+00 -1.01967752e+00
5.61077356e-01 4.02047932e-02 3.69345605e-01 -1.91631354e-02
5.41118860e-01 -8.58236253e-01 6.74843371e-01 6.11963212e-01
7.30030954e-01 8.41867566e-01 -2.58319765e-01 6.96838601e-03
1.02527931e-01 -1.48483682e+00 7.88230360e-01 -6.10149264e-01
-6.39705896e-01 -4.30071503e-01 -1.17056942e+00 -5.16079724e-01
1.87321103e+00 -6.57353401e-01 -5.83631814e-01 8.37198377e-01
-1.24712658e+00 -7.92217731e-01 3.56953979e-01 1.31510377e+00
1.51386452e+00 -1.28701437e+00 -1.76412416e+00 5.29767394e-01
-9.21878994e-01 9.89742577e-02 1.99365720e-01 9.96522844e-01
-6.82993591e-01 6.46403670e-01 -4.31049287e-01 -7.18477964e-01
-2.78464270e+00 8.85692239e-01 -5.46281636e-01 -4.84568983e-01
-5.07109344e-01 8.00234616e-01 -1.12094951e+00 2.50397384e-01
4.40740228e-01 -3.84086907e-01 -9.80630398e-01 2.61548936e-01
9.25913215e-01 7.83098042e-01 1.38041049e-01 -1.56864536e+00
-7.76668668e-01 1.93462694e+00 -7.44410872e-01 3.02077588e-02
4.22056288e-01 -6.08609021e-01 -7.36885250e-01 -1.85747921e-01
9.96662259e-01 1.18440509e+00 3.27363759e-02 1.01650357e+00
-8.40431690e-01 -4.16971654e-01 -6.61221325e-01 -1.36179245e+00
-7.62681484e-01 3.40834320e-01 9.69545543e-01 -5.74364305e-01
8.18293452e-01 -9.53267097e-01 1.11395800e+00 5.43619275e-01
1.94992077e+00 -1.90510762e+00 -8.60220790e-01 5.50628066e-01
8.08158040e-01 -1.80541432e+00 5.91359794e-01 -7.77457178e-01
7.19087347e-02 1.66220140e+00 7.25833178e-01 -3.18081588e-01
1.56327415e+00 1.29273254e-02 -3.74368578e-01 -1.11164510e-01
-2.69762546e-01 -4.59793895e-01 -1.51549116e-01 1.20871294e+00
4.82084215e-01 5.11918031e-02 -1.20938003e+00 1.00491025e-01
-1.15464616e+00 2.27640659e-01 1.24386346e+00 1.26594746e+00
-9.70205903e-01 -6.54624879e-01 -1.49153960e+00 2.69042440e-02
-5.74439645e-01 6.55595720e-01 1.38626382e-01 1.02993631e+00
5.75986266e-01 2.49581623e+00 -7.75984526e-01 -1.09153318e+00
1.57304704e+00 -5.56787848e-02 7.96386600e-01 -5.03460109e-01
-1.22741711e+00 -1.83220968e-01 1.03075601e-01 6.92844450e-01
-7.03544796e-01 -3.06396812e-01 -8.18619907e-01 -1.43558168e+00
7.54627138e-02 1.76523471e+00 1.27789724e+00 5.81919611e-01
-7.98195824e-02 9.93509114e-01 1.10672581e+00 -6.41301155e-01
-1.19141614e+00 -8.84085894e-01 -1.05091226e+00 -4.62372482e-01
2.59955704e-01 -1.21530378e+00 -1.66326165e+00 -9.72822249e-01] | [-3.3161251544952393, 6.907712459564209] |
fa51ad89-02bd-48e8-83b4-0af321e680e3 | minimal-solvers-for-single-view-lens | 2011.08988 | null | https://arxiv.org/abs/2011.08988v1 | https://arxiv.org/pdf/2011.08988v1.pdf | Minimal Solvers for Single-View Lens-Distorted Camera Auto-Calibration | This paper proposes minimal solvers that use combinations of imaged translational symmetries and parallel scene lines to jointly estimate lens undistortion with either affine rectification or focal length and absolute orientation. We use constraints provided by orthogonal scene planes to recover the focal length. We show that solvers using feature combinations can recover more accurate calibrations than solvers using only one feature type on scenes that have a balance of lines and texture. We also show that the proposed solvers are complementary and can be used together in a RANSAC-based estimator to improve auto-calibration accuracy. State-of-the-art performance is demonstrated on a standard dataset of lens-distorted urban images. The code is available at https://github.com/ylochman/single-view-autocalib. | ['James Pritts', 'Rostyslav Hryniv', 'Oles Dobosevych', 'Yaroslava Lochman'] | 2020-11-17 | null | null | null | null | ['scene-parsing', 'scene-labeling', 'camera-auto-calibration'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [-5.99212162e-02 -1.49559751e-01 1.37297332e-01 -5.72163820e-01
-6.96074724e-01 -9.47664857e-01 5.30019701e-01 -7.64799893e-01
4.08000238e-02 5.05346239e-01 1.83778182e-01 1.20632676e-02
-2.48665944e-01 -3.64861608e-01 -1.01392627e+00 -5.88085830e-01
5.95574856e-01 6.88688278e-01 1.60476584e-02 8.89460091e-03
3.09054703e-01 6.22899354e-01 -1.27089500e+00 -5.51023111e-02
7.84646392e-01 5.87538302e-01 8.59601572e-02 7.82021344e-01
3.49819094e-01 5.04173398e-01 1.12640694e-01 -4.96522635e-01
1.06761539e+00 -1.43984929e-01 -4.76344049e-01 5.60929716e-01
1.41006112e+00 -7.03301251e-01 -4.49322522e-01 1.08101046e+00
3.39909434e-01 -2.77535826e-01 3.58077377e-01 -9.80644107e-01
-2.53889889e-01 -1.31677585e-02 -8.37761819e-01 -3.64439577e-01
7.00887978e-01 1.40105635e-01 6.81320488e-01 -1.14223158e+00
8.09387624e-01 1.07409978e+00 7.85101712e-01 -1.57772258e-01
-1.27757776e+00 -6.33786082e-01 -6.09515943e-02 6.74149320e-02
-1.45488727e+00 -6.99318707e-01 6.81005239e-01 -6.41633034e-01
8.49344969e-01 2.02325642e-01 8.26090157e-01 7.55140662e-01
3.54982793e-01 2.08447173e-01 1.31196225e+00 -3.65990371e-01
-2.20906705e-01 1.22826204e-01 7.27306977e-02 7.11409092e-01
5.35885513e-01 3.22026759e-01 -7.04808474e-01 -1.68941155e-01
1.22018719e+00 -1.12215608e-01 -6.29701138e-01 -9.53124046e-01
-1.48118484e+00 6.89612985e-01 2.30435863e-01 -1.51037276e-01
6.37977719e-02 1.10798411e-01 -3.60205352e-01 1.15661331e-01
3.23745191e-01 9.35281515e-01 -3.74187231e-01 3.15470584e-02
-9.62283552e-01 4.60838765e-01 7.56357431e-01 1.73706245e+00
1.08338213e+00 1.34579509e-01 4.72658187e-01 6.05453968e-01
2.29289532e-01 1.15861428e+00 4.82091978e-02 -1.78847229e+00
2.64665157e-01 2.86870271e-01 4.04860318e-01 -9.09303129e-01
-3.99511933e-01 -5.45845270e-01 -5.75765312e-01 1.29156768e-01
3.75553131e-01 -1.51175007e-01 -7.78793454e-01 1.15856504e+00
2.85474986e-01 4.81068611e-01 -9.33318660e-02 1.13515842e+00
4.79202956e-01 2.53072739e-01 -1.25317609e+00 -1.31648868e-01
1.18085790e+00 -9.60442662e-01 -6.57045782e-01 -3.86113405e-01
1.81836531e-01 -1.62921989e+00 4.91881490e-01 6.64867699e-01
-1.33308303e+00 -2.72141755e-01 -1.23499715e+00 -5.04698038e-01
1.47419572e-01 4.22163576e-01 5.61855316e-01 4.73351598e-01
-1.01398420e+00 2.94682950e-01 -8.32410097e-01 -3.00248653e-01
-3.49605307e-02 3.48052889e-01 -4.43659484e-01 -1.65127963e-01
-1.97031632e-01 8.69303167e-01 -1.04076132e-01 -2.10958011e-02
-4.39973891e-01 -1.02023542e+00 -9.00457144e-01 -4.13487107e-01
4.77873892e-01 -1.05403852e+00 1.33660042e+00 -6.60023391e-01
-1.66180730e+00 7.82944262e-01 -4.56984937e-01 -2.42339998e-01
8.44521642e-01 -6.86521351e-01 1.60048097e-01 1.15981765e-01
-3.68422270e-02 4.05240685e-01 8.15529346e-01 -1.49402070e+00
-3.24697047e-01 -3.03389341e-01 -2.01680958e-02 3.59003633e-01
1.83191761e-01 -2.65729159e-01 -7.13187754e-01 -2.55979478e-01
8.61514509e-01 -1.29590714e+00 -1.11369438e-01 1.47869974e-01
-6.50757730e-01 7.61578619e-01 7.20832705e-01 -6.69981480e-01
4.91386592e-01 -1.76198399e+00 4.67925400e-01 2.55748600e-01
7.47679174e-03 -1.86896726e-01 -4.52859960e-02 3.00388604e-01
-1.19501069e-01 -4.15617883e-01 -2.06154928e-01 -6.15341961e-01
-3.41100603e-01 1.29570663e-01 -4.71079767e-01 1.07935345e+00
-1.22990973e-01 4.77912813e-01 -5.41906893e-01 -7.04910457e-02
7.63968587e-01 7.68055022e-01 -7.07903624e-01 1.80962652e-01
-5.81944399e-02 7.84470975e-01 -2.13494040e-02 5.72684288e-01
1.26978850e+00 7.73053318e-02 1.80190936e-01 -7.08976209e-01
-5.30038595e-01 1.26175821e-01 -1.74077582e+00 1.98665071e+00
-3.80048156e-01 8.46068382e-01 2.14445695e-01 -2.42189869e-01
7.40753949e-01 5.21121323e-02 5.05896628e-01 -3.80719334e-01
2.35314623e-01 3.83127630e-01 -2.12528616e-01 -2.24458411e-01
5.11385202e-01 2.90314108e-01 4.74187255e-01 -1.04343683e-01
1.99837953e-01 -1.11518657e+00 1.50700867e-01 1.04390860e-01
7.56301582e-01 6.53016090e-01 4.07165349e-01 -5.05151689e-01
5.54233253e-01 1.07173204e-01 6.55009866e-01 3.83172423e-01
5.08578598e-01 1.09014201e+00 1.53625831e-01 -4.04074043e-01
-1.56549585e+00 -1.07189882e+00 -5.93899786e-01 -1.31129816e-01
2.54975438e-01 -5.32737315e-01 -5.31357884e-01 1.39816433e-01
1.15855590e-01 5.85331917e-01 -1.65496409e-01 4.52266753e-01
-7.40446866e-01 -4.04354513e-01 1.11683970e-02 2.33843699e-01
6.48301899e-01 1.62588865e-01 -6.65674210e-01 -3.42231303e-01
-1.30363613e-01 -1.50504482e+00 -4.20074224e-01 -1.95563346e-01
-7.92960644e-01 -1.48395371e+00 -5.42323172e-01 -2.96886533e-01
9.25498605e-01 8.42987657e-01 1.09499347e+00 -9.17927548e-02
-7.75640756e-02 8.34160507e-01 9.96294171e-02 -2.23134875e-01
1.10563390e-01 -2.24524483e-01 2.31586233e-01 -6.66829869e-02
-2.86985282e-02 -7.64464915e-01 -5.24083257e-01 7.67618895e-01
-3.65865529e-01 6.20853007e-01 1.36617184e-01 6.91804171e-01
6.81117654e-01 -4.41513062e-01 -7.62714207e-01 -8.01140606e-01
-3.25571835e-01 4.67517301e-02 -1.61567688e+00 -2.51074228e-03
-4.68416333e-01 8.96645486e-02 2.67702729e-01 -2.18333915e-01
-1.30932546e+00 5.65377235e-01 4.38822150e-01 -8.90983701e-01
3.01383645e-03 -6.42544869e-03 -7.80190602e-02 -6.96092904e-01
5.50649405e-01 -1.01015061e-01 -8.47550482e-02 -4.34934735e-01
2.99784094e-01 1.93519816e-01 7.00922728e-01 -5.77491999e-01
1.42356777e+00 1.01577449e+00 5.05799592e-01 -9.62069750e-01
-7.93678403e-01 -7.50237286e-01 -9.94028986e-01 -2.09865302e-01
7.13361502e-01 -1.36588991e+00 -5.47892869e-01 5.31038582e-01
-1.22648251e+00 -1.91202715e-01 2.22013276e-02 9.36856031e-01
-7.69856215e-01 4.73585695e-01 -2.99856156e-01 -4.74077433e-01
1.66028500e-01 -1.36476374e+00 1.28613520e+00 2.74937481e-01
6.99836388e-02 -6.94174588e-01 2.25299448e-01 6.86420381e-01
2.04976141e-01 1.58821307e-02 8.20059255e-02 3.84460747e-01
-1.37948418e+00 1.64568406e-02 -1.74176112e-01 2.23267525e-01
8.15283414e-03 4.45323706e-01 -1.14056337e+00 -4.52834010e-01
2.34662980e-01 1.37375250e-01 6.18088424e-01 5.95671237e-01
6.19453430e-01 -1.85230806e-01 -1.59694329e-01 1.59989250e+00
1.97689378e+00 -1.51087090e-01 7.78257370e-01 4.14050728e-01
9.78256881e-01 5.19621909e-01 5.64869761e-01 4.49455082e-01
4.86891061e-01 9.80580568e-01 4.97849673e-01 5.51382378e-02
-2.29412124e-01 4.86799367e-02 1.40804142e-01 8.47484827e-01
-4.60118294e-01 8.57214257e-02 -1.02866292e+00 3.93270075e-01
-1.84380412e+00 -7.08285511e-01 -8.18208337e-01 2.47808909e+00
4.84798282e-01 -3.19418937e-01 -4.19658035e-01 -3.12351197e-01
2.00616330e-01 -7.20269531e-02 -4.36946392e-01 3.29956450e-02
-4.21537966e-01 -4.77939099e-02 1.33643663e+00 1.20365715e+00
-9.13344681e-01 8.15443814e-01 6.04518938e+00 8.70467350e-02
-9.91115928e-01 4.08919416e-02 2.48695109e-02 -2.84639776e-01
-4.24061269e-01 5.75336099e-01 -1.00503290e+00 -3.07928585e-02
4.32930440e-01 3.60834375e-02 6.81073487e-01 6.44463241e-01
9.93651673e-02 -3.64567369e-01 -1.10833943e+00 1.19174159e+00
4.64490205e-01 -1.36554480e+00 -2.67566621e-01 2.42320582e-01
1.17593098e+00 2.26129755e-01 -7.73441866e-02 -5.45498431e-01
3.34904253e-01 -5.14034152e-01 6.54799879e-01 7.72057474e-01
6.47947371e-01 -3.09841931e-01 6.69451416e-01 1.23895042e-01
-1.03363621e+00 2.37140462e-01 -5.44172287e-01 -1.21760093e-01
2.65542775e-01 7.11293161e-01 -8.37321818e-01 1.06107116e+00
7.53237009e-01 1.02677929e+00 -6.32787704e-01 1.25988233e+00
-2.34135404e-01 1.34814367e-01 -5.89762807e-01 8.69264364e-01
-3.26699287e-01 -8.75492334e-01 7.56279469e-01 7.54989684e-01
7.80860066e-01 1.98110715e-01 -4.07754853e-02 8.86579692e-01
3.54922175e-01 -2.36610651e-01 -6.60904229e-01 6.28227353e-01
2.90532559e-01 1.30895221e+00 -3.24085146e-01 -6.53710514e-02
-6.93652213e-01 8.62478733e-01 -8.99209529e-02 5.05810022e-01
-7.14495718e-01 2.63525158e-01 8.66181195e-01 4.25172478e-01
1.43965974e-01 -7.04947293e-01 -3.61427069e-01 -1.73759556e+00
1.86332166e-01 -8.36108804e-01 -2.75790542e-01 -1.35725355e+00
-9.04658735e-01 2.12891638e-01 3.75679076e-01 -1.54452300e+00
-2.50063956e-01 -9.04854953e-01 -2.64500201e-01 8.98575604e-01
-1.55802679e+00 -1.44192386e+00 -8.35614562e-01 6.39831543e-01
4.45631206e-01 -4.66496348e-02 6.77979171e-01 2.59701572e-02
-2.16756657e-01 3.30536403e-02 1.33476764e-01 -1.62792489e-01
1.14265931e+00 -1.28895175e+00 1.21674128e-01 1.13892913e+00
1.03673160e-01 8.22376609e-01 9.24370408e-01 -5.53237200e-01
-1.96383047e+00 -6.61632001e-01 4.44036365e-01 -6.10166013e-01
4.35606569e-01 -4.88439381e-01 -4.24418062e-01 1.31705964e+00
6.00840807e-01 -1.67811736e-02 2.28807285e-01 -9.06701759e-02
-4.70650405e-01 -3.14543247e-01 -8.82257879e-01 3.59012693e-01
1.00269580e+00 -4.18964773e-01 -3.05369467e-01 5.97018898e-01
3.35087657e-01 -1.29964995e+00 -8.18908513e-01 4.17108953e-01
6.83101773e-01 -1.34817433e+00 1.23849177e+00 1.73312798e-01
2.06514642e-01 -7.01066256e-01 -4.44658935e-01 -1.13506746e+00
-2.64653206e-01 -8.26652050e-01 2.34782428e-01 1.03181827e+00
2.89113224e-02 -8.91079068e-01 5.05607188e-01 4.70602721e-01
-2.89766639e-01 -1.73275664e-01 -8.00708413e-01 -7.07534850e-01
-1.12033859e-01 -1.72436014e-01 4.47749615e-01 9.70805287e-01
-4.54503596e-01 1.83120891e-01 -5.45914829e-01 8.36075723e-01
1.04345107e+00 2.77080715e-01 1.43486953e+00 -1.12739027e+00
-3.73051018e-01 7.75714219e-02 -5.51550925e-01 -1.13609684e+00
-2.88992990e-02 -5.30190885e-01 -2.15220779e-01 -1.19932163e+00
2.26441875e-01 -2.08624825e-01 6.54799581e-01 1.22937262e-01
5.07944405e-01 2.25722402e-01 3.17430168e-01 3.57463419e-01
1.93265472e-02 2.40137875e-01 1.23581827e+00 -1.48107437e-03
1.41860656e-02 -1.42570406e-01 -2.86822587e-01 1.07508469e+00
5.57480991e-01 -1.10225119e-01 -2.79561043e-01 -8.80849302e-01
5.73175192e-01 9.83999297e-02 5.69804490e-01 -1.27273905e+00
4.30542707e-01 -2.21353367e-01 4.73784864e-01 -6.78312838e-01
9.06676948e-01 -1.12377930e+00 8.20492983e-01 2.33967211e-02
1.01117894e-01 8.23609605e-02 2.05772981e-01 4.69927415e-02
6.38743937e-02 -1.10739648e-01 7.03589022e-01 -8.32959041e-02
-2.01410323e-01 1.41729429e-01 1.72950327e-01 -2.25306258e-01
8.09305668e-01 -2.87995785e-01 -6.46037340e-01 -5.93309760e-01
-2.30325669e-01 1.35179982e-01 1.31475019e+00 2.30373532e-01
6.11770809e-01 -1.21313179e+00 -7.86577642e-01 5.10462105e-01
-2.16012932e-02 2.50075817e-01 -5.01373177e-03 1.05095184e+00
-1.17152810e+00 4.88163531e-01 -1.30302563e-01 -1.12571394e+00
-1.49285555e+00 1.11849703e-01 7.64335036e-01 2.76954353e-01
-5.19769669e-01 7.40648389e-01 1.96000755e-01 -6.87719703e-01
-1.63856074e-01 -4.90863860e-01 5.14656842e-01 -3.40089142e-01
3.33983600e-01 4.19964552e-01 6.63205609e-02 -9.32025850e-01
-3.39793712e-01 1.52161050e+00 1.86401054e-01 -2.30793819e-01
1.42470777e+00 -2.85928458e-01 -2.64733672e-01 2.35353589e-01
9.60807025e-01 6.49733961e-01 -1.51187229e+00 -2.39636406e-01
-6.04676247e-01 -1.10005820e+00 2.96260357e-01 -4.81442809e-01
-1.31358576e+00 6.45511329e-01 5.62851429e-01 -4.94988829e-01
9.61705267e-01 -1.57107934e-01 2.63866097e-01 4.34021294e-01
5.67066431e-01 -6.95952415e-01 -3.46647024e-01 4.34722900e-01
1.19434273e+00 -1.25584197e+00 8.21788430e-01 -1.10220766e+00
-2.48000816e-01 1.42698801e+00 5.90896606e-01 -4.77108389e-01
7.05191493e-01 6.09037697e-01 1.05031990e-01 -1.05522819e-01
-3.02134812e-01 4.76398654e-02 5.30822635e-01 4.39726949e-01
4.53862071e-01 -4.28126790e-02 -7.59927090e-03 -1.91604674e-01
-6.52183056e-01 -2.47165248e-01 9.68228877e-01 7.53379345e-01
3.10646314e-02 -1.22215366e+00 -7.80548751e-01 2.27126986e-01
1.65680200e-01 -3.59380431e-02 -2.86836952e-01 7.42483974e-01
1.07612744e-01 7.56468654e-01 1.87984854e-01 -1.60241008e-01
4.05838102e-01 -3.95407557e-01 1.07982051e+00 -4.30024743e-01
-3.09241414e-01 6.19641006e-01 1.82916686e-01 -1.09448111e+00
-8.21921229e-01 -1.04194760e+00 -8.77316654e-01 -5.06732583e-01
-4.49181050e-01 -3.62966955e-01 6.90188885e-01 5.84281921e-01
3.28470051e-01 6.65072575e-02 4.04460281e-01 -1.16270745e+00
-2.10328773e-01 -7.76312232e-01 -4.17378396e-01 -5.86129688e-02
6.51501834e-01 -8.14471185e-01 -6.56948447e-01 2.92457908e-01] | [8.389333724975586, -2.4731714725494385] |
04ee89eb-ac27-44e2-860c-edcd8da24855 | overview-of-the-wanlp-2022-shared-task-on | 2211.10057 | null | https://arxiv.org/abs/2211.10057v1 | https://arxiv.org/pdf/2211.10057v1.pdf | Overview of the WANLP 2022 Shared Task on Propaganda Detection in Arabic | Propaganda is the expression of an opinion or an action by an individual or a group deliberately designed to influence the opinions or the actions of other individuals or groups with reference to predetermined ends, which is achieved by means of well-defined rhetorical and psychological devices. Propaganda techniques are commonly used in social media to manipulate or to mislead users. Thus, there has been a lot of recent research on automatic detection of propaganda techniques in text as well as in memes. However, so far the focus has been primarily on English. With the aim to bridge this language gap, we ran a shared task on detecting propaganda techniques in Arabic tweets as part of the WANLP 2022 workshop, which included two subtasks. Subtask~1 asks to identify the set of propaganda techniques used in a tweet, which is a multilabel classification problem, while Subtask~2 asks to detect the propaganda techniques used in a tweet together with the exact span(s) of text in which each propaganda technique appears. The task attracted 63 team registrations, and eventually 14 and 3 teams made submissions for subtask 1 and 2, respectively. Finally, 11 teams submitted system description papers. | ['Preslav Nakov', 'Giovanni Da San Martino', 'Wajdi Zaghouani', 'Hamdy Mubarak', 'Firoj Alam'] | 2022-11-18 | null | null | null | null | ['propaganda-detection'] | ['natural-language-processing'] | [ 2.11859077e-01 1.27010375e-01 -1.46696165e-01 -1.38181180e-01
-4.35963660e-01 -8.06547880e-01 1.22593033e+00 6.82200670e-01
-5.16810954e-01 6.61182404e-01 5.50535321e-01 -5.71066976e-01
2.55774975e-01 -5.80429673e-01 -2.54701048e-01 -6.59141719e-01
2.47998625e-01 2.38550752e-01 2.90839095e-03 -5.14142513e-01
9.92278934e-01 1.81790084e-01 -9.91784811e-01 4.28795755e-01
7.46271491e-01 2.88227141e-01 -1.78796053e-01 4.42131311e-01
-5.77429235e-01 1.54669344e+00 -1.18872297e+00 -2.96725750e-01
-3.19110364e-01 -8.25870812e-01 -1.14141870e+00 1.28439784e-01
4.64028418e-01 3.55459124e-01 6.92056259e-03 1.18249691e+00
4.41096425e-01 2.42881346e-02 8.02279711e-01 -8.88841271e-01
-6.25396967e-01 1.27070081e+00 -7.10000098e-01 4.79215086e-01
7.61350811e-01 -2.81816393e-01 8.14200222e-01 -7.51608312e-01
7.98973203e-01 1.49852836e+00 3.17668736e-01 4.14184541e-01
-1.28863800e+00 -7.40893781e-01 -1.22526744e-02 -7.34072849e-02
-1.04944074e+00 -5.91197252e-01 8.15775514e-01 -1.00367725e+00
5.55482030e-01 3.02254587e-01 4.63733971e-01 1.21665132e+00
2.77964771e-01 7.07499623e-01 1.50907898e+00 -8.04998696e-01
9.14823040e-02 5.33318341e-01 4.90880042e-01 7.67883062e-01
-1.25029266e-01 -4.47493196e-01 -7.09080458e-01 -6.01706862e-01
-5.74162379e-02 -6.97906971e-01 -3.41676831e-01 4.09698546e-01
-1.27570629e+00 1.31621671e+00 3.78538519e-01 7.96395481e-01
-1.62839144e-01 -1.43824846e-01 7.28185296e-01 5.14426470e-01
9.25082624e-01 9.46942091e-01 8.18049461e-02 -1.31202132e-01
-8.07088733e-01 4.99638677e-01 1.09415448e+00 2.96718836e-01
4.80711788e-01 -2.16016188e-01 -3.14067841e-01 8.53029549e-01
-9.92183294e-03 6.88512325e-01 3.38251561e-01 -3.16452235e-01
5.69818139e-01 7.88897038e-01 2.14332834e-01 -1.60410547e+00
-5.60741007e-01 -2.99870390e-02 -5.33963859e-01 -5.44090495e-02
5.80188990e-01 -3.78299326e-01 -4.84618872e-01 1.56030095e+00
3.62139732e-01 -6.57320201e-01 -5.33067822e-01 8.20792675e-01
1.01492333e+00 7.55614638e-01 1.36920989e-01 -5.33168852e-01
1.35352838e+00 -7.54027963e-01 -8.93584847e-01 -1.06124103e-01
1.12278497e+00 -1.40809870e+00 7.24111497e-01 2.20179588e-01
-8.01118433e-01 -1.06684148e-01 -7.27654219e-01 3.97217087e-02
-4.77854371e-01 -7.98166618e-02 2.77346313e-01 6.47490859e-01
-5.77158391e-01 2.21415728e-01 -1.62026361e-01 -4.46921796e-01
1.62475437e-01 -2.45953545e-01 -1.86802492e-01 3.64005536e-01
-1.35628521e+00 1.31414115e+00 1.04044184e-01 -1.89587399e-02
-3.44749957e-01 -4.21681017e-01 -7.52026916e-01 -5.44179678e-01
4.41293538e-01 -2.01434359e-01 1.16932380e+00 -1.32132709e+00
-1.25481844e+00 1.44416666e+00 -3.46392170e-02 -3.29335809e-01
4.81074423e-01 -1.05664484e-01 -6.84320569e-01 -1.98556170e-01
6.23238206e-01 1.82886526e-01 1.17241704e+00 -8.54931831e-01
-4.77115750e-01 -2.59358257e-01 1.20819986e-01 -3.19922864e-02
-3.13293338e-01 9.81861711e-01 4.85736072e-01 -8.19136322e-01
-8.86016116e-02 -1.03931582e+00 2.03863859e-01 -6.29473209e-01
-7.17379272e-01 -8.63255143e-01 8.18383992e-01 -6.85649693e-01
1.53791606e+00 -2.00768876e+00 3.25917423e-01 2.39076480e-01
6.38215780e-01 4.41068947e-01 4.25451323e-02 8.80985439e-01
2.16297600e-02 5.84422350e-01 1.72355965e-01 1.10046342e-01
5.72015531e-03 -3.27452928e-01 -6.45304799e-01 7.41145790e-01
-3.62567082e-02 6.08476162e-01 -1.13484371e+00 -4.97607142e-01
-1.15993157e-01 -4.67230678e-02 3.72527204e-02 4.29207197e-04
-2.09858924e-01 4.50819105e-01 -7.36563861e-01 2.73399353e-01
2.69135237e-01 -3.30762357e-01 1.00963093e-01 2.15485945e-01
-6.86862111e-01 8.62928450e-01 -4.88948107e-01 9.21094060e-01
-1.05753787e-01 9.81782556e-01 3.10520083e-01 -7.16762722e-01
7.46371686e-01 2.49510869e-01 2.24789932e-01 -5.85344195e-01
5.36695123e-01 2.46329889e-01 1.17159329e-01 -5.21319807e-01
5.28395653e-01 -1.79714918e-01 -5.18980324e-01 1.10704923e+00
-5.57456493e-01 -1.62322924e-01 5.67058682e-01 6.16122186e-01
1.30480862e+00 -4.49964285e-01 7.09901512e-01 -5.51662624e-01
7.00998485e-01 4.33707267e-01 2.24019140e-02 9.91351128e-01
-2.52493680e-01 -1.05525382e-01 5.69439888e-01 -6.39082134e-01
-5.49156308e-01 -5.31311750e-01 -1.41399041e-01 1.52687812e+00
2.83886809e-02 -5.82033575e-01 -4.87758398e-01 -9.71045494e-01
-2.62194842e-01 7.85265803e-01 -9.28073108e-01 1.88388139e-01
-7.23732591e-01 -5.33116758e-01 6.94218814e-01 -3.32556546e-01
6.17658019e-01 -1.42331278e+00 -7.21434414e-01 3.68258178e-01
-3.85140270e-01 -7.83233881e-01 -5.40966213e-01 8.67715925e-02
-1.48710817e-01 -1.27675176e+00 -3.90663087e-01 -9.33124006e-01
6.43259406e-01 3.75772536e-01 1.11599362e+00 3.88744086e-01
-3.35776084e-03 -2.10482225e-01 -8.79688442e-01 -6.74271226e-01
-9.51124489e-01 3.23698103e-01 5.19314967e-02 -8.45950842e-02
3.08087796e-01 -7.07681403e-02 1.25546589e-01 1.35032028e-01
-8.12264085e-01 3.26894283e-01 8.94865990e-02 6.89472675e-01
-3.73535782e-01 -2.02086344e-01 4.32637721e-01 -1.38132894e+00
1.27004993e+00 -3.84688824e-01 -2.94419646e-01 1.65296897e-01
-3.34939420e-01 -1.13152824e-01 4.30761367e-01 -5.11686921e-01
-7.09744573e-01 -3.12408805e-01 3.20420638e-02 5.82323968e-01
8.73304680e-02 9.07511175e-01 3.92643332e-01 -1.50981164e-02
1.23999500e+00 -1.85500272e-02 -1.06183909e-01 -2.69287616e-01
2.86987990e-01 9.14358914e-01 4.82628308e-02 -4.53044146e-01
9.43158269e-01 1.08722672e-01 -3.48431081e-01 -9.88050282e-01
-1.52690721e+00 -6.88947082e-01 -2.92138994e-01 -4.68399107e-01
7.67023206e-01 -5.32112598e-01 -3.45794410e-01 5.36369324e-01
-1.34146631e+00 -1.52050361e-01 3.33946407e-01 9.30491164e-02
1.83376878e-01 3.27355325e-01 -5.70098758e-01 -8.40355158e-01
-5.12787223e-01 -6.93531096e-01 5.38143873e-01 -3.01521011e-02
-9.20809209e-01 -1.02962029e+00 3.06981444e-01 4.92474020e-01
3.20769697e-01 3.04460108e-01 1.06235135e+00 -8.19856882e-01
2.32771605e-01 -3.29509288e-01 -1.77400708e-01 -1.25140082e-02
3.82540166e-01 7.71859735e-02 -6.11003935e-01 -6.58641085e-02
6.43172190e-02 -5.74900270e-01 6.55152500e-01 6.53567389e-02
4.81229365e-01 -7.74021983e-01 -2.41256744e-01 -4.69581902e-01
5.12585044e-01 -6.13858066e-02 2.78436124e-01 4.67291623e-01
4.60349858e-01 9.38820243e-01 5.20880580e-01 3.02999288e-01
4.11198251e-02 5.92433035e-01 1.37525469e-01 1.44523755e-01
1.61361471e-01 -9.40271467e-02 6.67773724e-01 9.12833571e-01
7.34548941e-02 -3.13758731e-01 -1.09913874e+00 4.36178267e-01
-1.72826135e+00 -1.22863209e+00 -7.69533098e-01 1.83070230e+00
1.01307452e+00 1.07292764e-01 2.66804248e-01 2.19635963e-01
8.93625081e-01 5.93257487e-01 2.06600592e-01 -7.55873203e-01
-1.76054746e-01 -1.42761683e-02 2.10071251e-01 6.34707868e-01
-1.26595712e+00 1.02601635e+00 5.98260880e+00 9.24245238e-01
-1.19256163e+00 1.97659627e-01 3.44488055e-01 1.41373947e-01
-1.67290926e-01 -3.38647366e-02 -6.48716688e-01 4.67703640e-01
6.83722675e-01 -3.68612885e-01 2.61979789e-01 4.45826739e-01
4.41388667e-01 -3.99603367e-01 -8.10224831e-01 5.56991816e-01
3.51843119e-01 -1.39069331e+00 -6.20523505e-02 -1.84853170e-02
7.60461509e-01 -1.48117259e-01 -1.04455285e-01 4.22256410e-01
3.03791851e-01 -9.54002500e-01 8.83509398e-01 7.63073489e-02
9.76680666e-02 -3.37150276e-01 5.95205963e-01 7.96807587e-01
-6.28466010e-01 2.27356240e-01 1.11880317e-01 -6.70774341e-01
3.62613112e-01 7.73754954e-01 -1.03293920e+00 -3.49518247e-02
2.87910461e-01 4.89636540e-01 -6.78230345e-01 6.49788260e-01
-8.94991875e-01 1.14894080e+00 2.36296505e-02 -8.60888362e-01
5.22246957e-01 1.45296156e-01 9.27935362e-01 1.49361563e+00
-3.32580447e-01 1.36847764e-01 4.08813983e-01 6.74250603e-01
-2.32897401e-01 5.41735768e-01 -5.53061843e-01 -6.33400977e-01
4.16651070e-01 1.38302982e+00 -7.26776123e-01 -3.67590159e-01
1.41834483e-01 6.69447064e-01 4.52567995e-01 1.31826535e-01
-7.13713288e-01 -3.17533344e-01 1.77659556e-01 4.19831514e-01
-3.53406996e-01 -2.18194395e-01 -9.24716517e-02 -9.07787144e-01
-2.27462694e-01 -9.09904838e-01 3.07609320e-01 -5.28060555e-01
-1.54336202e+00 6.00394189e-01 -2.61824965e-01 -5.99474013e-01
-3.00177634e-01 -4.08505350e-01 -5.76834261e-01 9.22505379e-01
-7.78618455e-01 -8.85721922e-01 -6.09368011e-02 2.23682076e-01
5.56805193e-01 -1.04737580e-01 8.87138963e-01 -1.63373668e-02
-4.27737266e-01 1.77788943e-01 -3.64295393e-01 4.22266722e-01
1.06126058e+00 -9.74321485e-01 -6.07112311e-02 6.81041300e-01
3.10335308e-01 8.36724520e-01 1.24384880e+00 -6.76572680e-01
-9.25085485e-01 -8.75604093e-01 1.82081091e+00 -8.21911335e-01
1.35021913e+00 -1.85120404e-01 -7.03962564e-01 3.45261931e-01
4.09699172e-01 -6.26992047e-01 6.99320138e-01 4.04854923e-01
-6.00832522e-01 6.33569241e-01 -8.55423868e-01 4.86737102e-01
8.72053921e-01 -6.06439948e-01 -8.65496218e-01 9.66878712e-01
3.19738328e-01 -4.09611046e-01 -1.71454355e-01 -1.29412293e-01
2.89764792e-01 -4.51961309e-01 3.19660485e-01 -8.22228789e-01
8.33071113e-01 -2.30035841e-01 2.10314319e-01 -1.34830511e+00
-4.57305968e-01 -8.03914964e-01 1.48069680e-01 1.24949920e+00
4.87477541e-01 -4.91841793e-01 2.29010507e-01 2.59395868e-01
6.17541149e-02 -1.41993642e-01 -7.98508823e-01 -3.21813732e-01
9.27557424e-02 -1.05986536e-01 -1.92335159e-01 1.35488069e+00
4.20834512e-01 1.10103810e+00 -4.48324412e-01 -2.86779135e-01
3.29703391e-01 2.71896422e-01 7.91603684e-01 -9.97919321e-01
2.10579872e-01 -7.76712477e-01 -1.24456108e-01 -8.61815155e-01
3.96002322e-01 -1.09493375e+00 2.11169660e-01 -1.35672617e+00
2.65561879e-01 -2.82960624e-01 2.25775793e-01 5.88114381e-01
-3.70111838e-02 2.59477496e-01 1.52047887e-01 4.49981391e-01
-6.49175465e-01 2.22532839e-01 9.97072995e-01 -4.35318530e-01
-4.83807951e-01 -4.91446517e-02 -9.34767306e-01 9.74538386e-01
7.12323666e-01 -7.45614469e-01 1.04348607e-01 -2.96522170e-01
7.64331877e-01 -3.05102557e-01 3.39096934e-01 -2.79013842e-01
1.76030681e-01 -4.03025389e-01 -1.83427364e-01 -3.95434260e-01
-1.93850443e-01 -1.14694461e-01 -2.11232990e-01 5.43870032e-01
-5.76711297e-01 -2.59797298e-03 -1.90904722e-01 3.23277026e-01
-2.16137901e-01 -3.48865449e-01 7.09215522e-01 -2.55026780e-02
-4.19489890e-01 -3.62678766e-01 -1.36504710e+00 2.88471878e-01
9.00852621e-01 2.10312456e-01 -7.75862455e-01 -4.39729273e-01
-4.45115894e-01 1.84449539e-01 2.06141889e-01 5.51158369e-01
2.19373852e-01 -1.07454097e+00 -1.19438326e+00 -3.98653626e-01
3.42409551e-01 -7.03728199e-01 -2.96431124e-01 1.12137067e+00
-5.55449665e-01 2.53531605e-01 1.17645234e-01 -1.88715965e-01
-1.63852918e+00 2.38140702e-01 2.08439063e-02 -3.04701567e-01
-4.62236375e-01 1.00291646e+00 -1.25048473e-01 -2.55008161e-01
-1.05862036e-01 4.04283613e-01 -5.60902059e-01 5.10422826e-01
8.77142131e-01 3.07007313e-01 -9.91843790e-02 -1.07748914e+00
-4.92371708e-01 -4.75004762e-02 -2.90898800e-01 -1.85289949e-01
9.29450631e-01 4.20801528e-02 -7.88665771e-01 7.24721193e-01
8.44744861e-01 5.68446159e-01 1.36252819e-02 -2.46121511e-01
2.25152478e-01 -3.73592854e-01 -2.69618183e-02 -1.13079941e+00
-2.79418141e-01 4.56644952e-01 -1.48799866e-01 8.80009890e-01
4.97861326e-01 3.02989274e-01 4.95402157e-01 4.13099378e-01
3.32277596e-01 -1.27477324e+00 2.83486724e-01 1.05819952e+00
1.35800195e+00 -9.99071360e-01 3.20581272e-02 -5.68359017e-01
-3.81302804e-01 9.71408129e-01 2.88525820e-01 1.18710160e-01
5.37998080e-01 -1.52080879e-01 2.81761348e-01 -6.38958395e-01
-5.15341640e-01 6.68013245e-02 4.71377879e-01 1.78344265e-01
1.00794971e+00 2.33549401e-01 -1.19028080e+00 2.57246763e-01
-3.73386025e-01 -4.76484686e-01 5.67395747e-01 9.99498308e-01
-7.76859581e-01 -9.25590396e-01 -5.70854664e-01 5.64251661e-01
-5.36900878e-01 -2.04266831e-01 -1.45194709e+00 5.63187122e-01
1.45681322e-01 1.57459116e+00 -3.48417073e-01 -4.62767869e-01
1.64482277e-02 -9.69947353e-02 4.00807500e-01 -1.00977850e+00
-1.33727968e+00 -2.65945464e-01 8.76107991e-01 -2.07715780e-01
-7.81497419e-01 -6.02083802e-01 -1.00200689e+00 -7.70422518e-01
-5.88374317e-01 5.68661809e-01 5.25182128e-01 1.26404560e+00
-1.92327932e-01 1.17417842e-01 6.17294788e-01 -4.30423558e-01
-6.70279384e-01 -1.47965932e+00 -4.84467804e-01 5.69415569e-01
-5.36473691e-02 -5.00364959e-01 -5.98072588e-01 -8.74531567e-02] | [8.54932689666748, 10.493205070495605] |
59199f7b-c658-43c7-af9f-5be5d18a7d4a | polar-relative-positional-encoding-for-video | null | null | https://www.ijcai.org/proceedings/2020/132 | https://www.ijcai.org/proceedings/2020/0132.pdf | Polar Relative Positional Encoding for Video-Language Segmentation | In this paper, we tackle a challenging task named video-language segmentation. Given a video and a sentence in natural language, the goal is to segment the object or actor described by the sentence in video frames. To accurately denote a target object, the given sentence usually refers to multiple attributes, such as nearby objects with spatial relations, etc. In this paper, we propose a novel Polar Relative Positional Encoding (PRPE) mechanism that represents spatial relations in a ``linguistic'' way, i.e., in terms of direction and range. Sentence feature can interact with positional embeddings in a more direct way to extract the implied relative positional relations. We also propose parameterized functions for these positional embeddings to adapt real-value directions and ranges. With PRPE, we design a Polar Attention Module (PAM) as the basic module for vision-language fusion. Our method outperforms previous best method by a large margin of 11.4% absolute improvement in terms of mAP on the challenging A2D Sentences dataset. Our method also achieves competitive performances on the J-HMDB Sentences dataset. | ['Qi Tian', 'Fei Wu', 'Lingxi Xie', 'Ke Ning'] | 2020-07-20 | null | null | null | null | ['referring-expression-segmentation'] | ['computer-vision'] | [ 2.29716420e-01 -5.95722173e-04 -1.33016884e-01 -6.88831151e-01
-7.68948913e-01 -4.94371951e-01 6.61629915e-01 1.48535460e-01
-6.42816782e-01 3.50673854e-01 3.66455853e-01 -9.07412544e-02
1.05346665e-01 -5.34004211e-01 -1.00261867e+00 -8.18831503e-01
2.14728937e-01 2.55342036e-01 5.11985123e-01 -1.55902728e-01
2.35686526e-01 4.47536856e-01 -1.26982999e+00 3.61927688e-01
5.01806200e-01 1.24886465e+00 6.96416199e-01 6.86283648e-01
-2.67922916e-02 8.47178698e-01 -5.98757982e-01 -2.14612693e-01
9.76661295e-02 -3.01553875e-01 -8.35719526e-01 3.41286153e-01
3.35384399e-01 -2.23319024e-01 -4.23318952e-01 1.10656703e+00
4.23524380e-01 3.01031649e-01 7.00754702e-01 -9.51007307e-01
-5.89356184e-01 5.95920026e-01 -9.23078120e-01 3.01714092e-01
4.25279140e-01 1.18363567e-01 9.85260129e-01 -8.56236577e-01
6.15419805e-01 1.37861252e+00 2.11158589e-01 4.29834038e-01
-1.09234548e+00 -2.29491994e-01 5.58254063e-01 4.77284700e-01
-1.35678208e+00 -3.79101604e-01 9.09771860e-01 -5.21222234e-01
7.92698801e-01 3.59610528e-01 5.77048302e-01 9.20558274e-01
1.14694007e-01 1.09034312e+00 6.57476306e-01 -2.98347712e-01
5.83155230e-02 8.29509497e-02 9.24586058e-02 3.81278813e-01
-1.89330563e-01 -4.75052565e-01 -4.13993388e-01 3.82680565e-01
6.30927265e-01 -1.39859512e-01 -3.87180865e-01 -3.32293600e-01
-1.44085526e+00 5.39242148e-01 6.27739191e-01 2.20798492e-01
-3.12616587e-01 2.92601228e-01 2.95130908e-01 -2.11033329e-01
3.47645044e-01 5.40087998e-01 -3.94878626e-01 -1.91715229e-02
-6.61482513e-01 3.02258104e-01 2.53192812e-01 9.27269816e-01
6.13880515e-01 -4.28774714e-01 -5.71597695e-01 8.22828233e-01
5.29752135e-01 4.38415945e-01 1.59423724e-01 -1.19160950e+00
7.00612664e-01 4.04622704e-01 2.02242866e-01 -1.29784036e+00
-3.67744565e-01 -1.67766780e-01 -8.36327553e-01 -1.83320835e-01
2.12024704e-01 1.53669519e-02 -8.87323558e-01 1.68170261e+00
4.77786481e-01 3.39355201e-01 8.42922106e-02 1.37616634e+00
1.01510143e+00 1.03965175e+00 8.28925222e-02 -2.53358781e-01
1.76157582e+00 -1.11639082e+00 -8.75710845e-01 -4.15634096e-01
4.58483636e-01 -6.69728577e-01 8.14551234e-01 1.54637992e-01
-1.03787374e+00 -6.64298117e-01 -8.77872527e-01 -4.53259468e-01
-2.56950676e-01 2.03702539e-01 3.45456034e-01 1.17890663e-01
-9.44229126e-01 1.54078811e-01 -6.99219108e-01 -2.87382185e-01
3.92519087e-01 3.94897223e-01 -3.56556386e-01 9.41961482e-02
-1.30121100e+00 7.08319664e-01 4.62064207e-01 2.99241096e-01
-3.88753623e-01 -3.96259755e-01 -1.08364010e+00 -4.17142622e-02
6.03520334e-01 -6.55288279e-01 9.96273279e-01 -9.05788898e-01
-1.21397638e+00 9.80739892e-01 -2.68319964e-01 -7.55590737e-01
2.13016748e-01 -2.85336137e-01 -9.36650112e-02 4.67233300e-01
1.14377767e-01 1.25561953e+00 7.13911772e-01 -1.21263659e+00
-7.00608194e-01 -4.59730268e-01 3.67128849e-01 6.14124238e-01
-1.01467602e-01 2.08119079e-01 -1.07567132e+00 -6.93233252e-01
1.85683772e-01 -8.77514541e-01 -3.15320879e-01 3.98915671e-02
-4.16154355e-01 -4.05050725e-01 9.23683941e-01 -6.88525140e-01
1.15480471e+00 -2.26942229e+00 5.66595137e-01 -2.70803571e-01
2.81122744e-01 2.95601249e-01 6.47632480e-02 -1.69241324e-01
1.72398776e-01 7.67996088e-02 -4.42868024e-01 -4.69190687e-01
-5.34471124e-02 2.17236042e-01 -3.34615022e-01 4.27968025e-01
5.30188203e-01 1.02408767e+00 -9.03762698e-01 -8.44080210e-01
4.63062942e-01 5.48767447e-01 -4.43394333e-01 2.13833362e-01
-4.86131221e-01 5.14113963e-01 -6.59746647e-01 4.36487377e-01
7.28493989e-01 -6.93018660e-02 -5.05084023e-02 -5.86577177e-01
-2.03694686e-01 2.85253618e-02 -9.74398315e-01 1.96540654e+00
-2.34232694e-01 9.70834672e-01 3.08565442e-02 -1.11611259e+00
8.79764736e-01 2.91068144e-02 4.83530790e-01 -6.01900041e-01
2.77550697e-01 -1.69087395e-01 -1.73438162e-01 -7.83193886e-01
6.67434931e-01 4.51515138e-01 -3.28831732e-01 6.58992082e-02
-1.73055172e-01 -1.34039298e-01 1.37831062e-01 1.81820408e-01
7.79110014e-01 3.35917205e-01 3.37489069e-01 -1.68817043e-01
8.14720392e-01 -1.35979056e-01 5.35957873e-01 4.77096438e-01
-3.98235500e-01 8.63550365e-01 7.86136746e-01 -4.51970577e-01
-7.90263534e-01 -8.15329015e-01 -1.10814184e-01 9.80170965e-01
6.53780162e-01 -3.18931967e-01 -6.77400649e-01 -6.31996036e-01
-2.82599449e-01 6.75428927e-01 -6.73858225e-01 -4.50254371e-03
-6.85980260e-01 -4.72113729e-01 3.06100488e-01 7.86293685e-01
5.49526513e-01 -9.54269528e-01 -7.37970591e-01 1.10724688e-01
-3.84400189e-01 -1.62358117e+00 -8.27552021e-01 1.70001358e-01
-3.90806556e-01 -8.07942808e-01 -7.19531715e-01 -1.03990042e+00
6.19657934e-01 3.89692187e-01 8.46070886e-01 -4.02409196e-01
9.56859440e-02 8.59148949e-02 -4.90944505e-01 -2.00860471e-01
5.19142374e-02 -1.81625888e-01 -8.62256512e-02 3.22393507e-01
4.74886358e-01 -1.15323901e-01 -7.85755754e-01 3.15446556e-01
-8.89816046e-01 3.76000673e-01 5.35086632e-01 7.22900450e-01
9.32530940e-01 -2.18039975e-01 5.84135912e-02 -5.44817269e-01
3.17858964e-01 -3.30576420e-01 -4.32698935e-01 3.32419097e-01
8.03512409e-02 2.12750360e-01 3.64308685e-01 -4.14285004e-01
-9.25741076e-01 3.73267293e-01 -1.49218172e-01 -5.30982912e-01
-3.38823259e-01 2.65256643e-01 -6.38397753e-01 3.05921853e-01
2.18945548e-01 2.80282438e-01 -3.04927558e-01 -2.86504030e-01
5.46745420e-01 8.36960971e-01 7.44989812e-01 -5.63991070e-01
3.39199543e-01 4.52428371e-01 6.72633126e-02 -8.00503671e-01
-9.50781465e-01 -5.18862605e-01 -9.40723360e-01 -1.65557131e-01
1.46338308e+00 -1.00727701e+00 -7.14310348e-01 3.18919957e-01
-1.54394424e+00 -8.56406391e-02 -2.48755515e-02 4.48406100e-01
-5.63574970e-01 3.18688363e-01 -3.12826276e-01 -6.90503299e-01
-1.22974239e-01 -1.62864351e+00 1.49448276e+00 3.23903471e-01
-4.84939963e-02 -7.09889054e-01 -2.78597534e-01 4.96224284e-01
-3.57587785e-02 2.38933340e-01 5.31965375e-01 -6.07367158e-01
-7.04908729e-01 1.91558272e-01 -4.79026556e-01 3.70483249e-01
4.28870842e-02 7.98483267e-02 -7.89513528e-01 1.73437059e-01
1.16284333e-01 1.18144095e-01 1.08364701e+00 6.75173581e-01
1.46653235e+00 -1.96699783e-01 -3.46632630e-01 8.17520678e-01
1.02699769e+00 4.73949105e-01 5.43276548e-01 3.39371979e-01
1.02898121e+00 7.33922780e-01 8.77108455e-01 2.69563496e-01
5.57059526e-01 1.04587746e+00 4.40352619e-01 9.52880159e-02
-6.72617108e-02 -4.16217707e-02 2.40498826e-01 4.73186672e-01
1.13342255e-01 -5.36742628e-01 -7.97797382e-01 5.18076420e-01
-2.10054922e+00 -9.64006662e-01 -1.06279843e-01 1.78773475e+00
7.08292007e-01 2.09878623e-01 -1.30482569e-01 -9.75623205e-02
8.83801401e-01 5.40427625e-01 -4.29689169e-01 -3.05520684e-01
-3.43289971e-01 -3.54687870e-01 5.43119311e-01 6.13984287e-01
-1.48073530e+00 1.03605819e+00 4.66243362e+00 7.98647106e-01
-1.18135560e+00 8.49530473e-03 9.23481286e-01 7.17925355e-02
-2.40164891e-01 -2.10918143e-01 -8.07684362e-01 5.39139211e-01
4.91357297e-01 1.06572233e-01 1.31458566e-01 5.05549729e-01
3.27459842e-01 -2.33222052e-01 -1.02410793e+00 1.21871805e+00
3.78115654e-01 -1.34593773e+00 2.80519202e-02 -1.54298007e-01
5.82504511e-01 -3.42486799e-01 1.58391580e-01 1.37119163e-02
-4.32293355e-01 -1.00789976e+00 9.55617726e-01 6.66548073e-01
6.01510584e-01 -7.48917520e-01 7.61048615e-01 3.29444617e-01
-1.23548877e+00 1.65137410e-01 -9.25032943e-02 1.21052839e-01
4.23229069e-01 3.10203135e-01 -5.97928882e-01 4.78383482e-01
7.55990386e-01 9.78189588e-01 -4.69549149e-01 8.08145225e-01
-3.43847811e-01 3.49585444e-01 -3.45849097e-01 -1.30044937e-01
4.80404109e-01 -3.13804567e-01 6.10191107e-01 1.29243290e+00
7.82105774e-02 3.25810850e-01 2.92860568e-02 6.61320329e-01
6.51799887e-03 3.80222574e-02 -5.02581894e-01 9.28996354e-02
3.95870388e-01 1.10836709e+00 -8.31556678e-01 -3.56170028e-01
-3.41763496e-01 1.15792286e+00 1.99821021e-04 3.71576279e-01
-1.24411988e+00 -2.33650491e-01 6.38717949e-01 3.83077115e-02
5.84543586e-01 -2.14827016e-01 -1.56712696e-01 -9.83954191e-01
2.55329758e-01 -4.97316629e-01 3.37900333e-02 -1.05943620e+00
-8.40639353e-01 8.20582330e-01 1.10869117e-01 -1.23361337e+00
-2.34355889e-02 -6.82207704e-01 -2.84122348e-01 7.97164381e-01
-1.19437242e+00 -1.12389266e+00 -2.53403872e-01 3.27083766e-01
9.56520557e-01 6.75580949e-02 3.76647025e-01 2.16933787e-01
-6.51038289e-01 2.65484691e-01 -1.31830394e-01 3.07756275e-01
6.94108725e-01 -1.17748940e+00 4.33549553e-01 8.48046005e-01
4.09663349e-01 3.97197276e-01 8.92480195e-01 -3.49442005e-01
-1.30406678e+00 -1.29309916e+00 8.73386919e-01 -4.48245257e-01
4.81612980e-01 -5.01567781e-01 -7.17464507e-01 6.39456451e-01
3.29636037e-01 1.88749522e-01 2.73670584e-01 -3.99539143e-01
-7.82793388e-02 -3.46331984e-01 -9.29126382e-01 7.04798281e-01
9.36380744e-01 -3.26880962e-01 -7.10478902e-01 3.88765872e-01
1.31915534e+00 -8.17203343e-01 -5.86644709e-01 5.00076592e-01
2.79233575e-01 -7.64492512e-01 1.18590093e+00 -4.36727226e-01
5.70031822e-01 -6.96879268e-01 -4.58055258e-01 -1.01270294e+00
-2.29184359e-01 -3.54464263e-01 -2.43995026e-01 1.21232820e+00
2.40135461e-01 -2.11867109e-01 4.34087783e-01 3.56717110e-01
-2.21163437e-01 -1.10860419e+00 -9.65592802e-01 -3.07369649e-01
-2.98592031e-01 -7.78214216e-01 6.02971673e-01 5.56349695e-01
-3.67788047e-01 5.32633245e-01 -4.92540210e-01 4.30935979e-01
4.25006330e-01 8.47779661e-02 5.93208432e-01 -7.34509647e-01
-1.38812795e-01 -5.99232376e-01 -7.92015195e-01 -1.87962914e+00
4.06015307e-01 -5.86382449e-01 2.47464910e-01 -1.67980683e+00
1.76983073e-01 -1.48582458e-01 -2.70099014e-01 1.11236729e-01
-2.19011247e-01 1.97709739e-01 3.01226825e-01 -3.78991775e-02
-1.05843258e+00 6.60988748e-01 1.30759442e+00 -4.90827024e-01
-1.92584470e-01 -8.28279629e-02 -6.40541136e-01 7.78902948e-01
6.61960423e-01 -1.62638918e-01 -3.38079125e-01 -8.92833531e-01
1.10630980e-02 2.61894315e-01 5.22677898e-01 -8.21750760e-01
3.99296731e-01 -2.84170479e-01 3.67139041e-01 -9.48332667e-01
6.62289262e-01 -8.50523651e-01 -1.00667447e-01 7.34041110e-02
-4.74256277e-01 -5.08331927e-03 2.58917771e-02 5.33136964e-01
-4.14046317e-01 3.54509428e-02 5.91617286e-01 1.89228743e-01
-1.07554555e+00 2.48241752e-01 -8.27430561e-02 -4.47802134e-02
1.40576077e+00 -2.13672146e-01 -2.34629288e-01 -2.29553223e-01
-7.36024916e-01 6.43578887e-01 3.62856746e-01 6.24669790e-01
8.45869303e-01 -1.36839485e+00 -4.90900815e-01 1.41759157e-01
2.47278452e-01 4.25158411e-01 4.37856197e-01 1.06296492e+00
-5.78988373e-01 5.24868131e-01 7.74429813e-02 -9.83418524e-01
-1.29969728e+00 6.36072814e-01 3.60892892e-01 1.58208653e-01
-5.56007504e-01 1.13324201e+00 5.03664792e-01 -3.52413245e-02
2.69221902e-01 -5.85170686e-01 -6.16427362e-01 1.69075698e-01
5.91781080e-01 -1.40231118e-01 -3.38062525e-01 -1.25326264e+00
-6.42672777e-01 8.64498675e-01 7.04574734e-02 -1.41304687e-01
1.15861261e+00 -3.13913763e-01 -1.85817942e-01 5.53598285e-01
1.31099498e+00 -1.13565817e-01 -1.55360150e+00 -3.28329802e-01
-1.72779769e-01 -5.34389734e-01 3.70091163e-02 -4.71983820e-01
-1.20487332e+00 8.34744930e-01 4.49289471e-01 7.27923363e-02
1.17350698e+00 5.41205406e-01 6.35422289e-01 1.65498525e-01
5.56639843e-02 -8.90728951e-01 -7.34770969e-02 4.57582176e-01
9.75619793e-01 -1.37862849e+00 -1.21535249e-01 -3.45928311e-01
-1.07283950e+00 1.11468434e+00 6.93545818e-01 5.23363166e-02
5.86619079e-01 1.75981134e-01 2.52736383e-03 -6.35782853e-02
-6.61728203e-01 -3.08017582e-01 6.57275438e-01 4.43588316e-01
3.52281898e-01 1.25467196e-01 -3.56997609e-01 6.75126672e-01
6.55450001e-02 -4.64478314e-01 1.76566333e-01 5.00293612e-01
-5.92431247e-01 -7.83154547e-01 -4.81444657e-01 2.38299280e-01
-2.37423569e-01 3.50312404e-02 -2.06812009e-01 4.93020087e-01
4.98223990e-01 9.30711448e-01 4.91532862e-01 -5.22186697e-01
3.56829733e-01 -2.36672819e-01 3.89687508e-01 -4.85749751e-01
-1.56814158e-01 3.62013102e-01 -1.36097446e-01 -5.31711102e-01
-8.40177178e-01 -7.36419976e-01 -1.57063508e+00 1.18562855e-01
-2.30586007e-02 6.97897226e-02 5.53779602e-01 1.04733109e+00
2.43414253e-01 6.84447110e-01 5.37026942e-01 -8.98279727e-01
-3.39590386e-02 -8.04023504e-01 -2.99573988e-01 3.86239678e-01
3.43666285e-01 -5.66753685e-01 -3.78003903e-02 2.11246684e-01] | [9.788094520568848, 0.7649635076522827] |
5f7f53a0-836b-46f2-81da-e8381e72fba6 | potential-penetrative-pass-p3 | 2302.10760 | null | https://arxiv.org/abs/2302.10760v1 | https://arxiv.org/pdf/2302.10760v1.pdf | Potential Penetrative Pass (P3) | To score goals in football, a team needs to move forward on the pitch and there are various ways to do so. Depending on the game plan & philosophy; some teams prefer to play long balls from either wings or defense. Others, prefer to penetrate in depth with passes and outplay the opponent players. To objectively & in an automated way evaluate how teams play penetrative passes compared to the number of times they had the potential to do so, the "Potential Penetrative Pass (P3)" concept is presented here. | ['Hadi Sotudeh'] | 2023-01-26 | null | null | null | null | ['philosophy'] | ['miscellaneous'] | [-3.67717236e-01 -2.56416239e-02 -3.34934928e-02 3.21065634e-01
-3.16243678e-01 -9.28899407e-01 2.42064551e-01 2.44168136e-02
-7.35401988e-01 1.03357720e+00 4.97540623e-01 -4.05742019e-01
-6.32866740e-01 -9.73851919e-01 -1.72557414e-01 -4.75128293e-01
-1.00010462e-01 5.99654257e-01 6.23316467e-01 -8.27205896e-01
7.34146118e-01 2.96076387e-01 -1.01603723e+00 2.20761135e-01
1.95524573e-01 2.84571201e-01 -4.45319079e-02 7.57564008e-01
-4.37533623e-03 1.19156718e+00 -8.05274665e-01 -6.77594364e-01
6.07983649e-01 -7.05332696e-01 -9.78843153e-01 -4.07877922e-01
-4.74398077e-01 -1.76516816e-01 -5.44648431e-02 7.14932919e-01
5.26297569e-01 2.81147212e-01 5.20337164e-01 -5.22494435e-01
3.08943331e-01 8.18378150e-01 -5.87094665e-01 7.32795954e-01
7.14570642e-01 1.58876806e-01 7.00915277e-01 -4.35546041e-01
5.27723551e-01 7.73159862e-01 6.56919420e-01 4.31134611e-01
-1.09614301e+00 -6.72298372e-01 -1.20573729e-01 -2.24796608e-01
-1.35771883e+00 -1.80009138e-02 4.64422494e-01 -5.82737505e-01
6.23971820e-01 4.65050459e-01 1.42680788e+00 4.90307063e-01
7.78467238e-01 2.87281543e-01 1.36210716e+00 -3.37188214e-01
2.99042851e-01 -1.34892687e-02 -2.11954057e-01 5.94066307e-02
4.18896347e-01 5.38357079e-01 -9.63058591e-01 -2.35562678e-02
6.92425072e-01 -7.61897504e-01 -1.35651171e-01 3.72060597e-01
-5.72698116e-01 6.49832308e-01 1.53438181e-01 1.63042739e-01
-3.26602578e-01 3.37473065e-01 3.88298452e-01 1.81635112e-01
2.72519261e-01 1.11967659e+00 -8.51634964e-02 -1.05363595e+00
-1.19860947e+00 1.21193743e+00 6.28253043e-01 1.27422273e-01
2.02316642e-01 -2.60702342e-01 2.57353522e-02 4.37686414e-01
1.74809203e-01 -2.25614324e-01 -3.76262441e-02 -7.09682703e-01
5.67390144e-01 6.67918146e-01 2.32463747e-01 -6.68125153e-01
-4.25628096e-01 -2.48835489e-01 1.06140979e-01 1.13605058e+00
6.53938174e-01 -5.54435313e-01 -3.85477096e-01 1.05333054e+00
9.16514248e-02 -4.43345606e-01 -2.33269334e-01 9.48673010e-01
5.48526585e-01 3.53268296e-01 1.95530783e-02 -8.21328089e-02
9.23153222e-01 -6.13549829e-01 -3.84491146e-01 -2.21128121e-01
5.09278297e-01 -9.92267549e-01 7.12049782e-01 1.04963505e+00
-1.79594862e+00 -3.39845240e-01 -1.01189041e+00 6.24322116e-01
1.42158955e-01 -5.33518851e-01 4.90723580e-01 1.06020224e+00
-8.23881686e-01 7.76910663e-01 -5.87173522e-01 6.58675358e-02
9.84716266e-02 3.24583262e-01 -3.19728643e-01 2.54193485e-01
-9.84626234e-01 1.15950298e+00 9.22959030e-01 4.14292254e-02
-6.10873163e-01 -6.76473260e-01 -4.84180421e-01 -1.83716238e-01
2.96441644e-01 -3.82352144e-01 9.05982435e-01 -3.91659558e-01
-1.65210474e+00 8.66274595e-01 6.82405472e-01 -2.12973982e-01
9.85305071e-01 -4.98099744e-01 1.56172097e-01 1.12770028e-01
1.06751956e-01 4.16142106e-01 -7.35815838e-02 -8.84753883e-01
-9.53744173e-01 -7.01594725e-02 9.46340919e-01 8.57036948e-01
2.08449870e-01 6.48935676e-01 6.05647117e-02 -3.01111162e-01
2.42058471e-01 -7.87879705e-01 -3.95045489e-01 -4.31152314e-01
-3.50792140e-01 -2.27548823e-01 2.33335599e-01 -4.28269267e-01
1.35764909e+00 -1.96523261e+00 3.12918335e-01 2.65971601e-01
2.91355193e-01 -3.40828374e-02 6.40998840e-01 1.08833456e+00
-2.18843780e-02 1.55208692e-01 5.30232370e-01 3.24400634e-01
-1.22818552e-01 2.09937006e-01 -4.06034514e-02 3.48621935e-01
-3.90171409e-01 4.29241538e-01 -8.00901473e-01 -4.63844299e-01
3.60946618e-02 -1.65704995e-01 -5.73519409e-01 -1.53110147e-01
3.86874616e-01 2.46547922e-01 -5.45437396e-01 5.82157373e-01
6.16523623e-01 6.80990696e-01 7.12653026e-02 5.38618028e-01
-8.85929585e-01 5.80789447e-01 -1.29900599e+00 1.07751596e+00
1.50141641e-01 5.07420063e-01 5.65117635e-02 -7.14755476e-01
7.99284339e-01 4.81948018e-01 3.46052170e-01 -3.46845835e-01
2.18458146e-01 4.42122340e-01 4.02563334e-01 -1.91331834e-01
8.86705875e-01 -1.12244081e+00 -2.95190722e-01 1.90163434e-01
-3.28501493e-01 -3.77144575e-01 5.60575187e-01 -2.58113965e-02
1.17055809e+00 5.31156480e-01 3.77743751e-01 -5.98878384e-01
9.51149035e-03 5.31142950e-01 5.28181911e-01 7.98791349e-01
-2.73426652e-01 4.74342227e-01 8.57031226e-01 -5.43319166e-01
-6.18760645e-01 -1.16524732e+00 1.38194814e-01 8.44769776e-01
3.51246446e-01 -4.05274421e-01 -6.63659930e-01 -2.14083359e-01
-3.05013388e-01 6.81674540e-01 -6.73638225e-01 -2.82520115e-01
-4.14656818e-01 -5.20593166e-01 7.16774702e-01 6.45851254e-01
3.59906197e-01 -8.50513875e-01 -1.03762221e+00 6.06025577e-01
-2.76182573e-02 -3.05319399e-01 4.43064161e-02 3.46490532e-01
-5.68039894e-01 -9.49224949e-01 -6.31491244e-01 2.69712433e-02
1.67316616e-01 -4.40048948e-02 9.46914136e-01 1.81638539e-01
-3.33189517e-01 1.17977396e-01 -9.50786054e-01 -7.76794314e-01
-3.62243429e-02 -2.51491338e-01 1.14782363e-01 -6.90874457e-01
2.91193247e-01 -4.14523810e-01 -5.83523691e-01 4.24243510e-01
-4.69823807e-01 1.06132120e-01 2.08644509e-01 4.51309800e-01
2.86876947e-01 6.30966604e-01 -8.01317245e-02 -3.89892846e-01
8.40497077e-01 -1.29493296e-01 -2.47268811e-01 -3.77215922e-01
1.65621459e-01 -6.49767756e-01 2.08998665e-01 -5.97189888e-02
-6.78371727e-01 -4.84063298e-01 -4.33212698e-01 1.70120537e-01
2.07489192e-01 6.20818079e-01 2.54444033e-01 -4.80645716e-01
1.04755712e+00 -8.86677131e-02 -3.03930670e-01 -1.85692251e-01
-1.72853291e-01 -9.92631260e-03 -1.14713544e-02 -9.87520158e-01
9.63445425e-01 3.45378518e-01 1.60167385e-02 -4.69769806e-01
-7.45506287e-01 -3.01745176e-01 -5.08431673e-01 -1.07405710e+00
1.15437615e+00 -4.30532783e-01 -1.03085411e+00 1.40525118e-01
-6.35811210e-01 -3.05181563e-01 -5.16993523e-01 1.06079364e+00
-5.89319766e-01 -7.75389001e-02 -3.36921692e-01 -1.17721319e+00
1.47928682e-03 -8.44705641e-01 2.14459985e-01 7.37138212e-01
-6.63876235e-01 -8.12755287e-01 5.25807321e-01 4.16344464e-01
6.26519024e-02 5.18130720e-01 2.93821007e-01 -3.75961930e-01
-2.03196988e-01 -5.86499631e-01 4.38165426e-01 1.96481541e-01
-2.01645121e-02 -1.41898124e-03 -7.73948953e-02 -3.05267394e-01
-3.89661007e-02 -2.94174731e-01 5.11929333e-01 5.85687518e-01
3.55009586e-01 4.84394245e-02 -2.32928827e-01 3.17138612e-01
1.50229490e+00 3.56319219e-01 8.37761402e-01 9.94282603e-01
-3.42496075e-02 8.78705442e-01 1.03917480e+00 4.59944338e-01
-2.59861946e-01 6.25281632e-01 4.10753816e-01 2.93125391e-01
1.70036703e-01 -2.95742780e-01 3.44893903e-01 6.14952087e-01
-1.13371992e+00 -2.13864639e-01 -9.10242140e-01 4.98960435e-01
-1.16265237e+00 -1.34618545e+00 -4.46431011e-01 2.08356190e+00
3.89021128e-01 1.08725727e+00 7.51300395e-01 6.50030196e-01
3.85173947e-01 2.43875291e-03 2.94213146e-01 -7.71051347e-01
1.77219778e-01 5.57557881e-01 8.62103224e-01 5.07199407e-01
-4.25378382e-01 8.14681888e-01 7.98724651e+00 1.03746963e+00
-7.77859569e-01 -9.99656394e-02 2.95463830e-01 -7.31555700e-01
-1.73536971e-01 6.21380985e-01 -9.60237026e-01 4.29583609e-01
4.64686155e-01 -4.28351521e-01 1.95182770e-01 3.76345456e-01
2.37828359e-01 -7.78776944e-01 -6.26632810e-01 2.72513151e-01
-3.50715131e-01 -1.30384922e+00 -3.14182550e-01 3.65707904e-01
2.39712119e-01 -5.16955078e-01 -1.34796172e-01 3.79575729e-01
6.65571868e-01 -1.06294501e+00 1.33656847e+00 5.58551490e-01
5.14654279e-01 -9.81459439e-01 6.32236779e-01 6.07590437e-01
-1.24075747e+00 -1.37830094e-01 -4.38735247e-01 -1.24834716e+00
4.59427297e-01 -9.92550030e-02 -7.14268148e-01 8.90730798e-01
9.16862607e-01 -3.57923001e-01 6.51481673e-02 1.39238977e+00
-3.16486418e-01 7.90730953e-01 -6.06478214e-01 -3.54874492e-01
6.94849372e-01 -4.12019938e-01 7.48919189e-01 5.20192087e-01
4.82008517e-01 6.96130991e-01 1.18598782e-01 6.35657847e-01
9.15371358e-01 4.31558341e-01 -2.70257145e-01 -4.33145799e-02
2.64532208e-01 9.03289676e-01 -1.04876494e+00 2.63887256e-01
-4.31316197e-02 4.29825455e-01 -2.26696998e-01 -1.06874853e-01
-6.15006626e-01 -3.57740194e-01 5.43672025e-01 8.12590420e-01
-2.09895059e-01 -1.95686251e-01 -4.32763338e-01 -4.47991610e-01
-5.34124896e-02 -5.76625705e-01 3.01004082e-01 -7.81758189e-01
-8.70404363e-01 5.63489377e-01 3.50921363e-01 -1.27237856e+00
1.32969081e-01 -7.27262616e-01 -1.10807109e+00 1.35539389e+00
-5.71551323e-01 -8.58388066e-01 9.59892198e-02 1.83626771e-01
5.33363283e-01 -5.44476770e-02 2.66140342e-01 1.37411868e-02
-7.66753703e-02 3.24305952e-01 -3.44229251e-01 -8.36401340e-03
1.37869254e-01 -1.21895528e+00 6.37504784e-03 5.24288535e-01
-3.67357820e-01 4.09866959e-01 1.19238234e+00 -7.31514871e-01
-7.66639709e-01 6.13472499e-02 5.33759534e-01 -3.71742606e-01
5.87944329e-01 1.87232837e-01 -1.34335265e-01 3.35349768e-01
2.54318155e-02 -6.92453027e-01 9.83504951e-01 3.73900592e-01
3.63167763e-01 1.55217588e-01 -9.04842675e-01 8.54286194e-01
7.68042743e-01 -9.23130959e-02 -1.07197404e+00 -3.99175920e-02
-8.84161890e-02 -7.60804474e-01 -6.74653649e-01 2.24996909e-01
7.44007885e-01 -1.56500185e+00 8.72006714e-01 -4.74849284e-01
6.28988326e-01 -2.76598424e-01 2.92376522e-02 -1.20269406e+00
-5.27575731e-01 -6.80461049e-01 1.00099051e+00 8.61161113e-01
3.19979221e-01 -4.47528839e-01 1.14143515e+00 4.66533035e-01
-3.90296161e-01 -9.50745642e-01 -1.13205945e+00 -6.93620026e-01
5.89451492e-01 -3.64950031e-01 1.76528722e-01 5.84811091e-01
7.85617054e-01 -1.72575563e-01 -4.77343589e-01 -1.71371073e-01
5.69747269e-01 -5.79551041e-01 7.26350844e-01 -1.25392354e+00
-8.32360446e-01 -8.47074270e-01 -7.53525317e-01 -8.42039585e-01
-8.51827264e-01 -4.13980693e-01 -2.65786678e-01 -1.67088723e+00
9.81984288e-02 -3.62323135e-01 2.24154303e-03 2.06483398e-02
-1.55205339e-01 4.12644148e-01 4.43056166e-01 8.08813274e-02
-7.48883700e-03 -2.14970499e-01 1.37936044e+00 6.00624979e-01
-3.24663430e-01 1.93803713e-01 -9.16562557e-01 8.70538235e-01
7.18414545e-01 -5.43694377e-01 -3.13589305e-01 7.05235302e-02
9.67303514e-01 6.41806722e-01 -4.10195813e-02 -1.24041808e+00
3.72606739e-02 -7.27864027e-01 2.42493525e-01 -6.75957859e-01
5.17072916e-01 -1.41147703e-01 5.32259524e-01 7.92048633e-01
2.34493334e-02 5.57337003e-03 3.74444336e-01 8.47851187e-02
-1.89537302e-01 -9.30657566e-01 7.13166714e-01 -4.77691293e-01
-1.56840533e-01 -1.63195580e-01 -7.91891575e-01 1.09523907e-02
1.42393386e+00 -1.02831113e+00 -1.63403265e-02 -3.37470591e-01
-1.12777543e+00 2.18405142e-01 4.83237803e-01 -2.67207235e-01
3.67160112e-01 -1.05844831e+00 -9.97900069e-01 -4.53521281e-01
-2.26798370e-01 -4.54537600e-01 5.48518598e-01 1.08902633e+00
-1.62723446e+00 1.75447054e-02 -5.87675750e-01 9.77941453e-02
-1.25004506e+00 1.10387102e-01 7.37855911e-01 -4.64525014e-01
-8.34976137e-01 1.63304508e+00 3.11049312e-01 9.45195779e-02
-3.85093898e-01 3.89938921e-01 -4.51572061e-01 1.38592173e-03
7.24435329e-01 6.08684659e-01 -2.01281369e-01 -4.87181216e-01
-3.10043126e-01 5.39490759e-01 -1.32858932e-01 -6.79211378e-01
9.65558946e-01 1.36737257e-01 8.33139792e-02 4.37375754e-01
-2.97277956e-03 6.58729792e-01 -1.42277157e+00 5.30531526e-01
-2.90202886e-01 -1.03187418e+00 -8.81279856e-02 -7.15507507e-01
-5.15001416e-01 6.73874676e-01 -1.83270246e-01 5.39572537e-01
9.07800436e-01 -8.34495872e-02 3.32343221e-01 -2.01606393e-01
6.92467034e-01 -1.43513036e+00 6.28288388e-02 3.25843304e-01
6.76536977e-01 -7.83025101e-03 2.95069069e-01 -5.22230446e-01
-7.88031340e-01 1.13407648e+00 5.29317617e-01 -5.59287608e-01
4.79124904e-01 4.84229773e-01 6.72961622e-02 -6.51908040e-01
-5.77621877e-01 -1.98657006e-01 -2.42041517e-02 5.51296234e-01
5.80197811e-01 -1.22938212e-02 -1.28029299e+00 5.15981734e-01
-1.00873041e+00 -8.96605477e-02 9.37574506e-01 1.09854436e+00
-7.86846697e-01 -1.31722176e+00 -8.49092722e-01 4.73042816e-01
-7.57327616e-01 1.26588270e-01 -5.81120849e-01 9.89422619e-01
6.47557497e-01 9.77200687e-01 -2.30232835e-01 -5.89769840e-01
6.35906279e-01 -1.91859692e-01 8.26851606e-01 -6.83533788e-01
-1.12725365e+00 1.46010011e-01 4.82687503e-01 -1.89081475e-01
-2.25986898e-01 -8.39570761e-01 -1.06500494e+00 -9.02173221e-01
-5.75367093e-01 5.68536103e-01 2.06211075e-01 1.24480307e+00
-9.29804623e-01 6.91842318e-01 3.06406766e-01 -9.18174863e-01
-4.39250976e-01 -7.55442142e-01 -1.26556027e+00 -1.94060817e-01
-5.25053740e-01 -1.13450861e+00 -3.34941149e-01 -6.75884902e-01] | [6.434110164642334, 0.41427457332611084] |
c458a7fe-2367-4768-a6cb-4bfef5f41d9f | coordinated-multi-agent-exploration-using | null | null | https://openreview.net/forum?id=MPO4oML_JC | https://openreview.net/pdf?id=MPO4oML_JC | Coordinated Multi-Agent Exploration Using Shared Goals | Exploration is critical for good results of deep reinforcement learning algorithms and has drawn much attention. However, existing multi-agent deep reinforcement learning algorithms still use mostly noise-based techniques. It was recognized recently that noise-based exploration is suboptimal in multi-agent settings, and exploration methods that consider agents' cooperation have been developed. However, existing methods suffer from a common challenge: agents struggle to identify states that are worth exploring, and don't coordinate their exploration efforts toward those states. To address this shortcoming, in this paper, we proposed coordinated multi-agent exploration (CMAE): agents share a common goal while exploring. The goal is selected by a normalized entropy-based technique from multiple projected state spaces. Then, agents are trained to reach the goal in a coordinated manner. We demonstrated that our approach needs only $1\%-5\%$ of the environment steps to achieve similar or better returns than state-of-the-art baselines on various sparse-reward tasks, including a sparse-reward version of the Starcraft multi-agent challenge (SMAC). | ['Alex Schwing', 'Unnat Jain', 'Iou-Jen Liu'] | 2021-01-01 | null | null | null | null | ['smac-1', 'smac'] | ['playing-games', 'playing-games'] | [-4.00773317e-01 -1.28743529e-01 -3.13957602e-01 6.80168271e-02
-9.77570117e-01 -4.21477973e-01 6.60641909e-01 3.08424294e-01
-8.81629109e-01 1.00671065e+00 2.34071970e-01 2.73805461e-03
-1.98280647e-01 -6.80018842e-01 -6.33003592e-01 -8.03426325e-01
-5.64894676e-01 8.88538718e-01 8.24873522e-03 -5.47153056e-01
3.66042912e-01 1.94227383e-01 -1.37634468e+00 -2.27176815e-01
9.54437554e-01 6.34462774e-01 2.51198083e-01 4.74676907e-01
1.14826962e-01 9.11195040e-01 -6.22238636e-01 1.61280975e-01
5.01964092e-01 -7.49034703e-01 -9.10442173e-01 -2.30521619e-01
-1.26220912e-01 -5.04344940e-01 -2.89495558e-01 1.17532253e+00
6.51035011e-01 7.13568807e-01 2.08677709e-01 -1.39644885e+00
-3.02814394e-01 9.82359707e-01 -6.73929751e-01 2.22897276e-01
1.91683084e-01 5.29845357e-01 1.16995621e+00 -3.27540815e-01
7.33092308e-01 1.36949170e+00 2.26762041e-01 6.83654368e-01
-1.19952762e+00 -6.82550311e-01 4.92210895e-01 2.02304527e-01
-7.97546446e-01 -2.48930216e-01 7.14345813e-01 8.25232342e-02
1.23834383e+00 -9.18152779e-02 1.00509572e+00 1.11266971e+00
9.38543379e-02 1.14931905e+00 1.31676888e+00 -1.71193495e-01
9.12922621e-01 -3.91220033e-01 -3.51887554e-01 8.66604328e-01
1.77964985e-01 6.86306655e-01 -7.66510129e-01 -2.43613854e-01
7.36224115e-01 -2.26256058e-01 2.22717613e-01 -7.17334330e-01
-1.31574321e+00 1.30847979e+00 5.77920377e-01 2.21076682e-01
-6.84982777e-01 5.82595944e-01 4.81307238e-01 3.84507686e-01
6.07411414e-02 1.06670260e+00 -1.87479734e-01 -8.05787981e-01
-6.36069477e-01 7.59235919e-01 8.14904392e-01 5.15226007e-01
8.14639032e-01 3.90887052e-01 1.47234052e-01 5.81058025e-01
2.73863852e-01 4.32626158e-01 6.66342318e-01 -1.35974479e+00
4.68957752e-01 5.72216749e-01 3.45892102e-01 -6.82806730e-01
-6.14274085e-01 -5.66809595e-01 -4.74358499e-01 9.38783407e-01
2.17545763e-01 -6.01348877e-01 -7.95493007e-01 2.00490832e+00
4.53751504e-01 6.94023743e-02 5.44118345e-01 1.10371518e+00
3.47998351e-01 3.35737199e-01 4.33355309e-02 -7.19124302e-02
8.86418879e-01 -1.35129464e+00 -5.31799614e-01 -7.69826353e-01
6.04721963e-01 -1.91656038e-01 8.97966981e-01 4.57928658e-01
-1.39964855e+00 -2.11358249e-01 -1.12397408e+00 5.83369792e-01
-1.03245057e-01 -4.35140193e-01 8.42965543e-01 4.13553566e-01
-9.59921539e-01 8.45743716e-01 -1.27400362e+00 -1.17233746e-01
5.72180986e-01 3.08824837e-01 -6.82766316e-03 1.97291553e-01
-1.03407586e+00 1.22800136e+00 5.49536705e-01 -1.78747758e-01
-1.63288295e+00 -1.81153610e-01 -8.03205788e-01 5.98349199e-02
8.40142965e-01 -4.17058617e-01 1.57864666e+00 -8.53496969e-01
-1.75800800e+00 7.22201094e-02 1.29568607e-01 -8.55871856e-01
4.03948337e-01 -4.48798865e-01 -6.74855784e-02 4.32294086e-02
2.55259544e-01 8.30455959e-01 5.68091571e-01 -1.34922981e+00
-8.48012090e-01 -2.15929672e-01 1.73753008e-01 7.97413707e-01
-1.69529602e-01 -1.88418701e-01 -1.25208214e-01 -4.34309810e-01
-6.88804686e-02 -1.14374936e+00 -8.76848757e-01 -4.38272446e-01
-1.97524771e-01 -2.58018196e-01 6.08478785e-01 -4.95205112e-02
8.63991737e-01 -1.69961262e+00 6.10090017e-01 2.38458320e-01
2.70664811e-01 2.09176347e-01 -5.90204298e-01 6.23542309e-01
4.46800739e-01 3.91732566e-02 -3.76787707e-02 -6.61347747e-01
3.35178226e-01 3.16320956e-01 -1.73339710e-01 3.54449034e-01
-1.04128249e-01 9.33806658e-01 -1.35791600e+00 -3.02589983e-01
1.84971422e-01 3.79506499e-02 -6.47508085e-01 2.64865518e-01
-6.76980913e-01 4.01773095e-01 -6.50361359e-01 6.84517026e-01
1.99836046e-01 -2.96492964e-01 3.74281585e-01 5.70173383e-01
-1.67547137e-01 2.25222304e-01 -1.32759750e+00 2.08518434e+00
-2.52170265e-01 3.90028685e-01 2.16334060e-01 -9.12686050e-01
8.15934658e-01 3.37523296e-02 8.11834753e-01 -1.01434696e+00
1.50217041e-01 3.07327181e-01 4.16719794e-01 -1.22559078e-01
7.77110219e-01 1.30149752e-01 -9.23411548e-02 8.46468985e-01
-1.98018029e-01 -2.96551526e-01 5.27951896e-01 9.38852802e-02
1.38672471e+00 3.91045988e-01 3.46307218e-01 -1.05022736e-01
-1.91701680e-01 5.29604316e-01 7.27317274e-01 1.17623973e+00
-5.15957952e-01 5.72068840e-02 5.86739540e-01 -5.97020984e-01
-8.69822621e-01 -9.01161253e-01 6.07628167e-01 1.17164409e+00
3.75022143e-01 -3.69097829e-01 -6.02704227e-01 -7.03576803e-01
-2.15838514e-02 8.02085638e-01 -8.08849275e-01 -1.55536458e-01
-8.56714368e-01 -7.28471935e-01 4.03573424e-01 5.03273129e-01
6.83334708e-01 -1.68232799e+00 -1.70763993e+00 6.05542064e-01
-4.76679020e-03 -6.34465158e-01 -1.88044995e-01 5.70015490e-01
-7.49701977e-01 -1.10401380e+00 -6.73095524e-01 -5.23030996e-01
3.79554063e-01 1.02685161e-01 1.15153456e+00 3.97034213e-02
-1.25223935e-01 2.83850938e-01 -5.20154953e-01 -3.22502583e-01
-4.29282904e-01 2.05188215e-01 2.66469419e-01 -6.44609690e-01
2.87710220e-01 -4.33417052e-01 -7.89101422e-01 2.16028720e-01
-5.54762959e-01 -1.22720733e-01 8.39919865e-01 1.04602003e+00
3.80759269e-01 -3.70410904e-02 8.65963042e-01 -1.73598066e-01
1.17411613e+00 -4.70594585e-01 -8.23215723e-01 -4.52502705e-02
-7.34723449e-01 4.21923578e-01 5.72996378e-01 -6.25084400e-01
-6.97226763e-01 -9.38615762e-03 5.00890389e-02 -4.47674930e-01
1.67801958e-02 6.76557124e-01 4.27931368e-01 -1.59314319e-01
7.95638919e-01 3.17320943e-01 3.66159707e-01 -1.23346992e-01
3.42413515e-01 -3.14713009e-02 1.51463553e-01 -7.38699079e-01
5.19524574e-01 3.59432459e-01 -5.35934754e-02 -2.61219889e-01
-4.21459138e-01 -9.69567448e-02 4.86658737e-02 -1.75710589e-01
5.37903786e-01 -9.92159903e-01 -1.10508239e+00 2.09362656e-01
-7.46450424e-01 -8.84166002e-01 -5.97681880e-01 6.16923034e-01
-8.00811648e-01 2.25490317e-01 -4.03316110e-01 -9.45349634e-01
-3.81081700e-01 -1.58957696e+00 5.65245986e-01 5.82921386e-01
-1.37978435e-01 -7.80925393e-01 6.49371743e-01 4.21896316e-02
6.79127097e-01 2.25177124e-01 4.77541417e-01 -7.12011993e-01
-6.79450154e-01 2.51015067e-01 2.19939008e-01 -2.02330336e-01
-7.35424533e-02 -5.87948024e-01 -3.60977650e-01 -8.32223058e-01
-3.59772861e-01 -1.06598341e+00 9.74435687e-01 4.79806393e-01
7.15170741e-01 -2.44995728e-01 -3.68426919e-01 3.67486775e-01
1.39085865e+00 4.92838532e-01 3.50594044e-01 1.03561008e+00
5.25865182e-02 1.65609673e-01 9.04526949e-01 9.50294435e-01
4.82309878e-01 5.53763866e-01 1.20143533e+00 1.24185242e-01
2.58407742e-01 -2.95025140e-01 5.14426649e-01 2.37308249e-01
-1.56035367e-02 -2.80661613e-01 -8.74792039e-01 7.51531005e-01
-2.24710298e+00 -8.95299494e-01 7.08447099e-01 1.89554667e+00
8.06461871e-01 2.09363595e-01 4.67077047e-01 -4.29319769e-01
2.39695325e-01 4.32927072e-01 -1.26115310e+00 -3.68408263e-01
-1.29381299e-01 6.54524043e-02 3.95589501e-01 5.23840904e-01
-9.94834542e-01 1.26358116e+00 6.52246571e+00 6.13064289e-01
-7.99371362e-01 -1.84726939e-02 3.97005528e-01 -6.88686848e-01
-2.55844742e-01 1.20180070e-01 -6.61190510e-01 2.34347522e-01
7.39412487e-01 -5.31533696e-02 9.57328916e-01 1.14694142e+00
8.34446028e-02 -5.00291586e-01 -9.09051180e-01 7.98127711e-01
-3.00760772e-02 -1.34612942e+00 -3.79988372e-01 2.70224929e-01
1.02281618e+00 5.36487222e-01 2.17424631e-01 7.48916328e-01
1.44120765e+00 -1.23308468e+00 5.84639072e-01 6.24795370e-02
1.88345656e-01 -9.55577314e-01 4.39127028e-01 3.96029055e-01
-1.08142960e+00 -4.88408983e-01 -3.47179234e-01 -2.29124323e-01
3.05252820e-01 -1.40705064e-01 -8.56607437e-01 3.05078834e-01
6.60404921e-01 3.24290246e-01 -3.50450516e-01 1.02069294e+00
-1.42927870e-01 2.31811494e-01 -3.07091147e-01 -6.31384671e-01
9.64167118e-01 -2.85128951e-01 7.32324541e-01 5.12793839e-01
3.69770080e-01 -1.54984280e-01 6.05765224e-01 9.14880514e-01
1.22161126e-02 -2.12508380e-01 -4.15130347e-01 -3.16726536e-01
6.53528988e-01 1.10589302e+00 -6.40292108e-01 -2.51108825e-01
-1.35602236e-01 7.27395177e-01 7.93649793e-01 2.30183765e-01
-8.69866490e-01 -1.57031432e-01 8.66175115e-01 -5.50927460e-01
3.43638390e-01 -4.51418996e-01 -9.58303362e-02 -1.00617051e+00
-3.15620869e-01 -1.32140946e+00 3.98635298e-01 -3.31905097e-01
-7.40639985e-01 7.23401189e-01 -2.50649750e-01 -1.07135570e+00
-7.78859079e-01 -7.14673400e-02 -6.43905759e-01 5.67480326e-01
-1.64770484e+00 -7.73599565e-01 -1.18204609e-01 4.23261851e-01
7.32561409e-01 -8.31299245e-01 9.45743620e-01 -2.92232871e-01
-5.70997775e-01 3.66005450e-01 4.28137600e-01 -1.35029495e-01
3.35175425e-01 -1.50745273e+00 5.41732550e-01 7.84501493e-01
1.97983831e-01 2.82522321e-01 9.18258846e-01 -8.37897897e-01
-1.68291712e+00 -6.41836047e-01 9.50663909e-02 1.74016375e-02
6.19878292e-01 7.25840256e-02 -5.69311023e-01 5.91216743e-01
6.30395889e-01 -6.57208711e-02 3.20719898e-01 3.73457134e-01
-4.57412750e-02 3.30497116e-01 -9.51857090e-01 1.00447202e+00
8.81035864e-01 3.43738608e-02 -5.48606455e-01 1.22093804e-01
4.39132541e-01 -5.60802102e-01 -6.28212750e-01 4.26762328e-02
4.08654273e-01 -1.06323493e+00 8.69489849e-01 -8.90179694e-01
3.76674891e-01 -1.60604551e-01 -1.30408809e-01 -2.12348151e+00
-3.51236910e-01 -8.79192352e-01 -3.85429621e-01 3.80534172e-01
2.70111114e-01 -5.21704137e-01 1.16302395e+00 4.34518814e-01
-2.74212450e-01 -9.41513598e-01 -1.17156208e+00 -9.30068493e-01
3.08732539e-01 1.57534018e-01 6.03142500e-01 6.09724224e-01
4.10611212e-01 1.23587891e-01 -5.69877326e-01 -6.61354279e-03
8.17575336e-01 3.81726593e-01 8.38180482e-01 -9.07938421e-01
-4.40483183e-01 -7.86158860e-01 1.94811285e-01 -9.66479421e-01
1.98490605e-01 -4.30216908e-01 2.45443389e-01 -1.66398716e+00
1.58851698e-01 -6.76012933e-01 -3.56684357e-01 5.47436774e-01
-6.79771826e-02 -3.84425312e-01 4.43219513e-01 1.94067955e-01
-1.24114144e+00 1.10187638e+00 1.25585616e+00 -8.90188962e-02
-5.25197446e-01 -3.70889544e-01 -7.18642771e-01 5.29570460e-01
1.16181278e+00 -4.13118869e-01 -3.30367982e-01 -5.14013648e-01
3.91164780e-01 3.32597494e-01 2.71267682e-01 -1.22292495e+00
4.08287764e-01 -6.01249337e-01 1.79296300e-01 -4.67017055e-01
8.04975688e-01 -4.45783824e-01 -1.57244548e-01 1.10257387e+00
-6.07707322e-01 4.22395170e-01 1.75317898e-01 7.01366723e-01
-9.34450254e-02 -3.92594337e-01 8.18616509e-01 -5.93585134e-01
-9.21123147e-01 2.22041726e-01 -4.94575053e-01 3.45758855e-01
1.17461050e+00 1.09408990e-01 -4.12369460e-01 -4.34207588e-01
-4.72931534e-01 7.90121853e-01 5.02618432e-01 3.15934539e-01
7.94586301e-01 -1.20596564e+00 -7.78397799e-01 -1.06390506e-01
-1.98344544e-01 5.51537648e-02 1.73025340e-01 4.67094362e-01
-2.80820191e-01 1.66533977e-01 -5.94265401e-01 -2.42421046e-01
-9.34125662e-01 5.45195937e-01 4.62416470e-01 -6.62718475e-01
-7.24646688e-01 7.51842558e-01 -4.25091296e-01 -3.71426195e-01
4.59465504e-01 -7.52598345e-02 -1.86109841e-01 -4.55308631e-02
4.43167120e-01 5.08916616e-01 -4.41917181e-01 -1.38408631e-01
-3.48025352e-01 1.97433025e-01 -3.00849527e-01 -6.40765727e-01
1.49158728e+00 8.31742585e-02 4.71924305e-01 4.62040789e-02
6.83246434e-01 -3.57470036e-01 -1.78565061e+00 -3.17825168e-01
-1.57070041e-01 -4.22541142e-01 2.12411970e-01 -1.01898515e+00
-9.59437191e-01 5.34770489e-01 6.33336663e-01 2.56257713e-01
7.32838154e-01 -1.25982419e-01 7.69038439e-01 8.53972018e-01
5.85888624e-01 -1.61402321e+00 6.57339513e-01 7.49851406e-01
9.28419590e-01 -1.66668534e+00 1.15832593e-02 5.96459568e-01
-1.13806176e+00 8.78404379e-01 9.87368166e-01 -3.04834783e-01
8.52653161e-02 1.09131448e-01 -3.39947157e-02 -3.51509273e-01
-1.12509608e+00 -6.07286692e-01 -4.00556535e-01 6.19148076e-01
-2.05318168e-01 4.94110137e-02 1.01243176e-01 2.07282007e-01
-1.08106397e-01 -2.25788027e-01 4.81374413e-01 1.28122497e+00
-8.82417500e-01 -9.91866231e-01 -2.15878978e-01 4.09059882e-01
-1.71844766e-01 1.32218793e-01 -9.31359753e-02 7.87079990e-01
-4.42259490e-01 9.52429593e-01 -1.91797577e-02 -2.51946777e-01
-2.62889974e-02 -3.13652962e-01 3.84484559e-01 -3.17770392e-01
-7.90349185e-01 6.49180189e-02 1.30709648e-01 -1.02008700e+00
-1.88191697e-01 -9.80530918e-01 -1.45887303e+00 -2.68230587e-01
-6.17603250e-02 4.70818073e-01 6.52135074e-01 7.50479043e-01
5.06647050e-01 5.99016011e-01 5.35707593e-01 -9.46819782e-01
-1.10695863e+00 -8.09374988e-01 -4.22009140e-01 1.57687902e-01
4.11979020e-01 -1.02262080e+00 -3.02427977e-01 -7.96826124e-01] | [3.8592233657836914, 1.8186182975769043] |
6692e0d6-2dc5-42a2-82d9-7eb32c43e0d0 | weakly-supervised-cell-tracking-via-backward | 2007.15258 | null | https://arxiv.org/abs/2007.15258v1 | https://arxiv.org/pdf/2007.15258v1.pdf | Weakly-Supervised Cell Tracking via Backward-and-Forward Propagation | We propose a weakly-supervised cell tracking method that can train a convolutional neural network (CNN) by using only the annotation of "cell detection" (i.e., the coordinates of cell positions) without association information, in which cell positions can be easily obtained by nuclear staining. First, we train co-detection CNN that detects cells in successive frames by using weak-labels. Our key assumption is that co-detection CNN implicitly learns association in addition to detection. To obtain the association, we propose a backward-and-forward propagation method that analyzes the correspondence of cell positions in the outputs of co-detection CNN. Experiments demonstrated that the proposed method can associate cells by analyzing co-detection CNN. Even though the method uses only weak supervision, the performance of our method was almost the same as the state-of-the-art supervised method. Code is publicly available in https://github.com/naivete5656/WSCTBFP | ['Chenyang Wang', 'Ryoma Bise', 'Junya Hayashida', 'Dai Fei Elmer Ker', 'Kazuya Nishimura'] | 2020-07-30 | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/1489_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123570103.pdf | eccv-2020-8 | ['cell-detection'] | ['computer-vision'] | [-2.40152776e-01 1.28731057e-01 -2.09221728e-02 -1.03745311e-01
-6.85301960e-01 -6.97845757e-01 4.50116962e-01 2.72738010e-01
-7.80995190e-01 9.16027844e-01 -2.96328038e-01 -1.97369680e-02
5.10634303e-01 -8.63341212e-01 -9.62615788e-01 -1.06159699e+00
5.86632565e-02 4.83905345e-01 4.70510423e-01 4.17280532e-02
8.75259005e-03 5.56457698e-01 -1.09786987e+00 3.17399919e-01
2.78323382e-01 7.58690536e-01 -4.20096330e-03 9.04424667e-01
7.61618167e-02 8.46514344e-01 -3.69498700e-01 8.35035555e-03
2.99857114e-03 -3.82556289e-01 -7.31767774e-01 -1.38268247e-01
1.85851589e-01 -2.96716690e-01 -3.97586107e-01 1.04442549e+00
3.90607655e-01 -4.38875645e-01 4.82780486e-01 -1.14444089e+00
-4.85360146e-01 3.05341572e-01 -6.14550054e-01 3.54821801e-01
-2.01087780e-02 1.01782240e-01 7.44812310e-01 -9.78957295e-01
9.60976481e-01 6.83912277e-01 8.33964288e-01 6.72507346e-01
-1.35201967e+00 -6.71777546e-01 -1.07731875e-02 -2.64992714e-01
-1.77367353e+00 -2.99874395e-01 4.14354116e-01 -5.22010803e-01
6.81765974e-01 8.15916136e-02 8.23068380e-01 7.08772242e-01
-6.88544139e-02 8.91170919e-01 8.74406517e-01 -3.80145073e-01
7.87840113e-02 -6.40847832e-02 2.33853295e-01 1.02201617e+00
1.36561289e-01 -9.80627537e-02 -2.39894584e-01 1.74646191e-02
1.12563217e+00 4.24001813e-01 -2.39294410e-01 -1.29905224e-01
-1.67631137e+00 4.44526643e-01 6.66018963e-01 6.39330685e-01
1.98364303e-01 4.68281329e-01 4.20270592e-01 1.93479434e-02
5.49220502e-01 7.84442872e-02 -7.71543860e-01 2.96602696e-01
-6.80228770e-01 3.84755656e-02 5.57865381e-01 7.84245372e-01
8.64490867e-01 -4.86888677e-01 -2.45174274e-01 3.96566659e-01
2.08026439e-01 3.12374264e-01 2.79391557e-01 -8.25065017e-01
-9.00480002e-02 8.20014298e-01 2.02889413e-01 -6.62313938e-01
-6.05837286e-01 -5.85588813e-01 -1.00562882e+00 2.21468434e-01
1.13996053e+00 -3.65979820e-01 -7.82590628e-01 1.77612102e+00
3.78365755e-01 6.17478669e-01 -2.46572793e-02 8.97972047e-01
8.49144578e-01 2.94069886e-01 -2.25114271e-01 -1.84447214e-01
1.19225812e+00 -1.09205401e+00 -7.42886603e-01 2.88082331e-01
1.40503359e+00 -4.74037826e-01 7.03633964e-01 -1.01866052e-01
-8.34458530e-01 -3.87476563e-01 -8.13731849e-01 -3.13578039e-01
-6.22756958e-01 5.74410677e-01 5.72735965e-01 -1.33879378e-01
-1.17797315e+00 3.68636072e-01 -1.10308266e+00 -4.99465466e-01
7.36299336e-01 4.30305123e-01 -4.80666816e-01 2.99778938e-01
-7.13508248e-01 4.70881194e-01 2.39069521e-01 1.62297711e-01
-9.40984786e-01 -6.20364189e-01 -7.30494916e-01 1.05940126e-01
1.09102286e-01 -6.28157198e-01 1.10599244e+00 -9.59855080e-01
-1.22669065e+00 1.37857592e+00 -5.18629730e-01 -2.67067760e-01
7.24629045e-01 -1.14248656e-01 2.94675287e-02 6.19560853e-02
8.18297118e-02 9.80016232e-01 -1.23723214e-02 -1.36639416e+00
-9.31922197e-01 -2.81993687e-01 -3.87166888e-02 -1.63880214e-01
-1.72398269e-01 -9.51786935e-02 -9.80783522e-01 -4.47063684e-01
1.89173043e-01 -1.06626523e+00 -1.15499333e-01 6.12048984e-01
-6.01915300e-01 -1.40651166e-01 1.09776008e+00 -3.32664579e-01
7.96059251e-01 -2.37206912e+00 -1.50449187e-01 3.54824811e-02
4.55817223e-01 8.37733373e-02 6.93826228e-02 8.82000998e-02
-1.29785324e-02 3.42457555e-02 -2.38027737e-01 -8.00323486e-01
-3.54759336e-01 8.70881751e-02 -1.47163302e-01 8.75841320e-01
5.68617135e-02 1.15675104e+00 -1.28052914e+00 -7.32311964e-01
-6.00064285e-02 7.23601818e-01 -2.95644939e-01 4.01530303e-02
-2.04732388e-01 7.89141357e-01 -1.07566208e-01 8.72860134e-01
4.22863811e-01 -7.95720577e-01 2.80737251e-01 -1.52311042e-01
-1.09795973e-01 -2.37997591e-01 -7.36947179e-01 1.41210592e+00
7.22337980e-03 8.02444160e-01 1.27407655e-01 -8.46175790e-01
6.96634412e-01 3.83942723e-01 5.03558874e-01 -3.34893793e-01
2.63915241e-01 2.56870866e-01 -1.25015378e-01 -9.69499126e-02
-6.25385270e-02 2.87765026e-01 2.06919387e-01 3.39270741e-01
2.67163217e-02 3.55054200e-01 2.76097268e-01 2.32771203e-01
1.03801298e+00 1.46686956e-01 9.96600911e-02 -2.93093890e-01
6.04404688e-01 -2.05354020e-02 7.83981323e-01 4.82319176e-01
-3.28884542e-01 7.89776981e-01 7.69010603e-01 -5.85172415e-01
-9.05603528e-01 -7.54862189e-01 -9.10181701e-02 9.64795887e-01
2.57256746e-01 -2.04012841e-01 -8.07740629e-01 -9.18823421e-01
5.22132702e-02 -1.23823352e-01 -1.11928654e+00 2.82666564e-01
-5.48527300e-01 -7.24665046e-01 7.73131311e-01 8.60353589e-01
4.36872840e-01 -9.63730097e-01 -3.16798866e-01 -7.31708035e-02
-1.37034757e-02 -9.93993640e-01 -4.75922823e-01 4.42858011e-01
-8.18208814e-01 -1.26581728e+00 -8.75117064e-01 -1.29049683e+00
1.32615268e+00 5.65539040e-02 7.65620947e-01 8.84469151e-01
-5.17840323e-04 -1.55208215e-01 -4.24060673e-02 -3.15612972e-01
4.51855622e-02 1.96842596e-01 -1.77560493e-01 -1.34161383e-01
3.56617332e-01 -3.49800915e-01 -7.41316378e-01 2.78928965e-01
-7.06711531e-01 2.43588656e-01 3.57766628e-01 1.00292826e+00
1.15525913e+00 -2.65917987e-01 5.47367334e-01 -1.38305545e+00
-1.23032607e-01 -2.05965459e-01 -7.48458326e-01 1.43780649e-01
-9.92201194e-02 -2.10287511e-01 7.10887849e-01 -1.99756712e-01
-6.72410369e-01 5.74480891e-01 -2.15415895e-01 -2.89980382e-01
-5.17273188e-01 3.27937156e-01 1.21044733e-01 -1.38516560e-01
3.66850227e-01 2.74217784e-01 2.78930087e-02 -1.71454653e-01
4.49684225e-02 2.01369032e-01 9.09897566e-01 -2.32729524e-01
5.69849133e-01 1.16957736e+00 2.43851036e-01 -2.41849601e-01
-9.64210153e-01 -6.30473316e-01 -9.86882865e-01 -1.37944028e-01
8.40698779e-01 -1.03488672e+00 -9.59236741e-01 8.96903098e-01
-1.27727485e+00 -8.62047374e-01 -2.49347780e-02 2.94120163e-01
-4.48034942e-01 1.23252019e-01 -1.28585446e+00 -4.61576909e-01
-1.58528671e-01 -7.61436045e-01 1.14090776e+00 1.97827101e-01
-1.12303987e-01 -1.26786482e+00 2.70785749e-01 4.80369925e-02
3.50956544e-02 5.07848620e-01 5.78436792e-01 -7.83612430e-01
-7.74843454e-01 -3.06629330e-01 -2.98912615e-01 -3.48547041e-01
2.06115186e-01 3.79295439e-01 -1.20362425e+00 -3.42378646e-01
-7.02917755e-01 -1.74163103e-01 1.11559451e+00 4.68515843e-01
1.40755498e+00 -1.02007881e-01 -1.09845436e+00 9.39092815e-01
1.37564504e+00 -2.30549891e-02 7.02170730e-01 3.76320183e-01
7.78280199e-01 3.17976207e-01 5.37389457e-01 1.88273713e-01
3.24396908e-01 4.98107761e-01 3.44642520e-01 -8.84505987e-01
-2.25021750e-01 -6.64122626e-02 -1.35440901e-01 6.23181701e-01
-8.64099041e-02 -8.70790631e-02 -9.65011954e-01 7.87022769e-01
-2.33165407e+00 -8.02440524e-01 -2.48437181e-01 1.83881545e+00
1.11776924e+00 1.67285725e-01 1.05795935e-01 1.21586144e-01
9.23283637e-01 -3.80576938e-01 -4.94112730e-01 3.10537696e-01
-1.91967994e-01 -2.37749398e-01 5.69582939e-01 6.14573777e-01
-1.20681131e+00 9.51234281e-01 5.88635445e+00 6.65891230e-01
-1.24826789e+00 6.58997744e-02 1.17482805e+00 1.00378662e-01
9.75058526e-02 -1.19122304e-01 -9.44992542e-01 7.16559231e-01
1.75662011e-01 3.55816424e-01 -1.09001316e-01 6.55562997e-01
5.48906252e-03 -3.34713496e-02 -1.28423846e+00 7.15234876e-01
-2.40782902e-01 -1.72112763e+00 -3.36814135e-01 1.73795283e-01
8.79015863e-01 9.51584578e-02 -1.30891919e-01 1.95473373e-01
3.81082088e-01 -9.21564400e-01 5.44716597e-01 6.87261999e-01
9.68764126e-01 -5.54818332e-01 1.19869542e+00 4.82082397e-01
-1.24765539e+00 2.28864238e-01 -1.60887912e-01 -1.33673400e-01
3.42989340e-03 7.21878231e-01 -7.81099319e-01 1.19765788e-01
7.10096180e-01 9.08244789e-01 -4.31513608e-01 1.06012154e+00
-2.93737501e-01 5.54259419e-01 -5.04890859e-01 9.58063826e-02
-1.05207741e-01 3.65133993e-02 1.18938424e-01 1.69126964e+00
3.31219912e-01 1.52182579e-01 1.97388813e-01 8.38655353e-01
-3.76764476e-01 -2.77117997e-01 -4.97463763e-01 2.74625123e-01
5.04155338e-01 1.60850596e+00 -1.36833203e+00 -5.72574973e-01
-4.31654721e-01 7.75367796e-01 7.88396478e-01 4.45500553e-01
-9.18130934e-01 -3.28973174e-01 3.33461195e-01 2.64334112e-01
3.85537326e-01 -1.27240762e-01 -4.96058404e-01 -1.01664484e+00
-3.97367448e-01 -1.53530031e-01 5.21703124e-01 -6.71272993e-01
-1.29749322e+00 2.80098587e-01 -5.35357773e-01 -1.12576663e+00
3.68344456e-01 -5.66128492e-01 -8.11532259e-01 4.14186895e-01
-1.55773735e+00 -1.32929254e+00 -4.28569824e-01 4.65377301e-01
1.71995193e-01 1.99980289e-01 7.75445282e-01 2.94727683e-01
-8.64845574e-01 6.87669456e-01 3.50361168e-02 7.92314887e-01
7.31509745e-01 -1.40408862e+00 1.50776565e-01 7.24884212e-01
5.08438097e-03 7.51307428e-01 2.50240743e-01 -5.87280452e-01
-7.44992197e-01 -1.36624575e+00 1.05763483e+00 -5.42273283e-01
5.81137478e-01 -5.31067908e-01 -9.24503624e-01 9.22967732e-01
7.40780607e-02 8.30646873e-01 7.37489164e-01 -1.18214063e-01
-1.34503663e-01 -6.78499267e-02 -9.12124395e-01 6.62174523e-01
9.76872802e-01 -5.62127233e-01 1.42077552e-02 3.75999063e-01
6.67416394e-01 -7.56318092e-01 -5.76792300e-01 3.78476679e-01
4.47372139e-01 -5.81956267e-01 8.51010263e-01 -1.29333615e-01
3.81355703e-01 -1.06919992e+00 2.47288406e-01 -8.55490267e-01
-4.57592100e-01 -1.12424074e-02 -3.80477011e-01 1.15913212e+00
6.46223605e-01 -7.32419372e-01 9.87304509e-01 6.76473081e-02
-1.46350592e-01 -1.08500290e+00 -7.59378493e-01 -6.31671250e-01
4.43998761e-02 2.49183312e-01 3.76576930e-01 1.23211730e+00
2.10898593e-01 1.91080883e-01 1.58814900e-02 3.92339647e-01
5.42711377e-01 5.22703715e-02 8.88998330e-01 -1.02506316e+00
-2.84233302e-01 -3.62535655e-01 -3.55615109e-01 -1.10088766e+00
2.91657329e-01 -9.28288639e-01 1.53515294e-01 -1.31877327e+00
5.37783086e-01 -2.72705704e-01 -7.49208033e-01 1.03416348e+00
-2.07422480e-01 8.38250935e-01 -1.67200878e-01 4.88077760e-01
-1.05491853e+00 5.58216609e-02 1.27052057e+00 -1.47439390e-01
9.62198339e-03 -4.21057761e-01 -3.36125106e-01 8.90148401e-01
9.23931837e-01 -4.82335985e-01 1.85266480e-01 -4.88575310e-01
2.79418141e-01 -9.65058357e-02 5.98460495e-01 -1.13200474e+00
7.77490735e-01 2.22849071e-01 8.72283340e-01 -8.33900869e-01
2.49593928e-02 -7.09844112e-01 8.08229223e-02 6.49940908e-01
-5.01129985e-01 -1.78587079e-01 1.07700273e-01 4.39097911e-01
-1.72008798e-01 1.47933304e-01 6.94033265e-01 -8.73631835e-02
-3.72566730e-01 4.11121666e-01 -3.45219851e-01 -2.77034014e-01
1.14808571e+00 -1.71026513e-01 -7.39542067e-01 -8.12931061e-02
-7.75649250e-01 4.30386722e-01 9.23965394e-01 -3.68553311e-01
2.54162759e-01 -1.29870236e+00 -3.80872071e-01 6.09718859e-02
2.90486336e-01 5.58300734e-01 -7.99861327e-02 1.21123481e+00
-8.21550548e-01 2.76719660e-01 1.90036651e-02 -8.74202847e-01
-1.25068736e+00 6.24988675e-01 5.70228934e-01 -3.78912151e-01
-4.03026283e-01 1.10446537e+00 4.96749490e-01 -4.82161462e-01
2.52790660e-01 -3.60722125e-01 -2.82898247e-01 -4.21391636e-01
4.22202170e-01 1.67619586e-02 -1.68369547e-01 -5.69174647e-01
-5.73910296e-01 5.69768727e-01 -1.69573128e-01 2.09320888e-01
1.13292050e+00 -4.00261171e-02 -4.12239552e-01 4.64109927e-01
1.25417888e+00 2.70036869e-02 -1.61965001e+00 -1.86593115e-01
-1.18830472e-01 -1.72413334e-01 -8.38382244e-02 -7.16610968e-01
-1.52524102e+00 6.88225269e-01 4.42046821e-01 -1.76530033e-01
8.08500350e-01 2.19451860e-01 7.89869308e-01 5.25121570e-01
2.24954978e-01 -8.84077489e-01 2.45239283e-03 6.94551051e-01
3.57335627e-01 -1.27575541e+00 -2.06468612e-01 -6.75362825e-01
-8.50298032e-02 1.06021380e+00 1.05569410e+00 -1.33553416e-01
8.66029859e-01 7.61570215e-01 4.10456538e-01 -3.83598030e-01
-7.83217072e-01 -3.57601374e-01 -2.03375876e-01 4.43514168e-01
8.14878047e-01 -1.34218261e-01 -1.66899040e-01 5.59688807e-01
2.45376632e-01 3.67256343e-01 3.26580673e-01 9.37563837e-01
-2.11931765e-01 -7.39796579e-01 -4.76117693e-02 1.90939903e-01
-6.68057978e-01 -3.02980691e-02 -4.73352104e-01 7.47040093e-01
4.06173587e-01 4.33013499e-01 5.18417120e-01 -1.95658192e-01
-1.23268977e-01 -2.74625599e-01 2.34543949e-01 -7.03371048e-01
-3.18689734e-01 2.01766193e-01 -3.30533803e-01 -3.38845342e-01
-5.00300348e-01 -4.09073353e-01 -1.90843427e+00 -1.80810526e-01
-4.16474313e-01 -1.15358355e-02 -7.35005215e-02 9.67188597e-01
2.73370475e-01 4.52934295e-01 3.58875424e-01 -6.37738705e-01
4.40757722e-01 -7.17225909e-01 -6.30420446e-01 4.78702903e-01
6.18955016e-01 -4.13731456e-01 -3.53595555e-01 5.67562699e-01] | [14.639287948608398, -3.1629228591918945] |
93446343-532f-4d89-bc1d-bd7617c67388 | edge-storage-management-recipe-with-zero-shot | 2307.04298 | null | https://arxiv.org/abs/2307.04298v1 | https://arxiv.org/pdf/2307.04298v1.pdf | Edge Storage Management Recipe with Zero-Shot Data Compression for Road Anomaly Detection | Recent studies show edge computing-based road anomaly detection systems which may also conduct data collection simultaneously. However, the edge computers will have small data storage but we need to store the collected audio samples for a long time in order to update existing models or develop a novel method. Therefore, we should consider an approach for efficient storage management methods while preserving high-fidelity audio. A hardware-perspective approach, such as using a low-resolution microphone, is an intuitive way to reduce file size but is not recommended because it fundamentally cuts off high-frequency components. On the other hand, a computational file compression approach that encodes collected high-resolution audio into a compact code should be recommended because it also provides a corresponding decoding method. Motivated by this, we propose a way of simple yet effective pre-trained autoencoder-based data compression method. The pre-trained autoencoder is trained for the purpose of audio super-resolution so it can be utilized to encode or decode any arbitrary sampling rate. Moreover, it will reduce the communication cost for data transmission from the edge to the central server. Via the comparative experiments, we confirm that the zero-shot audio compression and decompression highly preserve anomaly detection performance while enhancing storage and transmission efficiency. | ['Myung Jin Kim', 'UJu Gim', 'YeongHyeon Park'] | 2023-07-10 | null | null | null | null | ['audio-super-resolution', 'super-resolution', 'anomaly-detection', 'management', 'audio-super-resolution', 'edge-computing', 'data-compression'] | ['audio', 'computer-vision', 'methodology', 'miscellaneous', 'music', 'time-series', 'time-series'] | [ 4.56680432e-02 -6.03440523e-01 1.01196170e-01 -1.13608159e-01
-4.51928020e-01 5.22950441e-02 5.74546494e-02 2.53117114e-01
-3.22382987e-01 3.52552950e-01 1.74228743e-01 -2.02706307e-01
-9.01958197e-02 -1.23258543e+00 -5.87125897e-01 -6.98828995e-01
-1.65856183e-01 -4.74441983e-02 5.04228592e-01 -1.51609600e-01
3.05442899e-01 4.17809188e-01 -2.29163170e+00 1.26129463e-01
9.07146215e-01 1.29495561e+00 4.39433664e-01 7.32319653e-01
-2.34769136e-01 5.74029028e-01 -4.95169014e-01 -1.43297896e-01
2.05584019e-01 -2.29531944e-01 -3.68473560e-01 -2.12502241e-01
-1.49059832e-01 -8.47028255e-01 -4.70968574e-01 1.11536956e+00
5.88260531e-01 2.58707404e-01 3.59668672e-01 -1.04826570e+00
-7.39814714e-02 5.14175415e-01 -4.43585783e-01 3.89900595e-01
3.41058880e-01 -1.77169621e-01 6.70750976e-01 -5.85311294e-01
2.47959644e-01 7.31545925e-01 5.30006707e-01 1.48993060e-01
-7.26372480e-01 -6.59866512e-01 -2.76168048e-01 8.79646778e-01
-1.68463707e+00 -6.08230054e-01 1.06268620e+00 -1.18085489e-01
7.14931846e-01 6.28254533e-01 1.01481056e+00 6.40110672e-01
1.48220181e-01 3.76475990e-01 5.25199592e-01 -4.58451182e-01
4.69324827e-01 3.74795645e-02 1.28057199e-02 2.91866481e-01
2.02645779e-01 4.34945077e-02 -6.68984115e-01 -1.41608447e-01
4.27478075e-01 2.05194771e-01 -4.43795830e-01 2.17621654e-01
-7.05982268e-01 4.80422437e-01 -4.14540954e-02 4.08747196e-01
-6.70111120e-01 6.85517937e-02 8.16292703e-01 4.21593785e-01
5.15537977e-01 -1.44017071e-01 -1.09940179e-01 -6.71625435e-01
-1.12355113e+00 -3.68752256e-02 4.89194900e-01 6.58274710e-01
6.79398775e-01 5.24860740e-01 2.77360618e-01 7.43160009e-01
2.06085309e-01 2.71959364e-01 7.90082157e-01 -9.40534592e-01
2.65501499e-01 4.43465710e-01 -3.14449072e-01 -1.38234520e+00
-3.42978053e-02 -2.41148695e-01 -1.04356349e+00 2.11015537e-01
-8.58030245e-02 1.82312444e-01 -3.87665719e-01 1.28530574e+00
4.72313881e-01 6.29112184e-01 1.49455518e-01 8.75103772e-01
4.66689378e-01 8.92408490e-01 -3.40366989e-01 -4.86697614e-01
1.35102391e+00 -5.24043024e-01 -1.06636393e+00 4.02768523e-01
6.54713333e-01 -7.64852643e-01 1.21002007e+00 6.79661691e-01
-1.11427164e+00 -5.69349885e-01 -1.25639462e+00 8.42730850e-02
-3.54579747e-01 1.67430565e-01 3.15415531e-01 7.15687394e-01
-8.91191661e-01 6.70961201e-01 -8.73945951e-01 -1.46136940e-01
1.40717760e-01 1.54059961e-01 -2.32847512e-01 2.02341154e-01
-1.23257446e+00 3.06403428e-01 5.24966359e-01 2.66021937e-02
-6.07417345e-01 -4.96433020e-01 -5.72955906e-01 4.04431105e-01
1.10601574e-01 -3.06208491e-01 8.71512055e-01 -7.59167254e-01
-1.67477453e+00 1.52222976e-01 -1.78232402e-01 -8.02713513e-01
6.66454211e-02 -2.54489154e-01 -9.15686965e-01 6.25853360e-01
-2.33627617e-01 1.36594400e-01 1.12343156e+00 -6.79131389e-01
-7.11737633e-01 -3.77323240e-01 -2.69643307e-01 9.62441713e-02
-1.21918094e+00 -4.70527895e-02 -2.12827533e-01 -7.36036539e-01
2.69859344e-01 -5.34488678e-01 1.10884309e-01 -1.68623269e-01
-1.28127933e-01 -2.61713751e-02 1.29254723e+00 -9.54840899e-01
1.79010904e+00 -2.47630215e+00 -2.90318489e-01 4.03889894e-01
1.06411450e-01 3.39266330e-01 3.10185432e-01 3.25225204e-01
1.95664950e-02 -1.12589441e-01 -7.72015750e-02 -2.26521358e-01
-4.38738376e-01 8.72577280e-02 -5.19988537e-01 2.76162207e-01
-3.08005571e-01 1.44807294e-01 -5.04891753e-01 -5.67544162e-01
2.85910845e-01 5.53825617e-01 -8.51721644e-01 1.70758948e-01
1.75501451e-01 2.21072480e-01 -3.15801084e-01 4.85817045e-01
7.84853041e-01 6.53434023e-02 -1.93456426e-01 -3.62924725e-01
-2.99589038e-01 3.72646034e-01 -1.52260554e+00 1.55755246e+00
-5.16551018e-01 6.39643788e-01 2.67317206e-01 -1.30935144e+00
1.04532564e+00 5.11180043e-01 6.75233662e-01 -6.64748669e-01
1.90299243e-01 2.89351791e-01 -3.45918834e-01 -7.28078127e-01
9.61742759e-01 2.16328055e-01 2.84836560e-01 3.91326159e-01
-4.34381485e-01 2.41804212e-01 1.77509133e-02 7.66835958e-02
1.08081031e+00 -3.35985452e-01 6.25476316e-02 1.02174476e-01
9.08317327e-01 -2.26265445e-01 5.68225384e-01 2.22176775e-01
-7.07780421e-02 2.82865942e-01 -1.07792415e-01 -2.10742056e-01
-1.13938928e+00 -7.30317116e-01 -1.90384075e-01 8.80646050e-01
1.07941166e-01 -8.62287164e-01 -8.26201618e-01 -5.42818792e-02
-2.62066007e-01 6.12784922e-01 1.26015082e-01 -4.85903293e-01
-5.26379287e-01 -2.97118515e-01 7.82559395e-01 3.94470215e-01
9.23336029e-01 -5.98886311e-01 -8.14265192e-01 3.57076973e-01
-4.01970863e-01 -8.31702948e-01 -2.17554197e-01 -8.77702311e-02
-1.12343931e+00 -6.61838531e-01 -4.15430546e-01 -4.56280917e-01
3.51860255e-01 5.74789524e-01 4.45446074e-01 3.06892186e-01
4.89456654e-02 3.34920406e-01 -7.09119797e-01 -3.35324854e-01
-3.49513233e-01 -1.33018076e-01 5.12134552e-01 1.43037841e-01
5.37793517e-01 -1.10666966e+00 -5.17865121e-01 -9.61866751e-02
-9.55587506e-01 -1.48111567e-01 2.30731145e-01 5.42859972e-01
7.75571942e-01 6.94524586e-01 6.61018014e-01 -1.98468119e-01
6.00286007e-01 -3.95637989e-01 -4.39605564e-01 -5.83020337e-02
-6.38046741e-01 -1.95175231e-01 1.17960417e+00 -3.45831662e-01
-8.96480918e-01 -1.32326812e-01 -4.89943266e-01 -7.07981646e-01
-1.76222369e-01 4.42317456e-01 -1.93031773e-01 -5.04335109e-03
4.44682509e-01 7.16867685e-01 8.21761563e-02 -6.76284790e-01
2.03863814e-01 1.27625406e+00 6.80818856e-01 -3.26420516e-01
6.87272549e-01 6.83178604e-01 -8.38304237e-02 -1.31626987e+00
1.39685839e-01 -3.32743526e-01 -2.49949843e-01 -4.15292531e-01
6.81727290e-01 -9.36340928e-01 -7.31144845e-01 3.89449388e-01
-1.00023723e+00 2.30362654e-01 -3.07039827e-01 7.47180998e-01
-4.36364383e-01 8.39324355e-01 -5.26707351e-01 -1.02876472e+00
-5.49019396e-01 -8.78688157e-01 8.30484390e-01 1.25867546e-01
9.00452211e-02 -4.15601403e-01 -8.53309035e-02 8.95001665e-02
5.54576814e-01 -1.61731526e-01 6.89281404e-01 -4.20317411e-01
-7.30192125e-01 -3.64733547e-01 3.43267657e-02 4.26359117e-01
-4.78040837e-02 7.99439102e-02 -9.53388035e-01 -2.24938989e-01
3.28578562e-01 1.62411302e-01 7.11609483e-01 2.83558846e-01
1.84943843e+00 -4.67651248e-01 -1.39479041e-01 6.83943391e-01
1.17046964e+00 2.55699933e-01 8.99625361e-01 2.71311700e-01
4.14401114e-01 2.80492783e-01 6.51842594e-01 7.52885342e-01
2.70575672e-01 7.02652514e-01 3.20152462e-01 4.22837406e-01
9.25148576e-02 -3.31585824e-01 5.06289363e-01 1.55322611e+00
-2.08688229e-01 -1.36630654e-01 -4.70780849e-01 3.99491459e-01
-1.49736249e+00 -1.27743661e+00 -2.13785276e-01 2.50236011e+00
7.07095146e-01 1.51041940e-01 2.38860711e-01 1.07268667e+00
7.22183406e-01 -6.83945939e-02 -1.90859199e-01 -5.39217949e-01
2.93556333e-01 9.15191919e-02 2.68405437e-01 1.22677073e-01
-7.59667516e-01 3.87300760e-01 5.53420162e+00 1.26949787e+00
-1.43738246e+00 2.31168106e-01 1.53091162e-01 -1.65953353e-01
-3.10530186e-01 -7.44915903e-02 -5.55232286e-01 1.05200052e+00
1.57590020e+00 -3.74438256e-01 5.48360288e-01 1.15669417e+00
4.76490974e-01 -1.12604536e-01 -4.91584897e-01 1.21214783e+00
-1.56328052e-01 -1.23051918e+00 8.14353079e-02 1.24015182e-01
7.97358081e-02 -3.94891202e-01 -1.91771761e-01 1.34527981e-01
-5.42630315e-01 -4.29607242e-01 5.96176565e-01 6.48894548e-01
8.48436832e-01 -1.02791321e+00 5.97219586e-01 5.24466515e-01
-1.68069696e+00 -2.13709801e-01 -7.64209926e-01 -2.63316602e-01
3.07999730e-01 9.44919407e-01 -5.14610648e-01 6.42333090e-01
1.06153679e+00 6.38723910e-01 -2.91439354e-01 1.01951301e+00
3.71700376e-01 8.05188656e-01 -7.08302677e-01 1.11987837e-01
-3.07889491e-01 -2.31282651e-01 7.95909941e-01 9.86476719e-01
1.03679192e+00 2.47309878e-01 -3.14175747e-02 4.83842641e-01
1.32435426e-01 3.26633900e-01 -8.57509255e-01 7.22894222e-02
9.30045903e-01 1.05877149e+00 -4.29388851e-01 -3.98977011e-01
-3.97099406e-01 8.95768106e-01 -1.39357388e-01 4.79341410e-02
-8.20338309e-01 -7.41404772e-01 5.49802840e-01 4.45775807e-01
1.85557961e-01 -3.69641274e-01 -2.15916097e-01 -1.00947821e+00
2.25358665e-01 -6.63826525e-01 2.82068610e-01 -6.96187615e-01
-7.19555676e-01 4.31191683e-01 -2.02524379e-01 -1.83653843e+00
-4.10286605e-01 2.15312354e-02 -6.56914175e-01 4.80312198e-01
-1.40332806e+00 -5.58497190e-01 -6.03046536e-01 9.60147381e-01
4.86705661e-01 -3.14158887e-01 9.77305472e-01 9.35317934e-01
-5.70321500e-01 6.77225590e-01 2.61390716e-01 -7.82931373e-02
4.10331368e-01 -6.49281859e-01 -2.15915725e-01 1.10461795e+00
7.18196779e-02 5.00528038e-01 7.09517717e-01 -7.06609428e-01
-1.50625622e+00 -8.46406043e-01 5.43467641e-01 3.32293272e-01
3.78327161e-01 2.12723352e-02 -1.33069623e+00 2.55044550e-01
-5.01230769e-02 -6.03324659e-02 8.49934638e-01 -3.05061072e-01
9.05884355e-02 -6.08972728e-01 -1.12871623e+00 5.00355065e-01
8.14456403e-01 -7.86601841e-01 -4.60339755e-01 -6.77672774e-02
8.03611577e-01 -2.86366135e-01 -9.50039148e-01 3.53147566e-01
4.14478213e-01 -1.03831327e+00 9.25747156e-01 -9.67601687e-02
4.09811229e-01 -5.62715948e-01 -3.61665845e-01 -8.89221430e-01
-3.60594653e-02 -5.46894073e-01 -6.30838692e-01 1.42112315e+00
-7.53902048e-02 -6.35553300e-01 7.73439109e-01 2.52612323e-01
-2.57323414e-01 -5.63849986e-01 -1.31446028e+00 -8.74223590e-01
-6.45543873e-01 -9.56468344e-01 7.43171751e-01 6.57000184e-01
2.99686044e-01 -1.56004503e-01 -5.61636686e-01 3.47126395e-01
4.54274625e-01 -3.85442264e-02 8.02676916e-01 -1.23710823e+00
-2.67783791e-01 -5.63617647e-02 -7.37059832e-01 -1.21586537e+00
-1.42433673e-01 -6.83524370e-01 -2.09411010e-01 -9.22781110e-01
-3.37631553e-01 -3.77508312e-01 -1.40883684e-01 -8.12244937e-02
2.55972117e-01 1.72691122e-01 -1.94617778e-01 5.37438750e-01
-2.36075953e-01 8.32635880e-01 7.03901052e-01 1.73218146e-01
-4.47139949e-01 9.68174487e-02 -3.98228586e-01 7.19732165e-01
8.72740984e-01 -4.25779432e-01 -5.44457436e-01 -2.87529707e-01
2.29058757e-01 1.33455977e-01 3.43053102e-01 -1.59862614e+00
5.71909308e-01 9.98551622e-02 1.19931065e-01 -8.61357927e-01
4.08224076e-01 -1.30281711e+00 2.34712437e-01 4.50971723e-01
4.58023883e-02 2.23476812e-01 7.91649446e-02 8.69887292e-01
-4.17461812e-01 -2.65202165e-01 5.73592842e-01 1.70923740e-01
-8.15013528e-01 9.92693454e-02 -7.14334965e-01 -4.18368012e-01
1.04971552e+00 -5.78483820e-01 -1.64443590e-02 -5.75595081e-01
-4.34793204e-01 -1.03082545e-01 4.59923118e-01 2.86100805e-01
8.60353231e-01 -1.44997609e+00 -4.12464023e-01 7.12396979e-01
-1.96278065e-01 -1.42098859e-01 5.87997735e-01 8.05602789e-01
-5.88211119e-01 1.94195420e-01 -2.65616626e-01 -6.25559986e-01
-1.39012635e+00 5.09918153e-01 -7.41592944e-02 8.45933259e-02
-9.52527225e-01 4.42184567e-01 -6.35257125e-01 2.53529847e-01
2.54026294e-01 -1.71836048e-01 -3.80132586e-01 1.51289897e-02
9.48444188e-01 8.00026178e-01 2.28240132e-01 -4.70472187e-01
-1.92148477e-01 5.52167773e-01 1.90241933e-01 -8.12269002e-02
1.19283950e+00 -5.29218733e-01 -1.56798497e-01 3.41776580e-01
1.12366211e+00 1.86014399e-01 -6.42899930e-01 -6.82777762e-02
-2.91516006e-01 -7.11911023e-01 5.41810155e-01 1.66723460e-01
-1.14913142e+00 9.13252473e-01 8.83513153e-01 5.92163742e-01
1.70832574e+00 -4.98630285e-01 1.39309311e+00 4.34738129e-01
5.03640354e-01 -1.50824213e+00 7.13815242e-02 2.32452944e-01
5.20995915e-01 -6.78952575e-01 -8.82890448e-02 -3.87369782e-01
-1.76393270e-01 1.33444524e+00 4.19879287e-01 -5.04513718e-02
8.73023510e-01 5.42357802e-01 -4.17364955e-01 7.09274858e-02
-5.29848337e-01 -1.27470762e-01 -7.79105648e-02 5.20322084e-01
1.49374306e-01 -1.21138752e-01 -5.08992255e-01 7.51699984e-01
-5.47155917e-01 1.93275854e-01 5.74540854e-01 8.02375972e-01
-8.34733129e-01 -9.98901129e-01 -5.34349382e-01 6.13786280e-01
-4.36659992e-01 5.90699725e-02 2.89332151e-01 1.54708594e-01
3.94474044e-02 1.00559509e+00 3.78364354e-01 -1.02898180e+00
1.75533578e-01 2.50515431e-01 -8.66276100e-02 -1.20359056e-01
-1.96835265e-01 8.84961523e-03 -2.21529320e-01 -5.98266065e-01
-1.87027931e-01 -3.91393125e-01 -1.41414773e+00 -9.34035480e-01
-5.89764118e-01 2.21192956e-01 8.02356541e-01 7.33406007e-01
6.71656311e-01 6.64553642e-01 7.58218467e-01 -6.10111117e-01
-3.89385849e-01 -7.80902445e-01 -6.78332627e-01 9.90537405e-02
1.86196893e-01 -4.07773793e-01 -4.90935713e-01 3.21565419e-02] | [7.885635852813721, 2.8368101119995117] |
0dc31f64-0f70-4993-bdfe-40f177f6637f | real-time-streaming-video-denoising-with | 2207.06937 | null | https://arxiv.org/abs/2207.06937v1 | https://arxiv.org/pdf/2207.06937v1.pdf | Real-time Streaming Video Denoising with Bidirectional Buffers | Video streams are delivered continuously to save the cost of storage and device memory. Real-time denoising algorithms are typically adopted on the user device to remove the noise involved during the shooting and transmission of video streams. However, sliding-window-based methods feed multiple input frames for a single output and lack computation efficiency. Recent multi-output inference works propagate the bidirectional temporal feature with a parallel or recurrent framework, which either suffers from performance drops on the temporal edges of clips or can not achieve online inference. In this paper, we propose a Bidirectional Streaming Video Denoising (BSVD) framework, to achieve high-fidelity real-time denoising for streaming videos with both past and future temporal receptive fields. The bidirectional temporal fusion for online inference is considered not applicable in the MoViNet. However, we introduce a novel Bidirectional Buffer Block as the core module of our BSVD, which makes it possible during our pipeline-style inference. In addition, our method is concise and flexible to be utilized in both non-blind and blind video denoising. We compare our model with various state-of-the-art video denoising models qualitatively and quantitatively on synthetic and real noise. Our method outperforms previous methods in terms of restoration fidelity and runtime. Our source code is publicly available at https://github.com/ChenyangQiQi/BSVD | ['Qifeng Chen', 'Xin Yang', 'Junming Chen', 'Chenyang Qi'] | 2022-07-14 | null | null | null | null | ['video-denoising'] | ['computer-vision'] | [ 4.18513715e-02 -6.45947397e-01 1.58547863e-01 -3.05565178e-01
-7.82468557e-01 -4.76826370e-01 4.40424830e-01 -1.96730539e-01
-3.38860005e-01 3.09131205e-01 3.67129922e-01 -2.08978385e-01
-2.54218969e-02 -6.07754171e-01 -8.06538880e-01 -7.33374715e-01
8.38899165e-02 -3.56193930e-01 4.95459229e-01 -4.34628464e-02
1.96493436e-02 2.44069412e-01 -1.64582181e+00 5.43320656e-01
6.32497013e-01 1.15621591e+00 4.92516398e-01 9.98746872e-01
1.51028112e-01 8.11349094e-01 -3.55438054e-01 -5.09989262e-01
3.27159166e-01 -5.32226443e-01 -4.57286052e-02 -7.48938546e-02
4.08628613e-01 -8.97839844e-01 -9.32627261e-01 1.10055816e+00
7.04318225e-01 6.24700636e-03 8.90203640e-02 -1.09117031e+00
-4.13680345e-01 3.82988006e-01 -4.83634800e-01 3.28546137e-01
4.84756976e-01 2.33507320e-01 6.31734848e-01 -8.66056144e-01
4.43579942e-01 1.14209759e+00 8.00458729e-01 5.33557177e-01
-9.51030731e-01 -7.49134541e-01 2.56134599e-01 6.24363899e-01
-1.22287095e+00 -1.04310060e+00 5.87432742e-01 -1.17104940e-01
8.79527271e-01 5.65809727e-01 9.05457139e-01 1.19548082e+00
1.12541914e-01 8.48870039e-01 7.08811283e-01 -8.89858976e-03
2.65434593e-01 -5.12478650e-01 -2.53084511e-01 4.33274984e-01
-2.68641472e-01 1.20945178e-01 -9.84234452e-01 -5.90653419e-02
9.62268591e-01 4.15045112e-01 -5.76562345e-01 3.42544258e-01
-1.18390596e+00 4.45776701e-01 1.82268307e-01 1.44209698e-01
-5.27869463e-01 3.26623231e-01 5.45736909e-01 6.11241758e-01
6.04599059e-01 -5.05434871e-01 -1.30712405e-01 -5.21705925e-01
-1.45016241e+00 2.18867898e-01 6.25306129e-01 9.23884034e-01
4.00643915e-01 1.06790021e-01 -3.27087551e-01 6.87670648e-01
3.10238510e-01 5.04899502e-01 1.79744735e-01 -1.43974423e+00
3.53408515e-01 -4.16182056e-02 8.73099267e-02 -1.14334202e+00
1.67772304e-02 -1.23440288e-01 -1.27158713e+00 2.16789782e-01
1.28169879e-01 8.60549882e-02 -8.51021707e-01 1.40103269e+00
2.69103676e-01 7.95310199e-01 -1.74764380e-01 1.30725527e+00
7.88254917e-01 1.02900195e+00 -2.47390583e-01 -6.00393832e-01
1.18768704e+00 -9.73790526e-01 -1.18314838e+00 1.77509487e-02
-8.61448571e-02 -1.03861034e+00 6.51006699e-01 8.65724564e-01
-1.57083237e+00 -5.18117785e-01 -9.29776013e-01 -3.15432608e-01
2.41502389e-01 6.72581345e-02 3.10399860e-01 3.98988634e-01
-1.24430847e+00 7.18887925e-01 -1.41910803e+00 -1.25381529e-01
3.27323943e-01 1.91848159e-01 -2.29564145e-01 -2.90722430e-01
-9.11034644e-01 2.87931085e-01 -2.95196235e-01 6.10720932e-01
-1.07591069e+00 -5.43293893e-01 -6.76204562e-01 6.65545277e-03
1.83478653e-01 -8.60876679e-01 1.21366358e+00 -1.05644870e+00
-1.73986363e+00 2.44562089e-01 -7.50919521e-01 -4.70552295e-01
7.67673910e-01 -3.41783077e-01 -3.64872634e-01 3.88542026e-01
-2.59153187e-01 3.28488022e-01 1.38434792e+00 -1.05845904e+00
-5.38615525e-01 -2.47994855e-01 4.58599515e-02 1.19149953e-01
-5.19703209e-01 1.96952894e-01 -1.07150793e+00 -1.10794282e+00
3.22652072e-01 -4.62871850e-01 5.29426932e-02 4.28093582e-01
-6.45867661e-02 2.09068790e-01 9.75487471e-01 -1.22280872e+00
1.43102169e+00 -2.48691463e+00 3.65745723e-01 -6.29376546e-02
1.78056195e-01 2.63839155e-01 -2.48648766e-02 4.52360928e-01
2.39327356e-01 -1.51579455e-01 -3.40652496e-01 -8.74418914e-01
-1.74221992e-01 2.88905591e-01 -4.84111756e-01 6.22167587e-01
-7.68000111e-02 3.90469283e-01 -7.50738442e-01 -3.01866651e-01
2.55225599e-01 9.18086648e-01 -7.57233143e-01 4.87272471e-01
7.20978305e-02 5.32928705e-01 -1.26995131e-01 7.35412478e-01
9.51197803e-01 1.61987111e-01 1.62917718e-01 -5.74583232e-01
-2.97477871e-01 3.01283330e-01 -1.40585756e+00 2.10374284e+00
-5.05806208e-01 7.01841354e-01 7.92665303e-01 -7.76903152e-01
5.97248375e-01 6.75828278e-01 2.26395890e-01 -7.32927203e-01
-7.79620409e-02 1.73678398e-01 -4.63600218e-01 -6.54339194e-01
4.86505210e-01 3.68961878e-02 5.88779032e-01 1.70629114e-01
-1.04098842e-02 1.80869102e-01 2.80989975e-01 2.57862926e-01
1.28978467e+00 3.31868649e-01 -1.76570252e-01 1.16780072e-01
4.52600509e-01 -4.87355620e-01 8.48735154e-01 7.04169512e-01
-7.74771646e-02 1.03154933e+00 3.31140220e-01 -9.47971120e-02
-1.07466555e+00 -1.21103978e+00 8.19080174e-02 9.31274533e-01
2.93268293e-01 -7.73285687e-01 -7.05672026e-01 -4.92350571e-02
-2.50545770e-01 2.96372294e-01 -1.35189161e-01 1.77825868e-01
-6.35370374e-01 -3.89033884e-01 4.55081075e-01 4.40849662e-01
4.30895358e-01 -6.90043867e-01 -4.76318419e-01 2.99666554e-01
-4.33155030e-01 -1.10271215e+00 -6.62925363e-01 -3.16076398e-01
-1.03855538e+00 -7.12665677e-01 -7.02818573e-01 -7.11226583e-01
4.29823250e-01 5.26363134e-01 7.14328229e-01 3.03141415e-01
-1.35912560e-02 3.96250963e-01 -4.41624105e-01 1.02835685e-01
-1.24959715e-01 -5.27994335e-01 7.79665634e-02 3.07478398e-01
5.24719842e-02 -1.01261592e+00 -9.82234895e-01 2.54855186e-01
-1.23451674e+00 2.75552869e-01 1.34535134e-01 8.75376225e-01
6.48063302e-01 1.21269166e-01 2.36355320e-01 -2.11630017e-01
4.40449119e-01 -5.11204541e-01 -7.21250951e-01 9.66592785e-03
-1.26588047e-01 -3.09711397e-01 7.54291534e-01 -3.91515732e-01
-1.20064104e+00 -1.39886677e-01 -4.59054798e-01 -8.11391830e-01
1.47152632e-01 3.95788699e-01 -1.00400306e-01 1.87050909e-01
1.25666678e-01 3.18858385e-01 2.22259268e-01 -8.07202637e-01
7.94591904e-02 8.40269864e-01 7.26171315e-01 -2.44404331e-01
5.54996669e-01 9.14464772e-01 -3.18274468e-01 -8.13818693e-01
-3.43702346e-01 -4.43085670e-01 -1.67710960e-01 -4.39889848e-01
6.66835904e-01 -1.34618163e+00 -9.08371449e-01 8.32026780e-01
-1.36441195e+00 -3.95197600e-01 2.04618409e-01 5.03471375e-01
-3.39544743e-01 6.57760918e-01 -1.15199983e+00 -7.80878007e-01
-4.28706229e-01 -1.35705221e+00 1.13152015e+00 1.49030134e-01
1.19724669e-01 -6.65062547e-01 -3.22003275e-01 2.29756787e-01
4.94913101e-01 -1.96915746e-01 2.38142818e-01 2.73900837e-01
-8.80270779e-01 -7.79380277e-02 -1.94870651e-01 5.70768595e-01
-3.84438820e-02 1.75455391e-01 -1.07585895e+00 -5.18795550e-01
3.80535543e-01 3.62555459e-02 1.11869085e+00 6.17596924e-01
1.36826885e+00 -3.86746317e-01 5.05845472e-02 1.01421320e+00
1.40850961e+00 -2.27281097e-02 8.89987767e-01 1.52530253e-01
4.77050602e-01 3.23669940e-01 5.07976651e-01 7.12158382e-01
3.01631361e-01 6.04763985e-01 4.91386116e-01 1.95809200e-01
-2.78811723e-01 -1.77769825e-01 9.07611847e-01 1.38018000e+00
-3.04368764e-01 -5.18453658e-01 -5.43935835e-01 6.43092752e-01
-2.01642799e+00 -1.15880442e+00 -1.45059988e-01 2.22978830e+00
7.73691058e-01 -9.61390231e-03 -1.01472862e-01 3.31620395e-01
6.58674717e-01 2.23794937e-01 -3.79372567e-01 -1.58456087e-01
-2.18668431e-01 1.48219362e-01 3.56754333e-01 7.20741808e-01
-9.47253287e-01 4.46535408e-01 5.21136999e+00 9.69712496e-01
-1.09238386e+00 5.50378084e-01 3.22158396e-01 -5.68569601e-01
-3.08398843e-01 -8.24185163e-02 -4.58537608e-01 5.18514991e-01
9.62543786e-01 2.01791912e-01 7.98878610e-01 1.67626202e-01
7.69874275e-01 -8.75778943e-02 -9.47177708e-01 1.37169468e+00
-5.90596534e-02 -1.45504367e+00 -8.92137215e-02 -2.84031123e-01
3.46180528e-01 5.25931939e-02 -8.06628540e-02 -2.05916926e-01
-2.90042251e-01 -7.04980075e-01 9.42269146e-01 7.67081439e-01
7.62692809e-01 -6.25361085e-01 5.72101891e-01 2.86499530e-01
-1.25864863e+00 -8.09415057e-02 -2.36302540e-01 -1.78981617e-01
6.51045263e-01 8.64267349e-01 2.31060814e-02 5.75818360e-01
1.32375920e+00 1.17656791e+00 -1.72211483e-01 1.04677761e+00
-1.72074094e-01 8.28362286e-01 -4.62426901e-01 4.59145010e-01
5.75387711e-03 -4.06242490e-01 7.37916052e-01 1.26838267e+00
6.91856861e-01 2.51417756e-01 -1.07924536e-01 4.43972141e-01
5.83949089e-02 -3.10303181e-01 -2.48467147e-01 1.74715966e-01
3.99056286e-01 1.02472556e+00 -3.99503052e-01 -3.20254713e-01
-7.04632878e-01 1.32654405e+00 -2.25980803e-01 5.39797187e-01
-9.72365379e-01 -2.09906772e-01 1.00365984e+00 8.59767348e-02
5.24021506e-01 -5.38763881e-01 -2.62732863e-01 -1.45972931e+00
4.00326341e-01 -9.66351032e-01 1.79527342e-01 -7.95103252e-01
-1.22921026e+00 4.90798622e-01 -3.80173683e-01 -1.37483716e+00
-3.62173170e-02 -3.08854222e-01 -4.18135732e-01 7.43612170e-01
-1.47974634e+00 -8.98521364e-01 -5.93420863e-01 9.62089002e-01
6.36341870e-01 1.21444233e-01 5.68406999e-01 8.64262462e-01
-4.98408496e-01 4.39093351e-01 3.54165077e-01 -1.01299994e-01
9.05366778e-01 -7.28537738e-01 4.94381815e-01 1.36877489e+00
-1.05121300e-01 5.31691432e-01 8.28472614e-01 -6.08017325e-01
-1.99992037e+00 -9.19557273e-01 5.39517045e-01 1.60971746e-01
5.85576415e-01 -4.10552084e-01 -1.05107903e+00 5.37887692e-01
4.54283565e-01 2.40365744e-01 2.89755821e-01 -4.28834200e-01
-3.24814707e-01 -5.13106465e-01 -1.01430202e+00 6.29939020e-01
1.25798249e+00 -7.22445548e-01 -1.44820422e-01 1.60361513e-01
7.33781040e-01 -4.62617725e-01 -8.84537578e-01 1.89156339e-01
6.14528835e-01 -1.37667477e+00 1.03861237e+00 1.71982899e-01
4.14913744e-01 -6.36221290e-01 -2.77201474e-01 -8.63800526e-01
-1.11601196e-01 -9.70667720e-01 -5.50238609e-01 1.27189827e+00
-1.10598661e-01 -4.64152128e-01 4.08283383e-01 1.33347437e-01
-1.56295314e-01 -4.22186911e-01 -1.14674377e+00 -4.78734821e-01
-6.05073273e-01 -9.58559990e-01 2.68385798e-01 5.30153275e-01
-3.13347608e-01 -1.84389025e-01 -7.69780397e-01 4.88284320e-01
8.50525200e-01 3.71694677e-02 5.09689152e-01 -5.60289145e-01
-5.56445181e-01 -1.63044125e-01 -4.26246405e-01 -1.49207151e+00
-1.70687303e-01 -4.79706675e-01 1.42766163e-01 -1.52100444e+00
3.11841294e-02 1.98008977e-02 -1.05344124e-01 2.61524379e-01
1.73648939e-01 4.22514230e-01 2.82806903e-01 2.59977728e-01
-4.10568535e-01 6.50858760e-01 9.96175408e-01 -8.28488618e-02
-3.97226438e-02 -1.64126858e-01 -1.28548265e-01 6.65862381e-01
5.41011512e-01 -4.24953341e-01 -2.90365279e-01 -1.09491122e+00
3.14369470e-01 4.55214024e-01 7.04798818e-01 -9.40259635e-01
7.15379357e-01 1.51376411e-01 2.07292795e-01 -5.10692060e-01
7.58895040e-01 -8.36283505e-01 4.17969376e-01 3.24441552e-01
1.10687301e-01 1.45934582e-01 1.50348499e-01 7.62268186e-01
-6.02988064e-01 4.22666930e-02 6.30016804e-01 6.72858581e-02
-6.28920615e-01 3.75590354e-01 -4.67748284e-01 -3.15532923e-01
6.37215674e-01 -1.15303226e-01 -2.03175828e-01 -7.11359739e-01
-7.74338245e-01 9.75580886e-02 5.25926709e-01 2.85858482e-01
1.04619253e+00 -1.12109423e+00 -8.31218302e-01 3.34858626e-01
-4.77440417e-01 -1.05514161e-01 6.96042895e-01 1.17669797e+00
-7.69594371e-01 -1.50815204e-01 2.62082424e-02 -7.58803785e-01
-1.42624545e+00 3.82174313e-01 1.90860987e-01 3.01191688e-01
-9.54696953e-01 8.81632507e-01 -7.56834820e-02 1.48206264e-01
5.25305450e-01 -2.60763705e-01 2.62267500e-01 7.23702833e-02
9.55518961e-01 5.44719398e-01 2.02830687e-01 -3.19248766e-01
-2.25033686e-01 4.36008006e-01 2.00926945e-01 -2.97383070e-01
1.50995469e+00 -5.11460423e-01 -5.08205831e-01 3.81575257e-01
1.15381753e+00 4.02279338e-03 -1.56301761e+00 -7.43000433e-02
-5.63856184e-01 -7.54167199e-01 4.62729931e-01 -3.86153370e-01
-1.35344458e+00 8.26485217e-01 7.79151082e-01 4.79732007e-02
1.80334532e+00 -4.21496361e-01 1.14745450e+00 -5.49343377e-02
2.92034119e-01 -8.81916702e-01 -2.76487052e-01 2.18753323e-01
9.45515275e-01 -1.01219928e+00 -8.99083097e-04 -4.60377097e-01
-2.73231745e-01 1.23818362e+00 1.23399660e-01 -2.31662486e-02
7.76751637e-01 6.56711876e-01 1.35941403e-02 2.67654181e-01
-9.30102289e-01 1.64972350e-01 -5.40216118e-02 3.10504377e-01
3.84815425e-01 -2.26308838e-01 -3.62794161e-01 5.57585299e-01
1.77733079e-01 2.95074165e-01 4.53142792e-01 1.01022649e+00
-1.09501138e-01 -1.01749778e+00 -5.35096109e-01 1.84354886e-01
-6.10723078e-01 -3.22527587e-01 3.33511233e-01 8.10357183e-02
6.27550036e-02 1.22610950e+00 1.62161231e-01 -3.55060250e-01
1.39914975e-01 -2.59285867e-01 5.12244582e-01 -9.58067272e-03
-6.81810021e-01 6.10060334e-01 -4.79362085e-02 -9.62103605e-01
-6.36265159e-01 -7.69305289e-01 -1.03117871e+00 -6.94938183e-01
8.15001950e-02 -1.35228574e-01 7.10660100e-01 6.47608995e-01
5.34846067e-01 5.08838534e-01 5.21371424e-01 -1.17187130e+00
-2.20691442e-01 -8.41376066e-01 -4.31995720e-01 2.98887074e-01
7.56464064e-01 -2.50207663e-01 -5.05160749e-01 4.32928562e-01] | [11.352167129516602, -2.104137420654297] |
a8e1308f-f234-4a6e-9440-57ff5d772946 | beyond-i-i-d-three-levels-of-generalization | 2011.07743 | null | https://arxiv.org/abs/2011.07743v6 | https://arxiv.org/pdf/2011.07743v6.pdf | Beyond I.I.D.: Three Levels of Generalization for Question Answering on Knowledge Bases | Existing studies on question answering on knowledge bases (KBQA) mainly operate with the standard i.i.d assumption, i.e., training distribution over questions is the same as the test distribution. However, i.i.d may be neither reasonably achievable nor desirable on large-scale KBs because 1) true user distribution is hard to capture and 2) randomly sample training examples from the enormous space would be highly data-inefficient. Instead, we suggest that KBQA models should have three levels of built-in generalization: i.i.d, compositional, and zero-shot. To facilitate the development of KBQA models with stronger generalization, we construct and release a new large-scale, high-quality dataset with 64,331 questions, GrailQA, and provide evaluation settings for all three levels of generalization. In addition, we propose a novel BERT-based KBQA model. The combination of our dataset and model enables us to thoroughly examine and demonstrate, for the first time, the key role of pre-trained contextual embeddings like BERT in the generalization of KBQA. | ['Yu Su', 'Xifeng Yan', 'Percy Liang', 'Brian Sadler', 'Michelle Vanni', 'Sue Kase', 'Yu Gu'] | 2020-11-16 | null | null | null | null | ['knowledge-base-question-answering'] | ['natural-language-processing'] | [-4.20914054e-01 -5.29202037e-02 -1.28111407e-01 -5.10021925e-01
-1.03867483e+00 -8.92039657e-01 3.32434326e-01 9.25541967e-02
-4.31242973e-01 9.49629486e-01 4.08092320e-01 -6.61075652e-01
-2.02755779e-01 -1.10965443e+00 -7.44331181e-01 -2.94536561e-01
3.90608847e-01 7.17307627e-01 5.85651457e-01 -6.18036509e-01
-2.29997322e-01 2.82006145e-01 -1.20295918e+00 3.37947130e-01
1.04602206e+00 9.76702869e-01 -3.00435200e-02 6.99508786e-01
-2.57944703e-01 9.88815427e-01 -7.01057255e-01 -8.28777134e-01
8.74410570e-02 -4.83736634e-01 -1.15832663e+00 -1.71739087e-01
6.25502467e-01 -6.42822444e-01 -6.20895445e-01 7.74550676e-01
4.43939418e-01 3.77072662e-01 7.47909665e-01 -1.25171173e+00
-1.58585203e+00 4.99163955e-01 1.70823801e-02 4.59393352e-01
3.53666216e-01 1.72745392e-01 1.60132802e+00 -8.42025459e-01
3.94815803e-01 1.07172275e+00 5.41105032e-01 8.03464472e-01
-1.17814600e+00 -4.15072709e-01 1.89409778e-02 4.65380937e-01
-1.31914079e+00 -2.58667976e-01 5.36657155e-01 -1.36528358e-01
7.85109520e-01 3.91549796e-01 4.79977369e-01 1.20594275e+00
-3.73211652e-01 1.02098548e+00 9.19849157e-01 -4.70718801e-01
4.07225281e-01 3.40209931e-01 3.97716135e-01 5.14866531e-01
3.88230920e-01 -3.91112655e-01 -4.04501230e-01 -4.56649721e-01
4.10532564e-01 -2.44421020e-01 -5.03128052e-01 -5.92732131e-01
-7.42498815e-01 1.01151514e+00 4.35535669e-01 1.23406306e-01
-1.85909003e-01 1.35503054e-01 1.98087811e-01 3.89193326e-01
1.72433496e-01 7.33017683e-01 -7.19342470e-01 -3.63821208e-01
-5.24489045e-01 5.90487301e-01 8.22673738e-01 1.31967258e+00
8.69424760e-01 -3.03596556e-01 -2.24253803e-01 9.65861559e-01
1.88261151e-01 4.97860461e-01 5.99974215e-01 -8.50379407e-01
5.32483339e-01 5.97154319e-01 3.93439591e-01 -5.52179277e-01
-1.92675114e-01 -2.38486528e-01 -3.97994637e-01 -5.94630361e-01
5.57989538e-01 -1.08691007e-01 -7.43373990e-01 1.78247774e+00
3.73337030e-01 -3.59365419e-02 3.41210455e-01 8.18572879e-01
9.59389567e-01 7.35218644e-01 9.30146724e-02 2.70932138e-01
1.46288490e+00 -1.04696321e+00 -6.94643974e-01 -1.95843503e-01
8.98581564e-01 -1.62228689e-01 1.86660528e+00 1.91339403e-01
-9.87756968e-01 -3.47105205e-01 -8.71742249e-01 -4.11004990e-01
-4.95561421e-01 -1.11662880e-01 5.66953599e-01 1.11608326e+00
-1.01761031e+00 -5.59514351e-02 -4.92983192e-01 -3.36577535e-01
4.21595812e-01 -2.08318800e-01 -3.71189564e-01 -5.49746156e-01
-1.61721075e+00 9.78625596e-01 2.78730899e-01 -7.20417425e-02
-9.72463787e-01 -8.92649889e-01 -7.43070364e-01 2.54050523e-01
4.02296811e-01 -7.86274910e-01 1.55851185e+00 -4.59118038e-01
-1.33017313e+00 5.02542019e-01 -1.73478529e-01 -3.85417700e-01
-1.19666690e-02 -4.31633592e-01 -4.56660777e-01 1.34700075e-01
-4.62183654e-02 5.06704867e-01 4.05354530e-01 -1.26924956e+00
-4.65951115e-01 -2.33449683e-01 6.30695522e-01 2.09773242e-01
-8.16165268e-01 -3.90644163e-01 -4.03976232e-01 -3.21929485e-01
-2.98061579e-01 -5.77494442e-01 1.40188023e-01 -1.59507200e-01
-2.35273000e-02 -5.68117142e-01 4.42024291e-01 -7.38096356e-01
1.59626174e+00 -2.07612538e+00 -2.27109998e-01 -5.80349304e-02
1.07063957e-01 3.48056525e-01 -2.78879285e-01 6.45705879e-01
2.89719105e-01 2.92698503e-01 -9.10853073e-02 -1.03939489e-01
2.80550420e-01 5.63301921e-01 -6.70966983e-01 1.79980826e-02
3.74660581e-01 1.22527158e+00 -1.14026070e+00 -3.64074349e-01
-3.76107544e-01 -8.50822926e-02 -9.91576850e-01 3.36655140e-01
-5.44140279e-01 -1.16784878e-01 -4.14595664e-01 6.20060921e-01
4.67709869e-01 -5.24430692e-01 6.07234985e-02 -1.63638890e-01
4.86687839e-01 7.06826448e-01 -8.76008153e-01 1.59621871e+00
-4.41920042e-01 4.92308408e-01 -3.26015234e-01 -8.89217198e-01
7.59347677e-01 3.38340998e-01 -8.93922225e-02 -6.95012331e-01
-3.15523684e-01 2.22699761e-01 1.63905770e-01 -6.35029972e-01
8.73379409e-01 -4.60097164e-01 -1.17365897e-01 5.70458889e-01
3.74958843e-01 -4.72276777e-01 1.98035568e-01 4.46974188e-01
1.22695208e+00 -2.63756752e-01 9.97512043e-02 -2.63340801e-01
1.97092384e-01 5.95090054e-02 4.05407101e-01 1.00574028e+00
-4.58457232e-01 3.86655748e-01 5.49525261e-01 -1.55576244e-01
-8.88799071e-01 -1.37915671e+00 -2.28419811e-01 1.35881400e+00
3.14207338e-02 -3.97276878e-01 -5.84739387e-01 -1.05295765e+00
1.26766711e-01 1.10791612e+00 -7.65837252e-01 -3.84825081e-01
-2.82825738e-01 -4.64698762e-01 8.41126561e-01 7.35945880e-01
2.84209490e-01 -7.64251411e-01 -2.01445907e-01 1.92797303e-01
-3.19919348e-01 -9.83737409e-01 -3.21805775e-01 -6.81776628e-02
-7.26534367e-01 -9.59958971e-01 -4.88632023e-01 -6.13340378e-01
3.94184291e-01 4.39450830e-01 1.64295065e+00 1.90833788e-02
3.58493179e-02 8.27180147e-01 -8.25096071e-01 -4.25195992e-01
-2.07193464e-01 1.58585936e-01 -9.26879868e-02 -4.17774111e-01
7.61120439e-01 -4.08591509e-01 -6.37181163e-01 3.20557892e-01
-1.20669770e+00 -3.56439024e-01 5.20824134e-01 1.07814467e+00
1.86348811e-01 -4.75188484e-03 1.23009479e+00 -8.39183390e-01
9.90516722e-01 -8.06041420e-01 -2.23535061e-01 7.01017082e-01
-3.45316023e-01 1.24275409e-01 6.27415359e-01 -5.61544538e-01
-1.14138830e+00 -7.56927788e-01 -4.37667221e-01 -1.73048213e-01
-4.57791984e-02 7.08871841e-01 -2.42193207e-01 2.73088187e-01
1.03999674e+00 2.60831416e-01 -2.48647094e-01 -3.78000319e-01
9.60824311e-01 9.27633345e-01 2.75259882e-01 -1.01471996e+00
5.80576181e-01 3.66390079e-01 -6.92878366e-01 -9.55521166e-01
-1.19589925e+00 -6.29861772e-01 -2.16426164e-01 1.55659989e-01
5.53034723e-01 -8.46772313e-01 -3.96562397e-01 1.14404850e-01
-9.50485647e-01 -6.02721870e-01 -6.16598070e-01 3.62497181e-01
-4.11934763e-01 4.00058508e-01 -5.85051119e-01 -8.15508604e-01
-1.89717889e-01 -8.14389586e-01 6.28229082e-01 2.05124319e-01
-2.82134563e-01 -1.10538638e+00 3.15343320e-01 8.53163838e-01
6.18518412e-01 -5.29503286e-01 1.35638893e+00 -1.02899790e+00
-5.07300019e-01 -8.70563164e-02 -2.02839598e-01 7.88974941e-01
3.61772358e-01 -2.10631177e-01 -1.02284479e+00 -1.31923243e-01
-4.06945795e-02 -1.13344169e+00 6.04836285e-01 -1.24481045e-01
1.27051020e+00 -3.16015631e-01 2.05644563e-01 1.46563590e-01
1.19453251e+00 3.93095277e-02 7.02479362e-01 2.41786033e-01
3.97660434e-01 4.18426245e-01 5.27138174e-01 3.11926365e-01
8.52504611e-01 3.98979217e-01 2.78940588e-01 4.07112956e-01
2.09482424e-02 -4.74511623e-01 2.20019728e-01 1.11225200e+00
2.82950610e-01 -4.43599522e-01 -1.07103252e+00 1.07398450e+00
-1.36854160e+00 -7.46060371e-01 2.01449156e-01 1.95170212e+00
1.36820054e+00 -2.88795866e-02 5.68508990e-02 -1.72648683e-01
2.00481385e-01 -7.50906765e-02 -4.68786031e-01 -3.28852504e-01
-6.87529445e-02 4.70614284e-01 1.19724255e-02 3.83585632e-01
-7.08132505e-01 9.11679745e-01 6.09704924e+00 8.38008463e-01
-6.52440727e-01 2.59056300e-01 3.84539187e-01 -9.73692089e-02
-8.29691172e-01 -2.26388406e-02 -8.82481396e-01 4.35308158e-01
1.06684697e+00 -9.86903608e-02 3.30508947e-01 9.51109588e-01
-6.08805120e-01 1.17275611e-01 -1.14509153e+00 7.10014224e-01
4.83004302e-02 -1.24640763e+00 3.89094591e-01 -2.21649542e-01
8.16062033e-01 -4.51282375e-02 8.82962495e-02 9.85265791e-01
5.53569138e-01 -1.01501477e+00 5.85343361e-01 2.90247798e-01
6.31876230e-01 -5.63086331e-01 7.88866401e-01 6.14498079e-01
-6.39306426e-01 -1.30062029e-01 -7.42067933e-01 5.60396910e-02
9.95667204e-02 3.50023150e-01 -8.26652169e-01 6.11107230e-01
8.34338784e-01 1.72412500e-01 -6.05499268e-01 7.47977316e-01
-3.27887595e-01 1.07197750e+00 -2.91740894e-01 -4.77645814e-01
3.50311786e-01 -9.60096270e-02 -5.85640268e-03 1.05062795e+00
3.09796870e-01 3.96948159e-01 -2.71685958e-01 7.03779638e-01
-3.53201866e-01 7.51946419e-02 -3.68863076e-01 -2.84028143e-01
7.36968160e-01 8.70257318e-01 6.02904819e-02 -4.00420547e-01
-8.45133901e-01 7.65959799e-01 8.56024921e-01 5.36818206e-01
-7.91809082e-01 -3.77361327e-01 8.24654400e-01 -7.42660463e-02
5.39483488e-01 -1.80986539e-01 3.42696868e-02 -1.30418420e+00
2.41902024e-01 -1.08044767e+00 6.34205878e-01 -8.15520644e-01
-1.89820969e+00 3.37724596e-01 2.56639137e-03 -6.41188323e-01
-1.46505669e-01 -8.55135798e-01 -2.70006269e-01 9.00382519e-01
-1.58084488e+00 -1.00964510e+00 -2.48670146e-01 7.53453434e-01
1.98865116e-01 1.01114012e-01 1.05754280e+00 3.47771227e-01
-3.49080265e-01 8.84261072e-01 3.05591494e-01 2.78470993e-01
9.69613969e-01 -1.47911286e+00 3.16779226e-01 4.02623832e-01
3.50966811e-01 1.04181147e+00 4.55292195e-01 -3.06711614e-01
-1.53848922e+00 -1.03550959e+00 7.01547503e-01 -1.11461067e+00
1.02755654e+00 -4.88001496e-01 -1.39995193e+00 9.01938021e-01
8.97849128e-02 1.14758432e-01 1.15908539e+00 5.61799288e-01
-7.07410455e-01 -2.91992128e-01 -1.01445341e+00 7.35642195e-01
7.34058559e-01 -8.90977144e-01 -1.12039292e+00 3.12637389e-01
9.86551523e-01 -1.94796190e-01 -1.10848963e+00 2.79441655e-01
3.38449001e-01 -7.94241965e-01 9.33138609e-01 -1.18001711e+00
3.25112015e-01 -1.07986413e-01 -4.83656794e-01 -1.35926998e+00
-2.74243444e-01 -1.46287680e-01 -5.16749918e-01 1.26313686e+00
6.96054041e-01 -6.13861084e-01 8.67339492e-01 8.75133097e-01
-2.29950964e-01 -1.07361233e+00 -1.05042958e+00 -9.53854740e-01
6.78584099e-01 -6.25228584e-01 9.64280665e-01 9.72562075e-01
5.59455156e-02 4.96469617e-01 -6.87655658e-02 2.93595046e-01
-7.34885549e-03 -1.89370662e-02 9.55125511e-01 -7.53183842e-01
-5.49948871e-01 -7.03947544e-02 -3.02992910e-01 -1.74489009e+00
2.44125072e-02 -6.20003223e-01 3.69977169e-02 -1.63214767e+00
1.80000126e-01 -8.51918757e-01 -3.85598630e-01 2.97185868e-01
-4.89796162e-01 -1.03318378e-01 -3.10989648e-01 -2.79460922e-02
-7.64076412e-01 1.00831211e+00 1.24121535e+00 4.38730989e-04
1.26697704e-01 -2.53176510e-01 -1.00929964e+00 2.51514494e-01
9.52412367e-01 -1.27881005e-01 -8.75648618e-01 -5.98305762e-01
4.87935036e-01 -1.22323036e-01 2.94128805e-01 -5.38195491e-01
1.29360080e-01 -3.11529875e-01 -1.07481070e-01 -3.11814457e-01
5.68834186e-01 -6.84957445e-01 -5.32898664e-01 -3.00564934e-02
-3.49565923e-01 -6.05695210e-02 2.28471801e-01 6.69255435e-01
-3.71634752e-01 -5.24469793e-01 4.08966005e-01 -8.58572498e-03
-6.95494354e-01 1.94289595e-01 -4.73179221e-02 8.63294184e-01
6.80061340e-01 8.66405591e-02 -8.71802509e-01 -6.57177091e-01
-2.19887093e-01 3.86682570e-01 3.21814239e-01 3.25043678e-01
5.44830203e-01 -1.47127855e+00 -6.69933558e-01 -2.47028414e-02
6.60108566e-01 2.73299187e-01 4.98165935e-01 4.96342987e-01
-5.75775564e-01 5.83753824e-01 1.68464527e-01 -2.57746309e-01
-7.22601235e-01 6.29304230e-01 1.29032910e-01 -2.68988103e-01
-1.86603755e-01 1.14846754e+00 2.32206613e-01 -8.34972024e-01
3.04213986e-02 -5.58081627e-01 5.70658855e-02 -9.84535590e-02
6.40050590e-01 1.89021468e-01 2.02066198e-01 -9.75006372e-02
-1.01613514e-01 -1.98773354e-01 -3.11627418e-01 -2.41156831e-01
1.14293981e+00 -8.56572203e-03 -1.75739259e-01 6.08698189e-01
1.01139295e+00 1.10975206e-01 -9.86451983e-01 -4.96605396e-01
1.17509039e-02 -4.43602473e-01 -2.92684048e-01 -9.09402907e-01
-5.01592219e-01 1.11060774e+00 -2.85867639e-02 4.58636373e-01
9.25602794e-01 2.83538669e-01 9.14629817e-01 8.83815587e-01
4.40205306e-01 -1.04203451e+00 3.39503855e-01 6.92156136e-01
8.85725200e-01 -9.97320652e-01 -2.07684115e-01 -5.01102470e-02
-7.56632268e-01 7.67126083e-01 1.01362181e+00 1.52374446e-01
6.43312633e-01 -3.61293286e-01 1.70663014e-01 -1.94782898e-01
-9.54559445e-01 -1.35165706e-01 3.68651181e-01 6.04251623e-01
3.96799445e-01 3.12508531e-02 6.94922656e-02 9.62148726e-01
-4.65368330e-01 -1.16455719e-01 4.94206846e-01 9.68769491e-01
-5.12488723e-01 -9.18524325e-01 -1.02110595e-01 6.36318862e-01
-2.84866780e-01 -2.78647780e-01 -2.27406964e-01 9.39703465e-01
-2.66962647e-01 1.02395260e+00 -9.64848548e-02 -2.52793610e-01
5.45186043e-01 5.63535154e-01 4.89600450e-01 -8.40768099e-01
-1.82062671e-01 -9.01587844e-01 1.87481612e-01 -1.05685338e-01
-7.44465515e-02 -1.90858439e-01 -7.59771168e-01 -2.61143625e-01
-6.53574944e-01 5.18761933e-01 3.04836899e-01 8.39102626e-01
4.97572958e-01 1.70813918e-01 2.25148067e-01 2.22763985e-01
-1.25772917e+00 -1.25056362e+00 -6.91773534e-01 6.62876368e-01
3.30021709e-01 -5.37900865e-01 -4.39397216e-01 -1.49719313e-01] | [11.03396224975586, 7.977259159088135] |
4223ce02-66a8-4a5e-9fcd-00f1f37de439 | efficient-skill-acquisition-for-complex | 2303.03365 | null | https://arxiv.org/abs/2303.03365v1 | https://arxiv.org/pdf/2303.03365v1.pdf | Efficient Skill Acquisition for Complex Manipulation Tasks in Obstructed Environments | Data efficiency in robotic skill acquisition is crucial for operating robots in varied small-batch assembly settings. To operate in such environments, robots must have robust obstacle avoidance and versatile goal conditioning acquired from only a few simple demonstrations. Existing approaches, however, fall short of these requirements. Deep reinforcement learning (RL) enables a robot to learn complex manipulation tasks but is often limited to small task spaces in the real world due to sample inefficiency and safety concerns. Motion planning (MP) can generate collision-free paths in obstructed environments, but cannot solve complex manipulation tasks and requires goal states often specified by a user or object-specific pose estimator. In this work, we propose a system for efficient skill acquisition that leverages an object-centric generative model (OCGM) for versatile goal identification to specify a goal for MP combined with RL to solve complex manipulation tasks in obstructed environments. Specifically, OCGM enables one-shot target object identification and re-identification in new scenes, allowing MP to guide the robot to the target object while avoiding obstacles. This is combined with a skill transition network, which bridges the gap between terminal states of MP and feasible start states of a sample-efficient RL policy. The experiments demonstrate that our OCGM-based one-shot goal identification provides competitive accuracy to other baseline approaches and that our modular framework outperforms competitive baselines, including a state-of-the-art RL algorithm, by a significant margin for complex manipulation tasks in obstructed environments. | ['Ingmar Posner', 'Jack Collins', 'Jun Yamada'] | 2023-03-06 | null | null | null | null | ['motion-planning'] | ['robots'] | [ 3.02103251e-01 -1.51426400e-04 -2.29666889e-01 7.49840811e-02
-9.65020478e-01 -7.01242387e-01 3.42204034e-01 -1.97188154e-01
-5.87966323e-01 6.72769010e-01 -2.87122995e-01 -1.55090213e-01
-4.65995014e-01 -4.10289884e-01 -9.68556046e-01 -5.82985163e-01
-2.79420733e-01 8.52332890e-01 2.13721409e-01 -5.20545304e-01
3.27349424e-01 6.53193116e-01 -1.59579957e+00 -2.68096834e-01
9.77781951e-01 7.05278695e-01 1.13817656e+00 8.79504085e-01
2.85994858e-01 7.67387867e-01 -3.26133400e-01 4.62156117e-01
6.32032692e-01 -1.81018695e-01 -7.93433487e-01 3.44351307e-02
4.86226939e-02 -7.05050170e-01 -4.04343873e-01 8.11993241e-01
4.62096810e-01 6.39333248e-01 5.67893326e-01 -1.54603457e+00
-1.53359354e-01 4.10886049e-01 -2.98873603e-01 -1.36962771e-01
5.05515695e-01 9.62496996e-01 7.35496521e-01 -5.65241516e-01
6.12133741e-01 1.27396083e+00 2.35335514e-01 5.53110480e-01
-1.11898005e+00 -5.48773110e-01 3.04754317e-01 -1.23955803e-02
-1.09374380e+00 -2.74040431e-01 4.95595843e-01 -3.93749863e-01
1.26805925e+00 -2.69669771e-01 6.90695107e-01 1.43151748e+00
6.32667959e-01 7.74138927e-01 8.44374061e-01 1.00436322e-02
3.80552769e-01 -4.79197651e-01 -3.99664283e-01 9.84870136e-01
2.30381936e-01 4.38378155e-01 -5.40916443e-01 2.56223083e-01
1.08307505e+00 7.73114488e-02 -4.19568531e-02 -1.07296562e+00
-1.44431889e+00 7.64467359e-01 5.73656023e-01 -1.22522213e-01
-5.26880026e-01 5.18098533e-01 1.88874722e-01 4.47461784e-01
-3.48844737e-01 1.12625980e+00 -5.96810997e-01 -5.06937444e-01
-4.22319561e-01 7.19216764e-01 7.63285160e-01 1.62155128e+00
7.25512207e-01 2.05058932e-01 -2.53305525e-01 4.82279718e-01
1.41570970e-01 6.64611936e-01 1.08833469e-01 -1.43555307e+00
7.73216426e-01 4.36920255e-01 4.21866447e-01 -4.84095007e-01
-7.32473195e-01 -3.74410272e-01 -2.28008091e-01 8.63920867e-01
4.59969938e-01 -2.35316053e-01 -1.12646592e+00 1.82451451e+00
4.62191343e-01 -7.79757425e-02 1.90528512e-01 1.02747047e+00
2.66492814e-01 4.01865870e-01 6.52862489e-02 1.39307007e-01
9.91283655e-01 -1.22034490e+00 -4.00278836e-01 -7.63083041e-01
5.65579593e-01 -3.96375805e-01 1.34462297e+00 6.55218422e-01
-1.03885770e+00 -5.59677303e-01 -1.08023477e+00 1.21158659e-02
3.24996784e-02 7.56012425e-02 8.83250177e-01 1.22303687e-01
-7.91201293e-01 9.04558301e-01 -1.25641572e+00 -2.99117714e-01
4.72524256e-01 6.14183128e-01 -5.34396112e-01 -3.22503030e-01
-5.37275732e-01 1.33524024e+00 5.19136965e-01 5.65545969e-02
-1.88437808e+00 -5.33355176e-01 -1.39066339e+00 -4.05791886e-02
1.13084531e+00 -7.53966331e-01 1.49318147e+00 -1.36194870e-01
-1.91919422e+00 2.80527502e-01 3.09760779e-01 -3.94587576e-01
5.18564045e-01 -6.75210297e-01 3.20349693e-01 2.28791773e-01
3.80802989e-01 8.69306326e-01 1.01061261e+00 -1.40805829e+00
-6.92429662e-01 -2.21723437e-01 3.33561629e-01 6.63077056e-01
3.97946298e-01 -4.92615759e-01 -4.61440265e-01 -1.04841270e-01
1.77813515e-01 -1.27428794e+00 -6.83136284e-01 -4.44341674e-02
-3.33539605e-01 -1.57073855e-01 6.14165127e-01 -3.46246183e-01
5.27677834e-01 -1.86609650e+00 6.89783275e-01 -5.41955717e-02
-4.74876314e-02 -2.15886161e-01 -4.31954086e-01 5.03610194e-01
4.23806190e-01 -4.87713486e-01 -5.60490526e-02 -4.81010705e-01
2.93491960e-01 3.74329299e-01 -2.45070800e-01 3.77311349e-01
3.19127649e-01 1.15106380e+00 -1.32725966e+00 -3.40926260e-01
4.68976796e-01 -5.79057075e-02 -6.75562203e-01 4.54678535e-01
-4.94495213e-01 8.83507669e-01 -5.68338335e-01 8.00427616e-01
1.64485678e-01 4.69854921e-02 2.35604376e-01 1.96570128e-01
-1.26359671e-01 3.02862197e-01 -1.17755020e+00 2.44941044e+00
-6.98323727e-01 1.15143076e-01 3.02666068e-01 -8.20662975e-01
8.62196326e-01 -1.12591632e-01 5.01253605e-01 -4.77604747e-01
3.26801866e-01 3.07512403e-01 1.18614964e-01 -5.28508306e-01
6.89458966e-01 1.31043077e-01 -5.60048819e-01 2.14381948e-01
3.17468762e-01 -1.06308949e+00 2.92347997e-01 5.29920794e-02
1.42011321e+00 1.02788472e+00 1.84770003e-01 -1.96923334e-02
-9.91587639e-02 4.68620658e-01 4.42626685e-01 1.18697381e+00
-3.08915257e-01 4.14752275e-01 1.24247611e-01 -4.24443968e-02
-1.06055009e+00 -1.32377720e+00 4.84694093e-01 1.09791088e+00
6.75154150e-01 4.36550379e-02 -4.93940353e-01 -6.94585800e-01
1.71207964e-01 8.66711915e-01 -2.68329710e-01 -4.78959143e-01
-9.49992657e-01 5.72608858e-02 1.04256086e-01 6.26192927e-01
4.22894925e-01 -1.41622508e+00 -1.21320951e+00 3.92191172e-01
-6.44075647e-02 -1.20363820e+00 -4.06570882e-01 6.52101040e-01
-7.52772212e-01 -1.22048724e+00 -4.19216454e-01 -9.80235457e-01
6.54749930e-01 3.59662205e-01 9.00739014e-01 -2.79565364e-01
-4.41921055e-01 6.20354652e-01 -3.43757391e-01 -3.69107515e-01
-4.87423480e-01 1.35831237e-01 4.61510628e-01 -8.42705488e-01
-8.31840411e-02 -4.44278896e-01 -5.51725686e-01 2.78749287e-01
-4.17424917e-01 1.57828689e-01 1.10573852e+00 8.10388863e-01
4.95444119e-01 8.02994892e-02 4.17204142e-01 -1.09944567e-01
5.29041469e-01 -4.13570672e-01 -8.08174670e-01 -7.27717951e-02
-4.25293297e-01 3.73711199e-01 5.57527184e-01 -7.34239459e-01
-9.27365601e-01 3.84422749e-01 1.09313607e-01 -6.28034472e-01
-1.64235130e-01 3.05921376e-01 -9.86741483e-02 -7.62636587e-02
6.71209514e-01 1.76171482e-01 3.18824798e-01 -2.53008425e-01
3.37820739e-01 9.29455832e-02 6.96111500e-01 -9.85970080e-01
9.26189959e-01 1.58046320e-01 1.85391262e-01 -3.84861469e-01
-5.80949545e-01 -4.55007970e-01 -6.71083987e-01 -2.68150449e-01
7.52751291e-01 -9.02880549e-01 -1.14400387e+00 4.24312830e-01
-7.32646644e-01 -1.13169682e+00 -4.71165597e-01 7.11163700e-01
-1.33545470e+00 1.10495269e-01 -5.33032775e-01 -9.49023783e-01
5.69715165e-02 -1.50886011e+00 1.33428812e+00 3.21816117e-01
-2.47283727e-01 -2.84732997e-01 -3.16976368e-01 2.01370090e-01
2.49836683e-01 3.45580965e-01 4.90858346e-01 -2.07586661e-01
-9.11793351e-01 -7.35475272e-02 2.37835675e-01 -9.13428217e-02
1.46061271e-01 -6.28048301e-01 -2.53108621e-01 -7.17951059e-01
-2.02540547e-01 -9.05484378e-01 4.94332850e-01 3.21081698e-01
7.92529345e-01 4.32282574e-02 -4.90178317e-01 5.15214026e-01
1.15821421e+00 2.53515810e-01 3.48607868e-01 6.07573867e-01
5.55274904e-01 5.03144562e-01 1.47538435e+00 4.80202585e-01
3.74114186e-01 7.31801867e-01 8.94972146e-01 3.49677563e-01
6.62134960e-02 -5.38761616e-01 4.64652240e-01 2.18361869e-01
1.06510939e-02 -5.93686216e-02 -7.63187170e-01 5.80061674e-01
-1.98152316e+00 -6.73405290e-01 2.42632508e-01 2.13298202e+00
7.27179646e-01 4.55710590e-01 6.95212334e-02 -1.94461182e-01
6.83473274e-02 -1.66378081e-01 -1.11623228e+00 -8.20681527e-02
4.38146561e-01 9.88306329e-02 6.89301848e-01 5.08602142e-01
-1.01628804e+00 1.27503026e+00 5.73824358e+00 6.25059605e-01
-6.83818519e-01 -1.89030766e-01 -1.83963239e-01 -2.66705722e-01
2.44509935e-01 -7.15634227e-02 -9.13471222e-01 8.97151157e-02
4.78798181e-01 -8.18943139e-03 7.92650878e-01 1.36505604e+00
1.49182156e-01 -5.26178539e-01 -1.44636488e+00 9.32391286e-01
-2.49373928e-01 -8.45617652e-01 -4.64096159e-01 -1.15632750e-02
5.37262857e-01 6.26266822e-02 2.49880135e-01 9.37833309e-01
1.03210127e+00 -1.13945580e+00 8.89959991e-01 4.00852263e-01
7.21849561e-01 -7.05695450e-01 4.58123624e-01 8.29143226e-01
-1.19463754e+00 -6.14766181e-01 -2.68633544e-01 -3.00270021e-01
3.83516848e-01 -2.88025528e-01 -1.20111132e+00 6.20585799e-01
6.24614894e-01 2.87947536e-01 8.56388435e-02 7.82243550e-01
-5.41273475e-01 2.28235554e-02 -3.69877309e-01 -2.40491807e-01
4.99678075e-01 -1.81245640e-01 6.99785411e-01 5.63755095e-01
3.74768347e-01 4.20584381e-02 6.60237849e-01 1.02384448e+00
4.58531886e-01 -4.18363214e-01 -8.42741072e-01 -9.67358649e-02
5.59021533e-01 1.13617790e+00 -6.06740832e-01 2.92406790e-03
2.34394282e-01 1.18450105e+00 6.60536587e-01 2.88817555e-01
-8.99092793e-01 -2.42508069e-01 8.82486641e-01 -2.72777826e-01
3.47951353e-01 -9.97782946e-01 -1.12283096e-01 -7.72474825e-01
-3.02203801e-02 -1.02140689e+00 4.72433455e-02 -9.01896954e-01
-7.90918767e-01 1.48787677e-01 2.69033045e-01 -1.27782226e+00
-5.76982319e-01 -7.86956906e-01 -1.88950196e-01 7.23074615e-01
-1.44332349e+00 -1.10482645e+00 -4.96250063e-01 4.57083464e-01
1.01080632e+00 -1.62290037e-01 7.61199355e-01 -3.05755228e-01
-2.75391757e-01 3.11210364e-01 -2.18442261e-01 -2.09291860e-01
5.30593991e-01 -1.30485201e+00 3.53344500e-01 6.54453278e-01
-2.09518477e-01 6.42085314e-01 8.73069704e-01 -9.75531995e-01
-2.03594875e+00 -8.28844905e-01 -1.43364770e-02 -7.03562677e-01
4.88825232e-01 -4.88574088e-01 -4.64245886e-01 7.62248993e-01
-3.07919562e-01 -1.88777328e-01 -2.86626630e-02 -9.76041108e-02
1.38373405e-01 1.96048409e-01 -1.07049108e+00 1.06986547e+00
1.47471070e+00 -7.49184787e-02 -7.49878109e-01 4.12686199e-01
9.38086987e-01 -1.04990017e+00 -5.98657846e-01 5.94773352e-01
6.54359877e-01 -4.60079372e-01 1.11788464e+00 -7.98899770e-01
5.55199385e-01 -4.89805192e-01 -3.71399708e-02 -1.49776912e+00
-3.98514599e-01 -8.48912597e-01 -2.55812973e-01 3.83133382e-01
2.25021333e-01 -3.71838629e-01 7.74504721e-01 4.51734811e-01
-6.26843691e-01 -7.25746870e-01 -7.88289189e-01 -1.19360018e+00
-9.08444822e-02 -4.92300928e-01 4.06454295e-01 2.92741776e-01
2.45474465e-02 2.56741852e-01 -4.37060624e-01 4.44119096e-01
6.88277781e-01 1.76525831e-01 1.35739148e+00 -9.03104544e-01
-4.47207600e-01 -3.09418917e-01 -1.18025661e-01 -1.43554473e+00
3.00481021e-01 -5.50854206e-01 9.80459034e-01 -1.75407028e+00
-1.09075703e-01 -8.51191700e-01 -1.77547988e-02 4.57818091e-01
-2.05966353e-01 -2.43645117e-01 4.15520579e-01 1.22755386e-01
-8.48818719e-01 6.70171678e-01 1.71909225e+00 -6.95294365e-02
-5.94808817e-01 -1.33665018e-02 -3.75215828e-01 5.76093853e-01
8.62409353e-01 -3.07921290e-01 -6.58156574e-01 -4.20790821e-01
1.32661656e-01 3.31426322e-01 2.93804497e-01 -1.31961107e+00
2.68912196e-01 -4.92767423e-01 2.52193630e-01 -5.32552242e-01
6.74462974e-01 -8.85788560e-01 1.92514199e-04 8.77907753e-01
-2.48574838e-01 -6.60671741e-02 2.59241968e-01 8.28883886e-01
4.17845815e-01 -3.19062293e-01 4.99800712e-01 -3.33591789e-01
-1.16864133e+00 3.79342318e-01 -4.32720453e-01 -8.02056305e-03
1.36170578e+00 -3.68738353e-01 -2.05481082e-01 -2.97281832e-01
-7.41154432e-01 7.81794786e-01 5.11249542e-01 7.10798740e-01
6.88439727e-01 -1.06069779e+00 -4.13332373e-01 1.47181571e-01
1.27988666e-01 5.84371865e-01 2.09986478e-01 8.29378247e-01
-4.58288193e-01 2.26694331e-01 -4.55523282e-01 -6.71015620e-01
-9.05549109e-01 7.00501919e-01 8.82411823e-02 -4.24164832e-01
-8.33292365e-01 1.02035880e+00 2.79852331e-01 -8.23955417e-01
2.33657062e-01 -5.75222731e-01 -9.58961062e-03 -7.07773387e-01
1.73552334e-01 2.10958779e-01 -3.25959533e-01 -1.26076341e-01
-1.32282868e-01 2.91640222e-01 -1.27980777e-04 -1.77101821e-01
1.25607204e+00 -9.20636132e-02 5.36392570e-01 4.23654944e-01
5.16747594e-01 -3.38745743e-01 -2.18506312e+00 -1.98775367e-03
-2.01921701e-01 -5.00491619e-01 -1.91518605e-01 -8.74277234e-01
-2.96464354e-01 4.98998135e-01 3.07415754e-01 -3.56439292e-01
6.17137194e-01 9.21454467e-03 7.73206353e-01 9.36862230e-01
1.19710529e+00 -1.39773417e+00 8.65233123e-01 9.25401509e-01
1.14239824e+00 -1.46627402e+00 -1.28800362e-01 -3.62853557e-01
-8.92948627e-01 8.86666954e-01 1.11202240e+00 -3.01367879e-01
-5.52106127e-02 4.48885828e-01 -1.79651946e-01 -2.07409754e-01
-6.22160554e-01 -4.75802124e-01 1.28907010e-01 9.30001199e-01
-3.83165747e-01 -1.09530194e-02 3.09655696e-01 4.06815052e-01
-4.17416781e-01 -2.64496028e-01 2.44948164e-01 1.55375612e+00
-8.62021863e-01 -8.46103251e-01 -2.94897199e-01 3.73793155e-01
1.58244193e-01 1.62534028e-01 2.78013796e-02 1.07548714e+00
-1.57116801e-01 9.83858228e-01 -5.18198274e-02 -3.34421307e-01
4.73543108e-01 -5.94235361e-02 1.07696307e+00 -1.07635283e+00
-3.80284727e-01 -7.93133676e-02 1.37985080e-01 -1.13729525e+00
1.34273782e-01 -7.57632971e-01 -1.72420955e+00 6.90673068e-02
-3.28189611e-01 -2.42493719e-01 7.73392975e-01 1.01535487e+00
3.36028427e-01 8.47942710e-01 3.89382482e-01 -1.51593816e+00
-1.10961807e+00 -9.87485528e-01 -4.32369977e-01 2.25530118e-01
3.36824030e-01 -1.31437719e+00 -1.06713980e-01 -2.77163684e-01] | [4.5919599533081055, 0.8206328749656677] |
5cc086a8-9fb7-44cc-806a-e1a35077db86 | explaincpe-a-free-text-explanation-benchmark | 2305.12945 | null | https://arxiv.org/abs/2305.12945v1 | https://arxiv.org/pdf/2305.12945v1.pdf | ExplainCPE: A Free-text Explanation Benchmark of Chinese Pharmacist Examination | As ChatGPT and GPT-4 spearhead the development of Large Language Models (LLMs), more researchers are investigating their performance across various tasks. But more research needs to be done on the interpretability capabilities of LLMs, that is, the ability to generate reasons after an answer has been given. Existing explanation datasets are mostly English-language general knowledge questions, which leads to insufficient thematic and linguistic diversity. To address the language bias and lack of medical resources in generating rationales QA datasets, we present ExplainCPE (over 7k instances), a challenging medical benchmark in Simplified Chinese. We analyzed the errors of ChatGPT and GPT-4, pointing out the limitations of current LLMs in understanding text and computational reasoning. During the experiment, we also found that different LLMs have different preferences for in-context learning. ExplainCPE presents a significant challenge, but its potential for further investigation is promising, and it can be used to evaluate the ability of a model to generate explanations. AI safety and trustworthiness need more attention, and this work makes the first step to explore the medical interpretability of LLMs.The dataset is available at https://github.com/HITsz-TMG/ExplainCPE. | ['Min Zhang', 'Zhenran Xu', 'Baotian Hu', 'Jindi Yu', 'Dongfang Li'] | 2023-05-22 | null | null | null | null | ['general-knowledge'] | ['miscellaneous'] | [ 2.48348191e-01 9.66533124e-01 -4.33986187e-01 -6.70802474e-01
-1.03373015e+00 -3.22998255e-01 4.38996732e-01 3.02529067e-01
1.69003685e-03 9.25105810e-01 8.21805358e-01 -8.69572639e-01
-3.13050866e-01 -4.88699019e-01 -6.56679273e-01 -1.32921726e-01
3.22620362e-01 8.95954192e-01 -2.44454339e-01 -1.55932859e-01
2.11915225e-01 -1.00859396e-01 -8.57803285e-01 1.06650448e+00
1.34782505e+00 2.36084223e-01 6.52466193e-02 7.05889165e-01
-2.01032877e-01 1.20423877e+00 -6.19023800e-01 -8.76705766e-01
-4.30334091e-01 -8.59434068e-01 -1.44134879e+00 -1.51292861e-01
1.08068086e-01 -1.27574608e-01 2.03275830e-01 6.01135552e-01
5.53902805e-01 -2.47256085e-01 5.55983007e-01 -1.40904260e+00
-1.18041682e+00 1.22477055e+00 1.40874535e-01 1.96358696e-01
6.22535110e-01 2.54642785e-01 1.12237859e+00 -6.02433622e-01
7.19648421e-01 1.40131044e+00 7.55422354e-01 1.14203262e+00
-9.87044632e-01 -6.16040885e-01 4.37929071e-02 4.35951144e-01
-9.30373609e-01 -3.99425924e-01 2.03417227e-01 -2.73354381e-01
1.07733345e+00 5.99763155e-01 5.26491165e-01 1.44002163e+00
3.57226044e-01 1.00094879e+00 1.27050591e+00 -6.08447492e-01
1.26325920e-01 3.93453807e-01 3.59320551e-01 7.51776695e-01
5.20062447e-01 -1.41859621e-01 -6.46507621e-01 -4.71440077e-01
4.26830739e-01 -3.10677260e-01 -5.24586618e-01 2.27516949e-01
-1.39236808e+00 1.18031061e+00 5.23878515e-01 2.38051265e-01
-2.15083778e-01 -1.00770276e-02 1.38014495e-01 4.10959646e-02
2.20618650e-01 1.10774624e+00 -7.60598123e-01 -3.75235826e-01
-7.25038946e-01 3.86056423e-01 9.93603826e-01 9.94412780e-01
1.47304550e-01 -3.79754871e-01 -4.16272044e-01 6.10394776e-01
2.19156921e-01 4.22212988e-01 4.88884151e-01 -1.15490675e+00
6.64400160e-01 6.40853822e-01 -3.67847160e-02 -9.12530661e-01
-6.26153111e-01 -1.71816289e-01 -5.33633947e-01 -3.63628685e-01
4.53436583e-01 -3.31712902e-01 -7.26879179e-01 1.72779155e+00
-5.31970561e-02 -2.66635805e-01 4.74448681e-01 8.71972799e-01
1.33806324e+00 5.02678871e-01 4.25205708e-01 8.94259810e-02
1.65619528e+00 -1.00768614e+00 -9.23313141e-01 -6.56906307e-01
1.00314558e+00 -7.53827751e-01 1.38249397e+00 2.97680885e-01
-1.18870997e+00 -1.49496034e-01 -5.85126936e-01 -3.67392629e-01
-1.06931157e-01 1.65720508e-01 1.02064574e+00 5.90659738e-01
-8.49664569e-01 2.85366565e-01 -8.61443758e-01 -4.81617004e-01
4.88328785e-01 2.15876363e-02 -2.12857082e-01 -4.55341756e-01
-1.46297133e+00 1.21946633e+00 4.71321374e-01 7.10369870e-02
-4.75611627e-01 -8.25112879e-01 -9.83354211e-01 -5.35543337e-02
4.28377688e-01 -1.15804327e+00 1.48155427e+00 -8.48399162e-01
-9.86268580e-01 7.35764205e-01 -4.71573502e-01 -4.68287557e-01
4.85127866e-01 -2.07412064e-01 -4.93416697e-01 3.12923566e-02
5.71049929e-01 9.21487331e-01 7.49681965e-02 -1.07216752e+00
-4.45517302e-01 -7.82728866e-02 1.57326892e-01 1.72144026e-01
1.93306357e-01 -8.43585655e-02 -2.94204026e-01 -6.69828773e-01
3.24742012e-02 -1.10082698e+00 -3.25962305e-01 -4.47574824e-01
-7.23834336e-01 -2.57395297e-01 -3.27835185e-03 -8.20149064e-01
1.28779650e+00 -1.69415808e+00 -3.20018455e-02 -8.95446017e-02
1.26294538e-01 7.11262450e-02 -1.63281206e-02 5.63191354e-01
-1.98463500e-01 7.68516243e-01 -4.06860441e-01 -1.16163410e-01
-8.18596631e-02 5.12842536e-01 -4.39647704e-01 -2.60322273e-01
5.84500313e-01 1.32658744e+00 -1.01720691e+00 -7.21389472e-01
-2.13299945e-01 3.12460303e-01 -8.55819285e-01 -1.07724190e-01
-3.47158760e-01 5.20051003e-01 -7.23209143e-01 6.15907550e-01
3.01027983e-01 -7.56379247e-01 1.95898041e-01 9.43288580e-02
2.60372788e-01 7.99707115e-01 -7.54044771e-01 1.50303149e+00
-2.63248503e-01 6.04596078e-01 -5.17417371e-01 -4.71650511e-01
3.68430436e-01 6.12477243e-01 -1.32169053e-01 -5.66135764e-01
-1.22399420e-01 5.62657714e-01 4.40921903e-01 -8.96078169e-01
3.94370347e-01 -3.88938636e-01 -1.08676724e-01 5.83410621e-01
-3.24001223e-01 -3.88750076e-01 1.95231065e-01 4.88931865e-01
9.45612669e-01 -8.33756998e-02 4.63901550e-01 -3.26209128e-01
3.89130771e-01 6.61936045e-01 5.74658871e-01 9.40203726e-01
2.23354712e-01 7.43162930e-01 6.38874054e-01 -5.06653130e-01
-5.87156534e-01 -6.55824900e-01 -3.93147826e-01 5.01750886e-01
-2.01117665e-01 -7.41329670e-01 -8.09557855e-01 -8.89882624e-01
-1.82124093e-01 1.72942424e+00 -7.83558249e-01 -1.04414172e-01
-4.41802621e-01 -8.42348337e-01 8.70439172e-01 6.76746130e-01
2.13597149e-01 -1.39208543e+00 -8.51502419e-01 1.79132193e-01
-9.28216577e-01 -1.09015250e+00 -3.23061287e-01 -2.51198566e-04
-1.11213112e+00 -1.27583253e+00 -3.43052953e-01 -3.95060986e-01
8.52463305e-01 -1.19831927e-01 1.60700822e+00 5.83589435e-01
2.23146901e-02 5.32464862e-01 -5.18030167e-01 -9.50404406e-01
-8.92792225e-01 2.15495363e-01 -4.08810437e-01 -9.01209831e-01
7.03939736e-01 7.46850967e-02 -5.42348504e-01 2.97725528e-01
-9.29605603e-01 6.49989486e-01 6.50763512e-01 1.03905869e+00
4.42101955e-01 -2.32947409e-01 5.74389100e-01 -1.43527198e+00
1.10597718e+00 -5.97154438e-01 1.81855366e-01 4.72145289e-01
-8.00473750e-01 3.05445105e-01 2.96844244e-01 -1.93540715e-02
-1.16321278e+00 -4.71006274e-01 -3.22588563e-01 4.95565921e-01
-1.63852707e-01 7.70111978e-01 4.86938432e-02 5.16985476e-01
9.52391982e-01 -2.15123281e-01 -6.67812973e-02 -1.94241330e-01
4.11221594e-01 5.27082741e-01 2.08807647e-01 -7.89581358e-01
4.17399555e-01 2.49096498e-01 -5.22743225e-01 -3.82754385e-01
-1.13690615e+00 1.96847245e-01 -4.22313325e-02 2.32025608e-01
1.04931593e+00 -8.13097298e-01 -5.42757750e-01 -3.91246229e-01
-1.23793662e+00 -5.70004821e-01 -1.85989320e-01 6.26148820e-01
-4.58632708e-01 2.55235881e-01 -7.28999496e-01 -4.46996301e-01
-4.58056033e-01 -1.33960974e+00 9.18252707e-01 1.67078480e-01
-1.25031877e+00 -1.20301533e+00 -2.67792314e-01 1.05667639e+00
1.46426693e-01 8.83018374e-02 1.48653555e+00 -8.80489647e-01
-4.48197961e-01 5.69966994e-02 -1.06272101e-02 -1.79884788e-02
2.14712352e-01 -2.08836585e-01 -7.09364653e-01 2.68899351e-01
2.24239811e-01 -5.75166821e-01 6.16777897e-01 5.07358849e-01
1.36935818e+00 -6.73227310e-01 -4.58899736e-01 2.82432348e-01
1.00154626e+00 1.77983001e-01 5.66797733e-01 4.29054081e-01
5.64173639e-01 8.06453586e-01 7.23580599e-01 8.35963637e-02
8.36207628e-01 2.88735479e-01 6.00221567e-02 -1.14954233e-01
-1.95379093e-01 -4.19794649e-01 1.23132832e-01 7.26793706e-01
-1.29709234e-02 -3.83892059e-01 -1.37319243e+00 5.52057385e-01
-1.80005407e+00 -7.43886054e-01 -5.03787518e-01 1.66006839e+00
1.10927165e+00 2.50067841e-02 -4.80512738e-01 -6.47456199e-02
1.52394667e-01 -3.34423959e-01 -4.59775209e-01 -7.61415839e-01
-2.61686713e-01 9.29379836e-02 7.41110072e-02 7.01308548e-01
-4.58238959e-01 6.88514292e-01 6.25342512e+00 5.44191599e-01
-6.29379034e-01 7.73704750e-03 1.06203401e+00 2.03323126e-01
-1.13322341e+00 1.16727687e-01 -6.00679457e-01 3.64757508e-01
9.38481450e-01 -1.00711584e-01 1.48326322e-01 6.69692039e-01
3.38749349e-01 -1.24697171e-01 -1.38243020e+00 5.72232246e-01
2.16090113e-01 -1.46159184e+00 2.84463197e-01 -1.63689256e-01
7.00098038e-01 -2.45530337e-01 -8.60693827e-02 4.19963926e-01
3.38671535e-01 -1.45485759e+00 6.38494492e-01 2.52956748e-01
4.73680258e-01 -3.07774216e-01 9.91614223e-01 4.49150294e-01
-4.78975296e-01 -7.29415119e-02 -3.64793688e-01 -1.72996879e-01
2.59314626e-01 3.58432591e-01 -1.41973829e+00 5.90339184e-01
6.52793527e-01 5.29372156e-01 -7.98036337e-01 6.07330024e-01
-8.00063550e-01 8.63190413e-01 1.71449929e-01 -8.27032849e-02
3.09949458e-01 2.33182997e-01 2.01342180e-01 1.24368632e+00
3.97489905e-01 5.02418280e-01 -1.15747519e-01 1.05060041e+00
6.49281312e-03 4.65682261e-02 -3.16113859e-01 -2.94090718e-01
3.16837341e-01 9.64070559e-01 -8.46106470e-01 -4.85526592e-01
-3.47413957e-01 7.58344173e-01 2.16850311e-01 2.01697320e-01
-9.45740342e-01 2.85925537e-01 3.38470966e-01 6.72764182e-02
-2.99763441e-01 3.74239862e-01 -8.47686887e-01 -1.12111652e+00
-1.96777359e-02 -1.56957746e+00 8.57385755e-01 -1.26558018e+00
-1.28235555e+00 6.83926463e-01 1.31593555e-01 -8.31188858e-01
-4.75272983e-01 -5.79641759e-01 -3.21686804e-01 8.43123734e-01
-1.33656895e+00 -1.12331975e+00 -3.59568864e-01 3.43008697e-01
9.09554183e-01 1.42805308e-01 1.08829033e+00 -1.12183228e-01
-3.36950511e-01 4.78760183e-01 -5.88833451e-01 3.57970521e-02
9.06548798e-01 -1.28711176e+00 5.75724542e-01 5.46325803e-01
1.89831018e-01 1.02991867e+00 8.90995920e-01 -1.04764068e+00
-1.01694703e+00 -7.83471107e-01 1.53251934e+00 -1.30270338e+00
2.51760423e-01 7.62681738e-02 -1.15486491e+00 9.94584024e-01
5.12354851e-01 -6.88318729e-01 1.14608896e+00 3.99859846e-01
-6.68393597e-02 5.68374217e-01 -1.01471186e+00 8.82775307e-01
9.68927443e-01 -2.84634590e-01 -1.19409132e+00 6.34138167e-01
8.81650448e-01 -7.25827217e-01 -7.28052497e-01 2.78245866e-01
2.82241166e-01 -7.87990034e-01 8.04153383e-01 -9.98845816e-01
1.06412542e+00 -1.54069975e-01 2.60428548e-01 -1.43153656e+00
-3.75558406e-01 -3.00627261e-01 3.41180950e-01 1.06886756e+00
1.25830567e+00 -6.22413933e-01 6.32352710e-01 1.35795617e+00
-3.39888662e-01 -1.02374709e+00 -5.72347105e-01 -2.11740032e-01
2.89481252e-01 -6.77138925e-01 7.31689095e-01 1.17637146e+00
3.14866275e-01 4.35721666e-01 -2.31879517e-01 1.46558329e-01
1.62453234e-01 -7.01646060e-02 3.67903292e-01 -9.06608224e-01
-3.69927347e-01 -2.29585811e-01 3.92503500e-01 -5.20516813e-01
1.96196467e-01 -1.21971154e+00 -7.10630715e-02 -2.29891658e+00
4.38365191e-01 -4.40439105e-01 1.88854530e-01 9.11445081e-01
-7.68316865e-01 -2.27241024e-01 2.65514076e-01 1.17635623e-01
-3.02827299e-01 2.43817702e-01 1.32480466e+00 -2.02124212e-02
-1.72171786e-01 -7.12015526e-03 -1.49463975e+00 8.75482678e-01
9.42128539e-01 -7.19855905e-01 -7.77653813e-01 -9.40076053e-01
4.55024898e-01 2.63109863e-01 3.38711321e-01 -4.79749441e-01
1.31610051e-01 -2.80490756e-01 5.19988477e-01 -1.93893343e-01
6.30052462e-02 -6.23871207e-01 3.31012338e-01 6.68424428e-01
-8.23180676e-01 4.81369823e-01 4.21041518e-01 2.07648352e-01
-1.11215226e-01 -4.19242352e-01 7.25907385e-02 -5.65292060e-01
-4.88901079e-01 -2.55048543e-01 -4.58471447e-01 4.75779474e-01
4.92540389e-01 -1.53930530e-01 -5.41441560e-01 -5.51686466e-01
-5.38932920e-01 5.98962843e-01 2.91332692e-01 5.71708798e-01
7.82004774e-01 -1.13131666e+00 -9.48084831e-01 -1.61935389e-01
3.20803463e-01 1.07342126e-02 4.22434121e-01 8.89899135e-01
-5.72465420e-01 8.16515744e-01 -2.87244003e-03 -4.21157658e-01
-1.05737185e+00 4.18696672e-01 2.42495045e-01 -3.20355713e-01
-6.83213949e-01 9.29946363e-01 1.91616103e-01 -5.37409067e-01
-1.12517148e-01 -5.69603682e-01 -2.59694815e-01 -2.21248612e-01
5.54657280e-01 1.85317785e-01 -1.09041639e-01 -1.59321502e-01
-4.56638426e-01 8.65839943e-02 -1.39943108e-01 -1.83217198e-01
1.18940020e+00 4.76533361e-02 -1.59073532e-01 3.40626985e-01
4.64901447e-01 1.13115393e-01 -5.71134210e-01 1.72033548e-01
3.81153792e-01 -2.93821126e-01 -4.98278260e-01 -1.43035316e+00
-5.07718563e-01 7.89137959e-01 -1.20379897e-02 -7.28050396e-02
9.30023134e-01 2.86758542e-01 6.13374352e-01 3.65729719e-01
1.07292756e-01 -8.54478657e-01 4.25781272e-02 3.61544490e-01
1.25305164e+00 -1.38124931e+00 -6.63907304e-02 -4.85410243e-01
-1.40006602e+00 1.01154327e+00 6.51973724e-01 6.49831891e-01
1.70687839e-01 2.35653156e-03 3.78141850e-01 -4.76446182e-01
-9.73168492e-01 1.77043214e-01 4.32888329e-01 4.52346146e-01
1.11919689e+00 3.58028144e-01 -5.77446699e-01 1.05583787e+00
-7.73417711e-01 1.50697917e-01 8.09211254e-01 5.87498009e-01
1.32543594e-01 -1.28825903e+00 -3.96576524e-01 6.40349865e-01
-6.05773568e-01 -5.41353047e-01 -6.88672483e-01 9.17754471e-01
6.56402111e-02 1.34915125e+00 -4.27843481e-01 1.27382092e-02
3.63179833e-01 3.18878621e-01 9.54459310e-02 -9.26309288e-01
-7.08972394e-01 -3.63930762e-01 7.14616120e-01 -5.02681911e-01
-4.10964489e-01 -6.42376363e-01 -1.55178702e+00 -1.69817492e-01
-7.69545883e-02 6.52270138e-01 2.54573822e-01 1.03855419e+00
4.46924359e-01 6.92192018e-01 -3.39853555e-01 1.66318625e-01
-3.63407522e-01 -8.20512712e-01 8.36769119e-02 5.34871280e-01
4.30674255e-02 -2.15591222e-01 -1.90866724e-01 -1.14173554e-02] | [9.539392471313477, 7.180451393127441] |
2f87e0a3-351c-4c72-acd9-d4d3ba0f627d | name-tagging-for-low-resource-incident | null | null | https://aclanthology.org/N16-1029 | https://aclanthology.org/N16-1029.pdf | Name Tagging for Low-resource Incident Languages based on Expectation-driven Learning | null | ['Heng Ji', 'Daniel Marcu', 'Boliang Zhang', 'Tianlu Wang', 'Ashish Vaswani', 'Kevin Knight', 'Xiaoman Pan'] | 2016-06-01 | null | null | null | naacl-2016-6 | ['cross-lingual-entity-linking'] | ['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] | [-7.431269645690918, 3.586265802383423] |
95ec8280-97a8-457e-8ad5-2aa70a33ee2e | upper-limb-movement-recognition-utilising-eeg | 2207.08650 | null | https://arxiv.org/abs/2207.08650v3 | https://arxiv.org/pdf/2207.08650v3.pdf | Upper Limb Movement Recognition utilising EEG and EMG Signals for Rehabilitative Robotics | Upper limb movement classification, which maps input signals to the target activities, is a key building block in the control of rehabilitative robotics. Classifiers are trained for the rehabilitative system to comprehend the desires of the patient whose upper limbs do not function properly. Electromyography (EMG) signals and Electroencephalography (EEG) signals are used widely for upper limb movement classification. By analysing the classification results of the real-time EEG and EMG signals, the system can understand the intention of the user and predict the events that one would like to carry out. Accordingly, it will provide external help to the user. However, the noise in the real-time EEG and EMG data collection process contaminates the effectiveness of the data, which undermines classification performance. Moreover, not all patients process strong EMG signals due to muscle damage and neuromuscular disorder. To address these issues, this paper explores different feature extraction techniques and machine learning and deep learning models for EEG and EMG signals classification and proposes a novel decision-level multisensor fusion technique to integrate EEG signals with EMG signals. This system retrieves effective information from both sources to understand and predict the desire of the user, and thus aid. By testing out the proposed technique on a publicly available WAY-EEG-GAL dataset, which contains EEG and EMG signals that were recorded simultaneously, we manage to conclude the feasibility and effectiveness of the novel system. | ['Ravi Suppiah', 'ZiHao Wang'] | 2022-07-18 | null | null | null | null | ['electromyography-emg'] | ['medical'] | [ 1.97364628e-01 -2.51984507e-01 -3.31494302e-01 -3.93266641e-02
-5.12510777e-01 3.92846018e-02 2.44378694e-03 -4.40685302e-01
-4.60008323e-01 9.15453553e-01 4.26933765e-01 7.99364299e-02
-5.15275955e-01 -5.47691464e-01 -4.02043372e-01 -7.37516642e-01
-1.05995134e-01 -2.16085045e-03 -2.79489338e-01 -3.47028583e-01
2.19866201e-01 2.85852492e-01 -1.89706194e+00 6.70323908e-01
7.93687046e-01 1.14394879e+00 5.79892814e-01 2.03957453e-01
2.04699785e-01 6.20313764e-01 -7.95936406e-01 4.15089309e-01
-8.58587027e-02 -5.37600756e-01 -6.57366812e-01 8.70542452e-02
-5.59035182e-01 -3.51776898e-01 -5.45202971e-01 9.28688765e-01
8.50640535e-01 -4.63917032e-02 5.19542813e-01 -1.20358610e+00
-2.21177518e-01 4.21955764e-01 -2.11880311e-01 4.29694921e-01
5.27540207e-01 1.53335109e-01 4.24137920e-01 -4.25823808e-01
5.03416657e-01 5.45705616e-01 2.38712519e-01 7.00658798e-01
-8.62461805e-01 -7.79010117e-01 -9.61409062e-02 1.06792867e+00
-1.04610574e+00 -8.65805969e-02 1.03197837e+00 -3.00170630e-01
8.05951238e-01 4.69819099e-01 1.10390162e+00 1.66560364e+00
7.49366224e-01 9.32022452e-01 1.12754035e+00 -1.78893641e-01
2.51414061e-01 -1.02579594e-01 1.74140424e-01 -9.35409367e-02
-1.20406359e-01 2.42434561e-01 -8.27848673e-01 6.57568350e-02
4.37101752e-01 2.61684388e-01 -9.51245666e-01 2.76158422e-01
-1.18093884e+00 2.10149556e-01 5.55512547e-01 7.44811893e-01
-1.20569646e+00 -1.78476334e-01 3.63230258e-01 5.61526179e-01
3.73903625e-02 9.80772823e-02 -3.46192896e-01 -6.55007124e-01
-6.61092401e-01 -4.46398109e-02 4.72533464e-01 6.39452636e-01
1.40638843e-01 3.81916203e-02 5.10810316e-03 6.73867762e-01
1.27095655e-01 2.75775015e-01 1.13097715e+00 -6.90103114e-01
3.05422604e-01 7.84589469e-01 -1.59206003e-01 -7.80982733e-01
-6.24909163e-01 -4.14501548e-01 -7.55669415e-01 4.30835664e-01
1.12784058e-02 -3.18721592e-01 -6.55654311e-01 1.48200989e+00
7.47901276e-02 -2.80390196e-02 -3.12240552e-02 1.44812226e+00
6.10080600e-01 3.15158844e-01 8.75184759e-02 -3.63731116e-01
1.40280783e+00 -2.04616517e-01 -1.09914947e+00 -3.90530676e-01
4.03528035e-01 -1.45878658e-01 8.52171302e-01 7.24318624e-01
-7.58203387e-01 -5.63175440e-01 -1.17277229e+00 4.77714449e-01
-3.12633589e-02 3.89608890e-01 6.34309769e-01 6.15348332e-02
-4.94015455e-01 9.16477025e-01 -9.60755944e-01 -3.17706704e-01
3.87814075e-01 7.83188701e-01 -7.01876223e-01 3.57949704e-01
-1.30197155e+00 1.22527277e+00 4.29646045e-01 5.81588328e-01
-2.64460921e-01 -2.17993230e-01 -2.99123943e-01 4.40032408e-03
-3.54888998e-02 -3.73634934e-01 7.45576739e-01 -8.45005691e-01
-1.42468786e+00 4.20435160e-01 1.32719144e-01 -9.36414078e-02
2.50699252e-01 -2.45760977e-01 -4.42609787e-01 1.46282911e-01
-2.14998379e-01 1.38408989e-01 7.04751134e-01 -7.14400053e-01
-6.65740073e-01 -9.46800232e-01 -4.72244143e-01 4.16287720e-01
-5.18017411e-01 -7.64261931e-02 2.11046740e-01 -4.96959031e-01
3.50977272e-01 -5.97852230e-01 2.35344902e-01 -5.04250824e-01
-7.07246214e-02 -2.99146503e-01 7.68114448e-01 -9.12520885e-01
1.02528131e+00 -2.30638933e+00 8.05075467e-01 2.44215101e-01
5.71783222e-02 -5.24964035e-02 2.39284355e-02 4.41693276e-01
-3.50372851e-01 -4.91754621e-01 9.83834565e-02 3.82334501e-01
-2.95091152e-01 4.06043857e-01 7.06861317e-02 5.40095568e-01
8.04095902e-03 6.97942436e-01 -3.69771063e-01 -1.55754834e-01
3.77230704e-01 3.77704710e-01 -2.21267506e-01 5.01534581e-01
3.15213203e-01 8.35853219e-01 -5.32571614e-01 6.52939141e-01
-2.50963103e-02 2.70146132e-01 9.56635848e-02 -6.16747141e-01
1.62953556e-01 3.72009814e-01 -1.16576850e+00 1.95215595e+00
-2.95334578e-01 6.04158282e-01 1.45234391e-01 -1.34272778e+00
7.21223891e-01 7.67456055e-01 1.03808081e+00 -1.10939610e+00
7.89690673e-01 3.70865136e-01 3.35292459e-01 -1.20831192e+00
-2.49032363e-01 -6.33563623e-02 1.64588511e-01 3.34769011e-01
1.91284895e-01 3.85222882e-01 -1.31858274e-01 -3.47988516e-01
1.36713302e+00 4.63947892e-01 6.20765463e-02 -3.59041318e-02
4.68036801e-01 -2.20349785e-02 5.19848049e-01 6.52278289e-02
-2.69767568e-02 1.29331231e-01 8.82754028e-02 -1.32381409e-01
-4.15926993e-01 -9.42575514e-01 -6.92919344e-02 8.09796035e-01
2.01954201e-01 -9.10752267e-02 -6.08502328e-01 -4.69333082e-02
-1.27518000e-02 5.31005800e-01 -4.78819579e-01 -8.20555747e-01
-4.58739787e-01 -8.71695340e-01 9.04729124e-03 7.68935084e-01
2.17414141e-01 -1.66067052e+00 -9.94564593e-01 5.22846341e-01
-6.12713516e-01 -8.70231390e-01 1.59852430e-01 6.04642451e-01
-1.01616633e+00 -1.26161301e+00 -4.82824594e-01 -8.57540786e-01
4.32632446e-01 -2.29376942e-01 1.28785893e-01 -4.01654579e-02
-2.96496183e-01 3.16325158e-01 -6.53695643e-01 -3.17341834e-01
-2.11998343e-01 -1.73366919e-01 3.25971633e-01 -3.38558690e-03
6.75942898e-01 -9.87030804e-01 -5.01801014e-01 1.76911041e-01
-5.50229847e-01 2.17912942e-01 9.74675357e-01 8.70371997e-01
2.94377267e-01 7.01914206e-02 9.89932954e-01 6.39897734e-02
1.22572494e+00 -5.18356800e-01 1.59710631e-01 9.50096399e-02
-4.10858810e-01 -2.27433980e-01 4.16686833e-01 -7.42105782e-01
-7.88651884e-01 -7.73758218e-02 -4.14531112e-01 -2.18720526e-01
-3.57540369e-01 6.98853314e-01 -2.76853472e-01 1.29343152e-01
6.87488317e-01 5.53410709e-01 1.78848788e-01 -4.53261048e-01
-2.82094687e-01 1.57003355e+00 1.07237709e+00 -3.14373136e-01
-1.11073248e-01 5.51788472e-02 -2.03949243e-01 -6.61407709e-01
-1.41262300e-02 -4.43842828e-01 -5.60431361e-01 -8.11891317e-01
7.57394314e-01 -7.55726993e-01 -1.10601866e+00 6.32045805e-01
-1.06668437e+00 1.78979151e-02 2.65685227e-02 1.23604190e+00
-9.23076153e-01 1.79660618e-01 -7.29626656e-01 -1.00407863e+00
-6.96959019e-01 -1.11091673e+00 7.70751715e-01 5.89943752e-02
-5.96450806e-01 -9.70098302e-02 -4.37325388e-01 2.99530655e-01
2.70589322e-01 2.94479072e-01 8.84862959e-01 -4.20195282e-01
9.96279791e-02 -6.16395116e-01 3.20476741e-01 3.76388937e-01
3.41895580e-01 -7.50427902e-01 -7.02275038e-01 -4.04211469e-02
8.22783649e-01 -2.35661373e-01 1.82469606e-01 3.80546063e-01
1.12335420e+00 -5.27401380e-02 -2.97175825e-01 2.57483155e-01
1.11496222e+00 6.41195893e-01 1.06068456e+00 3.72277200e-01
3.92523438e-01 8.97584856e-01 5.04835486e-01 2.65664697e-01
6.88386187e-02 5.31956553e-01 6.05837226e-01 1.67506024e-01
9.26660225e-02 2.35325724e-01 5.12038827e-01 1.05781281e+00
-5.88300526e-01 1.02529332e-01 -4.79524642e-01 2.88122356e-01
-2.02783942e+00 -9.44513321e-01 -4.14931476e-01 1.88141704e+00
7.11925328e-01 -2.60948436e-03 -7.69342259e-02 8.93741667e-01
4.88272995e-01 -6.71242595e-01 -7.49243081e-01 -5.57590090e-02
8.65214393e-02 5.75078130e-01 1.43832713e-01 -2.47177616e-01
-4.55449104e-01 2.75648504e-01 5.21544027e+00 5.67927420e-01
-1.39828384e+00 1.29659504e-01 -2.45480925e-01 -3.58298510e-01
2.66992509e-01 -5.06772161e-01 -7.94593021e-02 8.36144924e-01
1.05684733e+00 -1.58011362e-01 7.89243937e-01 5.51418841e-01
6.64934158e-01 -4.09802020e-01 -1.30554402e+00 1.34424996e+00
-9.21403766e-02 -8.80470574e-01 -4.59846497e-01 4.30444255e-02
-3.02927066e-02 3.09752315e-01 -5.56993306e-01 1.97183937e-01
-7.03426480e-01 -8.23844314e-01 6.56179249e-01 8.71802270e-01
3.27027857e-01 -6.21025860e-01 1.00206113e+00 9.01992202e-01
-8.37109506e-01 -5.73905408e-01 7.92643279e-02 -3.60333264e-01
2.12795660e-01 1.35974074e-02 -2.06776068e-01 6.69876933e-01
9.06290352e-01 8.22980702e-01 -1.72380894e-01 8.37499976e-01
-4.35002953e-01 3.78787130e-01 -4.74710912e-01 -2.98289627e-01
-2.35095903e-01 -2.43040338e-01 6.43081486e-01 4.87902582e-01
6.20165408e-01 3.88367295e-01 -2.93997258e-01 8.95693302e-01
3.48359853e-01 2.31642593e-02 -4.73926306e-01 -3.43391337e-02
-1.01728470e-03 1.07460332e+00 -2.70082563e-01 8.80037900e-03
-2.33264968e-01 1.07347727e+00 4.21119621e-03 3.96644861e-01
-5.62670529e-01 -3.58768702e-01 5.50269008e-01 3.73086333e-02
-3.87690008e-01 -6.63466305e-02 -4.82107401e-01 -1.12611103e+00
6.40346050e-01 -9.23016191e-01 1.33096412e-01 -1.10388935e+00
-1.05301404e+00 5.23262024e-01 -1.82735492e-02 -1.54646599e+00
-5.90300083e-01 -7.05752254e-01 -6.58876240e-01 1.01536775e+00
-9.07749116e-01 -5.13978839e-01 -4.64601934e-01 8.34997714e-01
4.89599466e-01 1.74036194e-02 9.29186225e-01 6.22894526e-01
-4.91396040e-01 -9.32046324e-02 -8.78687501e-02 5.12254760e-02
3.36522460e-01 -6.54888570e-01 -5.75651765e-01 6.74196541e-01
-1.97786406e-01 4.26928341e-01 6.49957776e-01 -6.48618639e-01
-1.71571171e+00 -3.87516171e-01 4.11305249e-01 1.03296481e-01
4.42046344e-01 -1.08768068e-01 -9.21486497e-01 2.24022254e-01
7.98361078e-02 -5.69986284e-01 6.29558742e-01 -3.21355045e-01
7.36050904e-01 -1.14515379e-01 -1.08836877e+00 5.89580357e-01
1.14033437e+00 -3.59063685e-01 -1.25526035e+00 1.44018710e-01
-1.17898330e-01 -2.10617915e-01 -1.09459054e+00 6.52750373e-01
9.90276456e-01 -6.26467943e-01 5.79505563e-01 -6.95507765e-01
2.49323934e-01 6.07167045e-03 -1.18790790e-01 -1.72520602e+00
-1.45226374e-01 -1.45272091e-01 -1.77514970e-01 6.69858396e-01
-2.78010350e-02 -5.85148752e-01 6.01972342e-01 5.20419300e-01
-3.84504408e-01 -8.35352719e-01 -1.16680574e+00 -5.15529156e-01
-4.20071632e-01 -7.75360167e-01 4.28621054e-01 5.61114252e-01
9.76854146e-01 3.24714988e-01 -5.07265031e-01 -1.19966954e-01
4.11258847e-01 7.05352500e-02 3.08905035e-01 -1.25890648e+00
-5.68732955e-02 -4.42833126e-01 -8.27591479e-01 -5.72402239e-01
1.24842867e-01 -1.16338611e+00 9.67379883e-02 -1.91922343e+00
-1.73599511e-01 3.24460924e-01 -5.19186080e-01 7.39417315e-01
1.65313721e-01 3.07995211e-02 2.86335940e-03 2.73837179e-01
2.25936860e-01 7.61372149e-01 1.15746498e+00 -1.57804534e-01
-6.10220909e-01 2.09251031e-01 -4.77316082e-01 5.91427147e-01
9.23932016e-01 -4.55720454e-01 -3.29220682e-01 -2.15234876e-01
-7.79351071e-02 4.99782026e-01 3.90391946e-01 -1.33609331e+00
4.91856545e-01 3.75814140e-02 7.31963336e-01 -6.95094943e-01
4.82976913e-01 -1.22816968e+00 5.35070002e-01 7.20610678e-01
-8.25827941e-02 -2.15089083e-01 5.27407490e-02 2.54838347e-01
-3.35817188e-01 3.68312523e-02 3.24633330e-01 9.51907039e-02
-8.11665595e-01 -3.41918953e-02 -8.16341221e-01 -5.96887708e-01
1.25277805e+00 -6.26879334e-01 -3.55097502e-02 -1.81695819e-01
-1.13972306e+00 3.46466005e-01 -3.22932214e-01 8.28324437e-01
8.94547462e-01 -1.56622577e+00 -5.69344521e-01 6.54480994e-01
7.87332729e-02 -4.97741967e-01 3.67698073e-01 1.40209329e+00
-1.42856156e-02 2.00408667e-01 -1.00108051e+00 -6.67984486e-01
-1.20013773e+00 1.68346912e-01 2.14555785e-01 4.62808430e-01
-1.00527442e+00 1.74391329e-01 -5.89002728e-01 2.54590139e-02
4.38703746e-01 -2.87266254e-01 -6.90443039e-01 1.77814201e-01
4.40247446e-01 2.68103749e-01 3.50181431e-01 -5.52935421e-01
-4.85375643e-01 2.24582881e-01 4.33813751e-01 -5.67111634e-02
1.63205993e+00 -7.64686093e-02 -5.76097630e-02 4.99663204e-01
1.03529084e+00 -7.60549128e-01 -6.84704304e-01 2.25441694e-01
-8.04942921e-02 -2.10237429e-01 3.43974769e-01 -9.52114284e-01
-1.06546462e+00 8.86124909e-01 1.07416761e+00 -5.67498431e-02
1.63742518e+00 -1.34378582e-01 9.17221487e-01 3.81052613e-01
9.79311049e-01 -1.39655197e+00 -2.46038780e-01 -2.49454096e-01
1.16770768e+00 -8.93542707e-01 -8.64697397e-02 3.04674394e-02
-6.01441741e-01 1.25794482e+00 5.46998858e-01 -1.93233818e-01
6.93085730e-01 2.20921919e-01 -3.26657295e-02 -2.02196971e-01
-6.13782585e-01 -1.94867194e-01 3.47998500e-01 5.98754466e-01
1.45570636e-01 8.52947459e-02 -9.95500445e-01 1.29418111e+00
-2.24264160e-01 5.71130097e-01 3.02158117e-01 1.00376129e+00
-4.25476730e-01 -1.00672555e+00 -5.41968942e-01 9.60834026e-01
-2.87226796e-01 4.10721391e-01 -2.43119359e-01 7.02635646e-01
1.62600681e-01 1.17599905e+00 -2.17029139e-01 -9.47240710e-01
7.70483375e-01 3.60296845e-01 6.21850729e-01 -3.28785717e-01
-6.68269932e-01 2.47375324e-01 -4.55347449e-02 -7.80796111e-01
-6.03235245e-01 -5.43854833e-01 -1.63680947e+00 9.96327475e-02
-3.22727799e-01 5.06849512e-02 1.00813568e+00 1.26189637e+00
2.17007220e-01 1.11319649e+00 5.50232291e-01 -1.02391696e+00
-6.47047639e-01 -1.56712461e+00 -7.62011170e-01 5.16447067e-01
1.92900132e-02 -8.45422268e-01 -3.36445153e-01 -1.74153849e-01] | [6.914714336395264, 0.21213054656982422] |
2a1651a8-1202-40dd-a788-3bb158961d56 | scat-robust-self-supervised-contrastive | 2307.01488 | null | https://arxiv.org/abs/2307.01488v1 | https://arxiv.org/pdf/2307.01488v1.pdf | SCAT: Robust Self-supervised Contrastive Learning via Adversarial Training for Text Classification | Despite their promising performance across various natural language processing (NLP) tasks, current NLP systems are vulnerable to textual adversarial attacks. To defend against these attacks, most existing methods apply adversarial training by incorporating adversarial examples. However, these methods have to rely on ground-truth labels to generate adversarial examples, rendering it impractical for large-scale model pre-training which is commonly used nowadays for NLP and many other tasks. In this paper, we propose a novel learning framework called SCAT (Self-supervised Contrastive Learning via Adversarial Training), which can learn robust representations without requiring labeled data. Specifically, SCAT modifies random augmentations of the data in a fully labelfree manner to generate adversarial examples. Adversarial training is achieved by minimizing the contrastive loss between the augmentations and their adversarial counterparts. We evaluate SCAT on two text classification datasets using two state-of-the-art attack schemes proposed recently. Our results show that SCAT can not only train robust language models from scratch, but it can also significantly improve the robustness of existing pre-trained language models. Moreover, to demonstrate its flexibility, we show that SCAT can also be combined with supervised adversarial training to further enhance model robustness. | ['Dit-yan Yeung', 'Junjie Wu'] | 2023-07-04 | null | null | null | null | ['contrastive-learning', 'contrastive-learning', 'text-classification'] | ['computer-vision', 'methodology', 'natural-language-processing'] | [ 5.21410465e-01 2.54840493e-01 3.76822203e-02 -3.19096357e-01
-1.10123932e+00 -1.20328367e+00 8.75389814e-01 2.66600579e-01
-5.29505253e-01 6.85149133e-01 -6.11721240e-02 -5.88509202e-01
4.62164253e-01 -9.67613339e-01 -7.75647700e-01 -4.65061575e-01
1.70509055e-01 5.32388031e-01 3.97099964e-02 -4.76876169e-01
-1.47988409e-01 7.42949963e-01 -7.35147059e-01 3.46394151e-01
8.73034596e-01 5.09712636e-01 -6.23773634e-01 6.86369002e-01
-1.31086707e-01 8.28933775e-01 -9.39844966e-01 -8.99600685e-01
4.65338707e-01 -1.20734118e-01 -8.50229263e-01 -2.19059438e-01
3.88745427e-01 -2.01213911e-01 -6.08006656e-01 1.16618538e+00
5.33998072e-01 1.47238553e-01 6.94595158e-01 -1.47447968e+00
-8.14469337e-01 6.97852135e-01 -3.09371203e-01 -1.39911398e-01
2.70338297e-01 2.99536288e-01 8.53777826e-01 -6.84006274e-01
4.37147021e-01 1.47073328e+00 5.65564990e-01 9.41431344e-01
-1.40885305e+00 -1.01826429e+00 1.55576468e-01 -1.66741431e-01
-1.24317575e+00 -4.00175393e-01 9.83834267e-01 -2.56277978e-01
5.98114967e-01 3.37353110e-01 -7.86266550e-02 1.63248610e+00
7.62555376e-02 8.84058058e-01 1.18245304e+00 -5.55176675e-01
1.69724286e-01 2.22455189e-01 -4.80772033e-02 6.92267776e-01
-1.08052492e-01 2.50934958e-01 -6.60514981e-02 -6.73336744e-01
2.70557940e-01 -1.27211586e-01 -2.69807696e-01 -1.47798449e-01
-1.04168081e+00 1.01068747e+00 5.39162576e-01 2.43058339e-01
-1.23016192e-02 3.59107554e-02 5.88096738e-01 3.74047041e-01
5.40675581e-01 6.60899043e-01 -4.79559988e-01 5.49827278e-01
-6.38621390e-01 2.38952830e-01 8.35497081e-01 7.13975966e-01
6.16676748e-01 3.04637313e-01 -1.69679433e-01 7.65811801e-01
3.19624811e-01 6.17998064e-01 5.39359570e-01 -5.43838203e-01
7.87554324e-01 4.54426616e-01 -5.70180193e-02 -1.07065153e+00
-2.54651368e-01 -2.03867123e-01 -1.06741452e+00 2.66323298e-01
4.73035932e-01 -3.13153595e-01 -9.52071846e-01 1.85856998e+00
3.13776165e-01 2.07403839e-01 5.16170204e-01 4.52793986e-01
4.01433438e-01 9.61373746e-01 3.46486419e-01 -1.91343552e-03
1.02796996e+00 -9.36329603e-01 -6.23952985e-01 -5.08268952e-01
7.67048717e-01 -7.93764532e-01 1.19443429e+00 3.09736669e-01
-8.81222546e-01 -3.41809601e-01 -1.03352368e+00 4.87967134e-02
-6.45351887e-01 -3.00905913e-01 3.92286509e-01 9.78752494e-01
-5.82959235e-01 4.73895252e-01 -8.72409582e-01 1.17344342e-01
5.72540283e-01 3.02476943e-01 -5.83123982e-01 -1.65030763e-01
-1.69152236e+00 8.26035738e-01 6.38474286e-01 -4.99870293e-02
-9.23714221e-01 -6.28194749e-01 -1.16566193e+00 1.23195620e-02
4.37902033e-01 -3.11176121e-01 1.17157722e+00 -1.03345740e+00
-1.69783032e+00 8.38111162e-01 2.40029380e-01 -7.07170188e-01
8.03588986e-01 -2.70592332e-01 -5.63971698e-01 1.48293465e-01
-1.70919597e-01 4.83064711e-01 1.16494370e+00 -1.43121660e+00
7.80962333e-02 -7.40870461e-02 2.63271749e-01 -1.89155683e-01
-7.36355186e-01 2.29867592e-01 -3.07535797e-01 -1.20126235e+00
-3.94753188e-01 -1.08687949e+00 -5.68018079e-01 1.12696677e-01
-7.17140079e-01 -1.92052592e-02 1.00562572e+00 -5.83991349e-01
8.87335539e-01 -2.14934444e+00 -8.68123770e-02 4.42599505e-01
4.87952307e-02 9.41623211e-01 -5.38244426e-01 4.69879955e-01
-1.63816974e-01 5.50558388e-01 -7.19337761e-01 -4.64361876e-01
2.64440894e-01 3.04184258e-01 -8.33762586e-01 4.57115382e-01
5.45269907e-01 9.79290307e-01 -9.00364459e-01 -4.86439437e-01
2.70627588e-01 5.31773388e-01 -6.22350037e-01 4.15776014e-01
-4.90658790e-01 5.60085535e-01 -4.51107830e-01 3.36709261e-01
7.85734534e-01 4.84709302e-03 1.43257052e-01 1.54480636e-01
6.55909181e-01 6.29194453e-02 -1.05286884e+00 1.32378733e+00
-6.64560497e-01 3.99378240e-01 9.81628895e-02 -1.15216506e+00
9.30756092e-01 4.03362364e-01 7.16372356e-02 -1.55420482e-01
2.42003247e-01 9.81072485e-02 -1.45584390e-01 -1.09305449e-01
2.32454732e-01 -3.30429882e-01 -5.42624235e-01 6.48158669e-01
-6.17680289e-02 -4.28272426e-01 -3.74413691e-02 4.78776813e-01
1.09518802e+00 -1.42034844e-01 3.28712851e-01 1.67086214e-01
1.01837528e+00 -2.40203097e-01 4.69375402e-01 7.90252447e-01
-1.53375357e-01 4.83230710e-01 4.50176418e-01 -2.60948420e-01
-9.77668166e-01 -1.12668025e+00 2.68167965e-02 1.01310170e+00
-2.97225952e-01 -3.43356669e-01 -7.54121721e-01 -1.50513971e+00
-6.13260875e-03 8.23484361e-01 -5.43990314e-01 -4.82201874e-01
-7.15134323e-01 -6.12926900e-01 1.25506568e+00 4.87355292e-01
5.28746009e-01 -1.24928856e+00 3.07430923e-01 1.21988066e-01
-2.58641001e-02 -1.33594346e+00 -5.51458657e-01 1.09151036e-01
-3.39371473e-01 -1.08799350e+00 -4.86043513e-01 -8.03180516e-01
9.23161328e-01 -2.07165644e-01 8.95551682e-01 2.69694388e-01
-6.32812902e-02 2.97370702e-01 -3.92562807e-01 -4.40305084e-01
-1.37132633e+00 2.61269003e-01 1.86970085e-01 8.45037922e-02
8.23352337e-02 -4.95733589e-01 -4.64537553e-02 2.63354599e-01
-1.48600984e+00 -2.83509344e-01 3.67221057e-01 1.01484561e+00
4.46981758e-01 3.53103131e-01 7.31833696e-01 -1.41041362e+00
6.89831913e-01 -3.68563950e-01 -7.05700159e-01 2.98404217e-01
-2.42307425e-01 1.42025098e-01 1.48407650e+00 -8.18765819e-01
-8.87456536e-01 1.08875185e-02 -6.06535435e-01 -5.30077279e-01
-2.73734093e-01 4.71458972e-01 -5.41586041e-01 -3.72981668e-01
9.30266440e-01 1.02744654e-01 -2.16011569e-01 -2.32900843e-01
6.44526899e-01 6.73509121e-01 5.81594586e-01 -7.08862543e-01
1.63843715e+00 3.05493891e-01 -9.97649431e-02 -5.61008990e-01
-1.12046754e+00 5.55999428e-02 -5.71560025e-01 3.59483898e-01
5.26214480e-01 -6.09936059e-01 -5.13345182e-01 5.29787004e-01
-1.16535425e+00 -5.03919780e-01 -1.70070112e-01 2.41248339e-01
-3.38624775e-01 6.44699037e-01 -7.29503334e-01 -6.76636815e-01
-4.95209932e-01 -9.04662192e-01 7.34260738e-01 -2.50274479e-01
-1.21489398e-01 -1.25496972e+00 2.31918544e-02 3.99314702e-01
3.62374008e-01 5.94644129e-01 9.86077189e-01 -1.32096732e+00
-1.35349125e-01 -7.18486488e-01 1.04993358e-01 8.09578538e-01
1.43846244e-01 -7.10849687e-02 -8.68515074e-01 -5.74986637e-01
-1.65395513e-02 -7.57259071e-01 5.62693119e-01 -5.26852548e-01
1.32497454e+00 -8.12206209e-01 -1.94535464e-01 5.25363803e-01
1.16176903e+00 -2.21613526e-01 6.05358303e-01 2.70714134e-01
9.40088570e-01 5.57978332e-01 3.46344680e-01 7.58215860e-02
-1.56041726e-01 5.58486283e-01 3.97202998e-01 -2.90605038e-01
2.65267313e-01 -4.33360696e-01 4.94167745e-01 4.68902647e-01
7.09840357e-01 -6.10103965e-01 -1.03526509e+00 2.42601097e-01
-1.57977521e+00 -8.92073333e-01 2.18847945e-01 2.03066301e+00
1.18735158e+00 4.21067923e-01 -2.71302700e-01 3.90446097e-01
7.77762413e-01 3.37396592e-01 -5.38240135e-01 -5.82305670e-01
-1.05581090e-01 6.25444531e-01 3.21086377e-01 7.06571519e-01
-1.51871443e+00 1.32927942e+00 6.19326448e+00 9.04886961e-01
-1.09585047e+00 1.96596771e-01 4.58455324e-01 1.16788045e-01
-2.91750431e-01 -2.10704684e-01 -5.48023403e-01 3.39752018e-01
8.74725223e-01 -2.88263261e-01 3.79770577e-01 8.73221219e-01
-1.27584517e-01 7.69767463e-01 -1.01730156e+00 5.07220864e-01
1.08537711e-01 -1.13180971e+00 4.85826164e-01 -2.25268811e-01
7.81226873e-01 -2.33182728e-01 1.35520086e-01 6.93494558e-01
8.25645149e-01 -1.13679147e+00 3.28053147e-01 -2.14385595e-02
7.37114966e-01 -1.00286126e+00 7.17592537e-01 6.27637386e-01
-8.71324420e-01 9.46816057e-02 -2.39601165e-01 1.80456147e-01
1.96650073e-01 5.21652639e-01 -8.66288781e-01 6.24173284e-01
1.23149246e-01 4.01680022e-01 -5.65915525e-01 3.02577585e-01
-6.67736709e-01 7.94594884e-01 -4.21353877e-01 2.30666518e-01
3.15812498e-01 2.65304267e-01 5.65427661e-01 1.21513379e+00
-2.01861113e-01 -4.51872908e-02 5.55642903e-01 7.43893504e-01
-6.91805840e-01 2.64216840e-01 -8.57758045e-01 -3.62209409e-01
5.23701429e-01 1.23481965e+00 -3.85587186e-01 -3.13238472e-01
-2.87001759e-01 1.16564131e+00 4.97097462e-01 3.12788606e-01
-9.18514490e-01 -4.82680827e-01 4.55059290e-01 -1.03429928e-01
-7.60687068e-02 -2.85845846e-01 4.57416400e-02 -1.31172311e+00
-8.39239284e-02 -1.35211909e+00 4.31982934e-01 -3.30944121e-01
-1.81543362e+00 7.68909037e-01 -2.70239383e-01 -1.06521976e+00
-3.47764045e-01 -5.78548193e-01 -6.92313612e-01 9.45151985e-01
-1.44141936e+00 -1.48849821e+00 1.62320599e-01 9.85797822e-01
2.34280124e-01 -4.41466808e-01 1.20964682e+00 1.32534787e-01
-5.89817822e-01 1.29460704e+00 1.16860762e-01 6.98871017e-01
8.89194369e-01 -1.16962016e+00 8.62652600e-01 1.12728286e+00
3.03735018e-01 6.51749194e-01 6.23317361e-01 -5.24279237e-01
-1.04609823e+00 -1.59571373e+00 6.31753802e-01 -4.92596775e-01
9.34131563e-01 -7.77162135e-01 -1.24674106e+00 9.97037470e-01
-5.41487932e-02 4.65745509e-01 9.05608118e-01 -2.71826863e-01
-8.75726759e-01 2.26446465e-01 -1.54955113e+00 8.53868186e-01
5.53700328e-01 -7.15066612e-01 -7.04126894e-01 6.55141115e-01
9.94012952e-01 -4.22357231e-01 -7.67419219e-01 5.34870744e-01
1.07573539e-01 -3.50571632e-01 1.07492900e+00 -9.61731315e-01
2.66799480e-01 -2.17694402e-01 -1.82952043e-02 -1.28791618e+00
4.03114818e-02 -9.40434277e-01 -1.78554669e-01 1.63227797e+00
4.00445610e-01 -1.02911222e+00 8.00208926e-01 5.97853959e-01
1.31925434e-01 -4.47116882e-01 -8.11547041e-01 -8.79275262e-01
7.59228408e-01 -5.55048287e-01 5.38613617e-01 1.31970406e+00
-1.96853116e-01 3.23316574e-01 -4.25302356e-01 5.77406645e-01
5.19660830e-01 -4.07846272e-01 9.85621750e-01 -9.92366254e-01
-3.74348700e-01 -1.24814920e-01 -2.98919797e-01 -6.57332897e-01
9.88564193e-01 -1.20910764e+00 4.15437147e-02 -9.83510971e-01
-3.52500975e-01 -6.23120070e-01 -2.04887509e-01 9.76090133e-01
-4.87893701e-01 4.88787860e-01 3.14484328e-01 1.50993839e-01
-2.34189704e-01 6.16384089e-01 9.86511230e-01 -4.80652094e-01
-8.83794427e-02 1.07677035e-01 -6.42508209e-01 8.86836648e-01
9.98981655e-01 -8.22850525e-01 -3.66108388e-01 -1.97179854e-01
5.61308526e-02 -2.66660452e-01 3.18332493e-01 -8.55846822e-01
-4.99152541e-02 -1.48634821e-01 2.36377209e-01 -1.31860957e-01
1.38413265e-01 -8.62983882e-01 -3.09018582e-01 4.87522990e-01
-6.42975211e-01 -2.47604892e-01 5.48740268e-01 5.88275850e-01
-1.12790398e-01 -3.52832019e-01 1.07133281e+00 2.41456274e-02
-1.44512683e-01 4.36054856e-01 -4.16644543e-01 2.30243728e-01
1.09183192e+00 5.72740734e-01 -3.87829483e-01 -2.83031285e-01
-6.53545976e-01 2.62539655e-01 4.84888047e-01 4.40668523e-01
4.16915834e-01 -1.23102868e+00 -9.03579891e-01 3.19903672e-01
7.46374428e-02 6.54542297e-02 7.63478130e-02 9.34056006e-03
-4.73114818e-01 1.01950638e-01 6.61305413e-02 -1.55763447e-01
-1.25500345e+00 1.23650718e+00 3.03113818e-01 -8.61112654e-01
-4.28431094e-01 6.26602232e-01 2.50657678e-01 -9.75604475e-01
1.84716240e-01 2.51864970e-01 -5.69378547e-02 -3.93666625e-01
6.10096037e-01 -8.60518962e-02 2.88573708e-02 -5.48202097e-01
-3.04589272e-01 2.26135343e-01 -5.47796309e-01 -2.33733177e-01
9.05010521e-01 3.19798082e-01 3.80488373e-02 8.78200494e-03
1.36661243e+00 4.66412932e-01 -9.26233113e-01 -5.59113324e-01
-2.17118874e-01 -2.82323480e-01 -2.53444225e-01 -8.46433640e-01
-9.97667432e-01 1.00812912e+00 2.22044289e-01 3.12880576e-01
9.75892484e-01 -3.83503467e-01 9.76948917e-01 7.21425653e-01
1.93220496e-01 -6.32328153e-01 3.39554995e-01 6.38247788e-01
9.20283616e-01 -1.22356188e+00 -1.50202617e-01 -6.14966154e-01
-6.06544912e-01 9.46455836e-01 5.32145619e-01 -2.45295882e-01
5.09917796e-01 4.60419089e-01 3.72083485e-01 3.75286013e-01
-3.64712328e-01 2.13012412e-01 1.91226259e-01 7.66489148e-01
1.43847391e-01 -2.32323259e-01 -3.15116830e-02 5.67924738e-01
-2.62643427e-01 -3.40483069e-01 3.44592720e-01 1.04756546e+00
1.36221021e-01 -1.68611121e+00 -6.22372389e-01 1.16805717e-01
-8.23081255e-01 -2.77673185e-01 -6.22261941e-01 7.46312678e-01
-2.11080372e-01 1.00155902e+00 -4.25350040e-01 -3.78148794e-01
2.88445354e-01 2.41986439e-01 1.82274655e-01 -8.02631438e-01
-9.54556227e-01 -4.70054775e-01 -1.02143027e-02 -1.85446277e-01
-8.57233927e-02 -2.03779802e-01 -1.21028519e+00 -4.97408211e-01
-3.24292451e-01 3.61060113e-01 4.33573544e-01 9.72519219e-01
9.23363268e-02 4.11575794e-01 9.70220149e-01 -7.94134796e-01
-1.00417888e+00 -8.88837099e-01 -1.65868506e-01 7.20451951e-01
2.74921238e-01 -3.54028434e-01 -6.19430900e-01 1.26900300e-01] | [5.9524736404418945, 8.047940254211426] |
49a9e8e9-de4f-4ed4-96a3-318f8b968ebd | latent-conditioned-policy-gradient-for-multi | 2303.08909 | null | https://arxiv.org/abs/2303.08909v1 | https://arxiv.org/pdf/2303.08909v1.pdf | Latent-Conditioned Policy Gradient for Multi-Objective Deep Reinforcement Learning | Sequential decision making in the real world often requires finding a good balance of conflicting objectives. In general, there exist a plethora of Pareto-optimal policies that embody different patterns of compromises between objectives, and it is technically challenging to obtain them exhaustively using deep neural networks. In this work, we propose a novel multi-objective reinforcement learning (MORL) algorithm that trains a single neural network via policy gradient to approximately obtain the entire Pareto set in a single run of training, without relying on linear scalarization of objectives. The proposed method works in both continuous and discrete action spaces with no design change of the policy network. Numerical experiments in benchmark environments demonstrate the practicality and efficacy of our approach in comparison to standard MORL baselines. | ['Chetan Gupta', 'Takuya Kanazawa'] | 2023-03-15 | null | null | null | null | ['multi-objective-reinforcement-learning'] | ['methodology'] | [ 2.54711229e-02 -3.35346043e-01 -2.81217575e-01 -7.79619887e-02
-7.72079766e-01 -6.71885788e-01 3.35452795e-01 1.22376524e-01
-7.19672382e-01 1.26357496e+00 -8.84271488e-02 -6.33309066e-01
-6.37781680e-01 -6.65334165e-01 -6.98766530e-01 -7.54934967e-01
-3.23795564e-02 5.10776222e-01 -2.57862955e-01 -3.93725514e-01
4.32998747e-01 3.52512687e-01 -1.27773917e+00 -1.34534925e-01
1.14383030e+00 1.22270012e+00 1.09185569e-01 6.72769308e-01
1.87213540e-01 7.04289794e-01 -7.78376102e-01 -2.28570655e-01
5.18917680e-01 -2.59198099e-01 -6.89252257e-01 3.31162545e-03
2.06683740e-01 -5.23952007e-01 -7.95185845e-03 1.22028136e+00
7.36084461e-01 4.72142190e-01 4.62582409e-01 -1.10967958e+00
-5.60886800e-01 6.56776905e-01 -7.35288441e-01 1.00443035e-01
-1.17898725e-01 5.20708680e-01 1.36930752e+00 -3.84387046e-01
1.17200814e-01 1.33979726e+00 3.01749021e-01 3.20702195e-01
-1.35905886e+00 -5.14466047e-01 3.59222800e-01 1.57591775e-01
-1.00362325e+00 -1.66167974e-01 7.63483465e-01 -1.44605279e-01
8.24447811e-01 4.97278310e-02 7.28484929e-01 9.46434557e-01
5.06832063e-01 8.37857485e-01 1.29500449e+00 -3.12729776e-01
5.58530688e-01 -2.30014622e-01 -4.57210511e-01 5.33003569e-01
4.31147188e-01 4.86387849e-01 -1.49084717e-01 -2.05555916e-01
5.83026230e-01 -8.31017047e-02 -3.26905021e-04 -5.95016956e-01
-9.62862909e-01 8.36087227e-01 3.06042165e-01 1.10368565e-01
-7.54508436e-01 6.06503665e-01 2.46456876e-01 5.71316540e-01
1.53186128e-01 9.62386489e-01 -3.61755282e-01 -1.50002986e-01
-8.43083739e-01 4.92364496e-01 4.69063550e-01 4.40601170e-01
5.50953150e-01 4.31625724e-01 -3.42584103e-01 5.92743099e-01
6.21735789e-02 3.35651010e-01 3.35165590e-01 -1.31102884e+00
5.55190802e-01 4.22757000e-01 7.81810522e-01 -9.68148947e-01
-3.14576209e-01 -7.96467781e-01 -6.46357536e-01 7.65423417e-01
3.96753132e-01 -7.48617589e-01 -5.81962347e-01 1.73015463e+00
2.88071752e-01 -1.00425474e-01 9.19519737e-02 9.33431864e-01
-1.95016161e-01 6.04798913e-01 -1.50829464e-01 -2.94400662e-01
6.00400686e-01 -8.25738072e-01 -6.34427130e-01 -2.92097539e-01
2.11045802e-01 -4.49382544e-01 9.93902028e-01 4.27039027e-01
-1.38319218e+00 -2.62829930e-01 -1.19852793e+00 5.28455615e-01
-1.56900719e-01 1.97034672e-01 6.25932515e-01 5.19894958e-01
-1.08597696e+00 9.38096583e-01 -6.68680549e-01 2.19114244e-01
5.57726204e-01 7.70510852e-01 3.82451326e-01 3.27691704e-01
-1.11279154e+00 9.24571633e-01 8.99980247e-01 2.79000133e-01
-1.19348955e+00 -5.99535584e-01 -4.18679029e-01 4.62844193e-01
1.06208241e+00 -5.96258402e-01 1.59503460e+00 -1.27840865e+00
-2.03422499e+00 1.22877629e-02 3.94423753e-01 -4.86418158e-01
7.90155172e-01 -3.24167013e-01 -1.77757040e-01 -1.43558219e-01
-2.20333755e-01 5.15483260e-01 9.79582012e-01 -1.20169234e+00
-1.09047508e+00 -4.62376438e-02 6.67005420e-01 4.52573001e-01
-3.93935829e-01 -1.35994717e-01 2.37156451e-01 -2.98579723e-01
-7.05237508e-01 -6.62535846e-01 -6.03319287e-01 -2.83166349e-01
-2.91405499e-01 -2.38248259e-01 3.14810038e-01 -2.82486767e-01
1.36891544e+00 -1.77231097e+00 3.31955910e-01 3.86973977e-01
-1.58764441e-02 2.98999071e-01 -3.45701486e-01 3.70655328e-01
1.98355466e-01 8.05604178e-03 -1.97495729e-01 1.94739401e-02
3.14854026e-01 1.60319209e-01 -3.26926082e-01 5.26690781e-01
1.96747527e-01 9.61366177e-01 -1.13544488e+00 -3.61263394e-01
8.06345716e-02 -8.22035149e-02 -6.99628115e-01 7.34276026e-02
-5.98197401e-01 2.24812746e-01 -6.20992303e-01 6.23288214e-01
2.71175325e-01 -3.26233268e-01 5.55721939e-01 2.27151051e-01
-3.10050607e-01 9.49239507e-02 -1.32639015e+00 1.41695821e+00
-5.77048719e-01 3.91686291e-01 1.54971868e-01 -1.44342899e+00
6.46374404e-01 1.03417933e-01 6.94216371e-01 -7.24803329e-01
5.16482949e-01 3.11736137e-01 1.96522102e-01 -8.13546926e-02
4.70074892e-01 -2.71334410e-01 -2.43561119e-01 6.57187045e-01
-9.37704444e-02 2.05164552e-01 5.34609437e-01 -5.99441648e-01
9.30750370e-01 1.58601031e-01 5.71196795e-01 -4.27379191e-01
5.61410844e-01 -1.26330964e-02 7.51805365e-01 1.02676857e+00
-3.38660777e-01 -1.21442199e-01 7.27009654e-01 -5.12016654e-01
-8.88770700e-01 -9.35029566e-01 3.11134398e-01 1.12161481e+00
2.30894223e-01 2.53725410e-01 -3.87668282e-01 -7.56266832e-01
2.34691948e-01 9.82656300e-01 -5.83628654e-01 -2.33330280e-01
-7.33288109e-01 -7.55230129e-01 2.16211885e-01 2.98043638e-01
4.84018803e-01 -1.06131887e+00 -1.30071926e+00 5.34839690e-01
9.64718089e-02 -7.61536121e-01 -4.91563410e-01 1.26180500e-01
-6.99677944e-01 -9.53072131e-01 -7.43110299e-01 -4.63089913e-01
5.28035760e-01 8.70368630e-02 1.05345058e+00 -2.13780731e-01
-1.35888802e-02 1.50560141e-01 9.38016549e-02 -3.53149354e-01
-4.20182437e-01 7.01171383e-02 2.00895026e-01 1.23011380e-01
-8.63094106e-02 -5.63504279e-01 -7.46451557e-01 2.33070076e-01
-7.98757911e-01 -1.96888506e-01 8.55737865e-01 9.92064834e-01
4.53593373e-01 5.00074387e-01 7.26824641e-01 -2.27133647e-01
1.23732328e+00 -4.55327570e-01 -1.35637534e+00 5.91204882e-01
-6.38397038e-01 4.83747572e-01 1.08611071e+00 -5.07332683e-01
-1.03505373e+00 -2.93731540e-01 3.53463322e-01 -4.18350577e-01
1.46375433e-01 5.17396867e-01 4.82743829e-02 4.90386672e-02
4.11842197e-01 2.52350807e-01 2.12852195e-01 -2.60742635e-01
3.35064203e-01 4.01003808e-01 3.16619307e-01 -1.05029261e+00
6.28961086e-01 2.31400549e-01 1.44245431e-01 -2.56995142e-01
-9.10497606e-01 3.72240320e-02 -1.86427206e-01 -4.45076346e-01
4.38626409e-01 -3.45130175e-01 -1.24839473e+00 1.67766526e-01
-7.87537515e-01 -3.67703229e-01 -4.62852418e-01 3.97546768e-01
-8.09480727e-01 8.95269439e-02 5.35059497e-02 -9.55981433e-01
-3.34785253e-01 -1.22924292e+00 5.34286678e-01 4.77483779e-01
-7.01887086e-02 -9.79575634e-01 1.99833572e-01 -9.53063443e-02
4.33676302e-01 4.04939264e-01 9.51163292e-01 -3.33449572e-01
-3.22654575e-01 3.50674875e-02 4.58256751e-02 3.05540472e-01
3.46423835e-01 -1.57740023e-02 -2.51049459e-01 -7.82270432e-01
-2.15889007e-01 -6.51969373e-01 6.11693203e-01 6.44857466e-01
1.20233071e+00 -7.15926170e-01 1.77539010e-02 3.77367437e-01
1.73465753e+00 8.94920826e-01 1.79547578e-01 6.62599266e-01
2.09063858e-01 4.35001194e-01 7.33366907e-01 8.49421799e-01
9.09154117e-02 4.61318880e-01 7.51035750e-01 1.37524948e-01
4.71598506e-01 -7.02859014e-02 4.14632291e-01 2.02355057e-01
-1.20391250e-01 -3.24810833e-01 -8.03685904e-01 5.63391089e-01
-2.09377360e+00 -1.03880823e+00 8.65516305e-01 2.33053732e+00
7.77387738e-01 2.43625283e-01 4.12547559e-01 -7.65259191e-02
8.24979961e-01 1.44177213e-01 -1.21329987e+00 -6.57617152e-01
1.15599893e-01 9.45539102e-02 6.73639238e-01 4.41605866e-01
-1.15445518e+00 7.19244719e-01 6.90313435e+00 8.05757463e-01
-1.34238088e+00 -2.33921126e-01 8.02223861e-01 -6.90341949e-01
-1.60686418e-01 -1.43532708e-01 -5.62336326e-01 3.58613342e-01
8.76817822e-01 -6.37560785e-01 8.03509235e-01 7.61231661e-01
4.61308450e-01 -6.03953302e-02 -7.83162296e-01 7.59003878e-01
-6.18579388e-01 -1.43821788e+00 -2.19661385e-01 8.12416598e-02
1.16021752e+00 -2.22122341e-01 3.23781013e-01 3.37991178e-01
1.06420147e+00 -1.21044350e+00 7.97289491e-01 3.19340497e-01
4.47426260e-01 -1.40365565e+00 6.09275579e-01 2.48646036e-01
-7.75085092e-01 -7.84842789e-01 -2.88395196e-01 -1.60064727e-01
-5.78804091e-02 1.56434312e-01 -6.23562753e-01 6.13876641e-01
5.08006454e-01 4.22705293e-01 -4.73876856e-02 1.01057768e+00
-2.31021181e-01 4.30957109e-01 -2.82123476e-01 -5.44855297e-01
1.03102612e+00 -3.71120840e-01 6.71214402e-01 7.40830183e-01
3.90339077e-01 -3.40541363e-01 3.44303578e-01 8.16770434e-01
-8.59510154e-02 5.66179901e-02 -2.54579306e-01 -4.82419550e-01
2.66434073e-01 9.95073020e-01 -5.81134498e-01 1.35501493e-02
-9.76319313e-02 4.86648351e-01 5.95998287e-01 5.46591938e-01
-9.82586980e-01 -2.50566930e-01 8.72704923e-01 -5.10278583e-01
5.86743116e-01 -2.83916861e-01 -2.88057953e-01 -7.59624660e-01
8.35058689e-02 -1.23065615e+00 4.94995624e-01 -4.16489653e-02
-1.26728392e+00 1.94937035e-01 -2.23737229e-02 -1.27127242e+00
-3.82421583e-01 -5.54466844e-01 -6.81386054e-01 6.40561759e-01
-1.64498270e+00 -3.54681939e-01 3.33129376e-01 2.84050196e-01
4.43131298e-01 -4.00851429e-01 3.18726629e-01 -6.83231130e-02
-7.56582499e-01 5.40373862e-01 8.36500049e-01 -3.74658227e-01
3.30744058e-01 -1.39098048e+00 -9.34250355e-02 7.03871846e-01
-3.85574281e-01 2.40780383e-01 9.09000039e-01 -2.72008061e-01
-1.41715646e+00 -8.79538536e-01 1.90430224e-01 2.51083821e-01
9.19207454e-01 1.66360393e-01 -4.09211457e-01 2.08440304e-01
5.40346265e-01 -3.90277445e-01 3.75786781e-01 3.29644606e-02
2.56124139e-01 -4.60571051e-01 -1.02141631e+00 1.03832245e+00
6.97831631e-01 1.48486808e-01 -4.17206019e-01 2.35242680e-01
4.88729566e-01 -2.49677539e-01 -6.28945887e-01 4.21015710e-01
5.09910524e-01 -7.28323221e-01 1.02157116e+00 -1.07619655e+00
7.35962093e-01 -2.32160777e-01 -1.57808483e-01 -1.81404805e+00
-3.97472471e-01 -1.07643700e+00 -4.13110197e-01 5.47257423e-01
2.74098933e-01 -7.46134460e-01 5.58631957e-01 4.22474474e-01
2.06796601e-01 -1.35890710e+00 -1.12650919e+00 -1.12105358e+00
3.60964477e-01 1.05415480e-02 8.72052133e-01 5.40033996e-01
-1.78387508e-01 1.15943000e-01 -5.70528984e-01 1.02745242e-01
9.02536988e-01 6.51925504e-01 3.66309017e-01 -8.94026220e-01
-6.07164919e-01 -1.13515699e+00 1.08124688e-01 -8.87730062e-01
3.34653348e-01 -5.68148971e-01 6.45197332e-02 -1.57053518e+00
-2.36907408e-01 -4.26186472e-01 -9.60710049e-01 5.24640441e-01
-2.27180108e-01 -5.12601972e-01 1.66795939e-01 -1.55201375e-01
-9.70736980e-01 1.04789960e+00 1.39000869e+00 -4.51468766e-01
-4.08564448e-01 1.05137989e-01 -1.01444852e+00 7.05125451e-01
1.20010710e+00 -4.34870481e-01 -6.12743735e-01 -4.10029173e-01
4.21569079e-01 2.88779616e-01 1.88943464e-02 -9.19932604e-01
1.89640671e-02 -9.71090138e-01 3.04116786e-01 -3.98232192e-01
9.23952535e-02 -8.51104915e-01 -6.91316202e-02 7.09524214e-01
-5.70366859e-01 3.02936137e-01 2.77240515e-01 6.13901734e-01
-3.51633951e-02 -3.21856350e-01 9.52277601e-01 -1.23192452e-01
-5.68702698e-01 2.60022998e-01 -3.28042477e-01 2.05979228e-01
1.16268778e+00 6.77626431e-02 7.24521466e-03 -3.12619478e-01
-2.52763212e-01 7.51958191e-01 8.06634501e-02 3.76855761e-01
4.69292939e-01 -1.30152524e+00 -6.86832368e-01 -4.34518367e-01
-4.95510668e-01 -3.24277818e-01 1.04882754e-01 5.24101436e-01
-3.02224785e-01 5.70116103e-01 -5.35834253e-01 -1.46474928e-01
-8.05386066e-01 4.77234095e-01 7.98947692e-01 -9.31492269e-01
-2.73371607e-01 4.39140677e-01 -6.72744587e-02 -3.72007310e-01
2.47562304e-01 -6.58795461e-02 -2.34406576e-01 1.97681621e-01
4.63484466e-01 6.63052917e-01 -2.63988584e-01 2.92588994e-02
-2.05111623e-01 1.71347663e-01 -3.77664268e-02 -3.39594156e-01
1.37739527e+00 1.02414437e-01 1.48738742e-01 2.32171625e-01
9.02577400e-01 -4.16249245e-01 -1.69401240e+00 -1.10360809e-01
9.34576839e-02 -5.98159015e-01 2.75274456e-01 -9.17195916e-01
-1.05065024e+00 4.90814418e-01 5.18511236e-01 1.95907474e-01
1.30444157e+00 -7.23328531e-01 6.01960957e-01 8.00371289e-01
3.78365189e-01 -1.62447679e+00 2.20487073e-01 5.24929404e-01
9.54108953e-01 -1.04255223e+00 9.67957312e-04 5.56286871e-01
-6.21556520e-01 1.21723163e+00 6.88527346e-01 -3.07690561e-01
3.12474072e-01 3.06167696e-02 -7.36917630e-02 7.88092464e-02
-9.76789474e-01 -2.23434448e-01 1.44559726e-01 1.13744214e-01
-6.44798428e-02 1.76976129e-01 -4.91661310e-01 1.66307837e-01
3.56105552e-03 -5.81734404e-02 4.68667448e-01 1.27495372e+00
-5.39624035e-01 -1.14153063e+00 -3.58105034e-01 5.13101995e-01
-6.21378362e-01 1.16961218e-01 -1.07375748e-01 5.87619424e-01
-2.51110494e-01 1.01537275e+00 -1.09258212e-01 -3.57186832e-02
2.15159714e-01 -1.96153536e-01 5.29001832e-01 -2.06624083e-02
-5.96814871e-01 3.27438191e-02 2.73715705e-02 -6.77155554e-01
-2.51656651e-01 -6.33854747e-01 -1.16836834e+00 -2.96700388e-01
1.23157732e-01 5.82472570e-02 4.62959856e-01 1.15917742e+00
4.30190533e-01 8.71671736e-01 9.19289708e-01 -6.22764289e-01
-1.37420845e+00 -4.38492119e-01 -4.53142017e-01 2.65130773e-03
5.27776659e-01 -9.84440386e-01 -1.21614143e-01 -6.59427822e-01] | [4.221834659576416, 2.4063339233398438] |
219c311f-7a25-4272-8313-73fc76e5523b | alap-ae-as-lite-as-possible-auto-encoder | 2203.10363 | null | https://arxiv.org/abs/2203.10363v2 | https://arxiv.org/pdf/2203.10363v2.pdf | Towards Device Efficient Conditional Image Generation | We present a novel algorithm to reduce tensor compute required by a conditional image generation autoencoder without sacrificing quality of photo-realistic image generation. Our method is device agnostic, and can optimize an autoencoder for a given CPU-only, GPU compute device(s) in about normal time it takes to train an autoencoder on a generic workstation. We achieve this via a two-stage novel strategy where, first, we condense the channel weights, such that, as few as possible channels are used. Then, we prune the nearly zeroed out weight activations, and fine-tune the autoencoder. To maintain image quality, fine-tuning is done via student-teacher training, where we reuse the condensed autoencoder as the teacher. We show performance gains for various conditional image generation tasks: segmentation mask to face images, face images to cartoonization, and finally CycleGAN-based model over multiple compute devices. We perform various ablation studies to justify the claims and design choices, and achieve real-time versions of various autoencoders on CPU-only devices while maintaining image quality, thus enabling at-scale deployment of such autoencoders. | ['Gaurav Bharaj', 'Nisarg A. Shah'] | 2022-03-19 | null | null | null | null | ['conditional-image-generation'] | ['computer-vision'] | [ 2.60035545e-01 2.25962877e-01 3.83343428e-01 -2.75293797e-01
-5.02621710e-01 -6.76380098e-01 3.07723284e-01 -4.70968455e-01
-3.57498139e-01 4.29760337e-01 1.85240768e-02 -4.45664406e-01
5.01750052e-01 -9.10032272e-01 -9.22233760e-01 -8.68309140e-01
1.69482350e-01 3.75086963e-01 -7.15316162e-02 -6.61389381e-02
-1.57965183e-01 6.72300041e-01 -1.64209986e+00 4.63037789e-01
2.89187253e-01 9.41835105e-01 9.57334936e-02 1.32553220e+00
2.97882855e-01 6.55289769e-01 -7.64300227e-01 -7.27558732e-01
6.49991572e-01 -4.20832157e-01 -7.37091362e-01 4.16576296e-01
5.19740283e-01 -7.74859786e-01 -2.89872438e-01 7.90379465e-01
6.11626089e-01 -1.76352933e-01 3.94308120e-01 -1.22664976e+00
-5.56006074e-01 6.30314410e-01 -5.48100531e-01 1.73002686e-02
-1.02244429e-01 6.92547262e-01 6.44657433e-01 -5.98675966e-01
4.69467074e-01 1.14304566e+00 5.50340712e-01 7.62463093e-01
-1.41434157e+00 -8.90981138e-01 -3.50426376e-01 -3.37370187e-01
-1.32275462e+00 -5.21346271e-01 4.73982215e-01 -3.36044490e-01
1.13024783e+00 1.07047550e-01 9.67083275e-01 1.04773009e+00
2.21284971e-01 4.98819143e-01 1.14932179e+00 -5.01237512e-01
2.78605044e-01 2.74957735e-02 -3.25273007e-01 9.90658760e-01
1.08851589e-01 -3.38075235e-02 -2.76658356e-01 -1.79310784e-01
1.19073057e+00 -1.33075774e-01 -1.17796310e-03 2.88223535e-01
-1.04163122e+00 7.78348923e-01 3.43755186e-01 9.98183861e-02
-3.75554770e-01 5.47998071e-01 3.70154291e-01 2.14123517e-01
3.41779619e-01 4.83166158e-01 -4.81320590e-01 -5.93843572e-02
-1.21531427e+00 1.92784131e-01 8.51307213e-01 1.16420293e+00
1.11003888e+00 5.02067447e-01 -1.99341238e-01 4.36151832e-01
-1.81549102e-01 3.92027080e-01 5.36618054e-01 -1.12915218e+00
6.79598516e-03 2.16287032e-01 -3.23450118e-01 -5.80038548e-01
-1.06902577e-01 -1.79178819e-01 -9.72317457e-01 4.21022922e-01
1.22746505e-01 -7.50608146e-01 -1.15768170e+00 1.31383681e+00
5.36882460e-01 3.37853581e-01 -5.03912009e-02 9.59789574e-01
5.67291439e-01 9.04587924e-01 7.75020719e-02 2.04757407e-01
1.75570619e+00 -1.19693184e+00 -2.97890484e-01 -3.18053253e-02
4.09908861e-01 -9.85481024e-01 1.09250045e+00 4.90668833e-01
-1.35844135e+00 -7.80744612e-01 -1.43023789e+00 -2.98035830e-01
-1.12152644e-01 4.82917428e-01 1.14228904e+00 1.03345823e+00
-1.39813995e+00 7.32614577e-01 -1.09442270e+00 2.72834767e-02
4.45640922e-01 6.26757205e-01 -2.01065600e-01 1.39985904e-01
-5.60307741e-01 1.43575862e-01 6.10850751e-01 -1.98153764e-01
-1.22371399e+00 -9.19940412e-01 -6.37528419e-01 2.61863858e-01
-1.17892087e-01 -1.09655118e+00 1.24359632e+00 -1.44612193e+00
-2.00819087e+00 8.09546173e-01 1.02830052e-01 -5.08860946e-01
1.08629063e-01 2.04253674e-01 -4.50997055e-01 3.36626947e-01
-2.03913882e-01 1.16932058e+00 1.53620505e+00 -1.03650200e+00
-6.38359785e-01 -1.25976950e-02 1.92170262e-01 5.32543324e-02
-5.42473316e-01 -4.60129008e-02 -7.55771756e-01 -6.01621509e-01
-2.77035445e-01 -1.16836178e+00 -2.90661037e-01 -2.18876034e-01
-3.99766862e-01 1.60921261e-01 9.83393013e-01 -5.88803828e-01
7.54331946e-01 -2.23278093e+00 -1.99243948e-01 2.57408470e-01
4.67543334e-01 1.62775204e-01 -1.64526552e-01 -3.56517769e-02
-2.65918612e-01 2.83925869e-02 -1.32086977e-01 -3.69126379e-01
-2.65267909e-01 3.12406600e-01 -3.88967454e-01 3.73939574e-01
5.21014512e-01 8.32529902e-01 -4.40661639e-01 -5.01160622e-01
1.96158335e-01 7.11636007e-01 -1.14621294e+00 3.16989690e-01
-2.51997888e-01 3.50041389e-01 -1.93946108e-01 4.92808908e-01
8.53375733e-01 -5.43424726e-01 3.18400294e-01 -5.30975342e-01
6.38538525e-02 -1.46731362e-01 -1.07092154e+00 1.50578582e+00
-8.11058521e-01 7.22357392e-01 3.74693662e-01 -5.16214192e-01
4.67941254e-01 2.79920697e-01 2.98081666e-01 -5.30685008e-01
2.67838687e-01 -7.11663365e-02 -1.06781796e-02 -2.10182473e-01
6.58937693e-01 -5.46953939e-02 4.52436395e-02 7.02551544e-01
4.97159511e-01 -4.41053331e-01 1.47580236e-01 2.49468341e-01
1.00642741e+00 1.26958743e-01 -1.81391045e-01 -2.92109668e-01
4.78402935e-02 6.04858175e-02 9.25058499e-02 5.61378360e-01
2.03240067e-01 7.10390687e-01 4.77022469e-01 -7.21541762e-01
-1.83247471e+00 -6.54288709e-01 1.26472831e-01 1.27988362e+00
-2.42164940e-01 -5.46950102e-01 -1.49042714e+00 -4.89284426e-01
-4.51891333e-01 3.14396739e-01 -5.94371796e-01 2.13257506e-01
-8.18104446e-01 -1.10528278e+00 9.96519446e-01 5.56116462e-01
6.75757587e-01 -8.29491794e-01 -8.41953814e-01 1.52882591e-01
3.47644120e-01 -1.24274039e+00 -5.35529494e-01 1.37339920e-01
-8.60870600e-01 -7.65775144e-01 -6.18175805e-01 -6.93661094e-01
9.34806764e-01 4.29220647e-02 1.20622098e+00 3.45858753e-01
-5.12681246e-01 1.61740944e-01 -4.52155340e-03 9.95267183e-03
-2.96210855e-01 1.32406466e-02 -1.17221065e-01 1.97065249e-02
-4.84439842e-02 -6.87718272e-01 -9.84103680e-01 -1.04070060e-01
-1.04597771e+00 2.81409144e-01 7.02013075e-01 1.00506043e+00
6.63712680e-01 3.14326406e-01 -2.55696744e-01 -8.04758132e-01
6.34619176e-01 -1.47215515e-01 -8.26152623e-01 -1.00592963e-01
-4.93882507e-01 3.29373986e-01 9.28614080e-01 -5.39015889e-01
-1.16485846e+00 4.08195049e-01 -4.06597316e-01 -5.74562848e-01
-1.20804533e-01 -1.77248687e-01 7.68624619e-02 -2.97821879e-01
8.76646042e-01 1.74818799e-01 -2.14550108e-01 -2.37875395e-02
6.07282877e-01 4.47181702e-01 6.05232894e-01 -8.76504600e-01
6.44168079e-01 5.85818291e-01 -6.89112619e-02 -8.11942279e-01
-2.87994385e-01 1.43711418e-01 -1.76874205e-01 -1.77427549e-02
1.04710615e+00 -1.19543409e+00 -1.27656329e+00 3.20533365e-01
-1.28931999e+00 -6.97368979e-01 -3.10532004e-01 2.16243446e-01
-4.14685160e-01 -3.64371985e-02 -9.08571780e-01 -4.78751928e-01
-6.75308824e-01 -1.44615471e+00 1.30835938e+00 5.47416329e-01
9.75800231e-02 -6.12475932e-01 -1.15002081e-01 2.45414719e-01
4.31329370e-01 1.29179046e-01 7.75417745e-01 -3.84631939e-02
-7.68323064e-01 2.73746494e-02 -3.98419827e-01 4.73548234e-01
-4.13075715e-01 2.60032296e-01 -1.14936256e+00 -5.55163205e-01
-6.01011775e-02 -4.79686201e-01 6.11060619e-01 2.48632997e-01
1.78043795e+00 -6.30711377e-01 -1.73561558e-01 1.15151358e+00
1.49409044e+00 1.55525589e-02 7.05239296e-01 -1.44050121e-01
8.24033737e-01 1.97341859e-01 -1.40619159e-01 4.22584802e-01
2.31196031e-01 4.12606746e-01 1.94658250e-01 -4.53737110e-01
-3.81823242e-01 -3.00545722e-01 1.54577151e-01 9.56219137e-01
-3.91014278e-01 -1.46469101e-01 -6.09668911e-01 3.95032823e-01
-1.33503222e+00 -5.79309583e-01 1.43944114e-01 1.88417041e+00
8.83768141e-01 -1.40506223e-01 1.80173274e-02 -8.95894840e-02
5.47727764e-01 -2.48304103e-02 -3.59773338e-01 -5.99011421e-01
-2.30985936e-02 1.01202416e+00 8.80321145e-01 4.99125957e-01
-9.14285421e-01 1.28808987e+00 7.06518364e+00 9.48522985e-01
-1.48038626e+00 5.61255038e-01 1.03720248e+00 -2.62369245e-01
-2.49745741e-01 1.01888217e-01 -7.47981369e-01 4.76697981e-01
1.17792928e+00 3.45456004e-01 9.04165566e-01 1.30688870e+00
-2.56652921e-01 5.68120703e-02 -9.87758756e-01 1.21837294e+00
-1.60987917e-02 -1.57938635e+00 -3.27339098e-02 1.14405705e-02
1.12347698e+00 -1.40743628e-01 8.84686410e-02 2.26990357e-01
7.23359227e-01 -1.05672240e+00 5.91474295e-01 -1.10990955e-02
1.05601084e+00 -9.84220982e-01 3.95903468e-01 1.05053447e-02
-9.69028533e-01 1.31701916e-01 -3.61544430e-01 -6.46367818e-02
-1.08107038e-01 6.78851902e-01 -1.00767899e+00 -5.32203130e-02
8.97483289e-01 -1.09979108e-01 -5.44870138e-01 3.96661490e-01
-1.97700396e-01 7.91778564e-01 -4.48335856e-01 -2.88080592e-02
2.51409054e-01 -3.14782299e-02 2.33059470e-02 1.37629020e+00
3.31996262e-01 4.70814854e-01 -9.11725238e-02 7.70890892e-01
-4.54370320e-01 -1.70300275e-01 -2.94055849e-01 -1.64472252e-01
3.73043537e-01 1.62240636e+00 -1.01407051e+00 -6.60905182e-01
-1.07630007e-01 1.66727769e+00 1.29204392e-01 4.04477209e-01
-9.19216514e-01 -4.96257037e-01 7.13717520e-01 5.18037267e-02
3.27337533e-01 -3.21764588e-01 -3.52093548e-01 -1.19295907e+00
-2.13285595e-01 -1.00365472e+00 4.12345752e-02 -9.49076116e-01
-5.97102165e-01 9.12012219e-01 -3.67186904e-01 -8.70001316e-01
-5.33673406e-01 -6.63790524e-01 -3.58571321e-01 8.24744940e-01
-1.04528129e+00 -1.27218270e+00 -3.58639359e-01 7.94983923e-01
4.69105095e-01 -2.83931196e-01 8.55107248e-01 4.75201011e-01
-5.48687398e-01 8.24531436e-01 -2.49387249e-01 1.42850325e-01
3.36935103e-01 -1.01515710e+00 6.66810095e-01 9.43510532e-01
6.92285821e-02 7.39042938e-01 5.66410065e-01 -5.16967297e-01
-1.76224911e+00 -1.12092829e+00 2.42434040e-01 -2.16433644e-01
3.18237752e-01 -5.91812372e-01 -3.40409994e-01 8.76969993e-01
6.88603222e-01 2.54811078e-01 5.23746014e-01 -1.20510630e-01
-1.92930505e-01 -1.22754931e-01 -1.13535166e+00 7.96223164e-01
7.88982213e-01 -5.02090514e-01 2.46201739e-01 4.87813085e-01
8.63924205e-01 -8.50063026e-01 -8.81939113e-01 -8.01270008e-02
6.21361792e-01 -8.92010212e-01 8.95744264e-01 -4.62613612e-01
8.74464095e-01 -4.19672787e-01 6.09293804e-02 -1.29391003e+00
-3.69344473e-01 -1.03710008e+00 -2.37286538e-01 8.87351692e-01
3.82170290e-01 -2.34069914e-01 1.18530715e+00 4.99859244e-01
-1.42707542e-01 -6.55597627e-01 -4.84486461e-01 -2.14214623e-01
-2.63931364e-01 -4.22984809e-01 1.11696088e+00 7.41512299e-01
-5.75902104e-01 3.55418682e-01 -5.81042647e-01 3.42451185e-01
4.98599052e-01 5.00133354e-03 9.96852815e-01 -4.22246307e-01
-9.59636986e-01 -1.23400472e-01 -2.84550488e-01 -1.17893541e+00
-1.27486929e-01 -5.89439750e-01 -1.16189584e-01 -6.57879591e-01
6.01213314e-02 -3.62787336e-01 3.82359385e-01 5.96806407e-01
2.09175097e-03 8.25509191e-01 2.41339505e-01 1.57138601e-01
-1.23642258e-01 1.10637859e-01 1.31846285e+00 -1.32985041e-01
-1.28851280e-01 -4.92801905e-01 -7.76726007e-01 6.35165393e-01
5.96745193e-01 -4.62438673e-01 -4.83945668e-01 -8.41055691e-01
1.72852069e-01 3.04163098e-02 4.85093623e-01 -1.20031810e+00
3.16092908e-01 1.55835096e-02 9.50928986e-01 -8.01701248e-02
3.79982948e-01 -7.36007273e-01 4.30521011e-01 3.97332460e-01
8.49314258e-02 2.48228401e-01 2.79060453e-01 3.64902318e-02
4.30585742e-02 -4.60453182e-02 9.09842432e-01 -2.67212182e-01
-5.14807343e-01 4.67135310e-01 -2.70099342e-01 -3.40031147e-01
7.98612714e-01 -1.10576510e-01 -1.38353005e-01 -2.60150552e-01
-7.04727769e-01 -2.76328564e-01 5.33579946e-01 2.00986471e-02
4.47355270e-01 -1.19174147e+00 -5.62459648e-01 6.01990998e-01
-2.62985021e-01 7.00515658e-02 4.30245966e-01 2.57972807e-01
-1.18916833e+00 2.27654159e-01 -4.40848321e-01 -5.84437311e-01
-9.75508928e-01 5.85175455e-01 2.26958200e-01 -1.52463496e-01
-6.11959398e-01 1.07963192e+00 2.72066772e-01 -1.88378885e-01
-2.82015353e-01 -4.60269153e-02 4.73555088e-01 -3.93545091e-01
6.74528062e-01 1.58680514e-01 4.00855273e-01 -2.00818926e-01
6.31787628e-02 3.60342950e-01 -7.91182071e-02 -4.63413328e-01
1.11010206e+00 1.50418937e-01 -8.94357562e-02 -5.23936093e-01
1.20429051e+00 1.61782708e-02 -1.64730537e+00 3.71447325e-01
-9.57698107e-01 -1.88355416e-01 4.25179780e-01 -5.64664304e-01
-1.62354243e+00 8.08228076e-01 7.77423024e-01 7.37958476e-02
1.47182024e+00 -1.89202875e-01 1.00676453e+00 2.07901672e-01
2.70240217e-01 -1.08669651e+00 1.05760425e-01 2.00498164e-01
4.90492940e-01 -7.73679852e-01 5.70432656e-02 -3.48077655e-01
-7.01723218e-01 1.09926820e+00 7.00147808e-01 -4.47515100e-01
4.80037093e-01 1.04044068e+00 -1.80512086e-01 -1.58464089e-01
-5.84392726e-01 6.75611570e-02 -1.23555250e-01 4.72691715e-01
3.27076286e-01 3.50391805e-01 3.70387398e-02 3.99715990e-01
-7.84343541e-01 6.33475482e-02 3.98313761e-01 6.82790697e-01
1.59841269e-01 -1.00954986e+00 -4.58206058e-01 3.36220443e-01
-6.02043509e-01 -3.12637269e-01 6.80458248e-02 6.07946277e-01
5.41563034e-01 5.53606093e-01 4.13275748e-01 -5.71388900e-01
-1.10859051e-01 -1.62361458e-01 7.77781427e-01 -3.08734566e-01
-8.70621443e-01 2.16070801e-01 -2.91787237e-02 -6.11041188e-01
-7.88537338e-02 -8.07736367e-02 -9.31344986e-01 -8.81131113e-01
-1.79124907e-01 -1.91970930e-01 1.00172937e+00 6.97437286e-01
6.96419179e-01 6.66300178e-01 6.21183872e-01 -1.21817791e+00
-8.58045146e-02 -6.56185925e-01 -3.05205584e-01 2.10568994e-01
1.19287878e-01 1.60961729e-02 5.22851534e-02 6.30223632e-01] | [11.334035873413086, -0.5330678224563599] |
32cbdf79-df74-4dd2-8cf8-6b39277dc79f | learning-to-annotate-modularizing-data | 1911.01352 | null | https://arxiv.org/abs/1911.01352v3 | https://arxiv.org/pdf/1911.01352v3.pdf | Learning from Explanations with Neural Execution Tree | While deep neural networks have achieved impressive performance on a range of NLP tasks, these data-hungry models heavily rely on labeled data, which restricts their applications in scenarios where data annotation is expensive. Natural language (NL) explanations have been demonstrated very useful additional supervision, which can provide sufficient domain knowledge for generating more labeled data over new instances, while the annotation time only doubles. However, directly applying them for augmenting model learning encounters two challenges: (1) NL explanations are unstructured and inherently compositional, which asks for a modularized model to represent their semantics, (2) NL explanations often have large numbers of linguistic variants, resulting in low recall and limited generalization ability. In this paper, we propose a novel Neural Execution Tree (NExT) framework to augment training data for text classification using NL explanations. After transforming NL explanations into executable logical forms by semantic parsing, NExT generalizes different types of actions specified by the logical forms for labeling data instances, which substantially increases the coverage of each NL explanation. Experiments on two NLP tasks (relation extraction and sentiment analysis) demonstrate its superiority over baseline methods. Its extension to multi-hop question answering achieves performance gain with light annotation effort. | ['Zhiyuan Liu', 'Yujia Qin', 'Ziqi Wang', 'Xiang Ren', 'Leonardo Neves', 'Qinyuan Ye', 'Wenxuan Zhou', 'Jun Yan'] | 2019-11-04 | null | https://openreview.net/forum?id=rJlUt0EYwS | https://openreview.net/pdf?id=rJlUt0EYwS | iclr-2020-1 | ['multi-hop-question-answering'] | ['knowledge-base'] | [ 5.68340421e-01 1.04405904e+00 -7.77580619e-01 -8.05955946e-01
-6.24411464e-01 -5.19623697e-01 6.02766871e-01 3.21303129e-01
3.20901386e-02 1.04193473e+00 3.45082939e-01 -6.80201650e-01
1.69714510e-01 -9.14610386e-01 -8.08380783e-01 -1.11371301e-01
3.66897017e-01 6.19491279e-01 1.27923012e-01 -3.00511390e-01
-1.14173010e-01 9.08677205e-02 -1.62599933e+00 8.12179327e-01
1.24859011e+00 1.13037634e+00 -2.69653369e-02 2.66490877e-01
-9.46448803e-01 1.00330007e+00 -5.88396966e-01 -7.78481960e-01
-5.27116023e-02 -4.43965822e-01 -1.19071519e+00 7.69192874e-02
4.28562820e-01 -2.84666151e-01 -9.68840905e-03 8.57214153e-01
-1.72912121e-01 1.11199141e-01 6.83902264e-01 -1.50968111e+00
-1.15542519e+00 1.23003531e+00 -2.58258641e-01 -1.98060483e-01
4.37113494e-01 -6.49502799e-02 1.41348529e+00 -8.54129553e-01
4.07735169e-01 1.31908047e+00 5.91488481e-01 1.01333725e+00
-1.02100074e+00 -6.87337697e-01 5.75422764e-01 1.53853491e-01
-6.69023514e-01 -1.35863885e-01 4.94185418e-01 -4.32811044e-02
1.36132157e+00 2.00542852e-01 3.65181863e-01 1.16103554e+00
-3.55247200e-01 1.07083929e+00 7.10395098e-01 -5.08045554e-01
1.54663965e-01 2.55154669e-01 5.94115198e-01 7.42346168e-01
4.58712488e-01 -2.06803009e-01 -5.33520818e-01 -1.47603406e-02
4.43740994e-01 1.89253286e-01 -2.27252409e-01 4.72190641e-02
-9.89016354e-01 1.11006606e+00 6.04819179e-01 1.26010522e-01
-2.51579314e-01 6.72855675e-02 4.53496754e-01 2.51509368e-01
4.12117660e-01 8.32961202e-01 -1.09109783e+00 1.53522417e-01
-3.53340298e-01 2.10889697e-01 9.80728507e-01 1.18638849e+00
8.96709621e-01 9.79599729e-02 -1.09824277e-01 7.20653951e-01
1.95919767e-01 3.27803642e-01 7.55748808e-01 -8.42455328e-01
8.34522724e-01 1.28517377e+00 -6.53730705e-02 -6.99573338e-01
-5.80106139e-01 -3.01862717e-01 -9.40543473e-01 -3.66222978e-01
3.62301081e-01 -3.12130094e-01 -9.11109984e-01 1.80143774e+00
3.93424720e-01 -6.44753650e-02 5.20955622e-01 6.91596508e-01
1.20541739e+00 7.01858222e-01 2.47285321e-01 -1.49725243e-01
1.55146813e+00 -1.34228134e+00 -8.33169699e-01 -7.71587431e-01
9.62133825e-01 -1.58724114e-01 1.49945176e+00 2.85061061e-01
-7.69887447e-01 -5.53162277e-01 -9.95018721e-01 -4.02888626e-01
-4.85682607e-01 3.79050002e-02 1.26412082e+00 4.05597538e-01
-6.16686285e-01 3.73722672e-01 -5.83426237e-01 -1.28361031e-01
5.34131348e-01 5.47712505e-01 -2.62655139e-01 -1.44817129e-01
-1.39828146e+00 5.69998264e-01 7.33623683e-01 -9.44816023e-02
-2.55732268e-01 -7.47591734e-01 -1.28825831e+00 3.75422627e-01
8.29252481e-01 -7.28594542e-01 1.40037048e+00 -1.09240508e+00
-1.41136372e+00 4.96240556e-01 -4.37608033e-01 -7.25911021e-01
1.29588783e-01 -4.18217480e-01 -4.69735801e-01 -3.09870783e-02
3.38713974e-01 1.09620595e+00 5.63505352e-01 -1.19712901e+00
-6.76439881e-01 -3.08305949e-01 4.51515704e-01 1.42421827e-01
-6.21555865e-01 -3.42623919e-01 -1.32185593e-01 -5.62572122e-01
2.10053548e-01 -6.45175159e-01 -4.09930527e-01 -3.63718837e-01
-6.81181073e-01 -6.46353424e-01 7.39468813e-01 -3.16020221e-01
1.00584888e+00 -1.84311080e+00 -1.90789804e-01 -2.34740123e-01
2.67783254e-01 2.04791397e-01 -6.99812025e-02 1.55557603e-01
-2.45861486e-02 4.78438258e-01 -3.91730189e-01 -2.49212861e-01
1.61882788e-01 6.93453491e-01 -1.07124770e+00 -5.32630086e-01
6.83653593e-01 1.20768070e+00 -8.26246321e-01 -3.34385753e-01
-1.08769841e-01 1.81748345e-02 -7.19066262e-01 3.12283814e-01
-1.03846908e+00 3.58134240e-01 -6.40584648e-01 6.50695026e-01
2.81252414e-01 -7.18486607e-01 2.28476167e-01 -1.51655143e-02
5.37771285e-01 6.80769265e-01 -8.90178204e-01 1.42385268e+00
-6.81672812e-01 3.61752033e-01 -5.41820586e-01 -1.07158411e+00
1.05889666e+00 3.02695811e-01 -1.33477837e-01 -3.43293846e-01
8.03857285e-04 2.93767840e-01 -1.98130116e-01 -7.37648427e-01
4.49472845e-01 -2.99077898e-01 -2.71941036e-01 6.28058612e-01
6.28263950e-02 -1.46093339e-01 3.34124953e-01 2.73793578e-01
1.03359520e+00 1.43166795e-01 5.60450971e-01 1.48605496e-01
5.67592740e-01 5.24415970e-01 6.73110366e-01 6.97767317e-01
2.70879000e-01 3.06719363e-01 7.24396884e-01 -9.10538018e-01
-6.64251745e-01 -6.67698562e-01 -5.98966563e-03 1.14022481e+00
2.76423216e-01 -4.11268145e-01 -6.32599533e-01 -1.39239216e+00
-3.40557918e-02 1.06385422e+00 -4.86571491e-01 -2.46137723e-01
-5.77742398e-01 -7.48034418e-01 4.79183346e-01 9.61499691e-01
5.71018219e-01 -1.38465834e+00 -4.90085185e-01 2.03941882e-01
-4.51100975e-01 -1.56236315e+00 -2.33437985e-01 4.78552729e-01
-1.11297393e+00 -1.08423436e+00 -2.96062813e-03 -8.03241193e-01
9.49240506e-01 9.78111327e-02 1.28170228e+00 3.97002369e-01
1.13387235e-01 -1.11192897e-01 -6.18386567e-01 -6.13906443e-01
-4.76944149e-01 3.16025823e-01 4.74156514e-02 -1.95355818e-01
8.59189689e-01 -2.99044609e-01 -2.37557232e-01 1.78394616e-01
-1.08338344e+00 2.99904019e-01 6.86740279e-01 1.01484060e+00
5.90236545e-01 -2.80087236e-02 1.02775967e+00 -1.52383840e+00
5.85697353e-01 -4.77721214e-01 -2.80562788e-01 3.00604582e-01
-8.39685917e-01 5.24504602e-01 1.11384749e+00 -4.57822949e-01
-1.38136697e+00 3.08966334e-03 -1.00914828e-01 1.93639342e-02
-5.13165474e-01 6.27566814e-01 -3.69936645e-01 6.33685291e-01
8.34405899e-01 -8.76133516e-02 -1.63564369e-01 -3.12138110e-01
6.52701914e-01 5.71398735e-01 5.97587943e-01 -6.42059565e-01
7.11094856e-01 2.94296443e-01 -1.84008300e-01 -3.76818627e-01
-1.79983819e+00 -2.73507219e-02 -5.25231063e-01 5.46441734e-01
8.98687363e-01 -6.61360204e-01 -5.32330811e-01 -3.05520952e-01
-1.33317637e+00 -3.87929499e-01 -5.63542664e-01 3.43248069e-01
-2.31326863e-01 2.69198388e-01 -5.87456942e-01 -6.07827902e-01
-3.77600133e-01 -9.36587572e-01 1.11006629e+00 3.80579472e-01
-6.10663950e-01 -9.45369303e-01 -4.32726860e-01 7.21403003e-01
1.40679497e-02 5.25024943e-02 1.42829955e+00 -1.32192564e+00
-4.83491004e-01 -2.93343008e-01 -3.59920025e-01 2.58374572e-01
1.28536656e-01 -3.91080856e-01 -1.01412129e+00 3.32191616e-01
-1.08038917e-01 -7.95372844e-01 7.28351057e-01 1.07016481e-01
1.46166217e+00 -8.54856372e-01 -3.79091293e-01 4.28922206e-01
9.17443156e-01 1.38323754e-02 1.58646390e-01 2.27530017e-01
6.88572228e-01 9.08616543e-01 7.63543487e-01 1.42907470e-01
5.71829021e-01 3.51083130e-01 6.11980557e-01 -2.14287996e-01
2.56088655e-02 -5.58777869e-01 2.94679910e-01 4.71109033e-01
3.30438256e-01 -4.22432214e-01 -7.83678651e-01 4.67730284e-01
-2.04382086e+00 -6.19276762e-01 -4.35034454e-01 1.69222999e+00
1.05806088e+00 1.66305229e-01 -2.59481847e-01 1.15974903e-01
4.87487793e-01 -1.20013379e-01 -9.39519167e-01 -4.76808459e-01
-1.02621756e-01 2.39380106e-01 5.59357144e-02 4.73558515e-01
-9.20300305e-01 1.20512164e+00 5.51446772e+00 5.85582852e-01
-7.42486417e-01 -2.49848627e-02 6.68259561e-01 2.40600139e-01
-5.49231112e-01 6.95550963e-02 -1.12695801e+00 4.03170809e-02
7.88793087e-01 1.09704740e-01 9.56035256e-02 1.09476507e+00
-2.34014332e-01 3.05021256e-01 -1.46359849e+00 5.83316386e-01
2.09497586e-02 -1.54977715e+00 7.44512796e-01 -1.43439382e-01
7.95670509e-01 -3.22713107e-01 -7.68258050e-02 7.25738168e-01
6.29421473e-01 -1.16565871e+00 4.23533171e-01 -1.96331620e-01
7.88308442e-01 -5.45725524e-01 1.09660256e+00 5.19469500e-01
-9.82904196e-01 -4.29662466e-01 -5.28209865e-01 -4.14005071e-01
1.81565374e-01 5.95262825e-01 -1.09836709e+00 4.72668380e-01
4.95299131e-01 6.50229216e-01 -4.66720253e-01 8.68930668e-02
-1.08902824e+00 6.07825875e-01 -1.49538204e-01 -2.74401933e-01
3.97990465e-01 1.23166598e-01 7.88047351e-03 1.00545907e+00
1.85463265e-01 4.18139338e-01 1.95430964e-01 9.85460758e-01
-5.44819713e-01 1.45556062e-01 -5.41580856e-01 -1.72389939e-01
4.50884819e-01 1.23961079e+00 -6.68998897e-01 -6.95512414e-01
-5.87406039e-01 7.59900451e-01 4.55453753e-01 3.47838014e-01
-7.69001126e-01 -2.40806669e-01 3.62149745e-01 -4.77914363e-02
-3.10474187e-02 2.61393785e-01 -6.82171822e-01 -1.36792839e+00
1.90801039e-01 -8.55885148e-01 7.79058218e-01 -8.22403550e-01
-1.36326039e+00 8.86609316e-01 -6.85220584e-02 -9.07207429e-01
-6.10416234e-01 -8.21195543e-01 -4.67401296e-01 6.14306509e-01
-1.54264796e+00 -1.15986872e+00 -4.37722892e-01 3.25270772e-01
1.02779198e+00 -2.31941536e-01 1.00712740e+00 -1.56936049e-01
-4.64949548e-01 5.80505252e-01 -5.81310630e-01 1.02778353e-01
4.88673002e-01 -1.46676898e+00 5.13538420e-01 6.75249517e-01
5.25564134e-01 6.59670949e-01 5.51670134e-01 -6.00706100e-01
-1.25907636e+00 -1.25558889e+00 1.24721396e+00 -5.86566389e-01
6.43200397e-01 -2.37703592e-01 -1.23019910e+00 1.04820132e+00
6.91894367e-02 1.04192579e-02 9.23754930e-01 4.70676452e-01
-6.02701306e-01 1.51643157e-01 -9.74200785e-01 7.20302701e-01
1.04311872e+00 -3.12718600e-01 -9.83629823e-01 5.69849610e-01
1.45748246e+00 -2.64055997e-01 -5.12558222e-01 6.05696678e-01
1.85241044e-01 -7.10801899e-01 6.97919309e-01 -1.15721226e+00
1.03456783e+00 -2.63573229e-01 -9.59753841e-02 -1.06875575e+00
1.72908291e-01 -4.08616781e-01 -5.12255847e-01 1.26198328e+00
1.07195687e+00 -6.50133669e-01 9.42449808e-01 1.08872879e+00
-2.95222133e-01 -1.09516287e+00 -4.51327592e-01 -6.21682644e-01
-1.54016688e-01 -6.26497388e-01 1.04939437e+00 1.07535982e+00
3.55620176e-01 1.05819118e+00 5.13724387e-02 1.70559362e-01
2.02286154e-01 4.36780483e-01 7.76656151e-01 -1.26917565e+00
-3.64154130e-01 -2.33579651e-01 1.02191560e-01 -1.42121339e+00
5.93928635e-01 -1.08713722e+00 3.71813849e-02 -1.67725527e+00
1.33442745e-01 -4.98438388e-01 2.14570910e-01 1.08813763e+00
-5.32917440e-01 -8.73476081e-03 -1.84297979e-01 1.63134053e-01
-5.31353235e-01 6.23941720e-01 1.05581272e+00 -1.84875041e-01
-2.53582597e-01 -3.71978730e-02 -1.20016217e+00 1.03337336e+00
7.48688757e-01 -3.52394074e-01 -7.01482654e-01 -7.35351205e-01
4.69549686e-01 4.43501659e-02 1.44837409e-01 -4.78640050e-01
-1.08366022e-02 -2.67848819e-01 1.76991910e-01 -3.21347386e-01
6.98335916e-02 -8.89030159e-01 -3.48271519e-01 3.20296526e-01
-8.27886879e-01 -1.06756307e-01 1.26593113e-01 6.55566871e-01
-5.03554225e-01 -4.80146199e-01 2.76634961e-01 -1.61609396e-01
-6.12377405e-01 3.89551401e-01 -1.95237324e-02 2.18499124e-01
7.48692572e-01 -1.21635318e-01 -4.71026242e-01 -3.49499166e-01
-6.98571980e-01 4.53460306e-01 1.68941960e-01 5.24862349e-01
4.48055685e-01 -1.17557716e+00 -4.81253088e-01 1.65375113e-01
2.93453008e-01 6.63192451e-01 2.38541141e-02 3.96509409e-01
-2.10758239e-01 5.73560894e-01 2.71093100e-01 -4.10101593e-01
-9.01126862e-01 7.54387140e-01 3.18266191e-02 -6.17048681e-01
-7.66132951e-01 9.98580754e-01 6.02850497e-01 -8.36122036e-01
1.51983187e-01 -6.35446072e-01 -4.07899559e-01 -1.05453491e-01
5.74849665e-01 -1.13574728e-01 -4.99516865e-03 -5.41212000e-02
-6.39412180e-02 1.70968190e-01 -2.01463804e-01 2.19894662e-01
1.33544540e+00 7.51678832e-03 -8.22959319e-02 2.68563420e-01
7.51643360e-01 -1.63179919e-01 -1.15622568e+00 -4.32489693e-01
4.46083486e-01 -3.86046208e-02 -4.38828230e-01 -9.00476277e-01
-7.94665873e-01 1.14806414e+00 -3.17390651e-01 4.33544904e-01
1.04858923e+00 4.19945478e-01 1.11266291e+00 8.11136067e-01
1.42234668e-01 -6.46974683e-01 1.44559532e-01 6.77937508e-01
8.15166891e-01 -1.44091618e+00 -2.12870941e-01 -8.42420816e-01
-1.00695980e+00 1.28018045e+00 1.10247731e+00 3.32185745e-01
9.82119292e-02 2.52669185e-01 5.41635305e-02 -2.79534191e-01
-9.80789363e-01 -5.83686270e-02 1.83891475e-01 5.33552229e-01
5.80880880e-01 9.36992653e-03 -3.37585732e-02 1.34778774e+00
-4.83093381e-01 -1.98287413e-01 4.55665737e-01 5.16342044e-01
-4.90782678e-01 -1.26390266e+00 7.24131381e-03 7.22821474e-01
-4.78510529e-01 -3.52733850e-01 -6.72159195e-01 8.21103275e-01
5.08994572e-02 1.03527319e+00 -1.28409520e-01 -1.28082484e-01
2.59466112e-01 4.32925314e-01 7.79745951e-02 -1.14573431e+00
-5.91241062e-01 -5.41818678e-01 4.12087411e-01 -3.81736785e-01
-3.39625269e-01 -3.96201551e-01 -1.94016290e+00 1.03613116e-01
-3.72577399e-01 4.60015446e-01 2.59965658e-01 1.48241854e+00
4.03537452e-01 5.24471343e-01 2.24525362e-01 -1.89357266e-01
-7.26290524e-01 -9.28598344e-01 -1.86598435e-01 5.26350617e-01
1.92884147e-01 -4.83370543e-01 -3.11426848e-01 3.54156822e-01] | [9.52722454071045, 6.909777641296387] |
2079096a-74ba-48c9-b35b-ba755bf1f557 | interpretable-multimodal-emotion-recognition-1 | 2306.02845 | null | https://arxiv.org/abs/2306.02845v1 | https://arxiv.org/pdf/2306.02845v1.pdf | Interpretable Multimodal Emotion Recognition using Facial Features and Physiological Signals | This paper aims to demonstrate the importance and feasibility of fusing multimodal information for emotion recognition. It introduces a multimodal framework for emotion understanding by fusing the information from visual facial features and rPPG signals extracted from the input videos. An interpretability technique based on permutation feature importance analysis has also been implemented to compute the contributions of rPPG and visual modalities toward classifying a given input video into a particular emotion class. The experiments on IEMOCAP dataset demonstrate that the emotion classification performance improves by combining the complementary information from multiple modalities. | ['Xiaobai Li', 'Puneet Kumar'] | 2023-06-05 | null | null | null | null | ['multimodal-emotion-recognition', 'emotion-classification', 'emotion-classification', 'multimodal-emotion-recognition'] | ['computer-vision', 'computer-vision', 'natural-language-processing', 'speech'] | [ 2.13859007e-01 -1.76833794e-01 -1.52271673e-01 -6.82970464e-01
-5.35454214e-01 -4.45242405e-01 5.06075561e-01 -9.45286900e-02
-3.61994654e-01 6.85223281e-01 6.17719531e-01 4.61119235e-01
-2.85327941e-01 -1.32119790e-01 -2.69203126e-01 -7.72233188e-01
-2.55458981e-01 -2.51483440e-01 -6.35628343e-01 1.13119371e-01
2.75932431e-01 5.46879530e-01 -2.25048685e+00 8.95532012e-01
5.28409958e-01 1.61708152e+00 -1.90594092e-01 6.97202981e-01
-9.77434367e-02 5.85272729e-01 -4.40110922e-01 -1.56253979e-01
3.29770707e-02 -5.55262506e-01 -6.45713925e-01 3.41359049e-01
2.80731082e-01 3.60573344e-02 1.34063400e-02 9.56410527e-01
4.13308740e-01 4.08303529e-01 7.69812524e-01 -1.90494418e+00
-3.19714963e-01 2.33506322e-01 -3.53027672e-01 2.29653344e-01
8.11080098e-01 -1.60376668e-01 9.14157808e-01 -9.36473072e-01
8.07118177e-01 1.19860172e+00 1.82775319e-01 3.18092346e-01
-8.54451478e-01 -4.03347641e-01 1.25804961e-01 6.69389784e-01
-1.23789883e+00 -3.02742541e-01 1.06865060e+00 -4.44601595e-01
9.74123538e-01 6.68263435e-01 6.69709086e-01 1.12288606e+00
1.13694966e-01 8.16386521e-01 1.44500732e+00 -4.08873975e-01
2.04552189e-01 4.09506947e-01 3.03140134e-01 6.40457571e-01
-2.49817535e-01 -5.22900000e-02 -1.11711836e+00 -1.38497904e-01
1.60829976e-01 -2.13695988e-01 -2.52810448e-01 -1.66899249e-01
-1.01295590e+00 7.43534684e-01 4.30766381e-02 5.13618469e-01
-6.17696583e-01 3.96593995e-02 7.64259696e-01 4.30658102e-01
4.17705595e-01 2.56294221e-01 -3.78291696e-01 -5.47569752e-01
-6.61402166e-01 -4.39117074e-01 6.14886165e-01 5.48826337e-01
6.71566963e-01 6.48168102e-03 -2.49377146e-01 6.16876006e-01
7.12902308e-01 6.24874830e-01 3.71770024e-01 -1.22912574e+00
1.01187147e-01 8.88056040e-01 -9.36761349e-02 -1.38398027e+00
-4.94359165e-01 3.04737747e-01 -3.66808414e-01 2.73593754e-01
-7.05769379e-03 -3.44176024e-01 -7.81010389e-01 1.71041870e+00
2.71227896e-01 1.30821448e-02 2.18187124e-01 1.05747962e+00
1.25914407e+00 6.15028799e-01 3.19786459e-01 -3.64173830e-01
1.52005696e+00 -5.81609666e-01 -1.01061821e+00 3.73485148e-01
1.66514724e-01 -4.71398503e-01 6.13632202e-01 7.74812281e-01
-6.49803638e-01 -7.12813020e-01 -9.53776658e-01 1.61965922e-01
-6.17065549e-01 5.50094187e-01 7.86503911e-01 9.13313091e-01
-8.79007220e-01 8.69247168e-02 -5.11736214e-01 -4.64006394e-01
4.66867834e-01 5.17106175e-01 -1.09721982e+00 1.40782654e-01
-1.09927773e+00 7.36341655e-01 5.50783873e-01 3.62995774e-01
-6.22585416e-01 -2.93013722e-01 -9.54330444e-01 2.14029662e-02
-1.89791754e-01 -3.31275523e-01 5.19745111e-01 -1.80753422e+00
-1.75091088e+00 6.81784272e-01 -3.82745773e-01 9.61748138e-02
-4.51708660e-02 -2.69091744e-02 -6.68627679e-01 9.87367153e-01
-3.93689871e-01 8.36475849e-01 1.05980003e+00 -1.16490602e+00
-5.90005040e-01 -6.74006462e-01 -2.26882830e-01 4.52772975e-01
-4.51916754e-01 3.33524406e-01 -4.12066668e-01 -3.38780314e-01
8.05365816e-02 -7.24825025e-01 2.04338163e-01 -2.70429254e-01
-4.19620387e-02 -7.40200728e-02 1.03810334e+00 -6.43283367e-01
8.38497162e-01 -2.29293823e+00 6.92167878e-01 6.44719303e-01
5.22773862e-02 -6.66903406e-02 -3.89171243e-01 4.26472425e-01
-2.91585654e-01 9.05144289e-02 2.75825977e-01 -1.35572791e-01
2.38928244e-01 3.04072231e-01 -9.82203037e-02 3.02937210e-01
3.52639407e-01 8.58696282e-01 -5.49982131e-01 -4.84425575e-01
4.49731648e-01 7.98865199e-01 -3.95110667e-01 2.58683473e-01
4.04346615e-01 5.80806851e-01 -3.48323077e-01 9.53279376e-01
6.22993171e-01 3.01473349e-01 3.90900940e-01 -7.30013490e-01
1.67291582e-01 -5.39772987e-01 -1.11961067e+00 1.61012220e+00
-1.72828957e-01 9.39909041e-01 9.52259973e-02 -1.01532376e+00
9.08660769e-01 5.73684812e-01 5.99214852e-01 -5.66154480e-01
5.91329455e-01 -2.34211937e-01 -1.78196490e-01 -1.07052708e+00
4.63645577e-01 -1.88952968e-01 -3.91208977e-01 2.47577429e-01
8.54306757e-01 3.18682402e-01 1.89304978e-01 -6.13766257e-03
6.05550349e-01 1.41596511e-01 4.05222774e-01 -3.71121652e-02
7.72138774e-01 -2.28467435e-01 4.08765167e-01 4.16723579e-01
-4.45510000e-01 2.17124507e-01 6.86982989e-01 -1.91993132e-01
-3.18967342e-01 -1.04347253e+00 -4.56069857e-02 1.34088409e+00
-5.37702627e-02 -3.92551899e-01 -4.19880778e-01 -8.44689071e-01
-1.11411966e-01 2.83650726e-01 -1.08876240e+00 -2.01003000e-01
3.02183151e-01 -7.43417978e-01 4.65032101e-01 5.87572336e-01
2.32877746e-01 -1.03789175e+00 -7.37977922e-01 -3.67814839e-01
-4.12701726e-01 -1.05382872e+00 2.32210800e-01 1.91853583e-01
-7.01239288e-01 -1.09276736e+00 -2.82195002e-01 -4.74290341e-01
8.32961261e-01 -4.34516706e-02 5.80822229e-01 -2.66469210e-01
-5.47887683e-01 1.30303097e+00 -7.55794466e-01 -3.22091371e-01
5.50034642e-02 -6.11481845e-01 1.05350353e-01 6.60006642e-01
6.65560007e-01 -2.75206655e-01 -2.73813754e-01 1.25020638e-01
-9.39066052e-01 -2.43351161e-01 4.01526481e-01 7.10432947e-01
4.37174827e-01 1.45723104e-01 5.14782250e-01 -2.43028179e-01
7.59751499e-01 -6.90080106e-01 -1.17138529e-03 3.23204845e-01
5.15440293e-02 -1.20053543e-02 1.04588076e-01 -4.12239134e-01
-1.37420058e+00 2.48966217e-01 9.09137651e-02 -5.51936865e-01
-6.48119390e-01 7.27314055e-01 -3.72159123e-01 -3.09460789e-01
1.71960413e-01 7.56432191e-02 1.34256601e-01 2.27542757e-03
5.93824089e-01 7.59606004e-01 2.46509269e-01 -5.34463823e-01
4.31859680e-02 5.19519746e-01 2.40931697e-02 -1.04274261e+00
-1.34760439e-01 -5.04710495e-01 -6.37457490e-01 -9.43880320e-01
1.05698800e+00 -7.99261749e-01 -1.23261094e+00 3.70437473e-01
-7.90514469e-01 5.29458344e-01 -1.30158782e-01 9.25439358e-01
-4.83921468e-01 5.56790829e-01 -4.19816494e-01 -1.01817441e+00
-1.35790393e-01 -1.14906251e+00 1.06124902e+00 4.11388963e-01
-3.05008948e-01 -8.70289803e-01 8.96845385e-02 5.46208739e-01
4.86711115e-02 2.50384897e-01 7.28185117e-01 -6.78509891e-01
9.32063721e-03 -3.35273534e-01 -2.84689814e-01 5.86678147e-01
1.05336696e-01 3.01127553e-01 -1.27491212e+00 1.86964542e-01
-7.46276677e-02 -6.55536890e-01 8.56461585e-01 2.92149961e-01
8.10487270e-01 -3.28294076e-02 -2.30654292e-02 2.00386256e-01
1.42185473e+00 3.23476940e-01 6.91692412e-01 -1.91611543e-01
5.65005660e-01 1.01274264e+00 7.65845418e-01 6.34997547e-01
8.67304653e-02 3.08589876e-01 6.25430703e-01 1.83371231e-01
3.76793832e-01 1.50824562e-01 5.80124974e-01 7.61392176e-01
-3.38561952e-01 -2.52915859e-01 -4.97539282e-01 1.52281091e-01
-1.81367743e+00 -1.17150629e+00 4.83056642e-02 1.67754900e+00
1.22172393e-01 -6.35839760e-01 2.61368938e-02 1.66274667e-01
5.60945094e-01 8.47493857e-03 -1.36469245e-01 -9.34262812e-01
-3.77037257e-01 4.14618813e-02 -1.03355318e-01 2.22684070e-01
-1.15850461e+00 5.34940481e-01 7.26312733e+00 5.38547516e-01
-1.14559329e+00 -1.44981667e-01 5.59570968e-01 -1.82910204e-01
-1.88833162e-01 -3.17229658e-01 -3.89365792e-01 3.31819832e-01
1.03409159e+00 8.62971917e-02 3.21397811e-01 7.43968844e-01
2.24768668e-01 -6.21036172e-01 -8.83632660e-01 1.38584876e+00
7.98481584e-01 -8.72441709e-01 3.69722992e-01 -7.43212104e-02
3.81112486e-01 -1.57787651e-01 1.79272532e-01 2.27001518e-01
-3.98731023e-01 -1.02186155e+00 3.97078246e-01 1.10028899e+00
3.66765082e-01 -9.37870264e-01 9.65972066e-01 -1.39778882e-01
-9.94576335e-01 -2.60121405e-01 -7.65753314e-02 -1.73262000e-01
-1.00283548e-01 6.52814955e-02 -7.47523785e-01 8.38795364e-01
7.89999902e-01 7.63380706e-01 -7.13965774e-01 7.79359639e-01
-1.58017531e-01 1.81848943e-01 -1.86562717e-01 -9.24780965e-02
-1.50693789e-01 -3.34209234e-01 4.69469994e-01 1.26973093e+00
4.11324710e-01 1.25331685e-01 -1.61657080e-01 4.58923042e-01
2.04652414e-01 5.11747658e-01 -7.31666148e-01 -6.32626832e-01
-3.71154211e-02 1.80751503e+00 -7.79350638e-01 -3.49110514e-01
-4.71968412e-01 1.11409974e+00 4.89959726e-03 4.32372421e-01
-8.23947549e-01 -4.14396524e-01 7.32598364e-01 -9.65431988e-01
5.32044649e-01 -4.94820625e-02 -1.35472948e-02 -1.15297866e+00
-1.81438521e-01 -7.21190810e-01 6.42817974e-01 -1.16290200e+00
-1.17996335e+00 7.50027537e-01 -8.29985365e-03 -1.38216424e+00
-2.11440504e-01 -1.05631602e+00 -4.21186209e-01 6.45160437e-01
-1.11434162e+00 -1.17046309e+00 -4.16984260e-01 7.61890709e-01
7.81354979e-02 -2.15136364e-01 1.04395127e+00 9.71027017e-02
-4.25411075e-01 4.31422979e-01 -3.04064482e-01 -4.72290590e-02
5.65795898e-01 -9.29802775e-01 -1.12216520e+00 6.60502136e-01
1.86049551e-01 3.72575790e-01 5.85883796e-01 -3.52788299e-01
-1.79394627e+00 -4.59481448e-01 5.76260448e-01 -3.88543397e-01
3.48047584e-01 -3.73688736e-03 -4.34560299e-01 4.13553745e-01
5.68101406e-01 -9.13839713e-02 1.52799857e+00 1.98250338e-01
-3.86771590e-01 6.31192476e-02 -1.33435690e+00 3.93765479e-01
4.54345137e-01 -8.06787610e-01 -9.04038608e-01 -4.65158492e-01
1.59420341e-01 2.15238661e-01 -1.29007363e+00 4.84391630e-01
9.10117388e-01 -9.61484492e-01 7.04683602e-01 -8.90389860e-01
4.78026122e-01 -4.33158189e-01 -7.28830516e-01 -1.30578601e+00
2.52556410e-02 -3.20495293e-02 5.02481237e-02 1.26447129e+00
2.33131692e-01 -3.14142197e-01 5.01076043e-01 6.34364128e-01
2.74005175e-01 -2.47174963e-01 -1.09425235e+00 -1.95549369e-01
-6.57223582e-01 -7.88018942e-01 1.83721259e-01 9.75499928e-01
8.09526742e-01 2.48348489e-01 -4.97401118e-01 1.42846465e-01
1.59498230e-01 2.17946902e-01 5.75690091e-01 -1.09308422e+00
-4.98051718e-02 -3.85988265e-01 -1.06941819e+00 -1.10056633e-02
4.11116630e-01 -9.47355807e-01 -1.75114542e-01 -1.23407960e+00
3.27145725e-01 5.13612688e-01 -8.58846247e-01 5.98326266e-01
-5.26276492e-02 7.96281755e-01 4.30634707e-01 -4.83202815e-01
-1.00926900e+00 7.30371177e-01 9.52033401e-01 3.08163203e-02
-9.39242244e-02 -6.00290477e-01 -6.88694477e-01 7.28779793e-01
7.23806977e-01 -1.04035698e-01 -4.93976027e-01 6.05417341e-02
3.24557692e-01 1.93067878e-01 4.11080569e-01 -9.13658977e-01
-3.25091206e-03 -5.33170626e-02 8.83923590e-01 -5.08554339e-01
8.55569303e-01 -1.14312947e+00 1.97227016e-01 1.38600037e-01
-3.35289210e-01 -6.78390190e-02 6.50142729e-01 6.85443640e-01
-7.11938560e-01 2.41896529e-02 3.24070364e-01 1.94469795e-01
-1.35577345e+00 -2.37706810e-01 -8.64877164e-01 -3.79273891e-01
1.15159810e+00 -4.69448477e-01 -1.32742509e-01 -4.52608943e-01
-1.34759748e+00 -6.43472299e-02 -6.70426860e-02 7.44086862e-01
9.56406295e-01 -1.55439115e+00 -2.73255825e-01 5.06951630e-01
4.56610292e-01 -1.47191501e+00 7.78830767e-01 1.22035480e+00
2.02410575e-02 4.12059426e-01 -1.14529192e+00 -4.88556832e-01
-1.90701520e+00 2.19125926e-01 -2.61338949e-02 2.52544105e-01
1.99186787e-01 7.58844435e-01 -9.35758352e-02 -3.50945950e-01
1.59867525e-01 -1.55252531e-01 -7.87392497e-01 9.34021711e-01
7.11452723e-01 4.55143124e-01 -1.34974688e-01 -1.17401540e+00
-6.10699654e-01 5.41613877e-01 2.99432516e-01 -4.47922289e-01
1.14537954e+00 -3.31865340e-01 -3.53426546e-01 7.36750841e-01
1.46611929e+00 3.74835171e-02 -7.86644995e-01 2.16330886e-01
-2.18482465e-01 -4.73591417e-01 2.99503151e-02 -1.06586397e+00
-1.06796551e+00 8.72387230e-01 9.72695231e-01 -4.73908223e-02
1.63998210e+00 -1.28820300e-01 -5.53616658e-02 2.25470260e-01
2.60157436e-01 -1.40047300e+00 1.29846588e-01 2.81208694e-01
1.02464497e+00 -1.35381246e+00 -1.32463664e-01 -2.96192020e-01
-1.24743164e+00 1.52037656e+00 2.74020344e-01 2.45265383e-02
7.80131698e-01 -1.16375655e-01 1.91310897e-01 -2.83963829e-01
-7.67404974e-01 -3.15950453e-01 8.39726210e-01 5.72971344e-01
2.61644781e-01 7.27719814e-02 -5.19278109e-01 9.21522319e-01
3.88938844e-01 1.80006444e-01 1.96158782e-01 8.40476573e-01
-2.97865152e-01 -9.79384661e-01 -4.14037406e-01 4.23946023e-01
-5.26749671e-01 1.99567795e-01 -7.99021661e-01 7.19147205e-01
2.01236695e-01 1.15388405e+00 3.39641199e-02 -8.25494289e-01
-1.89754561e-01 5.13471842e-01 5.81175625e-01 -1.69107109e-01
-6.49767339e-01 7.70471664e-03 4.37004805e-01 -9.62157607e-01
-1.23283052e+00 -6.01312459e-01 -1.17673063e+00 2.17240825e-01
1.00347564e-01 2.59516805e-01 8.70276690e-01 9.16631460e-01
7.70810008e-01 4.51123327e-01 5.98337829e-01 -1.13984883e+00
3.01937222e-01 -7.14705706e-01 -7.04278290e-01 3.90501648e-01
1.83130279e-02 -8.73491347e-01 -6.36499643e-01 1.15767568e-01] | [13.243236541748047, 5.166940212249756] |
e38dc572-8ee7-406c-b0bb-d090db038f4e | powergan-synthesizing-appliance-power | 2007.13645 | null | https://arxiv.org/abs/2007.13645v1 | https://arxiv.org/pdf/2007.13645v1.pdf | PowerGAN: Synthesizing Appliance Power Signatures Using Generative Adversarial Networks | Non-intrusive load monitoring (NILM) allows users and energy providers to gain insight into home appliance electricity consumption using only the building's smart meter. Most current techniques for NILM are trained using significant amounts of labeled appliances power data. The collection of such data is challenging, making data a major bottleneck in creating well generalizing NILM solutions. To help mitigate the data limitations, we present the first truly synthetic appliance power signature generator. Our solution, PowerGAN, is based on conditional, progressively growing, 1-D Wasserstein generative adversarial network (GAN). Using PowerGAN, we are able to synthesise truly random and realistic appliance power data signatures. We evaluate the samples generated by PowerGAN in a qualitative way as well as numerically by using traditional GAN evaluation methods such as the Inception score. | ['Alon Harell', 'Stephen Makonin', 'Ivan V. Bajic', 'Richard Jones'] | 2020-07-20 | null | null | null | null | ['non-intrusive-load-monitoring', 'non-intrusive-load-monitoring', 'non-intrusive-load-monitoring'] | ['knowledge-base', 'miscellaneous', 'time-series'] | [ 2.82926798e-01 5.42702258e-01 -3.87015827e-02 -2.91993231e-01
-1.01539612e+00 -8.58815014e-01 6.08489990e-01 -3.77550155e-01
4.05297130e-01 1.01800752e+00 3.33535373e-01 -1.84597954e-01
2.29335874e-01 -1.09931195e+00 -4.18398470e-01 -8.68860602e-01
1.18369475e-01 6.98585153e-01 -6.38072312e-01 -9.54082422e-03
-4.32513833e-01 3.73846382e-01 -9.39920783e-01 -2.46132970e-01
9.01920617e-01 7.37870812e-01 -4.34312701e-01 8.62058759e-01
5.16071975e-01 1.11602879e+00 -1.14336884e+00 -2.78285652e-01
4.32274759e-01 -1.01903188e+00 -4.87833858e-01 -3.49376760e-02
-2.62046922e-02 -7.31117010e-01 -3.36000994e-02 1.01629317e+00
7.34507084e-01 1.29240621e-02 5.62537968e-01 -2.14844871e+00
-5.97046196e-01 1.09797108e+00 -2.66118437e-01 -3.48695993e-01
6.45758510e-01 7.65300333e-01 8.71046364e-01 2.82916754e-01
1.95563883e-01 8.18176806e-01 8.90367508e-01 7.30886459e-01
-1.85578215e+00 -7.27642655e-01 -4.36204255e-01 -9.44711342e-02
-1.22620332e+00 -4.11549836e-01 1.29228926e+00 -4.65284847e-02
1.01647496e+00 7.82985985e-01 7.95806348e-01 1.66010523e+00
-1.94922730e-01 8.79043281e-01 1.27290344e+00 -4.25669014e-01
6.70703053e-01 1.23051934e-01 -2.75467783e-01 2.05908924e-01
2.69616753e-01 1.18458550e-02 1.33225590e-01 -5.23655176e-01
3.31508815e-01 -1.22316688e-01 -1.85265377e-01 -9.77820009e-02
-7.29771197e-01 6.86580539e-01 2.85496593e-01 4.67607796e-01
-4.58281219e-01 5.37708879e-01 2.42810994e-01 1.06292456e-01
3.87512535e-01 5.86445332e-01 -3.18054467e-01 -5.23573220e-01
-1.17711222e+00 8.23844224e-02 9.48879480e-01 9.49978650e-01
4.73969817e-01 5.17875016e-01 -1.89505994e-01 2.95316964e-01
2.73537487e-01 8.35637510e-01 5.59202850e-01 -1.11911368e+00
1.88027412e-01 4.59620118e-01 1.62810609e-01 -4.00111169e-01
-2.74430543e-01 3.20032179e-01 -1.08775985e+00 3.45310897e-01
1.22767776e-01 -5.29644072e-01 -7.20367789e-01 1.82124817e+00
2.80517358e-02 8.53468552e-02 6.92161992e-02 1.55434057e-01
2.81302631e-01 6.20611668e-01 2.31021971e-01 -2.62928456e-01
9.08246934e-01 -5.74999392e-01 -7.33914375e-01 2.45698795e-01
4.44444001e-01 -2.40753174e-01 1.16744375e+00 2.77157754e-01
-1.12548125e+00 -1.45915419e-01 -1.12224948e+00 2.21502557e-01
-3.67584825e-01 -5.53782582e-01 4.43488806e-01 1.14395273e+00
-1.03807926e+00 7.18568146e-01 -1.06253374e+00 -3.00221711e-01
8.48441780e-01 1.54097185e-01 -8.53165686e-02 1.09550498e-01
-9.33850110e-01 7.74797142e-01 3.08969557e-01 -3.26713979e-01
-1.03424394e+00 -9.35986340e-01 -8.43252003e-01 2.67701875e-02
-1.39977515e-01 -6.10353470e-01 1.49097514e+00 -8.22011232e-01
-1.71187675e+00 3.60031724e-01 3.56268764e-01 -5.45093656e-01
9.63532031e-01 1.91800311e-01 -5.96100211e-01 -1.20657593e-01
5.17690778e-02 3.82085294e-01 8.05223882e-01 -1.48761153e+00
-2.05642208e-01 -1.93206713e-01 -3.49043310e-01 -3.51804048e-01
-8.98912996e-02 -5.22539496e-01 5.95072508e-01 -5.98236084e-01
-6.73120499e-01 -7.70035505e-01 -1.84384003e-01 -5.12853622e-01
-1.10816622e+00 -8.15010965e-02 1.27459145e+00 -9.37897623e-01
7.17147052e-01 -2.00827289e+00 -6.19100213e-01 4.63008016e-01
1.29770478e-02 1.11232646e-01 -3.85250300e-02 4.85350400e-01
-2.06601843e-01 4.52784598e-01 -4.05947834e-01 -5.83196521e-01
7.78852761e-01 3.60610068e-01 -3.44853580e-01 3.54282647e-01
-1.09551966e-01 1.54048133e+00 -1.07031286e+00 -1.27260342e-01
7.73167670e-01 6.22691333e-01 -2.04391599e-01 5.55992007e-01
-4.55017269e-01 5.38538277e-01 -2.90527433e-01 6.36645615e-01
6.36352897e-01 -6.56219795e-02 4.05823648e-01 -1.97288036e-01
4.63211775e-01 2.32296303e-01 -7.25948751e-01 1.38643813e+00
-7.82313466e-01 5.13535321e-01 -1.90261170e-01 -7.54991055e-01
6.05463445e-01 4.21614081e-01 8.73646259e-01 -7.75409102e-01
2.90111691e-01 -8.13165456e-02 -3.98648560e-01 -1.37089804e-01
1.08014800e-01 -2.75312215e-01 -3.99917424e-01 1.02440369e+00
-4.34055924e-02 -6.06853962e-01 -5.51922247e-02 -3.52305807e-02
1.68455994e+00 -5.76994494e-02 3.84913087e-01 -1.82567716e-01
-1.40968576e-01 -2.26017088e-01 2.45396867e-01 5.11043251e-01
-7.13243410e-02 4.68019545e-01 4.37828958e-01 -1.70780659e-01
-1.39240885e+00 -1.39026845e+00 3.33485678e-02 4.10831392e-01
-5.27973235e-01 -2.36856297e-01 -1.28839779e+00 -9.69758093e-01
1.20105438e-01 1.69715643e+00 -5.87654471e-01 -3.06123972e-01
-4.90901142e-01 -1.12674987e+00 8.30335557e-01 7.73191929e-01
7.38496065e-01 -1.32711577e+00 -7.61936307e-01 3.90048116e-01
-6.31172657e-02 -9.50306177e-01 -4.57102150e-01 4.25945371e-01
-4.98941571e-01 -1.21812952e+00 -3.34342808e-01 -1.13766134e-01
7.61274338e-01 -4.96633381e-01 1.58765614e+00 -3.01211298e-01
-5.60231745e-01 6.59985781e-01 -3.19514304e-01 -5.76297879e-01
-9.83265698e-01 1.81418851e-01 -1.92540661e-02 -2.02920109e-01
6.37067258e-01 -1.39515793e+00 -5.86385012e-01 9.70951095e-02
-7.33503282e-01 -1.52898356e-01 2.72187024e-01 5.38153589e-01
2.72208661e-01 6.78342760e-01 5.86888731e-01 -9.69959199e-01
8.42368066e-01 -4.49189842e-01 -7.75956392e-01 1.50669813e-02
-1.12122750e+00 -6.17263466e-02 1.07405841e+00 -4.31435674e-01
-8.49638462e-01 -1.46806017e-01 -2.74116755e-01 -5.51615953e-02
-4.86626923e-01 -5.67007184e-01 -7.04784691e-01 2.86243200e-01
5.11503875e-01 2.41463900e-01 -3.38471264e-01 -3.74268293e-01
7.26927519e-01 4.49877709e-01 7.12147176e-01 -4.22954381e-01
1.30236173e+00 3.40458155e-01 -4.14689891e-02 -5.59822083e-01
-3.31081659e-01 2.56061286e-01 -3.64859253e-01 7.60867149e-02
9.67645466e-01 -6.05797529e-01 -1.09555542e+00 7.06922114e-01
-6.78969026e-01 -1.16120195e+00 -1.41157424e+00 -1.94700122e-01
-7.89693236e-01 -6.51419908e-02 -3.81488562e-01 -9.29482162e-01
-6.66464984e-01 -7.03090191e-01 1.04219759e+00 2.32380722e-02
-9.51625228e-01 -1.28012264e+00 5.36608100e-01 2.07853749e-01
6.41381979e-01 1.08145380e+00 1.05318582e+00 -6.83734417e-01
-5.08352280e-01 -4.91090655e-01 2.59331971e-01 8.14890325e-01
6.39860392e-01 -1.15094900e-01 -1.33653414e+00 -4.01765794e-01
1.85073912e-01 -4.13633734e-01 -3.94002385e-02 2.43203312e-01
1.27729595e+00 -9.05909538e-01 -1.37163997e-01 6.49644196e-01
1.56283176e+00 3.02476048e-01 1.02503216e+00 -3.45143937e-02
7.47028053e-01 -1.20900288e-01 -2.97986627e-01 6.07727408e-01
4.08170372e-01 3.20672572e-01 3.31130713e-01 -3.26450795e-01
-3.63848143e-04 -9.38920438e-01 3.55020046e-01 5.99778354e-01
4.74403724e-02 -5.88966846e-01 -4.90724891e-01 5.10334253e-01
-1.46534419e+00 -1.19109893e+00 3.21818024e-01 1.88452005e+00
8.17000091e-01 5.63980229e-02 2.98683167e-01 6.35477662e-01
1.81535691e-01 3.17471594e-01 -9.24813390e-01 -2.26714775e-01
5.94219416e-02 6.54310584e-01 6.79265559e-01 4.40784127e-01
-6.28865480e-01 7.00330362e-02 7.22325230e+00 5.42057753e-01
-6.46513522e-01 3.71474117e-01 7.59658456e-01 -1.92193329e-01
-6.49805546e-01 -3.70001614e-01 -1.83168367e-01 9.67240751e-01
1.18798554e+00 -3.52534652e-01 9.65270340e-01 1.10316718e+00
3.28924865e-01 -2.13051096e-01 -1.59852779e+00 1.11366975e+00
-6.83586001e-02 -8.24987948e-01 -4.40142006e-01 2.87526429e-01
1.10088253e+00 -8.64174813e-02 -1.11086585e-01 2.82609135e-01
1.31194437e+00 -1.28147626e+00 3.46977413e-01 5.43204248e-01
9.45750415e-01 -7.80006707e-01 6.26125455e-01 2.85320699e-01
-1.10314441e+00 5.59487753e-02 3.70623261e-01 2.32133612e-01
4.15410489e-01 8.01618934e-01 -1.04854238e+00 3.57994616e-01
5.46166658e-01 3.37637544e-01 -3.65021348e-01 1.60696462e-01
-5.26977539e-01 9.08473849e-01 -6.89823568e-01 1.67749882e-01
-2.52197325e-01 -2.58924037e-01 1.37336001e-01 7.91106880e-01
4.06222790e-01 -3.88510600e-02 -2.71636844e-02 1.38985121e+00
-5.27791739e-01 -5.28049171e-01 -8.49391818e-01 -1.19304590e-01
4.35381889e-01 1.32009828e+00 -4.38778818e-01 -3.83756876e-01
1.40336370e-02 1.32293367e+00 -2.82902211e-01 2.88428813e-01
-8.76631320e-01 -2.70678490e-01 8.40437055e-01 1.47974595e-01
-1.23645894e-01 1.62690833e-01 -4.53384221e-01 -1.02089930e+00
1.64247647e-01 -9.29159462e-01 8.33196193e-03 -9.85543668e-01
-1.63512433e+00 1.47593468e-01 1.15014389e-02 -7.17517197e-01
-9.15682971e-01 -5.56931086e-02 -1.22416818e+00 8.55075538e-01
-1.09885061e+00 -1.35385811e+00 -6.46058440e-01 6.27610326e-01
2.35974208e-01 -1.50795206e-01 1.32749593e+00 1.27177536e-01
-4.72361565e-01 8.43774974e-01 3.27214152e-01 2.78412014e-01
-2.10571177e-02 -1.76681268e+00 9.33750093e-01 6.41664684e-01
8.59971866e-02 -1.23773515e-01 8.30542505e-01 -4.54733640e-01
-1.26860833e+00 -1.32720971e+00 2.73544639e-01 -9.45105493e-01
4.73364264e-01 -6.97461724e-01 -4.52449530e-01 1.05752015e+00
5.56994557e-01 -1.48087084e-01 7.30481923e-01 -3.66435856e-01
-1.75918434e-02 -2.99034685e-01 -2.06596255e+00 4.04861629e-01
7.83131063e-01 -6.45200789e-01 -4.08897132e-01 4.10265297e-01
3.32429647e-01 2.36737281e-01 -1.20933032e+00 1.79922462e-01
2.80293018e-01 -9.08482969e-01 6.22754216e-01 -2.11013675e-01
2.77765859e-02 -2.03256637e-01 -2.14474306e-01 -1.68645203e+00
-2.62418501e-02 -1.13307464e+00 -5.37741840e-01 1.81158698e+00
1.32910147e-01 -9.12487268e-01 9.11707878e-01 1.20547497e+00
4.90706980e-01 -2.04250559e-01 -7.37958252e-01 -7.71022677e-01
-1.01176247e-01 -5.92777133e-01 1.55246711e+00 9.89032149e-01
-1.48567539e-02 1.08735256e-01 -3.06708664e-01 -6.88308850e-02
1.04301918e+00 -2.01522425e-01 9.39381123e-01 -9.18156326e-01
-4.05914128e-01 -3.42733443e-01 -3.69564295e-01 -3.78321469e-01
2.27056205e-01 -7.25587308e-01 1.15757905e-01 -1.47814095e+00
1.88754871e-01 -4.65353251e-01 -1.83582246e-01 9.16814029e-01
2.98643231e-01 6.54410303e-01 1.57576993e-01 -1.86928958e-01
-8.76941830e-02 5.09706020e-01 6.43575609e-01 -4.40478057e-01
-1.49322167e-01 1.85242742e-02 -6.43039644e-01 6.88227713e-01
1.19760025e+00 -3.56521726e-01 -7.81211555e-01 1.38461903e-01
-5.97184971e-02 -2.45738223e-01 5.64382911e-01 -1.05009890e+00
-3.85844052e-01 3.32697271e-03 7.33377337e-01 -3.55904669e-01
1.45840988e-01 -1.13669109e+00 8.69805038e-01 3.99596453e-01
-7.83223733e-02 8.50731954e-02 1.92197245e-02 4.32423130e-02
3.11498761e-01 2.05618665e-02 8.05935323e-01 -2.78820276e-01
-5.66522814e-02 1.00953363e-01 -1.54972255e-01 1.92937642e-01
1.18969417e+00 1.11716032e-01 -3.52770418e-01 -6.73250914e-01
-5.65491557e-01 2.03639179e-01 9.68905985e-01 1.62510008e-01
-2.96953768e-02 -1.77305257e+00 -4.59732920e-01 4.02923763e-01
-2.94943094e-01 1.46619126e-01 -3.88185121e-02 2.16907904e-01
-3.25933486e-01 9.59661696e-03 -9.03204754e-02 -2.13225961e-01
-7.03985393e-01 6.41424716e-01 4.90703136e-01 -4.62425411e-01
-8.72067273e-01 6.58430085e-02 -3.25403243e-01 -2.52306819e-01
-1.03844024e-01 -4.58563775e-01 5.03999233e-01 -1.73074469e-01
5.17358661e-01 8.01187336e-01 7.21485391e-02 -2.41252646e-01
-1.15878142e-01 2.16529489e-01 5.51940084e-01 -1.67262241e-01
1.51426947e+00 1.79810226e-01 -1.02849849e-01 4.74788934e-01
1.28848577e+00 -9.07016322e-02 -1.24225891e+00 5.72089672e-01
-2.51863807e-01 -1.81811765e-01 -1.83165655e-01 -1.13375425e+00
-1.17550159e+00 3.58180165e-01 8.55856657e-01 9.67799664e-01
1.36651897e+00 -9.72239524e-02 1.27519369e+00 4.38253060e-02
6.11569822e-01 -1.07660592e+00 -2.20876336e-01 -4.82247561e-01
5.22504807e-01 -1.05550158e+00 -1.33891270e-01 2.84745425e-01
-2.08884433e-01 5.99273026e-01 2.78896421e-01 -1.56302705e-01
5.42267323e-01 9.29469466e-01 1.33950621e-01 5.10519519e-02
-3.70799810e-01 3.21832240e-01 -3.15121740e-01 1.15897429e+00
-9.30620953e-02 5.72198510e-01 4.15112942e-01 4.54904169e-01
-7.56881714e-01 1.50558680e-01 4.94760305e-01 8.60777199e-01
3.59938353e-01 -1.27130377e+00 -3.83990556e-01 6.90330863e-01
-3.39518636e-01 1.38763845e-01 -1.97023392e-01 7.24352002e-01
1.46993008e-02 9.41194952e-01 8.32056180e-02 -4.08499777e-01
4.92262721e-01 3.49384457e-01 5.08657515e-01 -1.09231636e-01
-3.12219441e-01 -4.85355884e-01 -6.12980127e-02 -5.71601152e-01
-2.04097718e-01 -6.32185757e-01 -8.51989746e-01 -8.52265656e-01
9.89866108e-02 3.82818608e-03 5.50831199e-01 7.27150679e-01
5.68507574e-02 7.50327587e-01 1.23286343e+00 -1.07931340e+00
-7.17114270e-01 -1.16978204e+00 -7.70186365e-01 9.42585945e-01
3.36908996e-01 6.90003633e-02 -5.61076820e-01 1.50762901e-01] | [16.051918029785156, 7.569221019744873] |
30c97348-7193-47a8-a1fe-6d05bee69b26 | cloud-based-deep-learning-of-big-eeg-data-for | 1702.05192 | null | http://arxiv.org/abs/1702.05192v1 | http://arxiv.org/pdf/1702.05192v1.pdf | Cloud-based Deep Learning of Big EEG Data for Epileptic Seizure Prediction | Developing a Brain-Computer Interface~(BCI) for seizure prediction can help
epileptic patients have a better quality of life. However, there are many
difficulties and challenges in developing such a system as a real-life support
for patients. Because of the nonstationary nature of EEG signals, normal and
seizure patterns vary across different patients. Thus, finding a group of
manually extracted features for the prediction task is not practical. Moreover,
when using implanted electrodes for brain recording massive amounts of data are
produced. This big data calls for the need for safe storage and high
computational resources for real-time processing. To address these challenges,
a cloud-based BCI system for the analysis of this big EEG data is presented.
First, a dimensionality-reduction technique is developed to increase
classification accuracy as well as to decrease the communication bandwidth and
computation time. Second, following a deep-learning approach, a stacked
autoencoder is trained in two steps for unsupervised feature extraction and
classification. Third, a cloud-computing solution is proposed for real-time
analysis of big EEG data. The results on a benchmark clinical dataset
illustrate the superiority of the proposed patient-specific BCI as an
alternative method and its expected usefulness in real-life support of epilepsy
patients. | ['Hamid Soltanian-Zadeh', 'Mohammad-Parsa Hosseini', 'Dario Pompili', 'Kost Elisevich'] | 2017-02-17 | null | null | null | null | ['seizure-prediction'] | ['medical'] | [-8.30679014e-02 -5.14166892e-01 6.21873498e-01 -4.31948960e-01
-1.98447794e-01 5.53690866e-02 5.79716591e-03 -6.22493960e-02
-2.79374570e-01 7.69252658e-01 -1.10827848e-01 -8.47708508e-02
-5.28388560e-01 -7.16386735e-01 -2.17817143e-01 -9.50801253e-01
-3.58261794e-01 4.13942397e-01 -1.48176208e-01 -1.16195805e-01
1.48998871e-01 7.27576017e-01 -1.75155818e+00 4.44907755e-01
1.07856464e+00 1.28159940e+00 5.74662209e-01 1.24889404e-01
-6.51110858e-02 4.75848913e-01 -9.28843558e-01 1.56233117e-01
5.08589484e-02 -2.45331928e-01 -3.11578363e-01 1.32828027e-01
-5.04387140e-01 -1.17005721e-01 -2.88707882e-01 1.02365649e+00
8.61992359e-01 -2.14316726e-01 4.80849564e-01 -1.25498450e+00
-1.66515969e-02 1.89495847e-01 -1.67000428e-01 4.85008836e-01
1.96207032e-01 -1.58439294e-01 2.37196892e-01 -8.93817604e-01
1.14755526e-01 4.57239628e-01 2.71544278e-01 4.79272336e-01
-7.43512094e-01 -9.54507113e-01 -3.24155033e-01 8.63888979e-01
-1.57824075e+00 -3.48869145e-01 7.90588379e-01 -5.55888295e-01
1.06702924e+00 3.94102663e-01 1.25828958e+00 9.67574298e-01
6.96258247e-01 4.73345578e-01 1.03161621e+00 -2.72320896e-01
6.40782058e-01 3.00480455e-01 3.48746777e-01 -8.56802091e-02
3.29046577e-01 6.34765029e-02 -6.89422965e-01 -5.70070967e-02
3.42125297e-01 4.99242663e-01 -6.42802715e-01 1.43045047e-02
-1.02927315e+00 5.16471267e-01 1.94625273e-01 8.45069468e-01
-8.65318477e-01 -3.44433933e-01 4.41881627e-01 4.23065722e-01
4.90491956e-01 3.84619236e-01 -3.84047002e-01 -2.36007094e-01
-1.02151704e+00 -9.63449553e-02 7.91575789e-01 7.96078861e-01
3.48881096e-01 2.55181253e-01 1.22368775e-01 7.36239612e-01
-9.34457183e-02 3.46560061e-01 1.17834949e+00 -9.36995223e-02
2.97197700e-01 7.09741235e-01 -1.37077332e-01 -1.10036194e+00
-6.79886341e-01 -7.46641099e-01 -1.33982873e+00 7.25229159e-02
-1.01241983e-01 -3.44080985e-01 -4.36277479e-01 1.15245855e+00
7.89657831e-02 4.54599351e-01 8.92349109e-02 9.04545069e-01
4.18095291e-01 5.49124539e-01 -3.30222726e-01 -4.92536366e-01
1.41599488e+00 -2.95206249e-01 -1.10296428e+00 -9.31209847e-02
5.92875004e-01 -3.65228325e-01 6.37718141e-01 8.34092200e-01
-7.39017725e-01 -2.17139080e-01 -9.49433982e-01 6.91976249e-01
-4.13645446e-01 1.33065552e-01 6.36864245e-01 4.40690011e-01
-1.09655547e+00 5.31722546e-01 -9.39724863e-01 -1.27847001e-01
4.40850586e-01 7.56573737e-01 -4.48232055e-01 4.72054668e-02
-1.05342162e+00 8.98438156e-01 4.83457029e-01 4.22107428e-01
-3.77173871e-01 -5.60257137e-01 -3.56413186e-01 4.15654123e-01
-1.85365632e-01 -2.31331393e-01 6.95894301e-01 -1.03775918e+00
-1.41040039e+00 1.60166144e-01 -9.84267518e-02 -5.69290876e-01
1.70718566e-01 -9.14523900e-02 -7.63819456e-01 1.72735557e-01
-2.25713074e-01 -1.17969573e-01 1.03279054e+00 -5.82798362e-01
-6.31957233e-01 -8.48878980e-01 -7.06238747e-01 6.27810284e-02
-9.48023200e-01 1.73794433e-01 -1.40264081e-02 -4.90005136e-01
1.30998075e-01 -8.39502573e-01 -1.69771072e-02 -6.37929261e-01
-4.08845916e-02 -9.71422046e-02 1.05261087e+00 -8.33910525e-01
1.36663651e+00 -2.21832466e+00 1.06009975e-01 5.04341424e-01
3.37035447e-01 4.12551880e-01 1.53431147e-01 1.75187603e-01
-4.79294568e-01 -5.89176774e-01 -1.29539538e-02 1.48299798e-01
-5.44573724e-01 -2.57569570e-02 -1.15480088e-02 3.33083689e-01
1.55471325e-01 4.83508497e-01 -4.51504767e-01 -9.46921036e-02
3.30452174e-01 7.62168229e-01 -4.48877811e-01 5.33337951e-01
4.78638053e-01 6.92120194e-01 -3.54211509e-01 3.62560391e-01
5.07823348e-01 -3.16785336e-01 -2.02814452e-02 -6.69499412e-02
-5.40973172e-02 -1.18540913e-01 -1.15720177e+00 1.35127032e+00
-4.99312669e-01 6.68651462e-01 4.95846160e-02 -1.56904566e+00
9.85689580e-01 7.32494533e-01 9.79934156e-01 -6.16563678e-01
4.44275856e-01 5.31586826e-01 4.18027669e-01 -7.38234878e-01
-1.30696356e-01 1.24064274e-01 3.68741572e-01 4.23350990e-01
-2.33422942e-03 2.77016610e-02 -1.45928055e-01 -1.02874711e-01
1.25035381e+00 -5.47987521e-01 1.50841057e-01 -4.83669072e-01
6.02536619e-01 -1.35041445e-01 7.43261635e-01 -5.36028892e-02
-1.04442172e-01 2.44459525e-01 -6.09962754e-02 -6.26113892e-01
-5.83860397e-01 -2.17358485e-01 -2.90082693e-01 4.03142154e-01
-7.17585608e-02 -3.63239378e-01 -9.36429858e-01 4.78390120e-02
-3.04356307e-01 5.34930289e-01 -2.84396350e-01 -3.68196577e-01
-3.82596493e-01 -1.03885591e+00 -1.87875301e-01 2.84862101e-01
3.69666308e-01 -1.31017232e+00 -1.30019224e+00 6.03661001e-01
-1.11383878e-01 -9.30899143e-01 2.99075954e-02 4.82021153e-01
-8.76785934e-01 -9.11272287e-01 -6.30348086e-01 -7.70747602e-01
5.61631143e-01 2.48382539e-01 6.84911966e-01 1.41235143e-01
-5.03791332e-01 -4.05482687e-02 -5.16276598e-01 -8.92582417e-01
6.11413158e-02 -2.27515593e-01 4.53879833e-01 3.25820386e-01
6.96296871e-01 -8.65705192e-01 -9.00609612e-01 1.29429072e-01
-7.48355329e-01 9.67754871e-02 5.25562108e-01 9.79398191e-01
4.43518549e-01 6.30573034e-01 8.92517328e-01 -2.64859110e-01
1.02191627e+00 -5.30123413e-01 -6.65292740e-01 1.81148797e-01
-6.03815675e-01 -2.47211918e-01 8.41567993e-01 -4.62230951e-01
-7.36251295e-01 -4.46661301e-02 -1.04219774e-02 -3.50718439e-01
-1.45940587e-01 5.66172123e-01 -2.07328945e-01 -1.35834903e-01
3.75057995e-01 5.42440712e-01 1.41922668e-01 -4.55197960e-01
-4.68313277e-01 1.43523073e+00 2.01106414e-01 1.30821988e-01
2.16731444e-01 5.04063033e-02 -2.04643220e-01 -1.07699037e+00
-1.10036463e-01 -4.78423268e-01 -5.58021009e-01 -2.30685890e-01
6.71165168e-01 -9.28757429e-01 -8.13988149e-01 6.58650517e-01
-1.11714375e+00 -4.21313159e-02 -8.52206051e-02 9.54949081e-01
-4.03493106e-01 7.74093643e-02 -3.04295182e-01 -6.89767063e-01
-9.32554662e-01 -1.40753412e+00 5.56188226e-01 5.55284694e-02
-1.00288905e-01 -4.57726568e-01 -1.52890071e-01 1.67362392e-01
6.44849062e-01 4.92850952e-02 7.44031370e-01 -7.91616380e-01
-3.95956695e-01 -6.36073768e-01 -1.36461258e-01 4.47756350e-01
3.18591595e-01 -4.85131741e-01 -8.28630269e-01 -5.55229485e-01
7.18306839e-01 5.80116436e-02 3.00396932e-04 5.12202024e-01
1.69637477e+00 -2.69431531e-01 -3.62970024e-01 6.08907282e-01
1.33676088e+00 7.53961563e-01 5.91283500e-01 2.62187302e-01
4.74118203e-01 4.01942521e-01 4.88327622e-01 8.39879930e-01
1.17305435e-01 4.64524388e-01 2.53295362e-01 2.31000841e-01
3.23658794e-01 6.19069636e-01 1.83040440e-01 1.37134898e+00
-1.39707088e-01 6.78039715e-02 -1.06045926e+00 5.55230737e-01
-1.72643530e+00 -9.28278148e-01 -2.62830015e-02 2.40957189e+00
6.35657668e-01 -2.84037352e-01 -7.02997521e-02 7.43776202e-01
6.00481212e-01 -6.25567615e-01 -4.95122105e-01 -1.79706380e-01
1.87441677e-01 4.10048515e-01 3.35204452e-02 -1.20766252e-01
-6.42135441e-01 6.06666267e-01 5.35679770e+00 6.70227408e-01
-1.78777027e+00 3.32336545e-01 3.25617641e-01 -4.00135815e-01
1.89171687e-01 -4.85109985e-01 -4.99642730e-01 1.01377654e+00
1.50245345e+00 -2.29019597e-01 7.49432385e-01 7.42586672e-01
4.38713819e-01 -2.63310056e-02 -8.94397199e-01 1.68490577e+00
3.20332916e-03 -1.20373678e+00 -3.03042442e-01 2.56472826e-01
2.00024992e-01 2.12649807e-01 -4.47469056e-01 -8.55458826e-02
-5.75544417e-01 -9.03254688e-01 4.60757256e-01 6.90343022e-01
7.22965479e-01 -1.00770104e+00 9.65300679e-01 6.74896598e-01
-7.90546954e-01 -4.45995063e-01 -4.80630785e-01 -2.68598139e-01
-6.21080324e-02 9.02707160e-01 -6.55780733e-01 3.47822577e-01
1.07178175e+00 6.27081215e-01 -3.77979010e-01 1.20593274e+00
1.44589171e-01 4.33256775e-01 -3.51989746e-01 -3.31811249e-01
-2.70543516e-01 -4.34530884e-01 3.61406565e-01 9.44015026e-01
8.09647977e-01 5.29227436e-01 -2.33655572e-01 3.60589772e-01
3.98085058e-01 2.78896093e-01 -7.24399686e-01 -8.41706097e-02
4.25165594e-01 1.27804518e+00 -5.83036125e-01 -2.93965518e-01
-3.43799651e-01 9.58840966e-01 1.95056200e-01 5.18112183e-02
-4.47378963e-01 -7.13717461e-01 5.30463994e-01 1.73657879e-01
-1.10658720e-01 -8.63529518e-02 -5.28376341e-01 -1.57370067e+00
1.99540988e-01 -1.00875890e+00 6.73973113e-02 -6.44546449e-01
-9.67854500e-01 1.25356817e+00 -2.54036129e-01 -1.68566775e+00
-4.78855491e-01 -5.27974963e-01 -6.63953125e-01 8.40099394e-01
-1.30884445e+00 -5.19459724e-01 -5.76891899e-01 1.26635313e+00
4.91883606e-01 -7.25229025e-01 1.24818599e+00 5.01284540e-01
-7.32841551e-01 2.52316594e-01 4.17030394e-01 -1.28866300e-01
2.96663940e-01 -7.73230910e-01 -3.22479993e-01 8.41171622e-01
4.26953584e-02 4.31578428e-01 5.21600723e-01 -3.64652783e-01
-1.66752148e+00 -9.39437330e-01 7.95207381e-01 8.67939889e-02
6.31489277e-01 -3.94841194e-01 -1.04687500e+00 2.57679582e-01
3.00659299e-01 3.04759368e-02 1.03339076e+00 -1.50039747e-01
5.20616949e-01 -5.93213081e-01 -1.12497473e+00 2.88731188e-01
6.29040778e-01 -2.53219992e-01 -6.40341222e-01 5.91165364e-01
4.90366295e-02 1.66763477e-02 -1.07655621e+00 2.62231469e-01
3.51556301e-01 -9.23159778e-01 5.88269055e-01 -2.49009743e-01
6.24793619e-02 1.28253205e-02 3.20494585e-02 -1.80435228e+00
-2.03650028e-01 -4.82268125e-01 -1.16332233e-01 7.20105350e-01
7.90298581e-02 -9.18349504e-01 6.07173681e-01 9.64060545e-01
-5.02907708e-02 -9.36993837e-01 -1.07080901e+00 -7.02843130e-01
-3.32795382e-01 -6.32047772e-01 9.15215254e-01 6.91956460e-01
7.16609836e-01 2.09273994e-01 -2.13119432e-01 1.48631990e-01
5.44395506e-01 2.32144982e-01 4.18989629e-01 -1.61591506e+00
2.36714371e-02 -6.17088713e-02 -9.40490782e-01 -9.87688601e-02
8.57465342e-02 -7.15094566e-01 -1.69584647e-01 -1.31862521e+00
-1.01160454e-02 -5.02754271e-01 -6.06185913e-01 3.34104508e-01
2.61391371e-01 1.32369641e-02 4.78958599e-02 4.29727674e-01
9.65217724e-02 5.94997823e-01 8.04610610e-01 -1.33564966e-02
-5.16793847e-01 2.77541786e-01 -4.17618901e-01 6.06624901e-01
9.25415754e-01 -4.75671738e-01 -6.62308812e-01 -4.98968989e-01
-1.30845800e-01 4.96740639e-01 -5.41104153e-02 -1.50452077e+00
4.97374266e-01 1.50577173e-01 6.15233302e-01 -6.93961084e-01
4.73543853e-01 -1.36231029e+00 3.34026903e-01 6.11199439e-01
3.53590876e-01 1.30781189e-01 1.37621671e-01 3.55142772e-01
-6.17930949e-01 -2.24623177e-02 5.99572062e-01 1.46684095e-01
-4.81568485e-01 3.66328686e-01 -6.00232899e-01 -5.17023146e-01
1.43560278e+00 -3.87776792e-01 2.42952853e-01 -2.31400698e-01
-7.78770387e-01 -5.39545044e-02 -1.73821107e-01 4.62721556e-01
1.02608931e+00 -1.15494704e+00 -4.85313416e-01 8.34627807e-01
5.10853566e-02 -4.57390286e-02 3.42818916e-01 1.07479024e+00
-4.57403719e-01 5.71419895e-01 -7.45745838e-01 -5.64547956e-01
-1.31174362e+00 4.85952765e-01 1.24442905e-01 8.02787617e-02
-1.04844773e+00 5.91655791e-01 -1.41821802e-01 1.68611497e-01
2.58458108e-01 -1.67921990e-01 -4.86076266e-01 1.81314602e-01
9.55229878e-01 1.24053791e-01 6.92067742e-01 -4.72132504e-01
-5.97435057e-01 2.28119105e-01 9.90185793e-03 7.25239590e-02
2.00234723e+00 7.67634436e-02 -4.64145631e-01 1.29026473e-01
9.55666006e-01 -5.71220815e-01 -6.78019285e-01 4.67060246e-02
7.10457787e-02 -5.28175414e-01 4.83463615e-01 -6.17209375e-01
-1.47440207e+00 1.19921827e+00 9.86655414e-01 2.63144642e-01
1.59986031e+00 -5.41949034e-01 6.53348207e-01 5.92375875e-01
9.32109237e-01 -1.18071437e+00 -3.25031161e-01 3.01319629e-01
1.14348507e+00 -1.08090055e+00 -3.08192462e-01 8.07542652e-02
-6.97204351e-01 1.27610719e+00 4.14308786e-01 -1.93548247e-01
1.27595985e+00 5.64669728e-01 -1.01396374e-01 -3.05013239e-01
-6.09381855e-01 2.30358094e-01 8.88162255e-02 4.54063535e-01
2.87056953e-01 2.05194429e-01 -5.43278217e-01 1.08563721e+00
-4.49918419e-01 4.99008238e-01 4.73568976e-01 9.12557065e-01
-2.49390751e-01 -8.19507420e-01 -4.83878165e-01 9.08294857e-01
-4.67870116e-01 -6.34568036e-02 6.97283223e-02 2.04025209e-01
3.19139548e-02 1.20755863e+00 1.42103106e-01 -6.81326866e-01
1.90330133e-01 1.43105701e-01 2.03316197e-01 -5.84814489e-01
-4.86121535e-01 1.53119937e-01 -5.07305324e-01 -5.99321067e-01
-1.13679342e-01 -3.93362939e-01 -1.14818776e+00 -1.15859933e-01
-5.81286132e-01 4.23045605e-01 1.23291802e+00 9.52775061e-01
9.36177313e-01 6.93720341e-01 9.07426775e-01 -9.19384301e-01
-4.42812145e-01 -1.20886016e+00 -1.03821409e+00 3.12746227e-01
1.90160032e-02 -6.23189211e-01 -2.45116562e-01 -1.05501227e-01] | [13.235640525817871, 3.4766790866851807] |
8d8e56d9-8db4-41ec-b910-1ef5560b25aa | how-do-we-answer-complex-questions-discourse | null | null | https://openreview.net/forum?id=shZUViLG-DS | https://openreview.net/pdf?id=shZUViLG-DS | How do we answer complex questions: Discourse structure of long form answers | Long form answers, consisting of multiple sentences, can provide nuanced and comprehensive answers to a broader set of questions. However, little prior work exists on this task. To better understand this complex task, we study the functional structure of long form answers on two datasets, Natural Questions~\cite{kwiatkowski2019natural} and ELI5~\cite{Fan2019ELI5LF}. Our main goal is to understand how humans organize information to craft complex answers. We develop an ontology of sentence-level functional roles for long form answers, and annotate 3.3k sentences in 542 examples. Our annotated data enables training a reliable role classifier that can be used for automatic analysis and thus reveals machine generated answers are structured worse than human written answers. Our data further yields an extractive summarization dataset for long form answers, giving models the ability to identify a concise answer to a complex query. | ['Anonymous'] | 2021-11-16 | null | null | null | acl-arr-november-2021-11 | ['extractive-summarization'] | ['natural-language-processing'] | [ 4.53165203e-01 6.61155999e-01 -1.76689088e-01 -3.70409489e-01
-1.16061056e+00 -1.10033023e+00 7.57519186e-01 7.94258356e-01
-3.43274832e-01 1.02220869e+00 1.07042122e+00 -2.82315999e-01
-4.17747974e-01 -5.19133270e-01 -4.27825063e-01 -1.44409314e-01
3.55368823e-01 6.15629971e-01 4.90717292e-01 -6.26385212e-01
8.37258339e-01 -1.04857869e-01 -1.44758546e+00 8.37444067e-01
1.09541821e+00 6.28207207e-01 2.82196134e-01 1.15190291e+00
-4.48327899e-01 1.38339067e+00 -1.16630971e+00 -6.74524486e-01
-3.64567816e-01 -6.60855055e-01 -1.53009784e+00 -1.51095778e-01
5.55954218e-01 -1.81496397e-01 8.48588254e-03 5.70736229e-01
6.35082424e-01 7.28372037e-02 8.27447474e-01 -8.81300330e-01
-5.70865452e-01 7.40174472e-01 5.03108241e-02 6.95590019e-01
1.28549325e+00 9.95612666e-02 1.72210205e+00 -4.51214314e-01
1.06152081e+00 1.38799012e+00 4.49740618e-01 8.06464314e-01
-1.19762528e+00 -1.33024588e-01 -1.28166705e-01 3.02207351e-01
-5.16520083e-01 -6.81581557e-01 5.82247138e-01 -3.51693749e-01
1.17811036e+00 6.72533333e-01 4.16587025e-01 8.45702708e-01
1.07704602e-01 1.12016571e+00 7.88601577e-01 -3.80460858e-01
-8.55517462e-02 -1.20073624e-01 8.55616629e-01 5.87209642e-01
2.82885075e-01 -8.95687819e-01 -8.73941839e-01 -4.58313584e-01
1.19136848e-01 -4.00444359e-01 -6.12682998e-01 4.33875054e-01
-1.18034112e+00 8.71889651e-01 1.98736507e-02 3.44802290e-01
-1.90644443e-01 7.96181411e-02 4.77509767e-01 7.97226429e-01
3.55188340e-01 1.49118745e+00 -9.26575184e-01 -3.95324349e-01
-5.24508238e-01 7.12087572e-01 1.43245101e+00 8.63559306e-01
6.40095711e-01 -5.24566472e-01 -4.81097072e-01 8.76181901e-01
-1.53309613e-01 1.92651123e-01 5.71760118e-01 -1.59794450e+00
6.20262742e-01 9.43950653e-01 2.84092814e-01 -1.01499295e+00
-3.94544125e-01 -1.38956249e-01 -3.69347930e-01 -7.85427153e-01
5.31158388e-01 -1.14297725e-01 -5.18759452e-02 1.49062586e+00
3.31662521e-02 -7.73022771e-01 2.55105227e-01 4.81284201e-01
1.53372049e+00 7.25955725e-01 -3.11286688e-01 -2.83456951e-01
1.51484668e+00 -9.81169999e-01 -8.66064608e-01 -2.01818958e-01
8.33125234e-01 -7.55338132e-01 1.42477024e+00 2.14662597e-01
-1.11395192e+00 -3.80694300e-01 -8.91936541e-01 -6.91421509e-01
-8.12352598e-02 7.60509297e-02 6.11980855e-01 4.70638275e-02
-1.10276079e+00 3.77827108e-01 -5.26921311e-03 -5.69013953e-01
1.74636632e-01 5.12375645e-02 -2.95550495e-01 -2.01603681e-01
-1.23844111e+00 9.01811004e-01 2.49472961e-01 -4.64723498e-01
-2.21031070e-01 -8.76212895e-01 -8.52497160e-01 -1.07052429e-02
7.22279787e-01 -1.03529537e+00 1.75402832e+00 -3.27827215e-01
-1.14649653e+00 1.02252364e+00 -5.51195979e-01 -4.97397780e-01
6.07322007e-02 -2.62667626e-01 -6.65338933e-02 8.72288883e-01
6.48724794e-01 7.35229313e-01 6.26269639e-01 -1.05679727e+00
-4.48498905e-01 -3.33142251e-01 6.13595545e-01 5.00901602e-02
-2.52840966e-01 3.54310453e-01 -1.11112788e-01 -7.23637521e-01
-1.38377801e-01 -3.19360435e-01 1.81719437e-02 -4.15116042e-01
-5.39237261e-01 -9.21669662e-01 2.33811483e-01 -8.81599247e-01
1.72771764e+00 -1.41709709e+00 2.56640166e-01 -4.75673586e-01
7.98758507e-01 -8.84115472e-02 -2.30322182e-01 9.63121295e-01
2.15071395e-01 6.55286849e-01 -4.08141315e-01 8.60026330e-02
2.17112973e-01 3.42591375e-01 -5.80600023e-01 -2.19327688e-01
3.58467668e-01 1.28413332e+00 -1.01811659e+00 -7.53447115e-01
-5.31309843e-01 -5.40281236e-01 -7.72383451e-01 3.97064000e-01
-8.10146153e-01 1.94020197e-01 -6.56047344e-01 5.07255256e-01
-4.60229069e-02 -3.26227874e-01 -1.82148546e-01 6.65811896e-02
5.49676456e-02 7.38629460e-01 -3.91032398e-01 1.45098066e+00
-3.11405778e-01 7.64437973e-01 2.78374642e-01 -7.74306178e-01
5.53389788e-01 3.14646453e-01 2.71873266e-01 -6.12850666e-01
-1.58760753e-02 2.29993269e-01 3.24127935e-02 -1.12512064e+00
7.40437925e-01 -5.85948303e-02 -6.72282994e-01 8.84566367e-01
1.38357863e-01 -8.88483465e-01 8.92899930e-01 7.10758150e-01
1.56398511e+00 -3.77016962e-01 5.53746164e-01 -3.71491939e-01
7.82724738e-01 3.72922599e-01 2.55713940e-01 1.12570107e+00
1.09771952e-01 4.96790379e-01 1.07789123e+00 -3.56743217e-01
-6.96502030e-01 -6.68254435e-01 5.61500713e-02 1.29530549e+00
-1.68441772e-01 -9.83527184e-01 -7.67480731e-01 -6.90410554e-01
6.92739561e-02 9.04380560e-01 -3.73122424e-01 -6.79761032e-03
-7.82033205e-01 -2.11119831e-01 7.75714397e-01 1.90484285e-01
3.80377412e-01 -1.31011069e+00 -6.85126960e-01 1.87038988e-01
-8.61661851e-01 -9.79153097e-01 -5.53965151e-01 -1.22581758e-02
-7.61412203e-01 -1.59226573e+00 -2.90952772e-01 -6.10019386e-01
3.20057750e-01 1.51423052e-01 1.81686044e+00 5.08972406e-01
-2.31095194e-03 8.04265022e-01 -6.05097294e-01 -5.56455553e-01
-6.41079009e-01 4.32775527e-01 -1.50770620e-01 -4.08532888e-01
2.28548124e-01 -5.40560067e-01 -3.73210490e-01 1.11610599e-01
-1.14729083e+00 -2.20808595e-01 5.03957391e-01 6.33258641e-01
1.44602284e-01 -2.97358543e-01 1.15925288e+00 -1.19879746e+00
1.52898824e+00 -5.50444841e-01 5.10969572e-02 4.87847984e-01
-3.01040620e-01 5.62614739e-01 7.72893071e-01 -8.12136009e-02
-1.18778408e+00 -4.92356777e-01 -4.72075641e-01 6.23848259e-01
-6.03381656e-02 6.18307829e-01 -1.35419458e-01 5.32500148e-01
9.94587541e-01 2.00505301e-01 -1.62720874e-01 -6.47815943e-01
4.90144640e-01 6.43235743e-01 6.86824203e-01 -9.23233867e-01
5.86540937e-01 1.73318293e-02 -3.76770586e-01 -1.26608515e+00
-1.45050979e+00 -8.20328414e-01 -4.55343932e-01 -1.38184085e-01
8.98960710e-01 -4.01616484e-01 -7.55805314e-01 -8.46403930e-03
-1.46332896e+00 -1.74392313e-01 -3.37087423e-01 -3.04796696e-01
-7.06643105e-01 5.15976369e-01 -6.73932910e-01 -5.55202842e-01
-5.94953358e-01 -4.20034528e-01 1.07141554e+00 1.26666874e-01
-1.10106659e+00 -9.23434138e-01 7.29330555e-02 1.17561388e+00
9.24753845e-02 1.95080042e-01 1.55820906e+00 -1.12369752e+00
-3.79081279e-01 -1.76330447e-01 8.77613276e-02 1.82058379e-01
-1.26563370e-01 -1.85608149e-01 -5.78910172e-01 3.20609272e-01
3.80879372e-01 -8.88808906e-01 1.10582769e+00 -6.44708872e-02
1.06484663e+00 -9.63834703e-01 7.59817287e-02 -3.38256478e-01
7.02947915e-01 -2.08727598e-01 5.27834356e-01 1.08804451e-02
2.90639460e-01 1.28003013e+00 5.53383589e-01 2.02562824e-01
7.15409458e-01 2.00459197e-01 -1.83103550e-02 6.79103613e-01
-8.15803632e-02 -3.88835490e-01 2.15508163e-01 1.07486129e+00
2.72201568e-01 -4.87590492e-01 -8.47908258e-01 6.04489267e-01
-1.72782934e+00 -1.28119576e+00 -6.30512476e-01 1.51304448e+00
1.27214265e+00 1.13550693e-01 -1.00929365e-01 3.11098993e-01
3.21043879e-01 4.53260630e-01 -3.62726241e-01 -3.13435614e-01
-1.19845204e-01 3.17867965e-01 -2.89865583e-01 5.93234420e-01
-6.96041346e-01 8.08927715e-01 6.42974281e+00 8.40060890e-01
-3.20125878e-01 -9.78707671e-02 2.66378045e-01 3.47787179e-02
-8.66998851e-01 1.58088002e-02 -6.64415300e-01 2.18850195e-01
8.45810413e-01 -4.39782947e-01 1.72673598e-01 3.72747958e-01
3.53267372e-01 -5.05278885e-01 -1.24832475e+00 6.64211035e-01
4.09532577e-01 -1.65216553e+00 2.55892336e-01 -4.27261353e-01
6.08801603e-01 -4.73959416e-01 -5.22427559e-01 4.77498084e-01
3.13951820e-01 -1.04119301e+00 5.99569678e-01 7.98328817e-01
4.84985650e-01 -3.99078965e-01 7.42161930e-01 9.43569422e-01
-8.33042085e-01 -3.43545496e-01 -2.61254460e-01 -4.80138570e-01
5.12543954e-02 3.83344859e-01 -7.62973070e-01 5.14183164e-01
5.03729939e-01 6.55766189e-01 -1.05631888e+00 6.26463532e-01
-7.66625047e-01 8.40802133e-01 -3.80049609e-02 -6.36718631e-01
-4.72105406e-02 1.79746136e-01 6.42432868e-01 9.57410455e-01
-2.74321586e-01 6.29820406e-01 1.96916297e-01 6.06868267e-01
-4.51557636e-01 1.51436374e-01 -5.08636057e-01 -5.32375395e-01
3.91602576e-01 1.08138764e+00 -6.65351808e-01 -3.71008366e-01
1.65217463e-02 8.69427681e-01 3.75212461e-01 2.01812178e-01
-1.46839410e-01 -5.45913994e-01 2.09940180e-01 3.61636698e-01
-3.46169293e-01 -8.98452029e-02 -2.94526964e-01 -1.36932838e+00
2.36275226e-01 -1.25525880e+00 7.39839435e-01 -1.08992803e+00
-1.40594518e+00 1.86914593e-01 1.92311063e-01 -5.58098733e-01
-6.03146255e-01 -4.30250198e-01 -8.14069033e-01 4.45233852e-01
-1.22268951e+00 -9.00144219e-01 -1.04654089e-01 2.66166449e-01
1.03875589e+00 1.20023809e-01 8.37315083e-01 -7.04035684e-02
-3.45661461e-01 1.51115298e-01 -3.23412299e-01 -1.35944057e-02
8.06271493e-01 -1.56112945e+00 8.01101923e-02 4.71652448e-01
4.44524847e-02 9.59976852e-01 9.22566652e-01 -5.76068878e-01
-1.35835230e+00 -6.63605154e-01 1.60420132e+00 -1.22851980e+00
7.41260469e-01 -1.52829021e-01 -8.04758906e-01 3.25575024e-01
6.02823496e-01 -7.75518715e-01 1.08434606e+00 -1.36718797e-02
-2.42643461e-01 2.02139780e-01 -1.01057422e+00 4.80902642e-01
1.09779894e+00 -7.07923055e-01 -1.60547423e+00 3.97488743e-01
1.29836690e+00 1.31053492e-01 -6.84451818e-01 1.54163599e-01
2.90027082e-01 -1.00914955e+00 6.87607288e-01 -1.00178134e+00
1.01073825e+00 -1.14171453e-01 -2.70158872e-02 -1.11286986e+00
8.52627382e-02 -8.64515483e-01 -3.25457662e-01 1.27234387e+00
6.75574243e-01 -4.88389850e-01 4.11594689e-01 6.48273945e-01
-4.39465106e-01 -9.04889226e-01 -7.00099766e-01 -2.46682465e-01
1.02335796e-01 -3.46044540e-01 2.68115819e-01 6.94923878e-01
2.09830120e-01 1.06163073e+00 2.13225245e-01 -4.63611722e-01
1.98336318e-01 2.61050075e-01 8.23361814e-01 -1.52910924e+00
-2.55512357e-01 -5.88979661e-01 1.83509871e-01 -1.29640079e+00
4.52125221e-01 -9.40584540e-01 -8.62933919e-02 -2.20293474e+00
3.39258790e-01 2.77051181e-01 6.13360882e-01 3.36081594e-01
-4.64167863e-01 -5.10517247e-02 5.22708930e-02 2.78484017e-01
-1.16711605e+00 5.76484203e-01 1.30238807e+00 -2.24617630e-01
1.06238397e-02 1.87271044e-01 -1.37085843e+00 9.78592753e-01
6.16376758e-01 -4.65894282e-01 -5.95601857e-01 -3.50806594e-01
8.01719904e-01 3.63234133e-01 1.35202110e-01 -4.40149426e-01
3.37377876e-01 -1.01943016e-01 1.10177658e-01 -6.09395683e-01
-2.52780076e-02 -2.11622685e-01 -6.16744816e-01 2.80919403e-01
-8.72568548e-01 2.67884761e-01 -1.52336851e-01 5.88497758e-01
-4.32893455e-01 -8.76252592e-01 1.07700020e-01 -5.28665364e-01
-3.86436373e-01 -2.10049525e-01 -7.25472808e-01 1.05561149e+00
4.35765117e-01 -5.72662167e-02 -6.72314823e-01 -9.67829287e-01
-4.15730089e-01 8.81534338e-01 2.31876135e-01 4.73573983e-01
6.58281446e-01 -8.59089971e-01 -9.13537860e-01 -4.98060584e-01
3.50730151e-01 6.39330968e-02 1.34318918e-01 4.81794596e-01
-6.85885787e-01 6.80450857e-01 3.21887821e-01 -1.86384574e-01
-1.22035313e+00 2.20860884e-01 5.32173812e-02 -5.81280410e-01
-2.33309284e-01 1.03292346e+00 -1.07358120e-01 -6.95781291e-01
-3.55750620e-02 -4.73090380e-01 -7.43907332e-01 8.06817055e-01
5.71580291e-01 4.10490036e-01 -9.89828333e-02 -2.50624061e-01
-7.42121264e-02 2.60694712e-01 4.60287705e-02 -1.81472659e-01
1.35499573e+00 -1.26264885e-01 -8.79573226e-01 4.54228759e-01
1.18576229e+00 4.53836232e-01 -5.89678586e-01 -5.59061803e-02
7.63000429e-01 -1.52682364e-01 -7.73288012e-01 -8.75379443e-01
9.43516269e-02 6.19772017e-01 -6.65772140e-01 6.76761150e-01
9.34095681e-01 6.02900326e-01 9.01688933e-01 1.16579032e+00
3.37782912e-02 -1.10050750e+00 8.68402541e-01 1.00124753e+00
1.53098118e+00 -1.04407048e+00 4.90825735e-02 -3.47798526e-01
-5.76282918e-01 1.23481989e+00 6.03393853e-01 3.32059592e-01
7.01363804e-03 -2.21524760e-01 -1.00016192e-01 -6.73254073e-01
-1.21782637e+00 -2.49808550e-01 3.69839191e-01 2.82004207e-01
4.36325163e-01 -3.38103086e-01 -7.44170070e-01 1.09917057e+00
-8.60156178e-01 -4.57755476e-01 1.01000059e+00 9.88721490e-01
-8.69798362e-01 -9.95182872e-01 -2.55874544e-01 9.14572179e-01
-7.19653666e-01 -1.21683456e-01 -1.31768203e+00 4.20934826e-01
-2.28912175e-01 1.65044999e+00 -4.23940539e-01 -4.31412570e-02
3.89099598e-01 5.24571359e-01 3.65673721e-01 -1.24321020e+00
-8.69182587e-01 -4.94075447e-01 8.86153102e-01 -2.33884811e-01
-4.72218305e-01 -5.70820510e-01 -1.35947287e+00 -1.00707553e-01
-4.24027778e-02 6.07151866e-01 1.28102591e-02 1.07670772e+00
4.20019835e-01 5.64634025e-01 2.76160359e-01 -3.65752131e-01
-7.75576234e-01 -1.16254878e+00 8.25351849e-02 4.30174083e-01
3.89922440e-01 -4.51665930e-02 -4.72260088e-01 8.51900950e-02] | [11.62386703491211, 8.205540657043457] |
88a9da2b-109f-43ad-ad86-ae551b560818 | pose-graph-optimization-for-unsupervised | 1903.06315 | null | http://arxiv.org/abs/1903.06315v1 | http://arxiv.org/pdf/1903.06315v1.pdf | Pose Graph Optimization for Unsupervised Monocular Visual Odometry | Unsupervised Learning based monocular visual odometry (VO) has lately drawn
significant attention for its potential in label-free leaning ability and
robustness to camera parameters and environmental variations. However,
partially due to the lack of drift correction technique, these methods are
still by far less accurate than geometric approaches for large-scale odometry
estimation. In this paper, we propose to leverage graph optimization and loop
closure detection to overcome limitations of unsupervised learning based
monocular visual odometry. To this end, we propose a hybrid VO system which
combines an unsupervised monocular VO called NeuralBundler with a pose graph
optimization back-end. NeuralBundler is a neural network architecture that uses
temporal and spatial photometric loss as main supervision and generates a
windowed pose graph consists of multi-view 6DoF constraints. We propose a novel
pose cycle consistency loss to relieve the tensions in the windowed pose graph,
leading to improved performance and robustness. In the back-end, a global pose
graph is built from local and loop 6DoF constraints estimated by NeuralBundler
and is optimized over SE(3). Empirical evaluation on the KITTI odometry dataset
demonstrates that 1) NeuralBundler achieves state-of-the-art performance on
unsupervised monocular VO estimation, and 2) our whole approach can achieve
efficient loop closing and show favorable overall translational accuracy
compared to established monocular SLAM systems. | ['Yang Li', 'Yoshitaka Ushiku', 'Tatsuya Harada'] | 2019-03-15 | null | null | null | null | ['loop-closure-detection', 'monocular-visual-odometry'] | ['computer-vision', 'robots'] | [-2.44692847e-01 4.31211516e-02 -4.24637049e-01 -3.75813335e-01
-3.58678102e-01 -4.38083053e-01 5.26458502e-01 -3.04196067e-02
-4.38141584e-01 7.06794500e-01 4.11136709e-02 8.87647718e-02
-4.41419855e-02 -4.59809333e-01 -9.62406456e-01 -4.55443531e-01
7.53600895e-02 7.43314564e-01 2.34524548e-01 -1.32327527e-01
9.25586447e-02 6.67916059e-01 -1.41249406e+00 -7.02748120e-01
9.88824487e-01 8.16664279e-01 1.98362112e-01 5.53678811e-01
3.74482334e-01 8.60595167e-01 -1.78124160e-01 8.80619287e-02
4.51041400e-01 -1.10520460e-01 -4.27549809e-01 2.18804494e-01
1.22875500e+00 -2.76110440e-01 -6.28116548e-01 1.00981319e+00
6.50220573e-01 1.59761280e-01 3.54644150e-01 -1.26205146e+00
-1.65004432e-01 -5.03333434e-02 -3.42461139e-01 -3.31030875e-01
4.27892298e-01 1.83414191e-01 1.01217306e+00 -9.94931281e-01
1.22087085e+00 9.59073961e-01 9.62134302e-01 2.17180625e-01
-1.48244143e+00 -4.44077194e-01 4.58789170e-02 -2.77920831e-02
-1.59241617e+00 -4.37071294e-01 8.64465415e-01 -6.06977344e-01
1.24505317e+00 -1.32099435e-01 8.35571051e-01 6.75401151e-01
5.31340659e-01 4.55738932e-01 8.04868460e-01 -3.75272274e-01
7.62817413e-02 -1.74354238e-03 -1.84551165e-01 1.22740173e+00
4.49631184e-01 3.73306185e-01 -8.54935288e-01 1.16881862e-01
6.26631439e-01 6.41656853e-03 -3.51384103e-01 -1.62666333e+00
-1.17586184e+00 8.00378382e-01 9.47301030e-01 -3.21148574e-01
-8.88093859e-02 4.33931172e-01 1.95218742e-01 2.07951099e-01
4.29642528e-01 5.40610492e-01 -2.90133506e-01 1.10059485e-01
-9.40429151e-01 3.44794214e-01 6.56541049e-01 1.15095556e+00
1.30934191e+00 3.02402586e-01 3.82932574e-01 7.11489201e-01
5.26269495e-01 7.16621578e-01 3.21273357e-01 -1.06869698e+00
4.87216711e-01 9.37018335e-01 -5.83736636e-02 -1.12991035e+00
-8.02779198e-01 -6.11824036e-01 -5.68692327e-01 4.32695478e-01
1.10252857e-01 -1.31981317e-02 -1.10549247e+00 1.57408917e+00
5.24815440e-01 -2.72206664e-01 -7.87422806e-02 1.16097188e+00
6.49875522e-01 7.21742064e-02 -5.50632060e-01 1.44591749e-01
8.14845502e-01 -1.02093685e+00 -5.03357828e-01 -7.19770849e-01
8.20776999e-01 -7.04018056e-01 5.64579487e-01 2.82837629e-01
-6.29550934e-01 -4.36129302e-01 -1.65002906e+00 -5.43551505e-01
-2.42780313e-01 2.26626217e-01 4.65999067e-01 4.43535715e-01
-1.03766584e+00 4.02137816e-01 -9.55097795e-01 -6.13570154e-01
5.74336795e-04 6.31835639e-01 -6.17816627e-01 -8.85318220e-02
-8.27941895e-01 1.01699293e+00 6.57111526e-01 7.26770982e-02
-7.50630260e-01 -4.29395676e-01 -1.51903832e+00 -4.88276899e-01
5.52124321e-01 -1.00268126e+00 8.80184650e-01 -5.05838633e-01
-1.69582772e+00 9.95876312e-01 -1.13056816e-01 -6.72012627e-01
6.10027492e-01 -4.70243722e-01 7.86303356e-02 2.39738580e-02
1.84997916e-01 9.42821324e-01 7.62377024e-01 -1.49336922e+00
-4.71747518e-01 -5.03069997e-01 -8.31502378e-02 6.22449875e-01
1.19979724e-01 -7.50899494e-01 -6.12152636e-01 -1.94171980e-01
1.01111317e+00 -1.49820054e+00 -2.27983356e-01 1.98738948e-01
-4.12512362e-01 3.91074687e-01 8.43805909e-01 -3.97852182e-01
8.97815406e-01 -1.81639779e+00 3.35526884e-01 2.43768707e-01
3.67706746e-01 4.39874046e-02 3.52154016e-01 3.66127849e-01
2.28279680e-01 -6.62168324e-01 -1.20520078e-01 -7.43191838e-01
-3.25570516e-02 5.08214653e-01 8.51950201e-04 1.09889805e+00
-8.40542614e-02 8.52422774e-01 -9.69562769e-01 -3.27828437e-01
6.49056077e-01 3.87169629e-01 -7.33374894e-01 1.13337263e-01
-2.63121128e-01 5.87504506e-01 7.72610903e-02 8.22613716e-01
6.09464824e-01 -6.38428107e-02 1.85313731e-01 -2.59182960e-01
-3.18920463e-01 3.50310475e-01 -1.25263250e+00 2.37418437e+00
-2.93626338e-01 7.41415441e-01 -3.01148463e-02 -4.85462785e-01
1.19582677e+00 -2.12092385e-01 4.13046122e-01 -6.50299370e-01
3.05387259e-01 4.19075370e-01 -2.97353685e-01 -7.97092170e-02
9.37495649e-01 4.88067195e-02 -1.31798789e-01 -1.40207440e-01
3.81976843e-01 -7.37169921e-01 -3.34624201e-02 2.18020558e-01
7.39419520e-01 9.09119487e-01 4.78616089e-01 -1.55433819e-01
6.17075622e-01 3.39724541e-01 7.70167232e-01 3.63926679e-01
-3.61644387e-01 7.87056804e-01 -5.29623125e-04 -5.69539487e-01
-1.08763111e+00 -1.08118308e+00 6.25058711e-02 4.27365929e-01
6.22727454e-01 -4.59361076e-01 -3.79256159e-01 -4.47299570e-01
4.02290136e-01 9.92738828e-02 -2.67201871e-01 -1.67044148e-01
-7.11872578e-01 -2.65063584e-01 4.82794374e-01 3.71200532e-01
6.02049351e-01 -3.79476756e-01 -6.41517639e-01 1.14834987e-01
-1.15403540e-01 -1.47109783e+00 -4.68970180e-01 4.26169187e-01
-1.00720620e+00 -1.00311708e+00 -3.35459143e-01 -7.08431840e-01
7.00550735e-01 3.91021669e-01 8.88423920e-01 -3.69831234e-01
-1.31727740e-01 2.14979038e-01 1.57295037e-02 -2.75023550e-01
-1.60714611e-01 3.17915261e-01 4.61269230e-01 -1.77564874e-01
1.12649426e-01 -6.01942360e-01 -4.77320552e-01 4.07823265e-01
-4.37223643e-01 1.01689234e-01 3.43312860e-01 8.28691125e-01
8.11889172e-01 -7.73791194e-01 -1.37187481e-01 -6.18551791e-01
-8.37297142e-02 -5.83989695e-02 -1.13070846e+00 -4.50734943e-01
-9.88485098e-01 2.95755237e-01 4.49845731e-01 -3.93599570e-02
-8.42476845e-01 6.05714202e-01 1.72625884e-01 -8.00166667e-01
1.57727525e-01 5.63398063e-01 1.04662019e-03 -7.08488405e-01
9.63115454e-01 6.75839633e-02 4.15506482e-01 -2.51067638e-01
4.49201226e-01 3.07170242e-01 7.14731336e-01 -2.08848730e-01
1.22059643e+00 8.81811500e-01 4.11446363e-01 -8.25084090e-01
-7.20549762e-01 -7.83977032e-01 -9.95640635e-01 -2.55293489e-01
7.95245051e-01 -1.38091528e+00 -5.76316118e-01 3.90695155e-01
-9.53337014e-01 -2.81780481e-01 -6.97219595e-02 8.18341911e-01
-8.29342961e-01 5.31737387e-01 -4.28064913e-01 -6.64379179e-01
-2.22932369e-01 -1.10440862e+00 1.15286243e+00 8.66376013e-02
-2.33687699e-01 -8.98726344e-01 5.20714998e-01 2.63991654e-01
-1.46086499e-01 6.66529119e-01 1.13992862e-01 9.32552945e-03
-1.12622058e+00 -2.60490417e-01 -1.28305584e-01 3.24886829e-01
-6.23680614e-02 -3.13083202e-01 -8.68351042e-01 -7.32334018e-01
-2.91621923e-01 -3.80330414e-01 7.70793438e-01 2.37210199e-01
5.85108027e-02 1.02657601e-01 -3.42157543e-01 1.41082871e+00
1.74124157e+00 -1.22502096e-01 3.90300483e-01 7.62912750e-01
1.33590496e+00 4.22245950e-01 7.25819290e-01 2.46368483e-01
6.31635904e-01 9.20438826e-01 8.27763021e-01 -2.18136311e-02
-2.18145072e-01 -6.56519234e-01 5.28684556e-01 1.01745903e+00
1.22943155e-01 1.50329128e-01 -9.52400267e-01 6.42946959e-01
-2.11023545e+00 -3.93418074e-01 -2.00165659e-01 2.36960554e+00
3.60202581e-01 2.41917297e-01 -9.81016830e-02 -1.44802094e-01
3.27770799e-01 5.52357376e-01 -5.77490747e-01 -2.76129633e-01
-2.98175097e-01 -2.78368354e-01 1.27439344e+00 8.07332098e-01
-1.12171257e+00 1.21277642e+00 5.11389542e+00 2.07798123e-01
-1.31511414e+00 3.90860587e-02 -3.71558636e-01 -5.93549237e-02
9.09711793e-02 3.39971155e-01 -1.07772803e+00 -9.53517333e-02
4.23612386e-01 2.31065467e-01 3.30048651e-01 9.46511507e-01
1.06839888e-01 -3.35052848e-01 -1.04008818e+00 1.08451021e+00
4.58055049e-01 -1.25598609e+00 -2.17392609e-01 2.42269307e-01
1.06685662e+00 6.87591970e-01 -3.34669709e-01 7.32232165e-03
4.33363795e-01 -6.12867475e-01 8.99979234e-01 2.71095783e-01
8.57796490e-01 -6.70365334e-01 5.83314240e-01 4.88190562e-01
-1.35390234e+00 1.02535643e-01 -2.45221421e-01 -3.53741109e-01
2.81584173e-01 5.04694521e-01 -1.17044449e+00 1.09538388e+00
4.18291777e-01 9.61741805e-01 -4.66226369e-01 1.11991584e+00
-5.27277768e-01 -1.49724156e-01 -5.23642838e-01 2.72583425e-01
3.80379200e-01 -2.68580914e-01 9.85081017e-01 7.78138518e-01
6.79891557e-02 -5.99447131e-01 4.34065819e-01 6.64068401e-01
-7.23657012e-02 -1.29245995e-02 -9.84594166e-01 3.50186825e-01
2.54362553e-01 9.95450735e-01 -3.20524991e-01 -1.20264322e-01
-2.97825187e-01 9.83462512e-01 6.70664310e-01 2.40240619e-01
-5.63571393e-01 -2.39630029e-01 7.74988711e-01 3.46144065e-02
2.11563796e-01 -7.99432278e-01 -3.59133720e-01 -1.45044172e+00
1.93453342e-01 -5.27675509e-01 1.28234133e-01 -8.41618419e-01
-6.13818944e-01 4.11939532e-01 -1.57581910e-01 -1.50217831e+00
-4.10301983e-01 -6.28436744e-01 1.37829900e-01 6.35755837e-01
-1.71761727e+00 -1.35278988e+00 -7.65990198e-01 3.63709569e-01
3.82805973e-01 3.35411169e-02 4.72886324e-01 2.49727353e-01
-2.48901755e-01 3.05229604e-01 4.34284136e-02 2.03850307e-02
1.01171744e+00 -1.34736180e+00 4.20876384e-01 1.06990159e+00
3.96715850e-01 5.66239417e-01 8.01703811e-01 -9.12396073e-01
-1.63862145e+00 -1.17332399e+00 1.02984381e+00 -6.05543435e-01
4.69143093e-01 -5.18256605e-01 -5.24782419e-01 9.90851581e-01
-2.48410970e-01 2.18162939e-01 -1.35773540e-01 -1.32880405e-01
-3.79352957e-01 -2.03822985e-01 -9.25872803e-01 5.75082898e-01
1.33143616e+00 -7.02293992e-01 -4.55720395e-01 1.91122279e-01
6.90321088e-01 -1.20718396e+00 -7.52890944e-01 5.68071067e-01
7.67679691e-01 -1.05833077e+00 1.01688111e+00 9.58944038e-02
-6.73234314e-02 -7.93454528e-01 -1.48585036e-01 -1.28040469e+00
-1.59859538e-01 -8.33248436e-01 -2.48541623e-01 7.42342234e-01
1.74521636e-02 -8.09587479e-01 1.20178401e+00 -1.11007923e-02
-3.49344313e-01 -4.76275235e-01 -9.21583056e-01 -1.05782187e+00
-4.21667308e-01 -2.73323685e-01 1.16908319e-01 8.62176299e-01
-2.45513484e-01 4.55062419e-01 -5.99144757e-01 4.06880051e-01
9.27199304e-01 -1.08614437e-01 1.47878873e+00 -1.23889673e+00
-1.19960785e-01 -1.12816766e-02 -8.93869102e-01 -1.46271300e+00
1.03333913e-01 -9.30601060e-01 4.28049624e-01 -1.41098452e+00
-3.20468336e-01 -2.81630427e-01 2.13650510e-01 1.78687006e-01
1.89094007e-01 4.98121500e-01 2.65588403e-01 4.86306846e-01
-5.18340170e-01 6.96060359e-01 9.98817325e-01 4.56239618e-02
-4.65783238e-01 -5.36314309e-01 1.83110327e-01 9.08560455e-01
4.18215394e-01 -4.95716363e-01 -3.79159391e-01 -5.25477529e-01
3.76680374e-01 -1.15585953e-01 4.11089092e-01 -1.34660399e+00
5.21314859e-01 1.03676565e-01 1.90695688e-01 -8.12186122e-01
5.40528655e-01 -7.95945764e-01 2.16850787e-01 6.16630793e-01
2.71317899e-01 9.03410986e-02 -3.17988396e-02 6.62353098e-01
-2.66585737e-01 1.50298283e-01 7.66891599e-01 -6.91809952e-02
-1.06479454e+00 4.33410197e-01 6.65775016e-02 1.64727587e-02
7.08549201e-01 -6.44762635e-01 -2.12676108e-01 -2.56950021e-01
-4.47655171e-01 3.18626106e-01 1.16875136e+00 4.63871956e-01
5.50588489e-01 -1.41462636e+00 -2.25899279e-01 4.70036417e-01
6.62271321e-01 4.93303627e-01 -9.64929983e-02 1.15175164e+00
-1.12100768e+00 6.27870142e-01 -1.50324568e-01 -1.13088799e+00
-1.08382404e+00 2.20801920e-01 4.34049398e-01 -2.87862301e-01
-7.56353855e-01 7.49598622e-01 4.77273054e-02 -9.99761403e-01
1.95825428e-01 -3.88935626e-01 1.45947188e-01 -1.02121942e-01
-1.95478246e-01 3.79271418e-01 2.08245710e-01 -1.10494256e+00
-5.69396377e-01 9.25429821e-01 2.72993952e-01 -1.10721655e-01
1.04119122e+00 -5.16689956e-01 8.22843984e-02 5.07224560e-01
1.36526513e+00 1.29109651e-01 -1.60607493e+00 -3.16874534e-01
5.90226166e-02 -5.45269310e-01 1.16046131e-01 -3.73940378e-01
-6.44978285e-01 5.79478323e-01 5.36661685e-01 -3.96582484e-01
5.98937869e-01 -3.14606905e-01 8.23964655e-01 4.95993018e-01
6.57528400e-01 -1.17604303e+00 2.09803022e-02 8.96294653e-01
6.42489314e-01 -1.35545433e+00 4.15281713e-01 -5.39448738e-01
-2.51725912e-01 1.01743472e+00 5.84868908e-01 -4.90957797e-01
1.40064463e-01 -3.60153243e-02 3.36382121e-01 -3.39646578e-01
-2.92399913e-01 -4.69682693e-01 4.19902503e-01 6.11100495e-01
3.22067812e-02 -1.52567536e-01 -6.15571514e-02 -4.89025414e-01
-3.75965267e-01 -7.62493489e-03 4.00252759e-01 1.17873001e+00
-5.10196924e-01 -9.47694838e-01 -4.56814915e-01 -1.65106937e-01
9.69778821e-02 1.64937764e-01 -3.76182735e-01 1.08261228e+00
2.05410227e-01 5.14465511e-01 -1.02223933e-01 -5.10428965e-01
4.72493768e-01 -8.68212283e-02 6.77741826e-01 -5.52912056e-01
-4.12945718e-01 5.02015129e-02 4.05476242e-01 -9.50128198e-01
-3.72206599e-01 -5.36014736e-01 -1.21439540e+00 -7.08884746e-02
-4.66581136e-01 -2.67441243e-01 9.39053476e-01 1.02171016e+00
5.01015663e-01 2.61092335e-01 3.36238414e-01 -1.11616337e+00
-3.10046315e-01 -5.94860256e-01 -3.71420264e-01 3.34803015e-01
7.30627239e-01 -1.04998362e+00 -2.89415240e-01 -8.65983963e-02] | [7.9466118812561035, -2.14603853225708] |
42a9b6ba-5f03-4a84-8191-e5dc34cc8521 | cross-modal-consensus-network-forweakly | null | null | https://arxiv.org/abs/2107.12589#:~:text=Cross%2Dmodal%20Consensus%20Network%20for%20Weakly%20Supervised%20Temporal%20Action%20Localization,-Fa%2DTing%20Hong&text=Weakly%20supervised%20temporal%20action%20localization%20(WS%2DTAL)%20is%20a,with%20video%2Dlevel%20categorical%20supervision. | https://arxiv.org/pdf/2107.12589.pdf | Cross-modal Consensus Network forWeakly Supervised Temporal Action Localization | Weakly supervised temporal action localization (WS-TAL) is a challenging task that aims to localize action instances in the given video with video-level categorical supervision. Both appearance and motion features are used in previous works, while they do not utilize them in a proper way but apply simple concatenation or score-level fusion. In this work, we argue that the features extracted from the pretrained extractor, e.g., I3D, are not the WS-TALtask-specific features, thus the feature re-calibration is needed for reducing the task-irrelevant information redundancy. Therefore, we propose a cross-modal consensus network (CO2-Net) to tackle this problem. In CO2-Net, we mainly introduce two identical proposed cross-modal consensus modules (CCM) that design a cross-modal attention mechanism to filter out the task-irrelevant information redundancy using the global information from the main modality and the cross-modal local information of the auxiliary modality. Moreover, we treat the attention weights derived from each CCMas the pseudo targets of the attention weights derived from another CCM to maintain the consistency between the predictions derived from two CCMs, forming a mutual learning manner. Finally, we conduct extensive experiments on two common used temporal action localization datasets, THUMOS14 and ActivityNet1.2, to verify our method and achieve the state-of-the-art results. The experimental results show that our proposed cross-modal consensus module can produce more representative features for temporal action localization. | ['Wei-Shi Zheng', 'Ying Shan', 'Dan Xu', 'Jia-Chang Feng', 'Fa-Ting Hong'] | 2021-07-27 | null | null | null | proceedings-of-the-29th-acm-international | ['weakly-supervised-temporal-action'] | ['computer-vision'] | [ 9.61031392e-02 -3.06748569e-01 -3.57109666e-01 -1.49206698e-01
-7.56777346e-01 -1.00992836e-01 5.73176324e-01 -3.24043512e-01
-4.91304606e-01 5.16006470e-01 5.78908980e-01 4.63277757e-01
-1.36786178e-01 -1.85215741e-01 -5.55735707e-01 -9.83547091e-01
7.71381184e-02 -1.12844683e-01 7.11732090e-01 -1.29991651e-01
1.17962882e-01 -7.42327198e-02 -1.59751582e+00 7.80344605e-01
8.58088195e-01 1.08920050e+00 3.62905174e-01 7.44589642e-02
5.98426722e-02 1.18535578e+00 -3.46223652e-01 2.60824692e-02
2.15994745e-01 -8.14301252e-01 -6.91823006e-01 1.82000369e-01
2.13471070e-01 -2.95401454e-01 -4.86271143e-01 9.71417964e-01
4.03511375e-01 2.27075949e-01 4.17844504e-01 -1.49847245e+00
-4.06200737e-01 5.03705680e-01 -7.14122713e-01 3.75162840e-01
3.17900866e-01 4.28860962e-01 9.15820897e-01 -8.07230890e-01
5.87084353e-01 1.32789338e+00 4.13646847e-01 4.16230589e-01
-9.45717216e-01 -6.92222536e-01 6.95960164e-01 7.50319600e-01
-1.38884866e+00 -4.88727808e-01 1.02093565e+00 -3.66783231e-01
7.18682468e-01 1.32295072e-01 6.01333141e-01 1.36016703e+00
3.03829432e-01 1.11022258e+00 1.13712859e+00 -3.98390628e-02
-2.02314835e-02 -1.05653755e-01 -1.08279653e-01 6.58194840e-01
-2.59118468e-01 -1.71538517e-01 -7.60432959e-01 4.92356047e-02
7.38417327e-01 2.79325306e-01 -5.24327397e-01 -3.78230274e-01
-1.64289486e+00 4.03780013e-01 4.80373263e-01 7.19088376e-01
-4.23091322e-01 1.91890091e-01 6.32989168e-01 2.59448975e-01
2.50602424e-01 -8.46198276e-02 -6.02694154e-01 -2.30985358e-01
-5.04134059e-01 -1.30699366e-01 1.48819864e-01 7.90655375e-01
7.92341352e-01 -1.96315393e-01 -6.13061309e-01 6.59360945e-01
3.83521020e-01 2.89332658e-01 1.02662301e+00 -9.07121241e-01
6.77956820e-01 8.79021049e-01 6.50536269e-02 -1.20212603e+00
-3.66021991e-01 -2.49120399e-01 -8.58019173e-01 4.18709554e-02
1.02447815e-01 -8.04594681e-02 -7.87337422e-01 2.16346192e+00
3.15281004e-01 5.65779448e-01 7.23828822e-02 1.15361595e+00
7.53783047e-01 4.30843443e-01 2.87271410e-01 -4.22437549e-01
1.10855758e+00 -1.34184551e+00 -8.13766003e-01 -2.23116308e-01
6.47523105e-01 -6.06002212e-01 8.38703454e-01 1.14436574e-01
-7.88185954e-01 -9.19153452e-01 -1.00069118e+00 6.54762760e-02
-1.32928938e-01 4.24654007e-01 4.99883801e-01 -3.06550741e-01
-5.56109667e-01 6.67564273e-01 -9.35565472e-01 -3.58701289e-01
3.39561224e-01 4.45447490e-02 -7.87178576e-01 1.17818482e-01
-1.28601837e+00 8.19574118e-01 6.66218162e-01 3.84682536e-01
-9.25531983e-01 -2.79194832e-01 -7.18351960e-01 -7.97926858e-02
5.89177847e-01 -5.12923896e-01 9.99547064e-01 -1.47295308e+00
-1.32907605e+00 4.92703736e-01 -5.38084283e-02 -3.01046409e-02
5.50301075e-01 -2.58538306e-01 -4.86997694e-01 1.37827203e-01
3.42929274e-01 7.19446778e-01 7.69519866e-01 -1.09099293e+00
-8.86845946e-01 -2.65399128e-01 1.66767493e-01 5.11456788e-01
-3.55485827e-01 -1.29310280e-01 -7.77283549e-01 -7.96980321e-01
2.92250663e-01 -8.69015932e-01 -5.84491640e-02 -1.36223566e-02
-2.19574258e-01 -4.76503074e-01 9.25310731e-01 -6.55356407e-01
1.31003273e+00 -2.28102708e+00 6.14684582e-01 -8.57779682e-02
2.25453153e-02 2.18199059e-01 -3.90792042e-01 3.78117114e-01
-2.96440452e-01 -1.44121185e-01 -1.65553287e-01 -3.39262247e-01
-2.05394715e-01 2.15118051e-01 1.28201544e-01 5.32791376e-01
3.82910281e-01 8.11253309e-01 -1.10282803e+00 -8.43949616e-01
4.09004033e-01 2.57133901e-01 -2.02606529e-01 2.45857850e-01
-7.51257762e-02 8.44851255e-01 -7.23963499e-01 6.62847877e-01
3.65504950e-01 -1.68784142e-01 1.31112933e-01 -7.82483220e-01
-1.80059299e-02 2.59477831e-02 -1.24717140e+00 2.29639554e+00
-2.40209118e-01 1.83827639e-01 -1.79973096e-01 -1.16071665e+00
5.30557394e-01 3.85077238e-01 1.01185286e+00 -8.02299380e-01
2.69019663e-01 5.52448519e-02 1.24691300e-01 -9.11713481e-01
-1.31640911e-01 7.93954730e-02 -3.46461795e-02 2.85132110e-01
4.38149959e-01 6.57165706e-01 2.39993364e-01 4.95765842e-02
1.21215916e+00 5.81102073e-01 2.77499586e-01 -1.26661792e-01
9.75997508e-01 -1.61123037e-01 1.15990758e+00 4.46434677e-01
-5.43935895e-01 6.26460135e-01 3.69281530e-01 -3.40597332e-01
-6.39432847e-01 -7.12119997e-01 1.73354149e-01 1.01748514e+00
5.05073726e-01 -5.44120669e-01 -4.33987498e-01 -1.18131804e+00
-2.66662687e-01 1.69094622e-01 -8.16803157e-01 -4.54046935e-01
-4.94414717e-01 -4.72803593e-01 3.00312132e-01 7.71271348e-01
8.44933391e-01 -1.35847807e+00 -4.75084722e-01 1.26722053e-01
-6.02998435e-01 -1.08669031e+00 -7.11198032e-01 3.16500328e-02
-5.89268267e-01 -1.24770892e+00 -7.21932471e-01 -6.41853809e-01
5.59612572e-01 4.89179105e-01 6.78247750e-01 3.00973710e-02
1.50993377e-01 3.32000226e-01 -6.52143717e-01 1.03218168e-01
5.40959686e-02 -1.28690913e-01 1.33941099e-01 6.43730164e-01
4.20532674e-01 -8.35376680e-01 -5.96346021e-01 5.82103312e-01
-8.48775268e-01 3.80075604e-01 8.68559599e-01 8.57521594e-01
5.56787729e-01 -8.33784118e-02 4.67973918e-01 -2.72587270e-01
3.66230875e-01 -5.01966298e-01 -9.97036994e-02 3.96549612e-01
-1.28141224e-01 1.31045789e-01 5.49977720e-01 -7.71648765e-01
-1.09761083e+00 3.15610886e-01 1.20002113e-01 -9.64855075e-01
-1.98202297e-01 6.61129236e-01 -5.49448252e-01 9.50510204e-02
2.81996489e-01 4.64042813e-01 -1.43904656e-01 -5.00327766e-01
2.05780640e-01 4.13635314e-01 4.45711643e-01 -3.98282886e-01
7.39022136e-01 4.87138957e-01 -9.89671350e-02 -2.32580751e-01
-9.84911859e-01 -5.75375915e-01 -8.40687513e-01 -4.22234654e-01
1.19365156e+00 -1.01474249e+00 -6.17162108e-01 6.12056077e-01
-1.25236177e+00 -4.08919267e-02 -1.88852683e-01 7.66305745e-01
-5.75459659e-01 5.43864071e-01 -3.54532391e-01 -6.18421316e-01
-7.39173368e-02 -1.24442422e+00 1.21959412e+00 2.63134360e-01
9.72525179e-02 -6.69790924e-01 2.83277541e-01 3.53823453e-01
-2.17846893e-02 3.14819455e-01 4.54947263e-01 -6.20405674e-01
-5.16291797e-01 4.26720753e-02 -2.40046069e-01 4.71854925e-01
4.03330922e-01 -2.14962214e-01 -8.70195508e-01 -1.60975650e-01
-5.23272641e-02 -4.24121678e-01 1.10486686e+00 1.01480119e-01
1.12173641e+00 -1.17188931e-01 -2.60477275e-01 4.21495408e-01
1.15018129e+00 1.24405034e-01 6.38756990e-01 1.84102461e-01
9.89324987e-01 4.45108473e-01 9.40832794e-01 3.46405596e-01
3.50497663e-01 8.63486767e-01 4.87506628e-01 -6.30348027e-02
-9.82040539e-02 -3.33131403e-01 7.96885550e-01 9.43996608e-01
-4.56050128e-01 6.54909983e-02 -4.41065550e-01 4.51409191e-01
-2.61914206e+00 -1.18620670e+00 -8.34488645e-02 2.11299586e+00
6.94661856e-01 5.71628213e-02 8.05877224e-02 -3.48722190e-02
7.44361520e-01 3.89373690e-01 -5.44461966e-01 3.50427449e-01
-1.92962602e-01 -3.02000821e-01 2.01081276e-01 8.66918359e-03
-1.44137132e+00 7.07868934e-01 4.80982971e+00 1.08363497e+00
-1.00289965e+00 3.47687513e-01 3.34826678e-01 -5.62178753e-02
6.65291473e-02 1.16913661e-01 -3.58621478e-01 7.74835050e-01
4.96968865e-01 5.09713627e-02 1.88518375e-01 7.11366236e-01
4.22427237e-01 -6.21978939e-02 -1.10257471e+00 1.10226524e+00
2.51714706e-01 -8.73885214e-01 4.72941808e-02 -2.02359110e-01
6.04674459e-01 -4.28064093e-02 -2.14992747e-01 6.01163089e-01
-5.27493432e-02 -5.78520894e-01 7.74766862e-01 8.84970784e-01
3.94238263e-01 -4.86758053e-01 9.21299815e-01 4.61301625e-01
-1.65010571e+00 -3.07263702e-01 -2.34111428e-01 1.29092738e-01
2.41823763e-01 2.16656163e-01 7.57898986e-02 1.02167678e+00
8.87437403e-01 1.38709259e+00 -6.49146140e-01 8.92251432e-01
-3.29825282e-01 3.43820661e-01 -1.18839718e-01 3.91039431e-01
4.11080241e-01 -1.29594743e-01 5.44819117e-01 9.07116413e-01
1.77711919e-01 1.08917855e-01 5.37398517e-01 5.78820407e-01
2.09536821e-01 2.16677085e-01 -3.57588112e-01 -4.90341103e-03
2.69995816e-02 1.28023398e+00 -3.95949781e-01 -3.28351766e-01
-6.87192380e-01 1.24552035e+00 2.70265341e-01 3.18648010e-01
-1.20253932e+00 -8.54519084e-02 6.24549091e-01 -2.98134267e-01
2.96281785e-01 -7.10777789e-02 3.14617902e-01 -1.63705432e+00
3.13070148e-01 -9.12111998e-01 5.32491744e-01 -9.11752164e-01
-1.44373202e+00 4.35796380e-01 1.40679955e-01 -1.87855721e+00
5.18332981e-02 -2.90973574e-01 -7.34652579e-01 6.82173491e-01
-1.32403278e+00 -1.55393553e+00 -5.68753183e-01 9.00727034e-01
7.50521541e-01 -9.41550508e-02 5.60479701e-01 4.85195339e-01
-8.39811921e-01 4.28030759e-01 -2.48784065e-01 2.29516700e-01
8.90452385e-01 -8.47208083e-01 -3.27751994e-01 8.51138055e-01
1.50248893e-02 4.89510328e-01 3.11337441e-01 -7.19482899e-01
-1.24635959e+00 -1.09321618e+00 4.72767919e-01 -2.62844652e-01
7.38245308e-01 -9.20887068e-02 -9.34938133e-01 6.53408051e-01
2.34896988e-01 2.57285774e-01 3.89476120e-01 -1.07557565e-01
-1.94945335e-01 -2.98733681e-01 -6.97267830e-01 5.00855684e-01
1.39693451e+00 -4.76893991e-01 -6.83417737e-01 2.76819319e-01
7.45602727e-01 -1.26728848e-01 -9.03891742e-01 7.11285830e-01
6.51278079e-01 -1.09172833e+00 7.46275127e-01 -7.24486530e-01
5.49668491e-01 -7.24442422e-01 -1.16505392e-01 -1.07828212e+00
-5.86525738e-01 -2.14284331e-01 -1.80962145e-01 1.37526441e+00
6.16909824e-02 -3.56018215e-01 4.22590554e-01 1.64405540e-01
-3.30926359e-01 -7.78810740e-01 -1.18560886e+00 -7.62673199e-01
-3.47749710e-01 -2.40636453e-01 3.81795377e-01 1.05548334e+00
1.53040633e-01 4.69296426e-01 -7.57715404e-01 5.59616787e-03
2.92519271e-01 7.17997104e-02 6.54730618e-01 -1.02020037e+00
-3.35503489e-01 -4.46022481e-01 -6.90012157e-01 -1.05541873e+00
3.70663166e-01 -6.99219882e-01 1.59984395e-01 -1.28703582e+00
5.68778872e-01 -1.23583384e-01 -9.27819073e-01 8.54250669e-01
-4.00239289e-01 7.94960931e-02 2.82146335e-01 4.14087951e-01
-1.09846997e+00 1.11054075e+00 1.37321866e+00 -3.20522904e-01
-1.34087056e-01 -2.83656836e-01 -4.81092244e-01 8.70155156e-01
5.77996075e-01 -4.24499720e-01 -5.94813943e-01 -3.85104746e-01
-3.02355647e-01 3.85231264e-02 4.95654672e-01 -1.30416644e+00
2.81741679e-01 -5.84628642e-01 3.82879049e-01 -6.56436980e-01
3.16744566e-01 -1.14230597e+00 7.33324513e-02 2.94710994e-01
-2.48212144e-01 -9.76326540e-02 -4.16192748e-02 7.90400386e-01
-4.37796235e-01 7.69078583e-02 5.37714601e-01 -2.13407397e-01
-1.04545629e+00 4.97304261e-01 -1.78227842e-01 -1.68103307e-01
1.16728497e+00 -8.14516917e-02 -3.62161458e-01 -2.73096204e-01
-7.63771534e-01 4.47002798e-01 3.49916518e-01 7.78633356e-01
4.72116143e-01 -1.79433537e+00 -6.78963482e-01 -2.55624913e-02
4.95008707e-01 -3.24553877e-01 6.69762433e-01 1.45504034e+00
1.38790324e-01 2.33887836e-01 -4.48592782e-01 -7.51477301e-01
-1.12418735e+00 7.16875315e-01 4.59845394e-01 -3.50282937e-01
-5.52166164e-01 5.63191652e-01 5.37499905e-01 -1.01911969e-01
9.46004689e-02 -3.37760717e-01 -5.34785151e-01 3.96545567e-02
5.06358683e-01 1.26265898e-01 -2.22223774e-01 -9.84501123e-01
-6.67938948e-01 6.77167416e-01 1.76423401e-01 -2.55168416e-02
1.24960077e+00 -2.08666682e-01 -7.47449398e-02 6.74830258e-01
1.20407999e+00 -3.60936791e-01 -1.36911976e+00 -4.26562279e-01
-2.33253121e-01 -4.53223318e-01 -8.97253156e-02 -5.09379983e-01
-1.41324437e+00 7.03950822e-01 8.39090765e-01 -2.18476027e-01
1.42258334e+00 1.21616749e-02 5.77994943e-01 2.42361665e-01
3.41689914e-01 -1.20703840e+00 3.77959281e-01 2.22933352e-01
1.08235347e+00 -1.33323240e+00 -6.99234530e-02 -1.61618516e-01
-8.38546336e-01 9.01897728e-01 1.16663122e+00 -6.35812655e-02
6.03370428e-01 -2.97667056e-01 -1.43343762e-01 -1.13447294e-01
-8.46243739e-01 -5.23591816e-01 5.01485884e-01 4.44555074e-01
2.05621555e-01 -2.61708379e-01 -6.80193484e-01 9.01838303e-01
6.99397504e-01 1.19551376e-01 2.18855757e-02 1.12172937e+00
-3.08686227e-01 -1.05093300e+00 -7.68877417e-02 2.35716060e-01
-4.50188257e-02 1.74059287e-01 -3.60792607e-01 6.73913658e-01
7.00497329e-01 9.27871227e-01 -1.59818068e-01 -9.43311870e-01
3.28961521e-01 2.51579867e-03 3.74874353e-01 -3.03922683e-01
-4.50474709e-01 2.22896397e-01 -5.23949489e-02 -1.03385794e+00
-1.07036746e+00 -6.83645070e-01 -1.24092793e+00 1.43222585e-01
-2.39736900e-01 -4.65938151e-02 4.07213941e-02 1.27431154e+00
4.83541012e-01 7.04709113e-01 6.73489988e-01 -9.30644274e-01
-3.39354992e-01 -1.23551250e+00 -3.89460593e-01 8.22900951e-01
6.68644682e-02 -1.07348263e+00 -3.38746428e-01 1.99087992e-01] | [8.543191909790039, 0.7138928771018982] |
0f8df7b8-4c79-482d-aeff-3b9f401efb87 | you-only-need-two-detectors-to-achieve-multi | 2304.08709 | null | https://arxiv.org/abs/2304.08709v1 | https://arxiv.org/pdf/2304.08709v1.pdf | You Only Need Two Detectors to Achieve Multi-Modal 3D Multi-Object Tracking | Firstly, a new multi-object tracking framework is proposed in this paper based on multi-modal fusion. By integrating object detection and multi-object tracking into the same model, this framework avoids the complex data association process in the classical TBD paradigm, and requires no additional training. Secondly, confidence of historical trajectory regression is explored, possible states of a trajectory in the current frame (weak object or strong object) are analyzed and a confidence fusion module is designed to guide non-maximum suppression of trajectory and detection for ordered association. Finally, extensive experiments are conducted on the KITTI and Waymo datasets. The results show that the proposed method can achieve robust tracking by using only two modal detectors and it is more accurate than many of the latest TBD paradigm-based multi-modal tracking methods. The source codes of the proposed method are available at https://github.com/wangxiyang2022/YONTD-MOT | ['Mingguang Huang', 'Ting Meng', 'Chunyun Fu', 'JiaWei He', 'Xiyang Wang'] | 2023-04-18 | null | null | null | null | ['3d-multi-object-tracking'] | ['computer-vision'] | [-3.16688716e-01 -5.67670286e-01 -4.52076405e-01 1.14353962e-01
-6.93143964e-01 -3.81569237e-01 6.04364336e-01 -3.70423906e-02
-2.89469630e-01 6.29148602e-01 -2.23235324e-01 -1.54769376e-01
-3.54930609e-01 -6.32455409e-01 -4.72216725e-01 -9.06857133e-01
2.05799341e-02 4.88039166e-01 1.03939462e+00 1.49939284e-01
-3.22023011e-03 4.60085422e-01 -1.65144217e+00 -1.34164304e-01
6.61073327e-01 1.11529207e+00 2.80065536e-01 6.23504460e-01
2.35024020e-01 5.69133699e-01 -3.84025097e-01 2.48913616e-02
3.06540340e-01 -6.13458939e-02 -1.56970680e-01 2.81798691e-02
4.33975309e-01 -3.42169791e-01 -4.24695462e-01 1.10156333e+00
5.87549090e-01 2.90972412e-01 5.34219682e-01 -1.42875302e+00
-1.95862278e-01 2.57089168e-01 -8.96018445e-01 4.90579307e-01
2.33395547e-01 1.93072543e-01 4.53791201e-01 -8.73872936e-01
2.83735543e-01 1.37957609e+00 8.43653142e-01 4.08418447e-01
-9.16727483e-01 -1.05963218e+00 2.52951324e-01 2.59017408e-01
-1.55588925e+00 -2.32228413e-01 5.60970485e-01 -6.10642791e-01
3.97212029e-01 3.42959762e-01 7.50010192e-01 6.24336421e-01
2.43141159e-01 8.98960769e-01 9.52229857e-01 -1.46737099e-01
-2.36469835e-01 4.77348082e-02 2.95055449e-01 7.89491117e-01
5.91176271e-01 5.80893815e-01 -2.64603943e-01 -2.25399822e-01
3.99949729e-01 1.98780537e-01 5.63614853e-02 -2.79698998e-01
-1.18642414e+00 6.94821715e-01 2.49298975e-01 2.63983279e-01
-2.97052890e-01 8.60169381e-02 2.71426350e-01 -1.04562357e-01
3.16195786e-01 -6.47364914e-01 -7.95991495e-02 1.63924217e-01
-8.95419419e-01 5.57382762e-01 1.88097119e-01 1.17605698e+00
5.43652773e-01 1.62372395e-01 -3.89332205e-01 2.83315927e-01
9.11980569e-01 1.01036549e+00 7.41029605e-02 -7.47333407e-01
2.60027647e-01 6.80103183e-01 2.57969916e-01 -8.03253114e-01
-5.33071399e-01 -2.11650088e-01 -5.91170132e-01 5.49721897e-01
4.09800410e-01 -2.63014823e-01 -9.13924217e-01 1.52808678e+00
9.61180866e-01 5.79725981e-01 -1.50933161e-01 8.45380545e-01
7.18155980e-01 6.63227379e-01 2.28332713e-01 -4.01857227e-01
1.48206592e+00 -6.55784607e-01 -9.71552968e-01 8.93771946e-02
2.61319816e-01 -8.59455705e-01 2.15022042e-01 2.20026612e-01
-9.36198294e-01 -9.38632667e-01 -1.03588259e+00 4.53184664e-01
-3.09358627e-01 1.47017017e-01 5.25788784e-01 4.90945846e-01
-5.17114043e-01 2.56662458e-01 -9.47652221e-01 -4.77999866e-01
4.11001921e-01 3.79708111e-01 1.19582869e-01 7.10193114e-03
-1.17338705e+00 1.12231672e+00 8.10514033e-01 1.41180977e-01
-9.95047450e-01 -6.30856276e-01 -6.00534141e-01 -5.48577964e-01
5.01617014e-01 -5.98583102e-01 1.07422817e+00 -1.25268668e-01
-9.82151151e-01 4.20742482e-01 -1.87420920e-01 -2.85656065e-01
5.81858099e-01 -3.83550227e-01 -8.07627141e-01 2.53730826e-02
6.96655288e-02 4.01082128e-01 9.02154863e-01 -1.24159503e+00
-1.32951999e+00 -3.18215877e-01 -2.87264228e-01 1.90696672e-01
-1.62199941e-02 2.37815723e-01 -7.75717974e-01 -4.45544481e-01
3.97394598e-03 -1.13398862e+00 -1.37133792e-01 8.69313702e-02
-2.92399555e-01 -5.38976252e-01 1.51907802e+00 -4.26498264e-01
1.65789902e+00 -2.06670213e+00 3.34565826e-02 1.18336588e-01
4.84016500e-02 4.21356976e-01 3.20905298e-01 1.01212665e-01
3.13956797e-01 -3.77018571e-01 1.16921313e-01 -1.64674684e-01
-2.37760842e-02 -5.34160808e-02 -3.03326808e-02 7.69800425e-01
1.61657587e-03 5.72879970e-01 -7.51537919e-01 -9.58156466e-01
8.02471101e-01 4.74475056e-01 7.95865506e-02 1.65792264e-03
-1.51148945e-01 5.51658630e-01 -7.73340702e-01 1.04130220e+00
9.47385132e-01 -1.00861341e-01 -5.95766455e-02 -4.32153970e-01
-6.63381875e-01 -3.80781889e-01 -1.67350078e+00 1.30090046e+00
2.32827798e-01 2.56635070e-01 2.01116502e-02 -4.68219727e-01
6.94730043e-01 5.04291892e-01 8.27946067e-01 -4.27189529e-01
3.73243153e-01 -1.44190295e-02 -6.55887742e-03 -2.91940689e-01
5.48098981e-01 -4.50657643e-02 -5.53013757e-02 1.57403812e-01
-1.15904830e-01 1.64853632e-01 5.27870357e-01 1.92095622e-01
7.68472612e-01 3.68250221e-01 1.64494589e-01 -1.97947696e-01
8.25367033e-01 2.87664086e-01 9.81244445e-01 7.78169155e-01
-5.75457692e-01 7.45698884e-02 -4.16545421e-01 -1.09183222e-01
-7.02891886e-01 -1.25246215e+00 -3.24624360e-01 7.53979623e-01
7.37244070e-01 -3.01651090e-01 -1.18864380e-01 -5.09852290e-01
3.38993669e-01 4.31271672e-01 -4.26785201e-01 -5.53402416e-02
-6.02863014e-01 -8.36833775e-01 4.71868902e-01 4.91718560e-01
3.72448623e-01 -6.04374468e-01 -5.82335532e-01 2.95046002e-01
-1.19766630e-01 -9.97119606e-01 -4.21668053e-01 -1.77778229e-01
-1.05823112e+00 -1.15174162e+00 -5.23443520e-01 -6.36616230e-01
4.85231370e-01 5.09917259e-01 5.72661817e-01 2.78609127e-01
-2.36000255e-01 3.92896563e-01 -2.44429842e-01 -6.93364084e-01
-2.78311998e-01 -4.48242098e-01 4.75671947e-01 -1.42163262e-01
4.49684829e-01 -1.24175875e-02 -5.13646603e-01 6.72587872e-01
-5.26464701e-01 -2.03625020e-02 6.17792666e-01 4.90504503e-01
6.90729737e-01 3.64662021e-01 4.78782088e-01 -2.03578964e-01
1.04331627e-01 -5.31078041e-01 -1.12249398e+00 1.65914863e-01
-6.28280461e-01 -4.37476158e-01 1.19073659e-01 -8.07255149e-01
-1.26696694e+00 2.07950637e-01 9.17821079e-02 -8.91831756e-01
-1.95717499e-01 6.33717105e-02 -4.49951030e-02 -1.18368834e-01
2.23848730e-01 1.68730885e-01 -2.14392841e-02 -4.60913211e-01
1.95748061e-01 2.96118051e-01 4.55342501e-01 -4.43464994e-01
1.18019438e+00 8.21812272e-01 1.78549439e-01 -5.08197248e-01
-7.03221142e-01 -7.20380127e-01 -6.89328313e-01 -9.41652179e-01
9.28340971e-01 -1.14255452e+00 -1.03520238e+00 5.65087736e-01
-7.57554710e-01 1.47419274e-01 4.21114452e-02 1.00292706e+00
-3.11854780e-01 4.28017944e-01 -4.36684787e-01 -1.24071097e+00
-2.39899382e-01 -9.27337646e-01 8.68561924e-01 6.54359758e-01
8.06797147e-02 -8.56391966e-01 2.98975438e-01 3.29368472e-01
1.42572075e-01 2.29182571e-01 2.60487884e-01 -4.25337940e-01
-9.30013299e-01 -3.40546668e-01 -3.55021879e-02 -1.89025477e-01
1.12221487e-01 2.86822677e-01 -6.21430874e-01 -6.36041641e-01
-1.73178405e-01 -6.80748895e-02 7.93614626e-01 7.20988393e-01
5.57498872e-01 1.71275362e-01 -9.03730690e-01 1.99411899e-01
1.51217842e+00 6.16027057e-01 2.19051063e-01 3.99280667e-01
6.18882775e-01 1.62595510e-01 1.45413077e+00 5.13378024e-01
3.78791422e-01 7.82057762e-01 5.88560820e-01 5.98955899e-02
-1.80155888e-01 6.77873343e-02 4.90151048e-01 6.27932906e-01
-2.07898915e-01 -3.25749695e-01 -6.91658974e-01 6.13926113e-01
-2.23020768e+00 -1.45201027e+00 -6.83927357e-01 2.13246679e+00
2.79482216e-01 4.28962052e-01 5.61985314e-01 -5.54618537e-02
1.07933319e+00 8.33906308e-02 -5.22380769e-01 3.42855990e-01
-2.66823489e-02 -3.56214195e-01 6.40581846e-01 3.94660860e-01
-1.56265986e+00 6.41582429e-01 6.04251528e+00 1.02748489e+00
-6.83528721e-01 4.16418701e-01 -2.31038243e-01 -2.78079718e-01
4.07322586e-01 -7.07960054e-02 -1.49481297e+00 7.14817822e-01
8.69740188e-01 -1.95792064e-01 -1.82611287e-01 5.60542881e-01
3.31316769e-01 -4.61283177e-01 -5.82751215e-01 8.03970098e-01
-1.70039535e-01 -1.19377267e+00 -2.86878526e-01 5.61451428e-02
4.13088739e-01 1.30983412e-01 1.52033307e-02 2.78099060e-01
4.35596436e-01 -3.23751986e-01 9.57286954e-01 8.32163572e-01
2.17294708e-01 -6.92851186e-01 5.23388505e-01 4.60401624e-01
-1.97202694e+00 -2.41146490e-01 -8.49998072e-02 2.86072791e-01
6.80003285e-01 4.62582797e-01 -3.87219161e-01 1.19297647e+00
9.20011520e-01 8.11590493e-01 -5.44448078e-01 1.36727035e+00
7.14634061e-02 4.99802083e-01 -6.64112866e-01 1.32448182e-01
1.81085132e-02 -1.41433522e-01 1.15555787e+00 9.98777390e-01
5.81859648e-01 8.51645246e-02 5.43564916e-01 4.99336898e-01
6.04819596e-01 -4.67324764e-01 -4.61284369e-01 3.46257627e-01
6.27005160e-01 1.42196834e+00 -7.17121303e-01 -5.55040300e-01
-6.47184849e-01 2.82147646e-01 -2.51059026e-01 -6.41791597e-02
-1.49163771e+00 8.48932043e-02 4.07064259e-01 -1.16744816e-01
6.60922468e-01 -2.28453830e-01 1.44098610e-01 -9.04807627e-01
-2.16749549e-01 -4.54184413e-01 9.26089346e-01 -6.77242637e-01
-1.14171350e+00 3.08094293e-01 3.10554653e-01 -1.84836555e+00
5.68217076e-02 -4.78691638e-01 -6.72602057e-01 6.40569210e-01
-1.25059175e+00 -1.49601531e+00 -9.43377167e-02 8.49716187e-01
4.69740987e-01 -8.57672542e-02 2.49912620e-01 9.16693747e-01
-7.45685637e-01 2.33852774e-01 1.70396507e-01 -8.35393965e-02
6.06322229e-01 -8.11311245e-01 -3.23712140e-01 1.22842765e+00
-3.84969294e-01 4.63603616e-01 9.00769830e-01 -1.21635902e+00
-1.51027071e+00 -1.20679450e+00 3.44785899e-01 -5.95385909e-01
8.00331652e-01 -1.96694676e-02 -7.13034213e-01 7.16175139e-01
1.75852448e-01 4.39318857e-04 3.70831490e-01 -9.84984562e-02
3.39959472e-01 -1.08788991e-02 -1.10106003e+00 2.31807634e-01
8.16394210e-01 6.87283128e-02 -6.56446755e-01 1.61319360e-01
4.79043096e-01 -5.27773142e-01 -1.14465439e+00 8.07576537e-01
7.29513705e-01 -6.25522792e-01 1.20485115e+00 -1.72961652e-01
-6.36200726e-01 -1.30833030e+00 -1.10416472e-01 -6.87038720e-01
-5.91396809e-01 -1.95484355e-01 -8.84712458e-01 1.44622290e+00
5.97760454e-02 -2.58038521e-01 5.09606183e-01 1.10330127e-01
-3.34282249e-01 -4.47208792e-01 -1.16363704e+00 -1.24948704e+00
-3.93658876e-01 -3.36468786e-01 3.43465149e-01 6.19752169e-01
-4.65030789e-01 -2.60012168e-02 -6.87814653e-01 7.59471238e-01
1.31923044e+00 3.21219862e-01 7.14741290e-01 -1.31965899e+00
1.16269989e-02 -1.80520192e-01 -3.43290538e-01 -8.27121139e-01
-2.20772997e-01 -6.30845487e-01 5.95600232e-02 -1.47074997e+00
3.46625090e-01 -4.70813960e-01 -5.75416446e-01 3.61367404e-01
-2.49033719e-01 2.48067811e-01 3.96671891e-01 4.15238023e-01
-1.04070449e+00 4.86296803e-01 1.14009798e+00 -1.30810603e-01
-5.47746792e-02 4.69929874e-01 -2.58769155e-01 8.70481014e-01
7.21602917e-01 -7.57465243e-01 -7.91049153e-02 -9.41902325e-02
-3.06184381e-01 1.40021592e-01 6.29401863e-01 -1.30364430e+00
6.10791266e-01 -1.87265724e-01 4.87553179e-01 -1.80971122e+00
5.03813088e-01 -1.22855198e+00 7.43435800e-01 8.51286411e-01
3.85651320e-01 1.34197354e-01 6.45935774e-01 8.09887528e-01
7.99561739e-02 -6.63759410e-02 1.03455961e+00 1.76465765e-01
-1.02530646e+00 5.58238924e-01 -5.18468857e-01 -4.25395370e-01
1.58711648e+00 -3.70415866e-01 -3.65806669e-01 2.07000732e-01
-7.11000800e-01 8.84154141e-01 2.18567029e-01 5.84503651e-01
5.37929952e-01 -1.81680655e+00 -5.69476426e-01 -2.55502492e-01
8.82348120e-02 -2.55538404e-01 5.95946312e-01 1.25192773e+00
6.15883172e-02 4.80236173e-01 -1.44847289e-01 -1.00974452e+00
-1.70636439e+00 7.70517945e-01 1.50473297e-01 -2.84807831e-01
-6.69654250e-01 3.89104664e-01 -1.40530765e-02 -1.76376224e-01
1.68866605e-01 -5.88231198e-02 -2.39421204e-01 1.40161529e-01
6.19199514e-01 7.78164327e-01 -3.61641258e-01 -1.02361882e+00
-7.44247496e-01 8.13742220e-01 -6.08471856e-02 -2.86449189e-03
1.01134920e+00 -4.33805853e-01 1.94443196e-01 5.68168104e-01
3.94673705e-01 -8.88747051e-02 -1.48215044e+00 -1.55165285e-01
-3.31435073e-03 -5.01125038e-01 3.04809138e-02 -6.86901689e-01
-8.28850329e-01 4.03451055e-01 1.16980457e+00 1.04528308e-01
1.00137210e+00 1.68100953e-01 7.54037261e-01 -6.43081311e-03
3.15905124e-01 -9.87784803e-01 -2.09871158e-01 4.91318882e-01
3.49174231e-01 -1.34085417e+00 3.49651456e-01 -1.25488147e-01
-4.09593999e-01 6.74831569e-01 1.03488481e+00 7.00728744e-02
7.88133502e-01 4.50045466e-01 -1.62380174e-01 -2.86757708e-01
-6.78666115e-01 -5.85381746e-01 4.28094089e-01 4.62682307e-01
-1.41829655e-01 -1.23903714e-01 -3.15115422e-01 2.10031614e-01
5.86285710e-01 1.34525955e-01 -5.94627997e-03 1.19229722e+00
-7.50810623e-01 -8.11579466e-01 -1.03931355e+00 4.45510030e-01
-3.41422111e-01 3.98770869e-01 2.44849160e-01 1.15804768e+00
4.01588827e-01 1.05418408e+00 9.54637229e-02 -5.67590714e-01
3.91262740e-01 -1.32662445e-01 5.62305570e-01 -2.62776226e-01
-4.25404519e-01 5.09836912e-01 -4.51447777e-02 -3.58830810e-01
-9.85753298e-01 -1.02059245e+00 -1.49509811e+00 -4.92677480e-01
-9.13228095e-01 7.70041645e-02 3.18603992e-01 7.73121178e-01
1.68195873e-01 6.80114925e-01 3.38846862e-01 -1.00374424e+00
-9.77581590e-02 -8.25301349e-01 -4.62821305e-01 1.11620434e-01
4.06748325e-01 -1.40660620e+00 -2.80321240e-01 -2.81725917e-02] | [6.512155055999756, -2.033923387527466] |
3202a06d-4703-4147-8e5f-18551677b606 | machine-learning-driven-language-assessment | null | null | https://aclanthology.org/2020.tacl-1.17 | https://aclanthology.org/2020.tacl-1.17.pdf | Machine Learning--Driven Language Assessment | We describe a method for rapidly creating language proficiency assessments, and provide experimental evidence that such tests can be valid, reliable, and secure. Our approach is the first to use machine learning and natural language processing to induce proficiency scales based on a given standard, and then use linguistic models to estimate item difficulty directly for computer-adaptive testing. This alleviates the need for expensive pilot testing with human subjects. We used these methods to develop an online proficiency exam called the Duolingo English Test, and demonstrate that its scores align significantly with other high-stakes English assessments. Furthermore, our approach produces test scores that are highly reliable, while generating item banks large enough to satisfy security requirements. | ['Masato Hagiwara', 'Geoffrey T. LaFlair', 'Burr Settles'] | 2020-01-01 | null | null | null | tacl-2020-1 | ['skills-assessment'] | ['computer-vision'] | [-2.67214566e-01 2.37164900e-01 -4.26616549e-01 -5.03767788e-01
-1.38636494e+00 -1.12511694e+00 1.26833066e-01 5.47471166e-01
-7.39688277e-01 1.09161282e+00 1.68421373e-01 -1.06099582e+00
-1.30795047e-01 -8.94663513e-01 -5.82937419e-01 3.72818708e-01
4.49800700e-01 5.23801923e-01 2.04045534e-01 -2.22753286e-01
4.84786987e-01 1.28891632e-01 -1.27827477e+00 2.51922548e-01
1.39005971e+00 3.31546515e-01 -2.08896235e-01 6.90569699e-01
-6.21553101e-02 8.42506468e-01 -8.08439672e-01 -1.11395431e+00
1.49214208e-01 -4.94631350e-01 -9.91803169e-01 -2.86722273e-01
5.70573688e-01 -8.15702975e-01 5.25299311e-01 1.03235376e+00
2.31669843e-01 1.28682986e-01 5.75048387e-01 -5.97489178e-01
-1.19152749e+00 7.00507104e-01 -3.56177300e-01 2.35335305e-02
1.11408937e+00 2.04951823e-01 1.17820919e+00 -7.55249262e-01
3.87790948e-01 8.58882308e-01 8.36744547e-01 6.11826003e-01
-1.06769812e+00 -7.82153487e-01 -7.81757459e-02 -1.21727571e-01
-1.03030252e+00 -3.10833246e-01 1.35190859e-01 -5.01988113e-01
7.54933655e-01 -1.28599584e-01 1.00771964e+00 9.67132926e-01
1.61340803e-01 3.97308916e-01 1.62348855e+00 -7.83218861e-01
2.93514043e-01 6.49174333e-01 4.47156101e-01 8.55916142e-01
7.12467670e-01 -3.22514363e-02 -6.16867304e-01 -7.83301592e-02
7.14865983e-01 -6.09324813e-01 1.49744064e-01 4.74763140e-02
-7.86603153e-01 8.89394522e-01 -4.17663634e-01 3.14415932e-01
-2.75379628e-01 -5.15775859e-01 1.40248969e-01 8.00508916e-01
-2.36802176e-03 8.46484065e-01 -5.86790860e-01 -6.48725450e-01
-9.66534495e-01 1.68258771e-01 9.01772678e-01 5.65829396e-01
2.89435059e-01 3.60439420e-02 -8.68258253e-02 7.41585195e-01
4.07884449e-01 4.95893896e-01 8.34192753e-01 -1.03076541e+00
5.58117509e-01 4.47484076e-01 3.94465476e-01 -5.50903440e-01
-1.03806235e-01 -9.34831500e-02 1.52072102e-01 5.16730487e-01
6.40877366e-01 -4.16003972e-01 -5.44102550e-01 1.90114033e+00
4.96627903e-03 -2.76244134e-01 1.93789080e-02 2.73125082e-01
3.64532828e-01 3.25243086e-01 5.90286136e-01 -1.56088710e-01
1.35975254e+00 -4.37586784e-01 -5.01713991e-01 -2.01267660e-01
9.66150165e-01 -6.19880617e-01 1.90138578e+00 8.33797097e-01
-1.73923039e+00 -6.56414211e-01 -1.16893291e+00 4.66805175e-02
-1.73698291e-01 -2.43941709e-01 7.79568017e-01 1.79823494e+00
-1.24350452e+00 2.68982768e-01 -5.79724610e-01 -4.57555242e-02
5.65313175e-02 3.92165750e-01 -6.72828615e-01 -4.46893871e-02
-1.11126268e+00 9.65311050e-01 4.33106512e-01 -4.74300891e-01
-5.90848863e-01 -3.89557213e-01 -1.16268253e+00 1.21641271e-01
-6.16667457e-02 -1.08830430e-01 1.45850432e+00 -8.41397643e-01
-1.83754575e+00 9.32230413e-01 1.00759352e-02 -1.47034535e-02
8.47069845e-02 -6.30473793e-02 -4.67172652e-01 6.78041577e-02
3.33024710e-01 3.28372806e-01 1.21869862e-01 -3.30704421e-01
-6.84580624e-01 1.75614282e-03 2.66645133e-01 1.59628764e-01
-1.02225423e+00 5.39025486e-01 4.01845202e-02 -3.62563759e-01
-1.93901554e-01 -3.49768639e-01 4.96177115e-02 -5.55844903e-01
2.73542702e-01 -3.89280885e-01 -3.34263682e-01 -1.15534806e+00
1.28002548e+00 -1.86341751e+00 -5.54361105e-01 7.42616653e-01
-8.04007612e-03 3.33530217e-01 -3.76601040e-01 1.23207204e-01
6.67709410e-02 5.26189089e-01 3.06823611e-01 1.87917113e-01
3.24641973e-01 -4.31858629e-01 4.06307913e-03 5.14579639e-02
1.13035835e-01 1.08417118e+00 -7.24862695e-01 -4.93065059e-01
-2.24528894e-01 -1.22691609e-01 -1.16103351e+00 2.75971413e-01
1.74218848e-01 2.22323596e-01 -1.30271330e-01 6.95457101e-01
1.56459972e-01 1.25846237e-01 2.53514439e-01 8.14621329e-01
1.03601567e-01 1.05802715e+00 -1.02061081e+00 1.37237549e+00
-5.27061939e-01 2.24955320e-01 -1.31112903e-01 -6.77877307e-01
9.65869427e-01 3.64076346e-01 -6.61589727e-02 -8.21079731e-01
-1.18343125e-03 5.16287923e-01 1.22293711e-01 -5.37560523e-01
4.48760033e-01 -3.79067093e-01 -3.26354772e-01 1.00813866e+00
1.95538089e-01 -5.76905489e-01 4.78925675e-01 1.65187910e-01
1.06198132e+00 8.72285441e-02 3.53832364e-01 -3.67317766e-01
5.12097538e-01 -1.22365758e-01 5.87289691e-01 7.91204154e-01
-3.61819655e-01 -5.19099869e-02 4.63670433e-01 3.04870568e-02
-1.03615022e+00 -1.23336899e+00 -2.67608445e-02 1.31060874e+00
-5.98073661e-01 -2.80192792e-01 -1.14522731e+00 -8.40678930e-01
-5.91705702e-02 9.82847154e-01 -1.66818529e-01 -1.57485276e-01
-2.68476605e-01 -1.53717414e-01 7.31168747e-01 7.05725074e-01
2.25820243e-01 -9.86813962e-01 -1.29787967e-01 2.72183985e-01
-2.35994756e-01 -9.03188050e-01 -3.85655999e-01 -2.20378101e-01
-5.61964393e-01 -8.44210982e-01 -1.77766889e-01 -1.14371562e+00
5.51375210e-01 -2.34042794e-01 1.29576600e+00 4.33599770e-01
2.36871809e-01 5.93376338e-01 -3.07810783e-01 -6.18459642e-01
-6.10076308e-01 1.61047913e-02 4.71490711e-01 -6.74540997e-01
1.17426538e+00 -3.05508196e-01 -3.21665369e-02 2.02325881e-01
-7.65480161e-01 -4.02850002e-01 6.03479922e-01 7.71335363e-01
6.97584674e-02 -8.23582187e-02 1.08711851e+00 -9.36665535e-01
1.26081622e+00 -2.91231334e-01 -7.88608134e-01 4.44613069e-01
-8.63716722e-01 -2.20185205e-01 5.45971692e-01 -5.64179659e-01
-9.17385459e-01 -3.72739971e-01 -7.26131082e-01 7.78558016e-01
-3.26539755e-01 8.74306858e-01 5.49927242e-02 -4.75265443e-01
9.52785790e-01 -1.45164430e-01 -2.02128887e-01 -6.14239648e-02
-9.98543054e-02 8.76296103e-01 5.31175911e-01 -1.08157706e+00
1.10254097e+00 -4.68329400e-01 -6.10671282e-01 -4.55775261e-01
-8.77383232e-01 1.13878315e-02 -4.92756486e-01 -1.01027330e-02
6.44517839e-01 -1.05238104e+00 -8.71499896e-01 1.20908819e-01
-5.90971708e-01 -7.25891531e-01 -4.31282669e-01 1.04239845e+00
-4.97977316e-01 3.56678575e-01 -1.13954318e+00 -8.35112929e-01
-1.40002996e-01 -8.89045894e-01 4.45202976e-01 4.27121967e-01
-7.64142811e-01 -1.12700427e+00 2.34834179e-01 8.43765736e-01
3.57508928e-01 -3.15221578e-01 1.08281243e+00 -8.35595608e-01
-1.45144939e-01 -5.71213543e-01 1.65899575e-01 5.38195074e-01
-2.08150193e-01 4.76751328e-02 -4.03737277e-01 -3.43646049e-01
1.72902241e-01 -1.24566424e+00 -3.81837152e-02 9.85477939e-02
6.26536071e-01 -1.94627330e-01 7.25281775e-01 3.12137336e-01
1.08589506e+00 6.45618141e-02 6.31668210e-01 4.59520131e-01
2.11286731e-02 6.92123413e-01 6.60672307e-01 -4.36506979e-02
7.00396657e-01 2.51892984e-01 -8.07536483e-01 2.81946599e-01
2.58126646e-01 -7.22892523e-01 1.06984067e+00 1.31700563e+00
2.58968532e-01 2.95507193e-01 -1.02463913e+00 5.89187324e-01
-1.01657462e+00 -8.91065538e-01 2.19568893e-01 2.32889056e+00
1.40446162e+00 5.84621489e-01 3.67004275e-01 1.23857155e-01
4.44780499e-01 -7.15623915e-01 -9.72996056e-02 -9.77270126e-01
5.31403646e-02 1.12176168e+00 1.73098996e-01 8.36545408e-01
-5.19800067e-01 1.08955717e+00 8.37266350e+00 3.97995561e-01
-4.92008269e-01 1.21312901e-01 7.72697449e-01 1.39041767e-01
-8.59576583e-01 -8.83047432e-02 -9.78570163e-01 1.26993448e-01
1.59788585e+00 -5.46705306e-01 3.62156689e-01 5.81341863e-01
7.73642883e-02 -9.21452269e-02 -8.54024112e-01 2.69915372e-01
-6.38029650e-02 -7.51512825e-01 -1.50684774e-01 4.11844999e-02
8.80752027e-01 -4.86146569e-01 1.49640560e-01 6.83958411e-01
9.87657189e-01 -1.11613083e+00 8.22910607e-01 2.37146184e-01
1.17122364e+00 -6.76908076e-01 7.61463225e-01 3.50906104e-01
-7.32484460e-01 -3.67354900e-01 -6.14855111e-01 -5.28702319e-01
-1.93513975e-01 2.58053422e-01 -7.76020467e-01 -1.12813585e-01
3.17477971e-01 -2.37819225e-01 -8.24048698e-01 1.01917410e+00
-7.99244404e-01 1.21529257e+00 -1.06870696e-01 -4.65901792e-01
4.00187215e-03 -1.94881245e-01 -3.93073767e-01 9.49209034e-01
5.40268481e-01 1.89767465e-01 1.03098378e-01 5.89639902e-01
-1.93937272e-01 8.02246451e-01 -6.52109146e-01 -2.38801241e-01
7.46419251e-01 1.01179731e+00 -4.42328572e-01 -3.47337067e-01
-7.40682542e-01 4.81626868e-01 5.86523414e-01 1.60833836e-01
-5.03737748e-01 -6.10306323e-01 1.68749899e-01 1.51932374e-01
-1.64312467e-01 -5.27741373e-01 -6.92960083e-01 -1.27502358e+00
2.44749457e-01 -1.50788522e+00 3.48948568e-01 -5.73626161e-01
-1.10013413e+00 5.97156920e-02 -1.95440590e-01 -8.91789734e-01
-6.45043433e-01 -9.54530597e-01 -6.27971590e-01 1.21536207e+00
-9.37019467e-01 -5.95403552e-01 -3.05182692e-02 6.61871254e-01
1.02595732e-01 -4.12469983e-01 1.03597713e+00 2.81880885e-01
-3.73513520e-01 1.38454509e+00 -3.76304030e-01 1.88589185e-01
9.69246149e-01 -1.36818767e+00 5.64994037e-01 1.14793575e+00
2.47196004e-01 1.09568810e+00 2.20669568e-01 -9.10292923e-01
-1.01472449e+00 -3.77318710e-01 1.13508666e+00 -7.95986354e-01
9.87453818e-01 -3.59262884e-01 -9.16632414e-01 7.76333749e-01
1.71147779e-01 -7.01269269e-01 1.27331090e+00 4.15771067e-01
-5.61513364e-01 8.55028331e-02 -1.40091324e+00 5.41472614e-01
8.66057217e-01 -1.00184810e+00 -9.70214427e-01 3.96441579e-01
5.49756408e-01 -1.53955907e-01 -1.17133975e+00 4.63914871e-02
7.11199641e-01 -8.24768960e-01 5.37231624e-01 -9.19271767e-01
4.23994541e-01 2.26035580e-01 3.08473170e-01 -1.15942264e+00
-5.69077849e-01 -6.16589427e-01 2.91656315e-01 1.30877674e+00
1.02373576e+00 -8.93058538e-01 9.66290236e-01 1.30423665e+00
1.26310602e-01 -2.90030569e-01 -5.40999949e-01 -7.13960409e-01
6.50891066e-01 -3.47709149e-01 6.85264945e-01 1.05039001e+00
7.38918662e-01 1.11782789e-01 1.06214158e-01 -7.42893200e-03
3.46951514e-01 -4.93179113e-01 7.40903497e-01 -1.33169186e+00
-7.23632693e-01 -4.13628668e-01 -4.89905983e-01 -7.01571763e-01
3.96316320e-01 -6.81339800e-01 -8.73815045e-02 -1.10761702e+00
1.31524369e-01 -3.34624082e-01 -3.69044989e-01 6.38672769e-01
-5.62429845e-01 4.28409904e-01 -1.43098429e-01 -5.21316886e-01
-4.01786089e-01 -1.36517927e-01 7.67451525e-01 1.84798151e-01
-1.30771279e-01 -6.88815042e-02 -1.69939494e+00 6.47596061e-01
1.08602750e+00 -2.11810067e-01 -5.50246596e-01 -4.35254514e-01
6.45020843e-01 1.64575502e-02 -3.92960310e-01 -1.12280846e+00
2.30227441e-01 -4.81287748e-01 6.60877943e-01 2.11928055e-01
-3.03419858e-01 -3.40678990e-01 -4.01913226e-01 1.68565243e-01
-4.66940105e-01 5.57213366e-01 3.78122658e-01 -2.88314134e-01
-2.56491929e-01 -9.69256222e-01 6.59257233e-01 1.73036456e-02
-2.94394016e-01 -3.46727706e-02 -6.41777456e-01 1.52164608e-01
9.70654666e-01 -2.03245252e-01 -2.45737880e-01 -7.27096558e-01
-3.00007522e-01 1.05983429e-02 6.91273689e-01 8.85983109e-02
4.77056205e-01 -1.24836075e+00 -1.01045048e+00 6.10278130e-01
-9.14165452e-02 -7.21320450e-01 -1.90041810e-01 3.37045580e-01
-7.27321267e-01 4.85621363e-01 -5.07171750e-01 1.39529660e-01
-1.04207277e+00 2.59019375e-01 3.86151075e-02 -5.61700106e-01
-8.12737122e-02 9.45487976e-01 -2.81382024e-01 -6.15366757e-01
7.16426894e-02 -7.38684610e-02 -3.80068004e-01 -4.08849567e-01
1.06859100e+00 -1.57474846e-01 1.86428547e-01 -2.07879335e-01
2.47091770e-01 3.82261574e-01 -9.70297158e-02 -8.93857419e-01
1.12444532e+00 1.49378955e-01 -5.12233637e-02 2.50865519e-01
6.18812561e-01 9.70830500e-01 -6.31398857e-01 -4.22618166e-02
3.77697974e-01 -5.19549072e-01 -3.01266551e-01 -1.02294660e+00
-4.03505713e-01 4.38365221e-01 2.47192353e-01 3.16149443e-01
1.08203280e+00 -5.81200778e-01 6.09529138e-01 7.68505931e-01
7.30843842e-01 -1.45147419e+00 3.78357656e-02 6.35544717e-01
2.80213356e-01 -8.12797904e-01 -1.35311820e-02 -1.98238343e-01
-5.33117294e-01 9.07542229e-01 9.33009803e-01 1.36554524e-01
3.08909386e-01 3.08782190e-01 2.20549121e-01 1.80151790e-01
-6.31117344e-01 2.17467666e-01 1.16583258e-01 7.32865989e-01
8.49007189e-01 2.21529365e-01 -9.35860217e-01 9.53071356e-01
-8.98650169e-01 2.13736758e-01 8.85841429e-01 1.00749958e+00
-6.87852263e-01 -1.58637178e+00 -6.64495468e-01 6.45965040e-01
-8.81234825e-01 -2.91363984e-01 -4.52528805e-01 5.85761666e-01
-2.98867166e-01 1.16905189e+00 -2.79379815e-01 -5.01586378e-01
4.24577415e-01 6.60251677e-01 6.35853410e-01 -1.02168167e+00
-7.63689578e-01 -4.00835693e-01 2.53760844e-01 -1.94296628e-01
7.74947973e-03 -7.77447224e-01 -8.33021164e-01 -6.32666349e-01
-3.24717075e-01 7.26712525e-01 5.02967119e-01 1.00510454e+00
-1.40372008e-01 -1.11229450e-01 5.03037453e-01 -9.14999396e-02
-9.47418213e-01 -9.95726526e-01 -6.75886929e-01 4.43392456e-01
-1.99270919e-01 -2.30981380e-01 -1.62443385e-01 -2.11031541e-01] | [11.052508354187012, 9.751728057861328] |
2af2f006-5ca3-4d79-8211-f15f79920d3a | interactive-evidence-detection-train-state-of | null | null | https://aclanthology.org/D19-6613 | https://aclanthology.org/D19-6613.pdf | Interactive Evidence Detection: train state-of-the-art model out-of-domain or simple model interactively? | Finding evidence is of vital importance in research as well as fact checking and an evidence detection method would be useful in speeding up this process. However, when addressing a new topic there is no training data and there are two approaches to get started. One could use large amounts of out-of-domain data to train a state-of-the-art method, or to use the small data that a person creates while working on the topic. In this paper, we address this problem in two steps. First, by simulating users who read source documents and label sentences they can use as evidence, thereby creating small amounts of training data for an interactively trained evidence detection model; and second, by comparing such an interactively trained model against a pre-trained model that has been trained on large out-of-domain data. We found that an interactively trained model not only often out-performs a state-of-the-art model but also requires significantly lower amounts of computational resources. Therefore, especially when computational resources are scarce, e.g. no GPU available, training a smaller model on the fly is preferable to training a well generalising but resource hungry out-of-domain model. | ['Chris Stahlhut'] | 2019-11-01 | null | null | null | ws-2019-11 | ['small-data'] | ['computer-vision'] | [ 8.54518041e-02 2.57032841e-01 -1.00577570e-01 -2.76821822e-01
-8.98955107e-01 -6.69309556e-01 6.47803545e-01 7.52427399e-01
-6.58600688e-01 7.87524223e-01 -7.58821666e-02 -7.99586773e-01
1.08750857e-01 -9.82232392e-01 -8.76688480e-01 -2.03738168e-01
2.30123118e-01 8.44675124e-01 6.59608185e-01 4.64296676e-02
7.13940680e-01 1.91095546e-01 -1.81949353e+00 4.37000722e-01
8.21117997e-01 4.85214651e-01 4.65488046e-01 8.17980051e-01
-4.97295648e-01 3.98897886e-01 -9.50346291e-01 -6.56317115e-01
5.87034076e-02 -3.42662543e-01 -9.13073659e-01 6.63215593e-02
6.72399759e-01 -1.40953317e-01 2.41992444e-01 9.82887447e-01
3.27269942e-01 1.06986374e-01 3.85391623e-01 -9.04448092e-01
-2.31630802e-01 3.94868106e-01 -5.06377041e-01 5.57880402e-01
6.61723971e-01 1.13424234e-01 5.35266042e-01 -7.75283396e-01
9.85738277e-01 9.22108471e-01 5.16755521e-01 4.01456445e-01
-1.17343462e+00 -5.94108880e-01 -9.72499792e-03 1.51067987e-01
-9.53080475e-01 -1.94203824e-01 5.68298995e-01 -6.01046264e-01
8.09554517e-01 3.78925145e-01 7.70923615e-01 1.15853846e+00
-1.55652836e-01 8.20607245e-01 1.37692821e+00 -7.82714963e-01
3.38586450e-01 8.45728278e-01 3.02326292e-01 9.23917770e-01
5.49242377e-01 -5.74755110e-02 -5.59815347e-01 -4.31221485e-01
6.99087560e-01 -2.65842527e-01 -1.53934490e-02 -6.72308207e-02
-1.07687092e+00 8.75576556e-01 -1.54466543e-03 6.05530560e-01
-4.36752349e-01 -3.20567310e-01 5.42856514e-01 3.84430736e-01
6.86248720e-01 6.01160765e-01 -4.99515384e-01 -4.72436339e-01
-1.45879507e+00 2.91036844e-01 1.07809269e+00 5.20869493e-01
6.21578634e-01 -5.70274830e-01 -8.18133280e-02 8.44398618e-01
7.95366392e-02 3.55987623e-02 5.99127650e-01 -6.42443776e-01
5.75700343e-01 8.01820159e-01 1.51636630e-01 -8.15484941e-01
-1.11060739e-01 -3.39033604e-01 -3.62021297e-01 3.86279523e-01
9.34015870e-01 -1.26615614e-01 -8.68530393e-01 1.23178411e+00
5.97100556e-01 -1.57517157e-02 -4.38968316e-02 7.97143102e-01
7.84988880e-01 6.69680238e-01 -9.19861905e-03 -1.90503612e-01
1.59653330e+00 -9.82992411e-01 -5.67401826e-01 -2.97357798e-01
8.94846201e-01 -8.16510975e-01 1.16816437e+00 8.41733575e-01
-1.25026751e+00 -6.24123275e-01 -9.63332236e-01 -1.26130924e-01
-7.94596851e-01 1.94854895e-03 4.14398760e-01 7.73263335e-01
-9.08483148e-01 6.69930041e-01 -4.87299234e-01 -4.58869457e-01
5.40457368e-01 9.34337676e-02 -3.63106251e-01 -2.57823110e-01
-1.03952968e+00 1.17341352e+00 5.06143332e-01 -3.68674695e-01
-6.65448368e-01 -6.31349504e-01 -6.17061973e-01 4.00052965e-02
7.38998592e-01 -3.85549337e-01 1.39661515e+00 -7.72364497e-01
-1.25152385e+00 1.10095286e+00 -2.07258239e-01 -1.46600321e-01
7.54524648e-01 -2.33717903e-01 -3.11271042e-01 8.14279094e-02
3.19717973e-01 2.81521946e-01 7.39006877e-01 -1.37251914e+00
-6.63627565e-01 -3.26562703e-01 1.89413399e-01 -8.77932087e-02
-3.06186974e-01 3.03048909e-01 -7.11322427e-01 -4.05782461e-01
-2.39234522e-01 -7.04470634e-01 -1.69593900e-01 9.42967087e-02
-1.64786413e-01 -4.62060452e-01 9.86542761e-01 -8.36583197e-01
1.16215432e+00 -1.80996871e+00 -3.61720026e-01 5.62394023e-01
4.12110649e-02 8.15216541e-01 8.30769613e-02 3.63404810e-01
3.63401957e-02 2.23465592e-01 -4.69817035e-02 -1.59690842e-01
-2.40518093e-01 2.12034687e-01 2.90551241e-02 9.55247208e-02
1.23598941e-01 4.49545979e-01 -1.33376193e+00 -8.50729764e-01
-2.45294496e-02 2.89452821e-01 -3.85862231e-01 2.11998492e-01
-4.49737877e-01 4.54730183e-01 -4.38600212e-01 3.75354141e-01
5.62477291e-01 -4.87191737e-01 2.36908168e-01 3.64486516e-01
-2.33995125e-01 4.79121894e-01 -1.54697740e+00 1.60243380e+00
-6.27817452e-01 7.17776597e-01 -1.28988683e-01 -1.04956663e+00
7.37126350e-01 3.71891379e-01 -2.82240976e-02 -7.43268132e-01
3.35693657e-02 3.48168284e-01 3.50417159e-02 -7.91330695e-01
5.49016356e-01 -5.47175333e-02 3.02637428e-01 9.96865392e-01
8.18020850e-02 -1.69200346e-01 6.23389721e-01 3.78453851e-01
1.43747318e+00 -6.68722615e-02 3.19559783e-01 3.19869071e-02
3.52269083e-01 3.27849776e-01 -2.12896876e-02 9.62273180e-01
2.06018895e-01 1.59469724e-01 4.58153874e-01 -5.03540397e-01
-1.07019246e+00 -5.47304988e-01 -3.23590875e-01 1.20885086e+00
-3.11383545e-01 -5.39261043e-01 -7.56868064e-01 -1.01920581e+00
-9.51470435e-02 1.09193528e+00 -4.37928826e-01 3.33182842e-01
-5.01495421e-01 -4.19861913e-01 2.13775665e-01 4.68657702e-01
5.19967616e-01 -1.20377588e+00 -9.55587685e-01 3.03866237e-01
-1.60471290e-01 -8.28260481e-01 7.01621920e-02 1.74672320e-01
-1.02858710e+00 -1.05389071e+00 -6.42856658e-01 -4.75638658e-01
9.65323567e-01 2.63629943e-01 1.38349295e+00 6.94544911e-01
-3.51797640e-01 2.12004393e-01 -4.81454283e-01 -8.53780508e-01
-7.00895131e-01 -2.33865023e-01 -2.96534777e-01 -4.51460391e-01
4.55135852e-01 -3.14760178e-01 -3.67553949e-01 3.25003445e-01
-1.07671976e+00 2.09803894e-01 8.25237930e-01 8.10826778e-01
2.46757671e-01 3.54170382e-01 2.01183498e-01 -1.44019055e+00
1.01916599e+00 -4.65533882e-01 -5.94476461e-01 3.45170945e-01
-7.43536174e-01 1.30543396e-01 3.67462963e-01 -6.44320965e-01
-1.08519328e+00 -2.22547621e-01 -9.63158309e-02 -1.41613111e-01
-3.06242466e-01 7.90644526e-01 3.13905120e-01 2.57393032e-01
8.57649505e-01 -1.68984562e-01 -2.44258970e-01 -4.54429030e-01
2.86849529e-01 7.31732786e-01 3.29293787e-01 -7.78094530e-01
7.80019164e-01 1.81550607e-01 -2.13243023e-01 -7.85404384e-01
-9.62577939e-01 -8.10641110e-01 -6.23443782e-01 -3.57124358e-01
5.07554114e-01 -3.20539385e-01 -3.50663990e-01 6.47055954e-02
-1.29632127e+00 -5.50311446e-01 -3.50035280e-01 2.29825854e-01
-1.40522882e-01 1.93008289e-01 -2.45976642e-01 -9.32154000e-01
-1.69261679e-01 -7.58518577e-01 9.44234192e-01 2.11265936e-01
-4.08592314e-01 -9.40592825e-01 3.13017994e-01 5.97806990e-01
1.53112039e-01 5.28859608e-02 7.75288999e-01 -9.83039379e-01
-3.92233282e-01 -7.23784626e-01 -3.04950476e-01 9.67991352e-02
-1.08765364e-01 -4.74128649e-02 -1.04364920e+00 -7.81417340e-02
1.69653282e-01 -4.17952925e-01 5.32217503e-01 -1.38738215e-01
1.44308281e+00 -4.22114313e-01 -6.53687298e-01 -2.49219000e-01
1.19103825e+00 1.69058871e-02 6.02942348e-01 6.30398989e-01
2.57555783e-01 6.36759520e-01 8.14549565e-01 2.19617829e-01
1.71537951e-01 7.39137709e-01 -7.13832453e-02 -1.11833252e-01
4.28432561e-02 -1.83954731e-01 3.24749276e-02 5.29766202e-01
-3.06242496e-01 -3.24654549e-01 -1.02281284e+00 6.07388020e-01
-1.94320118e+00 -1.22848046e+00 -3.72529507e-01 2.31479883e+00
1.16254699e+00 5.81384242e-01 2.84738928e-01 3.84670377e-01
5.37286699e-01 -2.85042435e-01 -2.23189309e-01 -6.36559129e-01
3.25216532e-01 4.92139578e-01 4.13874835e-02 2.61606574e-01
-7.79879570e-01 4.56196934e-01 5.84843826e+00 9.55444872e-01
-9.30694342e-01 3.97268295e-01 5.94918072e-01 6.65626721e-03
-4.29139063e-02 6.42582476e-02 -6.75980151e-01 7.09244609e-01
1.32388473e+00 6.60342500e-02 3.86918299e-02 1.07219851e+00
1.68784320e-01 -8.72961342e-01 -1.22143161e+00 7.33963907e-01
1.18425056e-01 -1.37139761e+00 -2.02936649e-01 2.24284008e-01
5.67159772e-01 -3.62880856e-01 -4.16046411e-01 3.36366922e-01
5.24384201e-01 -7.29416609e-01 6.04567528e-01 4.39912587e-01
4.25305665e-01 -4.29815114e-01 9.32347536e-01 9.79946911e-01
-6.48860991e-01 2.06958070e-01 -2.38628134e-01 -9.48000103e-02
1.50526240e-01 9.53275442e-01 -1.17505431e+00 3.60922456e-01
6.22646153e-01 2.78498102e-02 -7.77422309e-01 1.35889101e+00
-5.16489625e-01 8.33312631e-01 -4.83885646e-01 -2.55138129e-01
1.48966946e-02 -1.74341677e-03 3.65253627e-01 1.67059028e+00
2.44495749e-01 2.83025295e-01 1.73372075e-01 7.56689191e-01
-1.62757680e-01 -3.27100083e-02 -6.16015375e-01 -3.78609709e-02
1.84323043e-01 1.32825720e+00 -9.87594545e-01 -9.95900333e-01
-4.86814141e-01 8.72444928e-01 5.00118792e-01 -8.61605927e-02
-5.29619634e-01 -1.07435621e-01 -1.99946523e-01 5.17207384e-01
2.68267214e-01 -1.26659006e-01 -3.08777720e-01 -9.50156271e-01
1.39066771e-01 -1.06981027e+00 5.50435007e-01 -1.05331802e+00
-1.19715607e+00 3.77673864e-01 1.64535061e-01 -9.83615279e-01
-4.01643693e-01 -6.11378968e-01 -8.45436037e-01 8.03500354e-01
-1.26296735e+00 -6.75489247e-01 -4.63832349e-01 2.43937865e-01
7.81182945e-01 2.05453560e-01 8.39435637e-01 3.12129468e-01
-3.16329420e-01 2.29142323e-01 8.14673316e-04 1.18727759e-01
6.07921183e-01 -1.36919868e+00 2.62046099e-01 8.53609502e-01
5.23422778e-01 8.64433944e-01 8.34912837e-01 -9.99731421e-01
-1.10769212e+00 -6.69075429e-01 9.95523512e-01 -7.01196671e-01
7.37192750e-01 -3.99478793e-01 -1.39515650e+00 2.92042613e-01
2.65041023e-01 -3.31373177e-02 7.14984953e-01 4.04824704e-01
-2.47797668e-01 4.72709119e-01 -1.30158305e+00 5.83420575e-01
8.61036122e-01 -4.87691820e-01 -8.90504420e-01 6.60531759e-01
9.27658603e-02 -6.80621922e-01 -3.95677358e-01 -1.13672569e-01
5.14936507e-01 -8.09186637e-01 7.85374105e-01 -6.60061359e-01
6.17283344e-01 -3.88243012e-02 3.45278114e-01 -1.31016290e+00
6.09583408e-02 -3.58054489e-01 -2.33837113e-01 1.27115023e+00
6.91522002e-01 -3.36483657e-01 9.05956268e-01 9.37337637e-01
1.03894539e-01 -7.39704192e-01 -5.14621079e-01 -8.41536283e-01
-2.32554212e-01 -6.86605811e-01 2.20546141e-01 1.06399477e+00
2.57726699e-01 3.43465209e-01 2.06970721e-01 3.22157294e-02
3.67724329e-01 1.96470553e-03 1.16757941e+00 -1.53782213e+00
-1.99957073e-01 -1.85824737e-01 -2.18743216e-02 -9.59856451e-01
-1.50066420e-01 -7.58687854e-01 2.78746426e-01 -1.83635199e+00
4.17608052e-01 -7.33382344e-01 -1.65838059e-02 5.47283649e-01
-3.54367614e-01 2.37044040e-02 -7.56778046e-02 1.90592661e-01
-4.41355199e-01 -2.45232195e-01 1.13865817e+00 1.48012295e-01
-2.29513258e-01 1.09676786e-01 -5.09898067e-01 8.83702934e-01
6.30012751e-01 -8.62696648e-01 -3.20360780e-01 -1.23833574e-01
4.17024642e-01 4.03832346e-02 6.51584089e-01 -9.60466027e-01
4.72318679e-01 -5.37623726e-02 6.17797792e-01 -6.85953796e-01
2.76348621e-01 -6.99659169e-01 -1.45424664e-01 4.52179164e-01
-2.07263932e-01 6.33435100e-02 4.22483861e-01 6.30398452e-01
7.47245327e-02 -1.18452442e+00 4.04985338e-01 -5.99355936e-01
-6.08955085e-01 -3.23605776e-01 -5.38982213e-01 -1.45253772e-02
7.93538272e-01 -6.16556168e-01 -5.24211109e-01 -1.40288770e-01
-6.24751866e-01 -1.20889716e-01 3.03926200e-01 1.16074570e-01
5.31202734e-01 -8.66129935e-01 -5.14255404e-01 1.62335023e-01
-4.64327224e-02 1.68249577e-01 -1.36552438e-01 6.26556039e-01
-4.43462729e-01 2.63020962e-01 -3.82745862e-02 -5.66828251e-01
-1.51663554e+00 6.82992160e-01 -3.66666585e-01 -6.90498769e-01
-7.39853442e-01 9.42206383e-01 -2.55072594e-01 -3.18137139e-01
2.09302366e-01 -1.49463385e-01 -2.63349414e-01 1.90807343e-01
6.55378938e-01 3.98725092e-01 4.11187381e-01 -6.57966807e-02
-1.11273378e-01 2.16389075e-01 -3.24989468e-01 -5.08459747e-01
1.57709289e+00 3.49365085e-01 1.13770299e-01 4.38572973e-01
7.22366214e-01 2.43754700e-01 -8.20519149e-01 -2.31032282e-01
2.27324769e-01 -9.02638614e-01 2.56555885e-01 -1.08851457e+00
-6.22026443e-01 7.70315051e-01 3.82947028e-01 6.37671888e-01
7.77803004e-01 2.20239967e-01 5.05299807e-01 5.65737605e-01
4.56078678e-01 -1.36012864e+00 5.23355484e-01 1.94258749e-01
8.72857094e-01 -1.60247576e+00 2.87597179e-01 -5.15013576e-01
-6.33316755e-01 1.23123538e+00 5.83620608e-01 -3.25794076e-03
5.91243565e-01 2.37795159e-01 -1.16200499e-01 -8.11031580e-01
-6.76541924e-01 -1.61202639e-01 3.68837893e-01 4.86164540e-01
5.07811069e-01 -3.32793444e-01 -3.61090630e-01 2.66861618e-01
-2.78002620e-01 4.83545274e-01 5.72629929e-01 1.32664704e+00
-4.29758757e-01 -1.27705443e+00 -7.15719759e-01 8.11591029e-01
-4.25289422e-01 -1.49610460e-01 -4.85889167e-01 1.01528394e+00
2.47883052e-01 1.08493984e+00 -4.06902805e-02 1.09279633e-01
3.58839214e-01 2.88586706e-01 5.27933776e-01 -1.01921654e+00
-7.08059072e-01 -1.59212068e-01 5.36389351e-01 -2.58737624e-01
-4.17234719e-01 -7.82969654e-01 -9.80246067e-01 -3.49460274e-01
-5.84597766e-01 2.51992524e-01 1.05392718e+00 1.25845397e+00
1.75983980e-01 3.63440454e-01 -4.39171260e-03 -7.85000384e-01
-1.76478341e-01 -1.07112467e+00 -1.04464665e-01 3.83657426e-01
-2.59846803e-02 -6.58258915e-01 -1.30618855e-01 1.87859342e-01] | [9.562148094177246, 8.145552635192871] |
3c0e13d3-5e5a-4f2c-a00b-ffe13608e6bb | bone-marrow-sparing-for-cervical-cancer | 2204.09278 | null | https://arxiv.org/abs/2204.09278v1 | https://arxiv.org/pdf/2204.09278v1.pdf | Bone marrow sparing for cervical cancer radiotherapy on multimodality medical images | Cervical cancer threatens the health of women seriously. Radiotherapy is one of the main therapy methods but with high risk of acute hematologic toxicity. Delineating the bone marrow (BM) for sparing using computer tomography (CT) images to plan before radiotherapy can effectively avoid this risk. Comparing with magnetic resonance (MR) images, CT lacks the ability to express the activity of BM. Thus, in current clinical practice, medical practitioners manually delineate the BM on CT images by corresponding to MR images. However, the time?consuming delineating BM by hand cannot guarantee the accuracy due to the inconsistency of the CT-MR multimodal images. In this study, we propose a multimodal image oriented automatic registration method for pelvic BM sparing, which consists of three-dimensional bone point cloud reconstruction, a local spherical system iteration closest point registration for marking BM on CT images. Experiments on patient dataset reveal that our proposed method can enhance the multimodal image registration accuracy and efficiency for medical practitioners in sparing BM of cervical cancer radiotherapy. The method proposed in this contribution might also provide references for similar studies in other clinical application. | ['Xiuming Zhang', 'Han Gao', 'Hanzi Xu', 'Kexin Gan', 'Jie Yuan', 'Ying Sun', 'Yuening Wang'] | 2022-04-20 | null | null | null | null | ['point-cloud-reconstruction'] | ['computer-vision'] | [ 3.74605358e-01 2.07989424e-01 -4.36716259e-01 -1.47903472e-01
-8.37987244e-01 -4.34287071e-01 2.76957214e-01 4.15572315e-01
-4.88286048e-01 5.66390574e-01 2.32043892e-01 -6.17681026e-01
-1.47995606e-01 -9.93957818e-01 -1.50701016e-01 -1.04036939e+00
3.64765048e-01 1.15644622e+00 4.73041832e-01 -2.51800030e-01
3.90450895e-01 1.08826840e+00 -7.44237781e-01 3.59220505e-01
7.15062320e-01 6.78928554e-01 5.55122674e-01 6.21955335e-01
-4.90371764e-01 5.29515803e-01 -1.94221392e-01 -3.10081929e-01
5.24124950e-02 -6.26658320e-01 -7.85608828e-01 -8.79599676e-02
1.17177274e-02 -7.71339983e-02 -1.83546975e-01 1.02758479e+00
7.33297944e-01 -4.03549559e-02 9.59902763e-01 -1.01010215e+00
-1.39971793e-01 4.66571510e-01 -1.00618124e+00 3.10642421e-01
2.01681942e-01 -1.85039714e-01 -1.25974312e-01 -5.59298456e-01
5.12025297e-01 7.65820384e-01 8.32658470e-01 7.28075683e-01
-6.94295347e-01 -5.63885987e-01 -6.23200595e-01 -8.39654207e-02
-1.36988389e+00 -9.96727198e-02 4.16917354e-01 -2.08421126e-01
4.07640606e-01 8.08427572e-01 5.14119744e-01 2.57087976e-01
8.90826344e-01 6.70855045e-02 9.79392707e-01 -3.04369688e-01
6.46914616e-02 -1.07488424e-01 -9.89893302e-02 6.87517285e-01
2.35256702e-01 -7.95736760e-02 -2.45514847e-02 -4.72682774e-01
1.04567432e+00 2.12155029e-01 -4.37109202e-01 -2.28560045e-01
-8.16021144e-01 5.82418025e-01 5.55269957e-01 9.83532369e-01
-1.82456121e-01 1.87892377e-01 4.28812146e-01 -5.15477002e-01
-4.48878333e-02 -3.21315676e-01 2.35139057e-01 2.06539288e-01
-7.88569748e-01 -1.91858411e-01 2.20870048e-01 7.15223730e-01
3.57129425e-02 -4.58219230e-01 -1.99366003e-01 4.66132939e-01
6.17680013e-01 7.00858057e-01 1.06315005e+00 -7.01873362e-01
1.67812183e-01 6.82499290e-01 -2.39525393e-01 -9.82780993e-01
-7.29197204e-01 -2.90683627e-01 -1.19743454e+00 -1.47553399e-01
4.36978281e-01 5.88216841e-01 -9.75567937e-01 1.07958317e+00
6.89581752e-01 2.26490036e-01 -1.41962171e-01 8.94852638e-01
1.08797371e+00 3.04551184e-01 3.26061636e-01 -4.45413053e-01
1.62764072e+00 -5.92000246e-01 -9.60272253e-01 2.63188779e-01
9.14568961e-01 -1.01363087e+00 5.40252686e-01 1.06902406e-01
-1.25076127e+00 -1.00256518e-01 -7.96354830e-01 -1.26484418e-02
1.24809913e-01 -2.29130313e-02 6.45495117e-01 1.07260656e+00
-1.04979146e+00 4.84976232e-01 -1.39658964e+00 -6.17028713e-01
3.04517925e-01 9.11885798e-01 -5.70465863e-01 -7.65880793e-02
-7.15779126e-01 1.23011959e+00 2.68107653e-01 2.54892260e-01
-5.77966213e-01 -6.28376245e-01 -7.13532627e-01 -4.33261067e-01
-1.10012859e-01 -7.24516869e-01 1.15327513e+00 -5.53446889e-01
-1.48004627e+00 1.05563116e+00 -3.88067484e-01 8.40635523e-02
4.54161137e-01 7.00521111e-01 -3.29793423e-01 5.66899717e-01
3.02807122e-01 4.15934831e-01 6.94996938e-02 -1.39417899e+00
-4.74185139e-01 -9.05172110e-01 -6.48049593e-01 5.37005424e-01
2.03306768e-02 1.02402240e-01 -7.23479092e-01 -2.50752240e-01
8.82947683e-01 -8.86639237e-01 -6.10014200e-01 5.30031100e-02
7.28747174e-02 1.59388259e-02 5.43753862e-01 -7.99162984e-01
1.07105947e+00 -1.80418980e+00 -1.79239318e-01 6.66483283e-01
-2.16331775e-03 -1.76095098e-01 3.81789804e-01 -4.43496555e-02
-1.16412945e-01 1.60060257e-01 -3.35716605e-01 -7.07836971e-02
-5.08187890e-01 3.69354069e-01 3.87689739e-01 1.16294646e+00
-7.89400816e-01 9.20944333e-01 -6.69167876e-01 -1.26994836e+00
3.80334705e-02 6.03913248e-01 -2.09836699e-02 -1.71581641e-01
4.58987296e-01 8.87350917e-01 -6.55783176e-01 7.18034565e-01
1.07788682e+00 -1.94279417e-01 3.57611239e-01 -3.88111830e-01
1.92438066e-01 -4.00555134e-01 -8.50204051e-01 1.83129585e+00
-1.20042041e-01 -4.36252169e-02 2.01981664e-01 -6.42521560e-01
7.60849655e-01 5.86308300e-01 9.76581395e-01 -9.92852867e-01
5.77104509e-01 3.74796897e-01 -7.26380199e-03 -6.42398179e-01
3.05135280e-01 -5.67464352e-01 3.02672505e-01 3.12934995e-01
-6.32691801e-01 -5.91262341e-01 -3.25315446e-01 7.20969886e-02
8.20315421e-01 -5.40735759e-02 3.68292183e-01 -4.34609413e-01
9.36508656e-01 2.55043149e-01 4.68698025e-01 1.67934939e-01
-1.53808728e-01 8.67646873e-01 2.15172432e-02 -1.85491756e-01
-7.52048850e-01 -9.75710511e-01 -7.52067804e-01 2.78171897e-01
5.83343267e-01 1.94399610e-01 -7.52873123e-01 -4.02302086e-01
-1.80256814e-01 4.01815742e-01 -6.69430137e-01 1.23762533e-01
-9.83753443e-01 -9.66762066e-01 7.43649065e-01 4.31413829e-01
2.61616111e-01 -6.77800000e-01 -4.54842269e-01 8.67731646e-02
-6.36548340e-01 -8.12110126e-01 -5.07773340e-01 -2.39502460e-01
-1.56779599e+00 -1.09181392e+00 -1.15686691e+00 -8.60488474e-01
1.11064446e+00 2.79804498e-01 9.55873132e-01 8.23329151e-01
-2.55708188e-01 1.88110426e-01 -4.98058386e-02 -1.34719595e-01
-7.21058309e-01 -3.90565604e-01 -2.76304573e-01 -3.43691617e-01
1.10978298e-01 -2.93099344e-01 -9.43176985e-01 6.01150751e-01
-1.13986337e+00 5.56855230e-03 4.91408050e-01 4.68003005e-01
1.00459802e+00 1.68307528e-01 -3.77721451e-02 -8.56480420e-01
2.66860515e-01 -4.17477757e-01 -2.64273137e-01 3.13815594e-01
-3.11527401e-01 -3.58469725e-01 -1.01616392e-02 -1.03347741e-01
-1.16580534e+00 -1.57223910e-01 -1.68477058e-01 2.66626894e-01
-6.46978468e-02 2.96770424e-01 -2.84351539e-02 -6.05148435e-01
7.53050208e-01 2.72767633e-01 7.42090791e-02 -1.14453092e-01
-2.18768910e-01 6.74393415e-01 8.49959850e-01 -6.32134616e-01
3.67208302e-01 1.00075579e+00 8.13852608e-01 -6.44201040e-01
-1.67998731e-01 -6.67594552e-01 -6.26565754e-01 -2.82807827e-01
9.64571476e-01 -5.34159720e-01 -1.04857790e+00 -2.30878796e-02
-1.09751081e+00 2.27305181e-02 9.19778794e-02 7.58183300e-01
-4.80611473e-01 7.82138467e-01 -9.47664976e-01 -5.28867006e-01
-6.29603982e-01 -1.27981079e+00 1.14004731e+00 5.33697903e-01
-7.46834874e-02 -1.18330026e+00 2.32831925e-01 5.25458813e-01
4.69671905e-01 5.09885550e-01 8.57619107e-01 -1.80776894e-01
-1.06198855e-01 -1.68869957e-01 -8.66497457e-02 -6.17776453e-01
1.41162843e-01 -1.73329607e-01 -7.64719903e-01 -1.24822274e-01
4.59943235e-01 3.05832744e-01 2.77150631e-01 8.45784426e-01
1.18680942e+00 3.62743288e-01 -9.81922448e-01 6.78714752e-01
1.73911202e+00 7.72294581e-01 1.01217377e+00 4.17321950e-01
5.37021875e-01 6.30303204e-01 7.67470658e-01 9.37155336e-02
3.87370169e-01 5.35684943e-01 5.92869580e-01 -3.79571170e-01
-4.90218028e-02 7.26850284e-03 -3.31494600e-01 6.35593235e-01
-3.80995601e-01 -1.18406996e-01 -1.32558346e+00 3.46118957e-01
-1.40814757e+00 -5.99754512e-01 -7.68847167e-01 2.04413700e+00
8.73166144e-01 -1.91371575e-01 -3.61325383e-01 1.25404537e-01
8.27610075e-01 -8.45675409e-01 -5.83385490e-02 -2.52643168e-01
2.23496601e-01 4.54359084e-01 8.27279449e-01 7.39329219e-01
-4.20837402e-01 4.04431522e-01 6.13089609e+00 1.24887490e+00
-1.18075395e+00 5.55981934e-01 7.76987731e-01 4.58033442e-01
-5.74313283e-01 9.62816328e-02 -7.86244690e-01 3.82437706e-01
6.73958421e-01 7.23080933e-02 -3.37096125e-01 1.85321152e-01
2.42790744e-01 -8.89206767e-01 -1.01608455e+00 9.99227166e-01
1.28029317e-01 -1.27527380e+00 6.86052293e-02 2.85283297e-01
4.84088242e-01 -4.62160051e-01 -9.27801281e-02 -2.37678394e-01
-2.95475513e-01 -1.60706508e+00 9.86192524e-02 7.32959092e-01
8.60399365e-01 -7.59632111e-01 1.25041997e+00 5.57077289e-01
-1.18203604e+00 6.38230860e-01 -5.18612266e-01 4.64279443e-01
1.55141398e-01 2.71356344e-01 -1.02513874e+00 1.07501745e+00
4.02780324e-01 -1.14249941e-02 -5.32583535e-01 1.19667196e+00
4.73703980e-01 7.15102255e-02 -4.01456237e-01 1.07904352e-01
-7.11612552e-02 -4.50073004e-01 -7.65727758e-02 9.01163816e-01
6.77738428e-01 8.27423871e-01 -2.24460103e-02 3.73214751e-01
5.24570346e-01 5.40701270e-01 -4.90625471e-01 4.18219060e-01
3.13137650e-01 1.21783948e+00 -1.56749403e+00 -6.83799386e-02
-1.17084019e-01 6.31579399e-01 -2.84768343e-01 -1.43753245e-01
-8.66401970e-01 2.58288026e-01 -5.90311587e-01 4.23642874e-01
-5.74232578e-01 3.84208076e-02 -7.47064292e-01 -4.54729408e-01
-4.73941982e-01 -5.28609097e-01 6.16042137e-01 -7.07154453e-01
-9.26478565e-01 6.77476287e-01 9.65370312e-02 -1.30535066e+00
2.38762140e-01 -3.43571931e-01 -6.37490630e-01 1.01895869e+00
-1.20214009e+00 -1.39664459e+00 -5.64672828e-01 6.76562846e-01
3.57105881e-02 2.24683106e-01 9.67649400e-01 3.40981960e-01
-1.12631671e-01 5.34788728e-01 -9.80221406e-02 7.13790487e-03
5.53357959e-01 -9.90861177e-01 -5.56523383e-01 6.39982596e-02
-5.89460075e-01 5.42609215e-01 5.75865686e-01 -1.00346446e+00
-1.29219651e+00 -5.44005215e-01 8.02513063e-01 -5.16973794e-01
1.23096652e-01 3.85410875e-01 -6.21857941e-01 5.91222048e-01
2.76802450e-01 9.15189683e-02 9.35193598e-01 -7.71514654e-01
7.34201789e-01 7.74676632e-03 -1.73621416e+00 5.02963901e-01
4.36182320e-01 5.43094613e-02 -2.86616147e-01 4.59846884e-01
1.81504950e-01 -1.13967288e+00 -1.26713419e+00 7.67186999e-01
4.37985569e-01 -7.75571704e-01 1.08525181e+00 -1.51434331e-03
1.06989071e-01 -4.96343881e-01 -6.78950772e-02 -5.69870174e-01
-2.25491617e-02 -2.09270064e-02 6.45728827e-01 1.05401552e+00
1.95450619e-01 -3.88643056e-01 1.33728135e+00 8.70768249e-01
-1.61772355e-01 -6.22848868e-01 -1.21997070e+00 -3.42669904e-01
3.93271655e-01 -2.12197214e-01 6.84665263e-01 9.88627672e-01
3.27239484e-02 -3.71795535e-01 3.85445833e-01 4.58398610e-01
6.85892344e-01 -1.16486616e-01 4.78152901e-01 -8.53226364e-01
-2.09438443e-01 -4.35627759e-01 -5.38009405e-01 -6.16029382e-01
-2.46687189e-01 -1.12670290e+00 -2.10056931e-01 -1.82052469e+00
8.11157346e-01 -7.65993178e-01 2.17075478e-02 6.27856031e-02
6.51821792e-02 4.81797785e-01 -1.47490159e-01 3.05809200e-01
-3.71548720e-02 3.37773979e-01 2.08687210e+00 -8.18983167e-02
3.90963703e-02 8.78683999e-02 -2.92109281e-01 9.11276877e-01
6.01631939e-01 -8.23852956e-01 -2.83357501e-01 -1.66859463e-01
5.02227731e-02 9.49323416e-01 2.72476040e-02 -7.52868950e-01
5.80456913e-01 -4.70586896e-01 7.10942686e-01 -1.05311608e+00
2.77837157e-01 -1.31116557e+00 7.47072101e-01 1.04719234e+00
2.01516107e-01 3.12488377e-01 2.55165756e-01 1.71007141e-01
-1.47074074e-01 -7.48127818e-01 9.15761590e-01 -3.98125350e-01
5.59154190e-02 3.28682214e-01 -3.48600149e-01 -3.91865134e-01
1.29031789e+00 -6.67948246e-01 -7.91695789e-02 -2.73061782e-01
-8.41600418e-01 3.04026991e-01 5.04269302e-01 -5.11573672e-01
7.02745140e-01 -1.60042489e+00 -6.04692161e-01 -1.85525849e-01
-2.59542465e-01 4.63635802e-01 4.60122764e-01 1.50905299e+00
-1.45735955e+00 4.99181122e-01 -3.02334160e-01 -1.01959991e+00
-1.52264273e+00 3.53409916e-01 8.36344242e-01 -7.12364554e-01
-4.67837393e-01 7.59475708e-01 3.87567192e-01 -6.12096131e-01
-7.17736706e-02 -2.48907208e-01 -6.46001816e-01 -4.77691293e-01
2.15909004e-01 3.96425575e-01 3.82987827e-01 -1.01057172e+00
-5.03046274e-01 1.06871176e+00 1.36805654e-01 -3.45078975e-01
7.52661586e-01 -4.19410944e-01 -5.71918488e-01 -1.02837913e-01
9.18016434e-01 3.01980853e-01 -3.98221433e-01 5.26496954e-02
8.17009881e-02 -5.45154095e-01 9.69386026e-02 -5.68978488e-01
-1.07862532e+00 7.05520093e-01 9.09889996e-01 -3.69332522e-01
1.21896958e+00 -6.55925795e-02 6.96275115e-01 -2.28459954e-01
5.00681341e-01 -6.27459884e-01 -1.66447371e-01 -8.56287107e-02
6.88835382e-01 -1.08270037e+00 2.97777414e-01 -8.38045716e-01
-5.35992980e-01 1.33775568e+00 5.62868357e-01 9.25003961e-02
7.09566772e-01 5.91264009e-01 2.49972373e-01 -3.11137497e-01
-3.25158518e-03 2.77913392e-01 1.76401988e-01 5.47074795e-01
7.86716521e-01 1.62725039e-02 -9.36144114e-01 5.75501978e-01
-1.93858311e-01 -3.06839440e-02 5.19153535e-01 1.14829266e+00
-4.64739025e-01 -1.32048225e+00 -1.11922097e+00 -3.54753505e-03
-7.82100737e-01 1.08243629e-01 -3.59105393e-02 9.56286550e-01
2.22402960e-01 8.89631569e-01 1.00047752e-01 3.73111397e-01
1.47619382e-01 -4.28017765e-01 9.41499472e-01 -4.79024202e-01
-7.36669838e-01 4.27742660e-01 -3.05031270e-01 -3.63909215e-01
-4.33232486e-01 -4.44166988e-01 -2.02415824e+00 -5.04717171e-01
-4.03111994e-01 4.07671571e-01 1.02661908e+00 9.82691228e-01
-5.63987792e-01 4.25330639e-01 4.17831689e-01 -4.97081399e-01
-1.09972201e-01 -7.83377171e-01 -7.61693418e-01 2.71782756e-01
-7.61928037e-02 -4.14845794e-01 8.28980431e-02 -5.00129238e-02] | [14.223434448242188, -2.7155704498291016] |
04527a5d-de6f-4632-82c7-d8ace046d6aa | geolocation-of-cultural-heritage-using-multi | 2209.03638 | null | https://arxiv.org/abs/2209.03638v1 | https://arxiv.org/pdf/2209.03638v1.pdf | Geolocation of Cultural Heritage using Multi-View Knowledge Graph Embedding | Knowledge Graphs (KGs) have proven to be a reliable way of structuring data. They can provide a rich source of contextual information about cultural heritage collections. However, cultural heritage KGs are far from being complete. They are often missing important attributes such as geographical location, especially for sculptures and mobile or indoor entities such as paintings. In this paper, we first present a framework for ingesting knowledge about tangible cultural heritage entities from various data sources and their connected multi-hop knowledge into a geolocalized KG. Secondly, we propose a multi-view learning model for estimating the relative distance between a given pair of cultural heritage entities, based on the geographical as well as the knowledge connections of the entities. | ['Marcello Pelillo', 'Alessio Del Bue', 'Diego Pilutti', 'Stuart James', 'Feliks Hibraj', 'Sebastiano Vascon', 'Hebatallah A. Mohamed'] | 2022-09-08 | null | null | null | null | ['multi-view-learning', 'knowledge-graph-embedding'] | ['computer-vision', 'graphs'] | [-5.39287031e-01 1.50415093e-01 -3.77959371e-01 -3.84111315e-01
-5.91833055e-01 -7.32391834e-01 8.07005882e-01 1.01777244e+00
2.35736165e-02 9.54343319e-01 6.65124416e-01 1.80649966e-01
-7.52270162e-01 -1.64484882e+00 -6.33743882e-01 -4.51785445e-01
-5.23369431e-01 5.78931153e-01 4.95328009e-01 -3.21014702e-01
8.35125595e-02 6.18285835e-01 -1.41407454e+00 3.34968060e-01
6.48184299e-01 6.94590867e-01 1.88534841e-01 1.57305613e-01
-4.57703203e-01 8.72411191e-01 -4.74347591e-01 -7.16081977e-01
-3.08289707e-01 1.42680138e-01 -1.14836287e+00 -2.28871301e-01
4.94033188e-01 -2.17027545e-01 -2.42318705e-01 8.14490020e-01
2.48240858e-01 3.10555212e-02 6.46717489e-01 -1.30647123e+00
-5.95972657e-01 9.51832175e-01 -3.67597580e-01 -1.73022240e-01
6.23374462e-01 -8.38106453e-01 1.24874830e+00 -7.72338271e-01
1.12121856e+00 1.09260440e+00 7.03884244e-01 -1.36539429e-01
-5.22523522e-01 -1.25334337e-01 2.85637558e-01 4.88554716e-01
-1.72989511e+00 -1.99942708e-01 8.47106576e-01 -3.33589554e-01
3.74011070e-01 1.47580281e-01 8.52524102e-01 5.74422419e-01
-3.23245138e-01 8.02639484e-01 6.30705774e-01 -4.57030952e-01
1.28553689e-01 1.97852567e-01 2.72010770e-02 8.92867506e-01
5.61732829e-01 -7.64237702e-01 -1.10581315e+00 -3.49595755e-01
1.00429571e+00 4.26834635e-03 -2.37476900e-01 -6.93058014e-01
-1.17662430e+00 6.13835037e-01 7.48717785e-01 4.88943756e-01
-2.24666595e-01 8.25659856e-02 1.96428224e-01 -1.35162383e-01
3.93295646e-01 2.13086940e-02 -3.13780963e-01 6.92250207e-03
-6.36959076e-01 -1.88662708e-01 8.54463756e-01 1.19857597e+00
9.72843945e-01 -3.02801937e-01 7.62231350e-01 7.88510501e-01
5.06319404e-01 1.96938097e-01 -1.61029056e-01 -6.59926534e-01
5.19419134e-01 9.88027990e-01 1.45694494e-01 -1.47285140e+00
-4.67537552e-01 -8.26407745e-02 -6.44854307e-01 -2.08776057e-01
2.28912011e-01 1.36730164e-01 -4.14307833e-01 1.29086423e+00
5.91395676e-01 1.99714482e-01 1.74378455e-01 5.24152160e-01
1.35256386e+00 3.86358976e-01 6.73727132e-04 2.06633091e-01
1.06029427e+00 -6.67463779e-01 -5.07522225e-01 2.09976926e-01
5.87649822e-01 -4.00874764e-01 5.34307063e-01 2.19567139e-02
-7.99186349e-01 -1.52571976e-01 -6.56967878e-01 -2.31841311e-01
-1.12228537e+00 6.63398728e-02 1.03306055e+00 4.93691057e-01
-1.00906384e+00 5.55371583e-01 -8.77715468e-01 -6.63931608e-01
4.54635710e-01 8.99537504e-02 -7.49460101e-01 -3.03968370e-01
-1.05321002e+00 7.09670842e-01 6.62732899e-01 -1.47627247e-02
-7.10066080e-01 -7.96378732e-01 -9.88154590e-01 -1.77473463e-02
4.43434358e-01 -5.61386466e-01 2.36176968e-01 -2.71689624e-01
-7.27908313e-01 7.35793471e-01 -5.68253584e-02 9.22758505e-02
2.87378699e-01 -4.04218405e-01 -1.05905688e+00 9.76057127e-02
4.17649001e-01 2.88644880e-01 1.24255717e-01 -1.71108425e+00
-9.79633331e-01 -6.45401835e-01 3.67995858e-01 4.08024132e-01
-5.44703066e-01 -4.03573841e-01 -7.92020500e-01 -4.85939920e-01
4.62529808e-01 -4.26704615e-01 7.71384388e-02 1.33974254e-01
-7.04807222e-01 -8.77376348e-02 1.02962530e+00 -7.63206661e-01
1.35058713e+00 -2.10291314e+00 2.86592036e-01 6.72490537e-01
2.02017978e-01 -2.13861004e-01 2.03087047e-01 9.02896821e-01
5.04276514e-01 2.23790511e-01 8.65694582e-02 9.61551219e-02
-2.31899083e-01 4.61543649e-01 1.23751275e-02 2.20903412e-01
-3.67998540e-01 5.36962748e-01 -1.47990453e+00 -6.20629132e-01
7.57868364e-02 7.11826265e-01 -4.99255583e-02 -4.35474068e-01
-8.92248452e-02 3.43272030e-01 -9.50287044e-01 1.12472010e+00
2.72244394e-01 -3.31098378e-01 5.85294425e-01 -2.29338199e-01
7.83551559e-02 -1.65000316e-02 -1.47337699e+00 1.92446709e+00
-3.59204113e-01 6.84703231e-01 -8.87621939e-02 -4.28367764e-01
9.85859334e-01 6.37304783e-02 7.32162118e-01 1.27791066e-03
-4.61254925e-01 1.28412649e-01 -9.97557163e-01 -4.87611532e-01
9.38214064e-01 7.85486773e-02 -1.73822597e-01 3.13964218e-01
2.05111504e-02 1.16054766e-01 1.99557260e-01 6.34126246e-01
8.32247674e-01 3.90523165e-01 4.82077777e-01 -1.58689797e-01
5.51241934e-01 9.81041044e-02 3.91591489e-01 2.83405513e-01
4.26859200e-01 3.60492170e-01 4.18544888e-01 -6.85758233e-01
-7.99283326e-01 -1.15256381e+00 -6.75909072e-02 8.29459906e-01
4.93112773e-01 -8.16959441e-01 -1.04706392e-01 -8.56484532e-01
3.16090077e-01 4.29011285e-01 -4.78987098e-01 2.91081786e-01
-2.90358186e-01 -4.94305670e-01 6.86868787e-01 7.09516287e-01
5.36924422e-01 -3.49099517e-01 -1.28682166e-01 4.05807234e-03
-5.09059548e-01 -1.09137416e+00 1.02099203e-01 -2.92094469e-01
-7.92761564e-01 -1.43119240e+00 -4.16572213e-01 -7.54699647e-01
8.82822454e-01 1.76532313e-01 1.17833436e+00 2.63708204e-01
5.51353060e-02 7.53844380e-01 -5.45544624e-01 -1.97917193e-01
2.21049390e-03 2.97613531e-01 2.50399429e-02 1.21744506e-01
3.69611323e-01 -7.22450018e-01 -3.21596652e-01 4.29527164e-01
-8.40178609e-01 -1.70906380e-01 1.86584786e-01 1.93064779e-01
9.74335253e-01 5.25645792e-01 5.71251929e-01 -1.08674896e+00
1.95997179e-01 -9.99608457e-01 -3.08140099e-01 9.34623241e-01
-2.20528692e-01 -6.67904839e-02 2.87028372e-01 1.56313851e-01
-1.10799611e+00 9.19789658e-04 2.74074048e-01 1.91476136e-01
-2.85482615e-01 1.06000626e+00 -6.14120483e-01 -2.60677934e-01
2.11157247e-01 -5.11292145e-02 -6.85037851e-01 -7.35034168e-01
5.95798731e-01 4.35824633e-01 4.62165058e-01 -8.42136085e-01
5.57362199e-01 1.02281141e+00 2.57893711e-01 -1.08798683e+00
-8.07636380e-01 -8.09415996e-01 -1.34885061e+00 -5.45365095e-01
6.50488794e-01 -1.04261172e+00 -4.56049085e-01 3.48020673e-01
-6.61610842e-01 2.04618305e-01 -1.26825064e-01 4.09817129e-01
-3.18760514e-01 4.32304442e-01 -2.70945609e-01 -4.88214403e-01
2.94643730e-01 -1.74152687e-01 1.12034702e+00 1.94654644e-01
4.86792773e-02 -1.59535658e+00 1.01427875e-01 3.68296266e-01
6.28960282e-02 9.05337393e-01 1.17127252e+00 -4.83286858e-01
-7.74301648e-01 -1.09713912e-01 1.14820730e-02 -3.14026624e-01
5.87788045e-01 4.04913947e-02 -5.95130146e-01 -7.49148205e-02
-8.51980925e-01 2.24413350e-03 6.17727339e-01 3.60506140e-02
5.80568075e-01 -1.88241303e-01 -9.88990724e-01 5.07097304e-01
1.84780276e+00 -1.04765274e-01 3.95113409e-01 6.31497383e-01
1.14647210e+00 6.94450021e-01 4.90202308e-01 5.84681273e-01
9.97551858e-01 5.06642044e-01 5.35925388e-01 1.33541554e-01
-1.26174867e-01 -7.55961895e-01 -4.61355969e-02 8.61279845e-01
-6.53281927e-01 -2.98399210e-01 -1.27371883e+00 1.03498590e+00
-1.86200202e+00 -9.70841825e-01 -5.24151444e-01 2.19201303e+00
4.54277456e-01 -3.90808254e-01 -5.45018464e-02 -6.34736568e-02
6.77486300e-01 2.13535592e-01 -4.50119078e-01 4.49873000e-01
-5.45091987e-01 -4.31874335e-01 6.27129614e-01 3.54805738e-01
-9.98291850e-01 1.05857468e+00 5.89223337e+00 5.25496840e-01
-2.83653498e-01 2.23837905e-02 -8.97932947e-02 3.46949063e-02
-7.95088112e-01 2.59818375e-01 -9.80248570e-01 6.45699352e-02
3.64664763e-01 -1.47709042e-01 -7.44037097e-03 9.57084835e-01
-2.76991725e-01 -4.21354681e-01 -8.79469454e-01 6.05473638e-01
1.78756431e-01 -1.62588954e+00 4.44940835e-01 1.33241907e-01
7.99029052e-01 -2.17396524e-02 -4.86599624e-01 -2.31355682e-01
7.65181541e-01 -7.32573628e-01 6.03954077e-01 9.68248248e-01
6.03955030e-01 -9.44685936e-01 6.59713566e-01 1.84983000e-01
-1.71346474e+00 1.05307177e-01 -3.43397021e-01 2.99085230e-01
5.05792201e-01 6.15670383e-01 -5.19423366e-01 1.28798068e+00
1.06186116e+00 1.18727481e+00 -6.61436856e-01 1.23780119e+00
-5.22091508e-01 1.42036244e-01 -5.09908974e-01 1.63178653e-01
2.05362514e-01 -2.61664867e-01 3.14358115e-01 7.25507438e-01
6.26604855e-01 1.57911703e-01 1.54635534e-04 2.48908684e-01
-3.11406672e-01 1.66189194e-01 -8.26002717e-01 1.52821317e-01
8.34421277e-01 1.15208650e+00 -9.80129957e-01 -1.57350570e-01
-5.05977273e-01 5.20979822e-01 4.64549273e-01 3.39430451e-01
-3.49487782e-01 -3.73534054e-01 4.90715802e-01 3.12024474e-01
2.61557013e-01 -5.90464532e-01 1.56927973e-01 -1.19085491e+00
3.99307795e-02 -3.99684273e-02 7.46802390e-01 -1.01978064e+00
-1.17663515e+00 2.19352677e-01 -7.58897662e-02 -1.19281256e+00
1.72831625e-01 -2.88088083e-01 -2.62342483e-01 3.29871088e-01
-1.74656475e+00 -1.82270694e+00 -4.89966244e-01 9.96047139e-01
-4.40844782e-02 -2.06767648e-01 1.00708151e+00 4.10515755e-01
-8.14686045e-02 -2.43547559e-02 5.59826553e-01 4.82332349e-01
5.24756074e-01 -1.29927981e+00 -9.19501018e-03 5.88620663e-01
6.52616203e-01 5.95232368e-01 1.30158871e-01 -1.07509100e+00
-1.29265881e+00 -1.11467648e+00 9.47775900e-01 -6.93966091e-01
7.85679102e-01 5.90234622e-02 -8.76648486e-01 1.07633090e+00
-2.16148600e-01 -3.92885245e-02 1.11812842e+00 7.33850360e-01
-3.29128861e-01 -9.31667462e-02 -9.97660398e-01 3.95274997e-01
1.32112503e+00 -5.70687830e-01 -4.41115022e-01 4.11511421e-01
3.64256948e-01 -2.47373641e-01 -1.61843085e+00 9.47878361e-02
4.75865781e-01 -8.17104161e-01 1.17297673e+00 -3.93331528e-01
2.70964235e-01 -4.67053533e-01 -4.06912208e-01 -1.31080806e+00
-1.04567818e-01 -6.72364607e-02 -2.51392841e-01 1.73316121e+00
4.73059148e-01 -1.94706157e-01 9.94346440e-01 4.19444650e-01
7.21784681e-02 -1.51935026e-01 -7.83504725e-01 -7.99039960e-01
-4.31641996e-01 -2.06638992e-01 1.05615008e+00 1.48747432e+00
1.42507702e-01 8.11570138e-02 -3.90459478e-01 8.42005432e-01
8.84419560e-01 1.42416403e-01 5.78368127e-01 -1.78063691e+00
3.29224139e-01 -1.59987867e-01 -9.59881842e-01 -5.00781298e-01
2.04719063e-02 -7.91547596e-01 -7.24399030e-01 -2.32592344e+00
-1.81066971e-02 -9.63909864e-01 -1.01133756e-01 6.01235092e-01
1.71583161e-01 -1.41257316e-01 -1.95360541e-01 4.19163734e-01
-7.47878492e-01 3.54950368e-01 8.19258213e-01 -2.14392096e-01
-2.09652796e-01 -3.84642482e-02 -3.59572887e-01 9.67015266e-01
6.97499931e-01 -3.67196828e-01 -4.24356163e-01 -8.53759289e-01
9.59886849e-01 6.21510260e-02 3.56227934e-01 -7.94360638e-01
7.89785266e-01 -1.75345406e-01 2.04869866e-01 -8.15690100e-01
3.88246000e-01 -1.12499595e+00 5.64410090e-01 -4.30844873e-02
-1.76370591e-02 -2.08127677e-01 -1.04637276e-02 1.12235975e+00
-7.21449435e-01 -2.81081557e-01 -1.40564531e-01 -4.77555096e-01
-1.47565281e+00 2.23957658e-01 2.29111493e-01 -1.31942898e-01
1.12287378e+00 -3.82619470e-01 -5.29137135e-01 -3.69276941e-01
-1.10158455e+00 6.39939725e-01 7.24924028e-01 5.26142538e-01
7.10867286e-01 -1.35787737e+00 -3.95861268e-01 -1.55499980e-01
3.48843634e-01 1.30176798e-01 3.14199597e-01 2.33615801e-01
-5.60637832e-01 1.86546132e-01 -3.49064946e-01 -2.17207313e-01
-1.41080046e+00 3.06380451e-01 3.03512048e-02 1.42840659e-02
-8.42699349e-01 7.72882581e-01 -5.21319509e-01 -5.08105218e-01
2.72537649e-01 8.47420096e-02 -8.65905404e-01 6.40737832e-01
3.38618070e-01 7.41557062e-01 -1.90798417e-02 -1.02131498e+00
-4.81721938e-01 9.22633767e-01 3.13411653e-01 9.03760716e-02
1.94946897e+00 -4.69546378e-01 -4.70696479e-01 6.90970302e-01
6.90652132e-01 3.72566938e-01 -8.37529838e-01 -5.33681095e-01
6.27453566e-01 -6.50887191e-01 -5.97082786e-02 -6.37559772e-01
-1.17747462e+00 5.83106220e-01 1.40437027e-02 2.70525724e-01
7.91164994e-01 4.97311741e-01 4.33695853e-01 5.20085990e-01
1.19097865e+00 -1.27843964e+00 -2.35589236e-01 1.39983848e-01
5.70372224e-01 -1.01307213e+00 4.61772323e-01 -7.64183581e-01
-6.43473506e-01 1.26912725e+00 4.03722495e-01 5.71243465e-01
1.04088879e+00 3.21089244e-03 -1.58887461e-01 -6.24453366e-01
-2.52793312e-01 -4.69793826e-01 7.86703750e-02 1.09374011e+00
1.18272856e-01 1.90796509e-01 1.18858002e-01 2.37980649e-01
-1.85642734e-01 -1.52265340e-01 6.42003477e-01 1.11097503e+00
-4.94233221e-01 -1.03272045e+00 -2.74416178e-01 3.41743857e-01
-1.83143497e-01 1.41115427e-01 -5.75190187e-01 1.01192117e+00
2.04811558e-01 6.66414618e-01 -1.01577044e-02 -1.87815130e-01
4.23181444e-01 -9.33599249e-02 2.14942202e-01 -4.92475033e-01
-2.08312944e-01 -3.15683931e-01 4.16270822e-01 -4.18836981e-01
-1.05834675e+00 -6.79014981e-01 -1.24817359e+00 -4.56627190e-01
-2.11317331e-01 2.09495902e-01 7.09811151e-01 8.67194057e-01
4.09695119e-01 2.93648064e-01 2.45004147e-01 -4.10547495e-01
7.37731099e-01 -2.80201256e-01 -1.12358224e+00 1.81420416e-01
-6.82663396e-02 -6.59201801e-01 1.22335874e-01 2.14583501e-01] | [8.87756633758545, 7.877692699432373] |
167a92a7-df17-4acd-b9b7-4caf6bd1428f | side-channel-assisted-inference-leakage-from | 2304.01990 | null | https://arxiv.org/abs/2304.01990v1 | https://arxiv.org/pdf/2304.01990v1.pdf | Side Channel-Assisted Inference Leakage from Machine Learning-based ECG Classification | The Electrocardiogram (ECG) measures the electrical cardiac activity generated by the heart to detect abnormal heartbeat and heart attack. However, the irregular occurrence of the abnormalities demands continuous monitoring of heartbeats. Machine learning techniques are leveraged to automate the task to reduce labor work needed during monitoring. In recent years, many companies have launched products with ECG monitoring and irregular heartbeat alert. Among all classification algorithms, the time series-based algorithm dynamic time warping (DTW) is widely adopted to undertake the ECG classification task. Though progress has been achieved, the DTW-based ECG classification also brings a new attacking vector of leaking the patients' diagnosis results. This paper shows that the ECG input samples' labels can be stolen via a side-channel attack, Flush+Reload. In particular, we first identify the vulnerability of DTW for ECG classification, i.e., the correlation between warping path choice and prediction results. Then we implement an attack that leverages Flush+Reload to monitor the warping path selection with known ECG data and then build a predictor for constructing the relation between warping path selection and labels of input ECG samples. Based on experiments, we find that the Flush+Reload-based inference leakage can achieve an 84.0\% attacking success rate to identify the labels of the two samples in DTW. | ['Han Wang', 'Houman Homayoun', 'Chongzhou Fang', 'Ning Miao', 'Jialin Liu'] | 2023-04-04 | null | null | null | null | ['ecg-classification', 'dynamic-time-warping'] | ['medical', 'time-series'] | [ 5.27478635e-01 -3.34756494e-01 -3.40576500e-01 -2.01266974e-01
-5.59411168e-01 -9.07146811e-01 -1.13151230e-01 1.34962961e-01
-4.78781164e-02 3.01509917e-01 -1.73868984e-01 -6.22940421e-01
-3.71010780e-01 -6.19061351e-01 -2.01502275e-02 -7.90072262e-01
-5.43708920e-01 -1.40451640e-01 -1.25615867e-02 7.79176131e-02
4.12261456e-01 5.06896913e-01 -6.40591800e-01 4.06077594e-01
4.42170292e-01 1.04106617e+00 -5.88447034e-01 8.18704724e-01
3.86193603e-01 4.36394274e-01 -9.40037489e-01 -2.74783015e-01
5.64440489e-01 -8.57450724e-01 -5.87677240e-01 -5.24675369e-01
-3.81013155e-01 -1.57654077e-01 -1.31391525e-01 8.24667394e-01
1.02009845e+00 -5.08176029e-01 1.97433978e-01 -1.39611530e+00
3.72433245e-01 6.88057959e-01 -6.20957434e-01 7.63097882e-01
4.34016526e-01 3.51477712e-01 7.20487356e-01 -4.97002661e-01
4.88245457e-01 5.12502134e-01 1.08947265e+00 4.22239184e-01
-1.33965671e+00 -1.03549635e+00 -7.32417226e-01 2.86807060e-01
-1.63005078e+00 2.11152397e-02 1.34497011e+00 -2.67206877e-01
4.67483640e-01 6.14154220e-01 8.70546877e-01 1.18734872e+00
8.49752665e-01 2.93329686e-01 1.20083177e+00 -3.02791506e-01
1.09868180e-02 -5.67526147e-02 3.06797355e-01 4.56476510e-01
6.39566481e-02 4.51101661e-01 -5.88733375e-01 -7.96571195e-01
5.09870589e-01 1.31224483e-01 -4.62660670e-01 2.42306888e-01
-1.30943036e+00 6.89143836e-01 5.28026819e-02 1.29096955e-01
-1.98757321e-01 -1.88629761e-01 7.68772542e-01 6.84200346e-01
8.20210278e-02 8.47430706e-01 -5.97340703e-01 -3.27072412e-01
-9.74383473e-01 -7.84181952e-02 7.05971062e-01 2.95360565e-01
2.31162593e-01 5.91198504e-02 -1.38099715e-01 2.34715313e-01
-6.47896752e-02 4.98074770e-01 5.36487103e-01 -5.63106954e-01
3.32350343e-01 3.57381731e-01 -3.12764883e-01 -1.41817546e+00
-5.80195010e-01 -4.53706443e-01 -8.49567533e-01 2.90968288e-02
3.88973624e-01 -5.52550554e-01 -4.76777077e-01 1.37802207e+00
2.57741302e-01 5.38669109e-01 3.67415473e-02 4.92093891e-01
2.25738093e-01 4.78439659e-01 -1.11693600e-02 -5.50902724e-01
1.36207855e+00 -1.28599867e-01 -8.41277361e-01 2.31626347e-01
8.72486770e-01 -5.96050620e-01 6.91142738e-01 6.26743376e-01
-5.69680274e-01 -4.90093827e-01 -1.49752438e+00 6.84194922e-01
1.08123690e-01 -3.49751472e-01 4.01229322e-01 1.17609656e+00
-4.29481447e-01 9.02735114e-01 -8.88668835e-01 1.12828150e-01
4.02628988e-01 4.10628021e-01 -7.13210702e-02 2.66427249e-01
-1.52679312e+00 7.50901103e-01 1.53301865e-01 4.02472056e-02
-6.89393461e-01 -9.56484079e-01 -4.25779819e-01 -3.73567730e-01
7.56860292e-03 -2.36425430e-01 5.04506052e-01 -7.07283616e-01
-1.18033564e+00 6.27922475e-01 2.84715027e-01 -7.12017119e-01
5.64556599e-01 1.21948272e-01 -1.03472853e+00 8.57725665e-02
-1.73793688e-01 -3.80213886e-01 1.21266246e+00 -5.86582482e-01
-3.69036168e-01 -3.34088653e-01 -4.95802611e-01 -4.06075507e-01
-2.97940373e-01 1.51605800e-01 3.23483676e-01 -8.39155197e-01
1.95800275e-01 -9.83858347e-01 -7.52456039e-02 -4.34675634e-01
-7.55595922e-01 3.63583297e-01 1.10087037e+00 -6.38050914e-01
1.86173606e+00 -2.54328156e+00 -3.04446757e-01 8.17167878e-01
2.70213604e-01 3.99777681e-01 2.04422101e-01 4.86999184e-01
-3.70692730e-01 2.94138372e-01 -3.01100761e-01 3.38982910e-01
-5.88029325e-01 1.19275399e-01 -8.68436873e-01 8.38654995e-01
5.11681177e-02 8.31785560e-01 -6.71031833e-01 -2.58481145e-01
-6.27313703e-02 3.91423136e-01 -2.30144128e-01 2.92183071e-01
5.78458369e-01 6.41563177e-01 -4.12735730e-01 5.84537327e-01
3.64852428e-01 5.71065806e-02 5.13033569e-01 -5.17349780e-01
1.03331856e-01 1.64877132e-01 -1.13109469e+00 1.41381609e+00
1.43562883e-01 5.34725189e-01 -6.41586602e-01 -9.15988326e-01
1.35068846e+00 6.74661636e-01 6.69437349e-01 -3.72712374e-01
2.05570385e-01 2.63101131e-01 4.88799095e-01 -7.60053694e-01
-2.48741686e-01 -1.69828624e-01 -2.20003739e-01 8.36618900e-01
-3.48340482e-01 1.11045457e-01 -6.27376854e-01 -1.65008664e-01
1.59607267e+00 -6.69176728e-02 4.95330423e-01 -3.96562994e-01
3.15809518e-01 -1.99124262e-01 8.77810359e-01 6.26651824e-01
-4.63589102e-01 5.15626371e-01 5.91583550e-01 -1.03338194e+00
-5.58617413e-01 -7.97779441e-01 -3.03323060e-01 3.00603360e-01
3.19801494e-02 -4.32018965e-01 -6.96306050e-01 -9.15719390e-01
-5.23161888e-02 3.79750848e-01 -4.47198510e-01 -6.87054634e-01
-8.19949627e-01 -1.06759143e+00 1.39405119e+00 5.19319236e-01
4.43644315e-01 -8.95633340e-01 -1.34960115e+00 3.84081006e-01
-2.81422406e-01 -8.00861061e-01 -4.57945973e-01 1.74505383e-01
-1.01870322e+00 -1.19514418e+00 -6.76285103e-02 -1.75439686e-01
3.90765011e-01 -2.93418556e-01 7.39677727e-01 2.40335301e-01
-9.20463860e-01 9.87974629e-02 -2.35237956e-01 -6.98451161e-01
-4.17754322e-01 -5.34588695e-02 7.38618225e-02 4.43041742e-01
2.59397030e-01 -6.67698860e-01 -8.13579559e-01 5.50693810e-01
-6.03999913e-01 -4.76443440e-01 1.87351238e-02 6.88893378e-01
3.30683827e-01 3.90604526e-01 8.11939299e-01 -9.38742816e-01
6.53672755e-01 -5.40172517e-01 -2.44197682e-01 7.27801248e-02
-1.11917222e+00 -2.02295542e-01 4.06090021e-01 -8.00268590e-01
-2.73925453e-01 -2.89689153e-02 -2.09314555e-01 -3.41156423e-01
1.88404128e-01 3.98060262e-01 -1.45384535e-01 -1.21203959e-01
7.43518114e-01 4.87331487e-02 1.14486240e-01 -1.47341684e-01
-3.31653476e-01 5.89186549e-01 4.18492973e-01 -2.39963204e-01
8.51440012e-01 5.05820334e-01 1.17158353e-01 -4.14287001e-01
-4.88641828e-01 -6.40310124e-02 -4.10115153e-01 -3.21934968e-01
8.16367924e-01 -4.16303664e-01 -8.00576091e-01 4.18252647e-01
-1.00780737e+00 1.28644230e-02 -2.61212468e-01 4.01549250e-01
1.67222656e-02 3.13705593e-01 -4.69917446e-01 -9.48626637e-01
-7.96240151e-01 -7.78708875e-01 3.30750167e-01 -2.52460629e-01
-9.20546412e-01 -8.48226190e-01 1.40169084e-01 1.89338606e-02
2.63844132e-01 1.06629777e+00 1.14010966e+00 -1.05476022e+00
-1.85236886e-01 -7.06294596e-01 5.30634940e-01 2.85742104e-01
9.33259800e-02 9.04799104e-02 -1.04504001e+00 -3.89046490e-01
7.33226717e-01 3.55902135e-01 1.39734820e-01 2.95374334e-01
1.09566414e+00 -2.48917535e-01 -4.17870432e-01 8.79769981e-01
1.15660584e+00 8.09615552e-01 8.65792990e-01 1.70642778e-01
5.69718063e-01 3.76442850e-01 7.08371818e-01 5.61561584e-01
-2.06087351e-01 5.38623691e-01 2.20523760e-01 -2.79802471e-01
4.18212235e-01 -1.22553326e-01 3.31990242e-01 8.53094161e-01
-2.67948419e-01 1.40192643e-01 -1.16817546e+00 1.05986983e-01
-1.32688582e+00 -9.94504035e-01 -2.13807419e-01 2.43125319e+00
9.48100150e-01 5.02744853e-01 1.07377842e-01 1.08281291e+00
7.10123181e-01 7.65270442e-02 -6.06049240e-01 -4.85776126e-01
1.03125624e-01 5.31988919e-01 4.79700863e-01 1.83682114e-01
-9.58893239e-01 8.03061724e-02 6.47070169e+00 5.49908042e-01
-1.47365642e+00 6.68997914e-02 7.11631477e-01 2.07028642e-01
-8.55136961e-02 -2.43741870e-02 -4.72483188e-01 6.95497096e-01
1.09358025e+00 -1.87382564e-01 3.42226867e-03 4.19763833e-01
2.06228986e-01 4.71162289e-01 -1.01450682e+00 1.21021223e+00
6.56205323e-03 -1.04718685e+00 -4.19114649e-01 4.78549860e-02
1.17450260e-01 -4.94549483e-01 -2.21370719e-02 -1.43431142e-01
-4.22267288e-01 -9.08201635e-01 3.80770594e-01 3.25583577e-01
1.12520683e+00 -8.69917333e-01 7.70089447e-01 2.10324541e-01
-1.15683746e+00 -4.23033267e-01 1.65336147e-01 -6.83994442e-02
7.32927397e-02 8.95303547e-01 -9.82069850e-01 4.83619004e-01
6.26517117e-01 6.51884317e-01 -1.94703624e-01 7.06203878e-01
-2.49486089e-01 1.23464155e+00 -1.27523229e-01 4.78122443e-01
-5.40488005e-01 2.37364247e-02 7.83230424e-01 9.23934102e-01
2.94858545e-01 2.60448188e-01 1.93537638e-01 4.60600317e-01
3.20527047e-01 -2.09221929e-01 -7.13101983e-01 8.59085247e-02
7.47344077e-01 1.04426873e+00 -8.12955141e-01 2.89617870e-02
1.22367278e-01 7.10556269e-01 -5.28955102e-01 1.45796388e-01
-1.10502028e+00 -8.25451910e-01 5.58007777e-01 3.66478711e-01
-3.39492500e-01 -1.09173946e-01 -7.48924971e-01 -7.36697555e-01
-1.11114874e-01 -1.06391907e+00 7.10013330e-01 -2.25838572e-01
-1.05180287e+00 7.26339161e-01 -2.63413310e-01 -1.68410063e+00
-1.97361454e-01 -4.69711684e-02 -8.79735231e-01 8.82808208e-01
-1.02911520e+00 -5.33598959e-01 1.54400291e-03 8.56413186e-01
1.91583484e-01 -7.93419182e-02 1.23186767e+00 2.77632415e-01
-5.33891976e-01 7.99569011e-01 -6.98071361e-01 6.12940073e-01
6.66946948e-01 -8.14225078e-01 3.13321233e-01 1.03559351e+00
8.62159729e-02 6.08904660e-01 6.93083525e-01 -6.93756640e-01
-1.34745073e+00 -8.19149375e-01 9.23458457e-01 -5.42671442e-01
4.62379038e-01 -1.84422210e-01 -8.46333146e-01 2.78304309e-01
-6.28429055e-02 2.52669543e-01 1.14869034e+00 -2.82429099e-01
-5.75171590e-01 -4.65925515e-01 -1.27058816e+00 3.70903105e-01
5.14720857e-01 -7.56180227e-01 -5.37827551e-01 -1.23400807e-01
3.83532196e-01 -1.89172000e-01 -1.25974476e+00 7.32363284e-01
1.01363170e+00 -7.67138064e-01 1.05652809e+00 -5.46517789e-01
1.89214736e-01 -1.72909200e-01 3.28992218e-01 -9.92163777e-01
-1.42894939e-01 -1.33243358e+00 -1.35483667e-01 1.21510053e+00
4.56653386e-01 -9.71565783e-01 4.36639875e-01 4.82887954e-01
3.21031421e-01 -8.09624970e-01 -9.32675421e-01 -6.93586946e-01
-4.46746975e-01 -4.47047830e-01 5.91639400e-01 1.47739708e+00
3.10417235e-01 1.48458287e-01 -6.58156872e-01 1.67615205e-01
6.33263111e-01 3.73984985e-02 2.76229739e-01 -1.23632467e+00
-4.20808822e-01 4.63159271e-02 -6.39720023e-01 3.06819025e-02
-4.60715055e-01 -9.70911384e-01 -4.10834789e-01 -4.05209094e-01
-1.53017923e-01 -5.33066452e-01 -8.30130577e-01 6.28547370e-01
-2.95261711e-01 4.07446831e-01 1.44399390e-01 4.37043518e-01
2.49068841e-01 -4.38278139e-01 8.61151874e-01 -1.97976213e-02
-6.11930013e-01 3.55169922e-01 -2.90829659e-01 5.58698952e-01
1.04375589e+00 -1.01028872e+00 -4.54267800e-01 1.71198368e-01
4.19742644e-01 4.08513486e-01 2.36143515e-01 -1.00939834e+00
-4.92652096e-02 1.11482307e-01 3.10505480e-01 -2.87079096e-01
-6.18900172e-02 -9.14622784e-01 6.06895030e-01 1.16195190e+00
-4.85290200e-01 4.41143632e-01 -1.83005109e-02 4.07891542e-01
-2.39196327e-03 1.42070636e-01 7.04804659e-01 1.51075900e-01
4.15046513e-02 3.18538785e-01 -5.07167995e-01 1.43598393e-01
1.15921617e+00 -3.52384567e-01 -4.67107669e-02 1.69075549e-01
-6.99463129e-01 -2.53918916e-01 -1.95709735e-01 2.35557556e-01
7.52688289e-01 -1.10695982e+00 -6.46607220e-01 7.41240144e-01
1.92072312e-03 -5.95947266e-01 4.85352874e-02 1.19667125e+00
-3.75376731e-01 -3.72326434e-01 -3.15215141e-01 -6.55728221e-01
-1.62863588e+00 4.17686582e-01 3.17466795e-01 -3.30305725e-01
-1.01869106e+00 6.54449522e-01 -5.24171710e-01 4.04730827e-01
1.25470325e-01 -8.10247809e-02 -1.03961468e-01 3.48353922e-01
6.26585126e-01 6.71320498e-01 2.06524789e-01 2.08004024e-02
-5.67092121e-01 6.18590534e-01 2.99525321e-01 1.00454316e-02
8.25898051e-01 -4.76647206e-02 -1.38734043e-01 6.42053604e-01
1.25542498e+00 4.39446084e-02 -7.77301967e-01 1.10329859e-01
2.44434789e-01 -6.05122447e-01 -2.63562709e-01 -8.04430246e-01
-1.23208082e+00 1.07521653e+00 9.42976058e-01 5.56712866e-01
1.45600629e+00 -6.50607705e-01 1.13510823e+00 -2.52107698e-02
5.48689067e-01 -6.83513582e-01 -5.30804973e-03 -2.25900680e-01
3.84738952e-01 -5.84349751e-01 9.13661346e-02 -1.94775924e-01
-7.64776826e-01 1.27076697e+00 -1.84191748e-01 -1.06453553e-01
1.05830336e+00 7.26840913e-01 4.21524704e-01 -1.22014143e-01
-3.93619657e-01 7.25296438e-01 -1.34988919e-01 6.57238662e-01
-2.88233254e-02 2.82358766e-01 -4.60361868e-01 6.34687603e-01
-1.42427340e-01 8.33277330e-02 4.73400980e-01 9.85136867e-01
-1.78633947e-02 -1.21363485e+00 -3.90133768e-01 5.35102963e-01
-8.33927274e-01 -1.09619796e-02 -2.17520997e-01 1.86891437e-01
3.01588207e-01 8.74491811e-01 -2.46699482e-01 -1.08258307e+00
2.86069512e-01 6.03904068e-01 -6.11069240e-03 -2.53747225e-01
-1.17286444e+00 9.89457890e-02 2.65374947e-02 -6.55204117e-01
-1.53107792e-01 -5.34272969e-01 -1.17145288e+00 -1.19423449e-01
-4.37688977e-01 1.25071453e-02 4.78177875e-01 8.36361289e-01
4.32873726e-01 6.23118877e-01 1.22236478e+00 -1.79425701e-01
-8.02736521e-01 -5.44885039e-01 -8.96101773e-01 2.96548694e-01
5.49945474e-01 -2.61523783e-01 -6.23978555e-01 2.16479614e-01] | [14.296346664428711, 3.201824426651001] |
e6130d09-a79b-4d26-8c8b-36af1e69a4ac | cycle-sum-cycle-consistent-adversarial-lstm | 1904.08265 | null | http://arxiv.org/abs/1904.08265v1 | http://arxiv.org/pdf/1904.08265v1.pdf | Cycle-SUM: Cycle-consistent Adversarial LSTM Networks for Unsupervised Video Summarization | In this paper, we present a novel unsupervised video summarization model that
requires no manual annotation. The proposed model termed Cycle-SUM adopts a new
cycle-consistent adversarial LSTM architecture that can effectively maximize
the information preserving and compactness of the summary video. It consists of
a frame selector and a cycle-consistent learning based evaluator. The selector
is a bi-direction LSTM network that learns video representations that embed the
long-range relationships among video frames. The evaluator defines a learnable
information preserving metric between original video and summary video and
"supervises" the selector to identify the most informative frames to form the
summary video. In particular, the evaluator is composed of two generative
adversarial networks (GANs), in which the forward GAN is learned to reconstruct
original video from summary video while the backward GAN learns to invert the
processing. The consistency between the output of such cycle learning is
adopted as the information preserving metric for video summarization. We
demonstrate the close relation between mutual information maximization and such
cycle learning procedure. Experiments on two video summarization benchmark
datasets validate the state-of-the-art performance and superiority of the
Cycle-SUM model over previous baselines. | ['Jiashi Feng', 'Li Zhou', 'Li Yuan', 'Francis EH Tay', 'Ping Li'] | 2019-04-17 | null | null | null | null | ['unsupervised-video-summarization'] | ['computer-vision'] | [ 4.73704636e-01 2.72952795e-01 -1.99209407e-01 -3.89679790e-01
-8.89864922e-01 -3.06381881e-01 4.75831658e-01 -3.55464995e-01
-2.12486476e-01 6.59693837e-01 6.05160415e-01 -7.44405296e-03
2.44072020e-01 -6.24660313e-01 -1.19305050e+00 -1.05685961e+00
1.45678788e-01 1.16264559e-01 -4.59271558e-02 1.98746949e-01
1.80155933e-01 1.85784362e-02 -1.06352377e+00 4.35389400e-01
9.22241390e-01 1.03278506e+00 3.82223755e-01 9.02613819e-01
1.28989920e-01 1.31097567e+00 -6.96005642e-01 -2.25210935e-01
1.00006074e-01 -1.13909280e+00 -7.65836179e-01 2.34259993e-01
4.31065828e-01 -6.75093949e-01 -9.88339186e-01 1.08054924e+00
4.98165250e-01 3.88978094e-01 6.68723702e-01 -1.26828039e+00
-9.35178101e-01 8.61420214e-01 -4.62685555e-01 3.18814188e-01
2.82619417e-01 2.49438897e-01 9.89953876e-01 -5.73678732e-01
6.06167614e-01 1.11247981e+00 3.69024336e-01 5.62503517e-01
-9.29638386e-01 -4.92310941e-01 1.75450489e-01 4.59111720e-01
-1.16914356e+00 -5.50461769e-01 9.95721757e-01 -2.85161287e-01
8.21005523e-01 3.65299493e-01 7.77982473e-01 1.03173375e+00
5.42279840e-01 1.11741447e+00 2.05036238e-01 -1.36000961e-01
2.74364650e-01 -1.81803450e-01 -2.13658184e-01 8.43900561e-01
-1.00756973e-01 -2.48966277e-01 -5.12665451e-01 1.19841501e-01
7.90667355e-01 1.26346827e-01 -5.43469250e-01 -3.72631371e-01
-1.21557403e+00 7.92581320e-01 5.25002480e-01 2.64192939e-01
-4.73119408e-01 3.37807626e-01 6.71877742e-01 3.93561929e-01
5.55204630e-01 2.04300925e-01 1.60142273e-01 1.06712095e-01
-1.28157783e+00 1.05388358e-01 5.41282356e-01 1.09251428e+00
6.15265787e-01 6.10779703e-01 -5.89037895e-01 6.15588546e-01
2.47571677e-01 3.42751801e-01 7.81557262e-01 -1.28196430e+00
5.44807792e-01 5.34084916e-01 -2.79153287e-01 -1.07611883e+00
2.00565636e-01 -4.69752431e-01 -1.18971050e+00 -2.00089946e-01
-3.10732871e-01 -2.38124833e-01 -7.25725532e-01 1.80480802e+00
-1.33008473e-02 5.44169962e-01 1.91798076e-01 8.66364062e-01
1.09336114e+00 1.15383410e+00 -2.92493459e-02 -6.48945510e-01
7.62221515e-01 -1.34213042e+00 -8.79742801e-01 -1.89825445e-02
2.99550384e-01 -4.43610758e-01 7.11085021e-01 -3.82415727e-02
-1.58212328e+00 -6.48483574e-01 -1.43296635e+00 -1.29359320e-01
2.60951459e-01 1.59617588e-01 4.06702310e-02 -4.13278043e-02
-1.40707231e+00 6.40476167e-01 -9.06774580e-01 -2.51959383e-01
4.56024766e-01 2.43693560e-01 -4.27132249e-01 2.64473379e-01
-1.04048598e+00 4.22191978e-01 8.31114948e-01 7.40839466e-02
-1.25225222e+00 -4.89622444e-01 -1.16832972e+00 2.97943562e-01
4.07677293e-01 -1.27468932e+00 1.17665875e+00 -1.52580154e+00
-1.64976692e+00 7.32392490e-01 -2.70529598e-01 -7.53071070e-01
5.53043962e-01 -2.37111181e-01 -8.40179995e-02 4.27743793e-01
1.48562372e-01 7.16808975e-01 1.26914370e+00 -1.23597240e+00
-6.05264664e-01 -6.18607439e-02 -1.50855333e-01 5.43971837e-01
-1.66562706e-01 -2.37767667e-01 -6.91835880e-01 -1.03879821e+00
-1.30317211e-01 -7.83510685e-01 9.41517651e-02 -4.67149854e-01
-6.76526606e-01 3.13677751e-02 1.12747169e+00 -1.05912232e+00
1.49225605e+00 -2.00381422e+00 7.51666069e-01 -1.65133134e-01
1.31286204e-01 2.35914156e-01 -2.12272897e-01 2.71746069e-01
-2.04834938e-01 -1.52457906e-02 -4.43274617e-01 -4.93728817e-01
-2.33472496e-01 1.29822314e-01 -4.20471638e-01 4.66122657e-01
1.34936228e-01 1.08011174e+00 -9.04526234e-01 -6.08353198e-01
2.54118323e-01 4.13170129e-01 -5.44907033e-01 9.02636766e-01
-3.49380672e-01 5.55557549e-01 -3.08830917e-01 3.39825302e-01
4.86667335e-01 -1.63580224e-01 3.02981645e-01 -5.14260888e-01
9.76404697e-02 -7.46973157e-02 -7.00465441e-01 1.99251592e+00
2.99861562e-02 8.07607532e-01 -2.20867425e-01 -1.34125841e+00
6.84163272e-01 3.24874729e-01 4.55008924e-01 -4.60319728e-01
3.20508033e-01 -1.67572647e-01 -3.77319962e-01 -8.14290106e-01
4.93899614e-01 6.39793426e-02 -1.16958404e-02 4.93451089e-01
4.08696651e-01 1.81201752e-02 1.76350385e-01 4.31711704e-01
9.22449172e-01 4.26651210e-01 3.68496567e-01 -1.52921602e-01
9.72560704e-01 -5.03491163e-01 7.89694607e-01 4.29133654e-01
-1.20374961e-02 7.16695130e-01 7.02375233e-01 -3.43532860e-01
-1.17046905e+00 -1.07838821e+00 7.18040884e-01 9.98229861e-01
3.42039138e-01 -2.36960992e-01 -1.25273800e+00 -8.21462393e-01
-4.93952572e-01 8.78042340e-01 -6.75407350e-01 -6.41767979e-01
-7.95862496e-01 -2.12126091e-01 3.97739619e-01 5.86661875e-01
9.78404641e-01 -1.13125443e+00 -3.53787571e-01 4.05240990e-03
-5.14184713e-01 -9.00885046e-01 -1.19459701e+00 -3.11893553e-01
-8.57348561e-01 -9.35198903e-01 -7.39196420e-01 -1.19412196e+00
7.01919913e-01 2.51705498e-01 9.39533651e-01 -7.40881115e-02
3.03385377e-01 3.30796659e-01 -2.53644854e-01 -7.89656788e-02
-7.55904436e-01 9.12990943e-02 -1.82565853e-01 3.71361375e-01
-7.27007687e-02 -6.26734853e-01 -6.91609204e-01 8.15872401e-02
-1.17580712e+00 3.72376531e-01 5.47775567e-01 7.78189301e-01
9.06323612e-01 2.37371097e-03 5.31781733e-01 -7.90256679e-01
5.11550367e-01 -5.75561225e-01 -1.90936878e-01 3.81598681e-01
-1.92131370e-01 2.93181896e-01 8.31524909e-01 -2.12559149e-01
-1.16508877e+00 -7.63677806e-02 1.37882959e-02 -7.93880403e-01
2.12903500e-01 4.04857486e-01 -6.18747771e-01 3.04621637e-01
1.14297338e-01 7.96792924e-01 5.69691285e-02 -1.63554158e-02
3.88019592e-01 4.73914534e-01 1.08866513e+00 -1.94548666e-01
7.01000035e-01 3.67646188e-01 -2.69010544e-01 -9.08372521e-01
-6.60737097e-01 2.99342879e-04 -5.83824873e-01 -3.69852781e-01
1.30485654e+00 -9.50424910e-01 -3.91245097e-01 6.90726519e-01
-1.18667793e+00 -3.04314256e-01 -4.91498232e-01 3.83859396e-01
-1.08633530e+00 5.28814554e-01 -6.68413818e-01 -3.78760666e-01
-8.22947562e-01 -1.15033233e+00 9.57931697e-01 4.83713776e-01
3.90143068e-05 -1.10127831e+00 9.08887237e-02 3.28249216e-01
1.79249793e-01 5.64196408e-01 9.23479617e-01 -5.66243649e-01
-7.32703745e-01 5.00173569e-02 4.87184934e-02 7.86573350e-01
1.58284754e-01 1.71460867e-01 -6.72174573e-01 -5.14845788e-01
3.55573744e-01 -1.35765910e-01 1.19707227e+00 6.67403042e-01
1.29978836e+00 -8.92987907e-01 -7.69997612e-02 1.01386476e+00
1.27030087e+00 5.34083545e-01 9.84288573e-01 1.86428294e-01
9.43446159e-01 1.83215111e-01 3.88300776e-01 1.89348668e-01
3.70311797e-01 2.82367408e-01 4.29701388e-01 9.67684016e-02
-1.65762961e-01 -6.02461576e-01 7.31106281e-01 1.31770623e+00
-1.07401073e-01 -6.16013646e-01 -1.87486708e-01 2.72872120e-01
-2.14008379e+00 -1.47745895e+00 5.48096120e-01 2.09562707e+00
7.31805384e-01 -3.03726234e-02 1.21265948e-01 -4.53571565e-02
9.60515022e-01 6.86519086e-01 -8.19500327e-01 -3.17371875e-01
-1.47755757e-01 -4.16588604e-01 2.11205199e-01 4.91505384e-01
-1.24454272e+00 8.48313093e-01 5.35137224e+00 8.50132585e-01
-1.12284493e+00 -7.69938459e-04 8.52057517e-01 -2.15095893e-01
-4.69359189e-01 -1.21765710e-01 -1.61125377e-01 7.69048750e-01
7.36911714e-01 -3.14968348e-01 3.56663018e-01 7.03858078e-01
4.45838600e-01 1.37070462e-01 -1.18943906e+00 9.70427454e-01
5.86751580e-01 -1.50980151e+00 6.62006080e-01 -2.93660164e-01
8.88416946e-01 -2.31855556e-01 5.63317724e-02 1.38844460e-01
-1.41796991e-02 -8.84196579e-01 9.10838902e-01 9.13731813e-01
8.67539287e-01 -1.03401840e+00 7.51169562e-01 3.15170437e-01
-1.25261223e+00 -5.60448878e-02 -3.11254740e-01 3.97488087e-01
2.97021866e-01 1.25882283e-01 -4.22116131e-01 7.60665715e-01
3.79634291e-01 1.19557011e+00 -3.03100020e-01 7.20050097e-01
-3.78570914e-01 5.21951616e-01 3.42564911e-01 3.46911758e-01
3.53341192e-01 -6.30566776e-01 9.28212225e-01 1.31137824e+00
3.46433133e-01 1.13832235e-01 8.86805654e-02 6.67192578e-01
-5.08079827e-01 -7.06234202e-02 -6.12728596e-01 -1.24677747e-01
3.59200686e-01 1.01605415e+00 -4.56857800e-01 -5.83352983e-01
-2.13688821e-01 1.41192555e+00 2.82156050e-01 5.74213624e-01
-1.14387989e+00 -3.59309018e-01 3.49878907e-01 2.57667042e-02
4.70335722e-01 2.28498746e-02 5.24059720e-02 -1.32334816e+00
1.14103265e-01 -9.03762639e-01 5.86373508e-01 -8.98240209e-01
-8.65051031e-01 7.12640703e-01 9.16586220e-02 -1.33532619e+00
-5.34620106e-01 1.14048257e-01 -1.19261038e+00 4.32752341e-01
-1.18619204e+00 -1.27298391e+00 -3.42801839e-01 7.07061052e-01
1.05800116e+00 -5.60484231e-01 3.19012612e-01 -2.78807394e-02
-9.39020991e-01 6.16263270e-01 2.55861878e-01 1.71202928e-01
4.33638871e-01 -1.14302897e+00 2.17195168e-01 1.14587402e+00
-1.34716511e-01 3.78032178e-01 6.92334235e-01 -6.92983449e-01
-1.34163880e+00 -1.50128722e+00 6.16473138e-01 1.18198015e-01
3.38182062e-01 2.93578431e-02 -8.32268119e-01 1.11096060e+00
7.92218506e-01 -4.23299789e-01 6.21008515e-01 -8.98051262e-01
-6.43175244e-02 -2.71191955e-01 -9.87991393e-01 5.68292081e-01
9.99362409e-01 -4.53184485e-01 -8.06214750e-01 1.84155390e-01
1.21158075e+00 -3.87242913e-01 -6.86153293e-01 3.32594395e-01
3.82680029e-01 -1.03648460e+00 8.05039704e-01 -5.77377021e-01
8.91770899e-01 -3.03053737e-01 -8.70358199e-02 -1.36098981e+00
-3.77398461e-01 -8.11401129e-01 -5.16700566e-01 1.38229227e+00
-9.00453329e-02 -2.25244567e-01 5.66737294e-01 2.38403916e-01
-4.54161853e-01 -8.04920077e-01 -6.64499938e-01 -5.11593282e-01
-2.33973488e-01 1.84775874e-01 4.95377153e-01 8.00189257e-01
-1.37351424e-01 5.60630739e-01 -6.28626049e-01 -1.03842244e-01
7.08383799e-01 1.51794717e-01 6.87136531e-01 -4.25036997e-01
-3.13046992e-01 -4.62448984e-01 -3.69471520e-01 -1.20691633e+00
5.59666097e-01 -9.35792983e-01 -2.35579889e-02 -1.62601936e+00
5.26123822e-01 4.36064750e-01 -2.34230876e-01 1.20253533e-01
-3.12515527e-01 -9.53691378e-02 1.95779666e-01 3.47865403e-01
-9.20121312e-01 1.04142654e+00 1.21119666e+00 -4.72848266e-01
-3.10704201e-01 -3.74675030e-04 -7.49914646e-01 8.57123494e-01
7.27044106e-01 -2.71147609e-01 -7.84069479e-01 -7.29966521e-01
-2.09767461e-01 3.63619536e-01 2.43680596e-01 -9.68957841e-01
3.78075421e-01 -2.92228386e-02 4.03049022e-01 -6.36024535e-01
7.18520507e-02 -5.92900991e-01 2.99114197e-01 5.35905123e-01
-7.28124440e-01 1.96535841e-01 -1.87122673e-01 7.94087589e-01
-3.60605061e-01 -2.14295432e-01 8.74115109e-01 -2.00836837e-01
-6.08374059e-01 5.18103063e-01 -2.01821208e-01 1.14976577e-01
1.16608894e+00 -2.24024147e-01 -8.15575346e-02 -7.97235429e-01
-4.70167756e-01 2.53487080e-01 4.57279861e-01 4.06054586e-01
9.95243847e-01 -1.56979263e+00 -9.57630694e-01 2.97381729e-01
-4.00169879e-01 1.72095999e-01 4.12151396e-01 4.64030653e-01
-8.47825170e-01 2.76109487e-01 -3.16519082e-01 -5.65590978e-01
-1.15404582e+00 6.47239745e-01 3.74287546e-01 -2.52408475e-01
-5.70359468e-01 8.16713929e-01 4.22515005e-01 1.98842362e-01
3.05314749e-01 -7.25735202e-02 -2.77758449e-01 -5.40422946e-02
6.22914553e-01 5.72143018e-01 -3.64158988e-01 -8.90045464e-01
-1.14273354e-01 3.02656949e-01 -1.28571659e-01 -1.47307653e-03
1.38142800e+00 -3.60733092e-01 -4.29415852e-01 3.07550699e-01
1.60569382e+00 -3.98415148e-01 -1.52838743e+00 -2.08574519e-01
-4.59456712e-01 -2.54843205e-01 -8.94791111e-02 -1.07324168e-01
-1.54118669e+00 5.43482423e-01 2.58925617e-01 -1.33008612e-02
1.50212181e+00 -1.91619366e-01 1.19438815e+00 1.84918284e-01
-1.25687152e-01 -1.03405559e+00 3.75501245e-01 2.27703169e-01
1.11309195e+00 -9.51035678e-01 -4.51668017e-02 1.29054621e-01
-9.00617778e-01 9.78423357e-01 6.52561605e-01 -4.48207468e-01
2.09142357e-01 -1.25575508e-03 -3.86034459e-01 -9.59108490e-03
-8.15061033e-01 3.14086407e-01 3.44764441e-01 4.29962784e-01
1.78283498e-01 -1.82175115e-01 -2.04735801e-01 6.25552773e-01
-2.14563847e-01 -4.90327068e-02 3.57416779e-01 6.68763220e-01
-4.86902922e-01 -5.54681718e-01 -5.71944378e-02 3.44270319e-01
-3.54353040e-01 5.37467562e-02 -3.21038872e-01 4.64965612e-01
-3.71846259e-02 7.85665989e-01 3.38717848e-01 -6.67854667e-01
1.16432816e-01 -6.64957687e-02 3.34805876e-01 -3.18534970e-01
-4.96894687e-01 1.93325266e-01 -2.76287943e-01 -6.72925711e-01
-6.01369858e-01 -6.20929599e-01 -1.20594096e+00 -4.01326537e-01
3.49478759e-02 2.44656771e-01 1.98637798e-01 9.52918530e-01
4.28071946e-01 8.65872860e-01 9.62024033e-01 -9.29842234e-01
-4.32067543e-01 -8.64120185e-01 -3.83407056e-01 4.34678197e-01
5.03928840e-01 -8.17545429e-02 -1.74580261e-01 7.30287552e-01] | [10.46061897277832, 0.4141937792301178] |
7bf221b6-35b3-4af0-956a-ffecac2fd423 | unicorn-a-unified-multi-tasking-model-for | null | null | https://dl.acm.org/doi/abs/10.1145/3588938 | https://dl.acm.org/doi/pdf/10.1145/3588938 | Unicorn: A Unified Multi-tasking Model for Supporting Matching Tasks in Data Integration | Data matching - which decides whether two data elements (e.g., string, tuple, column, or knowledge graph entity) are the "same" (a.k.a. a match) - is a key concept in data integration, such as entity matching and schema matching. The widely used practice is to build task-specific or even dataset-specific solutions, which are hard to generalize and disable the opportunities of knowledge sharing that can be learned from different datasets and multiple tasks. In this paper, we propose Unicorn, a unified model for generally supporting common data matching tasks. Unicorn can enable knowledge sharing by learning from multiple tasks and multiple datasets, and can also support zero-shot prediction for new tasks with zero labeled matching/non-matching pairs. However, building such a unified model is challenging due to heterogeneous formats of input data elements and various matching semantics of multiple tasks. To address the challenges, Unicorn employs one generic Encoder that converts any pair of data elements (a, b) into a learned representation, and uses a Matcher, which is a binary classifier, to decide whether a matches b. To align matching semantics of multiple tasks, Unicorn adopts a mixture-of-experts model that enhances the learned representation into a better representation. We conduct extensive experiments using 20 datasets on seven well-studied data matching tasks, and find that our unified model can achieve better performance on most tasks and on average, compared with the state-of-the-art specific models trained for ad-hoc tasks and datasets separately. Moreover, Unicorn can also well serve new matching tasks with zero-shot learning. | ['Song Gao', 'Xiaofeng Jia', 'Xiaoyong Du', 'Guoliang Li', 'Peng Wang', 'Nan Tang', 'Ju Fan', 'Jianhong Tu'] | 2023-05-01 | null | null | null | sigmod-pods-2023-5 | ['data-integration', 'zero-shot-learning', 'entity-resolution'] | ['knowledge-base', 'methodology', 'natural-language-processing'] | [ 3.36565286e-01 -1.37967840e-01 -7.87041843e-01 -4.41815048e-01
-7.81098425e-01 -3.57492477e-01 4.39847916e-01 4.26986545e-01
-4.05180305e-01 4.64986503e-01 2.41760343e-01 -2.99973935e-02
-2.73618042e-01 -9.91513073e-01 -9.74377453e-01 -1.66276202e-01
3.05739671e-01 6.24716163e-01 4.53205436e-01 -2.62360513e-01
-8.90192688e-02 -1.04919679e-01 -2.02700138e+00 8.40682089e-01
8.88027787e-01 1.18252075e+00 4.23494458e-01 1.57920301e-01
-6.97503448e-01 6.86263919e-01 -4.24743563e-01 -8.93739283e-01
3.07227284e-01 1.36804953e-01 -1.13447404e+00 -4.19668823e-01
6.65615559e-01 8.22200179e-02 -4.04754698e-01 9.92329240e-01
4.22271132e-01 2.11099505e-01 4.54168707e-01 -1.72456264e+00
-1.02609277e+00 7.61515677e-01 -2.68753648e-01 1.92966033e-02
4.04598892e-01 -5.98221272e-02 1.21619117e+00 -6.97644889e-01
9.07195330e-01 1.19717276e+00 6.85867250e-01 4.64730859e-01
-1.08732986e+00 -9.98468935e-01 -7.25254370e-03 5.77293992e-01
-1.34426725e+00 -5.46604216e-01 3.11123669e-01 -4.30138022e-01
1.20439041e+00 2.55309910e-01 1.57174066e-01 1.17449391e+00
1.03969216e-01 8.49122286e-01 5.82840502e-01 -3.22422862e-01
7.96579048e-02 1.05983399e-01 3.47256750e-01 3.91504407e-01
3.63907963e-01 -2.97831874e-02 -6.07939959e-01 -2.30004996e-01
2.08551869e-01 5.48372984e-01 -2.37176627e-01 -5.67439675e-01
-1.15630698e+00 4.62141693e-01 6.25770628e-01 4.76519287e-01
-6.81635067e-02 -9.44736376e-02 6.24847949e-01 7.27333665e-01
3.70309018e-02 5.61000943e-01 -6.16042376e-01 1.22433938e-01
-5.22572279e-01 3.95906687e-01 9.52542484e-01 1.58529103e+00
1.31021738e+00 -5.77615082e-01 -4.09634590e-01 1.10782969e+00
1.22269884e-01 2.68086642e-01 7.37618148e-01 -7.59795129e-01
8.47403944e-01 1.03187251e+00 5.38039301e-03 -7.58784890e-01
-2.68890619e-01 -3.66915941e-01 -8.33620369e-01 -3.29724014e-01
1.15900032e-01 1.02704197e-01 -8.38590145e-01 1.90087020e+00
2.81505436e-01 4.61593270e-01 4.80807513e-01 6.31493688e-01
1.14126694e+00 4.54575062e-01 4.14608955e-01 8.15493762e-02
1.47206032e+00 -1.10173690e+00 -6.73611820e-01 -6.31729305e-01
9.54190314e-01 -7.33074129e-01 9.40378606e-01 -1.57446623e-01
-6.46197200e-01 -8.93471003e-01 -1.09104395e+00 -4.10283238e-01
-8.79042804e-01 -3.23303670e-01 6.08596265e-01 3.21677595e-01
-4.38526183e-01 6.37239754e-01 -2.38782004e-01 -5.72018325e-01
2.90897548e-01 1.24277227e-01 -6.70400321e-01 -5.22315800e-01
-1.55626106e+00 9.32762980e-01 7.48217940e-01 -6.39042914e-01
-5.08869290e-01 -1.15469706e+00 -1.07073462e+00 1.62205264e-01
7.42879689e-01 -9.71487761e-01 1.28404880e+00 -7.02328622e-01
-6.73086584e-01 1.06060040e+00 -2.48727173e-01 -2.62326777e-01
1.31207928e-01 -1.36639431e-01 -9.62239623e-01 -5.77929497e-01
3.00373703e-01 5.83158612e-01 6.07999384e-01 -1.10662532e+00
-9.93534148e-01 -4.90901023e-01 1.13970503e-01 1.11036234e-01
-4.09392834e-01 1.40088573e-01 -6.20643198e-01 -5.78295708e-01
-1.17454439e-01 -6.39954448e-01 3.23982686e-02 2.85027232e-02
-8.67961869e-02 -5.21100342e-01 6.80862904e-01 -4.44853544e-01
1.35943580e+00 -2.17865491e+00 1.03254624e-01 -6.45496026e-02
1.57822207e-01 4.19574678e-01 -3.37772727e-01 5.22119701e-01
-1.50872812e-01 -2.46466957e-02 -1.50566354e-01 -3.67378205e-01
2.15637207e-01 5.53190947e-01 -3.73198301e-01 -1.43731073e-01
-2.90731322e-02 1.21542084e+00 -1.03022277e+00 -3.95539522e-01
-2.29886010e-01 -1.66956149e-02 -1.89448133e-01 5.69566190e-01
-4.54057395e-01 -2.47792900e-01 -2.25157902e-01 7.53205299e-01
7.18279004e-01 -5.54541588e-01 4.25851524e-01 -5.98997474e-01
2.09540099e-01 1.50138944e-01 -1.62541687e+00 2.18948126e+00
-6.81151688e-01 1.94275007e-01 -2.23865524e-01 -1.13443637e+00
1.11137807e+00 3.84272426e-01 4.09155577e-01 -9.91802275e-01
-1.42595753e-01 5.08155107e-01 -1.09066717e-01 -6.84783638e-01
4.66005147e-01 -4.86388505e-02 -3.62369597e-01 6.04336977e-01
5.82694292e-01 4.01394725e-01 1.79348812e-01 6.76998124e-02
1.27407300e+00 -1.19239829e-01 5.70801914e-01 6.40471280e-02
2.69825816e-01 -5.32882363e-02 9.80472803e-01 8.78551066e-01
1.08154818e-01 2.35672593e-01 8.14092830e-02 -4.63246346e-01
-8.05642247e-01 -8.28764677e-01 -1.61271721e-01 1.58909309e+00
5.26093245e-01 -6.33506417e-01 -8.86327326e-02 -7.20560789e-01
7.20419645e-01 6.40787601e-01 -5.39725780e-01 -6.05806649e-01
-2.45445639e-01 -3.04391533e-01 4.63131607e-01 7.40368962e-01
3.12406301e-01 -1.06397808e+00 -3.52069318e-01 3.97356033e-01
-3.00891638e-01 -1.12635326e+00 -5.70125341e-01 3.44150484e-01
-4.75114077e-01 -1.24330235e+00 -3.31364453e-01 -8.31223667e-01
1.93242431e-01 6.99133277e-01 1.38366425e+00 4.74086292e-02
-2.64872760e-01 2.04733044e-01 -4.78854746e-01 -4.19299543e-01
-2.85956413e-01 2.27647722e-01 -2.26277276e-03 5.13192602e-02
8.17431629e-01 -4.87382323e-01 -2.23164007e-01 7.74129093e-01
-1.12878394e+00 6.19450444e-03 5.57314634e-01 7.32252359e-01
5.46272814e-01 -1.99342310e-01 7.70927012e-01 -1.32711184e+00
4.66869146e-01 -1.04006612e+00 -2.08468035e-01 1.02670383e+00
-8.87268782e-01 2.88798124e-01 4.99482125e-01 -4.74419534e-01
-1.01878834e+00 -2.71575004e-01 2.89601415e-01 -8.04878235e-01
-1.14765815e-01 6.29537046e-01 -5.05136847e-01 1.67538524e-02
7.95036554e-01 -5.05150557e-02 -2.22412452e-01 -8.23159099e-01
6.04733825e-01 1.02677834e+00 8.40502322e-01 -8.73125553e-01
5.41060269e-01 8.43539387e-02 -2.77383536e-01 1.17020883e-01
-9.55779016e-01 -1.03516459e+00 -5.44536471e-01 2.58570284e-01
6.42113388e-01 -1.17378235e+00 -3.66443962e-01 4.71484721e-01
-1.22567725e+00 2.47525442e-02 -2.62092948e-01 1.10044703e-01
-3.34951043e-01 1.11218877e-01 -2.50743240e-01 -2.20285684e-01
-3.87973160e-01 -1.03486514e+00 8.99784625e-01 4.12158072e-01
-2.05126956e-01 -8.50287318e-01 1.44082993e-01 5.10117948e-01
4.65695649e-01 -2.62950003e-01 1.31713593e+00 -1.13479400e+00
-4.99997675e-01 -4.78480682e-02 -4.57517177e-01 -2.44728457e-02
3.11635405e-01 -3.83206666e-01 -9.09094691e-01 -3.13675940e-01
-4.62525338e-01 -6.71598494e-01 9.24300432e-01 -4.10145819e-01
1.30893028e+00 -1.52312905e-01 -8.23444963e-01 7.93000102e-01
1.48711729e+00 2.31332242e-01 5.47518194e-01 3.66295904e-01
6.75201714e-01 6.23048365e-01 8.47489297e-01 4.16730970e-01
7.00415492e-01 1.05605292e+00 2.73509234e-01 1.62051663e-01
-3.19531143e-01 -5.54651558e-01 -3.50644998e-02 5.60660005e-01
4.09243047e-01 -3.07222962e-01 -8.37900996e-01 6.64285243e-01
-2.45170188e+00 -9.85452652e-01 6.88468814e-02 2.32096982e+00
1.05214429e+00 -3.07633966e-01 -1.86039880e-01 -2.70216912e-01
8.59947324e-01 1.49086550e-01 -8.06502402e-01 -1.42704055e-01
-1.24953404e-01 3.72416914e-01 3.24174792e-01 1.79819241e-02
-9.81745005e-01 8.42213750e-01 5.48223639e+00 9.49708521e-01
-8.22803736e-01 3.60535234e-01 -3.32256407e-02 1.11393981e-01
-7.19600797e-01 2.44497359e-01 -1.10808301e+00 6.40550554e-01
6.06366515e-01 -6.02152824e-01 3.72835815e-01 8.31235647e-01
-8.89214516e-01 3.49102020e-01 -1.62827361e+00 1.23399520e+00
4.78143059e-02 -1.58838165e+00 2.31089264e-01 -2.15986997e-01
7.47628748e-01 3.98961678e-02 -2.31577188e-01 8.89370263e-01
4.55355883e-01 -9.46956754e-01 3.98316592e-01 7.29817152e-01
8.79366815e-01 -3.02318156e-01 7.73606062e-01 4.84831989e-01
-1.45862567e+00 -2.60593504e-01 -6.79826558e-01 2.88497657e-01
-5.57024078e-03 3.58806103e-01 -4.82091367e-01 1.20094013e+00
7.52095759e-01 9.90267038e-01 -5.48135757e-01 1.09478486e+00
1.22426458e-01 -2.17778817e-01 -4.46413681e-02 3.80202860e-01
-2.08620578e-01 1.74942061e-01 1.69782698e-01 1.17190158e+00
4.75680411e-01 -1.69982731e-01 4.90790844e-01 7.37002850e-01
-5.41730464e-01 1.88770220e-01 -6.72898948e-01 6.07030354e-02
1.03116989e+00 1.19563174e+00 2.41251886e-01 -5.16649544e-01
-9.19683933e-01 8.19976687e-01 8.34656477e-01 2.65450209e-01
-5.12132466e-01 -4.83675003e-01 1.03930831e+00 -1.21124461e-02
1.21308021e-01 4.64597940e-01 -3.57174804e-03 -1.44124007e+00
2.75378883e-01 -1.05485296e+00 1.09671819e+00 -7.38826752e-01
-1.75141847e+00 3.31199616e-01 2.26986017e-02 -1.30021393e+00
-2.10302129e-01 -3.41926426e-01 -6.44346654e-01 9.14068043e-01
-1.58144557e+00 -1.17513394e+00 -5.97425997e-01 7.93865979e-01
3.07487160e-01 -5.64445317e-01 9.45930719e-01 7.72037327e-01
-6.15646124e-01 9.88716304e-01 -1.26000836e-01 9.96042117e-02
1.25103188e+00 -1.00804162e+00 5.80645323e-01 5.48007488e-01
3.06185424e-01 6.43854260e-01 2.34305620e-01 -7.44310260e-01
-1.51998210e+00 -1.34822690e+00 1.18993366e+00 -3.63095999e-01
7.09370792e-01 -1.54836968e-01 -1.48637068e+00 8.49312901e-01
1.16679147e-01 3.65916580e-01 1.13643885e+00 5.52145660e-01
-1.07664776e+00 -3.65883738e-01 -1.03204322e+00 2.69560575e-01
1.59080696e+00 -7.58872926e-01 -1.00115848e+00 2.45328218e-01
1.04949498e+00 -3.33016425e-01 -1.28623497e+00 8.12236726e-01
5.02871931e-01 -7.65030682e-01 1.01319575e+00 -1.31524134e+00
3.25666726e-01 -2.75837898e-01 -6.02463961e-01 -1.30583191e+00
-5.58892250e-01 -3.09430331e-01 -3.87257546e-01 1.50033116e+00
4.39496309e-01 -5.09326458e-01 4.88099068e-01 8.46403003e-01
-2.18601197e-01 -4.89112318e-01 -8.77197087e-01 -1.05344021e+00
-2.74615109e-01 -3.79249632e-01 1.28144670e+00 1.49743009e+00
2.04837993e-01 4.21100944e-01 -3.86778176e-01 3.30459587e-02
4.51337814e-01 6.09459221e-01 9.55431521e-01 -1.48794770e+00
-3.94381523e-01 -3.37377846e-01 -4.56781685e-01 -8.92767370e-01
5.19418180e-01 -1.53531468e+00 -3.01567167e-01 -1.44964552e+00
4.86664414e-01 -9.10778821e-01 -7.28595018e-01 9.25849140e-01
-4.62449938e-01 -4.14703935e-01 1.85368806e-01 4.56375957e-01
-8.48333359e-01 3.58871460e-01 8.53835762e-01 -4.74841088e-01
-1.70699060e-02 -5.70587330e-02 -9.24041033e-01 1.09440148e-01
4.35595065e-01 -6.61764801e-01 -4.83328402e-01 -7.00883865e-01
3.89818192e-01 1.23886131e-01 1.69152394e-01 -9.54135537e-01
7.05773354e-01 -2.87629932e-01 -1.11074997e-02 -1.62764311e-01
6.25684783e-02 -9.41313386e-01 7.27340996e-01 1.41890183e-01
-4.59606498e-01 -4.36501652e-02 -8.33739992e-03 6.98954225e-01
-4.53543633e-01 -5.12475073e-01 4.84350979e-01 -9.50319543e-02
-1.44932914e+00 6.46023452e-01 6.05206192e-01 3.67740095e-01
1.09932983e+00 5.43716326e-02 -9.38864827e-01 1.47999838e-01
-4.96977985e-01 6.06634796e-01 5.49155533e-01 1.11609280e+00
4.10452187e-01 -1.68728888e+00 -5.30700564e-01 2.47868776e-01
1.09456158e+00 -6.16902858e-03 3.24935257e-01 3.52114797e-01
2.31922954e-01 1.55171186e-01 -5.09074330e-01 -3.36263269e-01
-1.17527640e+00 9.66028452e-01 2.60972351e-01 -3.35642844e-01
-4.08781290e-01 5.96378148e-01 7.67129064e-02 -8.30305755e-01
3.15220594e-01 8.72901874e-04 -8.72619152e-02 2.59658903e-01
6.85938478e-01 2.69241273e-01 3.58788401e-01 -3.45401138e-01
-3.48543167e-01 4.17669803e-01 -4.12178576e-01 5.65616012e-01
1.02918839e+00 2.36660525e-01 -1.86891288e-01 4.49382007e-01
1.25544405e+00 -5.14256179e-01 -8.35711539e-01 -1.12852573e+00
2.68895477e-01 -7.65180707e-01 -2.99653381e-01 -7.20317245e-01
-1.07980502e+00 6.72154546e-01 3.89604181e-01 -5.41598834e-02
8.42202961e-01 3.07638556e-01 1.01875186e+00 6.18514180e-01
6.41801357e-01 -1.03609669e+00 -9.42030922e-02 4.56536382e-01
6.82926595e-01 -1.42473304e+00 -3.28248918e-01 -4.69151974e-01
-6.24180317e-01 8.53309333e-01 1.09253955e+00 4.52481359e-01
6.14399910e-01 1.63946390e-01 -5.38000911e-02 -1.13849834e-01
-1.22424614e+00 -5.35825491e-01 5.52549720e-01 7.10596919e-01
3.34874243e-01 1.84321493e-01 -8.41441080e-02 1.05756819e+00
2.16999754e-01 2.91525602e-01 -1.88228488e-01 1.03525805e+00
-4.57889438e-01 -1.62858796e+00 4.20248546e-02 7.89593279e-01
1.10119320e-01 2.66108550e-02 -1.46632522e-01 5.06704152e-01
5.47149777e-01 7.97613919e-01 3.16954464e-01 -7.23322332e-01
6.62421227e-01 4.04826611e-01 2.54124224e-01 -9.17643845e-01
-7.65712619e-01 -7.06051052e-01 7.94778317e-02 -6.45013392e-01
-2.50083536e-01 -4.47054863e-01 -1.08886218e+00 -3.19382906e-01
-4.12583910e-02 1.12529881e-02 2.13326722e-01 1.02132809e+00
7.82008767e-01 3.35182220e-01 3.39353383e-01 8.76094215e-03
-5.67254126e-01 -7.49582291e-01 -4.85939652e-01 1.16039681e+00
-8.72271433e-02 -9.84828055e-01 1.69052362e-01 -1.59696907e-01] | [9.533543586730957, 8.527593612670898] |
49dfc015-6ef8-4bb8-b622-8fcdd3a8c410 | p-stmo-pre-trained-spatial-temporal-many-to | 2203.07628 | null | https://arxiv.org/abs/2203.07628v2 | https://arxiv.org/pdf/2203.07628v2.pdf | P-STMO: Pre-Trained Spatial Temporal Many-to-One Model for 3D Human Pose Estimation | This paper introduces a novel Pre-trained Spatial Temporal Many-to-One (P-STMO) model for 2D-to-3D human pose estimation task. To reduce the difficulty of capturing spatial and temporal information, we divide this task into two stages: pre-training (Stage I) and fine-tuning (Stage II). In Stage I, a self-supervised pre-training sub-task, termed masked pose modeling, is proposed. The human joints in the input sequence are randomly masked in both spatial and temporal domains. A general form of denoising auto-encoder is exploited to recover the original 2D poses and the encoder is capable of capturing spatial and temporal dependencies in this way. In Stage II, the pre-trained encoder is loaded to STMO model and fine-tuned. The encoder is followed by a many-to-one frame aggregator to predict the 3D pose in the current frame. Especially, an MLP block is utilized as the spatial feature extractor in STMO, which yields better performance than other methods. In addition, a temporal downsampling strategy is proposed to diminish data redundancy. Extensive experiments on two benchmarks show that our method outperforms state-of-the-art methods with fewer parameters and less computational overhead. For example, our P-STMO model achieves 42.1mm MPJPE on Human3.6M dataset when using 2D poses from CPN as inputs. Meanwhile, it brings a 1.5-7.1 times speedup to state-of-the-art methods. Code is available at https://github.com/paTRICK-swk/P-STMO. | ['Wen Gao', 'Siwei Ma', 'Shanshe Wang', 'Xinfeng Zhang', 'Zhenhua Liu', 'Wenkang Shan'] | 2022-03-15 | null | null | null | null | ['monocular-3d-human-pose-estimation'] | ['computer-vision'] | [ 1.24835335e-02 -1.67268813e-02 -2.38606766e-01 -3.31737369e-01
-8.06661189e-01 3.02835070e-02 1.46650702e-01 -4.68366057e-01
-4.65539843e-01 4.52411622e-01 2.90877134e-01 2.39160508e-01
3.00471216e-01 -5.54840446e-01 -8.93436074e-01 -5.54070771e-01
-3.15165482e-02 3.71282011e-01 5.29912472e-01 -1.79033861e-01
-1.25789866e-01 1.88197955e-01 -1.30022526e+00 2.01887503e-01
5.33452928e-01 1.19905186e+00 4.88160342e-01 6.00415587e-01
2.39218250e-01 7.23170102e-01 -5.13824105e-01 -2.04411209e-01
2.54194945e-01 -3.84709299e-01 -5.76807320e-01 2.98790395e-01
1.52531117e-01 -5.43367743e-01 -6.39835060e-01 7.53126562e-01
7.54036009e-01 2.12671906e-01 2.95776308e-01 -1.02170491e+00
-2.12872028e-01 1.23510092e-01 -9.01912868e-01 5.02880551e-02
4.37609285e-01 2.66600758e-01 5.46468973e-01 -1.18805552e+00
5.99592388e-01 1.41781187e+00 6.16807759e-01 5.82596660e-01
-1.02154636e+00 -6.89707816e-01 1.13861747e-01 1.81071937e-01
-1.43140507e+00 -3.25160563e-01 8.58463347e-01 -4.14644867e-01
8.76453280e-01 -2.52649151e-02 6.98307693e-01 1.19149780e+00
5.43562591e-01 9.96146321e-01 7.90940523e-01 -1.95252046e-01
-8.97789001e-02 -6.15350008e-01 -2.07236752e-01 7.93694496e-01
-1.12174049e-01 2.56451309e-01 -1.01388228e+00 1.09731846e-01
1.03399909e+00 1.46309853e-01 -1.32549450e-01 -1.44956902e-01
-1.35774732e+00 4.97296125e-01 5.88917553e-01 1.14388667e-01
-5.93160868e-01 4.09691542e-01 5.24146974e-01 1.22940496e-01
5.00360191e-01 -1.66623667e-01 -4.23237830e-01 -2.83356041e-01
-8.56445134e-01 3.26669723e-01 2.89080888e-01 9.81657147e-01
5.61306834e-01 -1.40238762e-01 -1.22600488e-01 8.57342064e-01
3.81248623e-01 4.57112044e-01 5.67675412e-01 -1.03864717e+00
7.62734830e-01 2.91420490e-01 1.43163115e-01 -9.65850532e-01
-5.13931513e-01 -4.99260992e-01 -9.84450221e-01 -5.40016741e-02
2.15507939e-01 -1.03760727e-01 -1.09882188e+00 1.62380922e+00
6.12812221e-01 2.31268957e-01 -2.22407103e-01 1.37716544e+00
6.50285542e-01 8.04880738e-01 -7.37793520e-02 -4.04495709e-02
1.62761438e+00 -1.38473082e+00 -7.01956511e-01 -6.28322482e-01
3.23094308e-01 -8.38626325e-01 8.46773326e-01 3.42784077e-01
-1.10264826e+00 -1.07689691e+00 -9.55794573e-01 -3.21005076e-01
2.49857441e-01 5.18457115e-01 2.31856421e-01 4.05808724e-02
-5.81146598e-01 4.91276592e-01 -1.34982729e+00 -1.37684703e-01
2.35651046e-01 3.16091120e-01 -5.35077453e-01 -5.90848401e-02
-1.27304399e+00 7.34906375e-01 4.38452631e-01 3.92221302e-01
-9.92834508e-01 -3.55406374e-01 -1.04386401e+00 -1.43890589e-01
6.33937240e-01 -8.08356941e-01 1.34377253e+00 -5.64168453e-01
-1.64588177e+00 5.99342585e-01 -5.82813323e-01 -2.39502534e-01
6.83133662e-01 -7.58524954e-01 -2.88526744e-01 3.08359504e-01
4.38429356e-01 7.91170180e-01 1.09512210e+00 -1.04563463e+00
-6.41279340e-01 -4.29905504e-01 -6.16497137e-02 4.08803850e-01
2.01361328e-02 -1.39991239e-01 -1.23460615e+00 -1.10335839e+00
4.14257973e-01 -1.05976486e+00 -2.63890088e-01 1.24508806e-01
-5.22769272e-01 -1.70552935e-02 5.70962310e-01 -9.36362922e-01
1.29794884e+00 -2.24579930e+00 5.26684940e-01 -2.25949679e-02
1.84546351e-01 1.57784924e-01 2.41676793e-02 4.17238414e-01
-4.23860140e-02 -4.99872684e-01 -8.93262550e-02 -9.06462014e-01
-9.21645164e-02 2.50200123e-01 -1.70606803e-02 4.71251279e-01
1.79501534e-01 9.01674986e-01 -7.30692387e-01 -6.68235123e-01
4.06436980e-01 6.14942729e-01 -4.54368055e-01 3.40611011e-01
9.22812615e-03 8.40741038e-01 -5.10102332e-01 5.99334776e-01
6.35233998e-01 -3.34912449e-01 9.31911767e-02 -2.97639847e-01
-8.06345046e-02 4.10317063e-01 -1.19297349e+00 2.09117174e+00
-3.82788360e-01 3.28110278e-01 -6.98869452e-02 -7.82672942e-01
7.41227567e-01 3.36180210e-01 5.15707910e-01 -7.58865654e-01
3.03245634e-01 2.13233456e-01 -2.81105489e-01 -5.63280821e-01
3.58949751e-01 9.20573547e-02 -2.50437677e-01 1.16071716e-01
9.44910124e-02 3.96063447e-01 1.24136545e-01 -1.21369176e-01
9.83223259e-01 6.67814970e-01 3.07665467e-01 7.71480501e-02
6.35782003e-01 -8.50061700e-02 1.02402782e+00 2.14007586e-01
-3.00444394e-01 8.42639804e-01 2.91193843e-01 -4.85783607e-01
-1.00972223e+00 -9.68881488e-01 3.09955776e-01 9.88778830e-01
3.57581019e-01 -6.00965679e-01 -8.03002357e-01 -3.91749740e-01
-2.21334975e-02 1.35945335e-01 -6.72178805e-01 -2.17335582e-01
-1.05814004e+00 -3.87245953e-01 4.52075750e-01 8.53214681e-01
7.96983719e-01 -9.76595104e-01 -8.46191764e-01 3.34837824e-01
-5.54951668e-01 -1.07604802e+00 -1.04195666e+00 -4.72153723e-02
-9.86870229e-01 -8.98190498e-01 -9.49966729e-01 -8.78932416e-01
6.70491219e-01 2.99202681e-01 7.52138734e-01 -1.46631628e-01
-1.31002665e-01 -1.73818380e-01 -4.58615541e-01 -9.35038626e-02
9.14539918e-02 1.74568102e-01 1.36225775e-01 1.21969812e-01
2.24052921e-01 -5.93592525e-01 -1.03722286e+00 6.91289246e-01
-6.34861529e-01 3.61975372e-01 9.81809914e-01 9.73255038e-01
9.06170368e-01 -2.26060182e-01 1.59707695e-01 -4.33882356e-01
1.24501087e-01 -1.98024526e-01 -3.35841000e-01 -4.11194973e-02
-7.33072683e-02 1.66440099e-01 5.59700310e-01 -6.01338148e-01
-1.09903216e+00 5.53065836e-01 -2.80311644e-01 -6.45304501e-01
4.60537188e-02 4.70154494e-01 -3.12036246e-01 2.81572342e-01
3.51918995e-01 3.52389365e-01 1.49290755e-01 -7.09900320e-01
6.66825995e-02 5.60162306e-01 8.30812871e-01 -4.64795679e-01
9.37116086e-01 4.77130830e-01 -7.23802000e-02 -6.06965005e-01
-9.63344216e-01 -5.32502294e-01 -7.64137924e-01 -3.30400616e-01
1.04238546e+00 -1.21504581e+00 -6.96895778e-01 7.34289765e-01
-1.18819427e+00 -4.05734539e-01 -4.86551458e-03 7.17059076e-01
-6.35159075e-01 3.63803118e-01 -8.90641272e-01 -4.42038894e-01
-3.84477019e-01 -1.23846757e+00 1.47175658e+00 1.28989354e-01
-2.25346446e-01 -3.52928877e-01 5.57717085e-02 6.70336843e-01
-2.09909827e-02 2.51000702e-01 1.63047880e-01 -3.46981250e-02
-4.43203032e-01 -2.02462658e-01 2.88893213e-03 2.44326621e-01
3.69333215e-02 -5.67161441e-01 -6.38636410e-01 -4.71306354e-01
9.60275084e-02 -3.23474079e-01 7.41276324e-01 4.46584970e-01
1.05809224e+00 -9.65519994e-02 -3.54059488e-01 6.33601785e-01
1.06570768e+00 7.85448849e-02 7.39573717e-01 3.82799417e-01
7.31145620e-01 3.83140028e-01 1.03652143e+00 4.48137760e-01
6.59284592e-01 1.01810038e+00 2.98079282e-01 2.34525353e-02
-2.03899920e-01 -6.03702724e-01 4.79851902e-01 9.53206062e-01
-3.69231552e-01 1.54648926e-02 -6.51542783e-01 3.02608699e-01
-2.08805203e+00 -8.31357718e-01 1.44798905e-01 2.10838795e+00
8.07793081e-01 3.80330294e-01 2.90568501e-01 1.68029279e-01
7.46126771e-01 4.48494852e-01 -5.59148431e-01 2.37266451e-01
2.58234680e-01 8.41994286e-02 4.03192341e-01 5.39457321e-01
-1.09207177e+00 8.82910311e-01 4.65630150e+00 8.40698600e-01
-1.02874887e+00 2.30527848e-01 4.84750956e-01 -3.20004404e-01
4.43198025e-01 -1.42749444e-01 -8.21891665e-01 7.52849519e-01
6.16106153e-01 3.04736346e-01 2.10438564e-01 7.66971231e-01
4.66668129e-01 -9.95412618e-02 -9.21040535e-01 1.11457920e+00
-4.70281625e-03 -8.45176399e-01 -2.85039723e-01 1.17228273e-02
4.11936074e-01 -1.16898969e-01 -2.49576822e-01 1.79790884e-01
-3.45938474e-01 -7.64325738e-01 1.03446734e+00 6.26255989e-01
8.51346016e-01 -8.78429353e-01 7.35781252e-01 5.68277001e-01
-1.73986530e+00 -4.85607982e-03 -2.70244300e-01 -9.00306702e-02
5.56790888e-01 6.67101204e-01 -2.40410984e-01 7.75990248e-01
9.91397917e-01 7.88999379e-01 -2.83618510e-01 6.89876556e-01
-5.36227643e-01 2.89248168e-01 -3.96191478e-01 3.06886733e-01
3.58157419e-02 9.13346037e-02 4.66439247e-01 9.17824984e-01
2.38141209e-01 2.47008786e-01 3.77546608e-01 2.75052309e-01
2.99980007e-02 -3.66580665e-01 3.28241512e-02 3.97471130e-01
5.18036067e-01 9.95908558e-01 -4.49564278e-01 -3.53465796e-01
-4.55443561e-01 1.46713948e+00 3.05347681e-01 1.73496753e-01
-1.20389366e+00 -4.32282805e-01 4.70787734e-01 2.49046326e-01
5.91084957e-01 -4.83822167e-01 1.87065694e-02 -1.24280095e+00
4.85899240e-01 -9.28477645e-01 2.88425475e-01 -6.90755069e-01
-9.31653738e-01 6.05886161e-01 2.11226027e-02 -1.44610023e+00
-3.70937169e-01 -3.84335577e-01 -2.81664073e-01 7.40045786e-01
-1.14988542e+00 -1.25677979e+00 -5.91940343e-01 5.51819623e-01
7.33482242e-01 2.07012698e-01 4.54913259e-01 4.79704559e-01
-6.17903709e-01 5.45024872e-01 -2.20815107e-01 4.09143090e-01
8.22118700e-01 -8.92303109e-01 7.21998870e-01 8.91669810e-01
-2.48952538e-01 5.56604803e-01 5.42211592e-01 -8.77755702e-01
-1.32801354e+00 -1.11529613e+00 1.07777584e+00 -2.06812114e-01
2.67518163e-01 -4.75333929e-01 -6.00901306e-01 7.55536973e-01
-1.85235560e-01 2.40913913e-01 2.64800161e-01 -3.52977008e-01
-1.23316690e-01 -2.34953523e-01 -7.59438753e-01 5.40999651e-01
1.30003190e+00 -4.78523761e-01 -7.60790825e-01 1.05722651e-01
7.69818723e-01 -1.04790163e+00 -9.97972846e-01 3.79686564e-01
8.10011804e-01 -9.34833944e-01 1.13280296e+00 -1.12347245e-01
6.15800917e-01 -5.88567615e-01 -1.22952566e-01 -9.93622541e-01
-3.53176802e-01 -7.66293705e-01 -4.70382273e-01 8.62647414e-01
3.63420621e-02 -4.47483987e-01 9.32227254e-01 1.25508174e-01
-2.20281258e-01 -1.21162820e+00 -1.07757616e+00 -7.97056019e-01
-3.24529111e-01 -3.80787134e-01 4.04029787e-01 3.79119277e-01
-3.19190502e-01 4.80317324e-01 -8.56898546e-01 2.51453519e-01
7.95047045e-01 3.68631817e-02 1.05294561e+00 -7.93092966e-01
-5.24712265e-01 9.79984105e-02 -3.54978710e-01 -1.87995219e+00
-2.14474201e-01 -3.12062085e-01 2.45197266e-01 -1.33034718e+00
1.00065298e-01 4.50745821e-02 -8.34253877e-02 4.09364879e-01
-3.09949785e-01 4.13005531e-01 3.07076097e-01 4.52031285e-01
-5.04549026e-01 7.81272769e-01 1.48860562e+00 3.46710458e-02
-2.39227206e-01 8.29487741e-02 -1.97105378e-01 6.86189294e-01
7.44254351e-01 -5.27036428e-01 -3.42536062e-01 -6.67509794e-01
-1.97976083e-01 4.17849660e-01 6.47555351e-01 -1.33172309e+00
2.82693475e-01 2.74111219e-02 6.85024559e-01 -1.08395255e+00
8.24272394e-01 -6.71076417e-01 3.26577812e-01 8.65779340e-01
-1.01768672e-01 2.64271319e-01 -4.62422445e-02 5.64152658e-01
-3.56703162e-01 1.13353670e-01 6.60740316e-01 -1.39377475e-01
-7.69080818e-01 5.27314723e-01 -1.02927037e-01 3.56667908e-03
9.36575353e-01 -2.48351097e-01 1.05631193e-02 -3.05392802e-01
-8.90034318e-01 4.06965911e-01 3.14164221e-01 4.99074399e-01
7.26218939e-01 -1.54564345e+00 -5.58809102e-01 1.09591492e-01
1.21531533e-02 4.66651231e-01 6.04707181e-01 1.01775289e+00
-6.93833828e-01 4.54937190e-01 -1.51117176e-01 -8.06655526e-01
-1.19200706e+00 3.81087661e-01 2.47790724e-01 -5.15837014e-01
-8.34211528e-01 8.40560257e-01 1.59423053e-01 -2.29464516e-01
3.39841366e-01 -2.60711491e-01 2.45598465e-01 -1.58949181e-01
5.22370398e-01 3.87188762e-01 -1.62214443e-01 -9.05742764e-01
-6.17576659e-01 7.88893819e-01 -5.15583195e-02 -2.59754539e-01
1.34860885e+00 -2.31316790e-01 1.42022222e-01 3.00609916e-01
1.31325555e+00 -1.61088035e-01 -1.75436604e+00 -3.31918687e-01
-3.20880234e-01 -5.64591825e-01 -2.72664934e-01 -4.74316806e-01
-1.12632656e+00 8.74431014e-01 4.82245594e-01 -4.63971108e-01
1.40053344e+00 -2.08332259e-02 1.33454883e+00 1.71091706e-01
5.76995850e-01 -1.17368269e+00 4.61459994e-01 4.85510468e-01
1.01405799e+00 -1.04835188e+00 1.05160393e-01 -5.79675615e-01
-7.03433812e-01 9.47481215e-01 8.41865122e-01 -3.23239535e-01
5.61419845e-01 2.07615063e-01 2.48429514e-02 -7.47230127e-02
-5.98053098e-01 -9.87357646e-02 4.12527502e-01 2.78735280e-01
3.48762929e-01 -8.16369727e-02 -3.10127407e-01 6.55723631e-01
-2.54627645e-01 2.36482129e-01 -1.09868631e-01 1.17460573e+00
-3.51696223e-01 -1.09023178e+00 -4.86732244e-01 8.59500691e-02
-3.19235861e-01 2.00772241e-01 -1.17649119e-02 8.10267687e-01
2.58616626e-01 6.73411369e-01 -9.86004919e-02 -7.74535835e-01
5.51595747e-01 -2.25848943e-01 4.09871310e-01 -3.96545589e-01
-4.38293010e-01 5.31407654e-01 1.01384431e-01 -1.12627077e+00
-4.15561438e-01 -5.92542231e-01 -1.34088898e+00 -2.03313410e-01
-1.46602318e-01 -4.98256758e-02 3.13331693e-01 8.84455502e-01
5.41015804e-01 7.39841998e-01 5.27796924e-01 -1.31341219e+00
-4.81742263e-01 -1.04741335e+00 -1.50804296e-01 3.70002091e-01
3.50951135e-01 -9.39363122e-01 -1.01007996e-02 1.33236900e-01] | [7.216409206390381, -0.6071715354919434] |
611f8233-428c-432b-a82b-4bae83b5ba87 | rord-rotation-robust-descriptors-and | 2103.08573 | null | https://arxiv.org/abs/2103.08573v4 | https://arxiv.org/pdf/2103.08573v4.pdf | RoRD: Rotation-Robust Descriptors and Orthographic Views for Local Feature Matching | The use of local detectors and descriptors in typical computer vision pipelines work well until variations in viewpoint and appearance change become extreme. Past research in this area has typically focused on one of two approaches to this challenge: the use of projections into spaces more suitable for feature matching under extreme viewpoint changes, and attempting to learn features that are inherently more robust to viewpoint change. In this paper, we present a novel framework that combines learning of invariant descriptors through data augmentation and orthographic viewpoint projection. We propose rotation-robust local descriptors, learnt through training data augmentation based on rotation homographies, and a correspondence ensemble technique that combines vanilla feature correspondences with those obtained through rotation-robust features. Using a range of benchmark datasets as well as contributing a new bespoke dataset for this research domain, we evaluate the effectiveness of the proposed approach on key tasks including pose estimation and visual place recognition. Our system outperforms a range of baseline and state-of-the-art techniques, including enabling higher levels of place recognition precision across opposing place viewpoints and achieves practically-useful performance levels even under extreme viewpoint changes. | ['K. Madhava Krishna', 'Michael Milford', 'Sourav Garg', 'Satyajit Tourani', 'Kinal Mehta', 'Aniket Gujarathi', 'Udit Singh Parihar'] | 2021-03-15 | null | null | null | null | ['visual-place-recognition'] | ['computer-vision'] | [ 8.73572230e-02 -4.01100785e-01 -1.87485665e-01 -5.54260194e-01
-7.55859494e-01 -7.76534855e-01 1.11231828e+00 6.24966733e-02
-4.84844357e-01 2.26651028e-01 3.99547815e-01 2.53589034e-01
-3.94473448e-02 -4.73845810e-01 -6.58948898e-01 -4.92018402e-01
-2.16444768e-02 3.83896708e-01 4.03501123e-01 -3.14068079e-01
6.19922459e-01 1.08170450e+00 -1.90797460e+00 5.16989864e-02
2.36643776e-01 7.43570030e-01 -1.38065964e-01 3.98134321e-01
4.29793566e-01 2.87162513e-01 -3.44428867e-01 -1.17446244e-01
9.33521450e-01 5.53390533e-02 -4.36454326e-01 1.18138596e-01
1.26287544e+00 -3.46131086e-01 -5.24620950e-01 7.76728988e-01
5.54616809e-01 3.81086320e-01 5.35021842e-01 -1.04637146e+00
-7.70692289e-01 -2.44057566e-01 -4.65952754e-01 2.20414340e-01
7.97291517e-01 1.38011888e-01 1.09295630e+00 -1.23150754e+00
1.04536569e+00 1.33849931e+00 7.09047556e-01 1.50472313e-01
-1.34414506e+00 -3.80131900e-01 2.17776328e-01 2.42925555e-01
-1.55522275e+00 -8.06680381e-01 8.08609724e-01 -4.60785836e-01
1.33542895e+00 1.90134287e-01 5.43141067e-01 9.39242363e-01
7.81073943e-02 5.66335976e-01 1.15783060e+00 -6.18423879e-01
1.91409767e-01 1.07526422e-01 -2.84961104e-01 4.61530328e-01
2.23177090e-01 3.62935513e-01 -7.95938551e-01 -1.26807377e-01
8.53078246e-01 2.56022692e-01 -1.79194748e-01 -1.10950530e+00
-1.52970958e+00 6.29516065e-01 6.94034696e-01 1.35161996e-01
-2.03155518e-01 -8.63357782e-02 7.93172121e-02 3.08340788e-01
4.04979527e-01 6.86959326e-01 -3.16265911e-01 9.30430293e-02
-8.74721944e-01 4.66994077e-01 6.18317664e-01 1.16571331e+00
7.47842073e-01 -2.74559855e-01 -2.11381271e-01 6.65296793e-01
4.67179984e-01 5.03186226e-01 3.91944230e-01 -7.84130156e-01
3.93075436e-01 6.97068036e-01 9.24091488e-02 -1.27621269e+00
-4.46019977e-01 -2.36121103e-01 -1.43952459e-01 2.03661621e-01
3.64858508e-01 4.55681324e-01 -9.93813753e-01 1.51954103e+00
4.46377724e-01 1.90463150e-03 -2.15124041e-01 8.85168910e-01
5.95824897e-01 1.63798690e-01 -3.68442714e-01 3.26351792e-01
1.24350488e+00 -9.64591801e-01 -3.32407296e-01 -4.06903297e-01
3.32539409e-01 -1.06002581e+00 7.67975271e-01 2.85392344e-01
-7.47120619e-01 -5.91965020e-01 -1.26040649e+00 -4.42891479e-01
-5.02299488e-01 1.76131859e-01 4.79440778e-01 4.96543825e-01
-1.31267893e+00 4.54603851e-01 -9.01054502e-01 -9.24986303e-01
1.92851827e-01 5.20349979e-01 -8.14049542e-01 -2.91160852e-01
-4.74960685e-01 1.21752238e+00 2.45164093e-02 -3.61304581e-02
-6.77025735e-01 -4.74206060e-01 -1.17046750e+00 -3.67243737e-01
4.96999994e-02 -5.26161313e-01 9.56622005e-01 -5.00725687e-01
-1.40224481e+00 1.08075321e+00 -1.31921053e-01 -1.88572407e-01
7.60083318e-01 -4.60973382e-01 -1.58201098e-01 9.72937644e-02
1.18012331e-01 9.51810062e-01 8.50689292e-01 -1.08315241e+00
-4.21808541e-01 -9.16265130e-01 -8.16085711e-02 6.06544554e-01
-2.47254044e-01 1.81929097e-01 -4.55444843e-01 -4.26071018e-01
5.73009729e-01 -1.34880221e+00 -1.21739097e-01 3.85282010e-01
-1.11529119e-01 -3.48395407e-02 8.82980466e-01 -4.58562553e-01
5.45655668e-01 -2.09807277e+00 1.34664506e-01 2.59021461e-01
8.76435297e-05 1.30947292e-01 -1.40914276e-01 5.53881168e-01
-8.18879232e-02 -2.74329156e-01 1.89017385e-01 -3.96630704e-01
-3.13915238e-02 7.16802031e-02 -3.02335560e-01 1.00170743e+00
2.37305954e-01 6.88555181e-01 -8.92147303e-01 -1.18116379e-01
6.50198162e-01 6.37440145e-01 -5.98549724e-01 8.53558928e-02
2.27780744e-01 6.08628392e-01 -1.85020030e-01 7.50073016e-01
6.98911250e-01 2.08419085e-01 -1.13296937e-02 -3.40834528e-01
-3.67152452e-01 3.58933747e-01 -1.25864208e+00 1.99565113e+00
-3.65473390e-01 8.11256468e-01 -3.89755547e-01 -5.57943046e-01
1.31237495e+00 -1.35241240e-01 3.82019967e-01 -5.26437521e-01
1.52162075e-01 2.78527647e-01 -2.41066322e-01 -1.58585966e-01
7.67793834e-01 3.20756882e-01 1.27198204e-01 -2.66447198e-02
2.83665985e-01 -1.87496707e-01 -4.56857979e-02 -2.35092685e-01
1.12147260e+00 6.18298769e-01 6.30432010e-01 -1.78729609e-01
5.32338142e-01 -1.10679053e-01 3.17169845e-01 5.96920073e-01
-3.95376116e-01 8.62757087e-01 -1.43097579e-01 -6.96584702e-01
-1.24551141e+00 -1.12545633e+00 -4.04529572e-01 1.04532838e+00
3.19735080e-01 -4.66650218e-01 -2.53461331e-01 -3.50095272e-01
2.47006938e-01 3.19434851e-01 -6.10460699e-01 -1.96206748e-01
-5.76921046e-01 -3.38639259e-01 4.08008963e-01 6.74226642e-01
6.70937181e-01 -5.84144115e-01 -7.83549666e-01 -1.23046584e-01
2.17306688e-01 -1.30476773e+00 -2.97356814e-01 -7.09893331e-02
-7.72995949e-01 -9.57308650e-01 -6.97376311e-01 -7.22647130e-01
7.56723583e-01 7.62160301e-01 8.06274593e-01 -2.88109660e-01
-4.37925190e-01 6.78263485e-01 -3.73920918e-01 -2.55494922e-01
1.47384912e-01 2.77035479e-02 3.15974832e-01 6.23061843e-02
6.63142502e-01 -5.51253736e-01 -6.52482569e-01 5.50711751e-01
-5.83892405e-01 -3.41561973e-01 7.49652922e-01 7.76159406e-01
5.08614540e-01 -8.36381614e-01 -1.74838558e-01 -3.34857285e-01
2.15604663e-01 -1.82283998e-01 -7.32373834e-01 1.98949590e-01
-4.70318675e-01 2.43458599e-01 3.69950980e-01 -1.94230720e-01
-7.95591354e-01 4.54097718e-01 2.79927701e-01 -5.86132705e-01
-3.19341928e-01 1.66539028e-02 -1.67555198e-01 -6.92690730e-01
9.95081544e-01 2.57276297e-01 -2.06776485e-02 -4.76429969e-01
6.39481604e-01 4.14771408e-01 6.46278918e-01 -4.00626898e-01
1.10584259e+00 9.09503698e-01 3.26020449e-01 -1.04489076e+00
-4.66558039e-01 -9.44740772e-01 -1.30273902e+00 1.87675133e-02
6.36999607e-01 -1.06055856e+00 -2.42659807e-01 4.71499532e-01
-1.02401233e+00 3.83704185e-01 -1.79057747e-01 6.27211273e-01
-6.45204008e-01 4.34387952e-01 -1.52273220e-03 -4.94479150e-01
3.09215244e-02 -1.09386516e+00 1.48593056e+00 2.07558379e-01
-4.98270094e-02 -8.72131228e-01 5.12147486e-01 1.20019332e-01
3.71709406e-01 4.17338669e-01 1.48159623e-01 -8.16507697e-01
-9.05363739e-01 -5.60049891e-01 -7.06453472e-02 1.26691386e-01
5.50139397e-02 1.07383393e-01 -1.19236076e+00 -6.98023319e-01
-2.80646533e-01 -1.86652496e-01 8.52240562e-01 5.96752390e-02
5.47905743e-01 3.51115800e-02 -4.78586555e-01 1.04659414e+00
1.30934358e+00 1.90109275e-02 5.08240223e-01 8.25159013e-01
6.52237236e-01 4.87207770e-01 6.21385813e-01 3.35015863e-01
6.59206927e-01 1.30701709e+00 3.47111344e-01 3.90232615e-02
-1.28625691e-01 -5.01895308e-01 3.31616640e-01 4.49262917e-01
-3.12042326e-01 2.76730120e-01 -9.79091763e-01 6.69318855e-01
-1.74415112e+00 -8.01705718e-01 3.35121661e-01 2.58930659e+00
2.75612950e-01 -1.46498471e-01 2.50519626e-02 -1.20113321e-01
5.44148266e-01 3.09720814e-01 -4.34527963e-01 -2.34723568e-01
-7.91653693e-02 -1.54071087e-02 7.64418602e-01 2.50442713e-01
-1.46929812e+00 9.95552361e-01 6.55391645e+00 7.99201950e-02
-1.33681703e+00 -1.75038889e-01 1.31572951e-02 4.65670079e-02
7.95929059e-02 2.54503399e-01 -9.12384450e-01 -1.46379068e-01
5.18865228e-01 4.90567870e-02 4.30318773e-01 1.09053779e+00
-2.05644384e-01 -4.18532938e-02 -1.49199152e+00 1.32135618e+00
6.94620788e-01 -1.02351749e+00 -1.16587713e-01 1.33821636e-01
8.75051439e-01 3.94064099e-01 3.71081918e-01 5.14549315e-02
4.06810828e-02 -8.51936162e-01 6.18330121e-01 3.66242975e-01
5.68003118e-01 -3.90588671e-01 5.53796947e-01 2.70261113e-02
-1.23285878e+00 -1.00149311e-01 -5.50800264e-01 -2.54168302e-01
-2.36826539e-01 -8.72475728e-02 -1.17509830e+00 5.96091747e-01
6.93773687e-01 1.14840794e+00 -1.13783109e+00 1.33105910e+00
-2.00861171e-01 -1.60432413e-01 -6.40225232e-01 1.84038162e-01
7.46161714e-02 -1.95599031e-02 7.15229452e-01 1.05591142e+00
2.15317830e-01 -3.21625918e-01 2.85699040e-01 5.14147878e-01
1.36848301e-01 1.92518860e-01 -1.22119033e+00 4.91400242e-01
7.42783546e-01 1.38779116e+00 -7.14769781e-01 -1.88430026e-02
-4.73516732e-01 9.63363230e-01 3.64998162e-01 1.65030122e-01
-4.25561428e-01 -2.58642763e-01 8.31820250e-01 4.90626954e-02
4.06041473e-01 -6.33079827e-01 -1.24603562e-01 -1.47462356e+00
2.86434233e-01 -7.42399037e-01 1.25510514e-01 -5.70639729e-01
-1.01493478e+00 5.16617537e-01 2.47048125e-01 -1.66180122e+00
-3.44177037e-01 -8.86428475e-01 -4.57171112e-01 7.10021675e-01
-1.60104454e+00 -1.55657494e+00 -6.04746938e-01 6.58139348e-01
4.63585734e-01 -2.46860713e-01 8.05210948e-01 1.99840769e-01
-5.98755255e-02 7.75394738e-01 2.26980180e-01 -1.16487883e-03
1.12469745e+00 -1.03886092e+00 7.36584067e-01 1.01372027e+00
5.94886959e-01 1.04120898e+00 6.29138172e-01 -3.59391838e-01
-1.59762383e+00 -9.30799127e-01 7.63853014e-01 -1.03304458e+00
3.66749406e-01 -5.31481147e-01 -5.29555917e-01 8.87346685e-01
-1.17621325e-01 4.33619112e-01 4.94514912e-01 1.28335878e-01
-7.93439567e-01 -2.00844735e-01 -1.18254423e+00 5.43941855e-01
1.26156437e+00 -7.66076148e-01 -7.85709143e-01 3.84959966e-01
1.61544055e-01 -7.50815630e-01 -7.50432312e-01 3.73038083e-01
7.78395355e-01 -9.06942368e-01 1.21816635e+00 -4.88468200e-01
5.51961064e-02 -5.42812526e-01 -5.22921324e-01 -1.26769125e+00
-4.27388042e-01 -5.56210637e-01 2.68362850e-01 9.85491097e-01
-2.72203069e-02 -6.03633821e-01 5.40583432e-01 4.72742915e-01
-1.16971299e-01 -3.78951997e-01 -1.06777632e+00 -9.36446130e-01
-2.21561119e-01 -4.70898822e-02 4.21881944e-01 9.33340490e-01
-7.54607618e-02 9.13487151e-02 -3.17508906e-01 2.47932971e-01
5.88801086e-01 -1.69555210e-02 1.25194991e+00 -1.19140649e+00
-7.93603994e-03 -2.86270171e-01 -1.41817868e+00 -1.12840390e+00
3.65880169e-02 -9.97872591e-01 1.44363595e-02 -1.25548053e+00
3.77963968e-02 8.47208221e-03 -1.44177139e-01 4.90994006e-01
1.72478929e-01 4.49738532e-01 5.29345870e-01 4.33126748e-01
-6.31795287e-01 5.36345482e-01 8.77385437e-01 6.63326308e-02
-2.64501244e-01 -2.62161642e-01 -3.93800408e-01 7.86534071e-01
3.41759712e-01 -1.78708136e-01 -2.54105866e-01 -5.07208049e-01
8.29837918e-02 -5.54602802e-01 5.90034604e-01 -1.26202226e+00
4.01794970e-01 1.88748469e-03 7.48865604e-01 -5.98275185e-01
5.86969972e-01 -8.33888650e-01 -2.60186568e-03 2.32530966e-01
-2.19838768e-01 5.56923449e-01 1.93552896e-01 5.94511032e-01
-1.70160100e-01 3.10707778e-01 7.59849906e-01 -1.19142886e-02
-1.07314277e+00 1.89139277e-01 2.28535295e-01 -1.79809123e-01
1.19379103e+00 -5.42543650e-01 -4.69314784e-01 -2.84339845e-01
-5.09942293e-01 -5.98911941e-02 1.05438781e+00 9.80112255e-01
7.53757238e-01 -1.50914037e+00 -6.31848037e-01 6.20654285e-01
6.24501705e-01 -1.51692748e-01 -1.57419220e-01 9.03721452e-01
-7.58463562e-01 5.88236451e-01 -6.26689136e-01 -9.90263999e-01
-1.31047702e+00 4.75429833e-01 3.02283168e-01 8.18625614e-02
-7.27617443e-01 7.58113265e-01 3.39674540e-02 -8.32081974e-01
2.99177459e-03 -4.34464872e-01 -1.75228417e-01 1.07507957e-02
4.44828898e-01 4.12747025e-01 3.61070037e-01 -1.19343555e+00
-7.37954140e-01 9.94022608e-01 -1.42448947e-01 -1.56289160e-01
1.38371289e+00 -2.35187873e-01 1.18538305e-01 3.38411689e-01
1.25780177e+00 5.97907603e-02 -1.36387956e+00 -4.55705285e-01
9.78962779e-02 -1.09328735e+00 -3.76741104e-02 -6.35633171e-01
-6.13072693e-01 8.19608808e-01 9.50749457e-01 -3.40481967e-01
7.84058034e-01 -1.85833592e-02 1.32384405e-01 7.01646090e-01
7.19014466e-01 -9.35028911e-01 2.22499622e-03 6.43697977e-01
1.03215206e+00 -1.37445176e+00 4.18761104e-01 -2.38992989e-01
-4.02004242e-01 1.20769095e+00 5.14335632e-01 -6.07995152e-01
5.24761856e-01 -2.32969552e-01 2.00703770e-01 -1.68416068e-01
-3.96474808e-01 -1.55824333e-01 7.81409144e-01 8.80223632e-01
3.12617540e-01 -1.30468190e-01 -6.91926032e-02 -2.79929459e-01
-3.69093359e-01 -3.58286262e-01 3.03536952e-01 1.17659390e+00
-2.96448708e-01 -1.10528374e+00 -5.89974225e-01 5.28446957e-03
-4.70407754e-02 1.09830052e-01 -6.48259819e-01 9.12146688e-01
-3.55047956e-02 6.32644117e-01 1.03779994e-01 -4.39347059e-01
5.40328443e-01 -1.75053868e-02 8.65116835e-01 -6.14638925e-01
-3.06909233e-01 -2.59152278e-02 -1.74469069e-01 -9.23675656e-01
-6.12643778e-01 -1.12073183e+00 -6.40340805e-01 -3.88222374e-02
-2.01446503e-01 -5.68431079e-01 9.15492296e-01 8.81624281e-01
4.90492672e-01 9.17164758e-02 6.20490432e-01 -1.53036296e+00
-8.57814729e-01 -7.64920950e-01 -3.75231892e-01 5.54914117e-01
4.42792356e-01 -1.13270891e+00 -1.78688809e-01 -1.17569692e-01] | [7.8513922691345215, -2.150010108947754] |
44c6ebde-79c6-4a65-b158-6c2a4399503e | mwptoolkit-an-open-source-framework-for-deep | 2109.00799 | null | https://arxiv.org/abs/2109.00799v2 | https://arxiv.org/pdf/2109.00799v2.pdf | MWPToolkit: An Open-Source Framework for Deep Learning-Based Math Word Problem Solvers | Developing automatic Math Word Problem (MWP) solvers has been an interest of NLP researchers since the 1960s. Over the last few years, there are a growing number of datasets and deep learning-based methods proposed for effectively solving MWPs. However, most existing methods are benchmarked soly on one or two datasets, varying in different configurations, which leads to a lack of unified, standardized, fair, and comprehensive comparison between methods. This paper presents MWPToolkit, the first open-source framework for solving MWPs. In MWPToolkit, we decompose the procedure of existing MWP solvers into multiple core components and decouple their models into highly reusable modules. We also provide a hyper-parameter search function to boost the performance. In total, we implement and compare 17 MWP solvers on 4 widely-used single equation generation benchmarks and 2 multiple equations generation benchmarks. These features enable our MWPToolkit to be suitable for researchers to reproduce advanced baseline models and develop new MWP solvers quickly. Code and documents are available at https://github.com/LYH-YF/MWPToolkit. | ['Ee-Peng Lim', 'Dongxiang Zhang', 'Yan Wang', 'Bing Tian Dai', 'Yunshi Lan', 'Qiyuan Zhang', 'Lei Wang', 'Yihuai Lan'] | 2021-09-02 | null | null | null | null | ['math-word-problem-solving', 'math-word-problem-solving', 'math-word-problem-solving'] | ['knowledge-base', 'reasoning', 'time-series'] | [-5.19155800e-01 -2.22481459e-01 -3.91925305e-01 -3.39950353e-01
-9.93852556e-01 -8.86496663e-01 2.81686187e-01 -1.50597870e-01
-2.89878268e-02 8.86085927e-01 1.11855805e-01 -4.77312446e-01
-8.77842456e-02 -1.00115430e+00 -5.90076208e-01 -4.18104887e-01
3.45365405e-01 7.13083625e-01 -2.15540111e-01 -2.33295470e-01
2.37699151e-01 3.36808890e-01 -1.26255798e+00 4.86270964e-01
1.14299905e+00 7.82651484e-01 1.20765872e-01 4.42480743e-01
-5.81951559e-01 7.49058425e-01 -5.15133679e-01 -6.66090071e-01
4.92446959e-01 -2.21641734e-01 -1.06033194e+00 -6.10659480e-01
5.24185896e-01 -1.10379718e-01 -3.26630950e-01 8.09874356e-01
5.76114714e-01 2.60169823e-02 3.76628041e-01 -1.62499785e+00
-7.71450162e-01 9.44794416e-01 -3.71353924e-01 9.44612473e-02
7.50019193e-01 3.92118394e-01 1.34804773e+00 -9.38516498e-01
6.08211100e-01 1.23164701e+00 7.80031562e-01 4.77417827e-01
-1.21797538e+00 -1.13388968e+00 1.57669023e-01 6.64200008e-01
-1.55454433e+00 -2.86078155e-01 7.05934465e-01 -1.70547232e-01
1.57231486e+00 4.33561265e-01 8.02783966e-01 1.16594076e+00
2.15939552e-01 9.59289968e-01 1.06995904e+00 -1.64705247e-01
5.20134792e-02 -1.87847127e-05 3.06066513e-01 6.84171736e-01
1.03777602e-01 -2.27618128e-01 -6.69836104e-01 -2.01450869e-01
5.46860814e-01 -6.65531576e-01 -2.26753592e-01 -6.28869236e-02
-1.01588750e+00 9.21243429e-01 1.70686424e-01 3.14254045e-01
9.65514854e-02 1.36714816e-01 2.02265829e-01 2.72107363e-01
4.12246823e-01 9.83316600e-01 -8.03516626e-01 -5.18484950e-01
-9.97594476e-01 1.08073270e+00 1.33132410e+00 1.18800986e+00
6.12818599e-01 1.12433285e-01 -6.79485500e-02 7.16689169e-01
8.69542062e-02 3.84261012e-01 5.20131826e-01 -1.22531223e+00
8.09914052e-01 5.91027856e-01 5.81705943e-02 -1.31356311e+00
-4.98999506e-01 -3.25637609e-01 -5.23843050e-01 -3.50536615e-01
3.62612993e-01 -3.83226752e-01 -6.52198911e-01 1.59466445e+00
2.69804657e-01 3.25846314e-01 -7.14766467e-03 1.01239276e+00
1.32758737e+00 1.43610096e+00 -1.93586886e-01 -1.82529822e-01
1.02467275e+00 -1.44398284e+00 -6.09469652e-01 -3.27554911e-01
9.68127012e-01 -9.54400659e-01 1.11127114e+00 7.19610035e-01
-1.61183631e+00 -2.39244729e-01 -8.22124004e-01 -5.73812068e-01
-6.26979172e-01 -1.71252899e-02 9.76724565e-01 4.18513119e-01
-9.52634633e-01 5.85115790e-01 -7.02088475e-01 -4.53756228e-02
1.02605797e-01 3.70763719e-01 8.42144266e-02 -1.95903167e-01
-1.53559113e+00 1.22585905e+00 3.88516515e-01 2.42919594e-01
-4.96647269e-01 -1.43873942e+00 -9.62809265e-01 2.22940743e-01
3.77574861e-01 -8.46347988e-01 1.65043497e+00 -7.47461677e-01
-1.70120084e+00 6.60681903e-01 -3.96524698e-01 -3.69622022e-01
4.19058383e-01 -2.19447494e-01 -3.33993793e-01 -3.18472177e-01
-7.58945346e-02 5.78342021e-01 4.81233597e-01 -8.50364685e-01
-2.15242118e-01 3.15005660e-01 2.41959199e-01 2.60705680e-01
-4.01813947e-02 3.92665446e-01 -5.88790178e-01 -6.30027890e-01
-3.68724883e-01 -5.92429996e-01 -2.14515537e-01 -3.88919622e-01
-3.34219187e-01 -3.99063528e-01 3.56368065e-01 -5.56275904e-01
1.49689198e+00 -1.71483278e+00 6.61453784e-01 -5.49421040e-03
2.62481421e-01 2.83648223e-01 -5.11633933e-01 8.42567086e-01
-3.19388986e-01 9.69053954e-02 -1.83465108e-01 -4.48568910e-01
4.99345720e-01 2.99520493e-01 -6.21371984e-01 9.90354419e-02
2.33823463e-01 1.17615795e+00 -7.15987444e-01 -4.90785539e-01
1.93385556e-01 2.16936767e-01 -7.22741723e-01 1.50759339e-01
-6.52091980e-01 1.30551457e-01 -3.07632953e-01 6.03041172e-01
9.40290630e-01 -2.36733377e-01 1.12429984e-01 -3.64862271e-02
-3.77107888e-01 7.23555326e-01 -1.57530260e+00 2.02350640e+00
-5.02435267e-01 6.67434692e-01 -7.73792109e-03 -8.81244183e-01
6.70877635e-01 8.29445273e-02 4.44856524e-01 -5.68916380e-01
2.05453709e-01 2.00760320e-01 -4.84176427e-02 -3.67980093e-01
5.94040811e-01 1.93696842e-01 -9.99050885e-02 4.75603729e-01
2.31464222e-01 -6.07201397e-01 9.81157482e-01 2.20852658e-01
9.55832124e-01 1.67127743e-01 1.26432568e-01 -3.29374760e-01
6.12481892e-01 2.76107401e-01 8.61323774e-01 4.62836564e-01
2.62554973e-01 6.90044940e-01 6.54581487e-01 -7.59491205e-01
-7.76917756e-01 -9.90659773e-01 -2.99707681e-01 9.91466880e-01
-1.68853596e-01 -1.12058902e+00 -5.87777197e-01 -1.25739723e-01
3.53790000e-02 9.68604565e-01 -2.15551600e-01 1.01533942e-01
-8.51875305e-01 -1.07130396e+00 6.63137674e-01 6.09579027e-01
5.57061136e-01 -9.31717813e-01 -1.55596584e-01 1.84630781e-01
-5.72760463e-01 -1.30925572e+00 -2.49669835e-01 -8.59308764e-02
-5.46255887e-01 -9.46765304e-01 -5.15915453e-01 -8.36067438e-01
4.01166558e-01 -4.01523858e-02 1.49692321e+00 7.56200925e-02
-5.00177264e-01 2.33417362e-01 -1.91071019e-01 -3.38513136e-01
-1.39665335e-01 4.79603499e-01 1.97348967e-02 -6.20129287e-01
4.77798462e-01 -4.78619784e-01 -2.49653235e-01 -5.84980994e-02
-5.09169042e-01 4.78925049e-01 1.91152275e-01 4.57651675e-01
2.58019865e-01 -1.16218023e-01 2.56671101e-01 -6.10553205e-01
8.03625762e-01 -5.11420846e-01 -1.15332425e+00 2.02552885e-01
-4.90033835e-01 -6.91709742e-02 7.57145166e-01 -5.12032807e-01
-9.39068615e-01 -1.65909991e-01 -4.06296223e-01 -1.37135237e-01
7.80497640e-02 8.42761576e-01 -1.00855432e-01 -2.02983797e-01
4.40357983e-01 2.20783696e-01 -5.66988707e-01 -3.34530592e-01
6.39943719e-01 2.25710899e-01 1.66638523e-01 -1.10321891e+00
9.57896054e-01 -3.90322387e-01 -2.64891535e-01 -4.93186742e-01
-1.07967234e+00 -2.08138913e-01 -2.78022498e-01 8.18510875e-02
5.08645177e-01 -8.95656765e-01 -7.27736771e-01 4.95558918e-01
-1.49955368e+00 -7.14324415e-01 9.43799876e-03 2.51888782e-01
-3.67967933e-01 2.88174033e-01 -9.62694764e-01 -2.61410296e-01
-5.59938610e-01 -1.29964423e+00 8.29765439e-01 3.25488985e-01
-5.60650587e-01 -1.06539965e+00 4.32999581e-01 6.64216638e-01
5.86953282e-01 2.47694328e-02 1.00361586e+00 -3.77683818e-01
-5.83301783e-01 -7.22885951e-02 -4.05638814e-01 1.44974425e-01
-2.44558558e-01 3.07555079e-01 -8.13559890e-01 7.92490542e-02
-1.21300273e-01 -5.84778190e-01 5.07205367e-01 3.85236084e-01
1.36104167e+00 -5.16765475e-01 -2.30964825e-01 1.00144339e+00
1.42727280e+00 -1.83040947e-01 5.79257846e-01 3.13976109e-01
5.14061451e-01 1.96987778e-01 2.87385106e-01 3.52042973e-01
6.03214502e-01 5.74759364e-01 8.77654627e-02 2.54337732e-02
2.95294046e-01 9.74269062e-02 4.57226008e-01 1.22978854e+00
-1.12574756e-01 -2.89685965e-01 -1.28894436e+00 3.78627121e-01
-1.98189890e+00 -5.90099812e-01 -3.23281676e-01 1.61452568e+00
1.44100475e+00 -1.78234354e-01 -9.96860415e-02 -1.99175984e-01
5.69015704e-02 1.96999297e-01 -3.61208171e-01 -8.34744573e-01
-1.39172431e-02 7.59016156e-01 1.70351550e-01 6.95656896e-01
-9.58986878e-01 1.33241355e+00 6.29701090e+00 1.02481115e+00
-1.05419278e+00 -6.15120493e-02 3.39095294e-01 -5.48179269e-01
-1.66422546e-01 -1.40836939e-01 -1.04328048e+00 3.61181587e-01
1.02083194e+00 -7.86747575e-01 7.85363436e-01 8.72845769e-01
5.98068163e-02 9.77232680e-02 -1.22335410e+00 1.07384121e+00
6.49217665e-02 -1.64897740e+00 -1.08128451e-01 -3.84132117e-01
8.84686708e-01 5.19861206e-02 1.03232794e-01 7.66225755e-01
4.74777669e-01 -1.41025329e+00 3.89761925e-01 1.50822684e-01
1.71275169e-01 -7.87281811e-01 6.32648110e-01 2.54458874e-01
-1.30401051e+00 6.06044829e-02 -4.64919776e-01 -5.83856404e-01
1.48831829e-01 5.32730579e-01 -2.50635952e-01 7.51646578e-01
7.71449566e-01 8.86369228e-01 -3.15265059e-01 1.00960743e+00
-4.71322656e-01 5.80953419e-01 -4.34624374e-01 -4.13954966e-02
4.01609451e-01 -3.17136168e-01 2.70198286e-01 1.29287612e+00
2.63035268e-01 3.46836627e-01 1.05684377e-01 1.44071114e+00
-1.00169718e-01 3.19256842e-01 -3.04671168e-01 -1.99446976e-01
4.24498290e-01 1.68498993e+00 -3.36237282e-01 -2.10356757e-01
-6.00973606e-01 6.01335764e-01 5.14711738e-01 3.75025362e-01
-1.46386409e+00 -3.16497028e-01 1.02607191e+00 -1.79315627e-01
8.10551494e-02 -6.27977550e-01 -3.84254456e-01 -1.59342086e+00
-4.01601056e-03 -1.33433425e+00 3.82882744e-01 -9.11320925e-01
-1.28730667e+00 2.81436950e-01 3.06186408e-01 -6.28324926e-01
-1.00368962e-01 -8.10037374e-01 -6.81208491e-01 1.00949657e+00
-1.60952282e+00 -8.43447745e-01 -3.95721406e-01 2.96857923e-01
6.90241039e-01 6.68206289e-02 1.07242060e+00 4.02365357e-01
-1.05813694e+00 5.53584158e-01 -4.63921539e-02 1.55334309e-01
7.61374831e-01 -1.23080087e+00 7.09833145e-01 6.58779383e-01
1.42495021e-01 8.61497283e-01 7.50025988e-01 -3.39722395e-01
-1.90965819e+00 -8.77786279e-01 1.06187046e+00 -5.61783016e-01
1.12456691e+00 -6.34699225e-01 -9.34671342e-01 8.84425044e-01
7.34961152e-01 -3.81738752e-01 6.38003886e-01 2.26972789e-01
-3.89328301e-01 -6.27322420e-02 -7.78419018e-01 6.67074382e-01
6.97644114e-01 4.41127792e-02 -7.67252505e-01 7.68205822e-01
6.39968634e-01 -1.18719137e+00 -9.77585375e-01 3.65762293e-01
4.04230177e-01 -6.04369342e-01 1.05252361e+00 -5.82638681e-01
9.37527835e-01 2.79820953e-02 2.30311304e-01 -1.29721045e+00
-3.77823770e-01 -6.85929894e-01 -4.86495703e-01 1.25944221e+00
7.40913272e-01 -8.84091496e-01 7.03082502e-01 1.07098317e+00
-9.13185924e-02 -1.11228430e+00 -7.64019549e-01 -6.78588867e-01
7.75336087e-01 -6.91392541e-01 7.69096553e-01 1.12467301e+00
1.78072572e-01 1.81169242e-01 -1.58563182e-01 7.27296993e-02
2.49958038e-01 6.59096003e-01 9.47205007e-01 -7.20231175e-01
-6.04815423e-01 -7.79104292e-01 9.91359055e-02 -9.80076194e-01
4.29297030e-01 -1.23338342e+00 -2.43209526e-01 -1.69987404e+00
2.75431037e-01 -1.21495731e-01 4.55811769e-02 8.49660635e-01
4.50159460e-02 -2.38795653e-02 3.43565285e-01 -2.20187292e-01
-4.33224022e-01 5.35036922e-01 1.12210846e+00 -2.18524501e-01
-2.27848485e-01 -4.04007822e-01 -7.33263552e-01 7.29494154e-01
1.23650765e+00 -3.83406103e-01 -3.00477773e-01 -1.00470805e+00
7.42647171e-01 -3.13748002e-01 1.27061918e-01 -9.55500484e-01
2.78964370e-01 -5.42996466e-01 1.83099389e-01 -5.37019253e-01
4.33859766e-01 -3.38704467e-01 6.39593154e-02 2.33289763e-01
-1.66773334e-01 2.17601836e-01 7.65983164e-01 -4.38388497e-01
-9.85200852e-02 -4.22337413e-01 4.39050466e-01 -3.07998866e-01
-5.80637097e-01 2.39401817e-01 -4.94319737e-01 4.66141880e-01
9.51020181e-01 2.56051838e-01 -6.10055327e-01 -2.14252681e-01
-1.98703915e-01 7.31423318e-01 1.94872126e-01 4.75042015e-01
4.58025217e-01 -1.16307092e+00 -7.22341061e-01 -9.92392376e-02
-2.55063415e-01 3.00139397e-01 1.57020554e-01 7.01264858e-01
-9.89186704e-01 7.91530252e-01 -5.19192554e-02 -1.00895151e-01
-1.11837733e+00 5.03845394e-01 4.50436473e-01 -6.14116013e-01
-6.00442410e-01 1.00115705e+00 -7.75045231e-02 -7.50081539e-01
3.14687788e-01 -9.56062198e-01 1.83925167e-01 -1.05653383e-01
6.78038776e-01 4.70276803e-01 4.16294597e-02 -1.03576951e-01
-2.56417930e-01 6.90451384e-01 -2.26423398e-01 2.30103806e-01
1.60936844e+00 4.47105408e-01 -4.35386539e-01 3.02556545e-01
8.74381602e-01 -1.30074546e-01 -4.85122472e-01 -2.58052535e-02
-2.21333697e-01 -1.36097834e-01 4.54705693e-02 -8.94017339e-01
-1.31035376e+00 7.46736169e-01 -1.75952747e-01 -6.79513142e-02
1.15445340e+00 -2.65878856e-01 1.22628605e+00 5.88030875e-01
2.72835284e-01 -1.06553495e+00 -2.58114696e-01 8.94053996e-01
1.07023406e+00 -8.69056821e-01 3.99557948e-01 -6.86945438e-01
-5.00156939e-01 1.23953843e+00 8.52270067e-01 -9.79993492e-02
3.02052051e-01 7.38468945e-01 -3.97717804e-02 3.09637450e-02
-1.08342409e+00 5.58793619e-02 3.84377122e-01 1.82135612e-01
6.85277343e-01 -1.96086019e-01 -4.41441566e-01 1.09040618e+00
-7.34841168e-01 2.31875092e-01 3.80542457e-01 8.12093794e-01
-2.35931631e-02 -1.46059394e+00 -2.91507512e-01 2.16515511e-01
-2.40895793e-01 -4.21795577e-01 -5.78777969e-01 9.13301826e-01
-1.92501117e-02 7.49967515e-01 -1.85180023e-01 -6.65409863e-02
1.41648486e-01 1.95501819e-01 8.57877791e-01 -5.92677176e-01
-8.46433818e-01 -3.53218645e-01 2.97369987e-01 -8.70036066e-01
-3.71033163e-03 -4.34380621e-01 -1.21329224e+00 -7.72597134e-01
-1.04920551e-01 1.81983128e-01 3.43954086e-01 9.48852301e-01
3.83462697e-01 3.44829351e-01 5.33985645e-02 -8.38047683e-01
-3.26519191e-01 -7.02437758e-01 -1.03084058e-01 -2.05304712e-01
-2.87102968e-01 -5.20517945e-01 -2.61573881e-01 -2.30936751e-01] | [9.74284553527832, 7.435670375823975] |
5fd64f65-a532-420b-ac8c-7672b15e81e3 | moet-interpretable-and-verifiable | 1906.06717 | null | https://arxiv.org/abs/1906.06717v4 | https://arxiv.org/pdf/1906.06717v4.pdf | MoËT: Mixture of Expert Trees and its Application to Verifiable Reinforcement Learning | Rapid advancements in deep learning have led to many recent breakthroughs. While deep learning models achieve superior performance, often statistically better than humans, their adoption into safety-critical settings, such as healthcare or self-driving cars is hindered by their inability to provide safety guarantees or to expose the inner workings of the model in a human understandable form. We present Mo\"ET, a novel model based on Mixture of Experts, consisting of decision tree experts and a generalized linear model gating function. Thanks to such gating function the model is more expressive than the standard decision tree. To support non-differentiable decision trees as experts, we formulate a novel training procedure. In addition, we introduce a hard thresholding version, Mo\"ETH, in which predictions are made solely by a single expert chosen via the gating function. Thanks to that property, Mo\"ETH allows each prediction to be easily decomposed into a set of logical rules in a form which can be easily verified. While Mo\"ET is a general use model, we illustrate its power in the reinforcement learning setting. By training Mo\"ET models using an imitation learning procedure on deep RL agents we outperform the previous state-of-the-art technique based on decision trees while preserving the verifiability of the models. Moreover, we show that Mo\"ET can also be used in real-world supervised problems on which it outperforms other verifiable machine learning models. | ['Sarfraz Khurshid', 'Rishabh Singh', 'Mladen Nikolic', 'Kaiyuan Wang', 'Andrija Petrovic', 'Marko Vasic'] | 2019-06-16 | null | null | null | null | ['game-of-go'] | ['playing-games'] | [-3.13196182e-02 9.03154671e-01 -3.21043611e-01 -3.41405034e-01
-2.82736301e-01 -5.68421423e-01 6.09722912e-01 8.11657012e-02
-3.53594571e-01 1.00135636e+00 -4.92979318e-01 -6.09451652e-01
-1.34541452e-01 -1.14605534e+00 -1.04168129e+00 -5.96702754e-01
7.22267777e-02 7.92388380e-01 3.96773666e-01 -1.53758496e-01
-2.60621756e-01 4.00823474e-01 -1.55382180e+00 1.57646820e-01
8.06478262e-01 1.16527867e+00 -9.11061391e-02 4.38400745e-01
3.23939055e-01 1.13080359e+00 -3.19192886e-01 -2.95158863e-01
3.24749917e-01 -2.10001424e-01 -8.68795276e-01 -1.63788348e-01
-1.31260753e-01 -5.90205371e-01 7.38694221e-02 9.18107033e-01
2.25112990e-01 1.97516512e-02 5.64144135e-01 -1.63172781e+00
-3.42830420e-01 7.67155170e-01 1.51272193e-01 -5.47777712e-01
2.36669078e-01 3.73467565e-01 1.16469741e+00 -8.71251300e-02
7.26479411e-01 1.24965060e+00 6.08417928e-01 9.78501320e-01
-1.38690007e+00 -7.42440820e-01 3.13323230e-01 3.75820130e-01
-1.01058650e+00 -2.91320503e-01 5.52447557e-01 -4.40997601e-01
1.13933587e+00 -1.11108266e-01 5.25482416e-01 1.31152594e+00
5.16855180e-01 9.44921255e-01 1.45506001e+00 -5.86215615e-01
8.17323685e-01 3.92587602e-01 1.81626920e-02 1.04172873e+00
1.59355968e-01 6.84746802e-01 -3.39957297e-01 -1.92965623e-02
6.78082585e-01 -1.51133865e-01 -2.38114782e-02 -5.72060287e-01
-8.44019532e-01 1.16871762e+00 3.58851194e-01 2.56774813e-01
-3.07917088e-01 4.94184047e-01 4.74238962e-01 5.71232378e-01
-8.32148567e-02 3.34072918e-01 -6.63927197e-01 1.31815359e-01
-5.44037163e-01 2.78881878e-01 9.93808270e-01 6.45675540e-01
7.14620590e-01 8.19766819e-02 7.71382526e-02 -9.13513303e-02
5.76919794e-01 3.23933899e-01 4.82565939e-01 -1.33707678e+00
-2.21203133e-01 5.93632400e-01 3.77292991e-01 -4.69217658e-01
-4.89742041e-01 -5.85803032e-01 -7.31618643e-01 9.11339462e-01
2.32169762e-01 -2.65596062e-01 -7.00116098e-01 2.06568289e+00
3.02682221e-01 1.17281817e-01 2.25955829e-01 5.27275324e-01
2.64269918e-01 5.30665934e-01 1.19396344e-01 -1.56862006e-01
9.75622475e-01 -7.16093838e-01 -3.87509853e-01 -5.78316189e-02
8.27705145e-01 2.17230916e-01 8.33718479e-01 9.44443464e-01
-9.02769089e-01 -6.87028766e-01 -1.21585989e+00 1.60534248e-01
-2.46413797e-01 -2.41670888e-02 7.30820656e-01 6.87237442e-01
-1.28913701e+00 7.45002568e-01 -9.87283230e-01 -8.45623687e-02
4.88194972e-01 8.36624146e-01 -4.48811769e-01 3.04751903e-01
-1.39291441e+00 1.28129935e+00 5.07297337e-01 -7.98013210e-02
-1.45621037e+00 -2.24041596e-01 -1.11876667e+00 4.94312719e-02
5.54484844e-01 -9.73719716e-01 1.63389361e+00 -1.18953013e+00
-2.17713499e+00 8.84326398e-01 -7.55544752e-02 -1.19314742e+00
8.85772467e-01 -7.86265284e-02 -2.93244034e-01 1.08653039e-01
-9.47792782e-04 8.41350257e-01 1.08909452e+00 -1.19189739e+00
-8.24962258e-01 -1.68903202e-01 5.52038491e-01 -4.03410077e-01
1.95050478e-01 -1.48998499e-01 4.01932716e-01 -1.53610602e-01
-4.77670074e-01 -1.03505623e+00 -3.71762872e-01 2.54961789e-01
-2.60113209e-01 -5.65205514e-01 6.43148899e-01 -1.45656869e-01
9.22508478e-01 -1.89270782e+00 -1.61352083e-02 2.77706325e-01
3.34726542e-01 2.53294528e-01 7.71527141e-02 2.80978709e-01
-1.08364895e-02 9.66497064e-02 -3.25592458e-01 -3.25550586e-01
6.27500713e-01 7.99687028e-01 -6.12639725e-01 3.50504667e-01
3.41787934e-01 1.02771676e+00 -9.17924225e-01 -4.68046248e-01
4.67018247e-01 1.79591894e-01 -5.34115255e-01 3.04170281e-01
-6.25803232e-01 4.68734741e-01 -6.57593727e-01 4.58477400e-02
1.99268609e-01 -2.57599235e-01 3.18077207e-01 4.72818375e-01
1.83668777e-01 3.18737805e-01 -1.12514269e+00 1.35375047e+00
-7.47262359e-01 2.96251297e-01 3.05874851e-02 -1.42994702e+00
9.10214543e-01 5.08545220e-01 1.60651840e-02 -4.82235938e-01
2.71915883e-01 9.33598205e-02 5.71001768e-02 -3.90690148e-01
-8.67028460e-02 -7.05096126e-01 -1.19628824e-01 5.07303476e-01
1.06657542e-01 -1.64130837e-01 -5.90346754e-03 -1.11442037e-01
1.14240623e+00 6.21372104e-01 4.08621073e-01 -1.60703287e-01
6.61769152e-01 9.13253240e-03 7.61558115e-01 8.76919687e-01
-2.21944422e-01 -8.12434405e-02 6.45330429e-01 -5.51778138e-01
-6.31432593e-01 -7.67211795e-01 4.10074741e-02 9.65600073e-01
-1.38139233e-01 -3.28329980e-01 -8.05808663e-01 -1.02310765e+00
-5.11777960e-02 1.19813490e+00 -9.06216741e-01 -4.11775053e-01
-4.20395583e-01 -3.04766834e-01 7.79218793e-01 7.19113946e-01
5.43227315e-01 -1.26505244e+00 -1.35637105e+00 3.90846670e-01
2.84327805e-01 -1.03238654e+00 3.20082754e-01 6.92345023e-01
-7.77727306e-01 -1.00848913e+00 8.42259079e-02 -5.47110319e-01
3.96238506e-01 -6.23242736e-01 1.10899222e+00 2.36905571e-02
1.33560002e-01 3.34101260e-01 -2.92613477e-01 -5.15435159e-01
-1.04048026e+00 -1.36003777e-01 2.83769608e-01 8.18281341e-03
1.61147654e-01 -6.10770285e-01 -2.50673532e-01 2.75008678e-01
-9.09765184e-01 9.84824747e-02 4.24640566e-01 9.39626038e-01
4.41184282e-01 4.01108831e-01 7.89879620e-01 -1.05229425e+00
4.51692671e-01 -1.34159863e-01 -9.06391025e-01 2.68722862e-01
-8.97452712e-01 5.61401784e-01 1.14988196e+00 -3.11294287e-01
-9.35990095e-01 1.29887342e-01 -1.95446163e-01 -2.14904174e-01
-4.72166061e-01 3.38577002e-01 -1.98133647e-01 2.64758747e-02
5.72255194e-01 1.51312174e-02 -3.37358844e-03 -8.54300559e-02
4.00668979e-01 4.44945067e-01 3.94503683e-01 -7.36343086e-01
6.14739060e-01 4.46449906e-01 2.38422737e-01 -1.61495328e-01
-6.06829345e-01 4.44466531e-01 -4.13406402e-01 4.71456647e-02
8.28960359e-01 -5.92237055e-01 -1.27918422e+00 1.60178661e-01
-1.14023376e+00 -9.57569599e-01 -6.68432534e-01 2.51834631e-01
-9.72713053e-01 2.92401165e-01 -6.38409972e-01 -1.10150909e+00
-1.46050066e-01 -1.10123456e+00 8.38596880e-01 -1.09408885e-01
-4.32139367e-01 -1.06330085e+00 1.31932031e-02 1.16740257e-01
1.39321536e-01 2.99916923e-01 1.21418631e+00 -9.55999970e-01
-2.95587808e-01 -2.83168495e-01 2.97711283e-01 6.92273021e-01
-1.33150727e-01 1.44502863e-01 -1.00723326e+00 -1.48740545e-01
1.96708739e-01 -5.16593814e-01 6.01329088e-01 2.83196688e-01
9.99047458e-01 -4.44197387e-01 -3.31602663e-01 7.34607279e-02
1.06691921e+00 3.76475722e-01 4.52693492e-01 4.49954957e-01
-1.18312305e-02 5.75575650e-01 4.17277515e-01 2.93798685e-01
4.16409314e-01 4.88752484e-01 7.46442199e-01 2.31470644e-01
3.41853499e-01 -5.09400964e-01 6.80661142e-01 2.11463973e-01
-2.59034876e-02 -6.76345453e-02 -7.87905872e-01 1.90362945e-01
-2.19954371e+00 -8.53030622e-01 4.08996433e-01 2.23653150e+00
8.51754367e-01 6.91979289e-01 2.07583252e-02 3.59328747e-01
4.28829402e-01 -3.49061668e-01 -7.64380276e-01 -7.91453540e-01
7.09534436e-02 6.64826572e-01 1.85412884e-01 6.33173227e-01
-1.04326963e+00 9.11120594e-01 6.52095842e+00 8.55208218e-01
-1.28408265e+00 2.48834014e-01 4.69949037e-01 4.15120572e-01
-1.78413302e-01 2.27449194e-01 -6.51327491e-01 2.18257785e-01
9.95938838e-01 1.02061808e-01 2.44724959e-01 1.11024582e+00
2.34721169e-01 -2.04682499e-01 -1.60723925e+00 6.15239143e-01
-5.08881509e-01 -1.19361341e+00 -1.79359183e-01 -6.83775917e-02
3.64291161e-01 -1.06772967e-01 -5.99612407e-02 6.67010069e-01
1.01185119e+00 -1.35380864e+00 1.01922965e+00 1.81041837e-01
5.43104529e-01 -6.57687902e-01 7.31779516e-01 8.25619280e-01
-7.08084106e-01 -3.54782432e-01 -1.40955582e-01 -4.68212157e-01
-1.78975150e-01 2.72716403e-01 -7.59925365e-01 6.65673435e-01
3.07939678e-01 6.26753807e-01 -1.55743569e-01 5.09073913e-01
-9.48273659e-01 7.79079437e-01 -5.42122543e-01 -6.27163947e-02
3.65969062e-01 2.16985960e-02 2.08319098e-01 1.02716196e+00
-3.41468416e-02 -6.36606514e-02 1.17506780e-01 1.03464925e+00
1.08593196e-01 -4.99355018e-01 -5.63706696e-01 2.37825140e-01
7.89405406e-02 1.07339978e+00 -3.99680078e-01 -5.24298489e-01
-1.10392325e-01 7.56968319e-01 4.03073817e-01 2.61192650e-01
-1.07711816e+00 -7.95515776e-02 3.42707932e-01 -5.93702383e-02
4.45250839e-01 6.07420970e-03 -1.07795276e-01 -1.03327036e+00
-2.39566509e-02 -1.02464247e+00 2.12513357e-01 -7.90691793e-01
-1.04865980e+00 8.43559742e-01 3.41679789e-02 -1.00248289e+00
-7.32165694e-01 -8.58604848e-01 -5.29493034e-01 4.69262898e-01
-1.56231785e+00 -1.13626003e+00 2.35602602e-01 7.26058781e-01
2.00242266e-01 5.25820628e-02 1.11823130e+00 -2.85720706e-01
-5.37609816e-01 4.93254244e-01 -2.53940850e-01 -5.92885427e-02
3.79956365e-01 -1.23643637e+00 8.08307007e-02 6.11549199e-01
-2.95235477e-02 3.68868828e-01 8.36494863e-01 -2.89601058e-01
-1.23362219e+00 -1.08839738e+00 7.17666864e-01 -3.38432640e-01
7.73806751e-01 -3.23759973e-01 -6.67060673e-01 1.11856055e+00
2.18241096e-01 -1.74387917e-02 3.91793042e-01 -1.38033375e-01
-4.18989062e-01 -2.91879803e-01 -1.30078101e+00 3.44837099e-01
7.79349923e-01 -4.37248081e-01 -9.62930739e-01 7.96035305e-02
6.16159856e-01 -1.01554237e-01 -7.30044842e-01 5.05462408e-01
4.41445410e-01 -1.13657200e+00 3.45313340e-01 -9.34220731e-01
3.59096557e-01 -4.94825423e-01 7.08115846e-02 -1.18380523e+00
-9.19510275e-02 -9.08943236e-01 -3.80154073e-01 6.02620542e-01
3.77291560e-01 -1.21708632e+00 6.51719272e-01 8.51791382e-01
-8.18534344e-02 -9.18027997e-01 -1.32258284e+00 -8.93828511e-01
2.79879332e-01 -7.45092332e-01 5.04071951e-01 6.51755273e-01
1.41209021e-01 3.50193918e-01 -3.42386395e-01 4.31126244e-02
5.13129413e-01 2.66148388e-01 5.18106699e-01 -1.28090847e+00
-8.31951559e-01 -5.03088951e-01 -3.28071386e-01 -1.04238355e+00
7.29774296e-01 -7.72526324e-01 2.47240424e-01 -1.40134108e+00
-2.43373036e-01 -6.80366874e-01 -3.32484126e-01 1.12403345e+00
4.39526916e-01 -1.56737164e-01 1.05413005e-01 -1.23081617e-01
-6.60645247e-01 5.40974796e-01 9.16047394e-01 -1.48222014e-01
-5.25959246e-02 2.14752078e-01 -4.57303375e-01 1.03347015e+00
8.53137493e-01 -5.55385768e-01 -2.89052933e-01 -9.24768522e-02
5.39073944e-01 1.27452046e-01 5.55792868e-01 -9.40164387e-01
2.30166525e-01 -1.52618378e-01 1.01411395e-01 2.63288647e-01
1.55573383e-01 -1.04342926e+00 2.36463144e-01 1.06242859e+00
-5.65724492e-01 -2.57111132e-01 4.38307412e-02 4.13299650e-01
-1.46763355e-01 -4.03158247e-01 7.16012239e-01 -7.60486051e-02
-6.26626909e-01 7.21750259e-02 -7.93283701e-01 -3.70007366e-01
1.23075414e+00 -1.12432480e-01 -1.11902331e-03 -4.88676071e-01
-9.11447763e-01 4.55524892e-01 4.36648220e-01 1.19726993e-01
4.46519315e-01 -8.08702767e-01 -5.33466339e-01 1.89952806e-01
-2.02424049e-01 1.64971858e-01 -8.33703876e-02 7.69276023e-01
-2.03799099e-01 3.44271541e-01 -2.90459424e-01 -3.50308478e-01
-9.15632784e-01 9.57087040e-01 6.47705197e-01 -6.90589309e-01
-5.49495220e-01 4.44286585e-01 5.47623515e-01 -4.68511820e-01
3.05255562e-01 -8.51765692e-01 -3.76684554e-02 -4.78656203e-01
2.64709711e-01 6.11697212e-02 -1.02963597e-01 -2.09242597e-01
-5.11957645e-01 5.07300198e-01 2.30248943e-01 -2.19965428e-01
1.19085789e+00 9.57569703e-02 2.37831511e-02 3.76934379e-01
5.29889524e-01 -3.40731382e-01 -1.37630117e+00 1.97719457e-03
5.54304291e-03 3.69367719e-01 -1.18889458e-01 -1.05985332e+00
-7.39436388e-01 1.09627664e+00 4.60071236e-01 2.55337328e-01
1.14930034e+00 -9.74790454e-02 6.44765317e-01 7.63165951e-01
8.09358895e-01 -1.05532575e+00 -3.61340970e-01 3.74076843e-01
7.06907272e-01 -1.24408829e+00 -2.75448531e-01 -3.16543989e-02
-6.40782595e-01 1.23492301e+00 4.28315431e-01 -4.24277067e-01
4.88690138e-01 4.55224305e-01 -1.40218556e-01 3.38416308e-01
-1.07986689e+00 -2.14638516e-01 -2.10893393e-01 6.95690453e-01
-1.51792020e-01 2.59025782e-01 2.58818977e-02 9.27407920e-01
-1.02266639e-01 6.27277017e-01 4.29238319e-01 9.49188471e-01
-5.17609417e-01 -1.42508662e+00 -3.69712301e-02 2.73157619e-02
-3.79837602e-01 1.29065916e-01 -1.95194900e-01 8.98538411e-01
4.96348411e-01 1.05800426e+00 -3.14264745e-01 -3.92908990e-01
1.33370578e-01 1.45012483e-01 7.56160319e-01 -4.61214572e-01
-7.71034062e-01 -3.18456292e-01 2.78804958e-01 -6.98821723e-01
-5.57728469e-01 -1.61411420e-01 -1.61717427e+00 -2.71059394e-01
-1.99218139e-01 3.33354652e-01 4.65815037e-01 1.30709434e+00
2.71475315e-01 3.90520930e-01 4.78674531e-01 -5.60073555e-01
-1.05065382e+00 -5.50320745e-01 -5.15653193e-01 6.85856789e-02
3.93339634e-01 -7.21986711e-01 -2.38869086e-01 -1.96006671e-02] | [4.195561408996582, 1.6402093172073364] |
551a0eb4-6af9-4969-832d-bc918553d0b1 | back-to-mlp-a-simple-baseline-for-human | 2207.01567 | null | https://arxiv.org/abs/2207.01567v3 | https://arxiv.org/pdf/2207.01567v3.pdf | Back to MLP: A Simple Baseline for Human Motion Prediction | This paper tackles the problem of human motion prediction, consisting in forecasting future body poses from historically observed sequences. State-of-the-art approaches provide good results, however, they rely on deep learning architectures of arbitrary complexity, such as Recurrent Neural Networks(RNN), Transformers or Graph Convolutional Networks(GCN), typically requiring multiple training stages and more than 2 million parameters. In this paper, we show that, after combining with a series of standard practices, such as applying Discrete Cosine Transform(DCT), predicting residual displacement of joints and optimizing velocity as an auxiliary loss, a light-weight network based on multi-layer perceptrons(MLPs) with only 0.14 million parameters can surpass the state-of-the-art performance. An exhaustive evaluation on the Human3.6M, AMASS, and 3DPW datasets shows that our method, named siMLPe, consistently outperforms all other approaches. We hope that our simple method could serve as a strong baseline for the community and allow re-thinking of the human motion prediction problem. The code is publicly available at \url{https://github.com/dulucas/siMLPe}. | ['Francesc Moreno-Noguer', 'Xavier Alameda-Pineda', 'Vincent Lepetit', 'Xi Shen', 'Yuming Du', 'Wen Guo'] | 2022-07-04 | null | null | null | null | ['human-pose-forecasting'] | ['computer-vision'] | [-5.66084944e-02 2.57002890e-01 -3.63276720e-01 -6.55795783e-02
-5.06193578e-01 -5.42567335e-02 5.51897943e-01 -2.89951175e-01
-7.41886914e-01 5.71590483e-01 3.51850390e-01 -2.75740653e-01
2.17238724e-01 -6.10425711e-01 -1.05477548e+00 -5.60086489e-01
-4.41751599e-01 5.16594946e-01 5.05632937e-01 -3.96465957e-01
-9.09455195e-02 2.97606468e-01 -1.35868800e+00 1.28174201e-01
1.56615958e-01 7.70793259e-01 -5.16273342e-02 9.91553724e-01
4.06460911e-01 1.20894575e+00 -1.83770120e-01 -4.69859362e-01
1.07560851e-01 -2.14907736e-01 -1.05096030e+00 -4.43194330e-01
2.76725322e-01 -4.45381731e-01 -9.00938988e-01 4.81786966e-01
7.43573487e-01 5.23004770e-01 3.94825190e-01 -1.15502906e+00
-3.09107989e-01 4.78378296e-01 -3.15448433e-01 2.23032475e-01
4.14423138e-01 3.05934846e-01 9.65528786e-01 -7.18885958e-01
7.82417178e-01 1.24796367e+00 1.21187913e+00 9.10410166e-01
-9.90232587e-01 -2.82168031e-01 8.62217173e-02 6.30789280e-01
-1.21599197e+00 -4.70808983e-01 6.98517442e-01 -3.87834430e-01
1.36694479e+00 1.53291654e-02 8.07339668e-01 1.61824524e+00
4.89015311e-01 1.22568250e+00 1.95842400e-01 -2.33671948e-01
2.01585200e-02 -6.39556527e-01 -3.13408934e-02 9.22494769e-01
-4.76595573e-02 1.24706842e-01 -5.26288271e-01 -3.77449878e-02
7.42422581e-01 8.82170629e-03 -2.19877973e-01 -2.54496664e-01
-1.43070042e+00 7.49871135e-01 5.44485331e-01 1.58292666e-01
-4.50171530e-01 8.99973571e-01 6.36150658e-01 3.24014314e-02
4.70235646e-01 -1.51213370e-02 -5.55622160e-01 -6.06321990e-01
-8.63731802e-01 6.71506763e-01 7.12031305e-01 5.56125820e-01
3.59019071e-01 1.12800069e-01 1.53876275e-01 6.47088170e-01
3.99302542e-01 1.88147634e-01 7.83621788e-01 -1.25676560e+00
4.59815651e-01 1.60402223e-01 6.04856759e-02 -1.09770072e+00
-8.83634448e-01 -2.57432669e-01 -1.02808774e+00 2.68219888e-01
5.99184036e-01 -4.19732928e-01 -9.72062469e-01 1.68593764e+00
3.25587243e-01 3.86515737e-01 -1.57284528e-01 1.11848533e+00
7.11115479e-01 7.59553552e-01 7.17835128e-02 3.06097448e-01
1.07338083e+00 -1.45582545e+00 -3.66281956e-01 -4.98367012e-01
8.24424922e-01 -4.51646000e-01 8.55145812e-01 3.97340536e-01
-1.29441273e+00 -8.18088710e-01 -9.59735811e-01 -3.32110703e-01
-1.33127332e-01 9.80377644e-02 7.20408976e-01 3.82358879e-01
-1.31777263e+00 1.24456406e+00 -1.57692301e+00 -5.63477695e-01
1.65403783e-01 4.88395005e-01 -3.66172612e-01 2.68814087e-01
-1.32086611e+00 1.01777434e+00 1.84113577e-01 4.22806799e-01
-8.63146305e-01 -3.83526325e-01 -9.69987631e-01 -1.62445053e-01
2.95495450e-01 -1.16154194e+00 1.39436984e+00 -6.44508958e-01
-1.78713262e+00 6.56888247e-01 -8.94376114e-02 -8.84756923e-01
7.13743985e-01 -8.47290516e-01 -1.75268337e-01 1.03732705e-01
-1.68638304e-01 8.70489240e-01 5.49602628e-01 -6.64014816e-01
-4.22533065e-01 -1.63040921e-01 -1.08629115e-01 9.10719931e-02
1.29983559e-01 -8.88314620e-02 -5.59140086e-01 -8.08789909e-01
-7.62352068e-03 -1.36157537e+00 -6.40194297e-01 -4.60755713e-02
-3.15790266e-01 -2.57825553e-01 4.31952387e-01 -9.33292866e-01
1.28699791e+00 -1.75486946e+00 5.48153400e-01 -7.91953132e-02
3.77604342e-03 3.70104164e-01 -7.67457113e-02 5.31789780e-01
6.98975800e-03 -1.62914410e-01 -2.64069468e-01 -5.36783457e-01
1.06185324e-01 3.10177863e-01 1.22458898e-02 7.54979491e-01
4.30890061e-02 1.15372896e+00 -8.35241795e-01 -3.50177705e-01
3.48443329e-01 6.41223490e-01 -6.28502369e-01 -3.92721519e-02
-3.48572969e-01 5.68828702e-01 -2.37375766e-01 3.58234227e-01
-1.50536913e-02 -5.31997442e-01 1.77810758e-01 -1.32441905e-03
1.32439196e-01 3.82249296e-01 -1.03834033e+00 2.04380703e+00
-2.54266322e-01 7.91211128e-01 -3.91663820e-01 -1.08638370e+00
6.34704769e-01 3.83245379e-01 7.19596267e-01 -4.12346452e-01
1.92731276e-01 8.62210169e-02 -1.39440671e-01 -5.96990347e-01
5.13459444e-01 1.47377074e-01 -5.98589964e-02 2.80324548e-01
1.87765211e-01 3.91702354e-01 4.91941012e-02 -2.09947631e-01
1.23705482e+00 7.30024755e-01 1.72071069e-01 1.47673979e-01
5.31185567e-01 -3.62608619e-02 6.47353709e-01 6.40356183e-01
-3.04466248e-01 7.09667027e-01 3.71501029e-01 -9.26829815e-01
-1.31646335e+00 -9.17609155e-01 5.31970739e-01 1.19974661e+00
-1.55651733e-01 -4.78914142e-01 -6.30061686e-01 -3.88923407e-01
-9.72888172e-02 3.81419122e-01 -5.73234499e-01 -9.09475982e-02
-1.33115292e+00 -6.91637576e-01 8.64830196e-01 9.38062608e-01
3.43730986e-01 -1.16511846e+00 -8.95287395e-01 4.17317927e-01
-2.78980970e-01 -1.07966101e+00 -3.19505095e-01 -1.09999321e-01
-9.92124081e-01 -8.53656650e-01 -1.09471655e+00 -7.36612916e-01
1.65834635e-01 -2.02205673e-01 1.16361666e+00 1.25340641e-01
-1.56067610e-01 2.49582052e-01 -3.45367670e-01 -1.04781631e-02
-2.18344718e-01 3.68148625e-01 1.62953541e-01 -3.43447983e-01
1.33011445e-01 -8.94363940e-01 -9.26161468e-01 6.94128200e-02
-4.69208270e-01 1.83943406e-01 4.61320281e-01 8.45142126e-01
4.02502477e-01 -5.28559208e-01 1.76289797e-01 -5.23558855e-01
2.24973589e-01 -5.01040399e-01 -2.62818485e-01 -7.23753795e-02
-3.90579462e-01 1.13011710e-01 5.61269760e-01 -5.28298974e-01
-7.25251675e-01 3.26485664e-01 -6.30676389e-01 -6.05021417e-01
-1.23049296e-01 5.37762880e-01 3.95281404e-01 -1.14088450e-02
6.68014884e-01 3.25686753e-01 9.96154640e-03 -5.87940872e-01
5.02392232e-01 1.60101578e-01 8.13211977e-01 -2.24876344e-01
5.08845031e-01 5.69476724e-01 1.53593913e-01 -7.52653301e-01
-5.58270693e-01 -3.88821393e-01 -8.49941552e-01 -3.87385309e-01
1.15927804e+00 -9.57888484e-01 -9.23165798e-01 7.00113773e-01
-1.25229537e+00 -8.93741786e-01 -9.33478773e-02 7.60553479e-01
-9.63565588e-01 6.41274393e-01 -1.16231310e+00 -8.21390271e-01
-4.49210763e-01 -7.25199163e-01 1.06969368e+00 1.88408922e-02
-5.55310011e-01 -1.06801736e+00 3.10548455e-01 4.75730777e-01
1.97646663e-01 5.11967003e-01 5.14638126e-01 -4.78423744e-01
-4.52998519e-01 -2.75320858e-01 4.37107682e-01 1.55830041e-01
-2.73746550e-01 -6.87372610e-02 -6.48775816e-01 -2.84672707e-01
-3.46411973e-01 -4.22850758e-01 1.01828802e+00 7.16146350e-01
1.03613973e+00 -4.02961969e-01 -3.94456685e-01 7.23761141e-01
1.00350726e+00 -6.71935529e-02 7.95059800e-01 6.55392766e-01
9.64476705e-01 4.03075159e-01 2.98784763e-01 5.12166083e-01
7.31334388e-01 7.14478552e-01 3.97620648e-01 -7.27527868e-03
-1.89055398e-01 -3.48558068e-01 4.92347836e-01 9.27383423e-01
-8.22936416e-01 -2.43302539e-01 -1.12762368e+00 5.16808569e-01
-2.26112604e+00 -1.16941786e+00 -9.94345322e-02 1.85931468e+00
4.89956111e-01 2.29241580e-01 5.15521646e-01 1.21674359e-01
4.59183991e-01 5.80579996e-01 -5.32585740e-01 -2.91498482e-01
1.84289679e-01 8.80142450e-02 6.61484838e-01 4.74427998e-01
-1.39257610e+00 1.04460335e+00 6.56756639e+00 5.78680098e-01
-9.65699136e-01 2.13080943e-02 4.75179285e-01 -2.80557275e-01
2.42771640e-01 -2.53076792e-01 -6.99018776e-01 4.02199000e-01
1.21336257e+00 6.29983917e-02 4.01173145e-01 7.92822957e-01
3.40430170e-01 2.28025839e-01 -9.32361186e-01 9.85629082e-01
3.11751664e-02 -1.29060447e+00 -1.48887619e-01 -4.48415652e-02
4.57425714e-01 6.66068375e-01 -6.19993843e-02 3.60026240e-01
3.85467529e-01 -1.01214516e+00 8.42192173e-01 8.31747293e-01
1.83138654e-01 -5.47486424e-01 7.60062754e-01 5.02433181e-01
-1.31806219e+00 -4.33550254e-02 -3.62651139e-01 -4.06628758e-01
6.23642862e-01 2.52445936e-01 -4.73017246e-01 4.68017757e-01
8.32526088e-01 9.19952333e-01 -3.69866729e-01 9.33260262e-01
-3.12255889e-01 6.44848049e-01 -3.37110966e-01 -1.29185706e-01
3.88858199e-01 1.00407131e-01 5.17850935e-01 1.17733908e+00
2.59536296e-01 9.06164870e-02 1.02952123e-01 3.43638331e-01
5.28744496e-02 -1.62481859e-01 -4.68319476e-01 1.11009851e-01
-7.34102130e-02 7.79230654e-01 -5.87990761e-01 -2.65606403e-01
-5.02256870e-01 1.15369844e+00 4.83551651e-01 3.14974844e-01
-1.04046559e+00 -1.23382501e-01 6.93302333e-01 3.04304808e-02
5.33450961e-01 -6.89568877e-01 2.57653836e-02 -1.11670005e+00
1.18445203e-01 -6.57834530e-01 5.38386762e-01 -6.43137991e-01
-1.07493782e+00 4.51116115e-01 -1.38365641e-01 -1.28155589e+00
-8.74362648e-01 -7.98723936e-01 -6.34747386e-01 3.99904579e-01
-1.03548563e+00 -1.16083896e+00 -1.94963142e-02 5.03183961e-01
6.06959045e-01 9.79081914e-02 6.77969992e-01 5.04076123e-01
-5.47944188e-01 4.83920336e-01 7.48438668e-03 4.12256002e-01
3.98343414e-01 -1.03269625e+00 1.20544696e+00 6.07583404e-01
1.86679319e-01 2.39407197e-01 8.72516811e-01 -6.15400076e-01
-1.46046853e+00 -1.01361513e+00 1.13210940e+00 -4.45338160e-01
8.82627666e-01 -1.16131902e-01 -7.73538113e-01 1.18221319e+00
-2.53962493e-03 1.79538414e-01 4.72229868e-01 9.08948332e-02
1.28855065e-01 2.73722738e-01 -4.46881413e-01 8.65448117e-01
1.47204602e+00 -2.22740263e-01 -7.12644398e-01 3.69144112e-01
7.01060355e-01 -6.54462755e-01 -8.75574648e-01 5.94432294e-01
9.50931609e-01 -9.46737647e-01 1.27468920e+00 -9.39946711e-01
6.07109249e-01 1.95838569e-04 4.85660359e-02 -1.09268606e+00
-5.65794170e-01 -7.11725771e-01 -6.26853764e-01 4.47895855e-01
4.36356813e-01 -4.11172748e-01 1.27587950e+00 4.16767448e-01
-1.74035624e-01 -1.08201241e+00 -1.06496775e+00 -8.97900343e-01
1.31414175e-01 -6.33197129e-01 1.41924635e-01 6.55801058e-01
2.22525988e-02 2.30920747e-01 -1.04124343e+00 -6.21500313e-02
4.55171853e-01 -2.47679293e-01 1.03432584e+00 -9.25240457e-01
-5.61021268e-01 -4.03049886e-01 -7.71664858e-01 -1.42161524e+00
3.20668668e-01 -6.44622982e-01 4.79127876e-02 -1.70101559e+00
-2.24592373e-01 1.62000507e-01 -2.12244704e-01 5.54617405e-01
-2.88001075e-03 3.14725846e-01 3.47492963e-01 2.10730731e-01
-7.38170266e-01 7.47898638e-01 1.13047087e+00 -1.48464724e-01
-1.97361872e-01 2.97059834e-01 8.18574652e-02 1.11856520e+00
8.87479544e-01 -3.99248004e-01 -2.38888748e-02 -5.96485853e-01
1.97559834e-01 3.03073347e-01 7.27697253e-01 -1.48307240e+00
5.86657941e-01 1.14874117e-01 4.67133611e-01 -7.84858048e-01
6.49556816e-01 -2.90587664e-01 3.33021075e-01 1.00957429e+00
-2.16315389e-01 3.41603398e-01 1.70980588e-01 5.52047491e-01
-1.26694798e-01 8.15636385e-03 2.56041914e-01 -3.64364803e-01
-9.70721126e-01 4.09248352e-01 -6.12663209e-01 -1.04510114e-01
6.78997815e-01 -2.59265661e-01 -1.45012274e-01 -5.84416270e-01
-1.10120404e+00 2.15273708e-01 1.94877133e-01 4.62523103e-01
5.49649894e-01 -1.47359586e+00 -8.10892999e-01 -1.81403488e-01
-3.06044430e-01 -1.40006110e-01 4.50364530e-01 9.51978624e-01
-8.03382337e-01 5.45096517e-01 -2.56254345e-01 -6.24715149e-01
-1.10164416e+00 3.98589075e-01 3.94920439e-01 -4.33705270e-01
-1.16400576e+00 9.06282485e-01 -2.00842232e-01 -3.82714570e-01
3.44546318e-01 -3.33858699e-01 -1.40911877e-01 -2.85059929e-01
2.52461940e-01 7.52099872e-01 -2.09421262e-01 -7.81235337e-01
-5.73516488e-01 5.69136918e-01 3.13674808e-01 -1.54513016e-01
1.47678709e+00 -6.73616752e-02 3.01942021e-01 6.45533264e-01
1.24841833e+00 -5.15087068e-01 -1.46216214e+00 -1.61354232e-03
1.87090054e-01 -1.26169603e-02 -3.13483864e-01 -3.70490611e-01
-1.09454572e+00 8.16163480e-01 4.67881024e-01 -8.90410691e-02
9.26084340e-01 3.97699103e-02 1.21640408e+00 7.11055517e-01
4.17107671e-01 -1.07250082e+00 1.10351285e-02 7.63588309e-01
7.45250106e-01 -1.08833969e+00 3.42823341e-02 2.26571714e-03
-5.19973040e-01 1.13453662e+00 5.02291143e-01 -5.70607126e-01
6.61130846e-01 2.51938868e-02 6.14367388e-02 -7.75154978e-02
-9.52079892e-01 -1.66062549e-01 4.40462172e-01 3.15583020e-01
6.90282404e-01 -3.58095802e-02 -3.02560210e-01 4.87600029e-01
-4.88005102e-01 3.13369960e-01 3.14980179e-01 1.00279844e+00
-3.89716834e-01 -9.41156864e-01 -1.93671376e-01 2.28354603e-01
-6.33162200e-01 9.95605160e-03 -4.68581356e-02 9.31374550e-01
-7.43962883e-05 6.23857200e-01 -7.54349399e-03 -6.32545173e-01
3.99124026e-01 1.36922181e-01 5.56861222e-01 -1.32199720e-01
-4.69309330e-01 -8.91403183e-02 3.52074534e-01 -1.09513271e+00
-6.30043983e-01 -5.95510602e-01 -1.22815871e+00 -5.64349532e-01
1.57642961e-01 -2.57010877e-01 3.11923832e-01 8.86383355e-01
2.15301767e-01 4.79102284e-01 5.96940704e-02 -1.41650307e+00
-5.35963297e-01 -1.03780103e+00 -2.78610826e-01 3.36486101e-01
4.68663096e-01 -6.33255303e-01 -1.33110762e-01 2.73063660e-01] | [7.254481792449951, -0.15771237015724182] |
49189313-13d8-4d40-ae3e-0d5661b0fa95 | superadditivity-and-convex-optimization-for | null | null | https://ieeexplore.ieee.org/document/9804854 | https://evoid.de/kostrykin2023-accepted-manuscript.pdf | Superadditivity and Convex Optimization for Globally Optimal Cell Segmentation using Deformable Shape Models | Cell nuclei segmentation is challenging due to shape variation and closely clustered or partially overlapping objects. Most previous methods are not globally optimal, limited to elliptical models, or are computationally expensive. In this work, we introduce a globally optimal approach based on deformable shape models and global energy minimization for cell nuclei segmentation and cluster splitting. We propose an implicit parameterization of deformable shape models and show that it leads to a convex energy. Convex energy minimization yields the global solution independently of the initialization, is fast, and robust. To jointly perform cell nuclei segmentation and cluster splitting, we developed a novel iterative global energy minimization method, which leverages the inherent property of superadditivity of the convex energy. This property exploits the lower bound of the energy of the union of the models and improves the computational efficiency. Our method provably determines a solution close to global optimality. In addition, we derive a closed-form solution of the proposed global minimization based on the superadditivity property for non-clustered cell nuclei. We evaluated our method using fluorescence microscopy images of five different cell types comprising various challenges, and performed a quantitative comparison with previous methods. Our method achieved state-of-the-art or improved performance. | ['Karl Rohr', 'Leonid Kostrykin'] | 2022-06-23 | null | null | null | ieee-transactions-on-pattern-analysis-and-25 | ['cell-segmentation'] | ['medical'] | [-6.69740140e-02 -1.29699916e-01 1.39666930e-01 -1.03693634e-01
-9.98928010e-01 -1.00121450e+00 -6.22399189e-02 3.13985169e-01
-5.65963984e-01 6.43123746e-01 -3.12972546e-01 -2.60840473e-03
-9.04869288e-02 -5.72857738e-01 -6.61371410e-01 -1.24298096e+00
1.53509080e-01 6.45282090e-01 4.34442103e-01 1.94246069e-01
3.06370527e-01 7.50130653e-01 -9.46957171e-01 -2.31946826e-01
1.04328990e+00 6.75111651e-01 2.38935933e-01 7.75695562e-01
1.74887314e-01 -2.97886264e-02 -1.93473086e-01 4.98845838e-02
2.09413007e-01 -3.30133140e-01 -9.32554364e-01 3.37191433e-01
3.05527776e-01 -3.27370465e-02 2.45793074e-01 1.12656450e+00
5.97367465e-01 1.91699252e-01 8.09947491e-01 -1.20012546e+00
-4.87491727e-01 1.02015376e-01 -6.78703725e-01 2.48210162e-01
-3.49381179e-01 -1.14179142e-01 7.12552309e-01 -8.94897819e-01
8.34101260e-01 8.71315718e-01 8.12661469e-01 6.17950201e-01
-1.57598972e+00 -1.78068772e-01 1.25618607e-01 -4.13905233e-01
-1.69040251e+00 -3.40685010e-01 3.39030862e-01 -4.81799424e-01
7.17058778e-01 3.49547416e-01 6.60107017e-01 1.66925520e-01
1.89403772e-01 5.49715877e-01 1.12666011e+00 -2.56761879e-01
3.97062063e-01 -2.33539432e-01 3.56659174e-01 1.03001595e+00
3.11721742e-01 -3.91026616e-01 1.63802534e-01 -4.70031023e-01
8.89274955e-01 2.09813900e-02 -4.07466829e-01 -4.41899598e-01
-1.15854204e+00 6.99771762e-01 1.24283358e-01 4.47971493e-01
-8.90106708e-02 1.56997666e-01 1.14318226e-02 -4.33409572e-01
4.40088362e-01 1.15364939e-01 -3.70186895e-01 8.84771198e-02
-9.10612762e-01 3.79515976e-01 7.68518388e-01 5.96476078e-01
8.11470151e-01 -2.68843919e-01 -1.19731568e-01 6.99392974e-01
4.98581111e-01 4.63934869e-01 1.28376484e-01 -1.27276969e+00
-3.60696241e-02 7.90552914e-01 2.20499411e-01 -1.19868612e+00
-6.41709626e-01 -2.26558432e-01 -8.02092433e-01 7.77587434e-03
7.82626212e-01 -1.57757014e-01 -8.87665808e-01 1.80224133e+00
8.09182346e-01 2.84520388e-01 -1.38574943e-01 8.96950245e-01
5.74545383e-01 5.50559163e-01 -8.43909681e-02 -5.78875422e-01
1.07930481e+00 -6.80812418e-01 -7.11354375e-01 3.88768911e-01
7.62421906e-01 -5.28035581e-01 6.25954211e-01 8.23295675e-03
-1.22019160e+00 3.35231163e-02 -7.24407673e-01 -1.73401296e-01
-1.50337458e-01 2.72830785e-03 1.39520228e-01 6.45332038e-01
-1.26263022e+00 6.82305932e-01 -1.30341256e+00 -3.61452848e-01
4.45006192e-01 7.46065140e-01 -1.91475675e-01 3.85628253e-01
-2.73771018e-01 4.10984844e-01 3.26304406e-01 1.31378651e-01
-4.94905412e-01 -7.88129628e-01 -6.66048408e-01 -1.05618015e-02
2.56880671e-01 -7.27293611e-01 8.54513824e-01 -4.39776927e-01
-1.55732846e+00 1.11902499e+00 -5.02080739e-01 -2.32438073e-01
3.79572839e-01 2.36521140e-01 3.45842808e-01 3.13646615e-01
-3.14553156e-02 6.16748989e-01 3.05583894e-01 -1.42818058e+00
-3.57102633e-01 -5.29934525e-01 -3.69828075e-01 1.75491676e-01
-2.50899404e-01 -7.24057481e-02 -6.15909040e-01 -4.75679040e-01
2.32369944e-01 -1.20660651e+00 -4.46329594e-01 -1.43790469e-01
-4.27171707e-01 -1.90532401e-01 8.62698138e-01 -5.81754446e-01
1.22952271e+00 -2.07321763e+00 4.35892969e-01 5.15803993e-01
3.81816387e-01 2.58704796e-02 2.29127407e-01 1.62237436e-02
3.52019757e-01 5.56899130e-01 -6.76140964e-01 -4.99395847e-01
-1.34129569e-01 2.56165206e-01 1.00473776e-01 9.11957562e-01
-1.19829677e-01 9.55502093e-01 -8.26371372e-01 -8.16935241e-01
4.88512106e-02 5.96401215e-01 -6.42741859e-01 3.51597667e-02
3.75823677e-02 4.25161391e-01 -3.95638973e-01 6.99274302e-01
8.80184710e-01 -5.19281328e-01 3.85449678e-01 -3.06785285e-01
-2.74815023e-01 -6.69662058e-01 -1.36226690e+00 1.21019542e+00
2.21664667e-01 1.68801710e-01 6.94840133e-01 -1.04454708e+00
6.23310268e-01 1.39820591e-01 1.01127350e+00 9.36493278e-02
4.58097845e-01 5.48841596e-01 -3.08489799e-01 -1.65210605e-01
1.69629306e-01 -3.31765473e-01 1.42205358e-01 2.80216634e-01
3.94801376e-04 -2.22917035e-01 3.23125362e-01 8.80672261e-02
8.62751961e-01 9.78715718e-03 2.29209915e-01 -9.55357790e-01
5.19004345e-01 -1.96505532e-01 7.23154426e-01 4.58151162e-01
-2.84104884e-01 8.71057451e-01 4.36227947e-01 -1.83100119e-01
-8.46908092e-01 -8.79981756e-01 -2.99454361e-01 7.83993244e-01
4.81643170e-01 -1.77011639e-01 -1.34702325e+00 -5.76488435e-01
-1.60313517e-01 9.40722898e-02 -5.64390004e-01 1.89057350e-01
-7.02157557e-01 -1.36608112e+00 3.87845248e-01 4.30326909e-01
2.44955644e-01 -6.18331969e-01 -4.67941284e-01 2.06158876e-01
-2.91247487e-01 -1.25229096e+00 -9.11797643e-01 -5.96419200e-02
-1.09878731e+00 -1.12789726e+00 -8.79746675e-01 -9.52392817e-01
1.11868012e+00 1.59264147e-01 7.80550480e-01 4.19196188e-01
-3.13742727e-01 5.02243340e-01 1.52584612e-01 -6.02409802e-02
-2.47874334e-01 -1.21060642e-03 3.93975116e-02 2.45180786e-01
-2.10239708e-01 -3.54967058e-01 -7.12967336e-01 4.96548593e-01
-1.01156461e+00 -2.56740093e-01 5.38495556e-02 7.13997483e-01
1.49386632e+00 1.13243811e-01 3.40583950e-01 -8.44662309e-01
3.58779103e-01 -3.34007740e-01 -7.92793095e-01 2.71590114e-01
-3.93560648e-01 7.36281350e-02 5.42668879e-01 -2.77092516e-01
-9.21792686e-01 4.58622217e-01 5.98887131e-02 -3.94229442e-01
1.45553276e-02 2.79029489e-01 -1.28515780e-01 -5.48447609e-01
2.66997665e-01 1.58118188e-01 1.36481330e-01 -2.55139619e-01
1.66067898e-01 2.81981230e-01 6.17629826e-01 -7.60181487e-01
5.82252860e-01 1.24529016e+00 3.22940052e-01 -8.50143015e-01
-5.25233626e-01 -6.39384389e-01 -6.37003243e-01 -1.46476552e-01
1.07911777e+00 -3.97438228e-01 -1.07550907e+00 7.29146719e-01
-1.01115525e+00 -5.65304697e-01 -1.69369221e-01 2.13363111e-01
-7.47986138e-01 8.39734495e-01 -7.70268440e-01 -7.69030154e-01
-4.58063155e-01 -1.11588347e+00 1.08843052e+00 3.26337576e-01
8.86359662e-02 -1.33221078e+00 4.13660765e-01 2.95965254e-01
1.49110526e-01 7.94687748e-01 8.13390374e-01 -4.36424583e-01
-4.94416833e-01 -6.52240887e-02 -1.90745533e-01 7.36782551e-02
5.59074022e-02 5.26375532e-01 -4.98657674e-01 -5.16162753e-01
2.92598852e-03 3.45830806e-02 9.04766440e-01 9.24995661e-01
1.10832226e+00 -3.02285939e-01 -8.22288334e-01 8.38749945e-01
1.77347243e+00 2.09697746e-02 4.76514369e-01 2.28678156e-02
8.25660288e-01 4.84890729e-01 3.17123711e-01 4.10149783e-01
5.08321226e-01 5.95006585e-01 3.04168642e-01 -2.45230108e-01
1.99314296e-01 4.31727469e-01 1.08018160e-01 9.34990585e-01
-4.30561036e-01 -4.29315150e-01 -9.66993213e-01 8.32408071e-01
-2.13148499e+00 -7.18195617e-01 -3.99868816e-01 2.17334533e+00
9.02487993e-01 -3.07150930e-01 3.77894819e-01 -1.98951572e-01
9.58810627e-01 -4.99266267e-01 -4.94692147e-01 -1.92009509e-01
-1.99417457e-01 2.89585013e-02 5.98765135e-01 8.33747089e-01
-1.23195815e+00 7.40689456e-01 6.84668970e+00 9.57285106e-01
-7.97933757e-01 1.70525953e-01 7.88082302e-01 2.63379551e-02
-2.49833032e-01 -1.50713116e-01 -9.65694189e-01 5.27489901e-01
4.80207711e-01 -7.24118203e-02 2.54415572e-01 3.16608667e-01
1.80887818e-01 1.58739351e-02 -8.26298773e-01 6.96470082e-01
-2.99868435e-01 -1.42168856e+00 -1.39243960e-01 5.30808687e-01
1.15405297e+00 -6.52444959e-02 1.56829953e-02 -2.93104559e-01
3.44483495e-01 -1.00029027e+00 4.65995431e-01 5.43201506e-01
4.80207682e-01 -9.05507684e-01 6.79029226e-01 3.90897661e-01
-1.42490125e+00 1.03285164e-01 -3.13408762e-01 3.78870368e-01
2.89496571e-01 6.11719728e-01 -4.77973461e-01 3.68382096e-01
5.70871055e-01 4.99564648e-01 -2.00139537e-01 1.08842003e+00
4.83020991e-01 4.30233598e-01 -8.48102093e-01 1.25598446e-01
5.85214011e-02 -5.57790160e-01 7.80431688e-01 1.49177015e+00
2.63270915e-01 5.31548202e-01 4.26204413e-01 1.06388056e+00
-1.57578856e-01 2.92468071e-01 -2.42640331e-01 3.58917117e-02
3.22111458e-01 1.64594042e+00 -1.52857101e+00 -4.57386412e-02
-2.37626582e-01 6.62819743e-01 6.34641767e-01 4.19114053e-01
-8.89443696e-01 -2.06498578e-01 3.38732272e-01 1.01992756e-01
4.19994295e-01 -3.05160880e-01 -2.97748536e-01 -9.61818695e-01
-1.08597890e-01 -3.76236528e-01 4.29912090e-01 -3.12583447e-02
-1.17524207e+00 2.18374893e-01 -6.58485070e-02 -6.75049961e-01
1.43140525e-01 -5.52993834e-01 -5.69975793e-01 4.49165761e-01
-1.24424636e+00 -1.08008575e+00 -2.89134890e-01 5.11876106e-01
9.32860374e-03 4.43901896e-01 6.52410090e-01 2.44945347e-01
-7.60410428e-01 5.01753509e-01 3.74701023e-01 1.17440531e-02
3.77618968e-01 -1.42589676e+00 -3.17883193e-01 7.37974167e-01
-2.84134239e-01 6.78354979e-01 4.97744232e-01 -6.42989814e-01
-1.17593706e+00 -1.08092904e+00 5.05516171e-01 -4.32773530e-01
4.25139964e-01 -1.21352717e-01 -1.09959483e+00 5.33247054e-01
1.45715121e-02 3.61224413e-01 8.61246109e-01 -4.86371815e-01
3.46010476e-01 2.11617947e-01 -1.43588471e+00 4.97419357e-01
7.95050919e-01 -5.82697466e-02 1.34524167e-01 3.75378042e-01
4.70121294e-01 -6.46227181e-01 -1.11547875e+00 5.52299619e-01
4.12092209e-01 -6.95743084e-01 9.39085901e-01 -2.56190926e-01
-1.28331631e-01 -5.89957535e-01 -1.35467112e-01 -8.98633778e-01
-4.11816925e-01 -8.91754329e-01 -3.00523996e-01 1.23739529e+00
2.59134233e-01 -5.48744023e-01 8.01979303e-01 6.74014151e-01
-2.89940625e-01 -1.24982524e+00 -9.98206139e-01 -8.68904710e-01
3.84347975e-01 1.71984121e-01 1.01937458e-01 9.07999396e-01
-4.88468707e-02 -1.38380498e-01 1.77493393e-01 3.07420731e-01
9.44060266e-01 2.20142066e-01 4.02039140e-01 -1.12984705e+00
-1.36767983e-01 -5.59375048e-01 -2.49636635e-01 -9.65699911e-01
2.87017375e-01 -9.71657276e-01 2.18928427e-01 -1.49378824e+00
7.60696888e-01 -3.73688310e-01 -2.55826563e-01 5.45336604e-01
-3.19568187e-01 5.04035652e-01 3.83575261e-02 2.60206074e-01
-7.84083128e-01 1.63416445e-01 1.41213810e+00 -2.26262659e-02
-4.97121155e-01 -1.67973876e-01 -6.01699352e-01 1.01536930e+00
8.55744123e-01 -4.87847298e-01 -6.80687949e-02 -1.28692076e-01
2.68135190e-01 -2.63552964e-01 4.38857853e-01 -6.70876145e-01
3.92701030e-01 -2.10258678e-01 9.01795030e-02 -4.80845600e-01
1.64894894e-01 -7.24466681e-01 1.90815821e-01 4.56256002e-01
1.25112191e-01 -3.38980615e-01 2.60227591e-01 6.70257866e-01
5.74483015e-02 -3.32233906e-01 1.26293623e+00 -7.10582808e-02
-1.85910687e-01 3.74934703e-01 -3.87737036e-01 3.18634659e-01
1.26448488e+00 -4.24124092e-01 -1.10769600e-01 2.30753813e-02
-8.80785823e-01 3.32000166e-01 8.24302256e-01 -3.77248049e-01
4.70467150e-01 -1.24381340e+00 -6.92423582e-01 -1.86392441e-01
-4.06361043e-01 3.86815190e-01 2.49784142e-01 1.35340333e+00
-7.76182950e-01 2.10529029e-01 5.43724969e-02 -9.09255028e-01
-1.56524050e+00 3.47564816e-01 6.99169874e-01 -4.88866001e-01
-1.19766742e-01 8.04866552e-01 3.55861992e-01 -5.86560249e-01
-2.42940813e-01 -6.51909173e-01 -1.90769404e-01 -8.30677822e-02
-1.61304567e-02 8.68769050e-01 -8.90184045e-02 -8.90936494e-01
-6.07581556e-01 1.22071016e+00 1.53891280e-01 2.00557411e-02
1.21260393e+00 -2.60000557e-01 -5.00281036e-01 1.72479108e-01
1.38188028e+00 9.69340056e-02 -1.40355647e+00 1.17994972e-01
-2.47734413e-01 -1.54495612e-01 5.34686968e-02 -2.91031867e-01
-1.26915920e+00 4.77593511e-01 4.97132480e-01 4.09559131e-01
9.76639152e-01 1.47726461e-01 9.55088735e-01 2.89597869e-01
1.91447243e-01 -1.23643112e+00 -3.67652136e-03 4.60755706e-01
4.10031348e-01 -9.63911772e-01 1.55649632e-01 -8.40185285e-01
-2.09906667e-01 1.01216710e+00 4.07412440e-01 -2.28961080e-01
7.31200516e-01 5.98234296e-01 -1.75362661e-01 -2.89769977e-01
-3.30918401e-01 -1.72028523e-02 2.46177822e-01 4.16828334e-01
2.92984664e-01 7.16052726e-02 -4.49665904e-01 7.44087577e-01
2.70430505e-01 -5.26129939e-02 3.34023833e-01 8.65784824e-01
-5.63229918e-01 -8.21308792e-01 -3.64192396e-01 7.54005983e-02
-9.01137352e-01 1.42709196e-01 -4.26930010e-01 6.55949354e-01
1.31421864e-01 7.62921989e-01 4.84284461e-02 2.13059232e-01
6.91776499e-02 -1.13133967e-01 6.98660970e-01 -4.27627355e-01
-2.99793690e-01 5.17453313e-01 -5.91616809e-01 -3.67234290e-01
-6.71371579e-01 -7.85815358e-01 -2.11487484e+00 -2.85350055e-01
-6.77738249e-01 -1.61095262e-02 2.75961071e-01 9.41242695e-01
5.70869625e-01 1.19794659e-01 4.41315532e-01 -9.69872713e-01
-2.11475164e-01 -3.76732409e-01 -6.58059537e-01 4.01732206e-01
2.53916621e-01 -4.49082673e-01 -4.84871715e-01 4.54249352e-01] | [14.478926658630371, -3.2107133865356445] |
7778100d-b4e0-4698-8e21-3da96d1825dd | scene-text-recognition-with-permuted | 2207.06966 | null | https://arxiv.org/abs/2207.06966v1 | https://arxiv.org/pdf/2207.06966v1.pdf | Scene Text Recognition with Permuted Autoregressive Sequence Models | Context-aware STR methods typically use internal autoregressive (AR) language models (LM). Inherent limitations of AR models motivated two-stage methods which employ an external LM. The conditional independence of the external LM on the input image may cause it to erroneously rectify correct predictions, leading to significant inefficiencies. Our method, PARSeq, learns an ensemble of internal AR LMs with shared weights using Permutation Language Modeling. It unifies context-free non-AR and context-aware AR inference, and iterative refinement using bidirectional context. Using synthetic training data, PARSeq achieves state-of-the-art (SOTA) results in STR benchmarks (91.9% accuracy) and more challenging datasets. It establishes new SOTA results (96.0% accuracy) when trained on real data. PARSeq is optimal on accuracy vs parameter count, FLOPS, and latency because of its simple, unified structure and parallel token processing. Due to its extensive use of attention, it is robust on arbitrarily-oriented text which is common in real-world images. Code, pretrained weights, and data are available at: https://github.com/baudm/parseq. | ['Rowel Atienza', 'Darwin Bautista'] | 2022-07-14 | null | null | null | null | ['scene-text-recognition'] | ['computer-vision'] | [ 8.07084367e-02 -1.71146259e-01 -2.02178076e-01 -2.79688567e-01
-1.29044187e+00 -3.66085887e-01 5.98818779e-01 -1.96360648e-01
-5.13369381e-01 4.65762526e-01 2.84469843e-01 -6.63427234e-01
3.99667203e-01 -6.38984799e-01 -8.02992105e-01 -5.54428756e-01
2.28878528e-01 4.45865363e-01 1.81882098e-01 -2.22311495e-03
5.37227333e-01 3.55308950e-01 -1.11746776e+00 3.67921770e-01
4.13066000e-01 7.41910100e-01 4.60669905e-01 9.70957339e-01
-4.52711254e-01 1.12230039e+00 -3.84798616e-01 -3.08702618e-01
9.24569666e-02 -5.82493190e-03 -8.26183736e-01 -4.61299300e-01
6.53470635e-01 -8.57994258e-02 -2.63697088e-01 6.47289157e-01
4.06963140e-01 1.31874502e-01 4.70592111e-01 -8.71303737e-01
-1.09442079e+00 5.95918119e-01 -7.81924009e-01 4.67875391e-01
1.59784034e-01 1.88364178e-01 1.34012997e+00 -1.47072327e+00
4.30967063e-01 1.42214906e+00 6.41978621e-01 5.97019374e-01
-1.55418384e+00 -9.89014030e-01 4.95316684e-01 1.02610841e-01
-1.81383657e+00 -6.03468537e-01 3.54786336e-01 -2.43547201e-01
1.45728040e+00 1.87127873e-01 1.98539287e-01 1.33840954e+00
6.62254393e-01 8.26033235e-01 1.07276297e+00 -5.49468398e-01
1.65941447e-01 -1.75184637e-01 3.08430910e-01 6.22590423e-01
1.85736883e-02 -1.99755982e-01 -7.88115203e-01 -7.84433726e-03
8.76140773e-01 -2.79915005e-01 1.54135302e-01 6.89358860e-02
-1.07452118e+00 6.97822094e-01 9.80514139e-02 -2.16582753e-02
-1.52191103e-01 6.37889385e-01 2.63734460e-01 2.10306570e-01
4.56961572e-01 3.76150131e-01 -5.49767852e-01 -3.29235673e-01
-8.27260911e-01 -9.89640057e-02 4.00186449e-01 1.04574978e+00
5.05005479e-01 4.36488718e-01 -1.57266259e-01 1.03214681e+00
4.18663472e-01 7.38725722e-01 5.88625193e-01 -1.00721872e+00
5.11210322e-01 1.94556758e-01 -1.57432243e-01 -7.73504078e-01
-3.08305234e-01 -5.82688332e-01 -7.74054348e-01 -6.79537058e-02
9.95038263e-03 1.94249645e-01 -1.32133007e+00 1.69621432e+00
-2.53699154e-01 6.09257460e-01 3.54159474e-02 4.78638858e-01
7.69824386e-01 7.69631624e-01 3.29175532e-01 -1.01730209e-02
1.18482924e+00 -9.86462057e-01 -5.93273282e-01 -4.96732056e-01
7.08352149e-01 -1.00243592e+00 1.36461806e+00 5.86868703e-01
-1.19192171e+00 -5.99299550e-01 -8.05272698e-01 -4.61764514e-01
-2.69895375e-01 2.26679489e-01 4.82070059e-01 3.67783695e-01
-1.42783880e+00 -4.63350536e-03 -1.09508300e+00 -2.09135979e-01
3.26381832e-01 4.77493972e-01 -1.53996781e-01 -7.27987140e-02
-9.09229517e-01 6.78547144e-01 1.04110412e-01 3.62815529e-01
-6.94068491e-01 -7.32573986e-01 -9.07381117e-01 -1.84877943e-02
5.43332696e-01 -6.25562310e-01 1.50583994e+00 -6.82356298e-01
-1.62695789e+00 6.20537937e-01 -6.74655497e-01 -6.22999787e-01
3.99812222e-01 -7.24072814e-01 -5.55593431e-01 -1.07633539e-01
2.17647627e-02 8.96978021e-01 9.77808595e-01 -9.81693447e-01
-5.02101123e-01 -7.23886117e-02 -4.63999689e-01 -5.39002232e-02
2.47679614e-02 3.46613601e-02 -8.00752401e-01 -7.79453933e-01
2.70534784e-01 -9.54969883e-01 -2.31179222e-01 -4.88319010e-01
-2.73491502e-01 -2.65566707e-01 7.49244273e-01 -5.40425181e-01
1.37739384e+00 -2.35309982e+00 -1.45394489e-01 1.35401949e-01
1.08007424e-01 1.93704322e-01 -5.12147307e-01 1.94219425e-01
-9.85455662e-02 5.74783981e-01 2.69098803e-02 -5.68720341e-01
-2.17086986e-01 3.40825051e-01 -8.30564797e-01 1.21077858e-01
5.14236093e-01 1.11544824e+00 -7.58506477e-01 -5.69068015e-01
2.16581106e-01 5.68397045e-01 -6.39064550e-01 5.58107346e-02
-1.90911189e-01 2.52281100e-01 -2.23159641e-01 4.98213977e-01
4.15582508e-01 -5.60116529e-01 1.01674281e-01 3.03744655e-02
-1.03903182e-01 4.84651744e-01 -1.25188613e+00 1.49429154e+00
-7.79735684e-01 8.45725536e-01 -3.80277753e-01 -4.99338061e-01
1.08092403e+00 2.67722197e-02 -5.49654402e-02 -8.61224711e-01
-1.91967025e-01 1.01952933e-01 -3.02023470e-01 -1.04849368e-01
8.63164783e-01 4.13227618e-01 -4.38343994e-02 3.92582238e-01
1.01462133e-01 -1.58511043e-01 9.02252197e-02 4.03618127e-01
1.09192812e+00 2.14832380e-01 1.87708780e-01 -1.81622989e-03
4.06149477e-01 -3.60700071e-01 7.27282345e-01 1.09694731e+00
2.74365246e-01 8.97274554e-01 3.24940830e-01 -6.19329572e-01
-9.95048583e-01 -1.45894527e+00 -1.14323460e-01 1.33631444e+00
-2.31740773e-01 -7.80558169e-01 -1.45304814e-01 -3.97888243e-01
-3.06436747e-01 9.55621898e-01 -4.98942345e-01 5.29647060e-02
-7.81791985e-01 -5.75772166e-01 5.39629102e-01 8.79712403e-01
3.99802625e-01 -1.10067773e+00 -3.44281465e-01 2.76815116e-01
-1.45379633e-01 -1.43061054e+00 -4.36590523e-01 4.22430634e-02
-9.20626163e-01 -7.56091177e-01 -3.11164945e-01 -4.41406727e-01
5.34614503e-01 2.83250540e-01 1.33005440e+00 3.48726399e-02
-1.81616247e-01 4.69810516e-01 -2.03397632e-01 -3.08470666e-01
-2.19264746e-01 1.93466723e-01 4.24149558e-02 -2.12995917e-01
5.14809608e-01 -3.39444220e-01 -2.25062311e-01 2.15088233e-01
-6.43595815e-01 1.95098892e-01 7.34678328e-01 9.99012351e-01
1.11208892e+00 -2.65693069e-01 3.27345669e-01 -9.13212717e-01
5.06249070e-01 -3.12739372e-01 -7.85110056e-01 2.63983905e-01
-6.55111730e-01 1.78923681e-01 3.96262467e-01 -3.89680862e-01
-1.22461474e+00 -8.94943699e-02 -1.35562867e-01 -4.89683419e-01
5.66389225e-02 4.69539702e-01 4.68261302e-01 3.14692795e-01
3.89615119e-01 3.95769954e-01 -2.74770796e-01 -4.16914403e-01
3.87638927e-01 4.75743324e-01 6.16121948e-01 -6.11922979e-01
5.89669585e-01 3.39677930e-01 -1.71910241e-01 -9.84780967e-01
-8.14394236e-01 -3.03778082e-01 -4.89688516e-01 2.20641181e-01
6.59153819e-01 -1.24544930e+00 -4.91646886e-01 3.59746814e-01
-9.77684438e-01 -7.38007188e-01 -1.73209310e-01 4.28973645e-01
-5.53897381e-01 1.83014631e-01 -9.21594262e-01 -8.70952606e-01
-6.19858265e-01 -1.02473533e+00 1.14393628e+00 -1.79606788e-02
-5.36660075e-01 -7.76317000e-01 -4.42144573e-02 1.94179967e-01
3.13196301e-01 -5.70081949e-01 9.37380135e-01 -4.94747609e-01
-8.83267879e-01 -3.07373144e-03 -2.92064875e-01 2.30536029e-01
-1.58209607e-01 4.66631144e-01 -1.02367735e+00 -2.33862191e-01
-5.40224910e-01 -2.75207967e-01 1.04973304e+00 4.99471247e-01
1.47238564e+00 -1.36052847e-01 -1.54213935e-01 5.25486827e-01
1.25194144e+00 -9.79901478e-02 8.67006719e-01 1.36001870e-01
6.37039244e-01 -1.92996731e-03 6.43497646e-01 2.03272685e-01
5.16270876e-01 5.90062618e-01 1.94118887e-01 -4.34457511e-02
-2.34446779e-01 -3.84246320e-01 6.79088831e-01 1.24858749e+00
3.75485383e-02 -3.60884637e-01 -1.26386273e+00 5.53410351e-01
-1.81357801e+00 -8.10637593e-01 -4.74076942e-02 2.27048135e+00
6.79453611e-01 5.06406486e-01 -2.77450442e-01 -1.30151302e-01
4.23245192e-01 1.10969700e-01 -4.12058026e-01 -8.59292567e-01
-3.06307167e-01 4.68015701e-01 5.42333305e-01 7.30451763e-01
-9.57293034e-01 1.37023079e+00 6.76699209e+00 9.00517166e-01
-1.31600058e+00 2.57635921e-01 8.42634618e-01 -3.61041188e-01
-3.75935584e-01 -1.81547821e-01 -1.29523551e+00 3.13589841e-01
1.21021974e+00 1.05361797e-01 2.18299747e-01 6.12537384e-01
2.95859307e-01 -1.34554550e-01 -8.17309916e-01 9.51273322e-01
1.03809595e-01 -1.50261807e+00 3.04446250e-01 -9.29451734e-02
6.23100519e-01 6.24143541e-01 4.11633611e-01 6.00360930e-01
7.36539900e-01 -1.20475793e+00 6.44703865e-01 6.49588943e-01
1.02878237e+00 -7.08061337e-01 8.12508166e-01 5.77999502e-02
-1.20941341e+00 -6.32691160e-02 -2.80780852e-01 -5.51875979e-02
-7.01742023e-02 3.78497750e-01 -8.31381261e-01 2.51268774e-01
9.55768108e-01 7.35942304e-01 -7.85858631e-01 5.44367492e-01
-5.33604264e-01 9.28001285e-01 -4.67825055e-01 2.13854209e-01
2.02110574e-01 1.23230562e-01 3.16801727e-01 1.50168431e+00
2.10795403e-01 8.84649456e-02 1.02593355e-01 7.63706028e-01
-1.01048544e-01 1.97766796e-01 -4.94895399e-01 2.13076305e-02
7.07126141e-01 1.00078321e+00 -6.12360358e-01 -2.31556013e-01
-4.90130186e-01 6.86428547e-01 5.20336032e-01 6.75503910e-01
-7.49235034e-01 6.27662316e-02 5.49184501e-01 1.29586477e-02
4.33196604e-01 -7.12303281e-01 -5.13141930e-01 -1.06710231e+00
-1.18611723e-01 -5.86985826e-01 5.72981477e-01 -1.11105800e+00
-1.03224289e+00 6.00264072e-01 -1.98833674e-01 -9.70657885e-01
-3.20191324e-01 -5.19286931e-01 -4.84200209e-01 8.87130380e-01
-1.70985007e+00 -1.26401067e+00 -9.50631723e-02 3.51296633e-01
1.05866885e+00 -2.81071573e-01 9.61683989e-01 4.44408432e-02
-8.00721824e-01 9.10761118e-01 2.56929491e-02 1.92496538e-01
9.53513920e-01 -1.02998149e+00 8.26420963e-01 9.97942805e-01
3.98412615e-01 9.64059770e-01 4.20548230e-01 -6.30097508e-01
-1.57124197e+00 -1.27026844e+00 1.06145418e+00 -7.56882250e-01
8.00265074e-01 -4.18861896e-01 -1.07989228e+00 1.20544064e+00
-4.20671739e-02 2.52905428e-01 6.52355611e-01 4.46102172e-01
-7.10102081e-01 -7.05392957e-02 -5.71858585e-01 9.73311841e-01
8.99627626e-01 -6.50606751e-01 -5.42144656e-01 -3.55519056e-02
7.63124466e-01 -5.47786593e-01 -7.25178063e-01 2.66820550e-01
6.27574623e-01 -4.63604689e-01 1.02988267e+00 -5.07615268e-01
5.64496160e-01 -2.18417540e-01 -2.94604331e-01 -7.82500982e-01
-5.31674743e-01 -5.82080424e-01 -3.78116161e-01 1.10004985e+00
6.97206199e-01 -8.43771756e-01 6.07597768e-01 4.76298004e-01
4.45107296e-02 -7.80007720e-01 -9.41638887e-01 -7.85473526e-01
1.34073123e-01 -9.53675807e-01 3.75273317e-01 9.00761724e-01
-5.47132850e-01 5.37485778e-01 -2.92377830e-01 3.20148617e-01
3.47643942e-01 -4.99969646e-02 8.36551905e-01 -8.94407630e-01
-1.21293254e-01 -3.78830135e-01 -1.54073834e-01 -1.16711056e+00
3.44440430e-01 -7.34140396e-01 -8.98436531e-02 -1.20159435e+00
-7.33232126e-03 -7.34584808e-01 -6.03164256e-01 8.06340992e-01
-5.35740145e-02 5.05981326e-01 2.60738492e-01 3.47266823e-01
-7.50844777e-01 3.83736253e-01 6.64583325e-01 -1.08332418e-01
-2.83167630e-01 -2.22366527e-01 -4.78752762e-01 7.73837924e-01
1.11165059e+00 -3.82602870e-01 -3.49075466e-01 -8.48909199e-01
6.41849875e-01 -2.73824394e-01 3.37941438e-01 -7.85362661e-01
3.66771519e-01 -2.25646093e-01 4.73909408e-01 -8.55979979e-01
6.45750880e-01 -5.56470096e-01 4.52427939e-02 1.83135226e-01
-5.61471343e-01 5.94264209e-01 4.72118706e-01 5.13189495e-01
-1.43867150e-01 -3.89348045e-02 6.32489204e-01 -1.26528680e-01
-1.03949654e+00 -4.51752022e-02 -4.72428441e-01 1.10006124e-01
4.96394157e-01 -3.13435256e-01 -3.94234866e-01 -4.06601846e-01
-5.86040020e-01 2.99408048e-01 2.70388573e-01 5.81830740e-01
8.75830591e-01 -9.73282516e-01 -7.95083761e-01 4.42451954e-01
5.55424877e-02 1.27583638e-01 2.82167614e-01 8.21687281e-01
-3.50497812e-01 6.64125502e-01 1.60280809e-01 -8.01131964e-01
-1.44467294e+00 2.79309213e-01 4.70889695e-02 -3.96336675e-01
-7.44258463e-01 8.76512527e-01 2.12984577e-01 -2.76600659e-01
7.08941072e-02 -3.02489489e-01 7.80124590e-02 -2.54832000e-01
7.09958553e-01 2.53287762e-01 3.34366299e-02 -3.99506360e-01
-1.34851918e-01 6.92721069e-01 -6.23618424e-01 -2.02557668e-01
1.20472491e+00 -2.60701895e-01 8.31865072e-02 7.06371546e-01
7.91561842e-01 2.43648261e-01 -1.04990923e+00 -4.67569768e-01
-4.59701419e-02 -4.25444067e-01 3.37291926e-01 -7.91161597e-01
-8.78098190e-01 1.02316952e+00 4.29355979e-01 -2.78041601e-01
1.01047623e+00 -1.27748355e-01 6.36984468e-01 6.69466257e-01
2.43067786e-01 -1.16773045e+00 1.01081781e-01 1.01196790e+00
9.05234337e-01 -1.12027144e+00 9.96757820e-02 -3.33546042e-01
-6.25160038e-01 1.02883458e+00 7.82970965e-01 -6.67755008e-02
3.56382102e-01 5.19674718e-01 2.57467210e-01 1.81628421e-01
-1.34551299e+00 1.89422607e-01 -7.05102086e-02 2.70837754e-01
5.10114551e-01 8.06037560e-02 6.10814430e-03 3.26784581e-01
-2.01431274e-01 1.30020067e-01 6.81461990e-01 8.28169227e-01
-1.31071344e-01 -1.07733417e+00 -2.69077271e-01 4.77186888e-01
-6.24487042e-01 -5.61885655e-01 -2.62779985e-02 7.71318316e-01
-5.76759040e-01 7.38851547e-01 3.89195979e-01 -2.55211025e-01
-2.23503504e-02 1.84840262e-01 3.00943017e-01 -7.16214895e-01
-3.85060847e-01 2.47128576e-01 7.27912933e-02 -6.14856303e-01
-5.65377660e-02 -5.60356796e-01 -1.39592934e+00 -3.59995008e-01
-1.59172401e-01 -2.19155759e-01 6.40890241e-01 6.55535758e-01
8.12825859e-01 4.39705729e-01 3.15959573e-01 -4.35665727e-01
-3.03381503e-01 -8.10922503e-01 -9.37446058e-02 4.98894937e-02
2.38322824e-01 -4.19697165e-01 -1.48504332e-01 8.33314061e-02] | [10.880827903747559, 7.213721752166748] |
a95c61ff-dbd3-47a3-85d8-7d8f8fc511a2 | transformer-based-assignment-decision-network | 2208.03571 | null | https://arxiv.org/abs/2208.03571v2 | https://arxiv.org/pdf/2208.03571v2.pdf | Transformer-based assignment decision network for multiple object tracking | Data association is a crucial component for any multiple object tracking (MOT) method that follows the tracking-by-detection paradigm. To generate complete trajectories such methods employ a data association process to establish assignments between detections and existing targets during each timestep. Recent data association approaches try to solve a multi-dimensional linear assignment task or a network flow minimization problem or either tackle it via multiple hypotheses tracking. However, during inference an optimization step that computes optimal assignments is required for every sequence frame adding significant computational complexity in any given solution. To this end, in the context of this work we introduce Transformer-based Assignment Decision Network (TADN) that tackles data association without the need of any explicit optimization during inference. In particular, TADN can directly infer assignment pairs between detections and active targets in a single forward pass of the network. We have integrated TADN in a rather simple MOT framework, we designed a novel training strategy for efficient end-to-end training and demonstrate the high potential of our approach for online visual tracking-by-detection MOT on two popular benchmarks, i.e. MOT17 and UA-DETRAC. Our proposed approach outperforms the state-of-the-art in most evaluation metrics despite its simple nature as a tracker which lacks significant auxiliary components such as occlusion handling or re-identification. The implementation of our method is publicly available at https://github.com/psaltaath/tadn-mot. | ['Konstantinos Karantzalos', 'Vasileios Tsironis', 'Athena Psalta'] | 2022-08-06 | null | null | null | null | ['visual-tracking', 'occlusion-handling'] | ['computer-vision', 'computer-vision'] | [ 9.59538370e-02 -1.85153604e-01 -1.12628974e-01 -1.90022424e-01
-6.89123988e-01 -5.73624730e-01 6.71298921e-01 1.60511211e-01
-7.87432611e-01 7.55010784e-01 -2.99957573e-01 -2.52138525e-01
-1.39675960e-01 -4.69701767e-01 -1.04843295e+00 -6.81008697e-01
-2.07142234e-02 8.80559504e-01 7.74437308e-01 1.79808587e-01
-1.24245487e-01 6.22712970e-01 -1.69769859e+00 -4.07014899e-02
5.29694557e-01 1.03766239e+00 3.11379761e-01 8.32552433e-01
1.17183737e-01 7.42622852e-01 -3.62001806e-01 -3.46556634e-01
4.85418379e-01 -1.85238153e-01 -3.94777775e-01 -7.70555809e-02
9.63204682e-01 -1.46905169e-01 -3.48855764e-01 8.78931880e-01
6.07281864e-01 3.38190705e-01 4.79071707e-01 -1.48172665e+00
1.17531471e-01 3.51314455e-01 -6.21143460e-01 3.60082150e-01
-7.12709352e-02 4.27832335e-01 8.65143776e-01 -9.25094843e-01
6.10814929e-01 1.01540911e+00 1.06605291e+00 5.62526762e-01
-1.47113311e+00 -7.05729663e-01 2.07473323e-01 3.77090812e-01
-1.40783286e+00 -5.48567116e-01 3.28969449e-01 -7.67711639e-01
6.73057199e-01 3.57502878e-01 7.11060703e-01 1.00834084e+00
-4.37565520e-02 7.99923301e-01 7.47788370e-01 -1.42253011e-01
1.01107899e-02 5.45062236e-02 1.82332352e-01 7.72330403e-01
3.86025816e-01 3.55400473e-01 -4.59682107e-01 4.30944841e-03
5.11781454e-01 -6.07988648e-02 5.78158237e-02 -5.92599034e-01
-1.28117728e+00 7.22130895e-01 5.23156464e-01 3.40175144e-02
-3.02153856e-01 5.40446699e-01 4.18224752e-01 6.71637505e-02
2.66925454e-01 6.73832968e-02 -2.61722982e-01 6.56632408e-02
-1.08604622e+00 3.88502985e-01 5.92379868e-01 8.69630158e-01
8.22146237e-01 4.14432809e-02 -5.69489121e-01 1.67116389e-01
4.01750624e-01 5.43548644e-01 4.76649925e-02 -8.38063240e-01
3.05566370e-01 4.10243452e-01 3.75784039e-01 -9.26007450e-01
-5.06888747e-01 -7.02132404e-01 -6.01796865e-01 4.77672756e-01
7.44649112e-01 -3.49405944e-01 -8.27874422e-01 1.85576034e+00
1.02129066e+00 7.40065038e-01 -1.28322840e-01 1.03462029e+00
6.94184721e-01 4.57962096e-01 3.46005410e-02 -1.93314657e-01
1.33615816e+00 -1.13323367e+00 -6.41586781e-01 -2.64736414e-01
5.30802369e-01 -6.64582849e-01 4.49234784e-01 2.45669842e-01
-1.01760411e+00 -7.57354438e-01 -8.32059205e-01 1.39110684e-01
-2.79914975e-01 4.38147426e-01 4.03715134e-01 5.07178128e-01
-8.85052681e-01 5.11473596e-01 -1.14214158e+00 -3.86193246e-01
5.49905300e-01 6.54475689e-01 -2.44163096e-01 2.51589805e-01
-6.72455847e-01 9.16412473e-01 6.07089877e-01 3.49219739e-01
-1.17253518e+00 -8.31210375e-01 -7.04035223e-01 -1.34002134e-01
8.66410494e-01 -9.49741423e-01 1.35239017e+00 -7.23885477e-01
-1.27347910e+00 5.11654377e-01 -3.22392762e-01 -9.25540447e-01
9.66928124e-01 -4.53780979e-01 -7.79931769e-02 -1.43469587e-01
4.14489061e-02 7.58157015e-01 9.28047180e-01 -1.07955801e+00
-9.89377618e-01 -3.57805081e-02 -1.20600043e-02 2.07153652e-02
-1.04358390e-01 9.07418728e-02 -6.40710175e-01 -4.94252592e-01
-2.76663303e-01 -1.15469265e+00 -2.62417555e-01 3.60282928e-01
-3.93531829e-01 -3.08080524e-01 1.06184566e+00 -4.20402706e-01
1.10941529e+00 -1.76143706e+00 4.34035361e-01 2.19270289e-02
4.47960556e-01 6.22015536e-01 2.10252386e-02 1.33074656e-01
2.81854928e-01 -4.56021398e-01 -1.75538197e-01 -1.00665820e+00
3.03006172e-01 2.06385076e-01 -8.07710737e-02 8.56078506e-01
1.75862849e-01 9.00890172e-01 -8.60072017e-01 -6.51957214e-01
6.17176175e-01 5.89862466e-01 -4.67507005e-01 2.24085882e-01
-5.61535418e-01 6.75682068e-01 -1.77566424e-01 4.25721198e-01
5.21430790e-01 -3.20269138e-01 1.87801253e-02 -3.20572376e-01
-6.02857232e-01 5.27355224e-02 -1.41936660e+00 1.63755631e+00
-2.57955879e-01 9.25513685e-01 1.08894393e-01 -1.03071618e+00
6.53730214e-01 2.06898406e-01 7.44987845e-01 -4.64535564e-01
2.37034619e-01 9.68209133e-02 6.64077327e-02 -2.48104841e-01
5.35110414e-01 1.15583986e-01 1.30988345e-01 3.07279136e-02
-1.45075610e-02 5.94946742e-01 5.59817195e-01 9.06108394e-02
1.16766739e+00 4.29183036e-01 1.59322873e-01 2.87025012e-02
7.15679884e-01 3.72796834e-01 8.98625910e-01 9.03763592e-01
-1.23149961e-01 3.31228375e-01 2.05994826e-02 -5.22100985e-01
-8.67453277e-01 -9.50631857e-01 -5.02963960e-02 9.88310218e-01
1.80230632e-01 -3.59899998e-01 -3.23758543e-01 -7.00391829e-01
8.56384188e-02 3.66981834e-01 -5.84858894e-01 2.71809846e-01
-8.13166142e-01 -5.93334734e-01 6.40474975e-01 3.29927117e-01
1.23919129e-01 -9.51474845e-01 -1.06258619e+00 4.32860225e-01
-1.29662395e-01 -1.42479277e+00 -4.78832155e-01 1.32378355e-01
-5.72321713e-01 -1.19689965e+00 -4.49615479e-01 -4.05041337e-01
5.33622265e-01 9.94169414e-02 9.88894224e-01 8.73371959e-02
-4.58821595e-01 3.08099121e-01 -1.76033244e-01 -4.93977219e-01
-1.97627693e-01 1.18255123e-01 2.12023318e-01 2.28667542e-01
1.55432120e-01 -3.68392527e-01 -5.76681554e-01 5.94039381e-01
-5.35585046e-01 1.47560075e-01 6.27426505e-01 5.07925212e-01
8.31198275e-01 -2.40093917e-01 1.86682463e-01 -4.57424849e-01
-1.09544255e-01 -4.70098197e-01 -1.32892764e+00 1.79324433e-01
-1.11671433e-01 5.37744164e-02 5.25126040e-01 -5.71543813e-01
-7.06224799e-01 7.46759355e-01 -2.67652925e-02 -1.01483262e+00
-8.36857557e-02 2.61790127e-01 1.63197756e-01 -3.37876111e-01
5.28957725e-01 1.38197497e-01 7.49732479e-02 -4.71409827e-01
2.41309181e-01 3.24864388e-02 7.48979270e-01 -4.13078099e-01
1.27769792e+00 7.09725559e-01 3.93084019e-01 -6.22128665e-01
-1.00040984e+00 -8.56240451e-01 -6.48625433e-01 -6.34986460e-01
1.01569891e+00 -8.97494793e-01 -1.25797951e+00 2.53395796e-01
-1.22842336e+00 -4.19487506e-01 -2.88392246e-01 6.10466480e-01
-4.24536169e-01 3.22848648e-01 -3.35257612e-02 -9.58019495e-01
-2.93860227e-01 -1.07487285e+00 1.06342006e+00 2.82991529e-01
4.73432913e-02 -9.75056231e-01 3.35610181e-01 2.91340441e-01
3.46826494e-01 6.09923244e-01 -1.96576864e-01 -4.90108132e-01
-1.05556822e+00 -5.80151081e-02 -2.13327304e-01 -1.77360669e-01
-1.06043711e-01 2.30365284e-02 -8.62699866e-01 -5.92514277e-01
-4.28605646e-01 -1.70005336e-01 9.50798810e-01 5.40792823e-01
9.48040128e-01 -3.49162638e-01 -6.78232431e-01 6.55717790e-01
1.48102248e+00 -1.01391278e-01 1.72118872e-01 3.09546292e-01
8.96326244e-01 3.91641349e-01 9.45809126e-01 4.81841207e-01
4.51347470e-01 1.32005250e+00 7.32835054e-01 -1.32829085e-01
-3.03366870e-01 1.07345946e-01 5.12560546e-01 2.06347540e-01
3.00458889e-03 -2.89374292e-01 -9.09728825e-01 7.18367755e-01
-2.33536363e+00 -1.21588910e+00 -6.12032056e-01 2.34880376e+00
5.54035366e-01 2.15831876e-01 5.52847147e-01 -1.28288165e-01
7.22179830e-01 -7.73449391e-02 -5.12675703e-01 2.74507165e-01
1.33660123e-01 6.76286891e-02 7.13549793e-01 7.31050968e-01
-1.45820296e+00 8.28444719e-01 4.33168030e+00 7.69275725e-01
-9.75362003e-01 2.99863636e-01 -1.30008817e-01 -4.33610350e-01
4.33378786e-01 -4.08082567e-02 -1.49757278e+00 4.82759804e-01
8.38666975e-01 1.25675350e-01 1.44987687e-01 5.53222179e-01
3.67689639e-01 -1.51601508e-01 -1.16188419e+00 8.82806480e-01
-3.13269794e-01 -1.56479120e+00 -3.63841981e-01 -6.63086830e-04
3.16076010e-01 4.16480511e-01 -1.30144864e-01 2.58363754e-01
2.92336792e-01 -7.15204775e-01 1.01026011e+00 5.64248800e-01
4.70813215e-01 -3.42263699e-01 5.63221514e-01 4.28397268e-01
-1.77157104e+00 8.69365130e-03 -2.16427922e-01 4.66954373e-02
4.13407564e-01 4.79351372e-01 -1.04811263e+00 8.19598854e-01
6.61515594e-01 7.43337870e-01 -5.46071947e-01 1.68844473e+00
3.29886973e-02 5.12580752e-01 -8.17599654e-01 1.47250265e-01
3.54392678e-01 -9.02145915e-03 1.03402722e+00 1.33433664e+00
2.83940405e-01 -3.65817189e-01 6.69499397e-01 7.35838830e-01
2.68737912e-01 -3.17605704e-01 -3.64436179e-01 2.42465019e-01
4.12206590e-01 1.51409578e+00 -8.23599041e-01 -3.19004923e-01
-3.68394405e-01 4.93082702e-01 4.18222964e-01 -1.05279371e-01
-1.40195322e+00 9.25433189e-02 7.33094096e-01 1.43393099e-01
8.96972239e-01 -2.58558631e-01 1.08072519e-01 -9.38549697e-01
8.55675936e-02 -5.74213028e-01 5.39931774e-01 -3.17154914e-01
-9.80947137e-01 5.04069448e-01 4.19249348e-02 -1.50956345e+00
-1.98167279e-01 -4.52694774e-01 -4.71029639e-01 5.43512523e-01
-1.66356063e+00 -1.22473037e+00 -5.50395250e-01 6.46717787e-01
5.22683859e-01 1.40867338e-01 2.07211345e-01 8.39571476e-01
-8.50162208e-01 5.45446217e-01 -1.29869774e-01 1.08115390e-01
6.48495436e-01 -1.27298999e+00 2.74079084e-01 1.19132960e+00
2.32320368e-01 2.22599879e-01 1.01823211e+00 -6.95563018e-01
-1.40178204e+00 -1.51050425e+00 7.02636838e-01 -5.64358354e-01
6.81943595e-01 -4.88757670e-01 -7.07504213e-01 8.61567795e-01
9.46916863e-02 4.04048264e-01 3.51825088e-01 -1.28404379e-01
9.47176069e-02 -1.38673335e-01 -7.52961814e-01 3.18777740e-01
1.04284036e+00 9.27394107e-02 -2.87532568e-01 4.54122812e-01
4.65821147e-01 -8.28599870e-01 -7.23986626e-01 3.91688675e-01
5.31934738e-01 -8.80993724e-01 1.12650752e+00 -2.71646470e-01
-2.29245052e-01 -1.05282378e+00 1.47754967e-01 -7.75614262e-01
-1.88126966e-01 -8.28967214e-01 -6.15460217e-01 1.08933401e+00
3.16315293e-01 -4.88482535e-01 9.64036942e-01 1.10119306e-01
-3.42314750e-01 -5.37309527e-01 -1.11038589e+00 -9.59065437e-01
-5.52184641e-01 -4.15098697e-01 2.23047465e-01 6.93354011e-01
-7.53039479e-01 2.57854640e-01 -6.42043710e-01 6.75127983e-01
1.19106293e+00 1.45268619e-01 1.34003925e+00 -1.36663473e+00
-6.03224516e-01 -2.27097288e-01 -3.77321243e-01 -1.03737068e+00
1.29941106e-02 -8.57738078e-01 3.08606207e-01 -1.44758832e+00
-6.40172735e-02 -7.31634021e-01 -6.89133182e-02 5.79080880e-01
-1.63083538e-01 3.48165244e-01 5.34862399e-01 2.04227701e-01
-1.08973324e+00 5.24531841e-01 8.71398509e-01 -1.22268744e-01
-2.28403479e-01 3.14058840e-01 -3.03647369e-02 4.82922018e-01
6.89109147e-01 -1.00353563e+00 -1.79019824e-01 -4.19996768e-01
6.87921271e-02 1.16408654e-01 9.19644594e-01 -1.53027868e+00
7.18682289e-01 -3.04179871e-03 2.51294911e-01 -1.01348269e+00
6.04140997e-01 -1.10397053e+00 6.65578842e-01 7.80811489e-01
5.35159037e-02 4.96981479e-02 5.67309678e-01 7.01763451e-01
6.91610128e-02 -1.76741064e-01 7.63701856e-01 2.52488017e-01
-9.38553929e-01 5.19512296e-01 -2.03675166e-01 -1.14048019e-01
1.40107024e+00 -3.49098653e-01 -1.93193540e-01 1.30451158e-01
-7.96818554e-01 6.77570701e-01 2.55710036e-01 3.41740042e-01
2.99574763e-01 -1.15933132e+00 -7.93607950e-01 -2.90340424e-01
-9.88402143e-02 2.10527048e-01 1.48588941e-01 1.46028185e+00
-2.92286783e-01 3.84389251e-01 -1.47183150e-01 -1.09794116e+00
-1.76662529e+00 4.14766431e-01 6.06575370e-01 -4.88007307e-01
-7.61907816e-01 7.30980694e-01 1.49158433e-01 -3.41183037e-01
5.10975838e-01 -1.99247777e-01 -1.86572358e-01 1.81795165e-01
4.32311863e-01 4.72933084e-01 -9.26758423e-02 -7.44882584e-01
-5.52502513e-01 5.74292600e-01 -1.65733043e-02 3.16923410e-02
1.22224164e+00 3.90075780e-02 2.77440518e-01 2.44047612e-01
7.02386200e-01 -2.63117284e-01 -1.67807078e+00 -2.82386571e-01
6.91833347e-02 -4.59602863e-01 6.39796257e-02 -5.60100555e-01
-1.08896565e+00 5.67590535e-01 8.14099908e-01 1.58222243e-02
8.33902538e-01 -2.52020489e-02 6.15279913e-01 5.45043945e-01
1.37789622e-01 -7.17047036e-01 -4.22487520e-02 4.23699170e-01
5.77650368e-01 -1.31477940e+00 1.50068238e-01 -2.88013101e-01
-2.23742902e-01 8.75333071e-01 8.72087061e-01 -1.74923074e-02
2.63470083e-01 3.38628560e-01 -1.12653814e-01 -2.63828158e-01
-7.72121787e-01 -6.98830128e-01 4.59299147e-01 3.63939255e-01
-1.75476130e-02 -2.45064482e-01 -8.97068605e-02 -1.81276187e-01
1.35589942e-01 7.89252967e-02 1.65455669e-01 9.96811986e-01
-3.67045969e-01 -1.01376057e+00 -5.93949258e-01 1.97698250e-01
-3.55296016e-01 2.67963141e-01 -1.93522945e-01 9.82154608e-01
3.02438438e-01 7.54383087e-01 6.67806789e-02 -2.02092350e-01
3.25782448e-01 -2.33845115e-01 5.05345702e-01 -3.51493955e-01
-9.22967255e-01 5.35042956e-02 1.68388888e-01 -6.78287327e-01
-8.67415905e-01 -1.04402816e+00 -1.22124398e+00 -2.75411665e-01
-4.91269171e-01 4.30410355e-02 5.26645541e-01 1.00326085e+00
3.73283982e-01 8.02477658e-01 1.21025868e-01 -1.19865167e+00
-2.71635830e-01 -7.90296197e-01 1.99303821e-01 7.53437951e-02
7.45245636e-01 -1.04671204e+00 -3.67893763e-02 6.26988942e-03] | [6.462277412414551, -2.0296738147735596] |
2576020c-a210-4d98-94a9-4595e6cfbfe2 | digest-deeply-supervised-knowledge-transfer | 2211.07993 | null | https://arxiv.org/abs/2211.07993v1 | https://arxiv.org/pdf/2211.07993v1.pdf | DIGEST: Deeply supervIsed knowledGE tranSfer neTwork learning for brain tumor segmentation with incomplete multi-modal MRI scans | Brain tumor segmentation based on multi-modal magnetic resonance imaging (MRI) plays a pivotal role in assisting brain cancer diagnosis, treatment, and postoperative evaluations. Despite the achieved inspiring performance by existing automatic segmentation methods, multi-modal MRI data are still unavailable in real-world clinical applications due to quite a few uncontrollable factors (e.g. different imaging protocols, data corruption, and patient condition limitations), which lead to a large performance drop during practical applications. In this work, we propose a Deeply supervIsed knowledGE tranSfer neTwork (DIGEST), which achieves accurate brain tumor segmentation under different modality-missing scenarios. Specifically, a knowledge transfer learning frame is constructed, enabling a student model to learn modality-shared semantic information from a teacher model pretrained with the complete multi-modal MRI data. To simulate all the possible modality-missing conditions under the given multi-modal data, we generate incomplete multi-modal MRI samples based on Bernoulli sampling. Finally, a deeply supervised knowledge transfer loss is designed to ensure the consistency of the teacher-student structure at different decoding stages, which helps the extraction of inherent and effective modality representations. Experiments on the BraTS 2020 dataset demonstrate that our method achieves promising results for the incomplete multi-modal MR image segmentation task. | ['Shanshan Wang', 'Yan Xi', 'Xiawu Zheng', 'Weijian Huang', 'Cheng Li', 'Haoran Li'] | 2022-11-15 | null | null | null | null | ['tumor-segmentation', 'brain-tumor-segmentation'] | ['computer-vision', 'medical'] | [ 4.13218141e-01 1.30624443e-01 -2.75279284e-01 -5.01491427e-01
-1.20711255e+00 -1.20119810e-01 1.99732661e-01 -2.16123313e-01
-5.38951159e-01 7.97890961e-01 1.13255166e-01 -3.74483019e-01
-2.63176829e-01 -6.17225528e-01 -7.22304821e-01 -1.03849292e+00
2.04826027e-01 7.24957228e-01 1.54270113e-01 6.48080856e-02
-1.27484426e-01 2.31415406e-02 -9.68030453e-01 3.46109778e-01
1.18440473e+00 1.01069856e+00 6.40179336e-01 2.91152120e-01
-1.79598153e-01 7.98271418e-01 -3.69281024e-01 -2.90320739e-02
-9.31129158e-02 -3.09283376e-01 -9.71236706e-01 2.06217423e-01
6.38877749e-02 -4.19882357e-01 -4.96792495e-01 1.30275464e+00
5.29150069e-01 -1.59378797e-01 7.11671948e-01 -1.20981932e+00
-4.71448064e-01 6.59732401e-01 -6.41784668e-01 1.43494174e-01
-1.74278259e-01 2.86515445e-01 4.19279873e-01 -4.27650571e-01
3.00084889e-01 7.39723086e-01 5.16916871e-01 7.37986982e-01
-9.68330860e-01 -7.97593355e-01 -2.74711102e-02 5.23699164e-01
-1.16617417e+00 -1.30630344e-01 8.50560308e-01 -4.21121508e-01
1.69984758e-01 1.41710103e-01 6.01760566e-01 1.19423771e+00
4.62180614e-01 1.03755355e+00 1.20611727e+00 -5.09386212e-02
4.06568572e-02 7.72510245e-02 9.19013098e-02 7.02842832e-01
-2.29376435e-01 -3.04017384e-02 -2.05447450e-01 -1.08948708e-01
8.38741839e-01 1.73817664e-01 -5.79954326e-01 -4.90235746e-01
-1.46428823e+00 6.15413547e-01 7.59167075e-01 4.68081087e-01
-4.90192443e-01 -7.83741474e-04 5.13754964e-01 1.68928087e-01
2.63001621e-01 -2.76617855e-01 -4.38522309e-01 1.90739751e-01
-8.71457219e-01 -2.70049393e-01 3.63892734e-01 8.21865678e-01
4.54184473e-01 -1.01314962e-01 -2.18472883e-01 9.28670585e-01
3.50727737e-01 5.53819776e-01 1.05991220e+00 -4.84056234e-01
4.83710080e-01 3.26966017e-01 -2.66852498e-01 -5.81087649e-01
-6.30123615e-01 -5.03235877e-01 -1.33282256e+00 -3.15697372e-01
4.40451413e-01 -4.97728065e-02 -1.25185084e+00 1.75381827e+00
4.19488668e-01 4.66281891e-01 -1.46392640e-03 1.04220879e+00
8.89804482e-01 2.56611288e-01 3.44279557e-01 -3.97120267e-01
1.50822806e+00 -9.63245809e-01 -6.82217240e-01 -8.58545601e-02
6.08164608e-01 -3.78999710e-01 8.40909600e-01 3.37768316e-01
-7.81101465e-01 -1.70592934e-01 -6.75441146e-01 2.92645693e-01
-1.54096577e-02 2.35164259e-03 7.63037324e-01 5.03587127e-01
-8.05398524e-01 1.01110324e-01 -1.15848970e+00 2.01009378e-01
8.49093795e-01 4.23203647e-01 -4.21587527e-01 -5.06688595e-01
-1.34952736e+00 7.62242317e-01 6.31046653e-01 2.84817159e-01
-1.28509235e+00 -1.10164618e+00 -7.68052101e-01 -2.23632917e-01
5.31629086e-01 -7.34146416e-01 1.20989001e+00 -9.43523407e-01
-1.33185029e+00 7.29716957e-01 9.10855457e-02 -2.48770148e-01
7.94137776e-01 2.76685029e-01 -4.45860982e-01 5.51629305e-01
5.79513386e-02 7.49621332e-01 9.01528478e-01 -1.31300354e+00
-3.70948851e-01 -6.33234859e-01 -3.65507901e-01 3.52765381e-01
-4.28502858e-01 -2.64981329e-01 -4.03354406e-01 -7.29684114e-01
3.75303030e-01 -9.68140423e-01 -3.41912031e-01 -1.99500114e-01
-5.61088860e-01 1.59698725e-02 6.92248762e-01 -9.62287009e-01
7.69906998e-01 -1.95400834e+00 3.02306175e-01 2.95568705e-01
3.33922982e-01 -7.47358650e-02 -1.17934749e-01 -4.47286546e-01
-2.02867210e-01 -1.20731756e-01 -6.31125867e-01 -1.18326664e-01
-3.17876369e-01 3.68766457e-01 1.20107904e-01 4.76817131e-01
-1.72929659e-01 1.05059755e+00 -8.51599872e-01 -7.58729100e-01
2.59240329e-01 5.23147106e-01 -3.37919623e-01 1.77308440e-01
-6.07965924e-02 1.16830635e+00 -7.24886477e-01 8.19286048e-01
5.68743587e-01 -3.99882913e-01 -5.18389419e-02 -4.25517321e-01
4.69567925e-01 -3.81548822e-01 -5.74555576e-01 2.17607474e+00
-5.50106227e-01 1.37906134e-01 1.36552632e-01 -1.44426548e+00
3.82679820e-01 6.28041804e-01 1.12012124e+00 -9.15013433e-01
3.96049261e-01 2.68818736e-01 2.32221603e-01 -7.38572121e-01
-4.64008981e-03 -5.54350495e-01 1.99541971e-02 2.66616821e-01
8.77781957e-02 -2.36233264e-01 -3.54080796e-01 1.28364965e-01
1.05160570e+00 -3.90453815e-01 -2.94488370e-01 -1.68310583e-01
5.36735892e-01 -4.71955165e-03 7.25160003e-01 3.99490446e-01
-7.12355971e-01 5.98804176e-01 1.63248569e-01 -2.02663898e-01
-7.17719913e-01 -1.19032049e+00 -4.58828032e-01 6.10432923e-01
3.00937295e-01 3.79288137e-01 -7.42566824e-01 -8.30870628e-01
-2.52832651e-01 5.46492100e-01 -7.69122243e-01 -5.21034122e-01
-4.71257806e-01 -1.24896395e+00 4.51031387e-01 4.69463289e-01
6.17711186e-01 -1.05468702e+00 -5.00914812e-01 2.31057867e-01
-6.96341157e-01 -1.17815113e+00 -4.34626639e-01 1.13666750e-01
-1.06579709e+00 -1.26554835e+00 -1.14612889e+00 -8.71969938e-01
9.10721779e-01 1.91582635e-01 8.08076262e-01 -8.37679207e-02
-4.64911371e-01 4.63987231e-01 -1.53547719e-01 -7.87009373e-02
-5.34460664e-01 -5.54581992e-02 -1.03583559e-01 4.72748950e-02
1.69387132e-01 -4.54405218e-01 -8.01304817e-01 3.28703523e-01
-1.23946524e+00 3.79509091e-01 8.65642667e-01 1.28310537e+00
8.28129172e-01 3.01724881e-01 6.40918553e-01 -8.22252512e-01
2.88985908e-01 -7.10123658e-01 -1.14582792e-01 5.01652539e-01
-2.99715757e-01 -8.44751373e-02 4.11764711e-01 -6.05948925e-01
-1.11886632e+00 3.58904563e-02 -1.99580610e-01 -7.06118822e-01
-2.92934000e-01 6.93062544e-01 -2.01566443e-01 -2.16672391e-01
2.39935577e-01 6.48893476e-01 3.10028315e-01 -6.58427626e-02
5.81502616e-02 5.71571708e-01 6.26661479e-01 -6.46222055e-01
5.64187109e-01 4.76124465e-01 -2.05149263e-01 -5.34574926e-01
-8.54813755e-01 -2.14052051e-01 -4.71751362e-01 -2.81268209e-01
9.73144114e-01 -9.50514495e-01 -6.02812290e-01 9.15656090e-01
-7.62521386e-01 -4.80097473e-01 2.84196064e-02 6.81297839e-01
-7.24436402e-01 3.29949826e-01 -7.86814570e-01 -2.73672014e-01
-4.31479424e-01 -1.85691941e+00 8.49376798e-01 3.18824351e-01
4.35518324e-01 -1.09261823e+00 -2.40881264e-01 9.01652575e-01
5.47638178e-01 1.57989591e-01 1.24128222e+00 -4.76455927e-01
-4.94893134e-01 3.77794206e-02 -3.30853909e-01 2.89195716e-01
3.01459134e-01 -5.81916034e-01 -6.87555373e-01 -4.37789887e-01
1.39026031e-01 -6.29215837e-01 8.80514503e-01 7.19762206e-01
1.73745370e+00 1.21205011e-02 -4.75385725e-01 6.32636964e-01
1.22033548e+00 1.78297102e-01 4.41086531e-01 1.26487419e-01
1.01687384e+00 4.53103453e-01 3.26203376e-01 2.60823369e-01
7.63313055e-01 3.52992147e-01 7.11722910e-01 -1.89622678e-02
-1.15929887e-01 -3.02059595e-02 -9.68300998e-02 1.03368354e+00
3.59272361e-01 1.20366901e-01 -1.14587343e+00 7.18731403e-01
-1.55260563e+00 -6.86821699e-01 1.24065332e-01 2.06865644e+00
1.15602160e+00 -1.54029261e-02 -1.95690632e-01 -4.67920452e-02
7.64193654e-01 -9.02608559e-02 -1.02869368e+00 4.51002479e-01
9.39664245e-02 -5.75037710e-02 5.19017041e-01 1.42621011e-01
-1.03199077e+00 7.25553036e-01 5.66913605e+00 1.16428435e+00
-1.31542182e+00 4.95062709e-01 8.34218383e-01 7.69012645e-02
-4.51940954e-01 -3.78682852e-01 -3.14278632e-01 7.17568934e-01
6.99939013e-01 6.54125400e-03 3.46358895e-01 4.70488995e-01
1.50088862e-01 -2.45312572e-01 -7.68311381e-01 1.06107998e+00
8.20842758e-02 -1.03780174e+00 3.18514444e-02 -1.16206650e-02
6.64873958e-01 2.42623091e-01 1.98400378e-01 3.68517637e-01
1.83205545e-01 -1.03764057e+00 4.29966420e-01 6.17847264e-01
9.37023163e-01 -7.33711839e-01 5.96979916e-01 6.16273463e-01
-7.43753791e-01 -7.34554455e-02 -7.74811953e-02 6.78316355e-01
2.72876680e-01 8.79179180e-01 -7.24779725e-01 7.64340878e-01
5.86813390e-01 5.98128557e-01 -3.64200175e-01 1.05120409e+00
1.64916292e-02 7.68709898e-01 -2.96088815e-01 4.89856184e-01
1.21723533e-01 -1.13601670e-01 4.11537856e-01 6.32786274e-01
2.19814658e-01 2.14212686e-01 2.93927193e-01 7.00247288e-01
-1.41962886e-01 -9.25199315e-02 -9.23337117e-02 1.74295828e-01
3.19424808e-01 1.28505278e+00 -7.57426143e-01 -2.58234024e-01
-4.13314223e-01 9.13201809e-01 1.91178352e-01 4.09463137e-01
-9.87151146e-01 1.23887584e-01 1.75493807e-01 -2.16678664e-01
-4.99411598e-02 -3.76540162e-02 -4.26446378e-01 -1.41964340e+00
-1.51183516e-01 -8.42696369e-01 5.69274366e-01 -6.13987625e-01
-1.42483926e+00 5.82300425e-01 -6.92373812e-02 -1.15652883e+00
7.73574226e-03 -3.87040555e-01 -6.40019655e-01 6.90626621e-01
-1.81006217e+00 -1.30737805e+00 -4.67180014e-01 9.70859408e-01
3.28209132e-01 -1.83364913e-01 6.77604556e-01 5.48751831e-01
-7.82245338e-01 7.34227002e-01 4.44968231e-02 2.19777897e-01
7.29964018e-01 -1.07133532e+00 -5.39231420e-01 3.21698725e-01
-3.16093177e-01 1.43341720e-01 4.81375217e-01 -5.39509058e-01
-1.39653277e+00 -1.13522375e+00 -4.21545357e-02 -6.44588470e-02
7.94998705e-01 2.25622088e-01 -1.20030630e+00 6.30588591e-01
-5.36941215e-02 3.90650481e-01 8.51159573e-01 -3.60461801e-01
1.50219398e-03 2.93695889e-02 -1.34089112e+00 3.93971860e-01
6.31103873e-01 -4.91280138e-01 -6.11315131e-01 5.77182233e-01
6.73265100e-01 -7.97652185e-01 -1.19377184e+00 8.04390550e-01
2.50031680e-01 -4.56407577e-01 9.09431577e-01 -5.58134437e-01
5.37639618e-01 -9.22810435e-02 1.35751396e-01 -1.63805854e+00
6.44105449e-02 5.01162782e-02 5.87414056e-02 7.97940969e-01
1.96473449e-01 -5.95833480e-01 8.49059880e-01 8.01146090e-01
-3.10115516e-01 -9.33806062e-01 -1.25473475e+00 -4.36917186e-01
5.31716704e-01 -4.02942061e-01 4.14691418e-01 1.16324496e+00
-1.89327642e-01 -9.97292995e-02 -2.36160249e-01 2.39650846e-01
9.20136273e-01 7.60633945e-02 8.87783915e-02 -9.36816692e-01
-2.46322274e-01 -4.85261232e-01 -2.46511728e-01 -8.15896273e-01
3.75697225e-01 -1.07852769e+00 1.39343411e-01 -1.50087476e+00
6.89418375e-01 -8.89562786e-01 -7.68947005e-01 6.46272540e-01
-3.30790997e-01 2.31724769e-01 -2.61683047e-01 3.00704122e-01
-5.82898736e-01 8.96607876e-01 1.89608228e+00 -5.12216687e-01
1.25486791e-01 2.60078162e-02 -5.03390610e-01 5.98932147e-01
6.53716743e-01 -4.58708167e-01 -5.41704118e-01 -6.40591145e-01
-3.58062804e-01 6.42164469e-01 5.61283946e-01 -9.45228755e-01
4.30953085e-01 -2.05543786e-01 4.02283996e-01 -5.03496170e-01
1.35925800e-01 -1.02772880e+00 -2.23592389e-02 6.27014220e-01
-2.67279953e-01 -4.52909052e-01 1.83318228e-01 6.57280028e-01
-4.74982321e-01 -5.52735031e-02 8.42742026e-01 -1.63604140e-01
-5.94331920e-01 8.89998257e-01 -1.26978517e-01 1.80330455e-01
1.13642573e+00 3.03847641e-02 -1.81931913e-01 -2.04227921e-02
-1.01150179e+00 6.55299664e-01 3.12689990e-02 4.54682529e-01
6.82110071e-01 -1.44319868e+00 -7.32873857e-01 1.63996786e-01
8.98064002e-02 4.35417235e-01 1.01085448e+00 1.42875791e+00
-1.62678748e-01 1.29692703e-01 -3.72143209e-01 -9.04766321e-01
-8.68048847e-01 3.92146051e-01 5.53708553e-01 -4.00381923e-01
-6.65808201e-01 6.30114794e-01 3.34774584e-01 -6.15586400e-01
2.10551009e-01 -1.75127134e-01 -1.62603796e-01 -1.55454978e-01
5.08633912e-01 -7.26197958e-02 2.10068509e-01 -6.62106991e-01
-2.95500964e-01 3.09231281e-01 -4.49749053e-01 2.31974553e-02
1.17032170e+00 -1.26043603e-01 1.11875068e-02 2.00831771e-01
1.15826738e+00 -6.24246836e-01 -1.25248313e+00 -5.11367857e-01
-3.51496547e-01 -8.70368630e-02 5.02269983e-01 -1.02083623e+00
-1.64727926e+00 9.79752481e-01 7.47364283e-01 -2.73394883e-01
1.33377671e+00 -2.01271139e-02 1.25061536e+00 1.69243306e-01
5.94505787e-01 -8.56262863e-01 3.46230380e-02 2.90593147e-01
5.89657485e-01 -1.70275998e+00 -4.39953834e-01 -3.05494606e-01
-7.62627661e-01 8.82890403e-01 7.54105568e-01 3.78010750e-01
8.03114891e-01 2.55081415e-01 3.15343589e-01 -1.00192435e-01
-3.27587068e-01 1.60031483e-01 9.94453132e-02 5.70500612e-01
4.17451747e-02 3.67969811e-01 3.58564965e-03 9.19504225e-01
1.08406097e-01 2.07090765e-01 2.52085447e-01 6.90209031e-01
-2.77831137e-01 -8.92072082e-01 -3.62983197e-01 7.20273495e-01
-2.57028937e-01 5.01057953e-02 3.48600358e-01 5.04917085e-01
-6.34302199e-02 5.62400699e-01 -1.71316206e-01 -1.93707198e-01
7.09666014e-02 -3.88723724e-02 6.42205715e-01 -3.20883691e-01
-3.17880273e-01 -2.68132091e-02 -5.80895662e-01 -4.62465286e-01
-3.63027692e-01 -5.07385015e-01 -1.59780312e+00 6.65481910e-02
-2.59797156e-01 1.49778008e-01 6.71602666e-01 1.31804872e+00
8.98116082e-02 1.04254508e+00 6.97566509e-01 -7.00726986e-01
-7.53383040e-01 -8.79088819e-01 -5.48356056e-01 5.47259808e-01
2.16250911e-01 -7.77870178e-01 -1.73412025e-01 3.25795775e-03] | [14.519948959350586, -2.291588306427002] |
47b826ce-1019-4800-9e5a-3dd3e694afde | deep-video-inpainting-detection | 2101.11080 | null | https://arxiv.org/abs/2101.11080v1 | https://arxiv.org/pdf/2101.11080v1.pdf | Deep Video Inpainting Detection | This paper studies video inpainting detection, which localizes an inpainted region in a video both spatially and temporally. In particular, we introduce VIDNet, Video Inpainting Detection Network, which contains a two-stream encoder-decoder architecture with attention module. To reveal artifacts encoded in compression, VIDNet additionally takes in Error Level Analysis frames to augment RGB frames, producing multimodal features at different levels with an encoder. Exploring spatial and temporal relationships, these features are further decoded by a Convolutional LSTM to predict masks of inpainted regions. In addition, when detecting whether a pixel is inpainted or not, we present a quad-directional local attention module that borrows information from its surrounding pixels from four directions. Extensive experiments are conducted to validate our approach. We demonstrate, among other things, that VIDNet not only outperforms by clear margins alternative inpainting detection methods but also generalizes well on novel videos that are unseen during training. | ['Ser-Nam Lim', 'Abhinav Shrivastava', 'Larry S. Davis', 'Zuxuan Wu', 'Ning Yu', 'Peng Zhou'] | 2021-01-26 | null | null | null | null | ['video-inpainting'] | ['computer-vision'] | [ 5.84189892e-01 2.20489502e-03 -1.47369608e-01 -1.07594028e-01
-8.39707196e-01 -3.16030502e-01 1.94877893e-01 -3.03197891e-01
-1.25065014e-01 8.76337588e-01 3.50879639e-01 1.64023697e-01
6.09235883e-01 -5.83107948e-01 -1.40119290e+00 -5.10740936e-01
-1.00959875e-01 -3.76166910e-01 2.33896866e-01 1.29175425e-01
2.90708482e-01 3.99014533e-01 -1.29519391e+00 9.58914042e-01
4.41562533e-01 1.22251308e+00 4.08686072e-01 7.78119266e-01
3.67966145e-01 1.34226537e+00 -6.95349395e-01 -2.31806457e-01
4.52163845e-01 -6.61989331e-01 -6.27068877e-01 3.89968634e-01
8.15871119e-01 -9.94545996e-01 -8.84524703e-01 1.18903148e+00
6.70046583e-02 -1.83147416e-01 1.55875951e-01 -1.02249801e+00
-9.71654654e-01 5.14227152e-01 -8.79863441e-01 4.93986219e-01
7.33965635e-01 4.46049005e-01 6.19134545e-01 -1.17095244e+00
6.75444603e-01 1.29867780e+00 6.10943615e-01 4.82901067e-01
-1.00530183e+00 -5.88061094e-01 1.14230484e-01 3.81310940e-01
-1.19136429e+00 -3.84886593e-01 1.06519425e+00 -2.61954576e-01
9.00823236e-01 2.28980631e-01 6.44174635e-01 1.28016591e+00
6.58758521e-01 1.09639359e+00 8.16373050e-01 -2.34326199e-01
1.36683181e-01 -2.76405185e-01 -6.84144497e-01 8.72211516e-01
-1.90832257e-01 3.06165606e-01 -9.13052499e-01 2.50226349e-01
1.08677614e+00 2.99233198e-01 -6.47869885e-01 4.47397344e-02
-1.28245366e+00 4.52510834e-01 6.41701579e-01 3.38879302e-02
-6.42282844e-01 5.75398445e-01 2.83369601e-01 5.56909502e-01
5.60829937e-01 2.78702468e-01 -3.00535232e-01 -6.08245749e-03
-1.55859029e+00 -1.31208837e-01 3.65399420e-02 1.06981897e+00
8.97724926e-01 1.61062643e-01 -4.10896838e-01 4.44741428e-01
1.28971234e-01 5.22757769e-02 4.86160249e-01 -1.30832529e+00
7.99540401e-01 4.46435034e-01 1.64657190e-01 -1.04323316e+00
4.04535718e-02 -2.02593476e-01 -7.74967313e-01 5.53868115e-01
-2.72944152e-01 -2.07599595e-01 -8.34801137e-01 1.50489032e+00
-9.77497920e-02 5.43398976e-01 -6.85779750e-02 1.28962886e+00
6.45643711e-01 7.61880755e-01 -1.74332097e-01 -3.21385145e-01
1.19113374e+00 -1.39087605e+00 -8.04320633e-01 -4.81067538e-01
1.90051734e-01 -8.23001266e-01 7.62515903e-01 5.05706608e-01
-1.57613313e+00 -8.98861885e-01 -1.18563950e+00 -4.28149670e-01
-6.54774904e-02 1.60006747e-01 4.32467228e-03 1.00374795e-01
-1.21031451e+00 8.24823856e-01 -7.51640558e-01 -2.04825416e-01
5.42041242e-01 5.30578941e-03 -5.31210661e-01 -1.64035439e-01
-1.21624470e+00 8.83037210e-01 3.18006307e-01 1.56704545e-01
-1.36989975e+00 -4.34994698e-01 -1.13860977e+00 1.57892510e-01
1.18242808e-01 -5.70815742e-01 1.03208101e+00 -1.60517359e+00
-1.32204723e+00 7.33282208e-01 -2.69846261e-01 -6.75945818e-01
7.98628032e-01 -3.05801719e-01 -5.72830617e-01 6.79807842e-01
3.10641974e-01 1.06373072e+00 1.52545404e+00 -1.14191592e+00
-6.92169964e-01 -5.96020445e-02 2.49798968e-02 2.91510373e-01
-5.20287156e-02 2.21475378e-01 -6.89635277e-01 -1.17847061e+00
-1.57001219e-03 -1.82029739e-01 6.26488179e-02 4.96769726e-01
-6.08982205e-01 3.13380003e-01 1.42429471e+00 -1.19876432e+00
1.15963340e+00 -2.39102459e+00 1.84863061e-01 -2.10561454e-01
3.19304049e-01 3.36340964e-02 -3.03767681e-01 3.29984814e-01
-1.62382588e-01 2.24673580e-02 -4.08971369e-01 -5.55861712e-01
-4.44471300e-01 2.00683266e-01 -4.40561205e-01 5.77245593e-01
8.93556118e-01 1.02047193e+00 -8.34488690e-01 -5.20491898e-01
3.93547386e-01 5.17546535e-01 -4.23031867e-01 3.22496891e-01
-2.29184166e-01 3.18955421e-01 -2.26813797e-02 1.10977912e+00
9.26247358e-01 -9.32220891e-02 -1.82038039e-01 -2.74965435e-01
-1.49663568e-01 9.19658020e-02 -7.75463164e-01 2.07698011e+00
-2.11390078e-01 1.13250875e+00 2.92751849e-01 -8.42792928e-01
6.44620597e-01 2.13349268e-01 2.49965310e-01 -7.76287496e-01
1.70784354e-01 -2.48136427e-02 -4.36152816e-01 -8.81047666e-01
6.92453623e-01 1.54734224e-01 1.67287827e-01 1.20821886e-01
1.74766764e-01 3.87496859e-01 1.05874605e-01 8.40749070e-02
1.31188262e+00 4.20347005e-01 2.16273665e-02 2.38464102e-01
3.30603838e-01 -3.47932160e-01 4.92242873e-01 6.55584455e-01
-4.26858574e-01 1.24512541e+00 4.88483310e-01 -4.80749309e-01
-9.04612303e-01 -1.05523825e+00 1.56660184e-01 8.60648572e-01
4.58448827e-01 -4.20264006e-01 -7.24114478e-01 -6.95076525e-01
-9.06159878e-02 4.60534871e-01 -9.96113598e-01 -2.85641879e-01
-6.55603886e-01 -4.75430042e-02 3.98122698e-01 7.06209302e-01
7.40402877e-01 -1.30403042e+00 -9.78957832e-01 2.31317773e-01
-3.53708357e-01 -1.17665482e+00 -7.17605352e-01 2.79131383e-01
-8.80900562e-01 -1.15884626e+00 -9.11689758e-01 -1.11180747e+00
5.98915577e-01 3.75832558e-01 9.36331689e-01 4.82079312e-02
-3.24636549e-01 2.60761548e-02 -4.63249743e-01 1.55495748e-01
-4.26619411e-01 -4.49167639e-01 -2.68134803e-01 2.16308087e-01
5.33124134e-02 -4.89710540e-01 -8.89142990e-01 1.21148296e-01
-1.06236649e+00 2.19318822e-01 8.43425155e-01 8.76007378e-01
5.52859724e-01 -6.51470525e-03 1.44832626e-01 -3.24537873e-01
3.53129745e-01 -5.78406930e-01 -2.94709623e-01 1.91723287e-01
4.10497971e-02 -1.48395270e-01 7.46884286e-01 -2.03111097e-01
-9.09436345e-01 1.10417947e-01 5.53974360e-02 -1.24271548e+00
-1.88377976e-01 1.68194816e-01 -1.33230045e-01 -1.15246058e-01
3.98204684e-01 5.04183292e-01 -9.38825607e-02 -4.00287300e-01
1.79784909e-01 5.10043144e-01 9.22823966e-01 -6.35762960e-02
6.18428171e-01 5.90731204e-01 -3.43630314e-01 -5.67209125e-01
-3.84209067e-01 6.75149783e-02 -5.40493667e-01 -4.34496194e-01
9.31718647e-01 -1.18358052e+00 -3.51067811e-01 3.80086094e-01
-1.35346591e+00 -2.88775861e-01 -3.27048093e-01 2.56724447e-01
-6.32355690e-01 7.02093542e-01 -1.14138830e+00 -4.67790335e-01
-9.05151963e-02 -1.23410118e+00 1.43490982e+00 1.32526070e-01
-8.64286125e-02 -6.47174656e-01 -3.65910113e-01 1.80906251e-01
2.41194502e-01 4.07524556e-01 4.76074457e-01 1.34928420e-01
-9.23356116e-01 -1.03272824e-02 -4.26817983e-01 5.80986798e-01
1.69403151e-01 5.14978617e-02 -9.49483156e-01 -3.53774697e-01
1.29799426e-01 -3.08818340e-01 1.30806184e+00 3.73225302e-01
1.43087232e+00 -4.63233322e-01 -1.93001479e-01 8.80240381e-01
1.42564726e+00 2.70077825e-01 9.59533513e-01 4.02422875e-01
5.98825753e-01 3.18824857e-01 6.19217873e-01 4.94615287e-01
-1.16789117e-01 4.14096385e-01 9.30083036e-01 -4.27368701e-01
-5.16092122e-01 -4.48875666e-01 8.90135944e-01 2.26889655e-01
6.50861710e-02 -4.29625332e-01 -2.03711316e-01 3.87278914e-01
-1.67757607e+00 -1.22591269e+00 1.63129210e-01 1.84658861e+00
8.50126445e-01 4.39767092e-02 -1.88735917e-01 6.23462275e-02
9.37509418e-01 4.16987777e-01 -6.58357203e-01 -4.19649899e-01
-5.29382050e-01 2.55817503e-01 5.07416546e-01 5.31981409e-01
-1.23855948e+00 9.80284214e-01 6.38020229e+00 6.24401212e-01
-1.20482504e+00 1.37430191e-01 1.03085589e+00 -2.21767470e-01
-5.60662486e-02 -1.27813339e-01 -6.33583125e-03 8.71264577e-01
3.73373747e-01 5.34334302e-01 4.51495469e-01 7.16713428e-01
4.24877375e-01 -2.96640664e-01 -1.19396794e+00 1.14445102e+00
3.97312403e-01 -1.51615226e+00 1.79786995e-01 -2.91053444e-01
8.18029583e-01 -5.07576838e-02 1.38485938e-01 8.94097798e-03
-2.84086555e-01 -9.47894156e-01 1.13170445e+00 5.88483334e-01
1.05144966e+00 -7.55916655e-01 5.32728136e-01 1.05762564e-01
-9.50275540e-01 -2.22857505e-01 -4.25563693e-01 -2.52236724e-01
2.52363235e-01 3.38208348e-01 -5.45532405e-01 3.15241665e-01
9.22671139e-01 1.07153499e+00 -5.79571187e-01 1.10319078e+00
-5.30242622e-01 4.64689970e-01 1.27988234e-02 6.81287944e-01
2.28692785e-01 -7.95287415e-02 4.92321491e-01 1.32492507e+00
6.11180723e-01 -1.49402633e-01 -1.93497613e-01 1.21531343e+00
-3.69732738e-01 -4.48864192e-01 -5.55256784e-01 2.88860023e-01
2.38066465e-01 1.06621540e+00 -4.54112113e-01 -5.25801837e-01
-5.27618170e-01 2.00786686e+00 1.05480343e-01 6.34675384e-01
-1.16511273e+00 -3.80168319e-01 6.85122848e-01 8.12731087e-02
6.25102162e-01 7.59262666e-02 1.91498362e-02 -1.27301919e+00
3.24034631e-01 -5.92849731e-01 1.37296289e-01 -1.44248033e+00
-9.56710398e-01 4.87284631e-01 -4.76915210e-01 -1.58453965e+00
-1.13906898e-01 -4.45734620e-01 -7.79985189e-01 7.00965941e-01
-1.56436062e+00 -9.72899973e-01 -5.35391569e-01 7.33465612e-01
1.10934353e+00 1.04257127e-03 3.31411541e-01 3.58965844e-01
-6.11686945e-01 5.17543733e-01 8.34344104e-02 3.62570733e-01
8.46370757e-01 -9.16450441e-01 4.70044881e-01 1.37342703e+00
-5.50964214e-02 1.84424743e-01 6.67685568e-01 -7.86995351e-01
-1.35162485e+00 -1.56286407e+00 6.15901589e-01 1.20335512e-01
4.11827594e-01 -2.09874853e-01 -8.59855950e-01 9.01855648e-01
7.50616014e-01 2.57976145e-01 -2.60374136e-02 -8.96814585e-01
-4.13369536e-01 -5.37888287e-03 -1.22920346e+00 3.99880826e-01
1.00824380e+00 -6.45051062e-01 -5.52911997e-01 3.26624840e-01
6.83955371e-01 -6.44730628e-01 -5.11239529e-01 2.59977996e-01
3.41137916e-01 -1.44756186e+00 8.30210984e-01 -4.21793193e-01
1.12434554e+00 -5.19470870e-01 -2.34562978e-01 -1.08779478e+00
-2.62418151e-01 -7.93782175e-01 -5.42597175e-01 9.01687145e-01
1.00033335e-01 -1.46800622e-01 8.93712938e-01 5.50105609e-02
-4.91068274e-01 -7.92497754e-01 -9.62234974e-01 -5.63436151e-01
-3.91176313e-01 -4.14948583e-01 1.96232513e-01 8.79057288e-01
-1.19262114e-01 -8.63268673e-02 -8.83707762e-01 3.24578941e-01
3.89898956e-01 1.26711816e-01 2.86569387e-01 -3.40956300e-01
-4.69022691e-01 -3.50684017e-01 -6.21716201e-01 -1.42584193e+00
-4.03294303e-02 -4.32054907e-01 1.32849231e-01 -1.32697415e+00
2.00032383e-01 4.35801715e-01 -3.70868921e-01 5.50883532e-01
-1.10262036e-01 8.07542920e-01 1.79302141e-01 2.04572007e-01
-6.46170139e-01 5.95099390e-01 1.24977791e+00 -4.42954153e-01
9.24903303e-02 -4.09689784e-01 -4.19129789e-01 6.18987083e-01
7.41123140e-01 -3.09634089e-01 -1.03060082e-01 -6.47747576e-01
-5.82716241e-02 4.55150098e-01 7.75168359e-01 -1.33296609e+00
8.68283287e-02 5.98807372e-02 1.20721340e+00 -7.77730823e-01
5.97407758e-01 -7.49343932e-01 -3.84779051e-02 4.68592107e-01
-3.96090806e-01 3.85107458e-01 1.52285382e-01 6.34873867e-01
-5.98015726e-01 -8.00096095e-02 8.66159737e-01 -1.87110603e-01
-9.86111999e-01 2.37239346e-01 -5.70927858e-01 -2.81891078e-01
1.14766872e+00 -6.22069001e-01 -1.97611913e-01 -4.76826698e-01
-7.94118822e-01 4.90232930e-02 6.82752013e-01 5.00377595e-01
1.22042978e+00 -1.36399591e+00 -6.31600499e-01 6.87741280e-01
-7.31874183e-02 -1.43841580e-01 5.22559464e-01 8.90465677e-01
-9.79729414e-01 7.12864473e-02 -5.80533028e-01 -5.59396088e-01
-1.13956368e+00 9.17499959e-01 2.75036603e-01 3.37890983e-01
-6.86448216e-01 1.08987832e+00 3.32951069e-01 6.08235478e-01
3.64332497e-01 -6.54449999e-01 2.75654137e-01 -1.05154879e-01
7.26047814e-01 1.83695570e-01 -2.12602410e-02 -6.19480431e-01
-2.49644965e-01 2.91964471e-01 -7.11322799e-02 3.86312529e-02
9.91678238e-01 -4.05350983e-01 -1.21791996e-01 1.21535860e-01
1.38185775e+00 -7.06212074e-02 -1.98006725e+00 -5.37563190e-02
-3.27837199e-01 -7.71985173e-01 -1.34658769e-01 -7.77165234e-01
-1.48641682e+00 8.85313511e-01 5.87455034e-01 2.42746435e-02
1.64988959e+00 -3.90309989e-02 9.95518506e-01 -2.14133367e-01
-2.64847688e-02 -1.07296658e+00 3.12348008e-01 1.75963566e-01
1.24350667e+00 -1.25194728e+00 -1.30935237e-01 -1.67851433e-01
-4.52840388e-01 1.42013991e+00 7.48521209e-01 -4.40756530e-01
1.45518348e-01 4.03797597e-01 -6.82163462e-02 5.27566448e-02
-7.74886012e-01 1.48795664e-01 -3.86045873e-02 5.40301979e-01
2.77581960e-01 -2.93550909e-01 2.02891320e-01 1.03846401e-01
1.36607647e-01 3.22332196e-02 5.99880099e-01 8.36075008e-01
-4.95850235e-01 -6.82347894e-01 -5.48629820e-01 1.80124089e-01
-3.29997808e-01 -2.76920944e-01 -6.12617671e-01 7.49428868e-01
5.23015738e-01 8.25734615e-01 2.50059485e-01 -6.20395243e-01
-4.75693382e-02 -2.61361688e-01 4.79132980e-01 -2.46071801e-01
-4.82001066e-01 1.80980131e-01 -2.19984233e-01 -1.04558182e+00
-3.40435058e-01 -4.78191733e-01 -1.08657193e+00 7.14173866e-03
8.82889777e-02 -2.87715435e-01 2.75596708e-01 6.80317521e-01
5.58506489e-01 7.20435083e-01 6.26003683e-01 -1.20783758e+00
-1.97509050e-01 -1.00896394e+00 -5.64446032e-01 3.84036899e-01
9.40696776e-01 -4.04761553e-01 -3.07318628e-01 3.49272668e-01] | [10.814719200134277, -1.3460184335708618] |
edf52b34-4b81-4c6d-9792-1e9e28cbd870 | trihorn-net-a-model-for-accurate-depth-based | 2206.07117 | null | https://arxiv.org/abs/2206.07117v2 | https://arxiv.org/pdf/2206.07117v2.pdf | TriHorn-Net: A Model for Accurate Depth-Based 3D Hand Pose Estimation | 3D hand pose estimation methods have made significant progress recently. However, the estimation accuracy is often far from sufficient for specific real-world applications, and thus there is significant room for improvement. This paper proposes TriHorn-Net, a novel model that uses specific innovations to improve hand pose estimation accuracy on depth images. The first innovation is the decomposition of the 3D hand pose estimation into the estimation of 2D joint locations in the depth image space (UV), and the estimation of their corresponding depths aided by two complementary attention maps. This decomposition prevents depth estimation, which is a more difficult task, from interfering with the UV estimations at both the prediction and feature levels. The second innovation is PixDropout, which is, to the best of our knowledge, the first appearance-based data augmentation method for hand depth images. Experimental results demonstrate that the proposed model outperforms the state-of-the-art methods on three public benchmark datasets. Our implementation is available at https://github.com/mrezaei92/TriHorn-Net. | ['Vassilis Athitsos', 'Razieh Rastgoo', 'Mohammad Rezaei'] | 2022-06-14 | null | null | null | null | ['3d-hand-pose-estimation', '3d-hand-pose-estimation'] | ['computer-vision', 'graphs'] | [-3.76261733e-02 -8.15861225e-02 -2.37878636e-01 -9.70746856e-03
-6.23658240e-01 -3.10790181e-01 3.77500743e-01 -3.20341736e-01
-3.19032401e-01 6.67714298e-01 1.72348052e-01 -1.29168838e-01
1.91128373e-01 -4.93306279e-01 -4.51334506e-01 -8.34153354e-01
2.37690508e-01 7.26053596e-01 2.80457288e-01 5.49801737e-02
3.43370497e-01 5.85319281e-01 -1.44773698e+00 -9.50800022e-04
7.19280183e-01 1.24309683e+00 5.65660954e-01 5.79826713e-01
-6.52452782e-02 4.81071115e-01 -3.55982542e-01 -3.45315009e-01
4.46406394e-01 -3.36290002e-02 -8.24948847e-01 8.21797550e-02
7.03463972e-01 -1.09974599e+00 -4.28823918e-01 8.40256870e-01
1.07490599e+00 -1.38775423e-01 5.34858167e-01 -1.19818711e+00
-5.11810005e-01 2.14077592e-01 -1.00108230e+00 -1.24249151e-02
3.24825287e-01 1.93842992e-01 8.83043587e-01 -1.03264368e+00
6.82793558e-01 1.30932593e+00 5.41444659e-01 5.90554833e-01
-8.39223802e-01 -5.75603962e-01 4.68133986e-01 3.51662874e-01
-1.57842588e+00 -6.13118336e-02 9.72208679e-01 -5.50007224e-01
1.00570595e+00 7.03301951e-02 7.52507508e-01 1.11313331e+00
-4.40264940e-02 1.29116201e+00 9.45057333e-01 -5.22498369e-01
-7.15768635e-02 -3.37245017e-02 -4.00821380e-02 7.32365906e-01
2.23844163e-02 -5.28098047e-02 -4.74174172e-01 3.23325675e-03
1.36647975e+00 1.61813542e-01 -5.39069355e-01 -6.00642741e-01
-1.07698739e+00 3.82711083e-01 5.92777252e-01 1.97089761e-01
-6.43287778e-01 1.22670360e-01 1.32099509e-01 -2.01670840e-01
6.59509897e-01 -8.23253021e-02 -6.01860881e-01 -2.57814437e-01
-7.02973425e-01 4.63806212e-01 4.67468590e-01 9.36529517e-01
3.69527131e-01 -3.74856502e-01 -2.63435304e-01 7.72363722e-01
4.49509561e-01 3.82648647e-01 1.05121717e-01 -5.79060018e-01
8.82933855e-01 5.38274109e-01 3.03352147e-01 -8.19187880e-01
-4.49665338e-01 -5.08608341e-01 -7.38931894e-01 3.53041470e-01
9.61735070e-01 -1.80590138e-01 -9.79785979e-01 1.62547171e+00
5.32210886e-01 -3.78315002e-02 -6.18095696e-01 1.33046746e+00
6.59101129e-01 4.24015671e-01 -8.36124420e-02 1.19179025e-01
1.24915314e+00 -1.22810650e+00 -7.56384134e-01 -4.28989619e-01
3.04272890e-01 -8.46176565e-01 1.05981898e+00 6.28561378e-01
-1.00322807e+00 -6.36382341e-01 -7.67039120e-01 -3.13373148e-01
-5.81944659e-02 5.13506591e-01 7.82098114e-01 6.48497283e-01
-9.18097734e-01 4.51855838e-01 -8.85057092e-01 -1.95608214e-01
5.38860381e-01 3.44660223e-01 -3.69940877e-01 -1.49394855e-01
-9.32148635e-01 8.94068658e-01 1.00828081e-01 5.34524381e-01
-4.75352108e-01 -5.12500703e-01 -6.08771205e-01 -2.72275478e-01
4.43532228e-01 -6.02236927e-01 1.33600020e+00 -4.54505533e-01
-1.37477934e+00 8.07210267e-01 -4.64329034e-01 1.75651237e-01
9.26432848e-01 -4.90100801e-01 3.23480815e-01 -1.06996641e-01
6.11558855e-02 8.16710591e-01 9.18834805e-01 -1.44817078e+00
-7.19111681e-01 -1.03342223e+00 7.86988661e-02 3.96833569e-01
-3.80235553e-01 -2.33015910e-01 -9.68920112e-01 -7.47466743e-01
4.30009156e-01 -8.98154438e-01 -1.39657304e-01 4.56987143e-01
-5.64764023e-01 -4.89259154e-01 6.26912832e-01 -1.06670463e+00
1.20800328e+00 -1.77183700e+00 5.25267661e-01 2.28940353e-01
3.25684994e-01 1.46756113e-01 -6.99337050e-02 2.66590506e-01
9.45134535e-02 -1.86982125e-01 -5.13662063e-02 -9.23798621e-01
-6.88419044e-02 -7.50912800e-02 -1.29985631e-01 4.52834249e-01
-8.48703235e-02 8.07656467e-01 -8.03690970e-01 -5.63823283e-01
3.85942489e-01 9.19592202e-01 -5.29889166e-01 3.08328807e-01
-2.11018875e-01 7.75850356e-01 -5.29363155e-01 8.20326090e-01
1.01852119e+00 -1.09558173e-01 1.28983706e-01 -3.03029805e-01
-1.24620542e-01 1.80009633e-01 -1.30915165e+00 2.10043955e+00
-3.61321688e-01 4.04477507e-01 -1.50736675e-01 -4.37978745e-01
6.31710052e-01 4.88655061e-01 4.92541850e-01 -4.37720865e-01
3.76035899e-01 2.57588983e-01 -1.00424059e-01 -3.89551520e-01
3.85009766e-01 1.33879051e-01 4.06866580e-01 4.62948620e-01
-1.21018685e-01 -1.63381785e-01 -1.44466534e-01 -1.22827351e-01
5.56198537e-01 4.54894811e-01 1.37516230e-01 2.10722640e-01
3.59208882e-01 -4.10863817e-01 3.90145719e-01 4.65753615e-01
-3.84430677e-01 9.19058561e-01 4.47790146e-01 -4.09871221e-01
-9.60683405e-01 -9.70309377e-01 -1.47058547e-01 7.46230066e-01
1.82278350e-01 -3.95618290e-01 -9.81264055e-01 -9.06725764e-01
2.66663969e-01 1.22308083e-01 -7.73904324e-01 3.82053226e-01
-5.07684469e-01 -3.95124108e-01 1.25142649e-01 1.05397367e+00
5.83682954e-01 -1.11696613e+00 -4.98800993e-01 1.63629755e-01
-4.65986758e-01 -1.02236116e+00 -8.26011419e-01 -7.81641155e-02
-9.65849578e-01 -9.53639567e-01 -1.42589092e+00 -8.48173499e-01
7.60263085e-01 2.80992448e-01 7.12197363e-01 1.28833324e-01
-4.34381336e-01 2.92663783e-01 -4.39721316e-01 -5.65851033e-01
2.84503013e-01 3.91128480e-01 9.97183621e-02 -1.78272381e-01
3.07759523e-01 -6.06843472e-01 -1.06562090e+00 4.31283206e-01
-3.52023840e-01 2.35345513e-02 6.52005792e-01 8.15655947e-01
6.76171124e-01 -3.50821942e-01 3.08172107e-01 -3.84987503e-01
3.87989461e-01 8.01882818e-02 -4.86340523e-01 1.12038806e-01
-3.44617873e-01 -3.27944718e-02 1.53776005e-01 -5.60907245e-01
-1.20869446e+00 4.16785568e-01 -5.56438565e-01 -3.59356433e-01
-1.52988434e-01 1.01690464e-01 -4.33420002e-01 -1.02022775e-01
5.25727093e-01 1.27794251e-01 2.57902592e-02 -8.88135612e-01
1.68680415e-01 9.32982206e-01 3.09651256e-01 -4.58444327e-01
5.09992599e-01 5.51149428e-01 -1.71764404e-01 -7.89293587e-01
-8.30239177e-01 -5.05577862e-01 -9.74959791e-01 -2.69812465e-01
6.83696210e-01 -7.04713821e-01 -7.72730947e-01 1.03573072e+00
-1.61435926e+00 -4.60469037e-01 1.24772489e-02 4.38176185e-01
-6.05584025e-01 6.48712158e-01 -4.83595818e-01 -1.13982117e+00
-5.34280300e-01 -1.22805989e+00 1.28470910e+00 1.56016264e-03
-1.91535383e-01 -6.92335010e-01 -4.69134003e-02 5.66067338e-01
-3.43584106e-03 -4.65023592e-02 5.29448688e-01 2.97041014e-02
-7.48869836e-01 -4.74280506e-01 -3.76852244e-01 1.74617022e-01
2.68437028e-01 -3.52260947e-01 -1.25632989e+00 -3.51965725e-01
-2.31839120e-01 -1.15134150e-01 7.57385492e-01 6.94083035e-01
1.41554284e+00 1.10007368e-01 -3.80337358e-01 4.73310202e-01
1.04543281e+00 1.73294738e-01 8.05261552e-01 3.40691030e-01
9.90586579e-01 5.37847400e-01 7.93544769e-01 7.15064168e-01
5.84599316e-01 1.06554174e+00 7.81702340e-01 -1.31499231e-01
-4.90442276e-01 -2.06420556e-01 -1.78083628e-01 3.89986426e-01
-7.77133226e-01 -2.09411994e-01 -9.60077047e-01 5.36445796e-01
-1.87126172e+00 -6.22687161e-01 3.53784710e-02 2.10492396e+00
7.09716022e-01 -2.28519402e-02 4.02882963e-01 4.45457846e-01
5.16444027e-01 1.89151585e-01 -7.66788185e-01 1.84287712e-01
2.77924389e-01 2.57474780e-01 2.24825501e-01 4.63910341e-01
-1.14978170e+00 9.50690091e-01 4.96126747e+00 7.62446940e-01
-8.13601255e-01 2.45345369e-01 3.33604395e-01 -3.00040487e-02
1.37198672e-01 -3.33164930e-01 -9.84187305e-01 3.32766712e-01
-3.12386334e-01 7.13585436e-01 3.41644228e-01 9.27085996e-01
-1.41684534e-02 -1.58106938e-01 -1.06829917e+00 1.26845753e+00
6.86623827e-02 -8.00849199e-01 -7.79464543e-02 2.14424416e-01
6.05322540e-01 -3.66768867e-01 2.09582388e-01 -1.12964787e-01
-5.12836933e-01 -7.84049332e-01 8.56686652e-01 4.42514837e-01
8.25730145e-01 -7.36130536e-01 9.07700360e-01 4.93058383e-01
-1.37646115e+00 -7.72073567e-02 -1.41848549e-01 -1.32099912e-01
2.36245513e-01 4.43224400e-01 -7.07130849e-01 2.40276739e-01
8.30321550e-01 6.05909228e-01 -1.88058108e-01 1.04693627e+00
-5.69080055e-01 2.47956486e-03 -2.26744652e-01 -2.18484197e-02
2.51464150e-03 1.36942133e-01 5.39202929e-01 9.00127947e-01
2.35049620e-01 3.15025784e-02 -9.06351767e-03 6.89744234e-01
6.05798177e-02 9.44682062e-02 -2.59825021e-01 2.71394134e-01
4.06994998e-01 9.55881357e-01 -5.74920356e-01 -4.91038561e-02
-2.42380142e-01 1.42296541e+00 4.11851108e-01 2.22218305e-01
-6.00718021e-01 -3.77030611e-01 8.71457160e-01 3.55810881e-01
3.11447084e-01 -3.79947811e-01 -5.65602660e-01 -1.11118960e+00
5.61050534e-01 -6.73093081e-01 1.48843095e-01 -6.44340575e-01
-1.22086799e+00 4.71068501e-01 -2.51930326e-01 -1.28394341e+00
-1.77583337e-01 -8.37869883e-01 -2.67003685e-01 1.24842048e+00
-1.57911575e+00 -1.28774571e+00 -7.22666502e-01 4.77609217e-01
6.21091485e-01 8.53836015e-02 8.50422740e-01 3.37926596e-01
-4.76841539e-01 9.57145154e-01 -1.39730394e-01 8.74834061e-02
7.29209781e-01 -1.17581725e+00 6.79136753e-01 4.19582963e-01
-1.34941652e-01 3.75900537e-01 4.22736228e-01 -7.70251513e-01
-1.33059323e+00 -6.33476555e-01 8.00841093e-01 -6.13549829e-01
1.02785110e-01 -3.65264773e-01 -6.41501129e-01 6.05267107e-01
-2.49882445e-01 -8.17198232e-02 2.36143202e-01 1.34580091e-01
-3.62739503e-01 7.17296153e-02 -1.20602441e+00 5.73110163e-01
1.50312483e+00 -4.53428894e-01 -1.82024926e-01 3.55635256e-01
1.95729136e-01 -8.29969168e-01 -7.60745585e-01 3.25478166e-01
1.20958459e+00 -9.38671649e-01 1.31169534e+00 -1.73337162e-01
4.04681087e-01 -1.42759830e-01 -7.40477294e-02 -1.14617896e+00
-2.04705134e-01 -1.63947836e-01 -5.54154098e-01 1.01837814e+00
1.20669745e-01 -5.52470386e-01 1.31273603e+00 5.59907794e-01
2.43298903e-01 -1.08775187e+00 -9.52827871e-01 -7.14064479e-01
-4.11348767e-04 -4.29467887e-01 6.43900156e-01 4.95209157e-01
-9.19141993e-03 2.64362451e-02 -5.40938318e-01 2.68371314e-01
9.32679474e-01 6.72103241e-02 9.57707465e-01 -1.25531292e+00
-2.85820127e-01 -5.50453603e-01 -5.00166774e-01 -1.79761362e+00
-2.59128679e-02 -4.43342060e-01 1.05875865e-01 -1.93987334e+00
2.03792319e-01 -4.61015373e-01 6.55493438e-02 5.00412405e-01
-3.34728390e-01 2.87096053e-01 2.85000622e-01 1.90558538e-01
1.39343873e-01 5.65787613e-01 1.78122759e+00 -6.10896349e-02
-4.13160115e-01 2.67528385e-01 -4.58836287e-01 7.84719110e-01
1.09088743e+00 -2.60382574e-02 -3.27258348e-01 -6.67027950e-01
-1.87650949e-01 4.65971231e-03 4.62217271e-01 -7.68956304e-01
1.04796983e-01 5.99007271e-02 5.59586823e-01 -1.12684453e+00
6.55616224e-01 -9.16297495e-01 -2.42325813e-01 4.84344214e-01
7.54607171e-02 -2.27323323e-01 1.59450889e-01 3.45703185e-01
1.38391808e-01 2.78420839e-02 6.00814044e-01 -3.04476526e-02
-7.64129758e-01 6.83283329e-01 9.93761122e-02 -4.09880251e-01
8.27307105e-01 -5.05111873e-01 -8.80934075e-02 -5.22266090e-01
-6.24607027e-01 1.16392978e-01 1.81715786e-01 4.63365227e-01
8.11217010e-01 -1.28467870e+00 -6.46377444e-01 2.18207076e-01
1.72010425e-03 4.43694055e-01 5.38799405e-01 9.20757711e-01
-3.71717662e-01 4.91251618e-01 -1.99386612e-01 -6.58535302e-01
-1.43196332e+00 3.75957638e-01 3.01045299e-01 -3.16093475e-01
-7.20277786e-01 1.24823308e+00 3.15733731e-01 -4.10522610e-01
8.45378458e-01 -4.11248326e-01 -1.15708083e-01 1.14748545e-01
6.69219673e-01 5.11900961e-01 7.37635344e-02 -4.81382310e-01
-4.71087217e-01 1.10347986e+00 -1.97068885e-01 -5.00920117e-02
1.18334818e+00 -1.51457369e-01 3.67894216e-04 5.73422909e-02
1.05009639e+00 -2.63033479e-01 -1.56940067e+00 -3.70915651e-01
-4.80008185e-01 -9.91801083e-01 2.50535458e-01 -1.02634382e+00
-1.21064138e+00 1.40005040e+00 8.52707028e-01 -3.30137312e-01
1.14821041e+00 3.25537510e-02 1.03945053e+00 1.87428668e-01
6.32484972e-01 -9.48596835e-01 2.15225428e-01 4.84443426e-01
1.39726055e+00 -1.30276287e+00 -6.17506765e-02 -7.53387868e-01
-3.72693956e-01 1.00231278e+00 1.00283384e+00 2.57690191e-01
7.33145237e-01 1.70216441e-01 6.17926791e-02 6.73654750e-02
-1.50691457e-02 -3.44684452e-01 4.93423104e-01 7.66406775e-01
4.21465427e-01 -7.91391823e-03 -1.97103009e-01 6.07651591e-01
7.81811699e-02 9.54613313e-02 -2.02933103e-01 1.02886057e+00
-1.90374196e-01 -1.24484253e+00 -6.31131232e-01 3.25332433e-01
-1.70092583e-01 -4.63093668e-02 -4.34526682e-01 8.45921457e-01
2.25239083e-01 6.31071091e-01 4.34365794e-02 -4.88960832e-01
6.06381893e-01 -7.13605732e-02 1.12048435e+00 -5.11156559e-01
-2.62807012e-01 1.58283323e-01 -2.33244926e-01 -4.46490049e-01
-1.20594263e-01 -6.75653815e-01 -1.00754547e+00 -3.86202097e-01
-4.77973431e-01 -3.85437727e-01 8.73836875e-01 7.29926050e-01
2.07307890e-01 4.11435276e-01 4.31734592e-01 -1.75054955e+00
-5.86160243e-01 -1.15136623e+00 -6.80886984e-01 1.33135011e-02
3.68589967e-01 -1.24960721e+00 -1.51246712e-01 -3.43471885e-01] | [6.61602258682251, -0.7122858166694641] |
70a8d8ea-65c5-49d7-81ab-65b7788519ee | demystifying-fraudulent-transactions-and | 2306.06108 | null | https://arxiv.org/abs/2306.06108v1 | https://arxiv.org/pdf/2306.06108v1.pdf | Demystifying Fraudulent Transactions and Illicit Nodes in the Bitcoin Network for Financial Forensics | Blockchain provides the unique and accountable channel for financial forensics by mining its open and immutable transaction data. A recent surge has been witnessed by training machine learning models with cryptocurrency transaction data for anomaly detection, such as money laundering and other fraudulent activities. This paper presents a holistic applied data science approach to fraud detection in the Bitcoin network with two original contributions. First, we contribute the Elliptic++ dataset, which extends the Elliptic transaction dataset to include over 822k Bitcoin wallet addresses (nodes), each with 56 features, and 1.27M temporal interactions. This enables both the detection of fraudulent transactions and the detection of illicit addresses (actors) in the Bitcoin network by leveraging four types of graph data: (i) the transaction-to-transaction graph, representing the money flow in the Bitcoin network, (ii) the address-to-address interaction graph, capturing the types of transaction flows between Bitcoin addresses, (iii) the address-transaction graph, representing the bi-directional money flow between addresses and transactions (BTC flow from input address to one or more transactions and BTC flow from a transaction to one or more output addresses), and (iv) the user entity graph, capturing clusters of Bitcoin addresses representing unique Bitcoin users. Second, we perform fraud detection tasks on all four graphs by using diverse machine learning algorithms. We show that adding enhanced features from the address-to-address and the address-transaction graphs not only assists in effectively detecting both illicit transactions and illicit addresses, but also assists in gaining in-depth understanding of the root cause of money laundering vulnerabilities in cryptocurrency transactions and the strategies for fraud detection and prevention. Released at github.com/git-disl/EllipticPlusPlus. | ['Ling Liu', 'Youssef Elmougy'] | 2023-05-25 | null | null | null | null | ['anomaly-detection', 'fraud-detection'] | ['methodology', 'miscellaneous'] | [-3.95718962e-01 -2.42094353e-01 -4.10935432e-01 2.97625721e-01
-3.80495489e-01 -1.16774035e+00 6.46395385e-01 2.28606537e-01
2.91226432e-02 5.30248344e-01 1.26733512e-01 -1.03789318e+00
2.55485158e-02 -1.09861338e+00 -6.46270454e-01 -5.13371229e-01
-7.56957531e-01 8.28050494e-01 2.68601775e-01 -2.24838138e-01
5.96821785e-01 3.87604266e-01 -5.55629790e-01 4.90690559e-01
3.99516523e-01 1.01776612e+00 -1.12126613e+00 5.32965541e-01
-9.68576148e-02 1.46699524e+00 -5.47627449e-01 -1.01082110e+00
5.32889783e-01 -5.85377097e-01 -8.59805107e-01 -5.91365337e-01
1.18234254e-01 -5.76079547e-01 -3.64943773e-01 1.14876258e+00
-1.53507426e-01 -9.69156027e-01 3.51357162e-01 -1.66278231e+00
-9.32324156e-02 1.07330287e+00 -8.39687228e-01 6.19737744e-01
1.00999020e-01 5.17334521e-01 1.50970542e+00 -6.77354991e-01
8.54830325e-01 7.33327389e-01 8.96618545e-01 2.93768365e-02
-1.14977884e+00 -1.01213241e+00 -6.92402720e-01 3.95987421e-01
-5.29253066e-01 -8.13653395e-02 8.72345150e-01 -6.13747954e-01
1.05166662e+00 2.62051165e-01 1.06586039e+00 1.02443218e+00
2.93442070e-01 5.57810664e-01 1.00743628e+00 -3.07022840e-01
6.81796297e-02 -3.28725129e-01 6.50690496e-01 4.61669534e-01
1.13156593e+00 2.54407555e-01 -6.83599472e-01 -1.24407995e+00
7.83221304e-01 2.15394706e-01 -3.96310426e-02 -3.80412608e-01
-1.28446746e+00 1.31738687e+00 1.74379274e-01 5.37226081e-01
-5.12056172e-01 5.95799804e-01 1.12970555e+00 9.53227937e-01
1.10153280e-01 3.83838385e-01 -6.35661364e-01 -4.25952256e-01
-8.56515169e-01 4.70559657e-01 1.59883201e+00 5.82364976e-01
7.84851551e-01 -1.30357161e-01 6.35050297e-01 6.19948469e-02
3.99059206e-01 1.58493415e-01 3.72563094e-01 -9.61149156e-01
9.78447080e-01 9.84166086e-01 1.93263032e-03 -1.14642739e+00
-4.31781188e-02 1.32022277e-02 -5.44561684e-01 -8.07999820e-02
1.14777398e+00 -5.90629391e-02 -2.70057470e-01 1.15382874e+00
1.61750197e-01 1.45918848e-02 -3.11326176e-01 7.01713562e-01
5.97416312e-02 1.20524228e-01 -2.77294159e-01 -2.44939744e-01
1.56777000e+00 -5.31508982e-01 -5.04081726e-01 2.50469714e-01
1.16080368e+00 -5.50136328e-01 1.85119554e-01 3.15678716e-01
-9.46505010e-01 6.52799428e-01 -7.21476257e-01 2.18460917e-01
-5.43793559e-01 -1.00185382e+00 8.37307751e-01 7.66108513e-01
-4.69762295e-01 4.09782976e-01 -8.35725307e-01 9.67298895e-02
7.13992596e-01 3.32387805e-01 -3.86570334e-01 1.89744398e-01
-1.34341681e+00 5.61464489e-01 4.18519646e-01 -1.55775338e-01
-8.03616285e-01 -5.45112729e-01 -5.95696509e-01 1.52151674e-01
5.08248568e-01 -3.10482293e-01 9.26061988e-01 -9.71945643e-01
-6.97340667e-01 8.53167713e-01 3.83645266e-01 -8.44624043e-01
8.56090724e-01 3.16346079e-01 -4.82648253e-01 2.74525642e-01
1.49510235e-01 -5.28481245e-01 5.51099598e-01 -9.12161469e-01
-6.22515023e-01 -6.62657201e-01 -3.20876449e-01 -5.93089104e-01
8.52622185e-03 4.01505679e-01 3.80189747e-01 -8.53230119e-01
4.11944449e-01 -7.53160417e-01 3.54975373e-01 -7.22225428e-01
-4.71054822e-01 -3.12955797e-01 1.28610492e+00 -1.39257908e+00
1.52198505e+00 -1.62965751e+00 -3.99238527e-01 9.46054578e-01
6.00129783e-01 2.19132364e-01 4.57154185e-01 1.01153886e+00
-1.46259025e-01 3.78676653e-01 -2.40019113e-01 2.07165971e-01
-4.91564162e-03 1.70308407e-02 -5.17589450e-01 8.35625410e-01
-2.39352599e-01 1.31134975e+00 -1.00572228e+00 -2.49509171e-01
-2.41318911e-01 -2.79961288e-01 -5.54311454e-01 -2.31613949e-01
-1.28857702e-01 5.11346698e-01 -4.19734299e-01 1.42338896e+00
5.87348402e-01 -3.37766916e-01 8.33238006e-01 8.45416710e-02
1.57918528e-01 5.90429187e-01 -8.78558993e-01 7.95581937e-01
1.91731751e-01 5.29708743e-01 3.06534380e-01 -1.02009702e+00
8.41959596e-01 4.76935834e-01 1.06420207e+00 -7.09381342e-01
1.92309022e-01 9.65518177e-01 2.46903658e-01 -3.30390662e-01
2.09325045e-01 1.93766244e-02 -4.24665570e-01 1.46191061e+00
1.19019970e-02 5.62540889e-01 2.32096419e-01 3.22160721e-01
1.68159950e+00 -2.85397917e-01 4.01169509e-01 -1.92541540e-01
2.95506358e-01 2.33565003e-01 8.82951438e-01 6.22345507e-01
-8.43737423e-01 -9.36465338e-02 1.35316253e+00 -1.22918320e+00
-1.34015989e+00 -8.49376082e-01 1.88181028e-01 6.75832152e-01
-1.69028059e-01 -4.09512043e-01 -5.08881986e-01 -1.27354980e+00
8.90718699e-01 1.99118510e-01 -2.63415664e-01 1.23425111e-01
-1.41743839e+00 -7.94325948e-01 1.05923021e+00 3.57692480e-01
6.54850602e-01 -1.22523236e+00 -5.67212284e-01 3.11418653e-01
-6.48306489e-01 -8.35069597e-01 -4.48406160e-01 -1.31480852e-02
-1.10017037e+00 -2.07908940e+00 -1.10342138e-01 -4.68702018e-01
2.95754403e-01 -2.53638983e-01 1.27131796e+00 4.12642986e-01
-2.81779557e-01 -7.08685294e-02 -2.06722915e-02 -4.50302586e-02
-7.15954244e-01 -1.62944749e-01 -7.14524835e-02 2.54284501e-01
9.53586340e-01 -7.45115936e-01 -7.47373760e-01 4.99413550e-01
-7.02061892e-01 -6.09122574e-01 4.66503799e-02 8.84487569e-01
-5.61517067e-02 -4.15057421e-01 7.92196751e-01 -9.75850224e-01
5.72247982e-01 -1.13980305e+00 -7.72239029e-01 1.09151028e-01
-1.02487493e+00 -3.99896443e-01 3.14581305e-01 -8.99158418e-02
-2.90574193e-01 -7.01268137e-01 6.08673871e-01 -5.12114227e-01
3.64567369e-01 5.23756921e-01 1.69638604e-01 -1.24375165e-01
4.84941363e-01 3.05600427e-02 1.50749445e-01 -4.47441161e-01
1.96079001e-01 7.42772818e-01 2.85702467e-01 -5.51587760e-01
6.40241265e-01 6.69736207e-01 -1.57090649e-02 -2.52090067e-01
3.60706538e-01 -5.69674671e-01 -3.72520357e-01 1.56091079e-01
1.85292482e-01 -4.46347535e-01 -1.28631008e+00 1.03725922e+00
-1.16707182e+00 4.11712341e-02 -1.67746916e-01 2.34533384e-01
-5.10859132e-01 7.08812118e-01 -1.46975970e+00 -8.00530434e-01
-5.38547873e-01 -6.79158926e-01 1.84663460e-01 -3.12189311e-01
-5.97360611e-01 -1.14392900e+00 4.92450297e-01 8.28705370e-01
2.71382511e-01 7.85456657e-01 1.04365075e+00 -1.39264119e+00
-9.65261459e-01 -5.78411698e-01 -3.36943001e-01 2.22558767e-01
9.86586362e-02 -2.35127345e-01 -4.24568921e-01 -5.98651230e-01
2.15101317e-01 -1.54779702e-01 6.80472016e-01 -1.03745580e-01
3.75913531e-01 -9.29017842e-01 -2.40100533e-01 5.80349743e-01
1.33693254e+00 3.31032455e-01 5.79304755e-01 8.20450664e-01
7.53328383e-01 5.04286170e-01 -1.45045951e-01 4.39418703e-01
4.55519468e-01 3.32186729e-01 5.99484384e-01 3.81655157e-01
4.54331160e-01 -5.16600907e-01 2.33793393e-01 4.66417611e-01
-3.86409581e-01 3.80758971e-01 -1.27763355e+00 9.08397853e-01
-2.10022736e+00 -1.18445003e+00 -6.49828732e-01 2.03900433e+00
5.65635085e-01 -4.48959507e-02 6.27024949e-01 4.35877919e-01
7.96067834e-01 -7.20018707e-03 -4.69898283e-01 -6.60651624e-01
-1.80284083e-02 2.34035760e-01 9.58586752e-01 2.21852511e-01
-5.97403526e-01 5.81390023e-01 5.39716625e+00 2.57214218e-01
-8.40916395e-01 3.13337207e-01 7.02359438e-01 -6.08258098e-02
-4.24550354e-01 6.92413270e-01 -2.19549477e-01 1.08538091e+00
9.74573970e-01 -3.31691541e-02 8.11855555e-01 4.47858512e-01
7.74830356e-02 -3.90478373e-02 -1.04696333e+00 9.70560491e-01
-4.00349021e-01 -1.69768214e+00 -1.29367605e-01 6.46100640e-01
5.91687500e-01 1.85407817e-01 -5.54532886e-01 -2.66433042e-02
6.35768294e-01 -8.38225782e-01 7.27077723e-01 3.05164397e-01
5.97091019e-01 -8.97774577e-01 9.47481394e-01 8.78909007e-02
-9.08981919e-01 -5.92943370e-01 5.16719222e-01 5.03730588e-02
-3.42698768e-02 5.28657973e-01 -6.89520299e-01 6.91603839e-01
9.45042133e-01 6.64994538e-01 -9.99647528e-02 6.69963777e-01
-9.84358415e-02 1.03515136e+00 -3.63899171e-01 3.45193416e-01
4.24218565e-01 -7.60664821e-01 7.53650129e-01 1.06065989e+00
2.40959749e-01 -3.57069641e-01 -2.89723217e-01 1.26114011e+00
-6.20599866e-01 -3.47433388e-01 -7.43626058e-01 -3.89648885e-01
6.62676692e-01 1.01235592e+00 -6.78573012e-01 -2.63898373e-01
-5.14843881e-01 5.44803917e-01 1.15488686e-01 4.31868792e-01
-3.02480549e-01 -5.98488271e-01 7.57355690e-01 5.55873096e-01
4.02633518e-01 -1.30774409e-01 -7.06570208e-01 -1.43882668e+00
1.86091408e-01 -1.24187744e+00 1.00274503e+00 1.89137667e-01
-1.43745852e+00 -2.94919431e-01 -4.37350571e-01 -8.79889190e-01
-4.50467408e-01 -2.33545631e-01 -1.11596751e+00 1.03069723e+00
-1.31820035e+00 -1.27669060e+00 3.77464205e-01 8.92375231e-01
-4.45467204e-01 -4.43624824e-01 3.82586658e-01 2.71586001e-01
-3.04464847e-01 3.82824689e-01 2.50223756e-01 9.98677969e-01
3.96436334e-01 -9.71896410e-01 8.43794465e-01 7.47168660e-01
5.86185493e-02 8.36917818e-01 2.21379876e-01 -1.14320683e+00
-1.62069213e+00 -5.07968724e-01 1.36838591e+00 -9.70043063e-01
1.29261112e+00 -4.15745944e-01 -9.24522161e-01 1.21963382e+00
-1.47627428e-01 3.23430181e-01 4.34175938e-01 -9.41735581e-02
-7.14564085e-01 4.34280559e-02 -1.53121495e+00 -1.49134416e-02
7.20743299e-01 -8.96632910e-01 -5.35642326e-01 1.28516080e-02
1.36811629e-01 3.06557771e-02 -7.64788806e-01 1.20342754e-01
7.70345747e-01 -1.42279804e+00 6.69125080e-01 -9.09127474e-01
3.78360361e-01 1.51173294e-01 3.83277461e-02 -5.87212920e-01
3.01431231e-02 -1.24593341e+00 -1.20320690e+00 1.10254073e+00
-1.06515653e-01 -1.29728389e+00 1.11247551e+00 2.69173175e-01
5.76654553e-01 -6.53058708e-01 -1.47696459e+00 -6.86191976e-01
3.13614637e-01 -1.73211381e-01 8.72173965e-01 1.73585975e+00
6.45660639e-01 -4.07197863e-01 -2.04497859e-01 -3.35878134e-01
1.04488063e+00 3.99308085e-01 6.89636707e-01 -1.14078879e+00
-2.36549959e-01 -6.33711457e-01 -5.42748570e-01 -3.18527490e-01
-2.86039382e-01 -1.04426503e+00 -1.03757393e+00 -8.62562120e-01
2.27965996e-01 -5.52497327e-01 -4.30605531e-01 4.22718465e-01
4.78466153e-01 1.43519834e-01 2.38902673e-01 1.16851914e+00
-9.29849222e-02 -2.77211726e-01 7.73717046e-01 -1.05836734e-01
-6.03739880e-02 4.74505350e-02 -5.40468097e-01 6.20078146e-01
5.53811312e-01 -6.15124762e-01 4.92793590e-01 2.18241266e-03
5.39099514e-01 4.43022490e-01 8.02109659e-01 -8.26892629e-02
1.57195255e-01 -1.58396170e-01 -4.31051515e-02 -5.77470243e-01
-4.76537734e-01 -6.25387132e-01 2.46962383e-01 1.04049957e+00
4.61732775e-01 4.00729358e-01 -4.57677543e-01 7.12092876e-01
-3.64044994e-01 4.62133735e-02 4.24757212e-01 -6.47972465e-01
-1.45732090e-01 1.99186221e-01 -1.18323922e-01 1.50972426e-01
1.09792578e+00 -2.83511132e-01 -8.21887851e-01 -2.68448383e-01
-5.38447380e-01 2.79653162e-01 6.90815210e-01 2.36410886e-01
2.37294450e-01 -1.27312624e+00 -6.98512912e-01 4.95580733e-01
-1.36467770e-01 -3.57319117e-01 -4.31314826e-01 1.32882667e+00
-1.05686879e+00 5.63726187e-01 -3.27170193e-01 -2.63273269e-01
-1.05586588e+00 1.09227136e-01 5.90177536e-01 -9.00755346e-01
-3.92091662e-01 1.93865791e-01 -5.52422047e-01 -4.07651812e-01
-4.75587510e-02 -2.00650170e-01 3.78660351e-01 3.56249392e-01
2.78130800e-01 1.19720709e+00 5.71881570e-02 -6.92529321e-01
-5.01514435e-01 8.70783031e-02 -1.83196828e-01 7.12757464e-03
1.29973030e+00 1.87042877e-01 -9.31217670e-01 1.10604584e-01
1.11422980e+00 1.81725472e-02 -7.21423209e-01 -3.51342350e-01
9.60117161e-01 -6.68779194e-01 -4.41071361e-01 -7.93398559e-01
-1.19597638e+00 4.98278201e-01 -4.92170919e-03 7.54542232e-01
2.99445808e-01 -8.41946006e-02 1.40553033e+00 1.82975516e-01
9.55491245e-01 -9.93228436e-01 8.84751696e-03 6.54060245e-01
3.75689954e-01 -9.52053130e-01 -3.96495044e-01 1.75089657e-01
-2.65464962e-01 1.24413359e+00 3.75576392e-02 -3.27836663e-01
6.43571019e-01 7.54060149e-02 8.30295384e-02 -6.76862895e-01
-3.68888348e-01 7.02173352e-01 -5.94636619e-01 2.26370528e-01
3.71610448e-02 3.62733513e-01 -5.89225411e-01 7.03784406e-01
-1.11769319e-01 1.89626813e-01 6.03248358e-01 1.18960202e+00
2.50978693e-02 -1.50929713e+00 -6.30768538e-01 7.19816506e-01
-1.01019835e+00 -4.20099199e-02 -8.49993527e-01 8.02248299e-01
9.56134871e-02 7.78342962e-01 1.05101071e-01 -1.09284118e-01
1.14610277e-01 4.74593163e-01 -6.03689142e-02 3.04008590e-05
-1.20630646e+00 -2.80382037e-01 2.98840672e-01 -8.03937078e-01
1.47680357e-01 -8.62077653e-01 -1.10455251e+00 -1.28273916e+00
-5.24759054e-01 3.09398472e-01 5.43445647e-01 5.72650075e-01
3.71391445e-01 -2.79492021e-01 8.11002314e-01 -1.34881839e-01
-7.17712224e-01 -8.58586550e-01 -9.44723845e-01 8.60014975e-01
6.36205137e-01 -1.36491150e-01 -8.81864488e-01 -8.32548365e-02] | [6.652081489562988, 6.447898864746094] |
81d9e1c1-84eb-4aa2-b046-6a24836eff5b | can-humor-prediction-datasets-be-used-for | null | null | https://aclanthology.org/2020.figlang-1.25 | https://aclanthology.org/2020.figlang-1.25.pdf | Can Humor Prediction Datasets be used for Humor Generation? Humorous Headline Generation via Style Transfer | Understanding and identifying humor has been increasingly popular, as seen by the number of datasets created to study humor. However, one area of humor research, humor generation, has remained a difficult task, with machine generated jokes failing to match human-created humor. As many humor prediction datasets claim to aid in generative tasks, we examine whether these claims are true. We focus our experiments on the most popular dataset, included in the 2020 SemEval{'}s Task 7, and teach our model to take normal text and {``}translate{''} it into humorous text. We evaluate our model compared to humorous human generated headlines, finding that our model is preferred equally in A/B testing with the human edited versions, a strong success for humor generation, and is preferred over an intelligent random baseline 72{\%} of the time. We also show that our model is assumed to be human written comparable with that of the human edited headlines and is significantly better than random, indicating that this dataset does indeed provide potential for future humor generation systems. | ['Kevin Seppi', 'Nancy Fulda', 'Orion Weller'] | 2020-07-01 | null | null | null | ws-2020-7 | ['headline-generation'] | ['natural-language-processing'] | [-1.55606613e-01 4.25789773e-01 9.33083221e-02 8.01585838e-02
-3.22936863e-01 -5.22080421e-01 1.04036677e+00 -1.66348904e-01
-1.17446311e-01 1.09626329e+00 1.00447667e+00 -3.73711675e-01
3.15868169e-01 -6.14771366e-01 -3.72954875e-01 -3.22479844e-01
5.68398774e-01 7.12410390e-01 6.40165135e-02 -7.34054208e-01
5.56618989e-01 3.70395519e-02 -1.16820514e+00 4.03338522e-01
7.45683908e-01 3.96460779e-02 -6.82044029e-02 9.25281286e-01
1.77598014e-01 2.19409895e+00 -9.79378343e-01 -1.01535749e+00
-4.72398326e-02 -1.15134156e+00 -1.22887146e+00 -9.75219086e-02
6.46544456e-01 -4.07907844e-01 -7.41601229e-01 7.42574275e-01
7.35901415e-01 1.57741591e-01 8.10735881e-01 -1.16509521e+00
-8.90247643e-01 9.83006120e-01 -2.69069374e-01 1.17490776e-01
5.78051567e-01 5.73916018e-01 1.28815615e+00 -1.03764117e+00
1.00646198e+00 1.26951301e+00 8.38958859e-01 7.73545742e-01
-1.18004274e+00 -2.49319673e-01 -9.33475971e-01 4.22311664e-01
-5.57434201e-01 -5.11958241e-01 9.53780949e-01 -7.40653515e-01
8.38563740e-01 3.61666977e-01 5.09221137e-01 1.48731530e+00
5.84212132e-02 1.13597298e+00 1.22867882e+00 -4.46624696e-01
-5.42553440e-02 3.47753257e-01 2.44604290e-01 5.05572677e-01
1.66030228e-01 -1.51022688e-01 -5.80217242e-01 -2.20753357e-01
3.66460323e-01 -5.26545048e-01 -5.86765289e-01 2.87413329e-01
-1.03466940e+00 1.13498497e+00 4.35493708e-01 3.75383139e-01
-2.78919071e-01 8.10085759e-02 5.39220214e-01 1.49338737e-01
1.60433531e-01 1.02856886e+00 2.90650696e-01 -4.27636951e-01
-1.23530614e+00 9.94192004e-01 1.46389627e+00 9.89027977e-01
2.33723789e-01 3.91494036e-01 -3.69762599e-01 1.04786599e+00
-3.21401507e-01 4.75020796e-01 6.22838736e-01 -1.08129871e+00
2.06050172e-01 4.02137965e-01 2.98013300e-01 -1.38331199e+00
-5.47387540e-01 -4.99604672e-01 -7.47735262e-01 2.23370731e-01
7.08209515e-01 -4.99496423e-02 -3.37374657e-01 1.48837137e+00
-4.73915130e-01 -5.90405047e-01 -3.31376135e-01 9.44658577e-01
9.84609842e-01 7.21971393e-01 -1.63475990e-01 -2.43744507e-01
9.21365798e-01 -1.26385069e+00 -5.98186195e-01 -4.49829042e-01
4.61639225e-01 -1.12959242e+00 1.41911364e+00 4.33246911e-01
-1.46768105e+00 -3.03453863e-01 -1.01717687e+00 -4.96056587e-01
6.01908937e-03 -2.58559734e-01 3.32031846e-01 4.64779556e-01
-8.14233065e-01 7.74157941e-01 -4.28379560e-03 -2.77404189e-01
3.53296816e-01 -4.19522732e-01 -1.49942845e-01 7.63658211e-02
-1.37752795e+00 1.62162721e+00 2.32399330e-01 -5.26314557e-01
-8.33016157e-01 -5.40030241e-01 -5.64722359e-01 -1.41724974e-01
-3.07099134e-01 -8.33575606e-01 1.65171862e+00 -9.10661519e-01
-9.39067960e-01 1.08838630e+00 -8.71742368e-02 -8.47147107e-01
8.16596210e-01 -3.22825521e-01 -1.17583908e-01 -5.47927059e-02
1.79539695e-01 2.57410854e-01 6.92421257e-01 -1.56962037e+00
-2.56207645e-01 8.98288637e-02 -1.35904104e-01 -8.64290297e-02
-3.74068886e-01 1.97519287e-01 3.05100709e-01 -6.20737612e-01
-5.36049128e-01 -7.82355070e-01 1.17672473e-01 -6.25774622e-01
-4.98548478e-01 -2.11119249e-01 5.24626672e-01 -1.05801606e+00
1.84275627e+00 -1.43229556e+00 9.04194638e-02 -1.89441055e-01
7.24575877e-01 1.31117344e-01 2.13879958e-01 8.88153732e-01
7.66532496e-02 1.39796466e-01 -1.46133766e-01 -3.49124044e-01
3.42697471e-01 -1.38367966e-01 -7.43490756e-01 2.48458982e-01
-2.37924650e-01 1.18416393e+00 -1.03455651e+00 -5.31030476e-01
6.59604073e-02 1.69447139e-01 -4.75974798e-01 2.68125445e-01
-2.54660726e-01 -1.87640414e-02 1.60804614e-01 2.75990814e-01
1.39604947e-02 -5.89679003e-01 -4.01958860e-02 2.17224941e-01
7.07085878e-02 7.21688807e-01 -2.11410865e-01 8.37491333e-01
-3.38816196e-01 1.28224695e+00 -4.68101859e-01 1.02828197e-01
1.09065294e+00 2.61650175e-01 4.05347198e-02 -7.10131407e-01
2.19651178e-01 4.57734108e-01 2.57499605e-01 -5.96295536e-01
1.11720765e+00 -7.94269741e-01 -2.24967264e-02 7.90938079e-01
-1.37239411e-01 -5.85084736e-01 3.15710545e-01 6.32945240e-01
1.42459941e+00 -9.51737911e-02 4.51571018e-01 -2.56296575e-01
3.31446975e-01 3.56030136e-01 8.76266137e-02 1.04714954e+00
-2.12182298e-01 9.01787758e-01 6.52653456e-01 -3.55151504e-01
-1.78655457e+00 -7.05679953e-01 1.44733880e-02 9.88051116e-01
-2.03223512e-01 -6.00284278e-01 -9.15899515e-01 -3.72072726e-01
-8.56853873e-02 1.73657787e+00 -5.49810231e-01 -1.41415354e-02
-7.93676138e-01 -7.77108133e-01 8.33594143e-01 3.24900091e-01
2.93420672e-01 -1.55907547e+00 -8.15909088e-01 3.68207961e-01
-4.72786129e-01 -6.77475929e-01 -3.63370895e-01 -3.43472324e-03
-4.12308455e-01 -9.73503351e-01 -6.44094050e-01 -5.48296332e-01
6.85501471e-02 3.70374262e-01 1.87521231e+00 5.48926651e-01
-1.45456251e-02 9.83809307e-02 -5.94868541e-01 -2.93947101e-01
-1.01430786e+00 -1.47905517e-02 -1.93969846e-01 -6.69768631e-01
5.19565523e-01 -7.16799378e-01 -4.97601897e-01 2.80837119e-01
-6.92726552e-01 3.60993564e-01 1.82942793e-01 1.29875946e+00
-5.82645237e-01 -1.88079998e-01 7.82084286e-01 -1.29518914e+00
1.14840388e+00 -7.49357522e-01 3.90092939e-01 -3.37778777e-01
-7.98428118e-01 -3.52170259e-01 1.10469329e+00 -1.53979868e-01
-7.52804935e-01 -4.97149616e-01 5.44155529e-03 -1.29993837e-02
1.31601110e-01 3.44226658e-01 1.90350607e-01 4.60202992e-01
1.59040141e+00 1.12569101e-01 7.34170526e-02 -2.46842459e-01
5.54996431e-01 8.19345355e-01 9.37279522e-01 -6.41076744e-01
9.71922338e-01 1.08400203e-01 -2.45389536e-01 -8.00535679e-01
-9.75355089e-01 -3.65126669e-01 -3.45830709e-01 -3.54414254e-01
5.32466590e-01 -7.02267170e-01 -6.35573089e-01 4.34706956e-01
-1.70871246e+00 -4.31598067e-01 -2.68731296e-01 -1.78866193e-01
-8.50101113e-01 4.14885074e-01 -1.01703250e+00 -9.36037600e-01
-6.08365774e-01 -5.42143404e-01 4.13718611e-01 2.19348595e-01
-1.22166884e+00 -1.13447440e+00 3.14510673e-01 8.28160107e-01
5.13330877e-01 2.51247495e-01 1.29695857e+00 -9.19549227e-01
-1.23322703e-01 -5.18056810e-01 -2.69490212e-01 3.68169308e-01
-2.56416947e-01 -1.43357471e-01 -1.08683014e+00 2.09831242e-02
1.85476661e-01 -8.05661321e-01 8.86497140e-01 -2.16101944e-01
4.53187853e-01 -9.19498742e-01 3.19586396e-01 -1.68359116e-01
9.52354312e-01 -1.30028993e-01 1.11892545e+00 5.02280712e-01
6.36784971e-01 7.27595448e-01 6.69534802e-02 7.92968929e-01
4.68721092e-01 5.08092344e-01 2.88071424e-01 1.12564370e-01
-4.50223148e-01 -9.36527073e-01 4.42621768e-01 9.54506695e-01
-1.58282518e-01 -5.17782986e-01 -1.09234321e+00 7.79019535e-01
-1.93754864e+00 -1.80858171e+00 -9.22190845e-01 1.93391800e+00
1.09964800e+00 3.70029509e-01 6.60936117e-01 2.35472664e-01
4.27620828e-01 4.21086311e-01 -2.18012724e-02 -4.50626493e-01
-5.46432614e-01 2.71151394e-01 2.78271530e-02 9.08505857e-01
-4.50553626e-01 9.34520185e-01 6.67130947e+00 6.36829734e-01
-5.89475870e-01 4.15194258e-02 5.21447062e-01 -2.21073642e-01
-7.34305501e-01 6.12087026e-02 -4.17025566e-01 6.99434519e-01
8.09814990e-01 -6.00528657e-01 6.33835912e-01 1.02259505e+00
5.14030337e-01 5.80315432e-03 -1.21046913e+00 9.45695400e-01
7.70303309e-01 -1.28401327e+00 2.40068510e-01 -2.33603597e-01
9.92190719e-01 -4.54802990e-01 -2.59334385e-01 5.71640968e-01
5.49343288e-01 -1.78835750e+00 9.96181667e-01 6.16845489e-01
1.19429953e-01 -5.16386986e-01 7.76236117e-01 6.99545562e-01
-1.07083566e-01 6.71282336e-02 -4.40081537e-01 -5.07162392e-01
4.87286329e-01 5.78362286e-01 -1.32213616e+00 -2.17576146e-01
1.72015890e-01 5.33015847e-01 -1.15308630e+00 9.89055753e-01
-5.27864695e-01 1.06820250e+00 2.43065819e-01 -5.64154267e-01
4.72925343e-02 2.58572072e-01 8.38829875e-01 1.65224814e+00
-2.52394620e-02 2.04480123e-02 -1.24427043e-01 1.14329624e+00
-3.37743938e-01 1.57923654e-01 -7.52333939e-01 -2.45945424e-01
3.82331818e-01 1.25328207e+00 -6.51042461e-02 -3.09037894e-01
-7.57283717e-02 1.14225709e+00 3.05758417e-01 -5.64034656e-02
-8.79820585e-01 -3.74862671e-01 -1.52801350e-01 6.13078773e-01
-4.27203387e-01 -7.93297365e-02 -9.62284982e-01 -7.70945251e-01
-2.24172443e-01 -1.43790007e+00 1.27281874e-01 -1.37170446e+00
-1.64480019e+00 6.90506995e-01 -4.58885252e-01 -9.09745753e-01
-5.74584305e-01 -5.72253525e-01 -8.87417555e-01 9.42402959e-01
-6.93462908e-01 -1.27363968e+00 -4.91986841e-01 1.57599270e-01
7.30384052e-01 -2.74667948e-01 3.77473086e-01 3.26660648e-02
2.60364637e-02 4.93192017e-01 -3.13581437e-01 1.64224699e-01
8.88190567e-01 -1.57785773e+00 3.47106516e-01 6.08934283e-01
4.13308851e-02 7.47696936e-01 1.60160589e+00 -6.89342320e-01
-8.89909983e-01 -5.74157894e-01 1.46420813e+00 -1.32805645e+00
1.07483077e+00 1.63738608e-01 -1.11248124e+00 6.14528120e-01
7.85738230e-01 -9.27783430e-01 4.26773012e-01 1.12204337e-02
-7.03899562e-01 6.38947725e-01 -8.57196271e-01 7.81066597e-01
9.36957955e-01 -5.42278707e-01 -9.56311584e-01 4.83958989e-01
4.84316528e-01 -2.46284559e-01 -5.26060998e-01 1.01723984e-01
5.80862701e-01 -1.38441503e+00 5.33125579e-01 -1.09487677e+00
1.93412697e+00 -1.45717040e-01 -1.64331287e-01 -1.29999447e+00
-7.68264711e-01 -8.16475868e-01 -2.42817476e-01 1.28752613e+00
4.37759846e-01 6.67903870e-02 7.79950380e-01 6.60828590e-01
-1.58123761e-01 -4.55567360e-01 -3.42421889e-01 -5.83847106e-01
6.18170679e-01 -9.61628333e-02 2.19333544e-02 1.21362996e+00
5.92792213e-01 1.05155408e+00 -1.08823812e+00 -8.52136552e-01
6.09501421e-01 -8.92326087e-02 1.25978994e+00 -9.48358774e-01
-7.14085340e-01 -1.02045584e+00 -2.20696405e-01 -1.08830917e+00
2.35325754e-01 -1.13075733e+00 1.90905258e-01 -1.51615143e+00
9.87894833e-01 2.37393573e-01 4.40380961e-01 3.87506783e-01
-4.01387006e-01 3.81867528e-01 5.39136231e-01 4.27010626e-01
-2.91824400e-01 4.46323276e-01 1.38857019e+00 -9.98482406e-02
-3.36022042e-02 -2.10146099e-01 -8.91245127e-01 7.09531724e-01
5.59871435e-01 -1.14219651e-01 -9.65741426e-02 9.94807556e-02
5.76502502e-01 2.49348089e-01 8.90972495e-01 -8.41996908e-01
1.54825538e-01 -2.79340625e-01 2.93089002e-01 -3.18075359e-01
-1.22001925e-02 -1.92454219e-01 1.78703383e-01 2.59960860e-01
-5.41174829e-01 8.51599723e-02 -1.39886767e-01 3.53323519e-02
7.01914951e-02 -7.46219397e-01 1.14557159e+00 -2.37463742e-01
-2.27471411e-01 -4.62303966e-01 -4.08611149e-01 5.54877818e-01
3.81204426e-01 -1.06753685e-01 -1.13415217e+00 -1.18432236e+00
-2.93910652e-01 -9.07849148e-03 8.78319919e-01 3.94507796e-01
4.10701007e-01 -1.26957154e+00 -1.15326011e+00 -6.83023930e-01
6.80111870e-02 -5.77231526e-01 -1.21687815e-01 7.83779323e-01
-6.87902391e-01 2.56948799e-01 -1.59039676e-01 -2.89933141e-02
-1.05333436e+00 4.66772079e-01 2.82641172e-01 -2.73164392e-01
-6.82205260e-01 4.98092949e-01 -1.11562766e-01 -1.37492985e-01
-1.67363614e-01 9.06136930e-01 -2.65180487e-02 6.26873737e-03
6.75459802e-01 9.69211519e-01 -2.44194537e-01 -7.16298759e-01
2.63073891e-01 -3.68608743e-01 -1.58031553e-01 -3.00652266e-01
1.23564029e+00 1.43195599e-01 -4.62175757e-01 7.55549073e-01
7.83704877e-01 3.83322924e-01 -6.75171316e-01 7.63637200e-02
1.56178549e-01 -5.87563515e-01 -4.70558345e-01 -1.20345557e+00
-2.08064988e-01 7.87720680e-01 -2.93404669e-01 4.58135724e-01
3.65552634e-01 -6.08130805e-02 1.14159000e+00 2.97618300e-01
5.02439499e-01 -1.01433921e+00 6.30929291e-01 9.08768535e-01
1.42446136e+00 -9.99991119e-01 1.48865715e-01 1.44084752e-01
-1.03738296e+00 1.18810380e+00 8.08476508e-01 -2.47841090e-01
-1.91832736e-01 -1.33757502e-01 -8.76099989e-02 -2.26200923e-01
-9.16940093e-01 1.76488340e-01 2.17589632e-01 2.65061885e-01
1.13555074e+00 -1.69340521e-02 -7.52627552e-01 7.52676427e-01
-1.13273275e+00 -6.20674640e-02 1.23382127e+00 2.86317229e-01
-9.98976588e-01 -6.11555517e-01 -3.93582195e-01 4.99307722e-01
-1.67147413e-01 -3.42354417e-01 -1.16456902e+00 7.76462495e-01
-2.48381466e-01 1.02111661e+00 -6.55398488e-01 -6.86884582e-01
1.35989249e-01 4.05560791e-01 4.62625384e-01 -5.54303765e-01
-9.76429522e-01 -4.52487141e-01 6.10878408e-01 6.06377460e-02
2.49175236e-01 -4.58916426e-01 -1.07053316e+00 -1.22363901e+00
-3.21766883e-02 1.10065110e-01 3.99532318e-02 8.02197456e-01
-3.40764999e-01 1.49341688e-01 7.51152933e-01 -7.46187985e-01
-9.43701267e-01 -1.20609558e+00 -4.04692113e-01 1.14927745e+00
5.81412278e-02 -1.20140329e-01 -7.87266076e-01 3.37518424e-01] | [8.903158187866211, 11.050493240356445] |
02fce129-a068-44da-b974-4c3e1aea0d22 | synpaflex-corpus-an-expressive-french | null | null | https://aclanthology.org/L18-1677 | https://aclanthology.org/L18-1677.pdf | SynPaFlex-Corpus: An Expressive French Audiobooks Corpus dedicated to expressive speech synthesis. | null | ["{\\'E}lisabeth Delais-Roussarie", 'Ga{\\"e}lle Vidal', 'Damien Lolive', 'Marie Tahon', 'Aghilas Sini'] | 2018-05-01 | synpaflex-corpus-an-expressive-french-1 | https://aclanthology.org/L18-1677 | https://aclanthology.org/L18-1677.pdf | lrec-2018-5 | ['expressive-speech-synthesis'] | ['speech'] | [-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] | [-7.1972150802612305, 3.8151073455810547] |
d4b086ed-282c-4d7c-aec2-6a87b5312cd5 | big-data-science-in-porous-materials | 2001.06728 | null | https://arxiv.org/abs/2001.06728v3 | https://arxiv.org/pdf/2001.06728v3.pdf | Big-Data Science in Porous Materials: Materials Genomics and Machine Learning | By combining metal nodes with organic linkers we can potentially synthesize millions of possible metal organic frameworks (MOFs). At present, we have libraries of over ten thousand synthesized materials and millions of in-silico predicted materials. The fact that we have so many materials opens many exciting avenues to tailor make a material that is optimal for a given application. However, from an experimental and computational point of view we simply have too many materials to screen using brute-force techniques. In this review, we show that having so many materials allows us to use big-data methods as a powerful technique to study these materials and to discover complex correlations. The first part of the review gives an introduction to the principles of big-data science. We emphasize the importance of data collection, methods to augment small data sets, how to select appropriate training sets. An important part of this review are the different approaches that are used to represent these materials in feature space. The review also includes a general overview of the different ML techniques, but as most applications in porous materials use supervised ML our review is focused on the different approaches for supervised ML. In particular, we review the different method to optimize the ML process and how to quantify the performance of the different methods. In the second part, we review how the different approaches of ML have been applied to porous materials. In particular, we discuss applications in the field of gas storage and separation, the stability of these materials, their electronic properties, and their synthesis. The range of topics illustrates the large variety of topics that can be studied with big-data science. Given the increasing interest of the scientific community in ML, we expect this list to rapidly expand in the coming years. | ['Kevin Maik Jablonka', 'Daniele Ongari', 'Seyed Mohamad Moosavi', 'Berend Smit'] | 2020-01-18 | null | null | null | null | ['small-data'] | ['computer-vision'] | [ 1.92180708e-01 -1.59831136e-01 -3.70557547e-01 -4.36259434e-02
-3.68862361e-01 -4.12760079e-01 3.45609903e-01 3.95894766e-01
-1.51377439e-01 1.19354320e+00 -1.25023574e-01 -7.62641951e-02
-1.62685737e-01 -1.28505361e+00 -9.16847587e-01 -1.03910911e+00
-1.10605091e-01 8.23151946e-01 3.98470640e-01 -3.08644414e-01
2.43532702e-01 7.72821784e-01 -1.83020282e+00 4.61416572e-01
9.55388129e-01 9.65858042e-01 6.53861284e-01 2.95753807e-01
-3.97086591e-01 5.76293349e-01 -3.43705386e-01 -2.44897917e-01
1.80581897e-01 -1.76765218e-01 -7.68701494e-01 -2.19464734e-01
1.88702419e-01 3.96503061e-01 6.09239750e-02 6.84384882e-01
6.04416311e-01 2.00338468e-01 9.55417097e-01 -6.67359352e-01
-2.67204136e-01 7.68030643e-01 -1.09883146e-02 -9.84746441e-02
5.00812054e-01 -4.44610193e-02 1.03054035e+00 -1.29081798e+00
8.01892757e-01 1.07371259e+00 2.66187608e-01 7.53363788e-01
-1.03106701e+00 -4.32150513e-01 -8.00611302e-02 1.86647058e-01
-1.35038376e+00 -5.20854712e-01 4.77517873e-01 -6.06357574e-01
1.39556336e+00 3.66056114e-01 8.97891045e-01 6.71413064e-01
4.10611600e-01 5.21726489e-01 8.99854302e-01 -5.19151449e-01
4.66468394e-01 3.24909091e-01 -1.01808030e-02 5.88247001e-01
6.59109116e-01 1.87799588e-01 -6.70962334e-01 -3.36808980e-01
2.33542576e-01 2.08127841e-01 2.28724673e-01 -5.69430172e-01
-9.41765010e-01 9.20651078e-01 4.07167792e-01 6.36339307e-01
-1.92123011e-01 2.45901616e-03 1.17040962e-01 7.37419650e-02
1.40645772e-01 9.61959064e-01 -6.22345984e-01 2.30782554e-01
-7.94310629e-01 6.04869545e-01 1.04284310e+00 8.55433762e-01
6.16042852e-01 1.00811504e-01 2.52311856e-01 1.00480652e+00
2.87704438e-01 4.46768850e-01 1.93376154e-01 -6.98830962e-01
4.44581717e-01 1.61120519e-01 4.99751419e-02 -4.65003848e-01
-5.59266448e-01 3.40359025e-02 -6.57527924e-01 1.37841031e-02
2.81597376e-01 6.00646101e-02 -7.91899323e-01 1.18136013e+00
3.10990989e-01 -7.15621233e-01 -7.47775882e-02 4.57642764e-01
1.18011105e+00 7.57983327e-01 6.90344125e-02 -5.10773659e-01
1.13251734e+00 -1.15200102e+00 -4.44261730e-01 5.47838397e-03
3.35256606e-01 -6.88350677e-01 8.02370846e-01 4.73382264e-01
-1.34678340e+00 -2.43131235e-01 -1.08046532e+00 1.38841301e-01
-9.92463291e-01 -5.31886220e-01 1.23711693e+00 7.24192381e-01
-4.22093958e-01 1.30673623e+00 -9.23437297e-01 -3.94182205e-01
3.11820298e-01 8.69311869e-01 -2.54121661e-01 -1.96701631e-01
-1.26215684e+00 8.11899960e-01 5.57768643e-01 -1.52903482e-01
-7.96151280e-01 -7.91990399e-01 -5.43000758e-01 -2.07913548e-01
4.18352634e-01 -8.89884412e-01 8.34041059e-01 -1.32225141e-01
-1.41717064e+00 5.51762998e-01 -1.82925150e-01 -7.60245025e-02
3.42258275e-01 2.29118336e-02 -3.29642147e-01 1.07188180e-01
3.36222537e-02 6.73079491e-01 3.76208901e-01 -9.12654519e-01
-4.21617329e-01 -1.30071193e-01 -3.68854225e-01 -6.90132156e-02
-2.75363982e-01 1.13286078e-01 -1.30262405e-01 -5.72181940e-01
1.82919875e-01 -8.83059561e-01 -7.80577362e-01 -3.03262413e-01
-5.16364694e-01 -2.07320273e-01 2.50204295e-01 1.54529691e-01
1.17724466e+00 -1.73392785e+00 3.66038054e-01 6.22848094e-01
3.55625600e-01 -1.43688008e-01 1.26226336e-01 9.56041694e-01
-1.06303893e-01 3.64277035e-01 -3.16895396e-01 2.71267056e-01
-3.37714672e-01 -1.11930259e-01 -1.11238763e-01 3.50909412e-01
7.07135424e-02 7.84364104e-01 -8.90496612e-01 -2.08648369e-01
4.74104106e-01 5.90543091e-01 -5.49493432e-01 -2.71508098e-02
-6.17907643e-01 5.55076480e-01 -5.71703255e-01 1.05866766e+00
5.34730971e-01 -3.99213433e-01 9.12917703e-02 -1.17959656e-01
-4.81777936e-01 4.57540542e-01 -1.12582624e+00 1.30094051e+00
-8.21028650e-02 4.49667163e-02 -1.69440106e-01 -7.60414302e-01
1.07992601e+00 2.44388953e-01 1.08692193e+00 -6.39175892e-01
-1.85176045e-01 9.52750504e-01 1.33187249e-01 -2.25980952e-01
3.09709996e-01 -6.31608188e-01 1.29323035e-01 2.35252693e-01
9.32909101e-02 -5.92583656e-01 8.01564932e-01 2.47828737e-02
9.42639470e-01 -1.63257569e-01 3.11383158e-01 -4.33766752e-01
5.02834737e-01 3.87247838e-02 9.42669064e-02 7.27787673e-01
3.59602898e-01 3.57258916e-01 9.32451412e-02 -8.20416927e-01
-1.30673170e+00 -6.40254974e-01 -7.02992260e-01 1.00200808e+00
-2.85873879e-02 -1.00132132e+00 -2.92589873e-01 1.41679540e-01
1.57698646e-01 1.41908184e-01 -4.61477876e-01 1.40788883e-01
-4.33023751e-01 -1.31400275e+00 -1.68773644e-02 3.51523399e-01
1.49135441e-01 -1.17157102e+00 -1.05718039e-01 5.15137851e-01
3.90344501e-01 -6.11542583e-01 4.34453934e-01 7.96067417e-01
-1.05798495e+00 -9.11027849e-01 -6.28393173e-01 -6.14966273e-01
3.49786520e-01 2.26824075e-01 1.07174039e+00 1.68543711e-01
-4.79950219e-01 -1.11252129e-01 -2.82009482e-01 -4.54976261e-01
-4.31611985e-01 3.31772625e-01 2.49117747e-01 -6.96109593e-01
3.08676571e-01 -7.12414563e-01 -4.41506624e-01 2.43490860e-01
-6.72949970e-01 4.53978404e-02 5.85488439e-01 7.35884309e-01
1.14033854e+00 2.60452002e-01 4.26795453e-01 -1.22597671e+00
3.96386772e-01 -6.25056744e-01 -5.23177087e-01 3.76588374e-01
-8.47906768e-01 4.42269057e-01 6.81735039e-01 -1.66820809e-01
-6.96965277e-01 2.42567852e-01 -4.55774605e-01 1.66040272e-01
1.21649709e-02 4.32652831e-01 -5.70124567e-01 -2.79667854e-01
8.11746776e-01 -3.12350690e-01 -2.95137733e-01 -5.76741040e-01
4.49177593e-01 5.22633553e-01 -1.11080624e-01 -1.10965252e+00
4.56022829e-01 4.23441350e-01 4.71438736e-01 -1.24452174e+00
-7.41610289e-01 -4.20170635e-01 -7.03288496e-01 1.45362904e-02
6.93427622e-01 -5.05004287e-01 -8.19819152e-01 1.75539553e-02
-6.57398999e-01 -3.53397638e-01 -4.65719908e-01 1.85589254e-01
-5.07725775e-01 -7.06421863e-03 -7.24262476e-01 -5.63885987e-01
-5.91996431e-01 -1.63800037e+00 8.68171871e-01 2.47828275e-01
-1.46567196e-01 -1.11096418e+00 1.07979819e-01 3.04995209e-01
5.01699269e-01 3.91118526e-01 1.06052220e+00 -4.35374320e-01
-7.17394173e-01 -1.07221611e-01 2.91663975e-01 -3.26521814e-01
3.58071208e-01 3.25379223e-01 -9.96582806e-01 -1.30828813e-01
-2.09797934e-01 -2.11670816e-01 1.29830146e+00 5.28441429e-01
1.26226258e+00 -1.17122307e-01 -8.60958636e-01 4.91056651e-01
1.36263025e+00 2.92652935e-01 6.81714833e-01 5.45409024e-01
9.57657814e-01 6.69756532e-01 7.73778200e-01 3.53571922e-01
-2.67767251e-01 5.96751332e-01 1.70861512e-01 2.84549836e-02
6.12604469e-02 -1.96643680e-01 2.39682704e-01 8.85695279e-01
-5.79507470e-01 -2.65286207e-01 -7.99685359e-01 -1.62160873e-01
-1.51879525e+00 -1.14784348e+00 -2.79943556e-01 2.34908652e+00
9.12223458e-01 2.10341379e-01 4.17014301e-01 1.08005024e-01
5.23352385e-01 4.08277027e-02 -7.21252263e-01 -3.78193945e-01
-5.00235856e-01 6.55275762e-01 6.59111559e-01 4.94078666e-01
-1.01378822e+00 1.13912547e+00 7.90668964e+00 8.84542942e-01
-1.37842441e+00 -1.89217702e-01 4.33571875e-01 -2.81182826e-01
-5.20186245e-01 1.48359567e-01 -1.11079633e+00 2.84603387e-01
1.23071218e+00 -7.14445561e-02 4.96021241e-01 8.39960277e-01
-1.62024163e-02 -1.40181020e-01 -1.27102184e+00 8.09314609e-01
-4.21378076e-01 -1.84547281e+00 1.36814669e-01 2.73070991e-01
8.29605877e-01 3.12484205e-01 -8.06590095e-02 -1.35104164e-01
1.83673464e-02 -1.27095020e+00 6.33090734e-01 4.72980142e-01
9.17624950e-01 -5.54964066e-01 2.49152854e-01 -1.45418674e-01
-1.43406975e+00 -1.07264012e-01 -8.72482538e-01 -9.58639979e-02
1.45544976e-01 9.59444344e-01 -7.09542990e-01 4.49126124e-01
6.09314203e-01 9.27177608e-01 -4.74531978e-01 1.20743358e+00
-1.17661962e-02 1.25362635e-01 -5.39957047e-01 -4.85945106e-01
-2.13850856e-01 -4.71994102e-01 2.46434957e-01 9.26434398e-01
2.19047457e-01 -3.70041355e-02 1.96555138e-01 8.60807657e-01
-4.53565037e-03 4.45594907e-01 -6.64577365e-01 -4.17891890e-01
6.29705012e-01 1.00825894e+00 -8.01477730e-01 -3.34879398e-01
-6.30380154e-01 2.19221637e-01 1.86331674e-01 -5.67210801e-02
-1.74798921e-01 -3.27347994e-01 3.16264868e-01 4.37306225e-01
1.74866930e-01 -8.99371132e-02 -1.40609086e-01 -9.85289633e-01
-3.01379353e-01 -7.27732420e-01 4.48482543e-01 -4.31786984e-01
-1.31153166e+00 2.86278337e-01 3.91840115e-02 -7.08324850e-01
1.30659625e-01 -1.03930664e+00 -2.73238957e-01 4.33649182e-01
-1.06179726e+00 -6.42953813e-01 2.49648467e-02 1.08188108e-01
4.94587183e-01 -1.75260171e-01 9.37200785e-01 3.97523671e-01
-6.23467982e-01 8.12602341e-02 7.94914961e-01 -7.03180850e-01
8.11106086e-01 -1.08599591e+00 8.94100592e-02 5.91759011e-02
-9.04201940e-02 7.95222580e-01 8.36911023e-01 -1.02058995e+00
-1.66273212e+00 -8.04545701e-01 5.50960779e-01 -4.54659969e-01
6.82920396e-01 -5.24997354e-01 -7.80731142e-01 2.52214491e-01
-3.42956066e-01 -1.85406730e-01 1.25053942e+00 1.52139977e-01
-4.46638055e-02 -1.15059257e-01 -1.18130469e+00 4.36189085e-01
9.38753724e-01 -2.77382731e-01 -1.11575769e-02 9.62214708e-01
7.54007757e-01 -3.77318978e-01 -1.39902520e+00 3.27269256e-01
8.12657833e-01 -7.21737325e-01 1.18660843e+00 -8.01818669e-01
4.57050562e-01 -1.55170247e-01 -3.66525292e-01 -6.90017045e-01
-4.27430451e-01 -5.24218202e-01 -2.39390433e-01 8.08265984e-01
7.99125910e-01 -5.97433090e-01 7.20385313e-01 1.22404170e+00
-2.34276444e-01 -1.00381184e+00 -4.67945993e-01 -7.06435263e-01
4.76965904e-01 -1.43857300e-01 8.39892685e-01 4.96905833e-01
4.20652062e-01 4.56276864e-01 -2.28194296e-01 -4.02897865e-01
5.06785810e-01 3.22640568e-01 3.52755457e-01 -1.54612947e+00
-2.80565798e-01 -2.87966967e-01 -2.21677199e-01 -5.63124835e-01
-1.86827451e-01 -1.09728277e+00 -3.11380506e-01 -1.52680051e+00
4.63534683e-01 -8.71029496e-01 -3.20890874e-01 1.44762889e-01
1.08404130e-01 1.24261165e-02 4.22783056e-03 5.20674527e-01
-6.42549098e-01 3.03829223e-01 1.35880828e+00 -2.91455895e-01
-5.18539488e-01 4.88425270e-02 -8.43332350e-01 2.58810431e-01
8.84940386e-01 -6.40130639e-01 -6.75347671e-02 1.69538274e-01
8.57290089e-01 -2.18666539e-01 -4.26008403e-01 -1.02000880e+00
-6.16304725e-02 -5.40196657e-01 5.36644101e-01 -6.66377544e-01
4.24577802e-01 -7.47318566e-01 5.43645263e-01 5.78439355e-01
-1.71234012e-02 -3.18468779e-01 1.41372774e-02 2.51607001e-01
-3.89759988e-02 -6.63955629e-01 7.79894590e-01 -7.55959392e-01
-4.06152099e-01 6.08312428e-01 -5.34099042e-01 -3.11411709e-01
1.02819216e+00 -1.05342962e-01 -2.52012283e-01 1.98218495e-01
-9.03535545e-01 1.20172329e-01 7.89142728e-01 1.09137949e-02
3.51037592e-01 -1.15323842e+00 -4.07686792e-02 -3.77541110e-02
7.35250264e-02 3.28744352e-01 7.83043653e-02 6.81544721e-01
-6.48335457e-01 8.59813929e-01 -3.11439544e-01 -4.14735705e-01
-1.10479152e+00 9.65091705e-01 3.42896760e-01 -1.08239666e-01
-1.69018984e-01 6.76922500e-01 1.90922432e-02 -2.82023937e-01
-3.77435498e-02 -1.66770950e-01 -2.37076223e-01 -7.21041337e-02
6.26392901e-01 4.88504469e-01 4.94264573e-01 -2.60349363e-01
-4.69676018e-01 7.15093434e-01 -2.22949281e-01 2.46698231e-01
2.03802752e+00 2.60854900e-01 -5.17204463e-01 5.60872734e-01
7.96384156e-01 2.55298436e-01 -7.44693935e-01 1.95687339e-01
-5.34714498e-02 -1.73217833e-01 -3.65620613e-01 -3.62556994e-01
-1.03698373e+00 1.05622184e+00 3.72637331e-01 1.61077142e-01
6.15959644e-01 5.38199961e-01 4.28932935e-01 8.49351823e-01
4.82702672e-01 -1.18132687e+00 1.71222970e-01 3.70194376e-01
8.12184572e-01 -1.09168971e+00 4.90370870e-01 -6.40906692e-01
-1.59315214e-01 1.41140485e+00 5.26723683e-01 2.61346370e-01
7.49313235e-01 3.00055683e-01 -6.95950985e-01 -6.36777222e-01
-6.45099044e-01 6.91043027e-03 -5.71634620e-02 5.85060596e-01
8.75329554e-01 2.66058862e-01 -3.28379631e-01 3.61478567e-01
-6.06542408e-01 -2.68453062e-01 4.65071797e-01 1.28824830e+00
-1.06184614e+00 -1.97504425e+00 -5.13176143e-01 8.81894469e-01
-4.10782158e-01 -1.85825616e-01 -5.81228077e-01 3.31899375e-01
1.45265147e-01 8.79274249e-01 -4.45197463e-01 -2.33930588e-01
1.94345549e-01 -1.33413160e-02 8.89786005e-01 -7.94106901e-01
-3.26782495e-01 2.28343681e-01 1.96216464e-01 -3.37132782e-01
-5.21931648e-01 -7.20584810e-01 -1.50315905e+00 -5.61890662e-01
-8.32215846e-01 5.50630808e-01 7.96637237e-01 7.37724364e-01
8.52388591e-02 2.70369649e-01 3.35847557e-01 -1.04440618e+00
1.37882739e-01 -5.49966872e-01 -8.87625575e-01 -3.01980134e-03
-1.76413450e-02 -1.03314090e+00 -1.07336685e-01 -2.79169559e-01] | [5.1634392738342285, 5.427289962768555] |
9e8a0003-4c6f-4c4c-b43c-4cb885e96499 | a-novel-time-varying-spectral-filtering | null | null | https://doi.org/10.3390/s16010010 | https://www.mdpi.com/1424-8220/16/1/10/pdf | A Novel Time-Varying Spectral Filtering Algorithm for Reconstruction of Motion Artifact Corrupted Heart Rate Signals During Intense Physical Activities Using a Wearable Photoplethysmogram Sensor | Accurate estimation of heart rates from photoplethysmogram (PPG) signals during intense physical activity is a very challenging problem. This is because strenuous and high intensity exercise can result in severe motion artifacts in PPG signals, making accurate heart rate (HR) estimation difficult. In this study we investigated a novel technique to accurately reconstruct motion-corrupted PPG signals and HR based on time-varying spectral analysis. The algorithm is called Spectral filter algorithm for Motion Artifacts and heart rate reconstruction (SpaMA). The idea is to calculate the power spectral density of both PPG and accelerometer signals for each time shift of a windowed data segment. By comparing time-varying spectra of PPG and accelerometer data, those frequency peaks resulting from motion artifacts can be distinguished from the PPG spectrum. The SpaMA approach was applied to three different datasets and four types of activities: (1) training datasets from the 2015 IEEE Signal Process. Cup Database recorded from 12 subjects while performing treadmill exercise from 1 km/h to 15 km/h; (2) test datasets from the 2015 IEEE Signal Process. Cup Database recorded from 11 subjects while performing forearm and upper arm exercise. (3) Chon Lab dataset including 10 min recordings from 10 subjects during treadmill exercise. The ECG signals from all three datasets provided the reference HRs which were used to determine the accuracy of our SpaMA algorithm. The performance of the SpaMA approach was calculated by computing the mean absolute error between the estimated HR from the PPG and the reference HR from the ECG. The average estimation errors using our method on the first, second and third datasets are 0.89, 1.93 and 1.38 beats/min respectively, while the overall error on all 33 subjects is 1.86 beats/min and the performance on only treadmill experiment datasets (22 subjects) is 1.11 beats/min. Moreover, it was found that dynamics of heart rate variability can be accurately captured using the algorithm where the mean Pearson’s correlation coefficient between the power spectral densities of the reference and the reconstructed heart rate time series was found to be 0.98. These results show that the SpaMA method has a potential for PPG-based HR monitoring in wearable devices for fitness tracking and health monitoring during intense physical activities. | ['Chae Cho', 'Seyed M. A. Salehizadeh', 'Duy Dao', 'Yitzhak Mendelson and Ki H. Chon', 'Jeffrey Bolkhovsky'] | 2015-12-23 | null | null | null | sensors-2015-12 | ['photoplethysmography-ppg', 'heart-rate-variability', 'heart-rate-estimation'] | ['medical', 'medical', 'medical'] | [ 4.94017363e-01 -2.45038986e-01 9.43433959e-03 4.50084805e-02
-5.73171318e-01 -1.88982487e-01 -2.33526781e-01 -2.18616545e-01
-2.23734498e-01 7.91467369e-01 -3.39338346e-03 -2.52640173e-02
-2.46321917e-01 -5.03002167e-01 -2.29662910e-01 -6.87598586e-01
-4.16180432e-01 -2.95898765e-01 -1.55032992e-01 2.74508506e-01
-6.02495857e-02 2.83324540e-01 -1.21305740e+00 -2.50154883e-01
8.00361812e-01 1.14687049e+00 -3.10207576e-01 1.11392152e+00
7.15378463e-01 2.66749978e-01 -1.04547095e+00 2.01827273e-01
2.20742777e-01 -1.16213870e+00 -2.46591747e-01 1.08995393e-01
9.99929830e-02 -1.52814001e-01 -1.14786744e-01 5.47977746e-01
1.05093980e+00 1.52075917e-01 2.12525815e-01 -1.02520216e+00
2.68066347e-01 2.03934520e-01 -5.25761068e-01 5.77230692e-01
5.69005251e-01 8.38664845e-02 7.98643604e-02 -5.10562003e-01
2.87395328e-01 4.41557109e-01 1.35016847e+00 6.89022467e-02
-1.40937471e+00 -6.39548540e-01 -8.77712071e-01 2.96112180e-01
-1.68374622e+00 -3.55092108e-01 1.08777094e+00 -4.93909836e-01
5.95036089e-01 7.89005995e-01 1.24613440e+00 7.74998486e-01
6.42510951e-01 -2.01005608e-01 1.49635386e+00 -3.69409204e-01
2.24773899e-01 3.76258940e-02 2.97106922e-01 3.47246945e-01
4.63178277e-01 1.84309006e-01 -5.47123373e-01 -2.55024046e-01
9.33741271e-01 -5.76232448e-02 -5.95013201e-01 3.09095651e-01
-1.24635613e+00 3.71306777e-01 -1.52030200e-01 4.50438380e-01
-5.51882088e-01 1.60979331e-01 4.20537978e-01 2.97566116e-01
2.08792642e-01 3.36805284e-01 -3.31085086e-01 -7.10702419e-01
-1.01618755e+00 9.05412287e-02 9.50927138e-01 3.09090585e-01
2.04884365e-01 3.68312567e-01 -1.46893069e-01 6.84357464e-01
1.84097841e-01 7.59849370e-01 6.51842773e-01 -1.18268800e+00
6.34666830e-02 2.18675375e-01 3.72218519e-01 -1.29619694e+00
-8.11363459e-01 -4.90297556e-01 -7.81336188e-01 -3.59296739e-01
7.70616770e-01 -4.22772795e-01 -2.10645810e-01 1.39629972e+00
4.83989358e-01 4.84774590e-01 -7.41649121e-02 1.12844574e+00
6.41228557e-01 4.80957687e-01 7.93824643e-02 -9.48788524e-01
1.59575009e+00 -6.68687820e-02 -1.18436384e+00 1.45780966e-02
2.79468447e-01 -5.53649127e-01 9.35248137e-01 4.81060892e-01
-1.03547382e+00 -1.10013020e+00 -1.10225034e+00 5.48673511e-01
3.12801242e-01 3.07433009e-01 -4.56339903e-02 1.11687911e+00
-4.68335897e-01 1.10063028e+00 -8.59950900e-01 -2.86017030e-01
-1.91317096e-01 3.34221981e-02 -7.47698499e-03 4.17482674e-01
-1.38579881e+00 6.13485098e-01 -7.51653314e-02 4.35981244e-01
-1.66133851e-01 -8.09245169e-01 -5.92743635e-01 -1.31341651e-01
-1.54778026e-02 -3.81822109e-01 7.85349846e-01 -6.44635975e-01
-1.86533654e+00 4.81663942e-01 -2.02167779e-01 -2.72392631e-01
5.12134671e-01 -3.52869093e-01 -8.70769739e-01 3.71496916e-01
-9.97684374e-02 -5.76979041e-01 8.38008404e-01 -4.37117636e-01
1.95423022e-01 -4.05562311e-01 -9.50148940e-01 9.60476473e-02
1.88316569e-01 -2.75872856e-01 1.10899009e-01 -4.77411449e-01
6.30680323e-01 -1.03389609e+00 1.65127382e-01 -4.35113490e-01
-2.06954643e-01 4.54400808e-01 4.29435074e-01 -1.03005362e+00
1.57883716e+00 -2.33451891e+00 -2.63624936e-01 4.46030498e-01
1.12267612e-02 1.78958595e-01 5.49805343e-01 3.04281861e-01
-1.17196463e-01 8.42047930e-02 -1.74359322e-01 3.33798856e-01
-3.69640380e-01 5.33945598e-02 -3.82355903e-03 9.04554188e-01
-3.40221971e-01 5.84729373e-01 -6.31329179e-01 -2.30868086e-01
4.03108805e-01 5.65613866e-01 2.17482626e-01 3.12030673e-01
7.15507686e-01 6.81338847e-01 -6.70777857e-02 4.63416070e-01
4.66395408e-01 1.40067026e-01 5.71349502e-01 -7.70348608e-01
-1.13611065e-01 1.22798659e-01 -1.41333473e+00 1.27782762e+00
-2.05440640e-01 7.07648933e-01 -2.65056640e-01 -8.17808330e-01
1.41705739e+00 8.19334745e-01 8.42459977e-01 -7.70532548e-01
9.20981914e-02 1.78344503e-01 3.29123676e-01 -9.93628144e-01
-4.67046816e-03 -5.28391719e-01 -5.50210029e-02 4.50645030e-01
-2.05918700e-01 -2.03269973e-01 3.86368968e-02 -2.21009970e-01
9.42803800e-01 1.17376208e-01 5.80267370e-01 -4.44293588e-01
5.20046473e-01 -3.40154320e-01 7.82913864e-01 5.81591845e-01
-6.85179710e-01 6.06998503e-01 3.44329089e-01 -5.43973207e-01
-5.28421700e-01 -9.89307344e-01 -3.07917297e-01 1.82522535e-01
-9.10861492e-02 -4.78695989e-01 -6.70410156e-01 4.51153070e-02
-4.72781323e-02 3.54791641e-01 -3.83753389e-01 -3.62458974e-01
-5.78916788e-01 -9.62863386e-01 7.99417257e-01 4.61161464e-01
7.85440266e-01 -7.93500364e-01 -1.01635444e+00 5.36444545e-01
-6.63320005e-01 -1.10637164e+00 -3.96453530e-01 -1.49836510e-01
-1.21799386e+00 -1.29280686e+00 -5.19269586e-01 2.44103689e-02
6.47607967e-02 -1.38739020e-01 9.43351269e-01 -1.08811937e-01
-7.26699650e-01 5.24081051e-01 -1.47000924e-01 -4.37998027e-01
-1.72296301e-01 -4.78784502e-01 4.19501096e-01 2.12642148e-01
1.44304425e-01 -7.53607929e-01 -9.09362257e-01 6.98228776e-01
-1.74174100e-01 -4.18388456e-01 4.50754426e-02 4.40673351e-01
6.99448228e-01 -3.34303565e-02 7.75217474e-01 -3.95289987e-01
8.14274848e-01 -2.77629256e-01 -4.68229473e-01 -3.71291339e-01
-6.77831650e-01 -7.91198134e-01 4.61654156e-01 -6.72703803e-01
-5.59503675e-01 7.99208283e-02 9.30045545e-02 -1.63041607e-01
-6.29662350e-02 2.13938624e-01 1.39465839e-01 2.18347356e-01
9.75225508e-01 2.58884847e-01 2.38073394e-01 -2.74488926e-01
-2.35840321e-01 5.04579842e-01 1.04893732e+00 -4.06487763e-01
4.55406785e-01 1.16162606e-01 5.15216827e-01 -1.35466874e+00
-4.14015859e-01 -5.32435238e-01 -3.11234623e-01 -7.66960680e-01
9.73980188e-01 -9.11738634e-01 -1.02773857e+00 7.30249107e-01
-5.93513608e-01 -3.29727739e-01 -4.23878700e-01 1.15330434e+00
-6.41207755e-01 6.69484317e-01 -7.44494557e-01 -1.30049682e+00
-7.44508862e-01 -3.37865591e-01 3.84885639e-01 4.88964856e-01
-8.56686532e-01 -8.14774752e-01 1.53029889e-01 5.80523133e-01
4.86821145e-01 1.21333265e+00 2.88119733e-01 1.26556456e-01
3.08728069e-01 -4.96600926e-01 4.71983522e-01 3.87718230e-01
3.86268258e-01 -1.47878751e-01 -1.03013277e+00 -3.42941254e-01
7.33615577e-01 1.85170442e-01 -8.16315636e-02 7.98593879e-01
6.11751914e-01 -1.85191914e-01 -8.36258903e-02 4.98021185e-01
1.47281229e+00 3.54037166e-01 1.18084335e+00 -1.89504370e-01
4.67975497e-01 4.77543503e-01 6.24791026e-01 5.14671028e-01
-1.75580099e-01 7.49304712e-01 -5.47585525e-02 -2.78038144e-01
1.30493781e-02 9.94115248e-02 5.18705547e-01 7.90897846e-01
-7.76178896e-01 2.84447312e-01 -5.67771316e-01 2.08110988e-01
-1.32520306e+00 -1.03676963e+00 -1.02754223e+00 2.72004890e+00
7.66114891e-01 -9.46048945e-02 5.72024703e-01 8.84181321e-01
7.78041005e-01 -8.79698023e-02 -6.20915413e-01 -7.18926668e-01
1.36069626e-01 4.78601187e-01 4.57269728e-01 2.25253761e-01
-7.61242449e-01 -4.00013119e-01 6.08641100e+00 -4.80425879e-02
-1.17860079e+00 2.42573526e-02 4.37734187e-01 -1.54121116e-01
3.93129617e-01 -9.65507701e-02 -3.63402188e-01 8.80705893e-01
1.73322248e+00 -4.49965298e-01 1.65326372e-01 4.71657723e-01
8.58834565e-01 -6.00552678e-01 -6.64715528e-01 1.28218865e+00
-1.57832745e-02 -6.44081593e-01 -1.15657401e+00 -4.93546091e-02
4.11634803e-01 -5.06999433e-01 -5.21000683e-01 -1.29119121e-02
-7.97483206e-01 -7.84443021e-01 3.13509822e-01 9.39771771e-01
9.86396968e-01 -5.74921906e-01 7.14301765e-01 2.92184114e-01
-1.36934829e+00 1.81803256e-01 3.79321389e-02 -3.71372610e-01
2.71220058e-01 1.27891207e+00 -5.26492596e-01 6.54072821e-01
4.17256504e-01 6.08298659e-01 -1.62798628e-01 1.03540337e+00
-2.48378098e-01 1.14229369e+00 -5.24181724e-01 2.17508167e-01
-6.74539685e-01 -3.70563596e-01 6.46832943e-01 8.14571977e-01
5.05433917e-01 5.33043385e-01 -1.52698457e-01 9.22598898e-01
5.03920078e-01 1.33400457e-02 -4.68505681e-01 1.43435314e-01
4.29681510e-01 1.07655954e+00 -4.64240670e-01 -3.36570799e-01
-2.16346025e-01 4.11301494e-01 -6.95165873e-01 3.25030446e-01
-1.08626151e+00 -8.52565110e-01 4.80503738e-01 5.07137775e-01
-4.40301210e-01 -3.01772565e-01 -4.36094910e-01 -6.21786237e-01
2.56018907e-01 -6.76724553e-01 4.80494171e-01 -7.19249308e-01
-7.41174400e-01 2.51451761e-01 4.82397527e-02 -1.38392365e+00
-4.03171271e-01 1.40228808e-01 -7.57847905e-01 1.34975767e+00
-7.79796660e-01 1.88197523e-01 -7.60615587e-01 4.90450680e-01
2.29451016e-01 6.89711571e-01 9.40466166e-01 5.27068436e-01
-6.91743791e-01 2.51539677e-01 -4.12347645e-01 -1.99166611e-01
4.54148173e-01 -9.43241119e-01 -8.69422802e-04 7.98202693e-01
-3.93977672e-01 6.01205826e-01 8.18063498e-01 -5.47962427e-01
-1.58636475e+00 -8.75562966e-01 9.77669597e-01 -1.22151643e-01
9.82940421e-02 4.75582890e-02 -9.71069872e-01 -1.64024327e-02
-2.43405163e-01 3.12730461e-01 8.12193811e-01 -4.73310620e-01
5.12456894e-01 -4.90438759e-01 -1.33216155e+00 6.08490892e-02
4.02074367e-01 -4.61045891e-01 -4.48256820e-01 -3.14480923e-02
-3.13538946e-02 -6.12472713e-01 -1.82664478e+00 4.94754493e-01
1.14304781e+00 -7.88050354e-01 7.51875281e-01 1.09231874e-01
-1.24515556e-01 -6.24172747e-01 1.29638761e-01 -1.01568294e+00
-1.03643090e-01 -1.21363938e+00 -3.31860691e-01 1.06122470e+00
-5.46651296e-02 -9.40363586e-01 3.71170074e-01 8.09110343e-01
1.70115620e-01 -4.69741911e-01 -8.15599084e-01 -8.89837146e-01
-7.08213091e-01 -4.78533059e-01 -4.04398292e-02 9.42232490e-01
3.27423275e-01 2.02739492e-01 -7.57665634e-01 -1.64481695e-03
8.21423531e-01 -5.94867654e-02 7.74609327e-01 -1.11650980e+00
-4.40037847e-01 3.58206481e-01 -4.16378766e-01 -3.01019490e-01
-8.31719816e-01 -2.86800385e-01 -9.76952538e-02 -1.33617818e+00
-2.02913776e-01 -4.39404417e-03 -3.62401843e-01 1.43590346e-01
-3.01197797e-01 6.08553350e-01 8.35997239e-02 9.02972221e-02
3.91579986e-01 -3.35639678e-02 9.62334573e-01 3.33081722e-01
-8.89671087e-01 4.45814669e-01 -1.84908643e-01 4.59933013e-01
8.41859162e-01 -5.54736435e-01 -6.04521394e-01 5.66873431e-01
1.09954653e-02 7.74803400e-01 3.50798279e-01 -1.57294130e+00
-3.62241060e-01 1.65618524e-01 6.23715162e-01 -3.57920051e-01
2.80939937e-01 -6.26310706e-01 1.30067694e+00 9.76635277e-01
5.17527610e-02 -7.97625333e-02 2.03679755e-01 2.46037543e-01
9.04543418e-03 2.39989683e-01 8.73392403e-01 -1.35439903e-01
-2.44781794e-03 -2.85997599e-01 -6.02720022e-01 5.99837936e-02
7.81120479e-01 -8.17098081e-01 -2.34795719e-01 -3.64807248e-01
-1.12732482e+00 -3.61408263e-01 -4.70012762e-02 -3.81770320e-02
5.84859014e-01 -1.41322780e+00 -5.53827465e-01 3.46333176e-01
-3.16349685e-01 -5.53990543e-01 6.75064027e-01 1.64698970e+00
-4.53742087e-01 8.44937712e-02 -3.06523204e-01 -8.41171503e-01
-1.29094112e+00 4.11403514e-02 7.29092717e-01 -2.70501189e-02
-8.21389496e-01 9.37529206e-02 -6.49880469e-01 6.64458692e-01
-9.10184756e-02 -4.95499521e-01 -2.38656044e-01 1.73480972e-03
5.52817047e-01 1.06835163e+00 1.68195546e-01 -3.82364959e-01
-4.44954723e-01 1.02044356e+00 8.06570292e-01 -4.74213995e-02
6.78427398e-01 -4.23698723e-01 1.37955591e-01 8.19016993e-01
1.13811839e+00 5.32676987e-02 -8.74804556e-01 2.49068692e-01
-1.42128721e-01 -3.75510007e-01 -1.10608175e-01 -7.80090392e-01
-1.02023542e+00 6.13775432e-01 1.22347569e+00 5.55981219e-01
1.60527694e+00 -7.69971013e-01 7.95671821e-01 -3.44943196e-01
1.97459519e-01 -1.21606672e+00 -5.88084236e-02 -2.33834118e-01
6.42838001e-01 -2.61664420e-01 3.94756496e-01 -5.05979478e-01
-4.83933777e-01 1.11577642e+00 2.05619201e-01 -2.23687842e-01
5.01504421e-01 -1.20289652e-02 2.60931373e-01 1.92487072e-02
-2.34610841e-01 2.38265663e-01 1.52626052e-01 6.81937814e-01
7.74783611e-01 2.17502192e-01 -1.13211024e+00 6.20498955e-01
-1.04063980e-01 6.24076605e-01 8.44827354e-01 8.34305704e-01
-2.48401046e-01 -5.23247778e-01 -7.43909001e-01 3.90773684e-01
-7.75175214e-01 5.08329630e-01 -3.70322727e-02 8.02381575e-01
-3.97191718e-02 1.33108270e+00 -1.70403093e-01 -3.49242628e-01
8.30317974e-01 5.26546717e-01 4.81753886e-01 -1.47141829e-01
-7.38899887e-01 2.49156162e-01 3.80139768e-01 -8.14863861e-01
-5.13827264e-01 -5.89236557e-01 -1.22003174e+00 -1.24344654e-01
-1.68399170e-01 2.76244640e-01 8.25334609e-01 4.10484135e-01
3.40899974e-01 7.13109672e-01 7.68027961e-01 -5.58469057e-01
-3.13717037e-01 -1.14983690e+00 -1.21423745e+00 4.76899683e-01
2.31600538e-01 -3.63395810e-01 -6.67000473e-01 3.35616350e-01] | [13.949774742126465, 3.001685380935669] |
301aa8b9-0b04-48bf-bc28-01d5115f9e6f | kgpt-knowledge-grounded-pre-training-for-data | 2010.02307 | null | https://arxiv.org/abs/2010.02307v2 | https://arxiv.org/pdf/2010.02307v2.pdf | KGPT: Knowledge-Grounded Pre-Training for Data-to-Text Generation | Data-to-text generation has recently attracted substantial interests due to its wide applications. Existing methods have shown impressive performance on an array of tasks. However, they rely on a significant amount of labeled data for each task, which is costly to acquire and thus limits their application to new tasks and domains. In this paper, we propose to leverage pre-training and transfer learning to address this issue. We propose a knowledge-grounded pre-training (KGPT), which consists of two parts, 1) a general knowledge-grounded generation model to generate knowledge-enriched text. 2) a pre-training paradigm on a massive knowledge-grounded text corpus crawled from the web. The pre-trained model can be fine-tuned on various data-to-text generation tasks to generate task-specific text. We adopt three settings, namely fully-supervised, zero-shot, few-shot to evaluate its effectiveness. Under the fully-supervised setting, our model can achieve remarkable gains over the known baselines. Under zero-shot setting, our model without seeing any examples achieves over 30 ROUGE-L on WebNLG while all other baselines fail. Under the few-shot setting, our model only needs about one-fifteenth as many labeled examples to achieve the same level of performance as baseline models. These experiments consistently prove the strong generalization ability of our proposed framework https://github.com/wenhuchen/KGPT. | ['William Yang Wang', 'Xifeng Yan', 'Yu Su', 'Wenhu Chen'] | 2020-10-05 | null | https://aclanthology.org/2020.emnlp-main.697 | https://aclanthology.org/2020.emnlp-main.697.pdf | emnlp-2020-11 | ['kg-to-text'] | ['natural-language-processing'] | [ 2.80901611e-01 1.00573502e-01 -2.74265766e-01 -2.09328234e-01
-1.36887133e+00 -4.56713468e-01 9.34822440e-01 -6.10619895e-02
-3.78550887e-01 9.29919958e-01 2.49194458e-01 -2.89098084e-01
1.66344851e-01 -9.01786685e-01 -7.24095047e-01 -5.84376454e-01
4.53616798e-01 5.61134994e-01 3.39968354e-01 -4.55312848e-01
9.57299024e-02 -3.09482723e-01 -1.38388503e+00 3.39351922e-01
1.27026665e+00 8.25131476e-01 5.94655812e-01 6.16425276e-01
-4.29786086e-01 5.19187987e-01 -4.47445422e-01 -4.62213457e-01
1.51362166e-01 -5.52003682e-01 -8.81417811e-01 3.17290872e-02
1.01660684e-01 -3.39515299e-01 -2.48040870e-01 8.98294330e-01
7.56290674e-01 1.09042257e-01 6.25282407e-01 -1.29846621e+00
-8.93181145e-01 8.40061307e-01 -4.87053037e-01 -8.62070620e-02
2.98127979e-01 2.08421841e-01 1.11759448e+00 -1.16737103e+00
5.08448899e-01 1.07083511e+00 4.08337206e-01 7.37819433e-01
-9.15647030e-01 -6.27926528e-01 8.68259743e-02 5.21533191e-02
-1.13275766e+00 -3.95091027e-01 5.23272753e-01 -3.57279897e-01
7.13602066e-01 -8.71148556e-02 2.54703313e-01 1.36473584e+00
-2.30522960e-01 9.71418142e-01 1.06911516e+00 -6.96827948e-01
2.33837172e-01 1.74573481e-01 3.15417349e-02 4.69324052e-01
2.97965795e-01 -8.89023095e-02 -5.52318275e-01 -2.09081158e-01
5.00997305e-01 1.52240507e-02 -3.64171684e-01 -1.26174092e-01
-1.35064018e+00 9.22543705e-01 2.88178116e-01 2.40648016e-01
-3.00479054e-01 2.20191143e-02 4.61045504e-01 1.37233093e-01
6.07310176e-01 4.75863755e-01 -5.70307553e-01 -1.32321328e-01
-8.62203181e-01 4.43236053e-01 7.44522333e-01 1.32323813e+00
8.43962252e-01 -7.36924261e-02 -6.30592585e-01 1.09163213e+00
-9.37181409e-04 5.34496307e-01 9.49574113e-01 -4.12659496e-01
1.03145778e+00 5.01863539e-01 2.68034190e-01 -5.04912257e-01
-7.42076859e-02 -2.81985015e-01 -9.25539196e-01 -3.87401998e-01
2.60634333e-01 -4.78789002e-01 -1.03598380e+00 1.67428958e+00
4.13704664e-01 1.63763463e-01 3.05878401e-01 7.41412759e-01
7.51599431e-01 9.16278183e-01 1.34713560e-01 -8.75130072e-02
1.36832464e+00 -1.30444729e+00 -6.06265724e-01 -2.84423679e-01
7.96591222e-01 -7.77773142e-01 1.61969662e+00 1.66169941e-01
-8.55441213e-01 -4.78572220e-01 -9.00093496e-01 4.65554781e-02
-3.99766296e-01 1.43446609e-01 3.08830798e-01 4.00883615e-01
-7.79722571e-01 3.98165435e-01 -5.92679620e-01 -5.06512344e-01
4.05630082e-01 -1.09711237e-01 -6.10459782e-02 -3.14363807e-01
-1.38795578e+00 4.64459836e-01 7.14490294e-01 -2.92462617e-01
-1.07554030e+00 -6.18067086e-01 -7.00668216e-01 9.36811045e-02
7.15350389e-01 -8.30685258e-01 1.71231198e+00 -6.19470298e-01
-1.45767355e+00 5.62688768e-01 -2.52129555e-01 -3.16266894e-01
5.90540588e-01 -3.69476527e-01 -3.46642852e-01 -2.44971253e-02
3.09374034e-01 5.78637064e-01 7.35363007e-01 -1.25351071e+00
-8.26075137e-01 -1.33912653e-01 1.24588154e-01 3.66846472e-01
-7.33998537e-01 -2.21769944e-01 -6.56036675e-01 -8.88883412e-01
-3.38010311e-01 -8.82060647e-01 -2.05426678e-01 -4.84076858e-01
-6.66330457e-01 -5.02598464e-01 7.84017503e-01 -4.78183031e-01
1.16538274e+00 -1.80691373e+00 -2.59177953e-01 -3.59262019e-01
1.01411140e-04 5.54705381e-01 -3.63013119e-01 8.35054696e-01
2.05335438e-01 3.90058100e-01 -2.38739848e-01 -2.48614609e-01
9.92380679e-02 -2.25805659e-02 -5.16250134e-01 -3.16599131e-01
2.63850063e-01 1.20628035e+00 -1.17424703e+00 -5.57041228e-01
-8.47809687e-02 2.41114050e-01 -3.63016844e-01 4.74644125e-01
-5.94983399e-01 2.36322224e-01 -7.47149706e-01 4.03807551e-01
4.51843768e-01 -6.15591705e-01 1.35007426e-02 1.77141484e-02
1.47585616e-01 3.24352413e-01 -8.59273493e-01 1.80449092e+00
-7.68458247e-01 2.26350680e-01 -4.38714951e-01 -9.05018747e-01
7.47658670e-01 5.23302376e-01 7.47113302e-03 -6.21047616e-01
4.23793169e-03 2.54374713e-01 -2.43445903e-01 -5.29029667e-01
5.64359069e-01 -2.49291524e-01 -2.13906378e-01 7.92321563e-01
1.99890509e-01 2.56419387e-02 4.62416708e-01 5.62169135e-01
1.10144413e+00 2.63309896e-01 2.47545332e-01 1.75304059e-02
2.92832702e-01 3.01674511e-02 4.72276688e-01 7.88752735e-01
1.41151875e-01 7.49587774e-01 3.73116553e-01 -4.00102772e-02
-1.03960121e+00 -8.12078357e-01 2.14632064e-01 1.49026704e+00
2.39526168e-01 -3.74770343e-01 -8.60005200e-01 -9.54018891e-01
-1.70362771e-01 9.34163570e-01 -5.18429637e-01 -1.36747360e-01
-2.93960720e-01 -9.39709067e-01 5.08078754e-01 5.11643529e-01
6.55225515e-01 -1.27008915e+00 -2.04445422e-01 1.70359701e-01
-5.56290329e-01 -1.26853561e+00 -6.59191191e-01 -8.81313905e-02
-7.45174587e-01 -8.90979707e-01 -1.03262913e+00 -8.19362223e-01
6.73153341e-01 5.83472311e-01 1.09259522e+00 -3.48882265e-02
-1.28462855e-02 1.45776078e-01 -9.71332729e-01 -5.04381418e-01
-5.09849906e-01 4.69615877e-01 -1.75113574e-01 -7.74074420e-02
2.68358707e-01 -4.64629203e-01 -6.35943532e-01 3.15535516e-01
-1.01173794e+00 4.07155901e-01 9.60337281e-01 9.97496128e-01
5.37999749e-01 -8.11489532e-04 1.10770214e+00 -1.12246203e+00
9.76455748e-01 -7.43874371e-01 -2.30165854e-01 4.82379824e-01
-7.62909770e-01 1.36345327e-01 1.02034950e+00 -4.18791473e-01
-1.35843265e+00 -1.47139311e-01 2.60827467e-02 -1.33517742e-01
-6.23236187e-02 6.33971989e-01 -2.71164000e-01 4.99322772e-01
7.68993139e-01 5.13563275e-01 -3.71433884e-01 -5.24385393e-01
5.78714371e-01 9.80095327e-01 3.66267145e-01 -8.25996757e-01
1.05228400e+00 1.89583138e-01 -6.39160216e-01 -5.93036711e-01
-1.26300716e+00 -4.79543746e-01 -4.37955707e-01 1.24963678e-01
5.90178728e-01 -1.03720379e+00 3.05253372e-04 4.48363543e-01
-1.02965820e+00 -6.33681834e-01 -2.53346533e-01 2.79804140e-01
-5.16040683e-01 3.62663746e-01 -4.44058269e-01 -6.66735947e-01
-9.67727065e-01 -8.18534017e-01 1.10242236e+00 1.79101527e-01
2.06069976e-01 -8.77845705e-01 1.12916149e-01 4.85263616e-01
3.22714716e-01 3.79069299e-02 8.92204285e-01 -9.73413408e-01
-3.59654635e-01 -1.86540112e-01 -3.39252949e-01 3.77075344e-01
2.84616798e-01 -3.10544193e-01 -9.58345175e-01 -2.56813735e-01
-4.13087547e-01 -8.60323787e-01 7.89136648e-01 -1.44339561e-01
1.27619052e+00 -5.07943511e-01 -3.38267922e-01 2.69894868e-01
1.41594589e+00 5.67385405e-02 4.82738227e-01 1.80978805e-01
6.18321240e-01 4.17753935e-01 8.38868439e-01 5.04501462e-01
5.22665381e-01 6.76299870e-01 2.43915930e-01 -6.83599710e-02
-1.74981490e-01 -6.52847707e-01 3.41428280e-01 9.00508225e-01
4.62225340e-02 -5.86492062e-01 -9.22820926e-01 7.52212763e-01
-1.91787541e+00 -9.06996071e-01 8.14998597e-02 2.23654962e+00
1.25004804e+00 1.96157381e-01 2.15022892e-01 -9.83322114e-02
9.00966823e-01 9.42897126e-02 -6.54615104e-01 6.43999875e-02
1.45196572e-01 9.94373485e-02 5.48945032e-02 2.31616929e-01
-9.29800749e-01 1.16172028e+00 5.25905514e+00 1.19531047e+00
-9.14181888e-01 3.30490798e-01 5.73769569e-01 -3.65075916e-02
-4.66846526e-01 -2.41970979e-02 -1.07519615e+00 7.72881329e-01
8.90230775e-01 -7.55308568e-01 2.36475542e-01 1.03473461e+00
2.42344335e-01 1.98140070e-01 -9.03408647e-01 7.81757236e-01
5.15486188e-02 -1.21643007e+00 4.33341503e-01 -6.84369430e-02
9.93496954e-01 6.66783229e-02 -1.70883849e-01 7.21589148e-01
7.77896941e-01 -8.68397951e-01 4.64995623e-01 1.10820368e-01
1.04151392e+00 -6.11559570e-01 8.16444218e-01 7.87608445e-01
-1.05636728e+00 9.27846506e-02 -4.74802554e-01 1.00077070e-01
3.71354282e-01 9.04831171e-01 -1.09330344e+00 8.46388817e-01
4.18449551e-01 4.20033813e-01 -3.85414809e-01 8.25417757e-01
-5.11979461e-01 8.10078561e-01 -1.34530038e-01 -2.26213768e-01
3.53552699e-01 8.58084708e-02 1.02583013e-01 1.23187637e+00
5.89837730e-01 6.75709248e-02 3.51528347e-01 7.17301488e-01
-5.16276002e-01 3.91831070e-01 -4.67841804e-01 -2.32646883e-01
7.42057145e-01 1.50794280e+00 -4.53222305e-01 -6.55489266e-01
-5.23288846e-01 1.00619602e+00 5.04153430e-01 4.26298112e-01
-7.46222734e-01 -7.89395630e-01 1.36956766e-01 1.72696620e-01
4.39376891e-01 3.57036106e-02 -6.93795756e-02 -1.35328126e+00
1.67572826e-01 -8.88767660e-01 3.15156579e-01 -8.36424530e-01
-1.49631655e+00 7.08812356e-01 -1.49961922e-03 -1.26794600e+00
-5.30975163e-01 -3.70981753e-01 -7.93945253e-01 8.60498488e-01
-1.64999938e+00 -1.24294388e+00 -6.69499874e-01 5.37298858e-01
8.52883160e-01 -2.01685324e-01 6.82021439e-01 8.46802145e-02
-6.53121412e-01 7.39690542e-01 2.24618942e-01 2.96207160e-01
9.30684626e-01 -1.31178188e+00 6.51071370e-01 8.12167406e-01
1.24210812e-01 4.59865600e-01 4.20137554e-01 -6.83224261e-01
-1.21658826e+00 -1.46016145e+00 8.95339072e-01 -3.61503214e-01
6.60743296e-01 -4.55558956e-01 -9.04579103e-01 6.58184290e-01
3.09736878e-01 -1.16225071e-01 6.58309698e-01 1.37355298e-01
-4.28242564e-01 -1.00890756e-01 -8.17445457e-01 6.50582194e-01
1.05370879e+00 -2.19415024e-01 -6.64448261e-01 6.61080480e-01
1.06934488e+00 -2.32164651e-01 -6.63597047e-01 3.08153778e-01
2.82802224e-01 -6.80909276e-01 6.16944432e-01 -6.17187202e-01
8.85697722e-01 -7.31774420e-02 4.63838205e-02 -1.59875774e+00
-1.82051614e-01 -7.65441716e-01 -1.72763884e-01 1.46382952e+00
6.43047571e-01 -6.63263857e-01 6.79662287e-01 2.80633032e-01
-2.25659177e-01 -9.57575202e-01 -5.15278101e-01 -1.03212762e+00
2.64070719e-01 -3.13127816e-01 7.43865788e-01 9.76068616e-01
1.15764566e-01 9.07078862e-01 -4.83575761e-01 -2.15035439e-01
5.48183203e-01 1.94023281e-01 1.10819292e+00 -9.35501337e-01
-3.37073028e-01 -1.98098183e-01 2.84889072e-01 -1.25842905e+00
-4.20213677e-02 -1.09783351e+00 3.06858599e-01 -1.84541428e+00
5.49394190e-01 -6.22777283e-01 -2.03148663e-01 8.08457851e-01
-7.46350825e-01 2.47799288e-02 9.74280685e-02 3.98486376e-01
-8.08143198e-01 8.71081412e-01 1.34014416e+00 1.61179081e-01
-1.43258929e-01 -1.47334859e-01 -1.06334364e+00 4.35354143e-01
1.03232527e+00 -4.27939236e-01 -7.10331976e-01 -5.31375110e-01
1.14582814e-01 -1.94907002e-02 1.37046576e-01 -9.09631789e-01
7.56444111e-02 -2.53868192e-01 1.12439305e-01 -3.85915607e-01
1.87345799e-02 -2.88151562e-01 -2.37676278e-01 2.75265098e-01
-4.91643578e-01 -1.58335507e-01 -1.13039715e-02 7.09369183e-01
-2.02362180e-01 -3.82499874e-01 6.30345643e-01 -1.84171587e-01
-6.93718731e-01 5.41289568e-01 4.49963473e-02 5.65071642e-01
9.65904236e-01 1.66961312e-01 -6.07456028e-01 -5.31896889e-01
-3.13763648e-01 3.75589699e-01 3.26570123e-01 5.83898723e-01
3.96467179e-01 -1.42541420e+00 -1.00273585e+00 -7.63937384e-02
4.74487573e-01 2.70078540e-01 3.50680143e-01 5.11070013e-01
-2.16911554e-01 5.02714276e-01 1.82503805e-01 -3.71373951e-01
-7.99932182e-01 6.23063982e-01 -1.21364892e-01 -7.53240287e-01
-5.30682683e-01 6.21805489e-01 4.13531125e-01 -5.87833524e-01
-7.71410540e-02 9.07207951e-02 4.78074513e-02 -8.02213773e-02
7.13740408e-01 2.48840898e-01 5.56166582e-02 -1.62645400e-01
1.08591951e-01 2.87124693e-01 -4.46257710e-01 -1.42252982e-01
1.09231198e+00 3.09478380e-02 4.01501447e-01 3.77644271e-01
1.00404859e+00 -2.45511919e-01 -1.15578055e+00 -5.84995687e-01
-6.95757046e-02 -3.07712108e-01 -1.49124846e-01 -9.98382151e-01
-8.07321966e-01 1.00332320e+00 6.14725277e-02 2.03466147e-01
1.02210379e+00 1.00488298e-01 1.28958952e+00 5.76257348e-01
5.79213798e-01 -1.14672029e+00 4.95861620e-01 6.00039542e-01
8.22647274e-01 -1.43004262e+00 -1.87253579e-01 -4.53113675e-01
-9.34330106e-01 7.23060191e-01 7.27098763e-01 1.58841178e-01
3.44457775e-01 -3.90445888e-02 9.75127593e-02 1.09480910e-01
-1.03722882e+00 -4.00010705e-01 1.92441538e-01 5.33539295e-01
4.19507086e-01 -3.10647674e-02 -3.19681346e-01 7.65938759e-01
-2.14907467e-01 3.34393680e-01 3.82276982e-01 8.78066063e-01
-6.31326497e-01 -1.16533017e+00 1.60665754e-02 6.84060216e-01
-3.60510021e-01 -3.16897541e-01 -2.77277470e-01 8.56161356e-01
-2.68439800e-01 1.07420588e+00 -3.12670946e-01 -4.10752684e-01
3.35145742e-01 3.21601331e-01 1.08464651e-01 -9.15555775e-01
-3.37996721e-01 3.49089615e-02 1.22868866e-01 -1.92411438e-01
-1.89353943e-01 -3.58331084e-01 -1.24004459e+00 -2.14166746e-01
-4.73514527e-01 3.88589799e-01 3.75484258e-01 8.84003639e-01
6.48703158e-01 3.58319372e-01 6.90432072e-01 -6.32866800e-01
-8.39051306e-01 -1.37478614e+00 -4.42360163e-01 5.75064838e-01
-1.05469875e-01 -5.42577207e-01 -3.16978633e-01 3.03047419e-01] | [11.72624683380127, 8.795275688171387] |
540b3440-668e-42d2-b8e2-58b0ceffe2d0 | my-view-is-the-best-view-procedure-learning | 2207.10883 | null | https://arxiv.org/abs/2207.10883v1 | https://arxiv.org/pdf/2207.10883v1.pdf | My View is the Best View: Procedure Learning from Egocentric Videos | Procedure learning involves identifying the key-steps and determining their logical order to perform a task. Existing approaches commonly use third-person videos for learning the procedure, making the manipulated object small in appearance and often occluded by the actor, leading to significant errors. In contrast, we observe that videos obtained from first-person (egocentric) wearable cameras provide an unobstructed and clear view of the action. However, procedure learning from egocentric videos is challenging because (a) the camera view undergoes extreme changes due to the wearer's head motion, and (b) the presence of unrelated frames due to the unconstrained nature of the videos. Due to this, current state-of-the-art methods' assumptions that the actions occur at approximately the same time and are of the same duration, do not hold. Instead, we propose to use the signal provided by the temporal correspondences between key-steps across videos. To this end, we present a novel self-supervised Correspond and Cut (CnC) framework for procedure learning. CnC identifies and utilizes the temporal correspondences between the key-steps across multiple videos to learn the procedure. Our experiments show that CnC outperforms the state-of-the-art on the benchmark ProceL and CrossTask datasets by 5.2% and 6.3%, respectively. Furthermore, for procedure learning using egocentric videos, we propose the EgoProceL dataset consisting of 62 hours of videos captured by 130 subjects performing 16 tasks. The source code and the dataset are available on the project page https://sid2697.github.io/egoprocel/. | ['C. V. Jawahar', 'Chetan Arora', 'Siddhant Bansal'] | 2022-07-22 | null | null | null | null | ['procedure-learning'] | ['computer-vision'] | [ 1.18987702e-01 -3.88632804e-01 -9.16592181e-02 -2.68296510e-01
-5.41090488e-01 -7.34545290e-01 3.75373363e-01 -2.35078946e-01
-3.59048396e-01 4.60148126e-01 2.80676931e-01 2.93482125e-01
-1.60451561e-01 -4.85302545e-02 -1.06667435e+00 -7.15193331e-01
-1.60450190e-01 -2.97445413e-02 -1.04608037e-01 6.24652147e-01
1.42149448e-01 1.17350884e-01 -1.55053174e+00 3.98929596e-01
3.54122639e-01 8.65462244e-01 1.86053932e-01 6.78769648e-01
5.78005135e-01 9.48832095e-01 -3.41461271e-01 -2.85198063e-01
3.07123035e-01 -4.73611265e-01 -6.45079494e-01 4.93490279e-01
8.09603989e-01 -6.65232897e-01 -6.80719435e-01 1.04748678e+00
3.84867877e-01 5.37127316e-01 3.97902519e-01 -1.19779742e+00
-3.11187416e-01 3.14658850e-01 -7.38054335e-01 1.96453422e-01
9.74995136e-01 1.25879258e-01 7.08176494e-01 -9.48677540e-01
7.96152413e-01 1.01484478e+00 5.22023261e-01 7.90899515e-01
-1.12147522e+00 -7.41541028e-01 2.78461397e-01 4.31582808e-01
-1.32578218e+00 -6.79410100e-01 8.43461752e-01 -6.50202751e-01
4.95221198e-01 1.50196865e-01 8.46657991e-01 1.80449510e+00
3.09065670e-01 9.65970457e-01 9.85850453e-01 -1.44474670e-01
2.86943674e-01 -3.52187827e-02 -3.33994888e-02 7.52347469e-01
2.58755535e-01 -1.19590200e-01 -8.93438578e-01 5.61234877e-02
7.90562868e-01 4.30890054e-01 -7.04025984e-01 -7.81836927e-01
-1.59022188e+00 3.21475565e-01 2.45077610e-02 9.26417112e-02
-5.50411344e-01 1.75780073e-01 5.77974021e-01 2.73907065e-01
2.75121957e-01 2.85839051e-01 -3.93295854e-01 -5.58191597e-01
-8.19876313e-01 3.03551644e-01 8.94158125e-01 1.25967431e+00
6.05124712e-01 -1.27326310e-01 -3.26104611e-01 3.73273194e-01
-1.89335957e-01 3.92177850e-01 4.65475142e-01 -1.18239021e+00
5.70153654e-01 2.55314022e-01 2.60818362e-01 -1.00728750e+00
-2.70217538e-01 -1.82725310e-01 -4.77093220e-01 -1.82545841e-01
7.74276733e-01 -1.88766658e-01 -6.07645988e-01 1.71360421e+00
4.15252686e-01 4.07686293e-01 -2.48171404e-01 9.06251848e-01
6.71654701e-01 2.39463776e-01 -2.94859707e-02 -4.40674424e-01
1.65194404e+00 -9.64350998e-01 -1.02946067e+00 -2.83187479e-01
4.95331824e-01 -6.85387075e-01 1.17037570e+00 5.10729969e-01
-1.04025567e+00 -6.90000892e-01 -6.81281328e-01 1.68487087e-01
1.10353073e-02 4.42827165e-01 4.78523821e-01 2.70230740e-01
-5.54102063e-01 5.95962644e-01 -1.00546694e+00 -5.33524871e-01
5.43246686e-01 5.50082773e-02 -7.86074400e-01 -3.41769934e-01
-5.00262976e-01 4.26187694e-01 3.52850139e-01 9.64757651e-02
-1.35456824e+00 -7.73172438e-01 -9.46299314e-01 -7.47307986e-02
9.85974312e-01 -5.96721768e-01 1.24785841e+00 -1.23970783e+00
-1.34982872e+00 1.04097486e+00 -1.92237884e-01 -1.46303937e-01
8.71295214e-01 -8.83642256e-01 -2.93631613e-01 7.05372810e-01
8.12708139e-02 2.79128611e-01 1.11345136e+00 -1.00980330e+00
-5.19089758e-01 -4.83678818e-01 2.17513457e-01 2.06163958e-01
-2.68322796e-01 -2.20828373e-02 -7.97973335e-01 -6.58999681e-01
-1.43372819e-01 -1.33031225e+00 4.87203747e-01 1.71398997e-01
-4.12121654e-01 -9.20163766e-02 8.03894341e-01 -8.74543905e-01
1.20668268e+00 -2.52738023e+00 4.71490681e-01 -2.13060513e-01
2.43150160e-01 4.96459007e-02 2.02607989e-01 4.34258521e-01
-3.03173751e-01 -3.93672198e-01 -4.55099018e-03 -6.09195232e-01
-2.57365525e-01 -6.67627994e-03 -4.25940473e-03 1.06878126e+00
-4.67077315e-01 6.40230596e-01 -1.15828919e+00 -3.64287764e-01
3.87057036e-01 4.39021319e-01 -3.53762031e-01 5.54746032e-01
2.03288168e-01 8.31393719e-01 -4.93340641e-01 6.62704587e-01
3.49259496e-01 -2.22222209e-01 2.29421571e-01 -5.28327286e-01
1.96500242e-01 -1.01118498e-01 -1.27170062e+00 2.19814277e+00
-2.22247660e-01 5.55556715e-01 -3.51364687e-02 -8.97558153e-01
2.63741016e-01 6.30147457e-01 7.01211214e-01 -5.02031624e-01
2.11088836e-01 7.47740036e-03 -1.85419992e-01 -1.07939315e+00
6.42125495e-03 4.18654919e-01 9.49012935e-02 3.42801809e-01
3.16367835e-01 1.49130955e-01 5.25909722e-01 1.65363833e-01
1.25437212e+00 7.87607729e-01 7.71863997e-01 -1.33480296e-01
5.54000914e-01 -4.25398380e-01 7.93590248e-01 6.55364037e-01
-5.04917681e-01 5.03889859e-01 3.88903022e-01 -4.92617726e-01
-6.96259558e-01 -1.07241988e+00 1.90009236e-01 7.75120974e-01
2.64147744e-02 -6.68148518e-01 -9.53818321e-01 -7.52060056e-01
-1.71222627e-01 5.94851434e-01 -9.08535123e-01 -1.15681082e-01
-7.44472861e-01 1.34433001e-01 1.67055994e-01 6.49548590e-01
7.20787048e-01 -1.08331776e+00 -8.74912560e-01 -2.07414012e-02
-5.46976507e-01 -1.59043694e+00 -1.07805490e+00 -4.38348413e-01
-7.91872799e-01 -1.51814556e+00 -7.72125781e-01 -5.14968753e-01
8.69481325e-01 6.65082455e-01 8.93223822e-01 -2.73216814e-01
-2.25909561e-01 1.08574080e+00 -2.96670914e-01 -1.12265818e-01
2.71192342e-01 -3.46898288e-01 2.91182727e-01 5.89277565e-01
4.14923221e-01 -6.84422791e-01 -9.71685290e-01 3.39441180e-01
-4.38035697e-01 1.23020031e-01 4.24098969e-01 5.76848686e-01
5.45772672e-01 1.64270531e-02 7.25062564e-02 -9.20861900e-01
1.59254998e-01 -4.44569886e-01 -3.86622548e-01 1.98804572e-01
-1.65430784e-01 -3.66238743e-01 5.52459002e-01 -7.87364244e-01
-1.14856315e+00 3.15128982e-01 6.46996915e-01 -1.11526453e+00
-4.24045801e-01 -3.30581851e-02 -1.20668903e-01 2.18131885e-01
3.87216985e-01 2.27348834e-01 -1.81209281e-01 -2.86717147e-01
4.39109690e-02 2.15431720e-01 6.29279137e-01 -6.11295998e-01
8.26669455e-01 8.64899218e-01 -1.69038773e-01 -7.43061423e-01
-1.09538245e+00 -8.75548005e-01 -6.95008159e-01 -6.62374556e-01
1.20896590e+00 -1.28951418e+00 -1.18839896e+00 4.98262346e-01
-8.76749575e-01 -3.67874980e-01 -1.21331096e-01 9.49544191e-01
-9.36147749e-01 4.30756450e-01 -4.72196400e-01 -7.05789685e-01
-2.34404221e-01 -1.06800532e+00 1.12527168e+00 2.29542390e-01
-4.41699654e-01 -8.94622624e-01 -2.28685558e-01 6.25261188e-01
-1.63998649e-01 4.32667196e-01 3.19098204e-01 -3.69615287e-01
-4.90987271e-01 -2.13356435e-01 9.21237171e-02 3.20207924e-01
5.64737201e-01 -2.71098614e-01 -1.01579058e+00 -4.78000730e-01
2.59809852e-01 -2.80200213e-01 4.09103811e-01 5.06642640e-01
1.53208947e+00 -2.17908487e-01 -2.45352536e-01 7.42995858e-01
1.20296454e+00 2.44188592e-01 5.16780138e-01 1.98859349e-01
1.01163781e+00 9.19582665e-01 7.94592321e-01 4.53240126e-01
3.10673982e-01 7.23127306e-01 8.75645727e-02 3.28266710e-01
-1.06226295e-01 -3.29465866e-01 6.68175936e-01 5.84837675e-01
-4.25042003e-01 2.75806561e-02 -5.19302845e-01 6.10247433e-01
-2.00729752e+00 -1.40917206e+00 -3.49148959e-02 2.47207594e+00
5.16783834e-01 5.21859415e-02 1.99358270e-01 1.63996387e-02
6.33503795e-01 2.93409437e-01 -5.50244391e-01 1.12948641e-01
3.10196429e-01 -1.66289702e-01 3.98267806e-01 -2.48462334e-02
-1.32326102e+00 5.06554186e-01 4.71242714e+00 4.33227360e-01
-9.05611336e-01 7.37088844e-02 1.47501126e-01 -5.18864214e-01
3.94948959e-01 -8.93775001e-02 -5.19627988e-01 6.57177925e-01
6.82432473e-01 -8.78301114e-02 6.42113090e-01 7.64885962e-01
5.62911272e-01 -2.67629504e-01 -1.64571953e+00 1.47157407e+00
5.64245522e-01 -6.19227707e-01 -3.15534949e-01 -9.57482532e-02
7.34916389e-01 -4.55000997e-01 2.81183468e-03 2.31727719e-01
-1.86712921e-01 -4.60495174e-01 8.83928061e-01 6.14705682e-01
6.95312858e-01 -5.15231073e-01 3.25525284e-01 2.11339414e-01
-1.39065850e+00 -3.35516393e-01 2.11815685e-02 7.88882673e-02
8.16632062e-02 3.75670493e-01 -4.26973671e-01 6.51609719e-01
1.11141348e+00 1.22786391e+00 -5.63455164e-01 8.87886524e-01
-5.25085330e-01 6.40807569e-01 1.65091544e-01 4.91944849e-01
-2.37021577e-02 -3.48133683e-01 7.87851393e-01 9.96393800e-01
2.37715438e-01 6.92092404e-02 2.81226367e-01 5.50625622e-01
-1.46748632e-01 -3.70788425e-02 -8.30084801e-01 -3.35403113e-03
1.93492800e-01 1.12924302e+00 -5.38665175e-01 -1.95484072e-01
-7.84997344e-01 1.28508496e+00 3.45847189e-01 7.79674888e-01
-1.06133842e+00 -4.34607081e-02 5.75752556e-01 -5.44602014e-02
5.24605989e-01 -3.60700399e-01 3.12332988e-01 -1.61224627e+00
5.84639668e-01 -1.14754248e+00 4.67193872e-01 -8.49344790e-01
-9.63637352e-01 2.08921462e-01 2.95624137e-01 -1.68443370e+00
-1.09733373e-01 -3.23064893e-01 -4.64254558e-01 2.98845977e-01
-1.15791345e+00 -1.08574176e+00 -7.75887132e-01 8.51243556e-01
1.12447953e+00 -2.93776672e-02 5.92611372e-01 4.34576273e-01
-7.27324486e-01 3.78558129e-01 -1.77270889e-01 2.31187448e-01
1.20095146e+00 -1.13008690e+00 -6.14383742e-02 9.07539487e-01
2.27363870e-01 8.69580269e-01 7.43756831e-01 -5.67046046e-01
-1.82256114e+00 -9.53360200e-01 5.12194753e-01 -6.71943843e-01
6.06846571e-01 -7.09766626e-01 -7.41617739e-01 1.11157501e+00
1.16034485e-01 2.64884949e-01 6.81697369e-01 -5.65243177e-02
-3.77909392e-01 -3.30325812e-01 -7.60885656e-01 6.71850204e-01
1.33607090e+00 -7.01158524e-01 -5.76257288e-01 5.17282724e-01
3.56347442e-01 -5.25499225e-01 -7.52038240e-01 9.38104764e-02
9.67736244e-01 -8.93998504e-01 9.11563039e-01 -4.66591626e-01
4.76677030e-01 -2.13656262e-01 -1.10200614e-01 -1.00449479e+00
-1.19427063e-01 -7.19326675e-01 -6.09128177e-01 1.24221349e+00
-2.86008716e-01 -3.55586946e-01 4.74192888e-01 4.37307000e-01
1.16798677e-01 -5.06444037e-01 -7.26600766e-01 -8.49753022e-01
-6.74976826e-01 -1.36082724e-01 -4.49026339e-02 9.29730833e-01
-2.77359754e-01 1.01590842e-01 -8.85469198e-01 1.04077294e-01
8.57031167e-01 3.16702314e-02 9.70674753e-01 -9.67132390e-01
-4.15181428e-01 1.69700623e-01 -2.75713056e-01 -9.68997419e-01
4.41338152e-01 -4.18684572e-01 -1.88698769e-02 -1.12558472e+00
4.54212278e-01 2.75481164e-01 -2.45532379e-01 2.72902250e-01
-3.05756897e-01 3.07875662e-03 1.92852572e-01 3.54420274e-01
-8.28859866e-01 4.35549021e-01 1.30064988e+00 -3.48348580e-02
-2.06716999e-01 2.18645811e-01 -2.96478331e-01 8.60351741e-01
4.84238505e-01 -5.48562229e-01 -7.11022139e-01 -5.42779207e-01
2.45378539e-02 7.81999007e-02 5.49640119e-01 -1.17940354e+00
3.64004880e-01 -5.18481508e-02 4.24557954e-01 -3.56240720e-01
5.89758575e-01 -1.08474755e+00 2.55837917e-01 4.15129602e-01
-2.73595989e-01 2.24460319e-01 1.05028965e-01 1.06894708e+00
-2.75594778e-02 -7.38982484e-02 5.25827467e-01 -3.64425063e-01
-7.92663217e-01 3.70804638e-01 -2.21079126e-01 6.14880919e-02
1.43548548e+00 -3.51849496e-01 -1.19022883e-01 -5.28877974e-01
-7.33810902e-01 1.36523873e-01 3.96988243e-01 4.07263786e-01
5.60514450e-01 -1.02630413e+00 -4.73378807e-01 2.48338357e-02
2.67207444e-01 -2.34869912e-01 7.31968403e-01 1.29792321e+00
-2.77139157e-01 1.42793894e-01 -2.41012290e-01 -6.44204915e-01
-1.60602140e+00 7.90778399e-01 2.04004526e-01 1.53348714e-01
-1.02751327e+00 5.22037387e-01 6.76885307e-01 -6.84233904e-02
4.33155566e-01 -1.03409722e-01 -2.30788067e-01 1.37267813e-01
6.95451200e-01 5.04601955e-01 -1.70990095e-01 -5.16908228e-01
-3.64098519e-01 8.04831147e-01 -2.47726813e-01 2.40814030e-01
1.17831588e+00 -1.98703066e-01 1.98266700e-01 7.88319111e-01
1.24280846e+00 1.20585717e-01 -1.59552670e+00 -3.69008511e-01
-2.80458540e-01 -8.81894946e-01 -1.90137520e-01 -3.11382860e-01
-1.02956963e+00 5.33670545e-01 5.36900520e-01 -4.20460105e-01
1.20174086e+00 -2.25528926e-01 5.48869133e-01 5.29538691e-01
5.16762674e-01 -1.18597054e+00 3.87914032e-01 -2.77865566e-02
9.13334012e-01 -1.21105814e+00 4.11846340e-01 -5.90103328e-01
-8.91016901e-01 1.04826880e+00 6.51772261e-01 -1.25890464e-01
4.94361728e-01 -2.34194309e-01 -2.89361060e-01 -3.71909291e-01
-6.52177870e-01 5.72362393e-02 2.77142763e-01 4.05126661e-01
2.47799575e-01 2.09363960e-02 -1.69971794e-01 1.43982098e-01
1.22313932e-01 1.27946928e-01 6.19527400e-01 1.11583507e+00
2.95393467e-01 -5.99390328e-01 -2.15930805e-01 2.89443821e-01
-5.80156267e-01 3.79637599e-01 -4.67479117e-02 9.12124097e-01
-1.70160402e-02 9.06440735e-01 -3.58315073e-02 -4.68014851e-02
6.65905535e-01 2.00321786e-02 7.83673048e-01 -6.34311199e-01
-3.56992811e-01 1.11103885e-01 1.18872583e-01 -1.07270873e+00
-8.56658220e-01 -1.04056513e+00 -7.99468338e-01 -2.46181414e-01
-1.35683138e-02 -1.63434282e-01 3.18307698e-01 8.28792453e-01
2.47118831e-01 7.16282010e-01 6.13252759e-01 -9.80212331e-01
-5.33544064e-01 -9.26856458e-01 -6.59634352e-01 8.69314134e-01
3.88113528e-01 -1.03046143e+00 -5.39946377e-01 6.19163692e-01] | [8.197724342346191, 0.3357035517692566] |
71032492-1eba-4d37-9292-46bd7ec1eac7 | a-generative-flow-for-conditional-sampling | 2307.04102 | null | https://arxiv.org/abs/2307.04102v1 | https://arxiv.org/pdf/2307.04102v1.pdf | A generative flow for conditional sampling via optimal transport | Sampling conditional distributions is a fundamental task for Bayesian inference and density estimation. Generative models, such as normalizing flows and generative adversarial networks, characterize conditional distributions by learning a transport map that pushes forward a simple reference (e.g., a standard Gaussian) to a target distribution. While these approaches successfully describe many non-Gaussian problems, their performance is often limited by parametric bias and the reliability of gradient-based (adversarial) optimizers to learn these transformations. This work proposes a non-parametric generative model that iteratively maps reference samples to the target. The model uses block-triangular transport maps, whose components are shown to characterize conditionals of the target distribution. These maps arise from solving an optimal transport problem with a weighted $L^2$ cost function, thereby extending the data-driven approach in [Trigila and Tabak, 2016] for conditional sampling. The proposed approach is demonstrated on a two dimensional example and on a parameter inference problem involving nonlinear ODEs. | ['Ryan Tsai', 'Giulio Trigila', 'Josef Sajonz', 'Daniel Pocklington', 'Isa Lyubimova', 'Alfin Hou', 'Noam Gal', 'Anupam Bhakta', 'Ricardo Baptista', 'Jason Alfonso'] | 2023-07-09 | null | null | null | null | ['bayesian-inference', 'density-estimation'] | ['methodology', 'methodology'] | [-2.03126483e-02 2.29027912e-01 -4.35695127e-02 -2.08896577e-01
-9.04386878e-01 -6.54030323e-01 9.27965164e-01 -3.30601007e-01
-2.51394987e-01 1.18323159e+00 -6.67490289e-02 -2.49197379e-01
-2.93855935e-01 -9.09165442e-01 -9.73300517e-01 -1.05526006e+00
4.77861892e-03 9.15620744e-01 -1.15929842e-01 1.43019304e-01
3.19915265e-01 6.55291915e-01 -1.02316773e+00 -4.54643011e-01
1.04969692e+00 7.90663898e-01 -1.21103078e-01 8.91161203e-01
-1.16171189e-01 4.33454603e-01 -5.36892414e-01 -5.31748056e-01
1.15592964e-01 -5.17121613e-01 -5.38626730e-01 -1.38617411e-01
2.32466683e-01 -3.54031891e-01 -2.60299176e-01 1.38169169e+00
3.51951629e-01 3.75399321e-01 1.55964828e+00 -1.51693738e+00
-7.91272879e-01 3.97201449e-01 -1.41262010e-01 -1.17141880e-01
-5.42154312e-02 6.08766526e-02 6.19378686e-01 -7.75081038e-01
5.33928931e-01 1.42032421e+00 7.22546101e-01 3.71077478e-01
-1.89656556e+00 -4.84877110e-01 -1.77971467e-01 -1.35820046e-01
-1.60262764e+00 -2.59999484e-01 5.85609257e-01 -9.32796538e-01
2.21386939e-01 7.01547265e-02 5.42535126e-01 1.30717194e+00
2.85628170e-01 5.78054667e-01 1.23415363e+00 -8.29675049e-02
6.44128680e-01 3.95076334e-01 -3.28374714e-01 3.64868641e-01
1.11263022e-01 1.40266910e-01 -1.84037641e-01 -4.24427986e-01
9.58271503e-01 -1.56086609e-01 -2.27746248e-01 -8.35081518e-01
-7.68734574e-01 1.27628493e+00 2.35022426e-01 -4.90333624e-02
-3.84306788e-01 2.41828725e-01 4.25354652e-02 -6.65311590e-02
6.11073017e-01 3.13892871e-01 -9.43650901e-02 -2.02054113e-01
-1.16323316e+00 7.45694876e-01 1.18180680e+00 1.17421055e+00
9.25030649e-01 3.72332960e-01 -5.12134492e-01 5.24912596e-01
6.22331619e-01 1.08141685e+00 -1.98025256e-01 -1.01787174e+00
1.65518329e-01 8.23988952e-03 4.81582344e-01 -6.94533706e-01
4.07124637e-03 -2.58792967e-01 -9.35834169e-01 3.96350116e-01
9.92188871e-01 -5.03592908e-01 -9.56989944e-01 1.91856945e+00
4.05350626e-01 2.27759615e-01 -1.75496131e-01 9.49773133e-01
1.12709470e-01 7.62447298e-01 2.11610585e-01 3.75937037e-02
7.09551573e-01 -2.02428684e-01 -5.66118002e-01 2.25351706e-01
-4.56271751e-04 -6.93156719e-01 1.02413344e+00 2.46577710e-01
-1.25825202e+00 -2.30708376e-01 -7.87199557e-01 3.07111163e-02
-4.07182872e-01 1.72258765e-02 1.97281614e-01 9.30512726e-01
-9.74021077e-01 8.45908761e-01 -1.05270851e+00 -1.67293236e-01
5.59590280e-01 7.43544474e-02 1.20114416e-01 -5.99997118e-02
-1.21557772e+00 8.97970974e-01 1.16922170e-01 1.19879618e-01
-1.55034387e+00 -1.30208647e+00 -7.52613962e-01 2.95061767e-02
2.14952454e-01 -7.90659666e-01 8.66908550e-01 -3.79000038e-01
-1.99004138e+00 3.44084918e-01 -1.00517601e-01 -5.66333652e-01
9.66030717e-01 -2.31757075e-01 4.75233346e-02 -1.48798376e-01
8.54995400e-02 4.61552411e-01 1.39153957e+00 -1.17434645e+00
-7.73238763e-02 -1.65742621e-01 5.02223633e-02 -1.14719719e-01
1.35228008e-01 -3.64173591e-01 -1.72312573e-01 -5.86441398e-01
-2.33859777e-01 -9.73875403e-01 -2.70427585e-01 1.63874015e-01
-9.33758080e-01 1.90821528e-01 8.07014704e-01 -6.50611639e-01
8.26914787e-01 -1.74487114e+00 6.53870404e-01 5.61624467e-01
1.58124743e-03 -6.97185323e-02 2.44087920e-01 4.83457386e-01
2.54720002e-01 8.70307162e-02 -9.22038913e-01 -4.51419473e-01
4.66553688e-01 1.81383789e-01 -4.60464567e-01 9.39471245e-01
4.62548852e-01 8.15424740e-01 -8.76101732e-01 -2.79131263e-01
4.77752715e-01 1.13585341e+00 -6.41288400e-01 3.82574946e-01
-3.92683327e-01 9.04793322e-01 -4.08195883e-01 2.46954948e-01
8.73658538e-01 4.00076732e-02 -1.63822100e-01 -1.46273538e-01
-9.35236216e-02 1.18104396e-02 -1.36834764e+00 1.46893001e+00
-6.25183344e-01 4.37721223e-01 3.25892210e-01 -1.07501817e+00
1.05230749e+00 3.91404852e-02 8.94640163e-02 3.64844166e-02
2.58972049e-01 8.89771655e-02 -1.41986176e-01 -1.57087117e-01
3.76841277e-01 -4.97984827e-01 -1.34548560e-01 3.01061183e-01
2.33992338e-01 -7.91891098e-01 2.07304150e-01 2.02680573e-01
5.80131948e-01 5.92625916e-01 2.71485317e-02 -7.85256922e-01
4.57755655e-01 -1.18233114e-01 2.12526336e-01 7.37405419e-01
1.27075732e-01 7.15534389e-01 9.58149135e-01 1.46662340e-01
-1.41975367e+00 -1.70122492e+00 -6.71197593e-01 5.33044398e-01
-7.46896490e-02 1.91440061e-02 -9.38189566e-01 -3.48566115e-01
3.04836035e-01 1.27236664e+00 -1.00677836e+00 -2.30885878e-01
-2.51922905e-01 -7.49554873e-01 5.86278021e-01 2.59565443e-01
2.26617560e-01 -5.91279984e-01 -2.33470500e-01 2.11661160e-01
1.64968818e-01 -7.30073690e-01 -5.35795450e-01 7.62594715e-02
-6.31075144e-01 -8.59413981e-01 -1.24390078e+00 -7.30963796e-02
8.00334454e-01 -8.54588687e-01 9.99253273e-01 -8.37753296e-01
-2.55160958e-01 4.67229873e-01 2.78715879e-01 -3.14985216e-01
-7.77322352e-01 1.69340596e-01 5.24093397e-02 3.38120341e-01
-8.52168500e-02 -6.95806623e-01 -3.81229401e-01 3.75535518e-01
-8.68690968e-01 -2.46873870e-01 2.01160893e-01 8.02789450e-01
7.43663967e-01 -3.20972472e-01 6.16606355e-01 -8.96851063e-01
8.33481431e-01 -8.46344769e-01 -1.03596044e+00 1.93636771e-02
-4.31628883e-01 5.41624665e-01 7.37324417e-01 -6.83793187e-01
-1.15905929e+00 -2.46026576e-01 -9.33745317e-03 -7.34426856e-01
-1.82931647e-01 2.94031143e-01 -2.10089236e-01 1.44772707e-02
6.63581669e-01 1.74229935e-01 1.95129469e-01 -4.21936929e-01
6.56147480e-01 2.11359084e-01 6.59457445e-01 -1.09987032e+00
1.08178306e+00 5.03689647e-01 5.70964456e-01 -6.90940499e-01
-7.31906950e-01 8.69750082e-02 -6.73516691e-01 -1.93052351e-01
1.03656864e+00 -4.38333213e-01 -7.89928734e-01 3.22674781e-01
-9.22134995e-01 -6.46623015e-01 -6.85537577e-01 5.03972530e-01
-9.96724069e-01 -1.47343785e-01 -4.37790573e-01 -1.18830121e+00
1.23097366e-02 -1.16773021e+00 1.04139316e+00 2.85246193e-01
-1.50989994e-01 -1.48149443e+00 3.49474758e-01 -4.24260139e-01
8.55638981e-01 5.24492681e-01 7.68421531e-01 -3.62098783e-01
-6.36498511e-01 -6.77005947e-02 -7.34741017e-02 4.90218818e-01
2.58826111e-02 2.82031506e-01 -8.33999336e-01 -1.05341308e-01
-7.03443438e-02 -2.09111035e-01 4.99610633e-01 8.96662056e-01
9.80276227e-01 -3.94593626e-01 -3.14309269e-01 9.26547647e-01
1.27300560e+00 5.03328778e-02 6.81917429e-01 -1.60894677e-01
6.99303508e-01 3.92201900e-01 2.23575786e-01 4.01396930e-01
1.04412846e-01 4.01668131e-01 3.11313510e-01 2.52848685e-01
9.60141197e-02 -5.44238210e-01 1.13358371e-01 2.74981558e-01
-8.34953487e-02 -2.04387456e-02 -7.49188960e-01 5.44124067e-01
-1.57035756e+00 -9.18745160e-01 -3.09222415e-02 2.46156907e+00
9.08701420e-01 -6.34138510e-02 1.05299957e-01 -2.07297608e-01
8.10107589e-01 -1.78275928e-01 -8.28948438e-01 -2.90451884e-01
4.28571068e-02 4.00608182e-01 5.93057752e-01 8.83542120e-01
-1.02987540e+00 6.09518409e-01 6.76149511e+00 9.37410891e-01
-9.80609596e-01 3.25599238e-02 5.42347074e-01 7.99409822e-02
-4.64630127e-01 -3.80146280e-02 -8.39343131e-01 7.23109841e-01
1.05020714e+00 -5.31466544e-01 7.44836390e-01 8.04447353e-01
1.66717976e-01 -1.47240147e-01 -1.08743942e+00 6.35452867e-01
-8.77195224e-02 -1.33073127e+00 -7.44346380e-02 1.17529400e-01
9.14038062e-01 -2.68831581e-01 3.66482049e-01 2.44396150e-01
7.86217093e-01 -1.39502335e+00 7.03286052e-01 1.11228883e+00
8.46516132e-01 -9.14146185e-01 4.15585518e-01 2.92838216e-01
-7.36364782e-01 4.84814584e-01 -2.62443066e-01 3.34334105e-01
6.69387341e-01 6.25182569e-01 -8.70940626e-01 2.66265124e-01
5.07707000e-01 4.50835139e-01 7.30235353e-02 9.41470802e-01
-2.63632834e-01 8.71117592e-01 -5.56962013e-01 -1.24318823e-01
9.20470282e-02 -9.82475221e-01 9.58711207e-01 1.19847703e+00
5.78561664e-01 -3.96926790e-01 -1.64207861e-01 1.81704557e+00
1.98530816e-02 -1.03235438e-01 -6.25524402e-01 -6.38316944e-02
2.86322922e-01 1.11913323e+00 -5.69749534e-01 -2.12233767e-01
4.06525694e-02 5.69521606e-01 1.60767093e-01 8.16105247e-01
-1.03994226e+00 -4.41095918e-01 8.78880501e-01 1.87907204e-01
4.79855925e-01 -2.77301162e-01 -2.52215564e-01 -8.89785230e-01
-3.80051881e-01 -2.80874163e-01 -2.99730077e-02 -6.22486174e-01
-1.47296989e+00 2.64953107e-01 5.99309623e-01 -9.08826590e-01
-5.40447176e-01 -5.15402913e-01 -6.24188364e-01 1.44317353e+00
-1.13818467e+00 -8.61364782e-01 8.75165090e-02 5.96480072e-01
1.30386427e-02 -9.87515226e-02 6.82056010e-01 2.79771119e-01
-3.80294800e-01 3.95264685e-01 4.33450639e-01 -2.44806372e-02
4.55651820e-01 -1.54140341e+00 3.04713517e-01 7.34079838e-01
-3.28045040e-01 7.52871275e-01 1.15488648e+00 -6.82731867e-01
-1.33840525e+00 -1.11844385e+00 2.61654913e-01 -4.04857546e-01
8.91307414e-01 -5.00447571e-01 -9.90072072e-01 7.97193885e-01
-8.31050500e-02 4.80106287e-02 2.38325059e-01 -2.17276171e-01
-1.92945719e-01 1.66353315e-01 -1.57920325e+00 5.90566874e-01
5.72953343e-01 -4.15443897e-01 -1.38982296e-01 1.77385539e-01
2.89044976e-01 -5.91677368e-01 -1.15537107e+00 -2.13083401e-02
4.35367495e-01 -7.48242617e-01 1.11822891e+00 -7.45606720e-01
3.63328069e-01 -4.11952317e-01 -1.69577152e-01 -1.74306977e+00
1.94364727e-01 -9.48579013e-01 -2.30879083e-01 1.17200375e+00
3.24427217e-01 -7.90666878e-01 6.23775303e-01 7.11752295e-01
2.08405536e-02 -5.62313557e-01 -1.08918345e+00 -7.73578525e-01
6.98927522e-01 -6.10431492e-01 6.18198693e-01 5.92085660e-01
-5.52337348e-01 2.38767471e-02 -4.62493151e-01 2.27440611e-01
1.19194484e+00 -2.76246846e-01 9.41904604e-01 -1.05015457e+00
-3.33763510e-01 -4.65916842e-01 -3.34198147e-01 -1.06252778e+00
3.13827336e-01 -9.90619242e-01 2.27760985e-01 -1.26320839e+00
-7.15168491e-02 -6.01673186e-01 3.81873608e-01 -1.80145055e-01
2.82831099e-02 6.22847304e-02 -3.49176861e-02 -1.60985991e-01
4.88312952e-02 9.27509010e-01 1.43169057e+00 -1.13928311e-01
-1.38510153e-01 2.99871683e-01 -3.24469239e-01 6.21461093e-01
7.02684343e-01 -5.44346511e-01 -6.02989972e-01 1.01072334e-01
1.15681641e-01 2.23898724e-01 5.82940757e-01 -6.80478752e-01
2.40874171e-01 -2.81761765e-01 3.06966811e-01 -3.71142685e-01
5.36938190e-01 -5.27334511e-01 4.16183561e-01 7.80545101e-02
-3.40811610e-01 -3.86816651e-01 2.02499196e-01 5.73371351e-01
5.83096361e-03 -4.37166005e-01 1.04419756e+00 2.56310016e-01
9.25935805e-02 4.40728158e-01 -3.19043368e-01 6.28188729e-01
8.49138796e-01 1.90857589e-01 -1.41418397e-01 -6.92898095e-01
-1.08112955e+00 -8.51058587e-02 2.58589417e-01 -1.97625667e-01
2.76910305e-01 -1.36221504e+00 -7.61017084e-01 2.21460670e-01
-3.50899011e-01 3.31892073e-01 1.44421071e-01 1.00887835e+00
-4.36612844e-01 1.83306143e-01 1.07040638e-02 -8.38988662e-01
-3.02198112e-01 2.55510509e-01 6.90024555e-01 -1.25709295e-01
-4.09046352e-01 7.19713390e-01 2.30093479e-01 -6.15209520e-01
4.37880345e-02 -3.98004889e-01 2.74878979e-01 -1.80596381e-01
2.90714979e-01 7.39702463e-01 -3.63225132e-01 -4.79564846e-01
-4.98035252e-02 4.26640302e-01 3.30115795e-01 -5.04690528e-01
1.06684303e+00 -3.16319987e-02 4.96255457e-02 6.29912376e-01
1.31646931e+00 -1.42520115e-01 -1.84557903e+00 4.61080298e-02
-4.27520603e-01 -4.34981465e-01 -1.03423126e-01 -5.62300503e-01
-1.01558363e+00 9.51541007e-01 2.76251405e-01 3.88886720e-01
4.58109885e-01 2.39516627e-02 1.95590436e-01 4.38460335e-02
1.32374614e-01 -7.24559426e-01 -3.09302449e-01 3.85733157e-01
1.03130674e+00 -8.49867463e-01 -1.84623927e-01 -1.92634180e-01
-4.98800486e-01 9.89759505e-01 2.09258914e-01 -5.04165292e-01
1.07245934e+00 5.58004260e-01 -2.85943151e-01 1.55026168e-01
-1.51627809e-01 -2.61069573e-02 4.01824951e-01 9.24594283e-01
1.52081192e-01 1.62734449e-01 -5.99805685e-03 2.07297012e-01
-3.83811414e-01 6.44271150e-02 4.56836969e-01 4.03956294e-01
6.51330575e-02 -8.65837038e-01 -5.48946142e-01 5.00597060e-01
-3.59028518e-01 -8.93067643e-02 3.73263478e-01 9.57082987e-01
-2.58259743e-01 5.66626132e-01 3.15123618e-01 3.14584255e-01
2.47621000e-01 2.37496197e-01 4.44621861e-01 -1.88755959e-01
1.36663010e-02 5.73271438e-02 -2.46679291e-01 -3.05606097e-01
-2.56377786e-01 -1.02349544e+00 -8.98339808e-01 -5.27734101e-01
1.16424486e-01 2.21474499e-01 9.11247075e-01 8.31731558e-01
-1.69786029e-02 5.75132847e-01 1.88159987e-01 -1.15091240e+00
-8.21697712e-01 -1.16315770e+00 -8.42539310e-01 2.73453593e-01
3.46135080e-01 -9.93313134e-01 -6.89251661e-01 1.25207976e-01] | [6.866064071655273, 3.823768377304077] |
944e2504-c8d7-49b3-9254-8b77a2a7fc91 | no-shifted-augmentations-nsa-compact | 2203.10344 | null | https://arxiv.org/abs/2203.10344v1 | https://arxiv.org/pdf/2203.10344v1.pdf | No Shifted Augmentations (NSA): compact distributions for robust self-supervised Anomaly Detection | Unsupervised Anomaly detection (AD) requires building a notion of normalcy, distinguishing in-distribution (ID) and out-of-distribution (OOD) data, using only available ID samples. Recently, large gains were made on this task for the domain of natural images using self-supervised contrastive feature learning as a first step followed by kNN or traditional one-class classifiers for feature scoring. Learned representations that are non-uniformly distributed on the unit hypersphere have been shown to be beneficial for this task. We go a step further and investigate how the \emph {geometrical compactness} of the ID feature distribution makes isolating and detecting outliers easier, especially in the realistic situation when ID training data is polluted (i.e. ID data contains some OOD data that is used for learning the feature extractor parameters). We propose novel architectural modifications to the self-supervised feature learning step, that enable such compact distributions for ID data to be learned. We show that the proposed modifications can be effectively applied to most existing self-supervised objectives, with large gains in performance. Furthermore, this improved OOD performance is obtained without resorting to tricks such as using strongly augmented ID images (e.g. by 90 degree rotations) as proxies for the unseen OOD data, as these impose overly prescriptive assumptions about ID data and its invariances. We perform extensive studies on benchmark datasets for one-class OOD detection and show state-of-the-art performance in the presence of pollution in the ID data, and comparable performance otherwise. We also propose and extensively evaluate a novel feature scoring technique based on the angular Mahalanobis distance, and propose a simple and novel technique for feature ensembling during evaluation that enables a big boost in performance at nearly zero run-time cost. | ['Tom Bishop', 'Unmesh Kurup', 'Marcel Ackermann', 'Mohamed Yousef'] | 2022-03-19 | null | null | null | null | ['self-supervised-anomaly-detection', 'supervised-anomaly-detection'] | ['computer-vision', 'computer-vision'] | [ 1.09923251e-01 1.70542687e-01 2.10819334e-01 -3.18434030e-01
-5.13622999e-01 -5.32833099e-01 8.43198240e-01 2.67016202e-01
-4.01575655e-01 4.07507300e-01 -1.47294328e-01 -1.04635477e-01
-3.89215350e-01 -5.89074969e-01 -6.00973666e-01 -1.07739079e+00
-4.79274571e-01 5.89002132e-01 2.50995070e-01 -7.89328292e-02
3.26692224e-01 8.86819124e-01 -2.08218741e+00 -1.78146735e-01
1.07791626e+00 1.23416567e+00 -1.57174811e-01 7.10514605e-01
-1.12903759e-01 3.90189946e-01 -7.68417299e-01 -1.93816591e-02
6.69550240e-01 -3.40744346e-01 -5.31111777e-01 2.58823097e-01
7.47072220e-01 -2.35870868e-01 -6.33918121e-02 1.02401221e+00
5.13695538e-01 2.68348038e-01 1.16124368e+00 -1.38944387e+00
-3.93780470e-01 -2.00503066e-01 -4.72055316e-01 2.95234710e-01
1.81813449e-01 1.31396085e-01 8.48227024e-01 -9.12344694e-01
4.77678180e-01 9.89918113e-01 5.67693472e-01 2.75239170e-01
-1.37714863e+00 -3.64823341e-01 -7.94398636e-02 1.53696820e-01
-1.22932696e+00 -3.89641017e-01 8.32430482e-01 -4.74170536e-01
1.00445926e+00 3.49203527e-01 2.08018482e-01 1.08820581e+00
1.53383315e-01 5.88283241e-01 9.48203444e-01 -6.45367801e-01
4.54786718e-01 3.65159661e-01 9.65630561e-02 6.40559137e-01
5.36324918e-01 9.64046642e-02 -2.74118006e-01 -4.02065217e-01
3.77412349e-01 -6.65971357e-03 -2.58702606e-01 -9.04670537e-01
-8.58808398e-01 9.03694510e-01 2.62492448e-01 1.86702639e-01
-1.41863182e-01 -4.04340535e-01 3.81999284e-01 6.06228590e-01
6.97856784e-01 6.53662324e-01 -3.55807155e-01 -6.32238016e-02
-7.82083452e-01 1.71377994e-02 8.54357958e-01 7.65249193e-01
8.79178643e-01 1.87657773e-01 6.57921433e-02 8.08498085e-01
1.45618632e-01 6.58495963e-01 6.24154866e-01 -6.46885872e-01
2.20178738e-01 6.49854541e-01 8.71176720e-02 -1.15571058e+00
-6.46069884e-01 -4.89204943e-01 -9.35149252e-01 5.41467905e-01
6.20519161e-01 7.30971321e-02 -1.02738297e+00 1.40111053e+00
5.40237069e-01 1.78775191e-02 1.18129097e-01 6.30050838e-01
3.30834091e-01 3.02962959e-01 -5.85665405e-01 -1.50581926e-01
9.77780044e-01 -5.25031924e-01 -3.35070968e-01 2.05993317e-02
8.58619809e-01 -5.45808613e-01 1.11946154e+00 7.18326569e-01
-4.68800694e-01 -3.27432573e-01 -1.17153895e+00 3.10861021e-01
-5.73454022e-01 -7.34699667e-02 3.94987345e-01 7.87820339e-01
-7.04628050e-01 8.66323948e-01 -7.79512644e-01 -4.70519274e-01
5.13548791e-01 3.97559732e-01 -6.14429235e-01 7.63783557e-03
-6.80000067e-01 7.75247216e-01 2.79959351e-01 -5.25524057e-02
-7.22037256e-01 -5.70721626e-01 -9.75254059e-01 -1.69597059e-01
3.64120483e-01 -1.90516651e-01 6.29817545e-01 -1.01970208e+00
-1.32367957e+00 8.78423810e-01 8.00145864e-02 -4.71641123e-01
6.57120645e-01 -2.85730094e-01 -5.08710504e-01 3.29795837e-01
1.72502194e-02 2.14405596e-01 1.48746741e+00 -1.30732870e+00
-3.95765871e-01 -6.14263296e-01 -2.97176659e-01 -3.99270914e-02
-6.83131337e-01 -1.41360536e-01 -3.58794965e-02 -7.03031421e-01
3.53177518e-01 -1.01303589e+00 4.48843390e-02 1.88551303e-02
-5.02351463e-01 -1.36839405e-01 1.08882129e+00 -2.73911923e-01
9.23563778e-01 -2.25548911e+00 -2.57263452e-01 7.10888207e-01
2.46625036e-01 3.61215800e-01 -3.89291421e-02 1.94128960e-01
-3.40427488e-01 -7.19197839e-02 -4.97860670e-01 -2.89190531e-01
9.86649543e-02 3.21513832e-01 -3.23313773e-01 9.44947720e-01
5.24759173e-01 2.72727102e-01 -7.64522552e-01 -1.71906576e-01
2.65719026e-01 2.77109504e-01 -5.73160946e-01 4.43454057e-01
1.62557185e-01 4.46465671e-01 -1.97533816e-01 5.99587619e-01
7.34403431e-01 2.22598761e-02 -4.81859386e-01 8.11234023e-03
-7.24618807e-02 -1.06538786e-02 -1.48883653e+00 1.19929063e+00
-3.46732169e-01 5.39461315e-01 -1.62902743e-01 -1.32634103e+00
1.13670838e+00 -1.05118547e-02 5.10844469e-01 -6.25955641e-01
4.98329028e-02 3.51669550e-01 8.80976990e-02 -4.88585830e-01
1.56353265e-01 -1.82967260e-02 1.57224566e-01 3.60505730e-01
3.12570304e-01 -1.03729896e-01 1.80708498e-01 -3.15469094e-02
1.21035445e+00 3.00576519e-02 4.48394239e-01 -6.63197100e-01
5.97608387e-01 -1.02690473e-01 3.34722251e-01 8.49658728e-01
-9.33898240e-02 8.92258465e-01 6.87884748e-01 -3.91324997e-01
-9.63778734e-01 -1.17654955e+00 -6.41301453e-01 8.63118470e-01
2.70276982e-02 -1.83609337e-01 -5.90637565e-01 -1.06086195e+00
1.70546040e-01 7.64115810e-01 -6.63808286e-01 -4.48413044e-01
-4.19104189e-01 -1.13231552e+00 5.04638314e-01 3.43357831e-01
3.14662069e-01 -7.80496240e-01 -5.99586189e-01 -5.47342002e-02
3.81443292e-01 -9.37360585e-01 -1.33765042e-02 7.82249272e-01
-9.50015724e-01 -1.12525678e+00 -5.77601850e-01 -4.99680817e-01
9.59791660e-01 1.11356102e-01 9.26240683e-01 -8.35853294e-02
-6.02685332e-01 6.33062959e-01 -4.20804411e-01 -5.05829453e-01
-2.96056271e-01 -1.94750533e-01 5.93272030e-01 2.53765196e-01
4.14358348e-01 -7.07554877e-01 -4.52109039e-01 5.08119822e-01
-1.06203687e+00 -7.08069503e-01 4.96812254e-01 1.06382155e+00
4.01987135e-01 2.04462305e-01 5.19510329e-01 -8.07533622e-01
2.16055796e-01 -4.86055553e-01 -6.34371996e-01 -2.21076071e-01
-8.48178029e-01 4.68564063e-01 9.82191563e-01 -2.86663145e-01
-7.88168490e-01 -4.31117378e-02 -8.19761008e-02 -5.01156867e-01
-6.41071677e-01 2.58429837e-03 -3.28890771e-01 -3.17661285e-01
1.01709425e+00 1.97702348e-02 3.60929698e-01 -5.92727840e-01
1.24939695e-01 7.70928025e-01 5.11135519e-01 -6.01819456e-01
1.25106812e+00 7.33662903e-01 2.77301192e-01 -1.32515168e+00
-7.20335305e-01 -7.68484473e-01 -8.75920177e-01 2.10380986e-01
5.39209008e-01 -6.09695196e-01 -4.09439176e-01 4.95717198e-01
-5.79304457e-01 -8.99106860e-02 -5.87292433e-01 4.83566791e-01
-5.67810476e-01 7.21218348e-01 -4.35742438e-02 -7.69669771e-01
-1.00720607e-01 -8.10972452e-01 9.43069220e-01 -5.01912870e-02
-2.52748057e-02 -1.06903648e+00 1.51225075e-01 -2.49862507e-01
3.53790283e-01 3.98848683e-01 9.01047170e-01 -1.36047482e+00
-1.46331370e-01 -3.63555402e-01 -1.82413861e-01 8.87047768e-01
2.95097589e-01 -5.48251271e-02 -1.28611577e+00 -4.96960342e-01
2.57563889e-01 -1.98348075e-01 9.43007171e-01 2.36319341e-02
1.13257790e+00 -2.99104840e-01 -5.88451587e-02 8.10844898e-01
1.27258718e+00 -7.04485178e-02 5.29186189e-01 5.81519902e-01
7.08312154e-01 6.60972416e-01 7.05449820e-01 6.45772159e-01
-1.23212509e-01 6.54748619e-01 6.17284656e-01 -1.93133980e-01
4.07860503e-02 1.91809624e-01 4.74775553e-01 4.65980440e-01
-7.68251135e-04 -1.33655682e-01 -9.32600439e-01 4.91202384e-01
-1.61303461e+00 -7.57930219e-01 -1.20913647e-01 2.68726182e+00
4.20449466e-01 3.17019343e-01 3.08231115e-01 4.28886801e-01
4.24668849e-01 -2.24197209e-02 -5.95609844e-01 -3.28018606e-01
-1.85135007e-01 4.12107527e-01 5.47515631e-01 2.62047023e-01
-1.53458321e+00 3.96072417e-01 5.47998762e+00 6.81967914e-01
-1.15855432e+00 -2.60248959e-01 3.60619724e-01 6.33544847e-02
2.63158195e-02 -1.79576725e-01 -6.65407240e-01 4.02372688e-01
8.41916084e-01 2.20971808e-01 1.90869614e-01 1.06379175e+00
-2.66647134e-02 -2.29852334e-01 -1.12860048e+00 1.02902472e+00
3.83840084e-01 -6.48668587e-01 -3.85709256e-02 2.87679940e-01
6.09359145e-01 1.04340680e-01 3.86345387e-02 1.70450836e-01
-8.33478570e-02 -8.46982539e-01 3.62199157e-01 3.64520133e-01
6.42913043e-01 -8.45497727e-01 8.69365573e-01 3.00887823e-01
-7.67957151e-01 -2.01206401e-01 -6.47800863e-01 3.58755216e-02
-4.66844231e-01 9.41451967e-01 -8.26830506e-01 7.05697715e-01
8.35515916e-01 7.01347649e-01 -9.88322794e-01 1.06982660e+00
2.39134282e-02 6.10314608e-01 -7.66578376e-01 3.17383468e-01
2.15494454e-01 -3.52884889e-01 9.67727423e-01 1.14266539e+00
4.30383652e-01 -4.92318958e-01 1.00382872e-01 6.41886652e-01
2.43281946e-01 2.56421655e-01 -9.87369061e-01 3.30651313e-01
-5.49844950e-02 1.25336957e+00 -7.97301769e-01 -1.12900950e-01
-4.10462350e-01 1.00208271e+00 1.48741692e-01 2.78749764e-01
-6.60863221e-01 -7.83546507e-01 7.20657408e-01 2.42207736e-01
5.97526908e-01 -1.21696867e-01 -4.84156311e-02 -1.24732399e+00
2.80400872e-01 -7.80075490e-01 5.57768524e-01 -1.60096809e-01
-1.68001068e+00 5.01799107e-01 1.09956056e-01 -1.70039010e+00
-4.08171594e-01 -1.00400567e+00 -6.89687490e-01 4.96353179e-01
-1.56616426e+00 -6.09551728e-01 -4.55186784e-01 6.83788776e-01
5.80277070e-02 -2.46479705e-01 8.68768156e-01 1.11953728e-01
-5.32176495e-01 7.96374559e-01 6.53858483e-01 1.34519860e-01
1.05019498e+00 -1.72044301e+00 1.24269769e-01 9.21057463e-01
3.09280008e-01 4.13052797e-01 7.63370037e-01 -3.84010524e-01
-1.20577788e+00 -1.14689052e+00 2.88870662e-01 -5.72502136e-01
6.22704268e-01 -6.54229164e-01 -1.04326320e+00 3.77450734e-01
-2.84572780e-01 5.37637115e-01 5.58001816e-01 5.77971861e-02
-4.62180346e-01 -2.81307191e-01 -1.40074468e+00 3.21761280e-01
1.00424135e+00 -2.57767379e-01 -7.38689482e-01 3.43014598e-01
2.49042407e-01 -9.10694003e-02 -7.40927517e-01 5.50192773e-01
2.23593369e-01 -1.24221826e+00 9.75824237e-01 -5.55355489e-01
-1.44281024e-02 -5.32877743e-01 -1.90897420e-01 -1.36751246e+00
-5.14366385e-03 -6.07903242e-01 -2.17863619e-01 1.17381692e+00
1.61289081e-01 -1.13629818e+00 6.27116144e-01 2.27480948e-01
-2.98644304e-01 -5.96158028e-01 -1.04292226e+00 -1.33796012e+00
-8.42680037e-02 -4.42716926e-01 1.49375170e-01 9.08091724e-01
-7.92232901e-02 8.03175718e-02 -1.36208281e-01 4.53656256e-01
7.69410908e-01 -5.79027012e-02 1.00078619e+00 -1.63715506e+00
-2.27291688e-01 -2.54257292e-01 -1.03922033e+00 -6.85128629e-01
8.02619681e-02 -9.69606638e-01 1.41382292e-02 -9.20561552e-01
-2.30577976e-01 -4.69028443e-01 -2.78181225e-01 4.32417154e-01
-8.89239684e-02 3.55194181e-01 -4.80348244e-02 3.12631786e-01
-5.08025765e-01 6.82314694e-01 7.08192229e-01 3.41358706e-02
-2.47387514e-01 2.38104798e-02 -4.28026378e-01 1.00207758e+00
6.18992269e-01 -6.32209599e-01 -2.87105441e-01 2.02811316e-01
-5.31646647e-02 -7.53030658e-01 4.76829976e-01 -1.30634081e+00
-8.98539647e-02 2.46638849e-01 5.92174947e-01 -3.15701604e-01
8.45367163e-02 -1.00510609e+00 -5.05042017e-01 2.09691703e-01
8.41491595e-02 -7.82522634e-02 1.89072579e-01 6.86327517e-01
-2.14951932e-01 -5.04520118e-01 9.15721595e-01 2.19579935e-01
-4.64491427e-01 8.70524645e-02 -1.59089431e-01 3.25725704e-01
1.09867299e+00 -4.52094167e-01 -1.83682010e-01 -2.73314267e-01
-6.16815448e-01 -7.09100291e-02 6.50971949e-01 2.37268910e-01
4.11890537e-01 -1.09984851e+00 -5.73847830e-01 7.75995612e-01
4.97766286e-01 1.37144282e-01 -2.29107048e-02 1.11216605e+00
-4.28553015e-01 1.67417482e-01 -1.82371840e-01 -9.85612452e-01
-1.03657794e+00 6.28147840e-01 2.43072286e-01 -1.32869661e-01
-8.78922105e-01 5.08763731e-01 2.95602858e-01 -5.55640638e-01
2.37509966e-01 -3.61445606e-01 -8.45154598e-02 2.82516181e-01
5.27216852e-01 4.73012626e-01 5.96656322e-01 -5.24406075e-01
-4.11037385e-01 6.18240952e-01 -1.13878295e-01 3.34554732e-01
1.38853753e+00 1.98752969e-03 3.86324823e-02 5.81706285e-01
1.42489839e+00 2.43409082e-01 -1.25253057e+00 -9.12990049e-02
2.63312340e-01 -5.73478222e-01 -5.25199436e-02 -4.38220322e-01
-7.18144178e-01 7.68932581e-01 8.29564154e-01 5.20464003e-01
1.16120982e+00 5.41222282e-02 2.86583126e-01 7.05776930e-01
2.12280184e-01 -1.18250048e+00 2.79813766e-01 5.53770065e-01
7.83919394e-01 -1.46670735e+00 6.05206303e-02 -1.79377452e-01
-5.07088661e-01 1.35233724e+00 4.76072580e-01 -5.28980970e-01
6.61797345e-01 8.71438384e-02 3.76967750e-02 -2.47617334e-01
-1.88208804e-01 -3.72688711e-01 5.42895198e-01 8.30074728e-01
-1.91647410e-02 -2.45123252e-01 1.63623750e-01 2.43728131e-01
-1.77620977e-01 -7.40708590e-01 5.16753972e-01 8.73465061e-01
-5.19617438e-01 -8.71631205e-01 -5.91780722e-01 8.76827538e-01
-2.75853276e-01 9.09470543e-02 -2.75144845e-01 1.03300118e+00
4.57038917e-02 6.70087159e-01 2.80969709e-01 -2.00923413e-01
3.64444554e-01 3.64049137e-01 2.55277127e-01 -5.52593231e-01
-1.84927523e-01 -7.42481425e-02 -9.43378583e-02 -7.21987724e-01
-2.43447721e-01 -7.28445113e-01 -1.21357203e+00 1.34498462e-01
-5.63687086e-01 9.24204662e-02 5.64385653e-01 9.68039930e-01
4.68580037e-01 2.25560501e-01 1.01532745e+00 -9.78216112e-01
-8.58124971e-01 -9.15733874e-01 -7.78738558e-01 6.61725760e-01
6.56967640e-01 -1.05731046e+00 -1.12197030e+00 -2.29221463e-01] | [7.701941013336182, 2.4318180084228516] |
ed068b42-8c15-4f37-97e0-4050630497cf | auto-lambda-disentangling-dynamic-task | 2202.03091 | null | https://arxiv.org/abs/2202.03091v2 | https://arxiv.org/pdf/2202.03091v2.pdf | Auto-Lambda: Disentangling Dynamic Task Relationships | Understanding the structure of multiple related tasks allows for multi-task learning to improve the generalisation ability of one or all of them. However, it usually requires training each pairwise combination of tasks together in order to capture task relationships, at an extremely high computational cost. In this work, we learn task relationships via an automated weighting framework, named Auto-Lambda. Unlike previous methods where task relationships are assumed to be fixed, Auto-Lambda is a gradient-based meta learning framework which explores continuous, dynamic task relationships via task-specific weightings, and can optimise any choice of combination of tasks through the formulation of a meta-loss; where the validation loss automatically influences task weightings throughout training. We apply the proposed framework to both multi-task and auxiliary learning problems in computer vision and robotics, and show that Auto-Lambda achieves state-of-the-art performance, even when compared to optimisation strategies designed specifically for each problem and data domain. Finally, we observe that Auto-Lambda can discover interesting learning behaviors, leading to new insights in multi-task learning. Code is available at https://github.com/lorenmt/auto-lambda. | ['Edward Johns', 'Andrew J. Davison', 'Stephen James', 'Shikun Liu'] | 2022-02-07 | null | null | null | null | ['auxiliary-learning'] | ['methodology'] | [ 3.83700043e-01 -2.23872252e-02 -6.26100823e-02 -3.40654522e-01
-9.32158411e-01 -5.10438502e-01 7.28859544e-01 3.24809790e-01
-6.66584790e-01 6.57355666e-01 -1.73115373e-01 3.07311229e-02
-7.91413128e-01 -2.26988599e-01 -8.07424963e-01 -8.18999290e-01
-9.58235711e-02 7.41444051e-01 1.81015372e-01 -1.47496685e-01
2.91207135e-01 -7.39654154e-02 -1.54417574e+00 3.02716702e-01
1.05397856e+00 9.12552238e-01 7.05335855e-01 6.52561784e-01
1.02963686e-01 5.53444624e-01 -3.79335493e-01 -5.77312052e-01
2.49651521e-01 6.93673175e-03 -1.04843104e+00 1.06404051e-01
3.52218628e-01 5.89301407e-01 1.66898817e-01 9.37199712e-01
3.96175981e-01 4.27808255e-01 9.11138177e-01 -1.52119613e+00
-2.66861320e-01 2.59256631e-01 -4.66663182e-01 1.09382018e-01
-6.64135888e-02 3.67901800e-03 1.31635988e+00 -7.87233889e-01
1.96362182e-01 1.12602937e+00 5.84777951e-01 3.65581602e-01
-1.51966918e+00 -5.38208544e-01 4.19197202e-01 2.84634680e-01
-1.02383566e+00 -3.43150347e-01 8.15367699e-01 -5.89174747e-01
7.58375049e-01 -2.86369622e-02 2.81114519e-01 1.31819522e+00
4.06429946e-01 1.09019589e+00 1.19246817e+00 -4.10291016e-01
-1.68447029e-02 1.45572662e-01 3.05168331e-01 6.17184281e-01
9.00264923e-03 -7.03142807e-02 -5.79239547e-01 -3.06114078e-01
3.26853007e-01 9.78007019e-02 -7.91666061e-02 -7.65536308e-01
-1.46036112e+00 6.95947111e-01 2.53475279e-01 2.18054011e-01
-2.99393475e-01 2.09057033e-01 6.41971052e-01 7.13859618e-01
7.59427369e-01 7.18979776e-01 -7.87135839e-01 -7.61228800e-02
-4.45804060e-01 3.35203022e-01 6.16179883e-01 6.52317464e-01
1.07613873e+00 -3.15989792e-01 -3.61649394e-01 1.24416602e+00
1.09601960e-01 4.24315520e-02 5.56932986e-01 -8.99009883e-01
7.26397455e-01 5.16479433e-01 -3.32824811e-02 -6.04655147e-01
-6.80918276e-01 -7.70332515e-01 -8.13753724e-01 1.65691808e-01
4.77033228e-01 -2.67991841e-01 -6.57646894e-01 1.89077806e+00
3.45481038e-01 1.35994017e-01 3.59091051e-02 7.02681839e-01
3.30095649e-01 2.90389836e-01 2.19536334e-01 -1.52245328e-01
1.29187799e+00 -1.33789802e+00 -5.22389770e-01 -6.22464299e-01
7.05760181e-01 -7.60978103e-01 1.32903087e+00 6.67882085e-01
-9.59204018e-01 -7.12338388e-01 -6.64445341e-01 5.11576645e-02
-2.79899776e-01 3.69091928e-02 6.27274156e-01 1.64144844e-01
-9.23187554e-01 5.95895767e-01 -6.06630445e-01 -2.97752738e-01
4.87863481e-01 4.15943831e-01 -1.09853886e-01 -1.35554656e-01
-9.71760809e-01 1.02622855e+00 5.81170321e-01 -6.12768084e-02
-1.10442317e+00 -8.12147498e-01 -6.96999550e-01 -6.03648163e-02
7.94713914e-01 -9.17048812e-01 1.45470393e+00 -9.31356728e-01
-1.25982559e+00 7.37991989e-01 -8.82791504e-02 -2.61320591e-01
5.96966505e-01 -5.43069363e-01 -1.21444769e-01 -3.19366634e-01
4.05543335e-02 4.58910733e-01 1.20437396e+00 -1.25910139e+00
-7.63099790e-01 -4.00212884e-01 2.63704479e-01 4.64707345e-01
-5.67065835e-01 -8.88886005e-02 -3.96101773e-01 -7.55867302e-01
-3.90586793e-01 -1.07101989e+00 -3.96514386e-01 -1.00337649e-02
-2.43414998e-01 -5.91383100e-01 7.85937607e-01 -1.80510879e-01
9.38640535e-01 -2.08556247e+00 6.92412019e-01 -8.25595111e-02
3.72553736e-01 2.45044485e-01 -3.38983208e-01 2.87311912e-01
-1.00603057e-02 -2.05374092e-01 -4.67877358e-01 -7.81899154e-01
-6.10494129e-02 3.73428017e-01 2.45661691e-01 3.68430138e-01
2.72792906e-01 9.25903082e-01 -1.09034991e+00 -4.81953532e-01
1.74215361e-01 3.27870965e-01 -3.38363320e-01 1.86170086e-01
-4.12069768e-01 5.70028484e-01 -6.17578626e-01 4.20378178e-01
2.61424661e-01 -6.01668954e-01 1.40080392e-01 -6.41786680e-02
2.74244189e-01 1.65470839e-01 -9.53283489e-01 1.95370018e+00
-1.11106706e+00 5.37158489e-01 1.77939281e-01 -1.47335839e+00
8.60516727e-01 3.08147192e-01 5.67357898e-01 -5.19176424e-01
-8.06875676e-02 1.09428130e-01 9.80790481e-02 -4.56952900e-01
3.06032956e-01 -4.24432009e-02 -1.14502847e-01 5.49943447e-01
3.42109233e-01 -6.37075678e-02 2.09492341e-01 -8.17793980e-02
9.63051677e-01 2.44269401e-01 4.56086636e-01 -3.88760209e-01
6.10304117e-01 -1.65457070e-01 3.98584634e-01 7.89954424e-01
5.50397187e-02 2.31681153e-01 3.47836524e-01 -4.59834307e-01
-8.80357325e-01 -7.98998833e-01 -3.40519845e-01 1.82703090e+00
1.21686643e-03 -2.43881866e-01 -2.58018225e-01 -9.18618977e-01
3.00838053e-01 5.30892193e-01 -7.52070189e-01 -2.69733369e-01
-4.24325317e-01 -9.39401507e-01 1.86719656e-01 7.51528069e-02
2.22042173e-01 -1.16693914e+00 -5.96824348e-01 4.25661802e-01
-2.44638517e-01 -9.82481480e-01 -4.36584860e-01 7.69705057e-01
-9.33906257e-01 -1.17193532e+00 -8.48746181e-01 -7.58761585e-01
4.76440936e-01 2.60347694e-01 1.29125667e+00 -1.80825382e-01
-2.51468301e-01 4.97418463e-01 -4.10730958e-01 -4.73361552e-01
-2.92330652e-01 5.96073210e-01 -4.66330461e-02 1.51665941e-01
1.99622996e-02 -7.72848248e-01 -4.15021956e-01 6.06150210e-01
-7.09124804e-01 -1.05795920e-01 8.46801996e-01 1.11195445e+00
5.16885519e-01 2.59308261e-03 7.04060674e-01 -1.07181919e+00
8.58167112e-01 -7.14717388e-01 -4.30469662e-01 5.03198504e-01
-8.22111249e-01 3.99260193e-01 6.98446274e-01 -7.66397476e-01
-1.02517533e+00 4.73677851e-02 3.61180723e-01 -6.41737878e-01
-2.34033410e-02 3.90637904e-01 5.70404194e-02 -1.24582127e-01
7.86090136e-01 2.14529768e-01 -6.53950423e-02 -6.55847549e-01
3.01674604e-01 3.39716643e-01 3.15001339e-01 -1.02172852e+00
5.92104077e-01 1.78879201e-01 1.01355717e-01 -5.30704141e-01
-1.26786947e+00 -7.38893151e-01 -6.41987443e-01 -2.66594887e-01
6.10029817e-01 -7.45755374e-01 -7.53072262e-01 5.35004795e-01
-1.07570589e+00 -7.42805362e-01 -1.20089762e-01 3.45273197e-01
-7.98337936e-01 1.23264872e-01 -1.68269023e-01 -6.89812541e-01
-1.72043815e-01 -1.28904247e+00 1.21828783e+00 -1.91149667e-01
-5.99697195e-02 -1.45497668e+00 1.71497479e-01 4.46568400e-01
1.94516912e-01 8.07681531e-02 9.26562011e-01 -7.90392697e-01
-2.16439486e-01 2.89596438e-01 -1.27035305e-01 3.76649171e-01
3.31105590e-01 -4.88049299e-01 -9.93088901e-01 -5.03058732e-01
1.04847357e-01 -7.19561577e-01 1.16396129e+00 4.02474761e-01
1.39800525e+00 -3.73123527e-01 -3.71921390e-01 4.99824524e-01
1.22124529e+00 -2.01185178e-02 1.98005602e-01 6.43790066e-01
6.98133171e-01 8.37728024e-01 8.31468165e-01 5.12947381e-01
5.24592042e-01 9.33657885e-01 5.80100000e-01 6.65172338e-02
2.32335225e-01 1.58607736e-01 2.69664735e-01 5.13592482e-01
-7.42875934e-02 -1.45944297e-01 -1.11954415e+00 6.69531167e-01
-2.30551505e+00 -7.08086312e-01 -9.92177948e-02 2.18068075e+00
9.03975964e-01 2.26102039e-01 3.50095272e-01 4.58428301e-02
6.40168011e-01 2.53827870e-01 -8.14977825e-01 -4.34057087e-01
1.92013398e-01 2.99078953e-02 3.62924278e-01 4.91195410e-01
-1.47765040e+00 8.44546139e-01 5.72535324e+00 9.81995761e-01
-8.56892228e-01 4.77453589e-01 4.54073995e-01 -2.17105821e-01
-9.09644067e-02 -2.33515784e-01 -6.91535890e-01 3.06589872e-01
5.69036603e-01 -3.22638124e-01 5.22903800e-01 9.03968513e-01
-1.15218520e-01 8.47255141e-02 -1.20155942e+00 8.60389411e-01
-2.51921769e-02 -7.51217067e-01 -1.34483114e-01 6.57946104e-03
7.86164463e-01 1.41593337e-01 2.98435897e-01 4.46454197e-01
6.36123061e-01 -6.78193271e-01 5.75768292e-01 3.74911726e-01
3.40877593e-01 -7.03270912e-01 4.74449009e-01 5.91597199e-01
-1.23757839e+00 -5.65842688e-01 -3.70442092e-01 7.55053898e-03
-4.33401540e-02 8.06933224e-01 -8.15707803e-01 6.53616071e-01
8.06084216e-01 9.01351929e-01 -7.06634164e-01 1.04620874e+00
-7.23203421e-02 2.54587978e-01 -2.29707174e-02 1.32858351e-01
4.02247459e-01 -1.13705203e-01 7.50506818e-01 1.23877954e+00
1.09414600e-01 -4.61972207e-01 5.19542158e-01 5.26579916e-01
4.65764627e-02 -7.53289610e-02 -5.84125578e-01 2.70049632e-01
1.86650515e-01 1.45219481e+00 -2.72610933e-01 -2.19704751e-02
-3.68991613e-01 9.94821727e-01 8.46089840e-01 3.74205232e-01
-6.11914396e-01 -1.59860954e-01 7.69642830e-01 -1.63913220e-01
2.71123558e-01 -2.44755611e-01 -2.88333863e-01 -9.76504147e-01
1.48818716e-01 -8.30891073e-01 6.29881501e-01 -3.99682939e-01
-1.57232928e+00 5.65050900e-01 2.79820085e-01 -1.32608712e+00
-3.64932775e-01 -5.69829464e-01 -5.12278020e-01 7.11972713e-01
-1.81852722e+00 -1.08231461e+00 -1.41805604e-01 7.37943232e-01
1.00538075e+00 -3.30066651e-01 5.86403072e-01 5.74144945e-02
-4.53022093e-01 5.87142289e-01 2.95887768e-01 -3.75914723e-01
9.82236445e-01 -1.45041096e+00 2.17762694e-01 3.20650876e-01
1.96843177e-01 1.83115408e-01 6.73002779e-01 -2.99765885e-01
-1.06149018e+00 -1.25985050e+00 7.17083573e-01 -5.45967579e-01
7.79780209e-01 -5.46359301e-01 -1.03127027e+00 6.80972934e-01
7.94642940e-02 -1.87201221e-02 5.39794683e-01 6.64143503e-01
-3.16494465e-01 -2.21752852e-01 -8.48284483e-01 3.89738798e-01
1.19716907e+00 -3.97424370e-01 -6.28471911e-01 7.63957620e-01
7.89144874e-01 -2.19597906e-01 -9.95330811e-01 3.79860908e-01
4.60412234e-01 -7.37117887e-01 1.05986762e+00 -6.57848895e-01
4.33937550e-01 2.69336589e-02 3.52757834e-02 -1.71901822e+00
-4.35858637e-01 -6.36845946e-01 -3.11549187e-01 9.81009245e-01
7.21184909e-01 -7.62820840e-01 4.81245399e-01 3.66289169e-01
-3.23136032e-01 -9.20250773e-01 -9.39864516e-01 -1.09120011e+00
9.83482674e-02 -5.08231163e-01 1.67821690e-01 1.02652001e+00
-1.74268156e-01 5.72007656e-01 -5.17416239e-01 7.88575858e-02
1.03265667e+00 -7.17968196e-02 6.40985847e-01 -1.59966314e+00
-6.10488236e-01 -5.55165887e-01 -4.52033915e-02 -8.00631940e-01
5.81242800e-01 -1.06709290e+00 7.83035997e-03 -1.29939258e+00
1.70397803e-01 -7.73588777e-01 -5.96178710e-01 7.64177978e-01
-5.35830259e-01 -5.81663623e-02 2.70586759e-01 4.67510998e-01
-1.02434742e+00 7.91860521e-01 1.48256719e+00 -2.22419247e-01
-3.37823898e-01 5.18719316e-01 -7.74113178e-01 7.12802410e-01
9.90947545e-01 -7.77156651e-01 -5.77638388e-01 -7.51880467e-01
2.59138823e-01 -1.07805736e-01 4.62708235e-01 -9.26616907e-01
3.32963407e-01 -2.54676580e-01 7.15276077e-02 -1.47973493e-01
5.42620242e-01 -9.51543987e-01 -7.22811967e-02 2.79924661e-01
-6.53059363e-01 1.27058253e-01 1.45590544e-01 7.79456496e-01
-1.61999032e-01 -3.24568391e-01 7.28325844e-01 -1.31837249e-01
-5.96901953e-01 3.42875510e-01 -9.82897431e-02 3.37117195e-01
1.01543105e+00 -3.81291173e-02 -1.70527950e-01 -3.77239659e-02
-7.87386894e-01 7.41384447e-01 6.55280799e-02 6.45463347e-01
3.59058082e-01 -1.24281657e+00 -8.43193233e-01 -1.15473002e-01
5.01739681e-01 3.18198979e-01 2.50801861e-01 8.34422886e-01
3.78260702e-01 2.60964304e-01 -1.48586720e-01 -9.49399889e-01
-1.24146831e+00 5.41312873e-01 2.84135967e-01 -8.42000484e-01
-3.90316039e-01 7.24116027e-01 4.54713225e-01 -7.81543434e-01
3.07588220e-01 -1.90479737e-02 -3.81338775e-01 2.95072615e-01
1.64509267e-01 4.76811230e-01 1.31460354e-01 -2.57613242e-01
-2.27125928e-01 6.82947755e-01 -3.72277766e-01 2.10553035e-02
1.39947903e+00 -2.99697835e-02 1.25363603e-01 7.24155009e-01
1.20447958e+00 -7.86752999e-01 -1.61848807e+00 -7.51502752e-01
3.69528949e-01 -3.43491614e-01 1.66084934e-02 -7.95224249e-01
-9.65927899e-01 8.37764323e-01 3.19611490e-01 3.27546000e-02
1.19111264e+00 2.85715491e-01 3.24013770e-01 7.85064518e-01
3.46330047e-01 -1.04423666e+00 6.14910483e-01 7.07651496e-01
1.00139606e+00 -1.39073849e+00 -9.72384587e-02 -9.33333188e-02
-9.73569393e-01 8.84119630e-01 6.59661949e-01 -1.64347813e-02
7.55205750e-01 7.71503523e-02 -2.37850830e-01 -1.95445180e-01
-1.17494428e+00 -3.03383440e-01 5.61476529e-01 7.02435672e-01
2.66270638e-01 -1.89735770e-01 -1.00759119e-01 3.30189317e-01
5.50056025e-02 -1.02222450e-01 -1.38359275e-02 8.81108642e-01
-3.66309941e-01 -1.44012046e+00 -1.94496349e-01 7.35764205e-01
-3.49810302e-01 -4.54439744e-02 -6.86396584e-02 6.86712563e-01
8.80041793e-02 8.21879804e-01 -6.80211931e-03 -2.47965455e-01
4.05822307e-01 1.76121652e-01 4.86649662e-01 -6.97788656e-01
-8.49983811e-01 -9.81802419e-02 1.00179046e-01 -5.06668627e-01
-6.27056360e-01 -7.45193303e-01 -6.44776285e-01 6.63078129e-02
-2.14871958e-01 5.44664524e-02 5.77297151e-01 1.00842237e+00
2.04480648e-01 6.59203291e-01 8.28785241e-01 -9.66708899e-01
-7.66366124e-01 -9.05699670e-01 -4.28334951e-01 3.63625854e-01
4.97078598e-01 -1.10565329e+00 -3.53439242e-01 -6.35607019e-02] | [9.284388542175293, 3.75567626953125] |
fbcaacfb-41ff-4612-abe0-ee7be79fe345 | aop-net-all-in-one-perception-network-for | 2302.00885 | null | https://arxiv.org/abs/2302.00885v1 | https://arxiv.org/pdf/2302.00885v1.pdf | AOP-Net: All-in-One Perception Network for Joint LiDAR-based 3D Object Detection and Panoptic Segmentation | LiDAR-based 3D object detection and panoptic segmentation are two crucial tasks in the perception systems of autonomous vehicles and robots. In this paper, we propose All-in-One Perception Network (AOP-Net), a LiDAR-based multi-task framework that combines 3D object detection and panoptic segmentation. In this method, a dual-task 3D backbone is developed to extract both panoptic- and detection-level features from the input LiDAR point cloud. Also, a new 2D backbone that intertwines Multi-Layer Perceptron (MLP) and convolution layers is designed to further improve the detection task performance. Finally, a novel module is proposed to guide the detection head by recovering useful features discarded during down-sampling operations in the 3D backbone. This module leverages estimated instance segmentation masks to recover detailed information from each candidate object. The AOP-Net achieves state-of-the-art performance for published works on the nuScenes benchmark for both 3D object detection and panoptic segmentation tasks. Also, experiments show that our method easily adapts to and significantly improves the performance of any BEV-based 3D object detection method. | ['Bingbing Liu', 'Yuan Ren', 'Hamidreza Fazlali', 'YiXuan Xu'] | 2023-02-02 | null | null | null | null | ['panoptic-segmentation'] | ['computer-vision'] | [ 3.60369653e-01 -2.94809788e-01 -1.54081583e-01 -5.83584547e-01
-5.71818829e-01 -5.51044881e-01 7.12869883e-01 1.94179684e-01
-5.69456756e-01 3.94971594e-02 -6.44346893e-01 -3.77497584e-01
1.50550619e-01 -1.13095677e+00 -8.69962871e-01 -6.68032467e-01
-1.94561593e-02 7.10499465e-01 8.74175966e-01 -2.59480756e-02
1.99574500e-01 9.67938781e-01 -2.06394958e+00 8.31027925e-02
7.36762047e-01 1.42067957e+00 6.96441650e-01 7.45451152e-01
-2.70777136e-01 5.48749678e-02 -1.67880356e-01 2.28701681e-01
8.09313476e-01 4.97288436e-01 -1.29915029e-01 2.29376748e-01
1.06721461e+00 -4.88560587e-01 9.31309387e-02 1.06305718e+00
4.26656097e-01 1.75615102e-02 8.50706100e-01 -1.18699253e+00
-1.67352393e-01 -2.71496922e-02 -9.88710165e-01 1.94904819e-01
-2.77933061e-01 4.84532684e-01 9.47214663e-01 -1.31614792e+00
1.40991017e-01 1.73192608e+00 7.80049980e-01 1.34844899e-01
-1.20920253e+00 -7.86811709e-01 1.84824303e-01 -1.53392673e-01
-1.15319335e+00 -1.26368538e-01 5.83620250e-01 -8.02392244e-01
1.17392063e+00 4.93574515e-02 5.66704452e-01 4.64239180e-01
2.13166982e-01 7.89976776e-01 1.09337938e+00 -8.22794810e-03
-3.80787440e-02 1.57788709e-01 3.18037242e-01 8.73359144e-01
5.08980334e-01 3.96235526e-01 -2.64360309e-01 3.23379964e-01
6.64826512e-01 3.04261334e-02 3.60958964e-01 -5.13756871e-01
-9.23065543e-01 9.66338038e-01 8.48543823e-01 -4.37421441e-01
-5.08705199e-01 1.68210864e-01 2.28208721e-01 1.70247629e-02
6.65734649e-01 1.32325590e-01 -5.12037337e-01 5.86815476e-01
-9.33912575e-01 4.01727021e-01 6.85437322e-01 9.66827512e-01
1.36151171e+00 -1.47969529e-01 -1.94329247e-01 7.78855085e-01
7.18380034e-01 9.94562209e-01 -3.18364024e-01 -1.01018894e+00
4.33409214e-01 9.77048874e-01 1.52634025e-01 -6.26910508e-01
-7.21981764e-01 -1.84357896e-01 -5.23106754e-01 9.22448039e-01
-1.10517010e-01 9.50536430e-02 -1.50933981e+00 1.06801867e+00
8.21972609e-01 2.58395001e-02 3.68889309e-02 1.03733468e+00
1.30835557e+00 8.99521887e-01 -1.56077174e-02 2.63575137e-01
1.62525678e+00 -9.82232869e-01 9.02054086e-02 -7.95151353e-01
-1.17272101e-01 -7.21516192e-01 7.14635432e-01 -1.59525141e-01
-8.48298550e-01 -1.03347790e+00 -1.24513853e+00 -3.10278922e-01
-8.02463055e-01 1.81669682e-01 4.93053198e-01 5.18810809e-01
-7.93041527e-01 1.12891816e-01 -6.11455619e-01 -4.34244782e-01
6.29964173e-01 2.99893796e-01 9.17658731e-02 5.84508963e-02
-6.73955858e-01 1.18645620e+00 8.69775057e-01 9.57700014e-02
-1.16742849e+00 -8.43443394e-01 -9.51238334e-01 6.37748763e-02
5.82310259e-01 -7.65016556e-01 1.26534486e+00 -9.09938142e-02
-1.19373000e+00 1.26825285e+00 -1.15666851e-01 -7.46812403e-01
2.78040558e-01 -2.62195885e-01 1.76976874e-01 2.86464900e-01
3.54806572e-01 1.60064113e+00 1.09651935e+00 -1.47990716e+00
-1.41261828e+00 -4.47100967e-01 -7.33466968e-02 3.67343247e-01
4.37669963e-01 -2.10403189e-01 -5.96755266e-01 4.43738066e-02
2.56035388e-01 -6.30409062e-01 -3.87347907e-01 4.00977284e-01
-3.98921221e-01 -5.19318521e-01 1.15907025e+00 -1.90150768e-01
2.39036277e-01 -1.95178831e+00 -3.15281421e-01 -6.25278130e-02
1.86233461e-01 3.45421612e-01 -2.40102559e-01 4.64916751e-02
4.34086323e-01 -1.45687282e-01 -5.37744880e-01 -7.19354212e-01
1.81730315e-01 4.00480419e-01 -4.43442494e-01 3.20221841e-01
7.09795356e-01 9.20257807e-01 -5.55538774e-01 -6.06065154e-01
8.53805721e-01 3.12084138e-01 -3.58900130e-01 2.94841260e-01
-8.91052246e-01 3.18779230e-01 -5.33275604e-01 1.12322128e+00
1.47354460e+00 1.24473162e-01 -5.39431036e-01 -5.74299768e-02
-8.57978463e-01 3.49770457e-01 -9.38006818e-01 1.39933193e+00
-5.10491848e-01 6.43172085e-01 4.35314000e-01 -8.63960981e-01
1.15620565e+00 -2.06233516e-01 2.81857818e-01 -6.56498373e-01
2.79330164e-01 3.33274484e-01 -3.74350429e-01 -5.69959342e-01
7.91288853e-01 9.37480200e-03 -9.32101831e-02 2.93930136e-02
4.64690812e-02 -9.09649491e-01 1.99462444e-01 -2.65616596e-01
4.03676033e-01 1.55787572e-01 2.00368762e-01 -1.60528123e-01
5.38373351e-01 3.74930263e-01 5.96610665e-01 1.21642923e+00
-3.50813150e-01 3.60404938e-01 8.78245458e-02 -4.22189057e-01
-9.45886910e-01 -1.32227564e+00 -5.57466507e-01 1.05007899e+00
4.77988303e-01 2.49967203e-01 -3.06462765e-01 -5.12563169e-01
6.94730699e-01 6.70706451e-01 -3.77711415e-01 3.78604412e-01
-4.36915278e-01 -8.37532222e-01 4.27032292e-01 4.59976077e-01
7.73213327e-01 -1.03413224e+00 -1.27165902e+00 2.45664448e-01
2.65245408e-01 -1.51885855e+00 1.19954221e-01 5.65292060e-01
-8.42846334e-01 -1.03406155e+00 -2.02774569e-01 -8.66839409e-01
2.98611134e-01 1.00534570e+00 8.02325964e-01 -4.87344712e-01
-4.22761112e-01 2.14369640e-01 -2.48894580e-02 -1.05936825e+00
-3.19576353e-01 3.11407428e-02 -9.58043411e-02 -1.01683766e-01
7.25998998e-01 -6.13013327e-01 -4.26112682e-01 3.45563740e-01
-5.00906050e-01 8.17667544e-02 9.18007612e-01 2.33125195e-01
8.81995738e-01 -1.79934859e-01 1.91402391e-01 -3.23069900e-01
1.06144287e-01 -3.54089200e-01 -1.44882965e+00 -8.65462869e-02
-4.78081435e-01 -2.45918319e-01 7.92246759e-02 -1.05306253e-01
-8.07673991e-01 4.32933360e-01 -1.47502303e-01 -7.16486931e-01
-5.73767781e-01 1.18117914e-01 1.02750519e-02 -2.58451492e-01
4.85520691e-01 2.66029239e-01 -1.05779350e-01 -8.02830458e-01
5.85171461e-01 9.22152102e-01 7.11362422e-01 -2.59674847e-01
1.12284338e+00 9.20491815e-01 3.31933379e-01 -1.17503726e+00
-1.23952031e+00 -1.09340799e+00 -9.17264342e-01 -2.21537471e-01
1.09860575e+00 -1.25375390e+00 -8.66501331e-01 6.22545719e-01
-1.48904312e+00 -2.50909775e-01 -2.77087122e-01 3.70201737e-01
-4.79361564e-01 4.10031043e-02 -1.54927298e-01 -1.14323735e+00
-6.16341531e-01 -1.27109480e+00 1.65504658e+00 4.74852651e-01
6.67519748e-01 -3.50955814e-01 -5.76408580e-02 4.41254735e-01
2.17866868e-01 2.30181351e-01 7.20404983e-01 -4.28708106e-01
-1.22453332e+00 -1.27828121e-01 -1.03751791e+00 4.71807331e-01
-1.96638837e-01 -7.56251859e-03 -1.29364085e+00 -1.34741530e-01
-3.09878998e-02 -3.66842240e-01 1.65992880e+00 6.28732264e-01
7.47384071e-01 1.56907022e-01 -5.56932926e-01 7.06678450e-01
1.41746879e+00 2.96897683e-02 -8.42552353e-03 1.39211014e-01
7.84851134e-01 8.35592389e-01 8.36831927e-01 1.27687663e-01
6.36172831e-01 4.01876450e-01 1.06142735e+00 -8.40942860e-02
-1.53823167e-01 -4.53085937e-02 1.74108595e-01 9.82592478e-02
2.69356132e-01 -7.81073943e-02 -9.33614612e-01 6.34946406e-01
-1.76272190e+00 -6.93894267e-01 -3.28782588e-01 1.87139034e+00
5.22628725e-01 6.12162054e-01 3.98077555e-02 -2.80710012e-01
7.30733514e-01 4.87315834e-01 -8.08522165e-01 -3.27993304e-01
-3.65136936e-02 1.94595113e-01 9.15611923e-01 5.87620199e-01
-1.62412846e+00 1.38180542e+00 5.55858564e+00 7.36453056e-01
-1.09046984e+00 2.15730995e-01 -2.14553140e-02 8.59345570e-02
2.94640958e-01 -1.55042931e-01 -1.58058190e+00 7.44469911e-02
4.85017151e-01 3.99775892e-01 2.69402526e-02 1.05169427e+00
4.35322553e-01 -4.06671852e-01 -9.08371150e-01 7.92163253e-01
-1.27404690e-01 -1.36857891e+00 1.70298532e-01 8.99082720e-02
5.06374300e-01 1.08921218e+00 -5.15451059e-02 1.56652808e-01
2.28035197e-01 -7.28600204e-01 8.64934683e-01 3.02328318e-01
6.34283662e-01 -5.83807170e-01 3.69746476e-01 5.94321609e-01
-1.54867923e+00 -1.94788396e-01 -7.15345442e-01 5.06388359e-02
2.04047993e-01 8.76665175e-01 -1.29570961e+00 3.57161313e-01
8.92740488e-01 6.35194361e-01 -4.80355859e-01 1.49687433e+00
-4.14659023e-01 3.33650976e-01 -7.89670408e-01 -7.39289969e-02
8.09665620e-01 -1.77243963e-01 1.11873245e+00 1.56476498e+00
7.42082819e-02 -3.30672294e-01 6.75886095e-01 1.31968153e+00
-1.02928437e-01 -4.20434356e-01 -7.85053194e-01 3.84521902e-01
5.58136046e-01 1.64360821e+00 -6.99783504e-01 -3.09208870e-01
-4.12347049e-01 2.60989875e-01 -3.24709639e-02 7.67620206e-02
-6.53790414e-01 -4.01005685e-01 9.81090426e-01 -7.34009594e-02
1.02712882e+00 -5.85853815e-01 -4.55031306e-01 -5.12057364e-01
-7.58068115e-02 -1.28152028e-01 2.14373142e-01 -7.55515635e-01
-1.20734572e+00 1.72487289e-01 1.24285623e-01 -1.05766261e+00
3.12514573e-01 -9.50381279e-01 -7.42338836e-01 9.32657778e-01
-2.44486141e+00 -1.57586026e+00 -6.39319718e-01 1.25446409e-01
8.90804410e-01 1.29880771e-01 2.22242251e-01 8.70596617e-02
-3.82638663e-01 -2.46320158e-01 -1.87308550e-01 -2.29204044e-01
3.19291204e-01 -1.25262630e+00 7.36745596e-01 8.92219663e-01
-1.75667748e-01 -1.17965765e-01 3.07421952e-01 -6.10982239e-01
-1.18290043e+00 -1.64253223e+00 7.04210520e-01 -4.14581120e-01
6.15733564e-01 -6.08258009e-01 -7.94591606e-01 4.77362722e-01
-2.17203237e-02 -1.27096832e-01 -3.56256738e-02 -2.85874039e-01
-2.98395395e-01 -3.91581535e-01 -1.13693571e+00 2.75851220e-01
9.13161993e-01 -3.75385642e-01 -9.00837421e-01 4.01059359e-01
1.12442911e+00 -5.31493902e-01 -4.77314234e-01 7.95920372e-01
4.50120866e-01 -9.24134552e-01 1.28765869e+00 -9.75120217e-02
1.45127431e-01 -6.56453311e-01 -3.86576474e-01 -7.59965062e-01
-1.32360846e-01 -4.24484938e-01 -1.29417852e-01 8.44723165e-01
3.80617172e-01 -6.16509676e-01 6.25748992e-01 -4.67713565e-01
-6.41876101e-01 -3.46233696e-01 -1.16804779e+00 -8.17802191e-01
-1.73090622e-01 -7.42354870e-01 4.22866166e-01 3.38096648e-01
-1.05181980e+00 7.38494873e-01 1.27653807e-01 8.76424253e-01
1.08545434e+00 7.75863528e-01 9.22403157e-01 -1.69086909e+00
2.51888007e-01 -6.86149120e-01 -1.28125995e-01 -1.61164391e+00
-2.39003310e-03 -1.10607743e+00 6.54189408e-01 -1.72442007e+00
-7.42125511e-02 -4.81414795e-01 -5.97895384e-02 4.00567085e-01
4.66342755e-02 3.36748570e-01 3.63823861e-01 3.02811950e-01
-4.61492807e-01 5.55430412e-01 1.21464670e+00 -4.04644251e-01
-5.10992229e-01 4.53743756e-01 -2.80977488e-01 8.99703145e-01
7.35799730e-01 -6.02113605e-01 1.19877513e-02 -5.99256456e-01
-1.30514443e-01 -3.86256546e-01 9.12693441e-01 -1.12189460e+00
3.50590914e-01 -1.00350551e-01 2.35333264e-01 -1.90488148e+00
8.12135577e-01 -7.03679979e-01 -8.29452813e-01 5.56769073e-01
2.75320888e-01 -5.05233824e-01 4.54673856e-01 5.37544608e-01
1.29985854e-01 -2.40934283e-01 1.14273918e+00 -2.89313972e-01
-1.05566370e+00 5.41292846e-01 -4.59158152e-01 -3.09299707e-01
8.96387160e-01 -4.57378983e-01 -3.30060452e-01 4.99149173e-01
-2.64051706e-01 9.36772466e-01 4.35943045e-02 5.58000386e-01
7.79551566e-01 -7.95176685e-01 -8.40785563e-01 2.84040034e-01
3.12345803e-01 8.46313238e-01 -1.88828468e-01 7.95641661e-01
-4.90215361e-01 7.27451384e-01 -1.28295928e-01 -1.26989460e+00
-1.10890186e+00 4.44855899e-01 4.58095759e-01 2.57890493e-01
-6.06578827e-01 9.00138557e-01 3.91424388e-01 -8.58294249e-01
3.70741695e-01 -8.70824099e-01 -2.72064626e-01 4.34673905e-01
2.26543605e-01 4.17936146e-01 -1.40005246e-01 -5.01740694e-01
-3.87341857e-01 8.30976784e-01 8.32243711e-02 1.30305558e-01
1.40694153e+00 -1.73412442e-01 -2.16243014e-01 6.25483215e-01
8.70466590e-01 -5.23614943e-01 -1.70509756e+00 -5.04069030e-01
-1.01013862e-01 -2.85171628e-01 2.34857082e-01 -6.61889195e-01
-7.06563234e-01 1.34059727e+00 6.60361767e-01 3.15347761e-01
8.17702234e-01 2.19501123e-01 7.34643877e-01 8.51368785e-01
-3.05228699e-02 -9.79193985e-01 -1.04630329e-01 9.89603877e-01
7.95583069e-01 -1.64910626e+00 3.10665704e-02 -6.15788698e-01
-1.22177213e-01 1.01210773e+00 7.27573395e-01 -1.74798712e-01
6.61533237e-01 6.01872727e-02 -5.44859953e-02 -4.64835197e-01
-4.60782588e-01 -9.59161401e-01 4.52863902e-01 6.98984802e-01
-5.45545340e-01 8.02513212e-02 2.21864238e-01 2.82673482e-02
1.01629041e-01 -4.28180099e-01 8.29564035e-02 8.02917182e-01
-1.46040833e+00 -4.24401253e-01 -5.20025313e-01 3.93498421e-01
3.57725114e-01 -1.78242445e-01 -3.31852108e-01 8.01372528e-01
6.86523080e-01 8.19329143e-01 5.10494709e-01 -2.27590725e-01
3.63613993e-01 -1.49946779e-01 3.47143054e-01 -8.02148342e-01
-4.27445233e-01 1.69385627e-01 6.16775602e-02 -4.64137137e-01
-5.65151930e-01 -4.91546303e-01 -1.23620868e+00 1.30290449e-01
-2.95050949e-01 -4.18422192e-01 1.10000885e+00 7.97789991e-01
2.98605293e-01 3.03138345e-01 7.52300799e-01 -1.62432361e+00
-5.05563200e-01 -9.22378600e-01 -2.78975695e-01 -5.51837683e-01
7.68173575e-01 -9.00573969e-01 -1.46405667e-01 -2.50590265e-01] | [7.91614294052124, -2.7033028602600098] |
d5efc299-2207-4065-a554-91246801dd55 | the-effects-of-hofstede-s-cultural-dimensions | 2301.04609 | null | https://arxiv.org/abs/2301.04609v1 | https://arxiv.org/pdf/2301.04609v1.pdf | The Effects of Hofstede's Cultural Dimensions on Pro-Environmental Behaviour: How Culture Influences Environmentally Conscious Behaviour | The need for a more sustainable lifestyle is a key focus for several countries. Using a questionnaire survey conducted in Hungary, this paper examines how culture influences environmentally conscious behaviour. Having investigated the direct impact of Hofstedes cultural dimensions on pro-environmental behaviour, we found that the culture of a country hardly affects actual environmentally conscious behaviour. The findings indicate that only individualism and power distance have a significant but weak negative impact on pro-environmental behaviour. Based on the findings, we can state that a positive change in culture is a necessary but not sufficient condition for making a country greener. | ['Csilla Konyha Molnarne', 'Szabolcs Nagy'] | 2022-12-26 | null | null | null | null | ['culture'] | ['speech'] | [-1.28746867e-01 -7.83991292e-02 -4.19366866e-01 1.93904880e-02
3.12597185e-01 -3.24514478e-01 6.60344064e-01 1.06110550e-01
-6.47042394e-01 6.17995679e-01 4.71895933e-01 -5.41995585e-01
-7.14723915e-02 -8.89341831e-01 -3.06595534e-01 -8.90481055e-01
6.41708434e-01 -6.46348596e-01 -3.66344094e-01 -3.19750875e-01
6.72637284e-01 2.00184405e-01 -1.16514719e+00 -5.61694980e-01
1.32277834e+00 2.21286695e-02 4.63181257e-01 2.36180708e-01
-1.85812078e-02 5.66735625e-01 -3.47770900e-01 -5.12387753e-01
9.58591104e-02 -5.96701205e-01 -4.19400424e-01 3.36457416e-02
1.26440555e-01 -1.36846125e-01 5.14231503e-01 1.51105297e+00
3.45462680e-01 3.36104184e-02 6.17013037e-01 -7.62985647e-01
-9.08120394e-01 8.06822002e-01 1.53024688e-01 -1.15318380e-01
2.13258550e-01 4.17621821e-01 6.69091105e-01 -3.37742448e-01
5.74834287e-01 1.07200408e+00 5.92569411e-01 1.03567854e-01
-1.27256691e+00 -1.16133225e+00 3.89740542e-02 1.03066929e-01
-1.28121603e+00 -6.48044825e-01 7.38663018e-01 -3.26298088e-01
7.33547866e-01 4.00824368e-01 1.42377150e+00 6.22388005e-01
8.26782942e-01 -6.78869605e-01 2.08052850e+00 -8.31430197e-01
6.50986612e-01 7.85314798e-01 -3.22132617e-01 2.36439221e-02
1.35445774e+00 9.17923376e-02 1.93180725e-01 1.36044353e-01
5.07774174e-01 -2.06187233e-01 1.26416549e-01 1.66356787e-01
-9.10267770e-01 6.81734383e-01 1.77871451e-01 9.39054072e-01
-3.95111144e-01 1.84958637e-01 1.60278872e-01 4.22962964e-01
-9.47551578e-02 4.94527072e-01 1.51854217e-01 -3.97447288e-01
-1.77057073e-01 -4.42820728e-01 9.13867891e-01 1.35325268e-01
6.39575362e-01 2.24395260e-01 5.02370775e-01 2.33732089e-01
6.65970922e-01 1.07064116e+00 -4.00397032e-02 -1.28896773e+00
-3.23977232e-01 4.87614363e-01 4.47257072e-01 -1.22618735e+00
-2.49886215e-01 -6.17487788e-01 -1.40105277e-01 5.25310576e-01
3.57255071e-01 -4.33129132e-01 -5.00056684e-01 1.64552879e+00
2.10932776e-01 -5.66586375e-01 2.30523184e-01 7.80397713e-01
2.14460760e-01 5.59915066e-01 8.66313994e-01 -6.49864733e-01
1.09463489e+00 -3.67638350e-01 -1.22590244e+00 -2.77102828e-01
6.21646285e-01 -8.47515881e-01 9.77460563e-01 2.79581755e-01
-7.30400503e-01 -4.92734760e-01 -8.23668242e-01 4.67736363e-01
-8.88514519e-01 -2.51179755e-01 6.04911268e-01 1.57216835e+00
-1.12482262e+00 2.18322337e-01 -3.63441348e-01 -1.08119488e+00
-3.72145116e-01 -7.46200932e-03 -3.59786958e-01 -8.82446468e-02
-8.90584528e-01 1.15406883e+00 6.04171276e-01 4.06855866e-02
-1.58773199e-01 -7.20663443e-02 -5.59242129e-01 -3.29680681e-01
3.21520656e-01 -5.29125810e-01 6.83636963e-01 -1.16189790e+00
-1.24109638e+00 4.28076863e-01 1.70148462e-01 -1.46670379e-02
2.24603400e-01 -9.80155542e-02 -1.26549900e+00 4.15671058e-02
3.67708296e-01 2.39208445e-01 1.12358063e-01 -1.53089058e+00
-8.91365290e-01 -4.71456677e-01 -7.98223466e-02 2.33616173e-01
-5.21466613e-01 2.44334042e-01 3.77761364e-01 -1.86153978e-01
-1.51977167e-01 -1.20403576e+00 -6.41015291e-01 -9.16305602e-01
-1.85109556e-01 -1.57034993e-01 5.13054848e-01 -6.47074163e-01
1.28143132e+00 -2.26595116e+00 -7.26680696e-01 3.19945991e-01
-4.26711142e-01 -1.16160199e-01 2.34188184e-01 7.88039625e-01
3.27000052e-01 6.11249268e-01 4.02088463e-01 7.04917908e-01
4.85723734e-01 4.80854332e-01 4.79247123e-01 6.39755309e-01
-1.26897752e-01 1.19533926e-01 -1.21612918e+00 -3.10545862e-01
4.41840678e-01 6.51300371e-01 -3.51679981e-01 -4.34745044e-01
3.71559799e-01 3.63702148e-01 -3.83807868e-01 4.18160796e-01
8.60967040e-01 2.65181035e-01 8.01587164e-01 3.40403467e-01
-1.21400356e+00 4.27960664e-01 -1.21557796e+00 9.83515859e-01
-2.81976432e-01 3.55375558e-01 2.14608043e-01 -2.84560382e-01
1.01627612e+00 4.53209996e-01 1.37660488e-01 -1.48017037e+00
1.39400885e-01 5.45760870e-01 6.75955772e-01 -6.81671143e-01
6.62770092e-01 -7.16630876e-01 4.63456772e-02 9.21027958e-02
-7.37188458e-01 -1.19855285e-01 5.03621876e-01 -2.15979174e-01
4.11909491e-01 1.79054871e-01 6.04935884e-01 -9.14147139e-01
4.76956069e-01 -3.25325206e-02 1.11573684e+00 9.26883146e-02
-6.47524059e-01 -5.85945904e-01 7.16460049e-02 1.59335554e-01
-6.32869780e-01 -6.43781245e-01 -1.45359725e-01 8.72771800e-01
1.65799037e-01 -9.44645554e-02 -5.52186489e-01 -9.01046693e-02
-1.23689383e-01 1.77271330e+00 -3.38497221e-01 -1.40649721e-01
-2.97641814e-01 -4.80873257e-01 -3.43244880e-01 7.50127062e-02
8.73099446e-01 -4.70079809e-01 -9.47104156e-01 1.44638807e-01
2.31983632e-01 -8.76104295e-01 -2.03790534e-02 4.26568270e-01
-9.09887969e-01 -8.22518170e-01 -2.20409364e-01 -6.11880004e-01
7.81133473e-01 7.60668874e-01 8.47829759e-01 -3.57352868e-02
2.88873136e-01 3.41300070e-01 -1.24978393e-01 -8.78111362e-01
-9.00501728e-01 -2.72377074e-01 1.02091350e-01 -5.13656497e-01
6.96478605e-01 -6.37448072e-01 -6.67281628e-01 1.29795313e-01
-5.19687414e-01 -2.91035414e-01 6.66001141e-01 -3.83231401e-01
3.66702139e-01 8.60724628e-01 8.09650362e-01 -7.67538548e-01
5.92691600e-01 -4.40740347e-01 -1.92253008e-01 -2.17372641e-01
-1.27816617e+00 -6.51528180e-01 7.78016329e-01 9.61499885e-02
-1.48694277e+00 -2.70624965e-01 -2.32254956e-02 7.89859831e-01
-7.42289126e-01 4.06296104e-01 -4.67578441e-01 -1.06356390e-01
3.14923614e-01 -9.54673216e-02 4.66973968e-02 -2.54829854e-01
1.29866883e-01 2.30996728e-01 1.21413194e-01 -8.69555697e-02
1.15075684e+00 4.34835404e-01 1.74441487e-02 -1.15575969e+00
-1.26651540e-01 -5.25588274e-01 -6.55516505e-01 -7.29805410e-01
1.34548557e+00 -1.17463267e+00 -5.31943440e-01 -2.47138932e-01
-4.09073681e-02 -4.52554286e-01 4.25755009e-02 1.05164731e+00
-6.86915219e-02 3.68577987e-02 1.85547888e-01 -1.19891250e+00
2.05778897e-01 -7.20934987e-01 -8.69616792e-02 4.94697690e-01
-4.06050920e-01 -1.55432546e+00 2.72753119e-01 4.27973390e-01
6.05704069e-01 7.88335443e-01 8.98021638e-01 4.87917569e-03
-2.28946269e-01 1.32414296e-01 4.49308068e-01 3.39371443e-01
7.79646039e-01 2.59445012e-01 -4.58409578e-01 -1.32028565e-01
3.01109076e-01 1.80936038e-01 3.67499441e-01 6.89691231e-02
-2.46797562e-01 -6.67003930e-01 4.47613820e-02 -2.60497518e-02
2.46349692e+00 8.00247073e-01 7.91078746e-01 1.13878536e+00
5.19655704e-01 6.62896395e-01 9.42258596e-01 1.30830318e-01
4.90463912e-01 9.77508202e-02 4.81245875e-01 4.32578847e-02
2.73280311e-02 9.29519013e-02 8.12596083e-01 1.05533302e+00
-6.22959316e-01 3.44828486e-01 -6.80457830e-01 6.42845631e-01
-9.91485596e-01 -1.22507465e+00 -8.34031165e-01 2.06442857e+00
4.34441477e-01 1.76920384e-01 5.27178645e-01 1.90357819e-01
4.92029279e-01 1.23536503e-02 -4.02891263e-02 -1.17019701e+00
6.67979643e-02 -2.46165782e-01 8.06431234e-01 3.77687871e-01
-5.70016563e-01 6.28860891e-01 6.90991926e+00 -1.69348106e-01
-1.07865858e+00 -1.89258531e-02 2.05383271e-01 2.03296646e-01
-9.85648990e-01 1.28394261e-01 -4.62622344e-01 2.99680114e-01
1.29438913e+00 -5.33776939e-01 -5.44218943e-02 5.81334412e-01
1.30536008e+00 -5.99393964e-01 -3.53091359e-01 -5.94515614e-02
-4.11704004e-01 -5.94625413e-01 -4.98006046e-01 9.51467454e-01
9.49221194e-01 -4.38068688e-01 2.91847676e-01 1.36094227e-01
4.01803195e-01 -7.36159265e-01 9.25309896e-01 3.11950356e-01
1.46724835e-01 -1.13017702e+00 6.08003259e-01 1.20031461e-01
-6.96519017e-01 -3.00561428e-01 -5.74150622e-01 -7.76869416e-01
-5.89354783e-02 5.73944092e-01 -5.94122291e-01 4.27529097e-01
6.47383630e-01 1.27514869e-01 -7.09225297e-01 6.53142095e-01
-2.00247258e-01 1.13846552e+00 -6.81811385e-03 -6.93964213e-02
5.93796253e-01 -1.12609506e+00 3.88209850e-01 1.22876775e+00
4.49798316e-01 6.98180273e-02 -5.27932465e-01 6.48261070e-01
6.96696520e-01 5.41369855e-01 -9.86676991e-01 -6.30340874e-01
4.82772946e-01 9.92039680e-01 -9.44061697e-01 -1.25289115e-03
-9.92340207e-01 5.87309599e-01 -4.08408880e-01 3.46049160e-01
-5.68142533e-01 3.16298306e-02 7.83992589e-01 7.19664916e-02
5.04907489e-01 -3.55623245e-01 -6.34327412e-01 -3.43721241e-01
-4.37248349e-01 -8.48199427e-01 7.91937634e-02 -6.16231933e-02
-4.32782501e-01 -4.56102699e-01 -3.54388863e-01 -6.72684968e-01
2.91982949e-01 -3.23178411e-01 -6.08362854e-01 8.46628129e-01
-1.40555525e+00 -8.82566690e-01 -1.32610604e-01 8.73795226e-02
-1.63023561e-01 4.74176675e-01 9.17702913e-01 -2.76925769e-02
-7.96926081e-01 -1.28588274e-01 5.12343109e-01 -5.18944383e-01
6.72717631e-01 -1.34196723e+00 -2.46146128e-01 1.12038529e+00
-7.29323864e-01 9.05772209e-01 1.19532669e+00 -1.07131016e+00
-1.52435446e+00 -9.46008742e-01 1.27293730e+00 2.83030011e-02
5.32257736e-01 2.01189592e-01 -4.19675082e-01 5.31480193e-01
1.13667381e+00 -1.26193595e+00 1.36966848e+00 1.41737953e-01
2.41995752e-01 -2.44292334e-01 -1.30323827e+00 9.96275842e-01
8.78751934e-01 -1.72306046e-01 -2.30718717e-01 -1.91033985e-02
5.83658278e-01 8.13447118e-01 -1.45926392e+00 -2.20049262e-01
4.77856904e-01 -1.42539954e+00 6.58757746e-01 1.79410085e-01
1.31730512e-01 -2.89918929e-01 -2.44219705e-01 -1.27444637e+00
-1.14333558e+00 -5.55286109e-01 6.89726472e-01 1.41857755e+00
2.22571641e-01 -1.23230731e+00 4.26823676e-01 1.00583661e+00
-3.61371487e-01 1.72704473e-01 -6.19395912e-01 -8.92674267e-01
4.23910379e-01 3.27527262e-02 7.67054915e-01 1.39595997e+00
2.47526854e-01 -3.29839773e-02 1.63506165e-01 1.49747342e-01
8.11918378e-01 -2.49508932e-01 8.44097197e-01 -1.22270226e+00
2.97926813e-01 -5.51736116e-01 -1.71684727e-01 3.82262468e-01
-2.71998346e-01 -4.49262023e-01 -3.02636832e-01 -2.34179950e+00
2.42421985e-01 1.49931274e-02 -2.69260347e-01 -4.45798822e-02
-8.97838622e-02 -1.50482897e-02 5.34265637e-01 -8.44165906e-02
-5.04616722e-02 1.88608631e-01 1.43638515e+00 2.72398531e-01
-3.99397552e-01 -3.99008304e-01 -1.62161458e+00 5.30981362e-01
1.27554870e+00 -1.31291851e-01 -4.49304372e-01 -9.75681655e-03
6.36257410e-01 -3.12075377e-01 -1.04940042e-03 -1.36890316e+00
-1.35670006e-01 -1.29718494e+00 2.63084888e-01 -1.95060879e-01
-1.23170689e-01 -1.53584564e+00 9.58549082e-01 1.32706308e+00
1.79166179e-02 1.04978077e-01 -5.41608445e-02 1.79250672e-01
1.63136497e-01 -2.06187844e-01 5.80705345e-01 -3.42035085e-01
-8.64211559e-01 -7.03496873e-01 -8.77681553e-01 -6.53749347e-01
1.25754285e+00 -8.82613897e-01 -3.70097697e-01 -7.18867332e-02
-4.37190503e-01 -1.41066387e-01 1.46925640e+00 3.69269885e-02
-9.45836827e-02 -1.36376441e+00 -4.81448293e-01 -5.00563204e-01
-2.08934203e-01 -9.01863873e-01 6.05629906e-02 1.22921205e+00
-4.89281148e-01 8.88142109e-01 -6.15224421e-01 1.53701738e-01
-1.09524620e+00 4.11727041e-01 9.67248529e-02 4.88565087e-01
-6.11112833e-01 -6.54023588e-02 -1.47847570e-02 -9.94459167e-03
-2.82059848e-01 -1.92865431e-01 -4.09644157e-01 1.57080859e-01
3.45845193e-01 1.20810604e+00 -3.21400791e-01 -1.08299983e+00
-3.74384642e-01 4.73873198e-01 4.73678887e-01 -3.61736089e-01
1.21074557e+00 -5.36997080e-01 -4.96863246e-01 9.23448086e-01
7.56096661e-01 6.70069158e-01 -8.56746078e-01 6.41287982e-01
-1.02313921e-01 -1.08140802e+00 1.16347216e-01 -1.05539846e+00
-5.01139164e-01 1.59747288e-01 8.82067025e-01 6.48881555e-01
1.34689569e+00 -5.61634541e-01 3.43583852e-01 1.86698854e-01
3.80720437e-01 -2.11108017e+00 -1.65752336e-01 -9.04571563e-02
6.51066542e-01 -8.87031972e-01 -9.20900702e-02 -6.09698176e-01
-5.56025088e-01 8.06075454e-01 5.41548014e-01 2.20228612e-01
6.82908893e-01 8.70670006e-03 2.98569232e-01 7.95661062e-02
-4.68151927e-01 -4.90897149e-01 -4.76933181e-01 1.07676244e+00
1.08594501e+00 8.77651274e-01 -1.33043480e+00 -1.53833285e-01
-3.13911378e-01 1.26838699e-01 1.07712805e+00 7.75697529e-01
-1.08641613e+00 -1.05089581e+00 -1.01233101e+00 7.83811510e-02
-6.78108513e-01 3.78883600e-01 -9.77177620e-01 1.52053893e+00
5.49701869e-01 1.38743544e+00 4.34088632e-02 -2.49734521e-01
1.70179009e-01 3.87180224e-02 2.52729475e-01 -2.71696687e-01
-7.62361467e-01 4.32363093e-01 7.25087583e-01 -2.59381235e-01
-9.25063193e-01 -1.12404466e+00 -1.33041382e+00 -9.19007778e-01
-2.06441358e-02 3.55109781e-01 1.04062390e+00 6.84077144e-01
8.79520997e-02 3.55238974e-01 6.84396505e-01 1.77147210e-01
-3.13807189e-01 -9.76969600e-01 -9.87547517e-01 1.34775981e-01
-6.64642528e-02 -4.73441221e-02 -4.67798114e-01 4.08740379e-02] | [9.062433242797852, 6.267648220062256] |
8a91d992-dbe9-4388-aedb-4eb98d04f6b5 | why-discard-if-you-can-recycle-a-recycling | null | null | http://openaccess.thecvf.com//content/CVPR2022/html/Chen_Why_Discard_if_You_Can_Recycle_A_Recycling_Max_Pooling_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Chen_Why_Discard_if_You_Can_Recycle_A_Recycling_Max_Pooling_CVPR_2022_paper.pdf | Why Discard if You Can Recycle?: A Recycling Max Pooling Module for 3D Point Cloud Analysis | In recent years, most 3D point cloud analysis models have focused on developing either new network architectures or more efficient modules for aggregating point features from a local neighborhood. Regardless of the network architecture or the methodology used for improved feature learning, these models share one thing, which is the use of max-pooling in the end to obtain permutation invariant features. We first show that this traditional approach causes only a fraction of 3D points contribute to the permutation invariant features, and discards the rest of the points. In order to address this issue and improve the performance of any baseline 3D point classification or segmentation model, we propose a new module, referred to as the Recycling MaxPooling (RMP) module, to recycle and utilize the features of some of the discarded points. We incorporate a refinement loss that uses the recycled features to refine the prediction loss obtained from the features kept by traditional max-pooling. To the best of our knowledge, this is the first work that explores recycling of still useful points that are traditionally discarded by max-pooling. We demonstrate the effectiveness of the proposed RMP module by incorporating it into several milestone baselines and state-of-the-art networks for point cloud classification and indoor semantic segmentation tasks. We show that RPM, without any bells and whistles, consistently improves the performance of all the tested networks by using the same base network implementation and hyper-parameters. The code is provided in the supplementary material. | ['Senem Velipasalar', 'Huantao Ren', 'Burak Kakillioglu', 'Jiajing Chen'] | 2022-01-01 | null | null | null | cvpr-2022-1 | ['point-cloud-classification'] | ['computer-vision'] | [ 1.35862589e-01 3.50393876e-02 -9.33833346e-02 -3.80593002e-01
-4.47225720e-01 -5.50882518e-01 5.23628473e-01 3.48790914e-01
-6.34970665e-01 5.79036713e-01 -1.73454538e-01 -1.87927917e-01
-4.27734852e-01 -9.58127201e-01 -1.00440681e+00 -6.67302132e-01
-2.15623707e-01 2.61411130e-01 7.21538603e-01 -2.10198417e-01
3.45910251e-01 9.68040466e-01 -1.72377908e+00 2.70831555e-01
7.49899089e-01 1.27000344e+00 2.21705124e-01 2.12205857e-01
-3.65148872e-01 1.43750608e-01 -5.07391512e-01 -9.62994322e-02
6.56611145e-01 1.23414762e-01 -6.52103126e-01 -3.88531864e-01
4.99462843e-01 -1.20958768e-01 -1.67352319e-01 8.17291141e-01
4.84150916e-01 4.87256527e-01 4.32509392e-01 -1.21307981e+00
-2.73037523e-01 5.56882739e-01 -5.44874430e-01 3.25686820e-02
-9.20781791e-02 -1.31653309e-01 9.73178983e-01 -9.20744419e-01
5.54846883e-01 1.09946179e+00 1.03402710e+00 3.74238551e-01
-1.10119784e+00 -6.90805256e-01 3.25332731e-01 -2.70182658e-02
-1.55417097e+00 -1.61849648e-01 8.69093180e-01 -6.49764836e-02
1.17681015e+00 4.13910478e-01 6.31115258e-01 8.18066537e-01
-8.11317042e-02 7.32834876e-01 6.66599691e-01 -3.39047730e-01
2.60660827e-01 2.28407040e-01 5.04148245e-01 5.80005467e-01
3.45925778e-01 -1.68070674e-01 -3.84611696e-01 -1.31547600e-01
5.09199500e-01 2.52719909e-01 -2.04371333e-01 -7.92865932e-01
-8.30245554e-01 7.31904387e-01 7.77324498e-01 3.36408883e-01
-2.88578153e-01 1.96280241e-01 1.27768159e-01 -1.16331512e-02
7.94228256e-01 5.19968808e-01 -8.13174188e-01 2.66802628e-02
-1.18224120e+00 4.53597605e-01 7.28343964e-01 9.62697029e-01
1.14015687e+00 -4.77926105e-01 -2.26757258e-01 9.74535882e-01
6.67150170e-02 1.33606523e-01 1.45748526e-01 -9.40356493e-01
5.37852585e-01 1.04478598e+00 -6.89021731e-03 -8.99267554e-01
-6.04307830e-01 -6.90483510e-01 -5.16873658e-01 4.10022646e-01
2.99942493e-01 -1.23181008e-01 -1.26108384e+00 1.62478578e+00
2.24067330e-01 2.99560457e-01 -3.59869659e-01 7.05550075e-01
6.60901129e-01 4.91715103e-01 -2.00768515e-01 4.34120685e-01
1.02371359e+00 -8.47935200e-01 4.62474935e-02 1.01322420e-02
4.15591478e-01 -3.96374255e-01 9.01228845e-01 2.62642831e-01
-1.01283181e+00 -5.19399464e-01 -1.25025511e+00 -1.55853748e-01
-7.90124953e-01 -5.52002341e-02 4.97721136e-01 7.36912906e-01
-1.08489144e+00 1.27572155e+00 -1.00991023e+00 -4.96281713e-01
9.31501210e-01 7.26277649e-01 -3.76535118e-01 -1.95444208e-02
-7.70405769e-01 6.05517030e-01 2.91713685e-01 9.37859043e-02
-2.75554538e-01 -1.10930216e+00 -6.04266465e-01 2.84352839e-01
4.86865163e-01 -8.51936162e-01 8.32225740e-01 -6.98354065e-01
-1.33494878e+00 3.78530860e-01 -2.15540111e-01 -5.20504415e-01
4.25672442e-01 -4.20757204e-01 2.85556018e-01 3.26539390e-02
5.61458729e-02 8.21872771e-01 6.55837953e-01 -1.40801358e+00
-8.75369430e-01 -5.41796207e-01 3.00341547e-02 -7.47737214e-02
-1.88841343e-01 -3.10393095e-01 -7.31071353e-01 -5.67382216e-01
3.71974230e-01 -1.13968503e+00 -3.83747160e-01 -7.16643333e-02
-3.69805932e-01 -3.07319731e-01 8.88915479e-01 -1.86279327e-01
7.69474745e-01 -2.37522721e+00 1.54718272e-02 7.37460375e-01
1.37253910e-01 3.39291185e-01 -3.55877072e-01 2.93532789e-01
-1.01233445e-01 4.80079919e-01 -5.03244519e-01 -8.57468486e-01
8.67510773e-03 2.36982986e-01 -2.45181888e-01 3.35077763e-01
4.38664407e-01 8.73612463e-01 -7.35734701e-01 -7.05923289e-02
5.08793414e-01 6.56734645e-01 -8.20352316e-01 -4.03591305e-01
-1.07964531e-01 2.55897880e-01 -3.65622342e-01 4.55357939e-01
1.13465858e+00 9.74913016e-02 -3.92899156e-01 -1.94738820e-01
-2.54582435e-01 4.21985358e-01 -1.28851008e+00 1.99406886e+00
-2.79807419e-01 3.18934977e-01 -9.63758454e-02 -8.11857939e-01
8.73010039e-01 -2.08598852e-01 7.89884269e-01 -5.00021636e-01
6.73958808e-02 3.20227057e-01 -2.76066631e-01 -5.98832220e-02
7.28738189e-01 1.35313064e-01 7.17054978e-02 9.00362730e-02
2.13326693e-01 -1.63714498e-01 7.99750630e-03 -9.90200564e-02
1.28709066e+00 3.32480937e-01 -1.48766279e-01 -1.86415002e-01
4.70687151e-01 4.52991351e-02 5.78690231e-01 1.06897032e+00
-3.28364410e-02 1.05395114e+00 4.67126071e-01 -3.07002425e-01
-8.73247623e-01 -1.09779477e+00 -9.69183668e-02 9.13857937e-01
2.23527104e-01 -3.36257458e-01 -4.80906188e-01 -9.93825555e-01
4.10666138e-01 7.45239913e-01 -6.00989580e-01 -3.17427106e-02
-7.02583194e-01 -7.93364167e-01 5.57528913e-01 5.73841155e-01
4.43892092e-01 -9.20279205e-01 -4.89696205e-01 6.68290481e-02
1.07630752e-01 -8.75478089e-01 8.84940103e-03 6.85919583e-01
-1.09700656e+00 -8.24317455e-01 -5.10928810e-01 -4.60441172e-01
5.71677983e-01 4.46892917e-01 9.07487631e-01 9.36010107e-02
1.03750363e-01 -2.78389715e-02 -5.68181396e-01 -5.24596453e-01
2.73364395e-01 6.66635990e-01 -2.51093060e-01 -1.73943192e-01
3.89035821e-01 -8.64280701e-01 -5.09920895e-01 1.27751306e-01
-7.16470957e-01 -2.41768658e-01 6.25632286e-01 5.13547957e-01
7.44790792e-01 2.06413835e-01 2.81850606e-01 -7.52819419e-01
4.47182536e-01 -3.60134751e-01 -4.29728687e-01 -1.76290676e-01
-2.98473090e-01 2.43481323e-01 6.39161468e-01 -6.06412888e-02
-5.71105719e-01 2.99953043e-01 -3.71077150e-01 -5.91991007e-01
-2.65491068e-01 1.92472950e-01 -1.38689578e-01 -3.55913192e-01
3.85028124e-01 -9.45657305e-03 -1.32324457e-01 -7.90697515e-01
4.31790352e-01 3.57993782e-01 2.55012691e-01 -5.13946712e-01
9.04798985e-01 7.92571366e-01 3.71220499e-01 -8.20098221e-01
-6.84627354e-01 -7.05890536e-01 -7.48075306e-01 2.21920013e-01
7.18356669e-01 -5.91930330e-01 -5.99741757e-01 3.40037644e-01
-1.12695599e+00 -1.37005404e-01 -8.18869233e-01 3.38792384e-01
-3.98994297e-01 2.93351471e-01 -1.82045653e-01 -6.98492587e-01
-1.63106412e-01 -1.03858232e+00 1.10975718e+00 1.51985347e-01
-4.61550988e-03 -2.96869934e-01 -2.83404328e-02 9.79586039e-03
4.52634871e-01 2.53336430e-01 9.46805000e-01 -7.79488623e-01
-7.08074749e-01 -2.33497441e-01 -4.49036002e-01 7.05122888e-01
3.48995738e-02 -1.15959257e-01 -1.10713077e+00 -5.83257601e-02
-1.69411469e-02 3.76894623e-01 1.47045875e+00 4.84686106e-01
1.42103386e+00 5.46384044e-02 -5.06438434e-01 8.25455725e-01
1.37634969e+00 -8.11037514e-03 6.05576873e-01 4.39311892e-01
7.68595874e-01 6.29899800e-01 1.41806096e-01 1.99211687e-01
2.68105686e-01 5.12623608e-01 7.31768727e-01 -6.88056368e-03
-3.95929702e-02 -1.26027822e-01 6.10193191e-03 2.97981203e-01
-4.08520520e-01 -2.39081100e-01 -7.83136547e-01 6.64710402e-01
-2.00198197e+00 -7.96600878e-01 -2.61284430e-02 2.14197898e+00
2.51652598e-01 3.24042946e-01 5.82530685e-02 3.02805036e-01
5.08774221e-01 2.75149763e-01 -3.15292507e-01 -2.36316785e-01
6.74024085e-03 8.04726064e-01 1.01871407e+00 3.26390028e-01
-1.25131047e+00 8.08398485e-01 5.18371725e+00 8.62680197e-01
-9.59324002e-01 8.59958008e-02 2.33281463e-01 -4.50674921e-01
-5.04436381e-02 -5.72415292e-02 -8.12828124e-01 3.60759407e-01
5.26277304e-01 5.12054622e-01 3.60652417e-01 9.38825846e-01
2.06782594e-01 -1.56437084e-01 -1.08167040e+00 6.84756458e-01
-8.12416002e-02 -1.37097180e+00 9.82286558e-02 4.27620299e-02
5.69741666e-01 5.36879599e-01 -5.65122627e-03 2.40281805e-01
7.64017180e-02 -7.57242978e-01 8.64671409e-01 5.91244936e-01
1.19460560e-01 -1.03959155e+00 1.01139402e+00 2.96756327e-01
-1.03739405e+00 -3.26426744e-01 -5.04116237e-01 1.51116624e-02
-3.21791880e-02 8.27198684e-01 -8.08394551e-01 9.06560957e-01
1.00277066e+00 7.08916306e-01 -7.68679142e-01 1.43067515e+00
-8.26248080e-02 3.97347897e-01 -8.88084829e-01 -2.77483910e-02
4.39175576e-01 -2.52305605e-02 7.97370255e-01 1.13708711e+00
5.36386371e-01 -3.93457353e-01 -1.89013347e-01 9.83030140e-01
-2.73499012e-01 -2.07097623e-02 -3.98117542e-01 4.68480349e-01
3.73646617e-01 1.17979288e+00 -9.76545990e-01 -3.77246886e-02
-3.95256251e-01 7.19102740e-01 2.95585215e-01 2.08710998e-01
-5.96358657e-01 -6.76181614e-01 8.04879248e-01 3.41740042e-01
8.85019064e-01 -4.26361501e-01 -7.17829525e-01 -8.36079121e-01
3.04346293e-01 -3.03840905e-01 3.74423969e-03 -4.74702835e-01
-1.28206837e+00 4.55784678e-01 1.47125497e-01 -1.10178030e+00
2.12610811e-01 -5.69227934e-01 -5.89905560e-01 8.99062395e-01
-1.71889102e+00 -1.02731073e+00 -3.69149089e-01 3.54375899e-01
2.83175230e-01 1.42301917e-01 3.64098966e-01 4.21025097e-01
-4.69349831e-01 4.47709560e-01 8.37828517e-02 1.33466721e-03
4.17547524e-01 -1.25590324e+00 6.66681051e-01 7.51457632e-01
9.50633362e-02 6.78729713e-01 4.32413638e-01 -5.43796241e-01
-8.45414698e-01 -1.25680029e+00 8.14939260e-01 -5.47247708e-01
2.92098373e-01 -6.95665717e-01 -9.26837802e-01 4.49617118e-01
-1.65438697e-01 1.59095228e-01 2.53922343e-01 2.13408604e-01
-1.41265675e-01 -3.08469146e-01 -1.15481544e+00 3.98952901e-01
1.35655415e+00 -6.94123209e-02 -5.95517635e-01 -1.54928952e-01
9.26605642e-01 -2.19268560e-01 -5.11359155e-01 7.53003478e-01
4.71819073e-01 -1.12224782e+00 1.30205703e+00 -3.12178671e-01
1.68520898e-01 -4.84214485e-01 -2.32989952e-01 -1.18580723e+00
-3.12904298e-01 -2.75707006e-01 1.45250425e-01 1.24515927e+00
5.35169244e-01 -7.21612096e-01 1.28483999e+00 4.15451318e-01
-5.53936303e-01 -8.67645860e-01 -1.28785491e+00 -8.77610207e-01
1.19089708e-01 -6.91873014e-01 1.00336194e+00 5.67429304e-01
-5.40123343e-01 -1.51912808e-01 5.01023196e-02 2.38931060e-01
4.57909256e-01 -2.71382201e-02 8.56014669e-01 -1.43626142e+00
6.88323751e-02 -6.02881134e-01 -4.87160146e-01 -1.05932820e+00
-6.39287084e-02 -1.05250037e+00 1.08787343e-01 -1.58785415e+00
-1.86071157e-01 -7.89813817e-01 -5.78105748e-01 6.22253001e-01
2.71521462e-03 3.23356271e-01 5.28406322e-01 9.73138064e-02
-5.24944067e-01 4.93620574e-01 1.00074124e+00 -6.81703305e-03
-8.06609690e-01 2.82759577e-01 -7.45445132e-01 8.58767331e-01
9.14709985e-01 -8.02992284e-01 -1.60263300e-01 -4.33275521e-01
1.84225261e-01 -8.14444959e-01 7.31825948e-01 -1.41884363e+00
2.82325596e-01 1.56923801e-01 5.12320638e-01 -1.02446496e+00
5.39243221e-01 -1.08919036e+00 1.79415923e-02 1.42548457e-01
8.37823153e-02 -1.68848068e-01 5.50817132e-01 4.47606623e-01
1.46646611e-02 -3.71829689e-01 5.31798601e-01 -1.29974887e-01
-5.29816031e-01 2.02342272e-01 -4.48125787e-02 -4.16817188e-01
8.49439979e-01 -5.01794696e-01 -3.81325752e-01 1.62499130e-01
-6.39496505e-01 1.07559681e-01 4.90854204e-01 5.06664276e-01
4.67169225e-01 -1.04157841e+00 -4.60338742e-01 3.56586963e-01
-1.40100166e-01 4.38266546e-01 1.61419898e-01 7.00454891e-01
-4.18205678e-01 3.41838509e-01 -1.57685116e-01 -6.20363414e-01
-9.64066982e-01 3.46311480e-01 3.13856840e-01 -3.09704453e-01
-6.79205894e-01 7.30471849e-01 -9.17838812e-02 -6.41178548e-01
1.29857704e-01 -8.36362362e-01 -9.71224010e-02 2.21820980e-01
-7.25643933e-02 6.03158593e-01 5.27578354e-01 -3.70590001e-01
-6.83911264e-01 7.07820296e-01 -1.57400649e-02 8.51106271e-02
1.84111595e+00 6.21873364e-02 -1.38707191e-01 2.61306375e-01
1.23365664e+00 1.53530970e-01 -1.22021401e+00 3.07804178e-02
9.74697527e-03 -4.21367556e-01 1.83404788e-01 -6.27004445e-01
-1.23735619e+00 6.35955632e-01 5.67713141e-01 2.08401054e-01
1.06623185e+00 8.95790830e-02 7.34078765e-01 4.19454038e-01
3.73017371e-01 -1.01377177e+00 -4.19627368e-01 6.94457114e-01
6.54071927e-01 -8.83968949e-01 1.02694795e-01 -6.41120911e-01
2.59281253e-03 8.81410003e-01 3.14308852e-01 -6.58024788e-01
9.91051137e-01 1.51990399e-01 -3.68278921e-01 -1.57600671e-01
-2.38737062e-01 -3.98221850e-01 2.82027900e-01 5.34597218e-01
-1.44437268e-01 -1.94579944e-01 -1.97007328e-01 8.33573639e-01
-2.83685416e-01 1.23718001e-01 9.72984582e-02 9.55301344e-01
-5.16058743e-01 -1.16098046e+00 -3.01706880e-01 6.31042600e-01
-3.86875868e-01 -6.98123723e-02 -3.95794898e-01 9.50354099e-01
6.19120121e-01 7.72013664e-01 3.21880251e-01 -5.64926207e-01
7.72961676e-01 2.41114900e-01 3.71771932e-01 -6.10585690e-01
-9.53523695e-01 -2.73390085e-01 -2.44458020e-01 -7.39139140e-01
-5.67477405e-01 -7.08808064e-01 -1.14828992e+00 -2.62516290e-01
-4.81560588e-01 -9.02740359e-02 9.25175428e-01 7.34966278e-01
7.98964620e-01 6.75045133e-01 4.52312738e-01 -1.34042335e+00
-2.04233646e-01 -7.79535294e-01 -4.70729887e-01 2.24459499e-01
3.75689596e-01 -7.81689465e-01 -4.08238113e-01 -6.49769425e-01] | [7.945199012756348, -3.4321563243865967] |
95722c57-956b-42d3-9181-97108a571c27 | multi-frames-temporal-abnormal-clues-learning | 2208.04076 | null | https://arxiv.org/abs/2208.04076v1 | https://arxiv.org/pdf/2208.04076v1.pdf | Multi-Frames Temporal Abnormal Clues Learning Method for Face Anti-Spoofing | Face anti-spoofing researches are widely used in face recognition and has received more attention from industry and academics. In this paper, we propose the EulerNet, a new temporal feature fusion network in which the differential filter and residual pyramid are used to extract and amplify abnormal clues from continuous frames, respectively. A lightweight sample labeling method based on face landmarks is designed to label large-scale samples at a lower cost and has better results than other methods such as 3D camera. Finally, we collect 30,000 live and spoofing samples using various mobile ends to create a dataset that replicates various forms of attacks in a real-world setting. Extensive experiments on public OULU-NPU show that our algorithm is superior to the state of art and our solution has already been deployed in real-world systems servicing millions of users. | ['Jin Gao', 'Jiarong He', 'Rongyu Zhang', 'Heng Cong'] | 2022-08-08 | null | null | null | null | ['face-anti-spoofing'] | ['computer-vision'] | [ 1.18156388e-01 -6.86605275e-01 -1.74954534e-01 -1.17966570e-01
-1.43071875e-01 -4.09717143e-01 4.18342590e-01 -5.38093388e-01
-1.35274798e-01 4.89593238e-01 -4.89114597e-02 -1.56685099e-01
2.20802575e-01 -7.15410233e-01 -1.67404041e-01 -7.79273331e-01
-3.89724016e-01 -9.26041752e-02 3.40496957e-01 -1.10807270e-01
2.06851780e-01 9.57983017e-01 -1.32550359e+00 1.42209396e-01
2.88859814e-01 1.10045063e+00 -3.57793570e-01 4.33122486e-01
2.50411659e-01 4.15230662e-01 -7.49236822e-01 -7.30572581e-01
5.42494893e-01 -2.60516495e-01 -3.82077873e-01 2.17849329e-01
3.07953447e-01 -4.72863197e-01 -6.56146526e-01 1.32271743e+00
9.45448935e-01 -3.10723841e-01 3.09329122e-01 -1.56380796e+00
-6.47803485e-01 4.60941717e-03 -1.18998432e+00 6.66250110e-01
8.07029843e-01 7.08477348e-02 1.00078851e-01 -1.03976059e+00
4.80896711e-01 1.79123998e+00 8.90922248e-01 6.30301237e-01
-7.76383877e-01 -1.28041136e+00 -3.96925002e-01 6.75926805e-01
-1.57715654e+00 -8.54304969e-01 9.22594070e-01 8.00233930e-02
4.24340189e-01 2.32580841e-01 6.73535407e-01 1.30224967e+00
2.05405384e-01 4.74522412e-01 1.13610876e+00 -2.19995856e-01
-2.01963127e-01 -3.03231329e-02 -1.61239371e-01 1.07431710e+00
4.08921361e-01 1.75338343e-01 -6.74508572e-01 -5.27783930e-01
8.21554005e-01 3.94028664e-01 -4.75685537e-01 1.20930098e-01
-8.06156993e-01 7.97402143e-01 3.70023176e-02 2.48516247e-01
-3.06654930e-01 -5.11546880e-02 2.24567801e-01 4.27596241e-01
5.72135091e-01 -4.96256381e-01 -9.99727845e-02 3.05752665e-01
-9.17102873e-01 -1.55159235e-02 6.51651025e-01 7.02937841e-01
4.56261575e-01 1.33839637e-01 6.17532283e-02 7.64874816e-01
6.68292224e-01 9.32538748e-01 4.95946378e-01 -9.03766215e-01
1.86908454e-01 4.33118343e-01 -2.16488466e-01 -1.83491874e+00
-1.75663561e-01 -1.43115014e-01 -8.16434860e-01 1.57603230e-02
1.71390995e-01 -2.32708007e-01 -7.32760370e-01 1.41488183e+00
7.74175584e-01 1.47720170e+00 -6.01908751e-02 5.87864101e-01
6.49431407e-01 5.43867469e-01 -4.45138097e-01 -7.48512924e-01
1.63684833e+00 -5.68000913e-01 -8.86831045e-01 1.75035641e-01
1.59950688e-01 -1.14706075e+00 5.36445916e-01 6.17320895e-01
-7.17502117e-01 -3.90443772e-01 -1.05638301e+00 4.40183401e-01
-1.95588455e-01 -6.88400716e-02 5.77653408e-01 1.52243507e+00
-1.13766372e+00 4.72435445e-01 -7.41823018e-01 -4.77715611e-01
9.00645554e-01 5.84454179e-01 -7.08886683e-01 -3.17201257e-01
-9.33547258e-01 3.22504908e-01 -3.13316546e-02 2.09455356e-01
-1.01156592e+00 -1.89450413e-01 -7.03804135e-01 -2.65851408e-01
2.77220935e-01 -1.37567788e-01 7.99847603e-01 -5.07890701e-01
-1.54348147e+00 9.46726561e-01 -4.83340800e-01 -2.51238436e-01
2.86129981e-01 4.31904718e-02 -1.12710702e+00 5.61816514e-01
-2.65821032e-02 1.71970770e-01 1.35779178e+00 -8.69666219e-01
-5.63699782e-01 -8.01630914e-01 -3.41266483e-01 -4.21663582e-01
-7.93550372e-01 5.63834846e-01 -4.23051983e-01 -7.85855472e-01
2.63130099e-01 -9.03813243e-01 8.25338885e-02 3.64407822e-02
-3.00431252e-01 -3.17879319e-01 1.81478024e+00 -6.85911179e-01
1.12242091e+00 -2.17179251e+00 -4.12512422e-01 4.19879287e-01
1.75214455e-01 8.14988077e-01 -2.13771686e-01 1.82210460e-01
-8.18647742e-02 1.63714662e-01 -4.70038243e-02 -3.38078141e-01
-4.71537113e-01 -5.78404441e-02 -2.12514445e-01 1.18404961e+00
1.50475223e-02 3.93280834e-01 -9.54623580e-01 -8.46405208e-01
1.75259098e-01 8.60442042e-01 -5.83734751e-01 7.54772360e-03
4.87418413e-01 4.83037889e-01 -4.23751116e-01 1.17409420e+00
1.32285583e+00 2.17371788e-02 8.08118135e-02 -7.09657744e-02
4.61821079e-01 -3.67327750e-01 -1.46844780e+00 1.19604838e+00
2.13424545e-02 5.85483134e-01 3.31751287e-01 -7.53455997e-01
8.37727427e-01 6.89170003e-01 4.98293072e-01 -2.36919805e-01
5.27886152e-01 1.17100872e-01 -2.67691702e-01 -8.65211964e-01
3.95981781e-02 3.27919960e-01 3.67090940e-01 5.13616681e-01
1.42169639e-01 3.64209801e-01 -1.02726715e-02 -1.31573537e-02
1.29327285e+00 -2.28282660e-01 1.24318495e-01 -1.64544225e-01
1.10051966e+00 -6.88994288e-01 9.97040331e-01 2.63736933e-01
-7.85527408e-01 2.35698074e-01 1.33291245e-01 -4.41201597e-01
-3.96309078e-01 -7.99244165e-01 -2.32125759e-01 6.24484241e-01
4.06440139e-01 -5.06993532e-01 -9.02784169e-01 -1.01788592e+00
4.99435142e-02 -1.87912107e-01 -2.86571741e-01 -5.00312250e-04
-8.36068928e-01 -8.62347424e-01 8.70715976e-01 -2.17235889e-02
1.05895245e+00 -1.04895699e+00 -3.31621319e-01 -5.58949355e-03
-1.57063738e-01 -1.33533204e+00 -5.84004402e-01 -9.15753543e-01
-6.56581402e-01 -1.45700634e+00 -6.43069148e-01 -1.05872214e+00
7.45545030e-01 8.31784666e-01 5.49899936e-01 4.41518098e-01
-5.11422932e-01 2.19294116e-01 -3.03568274e-01 -4.39430147e-01
-7.17707649e-02 -4.77420956e-01 6.72369301e-01 6.36571944e-01
6.52851522e-01 -6.62892222e-01 -7.10030735e-01 5.28180122e-01
-7.04799175e-01 -7.53445148e-01 1.38711050e-01 6.54386282e-01
-1.05812758e-01 3.28926533e-01 6.44061983e-01 -6.80834413e-01
6.03174686e-01 -6.54723763e-01 -4.17424023e-01 1.70015126e-01
-4.07652736e-01 -4.66709584e-01 4.11097795e-01 -6.49695218e-01
-9.09414470e-01 -1.83833063e-01 -1.08960636e-01 -6.18461132e-01
-9.92261767e-02 -1.30474970e-01 -1.47775456e-01 -7.90934861e-01
4.00873840e-01 1.64947331e-01 1.23612851e-01 -4.78220910e-01
3.64155918e-02 1.00958490e+00 5.32426655e-01 -1.55583411e-01
1.17144454e+00 8.58432353e-01 2.32670009e-01 -1.09467006e+00
-3.49551030e-02 -3.96244824e-01 -1.91139087e-01 -3.61054331e-01
3.81365746e-01 -8.23393166e-01 -1.34641123e+00 9.42934811e-01
-1.35062027e+00 5.69128931e-01 4.52845991e-01 4.93559539e-01
-4.56594229e-02 9.19318616e-01 -9.07326400e-01 -1.06078815e+00
-3.79040003e-01 -1.23687363e+00 9.71560240e-01 6.90220714e-01
3.48188639e-01 -6.95722222e-01 -1.16093650e-01 9.59582478e-02
3.43307287e-01 3.90274674e-01 9.22073722e-02 -4.56345052e-01
-3.64993274e-01 -3.19608539e-01 -2.09571764e-01 2.84689903e-01
4.83876854e-01 2.55904555e-01 -1.13261783e+00 -7.23906636e-01
4.98584598e-01 -9.31672454e-02 5.99685013e-01 4.13714275e-02
1.24656677e+00 -5.32003284e-01 -6.04495525e-01 7.16752172e-01
1.12315857e+00 4.70881373e-01 6.20411932e-01 -1.77054122e-01
3.90578419e-01 6.32125735e-01 3.81196737e-01 6.26135409e-01
-4.80998755e-02 5.04995704e-01 3.97654176e-01 1.92174837e-01
-3.84765416e-02 -5.73315434e-02 6.05666935e-01 5.77208221e-01
-6.79432526e-02 -2.33240694e-01 -5.93468010e-01 1.79261640e-01
-1.39194357e+00 -1.35318446e+00 1.81316271e-01 1.95651209e+00
4.78798628e-01 -7.78945759e-02 1.86099306e-01 5.20757675e-01
1.16454291e+00 4.32420492e-01 -1.82136774e-01 9.01137292e-02
-9.43241790e-02 3.68183583e-01 4.12066340e-01 3.16615641e-01
-1.04831779e+00 1.00954163e+00 6.57546806e+00 1.13891387e+00
-1.33965766e+00 4.82961237e-01 7.72041976e-01 1.13731675e-01
3.52012873e-01 -4.04129386e-01 -9.35946405e-01 7.76351273e-01
8.34914625e-01 8.05727541e-02 5.03605902e-01 7.45584011e-01
1.74284399e-01 2.24450812e-01 -4.68988538e-01 1.59134269e+00
4.88014042e-01 -1.23578048e+00 -2.55221874e-01 2.49809220e-01
5.51979423e-01 -3.06335419e-01 2.05396906e-01 -2.06788361e-01
2.77772158e-01 -8.96540940e-01 8.89907256e-02 1.35183841e-01
8.08557749e-01 -7.85726368e-01 7.44140506e-01 1.28257275e-01
-1.49914491e+00 -2.05922514e-01 -3.92213523e-01 1.98677406e-01
2.03298166e-01 5.54350615e-01 -7.54423797e-01 4.77383435e-01
8.18774164e-01 6.93890750e-01 -5.70567667e-01 9.78345871e-01
-5.64305931e-02 7.86028326e-01 -5.98136783e-01 -4.38466929e-02
-1.01030260e-01 -8.03249776e-02 6.85094237e-01 1.10676420e+00
5.61915398e-01 2.03481644e-01 -1.59173254e-02 9.44947526e-02
-2.89407283e-01 5.67155965e-02 -9.48876083e-01 1.65439337e-01
8.33660543e-01 1.47093141e+00 -8.04934323e-01 -2.47623190e-01
-3.48071337e-01 1.03346121e+00 -1.72183722e-01 2.16215000e-01
-9.10779059e-01 -5.28017938e-01 7.40140438e-01 -2.51012534e-01
1.62824064e-01 -1.84383646e-01 3.16706002e-01 -1.30562317e+00
-1.18184481e-02 -9.60877538e-01 4.66109693e-01 -2.57403165e-01
-1.14168859e+00 6.53438509e-01 -3.97559166e-01 -1.12007749e+00
-1.16382845e-01 -7.47343063e-01 -6.88720226e-01 4.97375220e-01
-1.52770901e+00 -7.54663348e-01 -4.37282294e-01 1.16431105e+00
4.17702615e-01 -5.99061489e-01 6.64044023e-01 6.61441505e-01
-9.92241740e-01 8.01981151e-01 -2.50004321e-01 4.22595292e-01
6.11080468e-01 -3.02319735e-01 3.28298450e-01 1.14052963e+00
2.22430423e-01 7.22458184e-01 4.84072238e-01 -7.53056586e-01
-1.63576484e+00 -8.91876400e-01 7.61462569e-01 -1.84855953e-01
3.03677917e-01 -1.95227012e-01 -4.97642606e-01 5.36091328e-01
2.57042557e-01 7.62274325e-01 6.96059525e-01 -6.69772029e-01
-4.67165142e-01 -4.79372948e-01 -2.08367610e+00 3.95509332e-01
1.33049428e+00 -5.41575789e-01 -2.66375899e-01 4.50416654e-01
3.87772024e-01 -1.42748713e-01 -5.57597339e-01 5.09339213e-01
6.99491084e-01 -1.20319569e+00 1.20700097e+00 -3.40785742e-01
-4.23831940e-01 -3.67326289e-01 -2.19396845e-01 -7.18094051e-01
-1.81074306e-01 -1.20310891e+00 -5.45528889e-01 1.54109609e+00
-2.82618314e-01 -1.10142660e+00 1.09322262e+00 -2.21868634e-01
4.96200651e-01 -7.10283041e-01 -1.29038596e+00 -7.68481255e-01
-8.49963546e-01 -3.15957934e-01 9.62264717e-01 1.00606084e+00
-2.42479399e-01 -1.06499821e-01 -6.16137445e-01 4.49939579e-01
1.06040096e+00 -4.52699900e-01 6.25400841e-01 -1.29789066e+00
1.46929011e-01 -8.29587951e-02 -9.13678527e-01 -8.04065466e-01
3.00636411e-01 -5.32244623e-01 -5.67708433e-01 -5.04334033e-01
-4.44274992e-02 -2.18664005e-01 -1.97575778e-01 2.73885936e-01
3.29017378e-02 1.08194184e+00 -2.78486928e-04 1.46620914e-01
-1.45505697e-01 2.80368745e-01 1.01144516e+00 -1.08280011e-01
1.21431813e-01 8.25925823e-03 -3.94248635e-01 9.69842374e-01
6.51687801e-01 -4.92999315e-01 -2.14847773e-01 -1.51941568e-01
-4.42053676e-01 8.66397917e-02 2.90121466e-01 -1.28791106e+00
2.38860950e-01 2.82073859e-03 9.41241205e-01 -5.11537075e-01
5.27610421e-01 -1.11457133e+00 7.42095858e-02 8.46073091e-01
3.45431417e-01 4.07126755e-01 -2.42796123e-01 5.60671449e-01
-6.18363321e-02 4.76431213e-02 1.09124112e+00 4.84885182e-03
-5.27590215e-01 9.65251088e-01 -1.58227999e-02 -3.66545677e-01
1.43652368e+00 -3.08735043e-01 -5.59490621e-01 -1.85779914e-01
-2.78281003e-01 -1.05447233e-01 2.48329923e-01 3.56754363e-01
1.07076287e+00 -1.64616847e+00 -7.25968957e-01 7.81929135e-01
-4.00404036e-01 -6.59990132e-01 4.76234481e-02 7.44444907e-01
-8.03480089e-01 2.72079706e-01 -2.22659364e-01 -6.98247790e-01
-1.76915514e+00 5.98161757e-01 4.33842279e-03 9.89278257e-02
-4.49942142e-01 8.32522571e-01 -5.49457431e-01 -5.68073951e-02
2.62426823e-01 4.72184539e-01 -4.05229896e-01 6.31572753e-02
1.11403525e+00 6.94263637e-01 -2.32888520e-01 -1.20695281e+00
-7.11312294e-01 8.09905887e-01 1.46322817e-01 -6.42767027e-02
1.12404191e+00 -2.63212651e-01 -3.51724684e-01 -3.03721279e-01
1.55049372e+00 3.19174439e-01 -9.30119038e-01 -3.03145111e-01
-1.30601540e-01 -1.14191926e+00 -1.69974491e-01 1.52621463e-01
-1.59989512e+00 6.88592732e-01 1.20154309e+00 3.68219733e-01
1.31751192e+00 -2.99463421e-01 1.22965300e+00 2.65128911e-01
7.45614290e-01 -4.92913157e-01 3.84551555e-01 1.22972243e-01
4.23221648e-01 -1.11346292e+00 -1.41412765e-01 -8.30649078e-01
5.58608072e-03 1.10280442e+00 4.18552488e-01 -3.29788983e-01
1.07931745e+00 3.31838936e-01 7.43316785e-02 -8.15290550e-04
-2.06396550e-01 1.37176380e-01 -2.42637143e-01 9.31305468e-01
-2.36512572e-02 -2.54738420e-01 -2.49741450e-01 1.00500502e-01
-1.04357526e-01 1.67128533e-01 2.01473802e-01 1.01269817e+00
-3.04812789e-01 -1.31964207e+00 -7.97568738e-01 3.00965220e-01
-1.00569880e+00 2.77616560e-01 3.36582847e-02 3.86472642e-01
4.01550382e-01 1.40562439e+00 -1.03032507e-01 -8.37832749e-01
-7.29490593e-02 -1.44107595e-01 5.42384505e-01 -2.76264995e-01
-2.03182846e-01 1.75507495e-03 -2.67131031e-01 -8.43231380e-01
-6.76194966e-01 -6.29071116e-01 -8.85023594e-01 -9.17377353e-01
-3.92218918e-01 1.88171640e-01 5.67584693e-01 6.71521246e-01
4.46462393e-01 -8.77567828e-02 1.48638535e+00 -1.03778362e+00
-2.70621330e-01 -8.37922931e-01 -7.16924250e-01 3.02031457e-01
6.81710482e-01 -8.91747355e-01 -5.15864491e-01 7.12007284e-02] | [13.006760597229004, 1.170379877090454] |
ee3d3a1a-b165-4427-813c-1d918f2f9cb8 | exploiting-expert-guided-symmetry-detection | 2112.09943 | null | https://arxiv.org/abs/2112.09943v3 | https://arxiv.org/pdf/2112.09943v3.pdf | Data Augmentation through Expert-guided Symmetry Detection to Improve Performance in Offline Reinforcement Learning | Offline estimation of the dynamical model of a Markov Decision Process (MDP) is a non-trivial task that greatly depends on the data available in the learning phase. Sometimes the dynamics of the model is invariant with respect to some transformations of the current state and action. Recent works showed that an expert-guided pipeline relying on Density Estimation methods as Deep Neural Network based Normalizing Flows effectively detects this structure in deterministic environments, both categorical and continuous-valued. The acquired knowledge can be exploited to augment the original data set, leading eventually to a reduction in the distributional shift between the true and the learned model. Such data augmentation technique can be exploited as a preliminary process to be executed before adopting an Offline Reinforcement Learning architecture, increasing its performance. In this work we extend the paradigm to also tackle non-deterministic MDPs, in particular, 1) we propose a detection threshold in categorical environments based on statistical distances, and 2) we show that the former results lead to a performance improvement when solving the learned MDP and then applying the optimized policy in the real environment. | ['Caroline P. C. Chanel', 'Nicolas Drougard', 'Giorgio Angelotti'] | 2021-12-18 | null | null | null | null | ['symmetry-detection'] | ['computer-vision'] | [ 2.14672774e-01 2.74867624e-01 3.47445644e-02 -1.94568649e-01
-2.57398456e-01 -6.70711637e-01 9.88960028e-01 5.87665319e-01
-8.06315780e-01 8.50799084e-01 -1.11016557e-02 -1.93833381e-01
-4.03818250e-01 -8.10400605e-01 -7.23340571e-01 -9.00876701e-01
-2.72851527e-01 1.08448315e+00 3.49503547e-01 7.35629201e-02
2.08896279e-01 9.39836025e-01 -1.81429577e+00 -3.57471257e-02
8.85632515e-01 7.79629350e-01 4.55746323e-01 8.68914306e-01
-2.76751578e-01 8.41099858e-01 -3.73291761e-01 3.45290117e-02
3.40516210e-01 -3.60564202e-01 -7.33136952e-01 -3.46136541e-04
-1.23370990e-01 -6.63617849e-02 -7.05358200e-03 1.15374625e+00
2.63837457e-01 3.50416064e-01 9.13797855e-01 -1.05271435e+00
1.46779269e-01 7.09278166e-01 -2.67479233e-02 6.48684129e-02
2.05734178e-01 3.83642346e-01 5.41334569e-01 -3.51925522e-01
7.94036925e-01 1.22405958e+00 4.45907384e-01 3.23075742e-01
-1.53099000e+00 -1.79270074e-01 2.58843809e-01 5.90571463e-01
-1.18016934e+00 -2.65327692e-01 6.79398417e-01 -6.68779433e-01
6.59276426e-01 -7.74512365e-02 9.65797544e-01 1.33803689e+00
1.66264042e-01 5.33497334e-01 1.55226302e+00 -4.24641222e-01
9.94632959e-01 3.57271075e-01 -1.68479130e-01 3.66110682e-01
6.36622086e-02 2.91314572e-01 -1.61862999e-01 1.92705870e-01
6.01791859e-01 -2.19972968e-01 1.72995795e-02 -6.94410801e-01
-9.37820733e-01 5.17517269e-01 1.61652997e-01 7.13111162e-01
-7.39520967e-01 3.40560302e-02 3.79455596e-01 4.51288402e-01
3.12292844e-01 4.62045640e-01 -3.53881001e-01 -5.57192087e-01
-9.86782670e-01 3.04079294e-01 1.11932218e+00 3.17576140e-01
1.14700305e+00 -1.54214233e-01 -1.77304506e-01 1.17072277e-01
1.12558059e-01 5.14857471e-01 4.70348150e-01 -6.88404322e-01
2.24979520e-01 6.20463610e-01 4.46384490e-01 -8.08124304e-01
-4.33851928e-01 -4.37225729e-01 -6.63749576e-01 5.85935414e-01
9.53538716e-01 -1.98682457e-01 -7.93422222e-01 1.82716584e+00
6.42502546e-01 3.19608659e-01 1.36010677e-01 5.32324731e-01
-2.86328077e-01 4.63467181e-01 8.45797211e-02 -2.66116381e-01
9.64853406e-01 -3.11835408e-01 -6.35880113e-01 1.43115837e-02
5.12743115e-01 -4.22336429e-01 1.00818932e+00 6.40651584e-01
-6.42534077e-01 -5.10766149e-01 -8.61204922e-01 4.80880767e-01
-5.12728989e-01 1.04217373e-01 -4.81027104e-02 6.23629451e-01
-1.05144107e+00 1.10893095e+00 -1.29293907e+00 -5.29195309e-01
1.93698168e-01 3.05767417e-01 -2.51635969e-01 8.90276358e-02
-1.10495818e+00 1.12772524e+00 9.52976465e-01 1.48358852e-01
-1.29405379e+00 -3.48360181e-01 -5.79845548e-01 1.57889038e-01
7.05444336e-01 -4.62016553e-01 9.30753827e-01 -1.12529933e+00
-1.85640013e+00 3.75675380e-01 -1.26432344e-01 -8.63868058e-01
1.18794274e+00 -1.82246059e-01 4.18320969e-02 1.86941832e-01
-2.10890070e-01 3.75734389e-01 1.23037040e+00 -1.18428636e+00
-6.91084504e-01 -3.69727224e-01 1.22563802e-01 1.18454918e-01
-3.08358312e-01 -4.86995548e-01 2.43690416e-01 -2.37514526e-01
-2.53519982e-01 -1.01379442e+00 -3.62015873e-01 -3.49331707e-01
-3.26067001e-01 -1.82732508e-01 7.50878811e-01 -6.07517004e-01
8.54174316e-01 -1.78179908e+00 5.50712526e-01 4.85040814e-01
-1.17262356e-01 2.39893153e-01 1.87581759e-02 5.07881105e-01
7.05438107e-02 -2.33630344e-01 -3.95421803e-01 -4.31462258e-01
5.89581244e-02 3.64828616e-01 -2.55442083e-01 5.94166517e-01
3.08796912e-01 4.34282303e-01 -1.00485694e+00 -3.94695491e-01
4.70458984e-01 3.33079666e-01 -2.81687707e-01 3.52705210e-01
-5.42104363e-01 1.10299528e+00 -2.51111925e-01 -2.26319730e-02
6.09061480e-01 2.65258104e-01 2.71148652e-01 1.11650258e-01
-4.68074173e-01 4.50463258e-02 -1.61039305e+00 1.56962597e+00
-8.51707160e-01 3.21605891e-01 -2.34474577e-02 -1.32073271e+00
1.14696014e+00 1.02994539e-01 6.09669983e-01 -2.01537758e-01
2.37167194e-01 2.38658890e-01 2.26615295e-01 -4.90713745e-01
3.50587308e-01 -7.54645243e-02 2.61273354e-01 4.75696564e-01
3.29686880e-01 -1.10933997e-01 4.81746107e-01 -2.72913367e-01
1.10294473e+00 5.18224716e-01 3.89412999e-01 -5.06745696e-01
8.83299768e-01 -1.31206796e-01 2.36435115e-01 9.39005017e-01
9.20494348e-02 -5.95321134e-02 8.95699561e-01 -4.66290802e-01
-1.03978646e+00 -1.02389479e+00 -6.29023090e-03 8.38386953e-01
-7.29016140e-02 6.59749880e-02 -8.62547338e-01 -7.62222886e-01
-2.16490477e-01 8.78492832e-01 -8.90321612e-01 -4.21071321e-01
-5.29829800e-01 -4.13377345e-01 1.40661180e-01 3.46412271e-01
3.17913622e-01 -1.15990078e+00 -1.03821647e+00 3.91229659e-01
1.33727863e-01 -1.03934455e+00 3.00778419e-01 7.16567874e-01
-9.77910340e-01 -9.73623335e-01 -5.96384168e-01 -6.85093030e-02
4.64362949e-01 -6.34047329e-01 7.16996968e-01 -6.07133389e-01
8.87870416e-02 5.30188382e-01 -4.14472282e-01 -3.91760558e-01
-1.00661600e+00 3.29627246e-01 1.95467621e-01 3.46891612e-01
2.29108214e-01 -7.89947212e-01 -3.70165586e-01 1.82665542e-01
-1.03305638e+00 -2.08006367e-01 7.06710577e-01 5.60496271e-01
4.89435434e-01 2.53133744e-01 3.62471730e-01 -6.83331788e-01
5.52805185e-01 -2.77023256e-01 -9.96651053e-01 2.19067559e-01
-5.27679980e-01 8.85276437e-01 9.06433463e-01 -6.74309075e-01
-1.21504486e+00 4.34976697e-01 -2.27781236e-01 -5.26498258e-01
-7.14397073e-01 4.29804534e-01 -2.04515666e-01 2.47595608e-01
6.73479319e-01 3.41152489e-01 1.28953569e-02 -5.95022619e-01
4.22147602e-01 3.10192645e-01 3.35137457e-01 -8.09125006e-01
7.84913540e-01 7.09448755e-01 4.01693821e-01 -7.06329346e-01
-4.56984848e-01 -3.44069570e-01 -1.29334342e+00 -3.99176836e-01
7.74869978e-01 -3.66114348e-01 -7.90442407e-01 4.60238069e-01
-1.06828284e+00 -7.63347268e-01 -9.76766348e-01 6.29390478e-01
-8.03180039e-01 3.66222441e-01 -1.92362577e-01 -1.23103654e+00
1.21643715e-01 -1.17546070e+00 9.40460980e-01 8.43556225e-02
3.05931792e-02 -1.13392961e+00 5.72706342e-01 -3.36737573e-01
2.97480792e-01 1.25141174e-01 8.77516091e-01 -1.05398977e+00
-3.84566545e-01 -1.59932405e-01 2.54470259e-01 5.50473392e-01
3.66981253e-02 4.54479605e-02 -9.74642992e-01 -9.32061896e-02
2.51738697e-01 -1.36662303e-02 8.93803596e-01 2.73281813e-01
8.40818882e-01 -1.43632531e-01 -8.52464810e-02 1.42259911e-01
1.24010277e+00 5.65845445e-02 5.93508184e-01 3.36304694e-01
4.07172292e-01 8.28809798e-01 6.46845162e-01 7.32733965e-01
6.27534762e-02 5.89907825e-01 6.91633821e-01 3.39571685e-01
2.69371748e-01 -3.08436632e-01 5.88738799e-01 3.94317269e-01
-2.06881151e-01 4.81545776e-02 -9.97521579e-01 4.98992175e-01
-1.94141114e+00 -1.07063925e+00 9.57314894e-02 2.54576445e+00
4.79705274e-01 3.44996899e-01 3.53986353e-01 1.62038013e-01
7.06507385e-01 1.64434724e-02 -6.41887724e-01 -4.08829153e-01
1.47985965e-01 3.63755703e-01 3.25052321e-01 5.02095759e-01
-1.18835473e+00 7.08245635e-01 4.91341639e+00 8.70971680e-01
-1.30830097e+00 3.91628668e-02 3.33872855e-01 1.86984256e-01
1.68924257e-02 3.64097618e-02 -7.29644775e-01 4.43644792e-01
1.20396352e+00 -8.78758058e-02 4.77948904e-01 1.10058928e+00
3.89013708e-01 -3.71388882e-01 -1.37678230e+00 5.49539208e-01
-3.29172850e-01 -7.92182624e-01 -1.63382515e-01 3.23963672e-01
4.49378163e-01 -2.12366611e-01 -7.08747581e-02 4.44418967e-01
3.97512555e-01 -7.54708171e-01 7.48173773e-01 9.68154907e-01
2.91275591e-01 -9.16031003e-01 9.15500939e-01 7.38960028e-01
-1.11614656e+00 -2.60336041e-01 -2.24540174e-01 -8.10757056e-02
2.24979550e-01 6.77417696e-01 -1.13890958e+00 5.23029804e-01
3.54204059e-01 6.69793010e-01 -5.12983143e-01 9.29621696e-01
-3.04077387e-01 5.81283450e-01 -6.08730793e-01 -2.36980468e-01
3.01072359e-01 -5.46471357e-01 1.05140352e+00 1.02090657e+00
4.74792331e-01 -7.14168906e-01 1.60799786e-01 1.01709735e+00
4.78557080e-01 1.57087639e-01 -7.53914535e-01 -1.80552695e-02
5.13114519e-02 1.24170208e+00 -1.07553720e+00 -3.00054610e-01
2.24721372e-01 9.29560542e-01 3.89314801e-01 2.30104476e-01
-6.55227184e-01 -5.29202260e-02 5.43814719e-01 3.91237736e-02
6.02801323e-01 -3.40735853e-01 1.85430169e-01 -1.10385025e+00
-1.49921343e-01 -5.90082347e-01 5.29683493e-02 -2.03136131e-01
-1.06176865e+00 5.28844059e-01 2.88335741e-01 -1.12515640e+00
-8.52229834e-01 -5.90573668e-01 -4.40451831e-01 7.92268515e-01
-1.19148219e+00 -7.63219714e-01 -1.67457297e-01 5.82993567e-01
2.07314223e-01 8.45359489e-02 7.62513518e-01 -3.89911495e-02
-4.07818556e-01 -5.39536029e-02 3.50643396e-01 -2.24029288e-01
4.87897724e-01 -1.75646818e+00 -1.16126262e-01 1.01281607e+00
2.59811908e-01 2.52788842e-01 1.13195229e+00 -6.49545014e-01
-9.91279185e-01 -1.00199175e+00 3.05924684e-01 -2.23795056e-01
9.22796011e-01 -4.48636711e-01 -1.11609852e+00 4.73748446e-01
3.71891744e-02 -1.84334755e-01 1.48832634e-01 -6.63728407e-03
1.10850409e-02 -2.34171674e-01 -1.21315813e+00 5.30508935e-01
6.78557038e-01 -3.92456561e-01 -5.19669592e-01 -1.30191946e-03
3.55246693e-01 -1.77952439e-01 -8.51585090e-01 2.08747447e-01
4.71067369e-01 -9.86214995e-01 5.40887356e-01 -7.80081213e-01
2.00787812e-01 -2.88776606e-01 7.97536671e-02 -1.50218487e+00
1.16156869e-01 -5.70749998e-01 -3.78684491e-01 1.13317704e+00
-8.38793907e-03 -7.00522065e-01 6.62630022e-01 3.12619835e-01
1.61384925e-01 -5.26771605e-01 -1.25639832e+00 -8.57073605e-01
6.22520745e-02 -5.04774332e-01 4.40121979e-01 5.42120457e-01
-3.28241020e-01 -8.09971243e-02 -2.52179563e-01 1.05901413e-01
4.49980110e-01 1.08063377e-01 8.54824007e-01 -1.47227526e+00
-5.89360714e-01 -3.82733107e-01 -6.79723799e-01 -8.36737752e-01
3.98500055e-01 -6.06886089e-01 2.34012201e-01 -1.16176236e+00
-2.32541978e-01 -3.80637109e-01 -3.58728170e-01 1.97820142e-01
6.98575526e-02 -3.98765773e-01 2.48579815e-01 -1.56873614e-02
-5.47136784e-01 6.62955940e-01 8.55541587e-01 8.99612680e-02
-4.15067047e-01 3.23317915e-01 5.90371825e-02 7.41666973e-01
8.02479267e-01 -6.16826475e-01 -2.90116340e-01 2.14563832e-01
3.22060347e-01 -1.27085343e-01 3.41498494e-01 -1.30987144e+00
1.02523535e-01 -1.44270003e-01 1.69963509e-01 -3.53750646e-01
2.17897847e-01 -1.12772667e+00 2.02696905e-01 9.18682456e-01
-4.91008639e-01 -2.08397672e-01 1.17779151e-01 9.01733816e-01
-5.95226549e-02 -8.00890803e-01 7.25269616e-01 -1.14890873e-01
-7.16450810e-01 -8.56405962e-03 -7.92714596e-01 -1.85418621e-01
1.13423693e+00 -3.53472047e-02 1.54334188e-01 -4.28854972e-01
-1.26847637e+00 -6.27712011e-02 4.78110850e-01 -3.27902921e-02
-2.43883003e-02 -8.38328242e-01 -3.32329601e-01 -1.17705286e-01
-7.97652155e-02 -3.03298738e-02 3.43559474e-01 1.15691590e+00
-3.95484269e-01 1.88574046e-01 -5.24265110e-01 -7.77700663e-01
-1.01203394e+00 7.36306489e-01 4.58680034e-01 -9.18412387e-01
-4.21815336e-01 1.43208057e-01 -2.63862640e-01 -3.17856550e-01
1.28158554e-01 -6.66170657e-01 -3.92822683e-01 4.63479489e-01
3.19095910e-01 5.62887609e-01 1.70326605e-01 -4.19108182e-01
-1.47655979e-01 2.26567179e-01 1.00869209e-01 -4.53245103e-01
1.37362874e+00 -1.27778515e-01 5.23922145e-02 8.68472815e-01
9.90437508e-01 -3.70852165e-02 -1.56634581e+00 -1.39028549e-01
3.79862666e-01 -1.61768749e-01 -1.89192057e-01 -5.46208441e-01
-5.72615504e-01 1.09586132e+00 9.47105467e-01 3.47702742e-01
9.94818807e-01 -1.50469467e-02 8.44558403e-02 6.42414868e-01
4.10554528e-01 -1.23519695e+00 6.83884323e-02 5.76520681e-01
7.08733618e-01 -9.64374900e-01 -3.03028256e-01 8.62694159e-02
-5.67707121e-01 1.46268129e+00 1.92224100e-01 -1.96015239e-01
6.61196887e-01 2.94743836e-01 -2.11848542e-01 2.91882396e-01
-4.67241436e-01 -7.52281070e-01 -5.15862331e-02 7.87470758e-01
-1.84565797e-01 -3.63156595e-03 -1.27593115e-01 2.17410833e-01
-9.14125293e-02 1.39117673e-01 4.73645627e-01 7.39776313e-01
-3.81874561e-01 -1.27941084e+00 -3.25898826e-01 2.31992438e-01
-9.32870209e-02 2.40747079e-01 -2.76520580e-01 8.71271670e-01
2.35497460e-01 5.61744452e-01 1.40246674e-01 -2.08016589e-01
2.76170492e-01 4.55257803e-01 5.83478868e-01 -4.39097553e-01
-4.79512751e-01 -1.97896864e-02 -1.87734485e-01 -6.33124352e-01
-5.23052037e-01 -9.99219954e-01 -8.94492745e-01 1.36805609e-01
-1.40260383e-01 1.01519383e-01 9.47445929e-01 1.25835836e+00
2.86351323e-01 5.95336556e-01 5.21815598e-01 -1.11482084e+00
-9.00698364e-01 -1.04705298e+00 -5.33487082e-01 4.22695339e-01
2.39818901e-01 -9.79535460e-01 -3.77084225e-01 2.23323237e-02] | [4.3667311668396, 2.1172163486480713] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.